Proc. Nat. Acad. Sci. USAVol. 70, No. 10, pp. 2814-2818, October 1973

Plasma Aggregating Factor (Bovine) for Human Platelets: A Marker forStudy of Antihemophilic and von Willebrand Factors

(macromolecules and hemostasis/platelet "receptor sites"/factor VIII/platelet adsorption/chromatography)

THOMAS R. GRIGGS, HERBERT A. COOPER, WILLIAM P. WEBSTER, ROBERT H. WAGNER, ANDK. M. BRINKHOUS

Department of Pathology, School of Medicine, University of North Carolina, Chapel Hill, N.C. 27514

Contributed by K. M. Brinkhous, June 20, 1973

ABSTRACT The human-platelet aggregating factorin preparations of bovine plasma and bovine antihemo-philic factor has been reported to be an activity of bovineantihemophilic factor. However, we show that activities ofaggregating factor and antihemophilic factor are presentin varying proportion in different preparations of anti-hemophilic factor. Aggregating factor activity is morestable on storage than is antihemophilic factor activityand can be selectively adsorbed by human platelets.Further, antihemophilic factor activity free of aggregatingactivity was eluted from DEAE-Sephadex. The factorswere also separated during agarose chromatography in0.25 M CaCl2, which yields a small-molecular-weightmoiety with antihemophilic factor activity; aggregatingfactor activity continued to elute in the void volume.Thus, aggregating factor activity is clearly separable fromantihemophilic factor activity. The usual physiologicplatelet-aggregating reactions require fresh platelets.Fresh, aged, and sonicated platelets, but not plateletsthat have been frozen and thawed, react with aggregatingfactor, indicating that the platelet "receptor sites" for theaggregating factor are unusually stable.

Aggregation of blood platelets is induced by several physio-logic substances. Divalent cation is needed for these sub-stances to be effective, and fibrinogen is an essential cofactorfor several platelet-aggregating reactions (1, 2). Another com-ponent of the clotting system, antihemophilic factor (AHF,factor VIII), has been suggested as being concerned withplatelet function (3). There is considerable species variationin the response of platelets to various aggregating agents (4,5). In the platelet-aggregating reaction that occurs whenhuman platelets are mixed with bovine plasma without addedcation, the aggregating activity appears to be concentratedin bovine-plasma fractions rich in fibrinogen and AHF. Thisreaction may further elucidate the role of certain plasma fac-tors, e.g., fibrinogen, AHF, and "von Willebrand factor"(vWF, "vascular" factor), in the hemostatic function ofplatelets.The bovine-aggregating factor for human platelets was

recognized when concentrates of animal AHF were first usedfor treatment of human hemophilia (6). Thrombocytopeniaoccurred promptly. The aggregating agent was believed to bea heterologous antibody (6). After the demonstration thatfibrinogen is an essential cofactor for several types of plateletaggregation, it was suggested that the fibrinogen content orthrombocyte-agglutinating activity (TAg) of preparations of

animal AHF was responsible for their platelet-aggregatingeffect (7, 8). With separation of fibrinogen from AHF by gelchromatography, the aggregating agent was found to elutewith the AHF activity. These findings, along with othercharacteristics of the aggregating factor (7, 8), led to the sug-gestion that the AHF activity and the platelet-aggregatingactivity were identical (9, 10).The studies reported here re-examine the association of

bovine platelet-aggregating factor (PAF) and AHF. Separa-tion of these two activities in bovine plasma and plasma frac-tions by several procedures is reported. Recent data on thecomplex nature of plasma AHF, in combination with the datareported here, indicate that PAF activity is distinct from theisolated AHF activity, but is closely associated with and couldeven be a function of a large carrier protein which binds asmaller AHF moiety (11).

MATERIALS AND METHODS

Bovine Plasma was obtained from blood collected in 3.2%trisodium citrate during slaughter of adult animals. The bloodwas immediately stored on ice and transported to the labora-tory, where platelet-poor plasma was promptly prepared andfractionated or frozen at -20°. Samples were thawed at 370immediately before use.

