Effect of hematocrit on adenosine diphosphate-induced aggregation of human platelets in tube flow

t3iortrcwlogy,

Pergamon

Vol 32, No. 5, pp. 537-552, 1995 (bpyright 0 1995 Elsevier Science Ltd Printed m the USA. Ail rights reserved 0oolc355X/95 $9.50 + .oo

0006-355x(95)00031-3

EFFECT OF HEMATOCRIT INDUCED AGGREGATION TUBE FLOW H. L. GOLDSMITH,t:

E. S. KAUFER:

fMcGil1 University Medical CANADA H3G lA4 t=Department

ON ADENOSINE OF HUMAN

of Medicine,

Clinic,

McGill

AND F. A. MCINTOSH: Montreal

University,

Please address correspondence to: Montreal General Hospital, Montreal,

DIPHOSPHATE PLATELETS IN

General

Montreal,

Hospital,

CANADA

Montreal,

H3G lA4

Dr. H. L. Goldsmith, Room Quebec H3G lA4, CANADA

ClO-148,

ABSTRACT Both chemical and physical effects of red cells are known to play a role in the adenosine diphosphate (ADP)-induced aggregation of human platelets in sheared blood. Using a previously described double infusion technique (Bell et al, 1989a), we studied the effect of increasing hematocrit from 10 to 60% on the rate and extent of platelet aggregation with 0.2 @i ADP in titrated whole blood undergoing tube flow. Blood and agonist were rapidly mixed in a small chamber and the suspensions flowed through lengths of 1.19 mmdiameter polyethylene tubing at mean transit times from 0.2 to 42.8 s at a mean tube shear rate = 335 s-1. Effluent was collected into 0.5% glutaraldehyde, the red cells removed by centrifugation through Percoll, and all single platelets and aggregates in the volume range l-lo5 pm3 counted and Both the initial rate (over the sized using an aperture impedance counter. first 8.6 s) and the extent of aggregation with time increased with increasing mean hematocrit up to 35.8%, being significantly greater than in titrated plasma (cPRP). However, at 61.5% hematocrit, the extent of aggregation decreased markedly to a level close to that in cPRP. We also studied the effect of washed red cells at 39% hematocrit on the aggregation of washed platelets in Tyrodes-albumin fibrinogen-free suspensions. It had previously been shown that, at > 335 s-l, washed platelets in platelet-rich Tyrodes (PRT) aggregated with 0.7 pM ADP. We found that red cells markedly increased the extent of aggegation from that in PRT, and promoted the formation of large aggregates, absent in PRT. Spontaneous aggregation in whole blood or washed cell suspensions in the absence of added ADP at = 42.8 s was < 10% of that in the presence of ADP. The results indicate that a physical effect of red cells, likely manifested as an increase in the efficiency of aggregate formation plays an important role at low and normal (Goldsmith et al., 1995), hematocrits; however, at high hematocrits, particle crowding impedes the formation of aggregates.

KEY WORDS:

Platelet aggregation; fibrinogen

ADP;

537

tube

flow;

red cells;

hematocrit;

538

RBC and ADP-induced platelet

aggregation

Vol. 32, No. 5

Introduction Both chemical and physical effects of red cells have been implicated in the agonist-induced, as well as in the spontaneous, aggregation of platelets in sheared whole blood. We previously showed that red blood cells markedly increase the rate and extent of ADP-induced aggregation of human platelets flowing through small tubes (Bell et al., 199Oa). A double infusion technique was used to mix titrated whole blood and agonist rapidly and flow the suspension through tubes for various times at mean tube shear rates from 42 to 1920 s-l (wall shear stresses from 0.14 to 6.5 Nm-*) . It was found that the initial rate of aggregation with 0.2 pM ADP, as measured by the disappearance of single platelets, was -7x greater than in titrated platelet-rich plasma (cPRP) at = 42 s-l, and -5x greater at = 335 and 1920 s-l. Spontaneous aggregation (in the absence of added agonist) was insignificant in cPRP (Bell et aL, 1989b); in whole blood, it was found to be appreciable only at the lowest shear rate. Even then, the rate was still much lower than that observed with 0.2 pM ADP. Addition of the ADP scavenger enzyme system, phosphocreatine-creatine phosphokinase (CP-CPK), completely abolished spontaneous aggregation (Goldsmith et aL, 1995), indicating that the reaction was due to small amounts of ADP, most probably released from the red cells (Fox et a,!., 1982; Saniabadi et al., 1984, 1985, 1987, 1989). The release likely leads to the formation of small aggregates which are stable only at the lowest shear rates and are therefore of little consequence in the circulation of the major arteries (Goldsmith et aZ., 1995). Release of ADP from the platelets is not likely to be involved in spontaneous aggregation except at shear rates > 2,300 s-l (shear stresses > 8 Nm-*), above the critical value for shear-induced release for platelets (Dewitz et al., 1978; Jen and McIntire, 1984). Thus, it was concluded that the observed increase in the rate and extent of platelet aggregation at shear rates < 2000 s-l was mostly due to a physical effect of the red cells, as further discussed below. The experiments cited above were all carried out in whole blood at hematocrits close to 40%. The purpose of the present work was to investigate the effect of increasing the hematocrit from 10 to 60% on platelet aggregation in titrated plasma. In addition, we report here on the effect of red cells on ADP-induced platelet aggregation in washed cell suspensions in the absence of added fibrinogen. Fibrinogen is known to bind to the activated membrane glycoprotein complex, GPIIbIIIa (Phillips and Baughan, 1985; Phillips et al., 1988; Frojmovic et al, 1991a). It is believed that aggregation then occurs via cross-bridging of the bivalent fibrinogen molecules between the receptors on adjacent platelets. We showed that there is still significant, shear-dependent aggregation ( 2 335 s-t) in fibrinogen-free suspensions of washed platelets (Goldsmith et al., 1994). Although more than 50% of the washed platelets expressed fibrinogen-occupied GPIIb-IIIa receptors, the ADP-induced aggregation was shown to be independent of this bound fibrinogen. Recent work strongly suggests that von Willebrand factor, expressed on ADP-activated platelets, is able to bind the cells at moderate to high shear rates (Frojmovic et al., 1995). It was therefore of some interest to ascertain whether the addition of washed red cells to platelets in fibrinogen-free suspending media could further enhance the rate and extent of ADP-induced aggregation.

