Increased Neonatal Platelet Deposition on Subendothelium under Flow Conditions: The Role of Plasma von Willebrand Factor

Abstract

In vitro platelet function of umbilical cord blood and neonatal peripheral vein blood from full-term newborns was compared with that of adults. Citrated whole blood was subjected to shear stress (1300 s^-1) on subendothelial extracellular matrix (ECM)-coated wells in a cone and plate(let) analyzer. Adhered platelets on the ECM were quantitated by image analyzer. Both umbilical cord and neonatal peripheral blood platelets demonstrated more extensive adhesion than adult platelets, and similar aggregate formation on ECM. The ability of neonatal platelets to form aggregates on ECM was confirmed by scanning electron microscopy. Similar activation of neonatal and adult platelets after subjection to shear stress, in the suspension phase, was established by flow cytometry, which showed an increase in fibrinogen binding and a decrease in glycoprotein Ib expression on platelet membrane. The difference in adhesion rates between neonatal and adult platelets was preserved even when the hematocrit level of the neonatal blood was adjusted to that of adults. Reconstitution of neonatal or adult platelet-rich plasma with autologous or heterologous red packed cells yielded no change in adhesion and aggregation. When von Willebrand factor-covered plates were used to prevent deposition of plasma von Willebrand factor on the surface, no difference in platelet adhesion was seen between neonatal and adult blood. In conventional aggregometry assay, the response to ristocetin of washed platelets of either neonatal or adult source was higher on addition of plasma from neonates than from adults. Our data suggest that the extensive neonatal platelet deposition on ECM is mediated by plasma von Willebrand factor, which is known to be more multimerized and, therefore, more active in neonates than in adults. This mechanism may provide balanced primary hemostasis in neonates despite the platelet hyporeactivity to agonists without application of shear stress.

Main

Impaired neonatal platelet aggregation in response to ADP, epinephrine, thrombin, and collagen has been previously described (1,2). Recently, whole-blood flow cytometric analysis has confirmed that neonatal platelets are less reactive than adult platelets to physiologic agonists, such as thrombin, combination of ADP and epinephrine, and thromboxane A₂ analogue (3); defective thromboxane synthesis and serotonin release has also been shown (4–6). Moreover, serotonin and ADP are present in dense granules at concentrations less than 50% of adult values (7), suggesting a storage pool deficiency in newborns. On the other hand, some researches have demonstrated an equal calcium content of the dense tubular system in neonatal platelets but impaired mobilization of this important mediator in response to collagen and thrombin (8).

Electron microscopy studies of umbilical cord platelets have revealed a similar structure to adult platelets with normal number and types of granules (5,9). The receptors responsible for platelet adhesion and aggregation (GPIb, GPIIb-IIIa, and GPIa-IIa) were found on fetal platelet membranes from early gestation (10). However, there are decreased numbers of α-adrenergic receptors and a lower response to epinephrine (11,12).

All these studies were performed with WB or PRP under static or flow shear-free conditions. Physiologic components such as shear flow and subendothelial ECM, on which platelets adhere and aggregate, were not considered. As such, the early models only partially reproduce the complex series of events occurring in vivo during thrombogenesis.

To formulate a model more compatible with the physiologic milieu, we used the cone and plate(let) analyzer, in which WB is subjected to ECM-coated plates under a defined shear rate and the degree of adhesion and aggregation are evaluated (13,14). Our studies showed that neonatal platelets are capable of extensive adhesion and aggregation on ECM and that this capability may be attributed to the high level of plasma vWF multimers in newborn infants.

METHODS

Patients and blood sampling. This study was approved by the local ethics committee. We tested UC blood of 56 healthy full-term neonates born by spontaneous delivery after uncomplicated pregnancies. The UC was double-clamped, and samples were taken from the segment between the clamps. Peripheral blood samples from 11 infants aged 2-5 d and from 42 healthy adult volunteers were also assayed. The volunteers had not taken medications known to affect platelet function for at least 10 d before blood sampling. The blood was drawn into plastic syringes and immediately transferred to plastic tubes containing trisodium citrate (final concentration: 0.013 M). The first 2 mL of blood were discarded.

