Diabetes Mellitus and the Peripheral Nervous System: Manifestations and Mechanisms

INVITED REVIEW ABSTRACT: Diabetes targets the peripheral nervous system with several different patterns of damage and several mechanisms of disease. Diabetic polyneuropathy (DPN) is a common disorder involving a large proportion of diabetic patients, yet its pathophysiology is controversial. Mechanisms considered have included polyol flux, microangiopathy, oxidative stress, abnormal signaling from advanced glycation endproducts and growth factor deficiency. Although some clinical trials have demonstrated modest benefits in disease stabilization or pain therapy in DPN, robust therapy capable of reversing the disease is unavailable. In this review, general aspects of DPN and other diabetic neuropathies are examined, including a summary of recent therapeutic trials. A particular emphasis is placed on the evidence that the neurobiology of DPN reflects a unique yet common and disabling neurodegenerative disorder. Muscle Nerve 36: 144–166, 2007

DIABETES MELLITUS AND THE PERIPHERAL NERVOUS SYSTEM: MANIFESTATIONS AND MECHANISMS

DOUGLAS W. ZOCHODNE, MD

Department of Clinical Neurosciences and the Hotchkiss Brain Institute, University of Calgary, Room 168, Heritage Medical Research Bldg., 3330 Hospital Drive N.W., Calgary, Alberta T2N 4N1, Canada

Accepted 25 February 2007

Diabetes mellitus imposes substantial burdens on including diffuse damage (polyneuropathy) and fo- the nervous system and is the most common cause of cal damage (mononeuropathy). Both contribute to neuropathy or peripheral nerve damage. Moreover, sensory and motor deficits and both are associated diabetic neuropathies are rising in prevalence with with significant disability in patients. In polyneurop- the growing global burden of type II diabetes melli- athy it is now recognized that impaired glucose tol- tus. Although this review emphasizes peripheral erance, even without overt diabetes mellitus, may be nerve disorders, there is now recognition that diabe- a risk factor. tes also targets the central nervous system, especially The San Antonio Consensus criteria are com- white matter (diabetic leukoencephalopathy).15,234 monly used to define diabetic neuropathy for re- 5 Within the peripheral nervous system alone, how- search purposes. For clinical neuropathy, the guide- ever, diabetes renders several types of nerve damage, lines require symptoms and signs, or one of these with abnormal testing (nerve conduction, quantitative sensory testing, or autonomic testing). Subclini-

Available for Category 1 CME credit through the AANEM at cal neuropathy is identified by abnormal testing www.aanem.org. only. More specific staging of diabetic polyneurop- Abbreviations: AGEs, advanced glycosylation endproducts; AR, aldose re- athy (DPN) has also been described by Dyck and ductase; ARIs, aldose reductase inhibitors; CASE IV, computer assisted sen- Dyck50: NO, no neuropathy; N1, asymptomatic neu- sory examination; CDT, cool detection thresholds; CMAP, compound muscle action potential; CMT 1a, Charcot–Marie–Tooth disease type 1a; CTS, carpal ropathy without (N1a) or with (N1b) findings on tunnel syndrome; DLSP, diabetic lumbosacral plexopathy; DPN, diabetic neurological examination; N2, symptomatic; N3, dis- polyneuropathy; DRG; dorsal root ganglion; HNDT, heat-nociception detection threshold; HSP-27, heat shock protein-27; IENF, intraepidermal skin abling. fibers; IGFs, insulin-like growth factors; MAG, myelin-associated glycoprotein; Both pathophysiology and therapy for diabetic NGF, nerve growth factor; PKC, protein kinase C; POEMS, polyneuropathy, organomegaly, endocrinopathy, monoclonal protein, and skin changes; QST, neuropathies remain challenging. There has been a quantitative sensory testing; RICF, resistance to ischemic conduction failure; long history of failed clinical trials for polyneurop- SFEMG, single-fiber electromyography; SNAP, sensory nerve action potential; SSRI, selective serotonin reuptake inhibitor; STZ, streptozotocin; UNE, athy, in part related to issues of what was targeted, ulnar neuropathy at the elbow; VPT, vibration perception threshold; WDT, what was being measured, and how well the trial was warm detection thresholds Key words: diabetes mellitus; diabetic neuropathy; neuropathy; peripheral designed. Despite these problems, there are new and neuropathy; polyneuropathy exciting thoughts about how these disorders develop Correspondence to: D. W. Zochodne; e-mail: [email protected] and what avenues may offer significant hope. Be- © 2007 Wiley Periodicals, Inc. Published online 27 April 2007 in Wiley InterScience (www.interscience.wiley. cause of the size of the topic, a number of aspects are com). DOI 10.1002/mus.20785 only covered briefly in this review and the bias is

144 Diabetes Mellitus and the PNS MUSCLE & NERVE August 2007 toward emphasizing aspects of its neurobiology. intensively treatment patients had developed it by 5 Three excellent and comprehensive texts addressing years. For patients in the secondary intervention co- diabetic neuropathy have been published6,81,230 in hort with retinopathy at baseline but not neuropa- addition to recent reviews addressing slightly differ- thy, the rates were 16.1% for conventional and 7.0% ent points of view,18,241 and diagnostic criteria have for intensive treatment. Overall, when looking at a recently been published by the American Diabetes variety of studies (summarized by Shaw et al.197), Association.19 type I diabetic prevalence figures vary from 13%– 17% in hospitalized patients based on symptoms and CLASSIFICATION AND PREVALENCE signs, and 8%–54% with more comprehensive batteries in primary care or population-based screening. Diabetic neuropathies comprise diabetic polyneu- For type II diabetic patients, similar figures run from ropathy (DPN), a symmetric diffuse disorder that 19%–58% in hospital-based studies with some ancil- particularly targets sensory neurons with long axons, lary testing and 13%–46% in primary care or popu- and focal neuropathies or mononeuropathies. The lation-based screening more heavily weighted toward latter include classic entrapment neuropathies that testing. There are likely significant flaws, however, are more common in diabetes such as carpal tunnel from relying on hospital-selected data. syndrome (CTS), ulnar neuropathy at the elbow With very comprehensive and extensive batteries (UNE), meralgia paraesthetica (entrapment of the of evaluation, such as that applied to the Rochester lateral femoral cutaneous nerve of the thigh) at the Diabetic Cohort (n ϭ 380), evidence of DPN was inguinal ligament, or peroneal neuropathy at the identified in 54% of type I diabetics and 45% of type fibular head. Other mononeuropathies much more 58 specifically identified in diabetic patients include II diabetics. Using the strict criteria of an abnormal intercostal and abdominal segmental radiculopa- neuropathy impairment scale (NIS) and seven ab- thies, oculomotor palsies, and lumbosacral radiculo- normal laboratory studies, 21% of the Rochester plexus neuropathies. Brown and Asbury 27 subdi- diabetic cohort had DPN. Symptomatic DPN was vided DPN clinically into subtypes, with the group of identified in a smaller proportion, 13%–15%. In mixed motor, sensory, and autonomic neuropathy other cohorts, such as the Pittsburgh epidemiology ϭ representing 70% of patients. A predominantly sen- of diabetes complications (n 400), DPN was iden- sory phenotype was found in 39% that was yet fur- tified in 34% of type I diabetics, whereas in the San ther divided into large-fiber, small-fiber, or mixed Luis Diabetes Study DPN was present in 26% of type ϭ neuropathies. Pure motor DPN or autonomic DPN II diabetics (n 279). In patients with impaired were uncommon (Ͻ1% each). In the author’s expe- glucose tolerance only, as a precursor of type II rience, pure sensory DPN on the basis of clinical diabetes, the prevalence figures have been more evaluation alone (some have subclinical electrophys- controversial. iological motor involvement) represents the large The prevalence of cardiovascular autonomic neu- majority of patients, particularly early in their ropathy detected by heart-rate interval studies (in- course. Some have added a category of an acute cluding the response to Valsalva’s maneuver, or deep ϳ sensory DPN with rapid onset (likely overlapping breathing) has ranged from 16%–25% in type I with a condition known as “insulin neuritis” or neu- and II diabetic patients, with a smaller proportion ropathy after the onset of insulin use), an association having symptoms. Several studies have suggested with acute hyperglycemia, the presence of promi- that cardiovascular autonomic neuropathy is a risk nent pain, and a shorter overall duration related to factor for increased mortality.65,147,158,197,238 For gas- control of hyperglycemia.18,19,241 trointestinal symptoms, prevalence figures are also The reported prevalence of DPN varies with the variable, with numbers for constipation or diarrhea type and the intensity with which it is sought. In the ranging between 3% and 35%. Impotence has been classic Diabetes Control & Complications Trial identified in 23%–57% of type I and II diabetic men, (DCCT) of diabetic complications in intensively with higher rates with increasing age.197 Overall, a rather than conventionally treated patients with type population-based study from the Rochester diabetic I diabetes mellitus,46 clinical neuropathy was defined cohort (n ϭ 231 diabetics) identified a prevalence of as an abnormal clinical neurological examination autonomic dysfunction (using a composite scale of plus either abnormal nerve conduction in at least laboratory-based autonomic tests known as CASS) of two peripheral nerves or unequivocally abnormal 54% in type I diabetics and 73% in type II diabetics, autonomic-nerve testing. In patients without neurop- with postural hypotension in 8.4% and 7.4%, respec- athy at baseline, 9.8% of conventional and 3.1% of tively.124