Vol.

32, No. 5

Materials

and

RBC and ADP-induced

jdatekl

aggregation

539

Methods

Hemutocrit-Adjusted Whok Blood Venous blood was slowly drawn from healthy volunteers via a 19-gauge needle and winged infusion set into a 60 ml plastic syringe containing l/10 volume sodium citrate whose concentration, cf, was adjusted to give a final concentration, cf, of 0.62% in the plasma based on the predetermined value of the hematocrit, H, of the undiluted venous blood (Bell d al, 1990a,b): Cf= Ci x [9(1 - H)

t

11.

The drawn blood was used for the “40%” hematocrit experiments without adjustment. For the “10 and 2O%‘O”hematocrit experiments, the drawn blood was diluted with appropriate volumes of titrated platelet-rich plasma (cPRP), the latter obtained by centrifuging the blood for 20 min at = 100 in 50 ml polycarbonate tubes. To obtain blood at 60% hematocrit, cPRP was prepared as described above, and the red cell layer centrifuged at = 2,000 for 15 min and let stand for 15 min at roon temperature to allow degradation-to AMP and then to adenosine-of any ADP known to be released during centrifugation (Mustard et aL, 1972). The hematocrit of the packed cells was then determined and the cells were diluted with the appropriate volume of cPRP. Due to the 1:lO dilution of the blood in the flow tube, the hematocrits in the syringe reservoirs were adjusted to 11, 22 and 66%, respectively. Washed Cell Suspensions Suspensions of washed platelets were prepared according to the procedure described by Mustard et al. (1972), slightly modified. Blood was drawn into l/7 volume of acid citrate dextrose, transferred to 50 ml polycarbonate tubes and centrifuged at = 100 for 20 min to yield cPRP. The lightly packed red cells were set aside to be washed, as described below. The cPRP to which 50 nM of the stable prostacyclin derivative, ZK 36374, was added (Shor et al., 1981) was centrifuged at = 800 for 15 min at room temperature. The platelet pellets (each about 0.15 ml in volume) were dispersed in 20 ml of modified, Ca++-, Mg ++- free Tyrodes solution, pH = 7.35 (concentration in mmol/l: NaCl, 136, KCl, 2.7, NaHCO 3, 11.9, NaHzP04, 0.36, glucose, 5.6) containing 0.35% bovine serum albumin, fraction V (Gibco BRL, Burlington, Ontario), 50 units/ml porcine heparin (Sigma Chemical Co., St. Louis, MO), 10 nM ZK and 24 mg/l apyrase. This and all subsequent washing operations were conducted at 37” C. After incubation for 20 min, the suspension was centrifuged at = 800 for 10 min, the pellets were resuspended in 20 ml of modified Tyrodes-albumin containing 24 mg/l apyrase but no ZK, again incubated for 20 min and centrifuged. The pellets were finally redispersed in modified Tyrodes-albumin containing 6 mg/l of apyrase. The above procedure yielded platelet-rich Tyrodes (PRT) containing from 0.7 to 1.8 x lo6 cells/~l. For the runs in the absence of red cells, the PRT was diluted to 3.33 x lo5 platelets/@ with normal Tyrodes-albumin containing 2.0 mM Ca++ and 1 mM Mg++ at room temperature. The red cells were washed at room temperature by suspension in modified Tyrodes-albumin solution containing 24 mg/l of apyrase, and centrifuged at = 2000 for 10 min. The packed red cells were again washed with modified Tyrodes-albumin containing apyrase, allowed to stand for 15 min, and then added to a known volume of platelet-rich Tyrodes so that the final suspension contained 3.33 x IO” platelets/@ PRT, 2.0 mM Ca++ and 1 mM Mg++, and had a

540

RBC and ADP-induced

platelet

aggregation

Vol.

32, No. 5

hematocrit of 44% before being mixed with ADP. All suspensions were transferred to 60 ml Luer-Lok syringes prior to the flow experiments. RfwFrozen aliquots of 2 mM adenosine-5’-diphosphate (ADP, Sigma Chemical Co., St. Louis, MO, USA) were thawed immediately before use and diluted in normal Tyrodes-albumin solution. Electron microscope-grade glutaraldehyde (JAB. EM Services, Pointe Claire, Quebec, Canada) was diluted to 0.5% (vol/vol) in Isoton II (Coulter Electronics Inc., Hialeah, FL, USA). ZK 36374 was a generous gift of the Schering Corporation, Berlin, Germany. Frozen aliquots of 10 pM ZK in saline containing 0.13% NaHCOs were thawed immediately before use and added to give the desired concentration.

