STUDIES ON THE THROMBOLYTICACTIVITY OF HUMANPLASMA*

By WILLIAM D. SAWYER,ANTHONYP. FLETCHER, NORMAALKJAERSIGAND SOL SHERRY

(From the Department of Internial Medicine, Washington University School of Medicine,St. Louis, Mo.)

(Submitted for publication September 15, 1959; accepted October 22, 1959)

The process of in vivo clot dissolution, termedthrombolysis,l is complex, involving the interactionof clot components with the surrounding plasma.Participating in this interaction are plasminogen,plasminogen activator, plasmin, the plasmin in-hibitory activity of plasma, and fibrin (1). Al-though much is known of the nature and quanti-ties of individual components, the complexity of thethrombolytic process has led to difficulty in assess-ment of the overall capability of the plasma for ef-fecting clot lysis under physiological conditions-an activity hereafter termed plasma thrombolyticactivity. Plasma thrombolytic activity is a prop-erty of major physiological significance (2, 3) buthitherto the technical methods used for its assayhave been imperfect. Assays involving determina-tion of whole blood or plasma clot lysis time arevaluable for detecting markedly enhanced thrombo-lysis but are inadequate for measurement of lowlevels of thrombolytic activity due to the very longlysis times found and the lack of a precise end-point. Other assays have involved physicochemi-cal alterations of the plasma such as dilution (4,5), isolation of the euglobulin fraction (6, 7) andextraction with organic solvents (8). Such physi-cochemical procedures isolate or affect to a variabledegree one or more components of the thrombo-lytic process and, although valuable for specializedpurposes, disturb the complex balance between theindividual plasma components (1). Consequently,assays of this nature are unsuitable for other thanqualitative determination of plasma thrombolyticactivity.

* This work was supported by grants-in-aid from theNational Heart Institute, Bethesda, Md. (H 3745), andthe Lederle Laboratories Division, American CyanamidCo.

1 The term thrombolysis is used to designate a processby which preformed clots are lysed, and its use does notimply specific reference to pathological states (1).

In order to circumvent these limitations, a sen-sitive measure of plasma thrombolytic activity wasdeveloped in which unaltered plasma is tested.In principle the method involves the determinationof radioactivity released from isotopically-labeledhuman plasma clots immersed in unaltered plasma.Although an in vitro procedure, the use of unal-tered plasma and preformed clots mimics in vivoconditions and permits the obtaining of physio-logically relevant measurements. Using this iso-topic clot assay, measurements of plasma throm-bolytic activity have been made on plasma obtainedfrom healthy adults, from adults following stress,after the administration of drugs, and from indi-viduals with disease. The results indicate thatplasma obtained from nonstressed normal adultsexhibits thrombolytic activity due primarily tothe presence of a plasminogen activator and thatthis activity varies as a result of stress and disease.

MATERIALS AND METHODS

I131-trace-labeled bovine fibrinogen 2 was prepared aspreviously described (1).

Human plasminogen was prepared by Kline's proce-dure (9) from human plasma fraction III.3

Human plasinin was prepared by spontaneous activationof human plasminogen in 50 per cent glycerol (10).

Thrombin mixture consisted of Parke-Davis thrombindiluted to 10 units per ml with equal parts 0.7 per centcalcium chloride and 6 per cent dextran. The use ofthe latter reagents aided in the production of more con-sistent and stable clots.

Urokinase 4 contained 5,100 Ploug units per milligramdry weight.

2 Salt-free bovine fibrinogen containing 96 per centthrombin-clottable nitrogen; kindly supplied by Dr. KentMiller, New York State Institute of Health, Albany, N. Y.

3 Human plasma fraction III was obtained from Dr.T. 0. Gerlough, E. R. Squibb and Sons, through thecourtesy of Dr. J. N. Ashworth of the American RedCross.

4 Kindly supplied by Dr. J. Ploug, Leo Pharmaceuticals,Copenhagen, Denmark.

426

THROMBOLYTICACTIVITY OF HUMANPLASMA

e-Aminocaproic acid 5 was chemically pure and chro-matographically homogeneous.

Buffered saline refers to 0.9 per cent sodium chloridecontaining 0.1 Mphosphate buffer at pH 7.6.

