Pathophysiologic importance of platelets in neonatal necrotizing enterocolitis

Data from murine models

We developed neonatal murine models using wild type and genetically modified pups to investigate the mechanism(s) of NEC-related thrombocytopenia. ^, The incidence of NEC peaks in premature infants at a postmenstrual age (gestational age at birth + postnatal age at onset of NEC) of approximately 32 weeks, and therefore we postulated that the pathoanatomy of NEC may represent a generic injury response of the intestine during a specific developmental epoch rather than reflecting the effects of specific pathogenic trigger(s). In this context we tested the immunogen trinitrobenzene sulfonic acid (TNBS; 2 doses of 50 mg/kg body weight in 30% w/v ethanol), in 10-day-old C57BL/6J mice by gavage and rectal instillation. ^, TNBS is useful because it can be standardized and titrated for causing injury in different batches of mouse pups and it induces a consistent, timeable, clearly localized, and bacteria-dependent inflammatory injury that resembles NEC in many ways (vide infra). ^,

TNBS has been previously used in adult mice to induce colitis, where it caused subacute inflammation with basal cryptitis in distal parts of the colon. However, there were concerns about the artificial nature of the injury caused by TNBS, which is a nitroaryl oxidizing agent derived from picric acid, and that the possibility of direct chemical damage could not be excluded. However, in P10 mice there were several reassuring factors that the mucosal damage was not due to a corrosive action. The instilled volumes of the TNBS solution were very small (1.5 µL/g body weight), its aqueous solution was not injurious, and no bowel injury occurred in germ-free animals. Here, the intestinal injury seems to occur when its alcoholic solvent created small foci of mucosal damage, and consequent translocation of luminal bacteria led to the recruitment and inflammatory activation of gut macrophages. The histopathologic changes also closely resembled those seen in NEC. ^,

In P10 mice, TNBS induced an acute, temporally predictable necrotizing ileocolitis that resembled human NEC and was distinct from the more subacute inflammation in the distal colon in adult rodents that was mediated via recruitment of Th1-Th17 lymphocytes. In neonates, TNBS caused a rapidly progressive intestinal injury with a strong ileocecal predilection, with prominent necrosis, macrophage-rich infiltrates, and systemic inflammation mediated via signaling networks shared with human NEC. ^, The rapid and highly predictable progression of bowel injury in the neonatal model made it useful for studying NEC complications such as thrombocytopenia. The histopathologic evidence of intestinal injury could be easily seen at 12 hours, and thrombocytopenia ensued soon thereafter.

We examined the platelet volume indices (mean platelet volume [MPV], platelet-to-large cell ratio, and platelet distribution width [PDW]) and the IPF, indicating increased thrombopoiesis. Consistent with human NEC, ^, the affected mice showed larger MPVs and IPFs in the bloodstream. The megakaryocyte number, ploidy, and CD41 expression were also higher, indicating increased megakaryocyte differentiation and platelet production. The results favored increased platelet consumption, not impaired production, as the likely mechanism of thrombocytopenia. Although we did not investigate the mechanism(s) that could increase platelet production during NEC, lipopolysaccharides are known to stimulate megakaryocyte differentiation and platelet production in conjunction with inflammatory cytokines such as interleukin-6. These findings are consistent with the clinical observations of Brown et al., who showed a modest increase in circulating RPs and megakaryocytic precursors in infants with NEC.

• 1.

  We had anticipated platelet activation to be a delayed, secondary consequence of bacterial translocation through the damaged mucosa. However, the mouse model showed it to be a thrombin-mediated, early pathogenetic event beginning as early as 3 hours after initiation of intestinal injury. Platelets can be an important source of inflammatory factors ( Fig. 15.1 ). Although the platelet counts were unchanged at 3 hours, there were morphologic signs of activation ( Fig. 15.2 ) and increased expression of the activated conformation of the integrin α _IIb /β ₃ ( Fig. 15.3 ) (detected in our studies as reactivity to the antibody clone JON/A). There was less ATP release following collagen stimulation (a sign of exhausted stores due to prior release), fewer dense granules, and increased aggregability upon collagen exposure; all these changes preceded obvious mucosal injury or endotoxemia. Unlike dense granules, the α-granule discharge was seen later, at 6 hours or more. Our mechanistic studies showed that platelet activation was triggered by thrombin, which in turn was activated by tissue factor (TF) released from intestinal macrophages. This TF was preformed and stored in microvesicles in neonatal, but not adult, gut macrophages ( Fig. 15.4 ). The changes seen in plasma thrombin activity were also age specific and seen only in neonates, not adults. Compared to adults, neonatal platelets were also more sensitive to thrombin due to higher expression of several downstream signaling mediators and the deficiency of endogenous thrombin antagonists. These characteristics reflected higher expression of Toll-like receptor 4.

