Neuropathic changes associated with insulin treatment of diabetic rats: electron microscopic and morphometric analysis

Superimposition of hypertension on diabetic peripheral neuropathy affects small unmyelinated sensory nerves in the skin and myelinated tibial and sural nerves in rats with alloxan-induced type 1 diabetes

Diabetic peripheral neuropathy (DPN) is a major complication of diabetes mellitus, and hypertension is considered to be a risk factor for DPN in patients with type 1 diabetes (T1DM). However, the morphological effects of hypertension on DPN are unclear. In this study, we investigated the effect of hypertension on DPN by investigating the changes in unmyelinated and myelinated nerve fibers in hypertensive rats with alloxan (AL)-induced T1DM. Thirteen-week-old WBN/Kob rats with AL-induced diabetes were allocated to receive tap water only (AL group), tap water containing 0.5% saline (0.5AN group), or tap water containing 0.75% saline (0.75AN group) for 15 weeks. Hyperglycemia was maintained for 15 weeks, and the animals were euthanized at 28 weeks. By 23 weeks of age, the systolic blood pressure was significantly higher in the 0.75AN and 0.5AN groups than in the AL group and was unchanged in all groups at 28 weeks. The number of intraepidermal sensory unmyelinated nerve fibers was significantly smaller in the 0.75AN and 0.5AN groups than in the AL group. The axonal size in the myelinated tibial and sural nerve fibers was significantly smaller in the 0.75AN group than in the AL group. Furthermore, luminal narrowing and endothelial hypertrophy were observed in the endoneurial tibial nerve vessels in the 0.75AN group. These findings suggest that superimposing hypertension on hyperglycemia may accelerate a reduction in the number of small unmyelinated sensory nerve fibers in the skin and induce mild axonal atrophy in myelinated tibial and sural nerve fibers in rats with AL-induced T1DM.

Extended Duration of Hyperglycemia Result in Human-Like Corneal Nerve Lesions in Mice With Alloxan- and Streptozotocin-Induced Type 1 Diabetes

Purpose

Previous experimental studies assessing corneal nerves as a measure of the severity of diabetic peripheral neuropathy have yielded discordant results; this may have been due to the effect of the short duration of the induced diabetes. We investigated whether increases in the duration of hyperglycemia result in the development of corneal lesions in a mouse model of alloxan (AL)- or streptozotocin (STZ)-induced type 1 diabetes. We further determined whether corneal nerve fiber density, intraepidermal nerve fiber density (IENFD), and sural nerve morphology can be used as morphologic markers of diabetic peripheral neuropathy in rodent models.

Methods

A total of 30 female ICR mice were divided into three groups: those with STZ-induced (STZ group) and AL-induced (AL group) diabetes, and a control group. Hyperglycemia was maintained in diabetic mice for 35 weeks. Animals were euthanized at 41 weeks of age.

Results

Subbasal nerve plexus density (SBNPD) and terminal epithelial nerve density (TEND) in the cornea, as well as IENFD, were significantly lower, and mean sural nerve axon sizes were smaller in mice in the STZ and AL groups than in the control group. There were significant correlations between IENFD and SBNPD, and between IENFD and TEND.

Conclusions

These results indicate that the TEND and SBNTD of the cornea may be useful morphologic markers for diabetic peripheral neuropathy.

Under-recognised paradox of neuropathy from rapid glycaemic control

Insulin induced neuropathy has been reported previously in people with diabetes treated with insulin, and subsequently reported in patients with insulinomas. However, neuropathy caused by rapid glycaemic control in patients with poorly controlled diabetes with chronic hyperglycaemia is not a widely recognised entity among clinicians worldwide. It is expected that this phenomenon of paradoxical complication of neuropathy in the face of drastic decreases in glycosylated haemoglobin concentrations will assume greater importance with clinicians achieving glycaemic targets at a faster pace than before.

