Platelet Activation by a Humanized Monoclonal CD40 Ligand (CD40L) Antibody.

Abstract Monoclonal antibodies are increasingly used as therapeutics, and some (anti-VEGF, anti-CD40L) have produced unexpected thrombotic side effects. For example, clinical trials with the humanized monoclonal CD40L antibody, hu5c8, in patients with systemic lupus erythematosus (SLE, a chronic inflammatory autoimmune disease) were prematurely terminated due to thromboembolic events. In SLE, elevated soluble CD40L (sCD40L) antigen levels may indicate in vivo platelet activation, putting patients at increased risk for thrombosis. We therefore hypothesized that anti-CD40L immunotherapy may represent an additional prothrombotic hit, leading to clinical manifestation of thromboembolic disease in patients with SLE. In this report, we provide experimental evidence that hu5c8, when combined with homotrimeric recombinant human sCD40L (rhsCD40L) in 1:3 molar ratio to allow formation of higher-order immune complexes (IC), activates platelets in an FcγRIIA-dependent manner. In a washed platelet system, IC formed of rhsCD40L (5 μg/mL) with hu5c8 (15 μg/mL) strongly induced platelet aggregation, while IC of rhsCD40L with hu5c8-Fab or hu5c8-F(ab’)2 did not. Similarly, IC of rhsCD40L with hu5c8-agly, an aglycosylated hu5c8 variant with reduced Fc domain binding to the low affinity platelet IgG receptor, FcγRIIA (CD32), failed to induce aggregation. FcγRIIA dependency was further demonstrated with the monoclonal CD32 antibody, IV.3. These results suggest that hu5c8-mediated platelet activation requires both antigen (CD40L) and the IgG Fc domain. Flow cytometric analysis of P-selectin (CD62P) confirmed that IC of rhsCD40L with hu5c8, but not with Fab, F(ab’)2, or hu5c8-agly, strongly activated platelets via FcγRIIA. When washed platelets were first sensitized with ADP, resulting in surface exposure of endogenous CD40L and sub-maximal aggregation of 5–25%, hu5c8 (10 μg/mL) induced strong platelet aggregation of up to 80–100% in the absence of exogenous rhsCD40L, indicating that physiologic levels of CD40L were sufficient for hu5c8-mediated platelet activation. ADP-sensitized platelets did not respond to equimolar concentrations of Fab, F(ab’)2, or hu5c8-agly. In the serotonin release assay (SRA), which is used for the detection of FcγRIIA-activating heparin-platelet factor 4 (PF4) autoantibodies in patients with heparin-induced thrombocytopenia (HIT), the experimental procedure generates close platelet-platelet contacts and causes release of α-granule proteins such as CD62P, PF4, and CD40L. In the absence of rhsCD40L, hu5c8, but not Fab, F(ab’)2, or hu5c8-agly, induced significant dense granule release, suggesting that exogenous rhsCD40L was not an absolute requirement for hu5c8-mediated platelet activation. In summary, our results show that the humanized monoclonal CD40L antibody, hu5c8, can strongly activate platelets in an Fc domain-dependent manner and that physiologically released CD40L is sufficient for this activity. This mechanism of activation is similar to that of HIT, in which homotetrameric PF4 released from prestimulated platelets provides the molecular basis for clustering of, and signaling through, FcγRIIA, resulting in a vicious cycle of platelet activation. Because thrombosis is a serious complication of HIT, it is reasonable to speculate that hu5c8-mediated platelet activation contributed to the pathophysiology of thromboembolic events in SLE patients receiving anti-CD40L immunotherapy.