Chemical Treatment of Tumors. VII. Nature of the Hemorrhage ...

Chemical Treatment of Tumors. VII.Nature of the Hemorrhage-ProducingFraction From Serratia marcescens (Bacil-lus prodigiosus) Culture Filtratel

By J. L. HARTWELL, chemist, M. J. SHEAR, principal biochemist, and JOHN R. ADAMS, JR.,assistant chemist, with the technical assistance of ADRIEN PERRAULT, National CancerInstitute, National Institute of Health, United States Public Health Service

A previous communication (2) fromthese laboratories described the prepara-tion of a polysaccharide fraction fromfiltraten of Serratia marcescens (Bacillusprodigiosus) cultures, which was highlvpotent in producing hemorrhage in mousesarcomas. The present paper reports thefurther purification of the product and theresults obtained to date on the nature andcomposition of the material as revealed bychemical degradation experiments.

Highly potent concentrates of the activepolysaccharide fraction (2) were submit-ted to tryptic digestion. This treatmentlowered the nitrogen content from about 4to 2 percent without detectable loss of po-tency. With the trypsin-digested material,hemorrhage was produced in 50 percent ofthe tumors by a dose of 0.1 microgram.About two-thirds of this material consistedof sugar residuen, of which the major por-tion was aldohexose. The active material,after tryptic digestion, also contained smallamounts of phosphoric acid, a solid fattyacid, glycerol, and a nitrogenous constitu-ent, i. e., the components of a phospholipid.The lipid was present in firm combinationand, up to the present, methods whichremoved the lipid also destroyed thepotency. No evidence was found for thepresence of protein or polypeptide.

1 Presented in part (1) at the Boston meeting of theAmerican Association for Cancer Research, March 31,1942.

MATERIALS AND METHODS

STARTING MATERIAL

The concentrates prepared in the pre-vious work were turbid, viscous solutions,neutral in reaction. The preparations hadbeen dialyzed in cellophane balts againstwater and therefore contained only non-dialyzable constituents. Lyophilization ofthese concentrates gave a light gray, fluffy,noncrystalline product which was readilysoluble in water.

The aqueous solutions gave negativetests for proteins by the biuret, ninhydrin,xanthoproteic, and Millon's reactions.The solid material also reacted negativelyto the Millon test. The Molisch test forpolysaccharide was strongly positive. TheBenedict test for reducing rugars wasnegative, but after short boiling of theconcentrate with dilute hydrochloric acidit was positive. The iodine-potassiumiodide test was negative, showing theabsence of such substances as starch,glycogen, and certain dextrins.

These concentrates gave precipitateswith several reagents. Phosphotungsticacid and silver nitrate each gave a sedi-ment within 3 hours; flavianic acid, tri-chloracetic acid, and sulfosalicylic acidyielded sediments only after a longertime; and tannic acid and potassiomer-curic iodide gave only turbidities. Picricacid, picrolonic acid, and Reinecke salt

107

at Pennsylvania State University on Septem

ber 16, 2016http://jnci.oxfordjournals.org/

Dow

nloaded from

http://jnci.oxfordjournals.org/

108 JOURNAL OF THE NATIONAL CANCER INSTITUTE

gave no precipitate. These preparationstherefore contained only small amounts ofproteinaceous materials, although the ni-trogen content was about 4 percent.

The active material contained about 6percent "acetyl," which was not identifiedas such but represented the volatile acidscalculated as acetic acid. Ash was presentto the eitent of about 3.5 percent. Only0.16 percent sulfur was found. This con-stituent was not in a labile or unoxidizedform, as shown by testing with lead acetateafter boiling with dilute sodium hydroxide.The molybdate test for inorganic phos-phate was negative, but elementary analysisgave a value of 1 percent phosphorus.

BIOASSAY METHOD

Assays of potency in producing hemor-rhage in tumors were carried out accordingto the technique described in the precedingpaper (3). Each step in the manipula-tions was correlated with its effect on thepotency in this way. In addition, assaysof important steps were performed two ormore timer for greater reliability.

RESULTS

FURTHER PURIFICATION BY TRYPTICDIGESTION

The partially purified starting material,which was chiefly polysaccharide in nature,was submitted to the action of a numberof proteolytic enzymes for 2 days at 37° C.Trypsin was employed at pH 8.5, pepsinat pH 2.1, and papain at pH 4.4. Asjudged by analysis of the dialyzates fortyrosine, 2 trypsin was the only one whichgave evidence of digestive action.

Treatment of the active polysaccharidefraction with amylopsin (pancreatic amy-lase) at pH 6.9 for 24 hours at 37° did notsplit off reducing sugars, nor was any

2 By the method outlined in Koch's manual (4, p. 248).

change in the tyrosine value obtained.The potency of the material remained un-altered. The action of this carbohydrate-splitting enzyme was therefore not studiedfurther.

In the tryptic digestion, the concentratewas treated with 0.008 of its dry weight oftrypsin 3, brought to pH 8.5 by the addi-tion of phosphate buffer, and set aside(under toluene) at 37 ° C. for at least 2 days.The solution, which had become much lessturbid, was then dialyzed against runningwater and precipitated by alcohol alter-nately from alkaline (pH 9.5) and acid(pH 2.0) solution. The final precipitatewas dissolved in water and dialyzed again.

Within wide limits of trypsin excess, thetyrosine value of the nondialyzable mate-rial could not be reduced below a certainminimum figure; this apparently repre-sented a reducing action of the polysac-charide itself on the phosphomolybdotung-state reagent, since d-glucose and d-glu-cosamine hydrochloride were found togive reduction corresponding to 0.09 and0.20 percent, respectively, of their weightof tyrosine. The buffer alone, in the ab-sence of trypsin, was ineffective in liber-ating tyrosine. Furthermore, progressiveaddition of trypsin to the concentrate, withor without intermediate removal of deg-radation products by dialysis, failed tolower the tyrosine value below this mini-mum. On previously tryptic-digestedmaterial, trypsin had no further actionthat was apparent.

After tryptic digestion, the material gaveentirely negative results with most of thereagents for proteins or polypeptides thathad been applied to the starting material.(See p. 107.) A small amount of sedimentwas obtained with trichloracetic acid, with

3 Trypsin 1 : 250 (Difco Laboratories). This was acrude extract of pancreas which was not purified furthersince it was considered desirable to retain enzymes otherthan trypsin itself.



Dow

nloaded from


NATURE OF POTENT BACTERIAL FRACTION

109

potassiomercuric iodide, and with silvernitrate; however, these sediments were ob-tained only on long standing. Thetryptic-digested material was also testedwith several additional reagents: a turbid-ity was obtained with barium hydroxideand with neutral lead acetate; coppersulfate, and mercuric chloride both gavenegative results.

The trypsin-digested material was opti-

cally active with a specific rotation [a] 2of + 128 ° . The determination was madeon an approximately 1-percent neutralaqueous solution in a 0.254-dm. cell.