Flow System and Mixing

Chamber All experiments were done at 23 f 1” C. Whole blood, or washed cell suspensions, and ADP flowing from independent infusion pumps (Harvard Apparatus, Bedford, MA, USA) were rapidly mixed in a stirred chamber of 50 ~1 volume at a fixed flow ratio, cell suspension: ADP = 9:l (Bell et al., 1989a). The suspensions flowed out into 2 to 400 cm lengths of polyethylene tubing (Clay Adams, Parsippany, NJ, USA) of radius & = 0.595 mm, at a volume flow rate, Q = 104 pl s-l, corresponding to a mean linear fluid velocity, = 93.5 mm s-l, a tube Reynolds number -30 (based on a blood viscosity = 4 mPa s), and a volume averaged mean tube shear rate = 324/151~R,,~ = 335 s-l, assuming Poiseuille flow (Bell et al., 1989a). The mean transit times, = X/ (X is the distance of flow along the tube), ranged from 0.2 to 42.8 s, corresponding to tube lengths X = 2.0 to 400 cm. Control runs were carried out by susbstituting Tyrodes-albumin for ADP in the appropriate infusion syringe. The aggregation reaction was permanently arrested by collecting known volumes (-300 ~1) of the effluent into 12x the suspension volume of 0.5% isotonic glutaraldehyde.

Separation of Single Plutelets and Aggregatesfern Red Cells Approximately 5 ml of the glutaraldehyde-hardened effluent suspension were carefully layered onto 30 ml of isotonic Percoll solution (density = 1.097 g ml-‘, Pharmacia, Dorsal, Quebec, Canada) in a 50 ml polycarbonate tube and centrifuged in a horizontal swingout rotor at = 4000 for 20-30 min. The hardened RBCs (density = 1.155 g ml-‘) formed a tight pellet at the bottom of the tube while hardened single platelets and aggregates (density = 1.04 g ml-l) formed a thin layer at the glutaraldehyde-Percoll interface. A total volume of 8 ml Percoll containing the platelet layer was carefully aspirated using a wide-tip polyethylene Pasteur pipette. The RBCs did not entrap significant numbers of single platelets or aggregates as they migrated through the Percoll; at most, 5% of platelets and 1% of aggregates rosedimented with the RBCs. Particle Concentration and Size The number concentration and volume of single platelets and aggregates were measured using an electronic particle and sizing system (Multisizer II, Coulter Electronics Inc., Hialeah, FL, USA) to generate 250 class log-volume histograms over the equivalent sphere volume range l-lo5 pm3, as previously described (Bell et al, 1989a). Particle concentrations were kept I 14 pl-’ to reduce coincidence to < 5%. The number concentration per histogram class is N(xi), the particle volume v(xi) and the volume fraction @(xi) = N(Xi) x V(Xi), where xi is the mark of the ith class. Computer integration of the log-volume histograms yielded the nmliber concentration and volume fraction of particles

Vol. 32, No. 5

RX

and ALIP-induced

platelet

541

agpgatim

between lower and upper volume limits. Individual histograms from multiple donors were averaged as previously described, resulting in a histogram of the mean class volume fraction, <@(xi)> = [(xi) v(xi) / (x,) v(x,)], where (xr) is the mean normalized particle concentration per histogram class, and (x,) and v(x,) are the respective mean normalized number concentration and volume of the class of maximum concentration, m, at = 0 (Bell, 1988; Bell et al., 1989a). The procedure provides an estimate of the changes in particle volume in relation to the mean single platelet volume and standard deviation. As previously shown (Bell et aL, 1990a), there was no visible presence of cells or aggregates on the walls of the polyethylene tubing; tests with aggregated suspensions demonstrated complete recovery of single platelets and aggregates after passage through the tubes, except in the case of the 60% hematocrit suspensionswhere the recovery fell to between 80 and 90%. Results Three female (31 IL 4 yr, s.e.m.) and six male (35.1 + 12 yr) donors were used in the plasma suspension experiments at mean flow tube hematocrits of 9.6, 19.1, 35.8 and 61.5% (Table 1). Two female (27 and 32 yr) and three male (32 f 2 yr, s.e.m.) donors were used in the washed cell suspension experiments at a mean flow tube hematocrit of 39.0. Table 1 gives the mean measured initial single platelet number concentrations in the plasma or PRT at each hematocrit, as obtained from the Coulter count of a sample from the inflow syringe corrected for the 1:lO dilution in the mixing chamber. As in previous work, the singlet particle count encompassesparticles in the volume range l-30 Their concentration in the plasma compartment was reasonably pm3. constant, the mean for 28 experiments being value 3.15 x 105/pl + 0.76 (s.d). To correspond with our previous experiments, the flow tube [ADP] was 0.2 pM in the plasma compartment of the blood (Bell et ah, 1990a) and 0.7 pM in the Tyrodes compartment of the washed cell suspensions (Coldsmith et aL, 1994). Table 1 Experimental Hematocrits and Initial Platelet Concentrations Mean Measured Hematocrit a % k s.e.m.

aHematocrit whole blood;

Mean Platelet Concentrationa cells/p1 plasma or PRT X 10d5 * s.e.m.

