proceedings of the fifty-fourth annual meeting of

PROCEEDINGS OF THE FIFTY-FOURTH ANNUAL MEETING OFTHE AMERICAN SOCIETY FOR CLINICAL INVESTIGATION, INC.,HELD IN ATLANTIC CITY, N. J., APRIL 30, 1962. . THE TRAININGOF THE CLINICAL INVESTIGATOR: Presidential Address

Henry G. Kunkel

J Clin Invest. 1962;41(6):1334-1414. https://doi.org/10.1172/JCI104595.

ASCI Presidential Address

Find the latest version:

https://jci.me/104595/pdf

http://www.jci.org

http://www.jci.org/41/6?utm_campaign=cover-page&utm_medium=pdf&utm_source=content

https://doi.org/10.1172/JCI104595

http://www.jci.org/tags/45?utm_campaign=cover-page&utm_medium=pdf&utm_source=content

https://jci.me/104595/pdf

https://jci.me/104595/pdf?utm_content=qrcode

Journal of Clinical InvestigationVol. 41, No. 6, 1962

PROCEEDINGSOF THE FIFTY-FOURTH ANNUALMEETINGOF THE AMAIERI-CAN SOCIETY FOR CLINICAL INVESTIGATION, INC., HELD IN

ATLANTIC CITY, N. J., APRIL 30, 1962

PRESIDENTIAL ADDRESS

THE TRAINING OF THE CLINICAL INVESTIGATOR

By HENRYG. KUNKEL

In this address, I would like to touch on a few facetsof the intricate problem of training clinical investigators.During the past year the Society has received severalrequests from representatives of different branches ofthe Public Health Service for aid in the evaluation of thevarious training-grant programs as they apply to clinicalinvestigation. Several discussions have been held withrepresentatives of the Society with the expected re-sult that no one could clearly define an optimal programof training for such a variegated individual as a clinicalinvestigator. The question arose and of course will longcontinue to arise: Should clinical investigators trainwith clinical investigators or should they work with morebasically oriented scientists during their training period?

However, before going farther, I would like to tell alittle of the story of Doctor Brightfool. He was a mosthighly trained doctor, senior resident in medicine at themain teaching hospital attached to Harvumbia MedicalSchool. His special forte was new steroid hormones andnone could rival him in his knowledge of their dosages.Naturally, he became curious about their mechanism ofaction, but unfortunately he was far too busy in the con-scientious pursuit of his hospital duties to even considerattempting any experiments. Suddenly, in a moment ofexhilaration, he made the momentous decision to go intoclinical investigation and tackle the steroid problem sodear to his heart. Of course he was no little swayed byan innate conviction that "research accomplishment isthe royal road to academic success." His joy knew nobounds when he was accepted as a research associate atthe Federal Institute for the Preservation of Health.There was no longer any doubt in his mind; he was de-termined to become an academic medicine man. The dayfinally arrived when he tried his first experiment in thelaboratory. Many were his frustrations, and how hewished he knew more chemistry, or even as much as hehad once known. It soon became apparent that all heneeded was more basic training in enzyme chemistry.He was with too clinical a group. So after two years heleft and went back to the biochemistry department athis own medical school, where he was given the problemof purifying the enzyme hydroxymethyl deoxycytidinemonophosphate kinase. It was a slow process but hestuck it out because he felt he was really learning en-zyme chemistry. Eventually he ended up in a joint ap-pointment with the department of medicine, but by nowhe knew much more about kinetics than corticoids.Things looked rather bleak after some five years of such

combined activity and it gradually dawned on him thathe was not really cut out to be an investigator. For-tunately, rescue did occur when he was suddenly ap-pointed director of a small new Institute for the Studyof Cancer, Arthritis, Arteriosclerosis and Allied Dis-orders (the latter referring to other diseases sponsoredby wealthy foundations). He was now content-he hadfound his calling-he had become an administrator, whichof course is where (according to Parkinson's law) hewould have ended up anyway, even if he had been a goodinvestigator. So what did it matter?

Unfortunately, this little tale is not pure fantasy buta sequence of events observed not infrequently in the areaof clinical investigation. The great tragedy in it is theloss of highly trained physicians from vocations wherethey are urgently needed. It has been estimated that thetraining of a good physician cost something of the orderof $50,000. This large sum, which represents only tangi-ble costs, plus the added expense of special training, em-phasizes numerically the tragedy involved when the endresult is simply a poor investigator. Then there is theimportant question of the shortage of scientific manpower,which is exemplified by weekly advertisements in TheNew York Times for mathematicians and physicists (upto now there has been none for clinical investigators).This is reflected in the alarming fall-off in the number ofapplicants to medical schools, and makes it all the moreimperative that the medical school commodity be put tothe right use.

With so much at stake, it would seem of the utmostimportance that the skilled physician should have some

evidence of an aptitude for investigative work beforetaking the drastic step of enrolling in a prolonged planof training in so different an area. Whenever possible,this should be ascertained at an early period, preferablyduring the resident training period or even in medicalschool. Perhaps greater freedom from an oppressiveduty load during the house officer training period, as isfrequently advocated but rarely enforced, would repre-sent one way to permit the inquiring individual an earlieropportunity to test, at least in a preliminary way, hisinclinations and even fitness for the research alternative.Interest and enthusiasm are scarcely sufficient withoutother attributes for a truly successful career in clinicalinvestigation. Few individuals possess the insight to ad-mit to ineptitude, especially since proficiency in this areais so intangible. It is too easy to justify failure withplausible reasons such as the vicissitudes of research work

1334

AMERICANSOCIETY FOR CLINICAL INVESTIGATION

or inadequate training. Scientific inquiry is a sort ofopiate that once experienced is not readily shaken off,particularly if the individual has tasted what has beentermed "the Elysian delight" of observing something new.

The competitive criteria generally regulating ad-vancement in most scientific fields are not so clearly opera-tive in the area of clinical investigation, particularly inthe present climate of plentiful funds and expanding clini-cal research centers. Sufficient opportunity exists to en-courage even the poor investigator to persist in his courseat relatively little self-sacrifice. In the long run, theburden of responsibility for the important decision ofwhether the young man stays in clinical investigation mustfall primarily on the shoulders of the established investi-gator who has observed him in action. This very diffi-cult yet essential assignment requires considerable thoughtand might well be considered an obligation to the youngassociate, although help and advice on this question arerarely sought.

There is no doubt that a place will always remain forthe untrained individual with ingenuity and keen observa-tional ability. It is almost stating a truism to say thatthese talents can never be replaced by simple experiencewith highly specialized techniques. However, neithershould the latter be avoided. The British geneticist, Dar-lington, has recently presented a strong indictment ofmedical training for failing to provide the foundation forunderstanding the advances in biology, and he cites an ex-ample as follows: "Chromosomes which have been un-derstood in agricultural research for forty years madetheir debut in medical research only four years ago-and, as far as medical teaching is concerned, the chromo-somes still exist only on paper."

Many of our distinguished predecessors in this So-ciety availed themselves of opportunities to work withprominent scientists abroad for variable periods of timeafter their clinical training. This has become less popu-lar in recent years but still remains a valuable opportu-nity, particularly in certain fields. In most instances theyoung physician can readily find the facilities for furtherbasic training within the confines of his own university ormedical school. Our research institutes can also play auseful accessory role. Theobald Smith once stated thateach new generation of "research workers must have morediversified training than the older generation possesses."Can this be acquired by physicians working solely inclinical investigation laboratories? Facts are continuallyaccumulating in many branches of science and in obscureareas that have application to human disease but whichare unfamiliar to most of the present generation of clini-cal research workers. Our training and background donot permit us to see or appreciate their relevance, and itis highly desirable that the young investigator have an op-portunity to explore at least a few of these areas duringhis training period.

In addition, there is an element of selfishness in thisera of large groups and teams where the project repre-sents the first consideration and the training of the youngmember much the second. The attitude that what is good

for the project is good for the younger individual in it isnot uncommon. With the effective sterilization of ourselected best investigators by a continuous series of com-mittee meetings, conferences, and symposia, an ever in-creasing responsibility for actual investigative work fallsto the junior members of the research team. All toofrequently they do not represent co-workers acquiringexperience and training, working shoulder to shoulderwith the established investigator, observing his approach,technique, and fulfillment of a problem. Instead they arehard workers struggling on their own, and this is im-portant, not with their own approaches but with the pro-jected plan of the senior investigator. He still managesto find the time to fill the junior worker with ideas andschemes and then to report the experimental findings atselect conferences where the junior member, unfortunately,happens not to be invited. This of course is an exag-geration, but the key question that must be faced is:How often is the senior investigator prepared to advise hisjunior associates to leave their essential position on histeam and seek more basic training in laboratories whichmight furnish them with new, difficultly acquired tech-niques and fresh perspectives to apply to clinical problems?

There are of course a number of dangers that arise inthe instances where very thorough grounding in the basicsciences is undertaken. Perhaps foremost of these is thevery real problem of the almost invariable sacrifice ofproficiencies in clinical medicine which remain such anessential ingredient of the clinical investigator, at leastin the classical sense. Much has been said on this point.There is also little doubt that clinical investigation needsat least some basically trained individuals. Perhapsideally it should encompass a whole spectrum of investi-gators operating at various distances from the bedside.

Another very real danger is that the physician under-taking prolonged basic training may be lost completelyto clinical investigation. He may well become so ena-mored of fundamental problems of biology as to forsakeentirely his disease orientation. In addition, he almostcertainly will encounter the rather widespread disparagingattitude of basic scientists toward clinical investigationwhich dampens the original appeal of a career in thisarea. Such individuals, particularly those participatingin various postdoctoral Ph.D. programs, are in consid-erable competitive demand from various branches of thepreclinical sciences where they frequently end up in im-portant positions, all connections with clinical investiga-tion severed. Perhaps this is an overstatement; one linkwith clinical investigation often does remain and that ismembership in this Society.

Here then, to digress for a moment, is one of the fac-tors behind our current Society dilemma. These indi-viduals and other full-time research workers, perhapssomewhat less removed from clinical investigation, repre-sent an ever expanding group. Membership in the So-ciety is readily gained under the present system whereresearch accomplishment is the dominant criterion for ad-mission. However, the important point is that it isgained at the expense of the classical type of clinical in-

1335

PROCEEDINGSOF THE FIFTY-FOURTH ANNUALMEETING

vestigator who is involved in patient care. Under our

present fixed quota system, someone must lose out. It isimperative that steps be taken to prevent this selectiveaction on the more clinically oriented individuals who inthe past have represented the backbone of the Society.The most ready remedy, although no solution, is to sup-

port the proposed increase in membership.There is one final important point regarding the edu-

cation of the clinical investigator about which relativelylittle is heard. This concerns training in what mightbe termed the philosophy of research and involves suchdiverse questions as scientific discipline of thought, intel-lectual integrity, the sanctity of the written word, andeven the ethics of research work. Somehow, medical

training does not regularly inculcate upon the investigatorthe basic principles involved in these more esoteric ques-

tions. Their acquisition is often a slow, difficult process

accomplished primarily through bitter experience, andothers must suffer while the young worker learns by hismistakes. It is an area where the physician most cer-

tainly requires development, and this then marks histransition into a scientist. Irrespective of the type oftraining program adopted by a clinical investigator, be itin part-time clinical research or in basic science labora-tories, or perhaps ideally in both, strong considerationshould also be given to the acquisition of this other, mostessential type of education.

1336

journal of Clinical InvestigationVol. 41, No. 6, 1962

PAPERSPRESENTEDAT THE FIFTY-FOURTH ANNUALMEETING 1962

1. Role of the Thymus in Development of Immunity.ROBERT A. GOOD,* CARLOS MARTINEZ, OLGA K.ARCHERand BEN WV. PAPERMASTER, Minneapolis,Minn. (1361)

2. Studies of an Antigen in Cell Walls of Group A Strep-tococci Possessing an Immunologic Relationship toHuman Heart Tissue. MELVIN H. KAPLAN,* Cleve-land, Ohio. (1370)

3. Effects of Humoral Factors on Antigenic HistamineReleased from Human Leukocytes. PAUL P. VANARSDEL, JR., Seattle, Wlash. (introduced by ClementA. Finch). (1407)

4. Reticuloendothelial Clearance of Circulating Fibrinin the Pathogenesis of the Generalized ShwartzmanReaction. LEUNGLEE, NewYork, N. Y. (introducedby Chandler A. Stetson). (1378)

5. The Heterogeneity of Particulate Fat in Plasma.EDWIN L. BIERMAN, ENOCHGORDIS and JAMES T.HAMLIN, III, NewYork, N. Y. (introduced by VincentP. Dole). (1345)

6. Studies on the Relationship of Angiotensin to Hyper-tension of Renal Origin. RUSSELL E. MORRIS, JR.,PATRICIA A. RANSOMand JOHN EAGER HOWARD,BALTIMORE, Md. (introduced by Samuel P. Asper, Jr.).(1386)

7. Angiotensin II and Renal Sodium Transport: Natri-uresis and Diuresis in Patients with Cirrhosis andAscites. JOHN H. LARAGH,* PAUL J. CANNON,RICHARD P. AMES, ALFRED M. SICINSKI, CARL J.BENTZEL and JAY I. MELTZER, New York, N. Y.(1375)

8. The Importance of Conjugation with Glutathione OnSulfobromophthalein (BSP) Transport from Bloodto Bile. BURTONCOMBES, Dallas, Tex. (introducedby Gladys Fashena). (1351)

9. Primary Hyperoxaluria: A Defect in GlyoxylateMetabolism. ELIZABETH XV%". FREDERICK, MITCHELLT. RABKIN and LLOYD H. SMITH, JR.,* Boston, Mass.(1358)

10. A Mechanism of Action of Colchicine in Acute GoutyArthritis. J. E. SEEGMILLER, R. RODNEYHOWELLandSTEPHENE. MALAWISTA, Bethesda, Md. (introducedby Gordon M. Tomkins). (1399)

11. Pseudofeedback Inhibition of Purine Synthesis by6-Mercaptopurine and Other Purine Analogues.ROBERTJ. MCCOLLISTER, WVALTERR. GILBERT, JR.and JAMESB. XVYNGAARDEN,*Durham, N. C. (1383)

12. Characteristics of the Genetic Code. OLIVER WV.JONES, J. HEINRICH MATTHAEI and MARSHALL WX.NIRENBERG, Bethesda, Md. (introduced by Ivan W.Brown, Jr.). (1369)

* Member.) Page number of abstract.

13. Abnormal Myoglobin Chromatography in ChildhoodMuscular Dystrophy. GERALDT. PERKOFF,* ROBERTL. HILL and FRANKH. TYLER,* Salt Lake City, Utahand Durham, N. C. (1391)

14. The Regulation of Spleen Growth and SequesteringFunction. HARRYS. JACOB, RICHARDA. MACDONALDand JAMESH. JANDL,* Boston, Mass. (1367)

15. The Structural Basis for the Action of Neurohypo-physeal Hormones upon the Amphibian UrinaryBladder. HOWARDRASMUSSEN* and IRVING L.SCHWARTZ,* Madison, W\is. and Cincinnati, Ohio.(1393)

16. Measurements of K+ Exchange in Separated RenalTubules. MAURICE BURG, MONES BERMAN andJACK ORLOFF,* Bethesda, Md. (1349)

17. The Effect of Renal Perfusion Pressure on the ActiveTransport of Sodium out of Distal Tubular Urine asStudied with the "Stop-Flow" Technique. LOUISTOBIAN,* KAREN COFFEE, DOROTHYFERREIRA andJUDITH MEULI, Minneapolis, Minn. (1404)

18. Competitive Inhibition of Dibasic Amino Acid Trans-port in Rat Kidney. LEON E. ROSENBERGandSTANTON SEGAL, Bethesda, Md. (introduced byRobert M. Chanock). (1394)

19. Periodic Hemodialysis for the Rehabilitation of Pa-tients with Terminal Uremia. BELDING H. SCRIB-NER,* Seattle, W\ash. (1398)

20. The Use of Synchronized Direct-current Countershockin the Treatment of Cardiac Arrhythmias. BERNARDLOWN, RAGHAVANAMARASINGHAM, JOSE NEUMANand BAROUHBERKOVITZ, Boston, Mass. (introducedby Samuel A. Levine). (1381)

21. Myocardial Metabolism during Epinephrine-inducedNecrosis. TIMOTHY J. REGAN, LAWRENCETROUM,PATRICK H. LEHAN and HARPER K. HELLEMS,*Jersey City, N. J. (1393)

22. Studies on the Pathogenesis and Clinical Features of"Alcoholic Hypoglycemia." N. FREINKEL,* D. L.SINGER, C. K. SILBERT and J. B. ANDERSON,Boston,Mass. (1359)

23. Retardation of Sodium Flux in Dog Erythrocytes byPhysiological Concentrations of Aldosterone In Vitro.DAVID H. P. STREETENand COLETTESPACH, Syracuse,N. Y. (introduced by Eugene L. Lozner). (1403)

24. The Secretion of 18-Dihydroaldosterone by the HumanAdrenal. STANLEY ULICK and KATHRYN KUSCHVETTER, NewYork, N. Y. (introduced by Solomon A.Berson). (1406)

25. The Role of Endogenous Glucagon in Blood GlucoseRegulation in Man. ROGERH. UNGERand ANNAM.EISENTRAUT, Dallas, Tex. (introduced by EliasStrauss). (1406)

26. The Hypoglycemic Effect of Ketone Bodies. DAVIDMEBANEand LEONARDL. MADISON,* Dallas, Tex.(1383)

1337

SECTIONAL MEETINGS

Section I: INFECTIOUS DISEASE AND IMMUNOLOGY

1. Hypogammaglobulinemia produced by the Administra-tion of 6-Thioguanine to Patients with Nephrosis.SHELDON M. WOLFF and HOWARDC. GOODMAN,Bethesda, Md. (introduced by James H. Baxter).ASCI (1413)

2. Measurement of Hemolytic Complement (C') inSynovial Fluids. T. PEKIN and N. J. ZWAIFLER,Washington, D. C. AFCR

3. The Mechanism of Death in Experimental Pneumo-coccal Infection. JOHN E. PERRYand LEIGHTON E.CLUFF*, Baltimore, Md. ASCI (1392)

4. Increased Peripheral Resistance in Patients withSeptic Shock. V. N. UDHOJI, M. H. WEIL, M. P.SAMBHI and R. L. SUDRANN, Los Angeles, Calif.AFCR

5. Impaired Leukocytic Bactericidal Activity Resultingfrom Cortisol Treatment. F. ALLISON and M. H.ADCOCK, Jackson, Miss. AFCR

6. Studies on Lysosomes. The Effect of Cortisone andEndotoxin. GERALDWEISSMANNand LEWIS THOM-AS,* NewYork, N. Y. ASCI (1410)

7. Susceptibility to Infection in Experimental Anemia.D. KAYEand E. W. HOOK, New York, N. Y. AFCR

8. Selective Inhibition of Enterovirus Nucleic AcidSynthesis by 2-(a-Hydroxybenzyl)-Benzimidazole(HBB). HANS J. EGGERSand IGOR TAMM,* NewYork, N. Y. ASCI (1356)

Section II: ENDOCRINOLOGY

1. Metabolic Abnormalities of Adipose Tissue fromPatients with Early Diabetes. J. A. OWEN, J. H.GASKIN, W. D. WARRENand G. HOLLIFIELD, Char-lottesville, Va. AFCR

2. Studies on the Mechanism of Ethanol-induced Hypo-glycemia. JAMES B. FIELD* and HIBBARD E. WIL-LIAMS, Bethesda, Md. ASCI (1357)

3. Studies to Determine whether Glucose Must BeMetabolized to Induce Insulin Release. CHARLESKILO, CHARLESL. LONG, JR., RITA M. BAILEY, MARYB. KOCHand LILLIAN RECANT,* St. Louis, Mo. ASCI(1372)

4. An In Vitro Influence of the Central Nervous SystemTissues on Thyrotropin Secretion. WV. D. ODELL,Bethesda, Md. AFCR

5. Studies on the Role of the Brain in the SteroidalInhibition of ACTH Release. JOHN W. KENDALL,KUNIO MATSUDA, CAROLINE DUYCKand MONTEA.GREER,* Portland, Ore. ASCI (1371)

6. Impaired Aldosterone Production in "Salt-losing"Congenital Adrenal Hyperplasia. G. T. BRYAN,B. KLIMAN and F. C. BARTTER, Bethesda, Md.AFCR

7. Urinary A-5-Pregnene-3,3,17a,20a,21-Tetrol in 33-ol-Dehydrogenase Deficiency. ALFRED M. BONGIO-VANNI* and ALBERT CLARK, Philadelphia, Pa. andMontreal, Canada. ASCI (1346)

8. Biologic Potential of Transcortin. AVERYA. SAND-BERG* and W. R. SLAUNWHITE, JR., Buffalo, N. Y.ASCI (1396)

9. Corticotropin and Gonadotropin Effect on the Bio-synthesis and Catabolism of Steroids Derived fromVirilizing Ovarian Tumors. A. L. ENNIS and S.RICHARDSONHILL, JR., Birmingham, Ala. AFCR

Section III: METABOLISM

1. The Effect of the Physician's Role upon Patient Re-sponse to an Unpleasant Task. M. D. BOGDONOFF,L. BREHMand K. BACH, Durham, N. C. AFCR

2. Abnormal Neural Regulation of Vasopressin andAldosterone Secretion in a Patient with CerebralHyponatremia. JOSEPH F. DINGMAN, AKIRA ARI-MURA, RALPH E. PETERSON*and ROBERTG. HEATH,Boston, Mass., NewYork, N. Y. and NewOrleans, La.ASCI (1354)

3. A Syndrome of Periodic Hypothalamic Discharge.S. M. WOLFF, R. C. ADLER, E. R. BUSKIRKand R. H.THOMPSON,Bethesda, Md. AFCR

4. In Vivo and In Vitro Studies of Effect of L-Leucine onCarbohydrate Metabolism. R. A. YERRY, S. PEN-SUWAN, F. A. ZACHAREWICZand T. F. FRAWLEY,Albany, N.Y. AFCR

5. Observations on Cataract Formation in the Newbornof Pregnant Rats Fed a High Galactose Diet. STAN-TONSEGAL and HOWARDBERNSTEIN, Bethesda, Md.(introduced by Jacob Robbins). ACSI (1399)

6. Site of Action of Androgen on Adrenal Steroidogenesis.JULIAN I. KITAY, Charlottesville, Va. (introduced byAlfred Jay Bollet). ASCI (1373)

7. Incorporation of 17-Hydroxy-Pregnenolone-H3 intoCortisol by the Human Adrenal In Vitro and by theDog Adrenal In Vivo. PATRICK J. MULROW,GEORGEL. COHN, ALBERT A. KULJIAN and WILLIAM F.GANONG, New Haven, Conn. and San Francisco,(introduced by Philip K. Bondy). ASCI (1388)

8. Synthesis of Fatty Acids by Mitochondria. W. R.HARLAN, S. BOONEand S. C. DEES, Durham, N. C.AFCR

Section IV: KIDNEY ANDELECTROLYTES

1. The Relationship between Electrical and HydrogenIon Gradients Across the Rat Proximal Tubule.N. BANK, NewYork, N. Y. AFCR

2. A Kinetic Analysis of the Antidiuretic Action of Vaso-pressin. ISIDORE S. EDELMAN*and MARTINJ. PETER-SEN, San Francisco, Calif. ASCI (1356)

3. Water-Water Interaction during Solvent Flow inIsolated Turtle Bladders. WILLIAM A. BRODSKY*and THEODOREP. SCHILB, Louisville, Ky. ASCI(1348)

4. Osmotic Water Transfer Across a Vasopressin-sensitiveand an Artificial Membrane. LAURENCEE. EARLEY,Boston, Mass. (introduced by Laurence B. Ellis).ASCI (1355)

1338


5. The Effect of Adenosine Derivatives on Ion Transportand Energy Metabolism in the Toad Bladder. ROB-ERT P. DAVIS, MITZY CANESSA-FISCHERand CHESTERM. EDELMANN,JR., New York, N. Y. (introduced byI. Herbert Scheinberg). ASCI (1353)

6. The Action of Streptolysin S on Renal BasementMembrane. E. M. TAN and M. H. KAPLAN, Cleve-land, Ohio. AFCR

7. Osmolality, pH, and Inulin Concentration in ProximalTubular Fluid of the Dog. JAMES R. CLAPP andJOHN F. WVATSON, Bethesda, Md. (introduced byRobert WV. Berliner). ASCI (1350)

8. An Assay for ADH in which Levels Were Demon-strated in Jugular Vein Plasma of Children withNephrogenic Diabetes Inisipidus (NDI). M. A.HOLLIDAY, C. BURSTIN, L. TAVOULARISand L. POPPE,Pittsburgh, Pa. AFCR

Section V: GASTROENTEROLOGY

1. Effect of Ammonium Chloride and Urea Infusions on

Gastric Ammonium Levels. B. FLESHLER and G.GABUZDA, Cleveland, Ohio. AFCR

2. Effects of Norepinephrine on Hepatic Circulation inDogs. F. A. BASHOURand D. P. SELLERS, Dallas,Tex. AFCR

3. Mechanism of Bilirubin Excretion by the AvianKidney. E. E. OWEN, I. WNILLIAMSON and R. R.ROBINSON, Durham, N. C. AFCR

4. Abnormal Rise of Plasma Amino Acids and theAppearance of a Polypeptide in the Plasma in AdultCeliac Disease after Oral Administration of Gliadin.0. DHODANANDKOWLESSAR, ELLIOT WESERandLORRAINE J. HAEFFNER, New York, N. Y. (intro-duced by Marvin H. Sleisenger). ASCI (1374)

5. Adult Celiac Disease: Rapid Sequential Changes inJejunal Mucosa with Alterations of Dietary Gluten.THEODOREM. BAYLESS, JOHN H. YARDLEY, JAMESH. NORTONand THOMASR. HENDRIX, Baltimore, Md.(introduced by Ivan L. Bennett, Jr.). ASCI (1344)

6. The I'3l-triolein Absorption Test. N. TUNA, H. K.MANGOLDand D. MOSSER, Minneapolis, Minn.AFCR

7. A Novel Metabolic Pathway of Fat Absorption in theIntestinal Mucosa. JOHN R. SENIOR and KURT J.ISSELBACHER,* Boston, Mass. ASCI (1399)

8. Direct Evidence for the Participation of EndogeneousLipids in the Formation of Dietary Chylomicrons.SAMUELJ. FRIEDBERG, MORTOND. BOGDONOFF,E.HARVEYESTES, JR., PAUL B. SHERand GITTA JACK-SON, Durham, N. C. (introduced by David T. Smith).ASCI (1359)

Section VI: HEMATOLOGY

1. Relationship of High Levels of "Bart's" Hemoglobinin Infancy to a-Thalassemia. DAVID J. XVEATHERALL,

BALTIMORE, Md. (introduced by C. Lockard Conley).ASCI (1409)

2. Evidence for a Second Iron-binding System in Plasma.RAYMONDJ. DERN, ALBERTO MONTI and MICHAELF. GLYNN, Chicago, Ill. (introduced by Richard L.Landau). ASCI (1354)

3. Initiation of the Action of Complement in a HumanAutoinmmune Hemolytic System, the Donath-Land-steiner Reaction. CARL F. HINZ, JR.* and ANNAMARIE MOLLNER, Cleveland, Ohio. ASCI (1365)

4. Correlation of Complement-fixing Antibodies withPlatelet Survival in Thrombocytopenic Patients.R. H. ASTER, R. H. LEVIN, H. E. COOPER, E. J.FREIREICH and E. E. MORSE, Bethesda, Md. AFCR

5. Biochemical Interaction between Two Populations ofRed Cells. E. BEUTLERand M. C. BALUDA, Duarte,Calif. AFCR

6. Genetic Implications of Diisopropyl Fluorophosphate(DFP)32-Erythrocyte Survival Studies in NegroFemales Heterozygous for Glucose-6-Phosphate De-hydrogenase Deficiency. GEORGEJ. BREWER, AL-VIN R. TARLOVand ROBIN D. POWELL, Chicago, Ill.

(introduced by Alf S. Alving). ASCI (1348)7. A New Method for Measuring Minimum In Vivo

Concentrations of Factor VIII Applied in Distributionand Survival Studies. N. RAPHAEL SHULMAN,*VICTORJ. MARDERand MERILYNC. HILLER, Bethesda,Md. ASCI (1401)

8. Different Patterns of Hematologic Remissions in AcuteMyelocytic Leukemia. R. H. LEVIN, G. M. BRITTINand E. J. FREIREICH, Bethesda, Md. AFCR

Section VII: CARDIOPULMONARY

1. An Early Pulmonary Physiologic Abnormality inProgressive Systemic Sclerosis (Diffuse Scleroderma).R. J. WILSON, G. P. RODNANand E. D. ROBIN,Pittsburgh, Pa. AFCR

2. Functional Characteristics of the Left Atrium in Man.R. R. JOHNSTON, C. RACKLEY, H. J. SAUTER andH. T. DODGE,Seattle. Wash. AFCR

3. Nucleic Acid Studies in Experimental Cardiomegaly.R. G. SUMNERand H. D. MCINTOSH, Durham, N. E.AFCR

4. Electrocardiographic and Hemodynamic Observations

during Selective Coronary Cineangiocardiography.RICHARD S. Ross, NEIL SCHWARTZ, ROBERT A.

GAERTNERand GOTTLIEB C. FRIESINGER, Baltimore,

Md. (introduced by A. McGehee Harvey). ASCI

(1395)5. Congenital Deafness, ECGAbnormalities and Sudden

Death-A Recessive Syndrome. G. R. FRASER and

P. FROGGATT,Seattle, Wash. AFCR

6. Metabolism of Arterioles and Venules as Studied in the

Cartesian Diver. RUFUS 0. HOWARD, DAVID XV.

RICHARDSON, MORRIS H1. SMITH, JR. and JOHN L.

PATTERSON, JR.,* Richmond, Va. ASCI (1366)7. Postural Hypotension in XNernicke's Disease: A Mani-

festation of Autonomic Nervous System Involvement.

R. I. BIRCHFIELD, Seattle, W\ash. AFCR

8. Diminished Biological Half-life of Angiotensin II in

Hypertensive Plasma. ROGERB. HICKLER, DAVID

P. LAULER and GEORGEW. THORN,* Boston, Mass.

ASCI (1365)9. Antihypertensive Principle from Renal Medulla.

E. E. MUIRHEAD, J. WV. HINMAN, E. G. DANIELS, M.KoSINSKI and B. BROOKS, Detroit and Kalamazoo,Mich. ASCI (1387)

1339

ABSTRACTS

Effect of Reserpine on Vascular Responses of Forearmto Tyramine and Norepinephrine in Man. FRANCOISM. ABBOUD AND JOHN W. ECKSTEIN, Iowa City,Iowa (introduced by Willis M. Fowler).

The vasoconstrictor response to tyramine is reduced,and that to norepinephrine is enhanced, in animalstreated with reserpine. These effects have not beendemonstrated in man. Observations on blood pressure(cuff) and forearm blood flow (plethysmograph) weremade in 7 healthy men before and after intravenousinjection of tyramine (0.1 mg/kg) and before andduring 6-minute infusions of norepinephrine in dosesof 0.032, 0.064, and 0.15 jug per kg per minute. Eachsubject was studied 3 times. Session A served as thecontrol. Session B followed 2 weeks of treatment withreserpine, 1 mg daily. Session C was held 3 to 6 weeksafter reserpine was discontinued. Analysis of varianceof the data revealed no significant differences in thelevels of blood pressure, blood flow, or resistance toflow during norepinephrine infusions in the 3 sessions.The values for pressure, flow, and resistance obtainedwith all 3 doses of norepinephrine averaged 109 mmHg, 3.39 ml per minute per 100 ml of forearm, and33.5 U, respectively, during session A. The corre-sponding values during session B were 107 mm Hg,3.26 ml, and 38.7 U. During session C these valueswere 111 mmHg, 2.69 ml, and 44.2 U. Resting levelsof blood pressure and heart rate were slightly reducedin session B while resistance was unchanged. Tlhepressor response to tyramine during session B (4.3 +

0.9 mmHg) was significantly lower than during sessionA (13.7 + 4.1 mmHg) or session C (21.0 + 3.77 mmHg). The responses to norepinephrine were similar in5 additional subjects who received reserpine, 3 mgdaily for 5 days. Conclusion: In these experimentsreserpine failed to cause vascular supersensitivity tonorepinephrine even though the pressor response totyramine was reduced.

Effect of Hypertension on Cholesterol Synthesis in Rats.HAROLDADEL, MARIE M. DALY, QUENTIN B. DEMING,*Lim BRUN AND VICTORIA RAEFF, New York, N. Y.

The development of atherosclerosis in animals and manis accelerated by hypertension. Increased concentra-tions of cholesterol in serum and tissues have beenobserved in hypertensive rats. In order to determinewhether these observations might be related to an effectof hypertension on cholesterol synthesis, the incorpora-tion of labeled precursors into cholesterol by tissues ofnormotensive and Goldblatt-hypertensive rats was meas-ured in vivo and in vitro. After intraperitoneal iniec-

* Member.

tion of acetate-i-C'4 into hypertensive rats and pairednormotensive controls, the cholesterol isolated from liver,serum, and residual carcass showed higher specificactivity in the hypertensives. The mean increase was,in liver 58%, in serum 57%, in carcass 35%; p, respec-tively, less than 0.01, 0.05, 0.05. The incorporation ofacetate-i-C1' into cholesterol by liver slices was greaterin slices from hypertensive animals than in those frompaired normotensive controls. The mean difference inspecific activity was 135% (p less than 0.005). No dif-ference was shown when mevalonate-2-C1' was usedas precursor. The incorporation of acetate or mevalo-nate into cholesterol by 6 pools of aortas isolated fromhypertensive rats and pools of paired normotensivecontrols was measured. With either precursor theincorporation of label into cholesterol (purified by re-peated bromination) was higher in the aortas of thehypertensives (p less than 0.025). Synthesis of choles-terol, whether measured in the living rat or in isolatedliver or aorta, is accelerated by hypertension.

The Mechanism of Puromycin Inhibition of HemoglobinSynthesis by Rabbit Reticulocyte Ribosomes. DAVIDW. ALLEN AND PAUL C. ZAMECNIK,* Boston, Mass.

Reticulocyte ribosomes contain the partially formedpolypeptide chains of hemoglobin, of varying chainlengths but each having the same N-terminal group,valine. These incomplete polypeptide chains are con-sidered bound to the ribosome by soluble RNA residuesattached to the peptide chain at its growing C-terminalend. Puromycin, a potent inhibitor of protein synthesis,is an analog of aminoacyl-soluble RNA. The effect ofthis drug on reticulocyte ribosomes, labeled with C14-amino acids by prior incubation of the intact reticulo-cytes, has been studied in a cell-free system capableof hemoglobin synthesis. Puromycin is found to releasethe polypeptide intermediates of hemoglobin synthesisfrom the ribosome. Unlike the more naturally occur-ring release of the completed hemoglobin molecule, thepuromycin-produced release of polypeptide from theribosome does* not require an ATP-generating systemor the supernatant fraction and occurs maximally in1 or 2 minutes. The polypeptides released by puromy-cin do not appear as hemoglobin chromatographicallyor electrophoretically but, like the completed molecule,have only N-terminal valine. C'4-puromycin synthe-sized by methylating its tyrosyl analog with C!-diazo-methane is bound to the soluble polypeptide it releases.Saturation of the binding site of the puromycin-releasedpolypeptide occurs at a puromycin concentration (200m/Amoles/ml) equivalent to that producing maximalinhibition of C1'-amino acid incorporation in a cell-freesystem and maximal release of C'4-polypeptide from theribosomes. One residue of puromycin is bound to the

1340


released polypeptide for every N-terminal valine or oneresidue per polypeptide chain. It is concluded thatpuromycin inhibits protein synthesis by displacing thesoluble RNA residues binding the chain to the ribo-some and releasing the incomplete peptide chain. Itappears possible to employ puromycin, during the course

of synthesis of a given protein molecule, to obtain nas-

cent peptides of varying chain length.

Leukocyte Alkaline Phosphatase in Mongolism; a Pos-sible Chromosome Marker. AARONA. ALTER, STAN-LEY L. LEE, MOHAMEDPOURFARAND GERALDDOBKIN,Brooklyn, N. Y. (introduced by Ludwig W. Eichna).

Polymorphonuclear leukocytes contain one or more

phosphomonoesterases active at pH 9.3. In chronicgranulocytic leukemia the activity of this enzyme(s) isreduced. In leukocytes of chronic granulocytic leuke-mia, absence of part of the long arm of one of theshort acrocentric chromosomes has been observed.These two observations may be related: the missingchromosomal portion may contain the gene for thelacking enzyme. It is generally recognized that mongo-

lism is associated with trisomy of a short acrocentricchromosome. If the same chromosome is involved inmongolism as in chronic granulocytic leukemia, an

effect of the increased dosage of the gene might beevident in increased leukocyte phosphatase activity.Leukocyte phosphatase activity at pH 9.3 was deter-mined in a group of 36 healthy children with mongo-

lism and in 35 healthy control children comparable inage. A mean value of 64 U (mg phosphorus/10'leukocytes in 1 hour) was found in the mongols; thecontrol value was 24 U (p < 0.001). When these valuesare expressed per 1010 polymorphonuclear neutrophilsinstead of total leukocytes, the mongols showed 139 Uand the controls 83 (mean values). This differenceis statistically significant at the 1% level of confidence.These findings are consistent with the hypothesis thatleukocyte alkaline phosphatase production is controlledby a gene on the chromosome for which mongols are

trisomic. The ratio of 139/83 of the phosphatase valuesapproximates the 3/2 that might theoretically beexpected. Other explanations for the observed phos-phatase elevation in the leukocytes of children withmongolism are, of course, possible.

Functional Differences in Lactic Dehydrogenase Activityas a Metabolic Basis for Cell Survival during Hy-Poxia. C. R. AMARASINGHAM, G. M. GREEN ANDK. L. STUART, Boston, Mass. (introduced by EdwardH. Kass).

Although isoenzymes of different tissue origin are

well characterized, the functional meaning of differencesin enzymic structure are not well understood. Lacticdehydrogenase (LDH) with different physiochemicalproperties may be extracted from mammalian tissues.At least two different general molecular species of LDH

clearly exist; there may be three additional intermediateforms, although in many tissues intermediate patternsof LDH activity are the result of mixtures of the twobasic types of LDH. A new difference in the proper-ties of the two enzymes will be presented. LDH ex-tracted from heart, brain, kidney, and lung of rabbitsis inhibited noncompetitively by end product (lacticacid), whereas LDH of skeletal muscle and liver isinhibited only by much greater amounts of lactic acid,and the inhibition is competitive. The activity andproperties of LDH in these two groups of tissues withrespect to substrate (pyruvate) saturation, substrateinhibition, effect of hydrogen ion concentration, andend-product (lactic acid) inhibition can be correlatedwith the known susceptibility of the respective tissuesto hypoxia, thus providing a possible basis for relativesusceptibilities of tissues to infarction. Tissues withthe readily inhibited LDH accumulate sufficient lacticacid to inhibit the use of anaerobic mechanisms for cellsurvival during hypoxia. Significant differences in bothLDH activity and the type of enzyme present were foundin studies of comparable infant and adult organs. Thechanges in enzymic activity and pattern in the matura-tion of certain organs may in part be related to thereplacement of cell types. In other organs, however,such changes occur despite the absence of significantalteration of cell types, suggesting the possibility of en-vironmental control of LDH activity.

Chronic Unconjugated Hyperbilirubinemia (CUH) withIncreased Production of Bile Pigment not Derivedfrom the Hemoglobin of Mature, Circulating Erythro-cytes. IRWIN M. ARIAs, New York, N. Y. (intro-duced by Helen Ranney).The role of extramedullary hemolysis in the produc-

tion of CUH is well known. This study indicates thatCUH can result from intramedullary hemolysis with-out overt signs of extramedullary erythrocyte destruction.A 19 year old white male of Italian descent has hadasymptomatic jaundice without hepatosplenomegaly for6 years. The serum bilirubin (1.3 to 4.5 mg per 100 ml)was unconjugated. Complete blood counts, analysis ofhemoglobin including A2 and F, erythrocyte glucose-6-phosphate dehydrogenase and pyruvic kinase activities,liver function tests, and liver histology were normal.The mean reticulocyte count was 1.3%. Bone marrowexaminations revealed erythrocyte hyperplasia (M: E,1:3). Erythrocyte survival (Cr51) has been normal onfour occasions, but fecal urobilinogen excretion hasranged from 580 to 1,110 mg per day (normal 80 to250 mg per day). In vivo study of glucuronide forma-tion (menthol tolerance test) and in vitro assay ofhepatic uridine diphosphoglucose dehydrogenase andglucuronyl transferase activities (using bilirubin, o-ami-nophenol, and 4-methyl umbelliferone as glucuronidereceptors) were normal. The transfer of bilirubin fromplasma to the liver cells was studied in vitro. Two gof the patient's liver was obtained at surgery. Slicesand homogenates were incubated with bilirubin in either

1341


the patient's or in normal serum. No significant differ-ence was observed in direct-reacting bilirubin formationwhen compared with similar studies on liver fromcontrol subj ects. The half-time of disappearance ofinj ected Fe"9 from plasma was 20 minutes (normal 90min). Ferrokinetic study revealed a decreased rate ofremoval of Fe59 from the bone marrow. The patientingested glycine-N"5. The isotope concentration wasserially measured in hemin and stercobilin. The aver-age erythrocyte life span was 110 days. At least 49%of bile pigment was derived from sources other thanmature, circulating erythrocytes. Chronic jaundiceprobably results from increased intramedullary hemol-ysis. An inherited abnormality may be present, sincea younger brother has chronic unconjugated hyperbili-rubinemia (1.2 to 3.2 mg/100 ml), normal completeblood counts, a mean reticulocyte count of 1.4%o, anderythroid hyperplasia in the bone marrow, and excretes880 mg of fecal urobilinogen daily.

Metabolism of Levo- and Dextroisomers of Thyroid Hor-mones in Man. WALTERL. ARONS, Palo Alto, Calif.(introduced by J. A. Luetscher).

The metabolic fate of I13'-labeled L-thyroxine (LT4),L-triiodothyronine (LT3), D-thyroxine (DT4), and D-tri-iodothyronine (DT3) has been compared by administer-ing each orally in representative pharmacologic dosageto 5 or more euthyroid subjects. Serum, urine, andstools have been serially analyzed up to 192 hours fortotal radioactivity and by paper chromatography fordistribution of I13'-labeled compounds. The followingfindings are of note. 1) Both D isomers are deiodinatedas readily as are the naturally occurring, more activeL forms. 2) After LT4 administration, conversion toT3 has been evidenced by the presence of T3 in theurine in several instances. Also after DT4, T3 has beenrepeatedly found in both serum and urine. 3) AfterLT4, DT, LT3, and DT, administration, evidence formonoiodotyrosine formation was found in the urine.Studies are currently in progress to define this activitymore specifically. Preliminary experiments indicate that theglucuronide conjugates of the administered isomers alsoare excreted in the urine. 4) About twice as much (78%)of ingested DT4 activity is excreted in urine and stoolin 96 hours as of LT, activity (45%). Similar totalamounts of administered LT3 and DT3 activity are ex-creted in 96 hours (80%o ). However, significantlymore DT4 and DTL than the L forms are excreted in thestools. These data, in combination with the serumactivity measurements, indicate a probably poorer gastro-intestinal absorption of the D isomers. 5) External bodycounting has indicated that all four isomers, but LT4and DT4 in particular, are concentrated in the liver.These data reveal both significant differences and simi-larities in the metabolic fate of the L and D isomers ofthyroid hormones and provide some basis for interpretingthe differing pharmacologic activity of these substancesin man.

The Mechanism of Steroid Granulocytosis. JOHN W.ATHENS,* OTTO P. HAAB, SPENCER 0. RAAB, DANER. BOGGS, GEORGEE. CARTWRIGHT* AND MAXWELLM.WINTROBE,* Salt Lake City, Utah.

The purpose of this study was to determine the mech-anism by which adrenocorticosteroids produce granulo-cytosis. To determine whether the granulocytosis isdue to a shift of cells from the marginal granulocytepool (MGP) to the circulating granulocyte pool (CGP),the size of these compartments and of the total bloodgranulocyte pool (TBGP) was measured in ten sub-jects. The TBGP increased from 48 to 89 X 107 cellsper kg after 10 days of oral prednisone (40 mg/day).The TBGP increased from 79 to 138 X 107 cells per kg5 hours after the intravenous injection of 200 mg ofhydrocortisol. In both instances it was found that thegranulocytosis was due to an increase in size of boththe MGPand CGP. To assess whether the granulo-cytosis is due to an increased marrow output of cells,the effect of the rapid injection of hydrocortisol on thespecific activity of granulocytes in the circulation wasevaluated in five subj ects. If the granulocytosis weredue to an increased influx of unlabeled cells from themarrow, a sudden decrease in specific activity of theblood granulocytes should have occurred. No suchdecrease was observed. To assess whether steroids im-pair the egress of cells from the circulation, the cellularityof inflammations produced in the skin was measuredbefore and after steroids. The cellularity of theseexudates was reduced 75% by steroids. From thesestudies it is concluded that steroids produce granulo-cytosis by decreasing the egress of cells from thecirculation and not by causing an influx of cells fromthe marrow or by producing an intravascular shift ofcells. These studies may have significance in regardto the propensity of adrenocorticosteroids to increasesusceptibility to infection.

Adrenomedullary Effects of Guanethidine. WILLIAM J.ATHOS, EDWIN FINEBERG, BURTON MCHUGHANDJAMES G. HILTON,* New York, N. Y.

Guanethidine effect on adrenal catecholamine secre-tion was studied: a) by perfusion into the arterial cir-cuit of isolated denervated adrenal gland preparationsof Hilton et al. (Amer. J. Physiol. 1958, 192, 525),and b) by single intravenous injection into intact dogswith cannulated adrenal veins. Control mean secretoryrates of norepinephrine and epinephrine from isolateddenervated preparations were 0.041 and 0.070 ,ug perminute, respectively. During guanethidine perfusion at10 ,ug per minute, there was no change in catechola-mine secretion. Glucagon perfusion at the end of eachexperiment showed a rise in norepinephrine and epi-nephrine secretion to mean rates of 0.37 to 0.61 lAgper minute, respectively, indicating viability of the glands.Histamine perfusion (0.36 mg/min) studied in a similarway was also without effect on medullary secretion.

1342


In the intact dogs with innervated adrenals, single intra-venous guanethidine injections in dosages from 1.0 to10.0 mg per kg induced an initial mean peripheral bloodpressure drop, followed in 2 to 3 minutes by a promi-nent pressure elevation lasting 20 to 60 minutes. Thiswas associated with a significant fall in adrenal epi-nephrine secretion to 50% of control. Norepinephrinesecretion remained unchanged from control. Intravenoushistamine (0.1 to 0.25 mg/kg) in similar intact prepara-

tions resulted in abrupt drops in mean peripheral bloodpressure associated with a 20- to 40-fold rise in epineph-rine and a 3-fold rise in norepinephrine secretion.

Induction of Acute Iodine Depletion and the Responsethereto in Normal Subjects and in Patients withGraves' Disease. RusSEL M. BARAKATAND SIDNEY H.INGBAR,* Boston, Mass.

To ascertain whether self-regulatory mechanisms existwithin the human thyroid, a means of inducing acute

iodide deficiency was sought. Nine patients, includingtwo with thyrotoxicosis, were fed low iodine, anion-exchange resin-treated milk (chloride cycle) for 2 days.On the second day, brisk osmotic diuresis was inducedby infusion of mannitol. Cations were replaced as

iodide-free chloride salts. Iodide kinetics were assessedwith I.3. during a control period, during the osmoticdiuresis, and 14 to 24 hours thereafter. Renal iodideclearances increased during mannitol infusion. Acuteiodide depletion was induced since, after diuresis, plasmaconcentration and total body content of inorganic I',calculated from the specific activity of urinary iodide,averaged 50% less than control values. Concomitantly,thyroid iodide clearance rates and 3-hour 131 accumula-tion increased by an average of 100%. Although 2 mg

of iodide, when acutely administered, does not normallyreduce thyroid I131 accumulation, inclusion of 800 ,lgof iodide in mannitol infusions prevented the subsequentdecline in plasma iodide concentration and the increasein thyroid clearance rates. This suggested that iodidedepletion per se was the cause of increased thyroidfunction. This response could not be ascribed to in-creased secretion of endogenous thyrotropin. It was

demonstrable within 4 hours after initiation of diuresis,sooner than the stimulation of iodine accumulation

induced by exogenous thyroid-stimulating hormones

(TSH). Furthermore, plasma protein-bound iodines

neither decreased during diuresis nor increased subse-

quently. Finally, separate experiments revealed no acti-

vation of the TSH-dependent, iodide-trapping mechanism

when hormonal synthesis was completely blocked by

methimazole for several days. These findings suggest

that there exists within the thyroid an intrinsic, extra-

hypophyseal mechanism for increasing function, which is

responsive to depletion of plasma, and hence intra-

thyroidal, iodide and which may be related to the

pathogenesis of diseases characterized by functional au-

tonomy of the thyroid gland.

Characteristics and Perf ormance of the Left Ventriclein Massive Mitral Insufficiency. F. A. BASHOURAND

J. D. KHALAF, Dallas, Tex. (introduced by CarletonB. Chapman).

Left ventricular volumes (end-diastolic and end-sys-tolic) were measured in 8 patients with massive mitralinsufficiency, using the biplane cineangiofluorographicmethod of Chapman and Baker and a modificationof Arvidsson's technique. In 3 patients a minimaldegree of aortic insufficiency was demonstrated by retro-

grade aortography. The end-diastolic volume (EDV)ranged from 116 to 271 ml as compared with the normalrange of 85 to 128 ml. The end-systolic volume (ESV)ranged from 32 to 76 ml as compared with the normalrange of 35 to 43 ml. The absolute stroke volume andstroke volume index averaged 144 ml and 87.8 ml perm2 respectively. A linear relationship was observedbetween the stroke volume and EDV. The efficiencyof the left ventricular ejection (ratio of ESV to EDV)ranged from 20 to 36%o and averaged 28%, as comparedwith the normal ratio of 39%o. The mean rate of ven-

tricular ejection was calculated in 7 patients. It rangedfrom 182 to 860 ml per second and averaged 475, as

compared with the values obtained in 2 normal subj ects

(264 and 272 ml/sec). The tension-time index rangedfrom 2,325 to 3,415 mmHg-sec per minute and aver-

aged 2,930 as compared with the values obtained in2 normals (1,940 and 2,390 mm Hg-sec/min). Themean thickness of the left ventricular wall during diastolewas slightly higher than in the normals, measuring 10and 7 mm, respectively. These results demonstrate:1) the existence of increased left ventricular volumesand muscle mass in mitral insufficiency; 2) the ade-quacy of the left ventricular ejection; and 3) a risein the tension-time index, a determinant of myocardialoxygen consumption.

Analog Computer Extrapolation of Indicator DilutionCurves. E. J. BATTERSBY AND ELLIOT V. NEWMAN,*Nashville, Tenn.

In the presence of left-to-right shunts, the measurementof pulmonary blood flow by indicator injection into theright heart with peripheral arterial sampling is usuallyprecluded by the early recirculation of indicator. Thismasks all but the initial portion of the primary circula-tion curve. Theoretically, the ideal method for recover-

ing the masked portion of the curve would involve a

function, mathematical or analog, that would accuratelydescribe the whole curve. When matched to the avail-able portion of the primary curve it would predict theremainder. Such a function can be produced by an

electronic analog computer: y (x) when dy/dx = [G (x)-y]/K; G(x) is a normal distribution function and K

is the time constant of a unit lag component. Visualmatch by repeated solution to whole primary curves

in man and dog has proved satisfactory. The abilityof this analog to predict whole primary curves from

1343


anterior segments was tested in dogs with brachio-cephalic artery-to-right atrium shunts. Indicator dyeinjections were made into the right atrium and sam-pled in the left subclavian artery for computer curves.Control curves between the right atrium and pulmonaryartery or left atrium and subclavian artery had theirareas calculated in the usual manner. The anteriorportions of computer curves were matched and extra-polated. Thirty-four computer-extrapolated curve areasexpressed as percentage of control curve area gavevalues of 103.1 + 11.7 (I SD). Forward triangle analy-sis of the same curves yielded values of 94.7 ± 17.0(K = 0.329). These data indicate that the computerextrapolation method is superior to methods employingempiric constants whose value changes with flow, vol-ume, injection site, and species.

Adult Celiac Disease: Rapid Sequential Changes in Je-junal Mucosa with Alterations of Dietary Gluten.THEODOREM. BAYLESS, JOHN H. YARDLEY, JAMES H.NORTONAND THOMASR. HENDRIx, Baltimore. Md.(introduced by Ivan L. ]lennett, Jr.).

Objective clinical improvement in adult celiac disease(nontropical sprue) with a gluten-free diet is welldocumented. On the other hand, reversibility of thejejunal lesion has been questioned. Peroral intestinalbiopsy during periods with and without gluten couldgive important information about the nature of thisdisorder. Eight patients with severe jejunal alterationswere studied. After gluten exclusion, improvement inmucosal histology was seen in all. The disordered sur-face epithelium became tall and well arranged. Recog-nizable villus structures appeared in the five patientson treatment for 3 or more months. Biopsies in onepatient after 2 years and in another after 3.5 yearswere virtually normal. In order to determine therapidity of this epithelial improvement, biopsies weredone in two patients 3, 7, and 10 days after glutenexclusion. The pretreatment mean cell height of thesurface epithelium was 21 g. After 7 days, meanheight had increased to normal (32 ,). One of thepatients with a virtually normal jejunal biopsy wasgiven wheat gluten for 6 days. Daily fecal fat excretionof 2.4 g became abnormal on the first day (8.7 g)and then rose stepwise to 17 g on the seventh. Biopsieswere done after 1, 3, and 6 days. Focal epithelial de-generation at villus tips with edema and congestion ofthe lamina propria was seen after 1 day and becameprogressively more widespread and severe until glutenwas withdrawn. By 6 days the mean height of thesurface epithelium had decreased to the untreated celiacrange and the number of mitotic figures in the cryptshad trebled. Gluten was in some way toxic to theexposed epithelial cells. These observations stronglysuggest that the epithelial lesion seen in celiac diseaseis indeed gluten-induced and is rapidly reversible ongluten exclusion.

The "Induction' of Dihydrofolic Reductase Activity inErythrocytes and Leukocytes of Subjects Treatedwith Amethopterin. J. R. BERTINO, M. E. FINK ANDP. CALABRESI, New Haven, Conn. (introduced by D.Seligson).