131-trace-labeled human plasma clots were formed fromaged human plasma (stored for at least 30 days at 40 C)enriched with human plasminogen (vide infra) and con-taining trace quantities of I1"-labeled bovine fibrinogen.Two-tenths ml of such plasma was clotted by the additionof 0.1 ml of thrombin mixture. The clot was caused toform along the internal surface of glass tubing (7 mmOD) by rotation of the tube in a horizontal plane at ap-proximately 350 rpm in the chuck of an electric motor.Clots formed in this fashion were small, compact andlargely free of entrapped plasma, a critical feature of thetechnique. When formed, the clot was floated from thetubing and placed in a test tube. Clots were washed over-night by gentle agitation in buffered saline. Becausethis washing procedure, necessary to produce clots ofsufficiently low blank radioactivity, eluted considerableplasminogen from the clots, it was necessary to enrichthe original plasma with plasminogen as noted above.The physical state of dispersion at pH 7.6 of the puri-fied plasminogen used was such that it was taken up fromplasma into the clot and not washed out (1). The con-centrations of 131 fibrinogen and plasminogen added weresuch that the final clot contained on the average 2.5 X10' ,ic per Ag fibrin (128 cpm per Ag under detection con-ditions used, side infra), 0.4 casein unit plasminogen and500 ,ug fibrin. I.31 bovine fibrin represented approximately2 per cent of the total. Such clots, designated as plas-minogen-rich, were used for all routine determinations.

Low plasninogen antd plasininogen-free I13"-labeledplasma clots were made as above except that no purifiedplasminogen was added. Plasminogen content was de-termined by previously described methods (1) modifiedonly insofar as multiple small clots were used for eachdetermination. Plasminogen-poor clots contained on theaverage 0.04 casein unit plasminogen per clot. Such clotsexhibited normal sensitivity to lysis by plasmin. Clotscould be rendered plasminogen-free by heating (800 C for30 minutes); such clots were unsusceptible to lysis byplasminogen activators and their sensitivity to lysis byplasmin was somewhat reduced (1).

Thrombolytic assay with IP'-labeled human plasma clotswas performed by draining the wash solution from aclot and then incubating the latter for 2 hours at 37° C im-mersed in 0.2 ml of the material to be tested for thrombo-lytic activity. The clot was then separated by filtrationunder air pressure (10 p.s.i.) using medium porosity frittedglass disks. An additional 2 ml of saline was passedthrough the filter to minimize loss. The radioactivity ofthe total filtrate was determined in a well-type scintilla-tion detector equipped with a pulse height analyzer. Theemission of the isotope (0.364 MEVpeak) was deter-mined utilizing a 5 v window. Overall radiation detec-tion efficiency under such circumstances was 25 per cent

5 Kindly supplied by Dr. J. Ruegsegger, Lederle Lab-oratories. Pearl River, N. Y.

1 2UROKINASEugm. per ML.

FIG. 1. LYsis OF I131-LABELED HUMANPLASMA CLOTS

AS A FUNCTION OF UROKINASECONCENTRATION. The lysisof clots was linearly related to urokinase concentrationover the range illustrated. The open circles indicate re-sults for urokinase in salin'A and the closed circles indicateurokinase added to aged human plasma. The intercept onthe abscissa for both urokinase in saline and aged plasmaindicates that the method is sensitive to the presence ofless than 1 Ag urokinase per ml. The differences of theslope and intercept of the lines for urokinase in salineand in aged plasma were consistent findings.

and background was consistently in the range of 12 to25 cpm. The use of a pulse height analyzer was con-sidered essential since it afforded considerable increaseof sample to background ratio making possible accuratedetection of the low levels of radioactivity used in thistechnique. The statistical error of counting was less than2 per cent. Thrombolytic activity was expressed as mi-crograms of fibrin lysed per 2 hours. Previous use ofP13`-labeled clots revealed residual clot nitrogen contentafter partial lysis to be highly correlated with residualclot radioactivity (1). Thus the radioactivity of the fil-trate appeared to be a valid measure of clot lysis.

The sensitivity of the I"' clot assay to a plasminogen ac-tivator, urokinase, is shown in Figure 1. The amount offibrin lysed is linearly related to urokinase concentrationover the range shown. Concentrations of urokinase insaline of 0.5 ,ug per ml and in aged plasma of 1 ,ug per mlresulted in significant lysis of the isotopically-labeled clots.The divergence between urokinase in saline and urokinaseadded to aged plasma was a consistent finding and pre-sumably reflects an inhibitory effect of plasma upon uroki-nase. The reproducibility of the method under condi-tions of thrombolytic activity such as are found in normaladults was 0.81 ,ug per 2 hours, as determined by thestandard deviation of the difference between duplicate de-terminations upon 59 plasma samples (11).