Diabetes Mellitus and the PNS MUSCLE & NERVE August 2007 145 Further work in the Rochester diabetic cohort mechanical stimuli) from bed covers or from walk- identified symptomatic CTS in 7.6%.58 In an older ing. Walking thus may be painful and hesitant or Mayo Clinic study by Mulder et al.,145 symptomatic unsteady. Although diabetic patients do not gener- CTS was identified in 8.7% of 103 diabetic subjects ally exhibit a wide-based classic sensory ataxia, loss of and 4.9% had UNE. touch and pressure sensitivity on the soles, and later Risk factors for the development of DPN have loss of more specific proprioceptive axons, may pre- included the degree and duration of hyperglycemia, dispose them to falling. The sensory abnormalities of smoking, and the presence of other complications DPN are “stocking and glove” in distribution, reflect- such as retinopathy and nephropathy.53,162,163,197 ing the susceptibility of longer nerves. Tesfaye et al.,223 representing 31 centers participat- Clinical neurological examination by experi- ing in the European Diabetes (EURODIAB) Pro- enced examiners has inherent variability but its ad- spective Complications Study, examined risk factors vantages are its low cost, multimodality evaluation, for the development of DPN in patients with type 1 and rendering of a direct examiner–patient interac- diabetes mellitus over a 7–10-year period. At end- tion. Sometimes this kind of interaction is essential point, DPN had developed in 276 of 1,172 patients in identifying early foot ulcerations or other prob- without it at onset (23.5%) and its cumulative inci- lems and should not be replaced by laboratory mea- dence was related to the glycosylated hemoglobin sures such as computerized QST (quantitative sen- value and duration of diabetes. Independent risk sory testing). An additional screening tool for factors were higher levels of total and low-density sensory loss is the Semmes–Weinstein 10 g monofil- lipoprotein cholesterol and triglycerides, a higher ament test. The monofilament is pressed against the body mass index, higher von Willebrand factor lev- dorsum of the large toe (plantar application has also els, urinary albumin excretion rate, hypertension, been described) until it bends and the patient (with and smoking. Cardiovascular disease at baseline in- eyes closed) is asked to identify when the stimuli are dependently doubled the risk of neuropathy. Over- applied (for example, 5–10 times on each toe). all, the findings identified a striking correlation be- In addition to sensory loss which may involve the tween risk factors for vascular disease and the toes, whole foot, or leg below the knee, for example, development of DPN. there may be loss of ankle reflexes despite reinforce- It is possible that the severity and impact of DPN ment, and in more advanced neuropathies, loss of on individual patients is easing, given better control the knee reflexes or upper-limb deep tendon re- regimens, more education, and improved pain ther- flexes. In DPN the earliest motor manifestation is apy. This possibility, however, has not been rigor- atrophy of the EDB muscle. Clinical weakness of toe ously addressed and such improvements may be un- and foot dorsiflexors occurs later. Subtle motor axon even among populations and regions. involvement may lead to abnormalities of foot pos- ture that contribute to foot ulceration and gait im- CLINICAL MANIFESTATIONS balance. Charcot joints, for example, at the ankle, To appreciate fully the clinical features of DPN, the are destructive arthropathies secondary to repetitive reader is directed to one of the more interesting and injury of which the patient may be unaware (Fig. 1). comprehensive articles on the subject, written in Symptoms of diabetic autonomic neuropathy in- 1945 by Rundles.175 Symptomatic DPN usually pre- clude distal anhidrosis or inappropriate truncal sents with spontaneous positive (paresthesias de- sweating and rarely gustatory sweating (facial and scribed as prickling, tingling, “pins and needles,” trunk sweating from certain foods).246 Constipation, burning, crawling, itching, abnormal sensation to sometimes alternating with diarrhea, is a symptom of temperature, pain), or negative (numbness, injury diabetic visceral autonomic neuropathy. Other gas- insensitivity) sensory symptoms in the toes. With trointestinal autonomic symptoms are less common time, such sensations may advance up the foot and and include frank fecal incontinence, dysphagia, leg and involve the fingers and hands. Neuropathic heartburn, nausea and vomiting, early satiety, and pain is a prominent early feature of DPN and can be bloating and fullness after meals. Patients may have severe despite minimal signs of DPN. Patients de- unawareness of hypoglycemia. Impotence is a com- scribe their feet as “tight,” having painful prickling, mon manifestation of autonomic neuropathy in men burning, electrical, sharp, or jabbing sensations. (vascular factors also play a role). Postural hypoten- Symptoms may be prominent at night, and overall sion with orthostatic dizziness or occasional fainting quality of life is significantly compromised. There results from loss of sympathetic control to resistance may be allodynia (provocation of pain by simple arterioles. Loss of bladder sensitivity occurs in diabe-

146 Diabetes Mellitus and the PNS MUSCLE & NERVE August 2007 UNE may have a significant impact on hand function if motor involvement occurs. While asymptomatic, mild slowing of ulnar motor conduction across the elbow is common in diabetic patients; more significant motor involvement is associated with wasting and weakness of intrinsic ulnar-innervated hand muscles. Thoracic intercostal and abdominal radic- ular neuropathies are well described in diabetes, with a differential diagnosis that includes zoster without rash or, less commonly, radiculopathy from a segmental structural lesion.216 The neuropathies generally span several contiguous segmental territo- ries and may be bilateral. They can present with severe thoracic or abdominal-wall pain, sometimes suggesting a visceral emergency. Distinguishing features include sensory loss if present (not always), descriptors of the pain that indicate a neuropathic etiology (tingling, pricking, radiation around the chest or abdomen), and sometimes the presence of allodynia. Other focal neuropathies in diabetes, such FIGURE 1. Charcot joint at the right ankle in a patient with DPN. as peroneal neuropathy at the fibular head, meralgia Charcot joints occur from repetitive unrecognized damage in paraesthetica, and pupil-sparing oculomotor palsy joints with sensory deafferentation. Such joints, however, may are indistinguishable from similar neuropathies in still generate pain in some patients. nondiabetics. Diabetic lumbosacral plexopathy (DLSP), some- tes leading to overflow incontinence, but this is more times also known as radiculoplexus neuropathy, di- strictly classified as a visceral sensory deficit. abetic amyotrophy, Bruns–Garland syndrome, or Diabetic foot ulceration is a major forerunner of proximal diabetic neuropathy, is less common but eventual amputation. DPN contributes to the devel- highly debilitating.12,61 Rarely, nondiabetics can de- opment of foot ulcers in several ways. There is loss of velop a similar syndrome that appears to respond to protective sensory sensation in the feet, with conse- high-dose corticosteroid therapy.60 DLSP is often as- quent repetitive injury. Patients may injure their foot sociated with significant weight loss or cachexia. Sur- without realizing it. Foot atrophy distorts how gravi- prisingly, it may emerge early in the course of dia- tational forces are transmitted to foot and leg bones betes or following institution of insulin therapy in and leads to undue pressure on bony prominences. type II diabetic patients. It is usually asymmetric and Loss of sweating in the feet leads to skin drying and associated with deep boring or aching pain in the cracking. Clinical evaluation, including use of mono- thigh. Over the next few weeks after development of filaments, is valuable in predicting the risk of foot pain, weakness and wasting of proximal thigh mus- ulceration and some have advocated yearly screen- cles develops in iliopsoas, quadriceps, and thigh ad- ing.1,19 ductors. Some patients become unable to walk. CTS is most commonly associated with positive Spontaneous recovery occurs but is slow, over a num- sensory symptoms and pain in the territory of the ber of months, and in some cases contralateral DLSP median nerve. Many patients complain of symptoms emerges. Variations of DLSP include symmetric in- in all digits of their hand. Symptoms may be prom- volvement, foot drop, or apparent worsening of gen- inent at night or on awakening and are provoked by eralized DPN. repetitive activities involving the hand or by preg- Other causes of polyneuropathy must always be nancy. Concurrent hypothyroidism is important to considered in a diabetic patient. Patients, for exam- exclude. A high index of suspicion for treatable CTS ple, with a more subacute and a prominent motor in diabetic patients who also have polyneuropathy is disorder may have chronic inflammatory demyeli- valuable: in some patients CTS may emerge later nating polyneuropathy (CIDP) with more promi- with increased use of the hands when walking is nent features of primary demyelination on neuro- difficult. Longer standing and more severe CTS may physiological testing (see below). CIDP is recognized be associated with hand weakness from denervated to accompany diabetes in some patients.217 In the thenar muscles. author’s practice, additional patients with diabetes