Tetrahydrofolic acid, formed from folic or dihydrofolicacid in a reaction catalyzed by the enzyme, dihydrofolicreductase, plays an important role in the biosynthesisof purines and thymidylic acid. This enzyme combinesavidly with amethopterin and is inhibited by low con-centrations (10' M) of the antimetabolite. After treat-ment with amethopterin, increased enzyme activity hasbeen detected in leukemic leukocytes and also in normalcirculating leukocytes and erythrocytes. The mecha-nism of this adaptive enzyme formation has been investi-gated in the peripheral blood cells of 14 subjects aftera single 20-mg dose of amethopterin, administered intra-venously. Dihydrofolic reductase activity appeared inthe erythrocytes in 3 to 4 days, reached peak levels at8 to 9 days, and was measurable for as long as 4 months.In normal leukocytes, enzyme activity appeared in 3to 4 days, reached peak levels at 9 to 11 days, and couldnot be detected after 19 to 25 days. When lysates ofleukocytes were assayed for ability to inhibit dihyro-folic reductase, the concentration of the amethopterin-like inhibitor paralleled the increased levels of enzymeactivity observed in the leukocytes. Agents that affectgranulocyte production, such as vinblastine and busulfan,suppressed the magnitude of induced enzyme activity.Folinic acid, a competitive inhibitor of dihydrofolicreductase in vitro, also depressed this response in vivo.These findings suggest that: 1) dihydrofolic reductaseactivity may be under feedback control of reduced folicacid compounds; 2) its appearance in circulating leuko-cytes is inhibited by myelosuppressive agents; and 3)once induced, the enzyme activity appears to parallelthe estimated life span of the cell and may provideanother method for measuring leukocyte and erythrocytesurvival.

The Effect of Ethanol-induced Fatty Liver on the Trans-port of Triglyceride Fatty Acids from Liver to Plasma.A. BEZMAN, P. J. NESTEL, J. M. FELTS, R. J. HAVEL*AND S. FRENCH, San Francisco, Calif.

Circulating free fatty acids esterified in the liver area precursor of triglyceride fatty acids (TGFA) con-tained in plasma very low density (< 1.006) lipoproteins(VLDLP). Isotopic studies in rabbits have shownthat TGFA contained in the subcellular fractions ofnormal liver and TGFA of blood plasma equilibraterapidly. Two human subjects were studied when theyhad acute fatty liver due to alcohol and later whenthe hepatic triglyceride content had decreased. Palmi-tate-l-C1' was injected intravenously. Serial bloodsamples were taken and a sample of liver obtained 2to 3 hours later. Specific activities of those TGFA con-tained in VLDLP and of hepatic TGFA were measured.In all experiments: 1) the peak specific activities of

1344


plasma TGFA occurred 2 to 3 hours after injection;2) the ratio of the peak specific activity in plasma TGFAto that of the liver was inversely related to the concen-tration of hepatic TGFA. Similar studies were per-formed in two rabbits fed ethanol for prolonged periodsof time. In these experiments small pieces of liver weretaken at intervals and separated into floating fat, mito-chondria, microsomes, and supernate. Plasma TFGAequilibrated rapidly with TGFA of microsomes andmitochondria; however, the floating fat of these livers,which contained most of the hepatic TGFA, had muchlower specific activity than other subcellular liver frac-tions and plasma TGFA. Thus, in ethanol-induced fattyliver, TGFA of the floating fat only very slowly equili-brated with the other subcellular fractions of liver.Nevertheless, the time of peak specific activities of theTGFA of VLDLP was found to be within normal limits.

The Heterogeneity of Particulate Fat in Plasma. EDWINL. BIERMAN, ENOCHGoaiS AND JAMES T. HAMLIN,III, New York, N. Y. (introduced by Vincent P.Dole).

Two distinct groups of fat particles in turbid plasmacan be identified by zone electrophoresis (starch gran-ules) and by differential flocculation. One fraction mi-grates with a2-globulin, forms a dense flocculate at thetop of a polyvinylpyrollidone (PVP) gradient (con-centration range 0 to 5%), and appears to be identicalwith lymph chylomicrons. These particles are rela-tively large (diameter> 200 mA). The other fraction(of smaller average size) migrates with p-globulin,flocculates at the bottom of the PVP gradient, and isbelieved to come from sources other than thoracic ductlymph. Both kinds of particle are low enough in densityto float in saline. Chromatographic analyses of thetwo groups of particles in plasma taken during alimentarylipemia have shown that the fatty acid composition oftriglycerides in "a,-particles" resembles that of thefeeding, although not identically. This fraction appearsearly and reaches a maximum 4 to 6 hours after thefeeding. The ",-particles" increase in number duringthis period, become the predominant form by 9 hours, andare the last to be detected at the end of absorption.Their fatty acid composition, initially closer to thatof body fat, shifts toward the composition of the fed fatas absorption proceeds. Changes in the particulatelipid zones on starch block electrophoresis are paral-leled by corresponding changes in the turbid zonesdistinguishable on the PVP gradient. All the fat parti-cles in human thoracic duct lymph, labeled by a previousfeeding of C14-plamitic acid, migrate in the a2-globulinregion. When this material is injected intravenously,label soon appears in the 8-particles. The findings areconsistent with the hypothesis that two kinds of fatparticles are formed during alimentary lipemia: thechylomicrons, produced by the intestine, and a secondgroup containing a mixture of triglycerides derived fromchylomicrons and pre-existing fat stores.

Altered Autonomic Nervous System and Renal Responsesin Primary Aldosteronism. EDWARD G. BIGLIERI,MALCOLMB. MCILROY,* ARNOLDNAIMARK AND PETERH. FORSHAM,* San Francisco, Calif.

Patients with autonomic insufficiency and five subjectswith primary aldosteronism had a prompt and sustaineddiuresis and natriuresis associated with a rise in glomer-ular filtration rate during infusion of isotonic salineat a rate of 50 ml per minute for 60 minutes. Thissimilarity in response led to the investigation of baro-receptor reflex activity in the latter group in viewof the known impairment of this mechanism in theformer. Patients with primary aldosteronism had asimilar impairment of circulatory reflex control: a lackof hypertensive overshoot and bradycardia followingValsalva's maneuver, and a fall in blood pressure andpulse pressure without change in pulse rate whenassuming an upright posture. These responses becamenormal as early as 2 weeks after surgical removal ofunilateral adenomata. Normo- or hypokalemic patientswith essential hypertension and normal subjects givenlarge doses of salt-retaining hormone for 2 to 4 weeksresponded normally to salt infusions and revealed physio-logic circulatory reflex responses. This suggests thatprofound electrolyte abnormalities and excess of aldo-sterone must coexist to produce the observed alterationsin function of the renal and automatic nervous systems.

Measurements of Leukocyte Intracellular pH in ChronicMyelocytic Leukemia. JEROME B. BLOCK AND DAVIDP. RALL, Bethesda, Md. (introduced by Wallace P.Rowe).

Intracellular pH (pH,) measurements may be usefulin characterizing cell populations. Viable leukocytepreparations obtained from patients with chronic myelo-cytic leukemia by techniques of plasmapheresis and dex-tran sedimentation were suspended in Krebs III bufferwith added glucose at 370 C. In 70 experiments atvaried buffer pH (pH,), pH, was determined with theweak acid C`'-labeled 5,5-dimethyl 2,4-oxazolidinedione(DMO) or the weak base C'4-labeled Tris by measuringthe concentration of DMOor Tris in the intracellularwater and extracellular buffer. pH, was found to be de-pendent on pH.. With DMO, below pH. 7.4, pH, wasmaintained higher than pH. by approximately 0.5 to 1.0U; above 7.4 this gradient was reversed. The apparent pHgradient was not altered by incubation at 50 C or in-creasing cell acid production with various drugs, norwas it related to the clinical hematologic status. How-ever, n-butanol, arsenate, or other cell inj ury resultedin reduction of the gradient. Determinations using Trisresulted in uniformly lower values of pH, as comparedwith the use of DMO. No cellular or protein bindingof Tris was found to account for this discrepancy.These observations suggest that a pH gradient may bemaintained across the leukocyte membrane. If pH, ishomogeneous throughout the cell, estimates of pH1 basedon nonionic diffusion should be similar with both DMO

1345


and Tris. The dissimilar results obtained suggest intra-cellular nonhomogeneity with respect to pH.

Hypersensitivity to Staphylococcal Filtrates ProducingFever in Rabbits. PHYLLIS BODEL AND ELISHAATKINS,* New Haven, Conn.

Although many toxins have been isolated from Staphy-lococcus aureus, their role in the pathogenesis and mani-festations of staphylococcal disease remains uncertain.When broth filtrates of staphylococci were injected intorabbits, fever was produced only by those from certainstrains. Rabbits initially unresponsive to filtrate from80-81 staphylococci acquired a febrile response afterintravenous or intradermal infection with the homologousorganism. Other responses to intravenous inoculationof filtrate, including lymphopenia and rapidly developingtolerance to repeated inj ections, are similar to thoseseen in systems involving delayed hypersensitivity. More-over, febrile response to 80-81 filtrate could be passivelytransferred to unreactive normal recipients with mono-nuclear cells from previously infected donor animals.Serum transfer was not effective. Filtrates of VA4and 80-81 strains were studied to isolate the pyrogenicfraction. By methods similar to those used to obtainpurified protein derivative, a heat-stable, antigenic pro-tein fraction was recovered which contained the pyro-genic activity. Carbohydrate fractions were inactive.The pyrogen did not resemble endotoxin. The relation-ship of filterable pyrogen to fever-inducing substancesobtained from both whole and disrupted staphylococcalcells as well as to known staphylococcal toxins such asa-hemolysin and coagulase will be discussed. Thesestudies indicate that delayed hypersensitivity to staphy-lococcal products, whether occurring naturally owing tounapparent infection or induced by infection in thelaboratory, produces fever in rabbits. They also suggestthat such hypersensitivity may play a significant rolein the manifestations of staphylococcal disease. Therole of the staphylococcal toxins may acquire new sig-nificance in view of the close resemblance of staphylo-coccal disease to such chronic diseases as tuberculosisand systemic fungus infections where hypersensitivity hasbeen clearly implicated.

Urinary A-5-Pregnene-3f3,17a,20a,21-Tetrol in 3f3-ol-De-hydrogenase Deficiency. ALFRED M. BONGIOVANNI*AND ALBERT CLARK, Philadelphia, Pa. and Montreal,Canada.

To date 7 cases of the adrenogenital syndrome due to3,6-ol-dehydrogenase deficiency have been studied. Asreported previously the major urinary steroids in allwere present, not in the fraction hydrolyzed by j8-gluco-ronidase but in the residue, after solvolysis and digitoninprecipitation. A novel compound, A-5-pregnene-3,8,17a,-20a,21-tetrol, has been isolated in this condition. Theidentity is based on chemical characterization, elementalanalysis, synthesis of the proposed compound, and suitablecomparisons between the synthetic and natural substances.

This finding is of interest in that it indicates the 21-hydroxylation of steroids with the A-5-pregnene-3,6-olconfiguration, a circumstance hitherto considered unlikelyunder natural conditions. Contrary to our earlier report,not all of these subjects are salt losers nor is the condi-tion uniformly fatal. A partial block with compensatorycortisol biosynthesis is evident in the milder form. Inthe two males studied one had cryptorchidism and theother was a male pseudohermaphrodite with anatomicalstructures of the female. Despite the ambiguity of theexternal genitals in the females, due to labial fusion,careful radiologic study reveals discrete urethral andvaginal orifices rather than a common urogenital sinus.The clinical implications of this will be discussed.Finally, one of the cases also excreted significant quan-tities of 3a,17a,20a-pregnanetriol, suggesting a doubleblock including 21-hydroxylase. In this last case thenovel compound herein described was not present.

An Abnormal Pathway of Steroid Metabolism in Patientswith Glucose-6-Phosphate Dehydrogenase Deficiency.ABRAHAMJ. BORKOWSKI, PAUL A. MARKS,* FRED H.KATZ, MARVIN M. LIPMAN AND NICHOLAS P.CHRISTY,* New York, N. Y.

Before hepatic conjugation with glucuronic acid, ster-oid hormones are reduced in the A-ring. TPNH isessential for this reduction. In subjects deficient inhepatic glucose-6-phosphate dehydrogenase (G6PD), theensuing relative insufficiency of TPNH might be sup-posed to result in a decreased capacity to reduce thesteroidal A-ring, with a consequent increase in excretionof compounds not so reduced. This hypothesis wastested. In five patients with G6PD deficiency, restingexcretion of a nonreduced metabolite of cortisol, 6,B-hydroxycortisol (6,8-OH-F) was studied. The steroidwas extracted and chromatographed by the method ofA. G. Frantz; after isolation, eluates were quantifiedagainst authentic standard by the Porter-Silber reaction.Losses were corrected for by measuring the loss ofradioactivity of added 4-C'4-6,8-OH-F (recovery, 25 to50%). Mean daily excretion of 6,8-OH-F by the 5patients was considerably higher than normal (751 Mug,range 487 to 1,147; 16 normal subjects, mean 349,range 156 to 555; p < 0.01), and significantly higherthan in nine patients with hepatic disease (mean 340Ag). In six subjects 100 mg cortisol was rapidlyinfused intravenously to tax the A-ring reducing mecha-nism acutely. In three, an excessive amount of adminis-tered steroid was metabolized to 6#-OH-F (19.6, 10.1,and 6.0 mg/24 hours; normal not > 3.5 mg); and inthree, a subnormal quantity of exogenous steroid wasexcreted as two major A-ring reduced cortisol metabo-lites, tetrahydrocortisol (THF) and tetrahydrocortisone(THE) (patients, 11 to 22 mg; normal 26 to 30 mg/24 hours). Plasma half-time of infused cortisol wasnormal. Glucuroconjugation was normal in rate andamount as measured by normal metabolism of a reducedsteroid (75 mg THE, intravenously). The data showincreased excretion of cortisol in the form of a com-

1346


pound not reduced in the A-ring. Although other hy-potheses are possible, utilization of this abnormal pathwaycan be explained by insufficiency of TPNH generationowing to hepatic deficiency of G6PD.

Electrophoretic Heterogeneity of Glucose-6-PhosphateDehydrogenase and Its Relationship to Enzyme De-ficiency. SAMUEL H. BOYER AND IAN H. PORTER,Baltimore, Md. (introduced by Victor A. McKusick).

The function of the X-linked genetic locus governingglucose-6-phosphate dehydrogenase (G6PD) deficiency isunknown. It is uncertain whether this locus governsstructure or is concerned with regulation of enzymeactivity through some unknown nonstructural mechanism.The detection of qualitative variation might suggest, onthe model of the hemoglobinopathies, that the X-linkedlocus specifies enzyme structure. We have detectedqualitative variation of G6PD after starch gel electro-phoresis of crude hemolysates from Negro subjects.Such variation was not detected in Caucasians. Bythis method two G6PD types, A and B, were distinguish-able in males; 0.169 (40/236) were A, 0.648 (153/236)were B, while 0.182 (43/236) were G6PD deficient andnot directly typable. Distinction between A and Btypes is also evident after electrophoresis of purifiedenzyme preparations. Crude hemolysates proved un-satisfactory for typing female subjects. Examination ofleukocyte homogenates indicates that three G6PD types,A, A-B, and B, are evident in Negro females. The Atype is less common in females than it is in males.A relationship between G6PD deficiency exists, in thatall examined Negro males with G6PD deficiency haveleukocyte G6PD which is type A. No differences in pHoptimum, Km-G6P, heat stability, Arrhenius constant orinactivation by dehydroisoandrosterone were detected oncomparison of purified preparations of A and B. Sta-tistical, genetic, and electrophoretic evidence suggeststhat the locus governing the G6PD A and B poly-morphism is X-linked and, moreover, appears to be thesame locus as that governing G6PD deficiency. Threealleles are postulated: A normal, B normal, and Adeficient. The existence of electrophoretic variationsuggests that the X-linked locus has a structural func-tion and, in turn, that G6PD deficiency is produced bya structural alteration of the enzyme.

Renal Hemodynamic Effects of Mannitol Infusion. WIL-LIAM E. BRAUN AND LAWRENCE S. LILIENFIELD,Washington, D. C. (introduced by Edward D. Freis).The clinical success reported with mannitol infusions

has prompted a re-evaluation of its renal hemodynamiceffects. Anesthetized dogs were studied after laparot-omy and insertion of ureteral and renal venous cannulae.Clearance and extraction of para-aminohippurate (PAH)and clearance of inulin (GFR) were measured duringcontrol periods, during periods of 20% mannitol in-fusion and during periods of hemorrhagic hypotensionrmean arterial pressure (MAP) = 75 mm Hg] with

and without mannitol infusions. Renal plasma flow(RPF) was calculated from PAH clearance and extrac-tion values. During mannitol infusions in normotensivedogs, the extraction of PAHconsistently and significantlydecreased while the clearance of PAH remained rela-tively unchanged. Inulin clearance consistently decreased.Decrease in extraction indicates change in blood flowdistribution in the kidney; i.e., more blood perfusingnonfunctioning areas such as the medulla. DecreasedGFR in the face of increasing RPF is explained bydilatation of efferent arterioles. In the anesthetizeddogs made hypotensive by hemorrhage, PAH clearance,inulin clearance, and calculated RPF fell markedly.When 20%o mannitol was then constantly infused intothese dogs RPF rose toward control values withoutsignificant changes in MAP. In dogs receiving a con-stant infusion of 20% mannitol during the induction ofhemorrhagic hypotension, PAH clearances and GFR fellto a lesser extent than in those with hemorrhagic hypo-tension alone. In all cases during the infusion therewas a significant reduction in PAH extraction. Thus,RPF was restored to, or remained near control values.These data support the hypothesis that a 20% mannitolinfusion dilates constricted efferent arterioles, especiallythose of the juxtamedullary glomeruli, perhaps by inter-rupting or abolishing renal sympathetic vasoconstrictorreflexes.

Determinants of Ventricular Dimensions in Intact Un-anesthetized Man. EUGENE BRAUNWALD,* ALLANGOLDBLATTAND DONALDC. HARRISON, Bethesda, Md.

Although ventricular volume has been considered animportant variable in circulatory regulation, little in-formation is available on the determinants of ventricularsize in man. Accordingly, a method was developed forthe precise measurement of ventricular dimensionsthroughout the cardiac cycle in closed-chest, intact, un-anesthetized human subjects. Three silver clips weresutured to the surface of either or both ventricles of25 patients undergoing cardiac surgery and, after recov-ery, cineradiograms were obtained at 30 frames persecond. The distance between the clips on the individualframes was measured and related to the intraventricularpressures, which in four subj ects were recorded simul-taneously through a catheter with a micromanometermounted at its tip. During systole, ventricular dimen-sions decreased by an average of 18.5% of the end-diastolic dimensions. In all six patients studied duringnormal inspiration, an increase in right ventricular end-diastolic dimensions, averaging 3.2%, occurred, but therewere no detectable or consistent changes in left ventricu-lar dimensions. During deep inspiration, right ven-tricular end-diastolic dimensions increased by an averageof 9%o of those existing during expiration; the magnitudeof the systolic decrease always varied directly with theend-diastolic size of the preceding cycle, an observationsupporting the concept that Starling's law of the heartoperates in intact man. In five patients isoproterenolconsistently decreased end-systolic size by values up to

1347


13%o of control, increased the rate at which the clipsapproached each other during systole, but did not con-sistently alter end-diastolic dimensions. In six patients,pedaling a bicycle ergometer in the supine position con-sistently decreased both end-diastolic and end-systolicdimensions by an average of 5.4 and 5.0% of controlvalues. In conclusion, the effects on ventricular dimen-sions of a variety of basic interventions have beenelucidated in intact man.

Genetic Implications of Diisopropyl Fluorophosphate(DFP) 32-Erythrocyte Survival Studies in Negro Fe-males Heterozygous for Glucose-6-Phosphate Dehydro-genase Deficiency. GEORGE J. BREWER, ALVIN R.TARLOV AND ROBIN D. POWELL, Chicago, Ill. (intro-duced by Alf S. Alving).Erythrocyte survival curves (DFP32-labeling method)

were obtained twice in each of 14 subjects, once 4months before and once during administration of 30 mgprimaquine daily for 84 days. Subjects were five nor-mal Caucasian females, two normal Negro males, oneNegro male and four Negro females with fully ex-pressed G6PD deficiency, and two heterozygous femaleswith intermediate G6PD deficiency (G6PD values 65%and 30% of mean normal, respectively). Drug adminis-tration to enzyme-deficient subjects (both full and inter-mediate expression) produced curves having two linearphases-an initial brisk fall, then a less rapid fall. Fromthe latter, the life span of those erythrocytes not initiallydestroyed by drug was determined during continueddrug administration and compared with predrug values.Results: 1.) In normal female subjects, primaquinecaused significant shortening of the erythrocyte lifespan (predrug average: 114 days; with drug, 96 days).2.) In fully expressed G6PD-deficient subjects receivingdrug, approximately 50% of the erythrocytes were de-stroyed initially; the life span of the remaining cellswas 63 ± 7 days. 3.) In the two heterozygotes receivingdrug, initial red cell destruction was 42 and 72%; thelife span of the remaining cells was 66 and 75 days,respectively; there was no segment of the erythrocytepopulation whose survival was comparable with thatof normal controls. Results in the two heterozygotesindicate that the "mosaic" hypothesis of X-chromosomalinactivation, which suggests two genetically differentpopulations of cells in heterozygotes, is not applicablein G6PD deficiency. This hypothesis, recently extendedto include G6PD deficiency, suggests a variable mixtureof populations of completely normal and mutant eryth-rocytes in heterozygotes. The studies reported heredemonstrate that all erythrocytes in the heterozygousG6PD-deficient subjects are involved with the geneticdefect.

Active "Downhill" Sodium Transport Across the Iso-lated Frog Skin. NEAL S. BRICKER * AND THOMASBIBER, Copenhagen, Denmark.

Frog skin epithelial cells represent a prototype ofsystems which actively transport sodium transcellularly

while simultaneously preserving their own intracellularfluid composition. These cells exhibit striking func-tional polarization: at the external surface, Na enterscell water passively; at the internal surface Na is activelyextruded and K moves bidirectionally-inward by activetransport, outward by passive diffusion. Active transportof Na and K often is attributed to a single pump with 1 :1linkage. The pump therefore would be nonelectrogenic,and transcellular potential difference (PD) would relatemainly to Na and K diffusion potentials. In the steadystate, Nap influx= NaA efflux = KA influx= KP efflux(P = passive; A = active). In the present studies, KPefflux was modified by exposing the internal surface ofthe isolated frog skin to potassium concentrations greaterthan estimated intracellular concentrations. Frog Ring-er's was used as the external solution. PD short circuitcurrent (SCC) and electrical resistance all decreasedinitially. PD stabilized at approximately 20%o of con-trol values, resistance remained low, but SCC rose,frequently exceeding control values. Net sodium trans-port not only persisted, but appreciably exceeded SCC.This unusual discrepancy between these two parameterswas related largely to transcellular movement of K.Because Na transport occurred down a steep electro-chemical gradient, the process could be entirely passive.However, during complete short circuiting, the ratio ofunidirectional sodium fluxes (using Na22 and Na') ap-preciably exceeded the ratio of sodium concentrations.Furthermore, after addition of strophanthin (1.4 X10'M), flux ratios promptly fell to equal concentrationratios. Finally, when the Na gradient was abolishedby using hyperosmotic internal solutions (Na 114, K 120mmoles/L), net Na transport continued. These datatherefore appear to document the phenomenon of down-hill active sodium transport.

Water-Water Interaction during Solvent Flow in Iso-lated Turtle Bladders. WILLIAM A. BRODSKY* ANDTHEODOREP. SCHILB, Louisville Ky.

Isolated turtle bladders, filled with hypotonic solutionsof sucrose, and immersed in oxygenated Ringer's solu-tions, were impermeable to serosal electrolyte and tomucosal sucrose. Consequently, a fairly pure process ofosmosis was observed. Osmotic flow of solvent (relativeweight loss/hour of bladder sac and content) was pro-portional to the square of the transmembrane osmoticgradient, or R= Rw (AOs)2/(AOs) .2, where R is rateof weight loss of hemibladder with sucrose divided bythe rate of its paired control containing 110 mMNaCl;Rw is rate of weight loss of a hemibladder with dis-tilled water divided by the rate of its paired control;AOs is serosal minus sac fluid osmotic activity in sacswith hypotonic sucrose; and (,Os) is serosal minussac osmotic activity for sacs containing distilled water.The approximate equality between theoretically derivedand experimentally obtained values of net relative waterflow confirms the concept that osmotic flow of solventis due to a two-force mechanism. One force, the chemi-cal potential gradient, initiates difflusional flow of water;

1348


the other force, frictional drag of flowing water uponsurrounding water, accelerates flow of solvent. Additionof salt to mucosal fluid adds a third force-that due toactive ion transport. Water movement occurs acrosssacs containing NaCl even though transmembrane AOs' 0, but does not occur across sacs containing iso-

osmotic sucrose alone.

Human Erythrocyte pH with Special Reference to Intra-cellular Acidosis in Hereditary Spherocytosis. P. A.BROMBERG,J. THEODORE, E. D. ROBIN * AND W. N.JENSEN,* Pittsburgh, Pa.Human erythrocyte pH was investigated by means

of the indicator substance 5,5-dimethyl-2,4-oxazoli-dinedione (DMO), and the results compared with simul-taneous measurements based on chloride distributionbetween cells and plasma and by direct glass-electrodepH measurement of packed cells lysed by a freeze-thawprocedure. In 10 simultaneous determinations by eachmethod in six subjects without demonstrable hematologicabnormality, erythrocyte pH was 7.13 ± 0.03 (DMO),7.15 ± 0.05 (electrode), and 7.25 + 0.03 (chloride) at amean plasma pH of 7.44 ± 0.01. The induction oftransient moderate respiratory acidosis and alkalosis andof metabolic alkalosis by NaHCO3 infusion induced pHchanges of similar direction and magnitude in erythro-cytes and plasma. Erythrocyte pH was measured infive patients with active hemolytic anemias [sickle cellanemia (3), paroxysmal nocturnal hemoglobinuria (1),acquired hemolytic anemia (1)]. There were no dis-cernible differences in pH between these and the controlgroup. In contrast, three patients with hereditaryspherocytosis and intact spleens had DMO-determinederythrocyte pH values of 7.00, 6.93, and 7.04 whenplasma pH values were, respectively, 7.40, 7.42, and7.41. Three additional postsplenectomy patients withhereditary spherocytosis had erythrocyte pH values of7.05, 7.13, and 7.13 when plasma pH values were, re-spectively, 7.41, 7.43, and 7.37. The normal erythrocytepH value of 7.13 ± 0.03 as determined by the DMOmethod implies that erythrocyte [H+] is approximatelytwice that of plasma. This normal [H+] gradient hasbeen attributed to Donnan equilibrium produced by non-diffusible, negatively charged molecules within theerythrocyte. The unique finding of an increased erythro-cyte-plasma [H+] gradient in four of six patients withhereditary spherocytosis may, therefore, be related to anexcessive intracellular concentration of nondiffusiblenegatively charged particles. The possible relationshipof the erythrocyte acidosis to genetically determinedmetabolic abnormalities remains to be investigated.

Measurements of K+ Exchange in Separated Renal Tu-bules. MAURICE BURG, MONES BERMANAND JACKORLOFF,* Bethesda, Md.A method has been developed for the separation and

study of viable segments of renal tubules from rabbitrenal cortex. Separation is achieved by treatment ofthe excised kidney with the proteolytic enzyme, col-

lagenase. The resultant suspension consists principallyof short lengths of proximal tubules. Preservation ofrespiration and active transport in the intact cells isindicated by: oxygen consumption 8.0 ± 0.6 ,uL mg'(dry wt) hour', K+ concentration 303 ±21 mEq kg'(dry wt), and para-aminohippurate tissue to mediumratio 21 +3. The use of this preparation affords ameans of directly measuring transcellular exchange rates.This has not been possible in the conventional, relativelythick, kidney-slice preparations, since penetration throughthe outer layers of the tissue may limit rate of exchangein the centrally located cells. In addition, use of thesuspension permits better sampling and more uniformdistribution of oxygen, substrates, and test substances.The steady-state transcellular K+ flux rate in suspendedrabbit renal tubules has been measured with K4. Atleast two tissue K+ compartments were observed undercontrol conditions, one large and slowly exchanging,the other small and rapidly exchanging. Addition of thecardiac aglycone strophanthidin (3 X 10' M) resultedin a 98% reduction of the steady state K+ flux in themajor K+ compartment. This observation is consistentwith the known inhibitory effects of digitalis-like steroidsupon active K+ transport. In addition, the rate constantfor K+ efflux decreased 85%, indicating either that the"passive" permeability to K+ was decreased or, morelikely, that there is in this tissue considerable facilitatedor exchange diffusion of K+ that is inhibited by the drug.

Visualization of Myocardial Infarcts In Vivo by Scan-ning. EDWARDA. CARR, JR., MARY ELLEN PATNOAND WILLIAM H. BEIERWALTES, Ann Arbor, Mich.(introduced by Fred M. Davenport).Direct demonstration of myocardial infarcts in 'ivo

has not previously been feasible. Nine dogs were sub-jected to ligation of the anterior descending branch ofthe left coronary artery under pentobarbital anesthesiaand the incisions were closed; 1 to 21 days later eachdog received 700 ,uc of Hg`3-labeled chlormerodrin(Neohydrin) intravenously. Five sham-operated dogswere treated similarly. All dogs were scanned in thesupine position, while still alive, 1.5 to 24 hours afteradministration of the radioisotope. A photoscanner witha 3 X 2 inch crystal and 19-hole collimator was used.After scanning, each dog was sacrificed and the hearttransfixed with needles through the anterior chest wall,which was then removed with the dog still in scanningposition. The heart outline was projected onto thescan. Each heart was then excised, emptied of blood,and rescanned. The radioactivity of tissue sampleswas subsequently determined in a well-counter and thesamples examined histologically. Areas of frank necrosisor of ischemia (i.e., severely compromised blood supplywithout frank necrosis) were clearly seen as "hot" areasof increased radioactivity in scans of living dogs. Theresults in sham-operated dogs were negative. Thefindings were confirmed in scans of excised hearts. Inligated dogs the mean relative Hg`'3 concentrations(concentration of Hg`"3 in tissue per concentration of

1349


Hg"3 in blood) in posterior left ventricle (normal) andanterior left ventricle (infarct or ischemia) were 0.85and 2.78, respectively. This difference is significant(p < 0.025). In sham-operated dogs the mean relativeHg203 concentrations in posterior left ventricle (normal)and anterior left ventricle (normal) were 0.60 and 0.56,respectively. This difference is not significant. Scanspermit diagnosis of myocardial infarcts in vivo by local-ization of Hg"3, given as chlormerodrin. The appli-cability of this technique to man is now being assessed.

Osmolality, pH, and Inulin Concentration in ProximalTubular Fluid of the Dog. JAMES R. CLAPP ANDJOHN F. WATSON, Bethesda, Md. (introduced byRobert W. Berliner).The following experiments were undertaken to adapt

micropuncture and microanalytical techniques to thestudy of renal function in dogs. Micropunctures havebeen limited to the proximal tubule, of which 50% isaccessible for puncture. Tubular fluid samples havebeen obtained for osmolality, pH, and inulin concentra-tion; and urine flow, inulin, and para-aminohippurate(PAH) clearances were simultaneously measured inthe experimental and control kidney. Sixteen samplesfor osmolality collected during antidiuresis were isos-motic to plasma, with tubular fluid/plasma (TF/P)ratios ranging from 0.96 to 1.04, while urine/plasma(U/P) ratios ranged from 2.90 to 5.82. Fourteensamples collected during water diuresis also were isos-motic to plasma with TF/P ratios ranging from 0.98to 1.03 when U/P ratios ranged from 0.11 to 0.37.pH determined by quinhydrone microelectrodes on 15tubular fluid samples during antidiuresis ranged fromisohydric to 0.37 pH units below arterial blood, with themajority of the samples exhibiting some degree ofacidification. Inulin concentrations determined on 11samples by a microanthrone method, although revealingconsiderable variation among values for individual tu-bules, had TF/P inulin ratios ranging from 1.5 to 5.5,with the highest ratios located between 28 and 50%of the proximal tubule. In the experiments in whichurine flow, inulin, and PAH clearances were simul-taneously measured, mean decreases of 7, 13, and10%o, respectively, occurred in the experimental ascompared with the control kidney. It is apparent fromthese experiments that micropuncture studies can beperformed without appreciably altering renal function.Under the conditions of these experiments, the osmo-lality, pH, and inulin concentrations of proximal tubularfluid of the dog do not differ significantly from thosefound in rodents.

Isolation and CharacterizationPhysical Techniques. ALANCALKINS, Boston, Mass. andduced by Chester S. Keefer).

of Amyloid Fibrils byS. COHEN AND EVANBuffalo, N. Y. (intro-

Due to difficulties in the separation of amyloid fromsurrounding tissue, previously reported chemical charac-

teristics of this substance have been based upon thecomparison of massively involved amyloid liver tonormal human liver. Recently, techniques have beendevised to take advantage of the insolubility of amyloidand its relative hydrophilia in order to separate it fromhepatic tissue. Amyloid-laden and normal livers ob-tained at autopsy have been homogenized at 0° C,washed in saline, and spun five times in a refrigeratedcentrifuge at 5,000 rpm for periods of 1 hour. In thisfashion three distinct layers were obtained from theamyloidotic liver and two from the normal. Afterseparation of all layers, aliquots of each were stainedwith phosphotungstic acid, shadowed, and examined inthe electron microscope. The top layer of the amyloidliver, in contrast to the other two layers and those ofthe control, consisted almost entirely of fine fibrils com-parable with those seen in ultrathin sections of amyloidtissue. The calculated fibril diameter was found to bedependent upon the type of material examined (freshlyseparated, frozen, or dialyzed lyophilized material) andupon the type of preparation utilized for electron micros-copy (unshadowed, chromium- or platinum-shadowed,or cleaved mica preparation). In all instances, individualfibrils were under 300 A in diameter. They variedconsiderably in length up to 5,000 A. X-ray powderdiffraction pattern studies of all layers after dialysis andlyophilization confirmed the presence of a componentin the top layer that was not demonstrable in theother two. Chemical analyses and enzyme digestionstudies were also carried out on the concentrated amyloidfibrils. No differences were found by any technique inamyloid fibrils isolated from patients with the primaryor secondary form of the disease.

Androgen Production and Biosynthesis by AdrenalGlands from Patients with Normal and HyperadrenalFunction. GEORGEL. COHNAND PATRICK MULROW,New Haven, Conn. (introduced by Seymour R.Lipsky).

Adrenal androgen production was studied in vitrowith slices from 6 "normal," 7 hyperplastic, 3 adenom-atous, 2 atrophic, and 2 carcinomatous adrenal glandsremoved from three patients with normal, and ninewith hyperadrenal function. Selected flasks receivedeither 10 U ACTH per g tissue, progesterone-H3,dehydroepiandrosterone-H3, or equimolar amounts of17a-hydroxypregnenolone-HW and progesterone-4-C"4.Certain flasks without radioactive substrates receivedC`4-labeled androstenedione and 1ip-hydroxyandrostene-dione and dehydroepiandrosterone-H3 at the conclusionof the incubation to correct for manipulative losses. The17-ketosteroids were isolated by paper chromatographyin three different solvent systems and quantitated byalkaline fluorescence or microspectrophotometry. Radio-activity was determined by liquid scintillation andplanchet counting. The range of androgen steroid pro-duction (Ag/g/3 hrs) in all tissues without ACTH was

1350


dehydroepiandrosterone, 0 to 14.3; androstenedione, 0 to5.4; 11 8-hydroxyandrostenedione, 0 to 11.3. WithACTH it was 0 to 5.3, 0 to 12.6, and 0 to 47.4, respec-tively. ACTH resulted in a 0 to 34-fold increase in1ilp-hydroxyandrostenedione release, but an inconstantincrease in androstenedione and dehydroepiandrosteronerelease. Adrenal androgen production was not signifi-cantly different between the "normal" and atrophictissue, as compared with the pathological tissues. Pre-cursor experiments revealed that 5 to 7% of 17a-hydroxy-pregnenolone-H3 was incorporated into dehydroepian-drosterone; and 5 to 13% of 17a-hydroxypregnenolone-H'into androstenedione and llp-hydroxyandrostenedione as

compared with 0.2 to 0.8% of C4- (or H')-progesterone.In addition, dehydroepiandrosterone-H' was incorporatedinto androstenedione and 1lp-hydroxyandrostenedione.Radiochemical purity was verified by countercurrentdistribution, paper and column chromatography of thefree C' steroids, and their 2,4-dinitrophenylhydrazonederivatives. The evidence suggests that the major adrenalandrogen biosynthetic pathway in the human adultis 17a-hydroxypregnenolone>dehydroepiandrosterone-androstenedione->1 lp-hydroxyandrostenedione.

A New Coagulation Factor Vulnerable to Thorium andRelated Elements. ROBERTW. COLMANAND BENJA-MIN ALEXANDER,* Boston, Mass.

The unique anticoagulant action of thorium and relatedrare earth elements was investigated in vivo and invitro. Colloidal thorium dioxide infused into rabbitsprolonged venous blood clotting, destroyed Factor V inplasma, and reversibly inactivated a hitherto undescribedfactor in serum. The same abnormality was rapidly andprogressively produced by adding ThCl4 to serum itn vitro(final concentration, 3.0 X 10' M). This factor, essentialfor thromboplastin generation, was not Factor IX, X, XI,or XII. Destroyed also by uranium, neodymium, and lan-thanum, this factor is adsorbed by BaSO4, destroyed at50' in 20 minutes, 370 in 2 hours, and 4' C in 2 weeks.The thorium-induced defect in serum can be reversedby oxalate or citrate, both of which bind thorium. Theentity has been separated from other coagulation factorsby BaSO4 adsorption, citrate elution, ammonium sulfatefractionation, and starch gel electrophoresis. An assay

procedure has been developed, and the factor was foundto be profoundly depressed in patients treated withcoumarin derivatives, in some patients with liver disease,and in newborn infants. Unlike Factors VII and X,it is not elevated in pregnancy. Its level under physio-logical and pathological circumstances does not paral-lel the concentration of other serum clotting components.

The data indicate the existence of a new factor essentialfor thromboplastin generation, and suggest an importantrelationship between the atomic structure common to

these elements and a structural site crucial to the bio-logical activity of this factor.

The Importance of Conjugation with Glutathione onSulfobromophthalein (BSP) Transport from Blood toBile. BURTONCOMBES, Dallas, Tex. (introduced byGladys Fashena).

An examination of the effect of glutathione on BSPmovement from blood to bile was undertaken to deter-mine whether conjugation of glutathione with BSP isimportant for, or merely incidental to, the hepatic up-take and biliary excretion of BSP. Advantage wastaken of the finding that hepatic glutathione content inthe rat decreased 81%o (from 8.3 to 1.6 mg/100 g bodyweight) after 2 days on a protein-free diet. Controlrats excreted more BSP into bile than did glutathione-deficient rats with intravenous doses of BSP (5 to10 mg/100 g). Maximal BSP excretion in bile col-lected for 45 minutes of 4.25 per 100 g (3.56 mg con-jugated, 0.69 mg free BSP) was reached in controlanimals when 10 mg per 100 g was injected. Only2.94 mg BSP per 100 g was excreted in the glutathione-deficient rats at the dose of 10 mg per 100 g. Thedecrease in total BSP output was accounted for by afall in excretion of conjugated BSP, 1.99 mg conjugated,0.95 mg free. Conjugation was not necessary forhepatic uptake of BSP. The amount of BSP takenup by the liver of rats with ligated common bile ductswas the same in control and in glutathione-deficientanimals. Furthermore, the extent of BSP uptake inthese livers was not affected, although the proportion ofconjugated BSP contained in the liver increased overa 45-minute interval. In other studies, analysis of thedistribution of BSP compounds contained in liver withthose just excreted into bile revealed that conjugatedBSP was preferentially excreted. Glutathione-deficientrats contained less conjugated BSP in liver and excretedless conjugated BSP into bile than did control animals.These observations indicate that conjugation of BSPwith glutathione is an important determinant of themaximal rate at which BSP is transferred from bloodto bile. Transport from the liver cell into bile appearsto be the conjugation-dependent step, conjugated BSPbeing transported more rapidly than is free BSP.

Anatomic Structure of Lymphatic Tissues and ImmuneFunction. CHARLES C. CONGDON,* Oak Ridge, Tenn.

For many years attempts have been made to relatehistologic changes in lymphatic tissues to function ofthe immune mechanism. The task of demonstrating thatimmune function resides in lymphatic tissues was finishedearlier and has been repeatedly confirmed. Immunefunction outide the lymphatic tissues we can attributeto disbursement of mobile cells from these tissues.Within the lymphatic tissues new histologic studies seemto indicate clearly that immune function begins in thecortex of lymph nodes and the white pulp of spleen.Changes in the medulla of lymph nodes and red pulpof spleen depend on the use of certain kinds of antigensand follow the primary changes in cortical and whitepulp areas of the lymphatic tissues. The primary

1351

PROC1?iEDINGS OS THE FIFTY-FOURTH ANNUALMEETING

changes are qualitatively the same after all kinds ofantigenic stimulation. For this reason the primaryhistologic changes constitute a potential method fordiagnosing the presence of immunologic function. Withinthe cortex of lymph nodes and white pulp of spleenit is necessary to account for changes after antigenicstimulation in marginal zone cells, lymphocytes, germinalcenters, tingible bodies, and stromal elements. Thepresent work indicates that the germinal center is themajor histologic structure associated with immune func-tion. The tingible body is also a key-anatomic structurewhose function we do not understand, but perhaps isconcerned with the recognition of the antigenic materialthrough phagocytosis of lymphocytes and granulocytes.The present work also requires a precursor cell for thegerminal center. The two possibilities seem to be afixed reticulum cell or the lymphocyte. The ultimateprecursor cell may be the thymus. All questions ofimmune function and disease will need eventually todeal with these several anatomic structures.

Vitamin-B,, Binding by Normal Human Serum. BER-NARD A. COOPERAND Louis LOWENSTEIN, Montreal,Canada (introduced by David V. Bates).

Since the demonstration that vitamin B. in humanserum is bound to protein, numerous studies have at-tempted to characterize the protein and measure thebinding capacity. Diverse techniques have been used tostudy this phenomenon, and the values obtained havevaried with the techniques used. Consequently, widelydivergent values for cyanocobalamin-binding capacity ofhuman serum have been reported. The physiologicalimportance of this binding is unknown, although wehave reported recently that much larger quantities ofcyanocobalamin bound to human serum protein thanof free cyanocobalamin are taken up by certain tumorcells incubated in vitro. The rate of clearance from theplasma in vivo of injected cyanacobalamin is altered bybinding to the binding protein of serum. Using severaltechniques simultaneously, we have studied cyanoco-balamin binding by normal serum, to evaluate techniquesand to reconcile the differing results reported previously.The values obtained for cyanocobalamin-binding capacityof normal serum, measured by a modified charcoal-adsorption technique, paper electrophoresis, ultrafiltration,and the stimulation of uptake of cyanocobalamin byEhrlich ascites cells in vitro, are approximately equal.The much lower binding capacity determined withEuglena gracilis seems to be caused by the reducedaffinity of the binding protein for cyanocobalamin atpH 3 to 4. The cyanocobalamin-binding capacity atpH 3 to 4, measured by the modified charcoal-adsorptiontechnique, is changed to a value comparable to thatobtained with E. gracilis. Physiochemical studies indi-cate that, at physiological pH and temperature, thebinding protein binds cyanocobalamin tenaciously, andpreferentially to pseudovitamin B,2. Under carefullycontrolled conditions, normal human serum has a finite

binding capacity for vitamin B. and specific physical,chemical, and biological properties.

Familial Pyridoxine-responsive Anemia. H. BROOKSCOTTON, Fairfield, Ala. AND JOHN W. HARRIS,*Cleveland, Ohio.

Severe pyridoxine-responsive anemia has been demon-strated in a 44 year old white female; moderatelysevere, pyridoxine-responsive anemia in her 43 year oldbrother; and hyperferremia in three (one female) offive children (6 to 18 years). The patients' mother,sister, six paternal, and two maternal cousins showedno hematologic abnormalities and serum iron values werenormal; their father died of "cirrhosis" (no recordsavailable). Anemia of 32 years is documented for thefemale with exacerbations during pregnancies requiringnine transfusions; the erythrocytes have consistentlybeen described as hypochromic and microcytic despiteover 2 g of parenteral iron. Pyridoxine therapy resultedon three occasions in subjective improvement, reticulo-cytosis, and hemoglobin increases (7 g/100 ml before,13 g/100 ml after). The serum iron was consistentlyelevated to the 200 to 300 ,ug/100 ml range duringrelapse, and the erythrocyte Cr` half-life was 24 dayswith no sequestration by the moderately enlarged spleen.Anemia of varying severity (8 to 11 g Hb/100 ml) for15 years is documented for the brother. The erythrocyteshave appeared slightly hypochromic and microcytic, theserum iron elevated to 200 Iug per 100 ml. Pyridoxinetherapy was followed by mild reticulocytosis, hemoglobinincrease (10 g/100 ml to 14 g/100 ml), and fall tonormal range in the serum iron; the spleen became nolonger palpable (formerly 3 finger-breadths enlarged).Both patients had hematologic responses to oral orparenteral administration of various forms of pyridoxinebut not to the usual hematopoietic agents. The erythro-cyte abnormalities persist despite achievement of normalhemoglobin levels but no abnormalities in tryptophanmetabolism have been found upon repeated loading tests.The hemoglobin showed normal electrophoretic migra-tion and no elevation of A. or F hemoglobin. Noanemia or erythrocyte abnormalities were found in thechildren. In the one child tested by pyridoxine adminis-tration the serum iron fell from 254 to 139 /Ag per 100 ml.

Fatty Acid Composition of Erythrocyte Phospholipidsin Anemic Patients. W. T. DABNEY, III AND C. W.CHAMBERLAIN, Richmond, Va. (introduced by G. Wat-son James, III).The fatty acid composition of the phospholipid fraction

of the erythrocytes from 40 individuals has been investi-gated. Included are 10 normal males, 9 normal females,and 21 patients with a hemoglobin less than 10 g per100 ml. The total lipid of saline-washed erythrocytesfrom heparinized venous blood was extracted and thephospholipid fraction separated on a silicic acid column.An analysis of the methyl esters of the fatty acids wasperformed by gas-liquid phase chromatography withethylene glycol succinate polyester as the stationary

1352


phase. At least 17 fatty acids are associated with theerythrocyte phospholipids. Palmitic, stearic, oleic, lino-leic, and arachidonic acid are the major fatty acidspresent. No qualitative difference was observed in thefatty acid composition of the normal or anemic group.

A quantitative comparison of the per cent composition ofthe major fatty acids revealed a significant difference.Analysis of the data revealed a significant decrease(p <0.001) in the percentage of linoleic acid in erythro-cytes from anemic subjects when compared with normalindividuals. No significant difference was found in theother fatty acids with the exception of a slight increasein oleic (p, 0.01). There is a significant positive corre-

lation (r = 0.715, p < 0.001) between the hemoglobinconcentration and percentage of linoleic acid if the two

study groups are combined, but no significant correla-tion was found within the anemic group or normal group

when analyzed separately. Current studies indicate a

similar quantitative change in the fatty acid compositionof the plasma phospholipids and neutral lipids in severe

chronic anemia. The decrease in linoleic acid does notappear to be related to the etiology or morphologiccharacteristics of the anemia.

Hemodynamic Consequences of Oxygen Breathing inLeft Ventricular Failure. WALTERJ. DALY AND RoYH. BEHNIKE, Indianapolis, Ind. (introduced by JohnHickam).

Breathing oxygen decreases the heart rate and cardiacindex of resting normal man. Using standard techniquesof cardiac catheterization and an indocyanine green di-lution method, we studied the hemodynamic effects ofhyperoxia in 15 nonhypoxemic patients with left ven-

tricular failure (mean arterial oxygen saturation, 94.62.3%). Air or oxygen was breathed in random se-

quence. Oxygen breathing was associated with a

decrease in cardiac index from 2.36 0.71 to 2.07 0.45L per minute per m' (p <0.001); heart rate, 84 + 25 to81 ±25 beats per minute (p < 0.001); stroke index,33 19 to 30 + 18 ml per mt (p, 0.005). The effects on

A-V 02 difference (6.87 vol %), 02 delivery index(arterial 02 content X cardiac index = 438 + 163 ml 0,/

min/m2), systemic mean arterial pressure (106 + 13 mm

Hg), pulmonary artery mean pressure (33 + 11 mmHg),wedge pressure (20 + 6 mmHg) were not statisticallysignificant. Contrary to observations made in othersituations of pulmonary hypertension, oxygen breathingdid not alter pulmonary vascular resistance (296 + 176dynes-sec-cm'). In normal man the decrease in cardiacindex during oxygen breathing occurs without a changein stroke index. In heart failure, the chronotropic effectof hyperoxic breathing remains effective; furthermore,a small but consistent decrease in stroke index is ob-served. These data suggest that oxygen breathingfurther decreases the already reduced cardiac output ofleft ventricular failure without impairing the delivery ofoxygen to the tissues. Such changes would be expectedto reduce the work of the failing heart without furtherimpairing its ability to meet the demands placed upon it.

The Formation of Free Hydroxyproline by Cartilage.WILLIAM H. DAUGHADAY* AND IDA KOZAK MARIZ,St. Louis, Mo.

Formation of free hydroxyproline by cartilage hasbeen examined because of its relevance to the importantreaction involved in collagen synthesis by which prolineis hydroxylated. Segments of costal cartilage fromrapidly growing rats were incubated in an enrichedmedium containing L-proline-U-C"4. After incubation for24 hours, free hydroxyproline in the medium and boundhydroxyproline in whole cartilage or cartilage fractionshas been isolated by Dowex-50 chromatography and itsspecific activity determined by the method of Prockop,Udenfriend and Lindstaedt. Between 8 and 10 ,ug of freehydroxyproline per 100 mg of wet cartilage was releasedinto the medium. This hydroxyproline had a specific ac-tivity 20 times that of hydroxyproline of collagen solublein 0.45 MNaCl and 65 times that of hydroxyproline fromcollagen insoluble in 0.45 MNaCl. Cortisol (2 to 10 Sag/ml inhibited incorporation of Cl' into cartilage-bound hy-droxyproline by about 30%, but caused a small (12%)stimulation of labeling of medium free hydroxyproline.Conclusion: Free hydroxyproline formed by cartilagein the system is not derived from soluble collagen butprobably arises from an activated hydroxyproline inter-mediate that exists prior to the stage of peptide syn-thesis. Formation of this intermediate apparently isnot inhibited by cortisol, although cortisol appears toinhibit collagen synthesis.

The Effect of Adenosine Derivatives on Ion Transportand Energy Metabolism in the Toad Bladder. ROBERTP. DAVIS, MITZY CANESSA-FIScHER AND CHESTERM.EDELMANN, JR., New York, N. Y. (introduced byI. Herbert Scheinberg).

The energy for active transport of sodium is derivedfrom high-energy phosphate bonds and at least in partfrom cellular respiration. Since the major regulator ofcellular respiration is adenosine diphosphate, we hopedto clarify the linkage between transport of sodiumand cellular respiration by studying the effects of thisnucleotide on these processes in the urinary bladder ofthe toad. Transport of sodium was measured by short-circuit current and respiration polarographically by thevibrating platinum electrode. Reduced pyridine nucleo-tide, through which high-energy phosphate bond forma-tion is coupled to the reactions of cellular respiration,was identified in the intact functoning bladder by itsfluorescence. Three phenomena were observed whenadenospine diphosphate was applied to the toad bladder:1) respiration was acclerated, 2) the concentration ofreduced pyridine nucleotide was decreased, and 3) therewas a pronounced decrease in the transport of sodium.In contrast, adenosine, adenosine 5'-monophosphate(AMP) and cyclic 3',5'-AMP produced a significantinitial increase in transport of sodium, presumably duringinitial phosphorylation reactions. There then followeda decrease in the transport of sodium comparable with

1353

PROCEEDINGSOF THE FIFTY-FOURTHI ANNUALMEETING

that produced by adenosine diphosphate. Since adenosinediphosphate leads to opposite changes in respiration andtransport of sodium, a direct linkage through a redoxpump between the electron transport aspect of respirationand transport of sodium appears unlikely. In contrast,the changes in transport of sodium produced by adenosinederivatives parallel closely the changes in the concentra-tion of reduced pyridine nucleotide. From these results,we infer that an intermediate in the reaction sequencebetween reduced pyridine nucleotide and the phosphory-lation of adenosine diphosphate is the point of couplingbetween energy metabolism and ion transport.

Evidence for a Second Iron-binding System in Plasma.RAYMONDJ. DERN, ALBERTO MONTI AND MICHAEL F.GLYNN, Chicago, Ill. (introduced by Richard L.Landau).

Plasma iron was found to be cleared at two differentrates that are apparently unrelated to feedback kinetics:a rapidly cleared iron fraction accepted as transferrin-bound, and a slowly cleared iron fraction. Radioiron-labeled plasma was obtained from donors 2 days afterinjection of Fe'9. This was incubated with Fe"5 in theaccepted method of labeling transferrin and then inject-ing the doubly-labeled plasma into recipients. The dis-appearance rate of Fe' in the recipient was one-halfto one-third of that observed for Fe5". Since the recipienthad no tracer in the labile erythropoietic iron pool,the slow rate was not related to feedback kinetics.Recipients were selected to provide a spectrum of serumiron levels and storage iron. In three donor-recipientstudies, the clearance rate of the second binding com-ponent was not related to the iron status of the recipientsbut appeared to be somewhat faster with plasma froman iron-poor donor. When plasma was obtained fromdonors within 2 hours after oral or intravenous adminis-tration of Fe' and incubated with Fe"5 as before, thedisappearance rates in the recipients for both isotopeswere the same. It is suggested that plasma iron isbound in two ways: the transferrin(s) system for in-jected and absorbed iron; and a second iron-bindingsystem representing, in part, recirculating iron. Theresults of current studies of erythrocyte iron utiliza-tion and differential behavior of the two forms of irontransport will be discussed.

Abnormal Neural Regulation of Vasopressin and Aldo-sterone Secretion in a Patient with Cerebral Hypo-natremia. JOSEPH F. DINGMAN, AKIRA ARIMURA,RALPH E. PETERSON* AND ROBERT G. HEATH. Bos-ton, Mass., New York, N. Y. and New Orleans, La.A male patient with cerebral metastases from broncho-

genic carcinoma developed symptomatic hyponatremiawhich was partially corrected by high Na and ad libfluid intake, and completely reversed by high Na andlow fluid regimen. Low Na intake resulted in continuedrenal Na loss and return of hyponatremia which couldnot be prevented by fluid restriction. Ad lib hydration

accelerated the onset of hyponatremia with Na restrictionand resulted in higher Na excretion than during dehydra-tion. Urinary aldosterone was depressed during highNa intake but failed to increase normally on low Nadiet. The adrenals failed to respond to methopyrapone.ACTH administered intravenously during low Na regi-men produced a normal steroidal response and a markedrise in aldosterone to levels observed in Na-depletednormal subjects. These findings suggest a derangementin the physiological regulation of aldosterone secretionrelated to a disturbance in the neural control of ACTHsecretion. The typical abnormal response to waterloading was apparent even when serum Na and bodyhydration were within normal range. Vasopressin wasextracted from acidified plasma, purified on IRC-50columns and silicic acid-impregnated glass paper chro-matograms, and specific vasopressin zones were bio-assayed in the alcohol-anesthetized hydrated rat. Plasmavasopressin was five times normal on ad lib hydration,failed to fall with water loading, and was much inexcess of the normal response after intravenous nicotine.Curiously, the patient maintained normal serum Na onad lib Na and water intake in the last few weeks oflife. Postmortem examination revealed bilateral destruc-tion of the septal region and subtotal destruction of theadrenals, but the hypothalamic-pituitary region was un-affected. These findings suggest that the hyponatremiaresulted from involvement of subcortical centers regulatingsecretion of hypothalamic, pituitary, and adrenal hormoneswhich control the plasma concentration and the renal ex-cretion of Na and water.

