Flow cytometric assessment of platelet function in patients with peripheral arterial occlusive disease Spencer W. Gait, M D , M a r t h a D. McDaniel, MD, Kenneth A. Ault, MD, Jane Mitchell, BS and Jack L. C r o n e n w e t t , M D Lebanon, N.H., White River Junction, Vt.,
and Portland Me. This study compared new and traditional measures ofplatelet function in 16 patients with severe peripheral arterial occlusive disease and 15 age-matched controls. Circulating platelets 'were characterized by the use of fluorescence flow cytometry to assess platelet aggregate formation and expression of the secretion-dependent alpha granule membrane protein GMP-140, by measurement of plasma [~-thromboglobulin (f~-TG), and by performance of platelet-rich plasma aggregation studies. In addition, blood samples were treated with graded concentrations of adenosine diphosphate (ADP; 0 to 10 ~tmol/L) to characterize by fluorescence flow cytometry the secretory and aggregatory responses to mild stimulation. No differences were detected between the two groups with regard to platelet fianction in unstimulated circulating blood by use of these techniques. Values (mean --- SEM) observed were: GMP-140-positive platelets, 11% -+ 3% versus 13% + 2%; platelet aggregates in circulating whole blood, 4% -+ 1% versus 9% - 3%; plasma [~-TG, 92 - 12 versus 94 - 22 ng/ml; and ED so (concentration of ADP required to produce half maximal aggregation), 3.8 +- 1.1 versus 3.1 + 0.5 ~mol/L in the patients with peripheral arterial occlusive disease and controls, respectively. Treatment with ADP caused a dose-related increase in GMP-140 expression in both groups, without significant differences in this parameter between the groups at any given concentration. However, stimulation with ADP concentrations greater than 1 ~mol/L resulted in more frequent aggregate formation in the control than in the peripheral arterial occlusive disease group (25% -+ 4% versus 11% -+ 2%, respectively at 5.0 izmol/L,p = 0.002). Further application of flow cytometry may improve our understanding of platelet function in arterial occlusive disease. (J VASC SURG 1991;14:747-56.)
Abnormal platelet physiology has been implicated in the parhogenesis o f thrombotic events, neointimal hyperplasia, and atherosclerosis. Historically, the functional status o f circulating platelets has been characterized by quantitation o f spontaneous or From the Sectionof Vascular Surgery, (Drs. Gait, McDaniel, and Cronenwett) Dartmouth-HitchcockMedical Center, Lebanon, NH, and White River Junction, Vt., and the Maine Cytometry Research Institute, (Dr. Atilt and Ms. Mitchell) and Maine Medical Center, Portland, Maine. Supported by BRSG SO7RR05392, awarded by the Biomedical Research Grant Program of Research Resources, National Institutes of Health, and by the American Heart Association, New Hampshire AJ[fihate,Grant No. NH-88-G-11. Dr. Gait was the recipient of the Tiffany Blake Fellowship of the HitchcockFoundation, and Dr. McDanielwas supported by the Depamnent of Veterans AffairsCareerDevelopmentProgram. Presented at the Fifth Annual Meeting of the Eastern Vascular Society Pittsburgh Pa. May 2-5, 1991. Reprint Requests: MalxhaD. McDaniel,MD, Sectionof Vascular Surgery, Dartmouth-HitchcockMedical Center, One Medical Center Dr., Lebanon, NH 03756. 24/6/33419
agonist-induced aggregation o f platelet-rich plasma (PRP), or by measurement o f plasma levels o f platelet-derived secretory products such as betathromboglobulin ([3-TG). Both techniques have similar deficiencies. First, the requisite in vitro manipulation o f live platelets can artefactually stimulate secretion and aggregation and thereby obscure the observation o f subtle alterations occuring in vivo. Second, these methods measure the overall status o f platelet population, and cannot detect heterogeneity among platelets. Consequently, it is not surprising that studies using these tests have produced conflicting information regarding circulating platelet "activation" both in patients with quiescent and with unstable atherosclerosis, a3 Fluorescence flow cytometry (FCM) is a powerful technique that can provide detailed information about platelet physiology. Conformational changes o f membrane receptors,4-Tdefects in membrane receptors, 8,9 surface expression o f alpha-granule mem747
748
lournai of VASCULAR SURGERY
Gait et aL
brane proteins, 4,6,10-12rebinding of secreted alphagranule proteins, ~3 platelet-bound factor XIII, ~4 platelet-bound fibrinogen, 4'6 membrane-associated ligand- and receptor-induced binding s i t e s , 6 and platelet aggregate formation ~° can be assayed by use of FCM. Because these parameters can be measured in blood fixed in paraformaldehyde immediately after phlebotomy, artefacts are minimized. Since data are obtained from individual platelets, separate subpopulations can potentially be identified. In addition, FCM affords the oppommity to study multiple parameters in a small blood sample, typically less than 1 ml. However, although FCM has proven to be a successful investigational tool for the detection of platelet activation, its clinical utility remains largely unexplored. Platelet alpha granules contain proteins both mitogenic and chemotactic for smooth muscle cells) s-~7 Since neointimal hyperplasia and early atherosclerosis both involve abnormal proliferation of smooth muscle cells, and since there may be instances in which platelets contribute to the development of these lesions, ~8 we elected to study a marker of alpha granule secretion in patients with peripheral arterial occlusive disease (PAOD) and in healthy age-matched controls. The 140 kilodalton alpha granule membrane protein GMP-140 is expressed on the surface membrane of platelets only after secretion of alpha granule contents. 19'2°Since it is not reinternalized, 2~ GMP-140 is a potential marker for alpha granule secretion in circulating platelets. In the present study, we characterized the function of circulating platelets by performing doseresponse aggregometry, measuring plasma levels of the alpha granule protein J3-TG, and using FCM to quantitate circulating platelet aggregates and platelet surface binding of the murine monoclonal antibody S1222 to GMP-140. To characterize the reserve capacity of the platelets, we assessed changes in the flow cytometric parameters after whole blood stimulation with graded concentrations of adenosine dlphosphate (ADP).
PATIENTS AND METHODS Subject selection Dartmouth-Hitchcock Medical Center outpatients between 40 and 90 years of age were eligible for inclusion in the study if they were medically stable, had not taken dipyridamole, aspirin, or other nonsteroidal antiinflammatory medication within 10 days, or warfarin within 1 month, had no history of a bleeding disorder, had normal renal function (serum creatinine _
-
+:~:+ ....