Hypoglycaemia causes degeneration of large myelinated nerve fibres in the vagus nerve of insulin-treated diabetic BB/Wor rats

The aim of this study was to find out whether dysglycaemia causes neuropathy in the vagus nerve of insulin-treated diabetic BB/Wor rats. Specimens were collected from the left vagus nerve proximal and distal to the level of recurrent laryngeal branch and from the recurrent branch itself in control rats and diabetic BB/Wor rats subjected to hyper- or hypoglycaemia. Myelinated and unmyelinated axons were counted and myelinated axon diameters were measured by electron microscopy. In controls, the vagus nerve proximal to the recurrent branch exhibited three regions in terms of fibre composition: part A was mainly composed of large myelinated axons, part B contained small myelinated and unmyelinated axons, and part C contained mainly unmyelinated axons. The distal level resembled part C at the proximal level and the recurrent branch resembled parts A and B. In hyperglycaemic rats, a normal picture was found at the proximal and distal levels of the vagus nerve and in the recurrent branch. In hypoglycaemic rats, signs of past and ongoing degeneration and regeneration of large myelinated axons were found at the proximal and distal levels and in the recurrent branch. We conclude that hypoglycaemia elicits degenerative alterations in large myelinated axons in the vagus and recurrent laryngeal nerves in diabetic BB/Wor rats. The absence of signs of neuropathy in unmyelinated and small myelinated axons suggests that the sensory and autonomic components of the nerve are less affected. In contrast, the hyperglycaemic rats examined here did not show obvious degenerative alterations.

Hypoglycemic sensorimotor polyneuropathy associated with insulinoma

Hypoglycemia-induced peripheral neuropathy due to insulinoma is unusual and, as far as we know, has previously been reported in only 34 patients. In this case report, we describe the clinical features, electrophysiological features, and pathological findings in a 37-year-old patient with polyneuropathy from repeated hypoglycemic episodes over a 9-year period that related to an insulinoma. The literature is discussed. The reported case is of special interest because the peripheral neuropathy led to the correct diagnosis.

Diabetic neuropathy--a continuing enigma

In this article we will review the clinical signs and symptoms of diabetic somatic polyneuropathy (DPN), its prevalence and clinical management. Staging and classification of DPN will be exemplified by various staging paradigms of varied sophistication. The results of therapeutic clinical trials will be summarized. The pathogenesis of diabetic neuropathy reviews an extremely complex issue that is still not fully understood. Various recent advances in the understanding of the disease will be discussed, particularly with respect to the differences between neuropathy in the two major types of diabetes. The neuropathology and natural history of diabetic neuropathy will be discussed pointing out the heterogeneities of the disease. Finally, the various prospective therapeutic avenues will be dealt with and discussed.

Peripheral neuropathy in B6C3F1 mice and SD rats induced by chronic intermittent insulin hypoglycemia

The effects of sustained insulin-induced hypoglycemia on peripheral nerves were examined in 9-10-week old female B6C3F1 mice and 9-10-week old female SD rats. Insulin was administered via osmotic minipumps at a dose of 81 IU/kg/day for 2 consecutive weeks. Mice and rats treated with this high insulin dose showed marked hypoglycemia, resulting in half the normal blood glucose level, hypothermia, impaired motor nerve conduction velocity, and an increased incidence of peripheral nerve lesions, consisting of nerve fiber degeneration characterized by irregular myelin sheaths and axonal atrophy.

Natural course of diabetic peripheral neuropathy in spontaneous-onset diabetic Chinese hamsters

We investigated metabolic and pathological changes in the peripheral nerve of the spontaneous-onset diabetic Chinese hamster. Electrophysiological examination revealed that the motor nerve conduction velocity was significantly decreased at 10 months and afterwards, however, the F-wave latency was significantly increased at 5 months and afterwards. Concerning sciatic nerve contents of sorbitol, myo- and scyllo-inositol, the content of sorbitol was not significantly increased at 5 months, but, myo- and scyllo-inositol were significantly decreased at 5 months and thereafter. At 10 and 15 months, however, sciatic nerve content of sorbitol was significantly increased. On morphological examination, loss of large myelinated fiber and reciprocal increase in degenerative fiber were also seen in sciatic nerve, but not in tibial nerve, at 5 months. At 15 months, these morphological changes were also found in the tibial as well as the sciatic nerve. Thus, we may hypothesize that F-wave latency is useful in the detection of initial diabetic neuropathy, and that the initial pathological changes in diabetic neuropathy of diabetic Chinese hamsters are predominantly found in the proximal site of peripheral nerves.