Preparation of platelet-rich plasma, platelet-poor plasma, and washed platelets. PRP was prepared by centrifugation of the WB at 180 × g for 12 min at 22°C. PPP was prepared by centrifugation of the remaining layer of WB at 1 200 × g for 15 min. In separate experiments, washed platelets were prepared from WB supplemented with prostacyclin (Sigma Chemical Co, St. Louis, MO) to inhibit platelet release and aggregation during centrifugation. WB was drawn into 15% acid-citrate-dextrose solution (65 mM citric acid, 85 mM sodium citrate, 111 mM dextrose, pH 4.5), and PRP was prepared as usual. Prostacyclin was added to PRP at 10^-6 M (final concentration), and the PRP was centrifuged at 1 200 × g for 15 min. The resulting platelet pellet was resuspended in Tyrode's buffer (145 mM NaCl, 5 mM KCl, 0.5 mM Na₂HPO₄, 1 mM MgSO₄, 5 mM glucose, 0.35% BSA, pH 6.5). Platelets were treated again with prostacyclin, pelleted at 1 200 × g for 10 min, and then resuspended in the same buffer without prostacyclin containing 10 mM HEPES, pH 7.4. The washed platelet preparations were left for 1 h at room temperature to recover their sensitivity to aggregating agents. The platelet count was adjusted to 2.5 × 10¹¹/L with buffer by electronic particle counting.

Platelet aggregometry. Platelet aggregation was monitored by a standard technique in which 0.225-mL aliquots of native or reconstituted PRP were incubated at 37°C and stirred at 1000 rpm in a four-channel aggregometer (Helena Laboratories, PACKS-4, Beaumont, TX). Aggregation was induced by ristocetin at different concentrations, and changes in light transmission were recorded for 5 min.

Cone and plate(let) analyzer procedure. Tissue culture four-well plates (Nunc, Roskilde, Denmark) were coated with ECM produced by bovine corneal endothelial cells (15). In separate experiments, polystyrene plates were covered with pure 0.25 U/mL vWF (Alexis Corp, San Diego, CA) dispersed in 0.1% BSA. After 2-h incubation, followed by washings with PBS, the plates were covered with 0.5% BSA for an additional 2 h and washed again. In both types of plates, 0.2 mL of anticoagulated WB was added per well and subjected to shear at a rate of 1300 s^-1 for 2 min using a rotating Teflon cone. The wells were then thoroughly washed with PBS, stained with May-Grunwald stain (Sigma Chemical Co.) and analyzed with an inverted light microscope (Olympus, Tokyo, Japan) connected to an image analysis system (Galai, Migdal Haemek, Israel). Platelet adhesion and aggregation on the ECM were evaluated by examination of 1) percentage of total area covered by platelets, designated as surface coverage, and 2) average size of ECM-bound objects. Separate samples for analysis by SEM were fixed in 2.5% glutaraldehyde and processed as described before (16).

Flow cytometry analysis of platelet activation. WB cytometric analysis was used (17) with some modifications. Within 15 min of drawing, 5 µL of WB was placed in tubes containing 35 µL of modified Tyrode's buffer (134 mM NaCl, 2.9 mM KCl, 12 mM NaHCO₃, 1 mM MgCl₂, 0.34 mM Na₂HPO₄, 5.5 mM dextrose, 0.35% BSA, 10 mM HEPES, pH 7.4) and either 10 µL of FITC-conjugated GPIb MAb AN51 or 10 µL FITC-conjugated rabbit antihuman fibrinogen MAb (both from Dako A/S, Denmark). After 2-min incubation at 22°C, the reaction was stopped by 20-fold dilution with cold Tyrode's EDTA (5 mM) buffer, pH 6.5. The samples were immediately analyzed in an EPICS XL Coulter Flow Cytometer (Coulter Corp., Miami, FL). The flow cytometer was equipped with a 500 mW argon laser operated at 15 mW and a wavelength of 488 nm. After platelet identification by gating of both FITC-positivity and characteristic light scatter, 5000 individual platelets were analyzed. Background binding obtained from parallel samples with FITC-conjugated normal IgG (Dako A/S, Denmark) was subtracted from each tested sample.

Statistical analysis. Unpaired t test was used to compare the mean ± SD values of the results obtained for adult and neonatal platelets. Statistical significance was accepted for p values < 0.05.

RESULTS

Adhesion and aggregation of neonatal platelets on ECM. NPV and UC blood platelet deposition on ECM was studied and compared with that of adult peripheral vein blood platelets (Table 1). Both NPV and UC blood demonstrated a similar extensive platelet deposition on the ECM, as reflected by the higher surface coverage and a slightly lower average size compared with adult platelets. The average size for neonatal platelets (approximately 45 µm²) was much higher than that of a single platelet (<20 µm²), caused by the formation of moderate and large aggregates by neonatal platelets. The similar adhesion and aggregation obtained with NPV and UC blood allowed us to perform all subsequent experiments with UC blood only. The formation of platelet aggregates by UC blood was further confirmed by SEM; both UC and adult platelets demonstrated spreading and extensive aggregation on ECM (Fig. 1).

Table 1 Platelet deposition on ECM under shear stress

Figure 1

Scanning electron micrographs of platelet aggregates formed on ECM. WB from adults (A) and UC (B) was subjected to shear stress (1300 s^-1) for 2 min on ECM coated slides. Samples were washed, fixed with 2.5% glutaraldehyde and processed (magnification × 1500).