Concentrates of Bovine-Plasma AHF were of three types: (1)"AHG (bovine)," obtained from S. Maw Son and Sons, Ltd.,England; (2) glycine-precipitated (Wagner method) (12),prepared to our specifications by Hyland Laboratories, LosAngeles; (3) a highly purified concentrate prepared by a three-step procedure consisting of (a) plasma concentration bydialysis against polyethylene glycol, (b) precipitation at 00,and (c) agarose chromatography (13).

Suspensions of Human Platelets were used for adsorptionstudies and assay of bovine PAF activity. They were preparedby two different procedures: (1) Gel-filtered platelets weremade by a modification of the method of Tangen and Berman(14). Human plasma rich in platelets was prepared from wholeblood drawn from normal, fasting adult men, and on oneoccasion from a man with severe classic hemophilia. For someexperiments, platelets in suspension were disrupted ultrasoni-cally at 00 (Biosonic III ultrasonic generator, BromwellScientific Co., set at 60, for 60 sec). (2) Washed platelets wereprepared from platelet concentrates (15) that had been madeat the North Carolina Memorial Hospital Blood Bank 48 hr

2814

Abbreviations: AHF, antihemophilic factor, factor VIII; PAF,platelet-aggregating factor.

Dow

nloa

ded

by g

uest

on

Apr

il 10

, 202

0

Human Platelet Aggregation and Bovine Factor VIII 2815

previously. These concentrates were pooled in siliconized glasstubes and centrifuged at 1000 X g for 10 min. The plateletswere washed three times in a solution of 3.2% trisodium citratediluted 1:20 with saline. For some experiments, several plate-let buttons were frozen quickly in a dry ice-acetone bath andthen stored at -20° for 24 hr before use. These suspensionswere designated "freeze-thaw" platelets.Adsorption of Bovine Plasma Preparations with Human

Platelet Suspensions. When gel-filtered platelet suspensionswere used, platelet buttons were prepared by centrifugation of2 ml of suspension as it came from the column (1000 X g, 3min). The supernatant was discarded, and the platelets weresuspended in 5 ml of dilute preparations of bovine AHF. Thesuspension was gently agitated (5 min, 25°) and then cen-trifuged as above; 0.3 ml of the supernatant was removed forassays and the remainder was added to another platelet but-ton. This procedure was repeated up to six times. Because ofof the small number of platelets available in these suspensions,dilute samples of AHF concentrates (0.02-0.1 bovine AHFunits per ml) were used. The adsorption procedure was thesame with sonicated platelets, except that after sonication thebuttons were prepared by ultracentrifugation (20,000 X g,20 min, 4°) and were washed once before they were used foradsorption, and only 1 ml of dilute AHF preparation wasadded.When washed platelet suspensions were used, the number

of platelets and the concentration of AHF in each adsorptionstep were much greater than in the adsorption studies with gel-filtered platelets. Each platelet button (1000 X 9, 10 min)contained the platelets from 1 unit of blood. AHF solutionsused contained about 1 unit/ml. The procedure otherwise wassimilar to that described above.

Chromatography. Both ion-exchange and agarose columnchromatography were used to study the AHF preparations.(1) DEAE-Sephadex A-50 (Pharmacia) was swollen in waterand charged with 0.5 M HCl and then washed with 0.5 MNaOH and water. The gel, after washing, was equilibratedin 0.1 M Tris HCl (pH 7.0). The gradient [0.1-1.0 M TriseHCl (pH 7.0) ] was delivered with a described apparatus (16).(2) Agarose gel (Bio-Gel A-15m, 4% agarose beads, 200-400mesh, Bio-Rad Laboratories) was used. Column void volumeswere determined with Blue Dextran 2000 (Pharmacia). Buf-fers were Tris-buffered saline [50 mM Tris * HCl-0.15 M NaCl(pH 7.35) ] and Tris-calcium dissociation buffer [50mM TrisHCI-10 mM NaCl-0.25 M CaCl2 (pH 7.35) ].Assay Methods. AHF in bovine-plasma preparations was

