ABSTRACTSJ Clin Invest. 1973;52(6):41a-80a. https://doi.org/10.1172/JCI107329.

Find the latest version:

https://jci.me/107329/pdf

148. Cholesterol Solubility in Bile: Retardation of Pre-cipitation and Dissolution by Liquid Crystalline Meso-phase. R. T. HOLZBACH,* M. F. OLSZEWSKI,* AND M.MARSH,* Cleveland, Ohio (introduced by Daniel L. Horri-gan**).

That cholesterol supersaturation is frequent in healthy manhas been recently shown. Thus this factor alone can no longerbe considered the central issue in gallstone pathogenesis.Identification of those factors that promote its nucleationand precipitation from abnormal supersaturated biles andthose which inhibit precipitation from equally supersaturatednormal biles is now required. We have demonstrated thathigh lecithin concentrations retard the rate of cholesterolprecipitation from supersaturated in vitro solutions composi-tionally patterned after bile. In addition, the gallstone-dis-solving effect of chenodeoxycholic acid may be limited by thekinetics of cholesterol dissolution. Other recent work hasshown that this too may be correlated with relative lecithinconcentration. Using a slow-flow self-diffusion technique, ourrecent studies support this claim of retarded cholesterol dis-solution into undersaturated solutions of high relative lecithinconcentrations. Using periodic polarizing microscopy, a pre-viously undescribed spontaneous liquid--> solid phase transi-tion was observed in supersaturated model solutions. Con-ditions included sterile equilibration at 37°C extending attimes up to 3 wk. In these systems phase transition phe-nomena and transition rates were strong functions of lecithinconcentrations. Liquid crystalline mesophases of 3-30 ja diam-eter were observed in both crystallite nucleation and dis-solution studies with solutions containing appropriate lecithinconcentrations. Identical spontaneous phase transitions wereobserved in supersaturated normal human bile specimens.These findings indicate that rapid nucleation of the meso-phase can explain the rate-retarding effects on cholesterolnucleation and precipitation, as well as upon its dissolutionin model solutions containing high lecithin concentrations.Finally, the observed spontaneous liquid ->solid crystallinephase transition may have in vivo relevance for the earlieststages of cholelithiasis.

149. Human Erythrocyte Redox States: Effects of Po2and pH on NAD+/NADH and NADP+/NADPHSys-tems. B. R. HORN,* J. THEODORE,*ANDE. D. ROBIN,** Stan-ford, Calif.

The state of oxygenation of hemoglobin appears to influ-ence erythrocyte metabolic processes. The cellular redoxstate is intimately linked to such processes. Free NAD+/NADHand NADP+/NADPHratios may be used to mea-sure the redox state of cells. These are calculated by deter-mining the ratio of the concentrations of oxidized and re-duced metabolites of suitable nucleotide-linked reactions,using the known Keq of the reaction. These ratios weremeasured in normal human red cells at Po2's of 150 (AIR)and 15(N2) torr, and at extracellular pH's of 7.4 and 7.05.The NAD+/NADH ratio was determined from the ratio oflactate to pyruvate. The NADP+/NADPHsystem was in-vestigated by utilizing the polyol pathway: GLUCOSE+NADPH+ H+ SORBITOL+ NADP+. The glucose/sorbitol ratio is reported as an index of the NADP+/NADPHratio since the Keq for this reaction is not known.

NAD+/NADH Glucose/sorbitol

Air N2 P Air N2 P

pH 7.4 12.8 8.6 <0.05 392 460 =0.05pH 7.05 1151 742 <0.005 620 840 <0.05

These findings suggest that (a) significant changes in the

redox state of red cells occur as a function of hemoglobinoxygenation; (b) the NAD+/NADH and glucose/sorbitolratios vary inversely with deoxygenation indicating that thepolyol pathway may act as an electron shuttle betweenNADPHand NADH; and (c) there is a marked changein the redox state of red cells with alteration in pH, espe-cially the NAD+/NADH ratio, which can only partly beaccounted for by mass action.

150. Selective T-Cell Killing of Human Lymphocytes byUltraviolet Radiation. SHELDON HOROWITZ,* DEREKCRIPPS,* AND RICHARD HONG, Madison, Wis.

The effects of ultraviolet radiation (UV) on human B-and T-lymphocytes were studied. Peripheral blood lympho-cytes from (12) normal adults were irradiated with agermicidal lamp (GL) in graded doses (1-20 min) and withmonochromatic UV of 254, 280, and 296 nm. B-lymphocyteswere defined as those bearing surface immunoglobulinmarkers detected by immunofluorescence and T-lymphocyteswere defined by formation of rosettes with sheep red bloodcells. Functional activity was assessed in vitro by mitogenicstimulation: PHA for T-cells and pokeweed mitogen(PWM) for B-cells. Viability was determined by eosin-dye exclusion. Low intensities of the GL had a minimaleffect on the viability of B- and T-lymphocytes and mitogenresponsiveness. Longer exposures (0.19-0.67 Ws/cm2) killedT-cells (50-100%) without significant effects on B-cells.For each individual, a UV intensity was found which re-sulted in loss of rosette-forming capacity and ability to re-spond to PHA but preservation of normal numbers ofsurface immunoglobulin-bearing cells and response to PWM.At higher intensities (0.67-2.0 W* s/cm2), in addition to100% T-cell killing, there was significant B-cell killing (50-100%7o). The most effective monochromatic UV for T-cellkilling was at 254 nm with less effect at 280 and 296 nm.It is concluded that graded doses of UV can selectively de-stroy human peripheral blood T-lymphocytes, while allow-ing functional B-lymphocytes to survive. (Research sup-ported by NIH grants 5-TO1-00317-07, AM-15086, andAM-09995.)

151. Effect of Beta-Adrenergic Blockade on Left Ven-tricular Function in Exercise. LAWRENCED. HORWITZ,*JAMES M. ATKINS,* AND STEPHENJ. LESHIN,* Dallas, Tex.(introduced by Robert L. Johnson, Jr.**).

The hemodynamic effects of propranolol (1 mg/kg intra-venously) were evaluated in 10 dogs running a graded exer-cise regimen on a treadmill. Measurements were made of leftventricular internal diameter with a sonocardiometer, leftventricular pressure with a solid-state intracardiac gauge,ascending aorta flow with an electromagnetic flowmeter, andarteriovenous oxygen difference with pulmonary artery andaorta blood. Peak work loads were reduced by propranolol.At the same levels of mild and moderate exercise mean totaloxygen consumptions were 31.3 and 35.5 ml/min per kg dur-ing exercise without blockade and 27.6 and 28.1 with pro-pranolol. Propranolol significantly reduced heart rate, strokevolume, left ventricular systolic pressure, and left ventriculardp/dt max, and increased left ventricular end-systolic diam-eter at all levels of exercise. During mild exercise left ven-tricular end-diastolic diameter and arteriovenous oxygen dif-ference were increased with propranolol. During moderateand severe exercise these differences were no longer present.It is concluded that beta-block limits exercise capacity byattenuating the heart rate response and the rate and extentof muscle fiber shortening. At comparable levels of sub-maximal exercise, propranolol resulted in reduced total oxy-

41 a

gen consumption, and the exercise loads were accomplishedwith lower cardiac outputs and developed tensions. (Re-search supported by a grant from the NHLI.)

152. The Role of Pulmonary Surfactant in the Bacteri-cidal Activity of Alveolar Macrophages During OxygenToxicity. GARY HUBER,* DENISE O'CONNELL,* AND MARCLAFORCE,* Boston, Mass. (introduced by Edward Kass **).

Although the alveolar macrophage (AM) is the key cell inantibacterial defenses in the lung, the bactericidal role ofpulmonary surfactant is controversial. Prolonged exposureto 100% oxygen resulted in progressive depression in vivo ofthe intrapulmonary inactivation of an aerosolized challengeof radiolabeled (8'P) Staphylococcus aureus. Oxygen expo-sure also induced progressive increases in alveolar surfacetension forces (quantified by pressure-volume relationshipsand correlation of alveolar bubble stability indices with ul-trastructural morphometry) and decreases in surface activityand phospholipid phosphorus recoverable by branchopulmo-nary lavage. However, AM from control and oxygen-toxicrats were comparable in number, viability, and phagocyticuptake of staphylococci in vitro. Bacterial precoating with anacellular surfactant fraction (SF), recovered by centrifuga-tion of lavage fluid (40,000 g), was required for intracellularkilling by AMwhen evaluated in vitro with lysostaphin, inspite of the presence of excess strain-specificity antibody andfresh complement. SF by itself had no bactericidal activity.Incorporation of SF from control animals provided equiva-lent bactericidal activity for both control and oxygen-toxicAM, whereas incorporation of SF from oxygen-toxic animalsresulted in no intracellular killing by either control or oxy-gen-toxic AM. In vivo adaptive tolerance to oxygen toxicity,induced by intermittent exposure to oxygen and characterizedmorphometrically by proliferation of granular pneumocytes(cellular origin of surfactant), attenuated the impairmentin bactericidal activity and in recoverable surfactant. Theseexperiments imply that bacterial coating with normal SFstimulates bactericidal activity of both control and oxygen-toxic AMand that the depressed bactericidal activity in theoxygen-toxic lung is secondary to alterations in SF ratherthan to a direct impairment in the AM.

153. Isolation and Characterization of DNA, RNA, andImmune Complexes from Systemic Lupus Erythemato-sus (SLE) and Normal Plasma. DICK HUNTER,* JEANETTEDILLEY,* ANDHALSTEDHOLMAN,** Stanford, Calif.

Soluble complexes of DNAand antibody to DNA appearto cause SLE nephritis. Antibody to RNA is often associ-ated with other clinical forms of rheumatic disease. Antigenicplasma nucleic acids and soluble complexes of those nucleicacids with antibody have not hitherto been isolated or char-acterized. A chromatographic method will be reported forisolation of DNAand RNAfrom plasma. Both nucleic acidsare present in a single peak trailing behind the plasma pro-teins. Approximately i of the peak is DNA. This elutionposition is quite different from that of DNA or RNAchromatographed alone. The DNA is very antigenic. Bychemical and immunological criteria it appears to be nativeDNA. The buoyant density (ca. 1.720) of the isolated, puri-fied DNA free of RNA is different from known mammalianDNA. The RNA forms unusual orcinol color reactions andis poorly reactive immunologically. DNA and RNA arepresent in normal plasma at a total concentration approxi-mately 5 mg/100 ml. They appear increased in SLE, espe-cially during disease activity, reaching levels between 10and 20 mg/100 ml. At times, isolated DNA and RNA areaccompanied by small amounts of IgG with high specificity

for native DNA. This IgG is presumably a component of acomplex. The plasma nucleic acids may be previously un-known human types or may be derived from either alterednormal human nucleic acids or microorganisms. Their par-ticipation in tissue damage may depend upon the type ofantibody produced in the disease. (Research supported bygrant from NIH.)

154. Sodium Chloride, Urea, and Water Transport in theThin Ascending Limb of Henle: Generation of OsmoticGradients by Passive Diffusion of Solutes. MASASHIIMAI * AND JUHA KOKKO,* Dallas, Tex. (introduced byDonald W. Seldin **).

Studies were designed to examine whether the thin ascend-ing limb of Henle (TALH) decreases its luminal solute con-centration by an active or a passive transport process. In allexperiments isolated segments of rabbit TALH were per-fused in vitro. When tubules were perfused with solutionsidentical with the bath, active transport of NaCl was ex-cluded by the following: (a) collected fluid osmolality re-mained unchanged and same as the bath; (b) net waterreabsorption could not be demonstrated; and (c) transtubu-lar potential difference was zero. Isotopic permeability co-efficients (X 10' cm/s) were calculated from the disappear-ance rate of the respective isotope added to the perfusate.These values indicate that TALH is moderately permeableto [1"C] urea (6.74±1.04) while having a higher permeabilityto GNa (24.9±1.9) and TMCl (116+9.8) than any other seg-ment similarly studied. Osmotic water permeability was zero.When TALH were perfused with a 600 mOsm/liter solutionpredominantly of NaCl against a 600 mOsm/liter bath inwhich 50%o of osmolality was NaCl and 50%o urea (to simu-late in vivo papillary interstitium), the collected fluid os-molality was decreased significantly below that of the bath(300 mOsm/liter per mmof tubule). The decrease in os-molality was due to greater efflux of NaCl as compared toinflux of urea. Weconclude that (a) TALH is not capableof active salt transport and (b) the unique membrane char-acteristics of TALH allows for generation of osmotic gradi-ents (lumen less concentrated than adjacent surroundings)on purely passive mechanisms when perfused with isosmolalsalt solutions in a bath with appropriate salt and urea con-centrations. These findings are consistent with the passivecountercurrent model previously proposed. (Research sup-ported by grants from NIH.)

155. On the Apparent Increase in Production of Natri-uretic Factor in Chronic Uremia. E. IPAKCHI,* JACQUESBOURGOIGNIE,* K. HWANG,* AND NEAL S. BRICKER, Bronx,N. Y.

Serum of patients and dogs with chronic uremia containsa natriuretic factor which is not detectable in normal serum.That the inhibitor is a physiologic mediator of the adaptivenatriuresis per nephron is suggested by its absence in sodium-retaining uremic patients with nephrotic syndrome and ure-mic dogs in which the natriuresis is obviated by "propor-tional reduction" of sodium. The present studies representan effort to detect the inhibitor in urine from normal anduremic subjects. If an inhibitor accumulates in uremia be-cause of failure of excretion, it should be less detectable inuremic than in normal urine. However, if production is in-creased or degradation decreased, excretion might be greaterdespite low glomerular filtration rates (GFR's). Quantita-tive urine samples, collected from chronically uremic patientsand normal subjects, were lyophilized, fractionated by gelfiltration, and bioassayed in rats. The total fraction excretedin 2 h was infused in 1 ml of Ringer's. A single fraction

42 a

from uremic patients produced natriuresis (AUN.V = 126,AEq/h); the same fraction from normal subjects was non-natriuretic (AUNaV = 24 uEq/h (P < 0.05 uremic vs. nor-mal). The natriuretic fractions from uremic urine anduremic serum have similar characteristics: (a) both appearin Sephadex G-25 eluates immediately after the salt andurea peaks, (b) both resist freezing and boiling, (c) bothare inactivated by leucine-aminopeptidase, and (d) neitherincreases GFR. If the inhibitors in urine and serum are thesame, elevated serum levels are not due to diminished excre-tion. Rather, the rate of synthesis must be increased and/orthe rate of inactivation decreased. In either event, feedbacksuppression is not evident. The cumulative data support theconcept that adaptation in sodium excretion in chronic ure-mia is associated with elaboration of a humoral natriureticagent.

156. PGE2 Acts As an Intrarenal Hormone RegulatingMedullary Vascular Resistance. HAROLD D. ITSKOVITZ,*NORBERTOA. TERRAGNO,* ANDJOHN C. McGIFF, Milwaukee,Wis.

The role of PGE2 as a local hormone in the regulation ofintrarenal hemodynamics was assessed in isolated blood per-fused canine kidneys. Solvent extraction and thin-layerchromatography, coupled with parallel bioassay, were used tomeasure "PGE," expressed as PGE2 equivalents. Perfusateconcentrations of PGE in six experiments increased from0.04+0.02 (SEM) ng/ml during the initial 100 min to 0.80±0.40 ng/ml by 200 min and to 2.23±0.66 ng/ml by 400 min.Renal blood flow increased pari passu. In four experiments,indomethacin (5 mg) reduced perfusate concentrations ofPGEwithin 30 min from 1.37±0.58 ng/ml to 0.31±0.23 ng/ml. These results suggest a high synthetic rate for PGE inisolated kidneys which could be inhibited by indomethacin.Before indomethacin the intrarenal distribution of blood flow,as measured by radioactive microspheres, changed from aratio of 79:21 (outer half; inner half of renal cortex) to64:36 (P < 0.01). Inhibition of prostaglandin synthesis byindomethacin reversed this trend with a resultant increasedfractional blood flow to the outer cortex (85 :15; P < 0.025);simultaneously, renal blood flow decreased from 221+16 to147±15 ml/min (P < 0.01), in spite of keeping renal per-fusion pressure constant. Infusion of PGE2 at rates as greatas 2 ,g/min did not prevent a redistribution of blood flow tothe outer cortex after giving indomethacin, pointing to theimportance of tissue synthesis rather than circulating levelsof PGE2 as the critical factor responsible for these hemo-dynamic responses. PGE2presumably acts as a local hormoneat its site of synthesis within the renal medulla to affectmedullary vascular resistance and thereby blood flow, both toinner cortical and medullary areas. (Research supported bygrants HL 12677 and HL 13624 from NIH.)

157. Synthesis of Antihemophilic Factor (AHF) by Cul-tured Human Endothelial Cells. ERIC A. JAFFE,* LEON W.HOYER,* AND RALPH L. NACHMAN,Farmington, Conn., andNew York.

Previous studies have identified AHF antigen exclusivelyin endothelial cells in human tissue sections by immuno-fluorescence using a monospecific rabbit antibody to humanAHF (1972. Fed. Proc. 31: 262, Abstr.). Evidence for syn-thesis was sought using cultured endothelial cells derivedfrom human umbilical veins. These cells were differentiatedfrom blood vessel medial smooth muscle cells and fibro-blasts by cell morphology and growth pattern, and thepresence of Weibel-Palade bodies, smooth muscle actomyosin,and ABH antigens. AHF antigen was identified in cultured

endothelial cells by immunofluorescence. Soluble AHF anti-gen was also detected in culture media from endothelial cellsby specific radioimmunoassay using 'I-labeled anti-AHF.Culture medium not previously exposed to cells was devoidof antigen as were control media obtained from culturedsmooth muscle cells and fibroblasts. To determine whetherthe AHF antigen in endothelial cell culture fluid representedrelease of previously stored or newly synthesized protein, thecultured cells were pulse labeled with radioactive aminoacids. High molecular weight, AHF antigen-rich protein wasseparated from the radioactively labeled culture medium byagarose-gel chromatography. In three separate experiments,monospecific anti-AHF precipitated 8.0±2.1% of the radio-activity in this fraction, whereas normal rabbit serum pre-cipitated 1.7±0.8%. Incorporation of radioactive amino acidsinto the immune precipitates was inhibited by culturingendothelial cells in the presence of cycloheximide. No AHFclotting activity was recovered in the tissue culture media,suggesting either that the endothelial cells synthesize an in-active molecule or that the functional activity is destroyedby a thrombin-like enzyme which we identified in the culturemedium fetal calf serum.

158. Sulfodibromophthalein Excretion in Man. NORMANB. JAVITT, DANIEL DHUMEAUX,* AND PIERRE BERTHELOT,*New York and Creteil, France.

Wehave found that sulfodibromophthalein (DBSP, pheno3,6-dibromophthalein disulfonate) is not metabolized by manand therefore provides new insights into the critical deter-minants of hepatic transport function particularly when com-pared to sulfobromophthalein (BSP), which is conjugatedmostly with glutathione. Eight individuals without liver dis-ease and nine with cirrhosis were given equimolar amountsof BSP and- DBSP in random sequence within 36 h. Eachstudy consisted of three different infusions of 30 min dura-tion. Storage (S, /Amole/,tmole per 100 ml) and maximumexcretion rate (Tm, /Amole/min) were calculated from simul-taneous equations (Wheeler method). In the normal group,S as measured by DBSP was 125±57 (mean+SD), morethan double that estimated by BSP (58±34). In cirrhosis,the reduction in S measured by DBSP (49±34) was propor-tionate to the reduction estimated using BSP (21±15). Anexcellent correlation (r = 0.96) was found between the esti-mation of S by both BSP and DBSP. Tm for BSP was9.0+1.2 (mean+SD), not significantly different (P > 0.10)from the Tm for DBSP (10.9±1.6) in the normal group.Comparable reductions occurred in cirrhosis (BSP=4.0+t1.2;DBSP= 5.5±1.2). These studies demonstrate that (a)DBSP can be used safely in man, (b) DBSP is distributedin man as a single species in plasma, liver, and bile, thusproviding a more accurate estimation of S and (c) regardlessof the possible role of glutathione as a determinant of BSPexcretion, a profound defect in membrane-mediated transportmust occur to account for the reduction in DBSP Tm and Sin liver disease. (Research supported by grant AM-13094from the NIH.)

159. Normal Plasma Renin Activity in Malignant Hyper-tension. JAMES G. JOHNSON,* SANDRA LEE,* SERGIO Ac-CHIARDO,* CESAR CUESTAs,* LEONARDSHARE,* AND FRED E.HATCH,* Memphis, Tenn. (introduced by E. E. Muirhead**).

The malignant phase of hypertension, defined as severehypertension, renal insufficiency, and hypertensive retinopathyincluding papilledema, is observed in all hypertensive settingswith the possible exception of coarctation of the aorta. Recentinvestigations have emphasized the importance of renin-angiotensin-aldosterone system in the pathogenesis of the

43 a

malignant phase. To evaluate this concept, 21 black patientswith malignant hypertension were studied, by measuringplasma renin activity (PRA) by radioimmunoassay ofangiotensin I before treatment with diazoxide and furose-mide. The age range of the 21 patients was 23-50 yr. Therewere 9 females and 12 males. Blood was obtained immedi-ately after admission and PRA was determined in twoseparate laboratories using different antibodies to angiotensinI. Of the 21 patients, 10 had PRA clearly within the normalrange for sodium-loaded upright patients, and 7 had PRAwithin the normal range for sodium-loaded recumbent pa-tients. The range of PRA for the "normal reninemic" groupwas 0.2-4.4 ng/ml per h, whereas the range for the "highreninemic" group was 5.6-58 ng/mg per h. Of the normalreninemic group, 3 were admitted with oliguric renal failureand 2 others subsequently died of cerebral vascular accidents(CVA). One other patient in this group experienced twoCVA's. Of the 11 patients in the high reninemic group, 3were admitted with oliguric renal failure and one suffereda CVA. These data fail to support the concept that an ele-vated PRA is mandatory for the development and mainten-ance of the malignant phase of hypertension. (Researchsupported by a grant from NHLI.)

160. Inhibition of Phagocytic Bactericidal Activity bySuperoxide Dismutase: A Possible Role for SuperoxideAnion in the Killing of Phagocytized Bacteria. RICHARDB. JOHNSTON, JR.,* BERNARDKxEmL,* LAWRENCEWEBB,*DALE KESSLER,* AND K. V. RAJAGOPALAN,* Birmingham,Ala., and Durham, N. C. (introduced by Max D. Cooper).

Enzyme systems capable of generating superoxide anion(OJ), a highly reactive free radical of oxygen, have beenidentified in many mammalian tissues. The physiological con-trol of 02- has been attributed to superoxide dismutase(SOD), which catalyzes the removal of superoxide anion (O,-+ 02 + 2H+SOD -- H202 + 02). The product of this reaction,hydrogen peroxide (H202), has been implicated in the killingof phagocytized bacteria; it has not been clear whether02 itself serves a useful biological purpose. Wehave foundSOD in leukocytes from normal humans and patients withchronic granulomatous disease and have demonstrated anti-bacterial activity of an O. -gene'rating system. In an attemptto determine if Oa- might be involved in phagocytic bac-tericidal activity, 'we studied the killing of E. coli and S.aureus by normal leukocytes in the presence of SOD. Whenthe enzyme was fixed to latex particles which could enterthe phagocytic vacuole along with bacteria, there was almostcomplete inhibition of phagocytic killing by SODcomparedto albumin or heat-denatured SOD as controls. Similar in-hibition of killing was effected by'catalase, which catalyzesremoval of H202. SOD did not inhibit ingestion, as mea-sured by uptake of '4C-labeled E. coli, by counting numbersof ingested bacteria and yeasts on smear, or by measurementof the respiratory burst which accompanies phagocytosis.These findings are compatible with the concept that both02- and H202 are required for the optimal killing of ingestedbacteria by phagocytes, perhaps through generation of highlyoxidative hydroxyl radicals (OH-) in the reaction 02 +H20S -OH-+OH-+02.

161. Lecithin-Cholesterol Acyl Transferase Deficiency inAcute Pancreatitis. DON PAUL JONES,* Detroit, Mich.(introduced by Benjamin Lewis**).

Patients with pancreatitis may have hyperlipemia with de-creased percentages of serum cholesterol esters. Decreasedcholesterol esters correlate with decreased serum cholesterolesterifying activity (CEA) in patients with liver disease.

We have studied CEA in pancreatitis and also have deter-mined whether decreased CEA reflects lecithin cholesterolacyl transferase (LCAT) deficiency or an opposing serumcholesterol ester hydrolase (CEH). CEA, LCAT, and CEHwere evaluated in 13 patients with liver disease and in 17patients with pancreatitis without liver disease. CEA assayquantitated the decrease in concentration of serum unesteri-fied cholesterol during 5 h incubation. LCAT assay deter-mined the rate of [H2] cholesterol ester synthesis in serumcontaining [H3] cholesterol-labeled lipoproteins. CEH wasassayed (a) under incubation conditions for hepatic CEHassay, and (b) by measuring [H3] cholesterol formationfrom a unique substrate containing [H3] cholesterol esters innatural serum lipoproteins. 90%o of [H3] cholesterol esterswere hydrolyzed when substrate was incubated with dogplasma containing known pancreatic CEH. Mean CEA andLCAT activity were decreased in both patient groups (pan-creatitis: CEA, 48%, LCAT, 66%; liver disease: CEA,72%, LCAT, 65% of normal). In both, CEA and LCATcorrelated positively with each other (pancreatitis: r = 0.819[P < 0.05]; liver disease: r = 0.869 [P < 0.01]) and withthe percent of cholesterol esters. No CEHwas found. Serafrom four patients with pancreatitis and LCAT deficiencydid not inhibit LCAT in normal sera. Patients with pan-creatitis acquire LCAT deficiency which may be relatedetiologically to pancreatic hyperlipemia. Low CEA asso-ciated with both pancreatitis and liver disease representsLCAT deficiency rather than the presence of CEH. Noserum LCAT inhibitor was found. LCAT deficiency in pan-creatitis merits study of mechanism, diagnostic and prog-nostic value, and relationships to hyperlipemia.

162. A Possible Bifunctional Role for Cystathiosine Syn-thase. OLIVER W. JONES* AND MARYA. GRIsHAvER,* LaJolla, Calif. (introduced by William L. Nyhan).

Patients with homocystinuria due to cystathionine synthasedeficiency fall into two major categories regarding bio-chemical responsiveness to pyridoxal phosphate: those whorespond by reversal of methioninemia and homocystinuriaand those who have no change in the specific aminoacidemia/aminoaciduria after pyridoxal phosphate therapy. We findan enzyme in human fetal tissue which catalyzes the re-

serineaction: H2S + serine - ) cysteine, bypassing the re-

sulfhydraseaction requiring cystathionine synthase. Serine sulfhydraseis present in human fetal skin, liver, and brain and has beenpurified 100- to 150-fold. There is an absolute requirement forserine and pyridoxal phosphate. Serine sulfhydrase is alsodemonstrable in human adult liver. Km values for pyridoxalphosphate with fetal serine sulfhydrase are 5- to 6-foldhigher than adult serine sulfhydrase suggesting that moreavailable pyridoxal phosphate is required to achieve maximalcatalytic activity with fetal enzyme. Serine sulfhydrase andcystathionine synthase have the following similarities: (a)identical purification through six steps including elution atthe same point by DEAE chromatography; (b) similartemperature sensitivity; and (c) the same mobility on gelelectrophoresis and identical response to serine and sulfhydrylanalogs. Our results suggest that in human fetal tissue cer-tain enzyme protein may have bifunctional catalytic proper-ties. If cysteine biosynthesis is crucial to normal development,alternate routes for a specific biochemical reaction might beadvantageous. In contrast, a point mutation on the enzymepolypeptide chain might render it incapable of 'catalyzing onereaction but capable of catalyzing another. It is possible thatsome homocystinuric mutants might still have serine sulf-hydrase activity and cysteine biosynthesis in response to

pyridoxal phosphate, while the catalytic site for cystathioninesynthase is rendered inactive by mutation. (Supported bygrants from NIH and The National Foundation.)

163. Modulation of the Immunologic Release of ChemicalMediators from HumanLung and Nasal Polyps. MICHAELKALINER* AND K. FRANKAUSTEN, Boston, Mass.

Fragments of human lung and nasal polyps passively sensi-tized with human IgE release histamine, slow reacting sub-stance of anaphylaxis (SRS-A), and eosinophil chemotacticfactor of anaphylaxis (ECF-A) upon antigen challenge. Theimmunologically activated secretion of these mediators fromboth tissues is suppressed by beta-adrenergic or prostaglandinstimulation and enhanced by alpha-adrenergic and cholin-ergic stimulation. In human lung tissue, the evidence impli-cating hormonally induced alterations in the tissue concen-trations of cyclic AMP in modulation of mediator releaseincludes: (a) the kinetic relationship between increases incyclic AMPand inhibition of mediator release after stimula-tion with isoproterenol or the prostaglandins; (b) the paral-lel rank order of potency (ID50) of isoproterenol (7.5X10-8M) --epinephrine (5X10'M) > norepinephrine (10'M)> PGE1 (5 X 104M) > PGR2a (5 X 10'M) in terms of sup-pressing mediator release and increasing cyclic AMP; (c)the synergism expressed between stimulators of adenylatecyclase and inhibitors of phosphodiesterase upon both phe-nomena; and (d) the association of depletions of cyclic AMPlevels with enhancement of mediator release produced byalpha-adrenergic and a low dose prostaglandin (5X10'M)stimulation or by the nonhormonal agent imidazole. Cholin-ergic enhancement (Carbachol 109-10'M) of mediator re-lease, presumably acting through increases in cyclic GMPlevels, is independent of measurable changes in cyclic AMP.Whereas the requirements for divalent cations and energy inthe antigen-activated secretory process conforms to the usualrequirements for secretion, the inhibitory effect of cyclicAMP is an exception. One could argue that the immuno-logic release of chemical mediators is antihomeostatic andthat the inhibitory actions of cyclic AMPon mediator re-lease favor homeostasis. (Supported by grants AI-07722 andRR-05669 from the NIH.)

164. Collagen-Induced Platelet Aggregation: Inhibitionby Benorylate. ANDREWH. KANG,* EDWIN H. BEACHEY,*ANDRICHARDL. KATZMAN,* Memphis, Tenn. (introduced byLester Van Middlesworth**).

It is well established that native collagen both in solutionand native-type fibril form induces platelet aggregation, andit has been suggested that the intact secondary and tertiarystructures of collagen are essential in the process. In viewof published evidence suggesting the involvement of plateletmembrane-bound glycosyltransferases in the process, wehave investigated the possible role of the primary structureof collagen. Contrary to several previous reports, our dataclearly indicate that denatured, purified polypeptide chain,al, is capable of inducing platelet aggregation in vitro usingstandard techniques. Furthermore, of the nine CNBr peptidesof al, only one peptide, al-CB5, containing the only residueof Glu-Gal-Hyl in al, is active. In addition, purified Glu-Gal-Hyl inhibited collagen-induced aggregation, indicatingthat the primary structure of collagen also plays a role inplatelet aggregation. Because aspirin is a known inhibitor ofnative collagen-induced platelet aggregation, it was of in-terest to test if it also interfered with al-induced aggrega-tion. In addition, benorylate, a new lipid-soluble ester ofacetylsalicylic acid and N-acetyl p-aminophenol, was alsotested, since the latter had been reported to possess anti-

inflammatory and antirheumatic properties comparable toaspirin but to cause significantly less gastrointestinal bleed-ing. Benorylate or aspirin in 100- or 200-mg doses was ad-ministered to prefasted (24 h) New Zealand White rabbits(2 kg). 2 h after ingestion, aggregation times of plateletsobtained from aspirin or benorylate-fed rabbits were (a)aspirin 100 mg, 8.5 min; 200 mg, no aggregation after 12min and (b) benorylate, 100 mg, 3.5 min; 200 mg, 9.0 min.Control platelets aggregated within 3 min. These resultsshow that, on a molar basis, the inhibitory effect of benory-late upon platelet aggregation is identical with that of aspirin,and that factors other than interference with platelet aggre-gation must account for the reportedly greater amounts ofgastrointestinal bleeding observed with aspirin than withbenorylate. (Supported by grants from NIH and VA.)

165. Crystalline IgMX Cytoplasmic Inclusions in ChronicLymphocytic Leukemic Lymphocytes: a New Syndrome?MANUEL E. KAPLAN, ROBERT E. RYDELL,* AND CONNIECLARK,* Minneapolis, Minn.

Cytoplasmic crystalline inclusions have been observed veryinfrequently in peripheral blood lymphocytes (PBL) fromapparently normal individuals and from patients with chroniclymphocytic leukemia (CLL). The purpose of this report isto describe, in 4 of 29 CLL patients, cytoplasmic crystallinematerial made up, in each case, of IgMX. Isolated PBL wereroutinely stained for membrane-bound immunoglobulin (M-Ig) and cytoplasmic immunoglobulin (C-Ig) with fluoro-chrome-conjugated (FC) antisera specific for t-, y-, a-, K-,and X-determinants. Patients were categorized on the basisof large populations of PBL containing monoclonal M-Ig:

M-Ig type

No. of patients KA" -YK aK p Mixed None

29 8 9 3 1 1 1 6

In four of nine M-IgMX (1X) patients, crystalline cytoplas-mic inclusions staining specifically with FC anti-s. and anti-Xwere detected. Electron microscopy showed the crystals with-in the rough endoplasmic reticulum. Similar crystalline in-clusions were found in none of the remaining 25 patients.The statistical probability that this finding would occur solelyby chance is 0.005. Double fluorescent staining of M-Ig andC-Ig, utilizing fluorescein- and rhodamine-conjugated anti-sera, revealed that essentially all PBL containing crystalsexhibited M-IgMX, whereas a considerably larger populationof cells had M-IgX but lacked crystals. In none of the pa-tients with crystals studied was a paraprotein demonstrablein serum or urine. These studies indicate that (a) function-ally different populations of neoplastic lymphocytes may existin a given CLL patient; and (b) the mechanism(s) resultingin membrane insertion and/or extracellular secretion ofIgMX by CLL lymphocytes is frequently abnormal and mayresult in its intracellular crystallization. (Research was sup-ported by grant AM 13717 from the NIH and grant 01/4828.1 from the VA.)

166. Platelet Adenine Nucleotide Metabolism DuringAdenosine Diphosphate (ADP)-Induced Aggregation.HERMANE. KATTLOVE,* Torrance, Calif. (introduced byKouichi R. Tanaka**).

Inhibition of platelet metabolism leads to decreased aggre-gation in response to ADP. This is thought to be related todecreased platelet adenosine triphosphate (ATP) content.The purpose of these studies was to determine the role ofATP in ADP-induced platelet aggregation. [3H] adenine-labeled human platelet-rich plasma (PRP) was incubated

45 a

with 5 mM2 deoxy-D-glucose (2 DG), together with either2 mMKCN or 20 ,g/ml oligomycin, and tested simultane-ously for extent of ADP-induced aggregation and plateletATP content. Aggregation declined and correlated closely(r = +0.93, P < 0.001) with the decrease in platelet ATP asmeasured by either chemical techniques or by radioactivetechniques after separation by high voltage electrophoresis.This correlation suggests an important role for ATP inaggregation. This role was demonstrated when [3H] adenine-labeled PRP was exposed to ADP in the presence of meta-bolic inhibitors such as 3.5 mMKCN. After 30-60 s, 10-20%of the ATP was converted to inosine monophosphate (IMP).This occurred whether or not the PRP was stirred. This didnot occur in PRP exposed to ADP in the absence of meta-bolic inhibitors. Further studies demonstrated that withinthe first 5 s after ADPaddition, 10-20%o of the platelet ATPwas first converted to ADP which was subsequently con-verted to IMP. The presumed effect of the KCN in thesestudies is to prevent the regeneration of ATP. Other in-hibitors, such as 2 DGalone or oligomycin alone, were alsoeffective. The ADP-induced breakdown of ATP was notprevented by 1 mMouabain. We conclude that ATP isutilized during ADP-induced platelet aggregation. This isconsistent with the hypothesis that ADP causes aggregationby stimulating the platelet contractile protein, thrombos-thenin, and we would postulate that the ATP breakdown weobserved is due to the ATPase activity of this protein.(Supported by NIH grant HE 14344.)