Demonstration of an Exchange between Free and In-trinsic Factor-bound Vitamin B12. ROBERT M. DON-ALDSON, JR. JAY H. KATZ AND JOSEPH D. DIMASE,Boston, Mass. (introduced by M. B. Strauss).The source of "intrinsic factor" (IF) used in these

studies was neutralized human or rat gastric juice,which was shown to promote the absorption of vitaminB,, in patients with pernicious anemia or in the gas-trectomized rat, respectively. Gastric juice, the bindingcapacity of which had been saturated with Co'-labeledvitamin B12, was mixed with free Co5"-labeled vitaminB12. When the IF-bound B12 was subsequently sepa-rated from free B12, a significant quantity of the initiallyfree Co5TB12 was now bound, while an equal amountof the originally bound Co°B,2 had become free. Ex-change between the two isotopes was demonstratedwhether bound and free vitamin B12 were separated bydialysis, by starch gel electrophoresis, or by dextrangel filtration. The rate of exchange was temperature-dependent. When free and bound vitamin B,2 weremixed in equal amounts at 370 C, 40% of the theo-retically possible exchange occurred within 2 hours.Neither the binding capacity of the gastric juice northe electrophoretic mobility of the bound radioactivitywas altered after exchange had occurred. Schillingtests were performed in pernicious anemia patients withmixtures of equal amounts of free Co'B12 and IF-bound

1354


Co"7B12. When these mixtures were incubated at 370 Cbefore administration, the urinary excretion of the in-itially free Co'B12 was regularly increased, while ex-cretion of the originally bound Co'TB.2 was correspond-ingly decreased. Previous in vivo and in vitro isotopestudies concerned with vitamin B12 absorption and IFfunction have assumed that the specific activity of IF-bound vitamin B12 is constant. The occurrence of ex-change between free and IF-bound vitamin B12 maycomplicate the quantitative interpretation of such studies.

Control of Cardiac Output in Anemia. MARTIN DUKEAND WALTERH. ABELMANN,* Boston, Mass.

Fifteen anemic patients, 31 to 86 years old, werestudied supine, before and after treatment with 5.9 and10.9 g hemoglobin per 100 ml, respectively. Meas-urements (Evans blue dye) of cardiac output and totalblood volume (TBV) were made, injecting through apolyethylene catheter into axillary or caval vein andwithdrawing via brachial artery. After treatment, therewere significant changes in the mean differences (p <0.01) of the following: cardiac index (CI) decreasedfrom 4.73 to 3.44 L per minute per mi, heart rate (HR)decreased from 82.8 to 69.5 beats per minute, strokeindex (SI) fell from 57.6 to 49.5 ml per beat per mi,total systemic resistance (TSR) rose from 1,017 to1,526 dynes-sec-cm5', and TBV (13 cases) rose from4.55 to 5.20 L. Central blood volume (CBV) remainedunchanged, 1.69 to 1.62 L. Accordingly, CBV/TBVfell from 0.378 to 0.316 (p < 0.01). Four anemic patientswere studied in supine and sitting positions. Uniformly,decreases occurred in CI from 4.45 to 3.41 L per min-ute per m', SI from 52.9 to 37.9 ml per beat per m',TBV from 3.35 to 3.12 L, CBV from 1.21 to 0.93 L,and CBV/TBV from 0.360 to 0.299, HR rising in threecases. In eight normal subjects and six anemic patients,methoxamine was administered intravenously (0.3 to0.6 mg/min, total 8.5 to 20.7 mg). In normals, CIremained unchanged, HR fell, SI rose from 59.6 to76.0 ml per beat per m', CBV rose 11%o from 1.83 to2.03 L, and TSR increased from 1,046 to 1,286 dynes-sec-cm. In anemia, CI fell from 5.73 to 4.62 L perminute per m', HR fell, SI rose from 66.6 to 74.9ml per beat per m', CBV was unchanged, and TSRrose from 801 to 1,046 dynes-sec-cmn. The circulationin anemia is characterized by tachycardia, increasedstroke volume, decreased TSR, hypovolemia, and shiftof blood centrally. This response appears obligatorybut is partially reversible by changes of position or bythe vasopressor methoxamine.

Urinary Total Hydroxyproline as an Index of Con-nective Tissue Turnover in Bone. THEODOREA. DULL,LYNN CAUSING AND PHILIP H. HENNEMAN,* JerseyCity, N. J.

Hydroxyproline occurs in the body only in collagenand connective tissue; on a meat- and gelatin-free diet

urinary total hydroxyproline may reflect connective tissueturnover. Using the method of Prockop and Udenfriend,we found that the daily urinary total hydroxyprolinewas 27 to 96 mg in five prepubertal children and 8to 31 mg in five normal adults on a meat- and gelatin-free diet. Seven adults with diseases not known toinvolve connective tissue also excreted 9 to 27 mghydroxyproline daily. In three hypopituitary patients,aged 15, 21, and 22 years, urinary hydroxyproline was34, 45, and 42 mg, respectively. Eight days' treatmentwith human growth hormone in the 22 year old hypo-pituitary patient increased urinary hydroxyproline from42 to 72 mg. A patient in the fifth month of preg-nancy excreted 45 mg daily. A patient with severehyperthyroidism excreted 91 mg daily. In a patientwith severe osteitis urinary hydroxyproline decreasedfrom 100 to 61 mg in the 5 months following cureof hyperparathyroidism. Highest values (116 mg daily)were noted in a patient with extensive, active Paget'sdisease of bone. These data suggest that 24-hour urinarytotal hydroxyproline excretion during a meat- andgelatin-free diet may provide as sensitive an indexof bone collagen turnover as does calcium excretion ofbone mineral metabolism.

Osmotic Water Transfer Across a Vasopressin-sensitiveand an Artificial Membrane. LAURENCEE. EARLEY,Boston, Mass. (introduced by Laurence B. Ellis).It has been suggested that net water transfer along

osmotic gradients, exceeding predicted differences indiffusion across semipermeable membranes, results fromhydraulic (bulk) flow through porous openings in thebarrier. Vasopressin is thought to increase hydraulicflow along osmotic gradients by increasing pore sizein sensitive membranes. If so, such hydraulic flowshould produce some increase in water movement inthe direction of large net transfers. Total water-Hamovement from mucosa to serosa was measured, andtotal movement from serosa to mucosa calculated, fromthe isotopic measurement and simultaneous net transfer(weight change) across toad bladder pouches in vitro.With isotonic media, movement in each direction aver-aged 459 jul per cm2 per hour and net transfer wasnear zero. Osmolality of the mucosal or serosal medium,or both, was altered to produce gradients of 0, +165, or-165 mOsm per kg, and each result was comparedto its own control (isotonic media). Net water trans-fws did not correlate with increased water movementin the direction of the predicted hydraulic flow. Aninverse relationship existed between osmolality andwater movement from the altered medium. This rela-tionship accounted for net transfer, while water move-ment from the medium (of constant osmolality) changedlittle. A similar relationship accounted for net watertransfer along osmotic gradients (sucrose) across cellu-lose dialysis tubing. Without gradients, solute sup-pressed water-H2 movement in an order inversely relatedto permeation of the solute species (sucrose> glucose>urea); 2 molal sucrose depressed the movement of

1355


water-H' and urea-C' 75 and 70%, respectively. Thesedata do not support the concept of hydraulic flowthrough pores and suggest that large net transfers ofwater and unequal fluxes of permeant solute acrosssuch membranes may result from asymmetric diffu-sional permeability, produced by unequal concentrationsof impermeant solute on the two sides of the membrane.

A Kinetic Analysis of the Antidiuretic Action of Vaso-pressin. ISIDORE S. EDELMAN* AND MARTIN J.PETERSEN, San Francisco, Calif.

The discovery of the similarity in the effect of vaso-pressin on the skin and urinary bladder of amphibiaand the mammalian kidney made it possible to definethe kinetics of vasopressin action in an isolated system,uncomplicated by other physiological events. A simplechamber was designed for measuring microflows ofwater across a sheet of cells continuously and accurately,enabling measurement of the magnitude and the time-course of the antidiuretic response of isolated toadbladders to vasopressin. At fixed osmotic gradientsof 200 mOsmper kg H120, the minimal concentrationof vasopressin (Pitressin) necessary to produce an in-crease in the rate of water flow across the toad bladderwas 0.5 to 1 mUper ml. Dose-response curves revealedtypical saturation kinetics, the maximal response beingattained at concentrations> 10 mU per ml. A simplehypothesis served to describe these results. The assump-tions were that: 1) the rate of flow at a fixed osmoticgradient was linearly proportional to the number ofvasopressin molecules bound per cm' of cell surface,and 2) the binding of vasopressin to the cell membranewas a one-step reversible reaction. Based on theseassumptions, it was predicted and shown in 12 experi-ments that the action of vasopressin is characterizedby a linear dependence of the reciprocal of the rateof osmotic flow on the reciprocal of the vasopressinconcentration in the extracellular medium. The co-efficients in this equation provide independent measuresof empirical vasopressin-membrane binding constantsand of the membrane binding capacity for vasopressin.In addition, these studies confirmed that the action ofvasopressin is reversible and that exposure to highconcentrations of vasopressin induces resistance to theantidiuretic action of the hormone. The kinetic analysisprovides a method for evaluating the mechanism ofaction of vasopressin inhibitors and of intermediatesin antidiuretic pathways and a quantitative bioassay ofantidiuretic activity.

Selective Inhibition of Enterovirus Nucleic Acid Syn-thesis by 2- (a-Hydroxybenzyl) -Benzimidazole (HBB).HANS J. EGGERSAND IGOR TAMM,* New York, N. Y.

HBB is one of the few synthetic chemical compoundsthat are capable of inhibiting virus multiplication atnontoxic concentrations. Most human enteroviruses aresusceptible to inhibition by HBB, but members of other

groups of viruses are insusceptible. An investigationof the precise mechanism of action of HBB was under-taken, since it appeared that HBB might assist inidentifying a particular step in the reproduction of entero-viruses and provide an approach to the study of thechemical basis of virus specificity. Evidence has beenobtained that HBB has no effect on the early phases ofenterovirus-host cell interaction (i.e., adsorption andpenetration). However, HBB does interfere with oneof the later processes-synthesis of virus precursor ma-terials or assembly of virus particles-as shown by thefact that the compound inhibited virus multiplicationeven when added during the exponential growth of thevirus. A detailed study of the effects of HBB on thesynthesis of infective viral RNA has been carried outwith three viruses, ECHO 12, Coxsackie A9, andCoxsackie B4. HBB was used at a concentration of0.2 mM, which is sufficient to stop virus production.The compound prevented the synthesis of infectiveviral RNA, but had no effect on host cell RNA syn-thesis. It should be emphasized that HBB had nodirect irreversible inactivating effect on the infectivityof viral RNA, nor did it interfere with its uptake bycells. In HBB-treated cells, in which replication ofinfective viral RNA was inhibited, the synthesis ofcomplement-fixing antigen was also inhibited. Theresults suggest that HBB inhibits enterovirus multi-plication by inhibiting the synthesis of viral RNA.However, the possibility has not been excluded thatHBB may also have other effects on the viral repro-ductive process.

Quantitative Differences in Osteoporosis-producing Ac-tion of Corticoids. EUGENEEISENBERG AND GILBERTS. GORDAN,* San Francisco, Calif.

The adverse results (so-called "side effects") of cor-ticoid therapy for various disease are, in fact, due toexpected physiologic actions of these compounds in thedoses required. One of these effects is increased break-down of bone which results in osteoporosis and fracture.Previous chemical modifications of steroid structurehave failed to eliminate this undesirable property, al-though another unwanted action, sodium retention, hasbeen quantitatively reduced (prednisone), or even re-versed (triamcinolone). Measurement of skeletal dy-namics with stable strontium as a tracer showed thatthe osteoporosis of Cushing's disease or of corticoidtreatment is characterized in its earliest phase by in-creased urinary loss of the bone-seeking tracer withoutdecrease in bone deposition rate. After prolonged treat-ment or late in the disease, the rapidly miscible volumeof tracer distribution shrinks and the bone depositionrate falls. Thus, the corticoids are primarily "catabolic"and secondarily "antianabolic." Administration of largedoses of corticoids (cortisol 120 mg/day; prednisone20 to 30 mg/day; triamcinolone, 12 to 20 mg/day; ordexamethasone 3 mg/day, each administered to 6 ormore normal subjects) increased urinary excretion ofthe tracer. This effect was significantly reduced by

1356


halving the dose of dexamethasone. A difluorinatedcorticoid, 6-a-fluorotriamcinolone, with approximatelytwice the anti-inflammatory activity of triamcinoloneor prednisone, was similarly studied. In a dose of 12mg per day this compound did not increase the urinaryexcretion rate to the extent found with large doses ofthe other corticoids but only to the lesser degree ob-tained with the small dose of dexamethasone. Thesefindings signify that the difluorinated steroid is lesslikely to induce osteoporosis in usual maintenance dosesthan are the other anti-inflammatory steroids.

Postural Changes in Plasma Volume in Hypoalbumi-nemic Subjects. SEYMOUREISENBERG, Dallas, Tex.(introduced by Ben Friedman).

Transcapillary fluid movement is governed largely bythe net pressure generated by opposing forces of hydro-static and plasma oncotic pressure. This study wasundertaken to investigate postural alterations in plasmavolume (PV) under circumstances of a distorted Starlingequilibrium; i.e., in hypoalbuminemic individuals. PVwas calculated from the Cr'1 red cell mass and thehematocrit. Cirrhotic and nephrotic subjects werestudied during recumbency and after quiet standing.By comparing data obtained from venous sampling siteslocated at and below the atrial level and from arterialsamples evidence was obtained, indicating that depend-ency of the venous sampling site leads to local transu-dation and, hence, to spurious changes in PV. In 19normal subjects, standing resulted in a 279 ml (11%)decline in PV (arm dependent), 216 ml or 8% (veinat atrial level), and 183 ml or 7% (arterial). In 23hypoalbuminemic patients assumption of the erect pos-ture resulted in a 621 ml or 18% decrease in PV (armdependent), 523 ml or 16% (arm at atrial level), and606 ml or 16% (arterial) (p <0.001). Serum proteinconcentration consistently rose, indicating fluid lossrather than distributional shifts. The usual decline inPV associated with standing is exaggerated in hypo-albuminemic subjects. Losses of this magnitude un-doubtedly affect cardiac output and arterial blood volume.The arm position data suggest that hematocrit variationsoccur throughout the body in response to hydrostaticchanges; attention to position of the sampling site is es-sential to avoid errors induced by local fluid transudation.

Erythropoietin and Testosterone In Vitro. ALLAN J.ERSLEV,* Philadelphia, Pa.

Whole anemic serum has not been found to stimulatethe uptake or utilization of isotopically-labeled buildingblocks in bone marrow suspensions. The recent avail-ability of purified sheep-erythropoietin has made it pos-sible to determine with greater specificity whether ornot erythropoietin is active in vitro. Bone marrowsuspensions were prepared from surgically removedhuman ribs or rabbit bones and aliquot samples wereincubated for 21 and 45 hours at physiologic tempera-

ture, pH, PO2, and pCO,. Erythropoietin was addedand the cellular uptake of Fe' was determined. It wasfound that erythropoietin increases the uptake of Fe'if, 1) the cellular concentration is low; 2) the suspend-ing fluid contains appreciable amounts of tissue culturemedium. The optimal results were obtained at cellularconcentrations of 2,000 cells per mm' and a suspendingmedium of two-thirds Eagle's MEMand one-third nor-mal serum. Under these conditions erythropoietin in-duced a definite log-dose response with 2.5 cobalt unitsper ml, producing an increase of 20 to 30% after 21hours' incubation and 40 to 100% after 45 hours' incu-bation. Crystalline testosterone added to the samesystem also increased the cellular iron uptake but onlyduring the first 21 hours' incubation. These resultssuggest that testosterone may act by accelerating thematuration or division of cells present in the bone marrowsuspension whereas erythropoietin has a delayed effect,possibly through the creation of new pronormoblasts.In parallel studies erythropoietin increased the cellularuptake of C14-glycine and C14-formate by nearly 100%.Similar studies have been made on reticulocytes, onfrozen or preincubated bone marrow, and on bone mar-row cultures. These studies indicate that erythropoietindoes not increase the iron uptake of individual cells, butthey have not yet disclosed the exact site of action.

Studies on the Mechanism of Ethanol-induced Hypo-glycemia. JAMES B. FIELD* AND HIBBARD E. WIL-LIAMS, Bethesda, Md.

Hypoglycemia after alcohol ingestion has been at-tributed to nonethanol ingredients. Hypoglycemia hasbeen repeatedly produced in one such subject and allof five normal controls by the following procedure.After 42 hours of fasting, two control blood sugarsan hour apart were obtained and then 35 to 50 ml ofethanol was given orally. Hypoglycemia developedwithin 1 hour and persisted for at least 3 hours, at whichtime glucagon did not increase blood sugar. Fastingwithout ethanol caused no hypoglycemia. Equal amountsof ethanol after an overnight fast did not produce hypo-glycemia or impair glucagon-induced hyperglycemia.Plasma free fatty acids remained constant at an elevatedlevel for the last 5 hours of a control 48-hour fast butrose progressively during the hypoglycemia after ethanol,indicating that hypoglycemia was not due to increasedinsulin. During the ethanol-induced hypoglycemia theblood glucose rise after 50 g fructose (given i.v. 2 hoursafter ethanol) was comparable with that during a controlfast without alcohol, indicating unimpaired conversionof fructose to glucose. However, ethanol inhibited theconversion of fructose to glycogen as indicated by thesmall response to glucagon (given 1 hour after fructose)compared with the marked hyperglycemia after fructoseand glucagon during the control fast. Similar inhibitionof glycogen synthesis was demonstrated in vitro withliver slices from ethanol-treated rabbits. Perfusion ofisolated rat liver (in collaboration with Dr. Glenn E.Mortimore), using an alcohol concentration similar to

1357


the studies in man, demonstrated normal conversion offructose to glucose but marked inhibition of glycogensynthesis from fructose. These studies suggest thatethanol inhibition of glycogen synthesis may be primarilyresponsible for the observed hypoglycemia and thatglycogen formation during gluconeogenesis may be aprerequisite for hepatic glucose release.

An Approach to the Prevention of Rh Hemolytic Dis-ease. R. FINN AND J. R. KREVANS, Baltimore, Md.and C. A. CLARKE, Liverpool, England (introducedby C. B. Thomas).

We have reported evidence suggesting that the lowincidence of Rh hemolytic disease of the newborn is dueto infrequent occurrence of transplacental hemorrhageof sufficient magnitude to induce sensitization. Thedetection of such hemorrhages in mothers after delivery,by the Betke technique, which demonstrates fetal redcells in the maternal circulation, enables one to pickout a small "high risk" group in whom sensitizationis most likely to occur. The work of Levine, showingthat ABO-incompatible pregnancies interfered with Rhsensitization, raises the possibility that sensitizationcould be prevented if it were possible to remove Rhpositive cells from the circulation of the "high risk"mothers. To test this hypothesis Rh negative maleswere given 5 ml of Cr5`-labeled Rh positive red cellsand one-half hour later varying amounts of anti-D anti-body were injected. Approximately 10% of controls whoreceived only Rh positive cells developed anti-D. Inthe group who received 35 to 50 ml of anti-D therewas very rapid removal of Rh positive cells (half-life,3 hours) and no immunization occurred. In the groupreceiving 10 ml of anti-D, the Rh positive cell half-lifewas 24 hours, and there was a remarkable enhancementof immunization; 75% developed anti-D. These resultssuggest that immunization to the Rh antigen can beprevented if Rh positive cells are rapidly removed fromthe circulation. The mechanism by which immunizationis enhanced by small doses of antibody seems to berelated to accelerated removal at an intermediate rate.

The Relationship of Cations and ATP to MitochondrialSwelling and Shrinkage. SIDNEY S. FORTNEY ANDWILLIAM S. LYNN,* Durham, N. C.

Mitochondrial swelling and shrinkage can be studiedby 1) turbidity, 2) wet weight of pellet after centrif-ugation, and 3) urea-C14 space of pellet after centrifuga-tion. By all of these methods it can be shown thatswelling is associated with an increase in water contentof the mitochondrial pellet. Swelling is also associatedwith an increase in ATPase activity in a linear fashion.Mitochondrial swelling occurs spontaneously in isotonicbuffered media at 250 C. This spontaneous swellingproceeds more rapidly in ionic media; e.g., KCl, NaCl,LiCl, or CsCl than in sucrose of the same total osmoticstrength. The rate of swelling induced by these ionic

buffers has an inverse relationship to the hydrated radiusof the cation. That is, mitochondria swell more rapidlyin potassium and cesium than in lithium or sodium.ATP can increase the rate of swelling in sucrose at25°, but this is completely blocked by EDTA at 10' M.However, swelling induced by ionic buffer-e.g., Na+,K+ or Lit-is partially inhibited by ATP or EDTA, butcompletely inhibited when both are added together. Inmitochondria partially depleted of calcium by washingwith lithium, and then reincubated in a sucrose sodiumbuffer, we have found a response to ATP on the move-ment of cations. Under these conditions ATP liberatesthe more tightly bound calcium, and the efflux of lithiuminto the buffer is increased. Simultaneously, the influxof sodium from the buffer is inhibited by ATP as meas-ured both by flame photometry and Na'. Thus, shrink-age induced by ATP is accompanied by a loss of intra-mitochondrial cations and a decreased permeability toextramitochondrial sodium.

Primary Hyperoxaluria: A Defect in Glyoxylate Metabo-lism. ELIZABETH W. FREDERICK, MITCHELL T. RABKINAND LLOYD H. SMITH, JR.,* Boston, Mass.

Primary hyperoxaluria is a rare genetic disordercharacterized biochemically by urinary excretion of exces-sive oxalate. The clinical manifestations include nephro-lithiasis, nephrocalcinosis, and often a generalized deposi-tion of calcium oxalate, termed oxalosis. The diseaseusually progresses to early death in renal failure. Previousstudies have established that oxalate is not metabolizedin man. The presumed genetically determined metabolicdefect must therefore precede oxalate synthesis. Themetabolism of glyoxylate, the only compound demon-strated to be an immediate precursor of oxalate, hasnot been previously studied in man. Two siblings withjuvenile primary hyperoxaluria demonstrated a markedblock in the metabolism of intravenously administeredcarboxyl-labeled glyoxylate-C", excreting respiratoryCO-C14 at a rate only 20% of that found in control sub-jects. The rate of glyoxylate-C1" metabolism in theparents of the patients was approximately 65% of controlvalues, suggesting a heterozygous defect. The patientswith hyperoxaluria excreted 78 to 130 mg oxalate perday, 3- to 5-fold normal levels. Normal daily urinaryglycolate excretion, measured by an isotope dilutiontechnique, averaged 43 mg. Both patients with hyper-oxaluria also exhibited persistent glycolic aciduria, aver-aging 137 mg per day-an increase comparable with thatof oxalate. Preparations of leukocytes from patientsand control subj ects synthesized glycolate and oxalatefrom glyoxylate-C"4 at similar rates. The oxidation ofglyoxylate to oxalate produces a useless and potentiallyharmful compound. The development of an effectiveinhibitor of this enzymatic reaction would furnish alogical therapy for hyperoxaluria. In a survey of struc-tural analogs in vitro, sodium hydroxymethanesulfonatewas found to be a potent competitive inhibitor of oxalatesynthesis from glyoxylate. The potential effectivenessof this inhibitor in vivo is currently being explored.

1358


Studies on the Pathogenesis and Clinical Features of"Alcoholic Hypoglycemia." N. FREINKEL,* D. L.SINGER, C. K. SILBERT AND J. B. ANDERSON, Boston,Mass.

In the past 2 years, admission to the Boston CityHospital of seven subjects with hypoglycemic coma andhistory of recent alcoholic ingestion afforded the oppor-tunity to study this poorly defined entity first describedby Brown and Harvey in 1941. Analysis of presentingfeatures disclosed that coma: 1) occurred after pro-longed imbibition without adequate diet; 2) was notrestricted to chronic alcoholics; 3) was accompanied byhypothermia; and 4) was terminated within minutesafter glucose administration. Liver functions and bi-opsies were normal to minimally abnormal; tests ofthyroid, adrenal, and pituitary status (including metho-pyrapone) were normal. All subjects exhibited abnor-malities in one or more of the following challenges toblood-sugar maintenance: insulin (i.v.); tolbutamide(i.v.); or fasting (72 hours). During prolonged obser-vations on a metabolism ward, with excellent alimentationand total abstinence, the abnormalities persisted. Inthe three subjects in whom it was examined, i.v. or

p.o. alcohol reproduced hypoglycemia after an overnightfast; more profound hypoglycemia was elicited after 72hours of fasting. That pure ethanol, per se, constitutesa hypoglycemic agent was confirmed in normal subjects;however, 72 hours of prior fasting was necessary forsuch documentation. Manifest rise in peripheral glucoseutilization and plasma insulin (immunoassay) did notaccompany the hypoglycemia. The suggestion of directeffects upon hepatic gluconeogenesis has been confirmedby in vitro studies wherein ethanol reduced the produc-tion of glucose-C' from pyruvate-C1' and alanine-C1' byliver slices concomitant with an increased formation oflactic acid-C1' (presumably due to the augmented DPNHfrom alcohol oxidation). The conjoint clinical and invitro observations indicate that "alcoholic hypoglycemia"constitutes a distinct syndrome which is readily un-

masked in subjects who are inordinately vulnerable todiversion of 3-carbon fragments from gluconeogeneticpathways.

Transfusion of Granulocytes from Donors with ChronicMyelocytic Leukemia to Leukopenic Recipients. E. J.FREIREICH, E. E. MORSE, W. BRONSONAND P. P.CARBONE, Bethesda, Md. (introduced by James A.Shannon).

The technique of plasmapheresis has increased theavailability of fresh platelets and granulocytes for trans-fusion studies. While sufficient quantities of plateletscan be obtained from normal donors for control of hemor-rhage, the relatively small numbers of circulating granu-

locytes preclude use of normal donors for granulocytetransfusion. However, patients with chronic myelocyticleukemia (CML) and white blood cell counts of 100,000to 300,000 per mm' can donate adequate numbers ofgranulocytes in relatively small plasma volumes. Five

patients with CML donated 500 ml of granulocyte-richplasma as often as 5 days a week for 3 weeks. Themedian dose was 5 X 1010 granulocytes per transfusion(range, 2 X 1010 to 1 X 1011). Fifty such transfusions weregiven to 19 leukopenic recipients, 17 of whom had acuteleukemia. This resulted in a median 1-hour post-trans-fusion increment in circulating granulocytes of 1,000 permm' (range, 0 to 11,000). The median recovery in thecirculation was 5% (range, 0 to 20) of the dose given.Circulating granulocytes returned to pretransfusion levelsover the next 1 to 4 days (median, 2). With repeateddaily treatment, granulocyte levels could be maintainedabove 1,000 per mm'. Transfusions have been repeated asoften as 14 times for a single recipient. Thirty-two granu-locyte transfusions were given during febrile episodes. In14 of these, defervescence occurred within 12 hours. Tenpatients with fever had positive blood cultures for pseudo-monas organisms prior to transfusion. Nine had receivedcolymycin or polymixin for 2 or more days before trans-fusion. In six of these patients fever returned to normalwithin 12 hours of transfusion, and they remained afebrilefor more than 2 weeks. Transfusion of granulocytes fromdonors with CMLhas raised the circulating granulocytelevels of leukopenic recipients and has resulted in cure ofpseudomonas septicemia.

Direct Evidence for the Participation of EndogenousLipids in the Formation of Dietary Chylomicrons.SAMUEL J. FRIEDBERG, MORTOND. BOGDONOFF, E.HARVEYESTES, JR., PAUL B. SHER AND GITTA JACK-SON, Durham, N. C. (introduced by David T. Smith).

Previously reported data have provided conflicting evi-dence concerning whether or not the fatty acid compo-sition of dietary chylomicrons resembles that of the fedfat or assumes the pattern of the body fatty acids. Thepresent investigation was designed to resolve this prob-lem. Fasting male subjects were given a 100-g fat loadin the form of cream. Three hours later, at a time whenalimentary lipemia was well established, albumin-boundradiopalmitate was injected i.v. Serial blood sampleswere analyzed for triglyceride activity and total tri-glyceride over the following 8 hours. Chylomicrons fromseparate aliquots of serum were isolated by ultracen-trifugation. The supernatant chylomicrons obtained werewashed three times by resuspending in saline and recentri-fuged. The pooled infranatant solutions and washedchylomicrons were extracted and total radioactivity de-termined. The results indicate that radioactivity in totalserum triglyceride appears in 30 minutes and reaches apeak in 2 hours. Samples of serum obtained 2 and 3.5hours after injection of radiopalmitate contain 4 to 24%of the total triglyceride activity in the chylomicronfraction. In interpreting the results of these studies,passive exchange of the palmitic acid-l-C4 with serumtriglycerides could be ruled out, because radioactivity inthe triglyceride fraction was virtually absent during the10- to 15-minute period when radiopalmitate activity wasmaximal. It is concluded, therefore, that endogenouslyreleased free fatty acid (FFA), represented by the

1350


radiopalmitate, is in part incorporated as triglyceridesinto chylomicrons. This exchange could take place in theliver which takes up chylomicrons and converts FFA totriglyceride. Endogenously produced triglyceride mightalso exchange with alimentary chylomicrons in blood,intestinal lymphatics, or gastrointestinal tissues.

The Pco2 of the Right and Left Side of the Circulationduring CO2 Inhalations and Exercise. HERMANF.FROEB AND BYONGMOKKIM, La Jolla, Calif. (intro-duced by Edmund L. Keeney).

To investigate the concept of a chemoreceptor on theright side of the circulation for the control of breathing,the following experiments were performed during cardiaccatheterization in 11 patients. There were two withfunctional murmurs (C.H., P.S.) and the remaining hadsevere cardiopulmonary disease. While breathing througha pneumotachograph, nine patients inhaled 8% CO2 andtwo 5% CO2. Continuous recordings of respiratory peakflow and volume along with blood samples from pul-monary and brachial artery were obtained at 10-secondintervals for 2 minutes. Three of these patients werefollowed through recovery. Similar experiments wereperformed on six patients who exercised while breathingroom air; two of three were followed through recovery.The results reveal that while the patients breathed 8%oC02, there were simultaneous rises in pulmonary andbrachial arterial Pco2 with rising ventilation. In SubjectC.H., who breathed 5% C02, there was no rise inbrachial arterial Pco2 with rising ventilation, but a sig-nificant correlation with increasing pulmonary arterialPCO2. During recovery from 8% CO2 in Subject P.S.,decreasing ventilation followed the slow decline of thepulmonary arterial Pco2 over 80 seconds while brachialarterial Pco2 fell precipitously in 30 seconds. In patientswith cardiopulmonary disease, the correlation of pul-monary arterial Pco2 and ventilatory changes was notclear-cut. In five subjects, during exercise, there was alinear correlation with increasing pulmonary arterialPco2 to rising ventilation while brachial arterial Pco2 fell.One subj ect with tetralogy and Blalock operation wasthe only case in whom ventilation followed the brachialarterial Pco2. These results support the hypothesis of achemoreceptor on the right side of the circulation.

Properties of Univalent 3.55 Fragments of 7S Aggluti-nating Antibodies. HUGHFUDENBERGAND A. NISON-OFF, San Francisco, Calif. and Urbana, Ill. (introducedby Ernest Jawetz).

Purified 7S rabbit antibody to human erythrocytes andto human 7S y-globulin was converted to 5S and 3.5Sfragments with pepsin and papain, respectively. The 5Sand 3.5S fragments have been demonstrated previously tobe bivalent and univalent, respectively. The 7S aggluti-nating antibody against human erythrocytes possessedapparent anti-H specificity. Its 5S derivative retainedfull agglutinating activity, but the 3.5S fragments lackedagglutinating activity. However, combination of the 3.5S

fragments with the erythrocyte antigens was demonstrableby agglutination upon the addition of chicken or goatantisera prepared against the fragments of rabbit y-globu-lin. The 3.5S univalent subunits of anti-H failed toagglutinate appropriate red cells by albumin, papain, ficin,or trypsin techniques, in contrast to the behavior of so-called "univalent" 7S incomplete antibodies such asanti-Rh,. Similar results were obtained with the rabbitantihuman y-globulin antibodies, using Rh, cells "coated"with incomplete anti-Rh0 as indicators. The 5S deriva-tives of the antiglobulin serum possessed full agglutinatingactivity for these coated cells, whereas the 3.5S fragmentsfailed to agglutinate the coated cells. Combination of the3.5S fragments with the antigen (human y-globulin) wasagain demonstrable by use of goat or chicken antibodiesspecific for the 3.5S fragments. Treatment of the 5Smaterials with sulfhydryl reagents resulted in productionof 3.5S fragments with serologic properties identical withthose of the 3.5S materials resulting from papain treat-ment of the intact 7S antibody. These results indicatethat the carbohydrate-rich "piece III" of rabbit antibodyplays no role in agglutination reactions and that theunivalent fragments possess sensitizing but not aggluti-nating activity.

Blood Angiotensin Levels in Normal Subjects andHypertensive Patients. J. GENEST, P. BIRON, M.CHRETIEN, R. BOUCHERAND E. Koiw, Montreal, Can-ada (introduced by J. S. L. Browne).

Using the method of Boucher, Biron and Genest basedpartly on the procedure of Palladini and co-workers,105 determinations of angiotensin blood level were done,mostly on brachial arterial blood, from 84 normal sub-jects and hypertensive patients. The average concentra-tion in 33 normal subjects (35 determinations) was 10m,ug per 100 ml of blood (range, 0 to 55). Thirty-nineangiotensin determinations in 28 patients with essentialhypertension showed an average concentration of 76mug per 100 ml (range, 0 to 820); 25 of these 39 de-terminations were within normal range. One patientwith essential hypertension and without renal arterialdisease had a concentration of 90 mgg per 100 ml in hisrenal venous blood. Two patients with hypertensionassociated with bilateral parenchymatous renal diseasehad angiotensin arterial levels of 0 and 100 m/ug per 100ml. Seven patients with hypertension associated withunilateral renal arterial stenosis had arterial levels of265, 80, 110, 10, 35, 26, and 25; and renal venous levelsof 100, 0, 130, 30, 0, 280, and 20 mlug per 100 ml. Eightdeterminations made in 7 patients with malignant hyper-tension gave an average concentration of 18 m~ug per100 ml (range, 0 to 70). Seven determinations werebelow 35 m/ug per 100 ml. Blood angiotensin levelswere normal in 5 patients with proven primary aldoste-ronism. No correlation could be found between theangiotensin blood level and the volume of the bloodsample, the diastolic blood pressure level at the time ofthe sampling or the simultaneously determined urinaryaldosterone. Although angiotensin may appear to play

1360


a role in the genesis of hypertension in a few patients,these studies strongly suggest that in the majority ofhypertensive patients, whether of the essential, renal(parenchymatous) or malignant type, angiotensin is notinvolved in the maintenance of hypertension.

Acute Hemoglobinuric Renal Failure without Ischemia.MARTIN GOLDBERG, Philadelphia, Pa. (introduced byJ. Russell Elkinton).

It has been difficult to assess the specific effects ofhemolyzed red blood cells (RBC) on renal hemodynamicsbased on previous work because: 1) production of experi-mental acute renal failure due to hemolyzed cells, hemo-globin, or methemoglobin was associated with severedehydration or hypotension; 2) studies of renal bloodflow under these conditions involved clearance or directFick-para-aminohippurate (PAH) methods difficult to in-terpret at low urine flows (V). To circumvent thesedifficulties, moderate quantities of homologous hemolyzedRBC (02 to 0.3 g Hb/kg) were infused into the aortaabove the renal arteries in 12 dogs. Renal blood flow(RBF) was estimated by an indirect Fick method usingkrypton8" not dependent on urine collection. The infusionof hemolysate resulted in acute reductions in V (50 to95%), creatinine clearance (Cc,; 6 to 99%), andradioiodinated hippuran clearance (CHIp,; 16 to 90%).The fractional changes in Cc, and V were relatedto V1ontrol such that the lower the Vcontrol the greaterthe decrease in both Ccr (r = 0.88, p < 0.001) andV (r = 0.73, p < 0.02). Despite these changes, RBFand arteriorenal 02 did not change after the infusion(control RBF= 296 ml/100 g/min, SD ± 14.3; postin-fusion RBF= 329 ml/100 g/min, SD ± 29.6). In fiveanimals, mean changes in renal extraction ratios (E) ofhippuran and creatinine were < 10%o despite significantdecreases in V, CCr, and CHIp; in three of these the meanchange in EPAHI was 0. In one experiment a 50% re-duction in EHl, and Ecr occurred. In three animalsdecreases in U/P Cr and U/P osmolality were observed,but in the remainder, increases in these ratios followedthe infusion. Histologically, tubular damage in varyingdegree was present in each animal. The data suggestthat blood pigment excretion may be associated withacute decreases in urine flow and clearances not relatedto renal ischemia. Intratubular obstruction and, in someinstances, concomitant toxic tubular damage can explainthese observations.

Mechanisms of Renal Acidosis. H. C. GONICK, M. E.RUBINI AND M. H. MAXWELL, Los Angeles, Calif. (in-troduced by M. I. Grossman).

The inability of patients with chronic renal disease toexcrete hydrogen ion may be due to: 1) diminished H+production by the kidney, 2) a limitation in availableproton carriers in the urine (ammonia or phosphatebuffer), or 3) deficient ionic exchange with fixed cation(Na+t =±H+). Acid excretion was studied as follows in

a group of patients with various types of renal disease:1) total acid excretion was measured after a standardizedoral ammonium chloride load, 2) the ability to increasetitratable acid excretion was measured during isotonicneutral phosphate infusions, and 3) the ability to ex-change Na+ for H+ was measured by infusing sodiumsulfate under conditions of maximal sodium-retainingstimuli. An approximately equal reduction in excretion oftitratable acid and ammonium was demonstrated duringammonium chloride loading. No limitation was seen intitratable acid excretion when neutral phosphate wasinfused, and the slope of titratable acid vs phosphate wassimilar to that of normal subjects. Although blood pHwas unchanged during sodium sulfate infusion, the dropin urinary pH was similar to that evoked by ammoniumchloride and accounted for the increase in titratable acidexcretion. However, ammonium excretion relative to pHwas invariably less than normal. Despite reduced H+excretion in exchange for Na+, some patients showed anormal Na+-K+ exchange, implying intact distal sodiumreabsorption and exchange. The data indicate that:1) there is no intrinsic limitation in H+ production whensufficient buffer is provided, 2) the distal ion exchangemechanisms may be normal but ammonia productionfaulty, and 3) acidosis in renal disease primarily re-flects a decreased ability to provide urinary protoncarriers. Parallel studies in renal tubular acidosis (con-genital and acquired) indicate that this disease is excep-tional in that a true defect in hydrogen ion productionmay exist.

Role of the Thymus in Development of Immunity.ROBERTA. GOOD,* CARLOSMARTINEZ, OLGAK. ARCHERAND BEN W. PAPERMASTER,Minneapolis, Minn.

Although the thymus has been subjected to extensivestudy, its role in the body economy has not been eluci-dated. The observation that an abnormality of thethymus (benign thymoma) occurs frequently in patientswith acquired agammaglobulinemia suggested a role forthis organ in immunologic reaction. Removal of thethymus in adult rabbits did not alter immunologic reac-tivity. Glick and co-workers showed that removal ofthe bursa of Fabricius at an early age interfered with theimmunologic development of chickens, and Mueller andco-workers showed more profound depression of im-munologic potential in chickens lacking the bursa athatching. We postulated that thymectomy in neonatalanimals might interfere with the normal development ofimmunologic competence. Studies in our laboratory,using rabbits, rats, and inbred mice, have shown that thethymus is indeed essential to normal immunologic de-velopment. Animals thymectomized in the immediateneonatal period have a morphologically defective lymphoidsystem, deficient homograft and tumor immunity, andfeeble capacity for formation of precipitating and phage-neutralizing antibody. It is postulated that the thymusis involved in the centrifugal distribution of immunologi-cally competent cells which are ultimately involved in theseveral forms of immunologic reactivity.

1361


Generation of Reduced Pyridine Nucleotides in the Liversof Normal and Alloxan-diabetic Rats. EDWIN E. GOR-DON, New York, N. Y. (introduced by Irving M.London).

A major defect in the metabolism of the diabeticorganism is the markedly diminished ability of the liverto synthesize fatty acids. In the present experiments anattempt was made to assess directly the hypothesis thatdiminished generation of reduced pyridine nucleotidesmight be a significant rate-limiting factor for fatty acidsynthesis in the liver of the diabetic rat. Nicotinamide-7-C"4, introduced intraperitoneally, labels the DPN andTPN of liver. The reduced forms of the pyridine nucleo-tides are generated from the corresponding oxidizedforms in the course of pyridine nucleotide-linked enzy-matic reactions. The specific radioactivity of each of thefour forms as a function of time after administrationof nicotinamide-C14 provides data relating to the ratesof equilibration of the pyridine nucleotides. The spe-cific activity of hepatic DPN increases progressively,while that of TPN does not change appreciably be-tween 2.5 and 20 minutes after administration of nico-tinamide-C14. The ratio of the specific activity ofDPNHto DPN reaches 0.5 within 2.5 to 5 minutes andapproaches unity at approximately 15 minutes in thelivers of both normal and diabetic rats. In normal liver,equilibration of TPNH with TPN does not occur untilafter 20 minutes, and the ratio reaches 0.5 between 10and 15 minutes. The rate of equilibration of the tri-phosphopyridine nucleotides in the diabetic liver is thesame as, or more rapid than, that found in the normal.Since the corresponding pool s.zns of each of the fourpyridine nucleotides are essentially equivalent in normaland diabetic rat livers, the equilibration data are a func-tion of the rate of generation of reduced pyridinenucleotides. These data therefore, suggest that the rateof generation of DPNHand TPNH is independent ofthe capacity of alloxan-diabetic rat liver to synthesizefatty acids. The generation of reduced pyridine nucleo-tides in the major intracellular compartments of normaland diabetic liver is currently under study.

Dynamic Dilution Studies of Substances Passing throughthe Portal Circulation in the Liver. CARL A. GORESKY,Montreal, Canada (introduced by Douglas G. Cameron).

Labeled red cells, serum albumin, sucrose, inulin,sodium, and water were given by a single rapid injectioninto the portal vein in the dog. Their subsequent con-centrations in the hepatic venous blood were sampledserially. None of these substances was appreciablyretained in the liver on one circulation, but the outflowpattern of each of the substances was modified relativeto the red cell pattern, showing a diminution in peakconcentration and delayed transit time. A substance,such as water, that might be expected to enter the totalorgan water, showed a greater peak depression and pro-longation of transit than did substances like sucrose thatone might expect were excluded from the intracellular

water. Theoretical considerations of flow through liversinusoids of substances that exchange into an extravascu-lar space led to a quantitative mathematical treatment interms of two parameters, the extravascular volume ofdistribution and the effective sinusoidal blood volume(i.e., the volume in the vessels across which exchangeoccurs). These parameters are calculated from a groupof indicator dilution curves. The two calculated param-eters accurately describe the changes in each curveinduced by extravascular exchange. The volume calcu-lated for water agrees with the steady-state value ob-tained from the extirpated organ. There is a rapidextravascular exchange of plasma albumin into a spaceabout equal to the intravascular albumin space. Thisfact probably accounts for the apparent low "wholeorgan" hematocrit reported for this organ. A quantita-tive method has been derived to describe the passage ofdiffusible indicator substances through the hepatic sinus-oids in terms of two previously inaccessible parameters,1) sinusoidal blood volume, and 2) the volume of distri-bution within the organ. This method may facilitate thecalculation of effective capillary blood volume and of thevolume of distribution of diffusible substances in otherorgans as well.

A Re-evaluation of the Mechanism of Endotoxin Toler-ance. SHELDON E. GREISMAN* AND FRANK A.CAROZZA, JR., Baltimore, Md.

After reticuloendothelial system (RES) blockade, ani-mals tolerant to bacterial endotoxin exhibit febrile reac-tions to endotoxin that are comparable with normalanimals. These observations initiated the hypothesis thatendotoxin tolerance is mediated primarily by the RES.It should be emphasized, however, that tolerant animals,deprived of a normally functioning RES, were beingcompared with normal intact animals. The fact that bothgroups react comparably to endotoxin actually suggeststhat tolerant animals with blockaded RES possess amechanism capable of inhibiting endotoxin in vivo aboutas effectively as does the intact RES of normal animals.To verify this concept, rabbits were rendered tolerant to2.0 ,ug Escherichia coli endotoxin. These animals and nor-mal controls were then given Thorotrast (3 ml/kg i.v.)and 15 hours later 2.0 ,ug E. coli endotoxin. The controlRES-blockaded animals exhibited significantly and mark-edly greater fever indices than did the tolerant blockadedgroup. To exclude the possibility that these reactionswere attributable to resistance of tolerant animals to RESblockade, normal animals were given Thorotrast intra-venously (3 ml/kg); 3 hours later 10 ml per kg ofpyrogen-free, fresh, heparinized plasma was administeredintravenously from donor animals tolerant to 2.0 ug perkg E. coli endotoxin. Control blockaded animals re-ceived comparable quantities of fresh normal plasma.Sixty minutes after plasma transfer, all animals receivedSalmonella typhosa endotoxin intravenously (0.05 ,ug/kg). RES-blockaded animals given tolerant plasmaexhibited tolerant-shaped fever curves and significantlylower fever indices than did blockaded animals given

1362


control plasma. These studies indicate that when propercontrols are employed, tolerance to endotoxin can beassociated with the appearance of a potent, nonimmuno-logically specific, humoral inhibitor of endotoxin that actseffectively in vivo despite RES blockade. Such findingssuggest that humoral inhibitors play a more importantrole in mediating the tolerant state than has been previ-ously suspected.

Isolation of d-Tubocurarine-binding Protein from Skele-tal Muscle. DAVID GROB,* TATSUJI NAMBA, NATHANA. SOLOMONANDDANIEL S. FELDMAN, Brooklyn, N. Y.

In an effort to identify the receptor in skeletal musclethat combines with the neuromuscular transmitter, acetyl-choline, studies have been carried out on the combinationof muscle constituents with d-tubocurarine, which blockstransmission by competing with acetylcholine for the re-ceptor. d-Tubocurarine-binding substance was identifiedin normal human muscle by C'4-labeled dimethyl-d tubo-curarine and concentrated by repeated precipitation withd-tubocurarine. The binding substance is a ribonucelo-protein, of which 60% is RNAand 40% protein. It gavecharacteristic ultraviolet and infrared absorption spectra.The amino acid composition of the protein was de-termined and the presence of quanine, adenine, cystosine,uracil, and pentose was proven. The binding by theprotein of d-tubocurarine or dimethyl-d-tubocurarine wasinhibited by compounds that antagonize the neuromuscu-lar blocking action of these agents, including acetylcholine,choline, succinylcholine, and decamethonium. The ad-dition of acetylcholine to the protein produced a promptchange in the infrared absorption spectrum in aqueousmedium, and a prompt increase in the proportion ofhydrogen atoms which were rapidly exchangeable withisotopic hydrogen. These changes indicate that acteyl-choline may alter the structure of the protein by ruptur-ing hydrogen bonds, increasing water binding, or both.d-Tubocurarine did not have this effect. The reactionsof d-tubocurarine and acetylcholine with d-tubocurarine-binding muscle protein were not specific for this proteinbut also occurred with casein and thrombin, although notwith albumin, globulins, hemoglobin, gelatin, or elastin.d-Tubocurarine also reacted with RNA, DNA, chondroi-tin sulfuric acid, and polyserine. Preliminary studieshave indicated some differences between the d-tubocu-rarine-binding protein isolated from the skeletal muscleof normal subjects and patients with myasthenia gravis,cardiac muscle, and muscle of other species.

Stimulation of Radiosulfate Incorporation into GuineaPig Harderian Gland by Thyroid-stimulating Hormone(TSH) and Exophthalmic Serum. HESKEL M. HAD-DAD, St. Louis, Mo. (introduced by David M. Kipnis).

The 24-hour uptake of SU04 by guinea pig orbital tis-sues and the Harderian gland in particular can be stimu-lated by TSH (1.0 U/day X 2). TSH stimulation ofHarderian gland sulfation can be made more sensitive byinjecting S'O4 intraperitoneally 72 hours before killing

the animals, during which time TSH is administered inthree daily doses. An increase of 24 to 56% inHarderian gland sulfation was obtained with 0.1 U ofTSH per day X 3. The effect of TSH was not repro-duced by large doses of growth hormone, prolactin,ACTH, or thyroid hormones. The sera of seven patientswith progressive endocrine exophthalmos and four withinactive exophthalmos were screened for TSH sulfation-like activity, injecting 0.1 to 0.2 ml of serum intra-peritoneally daily for 3 days. The sera of sevenpatients with progressive exophthalmos produced an 18to 44% increase in Harderian gland radiosulfate incor-poration. Of the four patients with inactive exophthal-mos, two sera caused no change and two induced an 11%orise in Harderian gland S3O4 uptake. The uptake ofS3O4 by the thyroid gland was persistently stimulated byTSH (average increase, 34%o). Sera of nine patients,who were currently euthyroid (four with inactive and fivewith progressive exophthalmos) caused no significantchange in thyroid 3504 uptake. The sera of twothyrotoxic patients with progressive exophthalmos stimu-lated thyroid S"O0 incorporation by 27 and 42%o. Thesefindings suggest that a sulfating factor exists in exoph-thalmic sera which stimulates the Harderian gland butnot the thyroid tissue of the guinea pig.

Unusual Properties of Synovial Fluid Hyaluronate inRheumatoid Arthritis. DAVID HAMERMAN*AND JOHNSANDSON, New York, N. Y.

Hyaluronate was isolated from synovial fluids obtainedfrom normal joints and from inflamed joints of patientswith rheumatoid arthritis. The fluids were passedthrough a column of hydroxylapatite and IRC-50 resin,and the hyaluronate isolated by ultrafiltration on Milli-pore filters (0.1 A pore diameter). Normal productscontained hyaluronate and about 2%o protein (NHP)with a high content of alanine, glycine, serine, andglutamic acid. Rheumatoid hyaluronate (RHP) con-sistently contained more protein (5 to 7%o) with a highercontent of basic amino acids. To show that protein andhyaluronate were firmly bound, the products were labeledwith iodine"~1 and subj ected to zone electrophoresis.Hyaluronate was located by analysis for hexuronic acid,protein by radioactivity. Both NHPand RHPmigratedtoward the anode as a single component over a pH rangeof 6.4 to 10.5. At pH 4.5 (0.075 M acetate) a strikingdifference was noted: NHP moved toward the anode,RHP remained at the origin. The immobility of RHPwas not reversed by 6 M urea. Upon deploymerizationof RHPby cysteine or ascorbic acid, RHPmigrated asa single component toward the anode at pH 4.5. Afterdigestion of the protein of RHPwith trypsin, hyaluronatemigrated toward the anode at pH 4.5. Some samples ofRHP (1 mg/ml became "gel-like" after dialysis in acetatebuffer (pH 4.5) and would not flow in Ostwald vis-cometers (25° C). Dialysis at pH 7.2 restored thesesamples to a free-flowing state. The gel-like state at pH4.5, which could account for electrophoretic immobility,appeared to be due to aggregation of RHP, since these

1363


samples were retained on 5 a Millipore filters that per-mitted rapid filtration of NHP. Neither the electro-phoretic immobility nor gel-like state of RHPat pH 4.5could be demonstrated with NHPeven when NHPwas:1) isolated from normal synovial fluid with added rheu-matoid serum, or 2) mixed with the basic proteinlysozyme. The unusual properties of RHP at low pHsuggest that its structure differs from that of NHP.

Clinical Syndromes with Elevated Blood Histamine andSerotonin. BERNARDJ. HAVERBACK, EUGENET. BALD-RIDGE AND BARBARADYCE, Los Angeles, Calif. (intro-duced by Clayton G. Loosli).

Specific, highly sensitive spectrophotofluorometric meth-ods for histamine and serotonin analyses permit precisestudies of these amines. The purposes of this study are1) to present three diverse clinical syndromes in whichwhole blood histamine and platelet serotonin levels weresignificantly elevated and 2) to determine interrelation-ships of histamine and serotonin in the brain, mast cell,and gastrointestinal tract of the rat. The means of manyblood histamine levels in eight patients with the malig-nant carcinoid syndrome were 0.12, 0.13, 0.18, 0.04, 0.06,0.09, and 0.27 ,Ag per ml (normal, 0.02 to 0.08). All hadelevated blood serotonin levels. The histamine levels incarcinoid tumor from four of these subjects were notelevated (0.4, 2.2, 0.2, and 12.5 iug/g). Two subjects withtransient flushes, abdominal pain, nausea, diarrhea, and noobjective evidence of carcinoid tumor had mean values ofplatelet serotonin of 1.5 and 0.7 ,ug per mg of platelet pro-tein (normal 0.2 to 0.6) and of blood histamine of 0.15and 0.21 ltg per ml. One subject who attempted suicideby ingestion of a large amount of a potent monoamineoxidase inhibitor (trans-DL-2-phenylcyclopropylamine)which produced coma, extensor rigidity, tremor, hyper-thermia, and a striking facial flush had a mean bloodserotonin level of 1.75 ,ug per mg and a mean bloodhistamine level of 0.10 gg per ml. Because of these find-ings, interrelationships of tissue histamine and serotoninwere studied. In the rat it was shown that 1) serotoninor 5-hydroxytryptophan produced no change in tissuehistamine; 2) reserpine depleted bound tissue histamineas well as bound tissue serotonin; 3) a-methyl histidinelowered gastric mucosal histamine but did not decreaseserotonin. It is clear that certain substances affect bothserotonin and histamine, and that in the clinical con-ditions described herein, histamine as well as serotoninlevels in the blood were elevated.

Reduction of Serum Cholesterol in Man by an OralAndrosterone Preparation. LEON HELLMAN,* BARNETTZUMOFF, GERALDKESSLER, R. S. ROSENFELDAND T. F.GALLAGHER, New York, N. Y.

The demonstration of the hypocholesterolemic prop-erties of androsterone raised the possibility that thisnormal steroid metabolite might have a role in thepathogenesis and treatment of atherosclerosis. Althoughthe hypocholesterolemic effect has been amply confirmed,

evaluation of the clinical utility of long-term androsteronetherapy has lagged, since it is ineffective when givenorally and parenteral administration may be accompaniedby local discomfort and occasional fever. It has beenfound that androsterone, when administered orally to-gether with ethyl p-chlorophenoxyisobutyrate (ImperialChemical Industries), produces a hypocholesterolemic re-sponse indistinguishable from that of the parenteralsteroid, but free of the untoward effects: 1) 20 to 30 mgandrosterone together with 1 to 1.5 g p-chlorophenoxy-isobutyrate was given daily by mouth for periods up to4 months to 13 hypercholesterolemic patients. Ten pa-tients showed a 22% fall (p < 0.01) in serum cholesterolfrom an initial mean value of 397 mg per 100 ml. Freeand ester cholesterol fell proportionately. 2) There wasa concomitant 28% fall (p < 0.01) in serum alkalinephosphatase. There was no change in the usual liverfunction tests. 3) Desmosterol could not be identified byvapor-phase chromatography in the sera of three sub-jects who had significant decreases in cholesterol levels.4) The androsterone was absorbed and then excreted inthe urine as a conjugated 17-ketosteroid. 5) No viriliza-tion, feminization, or changes in libido were observed.Since this combination is well tolerated and no toxicityhas been observed thus far, the exploration of the long-term effects of hypocholesterolemic therapy with andros-terone is now feasible.

