Diabetic peripheral neuropathy: evidence for apoptosis and associated mitochondrial dysfunction

Abnormal PI3 kinase/Akt signal pathway in vagal afferent neurons and vagus nerve of streptozotocin-diabetic rats

The PI3 (phosphatidylinositol-3) kinase/Akt (protein kinase B) signal pathway is involved in the molecular signaling that regulates retrograde axonal transport of neurotrophins in the nervous system. Previous work showed that a reduced retrograde axonal transport of endogenous nerve growth factor (NGF) and neurotrophin-3 (NT-3) in the vagus nerve of diabetic rats occurred in the presence of normal production of neurotrophins and neurotrophin receptors. To assess the potential involvement of an impaired PI3 kinase/Akt signal pathway in the diabetes-induced reduction in retrograde axonal transport of neurotrophins in the vagus nerve, we characterized diabetes-induced changes in the PI3 kinase/Akt signal pathway in the vagus nerve and vagal afferent neurons. Control and streptozotocin (STZ)-induced diabetic rats with a duration of 16 weeks, kinase assays, Western blotting, and immunocytochemistry were used to show that diabetes resulted in alterations in activity and protein expression of the PI3 kinase/Akt signal pathway in the vagus nerve and vagal afferent neurons. Diabetes caused a significant decrease in enzymatic activity of PI3 kinase and Akt (52 and 36% of control, respectively) in the vagus nerve. The reduced enzymatic activity was not associated with decreased protein expression of the p85 subunit of PI3 kinase, Akt and phosphorylation of Akt (ser473). In contrast, there was a significant increase in the phosphorylation of p70s6 kinase (thr421/ser424) along with a normal protein expression of p70s6 kinase in the vagus nerve of diabetic rats. However, diabetes induced an overall decrease in immunoreactivity of the p85 subunit of PI3 kinase, phospho-Akt (ser473) and phospho-p70s6/p85s6 kinase (thr421/ser424) in vagal afferent neurons. Thus, impaired PI3 kinase/Akt signal pathway may partly account for the reduced retrograde axonal transport of neurotrophins in the vagus nerve of STZ-induced diabetic rats.

Aldose reductase mediates cytotoxic signals of hyperglycemia and TNF-alpha in human lens epithelial cells

Chronic hyperglycemia and cytokines such as tumor necrosis factor alpha (TNF-alpha) cause oxidative stress leading to dysregulated cell growth or apoptosis that contributes to the development of inflammation and secondary complications of diabetes. However, the mechanisms regulating hyperglycemic or cytokine injury are not well understood. Herein we report that inhibition of the polyol pathway enzyme aldose reductase (AR) by two structurally unrelated inhibitors--sorbinil and tolrestat--prevents, in the human lens epithelial cell line B-3, the apoptosis and activation of caspase-3 caused by exposure to high glucose levels or TNF-alpha. Inhibition of AR attenuated TNF-alpha and hyperglycemia-induced activation of protein kinase C (PKC), phosphorylation of the inhibitory subunit of nuclear factor-kappaB (NF-kappaB), and stimulation of NF-kappaB, but it did not prevent the activation of NF-kappaB and PKC by phorbol ester. Inhibition of AR also attenuated the increase in p38 mitogen-activated protein kinase and c-Jun N-terminal kinase phosphorylation. These signaling pathways were also inhibited in cells in which the expression of AR was reduced by antisense ablation. Collectively, these results identify a new participant in apoptotic signaling and suggest that AR is an obligatory mediator of the apoptotic events upstream of PKC. These observations could provide new insights into the pathophysiology of diabetes and the role of aberrant glucose metabolism in apoptotic cell death.

Apoptosis in diabetic neuropathy

It has been proposed that apoptotic death of some sensory neurons and Schwann cells occurs in, and may be causal for, diabetic neuropathy (DN). Some tantalizing but incomplete evidence for this has emerged from studies of rat models of diabetes and in vitro studies of sensory and sympathetic neurons and Schwann cells exposed to very high concentrations of glucose. This article reviews the evidence and suggests that most studies to date are far from being conclusive. Hence there is room for proper studies of apoptosis in DN, as such studies may reveal hitherto unexplored drug targets that may improve management of the disease.