167. Epinephrine-Induced Enhancement of MyocardialContractility: Possible Mediation by p-Receptor:Adeny-late Cyclase:Adenosine 3', 5'-Monophosphate-DependentProtein Kinase :Calcium Transport System Located onthe Sarcoplasmic Reticulum. ARNOLDM. KATZ, MADELEINEA. KIRCHBERGER,* MICHIHIKo TADA,* AND DORIS I. REPKE,*New York.

The positive inotropic effect of epinephrine may resultfrom stimulation of adenylate cyclase (ACase), which cata-lyzes adenosine-3' ,5'-monophosphate (cAMP) formation.cAMP has recently been shown to stimulate cAMP-depen-dent protein kinase (PK) to phosphorylate cardiac micro-somes that are rich in sarcoplasmic reticulum (SR) mem-branes, concurrently enhancing both Ca uptake and Ca-activated ATPase. This more rapid Ca transport into SRcould effect the classical epinephrine effects on the heart:inotropy and abbreviation of systole, the former due to in-creased intracellular retention of activator Ca++. In thepresent study, we examined the possibility that all compon-ents of this control system are located on the SR. Cardiacmicrosomes prepared in dilute bicarbonate buffer representenriched SR, viz. 45 nmoles/mg Ca-binding sites, and haveminimal contamination by plasma membrane fragments, viz.(Na++K+) -activated ATPase < 1 umoles/mg per h. ACaseactivity (111±9 [SEMI pmoles/mg per min), which wasstimulated by epinephrine (Km-5X10' M epinephrine), wassignificantly greater than ACase of microsomes prepared insucrose (90±4 pmoles/mg per min) which have fewer Ca-binding sites (13 nmoles/mg) and are rich in plasma mem-brane markers, viz. (Na++K+) -activated ATPase > 4,umoles/mg per hr. This indicates that SR possesses both 8-receptors and ACase. Enriched SR also contain active PK,viz. histone phosphorylation is at least doubled by addedcAMPand is phosphorylated by endogenous PK (0.3 nmolesP per mg per 10 min) with twofold stimulation by addedcAMP (Km~107 M cAMP). Furthermore, these endogen-ous systems can enhance Ca uptake after epinephrine. Epi-nephrine-induced inotropy thus may result from sequential

stimulation of a-receptor, ACase, PK, and a Ca transportsystem all located on the SR, and need not require increasedcytol)lasmic cAMP. (Supported by NHLI and NYHA.)

168. Control of Aldosterone Secretion in Supine Man.FRED H. KATZ,* GARY0. ZERBE,* AND STROTHERH. WALK-ER,* Denver, Colo. (introduced by Karl E. Sussman).

Striking coincidence between peak plasma levels of aldo-sterone and cortisol in the early morning in two normalsupine subjects was recently reported from this laboratory.Eight additional supine normal men and women have nowbeen studied on the metabolic ward and the relationship be-tween plasma aldosterone, renin activity, and cortisol in the10 experiments investigated by regression analysis. Two ofthe subjects received 10-meq sodium diets, and the rest 110-meq sodium diets. One of the four women on normal sodiumwas taking oral contraceptives. Plasma was sampled every30 min for 24 h or every 10 min from midnight to 8:00 a.m.Peak plasma aldosterone in all was seen in the early morninghours. The regression analyses revealed that of the eightsubjects on normal sodium only four showed a statisticallysignificant (P = 0.05 or less) correlation between renin andaldosterone levels. The correlation was also significant in thetwo on sodium restriction. The correlation of cortisol andaldosterone however was significant in all 10 subjects. Twoother men were given dexamethasone., 0.75 mg every 12 hfor 5 days, and subjected to a 24 h study while receiving anormal sodium diet. Although their plasma cortisol concen-trations were constantly below the normal range, the earlymorning peak aldosterone levels, like those in untreated sub-jects, persisted on dexamethasone, and were correlated withrenin. These results suggest that aldosterone secretion isunder the control of the same central nervous system bio-rhythm which governs ACTH-induced cortisol secretion,although suppression of ACTH by exogenous steroid doesnot affect the bio-rhythmicity of aldosterone secretion. (Sup-ported by grants from The Population Council, NIH, andthe VA.)

169. The Low Renin State: Definition and Implications.DAVID C. KEM,* NORMANJ. KRAMER,* CELSO GOMEZ-SAN-CHEZ,* MARTIN WHITE,* AND NORMANM. KAPLAN, Dallas,Tex.

Low plasma renin activity (PRA) has been used as amarker for determining diagnosis and prognosis in the hy-pertensive population. Wehave (a) simplified the screeningtechnique for recognition of low PRA, (b) compared whiteand black normotensives and hypertensives, and (c) ex-amined the relation between low PRA and hypertensivevascular complications. (a) PRA was performed by radio-immunoassay after generation at pH 5.5. PRA 30 min after40 mg Lasix intravenously plus upright posture (Lasix test)was highly correlated (P < 0.01) to PRA after 2 (r =+0.79) and 4 h (r = +0.58) upright and PRA after 3 dayslow sodium diet plus 2 h upright (r = +0.81, P < 0.01).Lasix values were higher than upright alone but lower thanlow salt plus upright. (b) Lasix test PRA's were similarin 30 normotensive whites (mean 5.3+0.57 SE) and in 18hypertensive whites (4.3±0.67) but significantly higher (P <0.02) than in 23 normotensive blacks (3.4±0.60) who had, inturn, significantly higher values (P <0.02) than did 27 hy-pertensive blacks (1.6±0.39). PRA's were lower in blacksdespite higher renin substrates. (c) Lasix PRA's were aslow (1.6+0.59) in 13 black hypertensives who had experi-enced a stroke or myocardial infarction as in the 27 withoutcomplications. Thus (a) the PRA state may be defined by afairly simple procedure, (b) blacks have lower PRA than

46 a

whites so that published comparisons between normotensivewhites and hypertensive blacks may be invalid; in additionto other factors, race must be considered in comparinggroups, and (c) low PRA does not seem to protect againsthypertensive complications. (Supported by NIH grant I-ROI-HL-14863-01.)

170. Demonstration of Fatty Acid-Binding Proteins As-sociated with the Brush Borders of Hamster IntestinalEpithelium. J. I. KESSLER AND S. MISHKIN,* Montreal,Quebec, Canada.

We have shown that the uptake of fatty acids (FA) bythe small intestine is accomplished through a reversible as-sociation of FA to the epithelium. Analogous results wereobtained with brush borders suggesting that the FA bindingmay be the property of the microvillous membrane. Thisstudy was undertaken in an attempt to define further the FA-binding properties of hamster brush borders. Micellar [14C]FA or ["C] FA dissolved in 50,ul dioxane:propylene glycol(2:1 v/v) was mixed with brush borders and kept for 1 hat 40C. After three washings the brush borders were dis-rupted and five fractions (A, B, C, C1, and D) were obtainedby discontinuous density gradient centrifugation (1965. J.Cell. Biol. 26: 687). The microvillous membrane fractions(C and C') contained 50-75% of the radioactivity and over98% of it was in the form of free FA. These fractions weresonnicated and the supernatant was chromatographed onSephadex G-75. The radioactivity was associated with thevoid volume peak and with a low molecular weight (10-12,000) protein. On addition of unlabeled FA, part of theradioactive FA from the low molecular weight protein wasdisplaced to the void volume peak. Fractions A, B, and Dand preparations of basal and lateral cell membranes (1969.Biochim. Biophys. Acta. 173: 456 and 1972. Biochim. Bio-phys Acta. 274: 336) exhibited only minimal FA binding.These results indicate that the FA binding by the intestinalepithelium may be related to FA-binding proteins in themicrovillous membrane.

171. Specificities of Peptide Hydrolases in the BrushBorder and Cytosol Fractions of the Rat Small Intestine.YOUNGS. KiM, YONGW. KIM,* AND MARVIN H. SLEISEN-GER,** San Francisco, Calif.

Peptide hydrolases have been shown to be localized in thebrush border and cytosol fractions of the small intestine. Inthe present study, the specificities of peptide hydrolases fromthe brush border and the cytosol fractions of the rat smallintestinal mucosa were examined with respect to the lengthof the peptide substrates and the amino acid residues hy-drolyzed. Partially purified enzymes were prepared as de-scribed previously (1972. J. Clin. Invest. 51: 1135). Thequantitation and identification of the products of enzymatichydrolysis was carried out using an amino acid analyzer.When heteropeptides and homopeptides up to six amino acidsin length were examined as substrates, the cytosol fractionwas found to be capable of hydrolyzing only the dipeptidesand tripeptides; the brush border fraction hydrolyzed thesepeptides and those of greater length. The cytosol preparationhad an apparent Km for L-ala-gly of 2.35 mM, and for L-ala-(gly)2, 0.95 mM. The brush border fraction had the follow-ing apparent Km values: L-ala-gly, 0.65 mM; L-ala-(gly)21.29 mM; L-ala-(gly)3, 1.01 mM; and L-ala-(gly)4, 0.84 mM.Timed incubation studies using the above and other hetero-peptides revealed that both brush border and cytosol peptidehydrolases cleaved amino acid residues sequentially from theamino termini of the peptides. Using dipeptide through hexa-peptide substrates, no carboxypeptidase or endopeptidase

activity was observed in either fraction. These results, in ad-dition to the zymogram studies reported previously, suggestthat the brush border and cytosol peptide hydrolases mayserve two different cellular functions. The cytosol hydrolaseactivity is probably only involved in the final breakdown ofpeptides, while the brush border has the capability of hy-drolyzing peptides of larger size which may be the primaryproducts of dietary protein digestion. (Supported by VAresearch grant and VA RETR-48.)

172. Increased Glucose-Induced Proinsulin Levels withIncreasing Age. ABBAS E. KITABCHI AND WILLIAM C.DUCKWORTH,*Memphis, Tenn.

Although deterioration of glucose tolerance with age hasbeen associated with increasing age, the contributory role ofthe circulating insulin and/or the less active precursor, pro-insulin, has not been elucidated. The total immunoreactiveinsulin (TIR) and proinsulin-like material (PLM) in re-sponse to oral glucose was studied in 70 nonobese subjectsof varying ages with normal glucose tolerance tests. Plasmafor glucose, TIR, and PLM was obtained before and at 30-min intervals after a 100 g oral glucose load. Although eachindividual subject had a normal glucose response, the meanglucose levels increased with increasing age. In the threeolder groups (ages 45-54, 55-64, and 6574) the integratedglucose areas above fasting were all significantly greater(P < 0.05) than in the youngest group (ages 15-24). Nosignificant differences in TIR responses were seen, butplasma PLM was significantly higher in the older-age sub-jects. All values except fasting were significantly greater insubjects aged 45-74 years, as compared with subjects 15-44years of age (P < 0.01). The total areas under the PLMcurves were also significantly greater (P < 0.01) in the threeolder groups, as compared with the youngest group. A sig-nificant correlation (P < 0.001) between the amount of pro-insulin and the age of the subject was seen while no signifi-cant correlation between TIR and age could be demonstrated.These findings may reflect a decreased conversion of proin-sulin to insulin in the aging pancreas which could play a rolein the glucose intolerance associated with aging. (Supportedby grants from VA and NIH.)

173. Erythrocyte Membrane Lipid Abnormalities in Hy-pophosphatemic Hemolysis. JOHN C. KLOCKAND STEPHENB. SHOHET, San Francisco, Calif. (introduced by Louis K.Diamond).

To evaluate possible distortions in lipid membrane compo-sition or renewal in hypophosphatemic erythrocytes, we havemeasured erythrocyte phospholipids from two severely hy-pophosphatemic patients with hemolytic anemia (serum phos-phate 0.4 and 0.2 mg/100 ml). Neither patient was jaundicedor hyperlipemic. The blood smear showed marked anisocyto-sis with many spherocytes, some echinocytes, and approxi-mately 20% normal discocytes (probably secondary to trans-fusions). Total erythrocyte lipids were markedly elevated(187, 165 -yP/cc cells vs. normal 125y). This was mostlydue to an increase in phosphatidylcholine (PC) (76; 71-yP/cccells vs. normal 35y5). Erythrocyte cholesterol was un-remarkable. The ratios of phosphatidylcholine to phos-phatidylethanolamine (PC: PE) were increased (2.54; 2.60vs. normal 0.95±0.08). Erythrocyte ATP levels were alsoreduced (0.91; 0.76 ,um/cc cells vs. normal 1.68+0.12) andosmotic resistance was increased, consistent with elevatedmembrane lipid (50% hemolysis at 0.36% saline). After 24 hincubation in autologous plasma, considerable loss of mem-brane lipid was observed (20% vs. normal < 7%). In con-trast to similar incubations of hereditary spherocytosis cells,

47 a

this loss was not wholly symmetric (PC: PE of 2.54->2.04). After parenteral phosphate therapy, all abnormalitieswere gradually corrected. Hemolysis in hypophosphatemichemolytic anemia has been demonstrated by others to be as-sociated with severe ATP depletion. Cell stiffness secondaryto loss of CA++ chelating effect of ATP has been suggestedas one mechanism for this hemolysis. Hemolysis has alsobeen demonstrated in a kindred of patients who have abnor-malities similar to those noted here (high erythrocyte PC,abnormal PC: PE, and increased osmotic resistance). Apossible additional mechanism for hemolysis in patients withsevere hypophosphatemia may be that low levels of ATPaffect the balance and stability of membrane lipids. (Sup-ported by NIH grants AM-16095 and AM-37337.)

174. On the Relationship of Hyperuricemia and Hyper-uricosuria to Muscle Injury and Their Possible Role inHeat Stress Nephropathy. J. P. KNOCHEL, Dallas, Tex.(introduced by Floyd Rector, Jr).

Myoglobinuria and hyperuricemia may be prominent fac-tors in the pathogenesis of acute renal failure during physicaltraining in hot climates. Indeed, hyperuricemia may attainvalues observed in patients with urate nephropathy asso-ciated with other conditions. Myoglobinuria may be an asso-ciated phenomenon and conditions favoring precipitation ofuric acid and myoglobin, such as dehydration and excretionof a concentrated, acid urine, commonly prevail under suchcircumstances. Studies were designed to determine if hyper-uricemia was due to decreased excretion or increased produc-tion of uric acid (UA) and if related to increased production,whether it was possibly associated with skeletal muscle in-jury. Weekly observations were made in hot and coolweather on 22 healthy subjects during 35 days of stereo-typed physical training under metabolic balance conditions.Dietary intake was constant (nitrogen = 23.4 g/day). Themost striking changes were noted during the second weekof training. Mean uric acid concentration [UA] in serumrose from 5.8(3.9-6.8) to 8.3(6.1-12.2) mg/100 ml (P <0.001). Mean uric acid excretion (UurV) rose from637(454-855) to 1031(792-1555) mg/day (P<0.001). Uricacid clearance (Cur) did not change and at no time couldserum [UA] be ascribed to a fall of creatinine clearance.Mean CPK activity rose from 7(0-16) to 58(9-304) U/ml(P <0.001). Mean creatine excretion rose from 52(32-73)to 92(37-124) mg/24 h (P<0.001). While there was noeffect of climatic temperature on the foregoing, those train-ing in hot weather had a mean [UA] in urine of 62(32-102)mg/100 ml; mean urine volume 1311(755-1842) ml/day andmean urine osmolality of 828(521-1128) mOsm/Kg H20. Incontrast, mean values for subjects in cool weather were[UA] in urine, 35(26-42); urine volume 2787(2080-3655)and urine osmolality 449(367-525). All values were signif-icantly different (P <0.001). We conclude that (a) hyper-uricemia during physical conditioning is due to overproduc-tion of uric acid and not to a net decrease of uric acidclearance; (b) overproduction of uric acid can be corre-lated with muscle injury; and (c) hyperuricemia and hyper-uricemia and hyperuricosuria may play a role in heat stressnephropathy.

175. Lipoprotein Changes in Pregnancy: a DistinctEndogenous Hypertriglyceridemia. ROBERTH. KNOPP*ANDMARIA R. WARTH,* Boston, Mass. (introduced by Ronald A.Arky).

Plasma triglyceride and cholesterol rise threefold and aquarterfold in normal pregnancy. To determine a mechanism

for these changes, triglyceride, cholesterol, and phospholipidwere measured in the very low density, low density, and highdensity lipoprotein fractions (VLDL, LDL, and HDL, re-spectively). Lipoproteins were separated by ultracentrifuga-tion and precipitation with MnCl2 and heparin. VLDL proteinwas determined by the Lowry method. In 10 nonpregnantsubjects, triglycerides in VLDL, LDL, and HDL, respec-tively, were: (mean±SEM) 23.7±3.6, 11.5±1.8, and 25.1±2.5 mg/100 ml. In 16 nondiabetic third trimester subjectstriglycerides were significantly elevated in each fraction:84.6+7.1 (VLDL), 67.5±4.4 (LDL), and 37.7±2.6 (HDL)mg/100 ml (P < 0.01 in each instance). These data pointto lipoproteins with larger neutral lipid cores. Phospholipidwas measured to determine if the polar surface coats werecorrespondingly increased. Phospholipid increased signif-icantly (72%o) only in the VLDL (P <0.01), as did chol-esterol (P < 0.001). Similarly, VLDL protein increased80%o (P <0.001). Thus total VLDL rose from 81.5+6.5 to154.3±+13.8 mg/100 ml in pregnancy. The increased levels ofVLDL in pregnancy are best explained by a balanced in-crease in the formation of all VLDL constituents. By con-trast only triglycerides increased in the LDL, suggestingan accumulation of light LDL particles seen in the St 12-20fraction (LDL intermediate). Paper electrophoresis of theVLDL fraction disclosed no beta mobility, excluding defec-tive LDL catabolism and over-flow into the VLDL. The datasupport an increased conversion of VLDL into LDL thatexceeds the capacity for triglyceride removal from the LDLintermediate. These adaptations can make triglyceride fuelmore readily available to lipid utilizing tissues in pregnancy.(Supported by NIH.)

176. Stem Cell Kinetics in RFMLeukemic Mice. W. H.KNOSPE,* S. HusSEINI,* F. E. TROBAUGH, JR.,* AND W.FRIED, Chicago, Ill.

RFMmice spontaneously develop myeloid leukemia whichis transplantable by a cellular mechanism into nonleukemicRFMmice. The leukemic process progresses to death within3-4 wk, and leukemic cells can be identified by a chromosomenumber of 39 as contrasted to 40 in nonleukemic cells.Colony-forming units (CFU) from marrow, spleen, andblood were studied by the spleen colony technique (suspen-sions of cells from these organs were injected into lethallyirradiated mice and the number of resulting macroscopicspleen colonies was counted as per McCulloch and Till) be-fore and after injection of nonleukemic RFM's with 106leukemic cells. The karyotype and histologic type of cellsin the resultant spleen colonies was studied. Marrow andspleen CFU increased 3- to 20-fold 2 wk after injection ofleukemic cells. 3 wk after injection, the number of marrowCFU dropped to near normal levels while splenic and bloodCFU remained high. A normal distribution of erythroid,granulocytic, and megakaryocytic colonies resulted from cellsobtained 1 wk after injection of leukemic cells, whereas amarked increase in granulocytic colonies and a marked de-crease in erythroid colonies resulted from cells obtained 2and 3 wk after injection of leukemic cells. Cytogenic studiesshowed all spleen colonies from cells 1 wk after injectioncontained cells with 40 chromosomes, whereas nearly allcolonies from cells obtained after 3 wk had 39 chromosomes.The latter ywere all of granulocytic type. No cells in eryth-roid colonies contained 39 chromosomes. We conclude thatthe leukemic RFMstem cells are capable only of differentiat-ing into the granulocytic series. As these cells increase innumber, growth of nonleukemic stem cells is inhibited. (Re-search supported by grants from Leukemia Research Founda-tion and NIH.)

48 a

177. Sleep EEG Stages and Growth Hormone (GH)Levels in Endogenous and Exogenous Hypercortiso-lemia. DOROTHYT. KRIEGER, MARKFINGERHUT,* AND SEY-MOURM. GLICK, New York.

Studies of nocturnal sleep EEG stages, plasma GH, andcortisol levels were performed on the following subjects:(A) six with active Cushing's disease (CD); (B) fivewith CD in remission; (C) one with Cushing's syndrome(adrenal adenoma) pre- and postadrenalectomy; (D) onewith CS (parasellar tumor) pre- and postradiotherapy; (E)three with Nelson's syndrome; (F) three with hypothalamictumors (normal basal endocrine status); (G) five receivingchronic high dose prednisone treatment; and (H) fournormal volunteers during a constant hydrocortisone infusion,after a loading dose 2 h before sleep onset. Compared tonormal subjects, stage II sleep was increased and slow wavesleep (SWS: stage III, IV) was decreased in five of sixgroup A patients; three also showed a decrease in REMsleep. Similar stage I and SWSchanges were noted in fourof five group B patients; patient C pre- and postadrena-lectomy; patient D pre- and postirradiation and all threegroup F patients. In contrast, percent times for all sleepstages were essentially within normal limits in group E,G, and H patients, save for a decrease in REMsleep in thelatter. Analysis of GH data has been completed in groupsA, B, C, D, E, H. A marked decrease in the nocturnal incre-ment (A < 5 ng/ml) was observed in all except patient Cpostadrenalectomy. Thus, abnormalities in sleep EEG stagesII and SWSand lack of nocturnal GH elevation in patientswith CD are not solely related to the hypercortisolemia,since such changes are absent in groups F and G and presentin CD in remission, as well as in normocortisolemic patientswith hypothalamic tumors. Normal sleep stage percentagesin group E patients may be related to decreased plasma corti-costeroid and/or markedly elevated plasma ACTH levels,whereas the significance of altered sleep stages in patient Cis unclear. A normal nocturnal GH rise was absent in groupE patients, indicating a dissociation between normal SWSand such a rise. (Research supported by NIH grants NB-02893, FR-71, and AM-09219.)

178. Familial Hyperbetalipoproteinemia (Type II Hyper-lipoproteinemia) in Children. P. 0. KwITEROVICH,* R. I.LEVY, AND D. S. FREDRICKSON,** Bethesda, Md.

236 children of 90 matings, hyperbetalipoproteinemia (II)X normal (N), were studied. They represented all but fourof the living offspring, age 1-19, from these matings (twowere not ascertained; two, being propositi, were excluded).Two log normal distributions were obtained for concentra-tions of both low density (CLDL) and plasma (C) choles-terol. Likelihood ratio test favored a bimodal distribution ofCLDL (x = 18.41, P < 0.0005) and C (x = 7.81, P < 0.025)over a unimodal distribution. The derived normal cutoff forCLDL of 164 mg/100 ml was used to calculate a segregationratio of 45: 55 (II: N), consistent with a monogenic traitwith dominant phenotype. If a frequency for II in the generalpopulation of 5% was assumed, a polygenic model washighly unlikely (x' = 65.80, P< 10-6) (Edward's test: ex-pected 22%, observed 45%). The percentage of abnormalchildren in the first decade (52) significantly exceeded thatin the second (39) (P <0.01). The mean triglyceride washigher in the hyperbetalipoproteinemics (P < 0.0025). Thechildren from 35 matings: type IIB (hyperbetalipoprotein-emia and hyperglyceridemia) X non-type II included 39 IIA,8 IIB, 1 with type IV, and 52 N. From 55 matings: type IIAX non-type II, there were 56 IIA, 2 IIb, 2 IV, and 76 N.A child of a parent with type IIB was more likely to have

IIB than a child of a IIA parent (x'= 5.55, P < 0.05). Thedata strongly support early expression of hyperbetalipo-proteinemia inherited by a monogenic model in these chil-dren. More than one genetic defect within the group is notexcluded.

179. Epidemiologic Studies of Cytomegalovirus (CMV)in Semen. DAVID J. LANG,* RICHARD B. FROST,* LAWRENCEJ. D'ANGELO,* AND JOHN F. KUMMER,* Durham, N. C. (in-troduced by Jerome S. Harris).

We have previously demonstrated the presence of CMVin human semen and suggested that this virus may in someinstances be venereally transmitted. The index patient wasrecuperating from a heterophile-negative mononucleosis andhad a history of recent gonococcal infection. Wepresent herethe results of our initial epidemiologic studies of CMVinsemen. CMV was not recovered from any of the semenspecimens of 19 healthy men. Samples of semen were ob-tained from 94 male members of couples seeking a fertilityevaluation; one semen was CMV positive. 16 universitystudents with a history of mononucleosis contributed speci-mens. Two yielded CMVon culture. 10 men with activebacterial venereal infections were also studied. One of these10 semen specimens contained infectious CMV. A medicalhistory was available from three of the four individuals withpositive semen cultures. All three had multiple sexual part-ners in the recent past. Two had a history of documentedbacterial venereal infections. Two of the three are homo-sexuals. These preliminary results support the thesis thatCMVmay be transmitted by venereal contact and that ex-posure to multiple sexual partners increases the risk of ac-quiring this virus infection as it does the defined venerealdiseases. Infections with CMVmay be subclinical and, evenif symptomatic, the infected patient may continue to shedvirus long after the illness has resolved. We have reportedelsewhere that CMV has remained demonstrable in thesemen of the asymptomatic index case for more than 10months. The full importance of the presence of CMVin themale and female genital tracts, beyond its transmission to thefetus and neonate, remains to be determined. (Research sup-ported by NIH NHLI-72-2910B and NIH 2 ROL-HD-AI-04183-02.)

180. Cytomegalovirus (CMV) in Human Semen. DAVID J.LANG,* JOHN F. KUMMER,* AND DAVID P. HARTLEY,* Dur-ham, N. C. (introduced by Samuel L. Katz).

In 1972 we reported the isolation of CMVfrom the semenof a young man recuperating from heterophile-negativemononucleosis. We report here further studies of CMVinsemen and document the persistence of virus for 10 months.Semen containing 107 * 7 tissue culture infectious doses(TCID50) per ml was studied. The cellular and fluid ele-ments of the semen were separated and the CMVpresentin each quantitated. 99% of the infectious CMVin semenwas found in the supernatant. Samples of semen were ex-amined by electron microscopy. Herpesvirus particles weredemonstrable in large extracellular aggregates. No virus wasseen within or adsorbed to spermatozoa or inflammatorycells. During early convalescence total sperm counts and themotility of spermatozoa were depressed, and the pH of thesemen was abnormally high. Subsequently, these values re-turned to within the normal range while CMV persistedin high titer. 10 months after its initial demonstration, CMVwas still present in semen although the titer was reduced to102 TCID5o/ml. The patient has been asymptomatic through-out this time and has remained sexually active. Cytomegalo-virus was recovered from the cervix of a sexual contact. To

49 a

determine whether multiplication of this virus can occur inspermatozoa or associated cells, semen from uninfected indi-viduals was inoculated with CMVin vitro. There was no evi-dence of virus adsorption or replication. Our studies indicatethat CMVmay persist in semen, primarily in extracellularfluids, while not replicating in spermatozoa nor necessarilyassociated with recognizable impairment of sperm function.However, the presence of CMV in semen may be an im-portant factor in the frequent demonstration of this virusin the female genital tract. (Research supported by NIHNHLI-72-2910B and NIH 2 ROL-HD-AI-04183-02.)

181. Clinical and Pharmacologic Studies of AdenineArabinoside (ara-A) in Serious Herpesvirus Infections.CARL B. LAUTER,* ELIZABETH J. BAILEY,* FRANCIS M. WIL-SON,* AND A. MARTIN LERNER, Detroit, Mich.

On account of in vitro and in vivo (animal model) activityof ara-A vs. herpesviruses of man and its relative lack oftoxicity in these systems, clinical trials are clearly warranted.Ara-A was administered parenterally in schedules of 2.5-20mg/kg per day to nine patients with serious virologicallyconfirmed herpes simplex (HSV), varicellazoster (V-Z),or cytomegalovirus (CMV) infections. Seven of nine pa-tients were debilitated or immunologically depressed. Non-interferon, nonantibody antiviral activity (AVA) due toara-A or its active metabolite, ara-hypoxanthine, wasassayed by a specific microbiologic method with a sensitivityof 50 Atg/0.4 ml. AVA in sera, urines, and vesicular fluidsranged from 50 to 200 ug/0.4 ml, and was found only infour adults receiving 20 mg/kg per day of ara-A. SporadicAVA was found in urine when intravenous ara-A was in-fused at a rate of 60 mg/hr, but was constant when de-livered more rapidly. Urinary excretion accounted for 17-38% of the daily dose. Improvement and resolution of clinicalfindings in four patients with disseminated V-Z occurred in5.7 and 8.7 days, respectively. Similar means for two patientswith generalized HSV were 2.5 and 6 days. A patient withherpes progenitalis relapsed after ara-A was stopped, anda newborn with CMVdid not respond. Toxicities were (a)a transient extrapyramidal syndrome in two patients withHodgkin's disease receiving the highest dosage of ara-A,(b) transient falls in hemoglobin (>2 g/100 ml) in sevenof the ten treatment courses, and (c) transient transaminaserise in a single patient. There were no white blood cell,platelet, or renal or cardiac side effects. In generalized HSVor V-Z infections ara-A appeared strikingly effective, and aclear advance over available agents (idoxuridine or cytosinearabinoside) which have limited usefulness in immunolog-ically depressed, thrombocytopenic, or leukopenic patients.(Research supported by grants from the NIH and Parke-Davis Co.)

182. Hereditary Deficiency of Sixth Component of Com-plement (C6) in Man. J. P. LEDDY, M. M. FRANK,*T. GAITHER,* R. S. HEUSKINVELD,* R. T. BRECKENRIDGE,*AND M. R. KLEMPERER,* Rochester, N. Y., and Bethesda,Md.

A young black female (D. B.) in good general health wasfound to lack serum hemolytic complement (C) activity. AllC components were normal except C6 which was undetect-able by both functional and immunoprecipitin assays. Thesefindings could not be accounted for by a C6 inhibitor. C6 wasalso absent in D. B. plasma. Hemolytic C (CH50) titerson family members were normal; however, both parents andfive of six available siblings had approximately half-normalC6 levels by functional assay. The other sibling was normal.Biologic properties of D. B. serum included; (a) absent

bactericidal activity against S. typhi 0 901 with or withoutadded rabbit antibody; (b) normal generation of chemotacticactivity for human neutrophils in the presence of endotoxinor aggregated IgG; (c) ability to sensitize appropriate cellsfor immune adherence or agglutination by anti-C3 Coombsserum; and (d) inability to lyse PNH red cells in eitheracid hemolysis or "sugar water" tests. An extensive coagula-tion workup was normal although clotting time (250C) andprothrombin consumption (370C) in plastic tubes were, re-spectively, at the upper and lower limits of the normalrange. Addition of endotoxin (40 /Ag/ml) or inulin (1-2 mg/ml) to blood from D. B. and a panel of normal donors hadno consistent effect on clotting time or prothrombin con-sumption. These endotoxin and inulin preparations wereshown to activate the alternate C pathway by several criteria.These studies document for the first time a human kindredwith C6 deficiency. This defect exhibits a classic Mendelianautosomal inheritance, with all three genotypes being recog-nizable. Heterozygotes would apparently escape detection bycommon screening methods such as CH50 titration. Unlikethe C6-deficient rabbits studied by others, the homozygousC6-deficient human exhibits chemotactic and coagulationfunctions within the range of normal. (Supported by NIHresearch grants.)

183. Circulating Prostaglandin A in Human EssentialHypertension. J. B. LEE, A. ATTALLAH,* V. K. VANCE,*C. ELLWOOD,* AND A. PREZYNA,* Buffalo, N. Y.

We have hypothesized that essential hypertension may bethe result of intrarenal cortical vasoconstriction secondary toa deficiency of renomedullary PGAwhose normal role maybe to promote cortical vasodilation by circulating frommedulla to cortex by either systemic or intrarenal circula-tions. Since little is known of endogenous human renal anti-hypertensive prostaglandins, the present study was designedto test this hypothesis by determining the presence of PGAin human kidney and in normotensive and hypertensive sub-jects. PGAwas determined by radioimmunossay after silicicacid chromatographic separation of PGA from PGE, andPGF2. (sensitivity: 100 pg/ml). Results are expressed aspicograms per milliliter plasma or milligram tissue±SEM.In fresh surgical specimens from five normotensive humans[PGA] in cortex was 26±13; outer medulla, 309+163; andpapilla 1683+34 (n = 5; P = < 0.005). Peripheral plasma[PGA] in 14 patients with essential hypertension was 461+49, which was significantly lower (P = < 0.01) than in 15normotensive subjects (724+81). There was a significantnegative correlation between PGA and 24 h urinary sodium(212±23 mEq; range 85-364 mEq). In six essential hyper-tensive patients there was no difference between right(RRV) and left (LRV) renal venous blood although RRV+ LRV was significantly higher than the inferior vena cava(IVC) so that RRV+ LRV contributed 299 pg PGA toIVC. These results show that (a) PGA is present in humanrenal papilla in high concentration; (b) the kidney contrib-utes to the pool of circulatory PGA in patients with essen-tial hypertension; and (c) despite this contribution, it ap-pears that a deficiency of circulating PGA in essential hyper-tension may be a major etiological factor underlying thegenesis of this disorder.

184. Interaction of [3H] Norepinephrine with SpecificAdrenergic Binding Sites in Cell Membranes of CulturedMyocardial Cells. ROBERT LEFKOWITZ,* DONALDO'HARA,*AND JOSEPH WARSHAW,* Boston, Mass. (introduced byEdgar Haber).

50 a

Chronotropic and inotropic effects of catecholamines onmyocardium are probably mediated by binding to specificadrenergic receptors. To examine the initial interaction ofnorepinephrine with intact target tissue, in an environmentfree of vascular and neural elements, the binding of [3H]norepinephrine to cultured chick embryo myocardial cellswas studied. These cells contract rhythmically and their rateof beating is increased by catecholamines. [3H] norepine-phrine (3-5 X 10' M) was incubated with myocardial cells(5 X 105/ml) at 370C for 1 h and the [3H] norepinephrinebound to the cells was quantified by Millipore filtration andliquid scintillation counting. Catecholamines and relateddrugs displaced [3H] norepinephrine from these binding sites.Half-maximal displacement of binding occurred at 1-5 X 10-7M for isoproterenol, epinephrine, and norepinephrine and10' M for dopamine. a-Active amines such as phenylephrineand metaraminol competed only at concentrations above10' M [cf. Ertel et al. 1971. J. Pharmacol. Exp. Ther. 178:73; half-maximal stimulation of rate of beating of thesecells by norepinephrine at 3 X 10' M, comparable effect byphenylephrine at 12 X 10' M]. Catecholamine metabolites,cholinergic agonists and antogonists, and glucagon did notcompete for binding sites. Propranolol inhibited binding by40% at 10' M; phentolamine was inert. 2.5 X 10' sites werepresent per cell. Norepinephrine covalently bound to agarosebeads displaced 70% of [3H] norepinephrine when present at5 X 10' M. Trypsin destroyed the binding sites as did trypsinbound to agarose. These findings suggest that the cell sur-face is the predominant location of these sites. Phospholipasedigestion of the cells increased binding, presumably by ex-posing previously inaccessible sites. Parachloromercuri-benzoate, EDTA, and EGTA at 1 mMmarkedly inhibitedbinding, suggesting that the binding site contains crucial-SH groups and that small amounts of divalent cations arenecessary for the binding reaction.

185. Hemodynamic Basis of Renal Failure in AlcoholicHepatitis. OSOTH LEKAGUL,* ARIE KEYNAN,* BERNARDOKOTELANSKI,* AND JAY N. COHN, Washington, D. C.