On the Presence and Character of the Increased Abilityof the Addisonian Patient to Taste Salt. R. I. HEN-KIN, J. R. GILL, JR., F. C. BARTTER* AND D. H. SOLO-MON, Bethesda, Md. and Los Angeles, Calif.

Adrenalectomized rats, unlike normal rats, can detectNaCl in very dilute solutions. In the present study thetaste threshold to NaCl has been determined in patientswith and without adrenal insufficiency. The subject wasrequired to choose the salt solution from a series of threefluids, two of which were distilled water. Eleven dif-ferent concentrations of NaCl were used, with a variedorder of presentation. Threshold was measured as thelowest concentration of NaCl that could consistently bedistinguished from distilled water. In 16 normal subjectsand 75 patients without evidence of disease of the adrenalglands, the median detection threshold for NaCl was12 mM (range, 6 to 60 mM). In 12 patients withadrenal insufficiency the median detection threshold wasmarkedly decreased to 0.4 mM (range 0.1 to 2.5 mM).Treatment with deoxycorticosterone (20 mg/day) alonedid not alter the taste threshold despite a gain in weightand return of serum sodium concentration to normal.Treatment with carbohydrate-active steroids (predniso-lone) raised the median detection threshold to the priornormal level of 12 mMwithin 18 to 48 hours; thisoccurred without change in body weight or in serumsodium. An increase in the sensitivity of salt-tasting toabnormal levels returned within 4 days, after treatmentwith carbohydrate-active steroids was discontinued. Anincrease in ability to taste NaCl in minimal concentra-tions thus appears to be a characteristic sign of adrenal

1364


cortical insufficiency. Ability to return salt-tasting sensi-tivity to lower, normal levels appears to be a specificproperty of carbohydrate-active steroids.

A "Messenger"-like RNA in Mammalian Cells. How-ARD H. HIATr,* Boston, Mass.

Transfer of information from the genes to the protein-synthesizing machinery of bacteria is now widely con-sidered to be mediated by a fraction of ribonucleic acidtermed "messenger" RNA. Its metabolic instability, thesimilarity of its purine and pyrimidine base ratios tothose of the corresponding bacterial deoxyribonucleicacid, the heterogeneity of its molecular size, and itscapacity to attach to the protein-forming particles withinthe cell are among its characteristics in accord with themessenger concept. To ascertain the applicability of thishypothesis to mammalian cells we are investigating invivo and in vitro RNAand protein synthesis in normaland regenerating rat liver and in human blood cells.From rat liver nuclei we have isolated a labile fractionof RNAwhich has a 20-fold higher specific activity thanhas the bulk of the nuclear RNA 10 to 40 minutes afterin vivo administration of orotic acid-C14 or P', butwhich, in contrast to the remaining nuclear and cyto-plasmic RNA, is no longer labeled 18 hours after injec-tion of isotope. We have shown by a sucrose gradientcentrifugation technique that the rapidly labeled materialhas the heterogeneity of size that would be required ofa fraction coding for the diverse cellular proteins. Thebase ratios of this fraction are much closer to those ofrat liver DNA than are those of the cytoplasmic RNAor the bulk of the nuclear RNA. When exposed toliver microsomes, the particles known to synthesize pro-tein, over 75% of this fraction becomes attached. Pre-liminary studies. indicate that nuclear RNA containingthe labile material stimulates the incorporation of aminoacids into protein in a cell-free protein-synthesizing sys-tem prepared from liver, but not in one prepared frombacteria. These findings are consistent with the presencein mammalian cells of an RNA analogous to bacterialmessenger.

Diminished Biological Half-life of Angiotensin II inHypertensive Plasma. ROGERB. HICKLER, DAVID P.LAULER AND GEORGEW. THORN,* Boston, Mass.

Interest in the metabolism of angiotensin in hyper-tension has resulted in an application of the rat bio-assay for the determination of the biological half-lifeof angiotensin II in human plasma. An in vitro admix-ture of angiotensin II and plasma is studied for theprogressive loss of pressor activity by serial injection ina 250-g vagotomized rat, prepared with pentobarbitaland pentolinium. The left jugular vein is intubated witha microsyringe containing a standard (1 ltg/ml) solutionof angiotensin II; the right jugular with a micro-syringe containing a mixture of 1 ml of the standardangiotensin solution and 1 ml of plasma. A stop watchis started at the time of the admixture. Room tempera-

ture is constant at 25 + 10 C. The arterial pressureresponse to the injection of the standard vs standardplus plasma solutions is alternately recorded every 3 or4 minutes. The dose-response curve for standard angio-tensin is linear in the range employed. Thus, theprogressive reduction in the pressure response for theplasma plus standard mixture (after correction for anyvariation in the standard response at each point duringthe assay) is equivalent to the rate of inactivation ofangiotensin by the plasma. The regression curve indi-cates a first-order reaction from which the biologicalhalf-life is determined. Angiotensin in the plasma of11 normotensive subjects so studied resulted in a meanbiological half-life of 30.2 minutes (SE ± 1.9 min) ascompared with a mean of 18.1 (± 1.5 min) in theplasma of 12 patients with primary hypertension(p < 0.001). The mean for 8 patients with hypertensionassociated with chronic renal disease was 24.6 (±t 1.7min) and was significantly greater than that for thegroup with primary hypertension (p = 0.01). The diag-nostic implications and the possibility that these changesmay represent an enzymatic adaptation to chronic angio-tensinemia are being investigated. A similar trendreported by Wolf and co-workers, using a radiochemicaltechnique, is acknowledged.

Initiation of the Action of Complement in a HumanAutoimmune Hemolytic System, the Donath-Land-steiner Reaction. CARL F. HINZ, JR.* AND ANNAMARIE MOLLNER, Cleveland, Ohio.Serum complement mediates the cytolytic reaction in

immune hemolysis including the autoimmune Donath-Landsteiner (D-L) reaction of paraoxysmal cold hemo-globinuria. The biphasic nature of the D-L reactionand the requirement for complement in both its initialcold and final warm phases, make it adaptable as amodel system entirely of human origin for studyingseparately the sequential phases of complement action.Previous studies from this laboratory indicated thatC'1, the first component of complement, or C'1-esterase,the enzymatically active form derived from partiallypurified C'1, is required in the early phase of the D-Lreaction, consistent with current concepts of the actionof complement. However, Mfiller-Eberhard has re-cently demonstrated that a normal serum protein withcharacteristics of an 11S 'y-globulin is required veryearly in the lysis of sheep cells. The current observa-tions indicate that this 11S factor initiates the D-Lreaction and acts before CT. The evidence indicatesthat the partially purified 11S factor and serum lacking11S are individually inactive in completing the initialphase, but combination of the two restores activity.Sequential exposure of erythrocytes and antibody tothe IlS factor and then to serum lacking 1 S also com-pletes the initial phase, but the reverse order does not,indicating that 11 S interacts before other componentsof complement. Subsequent to interaction with 11S,the cell-antibody complex can also interact with C'1-esterase to complete the initial phase, but C'1-esterase

1365


cannot precede 11S. The activity of 11S is not affectedby either Na3H-EDTA or serum inhibitor of C'1-esterase, whereas both block the action of C'1-esterase.Thus, the order of reaction of the components of com-plement in the D-L system is 11S factor followed byC'1 or C'1-esterase in the initial cold phase, after whichC'4, C'2 and C'3 cause lysis in the warm phase.

Blockage of D-Cycloserine Antibiosis by D-Alanine.PAUL D. HOEPRICH, Salt Lake City, Utah (intro-duced by M. M. Wintrobe).

While often clinically effective against several kindsof bacteria, D-cycloserine (the isomeric form of cyclo-serine produced by Streptomyces orchidaceus) is usuallyineffective according to in vitro testing with conventionalculture media. When 262 bacterial isolates were simul-taneously tested for cycloserine susceptibility in twosynthetic broth media (which differed only in the pres-ence of DL-alanine), Escherichia coli, Salmonella sp.,providence bacilli, enterococci and Staphylococcus au-reus were significantly more often susceptible in theabsence than in the presence of alanine. With 60 E.coli isolates, susceptibility to benzylpenicillin, dihydro-streptomycin, neomycin, polymyxin B, chloramphenicol,tetracycline, sulfadiazine, and nitrofurantoin were simul-taneously and identically assayed. Comparison of thein vitro effectiveness of these antibacterial agents atconcentrations obtainable in the blood of humans receiv-ing clinically practical dosages of these agents revealedthat cycloserine was more often inhibitory to E. coli thaneither dihydrostreptomycin or tetracycline and was aseffective as neomycin. These susceptibility data are inkeeping with the clinical experience that cycloserine isa useful agent in treating E. coli infections. D-alanineis the isomeric form of alanine that blocks D-cycloserineantibiosis. Using a strain of E. coli, the quantitativeaspects of D-alanine:D-cycloserine antagonism were stud-ied. Cycloserine antibiosis was subverted to a degreedirectly proportional to the quantity of D-alanine pro-vided, up to the point of equimolarity. When the con-centrations were equal, or when D-alanine was presentin excess, cycloserine had no antibiotic effect. L-alanineis without effect on D-cycloserine antibiosis. Yet testing,by means of conventional culture media, usually failsto reveal significant cycloserine susceptibility. WhenL-alanine containing synthetic broth medium was sub-jected to heat, as in ordinary sterilization by autoclave(1210 C for 15 minutes), it was rendered antagonistic tocycloserine antibiosis. The assumption that racemiza-tion of L-alanine had resulted with heating was borne outby direct analysis of the broth for D-alanine.

Studies of Lipid Transport in the Tissues. WILLIAM HOL-LANDER, RICHARD N. KAPLAN AND RussELL P. SHER-WIN, Boston, Mass. (introduced by Robert W. Wilkins).Extraction of nonatherosclerotic arterial tissue with

petroleum ether yielded less than 5%o of its contained

cholesterol, whereas extraction of atherosclerotic tissueyielded from 26 to 72% of its cholesterol. Extraction ofnormal plasma or of prepared 8-lipoprotein yielded lessthan 5% of its contained cholesterol, whereas plasma or,8-lipoprotein to which free crystalline cholesterol hadbeen added yielded all of the added cholesterol. Thesestudies indicated that atherosclerotic tissue containssizable amounts of unbound cholesterol as well aslipoprotein-bound cholesterol. The fate of unbound andbound lipid upon injection into living tissues was there-fore studied. C`4-labeled cholesterol and ,8-lipoprotein.labeled with C"4-cholesterol, were injected intradermallyinto separate sites on the normal human forearm. Skinbiopsies of the injection sites were obtained 3 to 7 dayslater. In the biopsies, over 90% of the C14-cholesterolradioactivity was recovered, whereas less than 10% ofthe lipoprotein radioactivity was recovered. Similarly,I"~-labeled triolein remained at the injection site, whereaslipoprotein labeled with I1"-triolein was removed within5 days. Removal of radioactive lipoprotein from alymphedematous limb was significantly slower than fromthe opposite normal limb, indicating that removal ispredominantly via the lymphatics. Histological studies onskin biopsies obtained 1 week after the intradermalinjection of nonisotopically labeled lipids revealed thatneither f-lipoprotein nor triolein caused any local in-flammation, whereas the injected unbound cholesterolcaused a chronic inflammatory response. After injectioninto the femoral arterial wall of dogs, C14-cholesterolradioactivity was 90% recoverable at the injection site5 days later, whereas lipoprotein labeled with C14-cho-lesterol had decreased to less than 5% of the amountinjected.

Metabolism of Arterioles and Venules as Studied in theCartesian Diver. RUFUS0. HOWARD,DAVID W. RICH-ARDSON, MORRIS H. SMITH, JR. AND JOHN L. PATTER-SON, JR.,* Richmond, Va.

The metabolism of the physiologically most active por-tions of the vascular bed, the arterioles and venules, hashad little attention. This report concerns application ofthe Cartesian diver technic to measurement of oxygenconsumption by arterioles and smallest arteries (130 to560 , in diameter), and of venules and smallest veins(300 to 900 is). Ninety-nine mesenteric arterioles and76 venules from 46 hamsters were studied in Ringer'sphosphate plus 200 mg%glucose as substrate. Resultswere similar whether the animal was anesthetized (pento-barbital) or decapitated unanesthetized. With 21%oxygen in the diver, oxygen consumption of arterioleswith in vivo diameters less than 300 /A averaged 0.9 ,1Aper mg per hour, double that of arterioles and smallarteries with diameters greater than 500 /A (0.4 ,ul/mg/hr;p <0.05). Metabolism of hamster venules averaged 0.5,0l per mg per hour, significantly less than that of thesmallest arterioles, but comparable with that of thelargest arterial tissue studied. There was no relationbetween venous size and metabolic rate. Both arteriolesand venules were highly resistant to hypoxia. Reduction

1366


in oxygen tension in the gas above the substrate in foursteps from 150 to 12 mmHg produced no change inmetabolic rate of hamster arterioles, but at 5 mmHg,arteriolar oxygen consumption was reduced to 60%o ofcontrol. Venules were even more resistant to hypoxia,maintaining normal metabolic rate at 5, although not at2 mmHg. Ten human vessels of both types from fivenormal mesoappendices showed lower metabolic ratesthan hamster vessels. Human arterioles, 130 to 350 a

in diameter, consumed 0.35; and venules of 260 to 900 IAconsumed 025 ul 02 per hour. The greater resistanceto hypoxia of venules as compared with arterioles mayreflect adaptation of metabolic processes to the loweroxygen tension (approximately 50 mm Hg) of thevenous environment. Working vessels in vivo might beexpected to show higher metabolic rates under normaloxygen tension, and depression of metabolism by lesserdegrees of hypoxia.

Responses of Isolated Toad Bladder to AnisosmoticBathing Media. JOHN A. ISOM, MEHMETN. KALAYCIAND JOHN D. CRAWFORD,* Boston, Mass.

Net sodium transport (short-circuit current) of toadbladder bathed in Ringer's solution is sharply reducedwhen the medium is made hypertonic (380 mOsm/kg)by addition of NaCl, sucrose, or urea. Potential differ-ence concomitantly falls and resistance rises. The effectis fully reversible. The influence of hypotonicity wasstudied in sodium-poor (55 mEq/L) Ringer's solution.The same fluid made eutonic by addition of sucrose servedas control medium and was shown to sustain satisfactorylevels of sodium transport. The hypotonic fluid causedfully reversible, large, and abrupt rises in short-circuitcurrent and potential difference with decline in re-sistance. Vasopressin provoked increments in short-circuit current both in membranes at the peak of theirresponse to hypotonicity as well as in those in whichprior exposure to hypertonic media had reduced elec-trical activity, in some instances almost to nil. In bothcircumstances, the extent of vasopressin response wassimilar to that of the control bladders in eutonic Ringer's.Vasopressin also rendered the otherwise unresponsivemucosal surface reactive to changes in medium tonicity.Measurements of tissue water content indicate 5.7 and2.2% losses after 30 minutes' incubation in Ringer's madehypertonic, respectively, with sucrose and NaCl; nochange with urea; and a 7.9% increase after incubationin the hypotonic medium. Vasopressin failed to influencetissue hydration. Changes in bioelectric behavior in-duced by anisosmotic media thus failed to correlate withalterations in tissue water. They may relate to changesin mitochondrial size, such as are known to occur withtonicity change, and recently have been demonstrated inrenal medullary mitochondria exposed to vasopressin.Regardless of mechanisms, the findings, especially shouldthey represent a general cellular phenomenon, appearto have interesting implications to volume control in thewhole animal.

Loss of Functional Activity of Platelets during Storage.DUDLEY P. JACKSON,* J. RICHARD GAINTNER ANDPHILIP ZIEVE, Baltimore, Md.

During storage, platelets lose their ability to supportclot retraction more rapidly than can be attributed to adecrease in the number of morphologically intact plate-lets. Platelet-rich plasma, prepared from human orcanine blood collected in plastic bags containing acidcitrate dextrose, was stored at 40 C. At intervals, thedegree of retraction of aliquots of the plasma wasmeasured after addition of thrombin, and the clots werefixed in formalin. The degree of retraction and thenumber of morphologically intact platelets remained con-stant for approximately 7 days. Sections of the retractedclots showed that the platelets had undergone extensivemorphologic changes of viscous metamorphosis, includingswelling, pseudopodial formation, and aggregation. Thefibrin strands appeared oriented to the platelets. Duringthe second week of storage, the degree of retraction pro-gressively decreased but the number of intact plateletsfell only slightly. Clot sections showed that as the degreeof retraction diminished, the number of platelets under-going viscous metamorphosis progressively decreased, andthere was a concomitant increase in the proportion ofsmall, discrete, rounded platelets. The fibrin strands ap-peared oriented to the swollen and aggregated plateletsbut not to the small rounded platelets. Clots of plasmastored for 21 days did not retract after addition ofthrombin even though more than 60% (>200,000/mm')of the platelets appeared intact. In sections of these un-retracted clots, the platelets were small and round, andnone appeared to have undergone viscous metamorphosis.These studies indicate that loss of the ability of storedplatelets to support clot retraction is due to progressiveloss of functionally active platelets and not to simul-taneous, gradual loss of activity of all of the platelets.

The Regulation of Spleen Growth and SequesteringFunction. HARRYS. JACOB, RICHARD A. MACDONALDAND JAMES H. JANDL,* Boston, Mass.

The mechanisms regulating the growth of reticulo-endothelial tissue are obscure. That growth of sucherythroclastic tissue is influenced by its "work load" issuggested, because splenomegaly often accompanies hemo-lytic anemias owing to intrinsic defects of red cells and,in animals, splenic atrophy may follow repeated venesec-tions. This subject was investigated by studying thegrowth, histology, and sequestering function of spleens,livers, and spleen autotransplants in four kinds of rats:sham-operated; splenectomized; splenectomized, with small,subcutaneously-placed spleen autotransplants; and hemi-splenectomized, with similar autotransplants. The func-tion of these various erythroclastic organs was assessedby injecting an "excess" of Cr51-labeled red cells that hadbeen so altered in vitro that they were susceptible tosequestration by the normal spleen and liver. The growthof spleen transplants and their capacity to sequester alteredred cells was three times as great in splenectomized as in

1367


hemisplenectomized rats. The increased size of transplantsin splenectomized animals reflected hyperplasia of reticulo-endothelial cells. Administration of actylphenylhydrazineinduced giant transplants with heightened sequesteringfunction of thrombocytopenia. The presence of hyperplas-tic spleens inhibited the sequestration of red cells by theliver, despite an excess of altered cells in the circulation,whereas in the absence of spleen tissue the sequesteringcapacity of the liver was accentuated. These studies dem-onstrate that the size of the reticuloendothelial system andits functional capacity with respect to sequestering redcells are governed by the work required of it. Thismay be through the extent of stimulation of reticulo-endothelial cell division by sequestered particles. Theysuggest that splenic hyperplasia caused by an abnormalityof one type of blood cell may produce increased sequestra-tion of another. Finally, a deficiency of spleen tissueappears to induce compensatory "hyperhepatism," a phe-nomenon that may account for relapses after splenectomyin hemolytic anemias.

Influence of Growth on Total Urinary Hydroxyproline.HUGOE. JASIN, CHESTER FINK, J. DONALD SMILEYANDMORRIS ZIFF,* Dallas, Tex.

Increased excretion of urinary hydroxyproline peptidesin children as compared with adults has been previouslydemonstrated in this laboratory. This has suggested a

relationship between the level of excretion of thesepeptides and growth. To investigate this phenomenonfurther, excretion of hydroxyproline peptides has beenstudied in dwarfism and acromegaly. Mean 24-hourexcretion of total hydroxyproline was 33.8 mg in 23normal and hospitalized adults. In 34 normal and hos-pitalized children, 5 to 17 years of age, it was 74.3 mg.Excretion was reduced in 8 children with dwarfism,ranging between 5.6 and 28.1 mg daily. An 8 year oldpituitary dwarf was subjected to a controlled metabolicstudy. Before treatment with growth hormone (GH),hydroxyproline excretion averaged 14.1 mg daily, one-

fifth the average of normal children of the same age.

Human GH was administered in a dosage of 2.5 mgdaily for 28 days. Growth occurred, and there was a

dramatic rise in excretion to a maximum of 63.5 mgper day. 'Twenty-four days after cessation of therapy,urinary hydroxyproline had fallen to 18.6 mg. In an

infant cretin, excretion was 5.6 mg; after 7 weeks oftreatment with thyroid extract it had risen to.47.5 mg.The mean excretion of a patient with acromegaly was

101.5 mg, three times the average for normal adults.Two months after irradiation it had fallen to 86.0 mg,

and after 3 months to 58.3 mg. In view of the 1) ele-vated excretion of total hydroxyproline in children as

compared with adults, 2) low excretion observed in non-

growing children, 3) rising values after administrationof GH and thyroid extract, 4) increased values in a

patient with acromegaly and subsequent fall after irradi-ation, it is concluded that there is a relationship betweenthe excretion of hydroxyproline peptides and growth. It

is suggested that this reflects the amount of metabolicallyactive, soluble collagen in the body.

Mechanism of Renal Effects of Glucocorticoids in Cirrho-sis. HERSHELJICX, JOSEPH L. COHENAND ROBERTS.MORRISON, Boston, Mass. (introduced by Joseph M.Hayman).

Depression of maximal urinary concentrating ability(Umar) occurs regularly in patients with cirrhosis. TheUmar was similar in six patients receiving at least 45mEq sodium per day [665 ± 40 mOsm/L (1 SE)] ascompared with nine patients receiving 9 mEq of sodiumper day (614 ± 29 mOsm/L). However, while the for-mer group exhibited a normal TemH2O (5.5 ± 0.78 ml/min), patients on marked sodium restriction had a lowTemH2O (2.5 ± 0.43 ml/min). In the latter group (aftercontrol values were obtained) the effects of acute (5 pa-tients) and chronic (3 patients) administration of methylprednisolone were studied. Administration of methylprednisolone for 8 days was associated with a rise inUmax of 300, 120, and 220 mOsmper L, and a rise inTCMHIO of 2.3, 3.0, and 4.7 ml per minute, respectively.Values returned to control levels 1 week later. Drugadministration was associated with a fall in CG. in two ofthe three patients. The five patients studied acutelywere in the dehydrated state, two without mannitol, andthree in a steady state of mannitol diuresis, with exoge-nous vasopressin. Intravenous steroid administration re-sulted within 3 hours in a small increase in urine volume(0.18 and 0.24 ml/min) in the dehydrated state alone, buta large increase (2.7, 4.3, and 5.0 ml/min) during steadymannitol diuresis. This rise in volume was accompaniedby an initial rise in urine osmolality, resulting in an in-creased T.mH2O. The Cz. remained unchanged or fellslightly in four of the five patients. The increasedTemH2Qwas independent of solute excretion and filteredor excreted sodium. The rise in volume and TcrnH2Qwithout a rise in Cin after acute steroid administrationstrongly suggests that glucocorticoids inhibit proximalreabsorption of fluid, and also, either directly or byincreasing the efficiency of the sodium pump in the loopof Henle, enhance maximal antidiuretic hormone effect.

Studies on Iron Transport in Surviving Intestinal Seg-ments of the Rat. JOSE M. JOB AND DORIS A. HoWEL,Durham, N. C. (introduced by Jerome S. Harris).

Contradictory conclusions concerning the intestinaltransport of iron using the "in vitro" surviving intestinalsegment technique of Wiseman and Wilson have beenreported. Using Fe', we have found that the concentra-tion of radioiron on the serosal side (C.) of the rat duo-denum and proximal jejunum becomes greater than thaton the mucosal side (Cm) after 3 hours of incubation influid of initially equal C. and Cm. However, the totalamounts of radioiron in both serosal and mucosal com-partments decrease during the incubation. The volumeschange but slightly. Although the presence of a concen-tration gradient is suggestive of iron transport, failure to

1368


find an actual accumulation of Fe's in the serosal com-partment leaves the proof of active transport by thistechnique still unsettled. Difficulties in the interpretationof ferrokinetic data obtained by this technique will bediscussed.

Experimenteal Coxsackie A-21 (Coe Agent) Infection inVolunteers. KARL JOHNSON, ANDERSONSPICKARD ANDHUGHEVANS, Bethesda, Md. (introduced by VernonKnight).

Recent evidence that Coxsackie A-21 was the cause ofa large outbreak of respiratory illness in a military popu-lation suggests that it may be a more frequent cause ofhuman illness than was previously appreciated. To sup-plement epidemiological studies, experiments in 50 adultvolunteers were carried out. They revealed an acuteinfluenza-like illness in susceptible volunteers given amoderate intranasal dose of the agent, but subjects withhigh titers of naturally acquired antibody exhibited im-munity to challenge as measured by virus recovery orproduction of illness. By means of a quantitative measureof nasal secretion it was found that inoculated volunteerswith low titers of pre-existing antibody developed acommon cold syndrome with a profuse rhinorrhea, whilethose with no antibody developed a more severe systemicillness, often with fever but with little or no nasal dis-charge. Despite its designation as an enterovirus, Cox-sackie A-21 was shed more profusely and for a longerperiod in the throat than in the feces. In corroborationof this apparent great affinity for the pharynx, inoculationof the agent into the small bowel (enteric-coated capsulesor tube) produced only sporadic viral infection, no illness,and no antibody response. Thus, by all the criteriaexamined, Coxsackie A-21 appeared to be primarily arespiratory viral agent, a finding that is in contrast tothe implication from its present classification as anenterovirus. This is but one of a number of such in-consistencies, and study of a larger number of new agentsin volunteers may permit some rationalization of dataon host-parasite relationships with those derived fromlaboratory studies now used for virus classification.

Effect of Parathyroid Hormone on Bone CollagenMetabolism In Vitro. C. CONRAD JOHNSTON, JR.,EDWARDB. MINER AND WILLIAM P. DEISS, JR.,*Indianapolis, Ind.

Both the clinician and the histologist have observedchanges in bone matrix produced by parthyroid hormone(PTH). In order to study changes in collagen metabo-lism of matrix at a biochemical level, calvaria fromcontrol rats and from rats injected with parathyroidextract (PTE; 75 U every 2 hours for 3 doses) wereincubated in Krebs-Ringer bicarbonate media containing1 ,Lc L-proline-C"4, and the specific activity (SA) ofmatrix hydroxyproline was determined. Active incorpo-ration of the label (74 cpm/,umole) is seen in an atmos-phere of 02- Incorporation is enhanced 50% by addingglucose to the media, inhibited 95% by anaerobiosis and

98% by iodoacetate (0.001 M), and is essentially un-changed by fluoroacetate (0.002 M). Bones from thePTE-injected rats demonstrate an early marked decreasein incorporation of labeled proline into the hydroxypro-line. Two hours after the last injection the SA was 57%of control, and at 12 hours it was 17% of control. TheSA of the hydroxyproline of the experimental group hadreturned to control level at 36 hours and had increasedto 142%o of control at 72 hours. The addition of glucoseto the media stimulated uptake in both groups, but thedifferences were maintained. These data indicate thatnormal bones can incorporate proline-C'4 into the hy-droxyproline of matrix in vitro without addition ofsubstrate. Bones undergoing PTH-induced reabsorptionshow a marked impairment of incorporation of the label.This is in contrast to the stimulated incorporation ofglucose-2-C1' into matrix hexosamine noted in bones ofsimilarly treated animals. Both observations demonstrateearly metabolic changes in bone matrix after the injec-tion of PTE.

Characteristics of the Genetic Code. OLIVER W. JONES,J. HEINRICH MATTHAEI AND MARSHALLW. NIREN-BERG, Bethesda, Md. (introduced by Ivan W. Brown, Jr.).

Knowledge of inheritable metabolic disease in manmust be preceded by determination of the mechanism ofprotein synthesis and the steps by which genetic instruc-tion for this synthesis is transmitted. Wehave recentlydeveloped a system that provides a sensitive assay for"messenger" RNA. The introduction of enzymicallyprepared polyribonucleotides into the system has resultedin detailed information regarding coding units corre-sponding to 15 amino acids. Characteristics of the codesuch as degeneracy, the existence of nonsense "words,"the minimal coding ratio, and universality of the codewere investigated. Polyuridylic acid specifically stimu-lates incorporation of L-phenylalanine into poly-L-phenyl-alanine. Incorporation of 14 other amino acids intoprotein was stimulated by a series of polymers contain-ing different ratios of uridine:adenine, uridine:cytosine,uridine: guanine, and uridine: cytosine: guanine. Sincecoding units corresponding to four of these amino acidsrequire three nucleotides, the minimal coding ratio mustbe three. Two coding units for leucine were foundwhich demonstrated that part of the code is degenerate.Certain polynucleotides did not code for any amino acids,suggesting the presence of "nonsense" coding units.Marked similarity between RNA coding units of twophylogenetically dissimilar species indicated that part ofthe code is universal.

The Pharmacologic Response of the Human Esophagusto Curare and Atropine. PAUL A. KANTROWITZ, M.JEROME STRONG, CHARLES I. SIEGEL AND DONALDSTEPHENSON, Army Chemical Center and Baltimore,Md. (introduced by Albert I. Mendeloff).Normal esophageal peristalsis progresses continuously

through the upper striated muscle and lower smooth

1369


muscle segments; moreover, the two types of esophagealmusculature cannot be distinguished by studying normalswallowing. Since esophageal involvement in neuro-muscular diseases has been demonstrated, ability to dis-tinguish between esophageal striated and smooth musclewould be useful. Intraesophageal manometric studieshave therefore been performed to determine whether thetwo musculatures differ in their response to pharmaco-logic stimulation. Pharmacologic agents selected wereatropine, an anticholinergic known to inhibit smoothmuscle, but without striated muscle effect, and d-tubo-curarine, a striated muscle inhibitor without smoothmuscle effects in usual doses. Earlier studies in theselaboratories demonstrated that peristaltic amplitude inthe lower esophagus is diminished by atropine in dose-related fashion, while upper esophageal peristalsis isunaltered. In the present study, simultaneous intra-luminal pressures were recorded, employing three open-tipped catheters with tips 5 cm apart and the upper tip1 to 2 cm below the upper esophageal sphincter. Sixnormal adult males were studied, before and afterintravenous administration of 9 to 10.5 mg d-tubocurarinechloride. After curare, peristaltic amplitude in thestriated muscle segment decreased by 30 to 50%, whileamplitude in the smooth muscle portion was unaffected.Intravenous administration of 1 mg atropine thenabolished smooth muscle peristalsis, without additionalstriated muscle effect. After intravenous neostigmine,peristaltic activity in both striated and smooth musclereturned to normal. These studies demonstrate that,although the striated and smooth muscle segments areindistinguishable in normal esophageal peristalsis, theydo differ markedly in their pharmacologic response.

Immunological Studies with Intrinsic Factor in Man.MANUELE. KAPLAN, RALPH ZALUSKY, JACK REMING-TON AND VICTOR HERBERT,* Boston, Mass.

It has been suggested that refractoriness to orally ad-ministered hog intrinsic factor concentrate (HIFC) inpatients with pernicious anemia may be immunologicallydetermined. To test this hypothesis parenteral immuniza-tion with purified HIFC was carried out in two patientswith pernicious anemia, who had demonstrated theirability to absorb vitamin B,2 when it was administeredwith HIFC, and in albino rabbits. Control and post-immunization sera were tested for anti-intrinsic factoractivity in vitro and in vivo. Despite the development,with immunization, of immediate skin sensitivity andweakly precipitating serum antibodies demonstrable byagar diffusion techniques, no refractoriness to orally ad-ministered HIFC appeared in either patient. However,the feeding of 40 ml of autogenous immune serum withHIFC-Co'B12 prevented B,2 absorption in one patient anddiminished it in the other. Under the same conditions,10 ml of immune rabbit serum, containing high-titerprecipitating antibodies of HIFC, totally suppressedCo6"B12 absorption. Immune, but not control, sera fromboth human and rabbit inhibited HIFC-enhanced uptakeof CoeB,2 by liver and gut homogenates and prevented

the migration of a HIFC-Co'7B,, complex on paperelectrophoresis. Serum from 5 of 6 patients refractoryto HIFC did not possess any of the in vitro propertiesdescribed above and exhibited by the immune sera.Serum from the other refractory patient possessed allthese in vitro properties except one: it failed to precipi-tate HIFC in agar diffusion systems. These studiesdemonstrate that the presence of serum antibody to HIFCdoes not necessarily affect the pharmacologic activity ofHIFC within the gastrointestinal tract. The refractorystate may be mediated by local factors within the in-testinal wall, possibly including cell-fixed antibody.

Studies of an Antigen in Cell Walls of Group A Strep-tococci Possessing an Immunologic Relationship toHuman Heart Tissue. MELVIN H. KAPLAN,* Cleve-land, Ohio.Experiments employing rabbit antisera to cell walls

have given evidence of a protein component in certainstrains of group A streptococci, which exhibits im-munologic cross reaction with human heart tissue. Thepresent report is concerned with the properties of par-tially purified preparations of this streptococcal antigen,the histologic distribution of the cross-reactive antigenin human organs, and the biologic variations of thestreptococcal antigen among group A strains. Partiallypurified preparations of antigen were derived from acidextracts of whole cells or cell walls, subj ected toribonuclease digestion, neutral salt fractionation, andchromatography on DEAE- and CM-cellulose. Theactive material was protein in nature, and was destroyedby trypsin, pepsin, and chymotrypsin. It could be obtainedseparately from the group A carbohydrate, essentiallyfree of glucosamine or rhamnose. Active fractions con-tained M protein activity, although the antigen waspresent in streptococcal strains of different Lancefieldtypes. The cross-reactive antigen as determined byfluorescent-antibody techniques is a constituent of cardiacmyofibers, skeletal muscle, and smooth muscle of vesselwalls, and was demonstrable in tissues of both rheumaticand nonrheumatic individuals. The antigen was alsodemonstrated in rabbit heart tissue. The converse crossreaction could also be demonstrated. Antisera to humanheart or rabbit heart exhibited precipitating antibodiesreactive with active streptococcal fractions. The hypothe-sis is proposed that in rheumatic fever an autoimmunereaction to a constituent resident in human heart maybe induced by exposure to a cross-reactive antigen ofgroup A streptococci. This concept is supported by ourprevious observation of bound y-globulin in myofibersand vessel walls of rheumatic hearts.

Bipolar Electrodes for Measurcment of Kidney Pulse bythe Impedance Method. YALE J. KATZ, Los Angeles,Calif. (introduced by Irving Gordon).Phasic volume change in the kidney during the cardiac

cycle allows the registration of the organ pulse by anelectric impedance method. The contour of the pulsewave is affected by changes in the caliber of the renal

1370


artery. For recording the kidney pulse, two small bi-polar electrodes have been recently developed. One con-sists of a plastic plate containing two needle points 1 cmapart. At surgery the electrode plate is fixed to thekidney capsule with the electrode points in kidney cortex.The recording apparatus consists of an AC generator,40,000 cycles per second, and a capacitance bridge; thesignal resulting from the unbalancing of the bridge bythe change in tissue impedance is fed to a DC recorder.The tracing of the kidney pulse wave resembles in formthe pressure curve of a peripheral artery. The kidneypulse of a patient with significant stenosis of the renalartery has a delayed onset, a prolonged upstroke time,and some decrease in amplitude. In one patient a suc-cessful endarterectomy was followed by conversion ofthe kidney pulse to one of a normal contour. The secondelectrode consists of a bipolar unit built into the trocarof the Vim-Silverman renal biopsy needle. It has beenused successfully for kidney pulse wave determination bythe percutaneous route in Goldblatt hypertensive dogs.Since stenosis of the renal artery is seen at times ataortography in the absence of hypertension, one shouldnot assume when hypertension is present that thestenosis of the artery is the cause of the hypertension.One can only properly assume this if the stenosis is asignificant one in a hemodynamic sense. Kidney pulsemeasurement will assist in determining the functionalsignificance of a stenosis of the renal artery.

Effect of Thyroid Hormone on Collagen Metabolism.HARRYR. KEISER AND ALBERT SJOERDSMA,* Bethesda,Md.

Body collagen, taken as a whole, has been consideredto be metabolically inert in adult animals. That this con-cept may require revision is suggested by the present find-ings of the production by thyroid hormone of increasedurinary excretion of peptide-bound hydroxyproline; thelatter is a convenient index of collagen degradation inman. Six normal adults and a myxedematous patientwere studied. During maintenance on a low hydroxypro-line diet, 3,5,3'-triiodothyronine (T,) was given orally indosage sufficient to produce hyperthyroidism in the nor-mal subjects and euthyroidism in the patient. Controlvalues for total urinary hydroxyproline were 16.4 to 40.5mg per day in the normal subjects. A 56 to 109%increase occurred during administration of T3 for 10 to21 days with a return to normal within 10 days of dis-continuing treatment. In all instances free hydroxypro-line was less than 5%7 of the total. Other changesduring the period of hypermetabolism were as follows:caloric intake, 0 to +31%; protein intake, 0 to +41%o;total urine nitrogen, +8 to +65%; free urine a-aminoacid nitrogen, +7 to +30%o; and free urine tyrosine, +30to +61%o. In the hypothyroid patient, control hydroxy-proline values averaged 3.1 mg per day, a very low level.An increase to 15.4 mg per day (+397%o) occurred after21 days on T3 in the presence of a constant protein in-take, total urinary nitrogen increasing only 8.8%. Thesedata indicate that collagen, as well as general protein

metabolism, was altered to a marked degree by T, ad-ministration. The rapid and marked increase in urinaryhydroxyproline could have resulted from increased syn-thesis of a rapidly metabolized form of collagen (i.e. solu-ble collagen), or increased degradation of collagen, or acombination of both. The precise mechanism is underinvestigation.

Studies on the Role of the Brain in the Steroidal In-hibition of ACTHRelease. JOHNW. KENDALL, KUNIOMATSUDA, CAROLINE DUYCKAND MONTEA. GREER,*Portland, Ore.

Previous studies have provided conflicting evidence re-garding the site of the receptor area at which adreno-cortical steroids inhibit ACTH release. To evaluate therole of the brain in this mechanism, the effect of corti-costeroid administration on pituitary-adrenal function wasdetermined in rats after removal of graded amounts ofbrain. A "hypothalamic island" group was prepared byremoving the cerebral hemispheres, thalamus, and hypo-thalamic connections to the posterior brain. "Medianeminence island" and "isolated pituitary" groups wereprepared by additional subtotal or total extirpation of thehypothalamus. Half of each group was given 100 ,ugdexamethasone subcutaneously 20 hours postoperatively;after another 4 hours adrenal corticosterone secretionrates were measured by fluorescences assay of a 3-minuteplasma specimen obtained by left adrenal vein can-nulation. Secretion rates were high in untreated intactand "hypothalamic island" groups owing to the stressassociated with obtaining the specimen. Approximatelyhalf of the untreated "median eminence island" and "iso-lated pituitary" animals had similarly high secretionrates indicating that they too responded to stress. Incontrast, adrenal secretion was markedly supressed in allanimals receiving dexamethasone. These findings indi-cate that administration of adrenocortical steroids pre-vents ACTH release in rats stressed by adrenal veincannulation even though the entire brain anterior to thesuperior colliculus and superior to the pituitary has beenremoved. This portion of the central nervous systemtherefore does not seem an essential part of the steroidalfeedback mechanism.

Intracellular Acid-Base Relations of Dog Brain withReference to the Extracellular Volume of the Brain.R. F. KIBLER, R. P. O'NEILL AND E. D. ROBIN,* Pitts-burgh, Pa.

Brain intracellular pH (pH,) was determined in vivoin dogs by the DMO(5,5-dimethyl 2,4-oxazolidinedione)technique under control circumstances and after the ad-ministration of NaHCO3, HCl, and CO2. The DMOtechnique requires estimation of the extracellular fluidspace (ECF) of the brain. Since there is no generalagreement concerning the magnitude of this space, bothchloride and sulfate spaces were measured as estimatesof brain ECF. These spaces averaged 41 ± 4%o and 6 ±1% of total brain water, respectively. Since calculations

1371


based on chloride space frequently resulted in negativevalues for intracellular DMO, chloride space could notbe used to determine brain pH,. In 21 dogs undercontrol conditions the mean sulfate space-derived pH,was 7.01 + 0.12 at a plasma pH of 7.36 ± 0.08. Onehour after NaHCO. infusion in 7 dogs, mean plasma pHrose 0.33 + 0.08 U while pH, rose 0.12 + 0.07 U; onehour after HC1 infusion in 4 dogs, plasma pH fell0.28 ± 0.13 U while pH, fell 0.04 ± 0.03 U. Inhalation of10% CO2 for 1 hour in 2 dogs caused a fall in plasmapH of 0.40 and 0.45 U and a fall in pH, of 0.35 and0.39 U, respectively. These results suggest that sulfatespace is a more reasonable estimate of brain ECF than ischloride space. The value of 6% is in close agreementwith recent estimates of brain ECF based on electronmicroscopic observations. Accordingly, it appears thatmuch of the brain chloride is intracellular, possibly withinglial cells, as previously proposed by others. The meansulfate space-derived intracellular pH of 7.01 is similarto that obtained for muscle and for mean whole bodyintracellular pH. Brain cells also behave like musclecells in their slow equilibration with HCO37 and H+ ionsand rapid equilibration with CO2.

Intrinsic Renal Regulation of Sodium Reabsorption.FREDRIK KIIL, Dallas, Tex. (introduced by Donald W.Seldin).Micropuncture studies have demonstrated that the thick

ascending limb functions as an obligatory diluting seg-ment. To test the hypothesis that this segment reducesthe concentration of tubular sodium to a hypotonic levelirrespective of sodium delivery, the quantity of sodiumreaching this segment in anesthetized dogs was increasedeither by raising glomerular filtration rate (GFR) orreducing proximal tubular reabsorption. Free waterclearance (CHso) was used as an estimate of sodiumreabsorption in this water-impermeable area. Startingwith water diuresis with minimal urine osmolal concen-tration (50 to 60 mOsm/kg), CH2o more than doubledwhen administration of hypotonic mannitol increased urineflow to two-thirds of GFR. When GFR was increasedby administration of hypotonic saline, C}I2o increasedto 12 ml per minute, nearly one-third of the GFR ob-tained during control water diuresis, or 18% of actuallyfiltered volume. Unchanged distal sodium reabsorptionassumed, sodium concentration of fluid issuing from thethick ascending limb did not exceed 80 mEq per L atmaximal flow rates. Sodium reabsorbed in the water-impermeable segment was calculated as half the productof CH2o and osmolal concentration of serum. Duringhypotonic mannitol diuresis sodium reabsorption de-creased in proximal tubules and increased in the water-impermeable segment. In this segment a larger increasein sodium reabsorption occurred during hypotonic NaCldiuresis, in accordance with hypothesis. These observa-tions suggest that the thick ascending limb alwaysfunctions to reduce the sodium concentration to a lowvalue, irrespective of the rate of sodium delivery, andwould thereby explain why changes in filtration rate or

serum sodium concentration result in corresponding di-rectional changes in tubular sodium reabsorption.

Electroencephalographic Evaluation of Uremia. JOHNE. KILEY AND ORLANDOHINES, Albany, N. Y. (intro-duced by Richard T. Beebe).

Dialytic therapy has been hampered by inadequateobjective criteria of the severity of uremia. Patients inuremia gradually lose their ability to think. Since thisimpairment has been a helpful indication for dialysis,electroencephalographic changes were investigated as anindex of the degree of uremia. Electroencephalogramswere performed on 16 uremic patients and serial electro-encephalograms were followed in 13. Patients wereevaluated clinically with particular regard to mentalstate and grouped as follows: A, alert and rational; B,lethargic or toxic; C, irrational (definite impairment ofthinking); D, semicomatose (responding only to pain));E, comatose (completely unresponsive). Changes inblood urea nitrogen and creatinine were followed as wellas the accumulation of generally unidentified anions, esti-mated as the difference of serum sodium plus potassiumminus bicarbonate plus chloride concentrations. Thediffuse slowing of the electroencephalogram observed inuremia was quantitated by enumeration in occipital leadsof both the per cent distribution of various frequenciesand the per cent of frequencies that were abnormal. Theelectroencephalogram slowed and the percentage of ab-normal frequencies increased as uremia became worse,and these changes reversed when uremia improved. Eitherhemodialysis or peritoneal dialysis decreased the per-centage of abnormal frequencies. The electroencepha-logram was a more sensitive indication of the effect ofuremia than was our evaluation of mental status. Ofseven patients with 90% or more of frequencies abnormal,all died except for two promptly treated by dialysis.Correlation between biochemical measurements and slow-ing of the electroencephalogram was poor. It is con-cluded that quantitative estimation of slowing of theelectroencephalogram by uremia provides a needed indexof the net effect of complex metabolic derangement oncentral nervous system function.

Studies to Determine whether Glucose Must Be Metabo-lized to Induce Insulin Release. CHARLES KILo,CHARLES L. LONG, JR., RITA M. BAILEY, MARY B.KOCHAND LILLIAN RECANT,* St. Louis, Mo.

Previous investigations using isolated pancreas, cross-circulation experiments in dogs, intravenous infusions ofglucose, and measurement of pancreaticoduodenal veinblood samples in dogs have shown that glucose is capableof inducing insulin release from the pancreas. The pur-pose of this investigation was to determine 1) whetherpenetration of glucose into the f8-cell was necessary forstimulation of insulin release; 2) the metabolic step inglucose utilization inducing insulin release. Substancestested were infused into pancreaticoduodenal and splenicarteries in intact anesthesized dogs. Portal venous speci-

1372


mens were obtained before and after infusions. Insulin-like activity (ILA) was measured in portal samples bya modification of the rat epididymal fat-pad assay.Experiments were performed on 62 dogs. Perfusion ofsaline or pyruvate resulted in no increase in ILA. Glu-cose, fructose, or tolbutamide produced significant in-creases in ILA (600 to 400 gtU/ml); 2-deoxyglucose, asugar metabolized only to 2-deoxyglucose-6-phosphate,resulted in no ILA response. Since 2-deoxyglucose-6-phosphate inhibits the- subsequent metabolism of glu-cose-6-phosphate, simultaneous infusions of glucose and2-deoxyglucose were studied. Complete inhibition ofinsulin release occurred when the glucose to 2-deoxyglu-cose blood levels were 1.5 or less. Delayed but significantILA increases were observed when ratios were increased.2-Deoxyglucose infusions resulted in increased glucoseA-V differences for 15 to 30 minutes, followed by pro-longed inhibition of glucose uptake. These studies showthat glucose, fructose, and tolbutamide, but not 2-deoxy-glucose or pyruvate, are capable of stimulating insulinrelease. They further suggest that glucose must pene-trate the 8-cell and be metabolized beyond the glucose-6-phosphate step in order that insulin be released.Studies are under way 1) to test additional utilizable andnonutilizable sugars; 2) to test metabolic products ofglucose; and 3) to evaluate the effect of 2-deoxyglucoseon release of ILA by tolbutamide.

Site of Action of Androgen on Adrenal Steroidogenesis.JULIAN I. KITAY, Charlottesville, Va. (introduced byAlfred Jay Bollet).

The influence of androgen on steroidogenesis wasstudied in adult intact and castrated male rats. Adrenaltissue for intact animals produced 0.031+± 0.002 /Amole(per 100 mg adrenal weight) of corticosterone whenincubated in vitro without ACTH. A significant decre-ment to 0.023 ± 0.001 gtmole per 100 mg was observedin adrenal slices from castrated rats. Testosterone ad-ministration to castrated rats in vivo enhanced in vitrosteroidogenesis to a mean value of 0.040 ± 0.002 Amoleper 100 mg. Studies were then undertaken to explorethe factors involved in this stimulatory effect ofandrogen. Adrenal glycogen concentration (as glucose)was found to be 0.758 ± 0.069 /mole per 100 mg in intactrats. Castration resulted in a 58% reduction to 0.317 +

0.052 /mole per 100 mg. Adrenal cholesterol levels were8.94 + 0.65 gmoles per 100 mg in the intact animals and13.73 ± 1.14 /Amoles per 100 mg after castration, an in-crement of 53%. The concentrations of three enzymesin adrenal tissue-phosphorylase, phosphodiesterase, and5'-nucleotidase-were found to be reduced significantly bycastration (decrements of 16, 32, and 24%o per mg ofadrenal protein, respectively) and restored to normallevels with testosterone replacement. On the other hand,adrenal glucose-6-phosphate dehydrogenase activity wasactually somewhat enhanced by castration, and ATPasewas unchanged. Finally, the decrement in steroidogenesisobserved in adrenal tissue from castrated rats was notobtained when adrenal homogenates were prepared and

incubated in the presence of excess TPNH. ACTHad-ministration has been shown previously to increase theactivity of adrenal phosphorylase, phosphodiesterase, and5'-nucleotidase. However, in the present studies, theaddition of a supramaximal dose (0.5 U) of ACTH invitro did not alter the decrease in adrenal steroidogenesisproduced by castration. These data suggest that tes-tosterone affects steroidogenesis by stimulation of thepathway leading to increased generation of TPNH.This stimulation is analogous to, but not identical with,that produced by ACTH.

Urea Transport in the Central Nervous System. CHARLESR. KLEEMAN,* HUGH DAVSON AND EMELIO LEVIN,London, England and Los Angeles, Calif.

Studies on urea transport in the central nervous systemof rabbits were done to investigate the mechanism of itsability to lower cerebrospinal fluid (CSF) pressure.Information was obtained on the site of the "barrier" tourea penetration and on the presence of a dynamicextracellular space of the brain. Sustained intravenousand intracisternal infusions of C'- and C14-urea wereadministered for 4 hours to normal rabbits; the rate ofC14-urea penetration in the brain, cord and CSF wasfollowed in nephrectomized rabbits up to 24 hours. Thepreinfusion content of urea in the water of cord and whitematter equaled that of plasma water. Urea content ofgray matter was 18% higher than in plasma, while ureain CSF was only 0.7 to 0.8 the plasma concentration.Despite the latter, the CSF had the same urea concen-tration as the choroidal secretion (ventricular fluid).Blood urea maintained at 200 to 300 mg per 100 mlcaused a 7 to 12% decrease in water content of CNStissues without affecting rate of CSF formation. Dehy-dration of the brain was the major cause of the fall inCSF pressure. Urea penetrated from blood to CNStissues aproximately one-tenth as rapidly as into muscle.The osmotic pressure differential explained tissue dehy-dration. It required 15 to 24 hours for plasma to tissueurea ratio to reach 1. Urea infused into subarachnoidspace penetrated CNS tissues four to five times morerapidly than when infused intravenously. Water contentof brain increased during subarachnoid perfusion. Be-cause of these results with urea, sucrose "space" in brainwas compared after i.v. and subarachnoid infusions ofsucrose; i.v. sucrose space was less than 2% of brain,while subarachnoid infusion space was 10 to 14%. Urealowers CSF pressure by decreasing brain volume. Themajor barrier to penetration is at capillary, not glial orneuronal cell membrane. A sucrose "space" exists whencapillary is "bypassed" by subarachnoid perfusion.

Physiological Changes of Erythrocyte Carbonic Anhy-drase Activity. SAMUEL KORMANAND MARIANNEKARA, New York, N. Y. (introduced by Robert G.Bloch).

The interaction of erythrocyte carbonic anhydrasewith Ca++, K+, acetazolamide, digoxin, and cortico-

1373


steroids was studied in vitro and in vivo. The additionof Ca++ or K+ to normal heparinized blood resulted ina 21% increase in enzyme activity. Mg++ increased theenzyme activity only in the blood of patients withhypoparathyroidism. The bloods obtained from two-thirds of normal individuals had an average enzymeactivity increase of 20% when incubated with digoxin.Acetazolamide (0.05 mg/ml) inhibited the activity 30%.The intravenous infusion of Ca++ over a 4-hour periodto patients with osteoporosis resulted in a 20% increasein carbonic anhydrase activity. Hypoparathyroid pa-tients showed an earlier and greater rise, whereaspatients with hypercalcemia had only a slight (or no)increase in their enzyme activity. The blood of pa-tients receiving corticosteroids or with Cushing's syn-drome did not show an increase in their erythrocytecarbonic anhydrase activity with either Ca++ or K+.Neither did the intravenous infusion of Ca++ to thoseindividuals result in an increase in activity. The incu-bation of hemolyzed red cells with K+ and digoxinresulted in no change in enzyme activity, whereas Ca++increased the activity approximately 13%, and acetazola-mide caused an even greater inhibition. Thus the intacterythrocyte is required to obtain the enhanced effectwith K+ or digoxin. The inhibition due to acetazola-mide could be prevented partially or completely by theaddition of Ca++, K+ or digoxin to whole blood butnot when added to hemolyzed erythrocytes. Advantagewas taken of the antagonism between digoxin and aceta-zolamide in diagnosing digitalis toxicity. The inhibitionof erythrocyte carbonic anhydrase was markedly de-creased or absent at acetazolamide concentrations of 0.05or 0.10 mg per ml when patients were digitalis-toxic.A digitalized but not digitalis-toxic individual exhibitedenzyme inhibition at the higher acetazolamide concen-tration; however, the inhibition might be absent or de-creased at the lower concentration. These studiescorrelate the interaction of digitalis, acetazolamide, andcorticosteroid therapy and abnormal Ca++ and K+ metabo-lism with erythrocyte carbonic anhydrase activity.

Abnormal Rise of Plasma Amino Acids and the Appear-ance of a Polypeptide in the Plasma in Adult CeliacDisease after Oral Administration of Gliadin. 0.DHODANANDKOWLESSAR, ELLIOT WESER AND LOR-RAINE J. HAEFFNER, New York, N. Y. (introducedby Marvin H. Sleisenger).The concentration of amino acids in the ultrafiltrates

of plasma before and at 2 and 3 hours after administra-tion of gliadin (350 mg/kg) was determined in sixnormal subjects and ten patients with adult celiac dis-ease (nontropical sprue). The amino acids were sepa-rated either by bidimensional chromatography-butanol:acetic acid :water (120:30:50), followed by butanol:pyridine: water (70:70: 70)-or by high voltage elec-trophoresis (HVE) at pH 2.2 (4% vol/vol formicacid; 0.3% pyridine) and 2,000 v for 70 minutes fol-lowed by chromatography in butanol: acetic acid: water(120:30:50). Chromatograms were stained with nin-

hydrin and the amino acids were eluted and measuredcolorimetrically. The results of these studies demon-strated: 1) that patients with adult celiac disease havelow fasting plasma amino acid levels; 2) concentrationof all amino acids in the plasma of both patients andnormal subjects increased after feeding gliadin, withthe peak level at 2 hours; 3) the percentage of increaseof levels at 2 and 3 hours after gliadin was significantlyhigher in the patients; and 4) unlike the normal sub-jects, the concentration of amino acids at 3 hours inthe patients was still greater than the fasting sample.After gliadin loading, the plasma ultrafiltrates werealso hydrolyzed with hog intestinal peptidase, and theconcentration of amino acids was redetermined. Thelevels of glutamine, glutamic acid, leucine, isoleucine,proline, valine, lysine, and arginine increased by morethan 2 mg per 100 ml in the patients, while only leucineand arginine rose in the six normal subjects. Further,a ninhydrin-staining spot appeared after HVE andchromatography, which yielded glycine, serine, threo-nine, lysine, valine, phenylalanine, leucine or isoleu-cine, or both, and glutamic acid after acid hydrolysis.These findings strongly suggest that patients with adultceliac disease absorb peptides when gliadin is fed. Thisphenomenon indicates a deficiency or absence in thesepatients of intestinal peptidases which are required forthe breakdown of gliadin polypeptides into amino acids.

