A reduced ability to regenerate peripheral axons may be partly responsible for diabetic neuropathy. The source of the impairment has not been narrowed down to axonal or Schwann cell failure. We used nerve grafts from control or diabetic donor rats transplanted into control or diabetic hosts to pursue this differential diagnosis. An isograft between the left sciatic nerves of inbred Lewis rats was performed 8 weeks after STZ treatment and on age-matched controls. The nerve exchanges were control-control, control-diabetic, diabetic-control and diabetic-diabetic. At postsurgical day 14, nerves were excised and analysed for levels of axonal markers, total and phosphorylated neurofilament, and Schwann cell receptors, ErbB2 and p75(NTR), using immunohistochemistry and Western blotting. The aim was to measure ingress of axonal markers into the graft and judge the appropriateness of Schwann cell phenotype changes. Transfer of nerve from diabetic to control rats resulted in a doubling in neurofilament, both phosphorylated and nonphosphorylated (both P<0.05). ErbB2 was decreased in grafts from diabetic rats (53% of control, P<0.05) and p75(NTR) levels were increased in both types of graft in diabetic rats (to 300-400% of controls, P<0.05). Schwann cells in diabetic nerve grafts showed receptor levels more similar to controls when placed into a normal environment and the converse also appeared to hold. TUNEL staining revealed increased apoptosis in diabetic nerve distal to the graft. The data show that alterations in Schwann cell phenotype in diabetes are reversed by transfer to control rats and develop in normal nerve after transfer to a diabetic host.