Plasma euglobulitz clot lysis was determined as previ-ously described and the results expressed as units ofactivity (6).

427

SAWYER,FLETCHER, ALKJAERSIG AND SHERRY

Assays for plasma proteolytic activity were performedusing the substrates, urea-denatured hemoglobin, as de-scribed by Anson (12) and casein, as previously de-scribed (1).

Fibrinogen was determined by the method of Ratnoffand Menzie (13).

Plasma samples were harvested from venous blood col-lected in oxalate tubes (4 mg potassium oxalate per mlof blood) and tested at once. Occasionally, samples werestored at 20 C for a short period prior to assay. In gen-eral, samples were obtained in the morning and in thefasting state.

RESULTS

I. Plasma thrombolytic activity of normal adults.Upon incubation of the isotopically-labeled clots inbuffered saline, radioactivity was slowly releasedfrom the clots into the surrounding media. Therange and mean of 134 control observations withbuffered saline are shown in Figure 2, the mean re-lease of radioactivity being equivalent in amountto 2 jug fibrin lysed per 2 hours. Since protein andother plasma constituents might affect the rate ofsuch release, advantage was taken of the labilityof plasma thrombolytic activity (vide infra) in

12loea

,' 14

8.

a

o4

2

27.6

1-

f 36.5

Mean- Range

TI-Soline

Controls

(134)

Aged

PlasmoControls

(126)

NormalAdults

(64)

531.4285.4

105-.9 10.

Ir IT

Exercise Nicotinic PyroeAcid

(16) (13) (S)

FIG. 2. PLASMA THROMBOLYTICACTIVITY OF NORMAL

ADULTS AND ITS VARIATION FOLLOWINGSTRESS AND DRUGADMINISTRATION. Figures in parentheses indicate thenumber of separate observations in each category. Therange of observations for each category and its mean are

displayed.

that aged human plasma was used as a secondcontrol. The results of 126 such tests are illus-trated in Figure 2, the mean release of radioac-tivity being equivalent to 3.6 jug per 2 hours.Such control determinations were made daily.

By comparison with saline, and especially withaged human plasma control values, it was possibleto assess the thrombolytic activity of fresh plasmafrom 64 normal adults. The results are illustratedin Figure 2. The range was 2.7 to 8.1 and themean 5.0 jug fibrin lysed per 2 hours. Althoughthe range of observations on normal adults over-lapped that of both controls, the mean thronmbolyticactivity of normal adult plasma was significantlygreater than either control (p < 0.001). Thesefindings indicate that unaltered plasma obtainedfrom normal individuals possesses the capacity tolyse fibrin6 under these in vitro circumstances.Presumably such plasma manifests a similar ca-pacity under in vivo conditions.

No sex difference was detected, the means formales and females being 4.9 and 5.2 ,ug per 2 hours,respectively. There was no correlation with agein the adult population studied. Thirteen pairedmorning and afternoon plasma samples providedno clear evidence of diurnal variation. Daily sam-pling in a small number of normal individuals dis-closed no striking fluctuations of plasma thrombo-lytic activity.

II. Response to exercise and to entotional stress.Enhanced thrombolytic activity following exercisehas been reported (6, 14) and the current studyis confirmatory. The range and mean thrombo-lytic activity of plasma obtained from 16 healthymales following vigorous sustained exercise (play-ing basketball) are shown in Figure 2. Twelve ofthe 16 had plasma activity considerably in excessof normal, the mean being 7 times that for normaladults. Inappropriate timing of sample collectionmay have accounted for the failure to find increasedplasma thrombolytic activity in 4 subjects.

Emotional stress has been implicated as a stim-ulus of enhanced plasma thrombolytic activity(15). Plasma samples were obtained from stu-dents before and after an examination. In-

6 In vivo thrombi and fibrinous deposits are invariablysubstantially contaminated with plasminogen (3) and thetest system itself was likewise contaminated (see Methodssection).