Representative analyses 4 of variouslots of concentrate before and after trypticdigestion are given in table 1. Digestionproduced significant changes in two re-respects: (1) The nitrogen content waslowered from about 4 to about 2 percent;and (2) the "acetyl" content was loweredfrom about 6 to about 2 percent. Nomethoxyl was found on analysis of thetrypsin-digested material.

Treatment with trypsin had no detecta-ble effect on the potency of the activematerial. Three portions of preparationK-15 were digested separately, at differenttimes. The activity of these products wascompared with that of undigested K-1 5 in

4 The microanalyses were performed by the ArlingtonLaboratories, Chagrin Falls, Ohio, unless otherwisenoted.

4 assay experiments (table 2). In Bach ofthe first 2 sets of experiments the compara-tive assays were carried out simultane-ously; 60 mice of the same batch of tumor-bearing animals were used in Bach experi-ment. The other 2 assays were performedon different days with tumor-bearing miceof different batches. The assays showedthat treatment with trypsin did not di-minish the activity.

All the data reported henceforth in thispaper were obtained with trypsin-digestedmaterial.

ALTION OF HEAT

The solid material or the dilute aqueoussolution could be preserved indefinitely at10° C. with no loss of potency. The solu-tion showed no appreciable change inactivity when kept at 37° for severalweeks. In the temperature range 60°-100° the activity rapidly decreased withtime. Loss of potency at 60° was notice-able in 1 hour, while at 100 ° there wascomplete destruction within one-half hour.Table 3 gives the data on the loss ofpotency when neutral solutions of theactive concentrate in test tubes wereimmersed in the vapors of various boil-ing liquids for hour to 16 hours.

Heating of the neutral aqueous solutionresulted in the separation of a brownish

TABLE 1.-Composition of the hemorrhage-producing polysaccharide fraction,' from Serratia marcescens culturefiltrate, before and after tryptic digestion

Sample Tryptiedigestion

C g N P Acetyl Ash

Percent Percent Percent Percent Percent PercentK-8d-------_-------_---_ Before------- 47.6 7.3 3.8 0.9 7.1 3.8K-15 - ----------------- -- ---

do------- -------------- ------------..-4.3 ------------ 4.6 3.2

K-15------------------ -lifter-------- -- ----------- ------------- - 2.1 _------------- 2.4 --------------

K-15b---------------- - -- 48.6 7.2 2.1 ------------ 2.2. 3.3K-15e- ------- -------- _ ----- do------- -------------- ------------- -------------- 1.1 3.0

-•--do------- 46.6 7.0 1.6 1.0 2.2 ------- --2 5.3

P-3 -- ------------------ -- - -do------- --_-...47.4 __..... 7.0 2.9 1.1 1.2 3.8

47.5 7.1 2.2 1.1 -2.2 3.5Avetage------_----_ After_----.-_

1 The solutions were evaporated to dryness and the residues dried to constant weight for analysis.2 Not included in average.



Dow

nloaded from



TAB1.F, 2.—Bioassays showing the resislance to tryptic digestion of the hemorrhage-producing activity of thepolysaccharide fraction from Serratia marcescens culture filtrate

Specimen ofconcentrate

Tryptiedigestion

Volume of1:10,000dilution

Mice with induced hemnrrhage in tumor _ End-pointdose 1

injected .... +.. ++ + —

Cubic Cubiccentimeter Number Number Number Number Number centimeter

7------ 0.25 ----------5 - ---------- 3 -- --------- ---------- 2. 2 0.15l}

5 7 2 1--- -------- -----------K-15

1Beiore-

1 ____ . ______ 1 2 1 6)After-------- 4 3 2 1 ______ .S{( .25

5 6 2 2------------ ----------

J}

( 1 1 3 1 1 4 l4 3 1 ____________ 2 } .1jl .2

4 6 1 1 1 1K-15a ----------------

1Before______.

J 1 ---------- 3 ------------ ------------After-------- .2 5 4

---------------- ------ 71

1.15l

4 3 5 ----------- ----------- 2

1 ------------ 1 ------------ ----------Betere_____-_ .25 -----_---___ 3 2 194 .25l

5 1 3 3 2 1K-15b

1 1 2 ____________ 3 4 lAfter-------- 3 7 .1

.2 2--------° ---------- 2_ 3 2

1 1 : 10,000 dilution.

substance, insoluble in water but solublein alkali.

ALTION OF ACID

Solutions of the active material werebrought to pH 5, 3, and 1 by means ofhydrochloric acid and kept for 24 hoursat room temperature; they were thendialyzed against distilled water. Thedialyzates were all found to be inactivewhile the nondialyzable fractions retainedthe original potency in full.

When submitted to the action of 2 Nhydrochloric acid at 37° C., the activitywas rapidly destroyed, about 90 percentbeing lost in 6 hours (table 4). At 48

TABLE 3.—Rate of destruction with heat of the hem-orrhage producingpotencyof the polysaccharidefractionfrom Serratia marcescens culture filtrate.

Time required for loss

Temperature (in Mice used in potency of-

degrees centigrade) in assays —J—

50 percent 190 percent

Number Hours Hours60---------------------151 4 >16

80-------------------- 130 2 8100_______________ 103 <0.5 1

hours, less than 1 percent of the potencyremained; more than half of the dissolvedmaterial was now dialyzable. The non-dialyzable residue, on analysis, yieldedonly 0.75 percent acetyl as compared witha value of 3.0 percent for this preparationprior to treatment with acid.

TABLE 4.—Rate of destruction with 2 N hydrochloricacid at 37° C. of the hemorrhage-producing potency ofthe piolysaccharide fraction from Serratia marcescensculture filtrate

Tempera- Mice Residual potency after various periodsture (in used ---------------degrees in as- 1.5 3 6 12 24 48

centigrade) says hours hours hours hours hours hours

Ier cent cent cent cent cent cent37__________ 517 >10 >50 10 5 2 <1

80---------- 30 0 ------ ------ ----- - -

The progress of the inactivation by hy-drochloric acid was followed with simul-taneous determinations of reducing sugar, 5

hexosamine, 6 and inorganic phosphorus 1

(table 5). The amounts of these compo-

5 By the Hagedorn-Jensen-Hanes method (5).e By the method of Palmer, Smyth, and Meyer (6).7 By the method of YToungberg, as Biven by Hawk,

Bergeim, Oser, and Cole (7, p. 454).



Dow

nloaded from



111

TABLE 5.—Rate of hydrolysis of the active polysaccharidefraction from Serratia marcescens culture filtratewith 2 N hydrochloric acid at 37 0 C.

Time (in hours) Reducingrugar Hexosamine Phosphorus

Percent 2 Percent 3 Percent 4

0___ 4.1 2.4 5.34.7 4.2 6.36.8 7.5 13.3

3---------------7.3 8.0 17.06_______________ 8.8 5.3 19.512__ 14.9 10 2524__ 25.2 19 3048______________ 38.0 38 32

1 Caleulated as glucose.2 Based on 64 percent total glucose obtained on maximum

hydrolysis.3 Based on 2.0 percent total hexosamine obtained on maxi-

mum hydrolysis.4 Based on 1.1 percent total phosphorus obtained on ele-

mentary analysis.

nents liberated were smaller than antici-pated; however, they all increased pro-gressively as the potency decreased.