Isoproterenol and Myocardial Function in Man. NOR-MAN KRASNOW, WILLIAM B. HOOD, JR., ELLIS R.ROLETT, PETER M. YURCHAKAND RICHARD GORLIN,*Boston, Mass.

The effect of catecholamines on myocardial energeticshas been infrequently studied in man. Ten subjects withheart disease, not in clinical failure, were studied duringcoronary sinus catheterization, with measurement ofcoronary flow (CF), myocardial oxygen consumption(Qo2), and left ventricular ejection; and pressure dy-namics at rest, during exercise, and during isoproterenolinfusion (2 to 3 ,ug/min, i.v.). Comparable meanrises were achieved during exercise or isoproterenolinfusion in cardiac and stroke index, heart rate, andsystolic mean pressure. Compared with rest, both statesinduced a fall in systolic ejection period and rise inmean systolic ejection rate (SER), somewhat greaterwith isoproterenol than with exercise. Systolic timeper minute was unchanged with isoproterenol but in-creased on exercise. Pressure-time per beat (PTB)increased on exercise, but decreased during isoproterenol.Decreases in PTB were associated with rises in SER.Isoproterenol caused coronary vasodilatation: increasedCF with decreased coronary A-V 02 difference,A-V/A 02, and rise in coronary venous oxygen (+1.2vol %o). Such changes did not occur with exercise.Contractility was increased so that normal or increasedstroke volume was delivered in a greatly shortenedejection time. Increased mean ejection velocity impliesincreased rate of fiber shortening. The chronotropicaction of the drug, contrary to exercise, did not alter

1374


total systolic or diastolic time per minute. MyocardialQo2 increased proportionately to pressure-time per min-ute (PTM) on drug and exercise. The unchangedPTM/Qo2 at this dosage of isoproterenol can be bestexplained by reduction in heart size (Rushmer) therebyreducing total wall tension (T = rr'P) and counter-balancing any "energy-wasting" effects of catechols.Isoproterenol does not simulate exercise in man. Thisdose of drug has primary as well as secondary vaso-dilator actions on coronary vessels, and enhances con-tractility and ejection velocity without increased myo-cardial energy cost. The reciprocal rise in SER andfall in PTB suggest a crude analogy to the force-velocity relation of isolated muscle.

Simultaneous Determination of Cardiac Output andRenal Hemodynamics in Decompensated Laennec's Cir-rhosis. RUBEN G. LANCESTREMERE,PAUL L. DAVID-SON, LAURENCE E. EARLEY, FREDERICK J. O'BRIENAND SOLOMONPAPPER,* Boston, Mass. and Richmond,Va.

Cardiac output (CO), measured by T-1824 dye or

by radioiodinated albumin dilution methods, and inulinand para-aminohippurate clearances (C1., CPAH), were

determined simultaneously in 20 patients with decom-pensated Laennec's cirrhosis. C1n ranged from 2 to225 ml per minute; CPAH ranged from 8 to 875 ml per

minute. The renal fraction of the CO was less than14%o in all cases. Seven cirrhotics in renal failure(Ci. less than 60 ml/min) had CO greater than 7.0L per minute. One patient studied on two occasions,had a Ci. of 114 ml per minute and a CO of 6.9 L per

minute in the first admission; 7 months later, in histerminal hospitalization, he had a CI. of 9 ml per

minute and a CO of 10.0 L per minute. Thus, CO was

generally elevated in patients who exhibited a broadrange of renal hemodynamics. Further, the findingsindicate that the diminished glomerular filtration rateand renal plasma flow found in some patients with cir-rhosis occur in the presence of an increased CO.

Thyrotropic Hormone Action on Carbohydrate Metabo-lism in Thyroid Slices. BERNARDR. LANDAU,* WIL-FRIED MERLEVEDE AND GALEN WEAVER, Cleveland,Ohio.

Beef thyroid slices were incubated with a thyro-tropic hormone (TSH) preparation in the presence ofspecifically labeled sugars. As reported by others, TSHincreased the oxidation of carbons 1 and 6 of glucoseto CO,; this effect was most dramatic in the first halfhour of incubation. From the randomization of carbon 2of glucose-2-C"4, as reflected in the distribution of radio-activity in glycogen, it is estimated that, of the totalquantity of glucose metabolized by the thyroid slices,approximately 4% was metabolized via the pentose cyclein the absence, and 12% in the presence, of TSH. Thisestimation depends upon isotopic equilibration betweenglucose-6-phosphate and fructose-6-phosphate. Support

for this assumption was obtained from the C-1 to C-6ratio of CO2 yields on incubation of fructose-l-C1 whencompared with that on incubation of glucose-i-C14 and6-C'4. The increase in pentose cycle activity on TSHaddition does not appear attributable to increased avail-ability of intracellular glucose, for while increasing theconcentration of glucose in the medium did result inincreases in the quantity of CO, from carbons 1 and6, the C-1 to C-6 ratio was less than that produced byTSH. TSH did cause a decrease in the incorporationof glucose into glycogen. While glucagon and insulinwere without effect, epinephrine also stimulated the oxi-dation of glucose to CO,. The possibility remains thatsome or all of the effects of TSH on the thyroid may beconsequent to an initial action on glycogen.

Comparison of Two Applications of the Stewart-Hamilton(S-H) Principle for Measurement of Central BloodVolume. RAMONL. LANGE AND HANS H. HECHT,*Milwaukee, Wis. and Salt Lake City, Utah.

In calculating subsegments of the "central blood vol-ume" (CBV), two theoretically equivalent methods formeasurement of mean circulation time (MCT) gave con-sistently divergent results. This led to systematic compari-son of CBV obtained by these two methods in 16humans with and without heart disease. The traditionalS-H method involves slug injection of indicator into thepulmonary artery (PA) and obtaining a single dilutioncurve from the femoral artery (FA) for measurementof cardiac output (Q) and MCT. Then CBV= QXMCT. In the second method, densitometer and (as afurther check) thermistor curves were obtained fromthe PA simultaneously with densitometer curves fromFA, after a rapid venous injection of cold dye. TheS-H principle is now expressed as CBV= QX AMCT,where AMCT is the difference between MCT of thePA and FA curves. Appropriate corrections were madefor the delay of densitometer curves due to samplingdead space. The traditional method consistently yieldedlonger MCT and, therefore, CBV that averaged 380ml or 31io larger than by the double-curve method(p < 0.001). It appears that PA slug injection (theoriginal S-H method) does not provide for efficientpulmonary mixing and that CBV is overestimated. PAthermistor curves recorded with PA injection (methodI) support this contention.

Angiotensin II and Renal Sodium Transport: oatri-uresis and Diuresis in Patients with Cirrhosis andAscites. JOHN H. LARAGH,* PAUL J. CANNON, RICH-ARD P. AMES, ALFRED M. SICINSKI, CARL J. BENTZELAND JAY I. MELTZER, New York, N. Y.

To study further a possible renal-adrenal interaction(angiotensin-aldosterone) for regulating sodium balance,the influences of angiotensin, compared with equipressoramounts of norepinephrine, were examined in normalsubjects and in a representative type of secondary hyper-aldosteronism (i.e., cirrhosis with ascites). Effects on

1375


aldosterone secretion, arterial pressure, electrolyte bal-ance, and renal clearances were observed during condi-tions of controlled metabolism. During continuousinfusions for up to 10 days, angiotensin stimulated aldo-sterone secretion in seven normal subjects (24 studies),but did not augment oversecretion of ten cirrhotics (24studies). In cirrhosis there was markedly reducedpressor responsiveness to angiotensin only, indicating thatincreased endogenous angiotensin could account for thishyperaldosteronism without causing hypertension. Innormals, both angiotensin and norepinephrine initiallyproduced salt and water retention, then re-establishmentof balance, In cirrhosis, both agents caused natriuresisand diuresis. Angiotensin caused more sodium retentionin normals and more natriuresis in cirrhosis; norepi-nephrine promoted less water retention in normals andmore free water excretion in cirrhosis. The natriuresisproduced by angiotensin in cirrhosis was marked andsustained. Urinary sodium excretion rose from valuesof less than 1 to over 500 mEq per day, leading toclearing of edema fluid. Effects were much greaterin magnitude than could be achieved with potent diureticagents. In ten clearances in cirrhosis, angiotensin in-creased mean sodium excretion by 1,367 IAEq per minute.CN./Cln increased from 0.2 to 10.9%o. Norepinephrineproduced lesser saliuresis of 63 /AEq per minute, withincreased GFR, and no change in CNa/CIn. Angiotensinthus markedly inhibited tubular reabsorption of sodiumchloride. Effects on Co.m and CHZo indicated that bothproximal and distal tubular transport processes weredepressed. The results suggest an intrarenal role forangiotensin in regulating sodium reabsorption. Themagnitude of the changes indicates that such an actionof angiotensin could be more important than its previ-ously described effect on aldosterone release.

In Vitro Endocrine Stimulation of Active Transport inthe Small Intestine of Hamster. LEONARD LASTERAND DAVID DANOFF, Bethesda, Md. (introduced byDonald S. Fredrickson).

Animal studies in vivo by Althausen indicated that"thyroid hormone has a direct stimulating effect on thepreferential intestinal absorption" of glucose, but notof alanine; that adrenocortical hormones do not exerta direct effect on intestinal absorption; and that hyper-thyroidism in man increases glucose absorption. Aspart of a systematic investigation of in vitro endocrineeffects on small-intestine function and metabolism, wehave demonstrated that 3,5,3'-L-triiodothyronine (T8)and L-thyroxine (T4) increase the rate of transportagainst a concentration gradient of D-glucose (metabo-lized) and 3-deoxy-D-glucose (minimally metabolized)by surviving everted segments of small intestine ofhamster. Glucose transport was not altered by 3,5-L-diiodotyrosine (I0r and 10' M). T, doubled the ab-sorption rate of L-alanine. Cortisol (10' M) ap-peared to triple the D-glucose absorption rate. Anindividual segment was studied during consecutive pe-riods with hormone absent, present, and again absent

from solution bathing mucosal surface, thus providingexperimental and control data from a single piece oftissue. Glucose transport was maximally stimulated,twofold by T8 (10 M) and three- to fourfold by T4(2.5 X 10' M), when the initial concentration of glucosein the solutions bathing mucosa and serosa was 2 X108 M. At lower (0.5 X 10 M) and higher (5 X10' M) initial glucose concentrations, stimulation wasminimal. After removal of an intestinal segment froma T.-containing medium to a T,-free medium, transportrates returned to basal levels, but T4-stimulated seg-ments continued to transport glucose at the acceleratedrate after exposure to T4 was terminated. T4 in solutionbathing serosa produced less stimulation of transport thandid T, in solution bathing mucosa. The observed hor-monal effects cannot be attributed to alterations in watertransport. Intestinal segments from animals 2 weeksafter thyroidectomy were not strikingly more sensitiveto T, than were segments from normal animals, althoughthe basal rate of glucose transport was lower in theoperated group.

Regulatory Mechanisms of Leukocyte Metabolism. JOHNLASZLO, Durham, N. C. (introduced by R. WayneRundles).

There are fundamental differences in the regulation ofenergy-yielding metabolism of lymphocytic and granulo-cytic normal and leukemic leukocytes. Plasma suspen-sions of leukocytes obtained after fibrinogen sedimenta-tion were studied in manometer vessels for 3 hours.Lymphocytes were characterized by high respiratoryrate (Qo2 ca 5), high anaerobic glycolysis (QN2' ca 25),and virtually absent aerobic glycolysis, indicating astrong Pasteur mechanism. Granulocytes have a slightlylower respiration (Qo2 ca 4), high aerobic glycolysis(Qo2_ ca 10), and high anaerobic glycolysis (QN2A ca 25)showing a, weaker Pasteur mechanism. Lymphocyticrespiration was not influenced by the presence or absenceof glucose, whereas glucose metabolism inhibited granu-locyte respiration. These different regulatory mecha-nisms carry over to lymphoblasts and myeloblasts andare helpful laboratory aids in differentiating types ofacute leukemia. High energy phosphate production(Seits method) was determined in leukocytes undervarying metabolic conditions. There was no inhibitionof ATP-ADP when normal and leukemic lymphocytesand granulocytes were incubated in glucose-free plasmafor 3 hours. There was an average of 14% inhibition(O to 47%) when leukocytes were incubated under an-aerobic conditions, with no detectable differences betweenlymphocytes and granulocytes. Combined anaerobiosisand glucose deprivation caused total ATP-ADP inhibi-tion. These studies demonstrate the remarkable adapta-bility of lymphocytes as well as granulocytes to anaero-biosis or glucose deprivation despite their metabolicdifferences under "normal" conditions. Preliminary sub-cellular leukocyte studies indicate that these metabolicregulatory mechanisms may operate at the mitochondriallevel.

1376

AMJtRICAN SOCIETY FOR CLINICAL INVgSTIGATION

The Kinetics of Renal Urate Transport. WILLOUGHBYLATHEM* AND GERALD P. RODNAN, Pittsburgh, Pa.

An analysis of the kinetics of renal urate transportunder conditions of intravenous urate loading was madein normal subjects and in patients with primary gout.At increasing filtered loads, urate excretion rose in acurvilinear manner although, as previously reported,gouty subjects excreted less urate at corresponding loadsthan did normal subjects. The rate of reabsorption as-sumed as the difference between load and excretion,a curvilinear reabsorptive curve (reabsorption plottedagainst load) was found in gouty subjects. This curvehad the following expression, which is consistent witha carrier-substrate reabsorptive process: T = TmaxC/(Kt + C), where T = reabsorptive rate, Tmax = maximalreabsorptive rate, C = urate concentration in filtrate, andK, =a constant. When corrected for changes in initial(filtrate) concentration: F = (l/f)ln[l/(l-f) ], whereF = correction factor and f = the fraction of urate re-absorbed, Tma. averaged 30.0 mg per minute and K,7.0 mg per 100 ml. In contrast to gouty subjects thereabsorptive curve in normal subjects appeared to involvetwo rate-limiting steps: one governing reabsorption atrates below Tm, the other that determining Tm. Thefirst of these was qualitatively similar to that foundin gout, but Tmax and K, were less, averaging 16.5 mgper minute and 4.4 mg per 100 ml, respectively. The ad-ministration of uricosuric agents altered the characteristicsof the curves in both normal and gouty subjects. Innormal subj ects reabsorption was linear rather thancurvilinear. This response, which argues against asecretory mechanism, is consistent with complete inhibi-tion of active reabsorption and suggests reabsorptionunder these conditions by back diffusion. In goutysubj ects reabsorption remained curvilinear, suggestingpartial rather than complete inhibition of active reab-sorption, and a Tm, not ordinarily found, was demon-strated. These results indicate that at least two rate-limiting steps are involved in urate reabsorption and thaturicosuric agents inhibit both.

Cytoplasmic Localization of Macroglobulin in Lympho-cytoid Marrow Cells in Macroglobulinemia of Wald-enstrom. PHILIP LEDER AND MUHARREMG6KCEN,Minneapolis, Minn. (introduced by C. J. Watson).

Small, lymphocytoid round cells have been repeatedlyobserved in the bone marrow of patients with macro-globulinemia of Waldenstrom. This cell-type has beenthought to be associated with the production of serummacroglobulin, but previous reports have failed to pro-duce consistent results regarding the localization ofcellular macroglobulin antigenic activity. To resolvethis difficulty, antiserum to purified macroglobulin, iso-lated from a patient with macroglobulinemia of Walden-strom, was produced and labeled with fluoresceinisothiocyanate and utilized according to Coons' technique.The antiserum was completely adsorbed with purifiedhuman y-globulin, removing cross-reacting components

and leaving specific antimacroglobulin activity. Fixed,concentrated bone marrow smears obtained from a macro-globulinemic patient were incubated with labeled anti-macroglobulin serum, both with and without previousblocking incubation with unlabeled antiserum. Otherswere incubated with labeled antihuman y-globulin. DL-penicillamine-treated smears were also incubated withlabeled antimacroglobulin. Striking fluoresence wasnoted in the cytoplasm of the lymphocytoid cells, both inthe antimacroglobulin and antihuman y-globulin-incu-bated smears, but most prominently in the former.Previous incubation with unlabeled antimacroglobulineffectively blocked the uptake of the labeled antimacro-globulin, confirming the specificity of the reaction.DL-penicillamine destroyed the antigenic property of thecytoplasmic macroglobulin in accordance with the an-ticipated effect on the macromolecule. Bone marrowsmears from two additional patients with macroglobu-linemia were also incubated with labeled antimacro-globulin and also showed (weak) cytoplasmic fluo-resence of the lymphocytoid cells. Thus, the cytoplasmof the lymphocytoid cell probably represents storage,and possibly the site of synthesis, of this protein. Inaddition, the macroglobulinemic serum was tested forthe presence of antinuclear and anticytoplasmic factors,with the use of peripheral blood, liver, and spleen asantigenic sources, and no activity was found.

The Effect of Osmolality and Sodium Concentration onthe Metabolism of Glucose-U-C1' by Rabbit KidneyCortex and Medulla In Vitro. JAMES B. LEE, VERNONK. VANCEAND GEORGEF. CAHILL, JR.,* Boston, Mass.

Establishment of osmotic gradients in kidney tissueappears fundamental to certain solute and water move-ments. Since maintenance of gradients requires energy,the present study was designed to investigate individualeffects of osmolality and sodium chloride on glucosemetabolism in cortex and medulla. Slices of kidneycortex and medulla of rabbit were incubated for 90minutes at 380 C in Krebs-Ringer-HCO3- buffer con-taining glucose-U-C14, 10 mM. Variations in osmolalityof the media were obtained by adding NaCl, sucrose,or urea to NaCl-free buffer. Measurements were madeof labeled glucose incorporation into CO2, glycogen, andfatty acids, as well as of lactate production, glucose dis-appearance, and levels of glycogen and fatty acids. Whenmedium osmolality was changed with sodium chlorideor sucrose, maximal cortical glucose oxidation, dis-appearance, and lactate production occurred between 67and 171 mOsmper kg H2O. Further osmolality in-creases resulted in progressive inhibition of glucosemetabolism to 50% of maximum at 360 mOsmper kgH2O. A secondary rise in glucose oxidation occurredat 550 mOsmper kg H20 before marked inhibition at1,000 mOsmper kg H20. In contrast, glucose oxidationby medulla was unchanged even at 1,000 mOsmper kgH2O. Urea-induced changes in osmolality had littleeffect on cortical or medullary metabolism. With me-dium osmolality constant (281 mOsm per kg H20),

1377


[Na+] was increased from 0 to 80 mmoles per L, pro-ducing a rise in glucose oxidation by cortex, withcharacteristics of a first-order reaction (half maximalglucose oxidation at [Na+] of 25 mEq per L). Inmedulla there was no effect on glucose oxidation by[Na+], although increasing [K+] depressed medullaryglucose metabolism significantly. These results suggestthat both "effective" extracellular osmolality and sodiumconcentration exert important independent effects onglucose oxidation in kidney cortex, but not medulla.

Reticuloendothelial Clearance of Circulating Fibrin inthe Pathogenesis of the Generalized Shwartzman Re-action. LEUNGLEE, New York, N. Y. (introduced byChandler A. Stetson).The formation of fibrin polymers in the circulating

blood during sublethal intravenous infusions of thrombinor thromboplastin in normal animals usually producesno detrimental effects, since the fibrin formed rarelyaccumulates in massive deposits in the peripheral vas-cular bed. The protective mechanism involved in theremoval of fibrin in these circumstances has generallybeen considered to be the fibrinolytic system, and inman and some other species active fibrinolysis may in-deed account for the uncommon occurrence of vasculo-occlusive lesions complicating generalized coagulationdisturbances. It has been found, however, that thereticuloendothelial system of the liver and spleen mayalso function as an efficient homeostatic mechanism inthe clearance of circulating fibrin polymers and may beof particular significance in the rabbit in which activefibrinolysis during intravascular clotting cannot bedemonstrated. This concept is supported by the findingin rabbits that interference with the phagocytic activityof the reticuloendothelial system during continuous intra-venous thrombin infusions results in massive fibrindeposition in glomerular capillaries with ensuing bilateralrenal cortical necrosis. The kidneys and other visceralorgans exhibit intravascular "fibrinoid" and tissue ne-crosis indistinguishable from the changes consideredpathognomonic of the generalized Shwartzman reactionelicited by bacterial endotoxins. Furthermore, severalexperimental situations have promoted fibrin accumula-tion after either a single injection of endotoxin or aninfusion of thrombin, indicating that similar mechanismsare involved in the action of these two substances. Fromthese and other experiments to be described, it is con-cluded that bacterial endotoxins are potent activatorsof the coagulation mechanism and that the generalizedShwartzman reaction in the rabbit results from the fail-ure of the reticuloendothelial system to adequately clearthe circulation of fibrin formed during intravascularcoagulation.

Observations on Hepatic DNASynthesis in Man. CAR-ROLL M. LEEvY, Jersey City, N. J. (introduced byHarold Jeghers).An in vitro radioautographic technique was used to

investigate DNA synthesis in percutaneous liver biopsies

from 38 normal subjects and 48 patients with liver dis-ease (17 fatty liver, 5 viral hepatitis, 4 metastatic ma-lignancy, 4 postnecrotic cirrhosis, 18 Laennec's cirrhosis).Serial studies were obtained in selected subjects and theresults correlated with clinical findings, liver functiontests, and routine histopathological studies. Specimensimmersed in 3 ml of Ringer's solution containing 50 lucof tritiated thymidine (specific activity 1.9) and 75 Uof hyaluronidase were incubated with constant agitationat 370 C for 1 hour in a 5% C02, 95 0°2 atmosphere.Radioautographs were prepared from paraffin sectionswith NTB3 emulsion and Kodak D-19 developer andstandard techniques. Preliminary rat studies demon-strated that in vitro labeling was equivalent to that ob-tained in vivo. Specificity of labeling was confirmed withDNAase. Temperature, pH, and length of incubationinfluenced labeling; addition of ATP and colchicine hadno effect. Normal liver and uncomplicated fatty liverwere characterized by approximately 1 labeled cell per10,000 cells; a maximal increase in labeling of 250-foldwas noted in malignancy, 50-fold in postnecrotic cir-rhosis, 20-fold in acute viral hepatitis, and 10-fold inLaennec's cirrhosis. DNA synthesis was related to thedegree of necrosis and inflammation, persisting withactivity in acute lesions and returning to normal ininactive cirrhosis. In Laennec's and postnecrotic cir-rhosis increased labeling was noted in both parenchymaland connective tissue cells. These results establish thefeasibility of assessing hepatic DNA synthesis in man,demonstrate that both connective tissue proliferationand parenchymal cell regeneration contribute to devel-opment of human cirrhosis, and provide a baseline forevaluating effects of medicinal agents.

A Comparison of Direct-current and Alternating-currentDefibrillation, under Conditions of Hypothermia.ARMANDA. LEFEMINE, RAGHAVANAMARASINGEAM,DWIGHT E. HARKEN, BAROUGHBERKOVITZ AND BER-NARD LOWN, Boston, Mass. (introduced by FredrickStare).Ventricular fibrillation is a frequent occurrence during

open-heart operations. The use of alternating current(AC) for reversion may require multiple shocks thatare associated with myocardial damage. At times de-fibrillation is impossible. A direct-current (DC) defi-brillator has been consistently effective in defibrillatingthe normothermic animal across the intact chest. Toevaluate DC defibrillation in closed-chest dogs, compara-tive studies with AC were carried out at esophagealtemperature levels ranging from 370 to 200 C. Coolingwas achieved by an extracorporeal system. Ventricularfibrillation was induced by single monophasic capacitordischarges synchronized to occur during the vulnerablephase of the ventricular cycle. Both types of counter-shock were compared under analogous conditions oftemperature, pH, Pco2, and electrolyte concentrations.DC defibrillation was tried in 143 episodes and wassuccessful in all but 3 (2o) ; these failures were notedin a single animal at 200 C. A total of 258 DC shocks

1378


was required, of which 154 (60%) were effective. Theenergy levels ranged from 40 to 400 watt-seconds. ACdefibrillation was tried in 91 episodes and was unsuc-cessful in 16 (18%o). A total of 267 shocks was givenof which 77 (29%) were successful. Voltages in therange of 500 to 1,000 were employed. AC defibrilllationbecame increasingly difficult as the esophageal tempera-ture fell below 30° C and was followed frequently byatrial and ventricular arrhythmias. Such arrhythmiasrarely followed DC countershock. Subcutaneous andintramyocardial temperatures were recorded duringcountershock. Rises of 2 to 4° C occurred after singleAC discharges. No temperature changes followed theuse of DC. Three clinical experiences during open-heart surgery confirmed the superiority of DC observedexperimentally. The first case was defibrillated readilywith DC when ten AC attempts had failed. Alternatingcurrent defibrillation has been abandoned for use inthe operating room.

Infections with Reozirus Type 2. A. MARTIN LERNER,JAMES D. CHERRY, JEROME0. KLEIN AND MAXWELLFINLAND,* Boston, Mass.

In the course of viral studies of febrile respiratoryand exanthematous illnesses, rises in hemagglutinin-inhibiting antibodies (HIA) to type 2 reovirus weredemonstrated in eight children observed in 1960; seven ofthem had a morbilliform rash but most of them hadno respiratory or enteric symptoms. The virus wasisolated from three patients-four times during the first4 days of illness in one of them. Among matched con-trols without illness or in children without illnesses,only one yielded this virus. Constantly elevated orfalling titers of HIA to type 2 reovirus were demon-strated in occasional asymptomatic children and in morethan one-fifth of those with exanthems or lower respira-tory infections considered to be due to other agents.In three of the patients with reovirus type 2 infection,the HIA titers obtained with prototype strains of types1 and 3 reovirus were also elevated in sera obtainedduring the first weeks after infection, but only the type2 HIA remained elevated in sera obtained 1 year later.About one-fourth of a small group of patients studiedin 1959 and of similar ones observed in 1961 were foundto have HIA to reovirus type 2. About one-third ofa larger number of patients studied over the 3-year pe-riod has serological evidence of infection with reoviruses,approximately one-half of them due to type 2. Theevidence suggested placental transmission with loss ofantibody within the first 6 months and reacquisition inabout half of the children before 10 years of age. Thedata also suggest that the type 2 reovirus may be asso-ciated with a rubelliform exanthem.

The Mechanism of Intestinal Absorption of Bilirubin.ROGERLESTER AND RUDI SCHMID,* Boston, Mass.

It has been shown recently that, in rats, intraduodenaladministration of either free or conjugated C14-bilirubinresults in biliary re-excretion of conjugated labeled pig-

ment corresponding to 10 to 30% of the administereddose. This indicates that bilirubin undergoes a signifi-cant enterohepatic circulation. The present studies wereperformed to determine whether both free and conju-gated pigments are absorbed from the gastrointestinaltract unaltered, and whether absorption occurs in hu-mans. Congenitally jaundiced (Gunn) rats retain freebilirubin but excrete intravenously administered conju-gated bilirubin in bile. In Gunn rats, intraduodenalinstillation of either free or conjugated C14-bilirubinresulted in significant radioactivity in free bilirubincrystallized from serum. Moreover, after intraduodenalinstillation of conjugated C14-bilirubin, virtually no radio-active bilirubin was recovered in bile. Estimates of thespecific activity and total miscible pool of free bilirubinrevealed that total intestinal absorption approximated15% of the dose irrespective of whether free or con-jugated pigment was administered. One and one-half/Ac of free C14-bilirubin was administered by duodenaltube to a child with congenital unconjugated hyper-bilirubinemia. From the specific activity of the serumbilirubin and the total miscible bilirubin pool, it wasestimated that approximately 80% of the administereddose was absorbed and retained by the subject at 24hours. A similar experiment was performed in a mildlyicteric woman in whom external biliary drainage hadbeen instituted; 15% of the intraduodenally administeredfree C14-bilirubin appeared as conjugated C14-bilirubinin bile collected over an 8-hour period. These studiesindicate that free bilirubin is absorbed by the gastro-intestinal tract of rats and humans. In contrast, littleif any conjugated bilirubin is absorbed unaltered, but asignificant fraction is hydrolyzed and absorbed as freebilirubin.

Analog Computer Analysis of Diffusing Capacity Meas-urements. B. M. LEWIS,* P. K. ADHIKARI, S. F.BOUSHYAND A. SAKAMOTO, Detroit, Mich.

This work is an attempt to apply modern techniquesof high-speed computing to the determination of diffusingcapacities of regions of the lung and from this toassess the accuracy with which over-all diffusing capacitycan be measured by methods currently in use. Sincethe accuracy of such methods is disputed in emphysema,patients with this disease were used as subjects for thisstudy. These patients rebreathed a mixture of neonand carbon monoxide in air for a 45-second period.Neon concentration was analyzed continuously by amass spectrometer, and the contents of the rebreathingbag were circulated through an infrared CO analyzer.Tidal volume was measured with a pneumotachograph.Residual volume and dead space were measured inde-endently. An analog computer was used to simulatea lung of two regions and a rebreathing bag. Theresidual volume and tidal volume of each region werethen varied until the mixing curve of the simulated lungmatched the mixing curve of neon in the patient. Asimulated diffusing capacity was then added to eachregion and these regional diffusing capacities altered until

1379


the patient's CO absorption curve was matched. In twopatients with slow mixing and a low diffusing capacitycalculated from the CO absorption curve ("calculated"DL), the sum of the diffusing capacities of the two regions("true" DL) agreed closely with the "calculated" DL.In two others with slow mixing and a normal calculatedDL, "true" DL exceeded calculated DL. In one patientwith moderate impairment of mixing and a normal cal-culated DL, true DL was slightly lower than calculatedDL.

Effects of Ethanol Metabolism on Serum Concentrationand Urinary Excretion of Urate in Man. CHARLESS.LIEBER, DON P. JONES, MONTY S. LosowSKY ANDCHARLES S. DAVIDSON,* Boston, Mass.

The traditional association between gout and alco-holism led us to test whether a relationship existsbetween ethanol and urate metabolism. Serum uratewas measured by the uricase method in 12 acutely intoxi-cated subjects. In 5, initial concentrations varied from9.7 to 11.5 mg per 100 ml, while in the others theywere below the upper limit of normal (7.5). Urateconcentrations decreased in each case to below 6.8 mgper 100 ml during recovery. The average decrease was38% (p <0.001). Ethanol administration (intravenous,2 subjects, 8 to 12 hours; oral, 5 subjects, 8 to 72 hours)was accompanied in each instance by a rise in serumurate (1 to 3.2 mg/100 ml), which returned to normalafter omission of ethanol. Blood ethanol concentrationsreached 200 to 340 mg per 100 ml. Glucose in iso-caloric amounts (oral) or saline (intravenous) givento two of these subjects did not change serum urateconcentrations. In three instances in which urinarystudies were performed, clearances and outputs of uratedecreased by 50 to 85% during 24 hours of ethanoladministration, and returned to control values there-after, while creatinine clearances remained unchanged.This deficit in urate output could account for the ob-served rise in serum urate. During ethanol adminis-tration, blood lactate rose to 16 to 23 mg per 100 ml;and similar levels were found in all 8 acutely intoxicatedsubjects (blood ethanol 311 ± 16.8 mg/100 ml) in whomsuch measurements were made. Sodium lactate ingestionon two occasions by one of the subjects formerly givenethanol achieved blood lactate concentrations, decreasesin urate excretion, and increases in serum urate similarto those after ethanol, while urinary pH changed in theopposite direction. Conclusions: 1) high serum uratelevels occur in acutely intoxicated subjects; 2) alchoholincreases serum urate; 3) this effect is mediated througha rise in blood lactate that decreases urinary urate.

The Regulation of Intragastric Hydrogen Ion Concen-tration: Evidence Regarding a Na-H ExchangeMechanism in Man. ARTHURE. LINDNER AND NA-THANIEL COHEN, New York, N. Y. (introduced byHenry D. Janowitz).

Intragastric mechanisms that regulate H+ concentra-tion were studied by instilling several isotonic hydrogen-

containing solutions and a dilution-indicator system intothe stomachs of 17 human subjects. Solution I containedH' and Cl- (160 mEq/L). Solution II contained H+(100 mEq/L), Na+ (60 mEq/L) and Cl- (160 mEq/L).Osmolarity and electrolyte concentrations were deter-mined in samples recovered at 15-minute intervals over1-hour periods. With the aid of the dilution indicator(phenol red) the volume of fluid added by secretion aswell as that lost through the pylorus by emptying couldeasily be calculated. Similar results were obtained withboth solutions. In all experiments H+ declined linearlywith time. The fall in hydrogen concentration wasassociated with a parallel decline in osmolarity andwith a reciprocal (straight line) rise in sodium concen-tration. Potassium in increasing concentrations en-tered the gastric fluid progressively. Cl- concentrationdeclined moderately. The electrolyte composition ofthe fluid entering and leaving the stomach was calcu-lated from the observed concentrations. The resultsthus obtained suggest that several mechanisms otherthan emptying are responsible for the lowering of H+concentration of instilled (and presumably secreted)acid: 1) entry of a fluid free of H+ and containing HCOS-(so-called alkaline component) with consequent fall inosmolarity by evolution of C02; and 2) an exchangemechanism (Na for H) is required to account com-pletely for the observed reduction in H+ concentrationin nearly one-third of the experimental periods. Theentry of K (as in our previously reported studies) isagain best correlated with the rate of gastric secretion.

Cell Proliferation Kinetics in Human GastrointestinalMucosa. MARTIN LIPKIN, PAUL SHERLOCKAND BER-TRANDBELL, New York, N. Y. (introduced by ThomasP. Almy).

Injection of tritiated thymidine into four human sub-jects has provided a means of studying the kinetics ofcell proliferation and migration in normal and abnormalgastrointestinal mucosa. Microradioautographs madefrom mucosal biopsies of inj ected patients have beenanalyzed to measure the total duration and componentportions of the mitotic cycle of the epithelial cells, aswell as the rate of migration and removal of the cells.In normal mucosa of the transverse colon, the percentageof labeled epithelial cells in mitosis reaches a peak at6 to 8 hours. The mean generation time of the epithelialcells is 22 hours. The duration of the DNA synthesis(S) phase is 15 hours, and the premitosis (G2) andmitosis (M) phases together comprise about 2 hours.After administration of 5-fluorouracil, abnormalities inchromatin pattern and size of nuclei are seen. Incor-poration of thymidine is rapid, but the durations of theG2 and Mphases are lengthened to 4 hours, and the peakof labeled mitoses is delayed. In a rectal carcinoma,earlier and heavier labeling is noted in tumor cellscompared with adjacent normal epithelial cells. In thestomach, earliest labeling occurs in the depths of thegastric glands, whence labeled cells migrate to the surfacein 48 hours. Labeling is less marked in zymogen cells;

1380


none was observed in parietal cells. Some epithelialcells in colon and stomach incorporate thymidine butdo not proliferate rapidly, and are retained in the mucosafor prolonged periods. These findings correspond closelyto kinetics of proliferation previously observed in themouse and rat and provide a basis for study of diseasestates in man.

The Effects of Alcohol in Experimental Infections inMice. DONALDB. LOURIA, New York, N. Y. (intro-duced by Robert F. Watson).Certain bacterial infections occur with inordinate fre-

quency in alcoholics. In the present studies, mice weregiven 0.5 ml of ethyl alcohol subcutaneously one-halfhour prior to infection. After intravenous inoculationwith staphylococci, 18 of 81 controls (22%) died in14 days, as did 93 of 142 (65%) which became ataxicbut not stuporous after alcohol, and 88 of 112 (81%)of those developing coma. Kidney staphylococcal popu-lations were similar in all groups at 2, 4, and 6 hoursbut were significantly higher in alcohol-treated comatoseor ataxic mice 24 hours after infection. Peritonealclearance of Escherichia coli and staphylococci wasnormal in alcohol-treated mice if small numbers of micro-organisms were introduced intraperitoneally, but withlarger inocula, clearance was markedly reduced. Popu-lations were similar in all groups during the first hour,but 2 to 6 hours after infection the microbial censusfell over 95% in 74% of controls, in 35% of ataxicmice, and in 13% of those that were stuporous. Quanti-tative measurement of the polymorphonuclear responseshowed that 2 hours after infection only small numbersof polymorphonuclear leukocytes entered the peritonealcavity of alcohol-treated mice and 4 hours after infec-tion, polymorphonuclear counts averaged 58% less inataxic animals, and 88% less in comatose mice, than incontrols. Since results were similar in ataxic and coma-tose animals, coma, hypotension, and shock cannot beimplicated in these experiments. If peritoneal inflam-mation was produced by injection of starch-aleuronatprior to infection, alcohol administration did not alterphagocytosis or intracellular killing significantly, sug-gesting that the enhancement of infection observed inthese studies was related primarily to delay in poly-morphonuclear mobilization.

The Use of Synchronized Direct-current Countershockin the Treatment of Cardiac Arrhythmias. BERNARDLoWN, RAGHAVANAMARASINGHAM, JOSE NEUMANAND BAROUH BERKOVITZ, Boston, Mass. (introducedby Samuel A. Levine).Some cardiac arrhythmias are refractory to all forms

of treatment. In principle, external electric counter-shock that can induce complete depolarization of theheart and permit resumption of sinus activity wouldbe an ideal form of treatment if it were safe. Experi-ments were conducted to determine what form of elec-trical stimulation would be most safe and effective.Nineteen dogs in sinus rhythm received a total of 438

alternating-current countershocks (AC) across the closedchest, with voltages effective for ventricular defibril-lation and with lower levels. There was a high inci-dence of diverse arrhythmias including 78 episodes ofventricular fibrillation (18%). Seven animals diedduring the first week after countershock. AC counter-shock was therefore considered unsafe for elective usein patients. A direct-current defibrillator (DC) deliver-ing a brief monophasic discharge was developed. In 30dogs receiving 858 DC countershocks there was a 1.0%incidence of ventricular fibrillation and no deaths. Thistechnique was consistently effective in defibrillating theheart even when AC failed. In the use of DC, ventricularfibrillation occurred only when the discharge fell duringa predictable brief vulnerable period of the ventricularcycle. To prevent this arrhythmia entirely, a syn-chronizer was developed which permitted the shockto be delivered outside this zone. Thus in 2,600 syn-chronized shocks, ventricular fibrillation never resultedunless the shock was triggered within the vulnerableperiod. Single synchronized DC shocks were appliedacross the closed chest in two patients with ventriculartachycardia resistant to all therapy and four patientswith long-standing atrial fibrillation. Reversion wasinstantly achieved in all except one with atrial fibrilla-tion. There were no complications. This extensiveexperimental and limited clinical experience suggeststhat a new simple and effective method of controllingsome serious arrhythmias may be available.

Hyponatremia in Acute Porphyria Probably Due to In-appropriate Secretion of Antidiuretic Hormone. GEORGED. LUDWIG* AND MARTIN GOLDBERG, Philadelphia, Pa.In the past, it has been noted that hyponatremia fre-

quently accompanies acute attacks of porphyria, but thecause has never been adequately explained. Eight pa-tients displaying the syndrome of hyponatremia, serumhypo-osmolality, natriuresis, and urine hypertonicity werestudied. Serum sodium and osmolality were as lowas 100 mEq per L and 220 mOsmper kg, respectively,in some cases. Edema, azotemia, and dehydration wereabsent. The syndrome probably results from inappro-priate secretion of ADH. Adrenal insufficiency, chronicrenal disease, and "true" sodium depletion were excluded.Central neurological involvement, such as coma, con-vulsions, or acute psychosis, was correlated with themost severe changes in serum sodium and osmolality.In selected patients, restriction of fluids to 800 to 1,000ml per day resulted in dramatic clinical improvement withreversal of serum sodium and osmolality toward normal.In one of these patients, increasing fluid intake to 3,000ml per day produced a sharp fall in serum sodium andosmolality that was accompanied by a convulsive seizure.With subsidence of acute attacks in these patients, theinitial greatly increased excretion of porphyrins andporphyrin-precursors decreased as normal values forserum sodium and osmolality were achieved, and theurine became hypotonic. Oral water loading then failedto reproduce the hyponatremic syndrome. This syn-

1381


drome provides an explanation for the oliguria andpossibly some of the central neurological effects ofacute porphyria. The syndrome apparently results fromcentral effects of porphyria, although exogenous non-osmotic stimuli of ADH secretion, such as barbiturates,analgesics, trauma, and dimercaprol, might have con-tributed in some instances. A new approach to therapyof acute attacks is suggested which differs radically fromthat previously recommended, since dimercaprol andEDTA appear to be contraindicated.

The Effects of Leukocyte Respiration in Leukemia.HAROLD J. LYNCH, JR. AND JOHN LASZLO, Durham,N. C. (introduced by Albert Heyman).

The increased oxygen consumption observed in pa-tients with leukemia is thought to reflect increased leuko-cyte metabolism. This report describes a simple methodfor measuring 02 uptake, CO2 and lactic acid production,and pH changes in blood resulting from leukocytemetabolism. Repeated observations were made on nor-mal controls, patients with leukocytosis of infection, and12 subjects with myelogenous leukemia. Arterial bloodwas collected in a sealed syringe and kept at 37° C.At 5- to 15-minute intervals blood 02 content, Pco2,lactic acid content, and pH were determined on aliquotsof the blood. Total body 02 consumption (TBC 02)and tolerance to local ischemia were also determined.In patients with leukemia, the blood 02 content fell rapidlytmean, 0.14 ml/100 ml/min in blood containing>100,000 white blood cells (WBC)/mm'). This fall wasblocked by cyanide and by cooling to 20 C. The 02content of controls fell less rapidly. The mean 02uptake was 56 Al per 1010 WBCper minute (range,23 to 87) in both leukemic and normal blood. Lacticacid content increased (as much as 1 mg/100 ml/min)while pH fell rapidly and Pco2 rose slightly in bloodcontaining > 100,000 WBCper mm3. The mean TBC 02of the leukemic patients was 212 ml per minute per m2(normal 120 to 160). Respiratory alkalosis was observedin all leukemic subjects as indicated by mean valuesfor pH and Pco2 of 7.48 and 29.3 mmHg, but severeacidosis developed in some patients terminally. Byestimating whole body leukocyte populations in patientswith chronic granulocytic leukemia, the observed 02uptake of leukocytes could account for 17 to 50% ofthe oxygen consumption of the total body. Decreasedtolerance to ischemia was observed in the leukemics,suggesting that relative hypoxia may occur in tissues,owing to competition with leukocytes for oxygen, espe-cially in areas with slowed circulation. The methoddescribed is useful for measuring leukocyte metabolismunder approximately in vivo conditions.

Disappearance Rates of Estrone, Estradiol-17p, and Es-triol from Plasma and Urine after Delivery of thePlacenta in the Human. F. DEBELE MANERAND JOHNR. K. PREEDY,* Atlanta, Ga.

Although the levels of urinary estrogens in the post-partum period have been estimated, observations have

been sparse. Plasma estrogens during this period havebeen little studied, and no correlation has been estab-lished between plasma and urine levels. Accordingly astudy was carried out as follows. Plasma samples weredrawn from normal female subjects at the time of de-livery of the placenta, and at intervals thereafter, togetherwith 4-, 8-, 12-, and 24-hour urine collections. Plasmaand urine samples were analyzed for total estrone,estradiol-17j8, and estriol content. Estradiol-17p dis-appeared from plasma to levels below the sensitivityof the method at a single rapid rate (ti r 0.8 hrs),whereas estrone and estriol each disappeared at twowell defined rates, a faster (ti 2.0 hrs for estrone,ti 1.9 hrs for estriol) for 3 to 5 hours, followed bya slower (ti 7 16.3 hrs for estrone, ti ~ 25.4 hrs forestriol) for longer than 45 hours. Estrone, estradiol-17,3, and estriol each disappeared from the urine attwo rates, a faster (tj i 3.1, 3.06, 3.45 hrs, respectively)for 5 to 10 hours, followed by a slower (ti ; 22.9, 22.9,14.5 hrs, respectively) for more than 80 hours. Since thetransfer of estrogens from the placenta to the maternalcirculation may be regarded as a long-continued infusion,it is reasonable to assume that all pools of distributionare in approximate equilibrium. The two rates of dis-appearance of both estrone and estriol from plasma couldbe accounted for in part by the presence of rapidly andslowly excreted components of the total plasma estroneand estriol, respectively. This possibility is supportedby the finding that, in the case of estriol, the renal clear-ance during the slow disappearance phase is less thanhalf that during the rapid phase. Possible slow mobili-zation of estrogen from stores or compartments otherthan plasma may also affect the rate of disappearance.

Impairment of Cardiovascular Response to Tyramine byReserpine Administration in Man. M. L. MASHFORDANDW. A. MAHON,Boston, Mass. (introduced by ShuChu Shen).Reserpine has not been shown to deplete tissue

norepinephrine in man, although this effect is wellestablished in animals. Since tyramine produces itspressor effect by releasing endogenous norepinephrineat sympathetic nerve endings, the response to tyraminehas been used to demonstrate that reserpine, in thera-peutic doses, depletes tissue norepinephrine. Restingsupine blood pressure was recorded continuously in 12patients; 0.2 mg per kg tyramine was given i.v. afterthe blood pressure had stabilized. Mean pressure wasgraphed for each patient against time, and the areaenclosed between this curve and the preinjection levelwas the statistic used for comparison. In eight patients0.03 mg per kg reserpine was given i.v. after the firststudy, and again 24 hours later, and the observationswere repeated 48 hours after the initial study. In fivepatients no reserpine was given. One patient wasstudied 3 times at 48-hour intervals; reserpine wasgiven between the second and third studies, and he thusfigures in both reserpinized and control groups. Cardiacoutputs (indocyanine green) were performed at intervals

1382


before and after tyramine on both days in 10 of the 12patients, and total peripheral resistance was calculatedfrom these data. The pressor response to tyramine wassignificantly less on the second occasion in patientsreceiving reserpine (mean difference in area, -888;p < 0.01) in contrast to the control patients whoseresponse was not significantly different on the twooccasions (mean difference in area, +84). Cardiacoutput changes after tyramine were variable and werenot significantly different after reserpine. In the reser-pinized patients the increase in total peripheral resistanceproduced by tyramine was significantly smaller on thesecond occasion (mean difference, 798 dynes-cm-secd;p < 0.025). Thus, reserpine has been shown to impaircardiovascular responses which are known to depend uponintact norepinephrine stores.

Pseudofeedback Inhibition of Purine Synthesis by 6-Mer-captopurine and Other Purine Analogues. ROBERTJ.MCCOLLISTER, WALTERR. GILBERT, JR. AND JAMES B.WYNGAARDEN,*Durham, N. C.

6-Mercaptopurine exerts a wide spectrum of anti-metabolic effects, but these are not fully understood.6-MP is sparingly incorporated into DNA and RNA.6-MP ribonucleotide is an inhibitor of reactions whichconvert inosinic acid to adenosine monophosphate andguanosine monophosphate. However, neither incorpora-tion into "fraudulent" nucleic acids nor inhibition of re-actions at the ribonucleotide level explains its inhibition ofde novo synthesis of purines. We have now found that6-MP ribonucleotide is a potent inhibitor of the firstenzyme specifically concerned with purine synthesis,phosphoribosylpyrophosphate (PRPP) -amidotransferase,This enzyme catalyzes formation of phosphoribosylaminefrom PRPP and glutamine, and is normally the site offeedback regulation by adenyl ribonucleotides. ATP,ADP, and AMP are competitive inhibitors of bothPRPP and glutamine binding. 6-MP ribonucleotide incomparatively low concentrations mimics the action ofadenyl ribonucleotides in a pseudofeedback inhibition ofpurine synthesis. K, (enzyme-6-MP ribonucleotide dis-sociation constant) is 5.3 X 10' M, much lower thanthat of any normal ribonucleotide monophosphate in-hibitor, and 1/5 the Km of PRPP and 1/11 the Kmof glutamine (pigeon liver enzyme). The ribonucleotidesof 6-thioguanine and 8-azaguanine are also effective in-hibitors of PRPP-amidotransferase: K, = 1.8 X 10' Mand 5.4 X 10' M, respectively. (The ribonucleotides andfree bases are noninhibitory.) Furthermore, in thissystem the inhibitory effects of these analogue ribonu-cleotides and of AMP are additive. Thus, the ribo-nucleosides of several purine analogues are potentiallyvery effective regulators of purine synthesis de novo.These compounds may therefore have multiple points ofaction at which they exert their antimetabolic effects.The concept that ideal chemotherapeutic agents may becompounds which in their primary action mimic or aug-ment normal intracellular regulators deserves furtherevaluation.

Remarkable Anticholinesterase Tolerance in Nonmyas-thenic Patients. MICHAEL P. MCQUILLENAND RICH-ARD J. JOHNS, Baltimore, Md. (introduced by RichardH. Shepard).

Five patients were studied who tolerated the dailyadministration of 15 to 75 times the usual toxic doseof anticholinesterase drug (neostigmine or pyridostig-mine) without ill effect. In the past, such pronouncedtolerance to these drugs was thought to be possible onlyin the presence of myasthenia gravis, and these patientswere diagnosed as having myasthenia on the basis ofthis erroneous belief. However, electromyographic stud-ies showed that this tolerance was not on the basis ofmyasthenia gravis, for these patients did not have ademonstrable defect in neuromuscular transmission.Nevertheless, their response to neostigmine was notnormal, for neostigmine did not produce its characteristicblock of neuromuscular transmission when administeredto these patients. The patients exhibited no dependenceupon the anticholinesterase drug on its abrupt with-drawal. Tolerance apparently developed with the pro-longed administration of gradually increasing doses ofanticholinesterase drug. A similar tolerance was pro-duced in rats and mice by the repeated administrationof sublethal doses of pyridostigmine or isofluorophate overa 5-week period. In the animal, tolerance to the directeffect of the drug on cholinesterase did not occur, forcomparable blood cholinesterase inhibition occurred intolerant and in nontolerant animals. Presumably thetolerance was developed to the excess acetylcholine itself.

The Hypoglycemic Effect of Ketone Bodies. DAVIDMEBANEAND LEONARDL. MADISON,* Dallas, Tex.

Although recent evidence suggests that ketones mayinhibit glucose utilization by the rat diaphragm, thein vivo effects of hyperketonemia on carbohydrate me-tabolism have not been adequately characterized. Thepresent study was designed to determine simultaneouslythe effect of infusions of f8-hydroxybutyric acid (BOH)and of acetoacetate (AA) on hepatic glucose output(HGO), peripheral glucose utilization (PGU), andarterial glucose concentration. Dogs with chronic end-to-side portacaval shunts were used, thereby permittingmeasurement of hepatic rather than splanchnic glucoseoutput. HGO [hepatic blood flow (Bradley) X hepaticvein-arterial glucose concentration difference] was meas-ured at 10-minute intervals during the 40-minute controlperiod, and for 135 minutes during and after ketoneinfusions (5.5 to 10 mmoles/kg/hr). In all 12 studiesboth BOHand AA produced a significant hypoglycemiceffect, arterial glucose concentration falling below 40mg per 100 ml in some studies. This hypoglycemia wasentirely the consequence of a marked reduction in HGO.During the last 30 minutes of ketone infusion, HGOaveraged only 24.5 mg per minute, a 60% fall fromthe control value of 61 mg per minute. The decreasein the glucose pool was not commensurate with thereduction in HGOas a consequence of a 40% decrease

1383


in PGU. These data indicate that induced hyperketo-nemia has a profound and immediate effect on glucosemetabolism. Not only is PGUimpaired but, as a resultof a greater reduction in HGO, hypoglycemia supervenes.Since the infusions of the ketone acids and their sodiumsalts, which resulted respectively in systemic acidosis(pH 7.11) and systemic alkalosis (pH 7.56), all pro-duced identical changes in carbohydrate metabolism,these changes can be ascribed to hyperketonemia per seand not to alterations in acid-base balance.

Studies on the Regulation of Iron Absorption in Ex-perimental Hypoplastic and Iron-deficiency Anemias.GERALDA. MENDELAND ROBERTJ. WEILER, Chicago,Ill. (introduced by Leon 0. Jacobson).

In our previously reported studies utilizing splenec-tomy and radiostrontium administration to suppresserythropoiesis in mice, evidence was presented that theaugmentation of iron absorption induced by atmospherichypoxia can occur independently of acceleration of eryth-ropoiesis. Confirmatory data have been obtained byexposing iron-deficient mice to reduced oxygen tension.Augmentation of iron absorption without measurablechange in erythropoiesis (as assessed by histologic ex-amination of the bone marrow, reticulocyte count, andmeasurement of the incorporation of intravenously ad-ministered iron into circulating red blood cells) wasalso found to occur in this preparation, thus verifyingthis observation. In order to evaluate whether anemicanoxia might play a role in the regulation of ironabsorption, which is independent of body iron contentand the rate of erythropoiesis, groups of splenectomized,radiostrontium-treated mice were further treated withphenylhydrazine. Enhancement of iron absorption wasfound to occur with the increased degree of anemia soproduced. To determine the effect of reduction in therate of erythropoiesis on iron absorption in iron-deficientmice, the effect of splenectomy and radiostrontium ad-ministration in such animals was studied. Iron absorp-tion in splenectomized, radiostrontium-treated, iron-defi-cient mice remained at a level greater than four timesnormal, in spite of the marked suppression of erythro-poiesis produced. In summary, these preliminary studiessuggest that: 1) anemic anoxia per se may play a rolein the regulation of iron absorption in hypoplastic anemia;2) iron-deficient animals continue to absorb iron in quan-tities several-fold greater than normal, in spite of sub-stantial reduction in the rate of erythropoiesis producedby splenectomy and radiostrontium administration; and3) the observation that the enhancing effect of hypoxiaon iron absorption is independent of acceleration oferythropoiesis is confirmed.

The Effect of Cytotoxic Antibody In Vivo. Observa-tions by a New Technique. JOHN P. MERRILL,* CO-LETTE HANAUAND DONALDE. OKEN, Boston, Mass.

It is now clear that damage to cells through the actionof cytotoxic antibody (CA) may occur before histologic

evidence of this is present in fixed tissue sections. Theaction of CA on cell suspensions is characterized bycell swelling, suggesting inability of the cell to maintainosmolality of its internal environment either throughchange in permeability of cell membrane or throughinterference with metabolic processes necessary for the"sodium pump." In an effort to study this phenomenonin the intact metabolizing tissue we have utilized measure-ments of the short circuit current developed by the frogskin as a function of active transport of sodium. Rabbitsimmunized by the injection of minced frog skin developantisera to frog tissues, as evidenced by a cytotoxiceffect which causes rapid lysis of nucleated frog erythro-cytes. These sera, when applied to either or both sur-faces of the frog skin, with or without complement andspleen cells from the immune rabbit, show no effect uponshort circuit current. When, however, the sera wereinjected into the frog ventricle so that they reached theskin by way of the blood vessels, a significant decreasein initial short circuit current was observed. No decreasewas observed when the contralateral half of the abdomi-nal skin was perfused in the same fashion with normalrabbit serum. These results suggest an effect upon cellpermeability or sodium transport by the intact wholetissue when CA reaches it by way of the blood vessels,but not when it is simply juxtaposed to the outside orinside of the tissue.