428

THROMBOLYTICACTIVITY OF HUMANPLASMA

1801

IS160-

o: 140-

a 120-

GD

DJ 100-

._

E 80-

CP

° 60-04,

40-

Ii4

20

,_- t- I a

40 80 120 160 200Time after Injection-Minutes

FIG. 3. COURSEOF ENHANCEDPLASMA THROMBOLYTICACTIVITY FOLLOWINGADMINISTRATION OF NICOTINIC ACID

ANDPYROGEN. Time relationships of plasma thrombolyticactivity to the intravenous injection of nicotinic acid(open circles) and pyrogen (closed circles) are illustratedfor two separate experiments. The rise in plasma throm-bolytic activity after nicotinic acid was abrupt and short-lived while that following pyrogen injection was delayedand of greater duration.

creased plasma thrombolytic activity was found in2 of 4 subjects after the examination.

III. Drug-induced changes in plasma thrombo-lytic activity. Of the several drugs known to in-duce increased plasma thrombolytic activity, ob-servations were made following administration oftwo, nicotinic acid 7 (16) and Pyrexal 8 (17).Nicotinic acid was given intravenously in a dose of1 mg per kg and in all instances induced typicalflushing. On 6 of 13 trials significantly increasedthrombolytic activity resulted. The mean andrange are shown in Figure 2. Pyrexal was givenintravenously in a dose of 0.2 ug. Antipyretics

7 Kindly supplied by Dr. R. Herting, Abbott Labora-tories, North Chicago, Ill.

x A protein-free lipopolysaccharide with potent pyro-

genicity prepared from the organism Salmonella abortusequii, kindly supplied by Dr. Schultze, Wander Company,Chicago, Ill.

were administered prior to and concurrent withpyrogen administration and at times suppressedrigor and fever customarily associated with ad-ministration of this agent. Six of 8 individualsreceiving pyrogen had increased thrombolyticactivity (range and mean are depicted in Figure2). Rigor and fever were not essential for theresponse. Neither agent, alone or when added toplasma, altered control assay values.

Typical responses to each of these agents areshown in Figure 3. The response to nicotinic acidis prompt but short-lived. With pyrogen the in-crease in activity followed a variable latent periodand was of greater duration. The reason or rea-sons for the observed failures of both agents toinduce increased activity in some patients and onsome occasions is not known, but such failureshave been reported (17, 18).

IV. Plasma thrombolytic activity of hospitalisedadults. Figure 4 illustrates the range and meanthrombolytic activity of 307 hospitalized adults.Determinations were for the most part on a singlesample from each patient drawn at random timesduring the hospital stay. Daily fluctuations of ac-tivity, described earlier in healthy persons as small,may have influenced the findings but, as indicatedby serial assays in some patients, the influence ofthis effect was probably unimportant. Only themajor diagnosis or diagnoses were considered.

The range and mean values noted in selecteddisease states are depicted in Figure 4. By com-parison with normal values it is apparent that in-creased plasma thrombolytic activity frequentlywas observed in association with hematologicmalignancies,9 liver disease and acute infections.Twenty-six, 33 and 25 per cent, respectively, ofthe patients in these categories had levels in excessof 95 per cent confidence limits (7.4 ,ug per 2hours) established from the data of normal adults.Isolated incidences of increased plasma thrombo-lytic activity were observed in patients sufferingfrom sickle cell crisis, collagen diseases, myocardialinfarction, cancer, diabetes mellitus, cerebral throm-bosis, polycythemia vera and amyloidosis.

Decreased levels of plasma thrombolytic activity(of less certain meaning in the I131 clot assay thanincreased levels) were observed in patients suffer-

Including leukemia of varied cell type, lymphomas,Hodgkin's disease, reticulum cell sarcoma and multiplemyeloma.

I

-

-

III

IIIIIIIIIIIII

429

IIIIIIIIIIIII l

160 .0I

SAWYER,FLETCHER, ALKJAERSIG AND SHERRY

16-

l4.a

-JI

N4-

, 1 2-

10a

._ ir8E

,, 6-0

4-4

" 2-

177.8 177.888.9 1

T r r 8.3

Normol Potients Hemotologic Liver AcuteAdults Malignoncy Diseose Infection(64) (307) (50) (15) (28)

Meon-{ Ronge

1; J

Cancer Heort Athero-Foilure sclerosis

(26) (21) (42)

FIG. 4. PLASMA THROMBOLYTICACTIVITY IN SELECTED DISEASE STATES.

