8/3/2019 8 Ftp

1/6

American Journal of Hematology 51 :296-301 (1996)

Disturbance of Plasma and Platelet ThrombospondinLevels in Sickle Cell DiseasePaul V. Browne, Deane F. Mosher, Martin H. Steinberg, and Robert P. HebbelDepartmentsof Medicine, Universityof Minnesota Medical School, Minneapolis, Minnesota; University of Wisconsin, Madison,Wisconsin; University of Mississippi School of Medicine and Veterans Affairs Medical Center, Jackson, Mississippi

Thrombospondin (TSP), a large protein found In platelet a-granules (as TSP-l), mediatesadhesion of sickle reticulacytes o cultured vascular endothelium. To further explore thephysiologic relevance of this observation, we have measured plasma TSP levels andplatelet TSP-1 content In subjects with sickle cell disease. Plasma TSP levels were similarfor normal controls (mean 491 nglml, range 331-723) and steady-state HbSS patients(mean 536, range 333-1107) and were slgnlflcantly (P = 0.012) but variably elevated forHbSS patients presenting with acute painful crisis (mean 868, range 442-2780). Someofthese elevated plasma TSP levels reached hose previously observed to support maximalred cell adhesion to endothelium In vltro. Compared to normals, both steady-state andln-crisis HbSS patients had significantly (P< 0.001) depressed platelet TSP-1 content(82.6 f 11.9,47.1 f 16.0 and 45.9 f 20.7 ng ll w platelets, respectively, mean=SD). HbSCdisease patients, all examined during steady state, had low-normal plasma evels of TSPand either normal or depressed platelet TSP-1 content. Serial observations on three sicklecell anemia subjects indicated a probable relationship between platelet TSP-1 release,elevated plasma TSP levels, and acute vasoocclusive episodes. These results suggesta state of ongoing release and depletion of TSP-1 from activated platelets In patientswith sickle cell disease.Key words: thrombospondin, platelets, sickle cell anemia

Q iggs Wiby-Liu. IN.

INTRODUCTIONAdhesion of sickle erythrocytes (RBCs) to vascular

endothelium is believed to be an initiating factor for thedevelopment of sickle vasoocclusion [1,2]. A number ofadhesogenic plasma proteins have been implicated asmediators of this event, including von Willebrand Factor,fibrinogen, and thrombospondin (TSP) [2]. Their partici-pation raises the possibility that fluctuations of theirplasma levels due to physiological stresses and concurrentillnesses would be of potential pathogenetic relevance.Recently it was suggested that adhesion mediated by TSPmay be a major initiating factor in the pathogenesis ofvasoocclusion in sickle cell disease [2-4].

TSP is a large, trimeric, modular protein having variousdomains implicated in cell binding [5-71. Several mem-bers of the TSP gene family have been identified [8,9],but most data pertain to TSP-1, the form of TSP foundin human platelets. Potential physiologic roles of thisprotein include participation in extracellular matrix func-

tions [6], platelet aggregation [10,11], adhesion of hema-topoietic progenitors [121,modulation of fibrinolysis [131,and inhibition of angiogenesis [141. TSP is produced bya variety of cultured cells, including monocytes/macro-phages, megakaryocytes, endothelial cells, smooth mus-cle cells, fibroblasts, glial cells, and various tumor cells.Significantly, TSP-1 is released from platelet a-granulesduring platelet activation and secretion [6,10,11].

The present studies were performed to determinewhether TSP levels are abnormal in patients with sicklecell disease. Results indicate that this is the case andsuggest that this is a consequence of in vivo platelet acti-vation.

Received for publication July 14, 1995; accepted November 8, 1995.Address reprint requests to R.P. Hebbel, M.D., Bo x 480, UMHC,University of Minnesota Medical School. 420 Delaware Street SE .Minneapolis, MN 55455.

0 1996 Wiley-Liss, Inc.