Oliguria and azotemia complicating the course of alcoholichepatitis (AH) superimposed on cirrhosis is almost in-variably fatal. The renal failure has been attributed to unde-scribed hemodynamic factors. Splanchnic and systemic hemo-dynamic studies therefore were performed in 28 patients withalcoholic hepatitis, including 13 with (group I) and 15 with-out (group II) renal failure. Prothrombin time was similarin the two groups, but bilirubin was higher, albumin lower,and ascites greater in group I patients. Group I had sig-nificantly lower mean arterial pressure (71.1 mmHg com-pared to 79.8 mm Hg) (P < 0.05) and higher cardiacoutput (11.0 liter/min vs. 8.9 liter/min) (P <0.05) than ingroup II. Hepatic blood flow (isotope dilution method) wascomparably high in both groups (2002 ml/min vs. 2208 ml/min), but hepatic oxygen consumption tended to be lower ingroup I (45.2 ml/min vs. 55.4 ml/min). Hepatic fraction ofcardiac output averaged 24.7% in group II but only 19.1%in group I (P < 0.02). 12 of 13 patients in group I had com-plete porta-systemic shunting of mesenteric inflow comparedto only 6 of 15 in group II. Hepatic vein wedge-free pressuregradient was lower in group I (14.3 mmHg vs. 19.3 mmHg) (P < 0.02), indicative of more effective portal decom-pression in the oliguric group. These data indicate that renalfailure in AH occurs in patients with higher cardiac output,lower systemic vascular resistance, and more extensive porta-systematic shunting. Hepatic anoxia could contribute to thehepatic failure. The pattern is consistent with functional

arteriovenous shunting, probably largely in the portal bed.(Supported in part by NIH Grant HL 11533.)

186. The Limits of Renal Calcium Conservation in Pa.tients with Idiopathic Hypercalciuria. JACOB LEMANN, JR.,AND EDWARDJ. LENNON, Milwaukee, Wis.

Renal calcium wasting has been proposed as a mechanismof hypercalciuria among patients with idiopathic hypercal-ciuria and stones. The patterns and limits of renal calciumconservation were compared in 6 idiopathic hypercalciurics(S; 80±8.6 SEMkg) and 15 adults without a personal orfamily history of urinary stones (N; 70.4±2.8 kg). All atesimilar normal diets for 3 days providing an average of23.6±2.3 mMcalcium per day and then an otherwise com-parable liquid diet providing only 1.2±0.1 mMcalcium perday. Urines were collected every 6 h. Control UcaV averagedN 3.54+0.45 and S 8.77+0.68 mM/day, P < 0.001. The nadirof Uc6V occurred 18-24 h after starting the low Ca diet inboth groups and averaged N 0.06±0.01 and S 0.36±0.08 mMevery 6 h, P < 0.001. On the first day of dietary Ca depriva-tion Uc.V declined to N 1.22+0.18 and S 3.60±0.50 mM/day, P < 0.001, or to N 35+4 and S 42±6 per cent of con-trol (NS). Uc1V did not fall further, despite continuationof the low Ca diet for 7 days in four N and for 3 days inall other subjects. Control serum total [Ca] was comparablein both groups and fell significantly on the first day of die-tary Ca deprivation while serum [PO4] did not change,serum [Mg] and UPO4V rose, and UmgVand UNIV fell. Forboth diets N Uc6V mM/day = 1 + 0.1 diet Ca mM/day(r = 0.89) and S Uc6V mM/day = 3.1 + 0.3 diet Ca mM/day(r =0.92). Idiopathic hypercalciurics cannot achieve quanti-tatively normal renal calcium conservation during maximaldietary calcium deprivation because of a renal leak and/orgreater fractional reabsorption of calcium from gastro-intestinal secretions. (Supported by NIH AM 15089 andRR00058.)

187. Increased Collagen Synthesis by Scleroderma SkinFibroblasts In Vitro. E. CARWILE LEROY, New York.

Events leading to the irreversible fibrosis of skin andviscera which characterizes scleroderma (systemic sclerosis)are not understood. To study the mechanism of this fibrosis,skin fibroblasts from 20 subj ects with scleroderma (histo-logically confirmed) were paired with fibroblasts from nor-mal subjects matched for age, sex, and biopsy site. Diploidfibroblasts propagated from explants were compared inearly monolayer subcultures for soluble and insoluble col-lagen content, rate of collagen synthesis ([14C] proline in-corporation), effect of ascorbic acid, glycoprotein, andproteoglycan content, and cell doublings. Media obtained atdays 1, 6, and 12 were analyzed; cells were estimated atday 12 by number, DNA, and protein content Sclerodermacultures consistently contained more soluble collagen (non-dialyzable hydroxyproline, 86 ,ug/flask+ 14, mean±+SEM)compared with matched normal cultures (30 ,g/flask+ 11,P < 0.001). Insoluble collagen was similar in sclerodermaand normal cultures. In 14 experiments, [14C] proline incor-poration into collagen hydroxyproline was increased inscleroderma cultures, expressed both as mean dpm/flask(26,414±5,408 vs. 17,743+3,886, P < 0.001), and as meandpm/10' cells, (5,042±584 vs. 3,425±291, P < 0.01), indi-cating increased collagen synthesis by scleroderma fibro-blasts. Ascorbic acid stimulated scleroderma and normalcultures without affecting differences in collagen synthesis.Glycoprotein content (sialic acid) was slightly higher inscleroderma cultures; proteoglycan (uronic acid) contentwas similar. The cell doublings in 12 days were slightly

51 a

higher in scleroderma cultures (3.1±0.2 vs. 2.5±0.2, P <0.02). Collagen from scleroderma cultures, fibroblasts fromuninvolved scleroderma skin, and culture-mixing experimentsare under study. The data suggest that fibroblasts fromscleroderma skin synthesize collagen more rapidly thanfibroblasts from normal skin. The mechanism of this in-creased collagen synthesis in scleroderma is unclear. (Sup-ported by the Lifschultz and RGK Foundations and theUSPHS.)

188. Histidine-Mediated Decreases in Total Body Zincand in Urinary Porphyrins and Porphyrin Precursors inTwo Types of Hepatic Porphyria. J. B. LEVINE,* R.AAMODT,* D. P. TSCHUDY, AND R. HENKIN, Bethesda, Md.

Zinc and porphyrin metabolism were studied in three pa-tients with hepatic porphyria (one with cutanea tarda [PCT]and two with acute intermittent porphyria [AIP] on ametabolic ward at NIH while ingesting a constant diet. Mea-surements of urinary uroporphyrin and coproporphyrin inPCT, serum and urinary porphobilinogen (PBG) and 8-aminolevulinic acid (5-ALA) in AIP, serum and urinaryzinc and copper and total body 5Zn were made during thefollowing 12-16 day sequential periods: initial control, oraladministration of f5Zn, ZnSO4 (300-400 mg/day), and L-histidine (8-32 g/day), and final control. ZnSO4 significantlyincreased serum and urinary zinc, but had no consistent ef-fect on copper, porphyrins, or porphyrin precursors. L-histi-dine (24-32 g/day) significantly lowered urinary uropor-phyrin and coproporphyrin in PCT and significantly loweredserum and urinary PBG and d-ALA in AIP; concurrently,significant decreases in biological tj of 5Zn, increases inurinary zinc excretion, decreases in serum zinc without sig-nificant change in fecal 5Zn, occurred in all patients com-pared to the ZnSO4 period or to the previous control period.Withdrawal of L-histidine resulted in the rapid onset of sig-nificant increases in urine and serum 5-ALA and PBG inAIP. In PCT, urinary uroporphyrin and coproporphyrin re-mained significantly decreased after L-histidine withdrawalfor the remainder of the study. Simultaneous decreases inserum and urinary porphyrin precursors in AIP suggest thatL-histidine is not acting via renal mechanisms. Thus, the ob-served decreases in porphyrins and porphyrin precursors maybe related to some direct or indirect effect of L-histidine onporphyrin metabolism or to an L-histidine-mediated mobiliza-tion and excretion of a tightly bound tissue zinc pool.

189. Erythrocytes in Chronic Renal Disease: DefectiveAdaptation to Anemia. MARSHALLA. LICHTMAN,* MARIONS. MURPHY,* AND APRIL A. WHITBECK,* Rochester, N. Y.(introduced by Lawrence E. Young **).

Red cells in patients with hypoproliferative anemia withoutazotemia increase their content of 2,3-diphosphoglycerate(2,3-DPG) and hydrogen ion and thereby increase the effi-ciency of hemoglobin (hgb) function. We have studiedthis adaptive mechanism in anemic subjects with and with-out chronic renal disease (CRD) In 10, nonacidotic (pH =7.38+0.018), anemic (hgb = 7.22±0.36 g/100 ml) subjectswith CRD on maintainance hemodialysis, red cell 2,3-DPG(16.2+0.92 Amoles/g hgb) and P50 (25.6±0.31 mmHg) weresignificantly less than in 14 subjects with other hypoprolifer-ative anemias (hgb = 7.49+0.46 g/100 ml, 2,3 = DPG= 27.2±1.3 atmoles/g hgb; Pso 28.3+0.45 mmHg), but were simi-lar to values in 10 healthy subjects (hgb = 14.6+0.22 g/100ml; 2,3-DPG = 14.6+0.39 ,tmoles/g hgb; P5o = 26.4±0.26mmHg). P50 was highly correlated (r = 0.76) with 2,3-DPG/hgb molar ratio in the 34 subjects, although patientswith CRDhave a slightly lower P50 at a given 2,3-DPG/hgb

ratio. In vivo Pw0 based on cellular pH, 2,3-DPG, and hgbconcentration was 26.5±0.62, 25.7+0.87, and 30.8±0.91 mmHg in healthy subjects, patients with CRD and hypoprolif-erative, anemic subjects, respectively. However, red cellsfrom patients with CRD increased their 2,3-DPG and P5owhen incubated with inosine (I), pyruvate (P), and phos-phate (Pi) indicating responsive metabolic pathways andhemoglobin. Moreover, red cell adenosine triphosphate(ATP) was elevated in CRD (6.30+0.50 gmoles/g hgb) ascompared to healthy subjects (3.48+0.68 Amoles/g hgb).This did not affect P50, since red cell magnesium was alsoelevated (3.2±+1.8 jumoles/ml RBC; healthy subjects= 2.2±0.08), preventing ATP-hgb interaction in vivo. An inversecorrelation of ATP and 2,3-DPG in all anemic subjects, anda fall in ATP as 2,3-DPG rose in response to IPPi in vitro,suggested that red cell 1,3-DPG is preferentially metabolizedby its kinase in an azotemic environment, favoring ATP pro-duction, whereas, by its mutase in other hypoproliferativeanemias, favoring 2,3-DPG production. This study identifiesa defect in the ability of red cells in CRD to adapt so asto facilitate oxygen transport. Although such patients maycompensate by increasing cardiac output, the heightenedneed for transfusion therapy so as to reduce cardiac workand enhance patient performance should be considered.

190. Hemoglobin Brigham (2s152'00 pro - "u): a New Hemo-globin Variant Associated with Familial Erythrocytosis.JACOB J. LOKICH,* WILLIAM C. MALONEY,* H. FRANKLINBUNN, SALLY M. BRUCKHEIMER,* AND HELEN M. RAN-NEY,** Boston, Mass., and Buffalo, N. Y.

Members of a large family were found to have erythrocy-tosis due to the presence of a functionally abnormal hemo-globin. The propositus had increased whole blood oxygenaffinity with a P50 of 19 torr at pH 7.40, 37°C (normal = 26torr). 18 other family members, in 9 of whom decreased P50values were demonstrated, had erythrocytosis. Red cell2,3-DPG and urinary erythropoietin levels were normal. NoHb abnormality was detected by electrophoresis on starchgels (pH 8.6), agar gels (pH 6.5), isoelectric focusing, orion-exchange column chromatography. No abnormal sub-units were demonstrated in the hemoglobin after treatmentwith p-chloromercuribenzoate or 8 Murea. Peptide maps ofthe jS-chains of the hemolysate revealed normal Tp 11 andan additional spot of the same electrophoretic mobility, buta greater Rr. The substitution of leucine for proline at posi-tion 100 (G-2) was established by amino acid analysis andby automated Edman sequencing of the total 8 Tp 11. Inthis 'new" variant, designated Hemoglobin Brigham, thesubstitution of a bulky leucine residue for proline at G-2 1may impair the participation of G-1 aspartate in hydrogenbonding with tyrosine C-7 a 1, thus decreasing the stabilityof the deoxy derivative. Oxygen equilibria of phosphate-freehemolysates of the propositus showed curvilinear Hill plotswith increased oxygen affinity, normal Bohr effect and nor-mal reactivity to 2,3-DPG. Heme:globin ratio and the Soretabsorbance of the deoxy derivative were normal. Of the 12known hemoglobin variants associated with familial ery-throcytosis, Hemoglobin Brigham and Hemoglobin Olympia(a2s220 val .> met) (1972. J. Clin. Invest. 51: 70a) are electro-phoretically silent. Measurement of oxygen affinity is impor-tant in the diagnostic evaluation of unexplained erythrocy-tosis. (Supported by grant C6516 from the NCI and AM15234 from NIH.)

191. The Relationship Between Cytomegalovirus andHepatic Function Abnormalities Occurring in the Post-transplant Period. JAMES P. LUBY,* WILLIAM BURNETT,*

52 a

ALAN HULL,* AND ATHOLWARE,* Dallas, Tex. (introducedby Jay P. Sanford).

45 consecutive renal transplant recipients have been studiedfor 3-18 months with an average study time of 9.5 months.Sera were collected before transplant and at 1, 2, 3, 6, 9, 12,15, and 18 months and tested for complement fixation anti-body titers against cytomegalovirus (CMV) and herpes sim-plex virus, type 1 (HSV1). Before transplantation, 34.5%had CMVtiters > 1:4, while the comparable figure for HSV1was 42.9%o. After transplantation, 73.4%o had fourfold orgreater CMV titer rises, while only 18.4%o had such titerrises to HSV1. 46.5%o of the patients (20/44) developedchemical evidence of hepatitis (2 consecutive SGOT's > 65International Technicon units [95%o confidence limit]) aftertransplantation. 5 of the 20 were HAA+. 12 of the remaining15 patients had the onset of hepatitis at a time when theywere undergoing CMVseroconversion. The HAA- hepatitisattack rate in persons with CMVtiter rises 2°-23 was 16.7%(3/18) as contrasted with 60% (12/20) in those with titerrises 24-28 (P < 0.01) 13 of the 20 cases, including 7 associ-ated temporally with CMVseroconversion, had evidence ofcontinuing hepatic dysfunction at and beyond 120 days. Thestudy indicates comparability of initial CMV and HSV1antibody rates despite transfusion of an average of 7.7 Uper person (predominantly leukocyte-poor blood) < 3 monthsbefore transplant. It suggests that CMVmay be an impor-tant cause of hepatic function abnormalities in the posttrans-plant period and that some of the changes induced may bechronic. (Research supported by NIH grant 5 T01 AI00030-13).

192. Altered Vitamin B6f Metabolism and a Mechanism ofConditioned Deficiency in Chronic Alcohol Abuse. LAW-RENCELUMENG* AND TINK-KAI LI,* Indianapolis, Ind. (in-troduced by Walter J. Daly).

Vitamin B6 deficiency occurs in patients with alcoholicliver disease, but its existence in alcoholics without liverdisease remains unexplored. Using the plasma concentrationof pyridoxal phosphate (PLP), the biologically active formof B6, as indicator, we have compared alcoholic patients freeof clinical liver disease with age and sex-matched controls.Vitamin users were excluded. Plasma [PLP], measured bytyrosine decarboxylase apoenzyme, of alcoholics in the agegroups 20-35 yr (n=20), 36-49 yr (n=40), and 50-69 yr(n = 6) were 6.0±+1.0 (mean±SEM), 5.8+0.4, and 5.2+0.5ng/ml, respectively, and were significantly (P <0.001) lowerthan the values for the corresponding control groups (11.9+1.0, 11.0±0.5, and 8.2+0.6 ng/ml). Although diet maypartly be responsible, alcohol ingestion has been shown toimpair PLP synthesis. To define the mechanisms of this in-terference, the PLP synthesizing capacity of erythrocytesand the effects of alcohol and acetaldehyde were examined invitro. Alcoholics with lowered [PLP] exhibited normal B6-kinase and B6-phosphate oxidase activities. Acetaldehyde,0.05-5 mM, but not alcohol, inhibited the net synthesis ofPLP from Be and pyridoxol phosphate by 20-40%. This in-hibition was abolished by 80 mMphosphate. Studies with dis-rupted cell systems indicate that acetaldehyde does not alterkinase and oxidase activities, both present in soluble frac-tions, but activates a membrane associated, phosphate-sensi-tive phosphatase which hydrolyzes phosphorylated B6 deriva-tives. These data provide a mechanism for conditioned vita-min B6 deficiency in alcohol abuse. (Supported by grantsfrom USAMRDCand Licensed Beverage Industries.)

193. The Use of Derivatized Nylon Catheters for Selec-tive Immunoadsorption In Vivo. LEONR. LYLE,* CHARLES

W. PARKER, AND BRENT M. PARKER,* (assisted by JoanPoll *), St. Louis, Mo.

Clear nylon tubing (type 6, 1.34 mmoutside diameter)was partially hydrolyzed with acid and reacted with humanserum albumin (HSA) or ovalbumin (OA) in the presenceof a water-soluble carbodiimide. The ability of the antigen-substituted tubing to react specifically with antibody wasdemonstrated by incubation with a mixture of 'I-labeledrabbit anti-HSA and "3I-labeled rabbit anti-OA antibodies.The capacity of the HSA-nylon for anti-HSA antibody was0.5 ,ug/cm with little nonspecific binding as indicated by ahigh I/13"I radioactivity ratio and inhibition of 'I bind-ing by soluble HSA. The specificity and capacity of the OA-tubing for anti-OA antibody was comparable. The ability ofthe derivatized nylon to adsorb antibody in vivo was demon-strated in heparinized dogs. Lengths of HSA- and OA-nylonwere passed into the inferior vena cava through the femoralvein and shown to selectively bind 'SI- and "3'I-labeled anti-body, respectively. Direct kinetic measurements of antibodyadsorption in vivo indicated that if a reel type arrangementwere used (in which fresh conjugated nylon was slowlypassed through the venous system) milligram amounts ofantibody could easily be removed in a 24 h period. Sincethe nylon conjugation is applicable to a variety of proteins,haptens, and hapten-protein conjugates, the nylon cathetersystem could provide a simple means of selective removed ofunwanted antigens, antibodies, and substrates in man. Assuch it has potential applicability in the treatment of drugtoxicity, allergy, autoimmune disease, and malignancy. (Sup-ported by grants from NIH).

194. The Multiple Mixed Lymphocyte Reaction: a Uni-versal Antigen for Testing Cellular Immunocompetence?RICHARD J. MANGI * AND FRED S. KANTOR, New Haven,Conn.

Cellular immunity is assessed in humans by skin tests andby two in vitro correlates: lymphocyte stimulation andlymphokine production. The use of specific antigen stimulantsis not always possible because identification of specific im-munity in humans is variable. The present study attempts touse irradiated human lymphocytes as a universal antigen totest cellular immunocompetence. Conventional one-way mixedlymphocyte reactions (MLR) will produce a response whichvaries depending upon the histocompatibility of the donor andthe responding cells. To obtain maximal responses we havevaried the number of stimulating cells, keeping the reactantcells constant, and have also varied the number of donorscontributing to a constant number of stimulating cells.MLR's were performed with 1 X 10' reactant lymphocytes.The total number of donor lymphocytes contained equalparts of cells from one to five different unrelated indi-viduals. Maximum stimulation occurred when the ratioof donor to reactant cells was 2 to 4:1. At each concentra-tion of donor cells, maximal response occurred with ir-radiated lymphocytes from three different donors. Morethan three different donor populations resulted in no increaseor a decrease in response. At each total concentration of ir-radiated cells, the response elicited by multiple donors wasequal to or greater than the response to any one donoralone. Weconclude that MLRmaximum response is obtainedwith donor cells from three unrelated individuals culturedwith reactant cells in a ratio of 2 to 4:1. The multiple MLRis being assessed as a universal antigen in studies of lym-phocyte competence in anergic states and neoplasia and dur-ing immunosuppressive therapy. (Research supported in partby NIH grants AI53886-01 and AI 06760-06.)

53 a

195. Familial Thrombosis Due to Antithrombin III De-ficiency. EWAMARCINIAK * AND CLAUDEFARLEY,* Lexing-ton, Ky. (introduced by J. William Hollingsworth).

At the present time antithrombin III (AIII) is recognizedas a blood antiproteinase capable of inactivating both throm-bin and factor Xa and is presumably identical with the so-called heparin cofactor. Only one family with hereditary de-ficiency of AIII has been reported. A second large kindredunduly susceptible to venous thrombosis is the subject of thepresent study. For evaluation of AIII a procedure involvinggel filtration of plasma or serum on Sephadex G-150 withsubsequent analysis of antithrombin and antifactor Xa ac-tivities in separated fractions was utilized. Low AIII titresranging from 26 to 50% of normal values were found inplasma of eight members in the four generations, and an-other six members were suspected of having the deficiency.In serum of the affected patients AIII was virtually absentwhich indicates its eminent involvement in thrombin neutral-ization. There was no decrease in antithrombin or antifactorXa activities residing in the macroglobulin region. The re-sponsiveness to heparin in vivo measured in three deficientcases was only slightly below normal when the intensity andduration of the hypocoagulable state was estimated by con-ventional assay procedures. However, when the post-heparinAIII activity was evaluated, very significant discrepancy be-tween normal and deficient plasmas was found, reflecting thebiological stimulation of AIII by heparin. In five affectedmembers the therapy with oral anticoagulants increased thelevel of AIII in plasma by 15 to 45%o and gave a markedincrease of residual AIII in serum. These findings substanti-ate the outstanding biological role of AIII in the support ofblood fluidity and suggest that a decrease in vitamin K-de-pendent coagulation factors abates the utilization of AIIIin vitro and in vivo. (Supported by grant from AHA.)

196. The Role of Phospholipids in Thyroid-StimulatingHormone (TSH) Stimulation of Adenylate Cyclase inThyroid Plasma Membranes. KEITH MASHITER,* KAMEJIROYAMASHITA,* ANDJAMESB. FIELD,** Pittsburgh, Pa.

The observation that treatment of thyroid slices with phos-pholipase C inhibited TSH stimulation of cyclic AMP,glucose oxidation, and 8P incorporation into phospholipidsindicated the importance of phospholipids in TSH action.Pretreatment of bovine thyroid plasma membranes withphospholipase A (5 U/ml) or C (1 U/ml) abolished TSHstimulation of adenylate cyclase activity but had no effecton basal or NaF stimulation. Phosphatidylcholine (25 ,ug)and phosphatidylserine (200 gig), but not phosphatidylethanol-amine (200 /ig), partially restored TSH responsiveness ofbovine thyroid plasma membranes which had been treatedwith phospholipase A. Phosphatidylcholine produced agreater effect than phosphatidylserine and smaller amountswere effective. Both TSH and NaF stimulation of adenylatecyclase activity in thyroid plasma membranes was diminishedby 0.02% Lubrol Px treatment. The effect on TSH wasconsiderably greater. Phosphatidylcholine (200 /g) partiallyrestored TSH stimulation but not NaF activation of adenyl-ate cyclase. These results indicate that phospholipids, espe-cially phosphatidylcholine, are probably essential componentsin the system by which TSH stimulates adenylate cyclaseactivity in thyroid plasma membranes. The effects do notseem to involve the catalytic activity of adenylate cyclase,but the data do not permit a distinction between decreasedbinding of TSH to its receptor or impairment of the signalfrom the bound hormone to the enzyme activity. (Researchsupported by grant from NIH.)

197. Adaption to Hyperoxia: Influence on Protein Syn-thesis by Lung and on Granular Pneumocyte Ultrastruc-ture. DONALDMASSAROAND GLORIA D. MASSARO,* Wash-ington, D. C.

We have previously shown that in vivo exposure of ratsto 98% 02 for 48 h decreases total protein synthesis bylung slices. The decrease in synthesis of protein in a surface-active lung fraction (SAF) is greater than that of total pro-tein. There is a concomitant decrease in the size of lamellarbodies (currently thought to be surfactant storage granules)in granular pneumocytes of 02 animals. The present studywas designed to examine the influence of adaption to 02on these processes. We exposed rats to 98%o 02 or com-pressed air at ambient pressure for 96 h. Lung slices wereincubated with [14C] leucine and radioactivity measured inacid-insoluble protein. The SAF was obtained by ultra-centrifugation of homogenized lung. Wemeasured free leu-cine, DNA, and protein and examined granular pneumocytesusing ultrastructural stereological methods. After 96 h ex-posure there was more total acid-insoluble radioactivity inthe 02 group per mg DNA, 12,409±836 cpm/mg DNA,(mean±SEM) than in the air group, 6492±850 cpm/mgDNA (P = < 0.001). Acid-insoluble radioactivity in theSAF, expressed per milligram of crude homogenate DNA is4459+389 and 3140±251, respectively (P = <0.025), in 02and air rats. These differences are not due to differences infree leucine in the tissue. Ultrastructural lineal analysis ofgranular pneumocytes reveal they are larger in 02-exposedrats and contain more rough endoplasmic reticulum. Thelamellar bodies are equal in size in 02 and air rats after96 h, whereas at 48 h, when protein synthesis was decreased,the lamellar bodies in 02 rats were smaller. We concludethat during adaption to hyperoxia the lung regains its abilityto synthesize protein including protein in a SAF; this isreflected in the ultrastructure of the granular pneumocyte.

198. Secondary Hyperparathyroidism and Hypocalcemiain Acute Renal Failure in Man: Evidence for SkeletalResistance to Parathyroid Hormone. SHAUL G. MASSRY,ALLEN I. ARIEFF,* JACK W. COBURN, GENAROM. PALMIERI,*ANDCHARLESR. KLEEMAN,** Los Angeles, Calif.

Skeletal resistance to parathyroid hormone (PTH) mayunderlie the hypocalcemia of chronic renal failure. In acuterenal failure, hypocalcemia was thought to be due to ele-vated serum phosphorus. It is unknown whether resistanceto PTH action develops in acute renal failure and causeshypocalcemia. Also data on blood levels of PTH in acuterenal failure are limited. In 10 patients with acute renal fail-ure, serum calcium, phosphorus, and PTH were measuredfrequently. Infusions of parathyroid extract (PTE) weregiven during the oliguric and diuretic phases of acute renalfailure and 2-3 months after recovery. Hypocalcemia of 5.3to 8.4 mg/100 ml occurred as early as 3 days after onset ofacute renal failure and persisted during diuretic phase whenserum phosphorus was normal or low. Blood levels of PTHwere elevated 2-5 times above normal in nine of ten patients.Infusion of PTE caused a 15- to 40-fold rise in PTH levelsbut without calcemic response in nine patients during oliguricphase and in eight restudied during diuretic period. Thechange in serum calcium was 0-0.3 mg/100 ml (normal, 1.0-2.2 mg/100 mil). The failure to respond to PTEwas unrelatedto serum phosphorus. Five patients restudied with PTEwhen renal function was normal had a normal rise in serumcalcium of 1.2-2.2 mg/100 ml. The data show that in acuterenal failure in man (a) blood levels of PTH are elevated,(b) skeletal response to endogenous or exogenous PTH isimpaired, and this abnormality is a major factor underlying

54 a

the hypocalcemia in these patients, and (c) normal skeletalresponsiveness to PTH returns after recovery of renal func-tion. (Supported by USPHScontract 43-68-1040.)

199. Jejunal Sodium and Water Transport in Uremic Man.EUGENES. MAY,* H. JOHN REINECK,* HAGOP S. MEHK-JIAN,* JAY H. STEIN,* AND THOMASF. FERRIS,* Columbus,Ohio (introduced by James V. Warren).

It has been suggested that the increased fractional sodiumexcretion by the kidney in uremia is due to the presence ofa humoral inhibitor of sodium transport. To investigatewhether other epithelial membranes are also altered, we haveevaluated sodium (Na+) and water (H20) transport in thejejunum of uremic man before and after hemodialysis. Fivepatients with severe renal insufficiency maintained on chronichemodialysis and ten normal controls were studied before andimmediately after dialysis. In uremic patients net secretionsof Na' (JNNa+) -400±+114 SEM ,tEq/min per 30 cm andH20 (JNH2O) -2.72+1.01 ml/min per 30 cm were consistentlyfound, while absorption of Na and H20 occurred in all con-trols, +174.4±92 and +1.35±0.93 (P < 0.001). This defectin Na absorption in uremics disappeared after dialysis: JNNa+= +487±107 gEq/min per 30 cm and JNH2O = +3.79±0.91.Studies of unidirectional Na+ flux revealed inhibition ofmucosa to serosa transport before, 386+55, compared to1029±+151 (P <0.001) after dialysis, whereas serosa to mu-cosa transport before and after dialysis did not significantlydiffer, 621+85 and 531+49. The correction of the sodiumtransport defect occurred even when body weight was notaltered by dialysis. In summary, (a) net secretion of waterand electrolytes into the jejunum occurs in the uremic statein man; (b) the major alteration in transport is a decreasein mucosa to serosa flux of sodium; and (c) the alteredtransport of sodium in uremia is corrected by hemodialysis.(Research supported by grants from NIH.)

200. Serum Viscosity and Protein Changes in EarlyDiabetes. DONALD E. MCMILLAN,* Santa Barbara, Calif.(introduced by John W. Farquhar).

Serum viscosity is elevated in diabetes mellitus, especiallywhen microangiopathy is detectable. The major cause of theelevated viscosity has been found to be an increase in theresistance to flow produced by the serum proteins, a propertymeasured as the serum limiting viscosity number (LVN).Changes in serum viscosity produced by fluctuations in serumprotein and glucose levels do not affect the LVN of dialyzedserum, making it a more stable measure of viscosity changein diabetes. The LVN of dialyzed serum was observed to beless elevated in latent and early diabetes than in establisheddiabetes but elevation did not increase with duration ofdiabetes. This apparent paradox was resolved by studies ofserum after removal of lipoproteins by ultracentrifugation.Lipid-free serum LVN values increased progressively fromnormal to diabetic levels as glucose intolerance advanced.Subjects meeting USPHS criteria for diabetic glucosetolerance had LVN values as high as those in well estab-lished overt diabetes, and subjects with lesser degrees ofglucose intolerance had intermediate values. Electrophoresisof lipid-free serum demonstrated a reciprocal correlation be-tween LVN and the percentage of albumin in serum proteinsin both diabetic and nondiabetic subjects. Albumin levels felland a,- and a2-globulin levels rose as diabetes became evi-dent, indicating that change in serum protein pattern is im-portant in producing elevated LVN values in diabetes. Re-moval of glucose and lipoproteins from serum has producedclear evidence that serum protein composition is disturbedearly in diabetes; the disturbance may be detected before

overt diabetes is present, and it produces an increase inserum viscosity which is more profound in diabetics withclinically recognizable microangiopathy. (Supported by DorisPalmer Fund, Kroc Foundation, and NIH grant AM13092.)

201. Preferential Association of ps-Globin with Stroma inSickle-Cell Disease. GREGORYMEARS,* CLAYTON NATTA,*JOYCE V. O'DONNELL,* ANDARTHURBANK, New York.

Membrane-bound sickle hemoglobin (HbSS) has beenassociated with abnormalities of the red cell membrane anddecreased cell deformability in patients with sickle-cell (SS)disease. The present study was undertaken to assess the roleof individual globin chains in the pathogenesis of the ab-normalities of the membrane of SS cells. The results indi-cate that newly synthesized 68-chains are preferentiallybound to the stroma of reticulocytes of patients with SSdisease. Peripheral blood cells from patients with SS diseaseand with other hemolytic anemias (AA) were incubatedwith [3H] leucine for 1 h at 370C. The cells were then lysedhypotonically with 20 vol 1 mMEDTA at pH 7.2, and thestroma washed 5-10 times until the supernatant was clear.The AA stroma prepared using this method contained lessHb than that from SS cells. Relative a- and p-globin chainsynthesis was determined using column chromatographyafter preparing globin by dropwise addition of either totalstroma or stroma-free hemolysate directly into acid-acetone.Equal amounts of newly synthesized a- and #'-chains wereassociated with both stroma and present in stroma-free hemo-lysates from AA cells. By contrast, twice as much radio-active p'- as a-chain was associated with stroma in patientswith SS disease, while equal amounts were present instroma-free hemolysates. The selective association of newlysynthesized p8-chains with stroma in reticulocytes in SSdisease could result from either precipitation of p8-subunitsin the stroma, precipitation of HbSS with rapid proteolysisof a-chains, or specific membrane binding properties of pS_subunits. The association of ,S'-chains with stroma may leadto preferential destruction of these chains as well as tomembrane abnormalities which contribute to the shortenedhalf-life of SS cells. (Supported by grants from NIH, NSF,and ACS.)

202. Crystalglobulinemia: Clinical, Morphological, andPhysicochemical Studies. IRA MELNICOFF,* HENRICKBOGAARS,* ALBERT KALDERON,* CHAN PARK,* FRANCISCUMMINGS,* STEPHENR. KAPLAN,* ISRAEL DIAMOND,* ANDPAUL CALABRESI, Providence, R. I.

A diagnosis of multiple myeloma was established in a 52-yr-old male who presented with generalized weakness, weightloss, and large, lamellar eschars involving the extensor sur-faces of the upper and lower extremities. Biopsy of the ulcersrevealed nonspecific acute and subacute vasculitis. Serumprotein electrophoresis revealed an M-component comprising65% of the total protein (12.4 g/100 ml). By immunoelectro-phoresis, the myeloma protein was found to be IgGi, withkappa light chains. The serum as well as the purified IgG1showed cryoprecipitation with spontaneous crystallization be-ginning at 31°C and cryogel formation at 4°C. This processwas completely and repeatedly reversed at 37°C and foundto be concentration dependent; 1: 7 dilution of the serumabolished crystallization. No evidence of participation ofother immunoglobulins in the crystal formation was detectedby immunofluorescence. Light microscopic examination re-vealed the cryoprecipitate to be entirely crystalline, thecrystals being fusiform, birefringent, and dichroic, 0.5-0.8mm in length. Electron microscopy of the IgG1 crystalsshowed a unique organized structure which was readily

55 a

recognizable in bone marrow plasma cell organelles whenprocessed at low temperature. Immunofluorescent studies ofclinically uninvolved skin and muscle showed no evidenceof immunoglobulin deposition in vessel walls although therewas specific interstitial staining for IgG in association withan active myositis. The patient underwent plasmapheresisusing continuous cell flow centrifugation. Serial cryocritdeterminations showed an initial sharp decline from 31%to 24% followed by a rise and subsequently reached a plateauat 29%o. Combination chemotherapy was then instituted usingvincristine, Alkeran, prednisone, and procarbazine. Evidenceindicates that the spontaneous crystallization of this proteinoccurs on the basis of serum saturation. Spontaneouslycrystallizing immunoglobulins are rare although not unique.The readily recognizable structure of this protein provides anopportunity to study its intracellular synthesis and transportas well as its ultrastructural subunit composition.

203. Enhanced Effects of Prostaglandin El (PGE1) uponHuman Lymphocytes in the Presence of Cortisol. JOHNMENDELSOHN,* MICHAEL MULTER,* AND ROBERT BOONE,*San Diego, Calif. (introduced by Mehran Goulian).

The inhibitory effects of cortisol and agents acting throughcyclic AMPhave been studied in cultures of purified humanperipheral lymphocytes. In both phytohemagglutinin (PHA)-stimulated and unstimulated cultures, [3H] thymidine in-corporation is reduced when cells are incubated with cortisol,PGE1, or dibutyryl cyclic AMP, and PHA-induced morpho-logic transformation is prevented. When cortisol and PGE,(or dibutyryl cyclic AMP) are added simultaneously tocultures, additive inhibitory effects are observed: ['H]thymidine incorporation is reduced > 90% in unstimulatedcultures and > 99% in PHA-stimulated cells. Peripherallymphocytes contain 2.2+0.2 pmoles cyclic AMP per 106cells (mean±FSEM). Within 20 min of adding 0.1 mMPGE1to cultures, the cyclic AMP levels rise to 7.2±1.1. Duringthe ensuing 1-6 h, the concentration falls to base line levels.Cortisol in concentrations as high as 0.1 mM does notaffect cyclic AMPlevels. However, when cells are incubatedwith PGE1 in the presence of cortisol there is a significantlygreater elevation of cyclic AMPlevels, to 13.4+2.2 pmoles/106 lymphocytes after 20 min, with a decline paralleling thatobserved with PGE1 alone. Stimulation of lymphocytes withPHA does not result in significant changes in cyclic AMPlevels, but the response to PGE1 parallels unstimulated cul-tures. The PHA-stimulated cells also show a further eleva-tion of cyclic AMPlevels when cortisol is added to cultureswith PGE1. The data provide evidence that cortisol enhancesthe response to PGE1 by human peripheral lymphocytes inboth unstimulated and PHA-stimulated cultures, and thatthis enhancement may be mediated by a synergistic inter-action with cyclic AMP. (Supported by NIH and DamonRunyon grants.)

