Proc. Natl. Acad. Sci. USAVol. 86, pp. 2839-2843, April 1989Medical Sciences

Localization of tissue factor in the normal vessel wall and in theatherosclerotic plaque

(thrombosis/atherosclerosis)

JOSIAH N. WILCOX*, KATHLEEN M. SMITH*, STEPHEN M. SCHWARTZt, AND DAVID GORDONt*Department of Cardiovascular Research, Genentech, 460 Point San Bruno Boulevard, South San Francisco, CA 94080; and tDepartment of Pathology,University of Washington, Seattle, WA

Communicated by Earl P. Benditt, December 28, 1988

ABSTRACT Tissue factor (TF)-producing cells were iden-tified in normal human vessels and atherosclerotic plaques byin situ hybridization and immunohistochemistry using a spe-cific riboprobe for TF mRNA and a polyclonal antibodydirected against human TF protein. TF mRNA and proteinwere absent from endothelial cells lining normal internalmammary artery and saphenous vein samples. In normalvessels TF was found to be synthesized in scattered cells presentin the tunica media as well as fibroblast-like adventitial cellssurrounding vessels. Atherosclerotic plaques contained manycells synthesizing TF mRNA and protein. Macrophages presentas foam cells and monocytes adjacent to the cholesterol cleftscontained TF mRNA and protein, as did mesenchymal-appearing intimal cells. Significant TF protein staining wasfound deposited in the extracellular matrix surroundingmRNA-positive cells adjacent to the cholesterol clefts andwithin the necrotic cores. These results suggest that depositionof TF protein in the matrix of the necrotic core of theatherosclerotic plaque may contribute to the hyperthromboticstate of human atherosclerotic vessels.

Tissue factor (TF) is a membrane-bound glycoprotein thatfunctions in the extrinsic pathway of blood coagulation byacting as a cofactor for factor VII (1, 2). TF binds to co-agulation factor VII, and the resulting factor VIIa-TF complexacts as a catalyst for the conversion offactors X to Xa and IXto IXa, leading to the formation of thrombin (3, 4). Thus, TFfacilitates both intrinsic and extrinsic pathways of coagulationand is a key protein in the activation of the coagulationcascade.TF activity has been isolated from a number of different

tissues and has been cloned from adipose (5), fibroblast (6),and placental (7, 8) cDNA libraries. A 2.3-kilobase bandcorresponding to human TF mRNA has been identified byNorthern blots using RNA from adipose tissue, small intes-tine, placenta, kidney (5), and brain (8). Cultured endothelialcells produce low levels of TF mRNA (5). TF procoagulantactivity in endothelial cells is enhanced by the addition ofendotoxin (9), thrombin (10), phorbol esters (9, 11), interleu-kin 1, or tumor necrosis factor (12, 13) to the culture medium.TF mRNA has also been identified in the monocyte cell lineU937 (8) and TF activity has been identified in normalmonocytes after activation with endotoxin or phorbol esters(14, 15).There is no information concerning the cellular distribution

of TF-producing cells within the tissues from which it hasbeen isolated. It has been postulated that TF is not exposedto blood elements but must be associated with the vascula-ture where it could act quickly in response to vasculardamage. With cDNA probes and antibodies specific for theTF protein, it is now possible to do more precise cellular

localizations using in situ hybridization and immunohis-tochemistry. We report here the localization ofTF-producingcells in the normal vessel wall and atherosclerotic plaques.

METHODSTissue Preparation. Normal human saphenous veins and

internal mammary arteries were obtained during coronarybypass surgery. Human atherosclerotic plaques were ob-tained from patients undergoing carotid endarterectomysurgery.The tissue samples were removed and immersed in freshly

prepared 4% (wt/vol) paraformaldehyde in 0.1 M sodiumphosphate (pH 7.4). The tissues were fixed at 40C for 3 hr toovernight and then immersed in 15% (wt/vol) sucrose/iso-tonic phosphate-buffered saline for 2-4 hr at 40C to act as acryoprotectant. The tissues were then embedded in optimalcutting temperature compound (O.C.T., Miles Scientific)blocks and stored at -70°C. Finally, the tissues were sec-tioned at 10 Am thickness by using a cryostat, thaw-mountedonto polylysine-coated microscope slides, immediately re-frozen, and stored at -70°C with desiccant.