Diabetes Mellitus and the PNS MUSCLE & NERVE August 2007 147 have been identiﬁed to have POEMS syndrome difﬁcult to establish, since CTS and UNE are fre- (polyneuropathy, organomegaly, endocrinopathy, M quently admixed and may cloud the interpretation protein, and skin changes), multifocal motor neu- of upper-limb studies. Conduction slowing or loss of ropathy, mitochondrial cytopathy, anti-MAG (mye- amplitude of the radial SNAP recorded at the base of lin-associated glycoprotein) neuropathy, and others. the thumb is a useful index of sensory DPN in the

Associated hypothyroidism or vitamin B12 deficiency upper limbs, and this nerve is not generally prone to should also be excluded. Although debated and ad- entrapment. Needle electromyography may detect vocated in some models, unequivocal clinical differ- distal denervation. Renal failure superimposed on ences in DPN of patients with type I or II diabetes DPN is thought to suppress abnormal spontaneous independent of age or duration of diabetes have not activity such that denervation may be more difficult been demonstrated. The combination of lumbar spi- to recognize.16 Motor unit recruitment is reduced in nal stenosis and CTS may present features that re- distal muscles of patients with DPN and remaining semble polyneuropathy with lower- and upper-limb units are remodeled and enlarged, indicating involvement. “Pseudopolyneuropathy” from this chronic denervation with reinnervation. combination can be distinguished by careful clinical, There is a large literature on the use of a number of neurophysiological, and imaging approaches. other neurophysiological techniques in DPN, but these are not reviewed here. They include elevated jitter on 25 NEUROPHYSIOLOGY SFEMG (single-fiber electromyography), F-wave latencies, H-reflex latencies, somatosensory evoked po- Neurophysiological measurements are gold stan- tentials, refractory period testing, macro-electromyo- dards in the evaluation of DPN. They are widely graphy, axonal threshold accommodation, resistance available, generally carried out in standardized for- to ischemic conduction failure (RICF), and motor unit mats, and have been rigorously examined in clinical number estimates. Extensive reviews of neurophysio- trials. Among the measures available, the amplitude logical testing in DPN and other diabetic neuropathies of the sural sensory nerve action potential (SNAP), have been published.22,45,251 measured behind the ankle 140 mm from the calf The neurophysiological characteristics of DLSP stimulating site with a surface skin temperature of at include focal denervation of muscles innervated by least 30°C, is the earliest alteration in DPN.22 A the lumbosacral plexus, e.g., iliopsoas, quadriceps, decline in amplitude below 6-␮V when measured and adductor muscles. CMAPs recorded over vastus from baseline to peak is considered abnormal, and medialis may be smaller on the side of involve- the amplitude correlates with myelinated fiber den- ment—their recovery parallels clinical improve- sity.22 Other sensitive indices are the declines in sural ment. Paraspinal muscle denervation is also com- nerve conduction velocity and peroneal nerve motor mon, indicating involvement of proximal roots. Foot conduction velocity. In clinical trials, centralized drop from peroneal nerve involvement may occur, monitoring of nerve conduction results and tech- and the overall pattern of neurological involvement nique with review of waveforms has substantially im- may be quite patchy. Most patients also demonstrate proved the quality of the material and of the associ- widespread changes of DPN, largely chronic but oc- ated trials.24 Most adequately powered and designed casionally exacerbated during DLSP. trials involve training over specific trial procedures Autonomic testing is important in identifying ei- and assessment of normal controls. High-quality tri- ther selective autonomic involvement or associated als have involved repeated measures at the trial onset autonomic dysfunction in DPN.71 Forms of testing and endpoints. include evaluation of cardiovascular autonomic neu- DPN is associated with some conduction slowing, ropathy (RR interval variation at rest, with deep although not as severe as in CIDP or CMT1a (Char- breathing, and with Valsalva maneuver; blood pres- cot–Marie–Tooth disease type 1A), suggesting pri- sure response to standing, Valsalva maneuver, or mary demyelination, and reduction in amplitude or isometric exercise; spectral analysis). Prolonged QTc loss of CMAPs (compound muscle action potentials) intervals in type I diabetics may predict an increased and SNAPs, indicating axon dropout. Distal motor risk of mortality.238 latencies may be asymptomatically prolonged at sites Quantitative sensory testing includes approaches of entrapment such as the carpal tunnel. Frank con- that span Semmes–Weinstein 10-g monofilament duction block is uncommon, except at entrapment sensation161,237 to sophisticated computerized inter- sites. Temporal dispersion can be identified, but faces such as the CASE IV (computer-assisted sensory usually in the setting of entrapment or severe dis- examination) developed by Dyck and colleagues.54 ease. DPN involving the upper limbs may be more Modalities tested for by QST include vibration per-