Intracellular Ions and Biologic Functions in Kwashior-kor. JACK METCOFF* AND SILVESTRE FRENK, Chicago,Ill. and Mexico, D.F., Mexico.

Previous muscle biopsy studies on 30 malnourishedMexican children indicated that death from chronic se-vere infantile malnutrition (kwashiorkor) was associatedwith intracellular accumulation of water and Na, plusreduced intracellular concentrations of K, Mg, and or-ganic P. Recovery was associated with reversal ofthese phenomena. A few biopsies suggested that musclepyruvate and a-ketoglutarate decreased, and lactate in-creased, as death occurred. It was hypothesized thatactivity of some ion-dependent enzyme system might beimpaired, contributing to the fatal accumulation of intra-cellular water and Na. To assess this hypothesis, musclebiopsies were obtained on 15 additional malnourishedchildren. The characteristic cellular electrolyte lesionswere verified. In eight children, including three of fivewho subsequently died, muscle contents of the interme-diates-high-energy bond phosphoenolpyruvate (- PEP),pyruvate, a-ketoglutarate, oxalacetate, and isocitrate-as well as activities of the enzymes, pyruvic kinase (PK)(catalyzing - P + ADP-> ATP - P), lactic, isocitric,and malic dehydrogenases, glutamic-pyruvic and -oxal-acetic transaminases, were measured and compared tothose of six control subjects. Reductions of - PEP,pyruvate, and a-ketoglutarate were found in repeat bi-opsies shortly before or at death from severe malnutri-tion. The enzyme activities (mmoles/min/100 g dry fat-free solids) often were increased, but rarely decreased.

1384


Apparently "normal" homogenate PK activity from a

typical malnourished child who subsequently died thenwas studied at substrate [- PEP] ranging from 33 to1,000 ,moles per L and at a) [K] 147, [Na] 16.8 equiva-lent to normal controls; b) [K] 147, [Na] 50.2; and c)[K] 63, [Na] 50.2 mM/L, characterizing fatal kwashior-kor. Activities were inversely proportional to [- PEP]and reaction velocities (Km) for the unpurified muscleenzyme in the above systems were a) 6.5, b) 7.5, c)8.0 X 10' M. Since activity was virtually optimal inthe control system, it seems likely that cellular - PEPdepletion and distorted ion concentrations inhibit PKfunction (or vice versa) contributing to death inkwashiorkor.

Distribution and Properties of Anaphylactic and Venom-induced, Slow-reacting Substance in Guinea Pigs.E. MIDDLETON, JR. AND G. B. PHILLIPS,* New York,N. Y.

Slow-reacting substance (SRS), which produces de-layed contraction (5 to 30 seconds) of guinea pig ileum,as well as of human bronchial muscle, in the presence

of atropine and antihistamines, is liberated from anaphy-lactically sensitized lung of several animal species, in-cluding man, on reaction with antigen. Perfusion ofguinea pig lung with snake venom also produces SRSactivity. Sensitized and normal guinea pig tissues were

incubated with antigen or snake venom (Naja naja),and the supernatants assayed on guinea pig ileum forSRS directly and after extraction or dialysis procedures.Tissue lipid extracts were examined for SRS activitybefore and after venom treatment. 1) Sensitized lungincubated with antigen or venom liberated SRS (alsonormal lung with venom); much less activity arose

from aorta on similar treatment and minimal, if any,

activity from heart, liver, spleen, ileum, uterus, trachea,brain, and diaphragm of sensitized guinea pigs. 2) Nor-mal lung lipid extracts displayed SRS activity, witha latent period before contraction of the ileum of about1 to 3 minutes. Lipid extracts of kidney and livershowed no SRS activity; heart and brain lipid extracts,however, showed typical SRS activity. 3) Normal lunglipid extracts exposed to venom produced typical SRSactivity. 4) Dialysis experiments showed that a) SRSin Tyrode's supernatants from lung-antigen reactionswas poorly dialyzable; however, when these super-

natants were lyophilized, extracted with 80% ethanol,dried, and re-extracted with Tyrode's, the SRS was

readily dialyzable; b) the atypical SRS of lung lipidextracts was not dialyzable; and c) SRS present inlung lipid-venom mixtures was poorly dialyzable. 5)Anaphylactic SRS after lyophilization and ethanol ex-

traction, and venom-induced SRS were extractable byether from acid solution. Antigen-antibody reaction andvenom are strikingly similar in their capacity to produceSRS, which appears to be an acidic, ether-soluble, smallmolecule.

Depression of Glucose Oxidation by Palmitate and byAcetoacetate in the Perfused Rat Heart. P. R. MIN-TON AND M. S. RABEN,* Boston, Mass.

The influence of palmitate and of acetoacetate on theoxidation of uniformly labeled glucose-C14 to C1402 wasstudied on contracting but nonworking hearts from fedrats during a 45-minute retroaortic coronary perfusionwith recycled modified Krebs-Ringer bicarbonate solu-tion, containing 2% albumin and 100 mg per 100 mlglucose. Production of C1402 was least during the first15 minutes (15% of total produced in 45 minutes),although loss of radioactivity from the medium wasgreatest in this period. Calculated from C140, recoveredin the 45-minute period, the glucose oxidized was 2.6mg per g wet weight per hour. Glucose oxidation wasreduced by 82% when the medium contained palmitate(1 mEq/L) and by 75% when it contained acetoacetate(4 mEq/L). Palmitate and acetoacetate inhibited oxi-dation of glucose without decreasing its uptake as cal-culated from the disappearance of chemically determinedglucose. Of the C14 lost from the medium, C1402 ac-counted for 32% with glucose alone, 10% with palmitatepresent, and 9% with acetoacetate. By contrast, withacetoacetate as sole substrate (4 mEq/L), C1402 fromacetoacetate-3-C4 accounted for 79%o of the C14 lostfrom the medium, representing use of 5.4 mg per g perhour; and of the total C1402 produced, 34% appeared inthe first 15 minutes. Glucose had little effect on theoxidation of acetoacetate or of palmitate in the concen-trations used. Both long chain fatty acid and ketoacidwere preferred to glucose as substrate for oxidativemetabolism in these experiments. The observed patternmay explain several phenomena occurring at times ofincreased availability of the preferred substrates, includ-ing the increase of cardiac glycogen content with fasting,and the cardiac "glycostatic" effect of growth hormone.

The Determination of Sodium in Body Fluids by Meansof the Glass Electrode. EDWARDW. MOOREANDDONALD W. WILSON, Boston, Mass. (introduced byThomas C. Chalmers).

Until the present time, there has been no method forthe direct determination of activity of alkali ions inbiological fluids, physiologic data having been expressedsolely in terms of concentration. The following datasummarize the application of the mathematical and elec-trochemical techniques for the determination of pH tothe potentiometric measurement of sodium ion in bodyfluids. Sodium activity (aNa+) was determined in atotal of 312 samples of urine, serum CSF, whole blood,and plasma from normal subj ects and hospitalized pa-tients, using a highly selective capillary electrode com-posed of Na2O-AL20,-SiO2 glass (Eisenman). Analysesrequired about 3 minutes in 0.5-ml undiluted samples,and duplicate variability was less than that for the flamephotometer (SD 0.68 vs 1.17 mEq/L in urine, respec-tively). In all fluids studied, aNa+ was linearly relatedto photometric concentration (CNap). In order to ex-

1385


press electrode activities in terms of concentrations(CN.E), it was necessary to derive appropriate empiricalmean activity coefficients ('yNa+): CNa.E -=Na+/YNa+. Thishas been accomplished for each fluid. The average de-viation of CNaE from CNap, without regard to sign, was2.6 mEq per L in urine, 2.2 mEq per L in CSF, and1.3 mEq per L in serum, whole blood, and plasma. Inurine YNa+ (ordinate) vs aNa+ was approximated by thehyperbola Y = 0.66 + (4.04/X). Suppression of urineaNa+ below that expected for aqueous NaCl solutionswas found to be related to potassium concentration,since CNaE was closely approximated by aNa+/'YNa+K.Glucose (2 to 6%) increased urine aNa+ slightly, whileurea, uric acid, and protein had no effect. Erythrocytesand serum proteins had no adverse effects on electrodepotential. The activity and concentration of sodium ionhave been determined rapidly and accurately with theglass electrode, and similar studies with other glassesindicate the feasibility of quantitating K+, NH,+, Ca++,and Mg++ in biologic fluids.

Splenectomy and Survival of Macrocytes Resulting fromErythropoietic Stimulation. R. R. MOORES, F. STOHL-MAN, JR.* AND G. BRECHER, Bethesda, Md.

Initially, macrocytic red cells are produced by ratsin response to intense erythropoietic stimulation, e.g.,phenylhydrazine-induced anemia, blood loss, exposure tosimulated altitude, or administration of erythropoietin.These macrocytes have a markedly shortened life span.Studies to date point to a dose-response relationshipbetween the intensity of stimulation and the degree ofresponse. Also, the initial cohorts of cells are moremacrocytic and have a shorter life span than ensuingones. The possible role of the spleen in removal of theselarge cells was evaluated by bleeding or giving phenyl-hydrazine to pathogen-free, female Sprague-Dawley ratsapproximately 2 weeks after splenectomy. Red cell sur-vival was measured with radioiron (Fe'9) and from cellsize distribution curves obtained with a model B Coultercounter. The phenylhydrazine-damaged cells were to-tally removed from the circulation, although at a slowerrate in the splenectomized than in the intact animals.However, neither the degree of macrocytosis nor thelife span of the macrocytic cells was affected by sple-nectomy in either the bled or phenylhydrazine-treatedanimals. These observations suggest that shortenedsurvival of red cells produced in response to intenseerythropoietic stimulation is not due merely to mechanicalfactors related to the macrocytosis.

The Nature of Pulmonary Capillary Blood Flow andGas Exchange. EUGENEMORKIN, 0. ROBERTLEVINE,FREDERICK 0. BOWMANAND ALFRED P. FISHMAN,*New York, N. Y.

Rhythmic fluctuations in intrapulmonary pressure,synchronous with the heart beat, have been recordedplethysmographically in man. However, these observa-

tions have not settled whether alveolar-capillary gagexchange, as well as the pulmonary capillary blood flow,is continuous or interrupted and pulsatile. In the pres-ent study, closed-chest dogs were used to test the effectsof arresting unilateral pulmonary blood flow and gasexchange on the contour of the ipsilateral intrapulmonarypressures. The experiment involved: 1) surgical ex-teriorization of the left bronchus for the ventilation ofone lung while the other was being tested, 2) placementof a balloon-tipped catheter so that blood flow throughthe test lung could be interrupted at will, 3) introduc-tion of cardiac and bronchial catheters so that intra-pulmonary pressure pulses could be related to cardio-dynamic events, and 4) pneumographic monitoring ofchanges in thoracic circumference. The intrapulmonaryoscillations were recorded during inflation of the testlung with air and with 80% nitrous oxide in oxygen;similar recordings were made when the blood flowthrough the test lung was interrupted. Separate spiro-metric and manometric measurements were used toverify that blood flow to the test lung had been arrested.It was found that: 1) the intrapulmonary pulsations onair and nitrous oxide differed principally with respectto the slope of their baselines, rather than to their con-tour; 2) the uptake of nitrous oxide fluctuated irregu-larly and continued without pause throughout the cardiaccycle; and 3) arrest of blood flow to the test lungabolished the slope of the nitrous oxide uptake so thatthe nitrous oxide and air oscillations became super-imposable. The results indicate that, under the conditionsof these experiments, alveolar-capillary gas exchangeand pulmonary capillary blood flow are continuous proc-esses and undergo only slight deviations from theirmean values during each cardiac cycle.

Studies on the Relationship of Angiotensin to Hyper-tension of Renal Origin. RUSSELL E. MORRIS, JR.,PATRICIA A. RANSOMAND JOHN EAGER HOWARD,Baltimore, Md. (introduced by Samuel P. Asper, Jr.).

The decreased flow of blood to a kidney with theproduction of ischemia can cause hypertension. It hasbeen postulated that this hypertension is due to a cir-culating humoral agent, angiotensin, produced by theaction of the kidney enzyme, renin, on plasma globulins.Whether angiotensin is produced in all types of kidneydisease or only by an ischemic kidney is unknown. Fortypatients have been evaluated in this study to correlatethe relation of angiotensin to hypertension. The diag-nostic studies used were renal arteriogram, split renalfunction tests, renal scintiscan, and peripheral arterialand renal venous blood determinations of angiotensin.Angiotensin was isolated by a method of chemical ex-traction developed in this laboratory and determined bybioassay in the rat. This procedure is sensitive to0.01 ,ug angiotensin in 50 ml of blood. Twenty patientswere diagnosed as having essential hypertension and inthese angiotensin was not found. Angiotensin was notfound in one patient with pheochromocytoma, one withprimary aldosteronism, and in five patients with malig-

1386


nant hypertension. After renal surgery, three of thesepatients with malignant hypertension showed no changein blood pressure. Thirteen patients were diagnosed ashaving renal hypertension with a demonstrated unilateralor bilateral renal lesion. In these, angiotensin wasfound both in peripheral arterial blood and in renalvenous blood from the ischemic kidney; angiotensin wasnot found in the renal venous blood from the uninvolvedkidney or in the peripheral venous blood. Six of thesepatients received unilateral nephrectomy and are im-proved; angiotensin was not found in the peripheralarterial blood postoperatively. From these data, it isbelieved that concentrations of angiotensin sufficient tocause hypertension by their direct effect on blood vesselsare present only in renal ischemia. Angiotensin has notbeen demonstrated in essential or malignant hypertension.

The Effect of Endotoxin on Cardiac Output and HepaticFlow in Normal and Tolerant Rabbits. JOHN M.MOSESAND WILLIAM J. MACINTYRE, Cleveland, Ohio(introduced by Robert H. Ebert).Endotoxins are known to alter the phagocytic function

of the reticuloendothelial system (RES). After largedoses of intravenously administered endotoxins, RESactivity is initially reduced. Several factors influencingRES activity have been implicated after endotoxin ad-ministration. One of these factors, organ perfusion, hasnot been emphasized in spite of the knowledge that endo-toxins induce profound systemic changes in circulation.The fact that certain toxic substances produce markedchanges in hepatic circulation and the observation thatrabbits resistant to the lethal effects of endotoxin mayhave impaired hepatic blood flow, suggested the possi-bility that local organ perfusion might be altered in theabsence of more widespread failure of circulation afterendotoxin. Both normal rabbits and those renderedtolerant to endotoxin by a series of daily injections ofbacterial lipopolysaccharide were studied before andafter the injectionof large doses of lipopolysaccharide.A simple method was devised to enable simultaneous andrepeated determinations of both cardiac output and hepaticblood flow by means of single injections of carrier-freecolloidal radiogold. An arteriovenous shunt was estab-lished between the carotid artery and contralateraljugular vein. The diverted arterial blood was ledthrough a scintillation counter by means of a poly-ethylene tube so that radioactivity in the blood couldbe monitored continuously. After the injection of gold,radioactivity in blood ejected from the heart inscribeda curve which permitted calculation of cardiac output.The subsequent disappearance of radioactivity establisheda measure of hepatic blood flow. The results showedthat reduced hepatic blood flow was associated with afall in cardiac output in both normal and tolerant rabbits.Impaired circulation was short-lived in tolerant rabbitsbut progressive in normals. The data suggest that altera-tions in organ perfusion may be of primary importancein influencing RES activity after the injection ofendotoxin.

Aryl Amine Acetylation in Human Red Cells. ARNOG.MOTULSKY* AND LAURA STEINMANN, Seattle, Wash.

Biotransformation of isoniazid (INH) varies consider-ably between individuals. Blood levels in family membersafter administration of INH indicate that a single genecontrols the major pathway of drug metabolism. Threeclasses of individuals can be defined: homozygotes withthe double dose of the gene, heterozygotes, and homozy-gotes lacking this gene. Pharmacologic evidence suggeststhat this pathway concerns hepatic acetylation of INH.It is uncertain whether para-aminosalicylic acid (PAS)is acetylated by the same mechanism. Since INH bloodlevels cannot characterize the enzymatic mechanismunderlying these genotypic differences, red cells wereselected as a potentially useful model for the direct studyof acetylation. Acetylation of PAS and of para-aminobenzoic acid (PABA) could be demonstrated inhemolysates using an acetate-CoA-ATP system to pro-duce acetyl-CoA as acetyl donor. When citrate wassubstituted for acetate, even stronger acetylation oc-curred. These reactions indicate the presence of acetate-activating enzyme, citrate-splitting enzyme, and arylamine acetylase in red cells. Levels of acetylating ac-tivity were high in reticulocytes but remained significanteven in mature cells. When isoniazid and sulfanilamidewere substituted as acetyl acceptors, no acetylation wasfound. INH and sulfanilamide also failed to inhibitacetylation of PABA, while PAS, aminopterin, andamethopterin strongly inhibited PABA acetylation bycompeting for available acetyl-CoA. Population studiesof PAS and PABA acetylation in both Caucasian andNegro subjects indicated that variability followed a nor-mal distribution as contrasted with the multimodal curveseen with INH inactivation. These data demonstrate thepresence of an aryl amine acetylase in human red cellswhich reacts with PAS and related compounds but notwith INH. This red cell acetylase is controlled by agenetic mechanism different from that regulating INHinactivation. These findings strengthen the suggestionthat hepatic acetylation of INH and PAS is carried outby different acetylating enzymes.

Antihypertensive Principle from Renal Medulla. E. E.MUIRHEAD,* J. W. HINMAN, E. G. DANIELS, M. Ko-SINSKI AND B. BROOKS, Detroit and Kalamazoo, Mich.

Hypertensive cardiovascular disease occurs in the dogafter renal ablation plus sodium and dietary protein.This form of renoprival hypertension is prevented byexplanted whole kidney or renal medulla, a saline extractof renal medulla given intravenously, and a medullorenalextract refined by 55%o ethanol extraction. Preventionof canine renoprival hypertension affords an assay pro-cedure for purifying the medullorenal antihypertensiveprinciple. Direct 55%0 ethanol extraction of canine andporcine renal medulla yielded a product preventive againstcanine renoprival hypertension. Subsequent preparationswere derived from porcine kidneys and tested in dogs.Removal of ethanol after 55% ethanol extraction, fol-lowed by Sephadex chromatography, permitted use of

1387


kidneys after partial cortical dissection. Elution of thecolumn yielded two fractions. The component activeagainst hypertension appeared to be of low molecularweight (<2,000 mol wt). The active principle was notretained on Dowex-50 (H+) resin and presumably wasnot an amino acid or peptide. The 55% ethanol extractwas extracted first with Skellysolve B and then withethyl acetate. The solids from ethyl acetate dissolved inethanol and given by mouth (1 mg/kg/day) protectedagainst renoprival hypertension. The active material wasslightly water-soluble and did not contain free aminogroups, phenolic hydroxyl groups, or reducing carbohy-drates. Further solvent fractionation yielded a prepara-tion active in doses of 50 Aig per kg per day. Protectionagainst hypertension also altered significantly other pa-rameters of nondialyzed dogs. The average change fromcontrol for 13 protected nonhypertensive dogs and for5 hypertensive animals was: for blood urea nitrogen,+188 and +316 mg per 100 ml; for serum potassium,+1.67 and +2.96 mEq per L; for reduced glutathione oferythrocytes, -2.6 and -24.5 mg per 100 ml red bloodcells. Cardiovascular injury was not observed whenhypertension was prevented. A principle from porcinerenal medulla protected against canine renoprival hyper-tension when given orally in relatively small amounts.

Incorporation of 17-Hydroxy-Pregnenolone-H' into Cor-tisol by the Human Adrenal In Vitro and by the DogAdrenal In Vivo. PATRICK J. MULROw, GEORGEL.COHN, ALBERT A. KULJIAN ANDWILLIAM F. GANONG,New Haven, Conn. and San Francisco, Calif. (intro-duced by Philip K. Bondy).

Studies of adrenal steroid biosynthesis indicate thatpregnenolone is incorporated into cortisol via pregnen-olone->progesterone >17-hydroxy-progesterone->1 1-de-oxycortisol->cortisol. However, 17-hydroxylation ofpregnenolone is considered the first step in adrenal an-drogen biosynthesis. The present study demonstratesthat a major pathway of cortisol biosynthesis is by wayof 17-hydroxy-pregnenolone in normal and hyperplastichuman adrenals. Slices of four "normal" adrenals ob-tained by surgery from two patients with prostatic cancerwere incubated with 17-hydroxy-pregnelone-H3. Eightto 19% of the radioactivity was incorporated into thecortisol isolated from the incubation media. Intermedi-ates, 17-hydroxy-progesterone-H3and 11-deoxycortisol-H3,were also isolated. Little or no corticosterone-H' wasdetected. Radiochemical purity of the free alcohols andtheir derivatives was achieved by paper chromatography.Slices of a hyperplastic adrenal gland from a patientwith Cushing's syndrome were incubated with equimolaramounts of 17-hydroxy-pregnenolone-H' and progeste-rone-C1'; 6%o of 17-hydroxy-pregnenolone and 9%o ofprogesterone-C1' were incorporated into cortisol. 11-De-oxycortisol-HW was also detected. Studies using proges-terone-H8 as the substrate in normal and in threehyperplastic adrenal glands demonstrated 7 to 16% in-corporation of progesterone-H' into cortisol. In threedogs who were receiving a constant infusion of ACTH,

equimolar amounts of 17-hydroxy-pregnenolone-H3 andprogesterone-C1' were infused directly into the isolatedadrenal gland in vivo. Cortisol containing H3 and C1'was isolated from adrenal vein blood. The relativeamounts of the two precursors incorporated into cortisol,expressed as the ratios of per cent incorporation of17-hydroxy-pregnenolone-H1/progesterone-C"4, were 0.65,0.76, and 1.8. In conclusion, a major pathway for cortisolbiosynthesis, bypassing progesterone, has been demon-strated to exist in the human adrenal in vitro and in thedog adrenal in vivo. A pathway via 17-hydroxy-preg-nenolone-->17-hydroxy-progesterone->1 1 -deoxycortisol->cortisol has been shown in vitro.

The Fate of the Platelet. J. F. MUSTARD, E. A. MUR-PHY AND G. A. ROBINSON, Toronto and Guelph, On-tario, and Baltimore, Md. (introduced by R. F. Far-quharson).Calculations of platelet life span from data obtained in

radioactive labeling experiments have varied widely be-cause the mode of destruction has hitherto been obscure.Estimates based on the assumption of random destructionhave been about 3 to 4 days; the same data yield an esti-mate of 10 to 12 days if death is assumed to be due toplatelet aging. It is clearly important to decide which ofthese mathematical models is the more appropriate, butattempts to do so by scrutiny of survival curves are frus-trated by even small experimental errors. We havetherefore had recourse to indirect methods. 1) If de-struction is due to senescence, then the shape of thesurvival curve should vary according to the age of theplatelets labeled. If, however, destruction is random, itshould be independent of platelet age. In swine and indogs, platelets were simultaneously labeled in vivo withS' (which is incorporated in developing platelets only)and diisopropyl fluorophosphate (DFP)2 (which ad-heres indiscriminately to circulating platelets of all ages).The plots of platelet survival curves on semilogarithmicgraph paper yielded two parallel straight lines, showingthat destruction occurred at the same rate in the twogroups of platelets. This suggests random destruction.2) By the DFP3 method, we found in man that plateletsurvival computed on either mathematical model wasprolonged by high doses of bishydroxycoumarin in 12subjects (p <0.001), by 8,000 U of heparin t.d.s. in 14(p <0.01), by restriction or change of dietary fat in 7(p < 0.01), and by restriction of smoking in 7 (p < 0.01).It seems most plausible to suppose that these varied ef-fects are mediated through the environment of the plate-let. It is concluded that the fate of the platelet isdetermined mainly by its environment rather than by itsintrinsic metabolism, and that its survival is best com-puted as an exponential half-life.

Digital Hemodynamics in Hypertension with ParticularEmphasis on Vealomotor Tone. RUDOLPHJ. NAPO-DANO, FRANCIS S. CALIVA, ROBERT M. STAFFORDANDRICHARDH. LYONS,* Syracuse, N. Y.There is ample evidence that the increased peripheral

resistance in hypertension results from narrowing of

1388


small arteries and arterioles. There is little informationconcerning the state of the venous circulation. Since thesmall veins and venules possess smooth muscle elementswhich are anatomically and physiologically similar, aninvestigation was undertaken to assay their role in theover-all increase in resistance. Ten normotensive and 15hypertensive subjects were studied three or more timesby the technic of digital rheoplethysmography. The pa-tients were divided into groups according to age, and theparameters investigated were digital blood flow, digitalarterial and venous pressures, vascular resistance, andtone in the small veins. Calculated peripheral resistancesin all the hypertensives were significantly higher than inthe normotensives. In hypertensive patients under 50years of age, there was a significantly lower mean bloodflow and higher venous pressure than in normals of thesame age group. Such differences were not evident in thepatients over 50 years. Analysis of the rate of inflowcurves for the first two pulse cycles revealed a moderatedegree of increased venomotor activity in the hyperten-sives. Criteria were designed for grading the status ofexisting tone from 0 to IV. Of the normotensives lessthan 50 years old, three of six showed no evidence of in-creased venomotor tone, while the other three wereclassed as grades I or II. None of these normotensivesdemonstrated grade III or IV venomotor tone. Of theseven hypertensives under 50, five demonstrated grade IIIor IV venomotor activity. In the age group over 50,three of four normotensives and all eight hypertensivesdemonstrated grade I to III venomotor tone. It thusappears likely that the increased peripheral resistance inhypertension, in part, includes the small veins. Experi-mentally, this is most evident in the subjects under 50years of age.

The Relationship of Increased Erythropoiesis to GlobinSynthesis in Man. DAVID G. NATHANAND FRANKH.GARDNER,* Boston, Mass.

This study was designed to compare the effects onheme and globin synthesis of androgen-induced polycy-themia with other states of increased erythropoiesis.Glycine-2-C"4 was administered i.v. to three individualsbefore and after sufficient androgen therapy to increasethe total red cell volume (patient group A). Daily de-terminations of C14-heme and -globin specific activity(SA) of peripheral blood were performed for 1 week.Similar studies were done in a patient who was poly-cythemic due to a virilizing adrenal tumor, a patient withpolycythemia vera, and a patient with iron deficiency(patient group B). These patients did not receiveandrogens. Hypochromia was a characteristic of theperipheral blood after androgen treatment (group A)and in group B. The curves of C14-heme and -globin SAin group A were compared both to similar data in normalindividuals and to the pretreatment curves. The data ofgroup B were compared to those of normal individuals.The results were very similar in groups A and B. Theheme curves differed slightly from pretreatment on nor-mal values in that the peak increment of SA occurred

2 to 3 days rather than 3 to 4 days after glycine adminis-tration. The globin curves were markedly altered. Theyassumed a hyperbolic rather than a sigmoid shape, sothat the peak increment of C14-globin SA occurred 1 or2 days rather than 5 days after glycine administration.These results indicate that the kinetics of heme and globinsynthesis due to androgen stimulation do not differ fromsuch kinetics in other states of increased erythropoiesis.Globin synthesis or "turnover" is enhanced out of pro-portion to heme synthesis or "turnover" when erythro-poiesis is stimulated. This enhancement is apparentlymost marked during the later stages of erythroid cellmaturation.

Differential "Aging" in Two Types of Bone. GEORGENICHOLS, JR.,* Boston, Mass.

Methods have now been developed in this laboratoryfor measuring in vitro the metabolic activity of bonecells, and for tracing incorporation of substrates intometabolic end products and components of bone organicmatrix. Thus, studies are possible of the cellular meta-bolic basis of the differential rates of growth and re-modeling which occur in different skeletal areas withincreasing age. Answers have been sought to threequestions: What changes occur in cell metabolic patternsand rates with age? Are these different in trabecularand cortical bone? Are these the result of changes in thenumber of cells, their metabolism, or both? Cell content,together with rates of incorporation of label from glucose,glycine, and proline into metabolic end products, cells,and collagen, have been measured in cortical and trabecu-lar bone from metaphyses of rats ranging in age from3 weeks to 8 months. Cell content, estimated byDNAand total N, is lower in cortical than in trabecularbone and decreases with age, while in trabecular bonecellularity changes little. Glucose consumption, calcu-lated per 100 mg fresh bone, is 25% lower in cortex at 3weeks and decreases steadily with age, while in trabecu-lar bone no decrease was observed. The percentage ofglucose label incorporated into various end products, cells,and collagen was the same in both bone types at all ages.Calculation on a cell N basis revealed that changes incortex were the result of changes in cell content, not incell activity. Studies of amino acids, however, showedprogressive decreases in incorporation into both cells andcollagen with age when calculated on either basis. Incontrast to glucose, proline incorporation was 30 to 50%ohigher in cortical than in trabecular bone. Thus, patternsof glucose and amino acid metabolism change with age indifferent ways and for different reasons depending onthe bone type.

Evidence for Abnormal Lipoprotein Synthesis in TwoVarieties of Idiopathic Hyperlipemia. WILLIAM H. R.NYE AND CHRISTINE WATERHOUSE,*Rochester, N. Y.

In order to explore the differences between fat-inducedand carbohydrate-induced idiopathic hyperlipemia, a typi-cal patient of each type was studied on a metabolic ward.The first patient has been observed at intervals during

1389


the past 17 months. Before diet was begun, plasma wasmarkedly hyperlipemic, and individual phosphatides werepresent in proportions that could not be explained by alter-rations in ratios among normal lipoprotein classes. Witha low fat, isocaloric diet, phosphatide proportions revertedto normal, and the hyperlipemia cleared. The second pa-tient, who was obese, has had two periods of study duringthe past 10 months, and was under continuous dietarysupervision for 4 months during the second period.Phosphatides were initially normal. On a low fat, highcarbohydrate diet, hyperlipemia increased and phospha-tides remained normal. On a low calorie, normal fat,low carbohydrate diet, hyperlipemia cleared, but pro-portions among individual phosphatides changed in di-rections not explainable by shifts in normal lipoproteinclass ratios. After the patient had lost 40 pounds over a3-month period, the low carbohydrate diet was madeisocaloric. Plasma remained clear, and phosphatideproportions became normal again. Glucose tolerance,which was decreased in this patient initially, became nor-mal after weight loss. Both patients showed a goodresponse to heparin, and in both cortisone caused adecrease in total lipids without altering phosphatide pro-portions. Wepostulate that the first patient synthesizedabnormal lipoproteins in response to exogenous fat, butwas able to synthesize normal lipoproteins and maintainnormal plasma lipid levels on a diet in which carbohy-drate predominated. The second patient synthesized ab-normal lipoproteins only when the major source ofenergy was from exogenous plus endogenous fat; how-ever, the hyperlipemia itself was correlated with ahigh carbohydrate intake, suggesting that this aspect ofthe patient's abnormality was quantitative rather thanqualitative in nature.

The Metabolism of Warfarin. ROBERT A. O'REiLLY,PAUL M. AGGELER* AND Lois LEONG, San Francisco,Calif.

Variable responses to coumarin anticoagulants have notbeen explained by detailed clotting factor analysis. Thesevariations are conditioned by differences in drug metabo-lism. While warfarin is widely used in anticoagulanttherapy, its metabolic behavior in man has not been previ-ously studied. Since no applicable method was available,we developed an assay based on ultraviolet absorptionspectrophotometry. Thirteen normal subjects were givensingle oral doses of warfarin, 1.5 mg per kg; 6 receivedthe same dose i.v. The following results were obtained(mean ± SD): biologic half-life (tj), 44 ± 13 hours;elimination rate, 1.6±0.5% per hour; apparent volumeof distribution (Vd), 12 ± 1%o body weight; time ofmaximum plasma levels, 4.6 ±2.0 hours. A significantcorrelation (p < 0.05) was found between tj and themaximum prothrombin time (Quick), and a highlysignificant correlation (p < 0.01) between the 48-hourwarfarin level and 48-hour prothrombin time. Variousparameters of warfarin and bishydroxycoumarin (Dicu-marol) metabolism were compared. Gastrointestinal ab-sorption: On the basis of plasma drug levels, biologic

effect, and drug recovery in the stool, warfarin is ab-sorbed rapidly and completely, and Dicumarol slowly andincompletely. Absorption rates calculated from theplasma data indicate higher maximum rates for Dicumarolthat become slower before the dose is completely absorbed.Elimination: The tj for oral and i.v. warfarin and forhigh oral Dicumarol doses is independent of dose size;for all i.v. and low oral Dicumarol doses, ti varies di-rectly with dose size. Excretion of warfarin in urineand stool is essentially 0, and of Dicumarol is 0 inurine and 20% in stool. About 0.25 of a warfarin doseis recovered from urine as a metabolite. Plasma proteinbinding of warfarin is 97%, and of Dicumarol 99%;these values may be related to the low Vd. In normalsubjects the prothrombinopenic effects of the coumarinanticoagulants correlate well with their metabolic fate.

A Serum Factor that Agglutinates Stored Erythrocytes.FARUK L. OZER, DAVID MILLER AND HUGHCHAPLIN,JR.,* St. Louis, Mo.

The authors have encountered a hyperglobulinemic hu-man serum which causes strong agglutination of storederythrocytes while fresh erythrocytes remain unaffected.Agglutinating activity was demonstrable in serum di-luted 1:100,000 and withstood heating at 600 C for 10hours. Activity was associated with a paraprotein frac-tion of gamma mobility and a sedimentation constant of19S. Protein precipitable at low ionic strength possessedstrong agglutinating activity when redissolved, withloss of activity after treatment with penicillamine.Agglutinating activity was also inhibited by exposure tosuch sulfhydryl-binding and disulfide-breaking agents asp-hydroxy-mercuribenzoate, Na arsenite, GSSG, GSH,and cysteine. Activity was diminished by absorptionwith aged erythrocytes and erythrocyte ghosts, and couldbe eluted from the washed cell membranes by heating at560 C. However, absorption of the serum factor wascompletely blocked by pretreatment of ghosts withp-hydroxymercuribenzoate. The alterations which ren-dered erythrocytes agglutinable by the serum factorprogressed rapidly at 370 C, slowly at 40, and could beprevented or diminished by addition of acid citratedeptrose or adenosine to the suspending medium. Ag-glutinability was hastened by exposure to methylene blue(preventable by adenosine in normal cells but not inglucose-6-phosphate dehydrogenase-deficient cells) andalso by inhibitors of glycolysis (2-deoxyglucose, Naarsenite, Na arsenate, Na fluoride, iodoacetate) and byp-hydroxy-mercuribenzoate. A close relationship of ag-glutinability to nonviability was supported by in vivostudies demonstrating markedly increased survival of thenonagglutinable cells in banked blood stored for 33 days.The present studies indicate that 1) the serum factor isa macroglobulin; 2) the agglutination reaction is inhibitedby exposure of the macroglobulin and the erythrocytemembrane to reagents which affect sulfhydryl and disul-fide groups; 3) agglutinability results from impairederythrocyte metabolism; and 4) separation of nonviablecells in stored blood may become possible.

1390


The Metabolism of d-Aldosterone: Isolation and Chem-ical Identification of Metabolites; Distribution of FreeAldosterone and Metabolites in Tissues and Fluidsunder Physiologic Conditions. MAURICE M. PECHET,*ROBERTH. HESSE AND HEINZ KOHLER, Boston, Mass.

From the incubation of the natural d-aldosterone witha homogenate of rat liver and a TPNH-generating sys-tem, the following metabolites were isolated and identifiedby precise chemical methods: 5cr- (4,5) -dihydroaldoste-rone; 3,8-hydroxy-5a- (4,5) -tetrahydroaldosterone; 3a-hydroxy-5a- (4,5) -tetrahydroaldosterone; and 3ar-hy-droxy-5p- (4,5) -tetrahydroaldosterone. Under theseconditions d-aldosterone is reduced primarily to metabo-lites with A/B trans (5a) configuration. Similar resultshave been obtained with the isolated perfused rat liver.The hemiacetal structure characteristic of aldosteroneremains intact. Experiments with tritiated aldosteroneand radioactive-trapping experiments with syntheticallyprepared dihydroaldosterone indicate that the sequentialsteps in the metabolism of aldosterone are: d-aldoste-rone--dihydroaldosterone--->tetrahydroaldosterone. In theintact rat, under basal conditions simulating the steadystate, liver demonstrated the highest metabolic activity.No unchanged aldosterone but very high concentrationsof conjugated metabolites are present in liver. Thehighest concentration of free aldosterone is found in thekidney, indicating a concentration of aldosterone bythe kidney. The level of free aldosterone is low inserum, of the order of 10'M. The plasma aldosteronepool is in equilibrium with a larger pool contained in thetotal water space. Muscle tissue metabolizes d-aldoste-rone to the tetrahydro metabolites, and because of thelarge mass of muscle present this tissue may make asignificant contribution to the metabolism of aldosterone.Experimental hemorrhage in normal and adrenalectomizedrats receiving a constant aldosterone infusion of 0.3 ,gper hour is associated with an increase in free aldosteronein plasma, indicating that under these conditions metabo-lism of aldosterone is markedly diminished.

Sickling Phenomenon Produced by Hypertonic Solu-tions: A Possible Explanation for the Hyposthenuriaof Sicklemia. PASQUALEE. PERILLIE AND FRANKLINH. EPSTEIN,* New Haven, Conn.

The effect of hypertonic solutions upon sickling wasexamined in an attempt to clarify the inability of pa-tients with sickle cell anemia to excrete a highly con-centrated urine. Blood from such patients was mixedwith solutions of NaCl, urea, sucrose, or mannitol withosmolalites varying from 300 to 1,200 mOsmper kg.All tests were done at the temperature and oxygen ten-sion of room air; the pH of the final mixture variedfrom 7.4 to 7.5. Sickling occurred in all solutions ofNaCl with a final osmolality greater than 600 mOsmperkg and could be reversed by diluting to concentrationsbelow this. Sickling also occurred when blood wasadded to concentrated solutions of sucrose and mannitol,but not when blood was added to solutions of urea alone.

Normal red blood cells did not sickle when placed in ahypertonic medium. The viscosity of whole blood fromsickle cell patients increased progressively when theosmotic concentration of the plasma was increased byadding NaCl. No change, however, was noted in theviscosity of sickle hemoglobin solutions when solute con-centration was similarly raised. The sickling of cellscontaining S-hemoglobin is promoted when they areimmersed in hyperosmotic solutions, probably as a resultof cellular dehydration. It is suggested that when suchcells enter the hypertonic milieu of the renal medullasickling is promoted, thereby increasing blood viscosity,restricting blood flow to the medulla, restraining activereabsorption of sodium in this region, and limiting theproduction of a high concentration of sodium in themedullary interstitial fluid.

Abnormal Myoglobin Chromatography in Childhood Mus-cular Dystrophy. GERALD T. PERKOFF,* ROBERT L.HILL ANDFRANKH. TYLER,* Salt Lake City, Utah andDurham, N. C.Normal human myoglobin fractionates into three major

(F,, F,, and F8) and three minor components on diethyl-aminoethyl cellulose columns. F, is separated further intosix components by a gradient elution technique. F, andF, comprise 70 to 85% of normal myoglobin and are acidand alkaline metmyoglobin, respectively. Their electro-phoretic mobilities, end groups, peptide patterns ("finger-prints"), amino acid compositions, and sedimentation-diffusion constants are identical. F, myoglobins com-prise 15 to 30% of normal myoglobin and are alkalinemetmyoglobins with more rapid electrophoretic mobilitythan F, and F,. One of the F, components, FD, has beenanalyzed and has a peptide pattern different from that ofF, and F,. In the visible spectrum at pH 8.6, FD showsa single peak at 540 m/c as does fetal myoglobin. Thespectral characteristics of whole Fs and fetal myoglobinare similar when these preparations are converted toferromyoglobin, oxymyoglobin, and cyanmetmyoglobin, re-spectively. Definitive chemical characterization of fetaland FsD myoglobin must be accomplished before anyconclusions concerning this spectral similarity are justi-fied. Myoglobin from skeletal muscle of two childhooddystrophy patients contained a preponderance of F, (50and 100% of the total myoglobin, respectively). In oneinstance F, and F,, although markedly reduced in amount,were present and showed normal structural character-istics. The spectrum of whole alkaline metmyoglobin ofchildhood dystrophy reflects the preponderance of F,.This probably explains the reported spectral "abnor-malities" in myoglobin from dystrophic patients. Thespectral characteristics, electrophoretic mobilities, andchromatographic patterns of myoglobins from two pa-tients each with facioscapulohumeral and myotonic dystro-phy were normal. These studies indicate an abnormaldistribution of myoglobins in one form of progressivemuscular dystrophy. The preponderance of F, in child-hood dystrophy myoglobin may in part represent per-sistence of fetal myoglobin, but this is unproved at this

1391


time. Elucidation of the mechanisms responsible for theobserved changes awaits further study.

The Mechanism of Death in Experimental PneumococcalInfection. JOHN E. PERRYAND LEIGHTON E. CLUFF,*Baltimore, Md.

Lethal pneumococcal infection in rabbits is character-ized by fever, leukopenia, bacteremia, and shock. Afterintraperitoneal inoculation of 5 X 10' type I pneumococci,56% of the animals died within 24 hours and the re-mainder within 5 days. The rapidity of death could becorrelated with the time of onset and degree of bac-teremia. None of the rabbits died if treated withpenicillin during the first 3 hours of infection. Whentreatment was delayed beyond 4 hours, the early mortalityrate was not affected; however, the animals were pro-tected against late mortality. The difference in mortalityduring treatment could be correlated with the rapidityin decline of bacteremia. The blood became sterilewithin 4 hours in those animals which survived; how-ever, bacteremia persisted for 6 to 8 hours in those thatdied. A similar delay in bacterial killing was observedwhen penicillin was added to broth cultures of pneu-mococci more than 4 hours after inoculation. Theseeffects may be related to the emergence of phenotypeswhich are relatively indifferent to penicillin. Hemody-namic studies revealed a progressively declining cardiacoutput, an increased peripheral resistance, and a normalvenous pressure. During fever, the cardiac output de-creased to 78% of control values and to 54% just priorto the onset of hypotension. The peripheral resistanceincreased to 131% of normal during the febrile period,and to 230% prior to shock. During the period of hypo-tension an even greater reduction in cardiac output andincrease in peripheral resistance was observed. It islikely that shock during this infection is due to regionalpooling of blood and decreased venous return to theheart, resulting in a decreased cardiac output and com-pensatory increase in peripheral resistance.

The Effect of Insulin-binding Antibodies on the Distri-bution and Excretion of Insulin in Diabetics with theNephrotic Syndrome. JOHNR. PETERSEN, NORMANH.ENGBRING, NORMANW. OSHER AND WILLIAM W.ENGSTROM,* Milwaukee, Wis.

The metabolism of insulin, as defined by the distribu-tion and fate of insulin-IX, is altered in insulin-treatedsubjects by the appearance of insulin-binding antibody.The clinical significance of this antibody in diabetics isuncertain. Decreasing insulin requirements often charac-terize the diabetic who develops the nephrotic syndrome.It might be anticipated that urinary loss of antibody orinsulin, or both, might occur in diabetics with proteinuria.Trichloroacetic acid (TCA)-precipitable radioactivity inplasma and urine was determined at intervals followingintravenous injection of insulin-I" in diabetics with andwithout proteinuria, in nondiabetics with a nephroticsyndrome, and in normal subjects. In normal subjects,

insulin was rapidly distributed throughout the extracellu-lar space, and free In, a product of degradation, wasrapidly excreted in the urine. In diabetics whose plasmacontained insulin-binding antibody, the greatest amountof insulin was retained within the vascular compartment,and free I"' was excreted slowly. In edematous nephroticpatients, both diabetic and nondiabetic, an expandedinsulin pool resulted in a slow turnover rate; however,the effect of insulin antibodies in diabetic nephroticswas to maintain higher blood levels of insulin. Using asalting-out method (modified from Grodsky and For-sham) for selective precipitation of antibody-bound in-sulin, proteins having insulin-binding capacity weredemonstrated in the plasma and urine of diabetic ne-phrotics. Excretion of TCA-precipitable radioactivity inthe urine, antibody-bound or free insulin, correlated withthe degree of proteinuria in both diabetics and nondia-betics. Patients without proteinuria excreted no sig-nificant precipitable radioactivity. The findings indicatethat insulin-binding antibody and probably free insulin arelost in the urine of insulin-treated diabetic nephroticpatients.

Enzyme Activity in the Anatomical Units of Nephronsof Healthy and Acidotic Dogs. VIcTOR E. POLLAKAND HERMANNMATTENHEIMER, Chicago, Ill. (intro-duced by Robert M. Kark).

The mechanisms responsible for the renal response tometabolic acidosis are not clearly defined. The possiblerole of enzymes in the increased ammonia and protonexcretion of metabolic acidosis was therefore studied inhealthy dogs with normal renal function. Renal biopsieswere taken from four dogs after 7 control days and afterNH4C1 administration for 7 days (110 mEq/day). Afifth dog served as a control. With NH4C1 loading, NH,excretion increased fourfold (19 to 79 mmoles/day), andproton excretion increased threefold (4 to 14 mmoles/day). Glutamic oxalacetic transaminase (GOT), glutamicpyruvic transaminase (GPT), glutamic dehydrogenase(GDH), carbonic anhydrase (CA), malate dehydrogenase(MDH), and lactic dehydrogenase (LDH) activitieswere assayed quantitatively by ultramicrochemical tech-niques in the individually dissected and weighed glomeruli,proximal and distal convolutions, from 16-Az frozen-driedsections. In healthy dogs GOTand MDHactivities werehighest in the distal convolutions. GDH, CA, and LDHactivities were similar in proximal and distal convolu-tions. During NH.C1 administration, GOT activity in-creased significantly in proximal convolutions of twodogs, in distal convolutions in one, and in both in one.CA activity increased in proximal convolutions of threedogs. Preliminary observations indicated increased GPTactivity in metabolic acidosis. No significant changesoccurred in GDH, MDH, or LDH activities. Glutami-nase and condensing enzyme activities, measured inhomogenates, did not increase with NH4C1administration.In rats, glutaminase activity increased in prolongedmetabolic acidosis. Our experiments confirm previousobservations that no glutaminase adaptation is necessary

1392


for the production of the increased NH3 excreted in dogs.They suggest that in dogs transamination reactions mightplay a role in increased NH3 transfer to amino acids,which could then be deaminated.

Dissociation of Insulin Action in the Forearm of Man.DAVID RABINOWITZ AND KENNETHL. ZIERLER,* Balti-more, Md.

During studies of insulin effect in patients with diseaseof the hypophyseal-adrenocortical axis we noticed that,although there was insulin resistance with respect to glu-cose uptake by muscle, there was either normal or in-creased responsiveness to insulin with respect to othermetabolic events. Closely related observations havebeen made in excised rat muscle and adipose tissue bySteinke, Taylor and Renold, Vallance-Owen and Lilley,and Lowy, Blanshard and Phear. These reports leaveunanswered the question of whether certain circulatingantagonists of insulin prevent its action on muscle butnot on fat or discriminate the action of insulin on glu-cose uptake from its action on free fatty acid (FFA) re-lease. This is a report of our observations on the effectof insulin, administered by local arterial injection, onmetabolism of human forearm in situ in five patients withactive acromegaly, two with Cushing's disease, and inthree normal subjects in whom endogenous growth hor-mone activity was stimulated by a 66-hour fast. Me-tabolism was assessed by measurement of forearm bloodflow and arteriovenous concentration differences. Glucoseand potassium uptake by muscle in response to insulinwere less than half normal. Insulin-induced glucose andpotassium uptake by adipose tissue was probably lessaffected but was probably reduced. Insulin-induced in-hibition of FFA release from adipose tissue was un-checked in these patients. Indeed, the 66-hour fastedsubjects were so exquisitely sensitive that arterial FFAconcentrations fell in response to an estimated maximumof 3 1uU of insulin per ml recirculating plasma, yet glu-cose uptake in the forearm, where the local concentrationof insulin was 100 times greater, was almost abolished.These studies may dissociate the action of insulin onpotassium and on glucose uptake from its action on FFArelease. They make a unitary explanation of insulinaction difficult.

The Structural Basis for the Action of Neuro-hypophyseal Hormones upon the Amphibian UrinaryBladder. HOWARDRASMUSSEN* AND IRVING L.SCHWARTZ,* Madison, Wis. and Cincinnati, Ohio.

When neurohypophyseal hormones are applied to theisolated amphibian urinary bladder (the homolog ofthe mammalian renal collecting tubule), they produce anincrease in its permeability to water. This phenomenonwas the basis of the method employed to determine theeffect of changes in hormone structure upon biologicactivity. Twenty-seven hormone analogs were tested.All produced the same increase in permeability, but theconcentration required varied from 10V to 10V M, de-

pending upon analog structure. None of them was activein the presence of 10' M cysteine, which did not inhibitthe action of cyclic AMPor theophylline (agents whichmimic the hormonal action). Angiotensin at 10' Mwasinactive even though it mimicked the action (an oxytociceffect) of these hormones on the rat uterus. The basicrequirements for hormone action upon the bladder were:1) an intact disulfide bond, and 2) a ring of 6 or 7amino acids (20-23 membered ring). Neither the tripep-tide side chain nor any of the charged groups on the ringwas essential. Thus, hormone-receptor interaction canbe considered a two-step process: 1) the multipoint at-tachment of hormone to receptor site through hydrogen,hydrophobic, and ionic bonds; and 2) a thiol-disulfideinterchange between receptor sulfhydryl and hormonaldisulfide. A change of structure in a constituent aminoacid of the hormone molecule leads to a change in theaffinity of hormone for receptor (step 1). The thiol-disulfide interchange (step 2) appears to be the catalyticstep that initiates the chain of events which lead to thepermeability increase. Previous data on the binding oftritium-labeled hormone to both bladder and kidney tis-sue, and a current analysis, in Michaelis-Menten terms,of the competition between protons and hormone for anionized group on the receptor are compatible with thisproposal.

Myocardial Metabolism during Epinephrine-induced Ne-crosis. TIMOTHY J. REGAN, LAWRENCETROUM, PAT-RICK H. LEHAN AND HARPER K. HELLEMS,* JerseyCity, N. J.

In view of the important regulatory role of the sympa-thetic nervous system on myocardial function, the rele-vance of its myocardial metabolic alterations in thetransition from an inotropic to a pathophysiologic effecthas been examined in the intact dog. Left coronaryartery infusion of l-epinephrine was used to minimizesystemic effects. An increase in contractility (first de-rivative of left ventricular pressure) was achieved with5 ,ug per minute for 15 minutes in eight animals (a modestdose since 2.5 ,ug per minute was ineffective). Sequentialsubstrate extractions were followed by arterial-coronarysinus sampling. In the presence of a 10% coronary flowincrement (N20 method), there was a shift in oxidativemetabolism as evidenced by a myocardial respiratoryquotient rise. An early appearance of excess lactateproduction, indicative of glycolysis, was associated witha decreased extraction of glucose. Similarly, free fattyacid extraction by the heart diminished. This deficit inavailable substrate was compensated by extraction oftriglyceride, a moiety not usually extracted in significantquantity. To evaluate whether these metabolic effectscould be related to the genesis of necrosis and lipidaccumulation reported with large doses, the same con-centration was extended over 90 minutes in a separategroup of eight animals, resulting- in changes analogousto ischemic necrosis. Glutamic oxalacetic transaminaseand K+ rose significantly in the coronary sinus abovethe arterial levels by about 2 hours, when ventricular

1393


ectopic activity usually appeared and contractility de-clined. The metabolic concomitants included a persist-ence up to 60 minutes of the altered substrate transportobserved during the initial positive inotropic effect andthen reversion to control levels of free fatty acid extrac-tion and lipid utilization. Assumed that the metabolicevents of the first hour are critical to necrosis, increasedcarbohydrate oxidation may be responsible for triglyc-eride accumulation during its accelerated transport.Thus, persistent epinephrine activity in modest concen-trations may produce cardiac necrosis without coronaryflow reduction or relevant ionic alterations. A meta-bolic genesis related to the interaction of lipid andcarbohydrate transport and utilization is suggested.

Adrenocortical Modification of the Interdependence ofCalcium and Sodium Reabsorption in the Kidney.BRIAN H. B. ROBINSON, EDWARDB. MARSH, JR.,JOHN W. DUCKETT, JR. AND MACKENZIE WALSER,*Baltimore, Md.

Administration of adrenal steroids is usually followedby an increase in urinary calcium and a decrease inurinary sodium. However, in other circumstances, cal-cium reabsorption tends to vary with sodium reabsorp-tion. We have reported that, in dogs, the percentagereabsorption of filtered calcium parallels that of sodiumduring a variety of diuretic procedures. By contrast, inthe hypercalcemia of malignancy, and in dogs (or humansubjects) infused with calcium salts, calcium clearanceregularly exceeds sodium clearance, often many-fold.This homeostatic response fails to occur in hypercalcemicadrenalectomized dogs; calcium clearance remains lessthan sodium clearance. In normal human subjects, hour-to-hour fluctuations in sodium excretion throughouteach 24 hours were accompanied by parallel changes incalcium excretion. During infusion of sodium sulfate,calcium clearance was highly correlated with sodiumclearance (r=0.87 for 16 observations in six subjects).By administering mineralocorticoids or aldosterone an-tagonists, and then varying sodium excretion over a widerange, it is possible to determine whether the action ofthese steroids on the tubule includes inhibition of calciumreabsorption at any given level of sodium reabsorption.Clearances of sodium, potassium, calcium, magnesium,and inulin were determined during diuresis induced suc-cessively by water, mannitol and saline, and sodiumsulfate. In one normal subject given fludrocortisone onone occasion and aldosterone on another, average ratiosof calcium clearance to sodium clearance were 2.5 and1.7, respectively, as compared with 1.2 during a controlstudy. In another normal subject the average ratiowas 1.4 while on spirolactone, and 4.0 on fludrocortisone.In a third, the average ratio was 0.4 on spirolactone,compared with 3.5 in a control study. Two Addisoniansmaintained without mineralocorticoids yielded 0.9 and1.1, respectively; after deoxycorticosterone acetate theratios were 1.8 and 1.7. These results indicate thatcalcium reabsorption is linked with that of sodium; thisdependence is altered by mineralocorticoid activity.

Competitive Inhibition of Dibasic Amino Acid Transportin Rat Kidney. LEON E. ROSENBERGAND STANTONSEGAL, Bethesda, Md. (introduced by Robert M.Chanock).Data from renal clearance studies in dogs, and from

urinary amino acid patterns in patients with cystinuriahave suggested that the dibasic amino acids-lysine,arginine, ornithine, and cystine-share a common renaltubular transport site(s). The present study, utilizingrat kidney cortex slices, was undertaken to define themechanism of the mutual interference for tubular reab-sorption exhibited by these compounds. Isotopicallylabeled L-lysine, L-arginine, and L-cystine were concen-trated against chemical gradients during 30-minute incu-bations, and mutual inhibition was demonstrated withequimolar concentrations of L-lysine, L-arginine, andL-ornithine, but L-cystine did not share this commontransfer process. The specificity of this shared mecha-nism was shown by experiments in which L-alanine,L-phenylalanine, and L-histidine failed to inhibit L-lysineuptake, and by studies indicating that transport of glycine,L-phenylalanine, and L-histidine was unaffected by L-argi-nine. The inhibition of L-lysine by L-arginine was studiedduring 5- to 120-minute incubation periods, and analysisof the results, using a two-compartment model, suggestedan equal reduction in influx and efflux rates in the pres-ence of inhibitor. Utilizing diffusion and Michaelis-Men-ten kinetics, the inhibition of L-arginine transfer by L-ly-sine and L-ornithine was shown to result from competitiveinhibition for the saturable component of the transportprocess, and was not related to alteration in physicaldiffusion rates. In agreement with previous in vivoobservations, L-arginine and L-ornithine had a greateraffinity for the shared transport site(s) than had L-lysine,while the transport capacity for L-lysine exceeded thatfor L-arginine. The failure of cystine to share the trans-port process (es) common to the other dibasic aminoacids studied contrasts sharply to observations in cystin-uric patients, and suggests that such transport mech-anisms are species specific and may be under geneticcontrol.