Flow cytometric analysis of platelet activation during CPA procedure. Adult and UC platelets in the CPA suspension phase were analyzed by flow cytometry before and at the end of the CPA procedure (Table 2). At baseline, surface expression of GPIb was similar in the adult and UC samples (95-97% of fluorescence positive cells in each case). Some background but equal fibrinogen binding was seen in adult and UC samples (6-8%) compared with values obtained with normal FITC-conjugated IgG. After the blood samples were subjected to shear stress, the expression of GPIb decreased to 61.8% in the adult platelets and to 79.0% in the UC platelets, as expressed in fluorescence intensity compared with baseline data. Both adult and UC platelets demonstrated similar increased fibrinogen binding at the end of the CPA procedure (186.6 and 169.5%, respectively, compared with baseline data).

Table 2 GPIb expression and fibrinogen binding to adult and UC platelets after CPA procedure

Role of neonatal PRP in the increased platelet adhesion to ECM. UC blood has a higher hematocrit than adult blood (43.0 ± 4.7 and 36.1 ± 2.9, respectively). To account for the importance of this parameter in platelet adhesion and aggregation under flow conditions (18–20), we performed the following experiment: PRP and PPP were prepared from UC and adult WB and added in different proportions to red packed cells to obtain different hematocrit levels while maintaining the same platelet count as in the original WB samples. The surface coverage obtained during CPA in both adult and UC samples decreased gradually in accordance with the decrease in hematocrit (Fig. 2). However, the increased surface coverage in UC samples was maintained for most hematocrit levels, suggesting that an intrinsic platelet or plasma (but not red blood cell) property is responsible for this difference. To further identify the contribution of blood constituents to the platelet adhesion, these fractions were prepared from either adult or UC blood, reconstituted either in autologous or heterologous fashion, and tested by CPA in parallel to native WB of adult and UC donors (Table 3). The results demonstrated that, on one hand, mixing a leukocyte-free layer of red packed cells with PRP did not change the measured CPA parameters, compared with intact WB, suggesting no contribution of leukocytes to this process. On the other hand, both surface coverage and average size were not affected by the source of the packed cells used for the reconstitution suggesting the increased surface coverage in UC blood samples derived from an intrinsic characteristic of the PRP. The next experiments were designed to explore the relative role of the neonatal plasma and/or platelets in increased platelet deposition on ECM.

Figure 2

Effect of hematocrit on platelet adhesion to ECM. PRP and PPP from adult and UC blood were prepared by differential centrifugation. The platelet count in PRP was adjusted to that in WB by dilution with PPP. Various hematocrit levels were obtained by addition of various amounts of autologous PRP to whole blood. The first point on the graph represents untreated WB, and the last point represents PRP. The data are mean ± SD of four adult and UC blood samples, each performed in duplicate.

Table 3 Contribution of PRP and packed red blood cells to platelet deposition on ECM

Mediatory role of plasma vWF in the increased adhesion properties of UC platelets. The effect of ristocetin at different concentrations on PRP was studied with routine aggregometry (Fig. 3). Although no difference between adult and UC samples was observed when 1 mg/mL of ristocetin was used as an agonist, a significantly higher response of UC than adult samples, was found at 0.8 and 0.6 mg/mL ristocetin. No aggregation was observed in both adult and UC PRP with 0.4 mg/mL ristocetin. These data suggest an increased sensitivity to ristocetin of GPIb-vWF mediated platelet aggregation in UC PRP compared with adult PRP. It should be noted that no difference was observed between adult and UC plasma samples in standard ristocetin cofactor assay where 1.0 mg/mL ristocetin was used (104.9 ± 5.6 and 113.7 ± 9.1%, respectively).

Figure 3

Ristocetin-induced platelet aggregation. Ristocetin at indicated concentrations was added to PRP, and the aggregation response was measured for 5 min. Maximal increase in light transmission is presented. Data are mean ± SD of four independent experiments.

To further explore the relative contribution of plasma vWF to the increased UC platelet adhesion to ECM, the effect of low ristocetin concentration (0.8 mg/mL) on reconstituted PRP was assayed (Fig. 4). Formalin-fixed platelets, washed adult platelets, or UC platelets were resuspended in either adult or UC PPP and reacted with ristocetin. The results showed a higher response to ristocetin for all PRP samples containing UC plasma, suggesting that the extensive adhesion to ECM of UC blood is caused by the vWF properties in UC plasma rather than an intrinsic platelet property. In addition, similar to other reports (21,22), we found higher molecular weight multimers in neonatal than in adult plasma (data not shown). These findings were further supported when polystyrene plates were precoated with commercial vWF and then reacted with adult or UC whole blood. The equal adhesion rate (surface coverage: 29.2 ± 4.7% in adult and 30.2 ± 3.8% in UC blood) in this model again suggested a role for UC plasma vWF in the increased adhesion of neonatal platelets on ECM.