The rates of destruction of potency andof liberation of reducing sugar, of hexosa-mine, and of inorganic phosphate areshown in figure 1. The data are given inpercentages, for each component, of theamount liberated in 48 hours, after de-duction of the values for the blanks, i. e.,the values obtained at zero time. Hydroly-sis released reducing sugars and hexosa-mine at rates which were approximatelyproportional to the time, and which didnot parallel the rate of loss of potency.

The hydrolysis curve for phosphate, how-ever, exhibited a sharp rise during thefirst 12 hours as did the curve for destruc-tion of potency. Whether there is a signif-icant relationship between loss of phos-phate and loss of potency remains forfurther investigation to determine.

The optimum concentration of hydro-chloric acid for maximum hydrolysis at100 0 was found to be 1.5 N, when 64 per-cent of reducing sugars (calculated asglucose) was released in 3 hours (fig. 2).

At 80 ° , 2 N hydrochloric acid causedcomplete inactivation within 1.5 hours(table 4) along with the separation of alight gray, flocculent precipitate, whichwas soluble in dilute alkalies. From thealkaline solution the precipitate could beproduced again by acidification. At 37 °a similar precipitate was obtained onstanding for a month. Further data onthis component are given on page 115.

ALTION OF ALKALI

Solutions of the concentrate werebrought to pH 9 and 11 with sodiumhydroxide and dialyzed against water atroom temperature for 24 hours; no lossof potency occurred in the nondialyzablefraction. Treatment with 0.1 N sodium

100

90

60

70

60

50

40

á 30

W 20

10

0

LEGEND:

❑ POTENCY LOSSX INORGANIC PHOSPHORUS0 REDUCING SUGARS RELEASEDA HEXOSAMINE

FICURE 1.—Effect of 2 N hydro-

chloric acid at 37 0 C. on the

active polysaccharide fraction

from Serratia marcescens culture

filtrate. The curves represent

the rate of destruction of

potency and the rates of 'liber-

ation of three components.

0 2 4 6 12 24 48

TIME IN HOURS



Dow

nloaded from


112

JOURNAL OF THE NATIONAL CANCER INSTITUTE

100

90

80

70

6 0

m 50

40

30

f 20

10

00 1 2 5 ♦ 5 6

TIME 1N 9,OU88

FIGURE 2.-Rate of hydrolysis of the active poly-

saccharide fraction from Serratia marcescens

culture filtrate with 1.5 N hydrochloric acid

at 100 0 C.

hydroxide at room temperature (28 ° C.)resulted in destruction of about two-thirds of the potency in 4 hours and ofabout nine-tenths in 12 hours (table 6);the solution turned light amber in color,and the turbidity cleared rapidly. Nosignificant amount of material was loston dialysis, but the nondialyzable solidon analysis showed a complete loss in "ace-tyl"; analyses of other components showedno significant change.

At 37° about two-thirds of the potencywas lost in 3 hours (table 6). At 56 °about 90 percent of the activity was de-stroyed in 1 hour; concomitantly the acetylcontent fell to zero while the nitrogencontent was unchanged. At 60 ° , aboutthree-fourths of the activity was destroyed

TABLE 6.-Rate of destruction of potency of the activepolysaccharide fraction from Serratia marcescensculture filtrate by treatment with 0.1 N sodium hydrox-ide at different temperatures

Tempera-Residual potency after various periods

Mice _tureTempera- _useddegrees

centigrafie) in 1 hours 3 4 u7 12

assaus hout hours hours hours hours 1 hours

Nu-- Per- Per- Per- Per- Per- Per-ter cent cent cent cent cent cent

28-_--- ----_ 250 ---_-- ------ 100 30 - 37-___-______ 330 __-___ 100 30 ------ 15 ______56--------- 30 110 ------ ------ ----- ------ ------60----------- 90 <1 --------------------------------

1 "Acetyl" content reduced to zero.

in 15 minutes, and almost all was gone in1 hour.

The progress of inactivation with 0.1 Nsodium hydroxide at 37° was followed withsimultaneous determinations of reducingsugar, hexosamine, and phosphoric acid;the amounts of these substances liberatedwere found (table 7) to be much smallerthan was the case with 2 N hydrochloricacid.

The action of 0.1 N sodium carbonateat 37° C. was similar to that of sodiumhydroxide but less drastic. After 24 hours,about four-fifths of the activity was de-stroyed; simultaneously the greater part ofthe acetyl was removed from the non-dialyzable fraction while the nitrogencontent remained essentially the same(table 8). The effect of 10 percent byvolume of pyridine, acting for 24 hours at37 ° C. was to destroy about half theactivity; about two-thirds of the acetylwas removed, while the nitrogen contentwas essentially unchanged. The onlyaction of 0.1 N sodium bicarbonate wasthe reduction of the acetyl to the lowestvalues obtained with full potency, viz,0.6 and 1.2 percent. After treatment for45 hours with sodium bicarbonate underthese conditions, potency was not de-creased while the nitrogen and phosphorusvalues remained unchanged.

TABLE 7.-Rate of hydrolysis of the active polysaccharidefraction from Serratia marcescens culture filtratewith 0.1 Nsodium hydroxide at 37 0 C.

Time (in hours) Reducing Hexosamine Inorganiesugar 1phosphorus

Percen 2 Percent 3 Percent 4

%-------------- 2.8 4.2 4.80----------------2.8 2.2 5.0

3---------------3.2 5.8 5.113/-------------- -3.1 5.7 4.8

7---------------3.6 10.9 5.412---------------3.5 5.4 5.724_____________ 4.1 7.8 5.4

1 Calculated as glucose.2 Based on 64 percent total glucose obtained on maximum

hydrolysis.3 Based en 2.0 percent total hexosamine obtained on maxi-

mum hydrolysis.4 Based on 1.1 percent total phosphorus obtained on ele-

mentary analysis.



Dow

nloaded from



113

TABLE 8. Effect of treatment of the active polysaccharidefraction from Serratia marcescens culture filtratewith weak alkalies for 24 hours at 37° C.

Alkali

Miceused

inas-

says

Resid-ualpo-

teney

Nitrogencontent

_Acetylcontent

_'Be- After Be- Afterfore fore

Nun'- Per- Per- Per- Per- Per-ber cent cent cent 1 cent 1 cent

Sodium carbonate'_ 100 1 20 2.1 E 1.9 3.0 1 0.2Pyridine 2 45 50 2.1 1 1.9 3.0 1 .9Sodium bicarbon- 145 {100 2. 1ate. 3 1.5 1 3.0 1.2

1 0.1 N.2 10 percent by volume.3 These data veere obtained in 2 sets of experiments with

different preparations of active material.