Figures in parentheses indicate the number of separate observations in eachcategory. The range of observations for each category and its mean are

displayed. The values for normal adults are shown for comparison.

ing from a variety of disease states but most com-

monly among those with cancer, congestive heartfailure or atherosclerosis (Figure 4).

V. Characterization of plasma thrombolytic ac-

tivity. In order to define at least partially the na-

ture of the observed thrombolytic activity ofplasma, observations were made on plasma sampleswith respect to the following properties: a) de-pendence on clot plasminogen content of the testsystem, b) lability of the activity on storage, andc) the effect of E-aminocaproic acid.

a) Dependence on clot plasminogen concentra-tion of the test system. Plasma thrombolytic ac-

tivity was determined on isotopically-labeled plas-minogen-rich, plasminogen-poor, and plasminogen-free clots (see Methods section). Typical resultsappear in Table I. The thrombolytic activity was

clearly a function of clot plasminogen content, be-ing greatest with plasminogen-rich clots, less withplasminogen-poor clots and not differing signifi-cantly from control values with plasminogen-freeclots. Such dependence is characteristic of plas-minogen activators (1).

b) Lability of the activity upon storage. Thethrombolytic activity of plasma was moderately

labile on storage. Serial assays revealed 50 percent loss of activity in 1 to 3 days, which was morerapid the higher the temperature. Experimentswith urokinase, a plasminogen activator, added toplasma, indicated loss of activity at a roughly simi-lar rate. Plasmin, however, was almost immedi-ately inactivated when added to plasma in otherthan large amounts, and when added in excess, a50 per cent loss of activity occurred during a pe-riod of a few hours.

TABLE I

Effect of varying clot plasminogen concentration upon thelysis of isotopically-labeled clots immersed

in test plasma *

1ill clot assay: pg fibrin lysed/2 hrs

Plasmino- Plasmino- Plasmino-Plasma gen-rich gen-poor gen-freesample clot clot clot

1 9.8 4.2 2.82 8.0 3.4 2.43 7.9 5.5 1.74 7.0 4.8 3.05 6.6 4.5 3.26 5.9 3.4 2.2

* Clot lysis was a function of clot plasminogen concen-tration, being greatest with plasminogen-rich clots, inter-mediate with plasminogen-poor clots and at or near controlassay values with plasminogen-free clots.

430

THROMBOLYTICACTIVITY OF HUMANPLASMA

TABLE II

Inhibition of thrombolytic activity bye-aminocaproic acid

Per cent inhibition by e-amino-caproic acid

Specimen 10-4 M 103 M 102 M

Plasma 1 34 47 81Plasma 2 30 41 85Plasma 3 25 46 80Urokinase 23 64 85Human plasmin 0 6 86

The results, expressed as per cent inhibition of the testsystem without e-aminocaproic acid, illustrate the similarinhibiting effect of e-aminocaproic acid on urokinase, aplasminogen activator, and on plasma thrombolytic ac-tivity; and the different effect of e-aminocaproic acid onplasmin.

c) Effect of E-aminocaproic -acid (19). Pre-sented in Table II are data indicating significantinhibition of plasma thrombolytic activity byE-aminocaproic acid at 10-4 Mconcentration. Re-sults are expressed as per cent inhibition of theassay value in the system without E-aminocaproicacid. Urokinase was similarly inhibited. Plasmin,however, was not detectably inhibited by E-amino-caproic acid until a concentration of 10-2 M wasreached.

Moreover, the independence of thrombolytic ac-tivity and plasma proteolytic activity was apparentas thrombolytic activity was increased in plasmasamples devoid of proteolytic activity, as measuredwith the substrates, casein and urea-denatured he-

y a 70.5 2 x - 5.84

t95%Cr -37

//7117

77-,

4, ,00

0es0 * ; 00 0 00

0 L 0 : .0 0040 00

.+ 174

A

7// I-/7

// /~~~~~~~~~~~~I-

711,'l

I,,

7-,7I7-17-171117

* 7

1-1

"Z~~

I I 1

.4 .6 .8Euglobulin Clot Lysis- Units Activity

FIG. 5. CORRELATION BETWEENTHE TWOMETHODSUSED FOR THE DETERMINATION OF

PLASMATHROMBOLYTICACTIVITY; ISOTOPICALLY-LABELED HUMANPLASMACLOT LYSIS AND

EUGLOBULIN CLOT LYSIS. The results of the I" clot assay are plotted against values forthe euglobulin clot lysis method. The shaded area represents the results on 192 plasmasamples in which the euglobulin method gave results of less than 0.1 unit (see text).These values were excluded from the statistical calculations. The regression line (solid)is shown along with the residual variance about the regression line illustrated as 95 percent confidence limits (broken lines).