Figure 4

Ristocetin-induced platelet aggregation in reconstituted PRP. Fixed lyophilized (1), washed adult (2), or washed UC (3) platelets were resuspended in UC or adult (A) platelet-poor plasma. Platelet count in all samples was adjusted to 2 × 10¹¹/L. Ristocetin was used at 0.8 mg/mL. Note that the response to ristocetin was increased in UC plasma compared with adult plasma independent of platelet origin. Representative aggregation curves of four independent experiments are shown.

DISCUSSION

Previous studies have shown that neonatal platelets are hyporeactive to different physiologic agonists. All these experiments were performed under static conditions using either platelet aggregometry (1,2), biochemical assays of the release reaction (4,7,8), or FACS analysis (3). In the aggregation studies, PRP was subjected to minimal shear forces (caused by stirring), which differs from those existing in the circulation. In the present study, we evaluated WB platelet deposition on ECM under a defined shear rate (1300 s^-1). Both UC and peripheral blood samples from full-term newborns were investigated with a new method, the cone and plate(let) analyzer, which is an alternative method to clinically evaluate platelet function under close-to-physiologic conditions (13,14). This system has the advantage of requiring only a small volume of anticoagulated WB (0.15-0.25 mL) and relatively short time for assay (5-10 min) while applying a predetermined shear stress; it can also be performed as a bedside assay to investigate platelet function in its physiologic complexity.

The results showed that under shear conditions, neonatal platelets have an increased adhesion and almost equal aggregation on ECM compared with adult platelets. There was no substantial difference between blood samples drawn from the UC and the peripheral vein of newborns in the first 5 d of life. This observation allowed us to conduct all subsequent experiments using UC blood. Scanning electron microscopy revealed that under shear flow, UC platelets are capable of the same adherence and aggregate formation as adult platelets. The similar decrease in GPIb expression and increase in fibrinogen binding observed by FACS analysis immediately after running the blood on ECM suggest a similar degree of platelet activation in the neonatal and adult samples.

It is known that erythrocytes release ADP and physically enhance platelet transport, leading to the formation of platelet-rich regions near solid surfaces (23). The important role of red blood cells in facilitating platelet deposition on ECM (18–20,24) and the higher hematocrit observed in neonates, were considered to be responsible for the increased adhesion of neonatal platelets to ECM. However, by modulating the hematocrit in reconstituted blood samples, we showed that despite the decreasing hematocrit in neonatal blood compared with adult blood there was still a substantial difference in platelet adhesion. A further decrease in hematocrit value to zero (i.e. by using PRP) was followed by a parallel decrease in platelet adhesion and aggregation in both neonatal and adult samples. These results confirm that red blood cells are a major factor in platelet adhesion, but the difference in hematocrit level between UC and adult blood is not the cause of the increased adhesion of neonatal platelets under shear stress.

Previous investigations have shown that leukocytes can release prostacyclin and platelet-activating factor, which modulate platelet function (25,26). To evaluate the possible role of leukocytes in platelet adhesion to ECM, we reconstituted blood with PRP and the leukocyte-free lower layer of packed red blood cells after blood centrifugation. In this case, platelet adhesion and aggregation was similar to that of intact WB samples in both neonatal and adult blood. Moreover, substitution of autologous packed cells by heterologous packed cells in both types of samples did not change the parameters of platelet adhesion and aggregation. The experiments with reconstituted blood led to the conclusion that the more extensive adhesion of neonatal platelets to ECM is an intrinsic property of the neonatal PRP.

The aggregometry studies demonstrated a higher response to ristocetin when neonatal plasma was tested, regardless of the source of the platelets (Figs. 3 and 4). This suggests a difference in functional properties of neonatal and adult plasma vWF. Numerous investigations have pointed out the crucial role of vWF in platelet deposition on ECM under shear stress (27–30). Binding of vWF to platelets elicits a unique mechanism that requires two platelets receptors: the GPIb/IX and the integrin αIIb/β3 (GPIIb-IIIa). It is also shown that the multimeric structure of vWF has a definite function in platelet adhesion (30,31). Therefore, it is conceivable that the presence of higher vWF multimers in neonatal plasma allows for the increased adhesion of the platelets to subendothelium under shear conditions. When we precoated the surface with pure vWF instead of ECM, no difference was observed in adhesion rate between neonatal and adult blood samples suggesting that under these conditions, plasma vWF did not contribute to the adhesion.

In conclusion, the previously described low response of neonatal platelets to agonists, as tested by routine aggregometry and FACS analysis, may not reflect the actual functional state of the platelets in their physiologic milieu. We suggest that the increased adhesion and sufficient aggregation, as shown here by image analysis and scanning electron microscopy, under shear conditions, may be a more accurate representation of the state of platelet function in full-term neonates.