8/3/2019 8 Ftp

2/6

Thrombospondin Levels in S ickle Cell Disease 297METHODSBlood Collection and Handling

Fresh blood was obtained from volunteer normal do-nors or from subjects with sickle cell anemia (HbSS) orhemoglobin (Hb) SC disease. Informed consent was givenin all cases. The sickle subjects were studied when theyfirst presented with acute painful crisis and/or when intheir steady state, at least one month remote from anyacute clinical event. Sickle patients with acute clinicalpresentations other than painful crisis were excluded.

Blood was drawn into an anticoagulant mix comprisedof citrate supplemented with adenosine, theophylline, anddipyridamole (Diatube H; Diagnostica Stago, France). Insome cases, we further supplemented this anticoagulantmi x with PPACK (D-phenylalanyl-L-prolyl-L-argininechloromethyl ketone), a high-affinity inhibitor of throm-bin-mediated platelet activation [151, sufficient to yield1 p,M after dilution in blood. In most cases blood wasobtained from a free-flowing, clean venipuncture using alarge needle after release of the tourniquet. An occasionalsample was obtained by indwelling venous access; thesesamples showed no tendency to yield different resultsfrom those obtained using our standard procedure.

Resulting blood samples were immediately centrifugedat room temperature at slow speed to obtain platelet-richplasma (PRP), which was then centrifuged at 18,OOOg at6C for 15 min. Using our anticoagulant mix, we foundlower TSP concentrations in platelet-poor plasma (PPP)prepared in this manner than at room temperature, a find-ing consistent with earlier studies [161. The resulting PPPwas saved by freezing. To make serum, blood was drawninto standard serum tubes without anticoagulant and wasincubated for 2 hr at 37C, after which serum separatedfrom clot was saved by freezing. Samples were shippedfor TSP assay frozen on dry ice.TSP Content

TSP content of PPP, expressed in ng/ml, was deter-mined using an enzyme-linked immunosorbent assay(EL1SA)-based test that employs a monoclonal antibody(MAb) raised against purified TSP-1, as previously de-scribed [171. Releasable platelet TSP-1 content, expressedin ng per 106 platelets, was calculated from the TSPcontent measured in serum and the platelet count deter-mined at the time of blood sampling. Platelet countswere measured using a Coulter STKS automated counter.Microvesicular red cell fragments have been demon-strated previously in sickle blood [181. However, the meanvolume of these microvesicles, based on previous reportsof their diameter [18,19], is estimated to be 0.05-0.1 fl ,well below the minimum platelet volume of 2 f l detectedby the automated counter in our study.

We chose the above method of quantifying platelet

TSP-1 content rather than assay of purified washed plate-lets in order to avoid potential artifact from inadvertent,selective isolation of different platelet populations fromnormal and sickle cell individuals. Moreover, this methodaffords the advantage of ensuring that depressed valuesfor platelet TSP content cannot be explained by artifactualloss due to platelet release during phlebotomy or subse-quent sample handling. In control experiments, we veri-tied that TSP concentration of serum prepared by ourmethod was within 10%of the total TSP content of PRP.This justifies using releasable TSP-1 as a measure ofplatelet TSP-1 content.Splenic Function

Fresh blood samples were added to an equal volumeof 0.1 M cacodylate buffer (pH 7.4) containing 1% glutar-aldehyde. These were examined microscopically usingNomarski optics to assess erythrocyte pit count as anindicator of splenic function [20]. We used values of>20% and >3.5% pitted RB C as the indicator of func-tional asplenia in HbSC [21] and HbSS [20] patients, re-spectively.Statistical Analysis

Data were analyzed by Students t-test. For plasmaTSP levels, the data were log-transformed before analysis,which is appropriate because plasma TSP level can rangefrom zero on the low side to perhaps an 100-fold increaseon the high side; moreover, the actual range of TSP levelswe found for sickle patients seems to confirm a log-normal distribution for this parameter (see below). Analy-sis of the same data without prior log-transformation didnot eliminate the statistical significance of the data re-ported here.