Nerve conduction changes and fine structural alterations of extra- and intrafusal muscle and nerve fibers in streptozotocin diabetic rats

Streptozotocin-induced diabetes mellitus is known to cause a reduction of both conduction velocity and axon caliber in sciatic nerves and also a decrease in muscle fiber size. The present study investigates whether the distal parts of the peripheral nervous system, including extra- and intrafusal muscle fibers, are more severely affected than the proximal segments in the diabetic state. Proximal and distal sensory nerve conduction velocities were monitored during a period of 3 months in rats rendered diabetic by injection of streptozotocin. Segments of the sciatic and ventral coccygeal nerves, and of the biceps femoris and lumbrical muscles, were studied by light and electron microscopy, including morphometric analysis. In contrast to previous studies, daily suboptimal insulin injections were given to prevent acute metabolic complications. Sensory conduction velocity in the ventral coccygeal nerve was significantly (P < 0.05) decreased in the diabetic rats compared to controls. Proximal and distal nerve segments were equally affected. Mean cross-sectional axon area of the sciatic nerve was moderately, but significantly (P < 0.05), smaller in insulin-treated diabetic rats than in controls. In both the sciatic nerve and the terminal, intrafusal nerve segments, occasional axons showed moderate dystrophic changes. Fibers of the intrafusal nerve segments appeared to be equally affected compared to the fibers in the sciatic nerve, although no quantitative comparison was made. The increase of small caliber skeletal muscle fibers in experimental streptozotocin-induced diabetes was confirmed. These findings indicate that proximal and distal segments of peripheral nerves are affected equally in the early stages of experimental diabetic neuropathy.

Ultrastructural, histochemical, and morphometric analysis of skeletal muscle in a murine model of type I diabetes

BACKGROUND

Since peripheral nerves are damaged in diabetes mellitus, morphological changes occur within the diabetic muscle in response to the diabetic neuropathy. The aim of this study was to examine the extensor digitorum longus (EDL) from a 42-day streptozotocin-induced diabetic Swiss Webster mouse (STZ) and compare the muscle morphology and histochemistry to age-matched, nondiabetic controls.

METHODS

The EDL was evaluated using electron microscopy in order to investigate the morphological integrity of the myofibers and neuromuscular junctions. Histochemical analysis was completed using the myofibrillar CA(++)-ATPase reaction of Doriguzzi et al. (1983. Histochemistry, 79:289-294) for use in computer-assisted morphometric analysis of fiber size using Bioquant System 4 software.

RESULTS

Ultrastructural analysis of the diabetic EDL (N = 5, 225 myofibers/animal) showed a significant number of abnormal myofibers, exhibiting various degrees of degeneration, signs of denervation, and necrosis. The STZ myofibers exhibited excessive lipid accumulations and abnormal mitochondrial arrangements. Histochemical analysis of the STZ EDL revealed a significant shift in fiber type profile (53.6% type 2A and 46.4% type 2B- STZ myofibers; 47.5% type 2A, 52.5% type 2B nondiabetic controls). Morphometric analysis of myofiber size by fiber type (200 myofibers/muscle/fiber type) indicated a significant decrease in myofiber size for both type 2A and type 2B fibers in the STZ diabetic mouse.

CONCLUSION

The degeneration and necrosis of myofibers concomitant with the sever atrophy of both the type 2A and 2B myofibers in the STZ muscle could account for the functional alterations seen in diabetic muscle.

The effect of pancreatic islet transplantation on experimental diabetic neuropathy

Quantitative light and electronmicroscopical morphometric techniques were used to determine the effect of pancreatic islet transplantation on experimental diabetic neuropathy. Groups of STZ-diabetic rats were given islet transplants at 3 weeks after diabetes onset (prevention) and at 6 months after diabetes onset (reversal). Comparisons were made with onset controls, age-matched non-diabetic controls and untreated diabetic controls 6 months later (n = 8 for all groups). Euglycaemia and normal levels of glycosylated haemoglobin were achieved in both groups of diabetics after islet transplantation. Loss of body weight in diabetic animals was prevented by early islet transplantation, but was only partially reversed following delayed islet transplantation. Normal growth of myelinated fibres and axons during development was retarded in untreated diabetics, but was normal in rats given islet transplants soon after the onset of diabetes (cross-sectional perimeter and area). Diabetics transplanted with islets after a delay had myelinated fibres and axons with diminished calibre. Teased fibre preparations of nerves from diabetics which had received islet transplants showed no excess of abnormalities. This study has shown that the development of certain structural abnormalities of peripheral nerve fibres is prevented in diabetic rats which receive transplants of islets of Langerhans soon after the onset of diabetes. However, once established abnormal fibre morphology can not be completely ameliorated merely by achieving and sustaining euglycaemia through delayed islet transplantation.