204. Human Interferon Applied Locally Inhibits Respira-tory Virus Infection in Man. THOMASC. MERIGAN, SYLVIAE. REED,* THOMASS. HALL,* AND DAVID A. J. TYRRELL,*Salisbury and Harrow, England.

Our in vitro studies have demonstrated that several strainsof influenza and rhinoviruses are as sensitive to the anti-viral action of human interferon as a known interferon-sensitive virus, provided that a yield-reduction type of assayis employed. In addition, influenza B/Hannover/1/70 wasshown to be equally highly susceptible to interferon in theciliated epithelial cells present in human fetal tracheal organculture. Human respiratory tract infections would be anideal target for the clinical application of interferon, since

the antigenic diversity of viral pathogens in this area pre-cludes immunological control. Randomized controlled double-blind studies were conducted on normal volunteers to evalu-ate the activity of human leukocyte interferon against tworespiratory viruses. 16 applications of interferon or placebowere given by nasal spray during the 24 h before influenzaB/Hannover/1/70 challenge of 22 volunteers who were se-lected on the basis of low serum antibodies and hence sus-ceptibility to that virus. However, a single day of localinterferon prophylaxis with a total dosage of 800,000 U ap-peared only to delay slightly the onset of clinical evidenceof infection but failed to prevent or diminish the severity ofdisease by standard clinical scoring, virus shedding or sero-conversion. Next, 39 applications of interferon or placebowere given by nasal spray to influence rhinovirus 4 challengein each of 32 volunteers who had low serum antibodies tothat virus. By employing a greater daily interferon dosageand combining a day of prophylaxis with 3 days of treat-ment for a total dosage of 14,000,000 U, a statistically sig-nificant prevention of clinical symptoms and virus sheddingafter rhinovirus infection was achieved. The frequency andextent of sero-conversions in the treated volunteers were alsolower than in the controls.

205. A Successful Approach to the Inhibition of GrowthHormone Secretion in Man. THOMASJ. MERIMEE ANDS. EDWIN FINEBERG,* Boton, Mass.

This is to report the use of diet for suppressing growthhormone (HGH) secretion in man. This suppression is sig-nificantly greater than that achieved by pharmacologic means.HGHsecretion was assessed in eight normal, nonobese sub-jects before and after each of four separate dietary re-gimens. With a high carbohydrate (525 g), normal pro-tein (75 g) diet of 3600 calories (diet 1), the mean maximalHGHconcentration in plasma after arginine decreased from21.5+3.5 miag/ml to 4.2±0.9 mutg/ml (P < 0.01). The meansum of responses to arginine (30, 60, and 90 min) was re-duced from 54.0±10.0 mug/ml to 10.5+3.1 m;g/ml (P <0.01). Total integrated 24 h secretion of HGH, assessed byhourly blood samples, was reduced to 19%o of that noted inthe control period. Virtually identical suppression of HGHsecretion occurred when carbohydrate and protein were pro-portional to diet 1, but caloric intake was reduced to 2300calories (diet 2). In diets 3 and 4, caloric intake was 3600calories with protein reduced to 30 g. Diet 3 contained allessential amino acids; diet 4 lacked lysine and tryptophan.Mean maximal HGHresponses to arginine before and aftereach diet were, respectively, 26.0±5.1 mng/ml vs. 10.0+2.1m/g/ml and 25.3+6.4 mjug/ml vs. 9.1+3.4 mug/ml. HGHsecretion was also examined after (a) 8-adrenergic stimula-tion, (b) acute elevation of plasma free fatty acid concentra-tions, and (c) induction of hyperglycemia. Suppressionachieved by these and other means was sporadic and didnot exceed 50% of the control state. This data establishesfor the first time an effect of prior diet on secretion ofHGHand supports a new approach to altering its secretorypattern. (Supported by NIH RR-533.)

206. Serial Studies of T- and B-Lymphocytes in Patientswith Systemic Lupus Erythematosus (SLE). RONALDP.MESSNER,* FOLKE D. LINDSTR6M,* AND RALPH C. WIL-LIAMS, JR., Albuquerque, N. M.

Peripheral blood lymphocytes from 24 patients with activeSLE were serially studied with respect to shifts in relativeproportions of T- and B-lymphocytes in association withclinical disease activity. Peripheral blood B-cells were identi-fied by direct immunofluorescence using enumeration of cells

56 a

showing surface Ig. T-cells were enumerated using indirectimmunofluorescence with rabbit antisera prepared againsthuman fetal thymocytes or thoracic duct lymphocytes. T-cell antisera were absorbed with B-cells from patients withchronic lymphatic leukemia to render them T-cell specific.T-cell specificity was indicated by 99% staining of humanthoracic duct cells, excellent correlation between lymphocytestaining and cold rosette formation (E-binding) of sheeperythrocytes, and only 18% staining of normal human mar-row lymphocytes. Lymphocyte studies also included deter-minations of E-binding by sheep erythrocyte rosettes, enu-meration of percentages of B-lymphocytes showing the C3receptor, and screening of plasma samples for lymphocyto-toxic antibody. Patients with relatively stable SLE showedno prominent serial shifts in peripheral blood B- or T-cellpopulations. However, dramatic changes in peripheral T-and B-cells were recorded in many patients with activeSLE. Most consistent was a rise in proportions of peripheralblood B-cells concomitant with exacerbations of SLE activ-ity. Sudden or gradual falls in proportions of peripheralblood T-cells were often noted in association with a risein B-cells from normal percentages of 22.9+7.1 to 50-70%.No correlations were evident between these changes andcirculating lymphocytotoxins. In active SLE, a marked dis-crepancy was recorded between a relatively normal valuefor percent T-cells by immunofluorescence (75.3±14.0%) andextremely low values for T-cells using E-binding (10-20%).Percentages of B-cells recorded by sums of peripheral bloodlymphocytes showing surface Ig was often 5 or 6 timeshigher than that recorded for B-cells showing identifiableC3 receptors.

207. Phytohemagglutinin Mitogenic Proteins: StructuralEvidence for a Family of Isomitogenic Proteins. BRUCEMILLER,* CLAUDIA NOYEs,* ROBERTHEINRIKSON,* HENRYS.KINGDOM,* AND STANLEY YACHNIN, Chicago, Ill.

Phytohemagglutinin (PHAP) mitogenic proteins consistof four subunits with mol wt of 34,000 as shown by SDS-acrylamide-gel electrophoresis. Wehave postulated (Yachninand Svenson. 1972. Immunology. 22: 871) that L-PHAP(low-hemagglutinating PHAP) is composed of four L sub-units, while H-PHAP (hemagglutinating PHAP) consistsof a series of molecules composed of 3L + 1R, 2L + 2R,IL + 3R, and 4R subunits. The amino acid sequence of bothL- and H-PHAP were determined by the Edman degrada-tion procedure. L-PHAP revealed the following sequence:NH2- Ser -Asn-Asp-Ile-Tyr- Phe-Asn- Phe-Gln-Arg-Phe- ?-Glu-Thr-Asn-Leu-Ile-Leu-Gln-Arg-Asp-Ala-Ser-Val- - -.Assuming a four subunit structure, recoveries ranged from50 to 80% of the theoretical yield. Similar analysis of H-PHAP revealed a major sequence, as well as a minor se-quence (the latter persisting only for the first seven aminoacids, major: minor molar ratio = 3.7: 1, total recovery 76%oof theoretical yield assuming a four subunit structure). Thesequences are: (major) NH2 - Ala - Ser - Gln - Thr - Ser -Phe - Ser. - Phe - Gln - - etc. (minor) NH2 - Ser - Asn -Asp - Ile - Tyr - Phe - Asn -. Dissociation of H- and L-PHAPby guanidine HCl, followed by isoelectric focusing in8 Murea, allowed isolation of the L (pI 5.3) and R (pI 6.0)subunits. Analysis of these materials confirmed that the L-PHAP sequence (and the minor sequence of H-PHAP) isthat of the L subunit, while the major H-PHAP sequence isthat of the R subunit. The latter, from the 8th to the 24thpositions, is virtually identical with the L subunit. These dataverify our isomitogen hypothesis and provide a structuralbasis for explaining the different biological properties of L-and H-PHAP.

208. Response of Plasma Dopamine Beta Hydroxylase(DBH) and Plasma Renin Activity (PRA) to Changesin Intravascular Volume (IVV). WALTER L. MILLER,*FRIEDHELM LAMPRECHT,* PHILIPPE V. CARDON,* ANDFREDERIC C. BARTTER,** Bethesda, Md.

Plasma DBH is principally derived f rom sympatheticnerves, where by hydroxylation of dopamine it producesnorepinephrine. Plasma renin is principally derived fromjuxtaglomerular cells. By its action on renin substrate itproduces angiotensin. Norepinephrine and angiotensin arethrought to be important in rapid adjustment to acutechanges in IVV. DBH and PRA were determined at 4-hintervals in healthy volunteers. DBH showed no circadianvariation, but PRA showed a significant rhythm with highvalues at 0400-0800 and low values at 2000-2400. Infusionsof 1000 cc 7.5% human serum albumin over 2 h loweredDBH by 30% (P < 0.01), while PRA decreased by 50%(P < 0.01). After a nadir at 3.5 h PRA returned to arhythmic pattern with a 12 h phase shift. Phlebotomy (5%IVV) increased PRA by 35% but did not alter DBH. Incats infused with 10%o dextran to increase IVV up to 30%o,DBHdecreased by 80%o (P < 0.001). Phlebotomy of up to30% IVV increased DBH by 120%o (P < 0.01). WhereasIVV expansion reduces the activity of the peripheral sym-pathetic nervous system and the renin-angiotensin system,volume depletion increases the activity of both systems. Thusboth systems operate to reduce the impact of acute changesin IVV.

209. Prostaglandin Binding to Beef Thyroid Membranes.WAYNEV. MOORE,* AND J. WOLFF,** Bethesda, Md.

The binding of tritiated prostaglandin E1 (['H]PGE,) tobeef thyroid membranes was investigated by a micro-discon-tinuous gradient centrifugation method. By studying theeffects of selected cations, pH, and temperature on the bind-ing of ['H] PGE1 to the membranes, two populations ofbinding sites were found. Binding of ['H] PGE1, occurringat pH 7.0, requires calcium, is heat labile, is inhibited byprostaglandin analogues and antagonists, and is inhibitedby ITP, GTP, and TSH but not by fatty acids. Half-maximal binding occurs with 1 mMcalcium. The bindingat pH 7.0 with 5 mMcalcium has an affinity constant of2.6 X 108 liters/mole. The other population of sites has apH optimum of 4.5, is heat stable, and has a partial calciumrequirement. The pK for binding at this pH is 5.1± 0.2,which is similar to that of long-chain fatty acids. Scatchardplot analysis of the binding at pH 5.2 with 5 mMcalciumrevealed two components of binding with affinity constantsof 1.5 X 10"° liters/mole and 1.6 X 108 liters/mole. Prosta-glandin analogues and antagonists are relatively less effectivein promoting debinding of ['H] PGE1 at pH 5.2. ITP, GTP,and TSH are ineffective in inhibiting binding at pH 5.2.The binding of ['H] PGE1 at pH 5.2 is inhibited by C18-24fatty acids with the degree of inhibition increasing with chainlength and degree of unsaturation. The results indicate thatthe binding of ['H] PGE1 is related to the ionic and lipidproperties of the prostaglandin at pH 5.2 but has more spe-cific structural requirements at pH 7.0.

210. On the Metabolic Pathogenesis of the Renal Dis-order of Hereditary Fructose Intolerance (HFI). R.CURTIS MORRIS, JR., ELISABETH MCSHERRY,* AND ANTHONYSEBASTIAN,* San Francisco, Calif.

In patients with HFI, fructose induces a dose-dependentdysfunction of the proximal renal tubule like that of Fanconisyndrome, hypophosphatemia, and hyperuricemia. The meta-bolic abnormality induced depends on accumulation of fruc-

57 a

tose-l-phosphate (F-1-P) in tissues genetically deficient inF-1-P aldolase activity, including renal cortex. For eachmole of fructose-- F-1-P, one mole of ATP-- ADP. Wepropose that in HFI, accumulation of F-i-P in the proximaltubule leads to impaired tubular function by acting both todeplete preformed ATP and to restrict ATP generation bydiminishing cellular concentrations of inorganic phosphate([P.]) and total adenine nucleotides. Substantial reductionsin [Pi] and [ATP] are known to remove the normal inhibi-tion of enzymes that catalyze the essentially irreversibledegradation of adenosine monophosphate to inosine mono-phosphate and adenosine, precursors of uric acid via inosine.Accordingly, our proposal predicts that prevention of cellulardepletion of [Pi] would attenuate the experimental renaldysfunction and the hyperuricemia. When fructose was ad-ministered to rats, (20, 40 mM/kg), simultaneously admin-istered sodium phosphate in amounts sustaining hyperphos-phatemia, significantly attenuated an otherwise strikingreduction in renal cortical [P.], [ATP], and total adeninenucleotides, despite a significantly higher concentration ofF-1-P. When the experimental renal dysfunction of HFIwas induced during water diuresis, experimental hyperphos-phatemia prevented an otherwise invariable increase inurine flow (an index of proximal tubular solute rejection)and significantly attenuated the acidification defect, amino-aciduria and the hyperuricemia, despite significantly lesseneduricosuria. These results provide strong evidence that inHFI, the major pathogenetic consequence of F-1-P accumu-lation is not inhibition of glycolytic enzymes but sequestra-tion of phosphate and greatly diminished tissue [Pi] thatleads to restricted ATP generation.

211. Location of Lipid-Binding Sites in Human PlasmaLipoproteins: Synthesis and Properties of Peptide Frag-ments of Apolipoprotein-Alanine (apoLP-Ala). J. D.MORRISETT,* J. T. SPARRow,* AND A. M. GOTTO, JR., Hous-ton, Tex.

ApoLP-Ala is an apolipoprotein which binds phosphatidylcholine (PC) and contains 79 amino acids of known se-quence. In order to locate the lipid-binding site(s) of thisprotein, we have undertaken its chemical synthesis by stan-dard solid-phase methods. We now report the synthesis offour fragments of this molecular corresponding to sequencepositions 41-79 (I), 48-79 (II), 55-79 (III), and 61-79(IV). The resulting peptides were cleaved from the resinwith liquid HF and purified by chromatography on G-50and DEAE-cellulose. Each purified peptide eluted as a sin-gle symmetrical peak, exhibited a single band on poly-acrylamide electrophoresis, and gave an amino acid analysisin excellent agreement with the theoretical value. Circulardichroism studies indicated that fragments I and II becamemore helical in the presence of PC, and also inhibited thereactivation of delipidated 8-hydroxybutyrate dehydrogenasewhich requires PC for activity. Electron microscopic studiesshowed that fragments I and II caused PC vesicles to becomealigned into linear arrays of stacked discs. PC complexes ofI and II could be isolated by density gradient centrifugation.Fragments III and IV exhibited none of these properties.These results indicate that residues 55-79 do not contain theminimum determinants required for the binding of PC. How-ever, extension of this peptide's N-terminus by seven resi-dues produces a molecule which binds PC. (Supported inpart by HEWgrant HL 14194 and by a grant from TheJohn A. Hartford Foundation, Inc.)

212. Immunoregulatory Alpha-Globulin (IRA) Selec-tively Affects T-Lymphocytes. J. H. MORSE.* A. M. MOR-

RIS,* AND R. MASCITELLI,* New York (introduced by V. P.Butler, Jr.).

Previous investigators demonstrated an immunosuppres-sive factor in plasma which inhibits antibody synthesis andgraft rejection. This immunosuppressive factor (IRA), with-out inherent cytotoxicity, inhibited lymphocyte transforma-tion induced by specific antigens, phytohemagglutinin, allo-geneic lymphocytes, and antilymphocyte serum. IRA, a crudepreparation of a-2-globulin, was prepared by DEAE chro-matography from pooled plasma (27 individuals) and as-sayed for its ability to inhibit blood lymphocyte proliferationinduced in vitro by the mitogens: phytohemagglutinin(PHA), concanavalin A (ConA), pokeweed (PWM), andE. coli lipopolysaccharide (LPS). All mitogens were usedat concentrations producing maximum uptake of ['H] thy-midine. Control cultures included cells without mitogen andthe substitution of serum albumin for IRA. Stimulation ofcultures (1 X 108 cells/ml) by PHA or ConA was inhibitedby concentrations of IRA varying between 0.25 and 4 mg/mlof culture medium. PHA and ConA are predominantly T-cell mitogens; therefore the effect of IRA on the action ofB-cell mitogens was also studied. IRA failed to inhibitblastogenesis by PWMand LPS but did suppress stimula-tion by PHA or ConA in simultaneous cultures taken fromnormal individuals. Thus, IRA had no effect on lymphocytesundergoing proliferation by mitogens classified as B-cellstimulators, whereas T-cell mitogen proliferation was in-hibited. Cultures stimulated by PWMand LPS incorporatedabout half the label seen in PHA- and ConA-stimulated cul-tures, consistent with estimates of 60% T and 30% B-lym-phocytes in peripheral blood. The specificity of inhibition of-fers a new tool for evaluating T- and B-cell function innormal and disease states and offers a possible mechanismfor the inhibition of PHA-stimulated cultures by plasmafrom patients with cancer and chronic renal disease. (Sup-ported by NIH and New York Arthritis and LE Founda-tions.)

213. Association of Adenovirus and E.Coli Infections withAcute Hemorrhagic Cystitis (AHC) in Children. MAUR-ICE A. MUFSON,* ROBERT B. BELSHE,* TERRENCEJ. HoRRi-GAN,* AND LOWELLM. ZOLLAR,* Chicago, Ill. (introduced byMark H. Lepper **).

AHC, characterized by gross hematuria and symptoms ofbladder inflammation, occurs in infants and children appar-ently as a self-limiting disease which must be differentiatedfrom serious renal disorders. Until recently no etiologic agentwas associated with AHC. Studies from Japan and from ourlaboratory have implicated adenovirus 11 in the etiology ofthis disease. From October 1968 through September 1972at the Cook County Hospital, 69 infants and children withAHCand 42 children without urinary tract disease (control)were tested for viruria and bacteriuria. Adenovirus 11 wasrecovered from the urine of 10 patients with AHC andadenovirus 21 from the urine of 2 patients with AHC.Children shed virus 5-7 days after the onset of illness. Twocontrol children had adenovirus 11 viruria and may representasymptomatic infection. Adenovirus 11 neutralizing antibodyrises were detected in 4 of 5 AHC patients with adenovirus11 viruria, 1 of 4 children with bacteriuria, and 1 of 14children fr9m whom no agent was recovered. None of 11control children tested had such rises. 12 patients with AHChad E. coli bacteriuria. Bacteria were not isolated from theurine of 18 control patients tested. In the adenoviruric group,males predominated, but the bacteriuric group consisted ofnearly all females. Hematuria was present in all and fre-quency and dysuria in nearly all cases. The average dura-

58 a

tion of AHC illness was 4.9 days. Overall, adenovirus or E.coli infections were detected in 39.1% of cases. These dataprovide evidence that AHChas a multiple etiology and thatadenovirus and E. coli infections are the etiologic agents ofsignificant segments of AHCillnesses.

214. Link of Porphyrias to Metabolism of Hemopexin(Hx), the Most Avid Serum Binder of Porphyrins. UR-SULA MfLLER-EBERHARD, La Jolla, Calif.

Two facts are established: heme is taken up by hepatocytesas the heme-Hx complex (1970. J. Lab. Clin. Med. 76: 426,and Hershko et al. 1972. J. Lab. Clin. Med. 80: 624) and theKd of Hx for heme and non metal-containing porphyrinsare ca. 10' and 10' M, respectively (1972. J. Biol. Chem.247: 7181). Whether interaction of nonmetal porphyrins withHx has pathophysiological significance has been investigated.Sera (samples) of patients analyzed were: 26 (33) withacute intermittent porphyria (AIP), 23 (34) with erythro-poietic protoporphyria (EPP), and 64 (67) with porphyriacutanea tarda (PCT). Hx concentrations in AIP remainednormal in 19 of 23 patients, in EPP and PCT approximatelyhalf were decreased. Therefore, porphyrin contents were re-lated to Hx levels in EPP and PCT. 14 patients (65 sam-ples) with active EPP showed an inverse -correlation be-tween levels of Hx and protoporphyrin (supplied by Dr. M.M. Mathews-Roth) in erythrocytes (correlation coefficient,r = 0.915), plasma (r = 0.490), and daily stool excretions(r = 0.518); significance level, P < 0.001. 20 patients (20samples) with PCTshowed r = 0.625, P < 0.01, for Hx levelsand urinary uroporphyrin (supplied by Dr. J. H. Epstein),but for coproporphyrin r = 0.232, P > 0.05. Moreover, for 30healthy individuals, concentrations of Hx and albumin werefound to be 69.3±2.0 mg/100 ml and 4.27±0.14 g/100 ml,respectively. Levels of Hx, 52.9±+1.8 mg/100 ml, werelowered simultaneously with those of albumin, 3.49±0.085 g/100 ml, in PCT (P < 0.001). In EPP, however, albuminlevels, 4.20±0.13 g/100 ml, were unchanged, whereas thoseof Hx, 53.3±3.1 mg/100 ml, were lowered, P < 0.001, whichwas not due to hemolysis, as haptoglobin levels were nor-mal. These data suggest an interrelationship between themetabolism of Hx and that of nonmetal porphyrins in PCT,very probably in EPP, but not in AIP. (Research supportedby grant HL-08660 from NIH.)

215. Immunoglobulin Light Chains in Diabetes Mellitus.FRANKLIN MULLINAX,* CHARLES 0. WATLINGTON,* ANDBETTY HIMROD,* Richmond, Va. (introduced by David W.Richardson**).

Pancreatic amyloid is found frequently in autopsied dia-betics and is apparently responsible for pancreatic hyaliniza-tion. Because of this relationship, the light chain content ofcertain amyloids, and the cryptic nature of diabetic vasculardeposits, we have studied the light chains in serum and urineof diabetes free of clinical renal disease and in control sub-jects. Serum light chain levels, measured by radioimmuno-assay, were higher (P < 0.05) in 32 diabetics (3.33±0.51SEMAig/ml) than in matched controls (2.08±0.27 /Ag/ml).The 12 pairs over age 39 were largely responsible for thisdifference (5.48±0.87 and 2.69±0.46 ,ug/ml, P < 0.01). Uri-nary excretion of light chains, estimated by radial immuno-diffusion, was also higher (P < 0.05) in 15 diabetics (9.20±177 mg/g creatinine) than in controls (4.84±0.72 mg/gcreatinine). The paired diabetics and controls over age 39were responsible for this difference (13.34±2.49 and 4.73±1.09 mg/g creatinine, P < 0.01). Serum concentration andurinary excretion of another low molecular weight protein,lysozyme, estimated by radial diffusion, were similar in the

diabetics and controls. Diabectics with clinically normal renalfunction over the age of 40 have elevated levels of serumand urine light chains apparently due to over-production offree light chains. The finding of elevated levels of serumlight chain provides a new approach to the etiology andcomplications of diabetes mellitus. (Supported by a grantfrom the Arthritis Foundation and NIH-AM 16564.)

216. Autonomy of Cortisol Secretion in the HumanFetus. BEVERLEY E. PEARSON MURPHY, SARAH JANECLARK,* ROBERT DIEz D'AUx,* IAN Ross,* MICHAELPINSKY,* AND DOREENVEDAHY,* Montreal, Canada.

In sheep, where the placenta appears to present a sig-nificant barrier to the maternal to fetal movement of cortisol,the fetal cortisol level rises a few days before parturitionand has been implicated as a factor in the initiation of labor.In human pregnancy it has been generally accepted thatcortisol as such crosses the placenta freely and that themechanism triggering parturition must therefore be dif-ferent. We have studied the levels of bound and unboundcortisol in the fetal circulation and tissues, and also the fateof tritiated cortisol injected into the mother and found thefollowing. High concentrations of cortisol [1540±1300 ng/g(SD)] were present in the fetal adrenal from 8 to at least18 wk of gestation. In early pregnancy (8-18 wk), thecortisol level of blood obtained from the fetal end of theumbilical cord (essentially arterial) exceeded that from theplacental end (P < 0.01). Mixed cord serum cortisol levelsrose from 7.0±1.2 ng/ml (SE) in the first trimester to23.3+3.4 ng/ml in the last trimester (elective Caesareansection) and were unchanged by induced labor (24.0±4.5ng/ml). Spontaneous labor was associated with a furtherrise to 72.6±+17.3 ng/ml at normal vaginal delivery and87.1+20.0 ng/ml when Caesarean section was done after theonset of labor. The maternal-fetal gradient of unboundcortisol (10 a.m. to 3 p.m.) was estimated to be about 8: 1in early pregnancy and about 3: 1 in late pregnancy. Tri-tiated cortisol injected into the mother was largely (about85%) converted to cortisone in the placenta. These findingssuggest (a) that the human fetal adrenal functions auto-nomously from early gestation, (b) that conversion of ma-ternal cortisol to cortisone serves as a mechanism to pre-vent suppression of the fetal adrenal, and (c) that, althoughthe nature of the placental "barrier" differs between the twospecies, the maternal-fetal cortisol relationships of man andsheep may yet be fundamentally analogous. (Supported bythe Medical Research Council of Canada.)

217. Thymic Function in Murine Lymphoid Leukemia.HIROSHI NAGAYA,* Durham, N. C. (introduced by Herbert0. Sieker).

Although 90% of AKRmice die of spontaneous lymphoidleukemia before the age of 12 months, the mice do not developleukemia until the age of 6 months, suggesting that a latentperiod is necessary for leukemia development. Moreover, theincidence of leukemia can be reduced to nil if thymectomizedbefore the age of 6 months. In order to study thymic influ-ence on leukemogenesis, 1- or 6-month-old AKR mice werethymectomized and received syngeneic 6- or 1-month-oldthymus grafts. When 1-month-old thymectomized mice weregrafted with thymus glands from 6-month-old mice, 13 of21 mice died of leukemia before the recipient mice reachedthe age of 6 months, while only 3 of 17 mice thymectomizedat the age of 1 month and grafted with 6-month-old spleensdied of leukemia. In contrast, none of 19 mice thymectomizedat the age of 6 months and grafted with thymus glands from1-month-old mice developed leukemia for at least 5 months

59 a

until the age of grafted thymus reached 6 months. Normalthymus cells can be stimulated by concanavalin A (ConA)to increase [3H] thymidine uptake 455-fold (median), whileleukemic thymus cells can be stimulated only 3-fold, thelatter response resembling the ConA response of normal bonemarrow cells (2-fold). Preleukemic thymus cells obtainedfrom 5-month-old mice were stimulated 269-fold. These re-sults suggest that 1-month-old bone marrow precursor cellscan become leukemic prematurely under the influence of6-month-old thymus grafts because of failure in the thymicfunction to convert ConA-unresponsive bone marrow pre-cursor cells to ConA-responsive normal thymus cells. (Sup-ported by grant from ACS.)

218. Monozygotic Twin Kinships: a New Model for theAnalysis of Quantitative Inheritance in Man. WALTERE.NANCE, PAO-Lo Yu,* PATRICIA I. BADER,* GLENN J.BINGLE,* AND PHYLLIS WINTER,* Indianapolis, Ind.

The genetic relationships within the families of mono-zygotic twins permit a more precise analysis of quantitativeinheritance than has hitherto been possible in man. The off-spring of identical twins are genetically related to each otherin the same way as half-siblings, and comparison of mono-zygotic twin, full-sib, and half-sib correlations in these kin-ships permits simultaneous estimation of the effects ofseveral degrees of genetic relatedness in the same body ofdata. In contrast to conventional half-sibs, monozygotic twinhalf-sibships have the same expected size and average age,and all the parents are usually available for study. Since thehalf-sibs live in different homes, the effects of genetic cor-relation can be largely separated from those of a common en-vironment. The model permits estimation of the additive anddominance components of the total genetic variance as wellas certain epistatic interactions, from observations on mem-bers of the same generation. Maternal effects can be de-tected by comparing the correlations of maternal half-sib-ships with those of paternal half-sibships, and the influenceof assortative mating on husband-wife correlations can evenbe distinguished from those of common environmental fac-tors by partitioning the covariance into within and betweenpair components. The model has been applied to theanalysis of multiple anthropometric, dermatoglyphic, and bio-chemical variables including weight, height, blood pressure,total ridge count, cholesterol, and uric acid in 18 monozygotictwin kinships, which include a total of 95 offspring rangingin age from 6 wk to 44 yr. (This work was supported byPHS RR-00750 and a grant from The John A. HartfordFoundation.)

219. Decreased p-Messenger RNAActivity in ErythroidCells in Heterozygous p-Thalassemia. CLAYTON NATTA,*JUDY BANKS,* GULZAR NIAZI,* PAUL A. MARKS,** ANDARTHURBANK, New York.

The decreased p-chain synthesis characteristic of erythroidcells of patients with homozygous p-thalassemia is also ob-served when mRNAisolated from these cells is added to acell-free system. The present studies demonstrate thatmRNAisolated from erythroid cells in peripheral blood andbone marrow of heterozygotes, as well as homozygotes forg-thalassemia, directs decreased amounts of ,-chain synthesiscompared to that of a-chains. RNA's were extracted fromtotal cells using phenol and SDS, and isolated by sucrosedensity gradient centrifugation. The 6-15S (mRNA) frac-tions were assayed in a Krebs ascites tumor lysate cell-freesystem. The ratio of a-chain synthesis to that of p-chains(a/,8 ratio) with mRNAfrom bone marrow cells of fournonthalassemic patients was 0.96, 0.95, 1.18, and 1.2, similar

to that of peripheral blood cells. mRNAisolated fromperipheral blood and bone marrow of a patient with homo-zygous 8-thalassemia who had no demonstrable ,-chainsynthesis in whole cells directed no synthesis of p-chainswhen added to a cell-free system. A patient with clinicalthalassemia intermedia had an a/,8 ratio of 1.9 in whole bonemarrow cells, and 1.45 using isolated mRNA. A patient withhigh A2 p-thalassemia trait had an a/p ratio of 1.0 inwhole bone marrow cells, while mRNAisolated from peri-pheral blood and bone marrow led to a/,p ratios of 1.6and 1.75, respectively. These studies indicate that there isdecreased p-mRNA activity in heterozygous as well as inhomozygous 8-thalassemia. Balanced globin chain synthesisin whole bone marrow cells of p-thalassemia heterozygotesmay reflect translational control mechanisms which lead tolimited a-chain synthesis or increased 8-chain synthesis inthese cells. (Supported by grants from NIH, NSF, andACS.)

220. Lack of Insulin Effect on Human Lymphocytes inCulture and During In Vitro Transformation. W. MAR-cUS NEWBERRY*AND MARIA G. BUSE,* Charleston, S. C.(introduced by Joseph C. Ross).

Specific insulin binding sites have been reported to appearon the lymphocyte surface during transformation, suggestinga possible physiological role of the hormone. Human bloodwas layered on a mixture of Hypaque-Ficoll and the lympho-cytes isolated by centrifugation. Lymphocytes were culturedfor 24, 48, or 72 h in Eagle's minimum essential medium con-taining fetal calf serum with or without added phytohemag-glutinin (PHA). The cells were transferred into Gey andGey's balanced salt solution containing 5.5 mMglucose, withor without ["C] leucine or ['H] thymidine and incubatedfor 1-4 h. Insulin (10 mU/ml) was added to alternate vials.Resting lymphocytes (2 X 106) cultured for 24 h incorporated0.3 nmoles ["C] leucine into protein in 2 h PHA-stimulatedcells 1.3 nmoles. Simultaneously, stimulated cells consumed6 times more glucose than controls. DNAsynthesis increasedonly after 48 h exposure to PHA, when stimulated lympho-cytes incorporated approximately 50 times more ['H]thymidine into trichloroacetic acid-precipitable material thancontrols. The addition of insulin during the 1-4 h incuba-tion to resting or to PHA-stimulated lymphocytes did notaffect either of the parameters measured; neither did theinclusion of insulin with PHA during 48 h transformationstimulate the utilization of glucose, ["C] leucine incorpora-tion into proteins, or DNA synthesis. By these criteria, in-sulin does not seem to regulate the metabolism of the un-stimulated or maximally transformed peripheral blood lymph-ocyte. (Research supported by grants from NIH [AM-02001] and SCTRDA.)

221. Lamellar Body of the Large Alveolar Cell-a Mis-nomer? ALBERT H. NIDEN,* Philadelphia, Pa. (introducedby Arthur B. DuBois **).

The terminal airways and alveoli of the lung are linedwith pulmonary surfactant (PS), predominantly saturatedphospholipids, capable of markedly reducing surface-activeforces on compression. Convincing indirect evidence indi-cates that the large alveolar cell (LAC) is probably the pri-mary source of PS and that its inclusion bodies (lamellarbodies, LB) are secretory granules containing PS. Althoughthere is some variation in the ultrastructural appearance ofLB contents, a characteristic finding is concentric osmiophiliclamellae separated by clear spaces. This appearance indicatesa partial extraction from and/or condensation of materialwithin the LB. In an attempt to reduce this known extrac-

60 a

tion of saturated phospholipids during preparation of tissuefor electron microscopy (EM),.heavy metals known to re-duce loss of lipids were combined with osmium (Os) andformaldehyde-glutaraldehyde (F-G) in varying proportions.A mixture of calcium, lead, Os, and F-G as a single primaryfixative followed by rapid acetone dehydration and embeddingin epon, resulted in good EMmouse lung preservation witha striking change in appearance of LB. Dense osmiophiliclamellae were no longer present. Most LB contained a uni-form fine granular material throughout with a dense osmio-philic border. The ultrastructural appearance was of anorganelle better preserved than that following standard fixa-tion. Lipids were extracted (Folch) and phosphorous phos-pholipid determined (Fiske) from the fixative, dehydrant,and fixed tissue for both the standard and Ca-Pb-Os-F-Glung preparation. 57% lung phospholipid had been extractedby the former and 43% by the latter procedure. It is con-cluded that the lamellae of the LAC are an artifact of tissuepreparation for EMdue in part to extraction of phospholipidand in part to poor fixation of remaining tissue phospholipid.(This work supported by USPHSand Council for TobaccoResearch.)

222. Plasma Fibrinopeptide A Concentration As an Indexof Intravascular Coagulaton. H. L. NosSEI, R. E. CAN-FIELD, ANDV. P. BUTLER, JR., New York.

Since thrombin cleaves fibrinopeptides A and B from theNH2-terminal region of the fibrinogen molecule, measure-ment of fibrinopepetide levels in plasma provides a direct in-dex of thrombin action. A sensitive radioimmunoassay (1971.Proc. Nat. Acad. Sci. 68: 2350) capable of detecting 0.2ng/ml fibrinopeptide A in plasma has been employed in clin-ical studies. Since fibrinogen cross-reacts with antifibrinopep-tide antibodies, an extraction method was used which wascapable of detecting more than 80%o of peptide added toplasma. Thrombin treatment of plasma cleaves two mole-cules of fibrinopeptide A per fibrinogen molecule. Strepto-kinase treatment of plasma results in cleavage of a largersegment of the NH2-terminal portion of the Aa-chain offibrinogen including the A peptide. The A peptide and theplasmin-produced fragment can be distinguished by gel filtra-tion and immunochemically. Extracts of plasma samplesfrom 20 healthy men showed immunoreactivity equivalent to0.1-0.8 ng/ml, some of which is thought to represent plasmincleaved NH--terminal portion of the Aa-chain. Plasma frompatients with intravascular coagulation contains peptide im-munoreactivity equivalent to 4-80 ng/ml. The immunoreac-tive material had the gel filtration and immunochemicalcharacteristics of the A peptide and not of the plasmincleaved fragment. Infusion studies indicated that the plasmatj for the A peptide is approximately 3 min. On the basis ofthis estimate, it is possible to calculate that thrombin pro-teolyses fibrinogen at the rate of 500 mg to 10 g/24 h inpatients with gross intravascular coagulation and at therate of less than 50 mg fibrinogen per 24 h in normal sub-jects. After heparin infusion in patients with intravascularcoagulation, the plasma fibrinopeptide A concentration fell toundetectable levels, implying that heparin therapy is associ-ated with decreased proteolysis of fibrinogen and that Apeptide levels may provide an index of the therapeutic ef-fectiveness of anticoagulant therapy. (Supported by grantsfrom the NIH and NYHA.)