0 (cPRP)

3.16 + 0.27 (n = 9)

9.6 t 0.3b 19.1 f 0.4b

3.48 210.59 (n = 5) 3.28 + 0.21 (n = 5)

35.8 t 0.9 61.5 ? 2.0b

3.02 f 0.17 (n = 9) 2.72 f 0.15 (n = 3)

0 (PRT)

2.70 + 0.14 (n = 4)

39.0 f 2.5<

2.29 + 0.31 (n = 4)

and number concentration ‘washed cell suspensions

of cells

in the

flow

tube;

bhematocrit-adjusted

542

RBC and ADP-inducedpkrtelet

Vol. 32, No. 5

aggqation

100 t

-o-

Control,35.8%Hct

-&

9.6% Hct

--a--

61.5% Hct

+

19.1% Hct

*

CPRP

-m-

35.8% Hct

0 0

5

10

15

20

25

30

35

40

Mean Transit Time, s Fig. 1. Effect of hematocrit on the time course of platelet aggregation in whole blood as measured by the disappearance of single cells. Plot of the mean values of the % of single platelets remainin * s.e.m.) as a function of mean transit time at = 33 !fi s- . The control run was carried out at 40% hematocrit FLairrepg Tyrodes-albumin instead of ADP into the mixing

Particle

Volume.

pm’

Fig. 2. Effect of hematocrit on the time course of aggregate growth in whole blood. Three-dimensional representation of histograms of the mean normalized class volume fraction against particle volume (+ s.e.m., dotted line). Singlet peaks (S) decrease in height with time as aggregates (A) appear. Between singlets and aggregates there is a peak due to white cells (WBC), whose height increases with increasing hematocrit.

45

Vol.

32, No.

5

RBC and ADP-induced jdatelef aggregation

543

Effect of Hematocrit on Aggregation in Whole Bbod rice of Sincde Platelets The time course of single cell disappearance is shown in Fig. 1 in a plot of the fraction of the number concentration of single platelets remaining against mean transit time, . As previously found (Bell et aZ., 1989a, 1990a), the curve for the cPRP has a characteristic sigmoid shape reflecting an initial delay in the onset of aggregation, whereas the curves in blood, with the exception of that at 9.6% hematocrit (dashed line), do not. It is evident that the initial rate of disappearance was significantly greater in 19.1 and 35.8% blood than in cPRP. Thus, as documented in Table 2, the mean rate at which single cells aggregated over the first 8.6 s increased from 1.45%/s in cPRP to 1.63, 3.22 and 3.72%/s at 9.6, 19.1 and 35.8% hematocrit, respectively. Even more striking is the marked decrease in the rate of aggregation from 35.8 to 61.5% hematocrit, when the initial rate of disappearance of singlets fell significantly, being only 26% greater than in cPRP. The differences in the extent of aggregation between blood and cPRP decreased as the rate of platelet disappearance increased in cPRP from = 8.6 to 17.1 s. At = 42.8 s the extent of aggregation was now the largest at 19.1% hematocrit, 89.7% of singlets aggregated, as compared to 35.8% hematocrit where 78.2% singlets aggregated, and to cPRP where 62.7% singlets aggregated. However, at 61.5% hematocrit, the extent of aggregation was not significantly different from that in cPRP; it was even markedly lower than that at 9.1% hematocrit. It should be noted that the control run, carried out in whole blood at 35.8% hematocrit with Tyrodesalbumin instead of ADP being infused into the mixing chamber, exhibited little aggregation; only 5% of single platelets had disappeared at = 42.8 s. Table Rate and Extent

Mean Hematocrit

of Platelet

Aggregation

Rate of Aggregation % single platelets /s f s.e.m.

% It s.d. 0

(n = 5)

2 as a Function

of Hematocrit

Extent of Aggregation % single platelets aggregated f s.e.m.

= 8.6 s

<1> = 8.6 s

= 17.1 s

= 42.8 s

1.45 + 0.20

12.5 + 2.6

38.3 f 5.2

62.7 f 3.5

9.6 (n = 5) 19.1 (n = 5)

1.63 * 0.19 3.22 IL 0.37

14.0 + 5.6 27.7 k 6.5

35.0 f 5.3 50.1 f 3.8

74.7 It 2.2 89.7 f 1.6

35.8 (n = 9)

3.72 + 0.43

32.0 f 4.1

55.1 f 5.2

78.2 f 4.7

61.5 (n = 3)

2.13 * 0.25

18.3 f 1.4

34.8 f 1.5

60.2 &I1.5

Formation of Aew The time course of aggregate growth at the four hematocrits is shown in the continuous volume fraction histograms of Figs. 2a-2d. Here, the mean normalized class volume fraction, @(xi), for each set of experiments has been plotted against particle volume over the range from 1 to lo5 pm3 in a three-dimensional diagram. Except at 9.1% hematocrit, the suspensions