Diabetes mellitus-associated decrease in nerve growth factor levels is reversed by allogeneic pancreatic islet transplantation

After an untreated 5-month duration of streptozotocin (STZ)-induced diabetes mellitus (DM), nerve growth factor (NGF) levels in BDE rats were decreased to 45-65% of control in the sympathetically innervated target organs iris and submandibular gland, in the superior cervical ganglion (containing NGF-dependent sympathetic perikarya projecting to the cranial targets), and in the NGF-transporting sciatic nerve. Successful allogeneic pancreatic islet transplantation (providing a physiological glucose homeostasis without immunosuppression) after 3-4 weeks of DM reversed the DM-related decrease in NGF levels 4 months after transplantation as compared with untreated diabetic rats. By contrast, NGF levels in the treated vas deferens (innervated by short postganglionic sympathetic neurons) remained increased as in the untreated diabetic rats (175% of control). Thus, DM-associated changes in endogenous NGF levels seem to be reversible by institution of metabolic control, at least at an early stage of DM when NGF-responsive neurons have not been deprived of NGF for a long time.

Duration and severity of hypoglycemia needed to induce neuropathy

The duration and severity of hypoglycemia needed to induce neuropathy is not known. To test these variables, the percentage of teased fibers of peroneal and tibial nerves showing graded pathologic abnormalities was estimated in groups of rats that had been made hypoglycemic for various times and severities one week earlier. The techniques used maintained core temperature, pO2, pCO2, and hematocrit within physiologic limits. A control group was anesthetized and mechanically ventilated but insulin was not given. A second control group underwent no experimental manipulation. Life could not be sustained with hypoglycemia below 1 mmol/l. In 23 rats that were hypoglycemic (1.4 +/- 0.2 mmol/l, mean +/- S.E.M.) for various times less than 11 h, the frequency of axonal degeneration of teased myelinated fibers (0%-1%) was not different than in controls. In 9 young rats that were hypoglycemic (1.4 +/- 0.0 mmol/l) for various times of 12 or more hours, the frequency of fiber degeneration was significantly higher than in controls (P less than 0.01) and increased to as high as 26%. By contrast, in 5 older rats that were hypoglycemic (1.5 +/- 0.1 mmol/l) for various times of 12 or more hours, the frequency of degeneration was not different from that of controls. Both duration and severity of hypoglycemia are risk factors for fiber degeneration. The peripheral nerves are more vulnerable to prolonged severe hypoglycemia in younger rats than in older rats.

Skeletal muscle in the diabetic mouse: histochemical and morphometric analysis

Despite the extensive literature concerning the neuropathy associated with diabetes, only limited information describes changes in the associated muscle. The objective of this study was to evaluate the histochemical and morphometric characteristics of diabetic muscle in the C57BL/KsJ db-m strain of mouse. The histochemical analysis of myofiber type for the diabetic mouse revealed that the extensor digitorum longus muscle consisted of 53.1% type 2a, 46.0% type 2b, and 0.9% type 1 myofibers, a significant shift from the percentages found in the nondiabetic litter mates (44.4% type 2a, 55.6% type 2b, no type 1). Computer-assisted morphometric analysis of myofiber size by fiber type indicated a significant difference in myofiber size for the type 2b fibers in muscles from diabetic mice. Similarly, there was a shift in the fiber size distribution to include a greater number of small type 2b myofibers when compared to controls. Skeletal muscle from diabetic mice exhibited a significant change in the percentage of fiber types, with an increase in the number of type 2a fibers, a fiber type grouping that implies possible denervation and reinnervation, and a decrease in myofiber size. These findings may explain why some diabetic patients complain of muscle weakness.

[Hypoglycemic polyneuropathy: report of a case with insulinoma]

A case of a young man who presented symptoms and clinical signs of polyneuropathy that occurred in connection with recurrent hypoglycemic episodes is reported. The hypoglycemia was probably caused by a pancreatic islet tumor. There were symmetric weakness and wasting of hands and feet, absent tendon reflexes and 'glove and stocking' loss of sensation. Electromyographic studies showed denervation potentials with slight reduction of nerve conduction velocities. Sural nerve biopsy studied by optic and electronic microscopy showed axonal degeneration without signs of demyelination or remyelination. There are only 30 similar cases reported in the literature. According to experimental findings, the authors believe that glucopenia is the mechanism responsible for the development of the neuropathy, and that at present time there is no evidence for a direct insulin effect.