torwaro angm light scatter
(a)
I~
~
~
~i
102
FSC-H...FSC-He i 9h t
i~ ~
l!i~
---}
log forward angle light scatter
(b)
Fig. i. Sample output from the flow cytometer, examining peripheral venous blood fixed in 2% paraformaldehyde then stained with PLT-1/FITC and biotin S12/streptavidin-phycoeccthrin, (a) immediately after phlebotomy or (b) after 5 minutes' stimulation with 20 ~mol/L ADP. Previous gating on FITC fluorescence ensures that all particles in both plots are platelets. Log phycoerythrin fluorescence, on the ordinate, is proportional to S12 binding; log forward-angle light scatter, on the abscissa, is proportional to particle size. Quadrants are established by the use of appropriate controls. The line at x = 191 is the size threshold; particles falling below this value are single platelets, whereas those exceeding it are aggregates. The line at y = 5.05 is the log phycoerythrin fluorescence threshold; particles above it exhibit greater phycoerythrin fluorescence than nonspecifically stained platelets. By summing the percent of total events that occur in the upper two quadrants, the total number of platelet-containing particles that are "GMP-140 positive" is determined. Similarly, by summing the percent of total events that occur in the two quadrants to the right, the percent ofplatelet aggregates is determined. In (a) there are fewer GMP-140 positive platelets and fewer aggregates than in (b), where secretion and aggregation have occurred in response to ADP stimulation. cally, the EDs0 (doses of ADP required to achieve half-maximal change in light transmission in PRP) were 3.8 + 1.1 and 3.1 + 0.5 t+mol/L, and plasma p-TG 92 + 12 and 94 + 22 ng/ml (mean + SEM; p > 0.05 for both variables) in PAOD patients and controls, respectively. Because the plasma levels of p-TG were higher' than the laboratory normal vdues of 25 _+ 3 ng/ml, observed when the blood of healthy young subjects is collected directly into ETP, plasma PF+ was measured on the plasma samples with elevated p-TG for determination of the p-TG/PF 4 ratio. Since the clearance of PF 4 from the circulation in vivo is faster than that of ~ - T G 26, the plasma level of p-TG is ordinarily at least twice that of PF427. Ratios lower than 2.0 are a soft indicator of alpha granule secretion induced in vitro, 28 under which condition the secreted PF 4 cannot be deared. The ratios were 3.2 _+ 1.0 and 2.0 ___ 0.1 in the patients and controls, respectively (p > 0.05).
Flow cytometric characteristics of peripheral venom blood collected into heparin then fixed immediately in paraformaldehyde were similar between groups: 11% _+ 3% versus 13% _+ 2% of platelets bound detectable quantities of S12, and 4% _+ 1% versus 9% -+ 3% of platelet-containing particles appeared as aggregates in PAOD patient and control groups, respectively (p > 0.05 for both variables). Relationships between the measured parameters were also explored. Since plasma p-TG is derived only from platdets, 29 one might expect a priori to observe a strong correlation between the platelet surface expression of GMP-140 (reflecting alpha granule secretion) and the plasma p-TG level. We did not observe such a correlation: the relationship between plasma p-TG and % S12-positive platelets in immediately fixed blood was characterized by r = 0 . 0 0 4 (p=0.8) in patients with PAOD and r = 0 . 0 3 2 (p=0.5) in controls. Similarly, since hy-
752
Iournal of VASCULAR SURGERY
Galt et al.
==sot
°1 40
30
-t--
t~ ~. 20o
controls
--B-- PAOD
o~ 10O. 0-
L
0
~
0.1
L
0.4
I
0.6
I
1
I
2.5
L
5
I
10
ADP Concentration (pM) Fig. 2, Percent of platelets binding S12 after exposure of whole blood to increasing concentrations of ADP for 5 minutes in vitro. Data are mean _+ SEM.
Table II. Percent SI2 binding in ADP-stimulated samples, patients with PAOD versus controls A D P concentration PAOD Control
0
0.1
0.4
0.6
1.0
2.5
5.0
10.0
14+4 15 + 3
I5-+ 3 I9+4
21+4 27-+4
25-+ 5 28 +-4
29- 5 37-+ 5
38+6 40+-6
42+6 50-+ 7
43+6 50+-6
Data are mean + SEM; p = NS between groups at all concentrations.
peraggregability might be associated with circulating platelet aggregates, the correlation between flow cytometric measurement of platelet aggregates and EDso was calculated and found to be described by r = 0.091 (p = 0.27) in patients with PAOD and r = 0.036 (p = 0.51) in controls. Platelet response to ADP stimulation In the PAOD patient and control groups, both S 12 binding and aggregate formation increased with increasing concentrations of ADP. $12 binding was not different between groups at any ADP concentration (Table II, Fig. 2). However, control subjects formed a significantly higher percentage of platelet aggregates than did patients with PAOD in response to stimulation by ADP at concentrations above 1 Ixmol/L (Table III, Fig. 3). DISCUSSION
The role of the platelet in the pathogenesis of neointimal hyperoplasia, atherosclerosis, and their thrombotic complications remains unclear. Suspicion that platelets might be etiologically important in these disease processes has stemmed largely from three observations: (1) that endothelial distruption frequently precedes accelerated growth of such le-
sions, (2) that platelets adhere to and degranulate at sites of endothelial disruption, 3° and (3) that platelet alpha granules contain factors that are chemotactic and mitogenic for smooth muscle cells, 15-17 which are a prominent component of such lesions. Accumulating evidence demonstrates that smooth muscle cell growth control may reside within the arterial wall independent of platelet participation, 31'32 but little is known at present about the behavior and responses of circulating platelets in the patient with artherosclerosis. Questions remaining to be answered include (i) whether platelets circulate in a "primed" or hyperaggregable state; and (2) whether platelets are stimulated to secrete chronically and submaximally within the circulation. Expression of GMP-140 on the external membrane of the platelet reflects previous alpha granule secretion by platelets. 2° After one hour in vitro, GMP-140 is not reinternalized or lost from the external platelet membrane. 21 Therefore, surface expression of GMP-140 is a potential indicator of previous alpha granule secretion by circulating platelets. Since granule secretion is a late stage in the platelet activation process, however,~ GMP-140 expression would not be expected to be a sensitive marker for platelet activation.
Volume 14 Number 6 December 1991
Flow cytometric assessment of platelet function
753
t 25-
¢n
controls
20
PAOD
E lOQ.
5-
I
0
I
0.1
I
0.4
I
0.6
I
1
I
2.5
1
5
10
ADP Concentration (~rnol/L)
Fig. 3. Percent platelet aggregates after exposure of whole blood to increasing concentrations of ADP for 5 minutes in vitro. Data are mean + SEM. *p < 0.05 versus controls; tP = 0.002 versus controls, by ANOVA.
Table III. Percent platelet aggregates in ADP-stimulated samples, patients with P A O D versus controls ADP concentration PAOD Control
0 6_+2 9+ 3
0.1 6±2 11±4
0.4
0.6
7+ 2 14±.4
8 +2 16-+4
1.0 "9+2 17±4
.2.5
5.0
10.0
"10-+2 20+ 3
711±2 25 ± 4
"14-+2 24_+ 3
Data are mean -+ SEM. *p < 0.05 versus controls; ?p = 0.002 versus controls, by A N O V A .