QUALITATIVE IDENTIFICATION OF SUGAR

COMPONENTS

In this work the scheme of Militzer (8)was found to be useful. Unless otherwisespecified, the tests were carried outaccording to the directions given byMilitzer or by Van der Haar (9). Thelyophilized solid, the aqueous solution,and the hydrolyzate prepared by heatingthe active material at 100 ° C. with 1.5 Nhydrochloric acid were all used in thetests. For controls, appropriate sugarswere used.

The polysaccharide did not reduceBenedict's reagent, whereas the hydroly-zate was strongly reducing. The followingtests all gave negative results: the iodine-potassium iodide test for starch, glycogen,and dextrin; Seliwanoff's resorcinol test forketone sugar; Bial's orcinol and theWheeler-Tollens phloroglucinol tests forpentose and uronic acid; the aniline andthe Bayer-Rosenthaler resorcinol tests forpentose; and Kiliani's test for desoxysugar. The negative results of the Bialtest also indicated the absence of heptose. S

The foregoing tests ruled out the presenceof most sugars except hexoses and methyl-pentoses. The Rosenthaler and theOshima-Tollens phloroglucinol tests formethylpentose, and the Elson-Morgan test(10) for amino sugar were strongly positive.

s Private communication from Dr. C. S. Hudson.

The Jatter was not an unusual finding, ford-glucosamine bas been reported severaltimes as occurring in bacterial polysac-charides. Inositol was shown by theSalkowski modification of the Schererreaction (11) to be absent.

Since quantitative data indicated thepresence of a large amount of reducingsugar which was not accounted for asmethylpentose or as amino sugar, the pres-ence of a considerable amount of aldo-hexose was indicated. Qualitative testsshowed the absence of the two aldohexosesother than glucose which are commonlyencountered in bacterial polysaccharides.Thus, the mucic acid test for galactose andthe phenylhydrazone test for mannose wereboth negative. The hydrolyzate never-theless yielded a crystalline osazone prod-uct which under the microscope was seento be a mixture of yellow needies andorange plates; the yellow needles bore astrong resemblance to a specimen ofglucosazone prepared under similar con-ditions. The aldohexose was thereforeprovisionally regarded as glucose. Furtherchemical work on identification of theindividual sugars present had to be post-poned because of factors beyond ourcontrol.

QUANTITATIVE DETERMINATION OF SUGAR

COMPONENTS

The determination of methylpentose wascarried out by the Krüger-Tollens-Kröbermethod (9). From a solution of concen-trate containing 214 mg. of polysaccharide,18.1 mg. of phloroglucide was obtained.According to the tables of Ellett and ofMayer, as given by Van der Haar (9),this corresponds to 35.4 mg. of anhydrousrhamnose, or 44.6 mg. of fucose, i. e., to16.5 or 20.8 percent, respectively. Thephenylhydrazone test applied to thehydrolyzate failed to yield the precipitatewhich fucose gives. Whether the methyl-



Dow

nloaded from



pentose was rhamnose or fucose was notestablished, but the available informationfavored the former of these two commonestrepresentatives of their class.

Aminohexose was determined by thecolorimetric method of Palmer and co-workers (6). With two different batchesof material, values of 3.3 and 2.0 percent,calculated as d-glucosamine, were obtained.As far as we are aware, no aminohexoseother than d-glucosamine has been identi-fied in bacterial polysaccharides.

The amount of aldohexose that is formedon hydrolysis may be estimated from thevalue of 64 percent found for the totalreducing sugars (calculated as glucose),less the determined amounts of methyl-pentose and aminohexose after conversionto their respective glucose equivalentsbased on their reducing power. By usingpure dry samples of the sugars, it wasfound by the Hagedorn-Jensen-Hanesmethod that 1 mg. of anhydrous rham-nose, 1 mg. of fucose, and 1 mg. of d-glu-cosamine gave glucose equivalents of1.04, 0.80, and 1.00 mg., respectively. 9

Under similar conditions of hydrolysis,rhamnose suffered no loss in reducingpower, while d-glucose lost only 6.7 per-cent. Thus, in the hydrolyzate, 16.5 per-cent of rhamnose would appear as 17.2

percent when calculated as glucose whilethe value of glucosamine (2.7 percentaverage) would appear unchanged. Sub-traction of the sum of these two valuesfrom 64 percent gave 44.1 percent whichwould represent the content of d-glucose.If fucose is assumed to be the methyl-pentose present, the percentage of glucosewould be 44.7.

Some support for the assumption maderegarding the identity of the component

9 This last value is different from that found by Freu-denberg and Eichel (12), who reported that glucosaminehas only 70 percent of the reducing power of glucose asdetermined by the Hagedorn-Jensen method.

sugars is given by data on optical rotation.After hydrolysis under optimum condi-tions, the observed optical rotation,

[a] D, changed from + 128 ° to + 36 ° .

The calculated values for the opticalrotation of mixtures of these sugars, inthe relative proportions postulated, agreedwell with the observed rotation for thehydrolyzate when rhamnose was assumedto be the methylpentose (table 9). Whenfucose was taken as the methylpentose, theagreement was much less satisfactory.

DETECTION OF COMPONENTS OTHER

THAN SUGARS

Nitrogen comprised 2 percent of thetrypsin-digested polysaccharide fraction.The 2.7 percent of hexosamine accountedfor 10 percent of the total nitrogen. Onhydrolysis an alkali-soluble product wasobtained which resisted further hydrolysis;it contained 1.9 percent nitrogen. Thismaterial appeared to be similar to the"humic acids" observed as byproducts inthe hydrolysis of other polysaccharides(13).

Heat alone also caused the separation

TABLE 9.-Comparison of the observed optical rotation ofthe hydrolyzate of the active polysaccharide fraction,from Serratia marcescens culture filtrate, with naluescalculated for mixtures of different simple sugars

Com- []of[] of posi-

tion [a1 hydrol-Mixture Sugar simpte of calcu- yzate,

sugars' mom _ lated2 obser-ture ved

De- Per- De- De-grees cent grees grees

[d-Glucose------ - -+52.5 69.7 +36.6A _________ Rhamnose 3 ____ + 9.9 26.1 + 2.6

d-Glucosamine. +44 4.2 +1. S

Total__ _________________ +36100.0 +41.0

d-Glucose______ +52.5 65.5 +34.4B _________ Fucose_________ -75.6 30.5 -23.0

d-Glucosamine. +44 4.0 +1.6

Total----------------------------- +36100.0 +13.0

1 Experimental values taken Erom the literature.1 Obtained by multiplying values in column 3 by these in

column 4. 3 Anhydrous.



Dow

nloaded from



115

of the component which was found to besoluble in alkali.