Diabetic and hypoglycemic neuropathy--a comparison in the BB rat

Functional and structural neuropathy was examined in hyperglycemic (diabetic) BB rats maintained on small maintenance doses of insulin, hyperglycemic BB rats receiving no insulin, and BB rats in whom hypoglycemia was induced by the administration of excessive insulin doses. The data were compared with those of non-diabetic age- and sex-matched BB rats. Functional deficits and structural abnormalities were comparable in diabetic rats with and without insulin supplementation, suggesting that the generally necessary insulin dosing in this model does not per se account for the neuropathy. Hypoglycemic neuropathy was characterized by slowing of nerve conduction velocity, marked loss of anterior horn motoneurons and Wallerian degeneration, as well as loss of large myelinated fibers, suggesting a neuropathy involving predominantly motoneurons. Diabetic neuropathy was not associated with nerve cell loss but showed marked axonal atrophy involving predominantly sensory fibers. Thus, diabetic and hypoglycemic neuropathies are two distinguishable entities under strict experimental conditions, but may overlap in human diabetic subjects in whom tight insulin control is desirable.

Slow axonal transport of structural polypeptides in rat, early changes in streptozocin diabetes, and effect of insulin treatment

The synthesis and transport of slowly transported polypeptides in sciatic nerves of rats was investigated by [35S]methionine pulse labeling and gel electrophoresis in control, diabetic, and insulin-treated diabetic rats. To detect very early changes diabetes was induced by streptozocin only 5 days prior to the labeling of the dorsal root ganglion cells. Fourteen days were allowed for axonal transport. In this experimental system, the neurofilament triplet is transported at an apparent velocity of 1.1 +/- 0.1 mm/day (mean +/- SD). The actin-related complex, including actin and two polypeptides of 87 kilodaltons and 37 kilodaltons, was transported at a velocity of 2.6 +/- 0.2 mm/day. For alpha- and beta-tubulin we found an apparent transport velocity of 2.2 +/- 0.1 mm/day, placing it between actin and the neurofilament triplet. The diabetic rats had a selective 32% decrease in the amount of the heaviest neurofilament subunit: 0.47 +/- 0.19% of trichloroacetic acid-insoluble radioactivity versus 0.69 +/- 0.17% in controls; 2p less than 0.05. This decrease was associated with a proximal accumulation of the two lighter neurofilament subunits. Insulin treatment of a diabetic group failed to normalize the changes of axonal transport and additional changes suggesting a hypoglycemic injury was observed.

Hypoglycemic neuropathy in experimental diabetes

Morphological and electrophysiological observations were made over 4 weeks on 5 groups of 8-week-old male Sprague-Dawley rats. These were comprised of controls, untreated diabetics, and diabetic animals in which sustained hypoglycemia, moderate hypoglycemia, or normoglycemia was induced by continuous subcutaneous insulin infusion (CSII) therapy. Teased fiber studies showed a marked increase in the number of myelinated fibers undergoing axonal degeneration and regeneration in the tibial nerve of severe hypoglycemic and also in moderate hypoglycemic animals but not in controls, untreated diabetic and normoglycemic groups. There was also a significant correlation between episodes of hypoglycemia (less than or equal to 2.0 mmol/l) and the prevalence of axonal degeneration and regeneration in CSII-treated diabetics. Motor nerve conduction velocity was significantly reduced in the moderate and severe hypoglycemic groups and also in untreated diabetic animals when compared with controls. However, it was significantly improved in the normoglycemic group over the untreated diabetic and severe hypoglycemic groups. In conclusion, this study has demonstrated that severe or even mild hypoglycemia produced a detrimental effect on peripheral nerve structure and function in experimental diabetes. Therefore, it may be desirable to avoid even asymptomatic hypoglycemia in the management of diabetes.

Cytoarchitecture of muscle in a genetic model of murine diabetes

Although diabetic neuropathy is well documented, diabetic myopathy is not, except for descriptions of diabetic patients with muscular weakness thought to be due to metabolic changes in the muscle. Muscle and nerve are dependent on each other for normal structure and function; since the peripheral nerve is damaged in diabetes, one would expect concomitant changes in the muscle. This study examines the cytoarchitecture of diabetic muscle. The extensor digitorum longus (EDL) muscles from 165-day-old C57BL/KsJ dbm mice were examined using electron microscopy. Morphological analysis of the diabetic EDL revealed that a significant number of the myofibers, examined within the midbelly region of the muscle, exhibited various degrees of degeneration, signs of denervation, and abnormal lipid stores. Both myoneural junctions and muscle spindles showed significant signs of degeneration, denervation, and abnormal structure. Thus the morphologic changes seen could account for the physiologic changes seen in diabetic muscle.