We observed no difference in GMP-140 expression on the drculating platelets of patients "with P A O D as compared with controls. At least three explanations for this finding exist: (1) platelet alpha granule secretion and P A O D are not casually related; (2) GMP-140 is a senescence marker, leading to clearance of GMP-140-positive platelets by the reticuloendothelial system; or (3) peripheral venous sampling does not yield platelets that have been stimulated to secrete. Our current study does not allow distinction among these possible explanations; membrane-active drugs, such as beta-adrenergic blockers, which some of the patients with P A O D were obliged to take, may also have influenced the results. Other investigators have used FCM to study the surface expression of different antigens on the platelets of patients with atherosclerosis. Devine et a1.14observed increased platelet-associated factor XIII in the circulating blood of patients with P A O D as
compared to age-matched controls; Ejim et al? 3 showed that platelets of patients with P A O D bound more fibronectin than platelets of controls after low-dose thrombin stimulation. The physiologic significance of these findings is not yet clear. If platelets were circulating in a hyperaggregable state, one might reasonably expect to observe circulating platelet aggregates, 34 or to be able to induce platelet aggregate formation at a lower concentration of agonist than with normally aggregable platelets. The present study revealed no difference in PRP aggregability between patients with P A O D and controls. As previously observed, however, this method may not be sensitive to small differences in platelet behavior because of the extensive manipulation required in vitro. The fact that we also observed no difference in the percentage ofplatelet aggregates between patients and controls in immediately fixed whole blood studied by FCM may mean that there is no physiologically significant difference in platelet
754
Journal of VASCULAR SURGERY
Gait et aL
aggregability between groups. However, the formation of aggregates in response to supraphysiologic doses of ADP was clearly less in the patients with PAOD than in the control population. This is consistent with the hypothesis that the platelets in the former group are subjected to chronic submaximal stimulation in the circulation and are therefore hypoaggregable; a~ alternatively, the more highly aggregable platelets may be removed from the circulation, perhaps at sites of endothelial damage. It should be noted that the aggregates observed by FCM are substantially different from those formed in PRP aggregometry. The former are much smaller (about 20 ~m) and are reversible, 1° in contrast to the latter macroscopic, irreversibly aggregated particles. The exact composition of the particles designated as "aggregates" by FCM in this work is not known. Each of these particles consists at a minimum of a large number of GPIIbflIIa-positive elements (i.e., platelets); whether they are adherent solely to each other or whether leukocytes or red blood cells are present is unknown. The parameter "percent aggregates" is actually the percentage of all GPIIbfllIa-positive particles that are larger than 99% of single platelets, and therefore considerably underestimates the proportion of platelets recruited into aggregates. The actual number of platelets recruited into aggregates cannot be counted by our method, but is thus far greater in the ADP-stimulated blood of controls than in that of patients with PAOD. The plasma level of [3-TG is a complex function of production (platelet secretion) and clearance, largely by the kidney. 26The high levels of [3-TG measured in our study raised the suspicion that collection of the blood into heparin before the addition of ETP might have caused alpha granule secretion in vitro, but the [3-TG/PF 4 ratios observed in selected samples are not consistent with this hypothesis. The lack of linear correlation between plasma [3-TG and GMP-140 expression was unexpected, and may reflect interindividual or interplatelet variation in the [3-TG/GMP-140 ratio. In this study, platelets were termed "S12positive" if their phycoerythrin fluoresence was greater than that of nonspecifically stained platelets. We chose this method of data tabulation because we felt it would be most sensitive to subtle differences in the secretion characteristics of minimally stimulated platelets. However, this parameter would not be effective in differentiating among populations of platelets expressing large number of GMP-140 molecules on the external membrane. Devine et al.~4have suggested calculating the platelet activation index,
obtained by multiplying the percent antigen-positive platelets by the mean fluorescence intensity of those platelets. Although this may be useful when a large proportion of the platelets studied have crossed the secretion threshold, it is less likely to be useful in populations of minimally stimulated platelets, since the mean fluorescence intensity would be calculated by use of a small number of data points. We purposely sought to examine the platelet reaction to minimal stimulation, and would not have detected differences between populations expressing large amounts of GMP-140. Use of the platelet activation index with higher ADP concentrations, or stronger agonists, such as thrombin, may be beneficial. The present study addressed the expression of a single secretion-dependent antigen on the external platelet membrane and the presence of platelet aggregates in peripheral venous blood of patients with chronic PAOD. Flow cytometry, which allows examination of platelets fixed immediately after withdrawal from the circulation, is a technique well suited to the study of platelets in other disease processes, such as hypercoagulable states, and to the study of earlier stages of platelet activation. We plan further experiments using FCM to extend our knowledge of platelet physiology during the development of arterial occlusive disease and its thrombotic complications. REFERENCES
1. Douglas IT, Lowe GDO, Forbes CD, Prentice CRM. Plasma fibrinopeptide A and beta-thromboglobulin in patients with chest pain. Thromb Haemost 1983;50:541-2. 2. McDaniel MD, Pearce WH, Yao JST, et al. Sequential changes in coagulation and platelet function following femorotibial bypass. J VASCSURG 1984;1:261-8. 3. Sobel M, Salzman EW, Davies GC, et al. Circulating platelet products in unstable angina pectotis. Circulation 1981;63: 300-6. 4. Abrams CS, Ellison N, Budzynski AZ, Shattil SJ. Direct detection of activated platelets and platelet-derived microparrides in humans. Blood 1990;75:128-38. 5. Jackson CW, Jennings LK. Heterogeneity of fibrinogen receptor expression on platelets activated in normal plasma with ADP: analysis by flow cytometry. Br J Haematol 1989;72:407-14. 6. Scharf RE, Tomer A, Teirstein P, Ruggeri ZM, Harker LA. Flow cytometric analysis of activated circulating platelets detected during percutaneous translumjnal coronary angioplasty (PTCA). Blood 1989;74(suppl 1):32a. 7. Shattil SJ, Cunningham M, Hoxie JA. Detection of activated platelets in whole blood using activation-dependent monoclonal antibodies and flow cytometry. Blood 1987;70:30715. 8. Adelman B, Michelson AD, Handin RI, Atilt KA. Evaluation of platelet Ib by fluorescence flow cytometry. Blood 1985; 66:423-7.
Volume 14 Number 6 December 1991
9. Marti GE, Magruder L, Schuette WE, Gralnick HR. Flow cytometric analysis of platelet surface antigens. Cytometry 1988;9:448-55. 10. Ault KA, Rinder ttM, Mitchell IG, Rincler CS, Lambrew CT, Hillman RS. Correlated measurement of platelet release and aggregation in whole blood. Cytometry 1989;10:448-55. 1I. Johnston GI, Pickett EB, McEver RP, George JN. Heterogeneity of platelet secretion in response to thrombin demonstrated by fluorescence flow cytometry. Blood 19817;69: 1401-3. 12. Tschoepe D, Spangenberg P, Esser J, et al. Flow-cytometric detection of surface membrane alterations and concomitant changes in the cytoskeletal actin status of activated platelets. Cytometry 1990;11:652-6. 13. Ejim OS, Powling M1, Dandona P, Kernoff PBA, Goodall AH. A flow cytometric analysis of fibronectin binding to platelets from patients with peripheral vascular disease. Thromb Res 1990;58:519-24. 14. Devine DV, BaadestadG, Nugent D, Carter CJ. Detection of platelet activation antigen expression in patients with peripheral vascular disease using flow cytometry. Blood 1989; 74(suppl 1):34a. 15. Kaplan DR, Chao FC, Stiles CD, Antoniades HN, Scherr CD. Platelet alpha granules contain a growth factor for fibroblasts. Blood 1979;53:1043-52. 16. Ross R, Glomset J, Kariya B, Harker L. A platelet-serum factor that stimulates the proliferation of arterial smooth muscle cells in vitro. Proc Nat Acad Sci USA 1974;71:120710. 17. Stiles CD. The molecular biology of platelet-derived growth factor. Cell 1983;33:653-5. 18. Swedberg SH, Brown BG, Sigley R, Wight TN, Gordon D, Nicholls SC. Intimal fibromuscular hyperplasia at the venous anastomosis of PTFE grafts in hemodialysis patients. Circulation 1989;80:17126-36. 19. Berman CL, Yeo EL, Wencel-Drake JD, Furie BC, Ginsberg MH, Furie B. A platelet alpha granule membrane protein that is associated with the plasma membrane after activation: characterization and subcellular localization of platelet activation-dependent granule-external membrane protein. 1 Clin Invest 1986;78:130-7. 20. Stenberg PE, McEver RP, Shuman MA, Jacques y v , Bainton DF. A platelet alpha-granule membrane protein (GMP-140) is expressed on the plasma membrane after activation, y Cell Biol 1985;101:880-6. 21. George JN, Pickett EB, Saucerman S, McEver RP, Kunicki TJ. Platelet surface glycoproteins: studies on resting and activated platelets and platelet membrane microparticles in
Flow cytometric assessment of plateler function
22. 23.