223. Intestinal Fatty Acid-Binding Protein (FABP):Immunochemical Quantitation and Effect of Diet. ROBERTK. OCKNER* ANDJOAN A. MANNING,* San Francisco, Calif.(introduced by Rudi Schmid **).

Recently we identified a protein (FABP, mol wt 12,000)which binds long-chain fatty acids in cytosol of intestine andother tissues, and suggested that it might participate in cellu-lar fatty acid transport. Wenow provide evidence supportingthis concept, based on studies using quantitative radial im-munodiffusion with specific antiserum. FABP ('90% pureby disc-gel electrophoresis) was isolated by gel filtrationand isoelectric focusing. FABP shared immunochemicalidentity with a protein in fractions ('-'12,000 mol wt) ofliver and myocardium cytosol which bind fatty acids. Noprecipitin reaction was observed with serum, intestinallymph, or partially delipidated lymph lipoproteins. Con-sistent with its proposed role in fat absorption, FABPconcentrations in mucosa from proximal and middle thirdsof small intestine (14.9±1.7 and 15.9±0.8 /Ag/mg solubleprotein; 794±81 and 828+49 /g/g tissue) significantly ex-ceeded that in distal third (7.3+0.9 gg/mg protein; 348+12Ag/g tissue). In liver, FABP concentration in cytosol wasS 1.0 pg/mg protein, S 70 gg/g tissue, and was barely de-tectable in particulate fractions solubilized with deoxycholate.The effect of diet was studied in rats maintained for 3 wkon chow diets supplemented isocalorically with sucrose (lowfat), corn oil (high fat), or medium-chain triglyceride(MCT). Weight gains were virtually identical. Concentra-tions of FABP in all three intestinal segments were signifi-cantly greater after high fat diet (14.4±1.6, 21.0+2.0, 10.5±0.9 pg/mg protein) than after low fat (9.6±0.6, 14.7±1.1,6.8+0.8) (P < 0.025 for all segments). Results expressed asmicrograms per gram tissue were similar. MCTvalues wereintermediate. The preponderance of FABP in proximal in-testine and its response to dietary fat intake support theconcept that FABP participates in intestinal fatty acid trans-port. Moreover, its immunochemical identity with FABP inother tissues is consistent with a more general role for thisprotein in fatty acid metabolism. (Supported by NIH grantsAM-13328 and AM-14795.)

224. The Effect of Sexual Maturation on Testicular Sensi-tivity to Luteinizing Hormone (LH) Stimulation ofTestosterone (T) Secretion in the Intact Rat. W. D.ODELL, R. S. SWERDLOFF,* F. WOLLESEN,*ANDP. K. GROVER,*Torrance, Calif.

Wehave previously shown that 5 days posthypophysectomythe sexually immature male rat (but not the sexually ma-ture) becomes markedly insensitive to LH stimulation oftestosterone secretion. Sensitivity may be restored by pro-longed pretreatment with FSH; constant doses of LH pro-duce a greater response dependent upon the time of exposureto FSH. In essence, a time-response relation exists. Wehavepostulated thus that FSH induction of LH sensitivity is amajor factor in sexual maturation in the male rat. Wehavenow studied sensitivity of the intact rat to exogenous LHduring sexual maturation. Varying doses (0.3-40 ,g/100 gbody weight) of NIH-LH B7 were administered intraperi-toneally to male rats at 10, 22, 42, and 62 days of age.Serum testosterone was quantified by radioimmunoassay 1 hlater, the time when the LH-induced testosterone rise wasmaximal. For all doses of LH which increased serum testos-terone, increases were least in the 10-day-old animals, greaterin the 22-day-old animals, and greatest at 42 and 62 days; thelatter were indistinguishable. Given as per cent increase overbase line in response to 30 gg LH/100 g body weight, thevalues were 10 days, 248±13%; 21 days 630±42%; 42 days2295±225%; 62 days 1873± 125. These studies demonstratefor the first time that the sensitivity to LH stimulation oftestosterone increases with age in the intact male rat, andoffers additional support for the hypothesis that changing

61 a

testicular responsiveness to LH is a major factor in sexualmaturation. (Supported by NICHD grant ROI HDO6701.)

225. Zinc and Sickle-Cell Disease. F. J. OELSHLEGEL, JR.,*A. S. PRASAD, G. J. BREWER, D. OBERLEAS,* C. KNUTSEN,*ANDE. B. SCHOOMAKER,*Ann Arbor and Detroit, Mich.

We hypothesized that the hemolytic process in sickle-celldisease (SCD) could lead to a state of zinc deficiency be-cause of hyperzincuria. In 14 patients with SCD, red bloodcell (RBC) zinc levels were normal in three (13.3-16.5 lsg/ml RBC), low normal in two (10.5 and 10.4), and abnor-mally low in nine (5.8-9.8). Thus it appears that a fair pro-portion of SCD patients may be zinc deficient, which couldpartially explain clinical problems such as leg ulcers, growthproblems, and hypogonadism. Since zinc is present in a num-ber of enzymes and binds to various amino acids includinghistidine, we thought it important to investigate effects ofzinc on RBC's before suggesting zinc supplementation inSCD. Surprisingly, zinc chloride added to either normal orSCD blood in a final concentration of 1.5 X 10' M causeda left shift of the erythrocyte oxygen affinity curve varyingfrom 1.5 to 3.5 mmin eight SCD samples studied. Higherzinc concentrations led to precipitation of plasma protein(s)and aggregation of RBC's. When added to normal blood,zinc levels equilibrated between plasma and RBC's within3 h. When comparable amounts of 2,3-diphosphoglyceratewere added to normal and zinc-treated hemolysates, the P50difference between the two samples did not change, sug-gesting the two are not competing for the same site. Thesestudies suggest that zinc supplementation in SCD, in additionto its potential role in correcting wound healing and growthproblems, may be beneficial in the basic pathological proc-esses. Current evidence suggests that even small leftwardcurve shifts may be quite helpful. It remains to be seen ifphysiologically obtainable zinc levels are of therapeutic valuein SCD. In vivo studies are planned. (Supported by a con-tract from, NIH.)

226. Nuclear Receptor Sites for Triiodothyronine (T3) in RatLiver: Kinetics of Binding and Evidence for the Induction ofNew Building Sites in the Hyperthyroid State. JACK H.OPPENHEIMER,HAROLDL. SCHWARTZ,* DIONA H. KOERNER,*AND MARTIN I. SURKS, Bronx, N. Y.

In vivo displacement techniques with tracer and carrier T3have previously been used to demonstrate specific Ta-bindingsites in the nuclei of rat liver and kidney. Additional isotopicexperiments were undertaken to characterize the kinetic prop-erties of these sites. Serial measurements of the subcellulardistribution of T3 in rat liver over a 4 h period allowed thecalculation of extremely rapid rates of T3 interchange betweennucleus and cytoplasm. Displacement experiments with T3and T4 showed that the nuclear sites bind T3 at least 13 timesmore avidly than T4. The estimated ka of T3 was 4.6 X 10"1M-'. Chromatographic studies indicated that the small degreeof cross-reactivity between T3 and T4 for the nuclear sitescould not be attributed to T4 to T3 conversion in vivo. Satura-tion experiments with tracer and carrier T3 were also used todetermine a nuclear binding capacity. Sharp plateau valueswere achieved in each of six experiments over a 10- to 20-foldrange of injected T3: mean (ng/g liver), 0.99; range, 0.91-1.41.Moreover, the nuclear sites appear close to saturation atendogenous tissue levels of T,3 in euthyroid animals. Whenanimals were rendered hyperthyroid by the daily injection of25 ;Lg T3/100 g body weight for 6 days, a 5.7-fold increase inthe capacity of nuclear binding sites was observed. Even 2days after the cessation of T3 injections the binding capacitywas still 3.7 times normal. These findings suggest the possi-

bility that induction of new nuclear sites is necessary beforethe effects of excess hormone can be manifested at the tissuelevel. The high degree of specificity of the nuclear sites for T3in liver and other tissues also supports previous indicationsthat the hormonal activity of T4 derives largely, if not exclu-sively, from its conversion to T3. (Supported by NIH grantAM15421-13.)

227. DNAof Epstein-Barr Virus Detected in Tissue of Bur-kitt's Lymphoma and Nasopharyngeal Carcinoma. JOSEPHS.PAGANO, M. NONOYAMA,*ANDGEORGEKLEIN,* Chapel Hill,N. C., and Stockholm, Sweden.

The Epstein-Barr virus (EBV) is the sole virus consistentlyassociated with malignant tumors in man. EBV is found incultures of explanted Burkitt's lymphomas (BL) but never inthe tumor itself. Having developed methods to detect the viralDNA in nonvirus-producing cells, we decided to search di-rectly in BL tissue and also in nasopharyngeal carcinoma(NPC) and other tumors from the same region of Kenya forEBV DNA. The tumor DNAwas extracted, denatured, fixedon nitrocellulose filters, and hybridized with EBV-specificradioactive complementary RNA (cRNA) to give the averagenumber of EBVgenome equivalents per cell. Of the 23 speci-mens of BL, 22 clearly contained EBVDNAin amounts rang-ing between 4 and 113 equivalents with a mean of 38. EBVDNAwas less regularly found in NPC: 18 of 23 specimenscontained detectable viral DNAranging from 5 to 85 equiva-lents with a mean of 19. In contrast, only 4 of 23 other Africantumors contained EBV DNA (5-12 genome equiv.). Thetiters of antibodies to viral capsid antigen (VCA) and to anearly virus-specific antigen (EA-R) correlated with the num-ber of EBV genomes in BL. These results strengthen the as-sociation of EBV with BL, although the single negative BLretested with the more sensitive DNA-DNA reassociationkinetics again revealed no EBV DNA (< 0.5 genome). SinceEBV infects only lymphoid cells, at least in vitro, the EBVDNAin NPCand the other tumors may be due to infiltrationof EBV-carrying lymphoid cells from a local site such as thetonsil rather than to EBV in the carcinomatous cells. (Sup-ported by the NCI [SVCP].)

228. Treatment of Acromegaly by Radiofrequency Hypophy-sectomy. JOHANNAA. PALLOTTA,* NICHOLAST. ZERVAS,* ANDLouIs M. SHERWOOD,Boston, Mass. (introduced by Louis M.Sherwood).

20 patients with acromegaly were operated on using stereo-taxic thermal hypophysectomy for definitive treatment.Humangrowth hormone (HGH) was lowered into the normalrange (10 mig/ml) in all but three patients. The tumors werelarge and invasive, extending deeply and laterally into thesphenoid sinus. Studies included pre- and posthypophysectomyglucose tolerance tests with glucose, insulin, and HGH.Patients were restudied from 3 months to 6 yr after operation.Headache, hyperhydrosis, heat intolerance, soft tissue swelling,and joint stiffness began to subside within 1 wk. Preopera-tively minor impairment of vision was detected in seven pa-tients, varying from quandranopsia to minor defects on redfield testing. Postoperatively visual fields were normal in all.All but one patient developed adrenal insufficiency. Two didnot require thyroid replacement. Permanent complicationsfailed to occur. The average hospital stay was 7-10 days. Themean preoperative level of HGH2 h after glucose load was 76m;ug/ml (SEM was 23.1). The mean postoperative HGHin18 patients tested for at least 6 months after the procedurewas 8.6 miug/ml (SEM was 2.6). HGHwas 10 m;ug/ml or lessin 15 patients and ranged between 0 m;g and 6 m;ug/ml in 11patients. The average percentage drop in HGHwas 88%

62 a

(P < 0.01). Six patients had definite impairment of carbo-hydrate tolerance improving postoperatively with decreasedinsulin secretion. The main features of this procedure werea rapid drop in HGH, improvement of clinical symptoms, andhalting of acral growth. Remission was distinctly related tothe absolute drop of growth hormone to the normal range. Thismethod of treatment is useful in patients with enlarged sellaturcicas, with high HGHconcentrations, and with seriousmedical complications which demand definitive and earlyintervention. The therapy of acromegaly has been multifacetedand controversial. Radiofrequency hypophysectomy offersa low morbidity procedure in reducing the oversecretion ofHGHto normal and diminishing the clinical features of thedisease.

229. Effect of the Kidney on Exogenous Parathyroid Hor-mone. Localization of PTHin Proximal Tubule Cells. G. M. A.PALMIERI* AND R. NoRDQuIsT,* Oklahoma City, Okla. (in-troduced by Jack Metcoff).

Parathyroid hormone (PTH) has a direct effect on renaltubular function. The kidney is also probably the main siteof destruction of PTH. The effect of the kidney on exogenousPTH was studied in dogs receiving a constant infusion of par-tially purified bovine PTH (TCA-PTH) through a catheter ina jugular vein. Determinations of PTHby radioimmunoassay,using serum antibovine PTH that did not cross-react withcanine PTH, were performed in plasma samples obtained witha catheter placed in the inferior vena cava (IVC) above thelevel of the renal veins. PTHbecame detectable in IVC within30 min of the initiation of the TCA-PTH infusion. A plateauwas reached at 60 min and remained constant for several hours.No detectable PTH was observed in urine obtained from therenal pelvis. Acute ligation of both renal pedicles provokeda 3- to 4-fold increase in IVC PTH, but only a slight increasewas observed when the ureters alone were ligated. After 5 h ofan intravenous infusion of [125I]PTH + TCA-PTH orTCA-PTH alone, the kidneys were removed and snap frozenor fixed in formalin. Immunofluorescent studies of tissue sec-tions using a 1:10 dilution of an anti-PTH guinea pig serumshowed a selective localization of PTH in the cytoplasm ofrenal tubular cells. Autoradiographic studies demonstratedthat PTH was only localized in the cells of the proximaltubules. Sections of other tissues of the same animals gavenegative results. Weconclude that in the dog, bovine PTH isselectively trapped by cells of the proximal tubules. (Sup-ported by VA and Oklahoma Medical Research Foundation.)

230. Role of RBCMembrane Lipid Peroxidation in Paroxys-mal Nocturnal Hemoglobinuria (PNH). N. V. PANIKER,* A.B. ARNOLD,* AND R. C. HARTMANN,Nashville, Tenn.

Increased susceptibility of PNHRBCmembrane unsatur-ated fatty acids to peroxidation has been cited as a basic defectin PNH. Since iron catalyzes lipid peroxidation, support forthis thesis has derived in part from the frequent occurrence ofgross hemoglobinuria after iron therapy. An equally tenablehypothesis is that iron therapy stimulates the production ofa greater number of "defective" cells. in iron-deficient PNHpatients. Supporting observations are: (a) frequent latentperiod with improvement in anemia long before the onset ofgross hemoglobinuria; (b) lack of correlation of serum ironlevels with degree of gross hemoglobinuria; (c) rise in reticulo-cyte count before and fall after onset of gross hemoglobinuria;and (d) failure of added iron to augment PNHin vitro hemoly-tic tests. Additionally, two of our PNHpatients treated formore than 5 months with the biologic antioxidant, vitamin E,failed to show any response with respect to hemoglobinuria andtransfusion requirements. Exposure of normal and PNHRBCto steady, low levels of H202 was carried out by diffusion tech-

nique and by the action of liver microsomes in the presence ofTPNHand ATP. Formation of large amounts of malonalde-hyde indicated adequate peroxidation, but there was no de-crease from base line RBC acetylcholinesterase activtiy.Despite extensive lipid peroxidation of normal RBC, theyshowed no hemolysis in a variety of in vitro PNH hemolytictests. Likewise, peroxidation after the inhibition of intracel-lular and membrane sulfhydryl groups in normal cells gavesimilar negative results. Moreover, addition of iron was with-out influence in these experiments. These data provide strongevidence against the thesis that PNH RBC membrane lipidperoxidation plays any significant role in this disorder.(Supported by grant HL-03509 from the NIH.)

231. The Origin of Glycoproteins in HumanAlveolar Proteino-sis. M. A. PASSERO,* S. N. BHATTACHARYYA,* R. W. TYE,*AND W. S. LYNN, Durham, N. C.

Two glycoproteins of mol wt 62 and 36 X 103 are found inlavage material from patients with pulmonary alveolar pro-teinosis. Each glycopeptide contains hydroxyproline (1%),hydroxylysine (0.5%), glycine (13%), and cysteine (5%).Each also contains about 8% carbohydrate, of which 1% issialic acid and 2% is glucose. Antibodies to the glycopeptidesdo not precipitate human serum proteins. These glycopeptideswere not found in isotonic lavage material from a normal hu-man volunteer. Material prepared from human or rabbit wholelung plasma membranes, endocytoplasmic reticulum, mito-chondria or soluble proteins, plasma membranes, and alveolarmacrophages also did not contain detectable amounts of theseglycopeptides. The observations of Nidin suggesting that Claracells are apocrine cells which may secrete "surfactant" haveled us to attempt to isolate these secretory products. Hypo-tonic washes of the rabbit airways with EDTAsolutions con-tain two proteins similar to those obtained from the proteinosispatients. Characterization of these normal glycopeptides willbe reported. Attempts to determine whether basement mem-brane or Clara cell secretions are the source of these glyco-peptides using fluoroscein-labeled antisera to the purifiedalveolar proteinosis glycopeptides will be reported. (Supportedby OH00302-NIOSH and ES00124-NlEHS, both fromUSPHS.)

232. Mechanism of Parathyroid Hormone-Mediated Stimu-lation of Uridine Incorporation in Isolated Bone Cells. WIL-LIAM A. PECK* AND KIRK MESSINGER,* Rochester, N. Y.(introduced by John R. Jaenike).

Treatment of isolated bone cells in primary culture withparathyroid hormone (PTH) is known to increase bone cellcyclic 3'5'-AMP (cAMP) and subsequently to enhance theincorporation of [2-i4C] uridine into the RNA precursorfraction and into RNA, an effect which is mimicked by exogen-ous dibutyryl cyclic 3'5'-AMP (Bt2-cAMP). To determine themechanism of this effect, we examined the influence of theseagents on [2-i4C] uridine incorporation into individual freeribonucleotide pools, and on activities of enzymes of uridinemetabolism in isolated bone cells. Exposure to PTH (100ng/ml) or to Bt2-cAMP (800 ,M) for 30 min significantly(P < 0.01) enhanced [2_i4C] uridine incorporation intocellular free uridine diphosphate (UDP) and uridine triphos-phate (UTP) pools (PTH; +29%, Bt2-cAMP; +25%) butdid not increase free uridine monophosphate (UMP) radio-activity. UMPphosphorylation was significantly (P < 0.01)augmented in loo,000 g supernatant fractions prepared fromsonicates of cells treated with PTH (1 ,g/ml) or Bt2-cAMP(800,uM) for 30 min (PTH; +37%, Bt2-cAMP; +38%). UTPdegradation was not suppressed, nor was uridine phosphoryla-tion enhanced in the same supernatant fractions, although in-creased conversion of UMPto UDPand UTP was apparent

63 a

when uridine was the substrate. The effects of PTH and Bt2-cAMPcould not be explained by alterations in bone cell ATP,as assays were performed in the presence of excess (6 mM)ATP, and neither agent increased bone cell ATP levels. Hence,PTH appears to enhance [2-'4C] uridine incorporation bycAMP-mediated stimulation of UMPkinase. Since UMPphosphorylation is the first common step in UTP formationvia salvage and de novo pathways, stimulation of this stepcould increase the availability of UTP for many metabolicprocesses, including RNA synthesis. (Supported by grantsfrom NIH.)

233. Studies with Oral Cyanate in Sickle-Cell Disease.CHARLESM. PETERSON,* YANGS. Lu,* JAMESM. MANNING,*PETER N. GILLETTE,* AND ANTHONYCERAMI,* New York(introduced by Attallah Kappas).

Cyanate diminishes sickling by specific carbamylation ofamino terminal valines of hemoglobin S (1971. Proc. Nat.Acad. Sci. U. S. A. 68: 1180). Aliquots of erythrocytes frompatients with sickle-cell disease have an enhanced survivalafter treatment in vitro with cyanate (1971. Proc. Nat. Acad.Sci. U. S. A. 68: 2791). Short-term administration of cyanateis well tolerated and decreases the hemolytic anemia (1972.Adv. Exp. Med. Biol. 28: 261). No adverse effects have beenobserved to date in cyanate-treated animals maintained at1.0-1.8 moles cyanate per mole hemoglobin tetramer(NCO/Hb) for 18 months (J. Pharmacol. Exp. Ther. In press).26 patients with sickle-cell disease (S/S) received sodium cya-nate orally in doses of 5-30 mg/kg per day for 8-15 monthsto determine toxicity and clinical efficacy. Maximal carbamy-lation was attained after 20-40 days, and the hematologicresponses were dose related. Circulating hemoglobin increased5-20% in patients maintained over 0.3 NCO/Hb, red cellmass increased 13% in six patients at 0.44 NCO/Hb, andoxygen affinity increased (APo - -3 mmat 0.5 NCO/Hb).Subjective toxicity was minimal, and no objective adverseeffects of cyanate were observed. 43 crises occurred; the meanincidence was 2.3/yr for 13 patients with carbamylation rang-ing from 0.05 to 0.26 NCO/Hb and 1.1/yr for 13 patients at0.28-0.54 NCO/Hb. This evidence suggests that cyanate haspromise as a possible therapeutic agent for sickle-cell diseaseand establishes the rationale for a randomized clinical trial.(Research supported by grants from NIH and NF.)

234. Differential Inhibition of Cholera and E. Coli Entero-toxins by Cholera Toxoids and Ganglioside. N. F. PIERCE,*Baltimore, Md. (introduced by C. C. J. Carpenter).

Cholera enterotoxin (CT) and the heat-labile component ofE. coli enterotoxin (ECT) each induce gut secretion by acti-vating mucosal adenyl cyclase. Yet the response to CT is slowin onset and of many hours duration, while that of ECT israpid in onset and of brief duration, suggesting differences inmucosal binding of the enterotoxins. Weinvestigated CT- andECT-binding characteristics by comparing the inhibitoryeffects of natural cholera toxoid (NCT), formalinized choleratoxoid (FCT), and ganglioside on the secretory response ofrabbit small bowel to CT and ECT. NCTacted as a competi-tive inhibitor of CT, being maximally effective when injected4 h before CT. Mucosal binding of CT was more rapid thanmucosal binding of NCT, but once bound the binding of eachappeared to be irreversible. By contrast, FCT did not inhibitthe secretory response to CT. FCT and NCTare each anti-genic, but neither stimulates mucosal secretion. These datasuggest that enterotoxic activity, mucosal binding capacity,and antigenicity are independent properties of CT. NCThadno inhibitory effect upon the secretory response to ECT, indi-cating different mucosal binding sites for ECTand CT. Priorstudies suggest ganglioside is the mucosal receptor of CT.

Premixing of CT with nanogram amounts of ganglioside com-pletely deactivated CT. By contrast ganglioside was relativelyinefficient in deactivating ECT, 1000-fold greater amounts ofganglioside being required. These findings suggest that ganglio-side is not likely to be the mucosal receptor for ECT. Thesedifferences in mucosal binding of ECT and CT may explain,at least partly, the marked differences in time of onset andduration of their secretory effects. (Supported by NI H grantAI-07625 and a Career Development Award.)

235. Mass Testing for Hematological Disorders. SERGIOPIOMELLI,* BERNARD DAVIDOW,* PATRICIA YOUNG,* ANDGISELLE GAY,* New York (introduced by Peter Elsbach**).

Microsamples of blood can be collected by untrained person-nel from finger puncture on S & S 903 filter paper for shipmentto a centralized laboratory. With a i" punch, aliquots equi-valent to 11 ul of blood are obtained from the spots. (Quanti-fication and reproducibility was confirmed with ilCr-labeledsamples of varying Hct.) Thus Hgb can be quantitated byelution into a modified cyanmethemoglobin reagent. (1000paired samples demonstrated an excellent correlation withexisting techniques.) Hgb electrophoresis can be performedeasily by direct implantation of the filter papers in agar gel.Free erythrocyte porphyrins (FEP) can be measured fluoro-metrically in 1.5 N HCl, after a rapid extraction in ethylacetate/acetic acid in the presence of Celite. (Paired measure-ments in 25 samples with widely differing FEP content agreedclosely with the standard technique of Wranne.) The ratioFEP/Hgb was an excellent indicator of Pb intoxication. In5000 samples analyzed for Pb levels by the New York CityBoard of Health, FEP/Hgb increased exponentially with theblood Pb level. The FEP exceeded 6.2 ,ug/g Hgb in all 290samples with Pb > 60 ,g/100 ml, but only in 5%of those withnormal Pb. FEP/Hgb was increased in 55 cases of Fe defi-ciency, but to a lesser degree than in Pb poisoning. A prelimi-nary field screening of 1100 children with these techniquesuncovered 42 cases of anemia (29 due to Fe deficiency) and 17cases of Pb poisoning. These simple microtechniques offera powerful tool for the mass detection of anemia, Fe deficiency,Pb poisoning, and hemoglobinopathies. (Supported by NIHgrant AM09274-08 and NYCHRCgrant U 2282. Dr. Piomelliis a Career Scientist of the NYCHRC,contract 1-383.)

236. The Effect of Histamine on the Cell-Mediated ImmuneResponse. MARSHALLPLAUT,* CHRISTOPHERS. HENNEY,*ANDLAWRENCEM. LICHTENSTEIN, Baltimore, Md.

Histamine inhibits lymphocyte-mediated cytolysis, as shownby in vitro experiments involving the killing (51Cr release) ofmurine DBA/2 mastocytoma cells by splenic thymus-depend-ent (T) lymphocytes from sensitized C57B1 mice. This inhibi-tion is related to and presumably caused by histamine'sability to increase lymphocyte cAMP levels. Standard anti-histamines (e.g., diphenhydramine) do not block this effect,but burimamide, a new type of antihistamine which affectshistamine 2 receptors, reverses both the inhibition of cytolysisand the rise in cAMP. The inhibition of cytolysis by a standard(10-' M) amount of histamine increased from 11 2%at day10 to 40±4% on day 14, while the ability of histamine to in-crease the cAMP level of the entire pool of splenic lympho-cytes remained constant throughout the immune response.The increasing inhibition observed was not an artifact ofaltered lytic, activity or of a change in the threshold sensitivityof the lymphocyte -to histamine, suggesting that the proportionof cytolytically active T-lymphocytes bearing histamine recep-tors increases with time after immunization. Histamine givenintraperitoneally on day 4 markedly decreased the subsequentcytolytic activity of spleen cells measured in vitro on days 11 or13, although the "killer" T-cells which remained still bore

64 a

histamine 2 receptors, as they were inhibited by histamine invitro. Histamine also inhibited antibody production: animalstreated with histamine on day 4 showed a decrease in masto-cytoma agglutination titers assayed on day 11. These data sug-gest that histamine may affect antigen-driven cellular differen-tiation, as well as having marked suppressive effects on effectorT-cells. (Supported by grants AI-10280 and AI-00423 fromNIH and a grant from the NSF.)

237. Removal of Bilirubin from Human Plasma by AffinityChromatography on Sepharose-Conjugated Albumin. PAuLPLOTZ,* PAUL BERK,* JOYCEGORDON,*ANDJOHNVERGALLA,*Bethesda, Md. (introduced by Jacob Robbins**).

Substances in human plasma that bind tightly to plasmaproteins cannot readily be removed. Unconjugated bilirubin(BR) is so firmly bound to albumin that it cannot be removedeven by dialysis. In a study of possible methods for removingBRor other potentially toxic albumin-bound substances fromplasma, we have coupled human albumin to Sepharose 6Bwith cyanogen bromide, attaining concentrations of 25-50 mgalbumin per g wet weight of gel. Plasma from a patient withCrigler-Najjar syndrome, containing 19.5 mg/100 ml. (195,ug/ml) BR, trace-labeled with [14C]BR, was passed over suchgels. 1 g of gel bound in excess of 90 Ag of BR, determined bychemical assay or radioactivity. The radioactive material waseluted quantitatively with 50% ethanol in water and thecolumn reused repeatedly without apparent loss of efficiency.An anion-exchange resin, Dowex 1-X2 (200-400 mesh), boundBR somewhat more efficiently per gram wet weight, and anuncharged resin, XAD-2, less efficiently than the Sepharoseadsorbent (cf. Willson et al. 1972 Gastroenterology. 62: 1191-1199). Two other substances which are partially bound toalbumin were studied. [25I] thyroxine, equilibrated withnormal serum, bound to the Sepharose-albumin and waseluted by 50% ethanol; [PCI salicylic acid, equilibrated withnormal serum, also bound to the gel and was eluted withbuffered saline. Protein adsorbents such as Sepharose-albuminmay prove useful in the removal of BRor other protein-boundtoxic substances from human plasma in patients with general-ized hepatic failure, specific metabolic deficits, or intoxicationwith protein-bound drugs.

238. Paradoxical Effects of L-DOPA Administration in Hunt-ington's Chorea. STEPHEN PODOLSKY* AND NORMANA.LEOPOLD,* Boston, Mass. (introduced by Belton A.Burrows**).

It has been reported that L-DOPA administration increasesthe adventitious movements of patients with Huntington'schorea (HC), possibly due to an increased sensitivity to intra-cerebral dopamine in these patients. We therefore studied itseffects on other parameters (glucose, insulin, and growthhormone). 10 patients with documented HC (mean age 48.5yr and mean duration of disease 6.8 yr) and 10 control sub-jects (mean age 51.5 yr) had glucose and arginine tolerancetests before and after being given L-DOPA (1.5 g/per day) for3 days. No clinical change was noted in controls or patientswith HC. L-DOPA was associated with a rise in mean glucoselevel in controls 2 h after oral glucose (before = 95.6±6.8mg/100 ml; after = 115.3±9.1 mg/100 ml), but a fall in thepatients with HC (before = 125.9±11.0 mg/100 ml; after= 116.648.9 mg/100 ml). In controls the mean insulin levelat 2 h was slightly lower with L-DOPA (before = 71.0±10.9M&U/ml; after = 61.2±8.8 MU/ml). In six HC patients whohad abnormal glucose tolerance, the insulin levels 2 h afteroral glucose were lower with L-DOPA (before = 197.5±41.8MU/ml; after = 64.5±9.7 MU/ml). However, in the fourpatients with normal glucose tolerance, the insulin levels 2 hafter glucose were higher with L-DOPA (before = 40.44±8.4

,uU/ml; after = 61.5±12.3 MU/ml). In all patients basalgrowth hormone was increased by L-DOPA (before = 4.2 ±0.8ng/ml; after = 17.044.9 ng/ml), but unlike normal controlswas completely suppressed by oral glucose. In 5/6 patientswith HCgiven L-DOPA, the peak growth hormone rise 60 minafter arginine infusion (0.5 g/kg) was attenuated (before= 34.447.5 ng/ml; after = 15.9±5.2 ng/ml). It is clear thatL-DOPA has significant effects on glucose, insulin, and growthhormone metabolism different in Huntington's chorea than innormal subjects. (Supported by a VA Clinical Investigatorgrant.)

239. Scanning Electron Microscopy of Normal and LeukemicHuman Blood Cells. AARONPOLLIACK,* BAYARDD. CLARK-SON, NINA LAMPEN,* AND ETIENNE DEHARVEN,* New York.

Until now, blood cells have been prepared for scanningelectronmicroscopy (SEM) mainly by air drying. This pro-cedure causes distortion and loss of surface detail and conse-quently precludes distinction between different leukocytesunder SEM. The present report describes the surface featuresof normal and leukemic blood cells dried either in the air or byC02 critical point. Cell suspensions obtained from buffy coatswere collected by aspiration and filtration on to silver filters,fixed with gluteraldehyde and osmium, and dehydratedthrough a graded series of alcohols. One half of the filter wasair dried, while the other was critical point dried after furtherdehydration in a graded series of amyl acetate. Comparison ofthese procedures showed that critical point drying was superiorand enabled one to distinguish different leukocytes. Aftercritical point drying finger-like surface projections were con-sistently present in most leukemic lymphocytes. Air dryingcaused shrinkage of cells and considerable loss of surface detailand did not allow ready characterization of the different cells.These findings indicate that SEMis useful in the recognitionof different leukemic cells prepared by critical point drying.(Research supported in part by grant CA-08748 from theNCI. Dr. Polliack is from the Department of Hematology,Hadassah University Hospital and Medical School, Jerusalem,Israel, on an International Fellowship from the NIH andUSPHS.)

240. Feedback Control of Cholesterol Synthesis in CirculatingGranulocytes and Deletion of Feedback Control in a Granulo-cytic Leukemia. FRED I. POLSKY,* MICHAEL S. BROWN,*AND MARIVIN D. SIPERSTEIN,** Dallas, Tex.

To demonstrate feedback control of cholesterol biosynthesisby dietary cholesterol in blood cells, a method was developedto isolate rat granulocytes, lymphocytes, and platelets and tomeasure their sterol synthesis from [14C] acetate. The ratesof acetate incorporation into cholesterol in granulocytes,lymphocytes, and platelets were 0.441, 0.263, and 0.001mjsmoles/h per 106 cells. However, the cholesterol syntheticrates per milligram of protein were nearly equivalent: 0.549,0.817, and 0.225 mmmoles/h. Feedback control was found incirculating granulocytes, but not in lymphocytes or platelets;a 5% cholesterol diet suppressed cholesterol synthesis ingranulocytes by 91%. The activity of HMGCoA reductase,the rate-limiting enzyme in hepatic cholesterol synthesis, wassimilar in microsomes from granulocytes and liver (0.357 vs.0.280 mjsmoles/min per mgmicrosomal protein). After 21 daysof cholesterol feeding, granulocyte HMGCoA reductase activ-ity decreased 1y 87%, a degree of depression sufficient to ac-count for the observed 917% inhibition of cholesterol synthesis.Since we have observed previously that feedback control ofcholesterol synthesis in liver cells is deleted after their malig-nant transformation, the effect of cholesterol feeding onmalignant granulocytic cells was examined. A spontaneous ratgranulocytic leukemia maintained in ascitic fluid synthesized

65 a

cholesterol at about the same rate as normal granulocytesisolated from peritoneal fluid (0.579 vs. 0.495 miumoles/h per106 cells). However, although the normal granulocytes showeda 90% inhibition of cholesterol synthesis after 21 days ofcholesterol feeding, no reduction occurred in the leukemiccells. This finding suggests that loss of the cholesterol feedbacksystem may be a more general characteristic of malignantchange than was generally believed.

241. Clinical and Genetic Heterogeneity of PseudoxanthomaElasticum. F. MICHAELPOPE,* Baltimore, Md. (introduced byVictor A. McKusick**).

Whereas pseudoxanthoma elasticum (PXE) was previouslyconsidered a single autosomal recessive disorder, we have foundclinical and genetic evidence of heterogeneity. An attemptedtotal ascertainment of affected persons in Britain provideddata on 182 patients. Pedigrees with two or more affectedpersons showed two patterns: one with multiple affected gen-erations typical of autosomal dominant inheritance, and theother with a single affected generation, typical of autosomalrecessive inheritance. When the pooled genetic data were ex-amined, correcting for bias of ascertainment, the affected-unaffected ratios were close to the one-to-one ratio expectedfor heterozygous dominant with normal matings, and close toone-to-three for the unaffected heterozygous matings of therecessive group. Clinical separation of two dominant and tworecessive groups was also possible. Dominant type I disease ischaracterized by classical, flexurally distributed rash, severeangina, hypertension, intermittent claudication, and verysevere degenerative retinopathy. In contrast, type I I dominantdisease is much milder with hardly visible cutaneous infiltra-tion, few vascular complications, and mild retinal disease.Type I recessive PXE is of intermediate severity, except thatthere is a particular liability to gastrointestinal bleeding. Itaccounted for 66% of such complications in this series. TypeII recessive disease is very rare and is characterized by a gen-eralized cutaneous infiltrate, but no evident ocular or vascularcomplications. Of the 122 index patients, 12 were dominanttype I, 52 were dominant type II, 54 were recessive type I,and only 3 were recessive type II.