assayed by a modification of the partial thromboplastin timeassay (17) with a kaolin activated system and substrateplasma from the Chapel Hill colony of hemophilia A dogs. Afurther modification was used when fractions were obtained inthe Tris-calcium dissociation buffer (18). One unit of AHFactivity is defined as that amount in 1 ml of normal bovineplasma. The identity test for AHF consisted of a modifiedneutralization procedure (19) in which plasma from a humanpatient with hemophilia A with a high titer of circulatinganticoagulant, an antibody against AHF, was incubated withthe test material,PAF activity in bovine-plasma preparations was assayed

by a modification of the macroscopic aggregating test (20).The test consisted of 0.2 ml of gel-filtered platelet suspension

TABLE 1. Ratio of antihemophilic factor to platelet-aggregating factor in bovine-plasma preparations

AHF PAF Ratio(units/ (units/ AHF/

Preparation ml) ml) PAF

Bovine plasmaPlasma 1.0 1.0 1.0Plasma stored 48 hr, 28° 0.01 1.0 0.01

Bovine AHF concerdrates*High purity (13) 8.0 4.0 2.0High purity (13)

Stored 24 hr, 280 1.8 1.9 0.95Intermediate purity(Maw AHG, bovine) 0.9 2.5 0.36

Intermediate purity(glycine-precipitated) 3.0 5.0 0.60(12)

* Refers to dilute solutions being assayed-

0.2 ml of bovine-plasma preparation (serially diluted withTris-buffered saline). Macroscopic aggregation time was mea-sured with a stop watch. The degree of aggregation was thendetermined microscopically at 120 sec. Platelet aggregatesthat contained more than 100 platelets per aggregate wererecorded as 4+. Titer values represent the greatest dilutionthat caused 4+ aggregation within 120 sec. One unit of PAFactivity is that amount present in 1 ml of normal bovineplasma.

RESULTS

Comparison of Concentration of AHF and PAF in BovinePlasma and Plasma Concentrates. Bovine plasma and varioustypes of AHF concentrates of bovine plasma were tested fortheir relative content of AHF and PAF activity. The resultsof illustrative experiments for each preparation tested areshown in Table 1. The relative concentrations of the twoactivities were arbitrarily considered as 1:1 in normal bovineplasma. In the AHF concentrate of high purity (purified2000-4000 times in relation to plasma AHF and plasma pro-tein), AHF was present in a relatively greater amount thanPAF. The converse was true for the two preparations ofintermediate purity, with the relative concentration of PAFbeing greater. PAF was more stable than AHF when plasmaor concentrate was stored at room temperature.

Selective Adsorption of PAF Activity from Bovine Plasmaand AHF Concentrates. PAF could be selectively removedfrom the preparations by adsorption with human platelets(Table 2). However, the adsorbed preparations, free of PAF,still contained AHF. All types of suspensions used were ef-fective in removing PAF except for the freeze-thaw type.The platelet suspensions were used from 2-96 hr after collec-tion of blood. They were effective as an adsorbing agent re-gardless of time of storage. Each of the preparations of bo-vine-plasma AHF listed in Table 1 was used in the adsorptionexperiments. Most of the experiments were performed withthe glycine-precipitated material. The number of adsorptionsrequired to remove PAF depended on the concentration ofplatelets and the concentration of PAF in the starting solu-

or platelet-rich plasma (400,000/mm3) to which was added tion. PAF was reduced with each successive platelet-adsorp-

Proc. Nat. Acad. Sci. USA 70 (1973)

Dow

nloa

ded

by g

uest

on

Apr

il 10

, 202

0

Proc. Nat. Acad. Sci. USA 70 (1973)

TABLE 2. Adsorption of aggregating factor (PAF) byhuman platelets

N AHF (units/mi)No. X 1012 PAFof afterad- Before After ad-

Human platelet Platelets/ sorp- adsorp adsorp- sorp-preparation mm3 tions tion tion tion