A quantity of the insoluble productwas prepared by hydrolyzing a solutionof the concentrate, containing 1.22 gm. ofsolid, with 1.5 N hydrochloric acid for2.5 hours at 100°. The gray product,which feit greasy, was obtained in a yieldof 0.191 gm. (or 16 percent). It gave astrong Molisch test for carbohydrate anda negative Liebermann-Burchard test forcholesterol. After treating it further withboiling alcoholic hydrochloric acid (equalvolumes of 95 percent alcohol and con-centrated acid) which apparently had noteffected further hydrolysis, a light-coloredsolid was extracted with ether. Thislatter solid dissolved in alkali and precipi-tated again on addition of acid to thealkaline solution. It thus appeared to bea solid fatty acid. In the quantitativemethylpentose determination, in whichthe concentrate was distilled with 12 per-cent hydrochloric acid, a white solidacid was allo obtained, which was presentin the steam distillate in about 4-percentyield.

The hydrolyzate, after removal of theacid-insoluble material, gave a stronglypositive molybdate test for phosphoricacid. A portion of this solution, whenevaporated to dryness and heated withdry potassium bisulfate, gave a strongodor of acrolein; this indicated thepresence of glycerol.

The filtered hydrolyzate was analyzedfor ammonium nitrogen and total nitro-

gen 10 and was tested for various nitrog-enous substances. Only traces of am-monium nitrogen and of amino acids werefound. The biuret test gave ambiguousresults, and the Millon test was negative.The Folin-Looney phosphomolybdotung-

10 By micro-Kjeldahl digestion, and determination ofthe ammonia colorimetrically with JVessler's reagent.

state test for tyrosine, the Sakaguchi testfor guanidines (arginine), and Ehrlich'sdiazo test were all positive. Benedict'smodification of the Hopkins-Cole reactionfor tryptophane, the Sullivan test forcystine, the bromine test for histidine, theEhrlich test for pyrroles, and the murexidetest for purines were all negative. Theabsence of purines was confirmed by Dr.Kahler, of the National Cancer Institutestaff, who found in the unhydrolyzed con-centrate no specific absorption in theultraviolet region at the wavelength char-acteristic of purines.

Of the 2 percent total nitrogen in thepolysaccharide fraction, 0.5 percent wasaccounted for by hexosamine and thealkali-soluble hydrolysis product. The re-maining 1.5 percent of the total nitrogenwas not accounted for quantitatively.Qualitatively, it was shown that thepolysaccharide contained materials whichgave positive colorimetric tests for certainamino acids even after tryptic digestion.The presence of a solid fatty acid, phos-phate, glycerol, and soluble nitrogenousmaterial indicated the presente of aphospholipid.

Elementary analysis of the polysaccha-ride fraction gave about 1 percent phos-phorus and about 0.2 percent sulfur. Themolybdate test for inorganic phosphatewas negative prior to hydrolysis but wasstrongly positive in the clear, filteredhydrolyzate, showing that the phosphorusin the polysaccharide was present inorganic combination. Not all of it, how-ever, was released on hydrolysis, for theinsoluble material obtained on digestionwith acid still contained the same propor-tion of phosphorus as the original material.

With barium chloride, the clear hydrol-yzate readily gave a small amount ofprecipitate, showing the presence of sulfatein traces.



Dow

nloaded from



ATTEMPTS TO SEPARATE THE POLYSAC-CHARIDE AND PHOSPHOLIPID MOIETIES

To determine whether the polysac-charide and phospholipid componentswere united in a molecular complex orwere present merely as a mixture, and inan attempt to ascertain which componentwas the Beat of the hemorrhage-producingactivity, several different types of experi-ments were carried out.

Morgan and Partridge (14) found thatthe polysaccharide moiety of a complexisolated from a dysentery bacillus could beseparated from the phospholipid compo-nent by simple solution of the complex inanhydrous formamide and precipitation ofthe polysaccharide with alcohol. We foundthat lyophilized active preparations fromSerratia marcescens dissolved readily in form-amide to give a solution clearer than theaqueous one. Unlike the finding of Mor-gan and Partridge, however, addition ofabsolute alcohol to the formamide solutionbrought down the original product appar-ently unchanged; the hemorrhage-produc-ing potency and the reducing value wereessentially unaltered, and no appreciablelowering of the phosphorous content waseffected.

The possibility that the phospholipidhad been dissociated in formamide solutionbut re-entered the complex after additionof alcohol was tested in two ways. Theformamide solution was dialyzed againstformamide for 27 days, after first establish-ing that a formamide solution of lecithinwould lose lecithin under these conditions.The formamide solution was also extractedwith hexane. Neither of these proceduresresulted in the removal of a phosphorus-containing component from the complex.

Fractional adsorption was also tried.The complex was shaken with successiveadditions of an adsorbing agent and wascentrifuged after each addition; the super-

natant solution was examined for phospho-rus, nitrogen, and for biologie activity aftereach treatment. The adsorbing agentsused were Lloyd's reagent, activated alu-mina, bentonite, norite B, calcium carbon-ate, and lead phosphate. There was noapparent adsorption of phosphorus ornitrogen or of the potent component withLloyd's reagent, activated alumina, cal-cium carbonate, or lead phosphate. Ben-tonite removed most of the activity alongwith a large proportion of the nitrogen andthe phosphorus. Norite B, however,removed most of the activity, but none ofthe phosphorus, and little of the nitrogen.Attempts to recover the potent materialfrom the norite by elution with water,with 0.001 N hydrochloric acid, or with0.001 N sodium hydroxide were unsuccess-ful. The experiments with norite indicated,however, that further work with adsorbentsmight be fruitful.

A sample of the concentrate was extract-ed continuously with hexane (b.p.65°-67°)at a temperature of 28 °-38 ° C. On theninth day it was made 0.1 N with respect tosodium bicarbonate, and the extractioncontinued for 6 more days. On the fif-teenth day enough hydrochloric acid wasadded to neutralize the bicarbonate andto bring the solution to 0.1 N with respectto the acid; the extraction was then con-tinued for 26 more days. Samples re-moved at intervals showed no significantreduction in phosphorus. This was goodevidence that the phospholipid was notpresent in simple mixture. Another ali-quot of the concentrate was brought topH 11 by the addition of sodium hydroxideand glycine, and extracted continuouslywith hexane for 24 days; about one-halfof the phosphorus was removed. It hadpreviously been found (p. 111) that treat-ment of the complex for 24 hours at pH 11did not produce any detectable diminutionin potency. Hence the removal by hexane



Dow

nloaded from



117

at this pH of a phosphorus-containingcomponent may possibly provide a methodof separation of the type sought. No datawere obtained on the activity of thesefractions, for the investigation had to bediscontinued at this stage.

MISCELLANEOUS EXPERIMENTS

In addition to the systematic workdescribed in the foregoing sections, a num-ber of miscellaneous experiments of anexploratory nature were carried out.''