148 Diabetes Mellitus and the PNS MUSCLE & NERVE August 2007 ception threshold (VPT; probably the most com- ation, and pericyte degeneration.56,113,133,228,252 De- monly used form of QST in clinical trials) but also clines in endoneurial capillary luminal area and rises touch-pressure, warm and cool detection thresholds in basement membrane thickness develop very early (WDT and CDT), and heat as pain (HNDT, or heat- in patients, even those with impaired glucose toler- nociception detection threshold). CASE IV can be ance prior to frank diabetes.132,226 Overall, the find- used for VPT, WDT, CDT, and HNDT.218 QST ab- ings confirm the importance of concurrent micro- normalities in clinical trials correlate with other mea- vascular disease in DPN, but do not necessarily sures of DPN.54 From a report of the Therapeutics indicate cause and effect, since it may represent and Technology Assessment Subcommittee of the parallel involvement of nerves and vessels. 198 American Academy of Neurology, QST has been Punch biopsies of the skin (3 mm), used to ex- considered a useful test for the diagnosis of DPN for amine epidermal innervation of axons, is a sensitive clinical and research purposes, but not alone and index of DPN.88,107,110 An alternative approach has the evidence for its sole use for diagnosis had not yet been the use of a skin blister.109 Samples are exam- achieved the highest level. ined using immunohistochemistry with an antibody directed to the axon marker PGP 9.5. Standardized MORPHOLOGICAL AND STRUCTURAL FEATURES approaches toward the sampling and processing of The most common abnormality of sural nerves stud- skin biopsies and the measurement of the linear ied morphologically in patients with DPN is axon density of intraepidermal skin fibers (IENF) have loss. The numbers of remaining myelinated axons been proposed by the European Federation of Neu- correlate well with the amplitude of the SNAP re- rological Societies.116 DPN is associated with a reduc- corded electrophysiologically prior to biopsy. In hu- tion in the density of IEFN110 and other changes, mans, it is uncertain whether such loss reflects re- such as the presence of degenerative axonal traction of branches of sensory neurons or loss of the ovoids.88,110 The loss of small epidermal axons is complete neuronal tree including perikarya at the important because it has a significant impact on ganglion. In some patients there may be evidence of wound healing.168 Reductions can also be observed, active axonal degeneration, but this is less common however, in patients with impaired glucose tolerance given the chronicity of the neuropathy. Axon loss is without frank diabetes.164,221 Kennedy107 also identi- patchy or multifocal (unless it is so severe that most fied gastrointestinal tract denervation in patients axons are absent), a pattern that has suggested an with autonomic neuropathy. ischemic or microangiopathic mechanism.57,59,95 In Recently, a sensitive and noninvasive approach some other forms of neuropathy, however, such as toward assessing DPN has been developed by exam- inherited neuropathy unrelated to microvascular dis- ining corneal innervation using corneal confocal mi- 123 ease, multifocal loss has also been identified. Un- croscopy. Malik et al.130 described the technique in myelinated axon loss and regeneration may also be 18 diabetic patients and 18 age-matched control sub- observed.132 It is probably a common misconception jects that involved scanning the cornea for the nerve that painful diabetic polyneuropathies are predom- plexus of Bowman’s layer. Corneal nerve fiber den- inantly due to loss of small fibers. Most pathological sity, length, and branch density were reduced in studies of DPN have identified widespread fiber in- diabetic patients, with a trend toward greater reduc- volvement, with both large- and small-fiber involve- tions in patients with more severe DPN. ment.74,135 In addition to undergoing axonal degen- Surprisingly, there are few morphological studies eration, myelinated axons also exhibit segmental demyelination178 and remyelination. Clusters of in humans of dorsal root or autonomic ganglia in small regenerating axons, with thin myelin sheaths, diabetes. A postmortem study, now dated, was car- 79 may superficially resemble onion bulbs (“pseudo- ried out by Greenbaum et al. involving 6 patients. onion bulbs”). Atrophy of axons, a feature of DPN Some, though not extensive, loss of sensory neurons models, has been difficult to demonstrate in human replaced by nests of Nageotte were reported. The biopsies, possibly because of preexisting dropout.74 authors also described vacuolar changes in neurons Alterations in the vasa nervorum have been an im- but their significance is uncertain, since vacuolar portant finding in human DPN nerve samples. A changes from swollen mitochondria can emerge as a number of alterations have been described includ- postmortem artifact.122 Schmidt et al.184 described ing perivascular basement membrane thickening, dystrophic changes of proximal sensory neuron ax- endoneurial capillary closure, microthrombosis, en- ons within dorsal root ganglia of diabetic humans dothelial cell reduplication, smooth muscle prolifer- but also with advanced age.

Diabetes Mellitus and the PNS MUSCLE & NERVE August 2007 149 NEUROBIOLOGY lead to ischemic damage of neurons and axons in 52,55–57,59,95,129,131,133,134,228,252,258 General Considerations. There are some very inter- diabetes. Mechanisms esting aspects of diabetes that reflect a unique neu- for microangiopathy center on endothelial damage robiology. Although some parts of the nervous sys- with features that include impaired vasodilation by nitric oxide (NO), damage from oxidative stress, and tem, such as sensory neurons, are targeted, others alterations generated by polyol flux.32,33,47,49,87,111,117,229,245 (e.g., motor neurons) are less involved. There have The abnormalities of NO release and function by the been substantial problems in bringing together di- endothelium are particularly important and include verse ideas and hypotheses as to how DPN develops. excessive quenching of NO by advanced glycosylation Investigators have not done a good job at interpret- endproducts (AGEs). Overall, diabetic vessels have ing the range of findings and their implications or in more prominent impairment of vasodilatory mecha- linking divergent ideas. At one end of the spectrum nisms, favoring excessive functional vasoconstriction. is the hypothesis that neuropathy is an exclusive Microangiopathy is then exacerbated by hyperviscosity, microvascular disease, brought about by reductions loss of red-cell deformability, increased platelet aggre- in nerve blood flow and ischemia. At another end, to gation, and alterations in local oxygen release, all likely which the present author leans, is the idea that the contributing to an eventual cascade of hypoxia and targeting is directly neuronal. At the level of the ischemia.20,28,29,82,114,126,127,209,214,215,222,245 neuron, the degeneration does not reflect, for ex- An important debate evolves around whether mi- ample, the pattern of alterations that occur after an croangiopathy and associated abnormalities of oxy- axotomy or with primary neuron dropout. It seems gen delivery explain the early development of poly- that neurons are beset by several forms of insult, neuropathy. As already discussed, structural changes including ischemia from microangiopathy, oxidative of microvessels are prominent in human sural nerve stress, and polyol flux. Neurons fail to resist such biopsies. Although some of this work involved pa- input because their support mechanisms are im- tients with long-standing diabetes and increased age, paired from attenuated growth failure support (in- work by Malik et al.135 has identified changes early cluding that of insulin). Some of these proposed during the course of DPN. Newrick et al.149 demon- mechanisms are considered below. strated that human sural nerves have decreased Experimental diabetes has been studied in sev- oxygen tensions in diabetic subjects. In contrast, eral models, particularly after administration of a Theriault et al.224 measured human sural nerve single or short course of streptozotocin (STZ) to blood flow (intraoperative laser Doppler flowmetry) mice or rats, a specific pancreatic beta-cell toxin that in patients with early DPN scheduled for research leads to hyperglycemia within a few days. The STZ trial–related nerve biopsies. Some patients later also model is closest (although not identical, since ro- underwent contralateral nerve blood-flow measures dents may survive for long periods without insulin) prior to follow-up biopsies, at the end of the year- to type I human diabetes. In rats, STZ models have long trial. Interestingly, blood flow did not decline been widely used, although often for only short pe- with increasing severity of DPN but tended instead riods, whereas a more complete repertoire of toward higher values in more severely affected pa- changes may take 6–12 months of diabetes. Simi- tients, or in the same patients over time (Fig. 2). It is larly, STZ diabetes can be studied in mice after unlikely that arteriovenous shunting explains such a somewhat shorter durations (e.g., 4–9 months). change. Patients with vasculitis of their peripheral There are a number of diabetic models, not all listed nerves, in contrast, had significant reductions in lo- here, including the BB/W rat with type I diabetes cal blood flow. discovered by Sima (discussed below), a more recent A number of investigators have identified very early BB/Z type II diabetic rat examined in that labora- reductions in nerve blood flow in animal models of tory, the ZDF (Zucker diabetic fatty) rat, the db/db DPN. Some of these studies can be criticized on tech- or ob/ob mouse with leptin receptor or leptin mu- nical grounds, as discussed in detail elsewhere,266 and tations, respectively, and others as reviewed else- not all investigators have identified such declines in where.100,118 nerve blood flow.34,35,93,125,167,224,227,235,253,259,269–272,277 An extensive battery of pharmacological approaches, Microangiopathy, Polyol Flux, Oxidative Stress. Mi- largely in short-term rat models, have reversed conduc- croangiopathy is a fundamental abnormality of diabe- tion and blood-flow abnormalities in tandem. It is un- tes. Functional abnormalities of the vasa nervorum may certain, however, whether at least some of these ap- develop early in models of DPN and are thought to proaches target other functional abnormalities, such as presage later structural changes. Such changes then blood flow or neuronal function in ganglia. Early

