Role of Platelet-Endothelial Cell Adhesion Molecule (PECAM) in Platelet Adhesion/Aggregation Over Injured but Not Denuded Endothelium In Vivo and Ex Vivo
Background and Purpose We previously demonstrated that a monoclonal antibody (MoAb) with anti-CD31, anti-platelet-endothelial cell adhesion molecule (PECAM)-like properties delayed platelet adhesion/aggregation at a site of minor endothelial injury. To our knowledge, this was the first in vivo demonstration of an effect of anti-CD31. There was no exposure of collagen or basal lamina at the injured site, and the modulation of adhesion/aggregation at such sites has not received much study. The present investigation attempted to replicate the first with the use of a different MoAb, definitely characterized as anti-PECAM. In addition, an ex vivo investigation was performed to see whether the in vivo action of anti-PECAM could have been caused by an effect of the MoAb on the platelets rather than on the endothelium.
Methods A helium-neon laser, in the presence of intravascular Evans blue, was used to injure the endothelium of arterioles on the surface of the mouse brain. Intravital microscopy was used to determine the number of seconds required for the light to initiate the first recognizable platelet aggregate forming at the injured site. Mice injected with vehicle were compared with mice injected with 2 mg/kg anti-PECAM through the tail vein. The injection was 10 minutes before challenge with the laser. Additional studies were performed of aggregation produced in vitro by arachidonate and by ADP added to platelet-rich plasma (PRP) prepared from blood taken from MoAb-treated and vehicle-treated mice.
Results Aggregation latency was significantly prolonged (P<.02) by anti-PECAM (121±59 versus 65±26 seconds in controls; n=10 each). Aggregation ex vivo was not affected.
Conclusions PECAM is an important modulator of platelet adhesion/aggregation at sites of minor endothelial damage in brain arterioles. The data are consistent with the hypothesis that PECAM sites on the endothelium are involved and may be exposed by the injury to promote adhesion/aggregation. Since the endothelial cell layer is intact at these sites, mechanisms such as this offer important alternatives to the more commonly studied pathways of platelet activation, which require exposure of collagen and are not applicable in this model.
Recently, anti-CD31 was reported to inhibit platelet adhesion/aggregation in vivo.1 Adhesion/aggregation was initiated at a site of endothelial injury that was very minor and had not yet resulted in separation or sloughing of endothelial cells; thus, basal lamina was not exposed.2 3 The anti-CD31 had properties of an anti-platelet-endothelial cell adhesion molecule (PECAM)4 but was conservatively characterized only as anti-CD31 or as anti–PECAM-“like” by the provider.1 We now confirm and expand the initial study using an anti-CD31 monoclonal antibody (MoAb) obtained from a different clone and definitely characterized as anti-PECAM.
Materials and Methods
All procedures were approved by the Institutional Animal Care and Use Committee of Virginia Commonwealth University. The preparation used for this study has been exhaustively described.1 5 6 7 Mice were anesthetized with urethane. The pial vessels were exposed, and the surface of the brain was continuously suffused with a mock cerebrospinal fluid (Elliott’s solution) at a constant pH (7.35). The arterioles were observed with a microscope with the use of epi-illumination from a halogen lamp and fiber-optic probe. The mouse was injected through the tail vein with 25 mg/kg Evans blue (0.5% solution in 0.9% NaCl). Thirty minutes later a segment of a preselected arteriole 30 to 50 μm in diameter was exposed to the beam of a 6-mW helium-neon laser (Spectra Physics) that was directed downward through the objective lens of a Leitz metallurgic illuminator. The study used a 20-power infinity-corrected objective lens, and the laser beam was 18 μm wide at the focal plane.
The laser was kept on until platelet adhesion/aggregation (“white body” formation)8 was noted at the exposed site. Propensity for adhesion/aggregation was defined as the time (seconds) required to elicit the first noticeable platelet aggregate. Ten minutes before laser challenge, each mouse was injected through the tail vein with 2 mg/kg of a MoAb (MoAb 390) directed against mouse CD31 (clone 390, catalog No. 09330D, PharMingen). The dose was the same as that found effective in the previous study, in which a different but effective anti-CD31 was used.1 The antibody was azide free, contained less than 0.02 ng/μg protein endotoxin, and was kept at −20°C until use, when it was diluted in buffered normal saline (pH 7.4).
The MoAb was screened for binding to L-cells transfected with murine PECAM-1. Further characterization was performed as follows. Immunoprecipitation showed that MoAb 390 bound to biotinylated surface antigens9 on a 120-kD extract from cells of the murine endothelioma eEnd.2. Immunohistochemical staining of mouse spleen also showed that MoAb 390 reacted with murine endothelium. Indirect immunofluorescent staining of endothelium in isolated tissues from BALB/c mice was also achieved. Finally, flow cytometry showed reaction with eEnd.2 cells, mouse splenocytes, and thymocytes. These observations are in agreement with the published tissue distribution and molecular size of murine CD31.4 10 MoAb 390 reacted with a distinctly different CD31 epitope than the MEC 13.3 clone used in the earlier study.1
MoAb-treated and diluent-treated mice were alternated. The data are expressed as the mean number of seconds required for the laser to induce the first noticeable aggregate. Standard deviations rather than standard errors are given so that the scatter in the data is easily seen; however, the experimental data points are compared with the control by the nonparametric Mann-Whitney test, which avoids assumptions about the distribution of the population.