Oxidative injury and apoptosis of dorsal root ganglion neurons in chronic experimental diabetic neuropathy

We evaluated the effects of chronic hyperglycemia on L5 dorsal root ganglion (DRG) neurons using immunohistochemical and electrophysiologic techniques for evidence of oxidative injury. Experimental diabetic neuropathy was induced by streptozotocin. To evaluate the pathogenesis of the neuropathy, we studied peripheral nerve after 1, 3, and 12 months of diabetes. Electrophysiologic abnormalities were present from the first month and persisted over 12 months. 8-Hydroxy-2'-deoxyguanosine labeling was significantly increased at all time points in DRG neurons, indicating oxidative injury. Caspase-3 labeling was significantly increased at all three time points, indicating commitment to the efferent limb of the apoptotic pathway. Apoptosis was confirmed by a significant increase in the percentage of neurons undergoing apoptosis at 1 month (8%), 3 months (7%), and 12 months (11%). These findings support the concept that oxidative stress leads to oxidative injury of DRG neurons, with mitochondrium as a specific target, leading to impaired mitochondrial function and apoptosis, manifested clinically as a predominantly sensory neuropathy.

Future treatments for diabetic neuropathy: clues from experimental neuropathy

Neuropathy remains a major complication of diabetes and there is no approved treatment that prevents its progression or alleviates the associated symptoms. Animal models of diabetic neuropathy are hampered by their short life span, which precludes the development of overt structural pathology, and they are best viewed as exhibiting early metabolic, neurochemical, and functional indices of nerve disorders that may predict progression to overt diabetic neuropathy. In this context, diabetic animals have use in both establishing potential etiologic mechanisms and for screening novel therapeutic agents. Treatment strategies are evolving in concert with a developing understanding of how hyperglycemia causes nerve dysfunction and recent or ongoing clinical trials are investigating this rational approach to drug design. It is only by the successful demonstration of clinical efficacy of a compound developed by this approach that the use of animal models of diabetic neuropathy can be validated.

High glucose-induced oxidative stress and mitochondrial dysfunction in neurons

The current study examines the association between glucose induction of reactive oxygen species (ROS), mitochondrial (Mt) depolarization, and programmed cell death in primary neurons. In primary dorsal root ganglion (DRG) neurons, 45 mM glucose rapidly induces a peak rise in ROS corresponding to a 50% increase in mean Mt size at 6 h (P<0.001). This is coupled with loss of regulation of the Mt membrane potential (Mt membrane hyperpolarization, followed by depolarization, MMD), partial depletion of ATP, and activation of caspase-3 and -9. Glucose-induced activation of ROS, MMD, and caspase-3 and -9 activation is inhibited by myxothiazole and thenoyltrifluoroacetone (P<0.001), which inhibit specific components of the Mt electron transfer chain. Similarly, MMD and caspase-3 activation are inhibited by 100 microM bongkrekic acid (an inhibitor of the adenosine nucleotide translocase ANT). These results indicate that mild increases in glucose induce ROS and Mt swelling that precedes neuronal apoptosis. Glucotoxicity is blocked by inhibiting ROS induction, MMD, or caspase cleavage by specific inhibitors of electron transfer, or by stabilizing the ANT.

Diabetes is not a potent inducer of neuronal cell death in mouse sensory ganglia, but it enhances neurite regeneration in vitro

We examined the effects of diabetes on the morphological features and regenerative capabilities of adult mouse nodose ganglia (NG) and dorsal root ganglia (DRG). By light and electron microscopy, no apoptotic cell death was detected in the ganglia obtained from either streptozotocin (STZ)-induced diabetic or normal C57BL/6J mice in vivo. Neurite regeneration from transected nerve terminals of NG and DRG explants in culture at normal (10 mM) and high (30 mM) glucose concentrations was significantly enhanced in the diabetic mice. Chromatolytic changes (i.e. swelling and migration of the nucleus to an eccentric position in the neurons, and a loss of Nissl substance in the neuronal perikarya) and apoptotic cell death (less than one-fifth of the neurons) in the cultured ganglia were present, but neither hyperglycemia in vivo nor high glucose conditions in vitro altered the morphological features of the ganglia or the ratios of apoptotic cells at 3 days in culture. By semiquantitative RT-PCR analysis, the mRNA expressions of ciliary neurotrophic factor (CNTF) in DRG from both mice were down-regulated at 1 day in culture. The expression in diabetic DRG, but not in control DRG, was significantly up-regulated at later stages (3 and 7 days) in culture. In summary, hyperglycemia is unlikely to induce cell death in the sensory ganglia, but enhances the regenerative capability of vagal and spinal sensory nerves in vitro. The up-regulation of CNTF mRNA expression during the culture of diabetic DRG may play a role in the enhanced neurite regeneration.