150 Diabetes Mellitus and the PNS MUSCLE & NERVE August 2007 blood flow, where the impact of ischemia might cause greater damage, can be identified in some models and we did confirm some shifts to lower oxygen tensions in both nerve and dorsal root ganglia.273,274 Both nerve trunks and DRG also exhibit abnormalities of blood-vessel function, for example, with prominent sensitivity to endothelin-induced vasoconstriction and ischemic damage.254,267,269 Overall, then, although microangiopathy is an undoubted accompaniment to DPN, the evidence for an exclusive role in the etiology of DPN is mixed. It seems unlikely that DPN patients, already fre- quently treated for vascular disorders, will experi- FIGURE 2. Measurements of blood flow (flags) in intact sural ence benefit from simple approaches such as the use nerves were made in patients using an intraoperative laser Dopp- of vasodilators. Such approaches also pose a risk of ler flowmeter following sural nerve conduction recordings and prior to nerve biopsy. The patients were participants in a clinical exacerbating postural hypotension from concurrent trial for mild DPN requiring sural biopsies at outset and on the autonomic neuropathy. opposite side at endpoint. While the plot indicates correlation Accumulation of Polyols. Accumulation of poly- between sural nerve myelinated fiber density and the amplitude ols, especially sorbitol, into peripheral nerves occurs of the SNAP, blood flow values tended toward higher, rather than lower levels in patients with more severe DPN (higher flow with by excessive flux through the aldose reductase (AR) lower action potential amplitudes and lower myelinated axon pathway. This critical metabolic abnormality has gen- density). Similarly, in the patient who completed the 1-year fol- erated intense attention for several decades and has low-up contralateral biopsy tended to have higher flow values led to a series of clinical trials of aldose reductase despite more neuropathy. Overall, the findings suggested that inhibitors (ARIs). AR is localized largely to glial cells declines in sural nerve blood flow were not a requisite for early DPN. Other mechanisms of disease likely contribute toward early (Schwann, DRG supporting cells; also endothelial axon loss. (Reproduced with permission from Theriault M, et al. cells), and polyol flux is thought to damage support- Brain 1997;120:1131–1138.) ing cells and then axon function.142 In axons, then, there are associated depletions of nerve myo-inositol, changes in protein kinase C (PKC) subunits, and nerve-trunk ischemia may therefore not account for dysfunction of nerve Na/K ATPase.17,20,80,82,141,172,207 diabetic neuropathy; instead, parallel forms of involve- The changes in Na/K ATPase activity, in turn, have ϩ ment may eventually become synergistic. In our labo- been linked to a rise in intra-axonal Na content ratory, we did not identify declines in nerve blood-flow with associated changes of channel properties ulti- (under strict physiological conditions with near-nerve mately manifest as slowing of nerve conduction ve- temperature control) in various models of DPN, using locity.43,80,207 Conduction slowing is a fundamental, a number of independent and blinded examiners widespread, and early marker of DPN that occurs in and a variety of blood-flow measurement tech- motor and sensory, myelinated and unmyelinated niques.104,269–271,273,274,277 We also observed that axons. ARIs or PKC inhibitors thus derive their sympathectomy and vasodilatation, in contrast to rationale for clinical trials.2,43,94,96,143,193,201,207,262 other work, was associated with worsened indices of Whether excessive polyol flux induces neurodegen- experimental DPN, not improvement.271 Moreover, eration, however, is unknown. Similarly, there are direct examination of unfixed vascular profiles in complexities in how PKC subunits are distributed DPN suggests that microvessels increase in caliber and behave in peripheral nerves that make it unclear and may exhibit angiogenesis.277 These findings whether inhibition is beneficial or harmful.256,257 have also suggested that the use of vascular endothe- Free Radical Oxidative and Nitrergic Stress May lial growth factor (VEGF), advocated for increasing Damage the Peripheral Nervous System. Several inves- nerve blood-flow in human clinical trials, may be tigators have provided evidence that free radical lib- unhelpful. For example, preclinical work addressing eration damages axons, Schwann cells, and probably its use in diabetic models has been criticized on perikarya of neurons through oxidative and nitrergic technical grounds regarding the approach used to stress,37,127,128,261 as reviewed elsewhere.239 Polyol evaluate both flow and vessels.196 There is also con- flux and advanced glycosylated endproducts are cern that VEGF may exacerbate diabetic retinopathy, thought to play a major role in the generation of free where elevated levels are thought to induce angio- radicals.153,157 This pathway involves auto-oxidation genesis.3 Declines in dorsal root ganglion (DRG) of glucose, and glycation products of glucose gener-

Diabetes Mellitus and the PNS MUSCLE & NERVE August 2007 151 ate oxygen free radicals that include the hydroxyl although excessive activation may mediate cellular ⅐ ⅐ Ϫ 144 213,248 radical (OH ), superoxide anion (O 2 ), hydrogen dysfunction (reviewed elsewhere ). There are peroxide, singlet oxygen, and organic analogs. Con- early indicators that polyol flux, nitrergic stress, and currently, antioxidant defenses are lowered in diabe- AGE–RAGE signaling are linked in important ways. tes, allowing targeting of lipids, DNA, and pro- For example, consumption of NO by AGEs in diabe- teins.249 The depletion of the amino acid taurine tes may contribute to loss of vasodilation (functional with antioxidant properties has been postulated to microangiopathy) in diabetic microvessels, as dis- aggravate oxidative stress in experimental diabe- cussed earlier.32,33,47,49,87,111,117,229,245 tes.120,154,156,165 Nitrergic stress from the free radical NO may Mitochondrial Dysfunction. Mitochondria may be arise from the targeting of proteins with thiol primary targets of oxidative stress and perhaps groups, DNA, or activation of PARP [Poly (ADP- growth factor deficiency in diabetes. Mitochondria ribose) polymerase] that depletes intracellular en- are essential for the bioenergetic reserve of cells ergy reserves. NO combines with superoxide to gen- including neurons and the formation of permeabil- erate peroxynitrite (ONOOϪ), a highly potent ity transition pores can release cytochrome-c and oxidizing agent that nitrates protein tyrosines and initiate an apoptotic cascade.83 The presence of mor- can eventually lead to cell death.136 It has been sug- phological changes in neuronal mitochondria such gested that nitrergic stress plays a significant role in as vacuolation is questionable, however, and may be the development of neuropathy. Nitric oxide syn- artifactual.122 Alternatively, molecular changes in mi- thase activity in ganglia and nerve is increased, and tochondrial function triggered by oxidative stress there are footprints of peroxynitrite toxicity in gan- may be critical. Along these lines, alterations in mi- glia (nitrotyrosine).40,102,279 PARP inhibitors or tochondrial depolarization and calcium flux, revers- PARP-null mice had attenuated experimental DPN ible with growth factors including insulin, have been but the mechanism was thought to be through ar- described.91,92 Interestingly, insulin receptors have resting microangiopathy.121,152,155 been localized to mitochondria in sensory neu- The idea that free radical damage underlies neurons.219 ron damage in DPN has thus tentatively linked several known abnormalities in diabetes. For example, Neurodegeneration. DPN is a neurodegenerative excessive AR flux may deplete the antioxidant GSH disease. It is truly a “dying back” disorder with loss of within cells.239 AGEs may generate reactive oxygen distal axons before dropout of the entire neuron species and deplete GSH. Reactive oxygen and ni- tree. The idea that large-scale catastrophic apoptosis trergic species may also target mitochondria.239 of parent neuron soma or perikarya develops early in AGEs (advanced glycosylation endproducts) are the disease has not been supported by a number of products of the nonenzymatic reactions (glycation) investigations.40,176,280 Instead, there is gradual with- of glucose with amino groups on proteins through drawal of terminals of peripheral neurons from their the Maillard reaction.13,28 During an initial glycation target organs, with later gradual loss of parent reaction that is potentially reversible, intermediates perikarya (Fig. 3). This does differ somewhat from known as Schiff bases are formed and then con- the original concepts of “dying back” where the le- verted into an Amadori adduct known as fructose- sion was considered exclusively distal or axonal. In- lysine. Further irreversible glycation and oxidation stead, there are concurrent alterations of perikaryal generates glycoxidation products or AGEs, which are function and gene expression that help support dis- permanently deposited. Proteins are structurally tal axons. Evidence for the idea of a unique diabetic modified. The prototypes (although over a dozen neurodegeneration has not been easy to acquire, but ⑀ are known) measured most often are CML [N -(car- has emerged from more recent models. For exam- boxymethyl) lysine] and pentosidine. AGEs are thus ple, in long-term experimental STZ-induced diabe- formed by oxidative stress but also contribute to it by tes in rats, neither neuron nor sural nerve axon catalyzing lipid peroxidation. In turn, AGEs may dropout can be unequivocally documented using alter myelin, the extracellular matrix, vasoreactivity rigorous three-dimensional counting approaches.280 (quenching of nitric oxide), and nerve structural Axons and sensory neurons in ganglia atrophy, and proteins. They also directly act on specific receptors, there may be loss of very distal epidermal skin axons. one of which is RAGE (receptor of AGEs) identified Concurrent with atrophy is loss of neurofilaments by Schmidt et al.179,180 AGE–RAGE interaction then investing axons, an important determinant of their activates NF-␬B, a nuclear transcription factor that is caliber. These changes do take some time to de- critical to survival signaling of peripheral neurons,68 velop; they may not be evident, for example, after