544

RBC and ALWidtud

jdatelet

aggregation

Vol. 32, No. 5

exhibited two sharp peaks, the first of which corresponds to single platelets (S) with a modal volume of -7 pm3 and a log-normal distribution with relatively few aggregates present. The second, smaller peak, with a modal volume of -250 pm3, corresponds to white blood cells (WBC) and increases in height with increasing hematocrit. The downslope of the second peak was skewed towards higher particle volume where it led to a third diffuse peak representing aggregates (A) whose particle size and number increased markedly with time. As previously observed in whole blood (Bell et al., 1990a), it is evident that the WBC peaks persisted up to = 42.8 s, and that these cells were not incorporated into platelet aggre tes. Fig. 3 shows the effect of 8”ematocrit on the formation of aggregates of discrete size, in a plot of the volume fraction of aggregates between lower (L) and upper (U) volume limits QL,U (co) normalized to the total volume fraction of particles at = 0 s, @1,105(O) at mean transit times of 17.1 and 42.8 s. It is evident that the volume fraction of the larger aggregates having volumes from 0.02 to 0.19 > lo3 pm3 increases with hematocrit up to 35.8%: at = 17.1 s, and from 0.22 to 1.16 at = 42.8 s. At 19.1 and 35.8% hematocrit, the total normalized class volume fraction of particles from l-lo5 pm3, which should remain at 1.0, now substantially exceeded that limit, most likely due to the entrapment of plasma in the large aggregates, thereby increasing the particle void volume. Such increases in total volume fraction accompanying the formation of large aggregates had previously been observed both in cPRP and in titrated whole blood (Bell et al., 1989a, 1990b). Although the extent of aggregation at 9.1% hematocrit at = 42.8 s was comparable to that at 35.8% hematocrit (Table 2), the size of the aggregates was smaller,

Fig. 3. Effect of hematocrit on aggregate volume distribution at = 17.1 and 42.8 s. The data are for the same experiments as in Fig. 2. Plot of the normalized volume fraction of aggregates between successive upper and lower volume limits from 102-103, 103-lo5 and l-10” (+ s.e.m.) against mean hematocrit. The high values of total volume fraction (l-lo5 pm3) at hematocrits 2 19.1% are likely due to trapping of plasma in the aggregates.

Vol. 32, No. 5

RBC and ADP-induced fdatelet aggregation

545

being limited to volumes I lo4 pm3 (Fig. Zd), yet their volume fraction was significant, increasing from 0.07 to 0.47 as CO increased from 17.1 to 42.8 s (Fig. 3)) presumably due to plasma trapping. In cPRP at = 42.8 s (not shown), the aggregates were located in a broad band extending only to a volume of 2 x IO3 pm3. Effect of Red Cells on Aggregation on Wbhed Cell Suspmsions To test the effects of red cells on aggregation in washed platelet suspensions containing no added exogenous fibrinogen, a series of runs was carried out in which the rate and extent of aggregation at 35.8% hematocrit with 0.7 pm ADP was compared with that in platelet-rich Tyrodes. As illustrated in Fig. 4, red cells have a marked effect on the extent of aggregation in fibrinogen-free suspensions: in washed blood, 50.5 f 9.1% (s.e.m.) and 66.1 f 6.8% of single cells had aggregated at = 17.1 and 42.8 s, respectively, compared with 21.1 + 2.9% and 45.4 ? 5.5% in PRT. The figure also shows that there was an initial delay in aggregation in PRT and at 39.0% hematocrit, which resulted in the sigmoid curves previously observed in fibrinogen-free suspensions. The control runs carried out in the absence of ADP exhibited 5.5 and 9.5% aggregation in PRT and at 39.0% hematocrit, respectively. The increase in the case of blood is probably due to ADP released from the red cells, the degree of aggregation being similar to that previously found in titrated whole blood (Bell et al., 1990a; Goldsmith et aZ., 1995). 120

I

I

I

I

‘2 E

I

I

I

Platelet-rich Tyrodes, no ADP

loo

F .+ C

I

---_

--30.0% hematocrit. no ADP

80

cL” 60 * 22 .-il.? 40 CL

s-”

30.0% hematocrit. 0.7pM ADP

20 I

0 0

5

I

10

I

I

I

I

I

I

IS

20

25

30

3.5

40

Mean Transit Time, s Fig. 4. Effect of red cells on the time course of platelet aggregation in washed cell suspensions as measured by the disappearance of single platelets. Plot, as in Fig. 1, of the mean values of the % of single platelets remaining (* s.e.m.) as a function of mean transit time in PRT and at 39% hematocrit. The control runs were carried out by infusing Tyrodes-albumin instead of ADP into the mixing chamber.

4.5

546

RBC and ALIP-inducfd

d

plutdet

aggregation

Vol. 32, No. 5

Hl,i L/‘Y1

Kg. 5. Comparison of the time course of aggregate growth in PRT (a) with that in washed blood (b). Three-dimensional representation of histograms, as in Fig. 2, of the mean normalized volume fraction against particle volume; S = singlet peak; A = aggregates.

Fig. 6. Corn arison of the time course of aggregate growth in PRT with at s9% hematocrit for the same data as that shown in Fig. 5. Plot of the normalized volume fraction of singlets and aggregates between lower and upper volume limits of increasing size (* s.e.m.) against mean transit time. As in Fig. 3, the increase in total volume fraction (l-lo5 pm3) to values as high as 1.8 at 39% hematocrit is likely due to trapping of Tyrodesalbumin in the aggregates.