100

0x

N

0cx 80-10

-J

*e 60-UA.

0)

7

. .1.0 1.2

431

SAWYER, FLETCHER, ALKJAERSIG AND SHERRY

moglobin. Conversely, increased plasma proteo-lytic activity was at times observed in the absenceof enhanced thrombolytic activity. Variations inplasma thrombolytic activity were observed with-out changes in plasma fibrinogen, plasminogenor antiplasmin concentrations, confirming previousdata (6). On occasion when plasminogen acti-vator concentrations were greatly increased, e.g.,after drug administration, moderate falls of plasmaplasminogen concentrations were recorded.

VI. Correlation of the I131-labeled clot assaywith plasma euglobulin clot lysis. The plasma eu-globulin clot lysis test has been frequently used asa measure of overall plasma "fibrinolytic" activity(6, 7) but more recently has been shown to be atleast a qualitative measure of plasma activator con-centration (6). Consequently it was of interestto attempt to correlate the results of the euglobu-lin test with the 1131-labeled clot lysis assay. Bothdeterminations were made on 302 plasma samplesdrawn under a variety of circumstances. The plotof results is shown in Figure 5. The euglobulinclot lysis time was expressed in units of activity.Lysis times greater than 300 minutes (customarilyexpressed as "normal") were recorded as less than0.1 unit. Such values were observed in a ma-jority of the samples, especially those in the rangeof normal adults as determined by the J131 clotmethod. In 110 instances the euglobulin techniqueresulted in values greater than 0.1 unit and it wasfrom these samples that statistical evaluation wasmade. The correlation coefficient is 0.677 which issignificant at the 0.001 level. However, the valuesfor the correlation coefficient and the residual vari-ance about the regression line (illustrated as 95 percent confidence limits in Figure 5) indicate that thecorrelation, although statistically significant, isnoticeably imperfect so that individual pairedreadings may be markedly divergent.

The significant although imprecise correlationbetween the isotopically-labeled human plasma clotassay and the plasma euglobulin clot lysis techniquefurther validates the euglobulin method as a simplyperformed qualitative test for enhanced thrombo-lytic activity (6, 7). However, one of the sev-eral disadvantages of the euglobulin test is dem-onstrated in Figure 5 in which over half the read-ings positive for low levels of thrombolytic activityby the I13'-labeled clot method were within the"normal limits" for the euglobulin test (6, 7).

DISCUSSION

The method for measurement of thrombolyticactivity used in this communication tests the ca-pability of freshly drawn unaltered plasma to lysea preformed clot. It closely mimics conditions invivo in which unaltered plasma will pass in contactwith thrombi and other fibrinous deposits. Con-sequently, the present observations, especially in-sofar as they refer to the consistent finding ofthrombolytic activity in plasma from normal adults,may be regarded as being physiologically relevantto in vivo thrombolytic mechanisms.

Variations in levels of thrombolytic activitywere demonstrated in health, following exercise,after drug administration, and in disease. Varia-tions in health were apparently small, but afterexercise, high levels of thrombolytic activity werefound, an observation confirming that of others (6,14). Similarly, the administration of nicotinicacid and pyrogens resulted in sharp rises in plasmathrombolytic activity (16, 17).

The data presented pertaining to disease, al-though limited, serve to focus attention on certaindisease states in which altered thrombolytic ac-tivity could be of significance. The instances ofincreased thrombolytic activity in patients withliver disease and hematologic malignancies con-firm previous reports (20-25) and the incidencesuggests that increased thrombolytic activity inthese states is more frequent than is commonlythought. Increased thrombolytic activity was pres-ent in some individuals with acute infections.Whether this is due to pyrexia and toxicity or isa part of the inflammatory response per se is un-known.

The suggestion of diminished thrombolytic ac-tivity in some patients suffering from cancer andcongestive heart failure is noteworthy in that uro-kinase is excreted in diminished quantity in theseconditions (26). Decreased thrombolytic ac-tivity could relate to the frequent thromboem-bolic complications of these states.