Gel-filtered (3) 400,000 1-4 1 1-3 NegHemophilic,

gel-filtered (1) 400,000 6 1 3 NegWashed, aged48-96 hr (3) 1000,OOO 2-5 100300 10-100 Neg

Washed (freeze-thaw) (2) 1,000,000* 5 t t 4+

Sonicated (2) 400,000* 1 1 1 Neg

Numbers in parentheses indicate number of experiments done.* Counts made before freeze-thaw or sonication step.t Not measured.

tion step until, finally, it could not be detected. In one experi-mnent, all the PAF was removed by a single adsorption proce-dure, but a very dilute bovine preparation with a concen-trated platelet suspension was used.

Quantitation of AHF after adsorption with platelets gavevariable results (Table 2) apparently because of interferingsubstances in the bioassay preparation resulting from expo-sure of the plasma and concentrates to large numbers ofplatelets. However, incubation with human antibody againstAHF resulted in prolongation of assay clotting times, veri-fying the presence of AHF.

Retention of PAF Activity After Neutralization of AHFActivity. After incubation of PAF-containing preparationswith a specific AHF-inhibitinig antibody in human hemophilicplasma, PAF activity was retained while AHF activity waslost. In two experiments, one with plasma and the other withglycine-precipitated preparation, AHF activity was com-pletely lost after 30 min of incubation, while the titer of PAFwas unchanged. Others have made similar observations (8,10).

Separation of PAF from Bovine AHF by Chromatography.The aggregating factor and its relationsip to bovine AHFwere better characterized in. three sets of chromatographicexperiments. DEAE-Sephadex chromatography was firstused (Fig. 1). The gradient between 0.2 and 0.5 M Tris - HCleluted fractions with a broad peak of activity containingboth AHF and PAF and a later peak of AHF activity withlittle or no PAF.

In a second group of experiments, agarose column chroma-tography was used to compare AHF elution behavior beforeand after removal of PAF from plasma preparations by ad-sorption by human platelets. Fig. 2A shows typical resultswhen unadsorbed bovine plasma is chromatographed onagarose. AHF and the PAF appear in the fractions in or closeto the void volume and, therefore, are of very high molecularweight. Fig. 2B shows the results of chromatography of asample of bovine plasma from which the PAF activity hadbeen removed by adsorption with washed platelets, compara-ble to the adsorbed sample (washed, aged) of Table 2. Again,AHF activity was excluded from the gel, and therefore wasof large size, even though PAF did not accompany it.

00.4

0.2

50 100 150 200 250 300 350 400 450EFFLUENT VOLUME (ml)

FIG. 1. Elution pattern of AHF and PAF on DEAE-Seph-adex A-50 ion-exchange chromatography. A 15-ml sample of aglycine-precipitated preparation of bovine AHF was applied toa column, 1. X 25 cm, at 23° and followed by 150 ml of initialTris buffer before the gradient was started. Flow rate was 20ml/hr. AHF, *, Tris * HC1, O.

In the third group of experiments, AHF in highly purifiedpreparations of bovine AHF was chromatographed in 0.25 MCaCl2, a concentration that dissociates canine AHF into asmall active molecule and a large inactive molecule (i8). Re-sults of an illustrative experiment are shown in Fig. 3. Dis-sociated AHF eluted from the column in late fractions, whilePAF was found in the void volume.

DISCUSSION

In this paper we show by several procedures that the AHFactivity and the PAF activity for human platelets of bovineplasma can be separated. Previous work has emphasized thesimilarities of the AHF and the bovine PAF (9, 10). Thesestudies pointed out the comparable lability of the two factorswhen exposed to heat (56i60o) and during clotting of bloodand the similar behavior of the two activities on agarose gelchromatography, both eluting as large-molecular-weight en-tities in the void volume. The unequal partition of PAF andAHF activities in different preparations of bovine plasma andplasma-concentrates first suggested to us that they might beseparable entities. There is a higher ratio of PAF to AHF inall of the bovine-plasma concentrates except for the fresh,highly purified preparation (Table 1). These findings couldin part be due to the greater lability of AHF, noted both instored plasma and in stored highly purified concentrate. Thegreater concentration of AHF than PAF in the bovine AHFpreparation of high purity does suggest that the two activitiesmight be separated by appropriate procedures. However,present methods for purification of AHF are inadequate forthis purpose, and it is impossible to prepare nonthrombocyto-penic bovine AHF concentrates for therapeutic use.