The effect of continuous subcutaneous insulin infusion therapy on morphological and biochemical abnormalities of peripheral nerves in experimental diabetes

Diabetes mellitus was induced in rats by the administration of streptozotocin and observations have been made over a period of 2 months in 3 groups of animals: controls, untreated diabetics and diabetics treated with continuous subcutaneous insulin infusion (CSII) therapy, using a 14-day Alzet osmotic minipump. Optimal control of day-to-day and 24-h blood glucose levels was achieved in diabetic animals treated with CSII. Body weight and skeletal growth, assessed by measurements of tibial length, were decreased in untreated diabetic rats and were normalized by insulin treatment. The concentrations of glucose, sorbitol and fructose in the nerves of diabetic animals were significantly increased and that of myoinositol significantly decreased; CSII therapy normalized these levels to those of age-matched controls. External myelinated fibre diameter in the tibial nerve was significantly less in untreated diabetic rats as compared with age-matched controls. In the insulin-treated group, fibre diameter significantly increased as compared with untreated diabetics and there was no significant difference between insulin-treated and control animals. Teased fibre preparations from the tibial nerve revealed very few abnormal fibres in all the three groups and no significant difference was detected between any of the groups. Continuous subcutaneous insulin infusion therapy, therefore, corrected biochemical abnormalities and also normalized myelinated fibre diameter in the peripheral nerves of experimental diabetic animals. The paradoxical excess of axonal degeneration that has been reported with conventional insulin treatment was not observed.

Effect of hyperglycemia and its prevention by insulin treatment on the incorporation of 32P into polyphosphoinositides and other phospholipids in peripheral nerve of the streptozotocin diabetic rat

The influence of varying doses of streptozotocin and preventive insulin treatment on phospholipid metabolism in sciatic nerve in vitro from diabetic rats was studied. Animals were given 30, 45, and 60 mg/kg injections of streptozotocin and 10 weeks later nerves were removed and incubated in the presence of [32P]-orthophosphate. The quantity of isotope incorporated into phosphatidylinositol-4,5-bisphosphate (PIP2) was progressively greater with increasing drug dosage, whereas uptake of label into other phospholipids was unchanged. Rats were made diabetic and within 72 h were implanted with long-acting, insulin-containing osmotic minipumps and the incorporation of [32P]orthophosphate into phospholipids of intact and epineurium-free nerves was examined 8 weeks later. For whole nerve, increased labeling in nerves from diabetic animals occurred only in PIP2 and phosphatidylinositol-4-phosphate (PIP) and was completely prevented by insulin treatment. Isotope incorporation into polyphosphoinositides was also markedly elevated (greater than or equal to 100%) in desheathed diabetic nerves, but not in nerves from insulin-treated animals. Other phospholipids in epineurium-free nerves displayed some rise in isotope uptake, but the increases were not prevented by insulin treatment and appeared unrelated to hyperglycemia. Morphological examination of nerves extended previous findings that prolonged insulin treatment produces axonal degeneration. These observations indicate that abnormal nerve polyphosphoinositide metabolism is at least in part a consequence of hyperglycemia. The metabolic alterations may be intimately involved in reduced nerve conduction velocity, which is characteristic of diabetic neuropathy.

The effect of insulin treatment on myelinated nerve fibre maturation and integrity and on body growth in streptozotocin-diabetic rats

Diabetes mellitus was produced in rats by the administration of streptozotocin and observations made over a period of 2 months. Four groups of animals were studied: onset and end controls, untreated diabetic rats and rats treated daily with a long-acting insulin preparation. Body weight increased in the end controls and insulin-treated diabetic animals to a similar degree over the observation period but was reduced in the untreated diabetic rats. Skeletal growth, assessed by measurements of tibial length, was also reduced in the untreated diabetic rats and partially corrected by insulin treatment. Myelinated fibre diameter in the tibial and sural nerves increased over the observation period in the controls, but the increase was less in the untreated animals and the growth deficit was not corrected by insulin treatment. Myelinated fibre numbers did not alter in the tibial or sural nerves between the onset and end controls. Numbers were significantly less in the tibial nerves of both the untreated and insulin-treated diabetic rats as compared with the two control groups; in the sural nerves, fibre numbers did not differ significantly between the four groups. Finally, the number of degenerating axons, assessed in teased fibre preparations, was very small in the control and untreated diabetic animals but was significantly increased in the insulin-treated group. Measurements of plasma glucose concentrations did not suggest that the axonal degeneration could be related to hypoglycaemia. The explanation for this paradoxical effect of insulin therapy is uncertain. It may be dependent upon fluctuations in blood glucose levels or other metabolic actions of insulin apart from its hypoglycaemic effect.