RESULTSPlasma TSP LevelThe average level of TSP in PPP was similar for nor-

mals and for either HbSS or HbSC patients in their steadystate, remote from an acute clinical event (Fig. 1). How-ever, for HbSS patients presenting with acute painfulepisodes, plasma TSP levels were significantly elevatedcompared to the steady-state HbSS patients(P= 0.012).Actual values for means and standard deviations are pre-sented in the legend to Figure 1. Of the 26 plasma TSPvalues obtained on the entire group of HbSS patients(regardless of clinical status), 38% were found to be abovethe range of the normals we studied. Additionally, in oneHbSS patient we followed plasma TSP levels longitudi-nally and found a high degree of variability (Fig. 2).Notably, each in-crisis value is higher than the followingsteady-state value.

An interesting, but unexplained finding is that the

8/3/2019 8 Ftp

3/6

298 Browne et al.

. 0 120

Normal Hbss HbSS H3sc1steady acute steadystate crisis state

Fig. 1. Plasma TSP levels In sickle cell disease. Similarplasma TSP levels are found for normal controls (mean 491nglml; values at 21 SD = 358 and 673; n = 8), steady-stateHbSC patients (mean 469; f 1 SD = 397 and 553; n = 14),and steady-stateHbSS patients (mean 536; 21 SD = 360 and798; 14 measurements on 9 patlents). However, plasma TSPlevel for incr is ls HbSS patients (mean 868; 21 SD = 527and 1,419; 12 measurements on 8 patients) is signlflcantlyelevated compared to the steady state HbSS patients(P = 0.012). Bars Indicate mean and f l SD for each group.

fsr:a

0 NL HbSS HbSS HbSCsteady acute steadystate crisis stateFig. 3. Platelet TSP content in sickle cell disease. PlateletTSP content, inferred from TSP-1 release Into serum (ex-pressed n ng perl@latelets), was slgnlficantly (P < 0.001)depressed or both steady-state HbSS subjects (47.1 f 16.0,mean *SD; 10 measurementson 9 patlents) and for in tr ls lsHbSS subjects (45.9 f 20.7; n = 6), compared to normalcontrols (82.6 f 11.9; n = 11). A wide range of platelet TSPcontent was observed for steady-state HbSC patients(60.5 2 32.4; n = 14; P = 0.033, compared to normal con-trols).

conferred no additional benefit over the basic anticoagu-lant mix.

2/16 3/26 7/17 9/17 1 1 6 l i lin in incrisis crisis crisis

Fig. 2. Longitudinal measurement of plasma TSP level inan HbSS subject. Plasma TSP level was measured repeatedlywhen this patient presented with acute painful crises and atsubsequent routine outpatient visits when In steady state.

plasma TSP values for the steady state HbSC diseasesubjects tended to be somewhat lower and to show muchless interindividual variability than those obtained fromthe steady-state HbSS patients (Fig. 1). For the log-trans-formed data, the standard deviation for the HbSC subjectswas only 0.072, while the data for normals, steady-stateHbSS patients, and acute-crisis HbSS subjects were0.137, 0.173, and 0.215, respectively.In several experiments we examined the effect of add-ing PPACK to our standard anticoagulantmix. The plasmaTSP level with PPACK was 94.9 * 15.1 percent of thatwithout PPACK, indicating that the addition of PPACK

Platelet TSP-1 ContentCompared to normal controls, both steady-state and in-

crisis HbSS patients manifested significantly (P < 0.001)depressed platelet TSP-1 content, as calculated fromTSP-1 released during blood coagulation in vitro (Fig.3). Values for the means and standard deviations arepresented in the legend to Figure 3. By contrast, HbSCpatients, all of whom were examined in steady state,exhibited either normal or depressed platelet TSP contents(P = 0.033 versus normals). Indeed, inspection of Figure3 suggests that all three sickle disease groups perhapsinclude some individuals that have normal plateletTSP-1 content: 1 of 6 of the acute-crisis HbSS patients,perhaps 3 of the 10 steady-state HbSS patients, and 6 ofthe 14 HbSC patients. The platelet TSP-1 content weobserved for our normal donors is equivalent to that pre-viously reported [22].Correlations Between Parameters

