Role of non-macrophage cell-derived HMGB1 in oxaliplatin-induced peripheral neuropathy and its prevention by the thrombin/thrombomodulin system in rodents: negative impact of anticoagulants

Background: Macrophage-derived high mobility group box 1 (HMGB1), a damage-associated molecular pattern (DAMP) protein, plays a key role in the development of chemotherapy-induced peripheral neuropathy (CIPN) caused by paclitaxel in rodents. Endothelial thrombomodulin (TM) promotes thrombin-induced degradation of HMGB1, and TM alpha, a recombinant human soluble TM, abolishes peripheral HMGB1-induced allodynia in mice. We thus examined whether HMGB1, particularly derived from macrophages, contributes to oxaliplatin-induced neuropathy in mice and analyzed the anti-neuropathic activity of the TM/thrombin system. Methods: CIPN models were created by the administration of oxaliplatin in mice and rats, and the nociceptive threshold was assessed by von Frey test or paw pressure test. Macrophage-like RAW264.7 cells were stimulated with oxaliplatin in vitro. Proteins were detected and/or quantified by Western blotting, immunostaining, or enzyme-linked immunosorbent assay. Results: Intraperitoneal administration of an anti-HMGB1-neutralizing antibody (AB) at 1 mg/kg prevented the oxaliplatin-induced allodynia in mice and rats. Antagonists of Toll-like receptor (TLR) 4, receptor for advanced glycation end products (RAGE) and CXCR4 among the HMGB1-targeted pro-nociceptive receptors, also mimicked the anti-neuropathic activity of AB in mice. Macrophage accumulation in the sciatic nerve was observed in mice treated with paclitaxel, but not oxaliplatin, and neither macrophage depletion nor inhibitors of macrophage activation affected oxaliplatin-induced allodynia. Oxaliplatin was 10- to 100-fold less potent than paclitaxel in releasing HMGB1 from macrophage-like RAW264.7 cells. Like AB, TMa at 10 mg/kg prevented the oxaliplatin-induced allodynia in mice as well as rats, an effect abolished by argatroban at 10 mg/kg, a thrombin inhibitor. The anti-neuropathic activity of TMa in oxaliplatin-treated mice was suppressed by oral anticoagulants such as warfarin at 1 mg/kg, dabigatran at 75 mg/kg, and rivaroxaban at 10 mg/kg, but not antiplatelet agents such as aspirin at 50 mg/kg and clopidogrel at 10 mg/kg. Repeated administration of the anticoagulants gradually developed neuropathic allodynia and elevated plasma HMGB1 levels in mice treated with a subeffective dose of oxaliplatin. Conclusions: Our data thus suggests a causative role of HMGB1 derived from non-macrophage cells in oxaliplatin-induced peripheral neuropathy and a thrombin-dependent anti-neuropathic activity of exogenous TM alpha and, most probably, endogenous TM.