242. The Role of Cell Swelling in Myocardial Ischemia andthe Protective Effect of Hypertonic Mannitol. WM. JOHNPOWELL, JR.,* JORGE FLORES,* DONALDR. DIBONA,* ANDALEXANDERLEAF,** Boston, Mass.

In brain and kidney, elevation of extracellular osmolality,through reduction of ischemic cell swelling, improves reflowof blood after arterial occlusion. Recently it has been demon-strated that hyperosmolar mannitol increases collateral bloodflow to ischemic myocardium with improved myocardial func-tion and decreased ischemia. The present studies in 55 dogsanesthetized with chloralose and urethane document and studythe mechanism of impaired reflow after coronary occlusion andexamine the possible salutary effect of mannitol in improvingreflow and in lessening the extent of myocardial damage. In-fusions of silastic into the aortic roots of 10 dogs after 40-60min of left circumflex artery occlusion and 15 min of reflowrevealed impaired distribution of silastic to the region of theposterior papillary muscle. Parallel electron microscope studiesdemonstrated myocardial and endothelial cell swelling withpartial closure of vessel lumina. This cell swelling was pre-vented by elevation of serum osmolality by 30-40 mOsmabovecontrol with hypertonic mannitol administered during andafter occlusion. These morphologic findings were associatedwith a 42±10 (SEM) % (P < 0.02) reduction of vascularresistance after occlusion if mannitol was administered. Micro-scopic examination of tissue obtained 12 h after the period ofocclusion revealed less damage in hearts which received manni-

tol. Thus, this study documents impaired reflow of blood aftercoronary artery occlusion and points to cell swelling as atleast a partial mechanism. Reflow is substantially increased,cell swelling reduced, and subsequent damage lessened by theadministration of hyperosmotic mannitol. (Research sup-ported by grants from NIH.)

243. DNASynthesis and Thymidine Kinase in Zinc-DeficientTissue. ANANDAS. PRASADANDDONALDOBERLEAS,* Detroit,Mich.

Growth retardation is a consistent manifestation of zincdeficiency in various animal species and man. Zinc is known tobe essential for many enzymes, but its precise role in cellulargrowth has not been established. In this study, effects of zincdeficiency on DNAand activity of thymidine kinase (TK)in rapidly regenerating tissue were investigated. After 1 wkon zinc-deficient diet, rats were implanted with polyvinylsponges subcutaneously. 6 and 10 days later, sponges wereremoved and connective tissue capsules studied in vitro. Con-tinuously pair-fed and ad lib.-fed rates were used as controls.DNA was measured by a modified method of Schmidt-Thannhauser. Activity of TK was assayed in vitro by themethod of Lieberman. Enzyme activity was expressed asnanomoles of thymidine monophosphate generated per hourper milligram protein. Incorporation of [2-14C] thymidine asa measure of DNAsynthesis was determined in vitro in 6 daytissue in zinc-deficient and pair-fed control rats. TK activityin sixth day connective tissue was as follows (units mean±SE): deficient, 0.58±0.02; pair fed, 2.440.59; and ad lib.,1.65±0.12. The difference between deficient and controls wasstatistically significant (P < 0.025). In tenth day tissue, defi-cient rats showed no measurable TK activity (pair fed,2.7±0.6; and ad lib., 2.68±0.7). [2-14C] thymidine incorpora-tion into DNAwas also significantly reduced in sixth day defi-cient tissue (DPM/mg DNA: deficient, (18±5.7) X 103 andpair fed, (136±15.2) X 103, P < 0.001). In conclusion, this isthe first demonstration of an adverse effect of zinc deficiencyon activity of thymidine kinase in animals. Inasmuch as thisenzyme is essential for cell division, our data suggest that thisearly metabolic defect may account for growth retardation inzinc-deficient rats. (Research supported by VA grant and NIHcontract.)

244. Tissue and Sera Factors Affecting Organic Anion andCation Transport in Hypertensive Rats (SHR). H. PREUSS,*M. ZMUDKA,* K. GRANT,* R. PARRIS,* L. SLOTKOFF,* ANDG. SCHREINER,** Washington, D. C.

Altered renal metabolism may be causally related to essen-tial hypertension. The present study was performed in a strainof spontaneously hypertensive Wistar rats (SHR), thought bymany to be the best animal model of this disease. Organicanion (PAH) and organic cation (TEA) transport in renalslices were compared in SHRand normotensive Wistars (C).In 14-wk-old SHRrats, blood pressure taken via tail plethys-mography averaged 146 mmHg compared to a control of 113mmHg. In nine experiments, the accumulation of PAH byslices from SHRexceeded control (T36%, P < 0.01), whileTEA transport was significantly decreased (17%, P < 0.05).In 7-wk-old SHR that had not yet developed hypertension(BP 110 mmHg to control 107 mmHg), the same pattern wasnoted: PAH accumulation T12% (P < 0.05) and TEAtransport 110% (P < 0.001). Within a given experiment, theABP between C and SHR correlated significantly with theAPAH (r = 0.56, P < 0.02). The presence of changes intissue are present before the development of elevated BP andconstitute a physiologic difference in renal function in SHR.Sera taken from SHRcompared to sera from control depressedsignificantly PAH transport in normal slices. When the sera

66 a

was fractionated on Sephadex 25, the depressor in SHRserato kidney slice PAH transport was found in the fourth frac-tion, the same fraction that suppressed PAHtransport and Naexcretion in azotemic sera (1971. J. Clin. Invest. 50: 303).Fraction 4 from C had no effect on PAHtransport. The obser-vation of renal metabolic abnormalities and factors in seraaffecting PAH transport early in the development of hyper-.tension may have implications in the pathogenesis of hyper-tension. (Supported by NIH grant 15458.)

245. In Vitro Hybridization of C-Type Viral RNAto HumanDNA. PETER M. PRICE,* NORMANGABELMAN,* KURTHIRSCHHORN,** ANDSAMUELWAXMAN,*New York.

We have previously reported the probable chromosomallocalization of human hemoglobin genes by in situ hybridiza-tion. With this technique, we subsequently detected sequencehomology between a D-group chromosome of uninfected hu-man cells and the RNAof a putative human oncogenic C-typevirus (RD114). This data lends credence to the "oncogene"theory proposed by Huebner and Todaro which proposes thatall mammalian genomes contain the information for transcrip-tion of oncogenic RNAtumor viruses. Wehave now investi-gated the product of in vitro hybridization of solubilizedhuman placental DNAwith the high and low molecular weightRNA components of Feline leukemia and RD114 viruses.Our observations have shown that there is a large degree ofhybridization of the high molecular weight RNAspecies (35to 70S) of both viruses with human DNA, and that the lowmolecular weight species (4S) extracted from these virusesdemonstrates much lower saturation kinetics with this DNA.These results indicate that the low molecular weight RNAspecies contained within C-type RNA tumor viruses is nota degradation product of the high molecular weight com-ponent. The low saturation kinetics also make it unlikely thatthe 4S RNAis a contaminating human cellular RNAspecies.This is supported by the finding that low molecular weighthuman RNA hybridized with human DNAdoes not reachsaturation at even much higher concentrations. Our resultsindicate that a portion of human (and perhaps most mam-malian) DNA contains closely related nucleotide sequenceswith mammalian C-type viral RNA, and that there is a sub-stantial number of oncogenic viral genome copies preexistingin the genome of uninfected human cells. (This study was sup-ported by grants from NIH.)

245a. "Little" Luteinizing Hormone (LH): the Appearancein Serum of an hLH-Related Peptide, Possibly a Subunit,After LHRHStimulation. D. RABINOWITZ,* R. BENVENISTE,*L. FROHMAN,J. BELL,* AND I. SPITZ,* Jerusalem, Israel (in-troduced by C. R. Kleeman).

This report describes the identification of a human luteiniz-ing hormone [hLH]-related peptide which can be preparedfrom pituitary hLH powder and is found in serum after LH-releasing hormone [LHRH] stimulation. Two purified pitui-tary preparations of LH, LER 960, and Hartree IRC-2 werelabeled with 125J and submitted to gel filtration on a SephadexG-100 column. The IRC-2 preparation showed two peaks,"standard LH" and "little LH" with Ka. = 0.24 and 0.40,respectively. LER-960 appeared as a single peak (Kay = 0.24)which could be partially converted to little LH [Ka, = 0.40]by 8 M urea treatment. Molecular weights for standard LHand little LH were estimated to be 32,000 and 18,000 respec-tively, and it is likely that the peaks represent "associated"and "dissociated" (i.e., individual subunits) forms of LH.Purified standard LH and little LH material were used astracers in a double antibody radioimmunoassay (RIA) withrabbit anti-HCG antibody. By choosing appropriate antibodyconcentrations, standard and little LH could be selectively

assayed by these RIA procedures. Selected human sera werechromatographed on Sephadex G-100, and the eluted fractionswere assayed in the two systems. Results follow. (a) LowLH sera gave negative results in both RIA's. (b) High LHserum, obtained at the time of the midcycle LH surge in anormal menstruating woman, gave a single peak in the positionof associated LH, which reacted only in the standard LHRIA. (c) High LH serum, obtained from a young patient withovarian failure, gave two peaks, the majority being standardLH, with a small but definite little LH peak. (d) Between 5and 20 min after LHRHstimulation in this patient, a largeincrease in little LH occurred, this little LH constituting 39%of total serum LH. Little LH had fallen to 18% of total LHby 90 min after LHRH. (e) Appearance of little LH has beenobserved also in two other patients with high LH levels afterLHRH. XWeconclude that little LH[ possibly dissociated LHsubunits, has been demonstrated in serum, and its appearancecan be triggered by LHRH.

246. Ultrastructural and Physiological Evidence for Cortico-steroid-Induced Alterations in Heptaic Production of VeryLow Density Lipoproteins. EVE P. REAVEN,* ORVILLE KOLT-ERMAN,* AND GERALDM. REAVEN, Stanford, Calif.

Werecently suggested that the cause of hyperlipoproteine-mia in corticosteroid-treated (CS) patients is overproductionof hepatic very low density lipoproteins (VLD). Wenow haveadditional evidence to support this observation. Methyl-prednisolone (0.3 mg/100 g) treatment of young adult ratsfor 8 days produced an increase in plasma levels of triglycerideof 148% (P < 0.01) and cholesterol of 72% (P < 0.01). InCS rats injected with Triton WR1339 (600 mg/kg), the esti-mated rate of accumulation of triglyceride into the measuredplasma compartment was 176 mg/2 h as compared to 117mg/2 h for control rats (P < 0.01). Liver samples were pre-pared for electron microscopic morphometric analysis of VLDlocalization, size, and abundance. The results show that inboth corticosteroid and control animals, VLD are primarilyassociated with hepatocyte Golgi complexes. The number ofGolgi complexes per area of hepatocyte cytoplasm does notchange with corticosteroid treatment. However, in 15 largeGolgi analyzed from each of five control and five CS animals,the mean (±SE) number of VLD per Golgi complex increased(P < 0.02) in CS animals (121+11 -+ 187-19) and theaverage diameter of VLD per Golgi increased (P < 0.01)in CS animals (535 A -. 707 A), representing a 120% increasein VLD volume. Thus, CS treatment results in (a) an increasein plasma triglyceride, (b) an increase in rate of accumulationof triglyceride after the inhibition of VLD clearance by Triton,and (c) an increase in the number and size of VLD present inhepatocyte Golgi complexes. These combined results suggestthat CS induces hyperlipoproteinemia through increasedhepatic production of VLD. (Research supported by grantHL 08506 from NIH.)

247. Proteases from Skeletal and Heart Muscle. M. K.REDDY,* J. ETLINGER,* R. ZAK,* D. A. FISCHMAN,* AND M.RABINOWITZ, Chicago, Ill.

Although intracellular muscle proteases are likely to be in-volved in the turnover of proteins in both normal and diseasedmuscle, their localization and mode of action remains obscure.Wehave studied the distribution and specificity of neutral andacid proteases in subcellular fractions. Recently it has been re-ported that a Ca2+-activated sarcoplasmic factor (CASF)causes the removal of the myofibrillar Z line. We find thatCASFcauses the removal from myofibrils of a-actinin withoutmyosin degradation as displayed on SDS-acrylamide gels andwithout loss of the EGTAsensitivity of myofibrillar ATPaseand contractility. Myofibrils digested with trypsin or papain

67 a

lose this EGTAsensitivity and degradation bands of myosinappear before complete Z line disappearance, indicating thatCASFZ line removal differs from trypsin or papain digestion.The CASFpreparation catalyzed a Ca2-activated proteolysisof casein and hemoglobin at neutral pH. Studies are in progressto ascertain whether Z line release is proteolytic in nature. Asecond neutral protease activity has been observed associatedwith the myofibrillar fraction. It is not Ca2+-activated. Mito-chondrial and membrane fractions are enriched in acid pro-tease activities. Unlike the neutral proteases, these show ten-fold activation in the presence of 0.2% Triton X-100. Thesestudies indicate that muscle contains highly compartmental-ized proteases which are activated by specific and differentconditions. (Supported by grants from MIRU, USPHS, AEC,and Chicago and Illinois Heart Association.)

248. Chemical and Biological Properties of Purified Nativeand Desialylated Human Thyroxine-Binding Globulin. S.REFETOFF,* V. S. FANG,* ANDJ. S. MARSHALL,* Chicago, Ill.,and Cleveland, Ohio (introduced by R. L. Landau**).

Thyroxine-binding globulin (TBG) and partially desialy-lated or slow TBG (sTBG) were prepared as previously de-scribed (1972. J. Clin. Invest. 51: 3173). Purified TBGwaselectrophoretically identical with TBG in serum. sTBGmoved more slowly. Both bound thyroxine (T4) equimolarly.sTBGhad lower T4 affinity and was less heat labile. IodinatedTBG and sTBG with from approximately 1 to 6 atoms of'3I1 or 1251 per molecule had increased anodal mobility. Never-theless, T4 binding, antiserum precipitation, and disappearancerates in rabbits of simultaneously injected [E1311] TBG (1atom I per mole) and [1251] TBG (6 atoms I per mole) re-mained unchanged. 15 min after intravenous injection of[1311] sTBG and [1251] TBGmixture in rats, 94% of sTBGwas in liver with an 1311/1251 ratio of 33.2-fold that of serum.With time, sTBG-derived 131I declined in liver and increasedin bile, intestinal content, and urine. In normal rabbits, serum[1251] sTBG/[1311] TBGratios declined from 1.0 to < 0.2 in10 min. The ratio remained unchanged for 1 h in acutelyhepatectomized rabbits. The serum tj of sTBG was 4-5 minand 2-3 days for TBG. Electrophoretic, immunologic, andchromatographic analyses identified 50% of sTBG-derived125I in bile as iodinated fragments and 98% as iodide in urine.In man, liver [125I] sTBG/E11311] TBG ratio was 51.8 timesthat of serum 30 min after intravenous injection during diag-nostic open liver biopsy. Since radioiodinated TBGand sTBGpreserve their biologic and immunologic properties, they areuseful for studies of TBG metabolism. In rat, rabbit, andhuman, sTBG is rapidly cleared by liver. Conversion of TBGto sTBG may regulate the metabolism of TBG. (Supportedby grants AM-15,070 and 15,308 from NIH.)

249. Description of a Serological Reaction Characteristic ofDermatomyositis. MORRISREICHLIN ANDMARTHAMATTIOLI,*Buffalo, N. Y.

Dermatomyositis (DM) is an inflammatory disease ofmuscle and skin for which no specific laboratory diagnostictest is available. Wereport a specific complement (C) fixationreaction involving human sera with calf thymus extract (CTE)as antigen which has thus far been found only in the sera ofpatients with DM. To identify the specificity of the immuno-logical reaction in which the "antigen" is a complex mixture,the following method was developed. Papain digestion ofwhole serum results in complete conversion of IgG to Fab andFc fragments. Serum containing C-fixing antibody to a specificantigen becomes a potent, specific inhibitor of that reactionafter papain digestion. Thus, papain-digested human sera canbe assayed for their ability to block the serum reacting witha specific antigen in a crude extract. Wehave found that the

sera of three cases of DMfix C with an antigen in CTE. Papaindigests of any one of these sera completely blocks the reactionof any of the three sera in their C-fixation reactions with CTE.Papain digests of human sera from patients with systemiclupus erythematosus (SLE) (20), rheumatoid arthritis (20),multiple myeloma and macroglobulinemia (9), polymyositissyndromes associated with other diseases (scleroderma,Sjogren's syndrome) (6), and normal human sera (19) do notblock this reaction characteristic of these DMpatients. TheSLE sera contained a diversity of antibodies to various tissueantigens. Despite a multiplicity of such antibodies, no inhibi-tory activity for the reaction characteristic of DMsera wasfound. Thus far, therefore, antibodies to this antigen appearspecific for dermatomyositis. Further studies of the specificityof this reaction for dermatomyositis are underway. (Supportedby research funds from VA, NIH, and MDAA.)

250. Positive Effect of Intravenous Triiodothyronine (T3) onThyrotropin (TSH) Secretion in Primary Hypothyroidism.E. CHESTERRIDGWAY,* BRUCED. WEINTRAUB,* ANDFARAHEMALOOF,** Boston, Mass., and Bethesda, Md.

That thyroid hormone exerts a negative feedback on pitui-tary TSH secretion has been well established; that it may alsoexert a positive feedback has been demonstrated in the follow-ing study. Eight hypothyroid patients were given intravenousT3, 100 jig on day 1, then 50 jAg intravenously daily for 9 days.Before and after T3 therapy, TSH metabolic clearance (MCR,nl = 48±15 cc/min) and production rates (PR, nl = 66442mU/day) were correlated with peak serum TSH concentra-tions (nl = 11.9,uU/ml) and pituitary TSH reserve (PTSHR,nl = 1.1 mU. min/ml), determined by integrating the totalarea under the TSH curve after intravenous thyrotropin-releasing hormone (TRH, 200 ,jg). In six of eight patients themean base line TSH was 107 juU/ml (nl = 1.6±0.6), TT4= 2.5 Ag/I00 ml, FT4 = 0.5 ng/100 ml, TT3 = 86 ng/100 ml(nl = 200±425), RAI uptake = 14%. After the initial dose ofT3, the serum TT3 rose to a peak of 500 ng/100 ml at 15 minand leveled at 112 ng/100 ml by 10 days. After 10 days, theserum TSH had decreased to 67 jiU/ml. In five of these sixpatients, TSH-PR decreased from 4254 to 2461 mU/day, andthe PTSHRdecreased from 45 to 32 (mU * min/ml) with littlechange in the MCR(from 33 to 34 cc/min). In two of eightpatients, the mean base line TSH was 81 ,uU/ml, TT4 = 0.5jig/100 ml, FT4 = 0.1 ng/100 ml, TT3 = 43 ng/100 ml, RAIuptake = 5%. After T3 therapy, the serum TT3 rose to a simi-lar peak at 15 min and leveled at 130 ng/100 ml by 10 days.At this time, the basal serum TSH in these two cases had in-creased to 106 and 110 ,uU/ml, respectively. In one patient,the peak serum TSH after TRH increased from 270 to 315,uU/ml, the PTSHRincreased from 38 to 41 (mU-min/ml),and the PRfrom 3046 to 4277 mU/day. In the other patient,the peak serum TSH after TRH increased from 180 to 590jIU/ml, the PTSHRincreased from 20 to 66 (mU.min/ml),and the estimated PR increased from 2220 to 3630 mU/daywith minimal changes in the MCR. In summary, intravenousT3 has been shown for the first time to exert a positive centraleffect on TSH secretion under appropriate metabolic condi-tions. (Research supported by USPH Research, MetabolicResearch Grant, and NIH Training Grant.)

251. The Interaction of Hemoglobin E and g-Thalassemia.RONALDF, RIEDER ANDROBERTFELDMAN,* Brooklyn, N. Y.

Hemoglobin E- is found with high frequency in SoutheastAsia and often occurs in association with a 30-thalassemia gene.The interaction of the two genes produces a moderately severethalassemia syndrome with complete absence of HbA. Thediscovery of a 5-yr-old girl with HbE-,3-thalassemia, who hadan Iranian father (fi-thalassemia trait) and a Burmese mother

68 a

(HbE trait), provided the-opportunity of studying the effectson hemoglobin production of the interaction of HbE withi3-thalassemia derived from a gene pool from Western Asia.Hemoglobin synthesis was studied in vitro, in blood and mar-row of the proposita, and in blood of family members. In thesubjects with HbE trait, the ratio in blood of the quantity ofHbA to HbE was 3:1. The 3,A/j3E synthetic ratio in reticulo-cytes was 1.5 and the specific activity of PSE was greater than(SA, suggesting instability of HbE with preferential destructionof abnormal hemoglobin. The blood of the proposita exhibitedonly HbF and HbE; reticulocytes and marrow showed noUAsynthesis. Therefore, this Iranian (3-thalassemia gene is alsoof the 300 type. The SE/a synthetic ratio (0.6) in blood of theproposita was similar to the 8A/a ratio in mildly affected rela-tives with thalassemia trait. The marrow BE/a synthetic ratiowas estimated to be 0.94. These results suggest that the sever-ity of hemoglobin E-3-thalassemia is attributable to instabilityand defective synthesis of HbE in association with absent jSAsynthesis due to a #0-thalassemia gene. In addition, the dataprovide further evidence that in some examples of fl-thalas-semia there may be greater balance of a- and (3-globin synthesisin marrow than in blood. (Research supported by grants fromNIH and Life Insurance Medical Research Fund.)

252. Scintiphotographic Evaluation of Left Ventricular Func-tion in Patients with Acute Myocardial Infarction. PIERRERIGO,* H. WILLIAM STRAUSS,* DEANR. TAYLOR,* DAVID T.KELLY,* MYRONL. WEISFELDT,* AND BERTRAM PITT,*Baltimore, Md. (introduced by Richard S. Ross**).

24 patients with acute myocardial infarction (AMI) werestudied noninvasively by the gated scintiphotographic tech-nique within 48 h of the onset of symptoms. Biplane scinti-photographic images in the right and left anterior obliqueprojections allowed calculation of left ventricular end diastolicvolume (EDV), end systolic volume, ejection fraction (EF),and extent of asynergy (expressed as a percentage of the enddiastolic circumference of the left ventricular free wall). Thesestudies were correlated with the patient's clinical class: classI-uncomplicated AMI; class II, mild to moderate congestivefailure; class III, pulmonary edema; class IV, cardiogenicshock; and the patients' hemodynamic evaluation. 22 patientshad akinesis ranging from 11 to 59% and two hypokinesis.Mean EDVwas 132 ml/m2 (range 60-214) and mean EF 37%(range 19-53). EF in class I (42.6%) was significantly greaterthan in classes II (33%), III (31%), and IV 19% (P < 0.001).There was an inverse correlation between extent of asynergyand EF (r = 0.87, P < 0.005). Correlation between EDVandpulmonary capillary wedge pressure (PCW) was, however,relatively poor (r = 0.56, P < 0.05), suggesting a change inleft ventricular compliance. These noninvasive studies demon-strate depressed function and regional abnormalities in leftventricular contraction in patients with AMI. Caution shouldbe used in the analysis of left ventricular function using PCWin patients with AMI because of possible changes in com-pliance. (Supported by PH 43NHLI 67-1444 from NIH.)

252a. Phospholipid Requirement of Jejunal Lipid-Reesterify-ing Enzymes. JOHN RODGERS*AND RICHARD O'BRIEN,* Al-bany, N. Y. (introduced by John Balint**).

Previously, we reported that lipid-reesterifying enzymeactivities were decreased in jejuna of bile fistula (BF) rats.This was associated with reduced esterification of absorbedfatty acid in vitro. As bile is rich in phospholipid (PL), it wasdecided to determine whether enzyme abnormalities in BFrats correlated with changes of PL content of jejunal micro-somes. Four groups were studied: controls (CD) and BF(BFD) infused intraduodenally with liquid diet for 48 h andcontrols (CS) and BF (BFS) infused for a similar period with

saline. Specific activity of monoglyceride-acyltransferase(MAT), total protein (P), total PL, and total lipid (L) werethen determined for jejunal microsomes. Data are presentedas ratios of microsomal lipid to protein content and nmolesproduct formed per milligram P per min. L/P ratios weresimilar in all groups. PL/P ratios were significantly reduced to0.139 and 0.095 for BFD and BFS, compared to 0.197 and0.199 for CD and CS respectively, (P < 0.01) and were as-sociated with significantly (P < 0.005) lower MATactivities:BFD = 87, BFS = 81, CD = 140, CS = 142. This associa-tion suggests that PL is necessary for normal activity of MATperhaps by altering the binding properties of microsomalmembranes. Bile diversion, by removing bile PL or by re-ducing absorption of dietary lipid, impairs synthesis ofmicrosomal PL, thus producing impaired enzyme activity. CDcontrols had greater P in response to diet compared to CS. PLwas increased proportionally so activity of MATand PL/Pratios were nearly identical. This again emphasizes the im-portance of PL to P ratio for normal microsomal enzymeactivity. (Research support from grants NIH AM11979 andAM15281.)

253. Deuterium-Labeled Folic Acid: Synthesis and Applica-tion to Studies in Man. I. H. ROSENBERG,D. HACHEY,* ANDP. D. KLEIN,* Chicago, Ill.

Quantitative studies of folate absorption and metabolismhave been limited in manand avoided in infants and pregnantwomen by the dose of radioactive folate, which can be safelyadministered. Stable isotopes offer an alternate tool for tracingintestinal absorption and metabolic pathways, and for deter-mination of body pool sizes without the hazards of administra-tion or radioactive compounds. Wehave prepared deuterium-labeled folate (D-PteGlu) as a nonradioactive tracer for studyof folate absorption and metabolism. D-PteGlu labeled in the3'-5' position of the para-aminobenzoyl moiety (PABA)was prepared by reductive dehalogenation with deuteriumgas of 3'-5' dibromofolate prepared from commercial PteGlu.The product was purified by column chromatography andmigrated chromatographically as tetrahydrofolate. D-PteGluwas administered orally to human subjects and the absorbedD-PteGlu was flushed into the urine with an intramusculardose of unlabeled PteGlu. The folate was concentrated fromthe urine using charcoal. Since folate is difficult to derivatizefor gas chromatography, the deuterium-labeled PABA wasisolated for isotopic analysis. PABAand glutamic acid (Glu)were released by oxidative cleavage and alkaline hydrolysis.Both were recovered as N-trifluoracetyl methyl (N-TFA-CH3)esters and quantified by gas chromatography-mass spectro-scopy-accelerating voltage alternation. The N-TFA-CH3-PABA gives a spectrum with a prominent peak at mass tocharge ratio (m/e) 2I, while the deuterated product wasdisplaced to m/e 218. A dilution curve of deuterated to un-labeled PABA was constructed and a dilution of 1:1000 wasreadily detected. This sensitivity and discrimination waseasily adequate for studies of folate absorption where isotoperatios were 1:100 or greater. Isolation of Glu allowed internalstandardization. Deuterium-labeled folates are versatile newprobes for studies of folate absorption, disposition, andmetabolism. (Research supported by NIH grant AM 15351and the U. S. Atomic Energy Commission.)

254. Transcobalamin II-Facilitated Uptake of Vitamin B12 byCultured Fibroblasts: Studies in Methylmalonicaciduria.LEON E. ROSENBERG,ANNE LILLJEQVIST,* AND ROBERTH.ALLEN,* New Haven, Conn., and St. Louis, Mo.

Weshowed previously that intact, cultured fibroblasts froma child with inherited methylmalonicaciduria are unable toconvert extracellular vitamin B12 to its methyl- and deoxya-

69 a

denosylcobalamin coenzymes, whereas broken fibroblast ex-tracts synthesize coenzymes normally. Wenow report studiesof uptake of [67Co] vitamin B12 (B12) by confluent fibroblastmonolayers incubated in serum-free medium for intervals of15 min to 8 h. In normal cells, the uptake of B12 bound tohomogeneous transcobalamin IL (TC II), purified 2 million-fold from human serum, was equal to that for B12 bound toTC II in unfractionated human serum. TC LI-B12 uptake wasmost rapid during the first 30 min, increased throughout the8 h interval, and proceeded at a rate 4-8 times greater thanthat observed for uptake of free B12 or B12 bound to a homo-geneous granulocyte B12-binding protein. Uptake of free andTC I1-B12 was saturable. The capacity for TC Il-B12 farexceeded that for free B12 (Vmax: 100 pg/mg protein per h forTC Il-B12; 10 pg/mg per h for free B12), but the apparentaffinity for the free and bound forms was the same (Ki: 1000p;/ml). Cells from the child with methylmalonicaciduria tookup free and TC I1-B12 normally during the first 2 h, but thenuptake plateaued. After preloading with TC I1-B12 for 15min, these cells released B12 normally; when preloaded for6 h, their rate of B12 release was 30% faster than normal. Weconclude (a) that TC IL facilitates B12 uptake by culturedhuman fibroblasts; (b) that no other serum protein is requiredfor this facilitation; and (c) that the site of the primary de-fect in this child with methylmalonicaciduria lies between theinitial binding of TC lI-B12 to the cell membrane and theintraceilular conversion of the vitamin to its coenzyme forms.

255. Altered In Vitro Granulopoiesis in the Presence ofLeukemic and Normal Marrow Cells. ALAN L. ROSENBLUM,*JOAN M. BULL,* AND PAUL P. CARBONE, Bethesda, Md.

Patients with leukemia and preleukemia manifest deficienthematopoiesis in vivo and depressed granulopoiesis in vitroalthough the bone marrow may show only minimal infiltrationwith abnormal cells. Colony formation is not linearly relatedto the proportion of abnormal to normal cells in the marrowof these patients, and normal colony formation in vitro occursregularly only when the proportion of blasts is under 10%.Using the in vitro assay for granulocyte colony formation(CFU-C), we examined the response of normal marrow tothe addition of small numbers of leukemic cells and unrelatednormal marrow cells. Human marrow was cultured in semi-solid methylcellulose, each plate containing 1.0-2.0 X 105normal test (recipient) cells to which 1.0 X 103 to 1.0 X 104unrelated (donor) cells were mixed before plating. Resultswere expressed as a percentage of control plates. Leukemiccells in concentrations of 0.5-5.0% uniformly inhibited growthof normal marrow cells to 48% of control levels (range 32-63%, 10 values from three patients). Addition of normalmarrow cells to a normal recipient marrow produced markedstimulation to 255% of control (range 210-290%, 10 valuesfrom six normals) if the donor/recipient pair was HLAidentical or resulted in inhibition to 48% of control (range45-56%, 11 values from five normals) if the donor/recipientpair was mismatched for 3/3 or 4/4 antigens. Mismatches for2/4 or 3/4 antigens resulted in intermediate responses. Thesepatterns were reproduced using irradiated or freeze-thaweddonor cells was well as supernates from the freeze-thawedcells. These results demonstrate that granulopoiesis in vitrocan be significantly altered by the addition of small numbers ofnormal or leukemic marrow cells. The relationship to HLAcompatibility suggests one mechanism of inhibition byleukemic cells in vivo may involve recognition and responseto neo-antigen.

256. Diminished Membrane Protein Kinase Activity inErythrocytes from Patients with Myotonic Dystrophy. ALLEND. ROSES* ANDSTANLEYH. APPEL, Durham, N. C.

Myotonic muscular dystrophy is a systemic disorderpossibly of membrane origin which is inherited as an autosomaldominant trait. Although muscle is one of the target organs,the delineation of the primary biochemical defect is madeextremely difficult on minimal biopsy material by the presenceof atrophy, fibrous tissue, and changes of denervation whichmay result in multiple secondary biochemical changes. If theerythrocyte membrane is found to express the same bio-chemical genetic traits, the difficulties with interpretation ofmuscle tissue are obviated. A significant difference was demon-strated in the phosphorylation of membrane proteins. RBCghosts were frozen at - 200 for 1 wk and examined underinitial rate conditions for transfer of p32 from [E-Y-32P]ATPto membrane proteins by SDS-gel electrophoresis. There were17.5 pmoles PI2/min per mg ghost protein incorporated ineight normal age- and sex-matched control patients, but only8.73 pmoles incorporated in seven myotonic patients fromthree different families (P < 0.001). Each individual mem-brane component which was phosphorylated reflected thisdifference. However, there was no difference in the cyclicAMPstimulation of normal and myotonic erythrocyte pro-tein phosphorylation. Furthermore, using purified F2Chistone as a substrate with membrane protein kinase, therewas no difference in 32p transfer between normal and myotonicmembranes. No differences were observed in membraneATPases, protein and lipid-bound carbohydrates assessed byGLC, phospholipids and gangliosides assessed by TLC, ormembrane proteins assessed by SDS-gel electrophoresis.Studies are presently in progress to characterize this mem-brane-associated enzyme further and to determine its signif-icance in myotonic dystrophy. (Research supported by grantNS07872 from NIH; and the National Multiple SclerosisSociety.)

257. The Effect of Prednisone in Paroxysmal NocturnalHemoglobinuria. WENDELLF. ROSSEAND J. R. MARION,*Durham, N. C.

Prednisone markedly reduced hemolysis in patients withparoxysmal nocturnal hemoglobinuria (PNH). When it wasgiven in high doses (30-60 mg/day) to six patients, the plasmaand urine hemoglobin and the endogenous carbon monoxideproduction fell abruptly. In one patient, hemoglobinuria wasdecreased from 22 g/day to > 0.05 g/day in 2 days. Withamelioration of the hemolytic episode, the hematocrit and theproportion of complement-sensitive cells all rose abruptly.The effect is not due to adrenal suppression since the sameeffect can be obtained with ACTH. In patients with nightlyhemolytic episodes, prednisone given in the evening abolishedthe nocturnal hemoglobinuria. Prednisone was administeredchronically (6-30 months) in doses of 15-30 mg every otherday to six patients. No response to 30 mg/day was seen in onepatient. An excellent response was obtained in five patients,all of whom maintained a normal hematocrit without trans-fusion. The proportion of markedly abnormal complement-sensitive cells (population 3) rose in each case (45-74%before, 55-94% after). Lysis of moderately abnormal (popul-tion 2) cells was also decreased in two patients. No side effectsof prednisone were observed. The mechanism of decreasedlysis was investigated in vitro. No change in the complementsensitivity of the cells or in their lysis in acidified normal serumwas seen when prednisone was administered in vivo or addedin vitro. Serum from PNHpatients or normal donors did notchange in complement content or in ability to induce lysisof PNH cells when acidified, reduced in ionic strength, ortreated with cobra venom or inulin. Prednisone may decreasein vivo lysis in PNHby altering the activation of the alternatepathway of complement activation in an unknown way.(Supported by a grant from NCI.)

70 a

258. The Ph' Chromosome: Evidence for a Specific Chromoso-mal Translocation. JANET D. ROWLEY,* Chicago, Ill. (intro-duced by Leif Sorensen).

Chromosomes from five consecutive Ph'-positive patientswith chronic myelogenous leukemia were examined withquinacrine fluorescence and the new Giemsa techniques.Except for the Ph' chromosome, chromosomes from threepatients appeared to be normal when conventional stains wereused. A change in the karyotype occurred in one of thesepatients (case 3) in blast crisis. Two other patients were firstseen in blast crisis and had chromosomal abnormalities inaddition to the Ph' chromosome. Cells of all five patientsshowed a consistent abnormality in addition to the Ph'chromosome. With quinacrine fluorescence, the abnormalityappeared as additional dull band on the terminal portion ofthe long arm of chromosome 9 (9q+). Giemsa-stained prepara-tions of the same cells showed an additional faint band at theend of the long arm of the same chromosome 9. The appearanceof this material is similar to that of the long arm of 22, sug-gesting that it may represent a translocation of the portionof 22 missing from the Ph'. Furthermore, the amount ofmaterial deleted from 22 is approximately equal to the addi-tional material on 9q+. Other findings, in addition to the9q+, noted in blast crisis were (a) the second Ph' observedin case 3 was not identical to the single Ph' seen before blastcrisis; (b) additional C group chromosomes were present inthree patients and, in two patients, appeared to resemblechromosome 8; and (c) a metacentric marker chromosome,noted in the same three patients, contained at least one armthat resembled the long arm of 17. These findings illustratethe value of analyzing human cytogenetic abnormalities withthe newer techniques of chromosomal identification.