We considered that elevated platelet counts in the sicklecell disease subjects [23] might contribute to elevatedplasma TSP levels. As expected, platelet counts tendedto be higher in patients with HbSS (396 2 171, mean+ SD, X109/L) and HbSC (439 2 294), than in normaldonors (327 2 81). However, we identified no significant

8/3/2019 8 Ftp

4/6

Thrombospondin Levels in Sickle Cell Disease 299DISCUSSIONABLE 1. Paired InCrisIs and Steady-State PlateletPlas maTSP Values in HbSS Subiects

Platelet countHbSS donor ( x 0 9 m1 Steady state 240Acute crisis 2732 Steady state 340

Acute crisis 3063 Steady state 325

Acute crisis 216

Plasma TSP level(ndml)

333714989278062559

PlateletTSP- content

(ng/106 platelets)75.833.465.638.262.256.0

correlation between blood platelet count and plasma TSPlevel for all subjects pooledor for any of the three individ-ual genotypes analysed separately (data not shown), con-sistent with the lack of such correlation observed earlierfor HbAA donors [22]. No r was there any correlationbetween plasma TSP level and platelet TSP-1 content(data not shown).

We were able to obtain paired acute-crisis and steady-state samples for both plasma TSP and platelet TSP-1 onthree HbSS patients (Table I). One of these patientsshowed no change in either parameter with change inclinical status. By contrast, the other two patients wereobserved to have lower platelet TSP- 1 content at presenta-tion with acute crisis than during steady state, and in bothcases an inverse relationship between plasma TSP leveland platelet TSP-1 content is apparent despite the factthat platelet count did not change significantly. This vari-ability in sickle platelet TSP-1 content cannot simply bean artifact of measurement variability, since we estab-lished that platelet TSP-1 content for normal donors ex-amined on a second occasion differed trivially from thaton the first occasion (by only 5.6% _+ 5.2%; n = 6paired observations).

Blood smears were checked visually for the presenceof platelet clumps or unusually large platelets, and noapparent correlation was noted between these featuresand plasma or platelet TSP levels.

Relatlonshlp to Splenic Fun ctionWe considered that TSP levels might be influenced bystatus of splenic function, which we assessed by per-forming RBC pit counts. Neither plasma TSP levelnor blood platelet count was significantly different forpatients with or without splenic hypofunction (data notshown). By contrast, although platelet TSP-1 contentclearly can be temporally variable (Table I) , we neverthe-less identified a tendency for it to be lower among patientswith splenic hypofunction (45.3 2 14.7 ng TSP/l@ plate-lets, mean ?SD, n = 10) than among patients with normalsplenic function (73.6 2 35.4, n = 8, P = 0.035).

We have documented that plasma TSP levels tend tobe elevated in HbSS patients presenting with acute painfulcrisis, compared to either normal controls or to HbSSsubjects in steady state, remote from an acute clinicalevent. Repeated measurements on two HbSS patients fur-ther suggest a relationship between plasma TSP level andthe acute vasoocclusive episode. One caveat is that forthe present study, we obtained samples at presentation ofacute crisis; it remains to be seen if a rise in plasma TSPlevel actually precedes crisis. Also, this clearly is notuniversally true since even some steady-state HbSS pa-tients had elevated plasma TSP levels and since someHbSS patients in acute crisis did not. To some extent, thisvariability may derive from the ambiguity of definition ofsteady state, which is a purely clinical judgment thatmay or may not reflect important biochemical differences[24].Overall, 38% of the plasma TSP levels we obtainedon the HbSS patients (regardless of clinical status) wereelevated above the range we found for the normalsubjects.