Type 1 and type 2 diabetic patients display different patterns of cellular microparticles

The development of vasculopathies in diabetes involves multifactorial processes including pathological activation of vascular cells. Release of microparticles by activated cells has been reported in diseases associated with thrombotic risk, but few data are available in diabetes. The aim of the present work was to explore the number and the procoagulant activity of cell-derived microparticles in type 1 and 2 diabetic patients. Compared with age-matched control subjects, type 1 diabetic patients presented significantly higher numbers of platelet and endothelial microparticles (PMP and EMP), total annexin V-positive blood cell microparticles (TMP), and increased levels of TMP-associated procoagulant activity. In type 2 diabetic patients, only TMP levels were significantly higher without concomitant increase of their procoagulant activity. Interestingly, in type 1 diabetic patients, TMP procoagulant activity was correlated with HbA(1c), suggesting that procoagulant activity is associated with glucose imbalance. These results showed that a wide vesiculation process, resulting from activation or apoptosis of several cell types, occurs in diabetes. However, diabetic patients differ by the procoagulant activity and the cellular origin of microparticles. In type 1 diabetic patients, TMP-procoagulant activity could be involved in vascular complications. Moreover, its correlation with HbA(1c) reinforces the importance of an optimal glycemic control in type 1 diabetes.

Hyperglycemia-induced apoptosis in mouse myocardium: mitochondrial cytochrome C-mediated caspase-3 activation pathway

Diabetic cardiomyopathy is related directly to hyperglycemia. Cell death such as apoptosis plays a critical role in cardiac pathogenesis. Whether hyperglycemia induces myocardial apoptosis, leading to diabetic cardiomyopathy, remains unclear. We tested the hypothesis that apoptotic cell death occurs in the diabetic myocardium through mitochondrial cytochrome c-mediated caspase-3 activation pathway. Diabetic mice produced by streptozotocin and H9c2 cardiac myoblast cells exposed to high levels of glucose were used. In the hearts of diabetic mice, apoptotic cell death occurred as detected by terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) assay. Correspondingly, caspase-3 activation as determined by enzymatic assay and mitochondrial cytochrome c release detected by Western blotting analysis were observed. Supplementation of insulin inhibited diabetes-induced myocardial apoptosis as well as suppressed hyperglycemia. To explore whether apoptosis in diabetic hearts is related directly to hyperglycemia, we exposed cardiac myoblast H9c2 cells to high levels of glucose (22 and 33 mmol/l) in cultures. Apoptotic cell death was detected by TUNEL assay and DAPI nuclear staining. Caspase-3 activation with a concomitant mitochondrial cytochrome c release was also observed. Apoptosis or activation of caspase-3 was not observed in the cultures exposed to the same concentrations of mannitol. Inhibition of caspase-3 with a specific inhibitor, Ac-DEVD-cmk, suppressed apoptosis induced by high levels of glucose. In addition, reactive oxygen species (ROS) generation was detected in the cells exposed to high levels of glucose. These results suggest that hyperglycemia directly induces apoptotic cell death in the myocardium in vivo. Hyperglycemia-induced myocardial apoptosis is mediated, at least in part, by activation of the cytochrome c-activated caspase-3 pathway, which may be triggered by ROS derived from high levels of glucose.

Assessment of the "common" 4.8-kb mitochondrial DNA deletion and identification of several closely related deletions in the dorsal root ganglion of aging and streptozotocin rats

The identification of several mitochondrial DNA (mtDNA) deletions and the accumulation of the "common" 4.8-kb mitochondrial DNA deletion (mtDNA(4834)) with aging and experimental streptozotocin-induced diabetes (STZ) were studied in the rat dorsal root ganglion (DRG). Twenty-one mtDNA deletions, including mtDNA(4834), were identified in rat L4-L6 DRG mtDNA of 15-month-old Spraque-Dawley rats with 13 months of STZ and age-matched controls. These deletions were flanked by breakpoints that ranged from 16-bp direct repeats to no direct repeats. The sciatic nerve contained undetectable levels of mtDNA deletions. Levels of mtDNA(4834) in rat DRG mtDNA significantly accumulated with age at a rate much higher than those reported in the brain, yet were not statistically different in STZ. Southern blot analysis demonstrated no significant accumulation of the total amount of mtDNA deletions in STZ over age-matched controls. The accumulation of mtDNA(4834) has not been studied in rat peripheral nerve tissue. Our identification of several mtDNA deletions with and without direct repeats at their breakpoint support the hypothesis that deletions can occur by both the slip-replication model and random recombination. Although there is a significant increase in accumulation of mtDNA(4834) associated with aging, the lack of significant accumulations of mtDNA deletions in STZ over age-matched controls indicates that this type of mtDNA damage is likely not a major alteration in STZ, although the changes could be confined to a small population of neurons that undergo apoptosis between 8 and 15 months.