Because of the sharp decrease in the"acetyl" content which usually accompa-nied loss of potency on hydrolysis, a fewtest experiments were performed from thestandpoint of acetylation and de-acetyla-tion. Several preparations were treatedwith acetic anhydride at room tempera-ture, but no increase in potency waseffected. Hydrolysis with 0.1 N sodiumhydroxide at room temperature for 22hours destroyed most of the potency;treatment of this product with aceticanhydride, likewise at room temperature,did not restore the potency to the originalvalue. The partial de-acetylation pro-cedure of Zemplén (15) was applied to thelatter material; this also was ineffective inrestoring the original activity.

Inasmuch as hexosamine was found tobe a component of the polysaccharide,d-glucosamine in doses up to 5 mg. wasinjected into mice bearing sarcoma 37; nohemorrhage was produced in the tumors.N-acetyl-d-glucosamine, in doses up to 0.5mg., likewise gave negative results.

Other carbohydrate substances whichwere found ineffective in producing hemor-rhage in tumors were: i-inositol; a, d-galacturonic acid; glycuronic acid; tre-halose; dihydroxyacetone; soluble starch;inulin; pectin; and a- and /3-lignin sulfonicacids.

A number of hemolytic agents, anti-

11 In part in collaboration with Dr. F. C.lTurner.

coagulants, and capillary constrictors ordilators, were tested in tumor-bearingmice. Several substances related to thesein structure or in physiologic action werealso tested. Intraperitoneal or subcuta-neous injection of the following materialsgave negative results as regards productionof hemorrhage in sarcoma 37 (the valuesin parentheses give the maximum doseinjected):

Cholic acid (0.5 mg.); desoxycholic acid(0.5 mg.); sodium taurocholate (0.5 mg.);sodium glycocholate (0.2 mg.); sodiumfluoride (0.2 mg.); sodium oxalate (0.2mg.); sodium citrate (0.2 mg.); hirudin(0.5 mg.); Pontamine Sky Blue 6BX (0.5mg.); Congo red (0.5 mg.); 3,3'-methylene-bis-(4-hydroxycoumarin) (0.5 mg.); hista-mine (0.2 mg.); penicillin (0.1 mg.);actinomycin (0.005 mg.); digitalin (0.5mg.); peptone (0.2 mg.); and acetylcholinebromide (0.2 mg.).

Evidence of hemorrhage production intumors was obtained with a few of thematerials tested by intraperitoneal injec-tion. Saponin, in a dose of 0.1 mg., killedhalf the mice but produced a moderateamount of hemorrhage in all the tumors;smaller doses gave negative results. Digi-tonin, in a dose of 0.5 mg., producedhemorrhage in the tumors of one-third ofthe animals. Three azo dyes, which werestated in the literature to be anticoagu-lants, were tested. The two which gavenegative results are listed above; Pont-amine Fast Pink BL, however, produceda moderate amount of hemorrhage intumors in a dose of 0.5 mg. Adrenalinehydrochloride was lethal in doses of 0.1mg. or more; 0.05 mg. was tolerated andproduced a moderate degree of hemor-rhage in some of the tumors. An old,weak, commercial preparation of heparinfrom dog liver produced hemorrhage insome of the tumors in doses of 0..5 and 2.0mg. Four relatively fresh preparations of



Dow

nloaded from



heparin were then tested. Two specimensderived from beef lung were obtained froma commercial source. Two other speci-mens obtained from the Connaught Lab-oratories in Canada were stated to behighly purified preparations derived frombeef lung and from sheep (organ notstated). In doses of about 1 mg., negativeresults were obtained with all four speci-mens of heparin. Fresh beef liver and lungwere fractionated in this laboratory ac-cording to the methods employed forpreparation of crude heparin; all fractionsyielded negative results in tests for hemor-rhage-producing potency. One possibleexplanation for the positive result givenby the first heparin preparation is that itmay have been contaminated, at somestage in its history, with bacteria possessingability to produce hemorrhage in tumors.

However, Antopol and Glick (16)obtained evidence of preparatory andprovocative potency, for the Shwartzmanphenomenon in normal rabbit skin, in thecase of sterile preparations from certainmammalian organs. Four of these prep-arations were obtained from Dr. Antopoland tested in mice bearing sarcoma 37.The dry specimens were dissolved in waterand injected intraperitoneally in doses of0.2 and 0.02 mg. One material gavenegative results, and two others each pro-duced a slight amount of hemorrhage inone of five tumors. The extract of steerpancreas, however, possessed considerablepotency, for 0.2 mg. produced hemorrhagein all the tumors, and in two of the fivemice the hemorrhage was severe. Evenwith 0.02 mg., hemorrhage was producedin some of the tumors. The pancreasemployed had been cultured and found tobe sterile both aerobically and anaerobi-cally. 12 Antopol and Glick (16) statedthat every step in the preparation of thefractions was of such a nature that sterility

12 Personal communication.

was automatically maintained. The posi-tive result with this material warrantsfurther investigation of mammalian tissueextracts.

An experiment was carried out to deter-mine whether the bacterial polysaccharidepossessed any hemolytic power. Washedrabbit erythrocytes were suspended inisotonic phosphate buffer solution (pH 6.7)containing the active polysaccharide inconcentrations ranging from 0.001 to 1.0mg. per cubic centimeter. A 0.9-percentsolution of sodium chloride, and the samebuffer solution containing saponin in thesame concentrations as the polysaccharide,were employed as controls. Hemolysiswas produced rapidly with concentrationsof saponin down to 0.03 mg. per cubiccentimeter, and progressively more slowlywith lower concentrations down to 0.008mg. per cubic centimeter. In contrast, nohemolysis was obtained with the polysac-charide solutions in any of the dilutionstested. The highest concentration of poly-saccharide employed contained 10,000mouse tumor units per cubic centimeter,i. e., 10,000 times the dose required toproduce hemorrhage in 50 percent ofthe tumors. This indicated that themechanism of hemorrhage production intumors with this polysaccharide did notinvolve hemolysis of erythrocytes.

Some of the substances mentioned pre-viously in this section were selected fortrial in an effort to obtain information onpossible mechanisms of hemorrhage pro-duction in tumors. The experiments ofAndervont and Shimkin (17) led them topostulate that bacterial filtrate, by lower-ing suddenly the ascorbic acid content ofthe tumor, weakens its fragile capillarieswith resultant hemorrhagic extravasation.This view was based on their Tinding thatascorbic acid prevented the production ofhemorrhage in tumors with bacterialfiltrate; similar but less pronounced inhibi-



Dow

nloaded from



119

tion was obtained with cysteine. Theirexperiments were carried out with sarcoma37 and with the crude filtrate from cul-tures of Serratia marcescens as the hemor-rhage-producing agent.

However, in our hands, neither ascorbicacid nor cysteine inhibited hemorrhage.The same technique was employed exceptthat even smaller doses of the hemorrhage-producing agent were used. The injec-tion of large doses of ascorbic acid or ofcysteine was therefore expected to exert amore pronounced inhibition than theseworkers reported, but no appreciable in-fluence on hemorrhage production wasobserved. The chief difference betweenthe two sets of experiments was that Ander-vont and Shimkin employed the crude fil-trate, whereas we employed a purified con-centrate obtained from the filtrate.Whether this was responsible for the dif-ference in the results is left for future workto determine. Similar lack of inhibitionof hemorrhage production was observedwith glutathione also. Shwartzman andSchick (18, p. 160) found that ascorbic acidfailed to exert any effect on the produc-tion of the Shwartzman phenomenon withbacterial filtrate.