Some potential explanations for elevation of plasmaTSP levels in sickle cell disease subjects can be excludedby available data, while others remain tenable. First, thelack of correlation between platelet count and plasmaTSP level argues that elevated plasma TSP levels are notsimply related to the higher number of circulating plate-lets manifested by some sickle cell disease subjects. Sec-ond, it is reported that plasma TSP levels tend to beelevated in HbAA donors who have been surgically sple-nectomized [25]. Thus, the autosplenectomy state ofsickle patients might contribute to the elevated steadystate TSP levels for some subjects. Against this, however,is the fact that plasma TSP levels for sickle cell diseasepatients clearly were not significantly increased for thosewith splenic hypofunction as defined by pit counts.Third, plasma TSP levels also tend to be elevated inHbAA subjects with chronic liver disease [25], so thismight contribute to our finding of elevated levels in somesteady-state HbSS subjects. Fourth, it is believed thatnonplatelet sources do make some contribution to plasmaTSP levels [22], so we cannot exclude a contribution ofenhanced endothelial release of TSP in response to localhypoxia, for example. We hope to identify the specifictype of TSP in sickle plasma in future studies. Finally, thetheoretical possibility that platelet TSP-1 content varies asa function of platelet age and somehow thus contributesto the relationships observed here seems unlikely, giventhe fairly uniform TSP-1 content reported for platelets inpatients with various types of thrombocytopenia [22].The age of platelets in sickle cell disease patients isunclear, as platelet survival has been reported to be bothshortened and lengthened in these patients [23].

Rather than the above possibilities, we believe that

8/3/2019 8 Ftp

5/6

300 Browne et al.elevated plasma TSP in HbSS subjects derives from invivo platelet activation resulting in release of TSP-1 fromplatelet a-granules. Our observation of diminished plate-let TSP-1 content in these patients is consistent withthis hypothesis. Furthermore, serial studies in two HbSSsubjects indicateda reciprocal relationship between plate-let TSP-1 content and plasma TSP level. In the bodyof older literature suggesting that sickle cell disease ischaracterized by in vivo platelet activation, the majorevidence for this has been demonstration of elevatedplasma levels of platelet factor 4 and/or beta-thrombo-globulin [23]. Our studies of both plasma and plateletTSP levels extend these observations. However, addi-tional evidence of ongoing platelet activation in thesepatients would be desirable. To date, only preliminaryreports have appeared describing abnormal expression ofactivation-specific antigens on the surface of plateletsfrom patients with sickle cell disease [26,27]. The extentof in vivo platelet activation among sickle patients, aswell as its temporal and causal relationship to clinicalevents, will require detailed, longitudinal studies of amuch larger number of patients.

The above interpretation of our results must be recon-ciled with the observed discordance among HbSS pa-tients, both in crisis and steady state, between fairly con-sistent depression of platelet TSP-1 contents and variablyincreased plasma TSP levels. Furthermore, 8 of 14 HbSC patients, all in steady state, showed diminished plateletTSP-I despite normal plasma TSP levels. This probablyreflects the fact that plasma TSP has an extremely shorthalf-life of 10-75 min [6], so that elevations of plasmaTSP will be recorded only if abnormal a-granule releaseis occurring virtually at the moment of venipuncture. Bycontrast, platelet TSP-1 content will reflect the averagestatus of circulating platelets, a longer-lived and time-averaged parameter. This would explain a lack of simplecorrelation between plasma TSP level and plateletTSP-1 content in our study subjects. Nevertheless, datafrom sickle patients on whom we were able to obtainpaired platelet/plasma TSP values in both steady state andacute crisis do suggest a reciprocal relationship betweenplatelet TSP-1 content and plasma TSP level.

Regarding this apparent instability of plasma TSP lev-els, it is intriguing that the blood levels of C-reactiveprotein are reported to be unstable in sickle cell diseasepatients, with the degree of instability being a predictorof the frequency of vasoocclusive involvement [28].This could be directly relevant to platelet release ofTSP-1, since C-reactive protein can stimulate tissue factorexpression [29], which in turn would help explain theenhanced thrombin generation and ultimate platelet acti-vationhelease in sickle cell disease [23,30].