24. 25.
26.
27. 28.
29.
30.
31.
32. 33. 34.
35.
755
normal subjects, and observations in patients during adult respiratory distress syndrome and cardiac surgery. J Clin Invest 1986;78:340-8. McEver RP, Martin MN. A monoclonal antibody to a membrane glycoprotein binds only to activated platelets. J Biol Chem 1984;259:9799-804. Stenberg PE, Shnman MA, Levine SP, Bainton D, Optimal techniques for the immunocytochemical demonstration of 13-thromboglobulin, phtelet factor 4, and fibrinogen in the alpha granules of unstimulated platelets. Histochem J 1984; 16:983-1001. Born GVR, Cross M1. The aggregation of blood platelets. J Physiol 1963;168:178-95. Vickers MV, Thompson SG. Sources of variabifity in dose response platelet aggregometry. Thromb Haemost 1985;53: 219-20. Dawes J, Smith RC, Pepper DS. The release, distribution, and clearance of human B-thromboglobulin and platelet factor 4. Thromb Res 1978;12:851-61. Kaplan KL, Owen J- Plasma levels of [3-thromboglobulin and platelet factor 4 as indices of platelet activation in vivo. Blood 1981;57:199-202. Files IC, Malpass TW, Yee EK, Ritchie 1L, Harker LA. Studies of human platelet (x-granule release in vivo. Blood 1981;58:607-18. Ludlam CA. Evidence for the platelet specificity of betathromboglobulin and studies on its plasma concentration in healthy individuals. Br J Haematol 1979;41:271-8. Lovaas ME, Gloviczki P, Hollier LH, Kaye MP. Quantitative effects of antiplatelet therapy on healing of the endarterectomized canine aorta. Am 1 Surg 1983;146:164-9. Golden MA, Au YPT, Kenagy RD, Clowes AW. Growth factor gene expression by intimal cells in heating polytetrafluoroethylene grafts. J VASCSURG 1990;11:580-5. ReidyMA. Proliferation of smooth muscle cells at sites distant from vasoalar injury. Arteriosclerosis 1990;10:298-305. White 1G. An overview of platelet structural physiology. Scanning Microsc 1987; 1:1677-700. Wu KK, Hoak JC. Spontaneous platelet aggregation in arterial insufficiency: mechanisms and implications. Thromb Haemost 1976;35:702-11. Umegaki K, Inoue Y, Tumita T. The appearance of "exhausted" platelets at the time of stroke in stroke-prone spontaneously hypertensive rats. Thromb Haemost 1985;54:
764-7. Submitted May 13, 1991; accepted Aug. 29, 1991.
DISCUSSION Dr. Kenneth Ouriel (Rochester, N.Y.). Dr. Gait has presented some interesting data on phtelet physiology, which is a topic that has been all but ignored by vascular surgeons both clinically and in laboratory investigaiton. We have been provided with the results of a battery of platelet function tests in two subgroups o f the population: one, atherosclerotic individuals, and two, normal controls.
Fluorescence FCM is probably the most elegant of these tests and represents an ideal method for identifying activated platelets. There exist two possible explanations for altered platelet fimction in the population with atherosclerosis. The first is that platelet function differences between the patient and control groups are primary, and the athero-
756
Gait et aL
sclerotic lesions are, in part, a function of hyperactive platelets. This hypothesis seem unlikely to be correct, as the present study quite clearly documents no increase in peripheral platelet activation in the individuals with atherosclerosis. In addition, platelet deposition, in general, is maximal in areas of high shear rate, whereas atherosclerotic lesions occur in areas of low shear rate. An alternate hypothesis is that the platelet differences are secondary and that they occur as a result of activation of the platelets as they come in contact with the deendothelialized surfaces, just like the type we find in atherosclerotic lesions, and this second hypothesis appears more likely to be correct. Why did the present study fail to find activated platelets in the peripheral blood of the atherosclerotic patient group? I feel there are two plausible explanations for this paradox. First, activated platelets may become irreversibly attached to the altered arterial surface and unavailable for analysis in the peripheral blood. The subpopulation of platelets with any degree of activation may in a sense be chelated from the circulation by the atherosclerotic surface, leaving a less active platelet population present in the peripheral blood and available for platelet function studies. Second, the nonulcerated atherosclerotic surface is not nearly as thrombogenic as one might expect, and we have documented this lack of thrombogeneity in perfusion studies of atherosclerotic and nonatherosclerotic human femoral arteries. Quite unexpectedly, we found that the surface of the atherosclerotic arteries was little more thrombogenic than the surface of their endothelialized counterparts. Thus, the mere presence of uncomplicated atherosclerosis may not comprise a surface thrombogenic enough to activate the platelet population. In this regard I have two questions for Dr. Gait. First, the platelet function profile of the atherosclerotic group appears to parallel that observed in patients on antiplatelet agents. Although patients were excluded from study if aspirin had been taken within 10 days of analysis, I wonder if it would have been beneficial to measure serum salicylate
Journal of VASCULAR SURGERY
levels in these patients just to be sure that they were not sneaking any aspirin. And second, is it possible that the small number of subjects contributed to an inordinately high probability of a type II error? Specifically, was the statistical power of the study sufficient to assure that true differences in platelet function were not missed as a reult of too few observations? Dr. Jon Cohen (New Hyde Park, N.Y). Given the relationship between heparin and platelet-derived growth factor, what do you suppose would happen to alpha granule secretion or particularly GMP-I40 in the setting of heparin, or do you plan to do that actual study? Dr. Gait. Had we not collected our samples on heparin? Is that your question? Dr. Cohen. No. The issue is, would you do the same study with heparin or do you plan to do it given heparin's relationship with PDGF? Dr. Gait. We used heparin as the anticoagulant to collect these samples, and I think it is possible that heparin does have an effect on alpha granule secretion. We thought it was necessary to use heparin as the anticoagulant because many other potential anticoagulants tend to disaggregate platelets, and specifically chose to aviod that effect. I thank Dr. Ouriel for his comments. We did indeed exclude anyone who had taken aspirin or any other nonsteroidal antiinflammatory drug, or any antiplatelet or anticoagulant medication. It may have been worthwhile to measure a salicylate level as well. As far as the number of patients is concerned, we did see a trend toward less expression of GMP-140 in the atherosclerotic than in the control population. The fact that we did not observe a statistically significant difference between the populations we studied means to us that differences between the populations with regard to GMP140 expression on platelets in the venous circulation are likely not pathologically significant. I think that we have quite a few other things to look at here, and we will continue on, following your suggestions.
