CD154 (CD40 ligand)-deficient mice exhibit prolonged bleeding time and decreased shear-induced platelet aggregates


Dr Alan H. Lazarus, Transfusion Medicine Research, St. Michael's Hospital, 30 Bond St., Toronto, Ontario, Canada M5B 1W8.
Tel.: +416 8645599; fax: +416 8643021; e-mail:

Dear Sir,

The co-stimulatory molecule CD154, expressed on activated platelets, may be involved in inflammation and hemostasis [1–5]. Ligation of CD40 on endothelial cells by platelet-expressed CD154 can induce the production of tissue factor [1], which contributes to the coagulation cascade, as well as proinflammatory cytokines, which facilitate thrombus formation [2,3]. Soluble CD154 contains an integrin αIIbβ3-specific binding motif [6] and it was recently demonstrated that the instability of ferric chloride-induced arterial thrombi in CD154 knockout (KO) mice can be ameliorated by infusion of soluble CD154; the authors suggested that this increased thrombus stability may be due to platelet integrin αIIbβ3–CD154 binding [5].

We have previously presented evidence of a bleeding diathesis in CD154 knockout (KO) mice [4]. We sought to determine if the absence of this molecule would affect hemostatic regulation. As shown in Fig. 1(a), examination of tail vein bleeding times (described in [7]) revealed that mice lacking CD154 had significantly prolonged bleeding times relative to strain matched control mice (n = 15 mice per group; P < 0.01), indicating a defect in hemostasis. Furthermore, subsequent bleeding time analysis on all mice at three later time points confirmed this hemostatic diathesis (data not shown). The platelet counts in both strains were not statistically different and within the normal range [8].

Figure 1.

CD154KO mice exhibit defective hemostasis. (a) Tail bleeding time: A portion of mouse tail was removed and the tail was placed into 37[0]C saline. The time taken to cessation of bleeding was recorded. n = 15 per group. (b) Shear-induced platelet aggregation (SIPA): citrated whole blood samples were subjected to the indicated degree of shear rate using a Cone and Plate analyzer and examined by flow cytometry for the percentage aggregates. The horizontal bar represents the mean (±SEM) basal percentage of aggregates. n = 6 mice per point. ●, C57BL/6 mice; □, CD154KO mice. (c) Platelet activation following SIPA: The samples in (b) were stained with an anti-CD62P antibody and analyzed by flow cytometry. The horizontal bar represents the mean (±SEM) basal percentage of CD62P expression. (d) Platelet activation following thrombin stimulation: citrated blood was either unstimulated or incubated with 0.2 U mL−1 thrombin in the presence of 2.5 mmol L−1 Gly-Pro-Arg-Pro. Samples were fixed and analyzed as above. Data are from four separate experiments. All data are represented as mean ±SEM. *P < 0.05, **P < 0.001.

As increased bleeding time may reflect a defect at the level of platelet function, shear-induced platelet aggregation (SIPA) was analyzed following exposure of blood to varying degrees of shear in a cone and plate (CAP) viscometer [9] in the absence of exogenous agonists. As shown in Fig. 1(b), fewer platelet aggregates were seen in CD154KO mice at all shear rates, particularly above 8000 s−1 (n = 6 mice per data point). The ability of platelets to aggregate is crucial for hemostatic regulation [10,11], but whether the lower number of aggregates in the CD154KO mice reflects the inability to form aggregates or increased deaggregation is unknown.

Expression of CD62P is a widely used marker of platelet activation [12,13]. Analysis of CD62P (Fig. 1c) following SIPA by flow cytometry (see [14]) revealed that the increase in CD62P at or above a shear rate of 8000 s−1 was lower in CD154KO mice compared with controls, although the difference was not significant. To determine if CD62P expression would be increased in agonist-dependent in vitro platelet activation, we examined the response of platelets to thrombin. Thrombin-induced expression of CD62P was significantly lower in CD154KO platelets, as compared with controls (Fig. 1d; P < 0.05). While the precise role of CD62P in platelet function is unclear, it has been demonstrated that CD62P deficient mice show increased tail vein bleeding time compared with controls [15]. The reason for this decreased thrombin-induced CD62P expression in platelets from CD154KO mice is unknown and its possible consequences for platelet function remains unclear.

We have found that mice genetically lacking CD154 have a significant bleeding abnormality, decreased in vitro platelet activation, and evidence of reduced platelet aggregates under high shear. A recent report by André and coworkers has demonstrated that CD154KO mice have only insignificantly prolonged bleeding times compared with control C57BL/6 mice [5]. However, we find a significant bleeding diathesis in CD154KO mice. This apparent discrepancy is not due to mouse strain related differences between the CD154KO mice or control mice used here and the ones used in the André study as both were obtained from the same commercial supplier. Although we do not know why André and coworkers only observed a minimal increase in bleeding times in CD154KO mice, they did report that CD154KO mice exhibited significantly delayed vessel occlusion as assessed by intravital microscopy. In addition, aggregation of blood platelets contributes to the arrest of bleeding at sites of vascular injury [10,11] and both our study and the André study measured deficiencies in platelet aggregate formation; we observe a decreased number of aggregates at a very high shear rates while Andréet al. reported abnormally small platelet thrombi in a collagen perfusion chamber at lower shear. Finally, we observed a significant decrease in thrombin-stimulated CD62P expression compared with control mice and a dramatic decrease or absence of CD62P expression has been linked to prolonged bleeding times in mice [15].

The strength of thrombus adhesion to the vascular wall is dependent on platelet interaction with the extracellular matrix (ECM)[16] and therefore the prolonged bleeding time observed in CD154KO mice may result from either the inability of activated platelets to bind to vascular endothelium or of platelets to bind each other. Neither the smaller thrombi resulting from high shear in a collagen-coated perfusion chamber [5] nor our observation of decreased shear stress-induced platelet aggregates in vitro is dependent on platelet/ECM contact. We therefore propose that the hemostatic anomaly observed in CD154KO mice is due to defective platelet/platelet interaction.


We thank Mr David J Allen for his excellent technical assistance and helpful discussion.