The evidence presented suggests that the throm-bolytic activity of plasma from normal adults andfrom individuals with disease is due primarily tothe presence of plasminogen activator. Two po-tential thrombolytic agents, plasmin and plasmi-nogen activator (exemplified by urokinase) whichcould account for the observed activity, were com-

432

THROMBOLYTICACTIVITY OF HUMANPLASMA

pared with plasma thrombolytic activity in termsof three different characteristics. In all three re-spects, dependence on clot plasminogen content,lability on storage, and inhibition by e-amino-caproic acid in 104 M concentration, plasmathrombolytic activity resembles plasminogen ac-tivator. Thus the present data both confirm andextend to other areas recent work on in vivo and invitro thrombolytic mechanisms (1, 27) indicatingthe critical role of plasminogen activator inthrombolysis.

The significance of the present observations liesin the provision of direct experimental data in aregion in which the evidence has, for technicalreasons, been largely inferential. Hitherto, lowlevels of thrombolytic activity could only be as-sayed by indirect techniques that disturbed thebalance of plasminogen system components inplasma and consequently the interpretation of suchassays was uncertain (3). The present methodfor assay of thrombolytic activity is direct, em-ploys unaltered plasma, is sufficiently sensitive fordetection of low levels of activity, and the interpre-tation of results is unambiguous. Consequently,the finding that plasma of normal individualspossesses demonstrable thrombolytic activity un-der these test conditions would seem to be of con-siderable significance in the elucidation of in vivothrombolytic phenomena. In the past it has beensuggested that the plasminogen-plasmin systemmay be in a state of dynamic activity (28, 29) andthe present evidence, obtained with acceptable tech-nical procedures, provides strong support for thiscontention.

The demonstration that increased plasma ac-tivator concentrations may exist without detectablechange in plasma fibrinogen, plasminogen or anti-plasmin (6) suggests either that plasma containsan inhibitor to plasminogen activator or thatplasma activator, because of physicochemical con-ditions, acts preferentially to activate thrombusplasminogen rather than plasma plasminogen.The evidence is, however, no more than suggestiveand work on this problem has been scanty (3).

It is of incidental interest that levels of plasmathrombolytic activity may be increased severalhundredfold over those found in health by theintravenous infusion of purified plasminogen ac-tivator (27), a procedure used for therapeuticpurposes (30).

SUMMARY

A sensitive measure of plasma thrornbolytic ac-tivity has been described which involves deter-mination of the release of radioactivity from iso-topically-labeled human plasma clots immersed inunaltered plasma.

Using the isotopic clot assay, thrombolytic ac-tivity was determined in plasma from healthyadults, from adults following stress and follow-ing the administration of drugs, and from individu-als with disease.

The results indicate that plasma from normaladults contains a plasminogen activator capable oflysing human plasma clots under conditions simi-lar to those in vivo and that the quantity of thismaterial varies in response to stress and drug ad-ministration and in disease.

ACKNOWLEDGMENTS

We record with pleasure our indebtedness to MissGeraldine Nelson, Mrs. Odessa Hill and Mr. John C.Walter for technical assistance. We also wish to thankthe physicians of Barnes Hospital who permitted us tostudy their patients.

REFERENCES

1. Alkjaersig, N., Fletcher, A. P., and Sherry, S.The mechanism of clot dissolution by plasmin. J.clin. Invest. 1959, 38, 1086.

2. Astrup, T. Fibrinolysis in the organism. Blood1956, 11, 781.

3. Sherry, S., Fletcher, A. P., and Alkjaersig, N. Fi-brinolysis and fibrinolytic activity in man. Physiol.Rev. 1959, 39, 343.

4. Biggs, R., and Macfarlane, R. G. Human Blood Co-agulation and Its Disorders, 2nd ed. Oxford,Blackwell, 1957, p. 11.

5. Fearnley, G. R., Revill, R., and Tweed, J. M. Ob-servations on the inactivation of fibrinolytic activityin shed blood. Clin. Sci. 1952, 11, 309.

6. Sherry, S., Lindemeyer, R. I., Fletcher, A. P., andAlkjaersig, N. Studies on enhanced fibrinolyticactivity in man. J. clin. Invest. 1959, 38, 810.

7. Kowalski, E., Kopec, M., and Niewiarowski, S. Anevaluation of the euglobulin method for the de-termination of fibrinolysis. J. clin. Path. 1959,12, 215.