Vol. 32, No. 5

RBC and ALIP-induced jdatdei

aggregation

547

The effect of the red cells in promoting the formation of aggregates is even more convincingly shown in the continuous volume fraction histograms of Figs. 5a and 5b. In PRT, the aggregates were contained in a broad band, confined to particles between lo* and lo3 pm3 volume, which only developed between = 17.1 and 42.8 s. By contrast, at 39.0% hematocrit, the suspensions already began to exhibit significant aggregation at = 8.6 s, and by 42.8 s a large peak containing particles between lo* and lo4 pm3 volume had formed, succeeded by a second smaller peak containing the largest measurable Fig. 5b also indicates that there were significant numbers of particles. aggregates whose apparent volumes exceeded the measurable limit of 105 pm’. The time course of the normalized volume fraction of particles between lower and upper limits of increasing size in PRT is compared with that at 39.0% hematocrit in Fig. 6. It is evident that volume fraction ( l-lo5 pm3) in PRT actually decreased by -2O%, an effect that could be explained by the formation of a few large aggregates having volumes > lo5 pm3, outside the measurement range. Discussion Effect of lhnatocn’t

on Platekt

Aggregation

The experiments have shown that platelet aggregation, as measured by the initial rate or the time course of disappearance of single cells, and by the volume fraction and size of the aggregates formed, increases with increasing mean hematocrit from 0 to 35.8%. The effect was not due to an increase in platelet number concentration in the plasma compartment with increasing hematocrit, since the concentration was kept constant in all suspensions (Table 1). Nor does activation of platelets by interaction with white cells appear to play a part (Rinder et al., 1991), since the white cell peaks in the mean class volume fraction histogram remained separate from those of the platelet aggregates and persisted to the end of each run (Fig. 2). As previously observed in whole blood (Bell et al., 1990a), the red ceils exert a much greater effect on the initial rate of aggregation (single platelet disappearance) than on the extent of aggregation at later mean transit times. Thus, while the initial rate at 35.8% hematocrit (measured over the first 8.6 s) is 2.5x greater than in cPRP, at = 42.8 s, only 15% more platelets had aggregated at 35.8% hematocrit than in cPRP (Table 2). Judging by the results previously obtained in whole blood, an even greater difference in the initial rates of disappearance of single platelets would have been observed at < 8.6 s: the rate of single platelet aggregation over the first 1.7 s had been measured to be 7x greater at 36% hematocrit (mean platelet plasma concentrations of 3.07 x 105/p1) than in cPRP (Bell et al., 199Oa). In this connection, it was interesting to find that the order in the extent of single cell aggregation at 19.1 and 35.8% hematocrit at = 4?.8 s (90 and 78% singlets aggregated, respectively) was the reverse of that at
548

R.BC and ALIP-induced

platelet

aggregation

Vol. 32, No. 5

aggregation in whole blood. However, comparison of the estimated two-body collision frequency due to random Brownian motion enhanced by the red cells, with the two-body shear-induced collision frequency, showed that, at CC> = 335 s-t, there is only a 19% increase in collision frequency (Bell et aZ., 1990a) which cannot account for the observed initial seven-fold increase in the An alternate explanation was therefore rate of single platelet aggregation. sought in terms of an increased collision efficiency, defined as the fraction of all two-body collisions/set/unit volume of suspension resulting in capture (formation of a permanent doublet). Such an increased collision efficiency might result from an increased time of interaction of two platelets during collision, due to the presence of the red cells. To test this hypothesis, we used a system consisting of an optically transparent suspension of reconstituted red cell ghosts in plasma (Goldsmith, 1971; Goldsmith and Marlow, 1979) containing 2.5 pm diameter latex spheres as models of platelets flowing through 100 urn-diameter tubes. Shear-induced two-body collisions between latex spheres at known local shear rates were viewed in the interior of the ghost cell suspension, captured on tine film and subsequently analyzed, frame by frame (Goldsmith el al., 1995). The results demonstrated that, at plasma concentrations of latex spheres in 40% ghost cell suspensions comparable to those of platelets in whole blood, the mean lifetimes of colliding doublets increased by about 60% over those predicted by theory (Goldsmith and Mason, 1964) to occur in the absence of ghost cells. Over 55% of observed collisions had lifetimes greater than predicted, and of these, 20% had lifetimes from 2.5 to 5x greater than predicted. In the case of activated platelets, such increased doublet lifetimes would increase the probability of forming a fibrinogen cross-bridge between the GPIIb-IIIa receptors on adjacent interacting cells, the likely mechanism of ADP-induced platelet aggregation in plasma (Frojmovic el al., 1991b; Sung et al., 1993). The time available for stable bond formation during collision is an important factor, since the cells express time-dependent changes in the degree of activation. The data appeared to show that impedance of doublet rotation in a crowded suspension is the likely mechanism by which the doublet lifetime is extended beyond the period predicted by theory for collisions between particles in very dilute suspensions. The jostling of the cells and the latex spheres in shear flow, due to the continual collisions of the ghost cells with each other, clearly seen on the tine films, appeared to prevent some doublets from rotating, and even led to momentary reversal of the direction of rotation. One would expect impedance of rotation to increase with increasing hematocrit. It appears, however, that there is a marked reversal in the initial rate and final extent of aggregation at high hematocrit. At 61.5% hematocrit, the rate of single cell disappearance over the first 8.6 s was 65% less than that at 35.8% hematocrit, and the extent of aggregation, as measured by the decrease in the fraction of singlets at CD = 42.8 s, was not significantly different from that in cPRP. Yet, as is evident in Fig. 3, the volume fraction of large aggregates was still significantly greater at 61.5% hematocrit than in cPRP. The decreased aggregation at high hematocrit is likely due to the particle crowding in such a concentrated suspension. The work in optically transparent suspensions of reconstituted red cell ghosts in plasma containing tracer normal red cells (Goldsmith, 1971; Goldsmith and Marlow, 1979) showed that particle crowding, especially at hematocrits > 40%, leads to continual deformation of the cells into a variety of shapes, the degree of deformation increasing with increasing hematocrit. One would therefore expect the red cell-induced increase in initial rate and extent of platelet aggregation with time eventually to be offset by cell crowding in the suspension. Such crowding likely lowers the

Vol. 34, No. 5

RBCandADP-inducedpldelef~tion

collision efficiency, thus impeding the formation the break-up of aggregates that do form.