152 Diabetes Mellitus and the PNS MUSCLE & NERVE August 2007 only 2 months of experimental diabetes but instead are found after 6–12 months. Both myelinated and unmyelinated axons exhibit this change, and its development correlates with downregulation of mRNA levels of all three neurofilament subunits in the cell body. Abnormalities of neurofilament phosphorylation, as well as deficits in neurofilament synthesis and investment, are also a feature of DPN.67 Paren- thetically, axon atrophy likely does not explain the much earlier slowing of conduction velocity demon- strable in the models, more likely to be related to changes in axon excitability, accumulation of intra- axonal sodium, and declines in Na/K ATPase activity.207 As discussed earlier, different models offer differ- ing insights into the pathogenesis of DPN. DPN in rats rendered diabetic with STZ is manifest as motor and sensory conduction velocity slowing, RICF, mild sural sensory axon atrophy, but relative preservation of neuron and axon numbers, even until late stages of the disease (e.g., 12 months). Conduction slowing develops early, within 2 weeks of the onset of hyperglycemia. Rises in “apoptosis or stress markers” without frank neuron loss may occur, and there are shifts of neuron sizes to smaller size categories with only an apparent loss of large neurons. Markers of cellular “stress” expression include activated caspase 3 and elevations of HSP-27 (heat shock protein-27), and PARP.40,98 Long-term models of STZ rat diabetes also highlight a number of molecular changes that pre- cede dropout. In addition to loss of neurofilaments, there is loss of t␣1-tubulin mRNA, the building block of microtubules; ␣CGRP, a peptide that has roles in pain neurotransmission and microvascular function; and GAP43/B50, the growth cone protein. Mole- cules altered in sensory neurons by diabetes also include declines in neurotrophin receptors, pep- ␤ tides, Nav1.8, PKC -II, and CREB with rises in ERK, JNK, p38, insulin receptor, RAGE, and Nav1.4, 1.6, and 1.9.231 Recent work using microarrays has also mapped early changes serially in rat DPN models for up to 8 weeks.166 Significant downregulation was noted for a ubiquinone subunit, MAG, and three as

4™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™ FIGURE 3. Hypothetical scheme of sensory neuron degeneration in progressive DPN. (A) The neuron structure is intact but there are early metabolic and electrophysiological changes such as slowed conduction velocity. (B) There is early axon atrophy and loss of epidermal axons. At this stage early changes in perikaryal gene expression occur. (C) There is loss of distal axon integrity and perikaryal atrophy. (D) There is further axon loss and loss of central branches of the sensory neuron. Perikarya at this stage have rises in apoptotic markers. (E) There is irretrievable neuron loss.

Diabetes Mellitus and the PNS MUSCLE & NERVE August 2007 153 yet unclarified molecules. There was upregulation of reinnervate collaterally may offer hope for patients tau, a sodium channel (type IV beta), a potassium- even with severe DPN. channel interacting protein, and two unclarified ad- From the transgenic mice studied to date with ditional molecules. Additional genes were catego- superimposed diabetes, several interesting observa- rized as being influenced by the duration of tions have emerged. In a mouse model in which diabetes, and changes were categorized for those axons were not invested with any neurofilaments, involved in glucose metabolism, synaptic vesicle for- electrophysiological and morphological features of mation, and extracellular matrix formation. Overall, DPN were accelerated.278 This is important because these findings are of considerable interest but re- it suggested that diabetic neurofilament abnormali- quire verification. A variety of approaches will be ties, rather than generating the changes of DPN, required that confirm protein expression and func- may instead protect axons to some extent from its tional significance in several long-term models and progression. DPN has been studied in thy1-YFP mice humans. Also, microarrays of ganglia do not strictly that have fluorescent axons: cutaneous axons from distinguish whether genes of interest are altered in these mice, evaluated noninvasively, progressively neurons or other cells within them, such as satellite disappeared during 6 months of experimental dia- 39 cells. betes. In contrast to the STZ rat model of DPN, STZ Experimental models of autonomic neuropathy diabetes in mice may be more akin to human disease in rodents have been extensively examined in the 181,183,185–195 despite lower polyol levels. The pancreatic ␤ cell Schmidt laboratory. A finding of partic- toxicity of STZ can be variable in mouse strains, with ular interest has been the discovery of axonal dystro- some resistance to its actions. Although this does not phic changes (neuroaxonal dystrophy), particularly indicate resistance to the development of DPN, dif- in proximal axons within sympathetic ganglia but also in distal axons. These dystrophic changes, also fering injection regimens of STZ may be needed to identified in humans, may represent accumulations induce comparable levels of hyperglycemia. Once of neurofilaments without frank cellular loss and are this is accomplished, however, the advantage of in keeping with a progressive neurodegenerative dis- mouse models is the ability to superimpose diabetes order.182,184 They may represent aberrant intragan- in mouse strains with specific molecular alterations. glionic sprouts. Abnormalities of diabetic autonomic STZ diabetes in simple Swiss–Webster mice or CD1 ganglia in models also correlated with changes in mice is associated with slowed motor and sensory gene expression. Overall, differential changes in- conduction velocity, loss of very distal sural axons, cluded transcripts involved in synapse and mito- and denervation of sweat glands.103,105 In models of chondrial structure and function, oxidative stress, sufficient duration (6–9 months), unlike STZ rats, and glycolysis.36 mice have overt and unequivocal loss of DRG sensory ϳ An important set of models of DPN in rats has neurons ( 20%). At the same timepoints, however, been discovered and generated by the Sima labora- motor neuron caliber and numbers are well pre- tory. Although more expensive and not studied served, highlighting how DPN can be selective. Fi- widely in other laboratories, these models have been nally, STZ mice with DPN also exhibit thinning of rigorously examined. The BB/W rat, a type 1 model, myelin sheaths, not generally observed in STZ rats, develops diabetes at about 70 days of age because of but a feature of human disease as well. Interestingly, an autoimmune-mediated destruction of pancreatic a small subset of mice given STZ recover pancreatic ␤ cells. These rats develop slowing of motor and ␤ cell function after a few months and revert to sensory conduction velocity as well as more marked euglycemia. This subset was of particular interest, structural alterations of axons (including loss) than since it offered some ideas on what humans under- STZ rats. A particular structural abnormality is axo- going successful islet ␤ cell transplant might experi- glial dysjunction, in which paranodal axon and ence. Electrophysiological changes, myelin thinning, Schwann cell contacts are disrupted.204 There are epidermal denervation, and sweat gland denervation associated electrophysiological and molecular changes were improved in the mice that recovered (Fig. 4). linked to this alteration.26,199,200,203,205,207,208 More re- Sensory neuron dropout, in the DRG, however, was cently, this BB/W model has been compared to a not reversed. These observations indicated that col- newer model of type II diabetes known as BB/Z. lateral reinnervation of target tissues, such as the Differences in this model include a slower evolution epidermis, from preserved sensory neurons can be of electrophysiological abnormalities, less axonal associated with clinical recovery. Strategies aimed at structural damage, no axoglial dysjunction, and preserving sensory neurons and allowing them to more targeting of Schwann cells and myelin.206,208

154 Diabetes Mellitus and the PNS MUSCLE & NERVE August 2007 FIGURE 4. Examples of epidermal axon innervation of the skin of footpads of mice with experimental DPN. (A) Diabetic (note loss of axons); (B) recovered from diabetes; (C) nondiabetic controls; (D) negative control (no primary antibody). Despite loss of ganglion sensory neurons in this model, collateral innervation permitted regrowth of epidermal axons. (Copyright © 2005 American Diabetes Association. From Diabetes, 2005;54:830–837. Reprinted with permission from The American Diabetes Association.)

