Platelets Mediate Cytotoxic T Lymphocyte-Induced Liver Damage.

We used hepatitis B virus (HBV) transgenic mice as recipients of cytotoxic T cells (CTLs) specific for the immunodominant Env28–39 epitope (Env28) of HbsAg, and C57BL/6J mice acutely infected with a recombinant replication-deficient adenovirus expressing lacZ (RAd35) to show that platelets accumulate in necroinflammatory foci of the liver during the course of a disease similar to acute hepatitis in humans. Liver disease was monitored by quantifying the activity of serum alanine aminotransferase (sALT), a hepatocellular enzyme that is released into the circulation by necrotic hepatocytes, and by evaluating histological changes at the time of autopsy. Livers from mice sacrificed 2 days after CTL transfer were stained with a mixture of rat monoclonal antibodies (α-PLT) against mouse glycoprotein (GP) Ibα, a receptor exclusively expressed on platelets and megakaryocytes. Livers from mice injected with saline (0.9% NaCl) alone served as controls. Platelets were detectable only inside vessels and hepatic sinusoids in saline-injected controls, while they accumulated alongside apoptotic hepatocytes and inflammatory cells within necroinflammatory foci of the liver in CTL-injected or RAd35-infected animals. The injection of α-PLT caused a >97.5% decrease in the number of circulating platelets (less than 2x10⁴ platelets/μl of blood) within 30 min, and similar low counts were maintained for up to 6 days. Mice that received either saline or an irrelevant antibody (α-Irr) exhibited stable platelet counts (8–10 x 10⁵ platelets/μl of blood). Previous platelet depletion in CTL-injected or RAd35-infected animals reduced the intrahepatic accumulation of virus-specific CTLs and the resulting liver damage, as evidenced by an 80% reduction in sALT activity (from 1600 to 300 U/L) and a reduction the size of necroinflammatory foci (from 6724 μm² to 1245 μm²) as compared to mice that received α-Irr. Infusion of washed mouse platelets expressing human GP Ibα (thus, insensitive to the depleting effect of α-Plt) within hours after CTL transfer or RAd35 infection restored CTL accumulation and liver disease severity, but not when the infused platelets were pretreated with the activation inhibitor, prostaglandin E1. Accordingly, platelet activation was required to promote CTL/platelet interactions under flow conditions in vitro. Fibrin deposition was abundant within hepatic necroinflammatory foci of CTL-injected or RAd35-infected animals, and markedly decreased in animals that were previously made thrombocytopenic. The platelet-dependent intrahepatic deposition of fibrin could be selectively prevented by anticoagulant treatment of the mice, which had no effect on the platelet count, but this did not ameliorate liver injury. Our findings indicate that activated platelets contribute to CTL-induced liver immunopathology by facilitating the accumulation of CTLs at the site of inflammation, independently of procoagulant function.