259. A Hemoglobin S-Specific Radioimmunoassay for theQuantitation of Sickle Hemoglobin in Fetal Blood. PETER T.RoWLEY,* RIcHARD A. DOHERTY,* CHERYLROSECRANS,*ANDELSA CERNICHIARI,* Rochester, N. Y. (introduced by RalphF. Jacox**).

The quantitation of hemoglobin S in fetal blood requires ananalytical method which is specific and highly sensitive. Aradioimmunoassay has been developed to take advantage ofthe specificity characteristic of immunological reactions andthe sensitivity possible with isotopic measurement. Hemo-globin S was purified by column chromatography and injectedwith complete Freund's adjuvant into goats. Each goat serumwas tested for reactivity against hemoglobins A and S byimmunodiffusion and by quantitative precipitation. Hemo-globin A reactivity was removed by cross-reaction. One serumso treated was specific for hemoglobin S and did not react withhemoglobins A or F. Hemoglobin S was labeled with 2261 bythe chloramine-T method. In the radioimmunoassay, com-plete precipitation of the antigen-antibody complex wasinsured by the addition of rabbit anti-goat gammaglobulin.This assay offers reliable and specific quantitation of as littleas 1 ng of hemoglobin S. Assuming that 10%of the hemoglobinin fetal blood at 14 wk gestation is of adult type, this assayis capable of detecting the amount of hemoglobin S in 10-vml of homozygous hemoglobin S blood. The prenatal diagnosisof sickle-cell anemia will require in addition a method fordemonstrating the absence of hemoglobin A and a methodfor obtaining fetal blood safely. (Research supported bycontract from NIH-NIHLI-71-2402B.)

260. A Therapeutic Role for Dichloromethylene Diphosphonate(Cl2MDP) and 25-Hydroxycholecalciferol (25HCC) in RenalOsteodystrophy. JEAN E. RUSSELL* AND Louis V. AvIOLI,St. Louis, Mo.

It has been well established that chronic renal disease isattended by alterations in 25HCC metabolism as well as anincreased bone resorption and defects in the maturation ofskeletal mineral and osteoid, each of which is resistant tovitamin D therapy. Diphosphonates (C12-MDP) have alsobeen shown to decrease bone resorption and stabilize itscrystalline integrity. Rats with experimentally induced renalfailure of 4 wk duration were treated with either C12MDPalone in doses of 2 mg/kg per day (DP), 25HCCat 500 IU/day(25HCC), or both compounds (DP-25HCC) for a 2 wk period.When compared to pair-fed nontreated uremic controls,plasma and tissue calcium and phosphate in all treatedgroups decreased with the greatest fall noted in the DP-25HCCgroup. Analyses of bone mineral and collagen maturation,using density gradient bromoform toluene techniques, demon-strated a reversal of the immature crystalline and collagenmoieties toward more mature forms. Changes were pronouncedin the DP group and greatest in the DP-25HCC group. Theresults strongly suggest that C12MDPand 25HCCeither aloneor in combination may prove to be therapeutic for the osteo-dystrophy and soft tissue calcification which attends chronicrenal insufficiency. (Research supported by NIH contract70-22 19.)

261. Synergistic Cytotoxicity of Bacterial Proteins Producedby Different Species. L. D. SABATH, S. J. WALLACE,* V.LORIAN,* ANDC. WILCOX,* Boston, Mass., and Bronx, N. Y.

The concept of two or more species of bacteria interactingto cause disease has been suggested, but details for a mech-anism of such synergistic pathogenicity have been lacking.A diffusible product from Staphylococcus aureus, the "fl-hemolysin" that by itself does not effect human red cells butdoes cause "incomplete lysis" of sheep erythrocytes, maycause synergistic hemolysis of human and sheep erythrocytesif it acts with diffusible substances from certain strains ofStaphylococcus epidermidis, Diplococcus pneumoniae, Strepto-coccus agalactica, Herellea sp., and diptheroids. The clearhemolysis produced on erythrocytes in agar is a result of lysisif the membrane and diffusion away of unchanged hemo-globin. The material produced by S. epidermidis has mol wt280,000, and it is neither the epsilon hemolysin nor a lipase(E.C. 3.1.1.3)-substances previously suggested; dialysisagainst tap water of pH 7 phosphate buffer causes loss ofhemolytic activity that can be restored by adding cysteine.Sheep erythrocytes treated sequentially with the S. aureusprotein and the S. epidermidis protein are more readily lysedthan when exposed in the reverse sequence. These resultsdemonstrate that identifiable proteins from different bacteriamay act together to damage specific cells in a way not possiblewith either alone. The possibility should be investigated thatapparent variations in host susceptibility or "changes" inpathogenicity of some organisms might be due to the simul-taneous presence or absence of apparent "nonpathogens"which are the source of factors necessary for the cytotoxicityusually attributed to the "pathogens." (Research supported bygrants from NIH.)

262. Immunological Cross-Reactions among Enterotoxins ofHumanEscherichia coli. R. BRADLEYSACK,* Baltimore, Md.(introduced by Douglas Carroll**).

Within the last few years enterotoxigenic E. coli have beenrecognized as distinct etiological agents in human diarrhealdisease. Previous studies from this laboratory have indicatedconsiderable immunological cross reactions between theenterotoxins of E. coli and V. cholerae. These studies now havebeen expanded to include additional serotypes of enterotoxi-genic E. coli from India and one strain isolated from a patientwith diarrhea in the United States. Heat-labile (100'C, 15

71 a

min) enterotoxins of six isolates of E. coli have been examinedfor immunological similarities. A single lot of antiserum wasprepared in rabbits by hyperimmunization with an entero-toxin preparation of E. coli 078:H12 (strain 408-3). Entero-toxin preparations consisted of dialyzed, lyophilized culturefiltrates of organisms grown in stationary shallow syncasecultures for 48 h. Twofold dilutions of antisera were incubatedwith 3 ED50 of enterotoxin for 1 h at 370C, and the residualenterotoxin activity assayed in 18-h rabbit ileal loops. 1 Uof antitoxin was defined as the reciprocal of the amount ofserum required to neutralize 1 ED50 of enterotoxin. Preim-mune rabbit sera had titers of < 10 U/ml for each of theenterotoxins studied. Enterotoxins from five different sero-types of E. coli from India (078:H11, 078:H12, 085:H7,0126:H12, 015:H11) were neutralized by the single hyper-immune sera at titers of 450-2500 U/mil. Enterotoxin fromthe E. coli from Whiteriver, Ariz. (serotype not yet deter-mined; not of "classical" enteropathogenic serotype), wasneutralized at a titer of 1505 U/ml. These studies suggestthat human E. coli of different serotypes and geographicallocations produce enterotoxins which are immunologicallysimilar, thus suggesting the feasibility of developing immuno-logical techniques to identify both the organisms and specificantitoxin responses in patients. (Supported by NIH grant 7R22 A111358-01 and NIH contract 71-2260.)

263. Quantitative Health Measurements in Randomized Trialof Nurse-Practitioners. DAVID L. SACKETT,* WALTER 0.SPITZER,* ROBIN S. ROBERTS,* MICHAEL GENT,* JOHN C.SIBLEY,* AND ANTHONY OLYNICH,* Hamilton, Ontario,Canada (introduced by Alvan R. Feinstein**).

Wehave now completed final quantitative analysis of endresults for patients in a randomized clinical trial of theefficacy and safety of nurse-practitioners in providing pri-mary medical care. In two nonuniversity family medicalpractices, 1500 families were randomly allocated to eithera control group, where they continued to receive care fromfamily physicians, or to an experimental group in which theirprimary medical care was provided by nurse-practitionerswho had completed a special training program in clinicaljudgment and in the evaluation and treatment of commondisorders. The experimental and control groups were initiallyidentical with respect to three measures of physical healthand satisfaction with previous medical care. During the 1 yrexperimental period, the drop-out rates among patientsrandomized to the family physicians and to the nurse-practi-tioners were 1.1% and 1.7%, respectively, and the mortalityrates in the two groups were 0.8% and 0.3%. In measurementsperformed at the end of the trial in the control vs. experimentalgroups, the following respective values were obtained: forunimpaired physical function, 88% vs. 86%; for ability toexecute usual daily activities, 90% vs. 90%; and for freedomfrom bed disability in the prior 14 days, 87% vs. 86%. For12 measures of positive social function, the average percentagesof maximum possible function were 76% vs. 75%, and fornine measures of positive emotional function, 74% vs. 73%.These results, together with concomitant studies of the qualityof clinical care delivered, indicate that a nurse-practitionercan be both efficacious and safe as a provider of primarymedical care. (Supported by research grants from the Prov-ince of Ontario and the Government of Canada.)

264. Fletcher Trait: Defects in Coagulation, Fibrinolysis,Permeability Enhancement, and Kinin Generation. HIDE-HIKO SAITO,* VIRGINIA H. DONALDSON,** AND OSCAR D.RATNOFF,** Cleveland and Cincinnati, Ohio.

Wuepper reported that the defective assays for coagulation,permeability enhancement (PF/Dil), and kinin generation

in Fletcher trait plasma were corrected by a plasma pre-kallikrein. All of these functions are dependent upon activa-tion of Hageman factor (HF), but Fletcher trait plasma hasnormal amounts of HF by immunologic and functional tests.Defective clotting, fibrinolysis, and PF/Dil generation inFletcher trait were corrected by small amounts of normal,HF-deficient, or PTA-deficient plasmas, and were partiallycorrected by partially purified HF, activated by kaolin orellagic acid. These agents were less effective in correctingdefective esterolysis and did not correct defective kiningeneration. HF fragments (mol wt approximately 30,OOOD),prepared by digesting partially purified HF with insolubletrypsin, induced kinin formation in normal, HF-deficient, orPTA-deficient plasmas, but not in Fletcher trait or 61'C-heated normal plasma. Partially purified plasma kallikreinrepaired the clotting and fibrinolytic defects in Fletcher traitplasma. Normal esterolytic and kinin-generating activitieswere eluted from Celite with which Fletcher trait plasmahad been adsorbed. All these experiments suggest that Fletchertrait plasma lacks the capacity to activate HF at a normalrate and there is the apparent deficiency of a plasma pre-kallikrein, but no single hypothesis can yet explain these data.Although Fletcher trait plasma appears deficient in a pre-kallikrein, a plasma prekallikrein may be present in a non-functional form, not activable by HF but activated by ex-posure to Celite. (This study was supported in part by grantsHL 11933 and HL 01661 from the USPHSand by grants fromthe AHA.)

265. Thyroid Hormone Action: Demonstration of NuclearReceptors and Transcriptional Control in Cell Culture.HERBERTH. SAMUELS* AND JIR TsAI,* New York (intro-duced by Saul J. Farber**).

We have previously demonstrated that triiodothyronine(T3) and thyroxine (T4) induce a 3-fold increase in the rate ofgrowth of GH1 cells in culture. Within 3 h of incubation ofcells with T3, nuclear template activity, as assayed with E.coli. RNApolymerase, increases 2-fold. To study further theaction of T3 and T4, we examined the binding of purified[l261] T4 and [125I] T3 to cellular fractions after 2 h of incuba-tion with intact cells in serum-free media. High-affinity, low-capacity, binding sites for T3 and T4 were demonstrated innuclear but not in mitochondrial or cytosol fractions. Chro-matographic analysis of the bound nuclear radioactivity fromcells incubated with [E2r1] T4 demonstrated 97% T4, 2%iodide, and 1% T3. Dissociation constants determined byScatchard analysis were 2.9 X 10-11 Mfor T3 and 2.6 X 10-10M fcr T4. The maximal binding capacity was identical forT3 and T4 with approximately 5000 molecules bound per cellnucleus. [1261] T4 binding was competitively inhibited by T3.These data suggest that T3 and T4 interact with identicalnuclear receptors, and that conversion of T4 to T3 may notbe a prerequisite for biologic activity. Similar high-affinity,low-capacity, nuclear binding sites were also demonstrated byincubation of [1261] T3 or [1261] T4 directly with isolatednuclei. Incubation of cells with increasing concentrations ofnonradioactive T3 resulted in a subsequent increase in bindingwhen [12561] T3 was then incubated directly with isolatednuclei. This suggests that the number of nuclear receptorsare not fiked, but increase after exposure of intact cells withhormone. This increase in nuclear receptor content may resultfrom the transfer of an unstable cytosol receptor to the nucleus.(Research supported by grants 1 K04 AM 46546-01 fromNIH and P-595 from ACS.)

266. Antibiotic Therapy in Experimental StaphylococcalEndocarditis. MERLEA. SANDE* AND MARKL. JOHNSON,*Charlottesville, Va. (introduced by Edward W. Hook**).

72 a

This study was undertaken to investigate factors influencingkilling of staphylococci by antibiotics in experimental endo-carditis. Rate of bactericidal activity was determined for eachantibiotic by standard techniques in broth using a strain ofStaphylococcus aureus (SA). Endocarditis was produced in 87rabbits by intravenous injection of 106 SA after 5 days ofaortic valve trauma (polyethylene catheter transarteriallyinto left ventricle). Within 24-36 h after challenge, cardiacvegetations containing 109-'" SA/g were present and animalshad fever and constant bacteremia (103-4 SA/ml). Anti-biotics were started 24 h after infection in doses that achievedsustained serum bactericidal levels of > 1:8. Bacteria invegetations were enumerated after therapy for 1-3 days(period I), 4-6 days (II), and 7-12 days (III). Penicillineradicated SA from vegetations in 1 wk (period I, geometricmean CLogio] 8.38 viable units SA/g vegetation; II, 2.86;III, 0). The combination of gentamicin and penicillin wassynergistic in vitro and eradicated bacteria from vegetationsmore rapidly (I, 2.56; II, 0; III, 0) than penicillin alone.Vancomycin was also rapidly bactericidal in broth and steril-ized vegetations more rapidly (I, 2.53; II, 0; III, 0) thanpenicillin. Rifampin was less effective than penicillin or van-comycin in vitro and in vivo (I, 7.61; II, 3.96; III, 4.32). SAhighly resistant to rifampin emerged in 3 of 14 rabbits treatedwith rifampin. Rifampin antagonized the bactericidal actionof penicillin in vitro and in vivo (I, 5.36; II, 3.15; III, 2.21).Clindamycin had a slow rate of bactericidal activity in brothand failed to consistently eliminate SA from vegetations within12 days (I, 8.27; II, 6.17; III, 2.32). The capacity of anantibiotic or combination of antibiotics to kill SA in brothseems to correlate well with ability to eradicate SA fromcardiac vegetations in experimental endocarditis.

267. Neutral Protease and Histonase Activity in HumanArticular Cartilage. ASHER I. SAPOLSKY,* J. FREDERICKWOESSNER,*AND DAVID S. HOWELL,** Miami, Fla.

Recently we reported increased activity of cathepsin Din osteoarthritic (compared to control normal) cartilage withcathepsin D, the predominant proteoglycan digesting pro-tease. The current report is, to the authors' knowledge, thefirst to show neutral protease activity in human cartilage notdue to cathepsin D or B. In addition the activity of thecathepsin D on proteoglycan was limited to the acid pHrange. Evidence for these statements is as follows. Humancartilage enzyme extracts prepared from postmortem patellae,as well as pprified cathepsin D from bovine uterus (Di),did not digest hemoglobin but degraded proteoglycan subunit(PGS) considerably at neutral pH. Pepstatin and 1-2-epoxy-3-phenoxypropane, new powerful inhibitors of cathepsin D,scarcely affected this degradation at neutral pH but inhibitedalmost all PGSdigestion at pH 5 and all hemoglobin digestionat pH 3.1 Moreover, highly purified bovine cathepsin D(D2) did not degrade PGSat pH 7. The human extracts andDi also degraded histone and casein at neutral pH, but Dadid not. Pepstatin did not inhibit the neutral histonase andcaseinase activity but completely inhibited their action atpH 5. This neutral activity was inhibited by human serumand chloroquine (20 mM). However, it was neither activatedby cysteine (10 mM) nor inhibited by diisopropylfluorophos-phate (10 mM), e-aminocaproic acid (50 mM), or trypsininhibitor (250 pg/ml); these findings differentiate it fromcathepsin B and neutral proteases derived from leukocytes.Thus, the current human articular cartilage contained uniqueneutral protease(s) capable of degrading the cartilage matrixproteoglycans at neutral pH and also histonase activity whichmight play a role in the chondrocytic proliferation believedto occur in advancing osteoarthritis. (This work was sup-

ported by grants AM-08662, AM-05038, and HE-11035 fromthe NIH and by USVA, Part I Research Program.)

268. Isotonic Volume Reabsorption and Transepithelial Po-tentials in Proximal Straight Tubules. JAMES A. SCHAFER*ANDTHOMASE. ANDREOLI, Birmingham, Ala.

These experiments were designed to evaluate factors affect-ing isotonic volume reabsorption (Jr, nI min-' mm-') andtransepithelial electrical potential difference (V., mV) inproximal straight tubules isolated from rabbit kidney. Theperfusing solution contained: 105 mMNaCl, 25 mMNaHCOa,10 mMNa acetate, 5 mMKCl, 4 mMNaH2PO4, 1.8 mMCaC12, 1.0 mMMgSO4, 8.3 mMglucose, and 5 mMalanine;the bath contained rabbit serum bubbled with 5%CO2-95%02. At 370C, J, was 0.5440.17 (SD; 27 tubules) and V.(lumen relative to bath) was - 1.5±0.8 (31). Both V. and Jywere independent of perfusion rate (2.5-50 nl min-') andperfusion pressure (5-30 cm H20). At 220C, J, fell 0.4640.17(6) to 0.04, and Vm rose + 1.7±0.6 (5) to + 0.1. Similarly,ouabain (1074 M) reduced J, by 0.52i0.08 (4) to - 0.02,and Vm rose + 1.7±1.0 (5) to + 0.1. These results agreequalitatively with those of Burg and Orloff (1968. J. Clin.Invest. 47: 2016; 1970. Am. J. Physol. 219: 1714) for proximalconvoluted tubules. Significantly, when the perfusion solutionwas altered by replacing NaHCO3and Na acetate with 17 mMNa2SO4 and glucose and alanine with 15 mMurea, J, wasreduced 30% and V. fell to - 2.6±1.0 (5). Weassume thatthese relatively small changes in V. and J, were referable tothe poorly reabsorbed sulfate counterion. Taken together,these data support the view that isotonic volume reabsorp-tion in the proximal straight tubule depends primarily onactive transport of Na+ from lumen to peritubular medium.(Supported by AHA Established Investigator Award, NIHRCDA, and research grants from the NIH, NSF, and AHA.)

269. Hyporeninemic Hypoaldosteronism. MORRIS SCHAMBE-LAN,* NORMABRUST,* AND EDWARDG. BIGLIERI,** SanFrancisco, Calif.

Previous reports from this laboratory suggested that isolatedhypoaldosteronism in adults may be secondary to renindeficiency. Additional experience indicates that this syndromeis common (20 patients recognized) and lends further supportto a primary hyporeninemic mechanism. Detailed studieshave been completed in 13 patients (age 41-81 yr, mean 64).All had hyperkalemia (5.9-6.9 mEq/liter) that was correctedby mineralocorticoid replacement. Creatinine clearance rangedfrom 18 to 72 ml/min, (mean 42). Plasma renin activity(PRA) (angiotensin I generation) and plasma aldosterone(PA) were determined by radioimmunoassay and urinarysteroids, by double-isotope dilution techniques. Generalizedadrenal insufficiency was excluded by demonstration of normalexcretion of tetrahydrodeoxycorticosterone, tetrahydrocorti-costerone, and 17-hydroxycorticoids, which increased normallyduring 3 days of ACTH. Urinary aldosterone was subnormal(3.0±0.7 (SE) /Ag/24 h, N 4-17) and failed to increasenormally during salt restriction (6.7±i1.4, N 20-67) or ACTH(6.5i1.0). Supine PA was subnormal (4.7±0.9 ng/100 ml,N 7.8± 1.6) and failed to increase normally with uprightposture (4.4± 1.0, N 22.0±3.4) or salt restriction (supine7.2±1.5, N 18.845.8; upright 12.144.5, N 50.4±8.6) inthe six patients so studied. Supine PRAwas subnormal (1.7±0.3 ng/ml per 3 h, N 5.5±0.9) and failed to increasenormally with upright posture (2.5±0.4, N 13.9±3.2), saltrestriction (supine 3.9±t0.7, N 12.1±1.2; upright 6.3± 1.3, N25.3±4.8) or intravenous furosemide, 20 mg (7.8±2.4). Ingeneral, PAcorrelated with PRAunder these study conditions.Four patients with Addison's disease on glucocorticoid re-placement had similar degrees of hyperkalemia and hypo-

73 a

aldosteronism but markedly elevated PRA (mean supine 82,right 124, salt restriction 118 ng/ml per 3 h). Thus isolated,hypoaldosteronism, in contrast to generalized adrenal in-sufficiency, is definitely associated with hyporeninemia andis a frequent cause of otherwise unexplained hyperkalemia.(Research supported by grants AM-06415 and HL-11046from NIH).

270. Parallel Discharge of Secretary Proteins from theExocrine Pancreas of the Guinea Pig. GEORGEA. SCHEELE*ANDALAN M. TARTAKOFF,* NewYork (introduced by JamesG. Hirsch**).

Weused an in vitro system of pancreatic lobules to studythe discharge of proteins from exocrine cells. Lobules remainedmorphologically intact for at least 5 h by light and electronmicroscopy, displayed linear incorporation of [3H] leucinethrough the period tested (3 h), and showed low levels ofresting discharge (- 5%/h). Stimulation with optimal dosesof carbachol (10-' M), pancreozymin (7 X 10-' M), andcerulein (10-9 M) caused discharge of 60 90% of secretoryproteins in 2 h and 5 h, respectively. 75 mMKCl (±i atropine)caused a 45% discharge of secretory protein within 2 h.Enzyme activity was used to monitor the kinetics of dis-charge of amylase, lipase, RNase, and the zymogens chymo-trypsinogen and procarboxypeptidase B (after activation withtrypsin). In addition, SDS-gel electrophoresis allowed analysisof secretory protein independent of enzyme activity. Wehaverecently reported the identification of the majority of gelbands with known secretory enzymes or zymogens. Maximalsensitivity was achieved by prior mass-labeling of the secre-tory proteins with mixed 14C-labeled amino acids. Rest-ing secretion and secretion induced by each stimulant(carbachol, pancreozymin, cerulein, and 75 mMKCl) col-lected over 2 h showed identical proportions of secretoryprotein by enzyme activity measurements and by measure-ment of radioactivity in gel bands. Furthermore, a detailedanalysis of the time course of carbachol-induced secretion,collected sequentially in 15-min periods over 3 h, showedconstant relative proportions of secretory protein by bothmethods studied. The evidence suggests that all secretoryproteins in the guinea pig pancreas are discharged in parallel.(Supported in part by USPHSand NIH Special FellowshipAM49448.)

271. Cooperative Multicenter Evaluation of Intra-AorticBalloon Pumping in Cardiogenic Shock. S. SCHEIDT,* G.WILNER,* H. MUELLER,* D. SUMMERS,* M. LESCH,* G.WOLFF,* J. KRAKAUER,* M. RUBENFIRE,* P. FLEMING,* G.NOON,* T. KILLIP, ANDA. KANTROWITZ,* Detroit, Mich., andNew York.

87 patients with rigidly defined cardiogenic shock refractoryto standard medical therapy were treated with intra-aorticballoon pumping (IABP) in 10 institutions under a commonprotocol. Favorable physiologic effects were observed in mostpatients. After 4 h of IABP, aortic systolic pressure ("after-load") fell from 76±22 to 57±17 mmHg (P < 0.001), peakaortic diastolic pressure rose from 53412 to 83±19 mmHg(P < 0.001), heart rate fell from 110424 to 103±21 (NS),cardiac output increased from 2.4 to 2.9 liters/min, urine out-put increased from 11±21 to 37434 ml/h (P<0.001), andmyocardial lactate extraction improved in all 19 patientsstudied. During IABP, oliguria was reversed in 70% ofpatients, acidemia in 69%, and need for pressor agents in64%, but abnormal mental status was improved in 44% andpulmonary congestion in only 34%. In spite of favorableclinical and physiologic responses, 52 patients died duringIABP. Of the 35 in whom IABP could be discontinued, 15were discharged from the hospital and 8 survived 1 yr. Neither

initial hemodynamic measurements nor responses to IABPwere useful in predicting the final outcome. Several patientssurvived after unsuccessful attempts to terminate IABP, butnone survived hospitalization if > 48 h of IABP was required.In a group of 22 patients with shock studied at one participat-ing institution, myocardial damage averaged 50.2% of leftventricular mass and could not be related to the duration ofshock or of IABP. It is concluded that refractory cardiogenicshock is associated with massive myocardial necrosis. AlthoughIABP is physiologically effective, few patients (18% in thisstudy) will survive shock, despite IABP. (Supported by TheJohn A. Hartford Foundation.)

272. Cycling Characteristics of Human Peripheral BloodLymphocytes in Culture. LEWIS SCHIFFER,* ALAN WINKEL-STEIN,* ARNOLDMARKOE,* ANDJANET NELSON,* Pittsburgh,Pa. (introduced by Jessica Lewis**).

By means of a new in vitro technique of demonstrating DNApolymerase in individual cells, it may now be possible toestimate the fraction of lymphocytes in phytohemagglutinin-stimulated peripheral blood cultures that are in cycle at anygiven time. The method entails coating unfixed lymphocytesmears with soft agar, with and without DNase-activated,sonicated DNA, and incubation of these slides with all fournucleotide triphosphates-one tritiated-at 370C in the pres-ence of Mg++. Slides are processed further by fixation,autoradiography, and staining. Insoluble, tritiated DNA isrepresented by exposed silver grains over the cells containingDNA polymerase. Unlike human cells in vivo, culturedlymphocytes have a short GL period once induced into pro-liferation and can thus be expected to maintain measurableamounts of DNApolymerase during 3 days of growth. Ourresults indicate that DNA polymerase-containing cells arepresent at least 10 h before there is significant DNAsynthesistaking place. Thereafter, the percentage of cells containingDNA polymerase assayed with exogeneous DNA primerincreases at 20%/day to reach 90% at 4 days. When theonly primer is the cellular DNA, the labeled cells increase at15%/day, starting from the same point, to reach 50% in 4days. If one assumes that the latter assay is equivalent to thegrowth fraction, as we have demonstrated in three animaltumor systems with short GI phases, the cell cycle time ofthese lymphocytes can be estimated at 33, 17, 14, and 23 hat 1, 2, 3, and 4 days, respectively. It may also be theorizedthat as few as 10-20% of the original, cultured lymphocytesmay actually enter into proliferation. (Research supported bygrants from the NIH.)

273. Mechanisms of Hypotension Induced by Quinidine andProcaine Amide. P. G. SCHMID,* F. M. ABBOUD,A. L. MARK,*ANDD. D. HEISTAD,* Iowa City, Iowa.

The antiarrhythmic drugs, quinidine (Q) and procaineamide (PCA), may cause significant hypotension. Effects onneurogenic transmission, adrenergic receptors, and vascularsmooth muscle might explain the hypotension, but the specificmechanisms involved are not known. In humans, norepine-phrine (NE) (0.075 and 0.15 ,ug/kg per min intravenously)produced increases in mean arterial pressure averaging14.5±2.2 (SE) and 21.8±2.1 mmHg before Q (200 mg baseintravenously) and 9.8±41.7 and 15.9±2.2 mmHg after Q(P < 0.05). Corresponding increases in forearm vascularresistance (plethysmograph) were 5.9±2.2 and 9.2±2.9 Ubefore Q and 2.4±-1.4 and 2.5±2.4 units after Q (P < 0.05).In the perfused gracilis muscle and hindpaw of the dogs, Qproduced marked, dose-related inhibition of vasoconstrictorresponses to postganglionic nerve stimulation (NS) and tointra-arterial NE, but not angiotensin and serotonin. PCA(20 mg/kg intravenously) inhibited reflex vasoconstrictor

74 a

responses to bilateral carotid artery occlusion and stimulationof carotid chemoreceptors with nicotine by factors of 25 and50, respectively; responses to preganglionic NS also werereduced (P < 0.05), but only by a factor of 2. Responses topostganglionic NS were not reduced. The results suggest thatQand PCAhave negligible direct relaxant effects on vascularsmooth muscle since intra-arterial infusions of the drugsdid not cause vasodilatation in denervated muscle and paw.Q inhibits stimuli which activate a-adrenergic receptors. PCAinhibits responses to reflex stimuli more than preganglionicNS, suggesting a predominant effect on vasomotor centers orafferent reflex pathways in addition to a ganglionic blockingaction.

274. The Mechanism of Renal Excretion of Potassium inUremia. DONALDA. SCHON,* PATRICIO SILVA,* ANDJOHN P.HAYSLETT,* New Haven, Conn. (introduced by L. R. Freed-man).

An increase in the fractional excretion of K occurs in uremicsubjects fed a normal diet rich in K. Recent studies in thislaboratory demonstrated that the augmented excretion of Kin normal rats on a high K diet (2.60 mEq/g) was associatedwith an increase in Na-K-ATPase in both cortex and outermedulla. In order to study the role of Na-K-ATPase in Kexcretion in uremia, normal and uremic animals (75% renalablation) were pair fed a normal diet (0.13 mEq/g K) for14-21 days, when fractional K excretion (FEK) during controlperiods and after the acute infusion of 0.5 M KCl was ex-amined. While serum K levels were similar in both groups,FEK during control periods was 6.7140.66% (mean-SE)in normal and 21.545.79 in uremic rats (P < 0.05). 60-90min after beginning KCl infusion, FEs rose to 27.94d5.22%in normal and 53.7145.01 in uremic animals. In contrast,control FEK in high K rats was 79.74424.52% and did notincrease further with KCl infusion. There was no differencein maximal FEK in uremic rats compared to high K rats. Thisincreased capacity to excrete K in uremic animals was as-sociated with elevated levels of Na-K-ATPase (cortex/medulla: 12.5141.2/33.3340.89 jsm Pi/mg protein per hr,normal; 16.8140.89/40.2342.32, uremic P < 0.05). Lower-ing control FEK in uremic rats to values present in normalrats, by decreasing dietary K (0.07 mEq/g), abolished boththe increased levels of Na-K-ATPase and the augmentedresponse to KCl loading. These data indicate that Na-K-ATPase plays an important role in the renal mechanism of Kexcretion in uremia to maintain K balance despite reducedGFR. (Supported by NIH and PHAgrants.)

275. Effects of Dietary Carbohydrate Exchange for Fat onCholesterol Balance in Man. PAUL H. SCHREIBMAN*AND E.H. AHRENS, JR.,** NewYork.

Carbohydrate (CHO) induction of plasma triglycerides(TG) has been extensively studied; the effects on plasmacholesterol (CH) are less well defined. To determine effectsof CHO feeding on sterol metabolism, 10 hyperlipidemicpatients were studied for 4-5 months on the metabolic ward.Weights were held constant. High-fat formula feedings fur-nished 45-70% of calories as polyunsaturated fat (PUF);high-CHO formulas contained 0-20% PUF, with isocaloricsubstitution of dextrose for PUF. Sterol balance studies werecarried out simultaneously with isotope kinetic analyses afterintravenous administration of [4-14C] CH. Since all feedingswere CH-free, CH absorption was not measured. In 7/10patients on the high-CHO diet, plasma TGrose (mean 147%)and also plasma CH'(45%). Total fecal steroid excretion de-creased'in four (-39%), was not significantly changed in two,and increased in one. Plasma CHspecific activity decay curvesflattened markedly in all seven. In the other three patients

plasma TG rose (163%), but plasma CH was unaffected intwo and fell in one. Sterol excretion increased in all (+39%);plasma decay curves were unchanged. These two kinds ofresponses were unrelated to differences in type or severity ofhyperlipidemia, body weight, age, or sex. Specific activity ofplasma CH and fecal neutral steroids remained identical inboth groups on both diets. Interpretation: In all patients, TGand CH synthesis increased on high-CHO diets with a con-comitant increase in isotopic CH exchange between plasmaand slowly-turning-over pools. In one group, the incrementin newly synthesized CH was excreted into the intestinal lu-men. In the other, CH retention in the plasma compartmentwas followed, in the new steady state, by feedback inhibitionof synthesis. (Research was supported by grants from NIH-NHLI and GCRC.)

276. Interaction of Calcium, Magnesium, and AdenosineTriphosphate (ATP) in Human Erythrocyte Ghost Endocy-tosis. STANLEYL. SCHRIER, MURIEL SEEGER,* IRENE JUNGA,*ANDKLAUS BENSCH,* Stanford, Calif.

Our purpose was to explore the nature of the interaction ofcalcium (Ca), magnesium (Mg), and adenosine triphosphate(ATP) at the interior of the ghost membrane as it undergoesendocytosis and vacuole formation. Endocytic vacuoles areinduced in ghosts by Mg ATP and blocked by 2.5 mMCa.The proposal that Mg ATP energizes the membrane to formvacuoles by displacing membrane-bound calcium was tested.Fresh ghosts were resealed with 2.5 mMATP and 1.0 or 2.5mMCa labeled with 4"Ca. After incubation, ghost ATP wasmeasured and ghost Ca was determined both by isotopicmeasurements and atomic absorption. In parallel tubes condi-tions were identical except that vacuoles were induced byaddition of 2.5 mMMg. The results were as follows. (a) Caaddition reduced ghost mean corpuscular volume perhaps byactivating an actomyosin. Electron microscopy confirmed thatthe Ca induced ghost contraction. Mg antagonized this Caeffect and produced vacuoles. (b) Mg addition reduced theamount of ghost ATP; the ghost preparation containing mostCa had least ATP. (c) Mg sharply reduced the amount ofcalcium resealed with ghosts between 3- and 10-fold. Therefore,it appeared that vacuole formation involved displacement ofCa from the ghost interior by Mg ATP. AT32P was resealedwithin ghosts and incubated for 10 min in order to discoverwhy less ATP was recoverable in ghosts containing both Caand Mg. AT32P did not leak from ghosts. Ca alone producedno loss of AT32P, but upon addition of Mg, as well as Ca, therewas loss of AT32P and production of 12Pi. The findings pointto the activation of a Ca Mg ATPase which could displaceCa from membranes and which may have a mechanical rolein endocytosis. (Work supported by grant RI AM 13682from the NIH.)

277. The Antiarrhythmic Effectiveness of IntramuscularLidocaine; Influence of Different Injection Sites. MORTIMERL. SCHWARTZ,* VIRENDER SETHI,* AND RAVINDER M.NARANG,* New York (introduced by Edward E. Fischel**).

To determine the effectiveness of intramuscularly admin-istered lidocaine on ventricular premature contractions(VPC's), 18 patients with at least 5%VPC's received injec-tions randomly in either the deltoid (nine patients) or vastusexternus muscle (nine patients). 10% lidocaine was admin-istered as 4.5 mg/kg. Statistically significant reductions from20 VPC's/min to 5 VPC's/min occurred within 15 min afterthe deltoid injections which were associated with concomitantrise in plasma lidocaine levels (3.5 ,g/ml) which remained at1.5 ,g/ml or above for 90 min. At this time the frequency ofVPC's was still significantly lower than control (8 VPC's/min). After injection into the vastus, no sustained reduction

75 a

in VPC's resulted. This correlated with generally lower lido-caine levels (less than 1.5 Mg/ml). Lidocaine levels above1.5 pug/ml were effective in causing a reduction of VPC's.Intramuscular injections of lidocaine into the deltoid causeda prompt increase in blood levels and prompt reduction ofVPC's. Obvious differences in results occurred when twodifferent muscle masses were used. If lidocaine cannot beadministered by vein, our results suggest that the intra-muscular injection of 4.5 mg/kg (deltoid) may be used toproduce a significant antiarrhythmic effect. (Research sup-ported by Astra Pharmaceutical Company grant.)

