Mice (Ins.Dd1) with hypoinsulinemic diabetes were created by increased expression of syngeneic major histocompatibility complex (MHC) class I protein in pancreatic beta-cells. The diabetic state was characterized in these mice by high glucose concentrations and islet pathology. To determine whether a neuropathy would develop, motor and sensory conduction velocities (CV) were determined in the sciatic nerves of 2-, 4-, and 7-month-old control and diabetic littermate male mice. Recording bipolar electrodes were placed in the plantar muscles of the hind foot of anesthetized (ketamine/xylazine) mice. Bipolar stimulating electrodes were positioned near the sciatic nerve at the sciatic notch or near the tibial nerve at the ankle. Motor CV from alpha-motor fibers and sensory CV from proprioceptive Aalpha nerves were measured and expressed as meters per second (m/s). Group data are reported as mean +/- SE and compared by analysis of variance. The CVs from nondiabetic mice (controls) were not different across the three ages and averaged 41.3 +/- 1.7 m/s for motor and 38.7 +/- 1.7 m/s for sensory. The motor CVs from diabetic mice at 2 and 4 months were similar to controls. Sensory CVs were unchanged at 2 months but were lower at 4 months (18.9 +/- 2.4 m/s). Both sensory (23.9 +/- 2.1 m/s) and motor (18.9 +/- 1.8 m/s) CVs were significantly reduced at 7 months, which is indicative of a polyneuropathy. NGF has well-known trophic effects on sympathetic and small sensory neurons. To determine whether NGF could influence this neuropathy, 6-month-old control and diabetic mice were divided into the following groups: 1) control + vehicle, 2) diabetic + vehicle, and 3) diabetic + NGF (1 mg/kg, 3x week, s.c.). After 1 month of treatment, motor and sensory CVs were determined. In some mice, the branches of the sciatic nerve were exposed and in situ recordings from the sural nerve were performed to determine compound C-fiber CV, integral, and amplitude. Sensory CV, determined via Hoffmann's reflex (H-reflex) (A-fiber), was decreased in diabetic compared with control animals as expected (P < 0.05), and NGF did not alter this parameter. Continuing diabetes reduced the amplitude (0.9 +/- 0.2 vs. 3.2 +/- 0.7 mV x 10(-2); P < 0.05) and integral (6.9 +/- 1.9 mV/ms vs. 18.8 +/- 4.4 mV/ms; P < 0.05) of the C-fiber response versus control, suggesting fiber loss. NGF treatment normalized C-fiber amplitude (2.9 +/- 0.8 mV x 10(-2)) and integral (21.2 +/- 6.5 mV/ms) in animals with established diabetes, with no effect on blood glucose. The C-fiber CV was similar in all groups, indicating that the animals had some normally conducting small fiber sensory nerves. These studies characterized a motor and sensory polyneuropathy in transgenic diabetic mice and are the first to demonstrate directly that NGF treatment can protect or restore abnormal sensory C-fiber function.