Coupled Nonionic Diffusion- A Hypothesis to Accountfor Apparent Changes in Permeability to Weak Elec-trolytes. JOSEPH B. ROSENFELD AND WILLIAM B.SCHWARTZ,* Boston, Mass.

Although the transfer of weak electrolytes by nonionicdiffusion has been extensively studied, little considerationhas been given to the circumstance in which diffusion ofboth a nonionized acid and nonionized base occurs simul-taneously (coupled nonionic diffusion, CND) from asolution in which pH is not autonomously regulated.Theoretical considerations suggest that transfer of asingle weak electrolyte by nonionic diffusion will belimited by the selective loss of the nonionic species anddisappearance of the nonionic concentration gradient.However, simultaneous diffusion of a weak acid andweak base in the presence of the ionic members of eachconjugate pair (a proton acceptor and donor, respec-

1394


tively) should lead to continuous regeneration of eachnonionic species with progressive loss of the ionic speciesvia CND. In order to test this hypothesis, 100 mMsolutions of ammonium chloride (pH 8), sodium bi-carbonate, and ammonium bicarbonate were instilled intodog bladders isolated by ureteral and urethral ligation.After instillation of ammonium chloride (pH 8), NH,+concentration fell by a mean of only 20 mEq per L in 8hours, whereas after instillation of ammonium bicarbon-ate, concentration fell by a mean of 65 mEq per L(p <0.001). Similarly, after the instillation of sodiumbicarbonate, bicarbonate concentration fell by only 10mEq per L, whereas in the ammonium bicarbonate ex-

periments it fell by approximately 58 mEq per L(p < 0.001). It is proposed that two processes were

responsible for the ionic loss from ammonium bicarbon-ate solutions. In part, ammonium and bicarbonate ionswere lost directly along their concentration gradients,just as appears to be the case when each is present indi-vidually as the salt of a strong electrolyte. In addition,a large loss occurred indirectly by the continuous trans-formation of NH.' and HCO3- into NH3 and CO2. Thesedata suggest that the migration of weak electrolytes froma solution in which a constant pH is not maintainedautonomously can be increased by CND, and that thiswill induce an apparent change in the permeability toeach ionic species.

Electrocardiographic and Hemodynamic Observations dur-ing Selective Coronary Cineangiocardiography. RICH-ARD S. Ross, NEIL SCHWARTZ, ROBERT A. GAERTNERAND GOTTLIEB C. FRIESINGER, Baltimore, Md. (intro-duced by A. McGehee Harvey).

The electrocardiographic and hemodynamic changesassociated with the injection of roentgen-opaque ma-

terials into the coronary arteries were studied in man

and dogs. The electrocardiogram was recorded duringthe selective injection of 5 ml of contrast material intoeach coronary artery in 40 patients. In all patients withangiocardiographically normal coronary arteries, the Twave vector shifted away from the part of the heartperfused by the injected vessel during the minute afterinjection. This pattern was altered in the presence ofcoronary artery disease. In two patients, collateral linksbetween the right and left coronary circulations were

demonstrated. In these, after injection into a singleartery, right and left coronary T wave changes ap-peared sequentially in an order corresponding to theobserved flow of contrast material. Serial measure-

ments of left ventricular pressures were made in 12patients during arteriography. An elevation of leftventricular diastolic pressure and an increase in the "a"wave preceded the onset of anginal pain in two patients.The administration of nitroglycerine was followed byrelief of pain and a drop in left ventricular diastolicpressure. These observations probably explain the ele-vated pulmonary wedge pressure and prominent "a" wave

of the apex cardiogram reported in angina pectoris. Inopen-chested dogs, the electrocardiogram, left ventricular

diastolic pressure, coronary blood flow, and myocardialcontractile force were recorded during selective injec-tion of contrast material into the coronary arteries.During the first 10 seconds, coronary blood flow, myo-cardial contractile force, and arterial blood pressure de-creased, and T wave changes developed. The T wavechanges persisted for 45 seconds, even though arterialpressure and myocardial contractile force had returnedto normal by 20 seconds, and coronary blood flow in-creased to 150% of control. The increase in coronaryflow persisted for 1 minute or longer.

Colloid Osmotic Pressure as a Possible Regulator ofAl4bumin Synthesis. M. A. ROTHSCHILD, M. ORATZ,E. C. FRANKLIN* AND S. S. SCHREIBER, New York,N. Y.A colloid osmotic regulatory mechanism was suggested

to explain the hypoalbuminemia following dextran in-fusions. It is suggested that a similar mechanism isresponsible for the hypoalbuminemia associated withhyperglobulinemia. To study this hypothesis further,hypergammaglobulinemia was produced in seven femalerabbits by the prolonged injection of a polyvalent pneu-mococcal vaccine. Albumin metabolism was measuredwith albumin-I' during a control period and again after4 to 6 months of vaccine administration when hyper-gammaglobulinemia and hypoalbuminemia had appeared.In four rabbits albumin metabolism was restudied afterthe plasma protein values had returned to normal andagain during the anamnestic response. After primaryimmunization, albumin levels fell from a mean controlvalue of 3.69 to 2.78 g per 100 ml and 'y-globulin levelsrose from 0.75 to 3.80 g per 100 ml. No change inplasma volume occurred. The albumin pool decreased by21.9% even though albumin degradation decreased by23.7%; thus albumin synthesis must have decreased evenfurther to account for this loss in total exchangeablealbumin. After the anamnestic response, mean albuminlevels decreased from 3.69 to 2.18 g per 100 ml. 'v-Glob-ulin levels rose from 0.84 to 6.07 g per 100 ml, with anincrease in the mean plasma volume of 24.5%. Albumindegradation failed to show any change until the plasmaalbumin concentration had fallen to low values, usuallyby 7 to 10 days after vaccine administration. Thereafter,albumin degradation decreased 19.5% below controlvalues. The fractional rate of albumin-I' degradationwas unaffected. Albumin synthesis decreased by amean of 23.2%, resulting in a decrease in exchangeablealbumin of 27.2%. The production of hypergammaglobu-linemia results in hypoalbuminemia and a decreased ex-changeable albumin resulting from decreases in albuminsynthesis. This alteration in albumin synthesis is con-sistent with the concept of an osmotic regulatorymechanism controlling albumin synthesis.

Investigation of the Mechanism of Duroziess Murmur.GEORGEG. ROWE* AND SKODAAFONSO, Madison, Wis.

Radiographic contrast substance has been infused witha constant injection syringe into the femoral artery and

1395


its motion has been recorded cineangiographically withan intensifier. In the human subject without aortic in-sufficiency or other abnormal retrograde flow of bloodfrom the central aorta, and in the normal dog, contrastmaterial during systole passes rapidly down the femoralartery; during diastole there is minimal backflow. Ifthe aortic valve of a dog is destroyed so that it becomesincompetent, the amount of femoral arterial backflowincreases greatly. In human subjects with aortic in-sufficiency or patent ductus arteriosus with a large aortic-pulmonary flow, there is considerable femoral backflowand a Duroziez's murmur is heard. In one subject with anormal heart but severe anemia, Duroziez's murmur andincreased femoral backflow were demonstrated. In gen-eral, the degree of femoral backflow seems to correlatewith the intensity of the Duroziez's murmur. Present ob-servations seem to afford visual confirmation of a longstanding clinical presumption.

Investigation of Possible Chemoreceptors in the Venousand Pulmonary Circulation for the Regulation of Res-piration. HERBERT A. SALTZMAN AND HERBERT 0.SIEKER,* Durham, N. C.

The presence of respiratory chemoreceptors in thevenous or pulmonary circulation is suggested from theventilatory response to exercise and by the recent demon-stration of carotid body-like structures in the lungs. Toinvestigate this possibility, changes in pulmonary (PA)and brachial (BA) arterial blood gases were correlatedwith minute ventilation (MV) and CO2 elimination(Vco2) following release of arterial tourniquets, therectal infusion of CO2, and exercise in 23 normal subjects.The release of leg arterial tourniquets was followedwithin a minute by transient increases in mean PAPco2 of 3.5 mmHg, MVof 0.6 L per min, and BA Pco2of 1.4 mmHg, while blood pH decreased in a reciprocalmanner. The rectal infusion of 1,200 to 2,500 ml of100% CO2 in 15 subjects produced statistically significantand prolonged (15 to 35 minutes) increases in mean PAPco2 of 3.0 mmHg, MVof 1.6 L per min, and Vco% of50 ml per min, while PA pH decreased 0.02 U. The BAPco2 and pH did not change significantly from controlvalues. Rectal infusion of comparable volumes of air orN2O did not produce significant changes in ventilation.Exercise was associated with increases in PA Pco, of6.0 mmHg, MV of 16.9 L per min, and Vco. of 400ml per min, while PA pH decreased 0.02 U. The BAPco, fell 1.2 mmHg, but the pH was essentially un-changed. Release of leg arterial tourniquets producedbrief changes in ventilation in association with blood gasalterations in both central venous and arterial blood. Thetime sequence of events suggests that the arterial chemo-receptors are responsible for this respiratory change.The increase in central venous Pco2, minute ventilation,and CO, infusion suggests the presence of chemoreceptorsin the venous system or pulmonary circulation whichcontribute to ventilatory regulation. Such a mechanismmay also explain in part the ventilatory response toexercise.

Cardiovascular Actions of Corticosteroids in NormalSubjects and in Patients with Acute Hypotension.M. P. SAMBHI, M. H. WEIL, V. N. UDHOJI AND L.ROSOFF, Los Angeles, Calif. (introduced by SamuelI. Rapaport).Observations on 11 normal subjects and 10 patients

with shock, receiving metaraminol and norepinephrine ingraded doses before and after a pharmacologic dose ofa corticosteroid, provided no support for the hypothesisthat the pressor effects of vasopressor agents are potenti-ated by corticosteroids. However, a striking hemody-namic effect followed within 0.5 to 3 hours of theadministration of corticosteroids. This led us to studythe circulatory effects of corticosteroids alone in ninenormal subjects and in five patients with shock associatedwith sepsis (3) and pancreatitis (2) In no instance wasa pressor effect observed and intra-arterial pressure andcentral venous pressure were essentially unchanged. Innormal subjects, after the injection of 500 mg cortisol orprednisolone and dexamethasone in equivalent amounts,the cardiac index was uniformly increased from 3.0 (2.4to 4.1) to 3.9 (3.5 to 4.8) L per min per m'. Thechange is highly significant (SED = 0.214, p < 0.01).Peripheral resistance declined by an average of 25%. Inpatients with shock, arterial pressure was 74 (62 to 88)mmHg and cardiac index was 2.3 (1.1 to 3.8) L permin per mi. After injection of 20 to 100 mg dexametha-sone, no change occurred in one patient, but a markedincrease in cardiac output ranging from 23 to 57%occurred in the remaining four, with a significant reduc-tion in peripheral resistance (10 to 53%). Pharma-cologic doses of corticosteroids increase flow and decreaseresistance. These actions oppose the primary defects ofreduced cardiac output and detrimental vasoconstriction,the latter particularly when augmented by vasopressortherapy during progression of shock.

Biologic Potential of Transcortin. AVERY A. SANDBERG*AND W. R. SLAUNWHITE, JR., Buffalo, N. Y.

Previous observations, based on indirect evidence inhuman subjects, indicated that transcortin-bound cortisolwas 1) biologically inert and 2) unavailable for catabo-lism. The partial purification of transcortin made it pos-sible to study two of its potentials. In a series of experi-ments it was shown that cortisol-induced glycogen depo-sition in the liver of adrenalectomized mice could becompletely abolished by the intravenous injection oftranscortin. Thus, transcortin rendered cortisol biologi-cally inert. The present study was undertaken to ascer-tain the second potential of transcortin. Human or ratliver homogenates or microsomes were incubated for 30minutes with C1'-cortisol (without carrier and as much as25 lsg of cortisol) and an excess of an electron donor(TPNH). Almost quantitative reduction of cortisoltook place, resulting in pregnane and allopregnane de-rivatives with human and rat liver, respectively. Theaddition of human serum albumin did not change theresults. In contrast, the presence of as little as 0.06 to

1396


0.1 mg of transcortin resulted in the recovery of sub-stantial amounts of unchanged cortisol (ca 30%). Therewas a linear relationship between the amounts of cortisolunmetabolized and the concentration of transcortin in theincubation medium. With 2 mg or more of transcortinlittle reduction of the cortisol nucleus occurred. Incuba-tion with normal plasma or plasma with high transcortinlevels (in pregnancy or after estrogen therapy) showedthe latter to contain three times as much transcortin asthe former. These findings indicate that the affinity forcortisol manifested by transcortin is higher than thatshown by the liver enzyme systems responsible for thereduction of the steroid. On the basis of these ex-periments, the hypothesis is advanced that variations intranscortin levels may lead to abnormal metabolic ef-fects of cortisol, not evident in the commonly measuredparameters.

Properties of Transferable Pyrogen in ExperimentalVaccinia Infection. GERALDSANDSON,ROSEPAVIA ANDA. I. BRAUDE,* Pittsburgh, Pa.

It is not known if microbial products are responsiblefor pyrogenicity of serum during infectious fever. Toresolve this question, circulating pyrogen in vaccinia in-fection was inoculated into immune rabbits because theresponse to microbial, but not endogenous, pyrogen mightchange. Vaccinia virus, propagated in cultures of rabbitkidney, was selected for study after it proved non-pyrogenic on intravenous injection into normal rabbits.Seventy-two hours after intracorneal or intravenous in-fection with 105 viable virus units, rabbits developed feverreaching 3.50 F and lasting 5 days. Intravenous injectionof 30 ml serum from febrile rabbits produced immediatemonophasic fever reaching 1.50 F in 1 hour in all 23 nor-mal rabbits examined. No viable virus was found in pyro-genic serum. In contrast to normals, immune rabbits (re-covered from intravenous vaccinia infection) eitherremained afebrile (66%) or exhibited immediate biphasichyperpyrexia (33%) after receiving intravenously 30 mlserum from febrile rabbits infected intravenously. Be-cause immune animals showed similar altered responsesto 105 viable virus particles, viral antigen was suspectedin pyrogenic serum. Influence of antibody on pyrogenicityof virus was next examined by mixing 105 viable vacciniaunits with serial dilutions of immune serum before intra-venous injection into normal rabbits. Virus plus immuneserum diluted up to 1:4 was nonpyrogenic, but virus plusimmune serum diluted 1:16 to 1:64 produced immediatemonophasic fever. Virus plus immune serum diluted be-yond 1: 64 to 1,024 produced no fever until infectiondeveloped at 72 hours. These results demonstrate thatfever of vaccinia infection is accompanied by a noninfec-tious serum pyrogen which produced different responsesin immune and nonimmune animals; and that its pyro-genic effect can be duplicated by mixing virus withappropriately diluted immune serum. It is possible, there-fore, that fever of experimental vaccinia results fromcirculating complexes of vaccinia antigen and antibody.

The Nongenetic Nature of the Alteration of Serum Cho-lesterol Induced by Deoxyribonucleic Acids. J. PHILIPSAVITSKY, New York, N. Y. (introduced by LouisLeiter).A single injection of DNA produces in rabbits a de-

crease in serum cholesterol and phospholipids lasting atleast 2 years. Microgram quantities of DNA fromhomologous and heterologous mammalian tissues areeffective in lowering cholesterol levels of animals withhigh normal values. Chemical, physical, and enzymaticevidence shows that the active material is purified, poly-merized, undenatured, 2-stranded DNA. It appeared im-portant to determine whether the prolonged alteration ofserum cholesterol regulation reflected the introduction ofspecific, transferable, genetic information into adult mam-mals. A standardized assay of DNAphysiologic activitywas developed: A citrate-saline solution of 150 ,ugDNA was injected intravenously into each of ten nor-mal rabbits with appropriate controls (citrate inhibitsthe plasma deoxyribonuclease). Serum cholesterol andphospholipids were determined prior to the injection andat monthly intervals for 6 months thereafter. Using thisassay, three groups of experiments were performed.1) Eighteen DNA preparations from five mammalianspecies and seven normal tissues have been tested. Thepreparations lowered the serum cholesterol with no speci-ficity observed as to species or tissue source of theDNA. 2) Individual normal rabbits were selectedwhich maintained levels of serum lipids for 6 months withminimal variation. The serum cholesterol levels of the"high cholesterol" group averaged fourfold that of the"low cholesterol" rabbits. DNA prepared from bothgroups was equally effective in the assay. 3) Normalrabbits were again selected for minimal variation ofserum cholesterol. Leukocytes were obtained from theseanimals and DNA prepared. Each rabbit was injectedwith its own DNA. The autologous DNA effectivelylowered the serum cholesterol and phospholipids, com-parably to all other preparations. The data indicate thatthe altered serum cholesterol regulation induced by heter-ologous, homologous, and autologous DNA is not associ-ated with the transfer of specific, genetic information.

Metabolism and Membrane Transport of ErythrocyteStroma. STANLEY L. SCHRIER AND LYDIE S. DOAK,Palo Alto, Calif. (introduced by Frederic L. Eldridge).Membrane transport frequently involves movement

against a gradient and therefore requires energy andspecific transport systems. Studies on metabolism andtransport systems of erythrocyte membranes (stroma)were undertaken with the intent of providing informa-tion on mechanisms of membrane transport and intra-corpuscular red blood cell defects. Stroma was pre-pared by dialyzing normal human erythrocytes againstincreasingly hypotonic saline solutions. The resultingmembranes were white, homogenous, biconcave disc-shaped, without rents, and contained approximately 1.5%oof hemoglobin by dry weight (or less than 0.03%o of theoriginal hemoglobin). Studies of carbohydrate metab-

1397


olism in stroma showed that stroma did not utilizeglucose, fructose, inosine, adenosine, ribose, or ribose-1-phosphate, even when appropriate cofactors were added.However, ribose-5-phosphate (R5P) was promptly me-tabolized to ribulose-5-phosphate (RU5P) indicatingpresence of pentose phosphate isomerase. We searchedfor further products of R5P metabolism because conver-sion of R5P to glyceraldehyde-3-phosphate (GA3P) wouldindicate the presence of a triple enzyme sequence of thepentose phosphate pathway in stroma: pentose phosphateisomerase, pentose phosphate epimerase, and transketolase.Our results demonstrated that stroma does catalyze theconversion of R5P and GA3P. Kinetic studies showedthat appearance of GA3P followed production of keto-pentose phosphate and that the rate of both reactionsdepended on stroma concentration. In addition, specificassays demonstrated relatively high concentrations ofthe enzyme glyceraldehyde-3-phosphate-dehydrogenase(GA3PD) in stroma. Our studies have shown the pres-ence in red cell membranes of a metabolic pathway whichusually involves three sequential enzymes in the pentosephosphate pathway leading to the production of GA3Pfrom R5P. The presence of GA3PD in stroma togetherwith a system for production of its substrate, GA3P,could result in an energy-yielding oxidative phosphoryla-tion in the membrane, and thus account for inorganicphosphate transport, partly explain potassium transport,and perhaps other energy-requiring systems.

Effect of Calcium and of Phosphate Intake on Radiostron-tium Absorption in Mice. ROBERT SCHWARTZ, Cleve-land, Ohio (introduced by George J. Gabuzda, Jr.).

Although strontium and calcium are metabolized simi-larly by bone, these ions are transported differently bykidney and by intestine. Absorption of strontium is con-ditioned by the calcium content of the diet. Since little isknown about the influence of dietary phosphate, the ef-fects of independently varying calcium and phosphateintakes on radiostrontium absorption were studied inmice. The progeny of mice, fed a vitamin D-free dietcontaining adequate calcium and phosphate, were weanedat 18 days. Groups of 10 to 15 litter mates were fedad libitum synthetic diets, the calcium and phosphatecontents of which were varied. Radiostrontium 85 (Sr8")was given after 18 to 21 days. Whole body counting ofeach mouse was then performed daily for 1 week. Bodycontent of Sr"5 given p.o. was inversely related todietary calcium level when the diet contained adequatephosphate. In the absence of phosphate, Sr' retentionwas increased. For example, Sr" retention at 24 hourswas 18 and 62% when the diets contained 0.94 and 0.02 g% phosphate, respectively, and 1.2 go calcium. Theseeffects were evident by 5 hours after administrationof Sr'. Approximately 80% of isotope given intra-venously was retained after 24 hours. Altering dietarycalcium and phosphate had only minimal effects; slightlyless isotope was retained when the diet contained littlephosphate but adequate calcium. The rates of loss ofisotope after the first 24 hours were similar for the two

routes of administration. These were not influenced bygiving vitamin D, although feeding the low phosphatediet caused mild chemical rickets. The differences notedbetween the oral and intravenous experiments indicatethat absorption of Sr" in mice is impaired by the presenceof phosphate in the diet, which presumably forms insolu-ble salts with strontium.

The In Vitro Metabolism of 6-Mercaptopurine in HumanLeukemic Leukocytes. JAMES L. SCOTT AND JOSEPHV. MARINO, Los Angeles, Calif. (introduced by JohnS. Lawrence).

In animal tumor cells and bacteria, 6-mercaptopurine(6-MP) has been found to inhibit de novo purine syn-

thesis, specifically affecting the conversion of other pu-rines to adenine prior to its utilization for nucleic acidsynthesis. In the present studies evidence of this effectwas sought in leukocytes isolated from the blood ofpatients with acute leukemia. De novo purine synthesiswas measured by the incorporation of formate-C" intothe nucleic acid purines. After incubation with apharmacological concentration (6 X 10' mmoles/ml) of6-MP, incorporation was lower than control values inthree cell populations, unchanged in two, and higher inthree others. Adenine and guanine incorporation wereinfluenced equally, so that the effect on adenine utilizationwas not a specific one. These in vitro effects could notbe clearly related to clinical responsiveness to 6-MP. Toclarify these results, C"-labeled 6-MP was incubatedwith acute leukemic leukocytes and the metabolic productsidentified by radiochromatography. Conversion of 6-MPto physiological purines and the utilization of these com-pounds for nucleic acid synthesis by the cells of patientswith 6-MP-resistant leukemia were observed. This capa-bility may represent one mechanism of drug resistance.Since preformed purine is utilized in preference to thatsynthesized by the de novo pathway, inhibition of purinesynthesis by 6-MP may reflect drug inactivation in thesecells.

Periodic Hemodialysis for the Rehabilitation of Patientswith Terminal Uremnia. BELDING H. SCRIBNER,* Se-attle, Wash.

A new, pumpless, low temperature hemodialysis systemand a technique of long-term cannulation of blood ves-sels have been developed and perfected to the point whereit is now possible to restore to good health selected youngadults with chronic kidney disease who would otherwisedie of terminal uremia. The first four patients werebegun on a weekly hemodialysis during the winter of1960. One of these patients, with a creatinine clearanceof about 1 ml per min, continues to work full time. Twoothers, who have been anuric for the last 18 months, leadcomfortable sedentary lives. The fourth died of a myo-cardial infarction after a year of hemodialysis therapy.Hypertension, which became malignant in one patient,has been satisfactorily controlled solely by a low saltdiet and ultrafiltration of extracellular fluid during

1398


dialysis. About a year ago two of the four patients de-veloped a slowly progressive peripheral neuropathy andall complained of uremic symptoms prior to each weeklydialysis. Switching from a long weekly dialysis to 12 to16 hours of dialysis twice a week has resulted in elimina-tion of predialysis symptoms and improvement in neu-ropathy. In the last 8 months four additional patientshave been added to the program. Each was carefully se-lected as being an emotionally stable, mature, cooperativeadult under the age of 40 who was free of serious com-plications of hypertension. Each was invalided byterminal uremia at the time treatment was begun.Hemodialysis once or twice weekly has restored each togood enough health to return to work; one patient mar-ried recently. All patients require weekly transfusionsto prevent severe anemia. Accumulation of iron mayrepresent a potential threat. Exhaustion of cannula sitesis another potential problem, but recent improvements incannula design may have solved this problem. It isconcluded that selected young adults with terminal uremiacan be restored to good health with periodic hemodialysis.

A Mechanism of Action of Colchicine in Acute GoutyArthritis. J. E. SEEGMILLER, R. RODNEYHOWELLANDSTEPHEN E. MALAWISTA, Bethesda, Md. (introducedby Gordon M. Tomkins).

A role for uric acid in the genesis of acute goutyarthritis has been generally denied, and the mechanismof action of the specific therapeutic agent, colchicine, hasremained an enigma. We have recently shown that theintra-articular injection of microcrystalline sodium urateproduces an inflammatory response in both gouty andnongouty individuals which is in some cases indistinguish-able from spontaneous acute gouty arthritis. A constantfeature of both the induced and the spontaneous inflam-matory reaction is a phagocytosis of the urate crystalsand production of large quantities of lactic acid by theleukocyte-rich effusion obtained from the involved jointwhen incubated in vitro. Wehave now observed that theamount of inflammatory reaction and phagocytosis is sub-stantially diminished by treatment of patients with thera-peutic amounts of colchicine prior to injection of thecrystals. Therapeutic levels of colchicine also resultedin a diminished phagocytosis by leukocytes when testedin vitro, as shown by direct count of intracellular crystalsas well as a diminished destruction of microcrystallinesodium urate-6-C"' by leukocytes to yield C1402. Thera-peutic amounts of colchicine administered in vivo alsohad a marked effect on the metabolism of leukocytes.The threefold increase in C1402 production from glucose-1-C'4 by leukocytes, which is induced by phagocytosisin vitro, was completely prevented by plasma from thesame subject after receiving a therapeutic dose of colchi-cine. Wepropose that the diminished metabolic activityof the leukocyte in response to colchicine results in adiminished phagocytosis and acid production which serveto interrupt the cycle of new crystal deposition, inflamma-tory response, phagocytosis of crystals, and increased acid

production that seems to be essential for continuation ofthe acute attack of gouty arthritis.

Observations on Cataract Formation in the Newborn ofPregnant Rats Fed a High Galactose Diet. STANTONSEGAL ANDHOWARDBERNSTEIN, Bethesda, Md. (intro-troduced by Jacob Robbins).The toxic effects of maternally ingested galactose on

the human fetus with congenital galactosemia are as yetundetermined. Since cataract formation occurs in galac-tosemia and is a known manifestation of galactose toxicityin rats, experiments were performed to evaluate in uteroeffects of this sugar by ascertaining the incidence ofcataracts in the newborn. A diet of 40% galactose wasstarted in pregnant Sprague-Dawley rats at differenttimes from mating until 3 days before term and con-tinued to delivery. Lenses of the newborn were examinedvisually and by slit-lamp. All animals in 15 litters hadnuclear cataracts in various stages of opacification. His-topathologic examination revealed degeneration of centrallens fibers similar to early nuclear galactose cataracts inhumans. Galactose-l-C14 oxidation to C140 in vitro wasassayed in liver mince from 14- and 20-day fetuses andfrom 1, 20, and 60 day old rats. In the 14-day fetus, 9%per 100 mg of liver was oxidized in 1 hour; in the olderanimals, the rate was 2 to 3%. The higher oxidationby the youngest liver was related to glucose oxidation,since the ratio of C402 from galactose and from glucose-1-C'4 was 1.2. This increased to 2.5 in the 20-day rat,indicating a maturation of galactose conversion to glucose.Although fetal liver can metabolize galactose, the resultsindicate that maternal ingestion of large amounts ofgalactose is toxic to the fetus in utero even after as littleas 3 days of feeding. The ease of cataract formation inthe fetus may be due to a lower total capacity of theliver to metabolize galactose due to its small size. Thisstudy indicates that in the human galactcemic fetus, nor-mal galactose intake by the mother mad be toxic andsuggests the need for galactose-free diet$ for pregnantmothers who have had galactosemic children.

A Novel Metabolic Pathway of Fat Absorption in theIntestinal Mucosa. JOHN R. SENIOR AND KURT J.ISSELBACHER,* Boston, Mass.

A new pathway has been demonstrated whereby themajor products of fat digestion may be resynthesized totriglycerides in the intestinal epithelial cells during lipidabsorption. After the intraluminal hydrolysis of the in-soluble dietary triglycerides to free fatty acids and lowerglycerides, especially monoglycerides, these products read-ily associate with conjugated bile salts to form ultra-fine aggregates, or micelles. Significanit quantities ofmonoglycerides are absorbed, and we have shown theirincorporation into triglycerides, the main component oflymph chylomicrons. An enzyme has been found in gutmucosal cells that catalyzes the direct recombination ofmonoglycerides and activated fatty acids to diglycerides,which then are esterified to triglycerides. This mecha-nism for the direct esterification of monoglycerides has not

1399


been found in the other tissues in which active triglycer-ide synthesis occurs, and therefore may be unique to theintestinal mucosa. Utilizing techniques developed re-cently in this laboratory for fractionation of jejunalepithelial cells, this enzyme, designated monoglycerideacylase, has been shown to be localized in highest specificactivity in the microsomes, as has been found also forthe long chain fatty acid activating enzyme, thiokinase.By means of chemically-synthesized palmityl-l-C"-CoAand glycerol-labeled monopalmitin-C14, very rapid andimpressive condensation of the activated fatty acid andmonoglyceride was demonstrated. Monoglycerides werefound to be better acceptors of the palmityl-CoA thanL-a-glycerophosphate, previously considered the "natural"acceptor. No significant amounts of phosphorylated in-termediates were detected, and neither ATP nor othercofactors were required. In summary, the enzymes neces-sary for triglyceride synthesis are associated with themicrosomes, a fraction principally derived from mem-branes of the endoplasmic reticulum, within which lipiddroplets accumulate during fat absorption, as seen byelectron microscopy. The observed monoglyceride "shunt"to triglyceride formation appears to be a mechanism bywhich considerable amounts of fat can be made availablerapidly for export to the rest of the body via lymphchylomicrons.

Increased Susceptibility to Pyelonephritis during AcuteHypertension Induced by Vascoconstrictor Agents.ALVIN P. SHAPIRO AND ROGERK. JONES, Pittsburgh,Pa. (introduced by Jack D. Myers).

Increased susceptibility to heniat genous pyelonephritiswas demonstrated previously in rats made hypertensiveby renal artery constriction or deoxycorticosterone.Chronic hypertension however may injure the kidney andfavor renal infection. To determine the effects of hemo-dynamic changes alone, susceptibility to pyelonephritisduring an acute hypertensive episode was evaluated. Tenug of angiotensin given intravenously to normotensive

rats elevated blood pressure to 172 + 14 mmHg within2 minutes, with return to baseline by 5 minutes. Threegroups then were studied: group A (44 rats) received10 ug of angiotensin i.v. in 0.5 ml of broth containing 10,Escherichia coli per ml; group B (45 rats), E. coli inbroth alone; and group C (16 rats), angiotensin in sterilebroth. Rats were sacrificed at 1, 4, and 24 hours, and1 and 2 weeks after injection. Kidneys from rats ingroup C were sterile and displayed no gross or micro-scopic alterations. At 1 and 4 hours, numbers of E. coliin kidneys of groups A and B were not different. How-ever, at 24 hours, 1, and 2 weeks, the mean logarithms ofbacteria per g of kidney in group A were 5.768, 4.554,and 3.706, respectively; in group B at these periods themean logs were 3.811, 2.130, and 1.020 (p < 0.001). Bac-teremia did not account for the differences. Gross pyelo-nephritic lesions were present in kidneys of group A asearly as 24 hours (11%), and were 54 and 67% at 1 and2 weeks, respectively. In group B, lesions were absentat 24 hours and present in only 8% at 1 and 2 weeks

(p = 0.001). Susceptibility similarly was increased with10 ,ug of norepinephrine plus E. coli, but to a lesser de-gree. Only 30% of kidneys in these rats demonstratedgross lesions at 1 and 2 weeks, and increase in renalbacteria was only tenfold. Enhanced susceptibility topyelonephritis during transient elevation of blood pressureby vasoconstrictor agents supports a primary role forhypertension in the pathogenesis of pyelonephritis.

Prognosis of Tropical Sprue. T. W. SHEEHY, E. PEREZAND M. H. FLOCH, San Juan, Puerto Rico (introducedby William H. Crosby).The prognosis of tropical sprue was evaluated by

studying the course of the disease and the fate of thejejunal lesion in 46 Puerto Rican patients. These pa-tients were divided into two groups. Group I consistedof 20 patients with an initial attack of classical sprue, andgroup II consisted of 26 sprue patients already treatedfor periods ranging from 3 to 5 years. Both groupswere hospitalized on a metabolic ward and subjected tobiochemical absorption studies (D+ xylose, vitamin A,carotene, Coe-vitamin B12), measurement of fecal fat,small bowel roentgenological motor meal, and j ej unalbiopsy. These studies were repeated every 6 months fora 2 to 3 year period. All patients were treated withfolic acid or folic acid and vitamin B12. The immediatetherapeutic results were partial amelioration of intestinalsymptoms, elimination of megaloblastic anemia, and clini-cal improvement. Intestinal function, as determined bychemical, histological, and roentgenological methods, wasrestored in 9 of the group I patients (45%o) after 6 to 36months of therapy. Seven patients from group II (27%)recovered after from 3 to 6 years of therapy. The histo-pathological changes, small bowel roentgenological pat-tern, and chemical tests correlate well in tropical sprue.Intestinal absorption is restored only when the jejunallesion heals. In contrast to celiac disease, the jejunallesion of tropical sprue is sometimes reversible, and sub-total villous atrophy is rare (2 of 46 cases). Moreover,even when subtotal villous atrophy is found, the villouspattern can revert to normal. Persistently impaired ab-sorption of radioactive vitamin B12, even in those patientswith recovery of all the other indices of intestinal func-tion, suggests permanent derangement of vitamin B12absorption in those who develop chronic sprue.

Arbovirus Infections of Puerto Rican Children with Ob-scure Neurological Syndromes. A. SHELOKOV, C. J.GIBBS AND D. MENDEZ-CASHION,Bethesda, Md. and SanJuan, Puerto Rico (introduced by Joseph E. Smadel).A serological survey performed on specimens from 145

healthy adult Puerto Ricans showed that 16% hadcomplement-fixing antibodies against B group arthropod-borne viruses, 3% had antibodies against A group, whilenone showed clearly positive reactions with C groupantigens. Ninety-four sets of paired sera obtained fromPuerto Rican children with obscure neurological syn-dromes were available for more definitive testing. The

1400


convalescent specimens from all of the cases were testedwith complement-fixing antigens representative of thethree groups-i.e., Eastern and Western (A); St. Louis,Ilheus, Dengue-2 (B); and Apeu (C)-prepared byCasals' technique. Only eight of these convalescent sera

reacted with any of the antigens. Seven of the eightcases were hospitalized with convulsions, and the otherwith meningitis. One of the infants with febrile con-

vulsions and bronchopneumonia displayed a diagnosticrise in Ilheus antibody. Another infant who had a

clinical diagnosis of purulent meningitis, and the onlypatient among the eight with pleocytosis, developed dur-ing convalescence antibodies against members of both Aand C groups. The serum obtained on the fourth day ofdisease was negative with Eastern, Western, and Apeu,while that obtained on day 18 had a titer of 1:32 withboth members of the A group and also with Apeu. Ofthe remaining six patients in the selected group, two hadlow levels of Apeu antibodies on admission and thesedid not change significantly during convalescence. Incontrast, the other four had no detectable antibodiesearly in the disease, but developed titers of 1:4 to 1:8against Apeu during convalescence. Two of these 4cases displayed hemiplegia during the acute illness. Itis evident from the serological studies that Apeu virus,or a related agent, is producing central nervous systemdisease in Puerto Rican children. In addition, membersof the A and B groups of arbovirus are producing in-fections among inhabitants of Puerto Rico.

Effect of Acetate and Pyruvate on Metabolism of Glu-cose-U-C14 and Palmitate-l-C1' in Isolated PerfusedRat Heart. J. C SHIPP, L. OPIE AND J. EVANS, Bos-ton, Mass. and Gainesville, Fla. (introduced by S. P.Martin).

The isolated rat heart perfused in a closed system withbicarbonate buffer permits a direct measure of the uptakeand certain cellular fates of labeled substrates; the C1402formed indicates substrate oxidation. In this study heartsfrom rats fasted overnight were used. Palmitate-l-C1'(0.5 mmole/L in 0.5% albumin) was readily extractedby the heart and about 80% of the uptake was recoveredas Co2 and "tissue fatty acids." The addition of glucose(5 mmoles/L) increased the uptake and the incorporationof palmitate label into CO2 and into "tissue fatty acids."Sodium pyruvate (5 mmoles/L) did not increase theuptake or fate of palmitate, and in certain conditions pal-mitate oxidation was reduced. Sodium acetate (5 mmoles/L) did not affect the uptake, but reduced the label in CO2while increasing the label in "tissue fatty acids." Glu-cose-U-C1' (5 mmoles/L) was extracted and recovered,primarily in C1402, lactate, and glycogen. The additionof palmitate (0.5 mMin 0.5% albumin), sodium acetate

(5 mmoles/L), or sodium pyruvate (5 mmoles/L) didnot affect glucose uptake, but the oxidation of glucose-U-C" to C"02 was reduced to 18, 8, and 36% of the control,respectively. The sparing effect on glucose oxidation ofall three subtrates was associated with more lactateformation, higher net glycogen and more incorporation

of glucose label into glycogen. These observationsstrongly suggest that palmitate, acetate, and pyruvate areoxidized by the heart in preference to glucose, and thatthe control for substrate oxidation is at the pyruvate-acetyl-CoA level.

A New Method for Measuring Minimum In Vivo Conz-centrations of Factor VIII Applied in Distribution andSurvival Studies. N. RAPHAEL SHULMAN,* VICTORJ. MARDERAND MERILYN C. HILLER, Bethesda, Md.

Small amounts of plasma (0.5 to 25.0 ml) containingnormal Factor VIII concentrations were in ected intosevere hemophiliacs. After a mixing period, blood wasdrawn for 2-stage prothrombin consumption determina-tions. In vivo Factor VIII concentrations from 0.02 to0.8% of normal, produced by dilution of Factor VIII inthe hemophiliac's plasma, were correlated with per centprothrombin consumed (15 to 85%). Standard curvesprepared in this way were identical in five different hemo-philiacs, and could be used to measure unknown in vivoFactor VIII concentrations with aproximately ± 20%error. This sensitive technique permitted measurementof Factor VIII survival in vivo for 12 half-lives over a5-day period. Previously available methods were usefulfor, at most, 6 half-lives, covering less than 2 days. Pro-longed survival measurements gave decay curves withtwo exponential components which were not previouslydiscernible. The initial component had a ti of 4 to 5hours (primarily extravascular diffusion) ; the secondcomponent had a tl of 9 to 11 hours (degradation andback diffusion). These two components, which did notvary with the amount of Factor VIII administered, in-dicated an extravascular compartment approximately 1.5times the plasma volume. Fresh plasma, infused at ratesof 3,000 ml per 50 kg body weight in 6 hours, gaveFactor VIII concentrations of only 40 to 50%, andmarkedly expanded the hemophiliac's plasma volume.Factor VIII concentration fell below the therapeutic 30%level within the next 7 hours. A recently available com-mercial Fraction I preparation (Merck Sharp & Dohme,1 g/50 ml), infused at rates of 1,500 ml per 50 kg bodyweight in 3 hours, gave Factor VIII concentrationsgreater than 100% without significantly increasing theplasma volume. Therapeutic levels greater than 30%persisted for the next 25 hours. The data support recentclinical observations suggesting that Fraction I is betterthan plasma for treating hemophilia.

Operation of the Renal Countercurrent Mechanism dur-ing Water Diuresis. MILES H. SIGLER, Philadelphia,Pa. (introduced by Earl S. Barker).

The operation of the renal countercurrent mechanismduring water diuresis in the dog was studied by meansof kidney tissue analysis. Kidneys were removed fromeight dogs during maximal water diuresis (urine osmo-lality 50 to 200 mOsm/kg) and frozen in liquid nitrogen(removal and freezing time 15 to 30 seconds). The samewas done in seven moderately dehydrated dogs (osmo-

1401


lality 900 to 1,700 mOsm/kg). Sections of inner papilla,outer papilla, outer medulla, and cortex were analyzed.Papillary sodium concentration was reduced in waterdiuresis (128 mEq/L) as compared to the hydropenicgroup (230 mEq/L) with a consequent diminution inthe papillary: cortex sodium gradient. However, totalsodium content of the inner papilla of the hydropenicgroup (1.45 ± 0.29 mEq/g dry solids) was only slightlygreater than the sodium content in the diuretic group(1.28 + 0.16 mEq/g dry solids). The difference betweengroups was not quite statistically significant (p = 0.06).Increase in inner papillary water content in the diureticgroup (90.4 vs 85.1% in hydropenia) appears to be aprimary factor in the decrease in papillary sodium con-centration. These was no evident correlation betweenGFR, serum sodium level, state of water balance, andmeasured papillary sodium content. During water di-uresis there was a slight increase in papillary K+ content,with a papillary: cortex gradient of 1.18 + 0.18 (vs 0.96+ 0.02 in hydropenia, p < 0.02). Water diuresis resultedin marked depletion of papillary urea (17.24 ,Amoles/gwet tissue vs 220 /Amoles in hydropenia). This occurredin conjunction with an increase in the urea: creatinineclearance ratios. It appears that 1) antidiuretic hormonehas little direct influence on the countercurrent multiplieras regards the establishment of a high papillary sodiumcontent, and 2) diminished papillary osmolality duringwater diuresis is due primarily to increased water con-tent and depletion of papillary urea.

The Role of the Liver in Clearance of Blood Coagula-tion Product I. THEODOREH. SPAET,* New York,N. Y.

The present study is an extension of previous investi-gations designed to test the hypothesis that in vivo bloodfluidity is preserved by a clearance mechanism wherebyclotting intermediates are removed during circulationthrough certain tissues or organs. Experimental animalswere Sprague-Dawley male rats, and biological reagentswere of rat origin. Product I was prepared by reactingadsorbed plasma with serum in the presence of calciumion. Intravenous product I injections failed to producea drop in plasma fibrinogen in intact animals, but similarinjections accompanied by clamping of the hepatic bloodsupply were followed by significant depression of fibrino-gen levels. Saline injection with hepatic artery ligationand product I injection with renal vessel ligation bothfailed to produce defibrination in control animals. Markedreduction of product I activity resulted from perfusionof this reagent through isolated rat livers, but activitywas preserved after perfusion through the hind legs.Reticuloendothelial blockade with carbon produced mini-mal and inconclusive depression of product I inactivationin the perfused liver, but addition of small amounts ofthe heparin-neutralizing drugs, protamine or Polybrene,to the perfusate almost completely blocked hepatic in-activation of product I. When perfused product I wasmixed with control material in vitro, no anticoagulanteffect was noted. Perfusion through the liver of pre-

cursor reagents showed that plasma thromboplastincomponent was inactivated, but antihemophilic factor,factors V, VII, and X were unaffected. Earlier studiesshowed that blood thromboplastin is cleared by the reticu-loendothelial system. The present data suggest thatproduct I is handled by a different mechanism, whichmay involve participation of hepatic heparin.

Pulmonary Function after Experimental Autologous Pul-monary Emboli. MYRONSTEIN, ISRAEL ALKALAY ANDISRAEL BRUDERMAN, Boston, Mass. (introduced byStanford Wessler).

After release to the lungs of endogenous thrombifrom the peripheral veins of the dog, arterial 02 satura-tion, plasma-dissolved oxygen after ventilation with99.6% 02, and cardiac output are decreased, while respir-atory rate, minute volume of ventilation, and arterial-alveolar (a-A) Pco2 differences are increased. Furtherstudies demonstrate that postembolic decreases in lungcompliance (CL) and increases in mean frictional resist-ance are transient, returning to control levels 15 to 30minutes after embolization. Measurements of lung vol-umes did not vary significantly from pre-embolic deter-minations. The a-A Pco, difference, which was notsignificantly augmented by bronchodilator drugs, wascompletely eliminated by inhalation of 2.5% CO2. Acutedecreases in plasma-dissolved 02 concentration duringventilation with 99.6% 02 were not reversed by positivepressure ventilation until 3 days after embolization.a-A Pco2 differences also returned to control levelswithin 4 days after emboli release. Necropsy examina-tion, however, revealed persistent massive pulmonaryembolization. The arterial blood 02 saturation studiesindicate that acute pulmonary embolism induced a per-fusion defect consistent with a right-to-left shunt. Flowthrough atelectatic lung seems unlikely since 1) no cor-relation exists between CL decrease and shunt size, 2)postembolic decrease in lung volumes did not occur,and 3) ventilation with positive pressure 02 failed toproduce a rise in plasma-dissolved 02 immediately afterembolization. Changes in lung mechanics appeared tobe related to endogenous release of a bronchoconstrictivesubstance. The return of arterial blood 02 saturation anda-A Pco2 differences to control levels 4 days after em-bolization could not be attributed to lysis of emboli orpulmonary edema or infarction. The anatomic locationof the shunt remains to be demonstrated.

Effects of Salt Depletion on the Concentrating Opera-tion in Hydropenic Man. R. M. STEIN, B. H. LEVITT,J. G. PORUSH, G. M. EISNER, M. H. GOLDSTEIN ANDM. F. LEVITT,* New York, N. Y.

Solute diureses were produced with hypertonic man-

nitol infusions in maximally hydropenic subj ects on

high salt diets (18 g NaCl/day) and after 7 daysof salt restriction (0.8 g NaCl/day). No differencein maximum Uo.m was noted but TcH2O, whenplotted against solute clearance (Cosm), rose more

1402

AMERICAN SOCIETY FOR CLINTICAL INVESTIGATION

slowly and attained a lower maximum in the salt-freegroup. In both groups TCH-O tended to stabilize be-tween a Cosm of 14 to 25 ml per minute. At Cosmgreater than 25 ml per minute TCH2O fell below maxi-mum in both groups, but hypotonic urine was not ob-served with levels of Co.m as high as 40 to 50 ml perminute. Control Na excretion (UNaV) averaged 1514Eq per minute in the salt-free subjects and 150 uEqper minute in those on the high salt diet. As Cosmincreased to 8 ml per minute, Na excretion rose slightly,but thereafter a more rapid increase was noted in bothgroups. Salt restriction produced a variable reduction inGFR, averaging 14 per cent. When GFRwas increasedat low levels of solute clearance (4 to 7 ml/min) inthe salt-free group by the administration of aminophyl-line, TCH2O rose to levels obtained in the high saltgroup. Aminophylline administered at Cosm greaterthan 12 ml per minute in both groups produced noappreciable change in TCHIG despite comparable incre-ments in GFR and UN5V. When distal Na supply isaugmented at low levels of solute clearance in salt-freesubjects, TCH2O increases to levels obtained in high saltstudies. The lower TCH2O levels obtained in salt-freesubjects may therefore result from diminished distal Nasupply (secondary to enhanced proximal Na reabsorptionor a decreased GFR, or both). Increasing distal Nasupply at high levels of solute clearance dose not aug-ment TcH2O-conceivably because a solute diuresis pro-vokes an increase in effective medullary blood flow.

Insulin-like Activity of Extracts from Large Sarcoma-tous Tumors Associated with Hypoglycemia. JURGENSTEINKE, J. STUART SOELDNERAND ALBERT E. REN-OLD,* Boston, Mass.

Association of large mesothelial tumors with severehypoglycemia is a recognized clinical syndrome whichcan be corrected by surgical removal of the tumor. Themechanism of the hypoglycemia is poorly understood.Large samples from five such tumors (all identified assarcomatous), weighing between 1,070 and 7,500 g,have been examined for insulin-like activity. Each sam-ple was 1) homogenized in buffer or water, centrifuged,and the supernatant assayed, and 2) extracted with acid-ethanol according to Scott and Fisher, followed bydialysis, lyophylization, and assay. The assay procedureemployed was that using the oxidation of glucose-i-C14by rat adipose tissue in vitro as an index of insulinactivity. Direct assay of the homogenate supernatantrevealed levels of insulin-like activity ranging from0.002 to 0.005 U per g of tumor (4.3 to 16.2 U pertotal tumor). The activity of the acid-ethanol extractsranged from 0.0001 to 0.006 U per g (0.2 to 6.4 U pertotal tumor). The values obtained after acid-ethanolextraction are within the range found for similar extractsof samples from human liver, kidney, adrenal, and lymphnodes, and are probably due to contamination with bloodinsulin-like activity. By way of comparison, extractsof normal pancreas assayed by the same technique ex-hibited 1.5 to 3.0 U per g, and seven pancreatic islet

cell tumors associated with hypoglycemia exhibited 2.7to 28.5 U per g. These findings on sarcomas withhypoglycemia contrast with the elevated insulin-like ac-tivity of a tumor extract assayed in this laboratory (asreported by August and Hiatt) and also with isolatedreports in the literature. It is concluded that hypo-glycemia associated with large mesothelial tumors isnot necessarily related to the presence of significantamounts of extractable insulin-like material in the tumor.

The Role of the Phenolic Hydroxyl Group of Thyroxinein Binding by Human Serum Albumin. KENNETHSTERLING* AND THORNTONS. WALKER, New York,N. Y.

Previous equilibrium dialysis studies have suggestedthat the binding of thyroxine by human serum albuminrequires interaction between the dissociated (anionic)phenolic hydroxyl group of thyroxine and cationic groupson the protein molecule, probably epsilon-amino groupsof lysine residues. To test the importance of the dis-sociable hydroxyl (phenolate) group of thyroxine in theinteraction, studies have been carried out with methoxy-thyroxine, the methyl ether of thyroxine, which cannotionize at this site. Addition of "cold" methoxy-thyroxineto human serum albumin had no detectable effect uponthe binding of I3"-labeled thyroxine in the equilibriumdialysis system, when 1, 2, or 3 moles of the methoxycompound were added per mole of thyroxine. Sincecomparable thyroxine additions reduced the fractionbound, it is evident that the phenolate group is requiredat this concentration range. However, the addition of4 or 6 moles of methoxy-thyroxine per mole of thyroxineresulted in a slight but definite diminution of Pl3l-thyroxine bound to albumin. Since large amounts ofmethoxy-thyroxine were needed to demonstrate displace-ment, the findings suggest that the anionic phenolategroup of thyroxine plays a facilitating role in bindingby human serum albumin.

Retardation of Sodium Flux in' Dog Erythrocytes byPhysiological Concentrations of Aldosterone In Vitro.DAVID H. P. STREETENAND COLETTE SPACH, Syracuse,N. Y. (introduced by Eugene L. Lozner).

Previous attempts to demonstrate effects of physio-logical concentrations of aldosterone on sodium exchangein isolated tissues have been unsuccessful. The possi-bility that such effects might be demonstrable in dogerythrocytes (which have unusually high sodium con-centrations) before and after adrenalectomy, was investi-gated. Samples of blood from a dog were incubatedin stoppered beakers to which Na'Cl and dextrose wereadded, and the mixture was agitated for 2 hours in anatmosphere of 95% oxygen and 5% CO2. Aldosteronewas added in varying concentrations to the blood inthree of four beakers, the fourth serving as control.Aliquots of blood were removed from the beakers at15-minute intervals for determinations of plasma radio-activity, stable Na concentration, and hematocrit. From

t4n3


the calculated regression line representing the rate ofdisappearance of radioactivity from the plasma, Na fluxrates were determined. Before adrenalectomy, a signifi-cant (p < 0.01) inverse dose-response relationship wasfound between D-aldosterone concentrations of from 30to 114 ,ug per L and Na flux rates. These concentra-tions are still considerably above physiological levels.Blood from the same dog after bilateral adrenalectomy,during maintenance therapy with cortisol and salineonly, disclosed a highly significant (p < 0.001) inversedose-response relationship between D-aldosterone withinthe physiological range of concentration (0.2 to 2.0 Ag/L)and Na flux rates. Above 2.0 jug D-aldosterone per Lthe dose-response relationship no longer obtained, andNa flux was reduced maximally to about one-fifth con-trol rates. In the absence of added aldosterone, themean Na flux was significantly (p < 0.001) higher inblood drawn after adrenalectomy (10.4 mEq/L of cells/hr) than before the operation (6.5 mEq/L of cells/hr).It is concluded that physiological concentrations of aldo-sterone have striking inhibitory effects on the movementof sodium into the erythrocytes of the dog.

Interconversion of Acanthocytes and Normal Erythro-cytes with Detergents. SAM SWITZER AND HOWARDA.EDER,* New York, N. Y.

Acanthocytosis (Bassen-Kornzweig syndrome) is a fa-milial disorder characterized by retinitis pigmentosa,progressive ataxia, steatorrhea, absence of low densitylipoproteins, and malformed erythrocytes. These cells,known as acanthocytes, are spherical with numerousprojecting spines, and are diagnostic of this disease. Wehave observed that acanthocytes (1% suspension) ex-posed in vitro to a nonionic detergent, Tween 80 (10' M),are rapidly converted to the biconcave disc shape ofnormal erythrocytes. Nonionic detergents contain ahydrophilic portion (usually a polyoxyethylene polymerof variable molecular weight, or a polyol, or both) inester or ether linkage with a lipophilic portion which isfrequently a fatty acidU derivative. In a series of com-pounds studied it was found that the hydrophilic portiondoes not influence acanthocyte reversal, but specificitywas demonstrated for the lipophilic portion; oleyl de-rivatives convert acanthocytes to normal, whereas, stearylderivatives do not. There is evidence for competitionbetween oleyl and stearyl derivatives, and studies withC4-Tween 60 (stearyl) indicate binding of the deter-gent by erythrocytes. No relation could be demon-strated between the surfactant properties of the non-ionic detergents and their ability to convert acanthocytesto normal. Ionic detergents do not reverse acanthocytes.Many such agents, however, produce in normal erythro-cytes a sequence of morphologic changes described byPonder as the "disc-sphere transformation." Interme-diate forms seen in this sequence are indistinguishablefrom acanthocytes. Furthermore, many agents unrelatedto nonionic detergents but known to reverse the disc-sphere transformation convert acanthocytes to normalshape and, conversely, nonionic detergents, which con-

vert acanthocytes to normal, reverse the disc-spheretransformation.

The Effect of Renal Perfusion Pressure on the ActiveTransport of Sodium out of Distal Tubular Urine asStudied with the "Stop-Flow" Technique. LouisTOBIAN,* KAREN COFFEE, DOROTHY FERREIRA ANDJUDITH MEULI, Minneapolis, Minn.