8. Greig, H. B. W., and Runde, I. A. Studies on theinhibition of fibrinolysis by lipids. Lancet 1957, 2,461.

433

SAWYER,FLETCHER, ALKJAERSIG AND SHERRY

9. Kline, D. L. The purification and crystallization ofplasminogen (profibrinolysin). J. biol. Chem.1953, 204, 949.

10. Alkjaersig, N., Fletcher, A. P., and Sherry, S. Theactivation of human plasminogen. L Spontaneousactivation in glycerol. J. biol. Chem. 1958, 233, 81.

11. Brownlee, K. A. Industrial Experimentation, 4th ed.London, Her Majesty's Stationery Office, 1949.

12. Anson, M. L. The estimation of pepsin, trypsin, pa-pain, and cathepsin with hemoglobin. J. gen.Physiol. 1938, 22, 79.

13. Ratnoff, 0. D., and Menzie, C. A new method forthe determination of fibrinogen in small samplesof plasma. J. Lab. clin. Med. 1951, 37, 316.

14. Biggs, R., Macfarlane, R. G., and Pilling, J. Ob-servations on fibrinolysis. Experimental activityproduced by exercise or adrenaline. Lancet 1947,1, 402.

15. Macfarlane, R. G., and Biggs, R. Observations onfibrinolysis. Spontaneous activity associated withsurgical operations, trauma, etc. Lancet 1946, 2,862.

16. Weiner, M., Redisch, W., and Steele, J. M. Oc-currence of fibrinolytic activity following adminis-tration of nicotinic acid. Proc. Soc. exp. Biol.(N. Y.) 1958, 98, 755.

17. von Kaulla, K. N. Intravenous protein-free pyrogen.A powerful fibrinolytic agent in man. Circulation1958, 17, 187.

18. Weiner, M., DeCrinis, K., and Redisch, W. Adapta-tion to nicotinic acid induced fibrinolysis. Fed.Proc. 1959, 18, 168.

19. Alkjaersig, N., Fletcher, A. P., and Sherry, S.e-Aminocaproic acid: An inhibitor of plasminogenactivation. J. biol. Chem. 1959, 234, 832.

20. Cooperberg, A. A., and Neiman, G. M. Fibrino-genopenia and fibrinolysis in acute myelogenousleukemia. Ann. intern. Med. 1955, 42, 706.

21. Pisciotta, A. V., Hinz, J. E., and Schulz, E. G.Fibrinolytic purpura in acute leukemia. Clin.Res. Proc. 1955, 3, 9.

22. Dameshek, W., and Gunz, F. Leukemia. New York,Grune and Stratton, 1958, p. 217.

23. Goodpasture, E. W. Fibrinolysis in chronic hepaticinsufficiency. Johns Hopk. Hosp. Bull. 1914, 25,330.

24. Ratnoff, 0. D. Studies on a proteolytic enzyme inhuman plasma. IV. Rate of lysis of plasma clotsin normal and diseased individuals, with particularreference to hepatic disease. Johns Hopk. Hosp.Bull. 1949, 84, 29.

25. Kwaan, H. C., McFadzean, A. J. S., and Cook, J.Plasma fibrinolytic activity in cirrhosis of theliver. Lancet 1956, 1, 132.

26. Smyrniotis, F. E., Fletcher, A. P., Alkjaersig, N.,and Sherry, S. Urokinase excretion in health andits alterations in certain disease states. Thromb.Diath. haem. 1959, 3, 257.

27. Fletcher, A. P., Alkjaersig, N., and Sherry, S. Themaintenance of a sustained thrombolytic state inman. I. Induction and effects. J. clin. Invest. 1959,38, 1096.

28. Astrup, T. The biological significance of fibrinolysis.Lancet 1956, 2, 565.

29. Sherry, S., Fletcher, A. P., and Alkjaersig, N. Thefibrinolysin system: Some physiological considera-tions in Connective Tissue, Thrombosis andAtherosclerosis, I. Page, Ed. New York, Aca-demic Press Inc., 1959, p. 241.

30. Fletcher, A. P., Sherry, S., Alkjaersig, N., Smyrni-otis, F. E., and Jick, S. The maintenance of a sus-tained thrombolytic state in man. II. Clinical ob-servations on patients with myocardial infarctionand other thromboembolic disorders. J. clin. In-vest. 1959, 38, 1111.

434