Extended Platelet In Vivo Survival Results in Exhausted Platelets

Platelet lifespan is limited to 10 days in humans and 5 days in mice. The intrinsic apoptosis pathway regulates the survival of platelets, where the pro-survival protein Bcl-x_Lrestrains the essential death mediators Bak and Bax (Mason et al., Cell 2007). Hence, platelet lifespan and platelet counts in mice are increased in the absence of Bak and Bak/Bax. While platelet production in mice is normal in the absence of intrinsic apoptosis (Josefsson et al., J Exp Med 2011), the function of these long-lived platelets has not been investigated. In the current study we examined the functional outcome of extended platelet survival.

We found that washed platelets from mice with a constitutive deletion of Bak and a platelet-specific deletion of Bax (Bak^-/-Bax^Pf4Δ/Pf4Δ) were fully resistant to apoptosis induced by the BH3-mimetic ABT-737, as demonstrated by lack of phosphatidylserine exposure (binding of AnnexinV) and unaltered mitochondrial membrane potential. Tail bleeding times into 37°C saline, were extended in the absence of either Bak alone or both Bak and Bax. Furthermore, the electrolytic thrombosis model showed that despite normal time to arterial occlusion, the thrombi formed in Bak^-/- Bax^Pf4Δ/Pf4Δ mice were unstable, a trend also observed in Bak^-/-Bax^fl/fl mice.

The formation of stable thrombi is dependent on the release of secondary agonists, such as ADP and Thromboxane, from activated platelets. To investigate potential defects in platelet signaling pathways in the absence of Bax and/or Bak, we performed in vitro platelet activation assays. Flow cytometric measurements revealed that activation of the PAR4 receptor (by PAR4-AP) or GPVI (by convulxin) led to reduced integrin activation (JON/A) and degranulation (P-selectin exposure) in the absence of Bak and Bak/Bax, while loss of Bax alone had no effect. In contrast, the response to activation with ADP, which does not induce granule release, was similar in platelets from all genotypes. Similarly, platelet aggregation in response to intermediate concentrations of PAR4-AP was severely reduced in the absence of Bak and Bak/Bax, but normal in response to ADP. We next investigated if abnormal degranulation in response to agonists could explain the aggregation defect. Platelet aggregation was performed with PAR4-AP and the platelet supernatants were collected after centrifugation. Dense granule release (ATP and serotonin) and alpha granule release (PF4) were significantly reduced from platelets deficient in Bak, Bak/Bax, but not Bax alone. Untreated resting platelets of all genotypes contained similar amount of granular proteins (ATP, serotonin and PF4). Hence, altered granule content was not the reason behind the abnormality.

We next explored if platelet age was a factor behind the observed functional differences. To be able to directly compare platelet function in Bak/Bax deficient mice and wild-type controls, we synchronized platelet age to ~3 days in all genotypes. Platelets were depleted in vivo by injection of anti-platelet serum (APS). Newly generated platelets were collected at 72 h post injection, a time-point were platelet counts had returned to normal. Remarkably, synchronized platelet age normalized PAR4-AP and convulxin dependent integrin activation (JON/A) and degranulation (P-selectin exposure) in the absence of Bak and Bak/Bax to control levels. Similarly, the platelet aggregation and release defects were rescued.

Lastly we investigated if synchronizing platelet age would revert the hemostatic defect of Bak/Bax mice in vivo. We determined tail bleeding times using mice, which were either untreated or depleted of platelets 72 h prior to the experiment. Strikingly, synchronization of platelet age to 3 days rescued the hemostatic defect in Bak^-/-Bax^Pf4Δ/Pf4Δ mice. We conclude that extended platelet survival leads to platelet exhaustion, with reduced ability to mobilize granular release. Our studies suggest that, in the context of blood bank storage, extending platelet survival times by pharmacologically inhibiting apoptosis may result in a hemostatically compromised product.