All the measurable PAF can be removed from bovine-plasma preparations by adsorption with human platelets,with AHF activity remaining (Table 2). The effectiveness ofthe adsorption procedure was related to the concentration ofthe platelets and the number of times the adsorption proce-dure was repeated, as well as to the concentration of PAF inthe preparations. Calculations of the number of humanplatelets required to remove the aggregating activity from theequivalent of 1 ml of bovine plasma gave a very large value,about 1 X 109-8 X 109 human platelets per mm' of bovineplasma. This is about 10,000-times more than the number of

2816 Pathology: Griggs et cd.

Dow

nloa

ded

by g

uest

on

Apr

il 10

, 202

0

Human Platelet Aggregation and Bovine Factor VIII 2817

L4-

5 10 15 20 25VoEFFLUENT VOLUME (ml)

12

10-B.I

08

o~~~~V

10 20 30 40 50 60Va EFFLUENT VOLUME (ml),

FIG. 2. Comparison of elution pattern before and after ad-sorption of bovine plasma with human platelets. (A) Agarose 15mchromatography of bovine plasma. 0.5 ml of plasma was layeredonto a 1 X 24-cm column and eluted with Tris-buffered saline at4.5 mI/cm' per hr at 230. (B) Agarose 15m chromatography ofbovine plasma adsorbed with human platelets. A 5-ml sample ofbovine plasma was serially adsorbed with washed human plateletsuntil no aggregating activity could be detected in the supernatant.This material was layered onto a 1.5 X 30-cm column and elutedwith Tris-buffered saline at 8 ml/cm' per hr at 23°. Void volume,Vo.

platelets normally present in that volume of plasma, eitherhuman or bovine.

Platelets are very labile cells, and in many of the physiologi-cal aggregating reactions, e.g., with ADP and thrombin, theyrapidly lose their ability to aggregate with ageing (21); intactplatelets are required. For the PAF reaction, in contrast(Table 2 h "receptor sites" on the platelets appear to berelatively stable, and fragmented platelets retain their abilityto react. Only in freeze-thaw platelets was there loss of ag-gregating and adsorbing capacity.

Since variable results were obtained in several AHF bio-assays of bovine preparations after they had been exposed tosuch large numbers of platelets, several control experimentswere done to verify that the clotting activity after adsorptionwas not due solely to some platelet derivative or other con-taminant. Neutralization tests with a human antibody againstAHF demonstrated that AHF was present after adsorption.Gel-filtered human platelets from a hemophilic donor, devoidof AHF, were used in one experiment (Table 2) to show thatthe AHF present after adsorption was not of human origin.Again, PAF was selectively removed and AHF activity re-mained. Finally, agarose-gel chromatography of the materialafter adsorption showed that AHF freed of PAF continued tobehave as before, eluting in the void volume (Fig. 2B). Theseexperiments would indicate that the AHF that remained afterplatelet adsorption was not greatly altered, if at all, in regard

'~8

2 -

PAF:4+'

20 40 60 0 100 120 140 160 180 200 220Vo

EFFLUENT VOLUME (ml)

FIG. 3. Separation of PAF and AHF by dissociation of bovineAHF with CaCl,. 15 ml of an AHF concentrate (Owen-Wagner)was brought to 0.25 M in CaCl2 and chromatographed on a 2.5 X40-cm agarose 15m column equilibrated with Tris-calcium bufferwith flow rate 8 ml/cm2 per hr at 40. Assays for PAF were per-formed after dialysis of selected fractions against Tris-bufferedsaline for 12 hr at 4°. Void volume, Vo.