Kidneys perfused at 170 mmHg excrete considerablymore sodium than those perfused at 100 mmHg. "Stop-flow" experiments helped to explain why. "Normal"rats without kidneys or adrenals supplied arterial bloodthrough a pump to isolated normal kidneys and simul-taneously received renal venous blood and urine. In 13experiments the pump perfused the isolated kidney at170 mmHg and, in 10 other experiments, at 100. Inall experiments the ureter was clamped after 60 minutesof perfusion. Ureteral pressures "plateaued" within thenext 4 minutes, after which clamping was continued for8 minutes. Then the ureter was unclamped and urinecollected in polyethylene tubing. Sodium concentrationwas determined in successive segments of the tubing.Each "stop-flow" collection showed the typical smoothdrop in sodium concentration. The segment with thelowest sodium concentration represents the maximumability of that kidney to pump sodium out of the distaltubular urine. The lowest sodium concentrations aver-aged 31.3 mEq per L in the 13 experiments with per-fusion at 170; the lowest concentrations averaged 14.2in the 10 experiments with perfusion at 100 (p <0.000006). Thus, kidneys perfused at 100 mmHg pumpsodium more avidly. The effect is not related to aldo-sterone since adrenal tissue was absent. It is not relatedto different rates of glomerular filtration because clamp-ing largely stops glomerular filtration. Medullary bloodflow is undoubtedly greater at the higher perfusionpressure; but under "stopped" conditions, this shouldnot prevent distal tubular cells from pumping sodiummaximally. The best explanation is that humoral sub-stances are secreted within the kidney itself, in greateror lesser amounts, depending upon the distention of therenal arterial system. These substances directly controlthe rate at which the distal tubular cells pump sodiumout of the tubular urine. This mechanism may play animportant part in various sodium-retaining syndromes.

Methotrexate-induced Intestinal Epithelial Damage Dem-onstrated by Electron Microscopy but not Revealedby Light Microscopy. JERRY S. TRIER, Bethesda, Md.and Seattle, Wash. (introduced by Cyrus E. Rubin).

After administration of 2 to 5 mg i.v. of the folicacid antagonist, methotrexate, per kg, light microscopy ofperoral biopsies of human jejunal mucosa revealed nochanges in the villi although there were some cryptalterations. The crypts exhibited almost complete mi-totic inhibition for 48 hours or more and intracytoplasmicDNA-positive bodies in some crypt cells. This paperdescribes the marked fine structural changes which were

1404


clearly revealed by electron microscopy, although theywere unapparent by light microscopy. Serial jejunalbiopsies in six patients at various intervals after metho-trexate administration showed striking fine structuralalterations in both crypt and villus which appearedwithin 6 hours, lasted for at least 48 hours, and revertedto normal within 96 hours. These structural changesinvolved absorptive, crypt, goblet, and Paneth cells al-though many cells of each type were completely spared.Damaged cells showed swollen, fragmented mitochondriaand strikingly dilated cysternae of the Golgi materialand endoplasmic reticulum. Altered microvilli were

seen in some absorptive cells. Some of the crypt cellsshowed clumping of the nucleoplasm or intracytoplasmicinclusions whose fine structure suggested cellular rem-

nants, or both. Despite these extensive fine structuralchanges, the histologic architecture of the villi was un-altered and the patients noted minimal or no gastro-intestinal symptoms. After methotrexate, severely dam-aged absorptive, crypt, and Paneth cells were seen

adjacent to morphologically normal ones. This cellsparing indicates a marked variation in sensitivity tomethotrexate among individual cells within a cell popu-lation. These doses of methotrexate arrest small intes-tinal cell replacement temporarily and produce profoundalterations in fine structure. These changes could easilybe missed clinically and by ordinary light microscopy.

Newer Lipids of Human Serum, Cerebrospinal Fluid,and Tissues. NAIP TUNA, HELMUT K. MANGOLD,RUDOLPH KAMEREKAND MARY L. LOUDEN, Minne-apolis and Austin, Minn. (introduced by Ivan D.Frantz, Jr.).

Human serum, cerebrospinal fluid, aortic atheromatousplaques, liver, kidney, spleen, perinephric fat, and bonemarrow were extracted with 2:1 chloroform-methanol(vol/vol), according to Folch, and the total lipid ex-

tracts were analyzed by a combination of silicic acidcolumn and thin layer chromatography. In every lipidanalyzed, in addition to the well known major lipidclasses (phospholipids, cholesterol, free fatty acids, tri-glycerides, and cholesterol esters), the following lipidcompounds, present in smaller concentrations, were sepa-

rated and identified by gas-liquid chromatography, thinlayer silicic acid chromatography, infrared spectroscopy,and identification of their oxidation and reduction prod-ucts. 1) Trimeres (trioxanes) have not, heretofore,been reported as occurring in animal or human tissues.They represent polymerized, long chain aldehydes, are

less polar than cholesterol esters, yield dimethyl acetalswith hydrochloric acid-methanol, and have Rf valuessimilar to those of "branched" fatty acid methyl esterswhen separated by gas-liquid chromatography. 2) Fattyacid methyl esters result, we believe, from the methyla-tion of free fatty acids during extraction with chloroform-methanol, since they were not encountered when theextraction was carried out with chloroform-ethanol (2:1,vol/vol). 3) Glyceryl ether diesters (alkoxydiglycer-ides) are less polar than triglycerides, and upon reduc-

tion with lithium aluminum hydride yield glyceryl etherswhich were separated by thin layer chromatography.4) Aldehydogenic triglycerides have not as yet beenseparated in pure form. Evidence will be presented fortheir existence. 5) Several sterols other than choles-terol were separated but not identified. The biologicalsignificance of these substances will be discussed.

Effect of a Single and Repeated Dose of "Pancreozy-min" on Pancreatic Exocrine Secretion. MALCOLMP.TYOR, DAVID A. GIORDANO AND EDWARDE. OWEN,Durham, N. C. (introduced by Julian M. Ruffin).

Considerable debate has centered about the uniformityof pancreatic enzymatic response to various stimuli. Thepresent report deals with such measurements obtainedfrom anesthetized dogs after administration of "pan-creozymin" (Vitrum). Pancreatic flow was maintainedby constant infusion of secretin, 0.88 to 1.08 U per min-ute, in ten dogs. Secretions were collected from thecannulated pancreatic duct for 15-minute periods, be-ginning 45 minutes after completion of surgery (bileduct cannulation, gastrostomy, and pyloric ligation) andonset of secretin administration. In each instance, 20 Uof pancreozymin was given intravenously after the eighthcollection period. Six dogs were given a second iden-tical dose 2.75 hours later. Each 15-minute sample wasanalyzed for protein concentration and amylase and lipaseactivities. Monoesters of p-napthyl actetate, caprylate,and laurate were used as substrate for lipase activity;solubilization of the latter two was obtained using sodiumtaurocholate. The pattern of response to the initial doseof hormone was similar in all dogs, i.e., an abrupt in-crease in volume (1.8-fold), protein content (13-fold),amylase activity (13-fold), and esterase or lipase ac-tivity: acetate (29-fold), caprylate (25-fold), laurate(27-fold). A fall to baseline was observed at the 45-minute post-hormone period. Thereafter, mean valuesfor all measurements were significantly lower than thosein the prehormone periods, despite a significant increasein volume. The immediate response to the second doseof pancreozymin was similar. These data confirm pre-vious observations on the magnitude of the enzymaticresponse to pancreozymin and characterize the post-hormonal effect. Although the enzymatic response wastemporally uniform, the data suggest that pancreozyminmay enhance lipase activity as compared with amylase,since the magnitude of the response to pancreozyminwas similar whether or not taurocholate was employedin the quantitation of esterase activity.

Metabolic Changes and Viability of Stored HumanPlatelets. P. UATHAVIKUL AND M. BALDINI, Boston,Mass. (introduced by William Dameshek).

Previous investigations in this laboratory have shownthat storage of blood platelets at 4° C produces rapidloss of their "viability," i.e., their capacity to recirculateand survive after infusion. Because of the known highrate of platelet glycolytic metabolism, the idea was

1405


formulated that interruption of regeneration of energy-rich phosphate compounds in the stored platelets mightbe totally or mainly responsible for their loss of via-bility. To this end, the rate of disappearance of the"7-minute phosphorus" (7 min-P), representing mainlyATP, in human platelets preserved in various media at40 C was determined by differential hydrolysis. En-zymes of glycolysis (PHADH, PGK, hexokinase) werealso measured. Correlation of these values with therate of decrease in viability of platelets preserved inthese same media was studied. Viability was meas-ured in vivo by the Cr`1-labeling technique. The 7min-P decreased rapidly in platelets preserved in salinemedium. After 5 days of storage, only 24% of the origi-nal value was obtained. The use of plasma medium im-proved preservation of the 7 min-P. Addition of adeno-sine to the plasma had favorable effect. Addition ofinosine plus adenine had the greatest effect, with 80%of the original 7 min-P found after 14 days of storage.The rate of decrease in the activities of the enzymestested grossly paralleled the rate of disappearance ofthe 7 min-P. Storage media which favorably affectedpreservation of the 7 min-P also had favorable effecton preservation of viability. However, with all themedia studied, the rate of decrease in viability wasstrikingly more rapid than the rate at which the 7 min-Pwas lost. This indicated that decrease of ATP in plate-lets during storage was not the only and probably notthe main metabolic damage which determined loss inviability.

The Antibody Response to Bacteriophage OX 174 inNewborn Premature Infants. J. W. UHR,* J. DANCIS,E. C. FRANKLIN,* M. S. FINKELSTEIN AND E. W.LEWIS, New York, N. Y.

Premature newborn infants and older children showa similar antibody response to injection of minuteamounts of bacteriophage pX 174 as judged by thefollowing criteria: 1) easily detectable neutralizing anti-body is produced during the first week; 2) the titerrises for an additional 2 to 3 weeks; 3) an efficientsecondary antibody response can be achieved by re-immunization 6 weeks later; 4) a change occurs in themolecular species of antibody, from 19S to 7S sedimen-tation constant 2 to 6 weeks after immunization. Theseobservations contrast with past studies demonstratingnewborn incompetency toward other antigens, and sug-gest that immunologic maturation may occur in two ormore steps.

The Secretion of 18-Dihydroaldosterone by the HumanAdrenal. STANLEY ULICK AND KATHRYN KuscHVETTER, New York, N. Y. (introduced by Solomon A.Berson).

A study of mineral hormone biosynthesis, using thebullfrog adrenal as a simple model of the zona glomeru-losa of the mammalian adrenal cortex, led to the isolation

of a new steroid in the aldosterone series. On the basisof chemical degradative studies and conversion to aknown steroid, lactone, the structure 18-dihydroaldo-sterone (18-hydroxycorticosterone) was assigned. Evi-dence for the production of the new steroid by thehuman adrenal was obtained when the previously de-scribed urinary metabolite of aldosterone was separatedinto two tetrahydro derivatives, 3a,18,21-trihydroxy-11,20-diketopregnane (I), and 3a-hydroxy-5,3-tetrahydroaldo-sterone (II). When aldosterone-H3 and the C"-labeled18-dihydro derivative were injected simultaneously andboth urinary products isolated, II was found to be ametabolite of aldosterone and I a metabolite of 18-dihydroaldosterone. Labeled 18-dihydroaldosterone wasinjected and the rate of production of the steroid esti-mated from the dilution in specific activity of urinarymetabolite I. Normal individuals secreted larger amountsof the 18-dihydro derivative than of aldosterone as didpatients with secondary hyperaldosteronism in whomsecretion of the 18-dihydro derivative reached values of4.0 mg per day. 18-Dihydroaldosterone exhibited lowactivity as a mineral hormone (courtesy of Dr. J. G.Llauredo). Its secretion appeared to reflect the activityof the renal pressor system, as was illustrated by astudy of malignant hypertension associated with unilateralrenal disease. The secretion of aldosterone was 600 and800 ,ug per day and of the 18-dihydro derivative, 2,900and 3,400 Aug per day. Removal of the diseased kidneyled to clinical improvement and a decrease in the secre-tion of both steroids (aldosterone: 150, 155 gig/day;18-dihydroaldosterone: 1,600, 1,800 /Lg/day).

The Role of Endogenous Glucagon in Blood GlucoseRegulation in Man. ROGER H. UNGER AND ANNAM. EISENTRAUT, Dallas, Tex. (introduced by EliasStrauss).Inability to measure circulating endogenous human

glucagon has prevented clarification of its hormonalstatus in man. Modifications of a previously describedspecific radioimmunochemical glucagon assay to permitthe recovery from serum of as little as 100 to 250 pqugper ml of added crude human glucagon made possible thefollowing studies, designed to examine the role of glu-cagon in blood glucose homeostasis by determining theeffects of starvation and hypoglycemia upon its serumconcentration. The fasting serum glucagon level in nor-mal humans averaged 100 Aug per ml (0 to 400).During hypoglycemia induced by 3-hour infusions ofglucagon-free (immunochemically established) insulin (0.1U per kg), rises in serum glucagon of 250 and 350 sqagper ml appeared in 2 of 3 subjects after an hour; modestreduction of glucagon followed glucose loading. Sixnormals and one mild diabetic were fasted for 72 hours;after 48 hours of starvation, all seven exhibited strikingincreases in serum glucagon from a mean of 112 (0 to400) to 326 Atg per ml (160 to 425) and at 72 hoursto 663 Al.g per ml (500 to 800) (p < 0.01); FFA in-creased by 530 to 1,960 /AEq per L, while blood glucosedeclined an average of 25 mg per 100 ml and body weight

1406


from 2.8 to 8.2 kg. Glucose loading produced within3 hours moderate but significant suppression (p<0.01)of serum glucagon to a mean of 488 gq~gEq per ml(375 to 600), and FFA declined; glucose tolerance wasimpaired in all subjects. In summary, immunochemicalmeasurements of serum glucagon in man reveal: 1) arise in glucagon during starvation and, less consistently,during hypoglycemia; 2) partial suppression of this risein glucagon during glucose loading; 3) essentially paral-lel fluctuations in glucagon and FFA. Glucagon appearsto be a true human hormone with a role in blood glucoseregulation, possibly that of maintaining glucose distribu-tion to glucose-dependent tissues during glucose depri-vation. The possibility is suggested that the hyper-glucagonemia of starvation may contribute to the im-paired glucose tolerance of "starvation diabetes" bylimiting the decline in hepatic glucose output which nor-mally follows glucose loading.

Effects of Humoral Factors on Antigenic HistamineReleased from Human Leukocytes. PAUL P. VANARSDEL, JR., Seattle, Wash. (introduced by Clement A.Finch).

When leukocytes from a person with atopic sensitivityare incubated with specific antigen, free histamine isreleased from these cells in proportion to the logarithmof antigen concentration up to an optimal point. Excessantigen releases less histamine. This phenomenon hasmade possible several quantitative, experimental observa-tions on the nature of human skin-sensitizing antibody.For the present study, heparinized blood from subjectssensitive to grass pollen was collected and used within30 minutes. In all instances, suspensions of thrice-washed leukocytes were prepared and incubated for 1hour with water-soluble pollen antigen at a concentrationof 0.01 Ag pollen nitrogen per ml. After centrifugation,histamine in each supernatant solution was measuredphotofluorimetrically. Two aliquots of leukocytes fromeach of eight sensitive donors were suspended in Tyrode'ssolution, and two others in isologous plasma. Theamount of histamine released by antigen varied from9.5 to 44 ,ug per L. Little variation was noted on repeatstudies of any one patient. Unrelated antigens releasedno histamine, nor was histamine released from normaldonor cells by antigen. These eight donors had receivedno parenteral immunization with pollen antigen. Thepresence of plasma did not alter the amount of histaminereleased even though each plasma contained a hightiter of skin-sensitizing antibody. The effect of plasmafrom patients who had been artificially immunized byrepeated injections of the same antigen was quite differ-ent. Antigenic histamine release from their leukocytessuspended in Tyrode's solution was in the same rangeas that of the untreated donors. However, almost nohistamine release occurred in the presence of isologousplasma, unless higher antigen concentration was used(0.1 Ag pollen nitrogen/ml). Serum from two of theartificially immunized donors has been fractionated bydiethylaminoethyl cellulose chromatography. The two

fractions containing yi- and '2-globulins inhibited hista-mine release in a manner comparable to that of wholeplasma. Another fraction, containing 18S 'y-globulin,was partially inhibitory. Two other protein fractionshad no effect. These observations suggest that spon-taneous atopic sensitivity is a cellular phenomenon, notdirectly mediated by free skin-sensitizing antibody. Con-trariwise, the effects of artificial immunization are pri-marily humoral, associated with the production of several'v-globulins, and perhaps other proteins, capable of neu-tralizing antigen.

Practical Growth Hormone from Human PituitaryGlands. W. P. VANDERLAAN,* U. J. LEWIS ANDA. E. JONES, La Jolla, Calif.

A simple procedure for preparing human growth hor-mone, in which hot glacial acetic acid extraction of ace-tone-dried powder was followed by precipitation withacetone and lyophilization, gave a 10 to 12%o yield. Thematerial was dissolved in dilute HC1 and administeredintramuscularly in clinical use. Assay showed no thyro-tropic or gonadotropic activity. Despite failure toremove corticotropin, the material showed greater homo-geneity than the Raben preparation when tested by discelectrophoresis on acrylamide gel. Clinically the ma-terial sustained over a 3-month period the rate of growthand elevated alkaline phosphatase activity of an 18 yearold dwarf who has now grown a total of 9 inches in 38months. 17-Hydroxycorticosteroid excretion remainedlow despite 2 mg intramuscularly on alternate days,and in the 24 hours following 5 mg intramuscularly norise in corticosteroid excretion was measurable. Afterovernight fasting, free fatty acids in two pituitary dwarfsrose 4 hours after 5 mg of the material was givenintramuscularly (338 to 920 1Eq/L; 665 to 968 uEq/L).Avoidance of alkali in the preparation accounts for thegreater homogeneity. Lack of corticotropic effect ap-pears to depend on the low initial corticolropin contentof human pituitaries and the inconsequential "flash"effect when it is given in aqueous preparation intra-muscularly. The preparation, therefore, is simple andsuitable for human use.

Factors of Rate and Kinetics of Cellular Reduplicationin Human Hair Roots. EUGENEJ. VAN SCOTT ANDTHOMASM. EKEL, Bethesda, Md. (introduced by C.Gordon Zubrod).

The rapid rate of hair growth suggests that the germi-native matrix of the hair manifests a rate of cellularreduplication equal to or exceeding that of any tissueof the body. The size of the germinative matrix ofhuman scalp hairs compared to the size of hair shaftproduced daily indicates that the matrix reduplicates itscellular population once every 24 hours. Some factorspossibly relating to the control of this benign orderlygrowth were studied. Ninety-two hair roots, from tenbiopsy specimens of normal human scalp, were histo-

1407


geometrically examined in serial sections 6 1A thick.The volumes of the entire germinative matrix andconstituent 30-A thick segments were determined. Thenumber of cells in mitosis in each segment was deter-mined and the mitotic index calculated. The volume andcell population of each connective tissue hair papillawere determined. Distances between component por-tions of the hair root were measured. The total numberof mitoses in the hair matrix was found to be proportionalto the volume of the matrix. A ratio of approximatelyeight connective tissue papilla cells per one mitosis inthe matrix was found in hair roots of different sizes.The mitotic index of the hair matrix was constant forhairs of different sizes, and comparable segments hadequal mitotic indices. Mitotic index was highest in themost proximal segment, 84 per 1,000 cells (1O6,uA), andlowest in the most distal segment, with a value of 2.The mitotic index of any segment of any sized hairroot was found to be expressed by the formula X =113-0.58Y, where Y is the distance in microns of thesegment from a midpoint in the base of the hair papilla.

Relation of Fecal Flora to Endogenous Infections of theUrinary Tract. KENNETHL. VOSTI, ARNOLDS. MONTOAND LOWELLA. RANTZ,* Palo Alto, Calif.

The serologic classification of Escherichia coli accord-ing to specific 0 groups has afforded an opportunityto further our understanding of the pathogenesis ofinfections of the urinary tract. These studies representanalyses of urine, stool, and blood specimens from 80patients with urinary tract infections, 24 of whom hadbacteremia. Stool specimens often contained multipleserologic groups; only a single group was usually iden-tified in urine and blood. Thirty-four different groupswere identified in stool, 17 in urine, and 8 in bloodspecimens. The same serologic groups were found inpaired stool and urine specimens from 75% of 62patients with infections of the urinary tract, in pairedurine and blood specimens from 12 of 13 patients withbacteremia, and from 6 of 7 with paired stool and bloodspecimens. In 5 of 6 patients the same 0 groups werepresent in blood, urine, and stool specimens. Groups1, 4, 6, 15, 25, and 75 comprised 18% of the total numberof different groups identified in stool, 35% in urine, and50% in blood specimens. These groups represented 43%of the total number of groups identified in the individualstool and 70% of those in the individual urine and bloodspecimens. Groups 4 and 6 accounted for 20, 37, and59%o of these values, respectively. These studies showthat: 1) individual urine and blood specimens usuallycontain E. coli of a single serologic group in contrastwith stool specimens where multiple groups may befound; 2) most of the patients with urinary tractinfections have the same serologic group in pairedurine and stool specimens, and in the blood when bac-teremia occurs; and 3) a few groups are more com-monly associated with infections of the urinary tractand bacteremia.

Dextran Gel Filtration of Serum Proteins Involved inHost Defenses Against Infection. SHELDONL. WAGNERAND GENE H. STOLLERMAN,* Chicago, Ill.

The molecular sieving effects of highly cross-linkeddextran gels (Sephadex) were applied to the purificationof serum proteins involved in host defenses against in-fection. In particular, the rapid removal of antibioticsand of other antibacterial substances (various preserva-tives) from serum was attempted in order to aid inthe study of acute-phase serological responses in patientswith fever. Two- to 3-ml serum samples were filteredat 40 C through G-75 Sephadex columns. This pro-cedure separated the effluent samples completely intotwo fractions, greater and lesser than molecular weightof 40,000, respectively. Serum proteins, which did notdiffuse into the gel strains, moved rapidly through thecolumn with minimal dispersion and were recoveredwithin 1 hour in buffers of desired pH and ionic strength.The total serum protein was diluted less than 20%.Peak protein concentrations in individual effluent samplesalmost equaled those of the native sera. The effluentfractions of serum proteins, collected in Tyrode's buffer,were compared with native serum for the followingbiological activities: 1) complement, 2) cofactors in hu-man serum required for efficient opsonization of group Astreptococci, 3) "gram-positive" bactericidins, and 4)"gram-negative" bactericidins. All activities were re-covered fully with minimal dilution. Human sera con-taining penicillin, erythromycin, chloramphenicol, tetra-cycline, or streptomycin, in various combinations, weresubjected to gel filtration. By assays against selectedsusceptible strains of bacteria, the gel-filtered serum pro-teins were found to be free of antibiotic activity. Theantibiotics were recovered, free of serum protein, in thesecond, smaller molecular weight fraction. The methodprovides a useful initial step in the purification of serumproteins associated with antibacterial activity, and sim-plifies the study of host defense factors in the blood ofpatients receiving chemotherapy for infections.

Aldosterone and Progesterone Secretion in Pregnancy.MAMORUWATANABE,OSCARV. DOMINGUEZ, C. IRVINGMEEKER, ETHAN A. H. SIMS, MARYJANE GRAYANDSAMUEL SOLOMON, Montreal, Canada and Burlington,Vt. (introduced by John C. Beck).Aldosterone and progesterone secretion rates (A.r and

P.r) were determined in the third trimester of pregnancywith isotope dilution techniques using H'-labeled hor-mones. Asr were measured in 59 normal pregnant sub-jects between 27 and 40 weeks of gestation who wereon known sodium intakes of 50 to 225 mEq per day.The range was 307 to 2,912 ,ug per day (normal non-pregnant subjects, 88 to 216 ,ug/day) with a mean valueof 1,127 ,g per day. Seventy %o of the Air were between500 and 1,300 ,ug per day. Na restriction markedly ele-vated the Air (to 10,000 ug/day) whereas Na loadingusually depressed it. In 52 patients on a normal Na in-take (120 to 190 mEq/day) there was a uniform rate of

1408


secretion between 30 and 35 weeks of gestation with arise thereafter. In severe pre-eclamptics the A.r wasfound to be in the normal nonpregnant range, whereasin mild pre-eclamptics the secretion rate was normal orslightly elevated. In 50 normal subjects in the thirdtrimester of pregnancy the Psr had a range of 188 to 563mg per day with a mean value of 329 mg per day. Whensimultaneous A., and P., were determined, no directcorrelation between the two secretion rates could bedemonstrated. With Na restriction the A., was markedlyelevated without much change in the P.,. Intrauterinefetal death invariably resulted in a fall of the As. tononpregnant levels, whereas the Pa, was not alwaysdecreased. The above findings are in harmony with theview that the marked stimulation of aldosterone secretionobserved in late pregnancy, especially with Na restriction,is due in part to the necessity for counteracting the largeamounts of progesterone produced, and in part in re-sponse to the need for maintaining a positive sodiumbalance previously described.

Relationship of High Levels of "Bart's" Hemoglobin inInfancy to a-Thalassemia. DAVID J. WEATHERALL,Baltimore, Md. (introduced by C. Lockard Conley).

Hemoglobin patterns of umbilical cord blood from 826Negro and 150 Caucasian infants were studied by electro-phoresis on vertical starch gel. Sixteen Negro infantsshowed 5 to 10% of a rapidly migrating hemoglobinwhich chemical studies, including "fingerprinting," identi-fied as hemoglobin "Bart's" (,y'). Ten of these infantshad repeated hematologic studies up to the age of 1 year.Bart's hemoglobin disappeared at about 6 months, butstriking abnormalities of the red cells and decreasedosmotic fragility persisted. One parent of each infantand 11 of 28 siblings showed similar although milderabnormalities of the erythrocytes. Traces of a hemo-globin with the mobility of hemoglobin H were demon-strated in nine of these apparently affected familymembers, but in no instance was the level of fetal hemo-globin or of hemoglobin A2 increased. Three siblingswho were heterozygous for hemoglobin S or C had morehemoglobin A than abnormal hemoglobin, but they dis-played erythrocyte abnormalities exceeding those usuallyseen in S-trait or C-trait and were presumably doublyaffected. Many normal newborn infants, both Negroand Caucasian, showed traces of a hemoglobin with anelectrophoretic mobility identical to that of hemoglobinBart's. Increased levels of Bart's hemoglobin ininfancy may result from an inherited defect in a-chainsynthesis (a-thalassemia). During rapid synthesis ofhemoglobin in fetal life this defect leads to an excessof -y chains, resulting in the production of hemoglobinBart's and traces of hemoglobin H(p4A). This de-ficiency of a-chains is further exaggerated during theswitchover from fetal to adult hemoglobin synthesis,since both -y- and 8-chains are competing for availablea-chains, a fact which may explain the small amountsof Bart's hemoglobin found in many normal infants.

In adult life this form of thalassemia is manifested byvery mild hematologic abnormalities.

New Approach to the Clinical Pharmacology of Digitalisin Man. ARNOLDM. WEISSLER, HARVEYE. GRODEANDWILLIAM G. GAMEL, Columbus, Ohio and Galveston,Tex. (introduced by James V. Warren).

Previous studies in normal individuals demonstratinga decrease in the duration of mechanical systole afteradministration of deslanoside (Cedilanid-D) have sug-gested the application of this physiologic effect in theassay of digitalis effect in man. In the present investi-gation this technique was employed in the study of theduration of deslanoside effect and the dose-response re-lationships in normal individuals. Serial determinationsof the duration of mechanical systole, as derived fromsimultaneous recordings of the first to second interval(S1-S2) on a phonocardiogram and the duration of leftventricular ejection (ET) on the carotid arterial pulsetracing, were made for 5 days in nine normal individualsafter administration of deslanoside, 1.6 mg i.v. Thedigitalis effects were compared to 4 control days ofobservation. Maximum decrease in Si-S2 and ET(corrected for heart rate) averaged, respectively, 21(p <0.001) and 18 msec per beat (p <0.001) and ap-peared 1 to 2 hours after drug administration. Significantshortening in the duration of mechanical systole con-tinued until 72 hours after administration of the drug.Comparison of the effects of 1.6 and 0.8 mg i.v. in thesame individual during the first 24 hours after drugadministration was made in seven subjects. A paralleltemporal effect of the two dose levels was observed.The degree of shortening in S1-S2 and ET after the 0.8mg dose was 66% and 84% of that of the 1.6 mg level.These observations demonstrate the applicability of thesetechniques in the clinical assay of digitalis effect, andsuggest the need for re-evaluation of our concept of thedose-response relationship of deslanoside in the human.

The Metabolism of Certain Isotopically Labeled Pyrimi-dines and the Rate of Pyrimidine Biosynthesis in Man.S. M. WEISSMAN, A. Z. EISEN, M. LEWIS AND M.KARON, Bethesda, Md. (introduced by Nathaniel I.Berlin).

The present studies were performed since quantitativeestimates of pyrimidine production rates and utilizationof salvage pathways by man have been lacking. 2-C14-or 6-C14-orotic acid or uracil was given intravenouslyto patients with slowly growing malignancies and thespecific activities of respiratory carbon dioxide, urinaryuracil, and urinary pseudouridine (5-ribosyluracil) werefollowed. Subject to the assumption that orotic acid deg-radation occurred principally through initial conversionto pyrimidine nucleotides, calculation of the ratio of dailypseudouridine excretion to cumulative isotopic excretionin urinary pseudouridine gave estimates of pyrimidinenucleoside production in three subjects of between 1.1and 1.6 g per day. Renal clearance of uracil exceeded

1409


50 ml per minute and plasma levels were estimated asless than 1 ,ug per ml. Little or no utilization of uracilfor pyrimidine nucleotide synthesis occurred at normalplasma levels, since there was a rapid disappearance ofradioactivity from plasma and urinary uracil and pseudo-uridine; thus, the potential salvage pathway of uracilmetabolism is of no physiologic importance for mosttissues.

Studies on Lysosonmes. The Effect of Cortisone andEndotoxin. GERALDWEISSMANNANDLEWIS THOMAS,*New York, N. Y.

Lysosomes are subcellular particles which release suchacid hydrolases as acid phosphatase, 83-glucuronidase, andcathepsins upon the death of cells. Previous work indi-cated that lysosomal enzymes were released from particlesin vitro by treatment with vitamin A or by irradiationwith the full mercury spectrum. Release of enzymes wassignificantly inhibited when suspensions were preparedfrom animals pretreated with glucocorticoids, whiledeoxycorticosterone acetate was ineffective. Becausesome of the effects of endotoxin in vivo might resultfrom the antemortem release of lysosomal enzymes intocell sap or into the circulation, young rabbits were in-jected intravenously with Aerobacter aerogenes endo-toxin (20 to 40 /Lg). Lysosome-rich suspensions fromlivers of these and control animals were prepared in0.25 M sucrose at various times after endotoxin adminis-tration. To test the "lability" of the particles, aliquotswere irradiated with a mercury vapor lamp, while otherswere simply incubated at 370 C. After treatment, sus-pensions were centrifuged at 15,000 G for 20 minutes andthe p-glucuronidase and catheptic activity of the super-natants determined. Five minutes after endotoxin in-jection, but not after 1 minute, depletion of the totalenzyme content of the suspensions was observed, withincreased lability of the particles to irradiation. Deple-tion of both enzymes reached maximum at 15 and 30minutes after endotoxin (66 ± 12% of control total),and activity released into the supernatant of irradiatedsuspensions was increased over controls by 20 ± 6%(cathepsins) and 48 + 12% (p-glucuronidase). Theseeffects lasted for 18 hours. Pretreatment of the animalswith cortisone acetate (50 mg/kg) for 3 days inhibitedthe depletion of enzyme and reduced the lability of theparticles to below control values. These experimentssuggest that one action of endotoxin is to release acidhydrolases from particulate form within cells, and thatglucocorticoids serve to stabilize such particles againstinjury by several agents.

Blood Viscosity Alterations in Dysproteinemia. ROE E.WELLS, JR. AND RICHARD D. PERERA, Boston, Mass.(introduced by Eugene C. Eppinger).That quality of blood which describes either its vis-

cosity or fluidity is dependent upon the concentration ofthe cellular elements and the forces creating the con-ditions of flow. Although plasma is known to have a

relatively constant viscosity under most conditions offlow, it has been demonstrated that a suspension of redcells in plasma has a viscosity that is both higher andmore flow-dependent (viscosity falling as flow rate rises)than a suspension of red cells in saline at equivalenthematocrit values. Present evidence suggests that globu-lin is the plasma protein component that accounts forthe greater part of this effect. Studies of blood viscosityhave been conducted in patients with abnormal elevationsof serum globulins including those with macroglobulins.Two viscometers were used: one, a Couette type re-quiring a 5 ml sample capable of recording shear stressand thereby viscosity over a shear rate range from 0.1to 20 sec'; the second, a cone plate type viscometerrequiring a 2 ml sample and providing a shear raterange from 24 to 240 sec'. All samples were measuredimmediately after collection at 37° C. Plasma or serumprotein analyses were done by biuret and paper electro-phoretic methods. Blood viscosity values of patients withelevated globulin levels were significantly higher thanthe values found in a large number of patients withnormal serum proteins of equivalent hematocrit levels.Plasma viscosity values were also higher than those ofnormal subj ects, and in the case of one patient withmacroglobulinemia, plasma viscosity also demonstrated asignificant degree of shear rate (flow) dependence.

Chromatograms of Individual Achlorhydric GastricJuice Specimens in Subjects With and Without Perni-cious Anemia. J. D. WELSH, J. HARTZOG, L. RUSSELLAND S. WOLF,* Oklahoma City, Okla.

Separation of the nondialyzable components of indi-vidual human gastric juice specimens on the IRC-50Amberlite resin column has been shown to yield repro-ducible and comparable patterns. The purpose of thispaper is to present data on the chromatograms of indi-vidual achlorhydric gastric juice specimens from patientswith and without pernicious anemia, the alterations in suchpatterns when serum albumin is present, and, for com-parison, the paper electrophoretic patterns on the samesamples. Separate chromatograms were performed onthe individual stimulated gastric juice specimens fromeight subjects with pernicious anemia. The patternswere distinct from those of stimulated normals, containingonly four of the usual six protein peaks. The twoproteolytic peaks were missing from all specimens. Fourof the specimens contained serum albumin. This couldbe localized in the last three protein peaks, which in turnwere unusually large. Paper electrophoresis of the samespecimens also revealed a characteristic pattern for per-nicious anemia, including the identification of serumalbumin in the gastric juice, confirming the work ofGlass. A similar "pernicious anemia" pattern on resincolumn chromatography or electrophoresis could beidentified in the fasting juice of an occasional subjectwithout pernicious anemia when no hydrochloric acidwas present. Even serum albumin was demonstrated inthe fasting achlorhydric juice from one patient withoutpernicious anemia. After adequate stimulation, however,

1410


the specimens of such patients yielded a normal chroma-togram and a normal pattern on paper electrophoresis.Serum albumin was no longer identified. In contrast,the gastric juice patients with pernicious anemia afterstimulation yielded the same abnormal pattern as thatencountered in the fasting juice and distinctly differentfrom the normal.

The Metabolism of Ring-labeled L-Thyroxine-C' in Rat,Dog, and Man. C. D. WEST, L. F. KUMAGAIAND M.J. GORTATOWSKI,Salt Lake City, Utah (introduced byFrank H. Tyler).In order to investigate the metabolic fate of the thyro-

nine ring of thyroxine, L-thyroxine-C1' (1.20 mc/mmole),with the radioisotope in the phenolic ring, was preparedfrom uniformly labeled C"4-phenol. The synthetic productwas proven to be L-thyroxine by mixed melting point,optical rotation, paper chromatography, and bioassay.When it was given intraveneously to intact rats, a totalof 67%o of the administered radioactivity was excretedin feces in 12 days with 87% of the fecal excretionoccurring during the first 3 days. Twenty per cent wasexcreted in urine with a half-time (tj) of 51 hours, andless than 0.1%o appeared as respiratory C1402. In manthe total amount of radioactivity excreted in the urinewas comparable to that observed in the rat, but the rateof excretion was slower (ti = 120 hrs). In the dogtotal urinary excretion was considerably less than ineither man or rat and much faster (ti = 24 hrs). Thedog also differed from man in that the rate of removalof the inj ected L-thyroxine-C" from the blood streamwas much more rapid. In rats with their common bileducts catheterized, the per cent of administered C"4excreted in bile varied from 57 to 80%o as a direct func-tion of the weight of thyroxine administered. Paperchromatography revealed that C4-thyroxine glucuronideand C4-thyroxine made up approximately 90%o of thebiliary radioactivity with the remaining 10%7o in an un-identified highly polar compound. The ratio of C14-thy-roxine glucuronide in bile was approximately 1.0 whenlarge amounts of thyroxine were administered but de-creased considerably with smaller doses. C14-thyroxineglucuronide was identified in the urine by paper chroma-tography, but over 75%o of the administered radioactivitywas in three compounds which have not yet been identi-fied. Neither C"-thyroxine nor C14-triiodothyronine wasfound.

A Demonstration of the Dynamic Nature of HypertrophicSubvalvular Aortic Stenosis. ROBERT E. WHALEN,ALLAN I. COHEN, ROBERTG. SUMNERAND HENRYD.MCINTOSH, Durham, N. C. (introduced by William M.Nicholson).

Patients with genetic, clinical, X-ray, and electro-cardiographic features suggesting hypertrophic subvalvu-lar aortic stenosis have been studied in this and otherlaboratories. When no pressure gradient is demonstratedacross the outflow tract of the left ventricle, the diagnosis

of idiopathic cardiac hypertrophy is usually entertained.Such a patient was studied on two occasions and whilein a basal state was found to have no significant gradientbetween the left ventricle and the systemic artery. How-ever, on both occasions, during an infusion of isopro-terenol (1.2 jug/min), this patient developed an 80 mmHg systolic gradient at the subvalvular level. (Thesubvalvular location of the obstruction was conclusivelydemonstrated by "pull through" tracings). A smallerbut definite gradient also developed during and afterexercise. Indicator dilution curves obtained when thegradient was presented indicated considerable mitral in-sufficiency, a feature commonly present in patients withhypertrophic subvalvular aortic stenosis. Similar curvesobtained when there was no gradient failed to revealmitral insufficiency. Repetition of the study in a groupof patients with left ventricular hypertrophy from othercauses, including aortic valvular stenosis, did not pro-duce a similar response. A study of laboratory animalswith left ventricular hypertrophy (turkeys, heart-blockdogs, and greyhounds) also failed to demonstrate such aresponse to isoproterenol. The inotropic effect ofisoproterenol presumably leads to a disproportionatestrengthening or asynchronous contraction of the leftventricular outflow tract. These changes result in hemo-dynamic findings similar to those observed in patientswith overt hypertrophic subvalvular aortic stenosis. It issuggested that similar studies may further elucidate thebasic defect present in patients with this syndrome, andthat isoproterenol infusion may serve as a provocativetest for the detection of patients predisposed to develophypertrophic subvalvular aortic stenosis.

Observations in Experimental Magnesium Depletion.R. WHANGAND L. G. WELT,* Chapel Hill, N. C.

Sprague-Dawley rats were fed a diet free of sodium,potassium, chloride, phosphorus, and magnesium. Controlgroups were gavaged with adequate quantities of theseions, whereas experimental groups were gavaged withidentical solutions devoid of magnesium. The observa-tions were as follows: magnesium-deficient animals hadhypomagnesemia, hypercalcemia, azotemia, an expanded"chloride space" in muscle, and a significant decrease inmuscle potassium. This deficit in muscle potassium couldnot be avoided despite the administration of as much as4 mmoles K per day throughout the study. In anotherstudy identical in description, except that neither con-trol nor experimental animals received potassium, it wasdemonstrated that from the seventh day on, the excretionof potassium was significantly greater in the magnesium-deficient group. This larger urinary excretion of po-tassium and the depletion of muscle potassium obtainedin magnesium-deficient animals, despite the fact that theserum levels of potassium were the same in the controland experimental groups. These data strongly suggestthat magnesium deficiency (or the associated hypercal-cemia) interferes with the ability of muscle cells tomaintain an appropriate concentration gradient for po-tassium between cellular and extracellular water,

141


Specificity in Fatty Acid Esterification during Fat Ab-sorption. H. MALCOLM WHYTE, ARTHUR KARMENAND DEWITT S. GOODMAN,Bethesda, Md. (introducedby Robert S. Gordon, Jr.).

The fatty acid specificity of several different ratintestinal esterifying systems has been studied duringfat absorption. Mixtures containing equal amounts ofC14-labeled palmitic, oleic, and linoleic acids, but withvarying ratios of unlabeled fatty acids, were given bygastric intubation to rats with cannulated thoracic ducts.The chylomicron lipid so obtained was chromatographedon silicic acid columns to separate glycerides (mostlytriglycerides), cholesterol esters, and lecithin. Gas-liquid chromatography was then employed to measurethe mass, total radioactivity, and specific radioactivityof the individual fatty acid components of these threefractions. These data provided quantitative informationon the specificity of incorporation of each of the threefatty acids into each lipid class, and on the relativeextent to which each fatty acid in each fraction wasdiluted with endogenous fatty acid during absorptioninto the chyle. All three exogenous fatty acids were in-corporated into the glycerides without preference, butwith evident endogenous dilution. Oleic acid was clearlypreferred for cholesterol esterification; palmitic and lino-leic acids were used to an equivalent extent althoughless than oleic acid. In contrast, there was marked dis-crimination against oleic acid and slight selectivity forlinoleic acid in the lecithin fraction. Dilution of the fedacids with endogenous fatty acids was greater in thecholesterol ester and lecithin fractions than in the glycer-ides; this was particularly true for palmitic acid. Inaddition, there were dramatic differences in the positionalspecificities of the different fatty acids in lecithin; thesedid not occur in triglycerides.

Studies on the Two Pathways of Cholesterol Synthesis.JEAN D. WILSON, Dallas, Tex. (introduced by M. D.Siperstein).

It has recently been demonstrated that the completionof cholesterol synthesis occurs by two separate pathways.Increased activity of either of these two pathways maybe responsible for derangements in cholesterol metabolism.One pathway (I) involves the completion of the sterolnucleus prior to saturation of the side chain (lanosterol->zymosterol->desmosterol->cholesterol). A second path-way (II) of cholesterol synthesis has been described ina tumor of the mouse preputial gland in which thesterol side chain is reduced prior to completion of thenucleus (lanosterol->dihydrolanosterol--*A8-methostenol->AT-methostenol->AT-cholestenol---cholesterol). The fol-lowing study was undertaken to devise a means fordetermining which of the pathways is present in normaltissues. Triparanol is known to block the reduction ofthe 24-25 double bond in pathway I, as a result of whichdesmosterol accumulates in serum and tissues. Sincepathway II also requires reduction of this double bond(lanosterol-*dihydrolanosterol), then lanosterol would

accumulate in tissues which contain pathway II if tripara-nol blocks the reduction of all 24-dehydrosterols. Utilizingboth silicic acid and gas-liquid chromatographic tech-niques, the preputial gland sterols from rats fed triparanolwere analyzed; three major components were found andidentified as cholesterol, desmosterol, and lanosterol,whereas untreated preputial gland contains predomi-nantly cholesterol, AT-cholestenol, and AT-methostenol.The accumulation of lanosterol in response to triparanoltreatment constitutes strong evidence that the preputialgland does in fact contain pathway II. Of the remainingorgans of the rat which have been similarly analyzed,only skin appears to contain a second pathway of majorsignificance. In summary: 1) A technique has beendevised to distinguish the presence and contribution ofthe two biosynthetic pathways of cholesterol synthesis inliving tissues; 2) evidence is advanced that in at least twotissues of the normal rat cholesterol can be formed viaboth synthetic pathways.

Achievement of Homografts with Selective LymphaticIrradiation.. H. S. WINCHELL AND M. POLLYCOVE,Berkeley, Calif. (introduced by John H. Lawrence).

The homograft rejection response appears to be a func-tion of the lymphatic tissues. Successful homograftingrequires the suppression of this function. We havedeveloped a method for delivering radiation more se-lectively to lymphatic structures than heretofore has beenavailable for the purpose of preparing large mammals,such as dogs and man, for tissue grafting. Yttrium9'chelated by diethylenetriamine pentaacetic acid (DTPA)was administered intravenously to dogs after whichthe dog's urine, which contained the rapidly excretedY9-DTPA, was continuously pumped into one of itsveins for a period of 6 hours. This procedure allowedfor maintenance of constant levels of radioactivity inthe body during the time of intravenous urine cyclingand for rapid clearance of radioactivity from the bodyof the dog when intravenous urine cycling was stopped.Dogs given sublethal doses of Y9-DTPA showed selectivedestruction of lymphocytes in both the circulating bloodand on histologic examination of lymph nodes, withoutconcomitant depression of platelets, granulocytes, orreticulocytes in the circulating blood. Lethally irradiateddogs showed depression of platelets, granulocytes, andreticulocytes with virtual absence of lymphocytes in thecirculating blood. The lives of the lethally irradiatedanimals could be saved by the administration of autol-ogous or homologous bone marrow with evidence ofproliferation of the marrow graft in both instances. Aseries of renal homografts in the lethally irradiated dogstreated with homologous bone marrow gave evidence ofinduced "compatibility" of the host and donor tissues.One animal had a renal homograft from an unrelateddonor which produced urine for a 34-day period. Pre-liminary evidence obtained in the treatment of humanacute lymphatic leukemia indicates that this procedurecan effect relatively selective depression of lymphocytesin man as well as in dogs.

1412


Back-diffusion of NH3: The Nephron's TerminalH+-generating System. KENNETH A. WOEBERANDA. GORMANHILLS,* Miami, Fla.

Urinary (total) ammonia arises from enzymatic de-aminations within certain renal tubular cells. Thedependence of its rate of excretion upon urine pH sug-gests diffusion of NH3 (but not NH,+) into tubular fluidat or below the site of H+ secretion. Acute experimentsin three healthy men show that 1) the logarithmic rateof urinary ammonia excretion is a direct linear function(about 0.018 AEq/ml) of urine flow throughout itsphysiologic range when urine pH is held constant at anyvalue. In the same experiments, provided urine pH is<6.0, 2) the logarithmic rate of titratable acid excretionis an inverse linear function of flow, and 3) the slope ofthis regression (sign ignored) is similar. In urines ofhigher pH 4) CO2 tension (pCO,) varies inversely withflow. These observations afford indirect but persuasiveevidence for the following formulation of the non-enzymatic determinants of urinary ammonia excretion:1) diffusion of NH3 to equilibrium in distal tubularfluid; 2) subsequent back-diffusion of NH3 varying as afunction of facultative water reabsorption, and entailing3) intraluminal conversion of NH41 to NH3 with HIgeneration. The value for the regression gives a biasedestimate of variation of the system's activity withchanging urine flow, for experimental conditions re-quired to stabilize urine pH doubtless affect distal tubu-lar H+ secretion. A better preliminary estimate, yieldedby experiments that were similar except that urine pHwas permitted to vary, is 0.0155 log /AEq per ml. Thissystem is clearly the major cause for the high pCO, ofalkaline urine, and its extra Ht-generation at low urineflows is ample to explain the correspondingly increasedpCO2.

Humoral Factors in Transplantation Immunity. PAULL. WOLF, CHIYO CHIBA, HERNANIO RAMOS, PETERSCHOLLMEYERAND R. J. BING,* Detroit, Mich.

Previous observations in this laboratory during thecourse of 250 canine cardiac homograft experiments re-vealed that in the second-set graft transplants a rapidexudative pericarditis and mural endocarditis occurredwithin 15 minutes after grafting. To demonstrate thepresence of humoral factors responsible for rapid rejec-tion, two types of experiments were performed. In thefirst series, first- and second-set canine cardiac trans-plants were perfused with radioiodinated serum albumin.Increased radioactivity of the tissue was present in thegrafts 15 minutes after transplantation, especially in thesecond-set hearts, demonstrating increased vascular per-meability. No cellular infiltrate was present. In the sec-ond group of experiments, humoral antibodies weredemonstrated in vitro. Canine hearts from Dog I wereperfused with Dog I's own blood. To this was addedultracentrifuged cell-free serum from a second dog(Dog II) which had been previously sensitized withspleen cells from Dog I. Radioactive iodinated serum

albumin was added 1 hour after perfusion was begun. In-creased radioactivity was found in the tissue of all 16grafts. Two grafts demonstrated a minimal immune-typegranulomatous myocarditis. The results suggest thepresence of humoral factors contributing to homograftrejection since increased vascular permeability appearsimmediately after exposure of grafts to sensitized cell-freeplasma, and graft rejection occurs without the presenceof cellular infiltrate.

The Role of Ion Size in Thyroidal Anion Transport.J. WOLFF* AND J. R. MAUREY, Bethesda, Md.

Iodide ion is actively concentrated by thyroid tissuewhereas Br ion is poorly concentrated and Cl- and Fions not at all. The possibility of a size requirement foranion transport by thyroid tissue was therefore investi-gated in sheep thyroid slices. Steady state Km values, aswell as K1 values for iodide transport, were determinedfor anions such as I-, SCN-, CIO-, NOS-, BF4-, ReO4-,and so forth. Where not available from the literature,partial molal ionic volumes of these anions were de-termined from the densities of their solutions accordingto the Masson relation. It was found that, despite thedifferent shapes of the anions, the pK values for aniontransport increased as a linear function of the partialmolal ionic volume of the ion. A maximum pK wasreached at anion volumes between 45 and 48 ml per mole.Beyond this volume pK values again decreased. Compari-son with other size parameters (e.g., from crystal orconductance measurements) showed no correlation withKm or KI values of anion transport. It is concludedthat univalency and a suitable size (as measured by thepartial molal ionic volumes) describe minimum, but notnecessarily sufficient, requirements for anion transportinto thyroid tissue.

Hypogammaglobulinemia Produced by the Administra-tion of 6-Thioguanine to Patients with Nephrosis.SHELDON M. WOLFF AND HOWARDC. GOODMAN,Bethesda, Md. (introduced by James H. Baxter).

During a study of the effect of 6-thioguanine (6-TG)on the idiopathic nephrotic syndrome, we have observeda marked decrease in serum y-globulin (paper electro-phoresis). Prednisone therapy alone was not associatedwith a significant change in -y-globulin in four of fivepatients. Three of this group and an additional "steroid-resistant" patient were treated with 6-TG alone or6-TG plus prednisone. In the two patients who re-ceived 6-TG alone 'y-globulin fell from 624 to 160 mgper 100 ml and from 506 to 356 mg per 100 ml(first 2 weeks of therapy). In the three patientstreated with prednisone plus 6-TG marked falls in-y-globulin levels of 83, 82, and 144 mg per 100 ml,respectively, occurred in all, although two had pre-viously shown no fall in 'y-globulin with prednisonealone. Decreases in y-globulin occurred despite diminu-tion or disappearance of proteinuria. Gel electropho-resis and immunoelectrophoresis confirmed the fall in

1413


v-globulin and suggested concomitant decrease in the,8.M (y-1 M) and fl2A (y-1 A) globulins. Isohemag-glutinin titers fell from 1:256 to 1:32 during 6-TGtherapy in one patient. The use of 6-TG has beenassociated with clinical remission in one patient (rela-tively steroid-resistant), with improvement in two, andno response in one (steroid-resistant patients). Temporarybone marrow suppression occurred in all 6-TG-treatedpatients, including two severe episodes. Plasma cellswere present on bone marrow examination at the endof 6-TG therapy. The above data demonstrate thecapacity of 6-TG to decrease the serum globulins withwhich antibody activities have been associated, andraise the question of a possible correlation betweenthe production of hypogammaglobulinemia and clinicalimprovement.

Detection and Quantitation of Left-to-Right Shunts byan Oximetric Inert Gas Technic. ROBERTC. WOODANDHIRAM W. MARSHALL, Rochester, Minn. (introducedby Earl H. Wood).In the presence of a left-to-right shunt, continuous

recording of the oxygen saturation of pulmonary arterialblood during a sudden transient change from breathingair to breathing a gas mixture containing no oxygen willreveal a decrease in oxygen saturation of pulmonaryarterial blood that occurs simultaneously with that insystemic arterial blood. This must result from anintracardiac left-to-right shunting of the reduced hemo-globulin, caused by transient reduction in oxygen tensionin the lungs. The reliability and sensitivity of this inertgas technic in relation to the highly sensitive but tech-nically more difficult method of venous dye-dilutioncurves were studied in six dogs having chronic atrialseptal defects. The magnitude of the left-to-right shuntsvaried from 9 to 65%,. No left-to-right shunt was de-tected in two dogs by conventional blood oxygen-satura-tion methods, although these shunts were demonstrableby both venous dye-dilution and inert gas curves. Maxi-mal sensitivity was attained when the animals werehyperventilated at a rapid rate with 10%o oxygen and5%o carbon dioxide in helium or nitrogen before andafter a transient change (two to six breaths) toventilation with 5% carbon dioxide in helium or nitro-gen. Correlation of the magnitude of the left-to-rightshunts by venous dye-dilution and inert gas curves wascomparable to that for venous dye-dilution and bloodoxygen-saturation methods. Experience with detectionand quantitation of left-to-right shunts with these

methods in patients has been similar. The need for onlya single catheter and the fact that no added electronicequipment is necessary are important practical advantagesof the inert gas method.

Studies on the Antigenicity of Natural and SyntheticPolynucleotides. STANLEY YACHNIN, Chicago, Ill.(introduced by Theodore N. Pullman).Rabbits were immunized with virtually protein-free

polyadenylic acid (Poly A), polycytidylic, polyuridylic,and polyinosinic acids, rat, guinea pig, and rabbit liverribonucleic acids (RNA) deoxyribonucleic acids (DNA),and ribosomes. Search for antibody to polynucleotidesby gel diffusion, complement fixation, and immunochem-ical precipitin techniques was negative. Previous workershave demonstrated the formation of soluble complexesbetween DNAand bovine serum albumin (BSA) underconditions of controlled pH (5.0 to 6.0) and low ionicstrength capable of protecting BSA from heat coagula-tion. Similarly, soluble complexes were formed be-tween heat-denatured BSA (HBSA)-Poly A andHBSA-RNA. Analysis of antisera formed to thesecomplexes by the Ouchterlony technique showed pre-cipitin lines which were absent when antisera reactedwith native BSA or HBSA. Absorption of theseantisera with HBSA removed most precipitating anti-body; a small amount of antibody remained, precipitableonly by polynucleotide-HBSA complex. Precipitin curveswith antisera to polynucleotide-HBSA complexes, nativeBSA, and HBSA, using all three antigens, showed thatin all three instances at equivalence homologous antigenprecipitated by far the greatest amount of antibody.Homologous polynucleotide alone was incapable of:1) precipitating antibody from homologous anti-poly-nucleotide-HBSA serum; 2) precipitating antibody fromhomologous antiserum absorbed with HBSA; 3) re-ducing the amount of antibody precipitable from anti-polynucleotide-HBSA serum. It is concluded that:1) synthetic and mammalian polynucleotides are notantigenic in rabbits; 2) antibody to mammalian liverribosomes will not react with protein-free homologousRNA; and 3) antibody formed to RNA- or PolyA-HBSA complexes differs from that formed to nativeBSA or HBSA, but that polynucleotide alone acts neitheras hapten nor as antigen. It is likely that the heat de-naturation of BSA in the presence of polynucleotideconfers upon it antigenic properties which are not presentin native BSA or in BSA heat-denatured in the absenceof polynucleotide.

1414

proceedings of the fifty-fourth annual meeting of

Documents

Transcript of proceedings of the fifty-fourth annual meeting of