Abstract Aims Increasing evidence indicates that extracellular high mobility group box-1 protein (HMGB1) is involved in the pathogenesis of inflammatory and autoimmune disease. Data from our laboratory demonstrates that HMGB1 contributes to nociceptive behavior in a model of rheumatoid arthritis-induced pain. HMGB1 binds to multiple receptors, including toll like receptor (TLR) 2, TLR4 and receptor for advanced glycation end products (RAGE). When the cysteine in position C106 is in the reduced thiol form and C23 and C45 are engaged in a disulfide bridge (disulfide HMGB1), the molecule functions as a cytokine-inducing TLR4 ligand. In contrast, when these three cysteines are all reduced (all-thiol HMGB1), HMGB1 exclusively potentiates chemotactic activity via CXCR4. It is currently not well understood which receptor and which redox form of HMGB1 that mediates pain hypersensitivity and is therefore the aim of this study. Methods All animal work was carried out in accordance with protocol approved by the local ethics committee for animal experiments in Sweden. Balb/c, C57B/l6 (WT), Tlr2–/–, Tlr4–/– and Rage–/– male mice were used for this study. Disulfide (ds) and all thiol (at) form of HMGB1 were injected intrathecally (1 μg) and mechanical hypersensitivity assessed by von Frey filaments. Lumbar spinal cords were collected after i.t. injection of atHMGB1 and ds HMGB1 and mRNA levels for cytokine and glia markers assessed by quantitative PCR. Results In Balb/c and C57Bl/6 WT mice, i.t injection of dsHMGB1, but not atHMGB1, led to a significant reduction in mechanical thresholds. dsHMGB1 induced mechanical hypersensitivity 6 h after i.t. injection, which lasted for 5 days, compared to i.t. injection of saline. When dsHMGB1 was injected i.t. to Tlr4 deficient mice it did not induce mechanical hypersensitivity. In contrast Tlr2 and Rage deficient mice were still susceptible to dsHMGB1-induced mechanical hypersensitivity. Analysis of mRNA for cytokines and glial cell-associated factors in lumbar spinal cords revealed increased levels of Tnf, Ccl2, Cxcl1, Cxcl2, Gfap and Cd11b in mice injected with dsHMGB1, but not atHMGB1, with exception for Il1β and Cxcr3 that was induced also by atHMGB1. Intrathecal injection of dsHMGB1 to Tlr4 –/– deficient mice, did not increase mRNA levels for Tnf, Il1β, Ccl2, Gfap and Cd11b. Conclusions We found the i.t. injection of the disulfide, but not the all-thiol, form of HMGB1 to induce pronouncedand long-lasting mechanical hypersensitivity, glial reactivity and cytokine induction in a TLR4-, but not TLR2- or RAGE-dependent manner. Thus our data indicates that, the redox state of HMGB1 is key for determining its nociceptive property and receptor usage and thus also the functional consequences of HMGB1 release. Agents interfering with extracellular HMGB1 may be considered in the development of new pain relieving therapeutics.