and Portland Me. This study compared new and traditional measures ofplatelet function in 16 patients with severe peripheral arterial occlusive disease and 15 age-matched controls. Circulating platelets 'were characterized by the use of fluorescence flow cytometry to assess platelet aggregate formation and expression of the secretion-dependent alpha granule membrane protein GMP-140, by measurement of plasma [~-thromboglobulin (f~-TG), and by performance of platelet-rich plasma aggregation studies. In addition, blood samples were treated with graded concentrations of adenosine diphosphate (ADP; 0 to 10 ~tmol/L) to characterize by fluorescence flow cytometry the secretory and aggregatory responses to mild stimulation. No differences were detected between the two groups with regard to platelet fianction in unstimulated circulating blood by use of these techniques. Values (mean --- SEM) observed were: GMP-140-positive platelets, 11% -+ 3% versus 13% + 2%; platelet aggregates in circulating whole blood, 4% -+ 1% versus 9% - 3%; plasma [~-TG, 92 - 12 versus 94 - 22 ng/ml; and ED so (concentration of ADP required to produce half maximal aggregation), 3.8 +- 1.1 versus 3.1 + 0.5 ~mol/L in the patients with peripheral arterial occlusive disease and controls, respectively. Treatment with ADP caused a dose-related increase in GMP-140 expression in both groups, without significant differences in this parameter between the groups at any given concentration. However, stimulation with ADP concentrations greater than 1 ~mol/L resulted in more frequent aggregate formation in the control than in the peripheral arterial occlusive disease group (25% -+ 4% versus 11% -+ 2%, respectively at 5.0 izmol/L,p = 0.002). Further application of flow cytometry may improve our understanding of platelet function in arterial occlusive disease. (J VASC SURG 1991;14:747-56.)
Abnormal platelet physiology has been implicated in the parhogenesis o f thrombotic events, neointimal hyperplasia, and atherosclerosis. Historically, the functional status o f circulating platelets has been characterized by quantitation o f spontaneous or From the Sectionof Vascular Surgery, (Drs. Gait, McDaniel, and Cronenwett) Dartmouth-HitchcockMedical Center, Lebanon, NH, and White River Junction, Vt., and the Maine Cytometry Research Institute, (Dr. Atilt and Ms. Mitchell) and Maine Medical Center, Portland, Maine. Supported by BRSG SO7RR05392, awarded by the Biomedical Research Grant Program of Research Resources, National Institutes of Health, and by the American Heart Association, New Hampshire AJ[fihate,Grant No. NH-88-G-11. Dr. Gait was the recipient of the Tiffany Blake Fellowship of the HitchcockFoundation, and Dr. McDanielwas supported by the Depamnent of Veterans AffairsCareerDevelopmentProgram. Presented at the Fifth Annual Meeting of the Eastern Vascular Society Pittsburgh Pa. May 2-5, 1991. Reprint Requests: MalxhaD. McDaniel,MD, Sectionof Vascular Surgery, Dartmouth-HitchcockMedical Center, One Medical Center Dr., Lebanon, NH 03756. 24/6/33419
agonist-induced aggregation o f platelet-rich plasma (PRP), or by measurement o f plasma levels o f platelet-derived secretory products such as betathromboglobulin ([3-TG). Both techniques have similar deficiencies. First, the requisite in vitro manipulation o f live platelets can artefactually stimulate secretion and aggregation and thereby obscure the observation o f subtle alterations occuring in vivo. Second, these methods measure the overall status o f platelet population, and cannot detect heterogeneity among platelets. Consequently, it is not surprising that studies using these tests have produced conflicting information regarding circulating platelet "activation" both in patients with quiescent and with unstable atherosclerosis, a3 Fluorescence flow cytometry (FCM) is a powerful technique that can provide detailed information about platelet physiology. Conformational changes o f membrane receptors,4-Tdefects in membrane receptors, 8,9 surface expression o f alpha-granule mem747
748
lournai of VASCULAR SURGERY
Gait et aL
brane proteins, 4,6,10-12rebinding of secreted alphagranule proteins, ~3 platelet-bound factor XIII, ~4 platelet-bound fibrinogen, 4'6 membrane-associated ligand- and receptor-induced binding s i t e s , 6 and platelet aggregate formation ~° can be assayed by use of FCM. Because these parameters can be measured in blood fixed in paraformaldehyde immediately after phlebotomy, artefacts are minimized. Since data are obtained from individual platelets, separate subpopulations can potentially be identified. In addition, FCM affords the oppommity to study multiple parameters in a small blood sample, typically less than 1 ml. However, although FCM has proven to be a successful investigational tool for the detection of platelet activation, its clinical utility remains largely unexplored. Platelet alpha granules contain proteins both mitogenic and chemotactic for smooth muscle cells) s-~7 Since neointimal hyperplasia and early atherosclerosis both involve abnormal proliferation of smooth muscle cells, and since there may be instances in which platelets contribute to the development of these lesions, ~8 we elected to study a marker of alpha granule secretion in patients with peripheral arterial occlusive disease (PAOD) and in healthy age-matched controls. The 140 kilodalton alpha granule membrane protein GMP-140 is expressed on the surface membrane of platelets only after secretion of alpha granule contents. 19'2°Since it is not reinternalized, 2~ GMP-140 is a potential marker for alpha granule secretion in circulating platelets. In the present study, we characterized the function of circulating platelets by performing doseresponse aggregometry, measuring plasma levels of the alpha granule protein J3-TG, and using FCM to quantitate circulating platelet aggregates and platelet surface binding of the murine monoclonal antibody S1222 to GMP-140. To characterize the reserve capacity of the platelets, we assessed changes in the flow cytometric parameters after whole blood stimulation with graded concentrations of adenosine dlphosphate (ADP).
PATIENTS AND METHODS Subject selection Dartmouth-Hitchcock Medical Center outpatients between 40 and 90 years of age were eligible for inclusion in the study if they were medically stable, had not taken dipyridamole, aspirin, or other nonsteroidal antiinflammatory medication within 10 days, or warfarin within 1 month, had no history of a bleeding disorder, had normal renal function (serum creatinine _
-
+:~:+ ....