DISCUSSION

The relevant literature has been summar-ized in a preceding communication (2).The most recent discussion of the chemicalnature of bacterial substances which pro-duce hemorrhage in tumors is that of Zahl,Hutner, and coworkers (79), who were ofthe opinion that these agents are probablyidentical with a polypeptide component ofcertain complete endotoxin 0 antigens.In a later paper Hutner and Zahl (20)substituted "protein" for "polypeptide" inreferring to their chemical nature.

The nitrogen content of our trypsin-digested potent preparations was about 2percent, which is too low a value for a poly-

peptide. Furthermore, part of this smallcontent of nitrogen has been accountedfor by hexosamine and by the nitrogenouscomponent of a phospholipid. Finally,the tests for proteins and polypeptides wereessentially negative, and those for aminoacids yielded only faintly positive results.In this connection, the findings of Freuden-berg and Molter (21) are relevant. Instudying the composition of the bloodgroup A substance from urine, they detect-ed some amino acids not of protein originin the hydrolyzate; the substance itselfwas thought to be a polysaccharide.

The qualitative tests and quantitativedata on carbohydrate substances showedthat our preparations were largely polysac-charide in nature since about two-thirds ofthe material consisted of sugar residuen.In addition, the components of a phospho-lipid were detected. The phospholipidmoiety behaved as if it were present in firmcombination and not in a loosely boundcomplex.

Zahl and Hutner (22) have reportedthat orally administered sulfanilamide iscapable of inhibiting the hemorrhagic effectof about 16 minimum hemorrhage-produc-ing doses of a fraction obtained from cul-tures of Salmonella typhimurium. They alsoreported (20) that sulfanilamide protectedagainst 2 to 10 minimum lethal doses andthat p-aminobenzoic acid interfered withthe protective action of sulfanilamide.

Gerber and Gross (23) found that asulfanilamide conjugate of a bacterialfiltrate possessed both preparative andprovocative potency in the Shwartzmanreaction. A meningococcus culture filtratewas coupled with p-aminobenzenesulfonyl-acetylimide, and a highly colored conju-gate was obtained. They concluded thatchemical treatment of the bacterial filtratein the syntheses of the conjugate apparentlydid not affect its ability to produce theShwartzman reaction.



Dow

nloaded from



The active polysaccharide material fromSerratia marcescens culture filtrate containedabout 1 percent phosphorus in organiccombination. Part, at least, of the phos-phorus was present presumably as a com-ponent of a phospholipid. Whether anyphosphate was present in combinationwith sugar residues is still an open question.

The potency was unaltered on standingat 37° C. However, the activity wasdestroyed at higher temperatures; the rateof destruction increased rapidly in thetemperature range 60° to 100°. Treat-ment with acid (2 N hydrochloric acid)caused rapid Toss of potency at 37 0 , whileheating with acid at 80° produced morerapid destruction. These findings are ingeneral agreement with those of Gardner,Bailey, and Hyde (24) who studied a toxiccarbohydrate complex from Salmonellaenteriditis.

However, they concluded that the hem-orrhagic reactions produced in the Walkerrat tumor 256 were due to the polysac-charide-free component of the complex.The low potency of their polysaccharidecomponent may well have been a conse-quence of almost complete destruction ofactivity by heat and by acid, since theirpolysaccharide was obtained by a hydro-lytic process involving heating at 100° C.for 4 hours in 0.2 N acetic acid. Detaileddata on minimum hemorrhage-producingdoses of the polysaccharide and of thetoxic component obtained in this waywere not Biven, but both had little potency,amounting in both cases only to aboutone one-thousandth that of the unheatedmaterial prior to hydrolysis. It would beworth wbile to attempt to devise a methodof obtaining these components by amethod which does not require such ahigh temperature or such strong acid.

Further work is required to establishwhether the entire polysaccharide fractionfrom Serratia marcescens filtrate is required

for potency, or whether it resides in somecomponent of the complex. In all casesin which the "acetyl" group was com-pletely removed, the potency was entirelydestroyed. In a few instapces, prepara-tions with a low acetyl content wereobtained which possessed high potency.It is not known at present whether thissmall residual amount of acetyl is essentialfor activity.

The production in this laboratory ofhemorrhage in tumors with 0.02 mg. ofthe extract prepared by Antopol andGlick (16) was all the more unexpectedinasmuch as high temperatures for manyhours were employed in the preparation ofthe extract. If this material is chemicallysimilar to the active polysaccharide frac-tion from Serratia marcescens, the potencymight be expected to be greatly enhancedby avoidance of such high temperatures.The heat employed may have been re-sponsible for the inconstant results ob-tained by Antopol (25) in producing theShwartzman reaction with analogousextracts of sterile tumor tissue. Thetesting of the effect on tumors of Shwartz-man-active extracts of sterile tumors isanother aspect of the subject which ap-pears to warrant attention.

Preparations obtained from a variety ofbacteria have a strikingly selective actionon tumor tissue. In the case of culturefiltrates from at least two organisms,Escherichia coli (26) and Serratia marcescens,polysaccharide fractions, which gave nega-

tive results in tests for protein, have been

obtained which induced hemorrhage in

mouse sarcomas in doses less than 1 micro-

gram. For conveniente these materials

are referred to as polysaccharide fractions,

not because it has been demonstrated that

the potency resides in the polysaccharide

moiety but because these preparations are

preponderantly polysaccharide in nature.



Dow

nloaded from



121

SUMMARY

The potent concentrate previously ob-tained from Serratia marcescens culture fil-trate was found to retain its activity afterdigestion with trypsin. The minimumdose which produced hemorrhage in 50percent of the tumors (sarcoma 37) wasabout 0.1 microgram. Elementary analy-sis of a number of preparations after trypticdigestion gave the following average re-sults in percent: Carbon, 47.5; hydrogen,7.1; nitrogen, 2.2; phosphorus, 1.1; acetyl,2.2; ash, 3.5. No methoxyl was found.A single determination for sulfur gave avalue of lens than 0.2 percent. The valuesfor nitrogen and "acetyl" were about halfthose obtained prior to digestion withtrypsin.

The active material, upon optimumhydrolysis, yielded aldohexose, hexosa-mine, methylpentose, and the componentsof a phospholipid. About two-thirds ofthe material consisted of sugar residues, ofwhich the major portion was aldohexose.The phospholipid was present in a firmlybound complex. The presence of neitherpolypeptide nor protein could be demon-strated.

The chemical behavior of the tryptic-

digested material was investigated in anumber of ways. The effects of the manip-ulations on the potency were determinedwith an improved bioassay technique, andthe bioassay results correlated with thechemical findings.