In Situ Hybridization. In situ hybridizations were carriedout as described (16, 17). Prior to hybridization the sectionswere pretreated sequentially with paraformaldehyde (10 min)and with proteinase K at 1 ,g/ml (10 min) and prehybridizedfor 1-2 hr in 100 ,ul of prehybridization buffer [50% (vol/vol)formamide/0.3 M NaCI/20 mM Tris-HCl, pH 8.0/5 mMEDTA/0.02% polyvinylpyrrolidone/0.02% Ficoll/0.02% bo-vine serum albumin/10% (wt/vol) dextran sulfate/10 mMdithiothreitol]. The hybridizations were started by adding600,000 cpm of the 35S-labeled riboprobe in a small amount ofprehybridization buffer. After hybridization the sectionswere washed with 2 x SSC (two 10-min washes; lx SSC =150 mM NaCl/15 mM sodium citrate, pH 7.0), treated withRNase (20 ,g/ml for 30 min at room temperature), washed in2x SSC (two 10-min washes), and washed at high stringencyin 0.1 x SSC at 520C for 2 hr. All SSC solutions up to this pointof the procedure contained 10 mM 2-mercaptoethanol and 1mM EDTA to help prevent nonspecific binding of the probe.The tissue was then washed in 0.5x SSC without 2-mer-captoethanol (two 10-min washes) and dehydrated by immer-sion in a graded alcohol series containing 0.3 M NH4Ac. Thesections were dried, coated with NTB2 nuclear emulsion(Kodak), and exposed in the dark at 40C for 4-8 weeks. Afterdevelopment, the sections were counterstained with hema-toxylin and eosin.A cDNA probe specific to human TF subcloned in the SP64

plasmid (Promega Biotec), provided by Karen Fisher andRichard Lawn (Genentech), was labeled by transcription (18)using 35S-labeled UTP (specific activity, 1200 Ci/mmol; 1 Ci= 37 GBq; Amersham). This was a 1.3-kilobase probe andincluded the entire coding sequence for human TF extending

Abbreviation: TF, tissue factor.

2839

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Dec

embe

r 13

, 202

0

2840 Medical Sciences: Wilcox et al.

*a t9.. .:

L.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~1.:* ¼ .: Ii..

4.

M

4

...1.*t

I,:>tj~F-. s

..<_ .'z'C

'I.. A; > .;

4F sw 1S

J Z~ ~ ~~,,:44f,,......

.1

W w-w 0* w~~

FIG. 1. (Legend appears at the bottom of the opposite page).

..s.

B"

.%. .-N

ii

L *h

1;,e;

.4..

O...I1

,01

E

II..m...1

....

4.f i.I.

..4.

go :.-

*9.-.. .

.....~~~~

>

Sy-F s*e...... ..... 1 Tim~~~~~~v

vzI

...7.. .. ..0.. .. 1-. I

.: 4 I.f

Proc. Nad. Acad. Sci. USA 86 (1989)

di.

Pol.

...... _. ;...."E

Dow

nloa

ded

by g

uest

on

Dec

embe

r 13

, 202

0

Proc. Natl. Acad. Sci. USA 86 (1989) 2841

from nucleotide 1 in the 5' flanking region to a HindIII site atnucleotide 1347 in the 3' untranslated region (5). The finalspecific activity of this probe was 300 Ci/mmol.Immunohistochemistry. TF antibody RD010 was prepared

by immunizing rabbits with recombinant human TF protein(19). The IgG fraction of the serum was purified by affinitychromatography on a recombinant human TF-Sepharosecolumn. This antibody was shown to react with a single bandat 42 kDa on a Western blot, neutralize TF activity, andimmunoprecipitate TF protein (20).Immunohistochemistry was performed according to the

manufacturer's direction using the Vectastain ABC alkalinephosphatase system (Vector). The final reaction product wasstained with the alkaline phosphatase substrate kit I to givea final stain that appeared red.The affinity-purified TF antibody RD010 was used at a

concentration of 4.4 ,ug/ml. An IgG fraction of the preim-mune serum was used as a control for the TF immunohis-tochemistry at the same IgG concentration as RD010. Thiswas prepared by passing the preimmune serum over a proteinA-Sepharose column. Additional antibodies specific for hu-man macrophages (ref. 21; HAM56, gift of Allen Gown,University of Washington), smooth muscle cells (ref. 22;HHF35, gift of Allen Gown), or human endothelial cells (ref.23; anti-Ulex lectin, Vector) were used on some serialsections to aid in cell identification.

