This study was performed in order to evaluate the angiogenic effect of implantation of either peripheral blood mononuclear cells (PBMNCs) or bone marrow mononuclear cells (BMMNCs) on diabetic peripheral neuropathy. Streptozotocin (50 mg/kg) was injected intravenously into 6-week-old male Lewis rats. Four weeks after the induction of diabetes, 6 x 10(7) of PBMNCs or 1 x 10(8) of BMMNCs were implanted into the left hindlimb muscle. Motor nerve conduction velocity (MNCV) was monitored before and after implantation. At the end of the experiment, bilateral nerve blood flow (NBF) was measured by laser Doppler and the number of vessels in the sciatic nerves quantified by Factor VIII staining of the sections. Diabetes resulted in an approximately 20% reduction (P < 0.01) in sciatic MNCV. Four weeks after implantation, MNCV was improved by 54% with PBMNCs and by 67% with BMMNCs (both P < 0.01). Moreover, the effects of implantation were almost abolished by administration of VEGF-neutralizing antibody. Sciatic NBF was reduced by approximately 50% by diabetes (P < 0.05). This reduction in perfusion was improved by 74% by implantation of PBMNCs and by 62% by implantation of BMMNCs (P < 0.05 and P < 0.01, respectively). These effects were observed only in the implanted limb. Immunohistochemical staining of sciatic nerve sections for Factor VIII showed no significant increase in the number of vessels in the sciatic nerve following implantation of either PBMNCs or BMMNCs. These data suggest that implantation of hematopoietic mononuclear cell fractions is associated with an improvement in MNCV as a result of arteriogenic effects in the sciatic nerve, and that VEGF may contribute to this effect. This improvement occurred in the absence of angiogenesis. Implantation of these cell fractions may therefore be a potential new therapeutic method for treating diabetic peripheral neuropathy.