Severity of Thrombocytopenia at the Time of Platelet Transfusion Influences Post-Transfusion Platelet Kinetics and the Effects of Platelets on Plasma Cytokines in a Model of Murine Neonatal Sepsis

Thrombocytopenia affects 20-30% of neonates admitted to the Neonatal Intensive Care Unit (NICU) and up to 70% of those born extremely prematurely, with sepsis being one of the most frequent causes. Since preterm neonates also have a high bleeding risk, it has been widely accepted that they should receive platelet transfusions at higher platelet count (PC) thresholds than older children or adults. However, there is no consensus regarding what the appropriate threshold should be, particularly in critically ill infants. Surveys and observational studies have revealed extraordinary world-wide variability in neonatal platelet transfusion thresholds, with North-American neonatologists typically using more liberal thresholds (i.e. PC<100x10⁹/L) than European neonatologists. Here we hypothesized that the severity of thrombocytopenia at the time of platelet transfusion would influence the post-transfusion platelet kinetics and the effects of platelets on the inflammatory response of septic newborn mice. To test these hypotheses, we used a validated model of neonatal sepsis/peritonitis, consisting of injecting cecal slurry (CS) or 15% glycerol (control) IP into post-natal day 10 (P10) mice. The CS dose used in these studies induced severe sepsis with a high mortality. c-MPL KO pups (which lack the TPO receptor, pre-infection PC 92±38 x 10³/µL) and WT B6 pups (pre-infection PC 748±169 x 10³/µL) were used to model severe and mild thrombocytopenia at the time of transfusion, respectively. Two hours after infection, mice from both genotypes were transfused with washed platelets from adult GFP mice (5x10⁷platelets/gram) or with Tyrode's buffer. PCs and GFP% were measured by flow cytometry 2 and 22 hrs post-transfusion. At 2 hours, as expected, transfused septic mice of both genotypes had higher PCs than non-transfused septic littermates (217±118 vs 24±9 x10³/µL for c-MPLKO mice, p<0.0001, and 764±193 vs. 549±155 x10³/µL for WT mice, p<0.01). By 22 hours post-transfusion, transfused septic c-MPL KO mice still had significantly higher PCs than their non-transfused septic littermates (127±71 vs. 21±12 x10³/µL, p<0.01; n=7 per group), but there were no significant differences between transfused and non-transfused septic WT mice (415±139 vs. 310±113 x 10³/µL, p=0.1;n=7 per group). Between 2 and 22 hrs post-transfusion, PCs dropped at a nearly three-fold higher rate in transfused septic WT mice than in transfused septic c-MPL KO mice (by 17±8 vs. 6±5 x10³ platelets/µL/h). In that interval, adult transfused platelets (GFP+) decreased by a larger percentage than neonatal platelets (GFP-) in transfused septic mice of both genotypes (by 64±14% vs. 46±21% in WT mice, and by 53±27% vs. 28±28% in c-MPL KO mice, both p<0.01), although these observations don't account for ongoing neonatal platelet production. Examination of a panel of plasma cytokines 24h after infection revealed significantly increased levels of pro-inflammatory (IL-6, TNF-alpha, and MCP-1) and anti-inflammatory (IL-10) cytokines in non-transfused septic neonates of both genotypes, compared to controls. Interestingly, platelet transfusions significantly reduced the levels of these cytokines in septic WT mice compared to non-transfused littermates, but had no effect on the cytokine levels of septic c-MPLKO mice. Platelets have been recently shown to reduce TNF-alpha production by murine macrophages at high LPS concentrations (Xiang, 2013) and to sequester both pro- and anti-inflammatory cytokines released by monocytes in response to LPS (Carestia, 2019). Our findings suggest that, in neonatal sepsis, this might require PCs to be above a certain threshold. The consequences of these differences on the outcomes of neonatal sepsis are being investigated.