RESULTSTF Localization in Normal Vessels. Normal human saphe-

nous vein and internal mamamary artery samples (ninesections for each tissue, three sections from each of threeindividuals) were examined for the presence of TF. Endo-thelial cells were negative for TF mRNA and protein in allvascular tissues examined. Many scattered cells in the tunicamedia of the saphenous vein contained TF mRNA as deter-mined by in situ cRNA hybridization (Fig. lA). Immunohis-tochemical staining of these cells, however, was very weakbut appeared to be cell-associated and correlated well withthe in situ results (Fig. 1B). The strongest labeling was seenover the adventitia where adventitial fibroblasts showedintense TF protein staining (Fig. 1B) and mRNA hybridiza-tion. The internal mammary artery samples also showedpositive mRNA hybridization and intense protein labelingover the adventitial fibroblasts. However, fewer medial cellscontained TF mRNA, as indicated by the in situ hybridiza-tion, and no protein could be detected in the media byimmunohistochemistry. Additional normal arteries obtainedfrom hearts discarded from recipients of transplants werescreened for TF mRNA by in situ hybridization (threesections each from the following tissues: aorta, n = 2; left

anterior descending coronary artery, n = 1; right coronaryartery, n = 2; left circumflex artery, n = 1). These arterialsamples came from idiopathic dilated cardiomyopathy heartsremoved at the time of cardiac transplantation. These heartstypically have no significant atherosclerotic disease, and thiswas confirmed in our samples. All of these vessels showedpositive adventitial cells but at most a single medial cellcontaining TF mRNA.

In general, more cells were found to be positive in themedia of the saphenous vein by in situ hybridization thancould be detected by immunohistochemical staining. Visualcomparison of the hybridization signal over the adventitialfibroblasts and that of positive cells in the media suggestedthat these cell types do not differ greatly in their content ofTF mRNA. However, a similar comparison of the immuno-histochemical staining intensity over these cells suggests thatthe adventitial fibroblasts contain much more TF protein thando the medial cells. This could suggest that there is somealteration in TF translation or secretion in the medial cellsresulting in the reduced protein content. Alternatively, anti-body RD010 may react differently with the TF in the medialor adventitial cells due to some change in the TF proteinitself.The morphology of TF-positive cells in the media of the

saphenous vein differed from that of typical smooth musclecells. The cytoplasm of the TF-positive cells stained poorlywith eosin and did not display a typical fusiform-shapedcytoplasm but rather appeared more cuboidal in shape, withsmall dense nuclei. Cells with this morphology do not stainwith smooth muscle a-actin antibodies (HHF35; data notshown) and must be considered undefined, since we lack apositive immunochemical marker for these cells.TF Localization in Atherosclerotic Plaques. Human athero-

sclerotic plaques obtained from carotid endarterectomy sur-gery were examined for TF mRNA and protein using theabove techniques. Extensive mRNA hybridization was seenin several regions of atherosclerotic plaque. Positive cellswere found scattered throughout the fibrous cap (Fig. 1C),the base and shoulder region of the plaque, as well as in thenecrotic core adjacent to the cholesterol clefts (Fig. 1F). Sixplaques were screened and cells showing TF mRNA hybrid-ization were seen in all ofthem. The normal media underlyingthe endarterectomy specimens did not contain any TF proteinor mRNA-positive cells.The necrotic cores of the plaques were characterized by

extensive TF protein localization in the extracellular matrixparticularly surrounding some cholesterol clefts (Fig. 1 E andG). Additional protein staining was seen in the macrophage-rich foam cell regions of many of the atherosclerotic plaquesexamined (Fig. 1D). Such foam cell-rich regions often layunderneath the fibrous cap and adjacent to the necrotic cores.