torwaro angm light scatter
(a)
I~
~
~
~i
102
FSC-H...FSC-He i 9h t
i~ ~
l!i~
---}
log forward angle light scatter
(b)
Fig. i. Sample output from the flow cytometer, examining peripheral venous blood fixed in 2% paraformaldehyde then stained with PLT-1/FITC and biotin S12/streptavidin-phycoeccthrin, (a) immediately after phlebotomy or (b) after 5 minutes' stimulation with 20 ~mol/L ADP. Previous gating on FITC fluorescence ensures that all particles in both plots are platelets. Log phycoerythrin fluorescence, on the ordinate, is proportional to S12 binding; log forward-angle light scatter, on the abscissa, is proportional to particle size. Quadrants are established by the use of appropriate controls. The line at x = 191 is the size threshold; particles falling below this value are single platelets, whereas those exceeding it are aggregates. The line at y = 5.05 is the log phycoerythrin fluorescence threshold; particles above it exhibit greater phycoerythrin fluorescence than nonspecifically stained platelets. By summing the percent of total events that occur in the upper two quadrants, the total number of platelet-containing particles that are "GMP-140 positive" is determined. Similarly, by summing the percent of total events that occur in the two quadrants to the right, the percent ofplatelet aggregates is determined. In (a) there are fewer GMP-140 positive platelets and fewer aggregates than in (b), where secretion and aggregation have occurred in response to ADP stimulation. cally, the EDs0 (doses of ADP required to achieve half-maximal change in light transmission in PRP) were 3.8 + 1.1 and 3.1 + 0.5 t+mol/L, and plasma p-TG 92 + 12 and 94 + 22 ng/ml (mean + SEM; p > 0.05 for both variables) in PAOD patients and controls, respectively. Because the plasma levels of p-TG were higher' than the laboratory normal vdues of 25 _+ 3 ng/ml, observed when the blood of healthy young subjects is collected directly into ETP, plasma PF+ was measured on the plasma samples with elevated p-TG for determination of the p-TG/PF 4 ratio. Since the clearance of PF 4 from the circulation in vivo is faster than that of ~ - T G 26, the plasma level of p-TG is ordinarily at least twice that of PF427. Ratios lower than 2.0 are a soft indicator of alpha granule secretion induced in vitro, 28 under which condition the secreted PF 4 cannot be deared. The ratios were 3.2 _+ 1.0 and 2.0 ___ 0.1 in the patients and controls, respectively (p > 0.05).
Flow cytometric characteristics of peripheral venom blood collected into heparin then fixed immediately in paraformaldehyde were similar between groups: 11% _+ 3% versus 13% _+ 2% of platelets bound detectable quantities of S12, and 4% _+ 1% versus 9% -+ 3% of platelet-containing particles appeared as aggregates in PAOD patient and control groups, respectively (p > 0.05 for both variables). Relationships between the measured parameters were also explored. Since plasma p-TG is derived only from platdets, 29 one might expect a priori to observe a strong correlation between the platelet surface expression of GMP-140 (reflecting alpha granule secretion) and the plasma p-TG level. We did not observe such a correlation: the relationship between plasma p-TG and % S12-positive platelets in immediately fixed blood was characterized by r = 0 . 0 0 4 (p=0.8) in patients with PAOD and r = 0 . 0 3 2 (p=0.5) in controls. Similarly, since hy-
752
Iournal of VASCULAR SURGERY
Galt et al.
==sot
°1 40
30
-t--
t~ ~. 20o
controls
--B-- PAOD
o~ 10O. 0-
L
0
~
0.1
L
0.4
I
0.6
I
1
I
2.5
L
5
I
10
ADP Concentration (pM) Fig. 2, Percent of platelets binding S12 after exposure of whole blood to increasing concentrations of ADP for 5 minutes in vitro. Data are mean _+ SEM.
Table II. Percent SI2 binding in ADP-stimulated samples, patients with PAOD versus controls A D P concentration PAOD Control
0
0.1
0.4
0.6
1.0
2.5
5.0
10.0
14+4 15 + 3
I5-+ 3 I9+4
21+4 27-+4
25-+ 5 28 +-4
29- 5 37-+ 5
38+6 40+-6
42+6 50-+ 7
43+6 50+-6
Data are mean + SEM; p = NS between groups at all concentrations.
peraggregability might be associated with circulating platelet aggregates, the correlation between flow cytometric measurement of platelet aggregates and EDso was calculated and found to be described by r = 0.091 (p = 0.27) in patients with PAOD and r = 0.036 (p = 0.51) in controls. Platelet response to ADP stimulation In the PAOD patient and control groups, both S 12 binding and aggregate formation increased with increasing concentrations of ADP. $12 binding was not different between groups at any ADP concentration (Table II, Fig. 2). However, control subjects formed a significantly higher percentage of platelet aggregates than did patients with PAOD in response to stimulation by ADP at concentrations above 1 Ixmol/L (Table III, Fig. 3). DISCUSSION
The role of the platelet in the pathogenesis of neointimal hyperoplasia, atherosclerosis, and their thrombotic complications remains unclear. Suspicion that platelets might be etiologically important in these disease processes has stemmed largely from three observations: (1) that endothelial distruption frequently precedes accelerated growth of such le-
sions, (2) that platelets adhere to and degranulate at sites of endothelial disruption, 3° and (3) that platelet alpha granules contain factors that are chemotactic and mitogenic for smooth muscle cells, 15-17 which are a prominent component of such lesions. Accumulating evidence demonstrates that smooth muscle cell growth control may reside within the arterial wall independent of platelet participation, 31'32 but little is known at present about the behavior and responses of circulating platelets in the patient with artherosclerosis. Questions remaining to be answered include (i) whether platelets circulate in a "primed" or hyperaggregable state; and (2) whether platelets are stimulated to secrete chronically and submaximally within the circulation. Expression of GMP-140 on the external membrane of the platelet reflects previous alpha granule secretion by platelets. 2° After one hour in vitro, GMP-140 is not reinternalized or lost from the external platelet membrane. 21 Therefore, surface expression of GMP-140 is a potential indicator of previous alpha granule secretion by circulating platelets. Since granule secretion is a late stage in the platelet activation process, however,~ GMP-140 expression would not be expected to be a sensitive marker for platelet activation.
Volume 14 Number 6 December 1991
Flow cytometric assessment of platelet function
753
t 25-
¢n
controls
20
PAOD
E lOQ.
5-
I
0
I
0.1
I
0.4
I
0.6
I
1
I
2.5
1
5
10
ADP Concentration (~rnol/L)
Fig. 3. Percent platelet aggregates after exposure of whole blood to increasing concentrations of ADP for 5 minutes in vitro. Data are mean + SEM. *p < 0.05 versus controls; tP = 0.002 versus controls, by ANOVA.
Table III. Percent platelet aggregates in ADP-stimulated samples, patients with P A O D versus controls ADP concentration PAOD Control
0 6_+2 9+ 3
0.1 6±2 11±4
0.4
0.6
7+ 2 14±.4
8 +2 16-+4
1.0 "9+2 17±4
.2.5
5.0
10.0
"10-+2 20+ 3
711±2 25 ± 4
"14-+2 24_+ 3
Data are mean -+ SEM. *p < 0.05 versus controls; ?p = 0.002 versus controls, by A N O V A .