278. Reduction of Infarct Size in Hypertensive Patients withAcute Myocardial Infarction. WILLIAM E. SHELL,* ALM A.EHSANI,* ANDBURTONE. SOBEL, La Jolla, Calif.

Prognosis after acute myocardial infarction (AMI) is af-fected significantly by the extent of myocardial damage. Ac-cordingly, this study was designed to reduce infarct size bydecreasing systemic arterial blood pressure (BP) and myo-cardial oxygen requirements by administration of trimetha-phan. 11 patients with AMI, hypertensive on admission (BP> 150/90 mmHg), were treated after infarct size had beenpredicted by analysis of early (7 h) serum creatine phospho-kinase (CPK) changes with a new nonlinear curve fittingtechnique. In each patient, infarct size predicted before tri-methaphan was compared to the extent of the completed in-farct determined from serum CPKchanges after reduction ofBP. Intraarterial BP, pulmonary artery wedge pressure(PAW), cardiac output (CO) (determined by dye dilution),and ejection fraction (EF) (determined by isotope angiocardi-ography) were measured serially. In 30 consecutive controlpatients with AMI, completed and predicted infarct size cor-related closely (r = 0.93, mean difference < 5%). In treatedpatients, completed infarct size was consistently less than thatpredicted (predicted = 84±20 [SE] CPK-g-eq; completed= 61 ±19, p < 0.001; mean difference = 31 ±i75%). Trimeth-aphan also decreased mean BP (22% average), heart size(4%), and PAW(20%) and increased CO (12%), EF (47%),and circumferential fiber-shortening velocity (40%). Heartrate was not altered appreciably. Mortality within 30 daysafter infarction in treated patients was < 50% of that in con-trols with corresponding infarct size. Thus, reduction of bloodpressure in patients with AMI and hypertension results notonly in improved ventricular performance but also in consider-able salvage of myocardium. (This research was supported byNIH MIRU contract PH-43-68NHLI-1332.)

279. The Role of Cellular Adenosine Triphosphate (ATP) inRegulating Erythrocyte Lysophosphatide Concentration, Ery-throcyte Shape, and Erythrocyte Membrane Stability.STEPHEN B. SHOHETAND JAMES E. HALEY, San Francisco,Calif.

To evaluate the effect of adenosine triphosphate ATP deple-tion on cell membrane renewal, erythrocytes were incubated inautologous serum at low hematocrit without added glucose for24 h. Cell and plasma levels of ATP, lysophosphatidylcholine(LPC), and cell LPC-acylation activity were followed. Red

cell shape was recorded by scanning electron microscopy. Cellmembrane stability was evaluated by an in vitro system inwhich the transfer of "4C-labeled erythrocyte membranelipids to human splenic macrophage monolayers was measured.When cell ATP levels fell below 0.25 Mm/cc cells, acylationactivity failed and membrane LPC levels rose (0.12 -* 0.33Mm/cc cells). Proportional to rising membrane LPC levels,progressive echinocytic transformation of the erythrocytesoccurred. These morphologic changes could be duplicated byadding sufficient LPC to fresh, ATP-replete cells suspended inheated serum in order tQ induce similar membrane LPC levels.

However, if cell metabolism was maintained by the addition ofglucose, these changes were reversed as membrane LPC levelsfell. Echinocytic RBS's with elevated LPC levels showed anearly threefold increase in membrane instability in compari-son to normal cells, as measured by membrane lipid transferto the macrophage monolayer. This was seen in both incubatedand fresh, LPC-treated echinocytes. In the latter, it was inde-pendent of ATP content and cell cation content. Both mem-brane instability and membrane LPC content of echinocytescould be reduced by modification of the membrane z± plasmaLPC equilibrium by the addition of defatted albumin to theincubation media. Membrane instability can be induced byincreased membrane LPC levels. Such deleterious levels, inturn, can be induced by the failure of cell LPC-acylationactivity consequent to ATP depletion. (Supported by NIHgrants).

280. Radioimmunoassay for Human 0I2-Microglobulin. J.SHUSTER,* P. GOLD, S. 0. FREEDMAN,AND M. D. POULIK,*Montreal, Quebec, Canada and Royal Oak, Mich.

02-Microglobulin (BMG) is a low molecular weight protein(11,600 daltons) which was first isolated from the urine ofhumans suffering from renal tubular malfunction. Amino acidsequence analysis has shown that BMGhas extensive aminoacid sequence homology with the heavy chain of human IgG,molecules. A preliminary, semiquantitative study, employinggel precipitin techniques, suggested that serum BMGlevelswere increased in the plasma cell dyscrazias. In order to fullyevaluate this problem, a quantitative assay for BMGwas nec-essary. Therefore, a radioimmunoassay to quantitate serumBMGwas developed. BMGwas radiolabeled with 125I by thechloramine-T technique. A standard inhibition curve was es-tablished using a double antibody procedure. This assayutilizes 100 ;l of patients serum, has a low limit of sensitivityof 0.2 ng, and can be completed within 18 h. BMGlevels weremeasured in the serum of 100 individuals (controls and hospi-tal patients) who had normal renal function as indicated byBUNand serum creatinine levels. In normal individuals theserum BMGvalues ranged fron 0.86 to 1.83 jg/ml. 17 of 21patients with multiple myeloma or Waldenstr6m's macro-globulinemia had evaluated u2M values varying from 2.6 to12 M&g/ml. The sera of two of seven patients with monoclonalgamma globulin peaks associated with disorders other thanmultiple myeloma had values of 3.75 and 2.76 Mg/ml. Severalpatients with a variety of other tumors also showed elevationsin BMG. The significance of BMGelevated in these disordersrequires the study of a larger number of patients. (This workis supported by grants from the NCI of Canada and the NIH.)

281. Heterogeneity of 5'-Nucleotidase in Chronic LymphocyticLeukemia (CLL). ROBERT SILBER, DOROTHEAZUCKER-FRANKLIN, MARYROSECONKLYN,* AND JOHN LOPES,* NewYork.

In view of the importance of the cell membrane in neoplasticprocesses, the membranes of lymphocytes isolated from bloodof normal subjects and patients with chronic lymphocyticleukemia (CLL) were compared. Membrane purification wasmonitored with electron microscopy and by the assay of twomembrane marker enzymes, NADHdiaphorase and 5'-nu-cleotidase. The membrane preparation contained no detectablenuclear or mitochondrial contamination. While the ultra-structure of normal and CLL membrane preparations ap-peared identical, a marked difference was found in the 5'-nu-cleotidase level of these preparations. The average activity ofthis enzyme in lymphocyte plasma membrane-from 12 normalsubjects was 6 ,umoles/h per mg (range 3-12 Mtmoles/per mg).In membranes from seven patients with CLL, the activity waseither not detectable or less than 0.4 ,umoles/h per mg, while

76 a

the level in the membranes of the other three patients waswithin normal limits. None of the patients had receivedtherapy; the values were unchanged for any given patienttested on multiple occasions. In contrast, NADHdiaphoraseactivity was not diminished in CLL membranes, indicatingthat the lack of 5'-nucleotidase was not associated with theloss of this other membrane function. 5'-nucleotidase couldnot be detected in crude lymphocyte homogenates from thesame patients whose lymphocyte membrane activity was lack-ing, indicating that the absence of the membrane enzyme didnot stem from a loss during purification. No evidence for aninhibitor in CLL cells was obtained. The absence of 5'-nucleo-tidase in 7 out of 10 patients with CLL is the first demonstra-tion of the lack of a plasma membrane enzyme in this disorder.Since the activity is present in the cells of some patients, itappears that lymphocytes from CLL patients are heterogene-ous with respect to this biochemical marker. (Research sup-ported by NIH.)

282. A New Role for Na-K-ATPase: Enhancement of Potas-sium Excretion by the Kidney. PATRICIO SILVA,* DONALDSCHON,* JOHN HAYSLETT,* AND FRANKLIN H. EPSTEIN,**Boston, Mass., and New Haven, Conn.

Adaptation to potassium loads requires an increase in renalpotassium excretion, thought to be mediated by distal tubularsecretion and influenced by mineralocorticoids. Na-K-ATPaseappears to play a major role in this process, since in the courseof potassium adaptation, renal enzyme activity increasesmarkedly. Tripling the usual intake of potassium for 7 daysincreased the specific activity of Na-K-ATPase by 30% in theouter medulla, but not in the cortex, of normal rats. Increasingpotassium intake and excretion to 10 times normal induced afurther increase in enzyme content of both medulla and cortex.The increase in renal Na-K-ATPase activity was specific inthat it was not accompanied by changes in Mg-ATPase or 5'-nucleotidase and did not occur in brain, liver, or diaphragm.Although potassium loading accelerates aldosterone secretion,stimulation of renal Na-K-ATPase activity was not exclusivelydependent on the adrenal, since it could be elicited by potas-sium loading in adrenalectomized animals. The rise in renalNa-K-ATPase seen in compensatory renal hypertrophy in ratsis also related to an increased excretory load of potassium pernephron, and this adaptive change after partial nephrectomywas also unaffected by prior adrenalectomy. Finally, sodiumrestriction did not induce enzyme changes, though it too in-creases aldosterone secretion. Studies of Q02 and PAH ac-cumulation by kidney slices suggest that the increase in Na-K-ATPase induced by potassium loading is localized to distaltubules. Potassium is thought to enter distal tubular urinefrom renal cells down an electrochemical gradient. An increasein Na-K-ATPase in the peritubular plasma membrane of distaltubular cells would increase the intracellular pool of potassiumand therefore accelerate its passive movement into the urine,thus playing a key role in the process of potassium secretionby the kidney.

283. Heterogeneity of Parathyroid Hormone: Clinical andPhysiologic Implications. ROBERTSILVERMAN* AND ROSALYNS. YALow,* Bronx, N. Y. (introduced by Stanley Ulick**).

Four fractions of immunoreactive human parathyroid hor-mone (PTH) were demonstrable in parathyroid glandulartissue extracted by three different solvents, 20%acetone in 1 %acetic acid (AA), 8 M urea (U), or normal saline (NS), andsubjected to Sephadex G-100 gel filtration and immunoassayusing two antisera (273 and C-329). The first (I), a voidvolume peak, was detected by both antisera, as was a second(II), which eluted from Sephadex and sedimented in the ultra-centrifuge with purified bovine PTH; a third (III) eluted be-

tween [125I] growth hormone and [1251] insulin, sedimentedsimilarly to [125I] insulin, and was detected primarily by 273;a final fraction (IV), detected primarily by C-329, eluted justbefore [l25I] insulin. The AAand U extracts were similar, con-taining fraction II as their major component, while in NS ex-tracts fraction III predominated. Three fractions, having gelfiltration and immunologic characteristics similar to fractionsII, III, and IV of NS glandular extracts, were detected in theplasma of patients with primary and secondary hyperpara-thyroidism. In every plasma, the intermediate component wasthe most abundant, and the final component, the least abund-ant form. The first eluted plasma fraction disappeared fromuremic plasma during calcium infusion with a half-time ofabout 20 min and probably represented biologically activehormone; the intermediate and final fractions had turnovertimes more than 100 times longer, remained elevated duringpostparathyroidectomy hypoparathyroidism, and were pre-sumed biologically inactive. The evidence favored a glandularorigin for the fragments. A striking advantage was observedusing antiserum 273 in the preoperative diagnosis of primaryhyperparathyroidism and was presumed due to the enhancedsensitivity occasioned by its detection of biologically inactiveas well as active hormonal forms.

284. Glucose Transport by Synovial Microvasculature. PETERA. SIMKIN* ANDJOSEPHE. PIZZORNO,* Seattle, Wash. (intro-duced by Mart Mannik).

The small molecules of synovial fluid are known to be inequilibrium with those of plasma, but the kinetics of theequilibria have remained unclear. In this study, rates of tran-synovial exchange were determined by a new experimentalmodel. 30 ml of saline were injected into one knee of 25 normalsubjects and sampled periodically over 35 min. Increasing con-centrations of physiologic molecules (urea, creatinine, urate,and glucose) and decreasing concentrations of marker sub-stances (tritiated water and 14C-labeled urea, urate, glucose, orsucrose) were analyzed by the Fick diffusion equation forDA/x (D = diffusion coefficient, A = area, x = path lengthof diffusion). DA/x, here termed the synovial diffusion capac-ity (SDC), is the volume of fluid equilibrating per minute.SCDapplies equally to plasma and intrasynovial fluid and isanalogous to the physiological concept of clearance. Respectivemean SDC values in milliliters per minuteiSD were: triti-ated water 0.94±0.29, [14C] urea 0.70±0.11, urea 0.81 ±0.29,creatinine 0.48±0.13, [14C] urate 0.6540.31, urate 0.51±0.19,[14C] glucose 0.36±0.08, glucose 0.83±0.36, and ['4C] sucrose0.42 ±0.13. Except for glucose, SDC correlated well withmolecular diffusion coefficients (r = 0.897, P < 0.001) sug-gesting that transynovial exchange of these molecules occursprimarily by free diffusion. Glucose, however, enters the kneemore rapidly. Within individual studies, comparison of bidirec-tional rates showed glucose ingress to be 177±19% (SD) of itsrate of egress, whereas respective values for urea and uratewere remarkably symmetrical: 100±7% and 98-±7%. Thesefindings indicate that specific glucose transport, consistentwith facilitated diffusion, occurs across the normal humansynovium. Since the synovial lining is discontinuous, this sys-tem is presumably within the endothelial cells of the synovialmicrovasculature. (Supported by grants from NIH andArthritis Foundation.)

285. Immunoreactive Arginine Vasopressin (AVP) and Argin-ine Vasotocin (AVT) in the Fetal Pituitary of Man and Sheep.RONALDSKOWSKY*AND DELBERTA. FISHER, Long BeachCalif., and Torrance, Calif.

AVT has been tentatively identified by bioassay in the fetalpituitary of the sheep and seal. To quantify the changes inpituitary AVPand AVT during gestation, we have developed

77 a

radioimmunoassay (RIA) systems for AVP using antiserawith and without cross-reactivity to AVT. Both RIA systemsare sensitive to 0.4 1AU AVPor AVT. Measurement of samplesin both RIA systems enables determination of the AVT con-tent of the sample. Using this dual RIA we measured the AVPand AVT content of 17 human fetal pituitaries (11-19 wkgestation) and 9 fetal sheep pituitaries (109-137 days gesta-tion). In the human fetus, AVPand AVTwere first detectableat 12 wk (0.6 and 1.6 mU/mggland weight) and tended to in-crease throughout 18 wk (AVP = 2.8 mU/mg and AVT =2.9 mU/mg). In these specimens AVT exceeded AVP (AVP/AVT = 0.71). In the fetal sheep AVPand AVTwere detectedin six of the nine pituitaries, but AVPexceeded AVT (AVP/AVT = 8.78). Mean (and SEM) of AVP and AVT in thesheep pituitaries was 6.944.7 mU/mgand 0.79±0.37 mU/mg.The data indicate that immunoreactive AVT is present in thefetal pituitary of the human and the sheep and suggests thatthe pituitary content of this peptide decreases throughoutgestation, while that of AVP increases. (Supported by USPHSgrants HD-04270 and HD-06335 from the NICHHD of theNIH.)

286. Physical State of Lipids of Atherosclerotic Lesions.DONALD M. SMALL, G. GRAHAM SHIPLEY,* CARSONR.LooMIs,* AND MARTIN J. JANIAK,* Boston, Mass.

Human atherosclerotic lesions are associated with the ac-cumulation of lipids, specifically cholesterol and its esters, andalthough the histology and biochemistry of aortic intima arewell documented, little is known of the physical state of thelipids. Using calorimetric, optical, and X-ray diffraction tech-niques, we have examined the physical state (crystalline,liquid-crystalline, liquid) of aqueous mixtures of the predomi-nant lipid types found in both normal and atheroscleroticintima. Multicomponent phase diagrams corresponding to dif-ferent temperatures were constructed and indicate the com-positional limits, molecular structure, and structural variationsof individual phases. Mixtures of the major intimal phospho-lipids (lecithin, sphingomyelin, cephalin) form lamellar liquid-crystals which can incorporate cholesterol to a molar ratio 1:1.The liquid-crystalline phase can incorporate traces of choles-terol esters (CE), the excess CE separating as liquid (oil),thermotropic liquid-crystal, or crystal, depending upon theCE and the temperature. The liquid and liquid-crystallinephases of CE solubilize small amounts of cholesterol but notphospholipid. Cholesterol, exceeding that solubilized by phos-pholipid and CE phases, separates as crystalline cholesterol.When the chemical composition and physical state of lipids indifferent stages of intimal disease are compared, we concludethat the development of atherosclerotic lipid lesions progressesas follows: first, cholesterol saturates the lamellar phospholipidphase (plasma membrane); second, CE is incorporated intothis lamellar phase, exceeds its solubility, and separates as aliquid or liquid-crystalline phase (microscopic fat droplets);third, the amount of CE increases producing the "fattystreak" lesion; finally, cholesterol exceeds its solubility inphospholipid and CE phases and separates as crystals. Thus,advanced plaques consist of lipids partitioned into severalphases; membranes saturated with cholesterol and CE, liquidand/or liquid-crystalline "oil droplets" of CE saturated withcholesterol, and crystalline cholesterol. (Supported by NIH-AM11453.)

287. Assay of Gentamicin with Gentamicin Adenyl Transferase(GAT). ARNOLDL. SMITH* AND DAVID H. SMITH, Boston,Mass. (introduced by Fred Rosen).

Previous studies have emphasized the wide variations inserum concentrations of gentamicin after a given dose, and the

potential toxicity of this antibiotic. Clinical managementwould be facilitated by an assay that is rapid, specific, andaccurate and utilizes small samples. Certain R factors mediatethe synthesis of gentamicin adenyl transferase (GAT). GAThas been purified from a mutant of E. coli W677/JR 66 se-lected for high gentamicin resistance. The enzyme was assayedby trapping antibiotic-adenylate produced from radioactiveATPon phosphocellulose paper. GAThas mol wt 35,000, a pHoptimum of 8.6-9.1, an absolute requirement for sulfhydrylgroups, and is stimulated by Mg++. It utilizes certain ribonu-cleotides and 2'-dATP, and adenylates only gentamicin, kana-mycin, and tobramycin. The Km for ATP is 61 liM; that fortobramycin, gentamicin, and kanamycin is 0.85, 1.54, and2.25 1AM, respectively. GAT is inhibited by high concentra-tions of these antibiotics and pyrophosphate, the other pro-duct of the reaction. This assay is not affected by anticoagu-lants, fluoride, urea, bilirubin, albumin, or other antibiotics;it can be applied to all body fluids and requires specimens of0.01 ml and 1 h to complete. With mock unknowns in 59 sera,half from patients with abnormal renal and/or hepatic func-tion, the arithmetic correlation coefficient (r) between the ob-served and theoretical gentamicin concentrations was 0.988;r = 0.945 for an 18 hr microbiological assay. With unknownsin 19 urines, r = 0.930 for both the enzymatic and microbio-logical assays. For 63 clinical specimens, the r between enzy-matic and microbiological methods was 0.956. These observa-tions suggest that the GATmethod meets the criteria for anideal antibiotic assay.

288. Phenotypic Distinction of R Factors Associated withEpidemic Bacilli. DAVID H. SMITH AND TERRY MARSH,*Boston, Mass.

A strain of Shigella dysenteriae is causing a pandemic inCentral America and a strain of Salmonella typhosa is epidemicin Mexico. Genetically similar, drug-sensitive isolates of eachstrain existed in the areas before the respective epidemics;both epidemic strains contain an R factor mediating resistanceto identical antibiotics. These and other epidemiological find-ings led Gangarosa et al. to propose that these R factors maybe genetically related and/or contain a locus-mediating "viru-lence." These questions and the possible use of phenotypicanalysis to distinguish R factors were studied with a strain ofE. coli infected with each of 12 coded R factors received fromGangarosa. The R factors could not be distinguished by pat-terns of drug resistance. None of the R factors restricted phagesT7, X, or P1, or mediated susceptibility to phage f2. On thebasis of 11 other tests, the R factors were categorized into fivegroups. When the codes were broken, it was found, for ex-ample, that all S. typhosa R factors mediated low conjugalfertility and low specific activities of chloramphenicol acetyltransferase (CAT) and possessed a CAT locus that was stablein S. typhimurium; all those from Sh. dysenteriae had identicalphenotypes which differed significantly in these properties.Both of these groups of R factors were fi- and incompatiblewith R factors of existing compatibility groups. Certain Rfactors associated with isolates of other genera recovered inrelation to the epidemic strains had an "epidemic R factor"phenotype. These results suggest different genetic origins forthe "epidemic R factors" or dramatic evolutionary divergenceand indicate the potential of phenotypic analysis of R factorsin epidemiological studies. (These studies were supported by agrant from NIAID.)

289. Effects of Antibodies Specifically Directed Against Na+,K+-ATPase. THOMASS. SMITH,* HENRY WAGNER, JR.,*MICHAEL YOUNG,ANDJACK KYTE,* Boston, Mass.

Antibodies have been obtained which specifically interactwith the transport enzyme Na+, K+-activated ATPase. The

78 a

antigen used was Na+, K+-ATPase from canine renal medullapurified to homogeneity as judged by polyacrylamide-gelelectrophoresis. Pre- and postimmunization sera from rabbitschallenged with this antigen in complete Freund's adjuvantwere fractionated by ammonium sulfate precipitation, DEAEcellulose ion-exchange chromatography, and prolonged dialy-sis to obtain the -a-globulin fraction. Gammaglobulin fromimmunized animals, but not from control animals or preim-mune serum, inhibited Na+, K+-ATPase activity of caninerenal medullary extracts up to 684A4 (SD) %in a concentra-tion-dependent manner with maximum inhibition occurringwithin 5 min at 370C. The Mg++-dependent, non-Na+,K+-activated, and non-ouabain-inhibited component of activ-ity (Mg++-ATPase) was unaffected. Fab fragments obtainedby papain cleavage of the -y-globulin fraction had similar in-hibitory activity and specificity. These antibodies also pro-duced concentration-related inhibition of canine myocardialmicrosomal and human red cell ghost Na+, K+-ATPase activi-ties by up to 50i6 and 98a5%, respectively. Control andpreimmune -y-globulin fractions had no effect on canine myo-cardial or human red cell ghost Na+, Kt-ATPase. Mg++-AT-Pase was unaffected by all preparations including those withmarked Na+, K+-ATPase inhibitory activity. Despite markedinhibition of Na+, K+-ATPase activity in these particulateenzyme preparations, experiments with canine renal slicesand human red cells showed no specific effect of antibody onouabain-inhibitable 86Rb uptake, indicating a lack of inhibi-tion of active monovalent cation transport in the intact cell.These experiments demonstrate immunologic cross-reactivityamong Na+, K+-ATPases from different organs and differentspecies. In addition, they indicate that the antibody responseis directed against an antigenic determinant or determinantsinaccessible to macromolecules at the outer cell surface.(Supported by grants from NIH and AHA.)

290. The Role of the Liver in Metabolism of Low DensityLipoproteins (LDL). A. D. SNIDERMAN,* T. E. CAREW,* J. G.CHANDLER,* S. HAYES,* ANDD. STEINBERG,** La Jolla, Calif.

Elucidation of the normal mechanisms for lipoprotein secre-tion into and removal from plasma is prerequisite to an under-standing of the metabolic errors in inherited hyperlipoprotein-emias. Our initial studies in pigs showed that autologouslabeled plasma LDL (d 1.019-1.063; > 95% 1251I in protein)was in rapid equilibrium with a pool of LDL in the liver, theonly major extravascular pool identified. To assess the role ofthe liver in LDL removal, the disappearance of [1251] LDLwas studied before and after total hepatectomy, each animalserving as his own control. Intact pigs and dogs showed a bi-exponential 1251 disappearance: phase I tj (equilibration withextravascular pools), 0.9 h and 1.7 h, respectively; phase IItj (irreversible removal from plasma), 20 h and 27.4 h, re-spectively. Three pigs and three dogs were re. tudied immedi-ately posthepatectomy (porto-caval shunt and en bloc re-moval). The disappearance was now monoexponential over theentire 13-20 h study period and more rapid: tj 8.8 h in pigs;11.3 h in dogs. In posthepatectomy pigs, LDL (d 1.006-1.063)was reisolated for measurement of protein specific radioactiv-ity. This did not change significantly, while absolute LDL pro-tein levels decreased by 55-73% over the intervals studied.The results show that under the conditions of study there wasno significant de novo LDL synthesis and secretion from extra-hepatic tissues. Barring the possibility that hepatectomy isaccompanied by substantial shifts in rates and patterns ofLDL removal in the periphery, irreversible LDL removal bythe liver seems to be quantitatively minor and the liver may,directly or indirectly, play a role in stabilizing LDL to prolongits lifetime in the plasma. (Research supported by NIH grantHL-14197.)

291. Pituitary and Thyroid Responses to Repetitive Adminis-tration of Thyrotropin-Releasing Hormone (TRH). PETER J.SNYDER* AND ROBERTD. UTIGER, Philadelphia, Pa.

Previous studies showed that the thyrotropin (TSH) re-sponse to thyrotropin-releasing hormone (TRH) is readily in-hibited by exogenous triiodothyronine (T3) and thyroxine(T4). In this study, repetitive TRH administration was usedto produce small increases in endogenous serum T3 and T4levels to determine if endogenously produced T3 and T4 in-hibited the TSH response to subsequent TRH doses. Eachsubject received 13 consecutive doses of 25 jug TRH at 4-hintervals. Serum TSH, T3, and T4 levels were measured before(basal levels) and for 4 h after the first, seventh, and thir-teenth doses of TRH. In 10 normal subjects, the mean increasein serum TSH levels fell from 14.6 /LU/ml after the first TRHdose to 6.9 and 3.0 IMU/ml after the seventh and thirteenthdoses. The reduced TSH responses were accompanied by risesin mean basal serum T3 levels from 81 to 115 to 114 ng/100 ml(normal range, 70-150) and in mean basal serum T4 levelsfrom 6.7 to 8.6 to 9.6 ,gIg/100 ml (normal range, 5-11) at thetime of the first, seventh, and thirteenth TRHdoses. In a pa-tient with primary hypothyroidism, repetitive TRHadminis-tration resulted in no increase in serum T3 and T4 levels andthe TSH response to TRH was not blunted. These resultsfurther demonstrate the marked sensitivity of TRH-inducedTSH release to suppression by small increases in serum Tsand T4 levels and show the suppression is not due to depletionof TSH from the pituitary. Inhibition of TRH-induced TSHrelease was preserved in two patients with hypothyrotropichypothyroidism. In one patient, with presumed TRH de-ficiency, the TSH response was blunted by repetitive TRHadministration, but only after the serum T3 and T4 levels in-creased to normal. In the other patient, with presumed pitui-tary insensitivity to TRH, the initially subnormal TSH re-sponse was blunted further by repetitive TRHadministrationeven though the serum T3 and T4 levels did not rise to normal.Preservation of sensitivity of TRH-induced TSH release toinhibition by small rises in serum T3 and T4 despite insensitiv-ity to TRHin this patient supports the concept that inhibitionand stimulation of TSH release are mediated by differentmechanisms.

292. RNA-Dependent DNAPolymerase of Erythroid Cells: aPossible Mechanism for Control of Hemoglobin Synthesis.ANTERO G. So,* JOHN J. BYRNES,* AND KATHLEEN M.DOWNEY,*Miami, Fla. (introduced by William J. Harrington).

There are numerous defined or suspected disorders of hemo-globin synthesis. Although the regulatory mechanism under-lying these disorders is not clearly understood, control of hemo-globin synthesis is usually held to occur at the translationallevel. Wewish to present an additional mechanism for controlof hemoglobin synthesis. To account for the exceedingly rapidrate at which hemoglobin is synthesized in developing ery-throid cells, amplification of the structural genes for the a-and 3-chains has been postulated. Amplification of the genesfor ribosomal RNAhas been found to occur in amphibian co-cytes and shown to involve RNA-dependent DNAsynthesis.We have found a RNA-dependent DNA polymerase in thecytoplasm of rabbit erythroid cells. The cytoplasmic synthesisof DNAusing the endogenous template in inhibited by ribonu-clease A, suggesting that RNAacts as the template for thispolymerase. The endogenous activity is also inhibited by highlevels of actinomycin D, suggesting that a DNA template isalso utilized. The purified enzyme prefers native DNAto syn-thetic DNA/RNAhybrid. The control of hemoglobin synthe-sis involving this enzyme will be discussed, with possible ap-plications to human disease. (Research supported by grantsfrom NIH and AHS.)

79 a

293. IgG Half-Molecules in a Patient with Plasma Cell Leu-kemia. HANS L. SPIEGELBERG* AND VICTOR HEATH,* LaJolla, Calif. (introduced by Frank J. Dixon**).

The paraproteins of a patient with plasma cell leukemia whoexcreted 300 mg/100 ml of a K, IgG paraprotein into the urinewere characterized. The serum contained two monoclonalproteins, one having a sedimentation rate of 7S (0.5 g/100 ml)and the other 4.3S (1 g/100 ml); the 4.3S protein was identicalto the urinary protein. The paraproteins were detected whenquantitation of IgG in the serum was attempted by commerci-ally available immunodiffusion plates, which showed two con-centric precipitin rings. Similarly, analysis of the serum by im-munoelectrophoresis showed a double line with anti-IgG, butonly a single line with anti-K antisera. Only the 7S protein pre-cipitated with anti-Fc fragment antisera; the 4.3S protein,however, inhibited this precipitin reaction. After reductionand alkylation, both the 7S and the 4.3S proteins dissociatedinto heavy and light chains, the mass ratios being 2:2 and 1 :1respectively. The molecular weights were similar to normal-y- and Kchains, excluding a large deletion in the heavy chain.The heavy chains of both proteins contained the peptide whichforms the inter-heavy-heavy-chain disulfide bonds, its aminoacid composition being identical to that of -yi-chains. Theelectrophoretic mobilities and the amino acid compositions ofthe polypeptide chains of both the 7S and 4.3S proteins wereindistinguishable, suggesting that the same clone of cells pro-duced both proteins. In contrast to normal IgG, the 7S proteindissociated into half-molecules (4.3S) after mild reduction andalkylation in aqueous solution at neutral pH. These data indi-cate that the paraproteins of this patient lacked the nonco-valent bonds normally present in the Fc portion of -y-chainsand suggest that the absence of these bonds resulted in im-paired assembly of 7S IgG molecules. (Supported by NIH andAHA.)

294. Studies on the Characterization of Transfer Factor.LYNNE. SPITLER,* DAVID WEBB,* CHRISTINE VONMULLER,*AND H. HUGHFUDENBERG,** San Francisco, Calif.

The nature of transfer factor was studied by immunologicaland biochemical analysis and by systemic transfers of cocci-dioidin reactivity in normal subjects using conversion of skinreactivity (SR), lymphocyte stimulation (LS), and productionof migration inhibitory factor (MIF) as measures of reactiv-ity. The crude dialyzable transfer factor preparation containeddeoxyribonucleotides, ribonucleotides, and peptides. Transferfactor which had been heated at 56°C for 2 h transferred allthree parameters of reactivity, and after heating at 80°C for 1h it transferred SR and LS in 4 normal subjects but trans-ferred MIF capacity in only 4. All parameters of reactivitywere transferred by a preparation which had been treatedwith pancreatic RNase and TI RNase at low molarity, andcontrols run in parallel indicated that only single strandedRNAwas hydrolyzed. Treatment with pronase destroyed thetransfer of SRand NIF, and LS was transferred in only 4 sub-jects who received this preparation, suggesting that a peptidemust be part of the active molecule. After passage through aUM10 Diaflo membrane, transfer factor did not transfer SRor MIF, but did transfer LS. Weconclude that there may bedifferent transfer factor for the transfer of SR, LS, and MIFand that these may be distinguished on the basis of physico-chemical characteristics.

295. Azurophil and Specific Granules Resolved for HumanPolymorphs (PMN): Immunochemistry, Biochemistry, En-zyme Histochemistry. J. K. SPITZNAGEL,* F. G. DALLDORF,*M. S. LEFFELL,* J. D. FOLDS,* Chapel Hill, N. C. (introducedby C. W. Gottschalk**).

Granules deliver antimicrobial substances to phagocyticvacuoles. Intraleukocytic bactericidal defects have been de-scribed in PMNlacking myeloperoxidase or lactoferrin, anti-microbial proteins normally associated with granules. Wehaveisolated normal granules and characterized their contents inorder to define antimicrobial capacities of granules and de-granulation and more readily to interpret studies with po-tentially abnormal PMN. Crude granule suspensions (from99% pure PMN) applied in 11% sucrose to linear (50-53%)sucrose density gradients and centrifuged at fotwldt = 2.3X 1010 rad2sec- with R. = 15.3 and Rmin = 6.4 cm yieldedthree turbid bands (I, II, and III in order of increasing sedi-mentation rate). Band III contained 88% of the myeloper-oxidase measured immunochemically (IC) and 77% of theneutral protease measured spectrophotometrically (SP).Band II had 65% of the lactoferrin (IC). Lysozyme (SP) wasbimodally distributed between II (50%) and III (46%).Alkaline phosphatase (85%) was in I. Crude granules appliedin 25% sucrose to gradients and centrifuged yielded in place ofIII two narrowly separated bands (Ills and IlIf). I and IIwere unchanged. The MPO(95%) and the neutral protease(95%) peaked in Ills with a shoulder in IlIf, while lysozymepeaked in IIIf with a shoulder in IIIs. Ultarstructurally bandsIlls and IIIf contained peroxidase-positive 0.3 IA granules.II contained peroxidase-negative 0.12 IA granules. In I wereempty vesicles with alkaline phosphatase positive membranes.Evidently the MPO-positive (azurophil) granules comprisetwo subgroups, one relatively richer in MPOand protease(Ills) and one (IIlf) which is richer in lysozyme. The MPOnegative granules (specific) are rich in lysozyme and lacto-ferrin. Perhaps IIIf granules arise fom the Golgi apparatus asits secretory cycle shifts from secretion of azurophil to specificgranules. (Supported by NIAID and AEC.)

296. Peripheral Nerve Myelin in Experimental Diabetes.NORTONSPRITZ, BARBARAGEYER,* AND HARBHAJANSINGH,*New York.

Abnormalities in Schwann cell function related to themaintenance of myelinization may play an important role inperipheral neuropathy in diabetic patients. Wehave quantifiedthe amount of myelin in sciatic nerves of rats and rabbits withexperimental diabetes, and have measured in vitro their in-corporation of isotopic precursers into myelin lipids. Sciaticnerves were incubated with either tritiated water or [1-C14]acetate and myelin isolated from the homogenate. Incorpora-tion into myelin phospholipid had first-order kinetics, andmyelin isolation was quantitative and uncontaminated. Inrats with diabetes of 1 wk incorporation was like that of con-trols. In four groups of 4-6 diabetic rats with diabetes of 3-12wk duration, specific activity ranged from 55 to 71 %of con-trols for phosphotidyl choline the main constituent andphosphotidyl ethanolaine and phosphotidyl serine-the otherphospholipids with significant incorporation. In rabbits, speci-fic activities were about 3 that of the rat and were not differentbetween diabetics and controls. Composition of isolated mye-lin did not differ between diabetics and controls in eitherspecies. In the rabbit, however, myelin content decreased incontrols with age (1.80 mg/cm of nerve at 6, 1.54 at 7, 0.9 at8, and 0.85 at 10 months of age). In age-matched diabeticrabbits (3-8 months duration) myelin content ranged from 65to 90%of controls. These studies indicate that defective mye-lin synthesis in the rat and decreased myelin content in olderrabbits occur in experimental diabetes. This suggests that, atleast in part, neuropathy in diabetic man may be a conse-quence of insulin deficiency per se and that experimental dia-betes provides a model for its study. (Supported by VA re-search and NIH grant AM13525.)

80 a