Aims/hypothesis: Dysregulation of biochemical pathways in response to hyperglycaemia in cells intrinsic to the nervous system (Schwann cells, neurons, vasa nervorum) are thought to underlie diabetic peripheral neuropathy (DPN). TNF-α is a known aetiological factor; Tnf-knockout mice are protected against DPN. We hypothesised that TNF-α produced by a small but specific bone marrow (BM) subpopulation marked by proinsulin production (proinsulin-producing BM-derived cells, PI-BMDCs) is essential for DPN development.
Methods: We produced mice deficient in TNF-α, globally in BM and selectively in PI-BMDCs only, by gene targeting and BM transplantation, and induced diabetes by streptozotocin. Motor and sensory nerve conduction velocities were used to gauge nerve dysfunction. Immunocytochemistry, fluorescence in situ hybridisation (FISH) and PCR analysis of dorsal root ganglia (DRG) were employed to monitor outcome.
Results: We found that loss of TNF-α in BM only protected mice from DPN. We developed a strategy to delete TNF-α specifically in PI-BMDCs, and found that PI-BMDC-specific loss of TNF-α protected against DPN as robustly as loss of total BM TNF-α. Selective loss of PI-BMDC-derived TNF-α downregulated TUNEL-positive DRG neurons. FISH revealed PI-BMDC-neuron fusion cells in the DRG in mice with DPN; fusion cells were undetectable in non-diabetic mice or diabetic mice that had lost TNF-α expression selectively in the PI-BMDC subpopulation.
Conclusions/interpretation: BMDC-specific TNF-α is essential for DPN development; its selective removal from a small PI-BMDC subpopulation protects against DPN. The pathogenicity of PI-BMDC-derived TNF-α may have important therapeutic implications.