Insulin Deficiency and Neuronal Growth Factors. To influence neuronal function, insulin receptors DPN may involve failed signaling from growth fac- signals share signaling cascades with neurotrophin tors that act on neurons21 but the results of clinical growth factors.31,99,220,255 Insulin-receptor expression trials using them have been disappointing.7,247 A is upregulated by injury, and its ligation by insulin number of neuronal growth factors share down- can promote regeneration of distal axon branches, stream signaling cascades that promote survival and even if given intrathecally.232,255 Insulin receptors outgrowth, but any presumed efficacy in treating have been identified on mitochondria, and insulin DPN would depend on whether the involved neu- can reverse their inappropriate depolarization dur- rons express relevant receptors. For example, in di- ing diabetes.91 Overall, then, insulin receptors are abetes it is unlikely that a single growth factor such as autophosphorylated by insulin binding, develop ty- NGF (nerve growth factor) protects all types of neu- rosine kinase activity, and signal through the IRS-1 rons against damage because most cases of DPN and 2 docking protein pathway.250 IRS-1 contains involve both large- and small-fiber involvement. Ele- multiple serine/threonine and tyrosine phosphory- gant work targeting DRG sensory neurons with her- lation sites that signal the survival kinases PI-3K-Akt, pes simplex-mediated transfer of NGF has, however, as well as Shc,Grb-2,S6 kinase, PKC⑀ kinase, MAP2 demonstrated substantial benefit in a mouse DPN kinase, Raf1 kinase, and c-fos.66,69,86 PI3K-PDK1-Akt model.78 block apoptosis by interacting with BAD, caspase-9, Insulin itself is a highly potent neuronal growth NF-␬B, and the forkhead transcription factor factor that is capable of acting on receptors ex- FKHRL130,62,146 (reviewed elsewhere70,84). Since in- pressed on most sensory neurons and on their axons. appropriate activation of NF-␬B by RAGE and

Diabetes Mellitus and the PNS MUSCLE & NERVE August 2007 155 caspases40 are implicated in diabetic neuron degen- Overall, these findings have suggested that nor- eration, this survival pathway is of particular interest. mal CSF insulin has an important role in the support The PI3k-Akt pathway is also locally activated in dis- of peripheral neurons. Among patients receiving in- tal axon terminals and growth cones to promote sulin therapy, insulin may yet be relevant to devel- outgrowth.70,139 Another important set of intracellu- opment of DPN because of inadequate doses or lar signals that influence survival, the MAPK pathway, is also altered in diabetes.140 Insulin is related to insulin-like growth factors (IGFs) in that they can cross-occupy each other’s receptors and share downstream survival transduction pathways. IGFs circulate but are also produced in glial cells,176,244,268 and IGF-1Rs are also expressed on Schwann cells,41 where they promote myelination.42 IGF-1 knockout mice have features similar to DPN.73 C-peptide, the cleaved fragment of the insulin prohormone, has also shown benefit in DPN in preclinical and clinical work.64,97,206,263,264 The mechanism may be through enhancement of insulin (or possibly IGF-1) signaling on neurons or axons. In rats with DPN, studied in our laboratory, low- dose intermittent insulin applied near nerve reversed sciatic motor conduction slowing indepen- dently of blood glucose levels; the contralateral nerve, exposed to the carrier, but not to insulin itself, applied identically in the same animals developed the expected conduction slowing.212 Similar reversal of conduction slowing, axonal atrophy, and epidermal axon loss in rats with DPN occurred when low-dose insulin was applied intrathecally31,233 (Fig. 5). Insulin applied intrathecally accessed motor and sensory neurons, yet had no impact on systemic hyperglycemia. Identical doses administered subcuta- neously over the back without intrathecal placement did not reverse DPN. In additional work, sequester- ing of endogenous local CSF insulin using an anti- insulin antibody generated conduction slowing and atrophy resembling DPN in nondiabetic rats, whereas unrelated antibody infusions had no impact.31

™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™™3 FIGURE 5. Intrathecal administration of insulin to rats is capable of accessing sensory neurons in lumbar ganglia (A). Insulin was labeled with the ﬂuorochrome FITC; sensory neurons in the ganglia are outlined by insulin labeling of their membranes. This route of administration reversed electrophysiological and structural features of experimental DPN without inﬂuencing systemic glucose levels. Identical section under light microscopy (B). Scale bar, 20 ␮m. Insulin can signal through receptors expressed on peripheral sensory neurons (C). Sections of mouse dorsal root ganglia are labeled with an antibody to the ␤ subunit of the mouse insulin receptor. Axons in the ganglia are also labeled. (A,B: Copyright © 2004 American Diabetes Association. From Diabe- tes, 2004;53:1824–1830. Reprinted with permission from The American Diabetes Association.)

156 Diabetes Mellitus and the PNS MUSCLE & NERVE August 2007 because the type of intermittent dosing applied is nonphysiological. Insulin may not be available to signal neurons partly protected by a blood–tissue barrier.69 Finally, resistance to insulin action, a defect in type II diabetes, could potentially operate at the neuronal level.

Impaired Regeneration. Diabetes mellitus imposes a “double hit” on the peripheral nervous system because it is associated with neuropathy with axon damage, and also impairs the ability of axons to regenerate. The topic of failed nerve regeneration in diabetes has been recently reviewed elsewhere.106 If therapy is identified that is capable of arresting diabetes-induced neurodegeneration, an important challenge will then be to support the regeneration of axons necessary to restore function. In addition, re- FIGURE 6. Tracings are CMAPs from tibial-innervated muscles in rats reared in cages with wire flooring. These cages generate covery from common diabetic focal neuropathies, pressure on distal tibial nerve branches in the hindpaw of the rats. such as CTS or UNE, may critically depend on how Longer duration of housing, greater weight, and diabetes wors- regeneration fares. Potential mechanisms of im- ened the neuropathy. Note the prominent dispersion from distal paired regeneration in diabetes, briefly listed, in- demyelination of the tibial nerves. (Reproduced with permission from Zochodne DW, et al. Brain Res 1995;698:130–136.) clude impaired sprouting of injured axons, a delay in how axons alter their synthesis of regeneration proteins in response to injury, slowed Wallerian and TRANSLATION THERAPY Wallerian-like degeneration including macrophage invasion of the distal stump of injured nerves, accel- Primary Forms of Therapy. No form of therapy in erated retrograde loss of neurons, alterations in ad- DPN has been identified that provides unequivocal, hesive extracellular matrix, failure of microvascular safe, and effective stabilization or reversal of the condition. There have been many trials, most pro- plasticity after injury necessary to support regrowth, viding only modest disease stability or no improve- altered elaboration of local NO for vasodilation and ment. The reasons for this disappointing record can myelin clearance, failure of insulin or IGF-1 support, be identified both in the preclinical work and choice and failed wound healing of neighboring tissues nec- of agent (e.g., short-duration conduction velocity essary to support axons. models, use of NGF despite its targeting of only small There has not been an extensive literature on the sensory axons or autonomic fibers) but also in trial development of mononeuropathies in experimental design and endpoints. Intensive control of hypergly- diabetes. Rats reared on wire cage flooring, instead cemia slows the progression of DPN and prevents its of sawdust or shaving covered plastic flooring, de- appearance in both type I and type II diabetic pa- velop electrophysiological evidence of an entrap- tients, discussed above. Successful pancreas trans- ment neuropathy of their very distal paw motor plantation associated with euglycemia has been asso- nerves (Fig. 6). Diabetics had earlier involvement, ciated with a slow and gradual improvement in similar to the predisposition to entrapment observed DPN.108,148 in humans.276 Nerves in experimental diabetes are Results of a number of recent trials are summa- also more susceptible to ischemia than those of non- rized in Table 1. Given the lack of current agree- diabetics, as demonstrated in nerve ischemia models ment about their benefits, none at this time are using microspheres or exposure to the vasoconstric- classified as providing level 1 evidence. Several tor endothelin.150,267,269 In our laboratory we have classes of agents have been examined. There has also examined a model of dorsal root ganglia isch- been, for example, several decades of experience emia from local endothelin exposure.254 The isch- with ARIs, aimed to prevent excessive sorbitol flux in emia induced neuron necrosis and apoptosis, intra- nerve. A meta-analysis has identified some improve- ganglionic axonal degeneration with neuron cell ments in conduction velocity with a number of body responses, and downstream degeneration of ARIs.4 These trials will not be reviewed in depth here sensory axons. Diabetics were more sensitive to dam- as they have been reviewed elsewhere.18,19,151 In gen- age. eral, ARI trials have been confounded by systemic