549 of aggregates

and promoting

Eflect of Red Cellr on Platelet Aggregation in Washed Cell Sus@nsions We recently showed that, in the absence of exogenous fibrinogen, activation of washed platelets in Tyrodes-albumin with 0.7 pM ADP leads to significant aggregation at shear rates 2 335 s-l (Goldsmith et aL, 1994). The reaction was shown to involve the membrane glycoprotein complex GPIIb-IIIa. Although the platelets had become activated during preparation as shown by the fact that they expressed GPIIb-IIIa receptors and appreciable prebound fibrinogen, it was found that the prebound fibrinogen played no role in the aggregation process. Relatively unactivated cells were then prepared by a single centrifugation process in the presence of ZK 36374 and the platelet button resuspended in Tyrodes-albumin at 5 x lo4 cells/@ to reduce [fibrinogen] to < 12 nM. These platelets were also shown to aggregate with ADP at 2 335 s-l. We therefore postulated that another protein such as von Willebrand factor (vWF), secreted during platelet isolation or in flow at sufficiently high shear rates, may be responsible for the shear rate-dependent aggregation without fibrinogen. Since red cells contain no fibrinogen or vWF, the increase in aggregation, whether measured by the rate or extent of singlet disappearance (Fig. 4), or by the growth of aggregates (Figs. 5 and 6), is further evidence for a physical effect of the red cells in promoting the ADP-induced reaction. Concluding Remarks The fact that red cells play a physical role in the agonist-induced aggregation of platelets is of some importance in the circulation. Thus, it has been shown that, at moderate wall shear rates < 800 s-l, platelet adhesion to subendothelium (Turitto and Weiss, 1982) and to a collagen-covered glass tube (Karino and Goldsmith, 1979) increases with increasing hematocrit up to 40%, after which it levels off. The effect has been ascribed to a physical role of the red cells, through an increase in platelet diffusivity in the flowing blood (Goldsmith, 1971; Turitto et al., 1972), which in turn leads to an increase in the rate of cell-wall collisions. At wall shear rates > 800 s-l, the chemical role of red cells becomes important and both the extent of adhesion and surface thrombus formation increase with increasing shear rate and hematocrit up to 70%. The present experiments, carried out at a moderate shear rate, have demonstrated that the degree of aggregation in bulk flow decreases at high hematocrit. In the circulation, this may counteract the adverse effects of enhanced platelet wall adhesion and thrombus formation which occur at high hematocrit in blood from patients with polycythemia (Wasserman, 1982; Turitto and Goldsmith, 1992). References BELL, D. N. (1988). Physical factors governing the aggregation of human platelets in sheared suspensions. Ph.D. Thesis, McGill University, Montreal, QC, Canada. BELL, D. N., SPAIN, S. and GOLDSMITH, H. L. (1989a). The ADP-induced aggregation of human platelets in flow through tubes: I. Measurement of the concentration and size of single platelets and aggregates. Biophys. J 56, 817-828. BELL, D. N., SPAIN, S. and GOLDSMITH, ADP-induced aggregation of human platelets

H. L. (198913). The in flow through tubes: II.

550

RBC and ALIP-induced

pbdet

Vol. 32, No. 5

aggregation

Effect of shear rate, donor sex, and ADP concentration. 829-843.

Biophys. J. 56,

BELL, D. N., SPAIN, S. and GOLDSMITH, H. L. (1990a). The effect of red blood cells on the ADP-induced aggregation of human platelets in flow through tubes. Thromb. Hatmost. 63, 112-121. BELL, D. N., SPAIN, S. and GOLDSMITH, H. L. (1990b). Extracellular Ca*+ accounts for the sex difference in the aggregation of human platelets in titrated platelet-rich plasma. Thromb. Res. 58, 47-60. DEWITZ, ‘I’. S., MARTIN, R. R., SOLIS, R. T., HELLUMS, J. D., and McINTIRE, L. V. (1978). Microaggregate formation in whole blood exposed to shear stress. Microvasc. Res. 16, 263-271. FOX, S., BURGESS-WILSON, M. E., HEPTINSTALL, S., and MITCHELL, J. R. (1982). Platelet aggregation in whole blood determined using the Ultra-F10 100 platelet counter. Thromb.

Haemost.

48, 327-329,

FROJMOVIC, M. M., WONG, T., and VAN DE VEN, T. (1991a). Dynamic measurements of the platelet membrane glycoprotein IIbIIIa receptor for fibrinogen by flow cytometry. I. Methodology, theory and results for two distinct activators. Biophys. J. 59, 815-827. FROJMOVIC, M. M., O’TOOLE, T. E., PLOW, E. F., LOF-IUS, J. C., and GINSBERG, M. (199lb). Platelet glycoprotein IIbIIIa (OIII$s integrin) confers fibrinogen- and activation-dependent aggregation on heterologous cells. Blood 78, 369-376. FROJMOVIC, M. M., GOLDSMITH, H. L., WONG, T., MCINTOSH, F., and BROWN, E. (1995). Von Willebrand factor binding to GPIb, but not to GPIIb-IIla RGD domain allows ADP-induced aggregation of human platelets in plasma-free suspensions at moderate shear stress. Thromb.