Additional studies were conducted on platelet-rich plasma (PRP) prepared from platelets harvested from carotid artery blood according to previously described methods.11 Aggregation was tested in a Chronolog aggregometer on aliquots of PRP containing 400 000 platelets per millimeter.3 PRP from two mice were pooled to supply enough sample to provide two aliquots: one tested with 0.5 mmol/L arachidonic acid and sodium salt (Nuchek) and one with 0.5 μmol/L ADP (Sigma). Aggregation was expressed as percent change in optical density.
In arterioles 36±2 μm in diameter from 10 control mice, it took 65±26 seconds (mean±SD) for the laser to produce endothelial damage sufficient to induce a recognizable platelet aggregate at the damaged site. In 10 mice injected 10 minutes earlier with anti-CD31, it took almost twice as long (121±59 seconds; P<.02) to induce aggregation in arterioles 34±4 μm in diameter (mean±SD).
Four pools of PRP from anti-CD31-treated mice were compared with four pools from vehicle-treated mice. The maximal change in optical density produced by either arachidonate or ADP was similar for both treatments: anti-PECAM, 39±16% (mean±SD) with arachidonate, 19±12% with ADP; vehicle, 54±8% with arachidonate (P>.14, Student’s t test), 19±14% with ADP.
These data confirm the single earlier report1 that in this model of endothelial injury, an anti-CD31 can delay platelet adhesion/aggregation at the injured site. The confirmation is important because the MoAb used here differs from that used previously and is more clearly characterized as an anti-PECAM. Thus, the present data reinforce the earlier conclusion1 that a CD31 site was involved in the adhesion and that this site represents PECAM.
In the present study no other MoAb controls were tested as controls for nonspecific effects of MoAbs. These tests were not conducted because in the earlier study1 from this laboratory, which used the same methods, two other MoAbs were tested, directed against glycoprotein IIIa and against intercellular adhesion molecule-1. Neither had any effect.1
In the earlier study it was possible that the effect of anti-CD31 was due to an interaction with either a site on the endothelium or one on the platelet. The present study showed that platelet aggregation induced in vitro by either arachidonate or ADP was not affected by treatment of the mice with anti-PECAM before the platelets were harvested. This supports the hypothesis favored in the original report that the PECAM site involved is on the endothelial cell. However, the in vitro data do not absolutely rule out a PECAM site on the platelet because of the timing of the in vitro observations compared with the in vivo observations. The in vivo observations were made 10 minutes after MoAb injection, and the blood obtained for the in vitro study was obtained 10 minutes after MoAb injection. However, because of the time required to prepare PRP, the test of platelet function was not performed until approximately 1 hour after the MoAb injection. There is no way to avoid this problem if ex vivo studies of PRP are to be conducted. Thus, it is theoretically possible that there was a MoAb interaction with platelets at 10 minutes that was no longer present at the time of in vitro testing.
If the inhibitory effect of the MoAb on platelet adhesion/aggregation is due to an action on endothelial PECAM, we must ask whether these are sites that became exposed only after the injury and that attract and activate platelets, thus contributing to the causes of adhesion/aggregation in the absence of an exposed basal lamina,2 3 or whether the affected sites are always exposed and merely modulate adhesion/aggregation initiated by some other change in endothelial properties. Among the latter is a local decrease in basally released endothelium-derived relaxing factor,6 12 a paracrine substance with potent inhibitory effects on platelet adhesion/aggregation. At the present time, too little is known about PECAM to permit further speculation concerning its role in promoting local platelet adhesion/aggregation. However, the data presented here, combined with what is known about the model of endothelial injury used here,6 12 permit the hypothesis that PECAM sites on injured endothelium are important modifiers of platelet adhesion/aggregation when injury has not been sufficient to denude the endothelium and/or expose basal lamina. In such instances conventional pathways that involve exposed collagen (basal lamina) and collagen-bound von Willebrand’s factor are not applicable. Since the induction of platelet adhesion/aggregation under conditions of minor perturbation of the endothelium is an often overlooked but potentially important cause of microvascular obstruction3 13 14 and/or thrombus initiation, further investigation of the role of PECAM in this phenomenon is warranted.
This study was supported by grant HL-35935 from the National Heart, Lung, and Blood Institute.
Reprint requests to William I. Rosenblum, MD, Division of Neuropathology, Medical College of Virginia, PO Box 980017, Richmond, VA 23298-0017.
Reviews of this manuscript were directed by Guest Editor W. Dalton Dietrich, PhD.
- Received October 10, 1995.
- Revision received December 26, 1995.
- Accepted January 5, 1996.
- Copyright © 1996 by American Heart Association
Povlishock JT, Rosenblum WI. Injury of brain microvessels with a helium neon laser and Evans blue can elicit local platelet aggregation without endothelial denudation. Arch Pathol Lab Med. 1987;3:415-421.
Said S, Rosenblum WI, Povlishock JT, Nelson GH. Correlations between morphological changes in platelet aggregates and underlying endothelial damage in cerebral microcirculation of mice. Stroke. 1993;24:1968-1976.
Rosenblum WI, Nelson GH, Povlishock JT. Laser-induced endothelial damage inhibits endothelium-dependent relaxation in the cerebral microcirculation of the mouse. Circ Res. 1987;60:169-176.
Nishimura H, Rosenblum WI, Nelson GH, Boynton S. Agents that modify EDRF formation after antiplatelet properties of brain arteriolar endothelium in vivo. Am J Physiol. 1991;261:H15-H21.
Rosenblum WI, Zweifach BW. Cerebral microcirculation in the mouse brain. Arch Neurol. 1963;9:414-423.