We observed no difference in GMP-140 expression on the drculating platelets of patients "with P A O D as compared with controls. At least three explanations for this finding exist: (1) platelet alpha granule secretion and P A O D are not casually related; (2) GMP-140 is a senescence marker, leading to clearance of GMP-140-positive platelets by the reticuloendothelial system; or (3) peripheral venous sampling does not yield platelets that have been stimulated to secrete. Our current study does not allow distinction among these possible explanations; membrane-active drugs, such as beta-adrenergic blockers, which some of the patients with P A O D were obliged to take, may also have influenced the results. Other investigators have used FCM to study the surface expression of different antigens on the platelets of patients with atherosclerosis. Devine et a1.14observed increased platelet-associated factor XIII in the circulating blood of patients with P A O D as
compared to age-matched controls; Ejim et al? 3 showed that platelets of patients with P A O D bound more fibronectin than platelets of controls after low-dose thrombin stimulation. The physiologic significance of these findings is not yet clear. If platelets were circulating in a hyperaggregable state, one might reasonably expect to observe circulating platelet aggregates, 34 or to be able to induce platelet aggregate formation at a lower concentration of agonist than with normally aggregable platelets. The present study revealed no difference in PRP aggregability between patients with P A O D and controls. As previously observed, however, this method may not be sensitive to small differences in platelet behavior because of the extensive manipulation required in vitro. The fact that we also observed no difference in the percentage ofplatelet aggregates between patients and controls in immediately fixed whole blood studied by FCM may mean that there is no physiologically significant difference in platelet
754
Journal of VASCULAR SURGERY
Gait et aL
aggregability between groups. However, the formation of aggregates in response to supraphysiologic doses of ADP was clearly less in the patients with PAOD than in the control population. This is consistent with the hypothesis that the platelets in the former group are subjected to chronic submaximal stimulation in the circulation and are therefore hypoaggregable; a~ alternatively, the more highly aggregable platelets may be removed from the circulation, perhaps at sites of endothelial damage. It should be noted that the aggregates observed by FCM are substantially different from those formed in PRP aggregometry. The former are much smaller (about 20 ~m) and are reversible, 1° in contrast to the latter macroscopic, irreversibly aggregated particles. The exact composition of the particles designated as "aggregates" by FCM in this work is not known. Each of these particles consists at a minimum of a large number of GPIIbflIIa-positive elements (i.e., platelets); whether they are adherent solely to each other or whether leukocytes or red blood cells are present is unknown. The parameter "percent aggregates" is actually the percentage of all GPIIbfllIa-positive particles that are larger than 99% of single platelets, and therefore considerably underestimates the proportion of platelets recruited into aggregates. The actual number of platelets recruited into aggregates cannot be counted by our method, but is thus far greater in the ADP-stimulated blood of controls than in that of patients with PAOD. The plasma level of [3-TG is a complex function of production (platelet secretion) and clearance, largely by the kidney. 26The high levels of [3-TG measured in our study raised the suspicion that collection of the blood into heparin before the addition of ETP might have caused alpha granule secretion in vitro, but the [3-TG/PF 4 ratios observed in selected samples are not consistent with this hypothesis. The lack of linear correlation between plasma [3-TG and GMP-140 expression was unexpected, and may reflect interindividual or interplatelet variation in the [3-TG/GMP-140 ratio. In this study, platelets were termed "S12positive" if their phycoerythrin fluoresence was greater than that of nonspecifically stained platelets. We chose this method of data tabulation because we felt it would be most sensitive to subtle differences in the secretion characteristics of minimally stimulated platelets. However, this parameter would not be effective in differentiating among populations of platelets expressing large number of GMP-140 molecules on the external membrane. Devine et al.~4have suggested calculating the platelet activation index,
obtained by multiplying the percent antigen-positive platelets by the mean fluorescence intensity of those platelets. Although this may be useful when a large proportion of the platelets studied have crossed the secretion threshold, it is less likely to be useful in populations of minimally stimulated platelets, since the mean fluorescence intensity would be calculated by use of a small number of data points. We purposely sought to examine the platelet reaction to minimal stimulation, and would not have detected differences between populations expressing large amounts of GMP-140. Use of the platelet activation index with higher ADP concentrations, or stronger agonists, such as thrombin, may be beneficial. The present study addressed the expression of a single secretion-dependent antigen on the external platelet membrane and the presence of platelet aggregates in peripheral venous blood of patients with chronic PAOD. Flow cytometry, which allows examination of platelets fixed immediately after withdrawal from the circulation, is a technique well suited to the study of platelets in other disease processes, such as hypercoagulable states, and to the study of earlier stages of platelet activation. We plan further experiments using FCM to extend our knowledge of platelet physiology during the development of arterial occlusive disease and its thrombotic complications. REFERENCES
1. Douglas IT, Lowe GDO, Forbes CD, Prentice CRM. Plasma fibrinopeptide A and beta-thromboglobulin in patients with chest pain. Thromb Haemost 1983;50:541-2. 2. McDaniel MD, Pearce WH, Yao JST, et al. Sequential changes in coagulation and platelet function following femorotibial bypass. J VASCSURG 1984;1:261-8. 3. Sobel M, Salzman EW, Davies GC, et al. Circulating platelet products in unstable angina pectotis. Circulation 1981;63: 300-6. 4. Abrams CS, Ellison N, Budzynski AZ, Shattil SJ. Direct detection of activated platelets and platelet-derived microparrides in humans. Blood 1990;75:128-38. 5. Jackson CW, Jennings LK. Heterogeneity of fibrinogen receptor expression on platelets activated in normal plasma with ADP: analysis by flow cytometry. Br J Haematol 1989;72:407-14. 6. Scharf RE, Tomer A, Teirstein P, Ruggeri ZM, Harker LA. Flow cytometric analysis of activated circulating platelets detected during percutaneous translumjnal coronary angioplasty (PTCA). Blood 1989;74(suppl 1):32a. 7. Shattil SJ, Cunningham M, Hoxie JA. Detection of activated platelets in whole blood using activation-dependent monoclonal antibodies and flow cytometry. Blood 1987;70:30715. 8. Adelman B, Michelson AD, Handin RI, Atilt KA. Evaluation of platelet Ib by fluorescence flow cytometry. Blood 1985; 66:423-7.
Volume 14 Number 6 December 1991
9. Marti GE, Magruder L, Schuette WE, Gralnick HR. Flow cytometric analysis of platelet surface antigens. Cytometry 1988;9:448-55. 10. Ault KA, Rinder ttM, Mitchell IG, Rincler CS, Lambrew CT, Hillman RS. Correlated measurement of platelet release and aggregation in whole blood. Cytometry 1989;10:448-55. 1I. Johnston GI, Pickett EB, McEver RP, George JN. Heterogeneity of platelet secretion in response to thrombin demonstrated by fluorescence flow cytometry. Blood 19817;69: 1401-3. 12. Tschoepe D, Spangenberg P, Esser J, et al. Flow-cytometric detection of surface membrane alterations and concomitant changes in the cytoskeletal actin status of activated platelets. Cytometry 1990;11:652-6. 13. Ejim OS, Powling M1, Dandona P, Kernoff PBA, Goodall AH. A flow cytometric analysis of fibronectin binding to platelets from patients with peripheral vascular disease. Thromb Res 1990;58:519-24. 14. Devine DV, BaadestadG, Nugent D, Carter CJ. Detection of platelet activation antigen expression in patients with peripheral vascular disease using flow cytometry. Blood 1989; 74(suppl 1):34a. 15. Kaplan DR, Chao FC, Stiles CD, Antoniades HN, Scherr CD. Platelet alpha granules contain a growth factor for fibroblasts. Blood 1979;53:1043-52. 16. Ross R, Glomset J, Kariya B, Harker L. A platelet-serum factor that stimulates the proliferation of arterial smooth muscle cells in vitro. Proc Nat Acad Sci USA 1974;71:120710. 17. Stiles CD. The molecular biology of platelet-derived growth factor. Cell 1983;33:653-5. 18. Swedberg SH, Brown BG, Sigley R, Wight TN, Gordon D, Nicholls SC. Intimal fibromuscular hyperplasia at the venous anastomosis of PTFE grafts in hemodialysis patients. Circulation 1989;80:17126-36. 19. Berman CL, Yeo EL, Wencel-Drake JD, Furie BC, Ginsberg MH, Furie B. A platelet alpha granule membrane protein that is associated with the plasma membrane after activation: characterization and subcellular localization of platelet activation-dependent granule-external membrane protein. 1 Clin Invest 1986;78:130-7. 20. Stenberg PE, McEver RP, Shuman MA, Jacques y v , Bainton DF. A platelet alpha-granule membrane protein (GMP-140) is expressed on the plasma membrane after activation, y Cell Biol 1985;101:880-6. 21. George JN, Pickett EB, Saucerman S, McEver RP, Kunicki TJ. Platelet surface glycoproteins: studies on resting and activated platelets and platelet membrane microparticles in
Flow cytometric assessment of plateler function
22. 23.