to clotting activity, reactivity with a specific antibody, andelution behavior during gel chromatography.The second approach to separation of AHF and PAF was

neutralization of AHF activity with inhibitor antibody. PAFactivity was not affected by this procedure, which yieldsPAF without AHF (8, 10). In these studies, as in the plateletadsorption experiments, one of the activities was retained atthe expense of the other.A last set of experiments with ion-exchange and agarose

chromatography was aimed at separation of PAF and AHFwithout loss of either of the activities. A requirement for sep-aration of the two activities by chromatography washigh ionic strength in the solution. With the DEAE-Sephadex column (Fig. 1), PAF and AHF appeared togetherin the midportion of the gradient. Unlike AHF, PAF appearedto be completely eluted at 0.2-0.5 M Tris - HCl. At higher ionicstrength only AHF appeared. If one assumes that an AHFcomplex is dissociated at these high ionic strengths, as hasbeen shown with canine and humanAHF (11, 18, 22, 23), thenthe AHF associated with PAF in the earlier parts of the gra-dient is in the large-molecular-weight, complexed form, whilein the latter part of the gradient, the PAF-free AHF is in thelow-molecular-weight, dissociated form.With agarose-gel chromatography, AHF was dissociated

with high concentrations of CaCl2. After dissociation, PAFeluted in the void volume, while AHF, free of PAF activity,was eluted later. These data differ from those of Donati et al.(9),. probably because of different experimental conditions.A common view is that the antihemophilic factor is of ex-

tremely large molecular weight, perhaps up to two million,and is possibly made up of repeating identical subunits (22-25). Several of us (11, 18) have presented evidence that AHFactivity is associated with a small-molecular-weight entity,as judged by chromatography, which is complexed with alarge carrier protein. The usual physicochemical attributesof AHF as found in plasma are due to the complex betweencarrier and active small subunit.Perhaps this AHF complex is associated with another

plasma factor, the "von Willebrand factor," which is neededfor platelet adhesion and normal hemostasis. Bouma et al.

Proc. Nat. Acad. Sci. USA 70 (1973)

Dow

nloa

ded

by g

uest

on

Apr

il 10

, 202

0

Proc. Nat. Acad. Sci. USA 70 (1973)

ONE MACROMOLECULETHREE FUNCTIONS

CP-AHFPAFvWF

vWF IPAF

TWO MACROMOLECULES o oTHREE FUNCTIONS PAF r or

vWF PAF vWF

CP-HF

THREE MACROMOLECULESTHREE FUNCTIONS

FIG. 4. Possible molecular and functional relationships ofAHF carrier protein, PAF, and "von Willebrand factor," ac-

cording to number of separate macromolecular moieties. CP-AHF, carrier protein for small subunit of AHF; vWF, "vonWillebrand factor."

(26) have shown that "von Willebrand factor" behaves likethe AHF complex on agarose column chromatography, elutingin the void volume. Thus, it too, like the AHF complex andPAF, is a macromolecule. There are then three functionalactivities that appear associated with the macromolecular or

void-volume fractions that bind the small subunit of AHF andplatelet membrane. PAF activity is a biological marker thatshould help in the separation and further characterization ofAHF and "von Willebrand factor." Fig. 4 summarizes some

of our speculations about the molecular identity or nonidentityof AHF carrier protein, bovine PAF and "von Willebrandfactor." While the postulation of a single macromolecule withmultiple functions is attractive on a biological economy basis,our platelet adsorption data taken at face value make thisunlikely. The three-macromolecules hypothesis, one macro-

molecule-for each separate function, is also compatible withthe data.Two of the functions, PAF and "von Willebrand factor,"

are expressed by reacting with platelets. If they were mo-

lecularly identical, the "'packaging" possibilities would begreatly simplified. There could be one molecule with two func-tional sites, one with AHF binding capacity and another withplatelet binding capacity. Alternatively there could be twoseparate molecules, one for each of these sites.