Oxidative stress and programmed cell death in diabetic neuropathy

Recent evidence in both animal models and human sural nerve biopsies indicates an association with oxidative stress, mitochondrial (Mt) membrane depolarization (MMD), and induction of programmed cell death (PCD). In streptozotocin (STZ)-treated diabetic rats, hyperglycemia induces typical apoptotic changes as well as swelling and disruption of the Mt cristae in diabetic dorsal root ganglion neurons (DRG) and Schwann cells (SC), but these changes are only rarely observed in control neurons. In human sural nerve biopsies, from patients with diabetic sensory neuropathy, there is transmission electromicrograph evidence of swelling and disruption of the Mt and cristae compared to patients without peripheral neuropathy. In human SH-SY5Y neurons, rat sensory neurons, and SC, in vivo, there is an increase in reactive oxygen species (ROS) after exposure to 20 mM added glucose. In parallel, there is an initial Mt membrane hyperpolarization followed by depolarization (MMD). In turn, MMD is coupled with cleavage of caspases. Various strategies aimed at inhibiting the oxidative burst, or stabilizing the DeltaPsi(M), block induction of PCD. First, growth factors such as NGF can block induction of ROS and/or stabilize the DeltaPsi(M). This, in turn, is associated with inhibition of PCD. Second, reduction of ROS generation in neuronal Mt prevents neuronal PCD. Third, up-regulation of uncoupling proteins (UCPs), which stabilize the DeltaPsi(M), blocks induction of caspase cleavage. Collectively, these findings indicate that hyperglycemic conditions observed in diabetes mellitus are associated with oxidative stress-induced neuronal and SC death, and targeted therapies aimed at regulating ROS may prove effective in therapy of diabetic neuropathy.

Morphometry of dorsal root ganglion in chronic experimental diabetic neuropathy

Chronic hyperglycemia results in a predominantly sensory neuropathy. Recent studies suggest that dorsal root ganglion (DRG) neurons comprise a specific target and may be responsible for the important complication of diabetic sensory neuropathy, since hyperglycemia for longer than 6 months results in a vacuolar ganglionopathy with associated radiculopathy and distal sensory neuropathy. We undertook morphometric analysis of L5 DRG neurons in seven diabetic rats and six age- and sex-matched littermates. Nerve conduction studies were also performed, and neuropathy was confirmed. Diabetes was induced with streptozotocin; duration of diabetes was 12 months. The DRG count for control rats was 15,304 +/- 991 neurons. Two of seven diabetic DRG counts were reduced, but the group mean count at 14,847 plus minus 1,524 was not significantly reduced. The number of small neurons (type B) considerably exceeded that of large neurons (type A), at a ratio of 71:29. The percentage of large cells was significantly reduced in diabetic compared with control rats (P = 0.01). The large-diameter population can be subdivided into two groups; with this subdivision, the number of neurons < 50 microm was not reduced in samples from diabetic rats, but the neurons of largest size (> or = 50 microm) were significantly reduced (by 41%).

GCPII (NAALADase) inhibition prevents long-term diabetic neuropathy in type 1 diabetic BB/Wor rats

AIMS/HYPOTHESIS

Hyperglutamatergic activity induced by ischemia is believed to underlie neuronal damage in a variety of neurological disorders, including neuropathic pain. Since ischemia is believed to be a prominent mechanism involved in diabetic polyneuropathy (DPN), we investigated the effect of the glutamate carboxypeptidase II (GCPII, EC #3.4-17.21; previously termed NAALADase), an enzyme responsible for the hydrolysis of the neuropeptide NAAG to NAA and glutamate, on the development of DPN in type 1 diabetic BB/Wor rats.

METHODS

Diabetic animals were treated with 10 mg/kg/day i.p. of the selective GCPII inhibitor GPI-5232 from onset of diabetes for 6 months. Hyperalgesia to thermal stimulation and nerve conduction velocity (NCV) were measured monthly. The effect on structural DPN was assessed by scoring of single, teased myelinated fibers, myelinated fiber morphometry and ultrastructural examination of C-fibers at 6 months.

RESULTS

GCPII inhibition showed significant but partial effects on hyperalgesia (p<0.001), nerve conduction slowing (p<0.01) axonal and nodal structural changes (p<0.001), small myelinated fiber atrophy, and degenerative changes of C-fibers.

CONCLUSIONS

GCPII inhibition has beneficial effects on hyperalgesia, nerve function, and structural degenerative changes in DPN, which are likely mediated by inhibition of ischemia-induced glutamate release.