24. 25.
26.
27. 28.
29.
30.
31.
32. 33. 34.
35.
755
normal subjects, and observations in patients during adult respiratory distress syndrome and cardiac surgery. J Clin Invest 1986;78:340-8. McEver RP, Martin MN. A monoclonal antibody to a membrane glycoprotein binds only to activated platelets. J Biol Chem 1984;259:9799-804. Stenberg PE, Shnman MA, Levine SP, Bainton D, Optimal techniques for the immunocytochemical demonstration of 13-thromboglobulin, phtelet factor 4, and fibrinogen in the alpha granules of unstimulated platelets. Histochem J 1984; 16:983-1001. Born GVR, Cross M1. The aggregation of blood platelets. J Physiol 1963;168:178-95. Vickers MV, Thompson SG. Sources of variabifity in dose response platelet aggregometry. Thromb Haemost 1985;53: 219-20. Dawes J, Smith RC, Pepper DS. The release, distribution, and clearance of human B-thromboglobulin and platelet factor 4. Thromb Res 1978;12:851-61. Kaplan KL, Owen J- Plasma levels of [3-thromboglobulin and platelet factor 4 as indices of platelet activation in vivo. Blood 1981;57:199-202. Files IC, Malpass TW, Yee EK, Ritchie 1L, Harker LA. Studies of human platelet (x-granule release in vivo. Blood 1981;58:607-18. Ludlam CA. Evidence for the platelet specificity of betathromboglobulin and studies on its plasma concentration in healthy individuals. Br J Haematol 1979;41:271-8. Lovaas ME, Gloviczki P, Hollier LH, Kaye MP. Quantitative effects of antiplatelet therapy on healing of the endarterectomized canine aorta. Am 1 Surg 1983;146:164-9. Golden MA, Au YPT, Kenagy RD, Clowes AW. Growth factor gene expression by intimal cells in heating polytetrafluoroethylene grafts. J VASCSURG 1990;11:580-5. ReidyMA. Proliferation of smooth muscle cells at sites distant from vasoalar injury. Arteriosclerosis 1990;10:298-305. White 1G. An overview of platelet structural physiology. Scanning Microsc 1987; 1:1677-700. Wu KK, Hoak JC. Spontaneous platelet aggregation in arterial insufficiency: mechanisms and implications. Thromb Haemost 1976;35:702-11. Umegaki K, Inoue Y, Tumita T. The appearance of "exhausted" platelets at the time of stroke in stroke-prone spontaneously hypertensive rats. Thromb Haemost 1985;54:
764-7. Submitted May 13, 1991; accepted Aug. 29, 1991.
DISCUSSION Dr. Kenneth Ouriel (Rochester, N.Y.). Dr. Gait has presented some interesting data on phtelet physiology, which is a topic that has been all but ignored by vascular surgeons both clinically and in laboratory investigaiton. We have been provided with the results of a battery of platelet function tests in two subgroups o f the population: one, atherosclerotic individuals, and two, normal controls.
Fluorescence FCM is probably the most elegant of these tests and represents an ideal method for identifying activated platelets. There exist two possible explanations for altered platelet fimction in the population with atherosclerosis. The first is that platelet function differences between the patient and control groups are primary, and the athero-
756
Gait et aL
sclerotic lesions are, in part, a function of hyperactive platelets. This hypothesis seem unlikely to be correct, as the present study quite clearly documents no increase in peripheral platelet activation in the individuals with atherosclerosis. In addition, platelet deposition, in general, is maximal in areas of high shear rate, whereas atherosclerotic lesions occur in areas of low shear rate. An alternate hypothesis is that the platelet differences are secondary and that they occur as a result of activation of the platelets as they come in contact with the deendothelialized surfaces, just like the type we find in atherosclerotic lesions, and this second hypothesis appears more likely to be correct. Why did the present study fail to find activated platelets in the peripheral blood of the atherosclerotic patient group? I feel there are two plausible explanations for this paradox. First, activated platelets may become irreversibly attached to the altered arterial surface and unavailable for analysis in the peripheral blood. The subpopulation of platelets with any degree of activation may in a sense be chelated from the circulation by the atherosclerotic surface, leaving a less active platelet population present in the peripheral blood and available for platelet function studies. Second, the nonulcerated atherosclerotic surface is not nearly as thrombogenic as one might expect, and we have documented this lack of thrombogeneity in perfusion studies of atherosclerotic and nonatherosclerotic human femoral arteries. Quite unexpectedly, we found that the surface of the atherosclerotic arteries was little more thrombogenic than the surface of their endothelialized counterparts. Thus, the mere presence of uncomplicated atherosclerosis may not comprise a surface thrombogenic enough to activate the platelet population. In this regard I have two questions for Dr. Gait. First, the platelet function profile of the atherosclerotic group appears to parallel that observed in patients on antiplatelet agents. Although patients were excluded from study if aspirin had been taken within 10 days of analysis, I wonder if it would have been beneficial to measure serum salicylate
Journal of VASCULAR SURGERY
levels in these patients just to be sure that they were not sneaking any aspirin. And second, is it possible that the small number of subjects contributed to an inordinately high probability of a type II error? Specifically, was the statistical power of the study sufficient to assure that true differences in platelet function were not missed as a reult of too few observations? Dr. Jon Cohen (New Hyde Park, N.Y). Given the relationship between heparin and platelet-derived growth factor, what do you suppose would happen to alpha granule secretion or particularly GMP-I40 in the setting of heparin, or do you plan to do that actual study? Dr. Gait. Had we not collected our samples on heparin? Is that your question? Dr. Cohen. No. The issue is, would you do the same study with heparin or do you plan to do it given heparin's relationship with PDGF? Dr. Gait. We used heparin as the anticoagulant to collect these samples, and I think it is possible that heparin does have an effect on alpha granule secretion. We thought it was necessary to use heparin as the anticoagulant because many other potential anticoagulants tend to disaggregate platelets, and specifically chose to aviod that effect. I thank Dr. Ouriel for his comments. We did indeed exclude anyone who had taken aspirin or any other nonsteroidal antiinflammatory drug, or any antiplatelet or anticoagulant medication. It may have been worthwhile to measure a salicylate level as well. As far as the number of patients is concerned, we did see a trend toward less expression of GMP-140 in the atherosclerotic than in the control population. The fact that we did not observe a statistically significant difference between the populations we studied means to us that differences between the populations with regard to GMP140 expression on platelets in the venous circulation are likely not pathologically significant. I think that we have quite a few other things to look at here, and we will continue on, following your suggestions.