  
  
  
  

The higher sensitivity of neonatal platelets to thrombin than those from adults seems to be developmentally regulated. Neonatal platelets expressed several thrombin-signaling proteins at higher levels than in adults, including the platelet glycoprotein-1b beta chain (GP1bβ), vasodilator-stimulated phosphoprotein, guanine nucleotide-binding protein subunit alpha 13 (G13), guanine nucleotide-binding protein (g[q] subunit α), cytosolic phospholipase A-2 (PLA2), phospholipase A2-activating protein (PLA2AP), and the synaptosomal-associated protein 23 (SNAP23). Neonatal platelets also expressed higher levels of the ras homolog gene family member A (RhoA), the Rho GDP-dissociation inhibitor 1 (ARHGDIA), and the Rho GDP-dissociation inhibitor 2 (ARHGDIB). Human infants also carried low plasma levels of antithrombin, which may be accentuated during critical illness and may potentiate the effects of thrombin generated during tissue injury. These findings are of translational importance because the administration of antithrombin perfluorocarbon nanoparticles, which can bind thrombin in nascent blood clots and prevent progressive activation of the coagulation cascades, were protective against intestinal injury.

The platelet counts and the timing of onset of thrombocytopenia were inversely related to the severity of intestinal injury. ^{, ,} Pups with mild injury showed some reduction in platelet counts at 18 hours and a nadir at 24 hours. In moderate-severe injury, platelet counts began dropping at 12 to 15 hours until 18 to 24 hours; survivors showed some recovery of platelet counts between 24 and 48 hours. Interestingly, mice with severe injury showed a transient rise in platelet counts at the 6-hour time point, probably reflecting the release of platelets from storage.

To determine the mechanism of thrombocytopenia, we compared the platelet indices in pups with intestinal injury with those in controls. The MPV is the quotient of the plateletcrit (ratio of platelet volume to the whole blood volume) and the platelet counts (= plateletcrit [%]/platelet count [× 10 ⁹ /L] × 10 ⁵ ). The PDW is the range of platelet volumes at 20% frequency (peak of the frequency histogram = 100%), and the platelet–large cell ratio (P-LCR) is the proportion of large platelets (>12 fL) in the total platelet population. The median MPV increased from 6.6 fL (range 6.1–8.3) in controls and increased to 7.5 (6.6–7.7), 7.2 (6.5–9.6), and 7.6 (6.4–9.9) fL in mice with mild, moderate, and severe intestinal injury. The MPV increased despite the drop in plateletcrits due to the concomitant drop in platelet counts. The PDW and P-LCR also increased in intestinal injury. The increase in platelet volume indices in murine intestinal injury indicated an increased number of larger, younger platelets in the circulation. ^, Pups with intestinal injury showed significantly increased IPFs, rising from a baseline of 7.4% (range 2.9–11.9%) to 9.4% (range 3.1–33.5). Pups with intestinal injury also showed increased number and ploidy of megakaryocytes in the bone marrow.

In the murine model, platelet depletion protected against neonatal intestinal injury. We subjected some mouse pups to antibody-mediated platelet depletion before inducing intestinal injury (intraperitoneal administration of rat monoclonal anti-GP1bα, 0.05 µg/g body weight), and this reduced the severity of bowel injury and improved survival without increasing the severity of hemorrhages into the injured intestine. These mice also showed less systemic inflammation, as evident from lower plasma C-reactive protein, CXCL2, and serum amyloid A, inflammatory markers known to be associated with human NEC and murine NEC-like injury. These findings in the animal model suggested that thrombocytopenia could be a secondary, protective event.

The use of mouse pups on postnatal day 10 (P10) to study NEC-like injury was acceptable because the murine neonatal intestine resembles the human midgestation intestine until P14 to P16. ^, However, the transition from fetal (hepatic) to adult (bone marrow) megakaryopoiesis may occur between P5 and P14. We confirmed our findings by measuring platelet counts, platelet volume indices, and IPF in a small cohort of P3 mice; the findings were similar. We concluded that murine NEC-like injury resembled human NEC at least in terms of decreased platelet counts. There was evidence of increased megakaryocyte differentiation and thrombopoiesis, which suggested that peripheral consumption of platelets was the likely mechanism of thrombocytopenia in these animals, not decreased platelet production.