The potency of aqueous solutions wasnot destroyed at 37 ° C. In the range 60 °to 100 e, potency was lost at rates whichincreased with the temperature. In thelight of the heat stability of the material,the effect of acid and alkali on the potencywas investigated at various temperatures,and the hydrolytic changes produced un-der these conditions were studied.

0Acknowledgment is made of the gener-

ous gift of specimens of the following sub-stances: extracts of mammalian tissues(Dr. William Antopol, Newark Beth IsraelHospital); heparin (Drs. A. F. Charles andL. B. Jaques, Connaught Laboratories,Toronto); a- and /3- lignin sulfonicacids (Prof. Harold Hibbert, McGill Uni-versity); coumarin derivative (Prof. K. P.Link, University of Wisconsin); and twoazo dyes, Pontamine Sky Blue 6BX andPontamine Fast Pink BL (DuPont Co.).

REFERENCES(1) HARTWELL, J. L., SHEAR, M. J., and ADAMS,

J. R., Jr.: Nature of the bacterial poly-saccharide which produces hemorrhage inmouse sarcoma. Cancer Research, 3: 122(1943).

(2) SHEAR, M. J., and TURNER, F. C.: Chemicaltreatment of tumors. V. Isolation of thehemorrhage-producing fraction from Ser-ratia marcescens (Bacillus prodigiosus) culturefiltrate. J. Nat. Cancer Inst., 4: 81-97(1943).

(3) SHEAR, M. J. PERRAULT, A., and ADAMS,

J. R., JR.: Chemical treatment of tumors.VI. Method employed in determiningthe potency of hemorrhage-producing bac-terial preparations. J. Nat. Cancer Inst.,4: 99-105 (1943).

(4) KocH, F. C.: Practical methods in biochem-istry. William Wood & Co., Baltimore(1934).

(5) HANES, C. S.: An application of the method

of Hagedorn and Jensen to the determina-tion of larger quantities of reducing sugars.Biochem. J., 23: 99-106 (1929).

(6) PALMER, J. W., SMYTH, E. M., and MEYER,

K.: On glycoproteins. IV. The estima-tion of hexosamine. J. Biol. Chem., 119:491-500 (1937).

(7) HAWK, P. B., BERGEIM, 0., OSER, B. L.,and COLE, A. G.: Practical physiologicalchemistry. Ed. 11. Blakiston's Son &Co., Philadelphia (1937).

(8) MILITZER, W. E.: The qualitative chemicalidentification of the natural sugars. J.Chem. Educ., 18: 25-28 (1941).

(9) VAN DER HAAR, A. W.: Anleitung zumNachweis, zur Trennung und Bestim-mung der reinen und aus Glukosiden usw.erhaltenen Monosaccharide und Alde-hydsauren. Gebrüder Borntraeger, Berlin(1920).



Dow

nloaded from


122

(10)

(11)

(12)

JOURNAL OF THE NATIONAL CANCER INSTITUTE

(19)ELSON, L. A., and MORGAN, W. T. J.: Acolorimetric method for the determinationof glucosamine and chondrosamine, Bio-chem. J., 27: 1824-1828 (1933).

SALKOwsKI, E.: Uber eine Verbesserung derSchererschen Reaktion auf Inosit. HoppeSeyler's Ztschr. f. physiol. Chem., 69:478-481 (1910).

FREUDENBERG, K., and EICHEL, H.: Uberspezifische Kohlènhydrate der Blutgrup-pen. Liebig's Ann. d. Chem., 510: 240-248 (1934).

(13) ABDERHALDEN, E.: Biochemisches Handlexi-kon. Bd. 2 and 10. Julius Springer,Berlin (1911, 1923).

(14) MORGAN, W. T. J., and PARTRIDGE, S. M.:Studies in immunochemistry. 4. The frac-tionation and nature of antigenic materialisolated from Bact. dysenteriae (Shiga).Biochem. J., 34: 169-191 (1940).

(15) ZEMPLÉN, G.: Abbau der ReduzierendenBiosen, I: Direkte Konstitutions-Ermitt-lung der Cellobiose. Ber. d. deutsch.chem. Gesellsch., 59: 1254-1266 (1926).

(16) ANTOPOL, W., and GLICK, D.: Organs as asource of factors capable of eliciting theShwartzman phenomenon. Proc. Soc.Exper. Biol. & Med., 38: 346-348 (1938).

(17) ANDERVONT, H. B., and SHIMKIN, M. B.:The effect Gf ascorbic acid upon the hem-orrhage produced by bacterial filtrate intransplanted tumors. Am. J. Cancer, 36:451-459 (1939).

ZAHL, P. A., HUTNER, S. H., SPITZ, S.,SUGIURA, K., and COOPER, F. S.: The ac-tion of bacterial toxins on tumors. I. Rela-tionship of the tumor-hemorrhagic agentto the endotoxic antigens of gram-negativebacteria. Am. J. Hyg., 36: 224-242 (1942).

(20) HUTNER, S. H., and ZAHL, P. A.: The actionof sulfanilamide compounds on the lethalfactor of bacterial toxins. Science, 96:563-564 (1942).

(21) FREUDENBERG, K., and MOLTER, H.: Thesubstance characteristic of blood group Ain urine. IV. Sitzber. heidelberg. Akad.Wiss., Math. Naturw. Klasse 1939,Abhandl. 9, 21 pp. [Abstracted in Chem.Abstr. 35: 759 (1941).]

ZAHL, P. A., and HUTNER, S. H.: Action ofbacterial toxins on tumors. II. Effect ofsulfanilamide on toxin-induced hemor-rhage. Proc. Soc. Exper. Biol. & Med.,51: 285-287 (1942).

GERBER, I. E., and GROSS, M.: Productionof the Shwartzman phenomenon with asulfonamide conjugate of a bacterial fi1-trate. Proc. Soc. Exper. Biol. & Med., 51:237-238 (1942).

GARDNER, R. E., BAILEY, G. H., and HYDE,R. R.: Hemorrhagic activity of toxic car-bohydrate complexes from bacteria on atransplantable rat tumor. Am. J. Hyg.,Sect. B, 29: 1-14 (1939).

ANTOPOL, W.: The use of tumor extracts inthe production of the Shwartzman phe-nomenon. J. Infect. Dis., 61: 334-337(1937).

SHEAR, M. J.: Chemical treatment of tumors.IV. Properties of hemorrhage-producingfraction of B. coli filtrate. Proc. Soc.Exper. Biol. & Med., 34: 325-326 (1936)

(22)

(23)

(24)

(25)

(18) SHWARTZMAN, G.: Phenomenon of local tis- (26)sue reactivity and its immunological,pathological, and clinical significance.Paul B. Hoeber, Inc., New York (1937).



Dow

nloaded from


Chemical Treatment of Tumors. VII. Nature of the Hemorrhage ...

Documents

Transcript of Chemical Treatment of Tumors. VII. Nature of the Hemorrhage ...