This work was supported by NIH Grants HL-01648 and HL-14228, SCOR Thrombosis Center, National Heart and -LungInstitute. T.R.G. was supported by Training Grant GM-00092

from the National Institute of General Medical Sciences. H.A.C.was supported by Training Grant HL-05652 from the NationalHeart and Lung Institute,

1. McLean, J. R., Maxwell, R. E. & Hertler, D. (1964) Nature202, 605-606.

2. Brinkhous, K. M., Read, M. S. & Mason, R. G. (1965)Lab. Invest. 14, 335-342.

3. Brinkhous, K. M. (1947) Proc. Soc. Exp. Biol. Med. 66,117-120.

4. LeRoy, E. C., Mason, R. G. & Brinkhous, K. M. (1960)Amer. J. Physiol. 199, 183-186.

5. Mason, R. G. & Read, M. S. (1967) 4xp. Mol. Pathol. 6,370-381.

6. Macfarlane, R. G., Biggs, R. & Bidwell, E. (1954) Lancet i,1316-1319.

7. Solum, N. 0. (1968) Scand. J. Haematol. 5, 474485.8. Forbes, C. D., Barr, R. D., McNicol, G. P. & Douglas, A. S.

(1972) J. Clin. Pathol. 25, 210-217.9. Doilati, M. B., de Gaetano, G. & Vermylen, J. (1973)

Thromb. Res. 2, 97-104.10. Forbes, C. D. & Prentice, C. R. M. (1973) Nature New Biol.

241, 149-150.11. Cooper, H. A., Griggs, T. R. & Wagner, R. H. (1973)

Proc. Nat. Acad. Sci. USA 70, 2326-2329.12. Wagner, R. H., McLester, W. D., Smith, M. & Brinkhous,

K. M. (1964) Thromb. Diath. Hemorrh. 11, 64-74.13. Owen, W. G. & Wagner, R. H. (1972) Thromb. Res. 1, 71-

87.14. Tangen, O., Berman, H. J. & Marfey, P. (1971) Thromb.

Diath. Hemorrh. 25, 268-278.15. Amer. Assoc. Blood Banks (1970) in Technical Methods and

Procedures (Chicago), 5th ed., pp. 184-187.16.' Fahey, J. L., McCoy, P. F. & Goulian, M. (1958) J. Clin.

Invest. 37, 272-284.17. Langdell, R. D., Wagner, R. H. & Brinkhous, K. M. (1953)

J. Lab. Clin. Med. 41, 637-647.18. Owen, W. G. & Wagner, R. H. (1972) Thromb. Diath.

Haemorrh. 27, 502-515.19. Roberts, H. R., Scales, M. B., Madison, J. T., Webster,

W. P. & Penick, G. D. (1965) Blood 26, 805-818.20. Brinkhous, K. M., LeRoy, E. C., Cornell, W. P., Brown,

R. C., Hazlehurst, J. L. & Vennart, G. P. (1958) Proc.Sac. Exp. Biol. Med. 98, 379-383.

21. Brinkhous, K. M., Mason, R. G., Read, M. S. & Shermer,R. W. (1968) in Metaboliam and Membrane Permeability ofErythrocytes and Thrombocytes, eds. Deutsch, E., Gerlach,E. & Moser, K. (Georg Thieme Verlag, Stuttgart), pp. 302-307.

22. Thelin, G. M. & Wagner, R. H. (1961) Arch. Biochem.Biophys. 95, 70-76.

23. Weiss, H. J., Phillips, L. L. & Rosner, W. (1972) Thromb.Diath. Hemorrh. 27, 212-219.

24. Graham, J. B., Barrow, E. M. & Roberts, H. R. (1965)Thromb. Diath. Hemorrh. Suppl. 17, 151-161.

25. Schmer, G., Kirby, E. P., Teller, D. C. & Davie, E. W.(1972) J. Biol. Chem. 247, 2512-2521.

26. Bouma, B. N., Wiegerinck, Y., Sixma, J. J., van Mourik,J. A. & Mochtar, I. A. (1972) Nature New Biol. 236, 104-106.

2818 Patholo' : Griggs et al.

Dow

nloa

ded

by g

uest

on

Apr

il 10

, 202

0