If platelet activation participates in acute vasoocclu-sion, it need not be by thrombosis per se. The plasmaTSP levels of a number of the HbSS patients in this study

reached or even exceeded the TSP concentration (1,000ng/ml) we earlier found to promote maximal sickle redcell adhesion to endothelium in vitro [3]. Indeed, it isworth noting that these TSP levels we measured in venousblood might well be lower than levels in microcirculatorybeds if the stimulus underlying their elevation is localizedrather than widespread. Thus, the present data perhapsargue for a convergence of the influences of coagulopathy[23] and erythrocyte adhesion (i.e., as mediated by TSP)on the pathophysiology of vasoocclusion in sickle celldisease. Parenthetically, an additional, intriguing possibil-ity relates to th e ability of TSP to inhibit angiogenesis[14]. We speculate that the well-known higher risk ofneovascularizing retinopathy among HbSC compared toHbSS patients could derive from the tendency of theformer group to have lower plasma TSP levels.

ACKNOWLEDGMENTSWe thank Renee Schultz and Stephana Choong fortechnical assistance. This work was supported by theNational Institutes of Health (HL30160, HL37528,HL49111) and by Research Funds of the Department ofVeterans Affairs.

REFERENCESI . Hebbel RP: Beyond hemoglobin polymerization: The red blood cell

membrane and sickle disease pathophysiology. Blood 77:214. 1991.2. Hebbel RP, Mohandas N: Sickl e cell adheren ce. In Embury SH, Hebbel

RP, Mohandas N, Steinberg MH, eds. Sickle Cell Disease: BasicPrinciples and clinic al Practice. New York: RavenPress,pp 217,1994.

3. Sugihara K , Sugihara T, Mohandas N. Hebbel R P Thrombospondinmediates adherence of CD36+ sickle reticulocytes to endothelial cells.Blood 80:2634. 1992.

4. Brittain HA, Eckman JR , Swerlick RA , Howard RJ, Wick TM : Throm-bospondin from activated platelets promotes sickle erythrocyte adher-ence to human microvascular endothelium under physiologic flow: Apotential role for platelet activation in sickle cell vaso-occ lusion. Blood81:2137, 1993.5. Lawler J, Hynes RO: Structural organization of the thrombospondinmolecule. Semin Thromb Hemost 13:245, 1987.

6. Mosher DF: Physiolo gy o f thrombosp ondin. Annu Rev Med 41%.1990.

7. Frazier WA: Thrombospo ndins. Curr Opin Cell Biol 3:79 2, 1991.8. Bomstein P Thrombospondins: Structure and regulationof expression.

FASEB J 6:3290, 1992.9. Lawler J, Duquette M, Whittaker C A , Adams JC, McHenry K, DeSi-mone DW: Identification and characterization of thrombospondin-4, anew member of the thrombospondin gene family. J Cell Biol120:1059, 1993.

10. Dixit VM, Haverstick DM, ORourke KM, ennessy SW, Grant GA,Santoro SA, Frazier WA: A monoclo nal antibody again st human throm-bospondin inhibits platelet aggregation. Proc Natl Acad Sci USA82:3472, 1985.

I I . Tuszynski GP, K owalska MA: Thrombospondin-induced adhesion ofhuman platelets. J Clin Invest 87:1387, 1991.

12. Long MW, Dixit VM : Thrombospondin functions as a cytoadhesionmolecule for human hematopoietic progenitor cells. Blood 75:23 11,1990.

8/3/2019 8 Ftp

6/6

Thrombospondin Levels in Sickle Cell Disease 30113. Hogg PJ, Stenflo J, Mosher DF: Thrombospondin is a slow tight-

binding inhibitor of plasmin. Biochemistry 31:265, 1992.14. Tolsma SS. Volpen OV, Good DJ, Frazier WA, Polverini PJ, BouckN: Peptides derived from two separate domains of the matrix proteinthrombospondin-1 have anti-angiogenic activity. J Cell Biol122:497. 1993.