FIG. 1. Localization of TF in the normal human saphenous vein (A and B) and in human carotid endarterectomy specimens (C-G). (A) Insitu hybridization using a specific 35S-labeled TF cRNA probe indicated that there were scattered TF-producing cells in the tunica media andadventitia. (B) Cells containing TF protein were detected by immunocytochemistry using antibody RD010 and the Vectastain alkalinephosphatase method (positive cells stain red). Scattered cells in the tunica media were lightly stained by antibody RD010, whereas strongimmunohistochemical staining was always seen in the adherent adventitial fibroblasts. Endothelial cells lining the lumen of the normal vesselwere always negative for TF protein and mRNA. L, lumen; M, media; A, adventitia. (C) Localization ofTF in the human atherosclerotic plaqueby in situ hybridization. Carotid endarterectomy specimens were hybridized to an 35S-labeled TF riboprobe and revealed many cells producingTF in the fibrous cap ofthe atherosclerotic plaque. (D) Localization oftissue factor protein in macrophage foam cells of the atherosclerotic plaqueby immunohistochemistry with antibody RD010. (E and F) Colocalization of TF protein and mRNA in the same cells of the plaque. (E)Immunohistochemistry with TF antibody RD010 indicated strong staining of the necrotic core region of the plaque particularly in areas adjacentto the cholesterol clefts. (F) In situ hybridization of a section immediately adjacent to the one shown in E indicated that cells containing TFmRNA were found adjacent to the cholesterol clefts, suggesting local synthesis of the TF protein detected in this region. Not all of the proteindetected here was cell associated. Arrows point to two cholesterol clefts followed on the serial sections. (G and H) Identification of cellsproducing TF in the necrotic core of the atherosclerotic plaque. Serial sections stained with the TF antibody RD010 (G) (same section as shownin E) or with a monoclonal antibody (HAM-56) to human macrophages (H) were compared and indicated that the TF-protein-producing cellsin the necrotic core may be macrophages. Additional adjacent sections in the series were screened with markers specific for smooth musclecells (HHF-35), endothelial cells (Ulex europaeus lectin binding), or T cells (anti-Leu4), all of which were negative in this region but labeledthe appropriate cells elsewhere in the section. Arrows point to two cholesterol clefts followed on the serial sections (E-H). (Magnifications areas follows: for A, x760; for B, x615; for C, G, and H, x310; for D, x1230; for E, x75; for F, x760.)

Medical Sciences: Wilcox et al.

Dow

nloa

ded

by g

uest

on

Dec

embe

r 13

, 202

0

2842 Medical Sciences: Wilcox et al.

Finally, as with the normal vessels, no TF mRNA or proteinwas detected in either the endothelium lining the vascularsurface or the small vessels within the plaques.

It was possible to confirm that the protein staining and thein situ hybridization labeled the same regions on serialsections (see Fig. 1 E-G). In afew instances we could identifythe same cell on two adjacent serial sections and show thatthis cell was positive with both the TF antibody and the in situmRNA hybridization. This observation served as a controlfor both the immunohistochemistry and hybridization anal-ysis. Additional controls were done on serial sections inevery experiment. The in situ hybridizations were controlledby hybridization of serial sections with platelet-derivedgrowth factor A chain or platelet-derived growth factorreceptor-specific riboprobes (17). Different patterns of hy-bridization were seen with these probes compared to TF. TFimmunohistochemistry was always controlled by incubationof serial sections with preimmune serum that failed to labelany cells at all.

All of the carotid endarterectomy specimens examined (n= 16) showed positive TF protein staining in some region ofthe plaque. Prominent staining of regions surrounding thecholesterol clefts in the necrotic core was a feature observedin five out of seven plaques in which this feature was present.The others showed variable staining in monocytes, macro-phage foam cells, or mesenchymal-appearing intimal cells(17). A femoral plaque obtained by endarterectomy wasexamined and showed TF protein staining consistent withthat seen in the carotid plaques. The mRNA hybridizationswere in agreement with the immunohistochemistry and all sixcarotid plaques examined by in situ hybridization had posi-tive intimal cells.

