Release of High Levels of Platelet Factor 4 (PF4) from Platelets Improves Survival after a Lethal Lipopolysaccharide (LPS) Challenge.

Infusion of activated protein C (APC) improves survival in sepsis. PF4 is a CXC chemokine predominantly expressed during megakaryopoiesis and stored in platelet alpha-granules whose biological function(s) are not well understood. We have shown that recombinant PF4 enhances APC generation by thrombin/thrombomodulin complexes both in vitro and in vivo. Would endogenous PF4 released from platelets activated during an inflammatory state similarly affect APC production? We addressed this issue using mice that were either completely devoid of PF4 (mPF4^−/−) or had a 6-fold excess of human PF4 (hPF4⁺). Using a lethal LPS challenge model (O111:B4, 40 mg/kg IP), we examined whether platelets become activated and release PF4 during endotoxemia. Two hours after LPS injection, the platelet count in mice decreased to ~70% of baseline levels (p=0.006). Serum PF4, as measured by ELISA, also dropped to ~60% of baseline from 47±5 kU/ml to 30±5 kU/ml (p=0.002). At the same time, plasma PF4 level was increased by 20%, consistent with LPS resulting in PF4 release. The smaller increase than expected suggests that much of the released PF4 binds immediately to the surface of vascular cells. Consistent with this, we have observed higher accumulation of PF4 in mouse lungs after LPS injection compared to uninjected mice (990±220 and 660±120 U/mg, respectively, p=0.017). APC generation was assessed 10 min after thrombin infusion (80 U/kg. IV) as a measure of endogenous platelet PF4’s effect in an inflammatory/procoagulant state. In mPF4^−/− mice APC levels were 72% of that in wild type (WT) mice (p=0.0006) while in hPF4⁺ mice APC formation increased to 178% (p=0.003). Survival of mice 24 hrs after LPS (25 mg/kg) challenge was then examined. hPF4⁺ mice had a mortality rate of 9% compared to ~40% in both WT and mPF4^−/− (p< 0.001). To examine the role of APC in this improved survival, we performed similar experiments with mice heterozygous for protein C deficiency (PC^+/−). More PC^+/− mice died 16 hrs after injection of 40 mg/kg LPS than WT mice (61% vs. 29% mortality respectively, p=0.005), while mortality for hPF4⁺/PC^+/− mice was significantly lower (14%, p< 0.001 compared to PC^+/− mice), supporting the hypothesis that the protective effect of PF4 is at least in part due to increased APC generation. Next we asked if infusion of platelets with high PF4 is protective in the LPS model. We injected either vehicle buffer or mPF4^−/− or hPF4⁺ platelets (3×10⁸ per 20 g mouse) into WT mice prior to treatment with LPS. Mortality of mice at 24 hrs after LPS injection with mPF4^−/−platelet infusion was not significantly different than mice with buffer infusion (86% vs. 96% respectively, p=0.4), but mortality was significantly lower when hPF4⁺ platelets were infused (58% vs. 96%, p=0.01). Our results suggest that PF4 is released from platelets after an inflammatory stimulus and that this may have a positive physiological role by enhancing APC generation. High endogenous platelet PF4 levels may have a survival advantage after exposure to endotoxins. Infusion of platelets containing high levels of PF4 in sepsis may be a novel therapeutic strategy that warrants further investigation.