Haemost.

73, 1046.

GOLDSMITH, H. L. (1971). Red cell motions and wall interactions in tube flow. Fed. Proc. 30, 1578-1588. GOLDSMITH, H. L. and MARL.OW, J. C. (1979). Flow behavior of erythrocytes. II. Particle motions in sheared suspensions of ghost cells. J. Colloid

Interface

Sci. 71, 383-407.

GOLDSMITH, II. L. and MASON, S. G. (1964). The flow of suspensions through tubes. III. Collisions of small uniform spheres. Proc. R Sot. (London}

A282,

569-591.

GOLDSMITH, H. L., FROJMOVIC, M. M., BRAOVAC, S., MCINTOSH, F., and WONG, T. (1994). Adenosine diphosphateinduced aggregation of human platelets in flow through tubes: III. Shear and extrinsic fibrinogen-dependent effects. Thromb. Haemost. 71, 78-90.

GOLDSMITII, II. L., BELL, D. N., BRAOVAC, S., STEINBERG, A., and MCINTOSH, F. (1995). Physical and chemical effects of red cells in the shear-induced aggregation of human platelets. Biophys. J. 69. JEN, C. J. and McINTIRE, L. V. shear-induced aggregation in whole 115-124.

(1984). blood.

Characteristics J.

Clin.

Invest.

of 103,

Vol. 32, No. 5

RBC and ADP-induced

plateki

mation

551

KARINO, platelets

T., and GOLDSMITH, H. L. (1979) Adhesion of human in an annular vortex distal to a tubular expansion.

Microvasc.

Rtx

17, 238-245.

MUSTARD, J. F., PERRY, D. F., ARDLIE, N. G., and PACKHAM, M. A. (1972). Preparation of suspensions of washed platelets from humans. Br. J. ffaematol. 22, 193-204. PHILLIPS, D. R. and BAUGHAN, A. K. (1985). Fibrinogen binding to human platelet plasma membranes. J. Biol. Chem. 285, 10240-10246. PHILLIPS, D. R, CHARO, J. F., PARISE, L. V., and FITZGERALD, L. A.. (1988). The platelet membrane glycoprotein IIbIIIa complex. Blood 71, 831-43. RINDER, H. M., BONAN, J. La., RINDER, C. S., AULT, K. A., and SMITH, B.R. (1991). Activated and unactivated platelet adhesion to monocytes and neutrophils. Blood 78, 1760-l 769. SANIABADI, A. R., LOWE, G. D., BARBENEL, J. C., and FORBES, C. D. (1984). A comparison of spontaneous platelet aggregation in whole blood with platelet rich plasma: Additional evidence for the role of ADP. Thromb. Haaost. 51, 115-l 18. SANIABADI, A. R., LOWE, G. D., BARBENEL, J. C., and FORBES, C. D. (1985). Further studies on the role of red blood cells in spontaneous platelet aggregation. Thromb. Res. 38, 225-232. SANIABADI, A. R., LOWE, G. D., BARBENEL, J. C., and FORBES, C. D. (1987). Effect of dipyridamole on spontaneous playelet aggregation in whole blood decreases with the time after venipuncture: evidence for the role of ADP. Thromb. Haemost. 58, 744-748. SANIABADI, A. R., TOMIAK, R. II., L,OWE, G. D., BARBENEL, J. C., and FORBES, C. D. (1989). Dipyridamole inhibits red cell-induced platelet activation. Atherosclerosis 76, 149-154. SHOR, K., DARIUS, II., MATZKY. R., and OHLENDORT, R. (1981). The antiplatelet and cardiovascular actions of a new carbacyclic derivative (ZK 36374)-equipotent to PGI:! in vitro. Arch. Pharm.

316, 252-55.

SMOLUCHOWSKI, Theorie der

M. von. ( 1917). Versuch einer mathematischen Liisungen. Koagulationskinetik kolloider Z. Physik. Chem. 92, 129-188. SUNG, P.-L., FROJMOVIC, M. M., O’TOOLE, T. E., PLOW, E. F., Determination of LOFTUS, J. C., and GINSBERG, M. (1993). adhesion force between single cell pairs generated by activated GpIIbIIIa receptors. Blood 81, 419-423. TURITTO, V. T., BENIS, A. M., and LEONARD, E. F. (1972). Platelet diffusion in flowing blood. fnd. Eng. Chem. Fundam. 11, 216-223. TURITTO, V. T., and WEISS, II. (1982). Platelet and red cell Ann. N. Y. Acad. Sci. 416, involvement in mural thrombogenesis. 363-376.

552

RBC and ADP-indd

Vol. 32, No. 5

platelet aggmgation

(1992). Rheology, TURITTO, V. T., and GOLDSMITH, H. L. transport and thrombosis in the circulation. In: Vascular Medicine - A Textbook of Vascular Biology and Diseases. J. Loscalzo, M. A. Creager and V. J. Dzau Eds., Little Brown and Co., Boston, pp. 193-197. WASSERMAN,

Sem. liematol, Rewired

15 Januaq

L. R.

(1976).

The treatment

of polycythemia

13, 57-78. 1995;

accepted in r-t-vised form 18 July 1995.

Vera.