DISCUSSIONThe purpose of these experiments was to localize sites of TFbiosynthesis in both the normal and atherosclerotic vessel. Inthe normal vessel, we have shown by immunohistochemistryand in situ hybridization that TF is synthesized by scatteredcells in the tunica media and fibroblasts in the adventitiasurrounding the vessels. We did not find any evidence of TFmRNA or protein localization in endothelial cells of anyvessel studied. Previous cell culture work has suggested thatinduction of TF synthesis by endothelial cells represents amajor procoagulant mechanism by which endothelial cellsparticipate in homeostasis (24). However, cultured fibro-blasts (24-26) and vascular smooth muscle cells (25) havebeen shown to produce TF at much higher levels. It is notclear if induction of endothelial TF biosynthesis is a normalin vivo mechanism by which the endothelium modifieshomeostasis or represents the response of the endothelium toinfection and endotoxin stimulation.

Atherosclerotic plaques were examined and found to havemany more TF mRNA-containing cells and stronger TFprotein staining than the media of normal saphenous veins,internal mammary artery, or regions of normal media under-lying the plaque. TF mRNA was found in mesenchymal-likeintimal cells in the atherosclerotic intima as well as inmacrophages and cells adjacent to the cholesterol clefts.Immunohistochemistry indicated that there was a consider-able amount ofTF protein trapped in the extracellular matrixof the necrotic core of the atherosclerotic plaque. This wasnot cell-associated but was probably synthesized locallysince cells adjacent to these regions contained TF mRNA.There is no evidence that TF is a secreted protein (5, 8);however, it is possible that the TF protein found in thenecrotic core may be shed from the surfaces of the synthe-sizing cells and trapped in the surrounding matrix. Alterna-tively, the TF protein in this region may originate from cellsthat have died and left TF-rich membranes behind.

The production of TF by macrophages has been demon-strated by other investigators (27, 28). The immunostaining ofmacrophage foam cells suggests that in these cells TF isintracellular as well as possibly cell surface associated. It isnot clear to what extent such stores of TF are macrophage-derived or whether this protein originates from phagocytosisof surrounding necrotic core debris. It is interesting to notethat Levy et al. (28), have found that certain lipoproteinfractions can induce procoagulant activity originating frommonocytes/macrophages.Evidence indicates (29) that TF-initiated coagulation is the

major in vivo coagulant pathway. It is a highly thrombogenicprotein and requires only phospholipid and factor VII/VIafor its activity. The TF-factor VIIa complex directly acti-vates factor IX and X leading to the generation of thrombin.Factor VII is normally present in blood but requires bindingto TF for activation offactors IX and X. Since normally thereis no in vivo coagulation in the absence of vascular damage,it is reasonable to assume that TF is not exposed to blood (7,29). This is consistent with our findings, since normal vesselendothelial cells in direct contact with the blood do notsynthesize or store TF. TF is found in scattered smoothmuscle-like cells in the tunica media and adventitial cellsadherent to the vessel. Vessel wall rupture into these areaswould be required to expose the blood to significant stores ofprocoagulantTF activity. Zaugg (30) has shown that damagedhuman aorta exhibits factor VII-dependent procoagulantactivity, in support of this hypothesis.Advanced human atherosclerosis is characterized by inti-

mal smooth muscle cell proliferation accompanied by accu-mulation of fats, inflammatory cells including macrophages,and T cells within the atherosclerotic plaque (24, 31, 32).Commonly, the critical event that converts an asymptomaticatherosclerotic plaque into a symptomatic one is thrombosis(33-35), whereas nondiseased arteries rarely become throm-botic. It has been suggested that plaque rupture is the integralevent that precipitates clot formation (36-39). An occlusivemural thrombus accompanies most cases of acute myocardialinfarctions (40-43). Plaque rupture or cracking that exposesthe necrotic core region to the lumen is usually found tounderlie such thrombi in both the coronary (33, 44, 45) andcerebral arteries (46). The source of the thrombogenicity ofthe plaque has not previously been determined, but it hasbeen suggested that coagulation occurs when blood compo-nents come into contact with fats or the collagen matrixwithin the plaque. Our studies clearly demonstrate that thereis significant synthesis of TF in atherosclerotic plaquesaccompanied by strong staining ofTF protein in the necroticcore and in foam cell-rich regions of the plaque. These resultssuggest that overproduction and/or trapping of TF protein inthe atherosclerotic plaque may play a significant role in thethrombosis associated with plaque rupture.