15. Harmon JT, Jamieson GA: Activation of platelets by a-thrombin is areceptor-mediated event. D-phenylalanyl-L-prolyl-L-arginine chlo-romethyl ketone-thrombin, but not Na-tosyl-L-lysine chloromethyl ke-tone-thrombin, binds to the high affinity thrombin receptor. J BiolChem 261:15928, 1986.

16. Files JC, Malpass TW, Yee EK, Ritchie JL, Harker LA: Studies ofhuman platelet a-granule release in vivo. Blood 58:607, 1981.17. Jaffe EA, Ruggiero JT, Leung LLK, Doyle MJ, McKeown-LongoPJ, Mosher DF: Cultured human fibroblasts synthesize and Secrete

thrombospondin and incorporate it into extracellular matrix. Proc NaUAcad Sci USA 80:998, 1983.18. Allan D, Limbrick AR. Thomas P, Westerman M P Release of spectrin-free spicules on reoxygenation of sickled erythrocytes. Nature295:612, 1982.

19. Wagne r GM , Chiu D T-Y, Yee MC, Lubin B: Red cell vesiculation-acommon membrane physiologic event. J Lab Clin Med 108:315, 1986.20. Pearson HA, Gallagher D, Chilcote R, Sullivan E, Wilimas J, EspelandM, Ritchey AK, and the Cooperative Study of Sickle Cell Disease:

Developmental pattern of splenic dysfunction in sickle cell disorders.Pediatrics 76:392, 1985.21. Lane PA, OConnell JL, Lear JL, Rogers ZR. Buchanan GR, WoodsGM , Hassell KL , Wethers DL: Delayed development of functionalasplenia in hemoglo bin SC disease. Blood 82(suppl 1):355a, 1993.

22. Dawes J, Pratt DA, Dewar MS, Preston FE: Do extra-platelet sourcescontribute to the plasma level of thrombospondin? Thromb Haemost59:273, 1988.

23. Francis RB Jr, Hebbel Rp: Hemostasis. In Embury SH, Hebbel RP,Mohandas N, Steinberg MH (eds): Sickle Cell Disease: Basic Principlesand Clinical Practice. New York: Raven Press, pp 299-310 , 1994.24. Em bury SH, Hebbel RP, Steinberg MH , Mohandas N: Pathogenesisof Vasoocclusion. In Embury SH, Hebbel RP. Mohandas N, SteinbergMH (eds): Sickle Cell Disease: Basic Principles and Clinical Practice.New York:Raven Press, pp 311-326, 1994.25. Trzeciak M-C, McGregor J, Renaux J-L, Ville D, Bermyer M, DawesJ, Dechavanne M: Plasma thrombospondin in patients with chronicrenal failure, liver disea se and splenectomy. Thromb Re s40: 21, 1985.26. Hebbel RP, Steinberg MH, Mosher D: Plasma and platelet thrombo-spondin levels in sickle disease. Clin Res 41:726A, 1993.

27. Tomer A, Harker LA, Eckman JR: Activation of platelets and plateletand red cell procoagulant activity in sickle cell disease (abstract 196).National Sickle Cell Disease Program, Nineteenth Annual Meeting,March 23-26, 1 994.

28. Akinola NO, Stevens SME, Franklin IM, Nash GB, Stuart J: Subclinicalischaemic episodes during the steady state of sickle cell anaemia. JClin Pathol 45:902, 1992.

29. Cermak J, Ke y NS, Bach RR , Balla J, Jacob HS , Vercellotti GM: C-reactive protein induces human peripheral blood monocytes to synthe-size tissue factor. Blood 82513, 1993.30. Kurantsin-M ills J. Ofosu FA, Safa TK, Siege1 RS. Lessin LS: Plasmafactor VII and thrombin-antithrombin I11 levels indicate increased

tissue factor activity in sickle cell patients. Br J Haematol8 1:539,19 92.