We thank Gary Kollman for excellent technical assistance, theDepartment of Vascular Surgery at the University of Washington fortheir help in obtaining the human tissues used in this study, KarenFisher for providing the TF probe, and Gordon Vehar, StuartBunting, and Richard Lawn for their helpful discussions. S.M.S andD.G. were supported by Grant P01-HL-03174 from the NationalInstitutes of Health. D.G. was also supported by a Robert WoodJohnson Minority Faculty Development Award.

1. Nemerson, Y. & Bach, R. (1982) Prog. Hemostasis Thromb. 6,237-261.

2. Nemerson, Y. (1983) Haemostasis 13, 150-155.3. Nemerson, Y. (1966) Biochemistry 5, 601-608.4. Osterud, B. & Rapaport, S. I. (1977) Proc. Nati. Acad. Sci.

USA 74, 5260-5264.5. Fisher, K. L., Gorman, C. M., Vehar, G. A., O'Brian, D. P. &

Lawn, R. L. (1987) Thromb. Res. 48, 89-99.

Proc. Natl. Acad Sci. USA 86 (1989)

Dow

nloa

ded

by g

uest

on

Dec

embe

r 13

, 202

0

Proc. Natl. Acad. Sci. USA 86 (1989) 2843

6. Morrissey, J. H., Fakhrai, H. & Edgington, T. S. (1987) Cell50, 129-135.

7. Spicer, E. K., Horton, R., Bloem, L., Bach, R., Williams,K. R., Guha, A., Kraus, J., Lin, T. C., Nemerson, Y. &

Konigsberg, W. H. (1987) Proc. Natl. Acad. Sci. USA 84,5148-5152.

8. Scarpati, E. M., Wen, D., Broze, G. J., Jr., Miletich, J. P.,Flandermeyer, R. R., Siegel, N. R. & Sadler, J. E. (1987)Biochemistry 26, 5234-5238.

9. Lyberg, T., Galdal, K. S., Evensen, S. A. & Prydz, H. (1983)Br. J. Haematol. 53, 85-95.

10. Brox, J. H., Osterud, B., Bjrklid, E. & Fenton, J. W., 2d (1984)Br. J. Haematol. 57, 239-246.

11. Stem, D. M., Drillings, M., Kisiel, W., Nawroth, P., Nossel,H. L. & LaGamma, K. S. (1984) Proc. NatI. Acad. Sci. USA81, 913-917.

12. Nawroth, P. P. & Stem, D. M. (1986) J. Exp. Med. 163, 740-745.

13. Nawroth, P. P., Handley, D. A., Esmon, C. T. & Stem, D. M.(1986) Proc. Natl. Acad. Sci. USA 83, 3460-3464.

14. Bjrklid, E., Holm, T. & Osterud, B. (1987) Thromb. Res. 45,609-624.

15. Lyberg, T., Nilsson, K. & Prydz, H. (1982) Br. J. Haematol.51, 631-641.

16. Rosenthal, A., Chan, S. Y., Henzel, W., Haskell, C., Kuang,W. J., Chen, E., Wilcox, J. N., Ullrich, A., Goeddel, D. V. &Routtenberg, A. (1987) EMBO J. 6, 3641-3646.

17. Wilcox, J. N., Smith, K. M., Williams, L. T., Schwartz, S. M.& Gordon, D. (1988) J. Clin. Invest. 82, 1134-1143.

18. Melton, D. A., Krieg, P. A., Rebagliati, M. R., Maniatis, T.,Zinn, K. & Green, M. R. (1984) Nucleic Acids Res. 12, 7035-7056.

19. Paborsky, L. R., Fendly, B. M., Fisher, K. L., Lawn, R. M.,Marks, B. J., McCray, G., Tate, K. M., Vehar, G. A. &

Gorman, C. M. (1989) J. Biol. Chem., in press.20. Paborsky, L. R., Tate, K. M., Harris, R. J., Yansura, D. G.,

Band, L., McCray, G., Gorman, C. M., O'Brian, D. P., Chang,J. Y., Swartz, J. R., Fung, V. P., Thomas, J. N. & Vehar,G. A. (1989) Biochemistry, in press.

21. Gown, A. M., Tsukada, T. & Ross, R. (1986) Am. J. Pathol.125, 191-207.

22. Tsukada, T., Tippens, D., Gordon, D., Ross, R. & Gown,A. M. (1987) Am. J. Pathol. 126, 51-60.

23. Holthofer, H., Virtanen, I., Kariniemi, A. L., Hormia, M.,Linder, E. & Mittine, A. (1982) Lab Invest. 47, 60-66.

24. Bevilacqua, M. P., Pober, J. S., Wheeler, M. E., Cotran, R. S.& Gimbrone, M. A., Jr. (1985) Am. J. Pathol. 121, 394-403.

25. Maynard, J. R., Dreyer, B. E., Stemerman, M. B. & Pitlick,F. A. (1977) Blood 50, 387-396.

26. Green, D., Ryan, C., Malandruccolo, N. & Nadler, H. L.(1971) Blood 37, 47-51.

27. Tipping, P. G., Campbell, D. A., Boyce, N. W. & Holdsworth,S. R. (1988) Am. J. Pathol. 131, 206-212.

28. Levy, G. A., Schwartz, B. S., Curtiss, L. K. & Edgington,T. S. (1981) J. Clin. Invest. 67, 1614-1622.

29. Nemerson, Y. (1988) Blood 71, 1-8.30. Zaugg, H. (1980) J. Clin. Chem. Clin. Biochem. 18, 545-549.31. Ross, R. (1986) N. Engl. J. Med. 314, 488-500.32. Jonasson, L., Holm, J., Skalli, O., Bondjers, G. & Hansson,

G. K. (1986) Arteriosclerosis 6, 131-138.33. Falk, E. (1983) Br. Heart J. 50, 127-134.34. Sherman, C. T., Litvack, F., Grundfest, W., Lee, M., Hickey,

A., Chaux, A., Kass, R., Blanche, C., Matloff, J., Morgen-stem, L., Aanz, W., Swan, H. J. C. & Forrester, J. (1986) N.Engl. J. Med. 315, 913-919.

35. Imparato, A. M., Riles, T. S., Mintzer, R. & Baumann, F. G.(1983) Ann. Surg. 197, 195-203.

36. Forrester, J. S., Litvack, F., Grundfest, W. & Hickey, A.(1987) Circulation 75, 505-513.

37. Constantinides, P. (1966) J. Atheroscler. Res. 6, 1-17.38. Friedman, M. & van den Bovenkamp, G. J. (1966) Am. J.

Pathol. 48, 19-44.39. Friedman, M. (1971) Hum. Pathol. 2, 81-128.40. DeWood, M. A., Spores, J., Notske, R., Mouser, L. T.,

Burroughs, R., Golden, M. S. & Lang, H. T. (1980) N. Engl. J.Med. 303, 897-902.

41. Davies, M. J., Woolf, N. & Robertson, W. B. (1976) Br. HeartJ. 38, 659-664.

42. Buja, L. M. & Willerson, J. T. (1981) Am. J. Cardiol. 47, 343-356.

43. Horie, T., Sekiguchi, M. & Hirosawa, K. (1978) Br. Heart J. 40,153-161.

44. Chapman, I. (1965) Arch. Pathol. 80, 256-261.45. Drury, R. A. B. (1954) J. Pathol. Bacteriol. 67, 207-215.46. Constantinides, P. (1967) Arch. Pathol. 83, 422-428.

Medical Sciences: Wilcox et al.

Dow

nloa

ded

by g

uest

on

Dec

embe

r 13

, 202

0