Diabetes Mellitus and the PNS MUSCLE & NERVE August 2007 157 Table 1. Some recent randomized* double-blinded controlled clinical trials for diabetic polyneuropathy. Agent Duration (n) Mechanism Impact (modality) Reference Therapy for neuropathy Ranirestat 12 wks (86) ARI Improved (TNS, VPT, CV) 23 Fidarestat* 52 wks (279) ARI Improved (F wave CV, 90 symptoms) Ruboxistaurin mesylate 52 wks (205) PKC␤ inhibitor Benefit in Subgroup only 242 Alpha-lipoic acid IVa 3 wks (1258) Anti-oxidant Improved (symptoms, 265 sensation) MIRE 4 wks (39) ? None 44 QR-333 4 wks (34) Topical flavanoid Improved (symptoms, 236 (?ARI), ascorbyl QOL) palmitate, Vitamin D3 rhBDNF 12 wks (30) Growth factor No benefit 247 Acetyl-L-carnitine 52 wks (1257) ?multiple Improved (morphometry, 202 VPT, ?pain, not CV) C-peptide 12 weeks (49) ?insulin sensitizer Improved (sensory CV, 64 VPT) Gamma-linolenic acid 52 wks (111) Unknown Improved (conduction, 101 sensation, strength) rhNGF 52 wks (1019) Growth factor No benefit 7 Pain Therapy Duloxetine† 12 wks (348) SSNRI 2pain scales 169 Duloxetine† 12 wks (457) SSNRI 2pain scales 77 Oxcarbazepine 16 wks (146) AED 2VAS 48 Oxcarbazepine 16 wks (347) AED Primary endpoint 14 negative CR Oxycodone† 6 wks (159) Opioid 2pain scales 76 Zonisamide 12 wks (42) AED none 8 Lamotrigine† 6 wks (59) AED 2pain scales 63 Pregabalin† 12 wks (42) Ca channel blocker 2pain scales but lumped 72 with nonDPN Pregabalin† 6 wk (246) Ca channel blocker 2pain scales 171 Pregabalin† 8 wks (146) Ca channel blocker 2pain scales 173 Pregabalin† 5 wks (338) Ca channel blocker 2pain scales 119 Sodium valproate 12 wks (43) AED 2pain scales 112 Sodium valproate 6 wks (31) AED No benefit 160 Venlafaxine 6 wks (244) SSNRI 2pain scales 174 Topiramateb 18–22 wks (1259) AED No benefit 225 Topiramate 12 wks (323) AED 2pain scales 170 Isorbide dinitrate spray 4 wks (22) NO donor 2pain scales 260 Tramadol† 42 days (131) Weak opioid, SSRI 2pain scales 85 Morphine,† gabapentin 5 wks (35 DPN) Opioid and AED 2pain scales 75 or both Paroxetine 2 wks (29) SSRI 2pain scales 211

AED antiepileptic drug; ARI, aldose reductase inhibitor; CV conduction velocity; MIRE monochromatic infrared energy; n total number of patients in trial, all groups; QOL quality of life; SSNRI selective serotonin and norepinephrine reuptake inhibitor; VAS visual analogue pain scale; VPT, vibration perception threshold. *This trial was not labeled as randomized; †Level 1 evidence of efficacy. aThis was a meta-analysis. bEndpoint criteria for this negative study have been criticized. An open-label trial showing benefit has been published in abstract form.243 toxicity (renal, hepatic, skin), lack of potency, lack of the results were disappointing, with benefits mainly penetration into the endoeneurium, and problem- in a subset of patients with less severe DPN.242 Two atic trial design (selection, duration, size). Only one antioxidants have been identified with some bene- ARI, epalrestat, is currently marketed in Japan.9 The fits: gamma linolenic acid101 or evening primrose oil most recent ARIs used in clinical trials have been and alpha lipoic acid.265 The use of neurotrophic fidarestat and ranirestat. family growth factors has not been beneficial in trials Roboxistaurin mesylate was used as a PKC␤ in- of rhNGF (Phase III) and rhBDNF.7,247 C-peptide hibitor in 205 patients in a recent Phase II trial but administered to human type I diabetic patients was

158 Diabetes Mellitus and the PNS MUSCLE & NERVE August 2007 associated with improved sensory conduction veloc- application of local anesthetics or capsaicin have had ity and VPT but no change in heat or cold sensation mixed benefits. 64 after 12 weeks in a small study of 49 patients. Supported by external grants from the Canadian Diabetes Asso- Multiple surgical decompression of peripheral ciation and Canadian Institutes of Health Research (CIHR). nerves in diabetic patients is of no proven benefit D.W.Z. is an Alberta Heritage Foundation for Medical Research and should be discouraged.38 Scientist. Brenda Boake provided expert secretarial assistance. Dr. For DLSP, effective therapy has not been estab- Aaron Vinik provided valuable thoughts on therapy. lished in published randomized controlled trials. A randomized controlled trial of methylprednisolone was reported in abstract form as improving symp- REFERENCES toms (especially pain) but not disability in DLSP.51 1. Abbott CA, Carrington AL, Ashe H, Bath S, Every LC, Grif- fiths J, et al. The North-West Diabetes Foot Care Study: Despite claims otherwise, no definite role for intra- incidence of, and risk factors for, new diabetic foot ulcer- venous gamma globulin has been demonstrated.115 ation in a community-based patient cohort. Diabet Med Our own experience in a small series of three pa- 2002;19:377–384. 2. Ahlgren SC, Levine JD. Protein kinase C inhibitors decrease tients with DLSP was disappointing, with overt pro- hyperalgesia and C-fiber hyperexcitability in the streptozo- gression despite immunosuppression and IVIG.275 tocin-diabetic rat. J Neurophysiol 1994;72:684–692. 3. Aiello LP, Avery RL, Arrigg PG, Keyt BA, Jampel HD, Shah ST, et al. Vascular endothelial growth factor in ocular fluid Therapy for Neuropathic Pain. It is hoped that ap- of patients with diabetic retinopathy and other retinal disor- proaches toward the reversal of DPN would concur- ders. N Engl J Med 1994;331:1480–1487. 4. Airey M, Bennett C, Nicolucci A, Williams R. Aldose reduc- rently improve neuropathic pain. As reversal DPN tase inhibitors for the prevention and treatment of diabetic therapy is currently not available, however, specific peripheral neuropathy. Cochrane Database Syst Rev 2000; therapy for neuropathic pain is required in many CD002182. 5. Anonymous. Report and recommendations of the San An- patients. Most currently available treatments are tonio conference on diabetic neuropathy. Consensus state- used for all forms of peripheral neuropathic pain ment. Diabetes 1988;37:1000–1004. and will only be briefly summarized here (for a 6. Anonymous. In: Dyck PJ, Thomas PK, editors. Diabetic neu- 240 ropathy. Toronto: WB Saunders; 1999. comprehensive and recent review, see Vinik ). Ta- 7. Apfel SC, Schwartz S, Adornato BT, Freeman R, Biton V, ble 1 lists some recent double-blind randomized con- Rendell M, et al. Efficacy and safety of recombinant human trolled clinical trials for newer agents, including nerve growth factor in patients with diabetic polyneuropathy: A randomized controlled trial. rhNGF Clinical Investi- those with level 1 evidence of benefit. Those treat- gator Group. JAMA 2000;284:2215–2221. ments not listed, but with level 1 evidence to support 8. Atli A, Dogra S. Zonisamide in the treatment of painful their use, include tricyclic antidepressants, mexili- diabetic neuropathy: a randomized, double-blind, placebo- controlled pilot study. Pain Med 2005;6:225–234. tine, and anti-epileptic agents including phenytoin 9. Baba M, Kimura K, Suda T, Yagihashi S. Three-year inhibi- and carbamazepine.63,137,138,159,177 Gabapentin, a tion of aldose reductase on development of symptomatic weak anti-epileptic agent, also with level 1 evidence neuropathy in diabetic patients. J Peripher Nerv Syst 2006; 11:176–178. for efficacy, is a safe treatment for neuropathic pain 10. Backonja M, Beydoun A, Edwards KR, Schwartz SL, Fonseca with a wide therapeutic range.10,11 Many clinicians V, Hes M, et al. Gabapentin for the symptomatic treatment consider it first-line therapy. In patients with signifi- of painful neuropathy in patients with diabetes mellitus. A randomized controlled trial. JAMA 1998;280:1831–1836. cant renal failure from diabetic nephropathy, the 11. Backonja MM. Gabapentin monotherapy for the symptom- dose of gabapentin should be titrated to a lower atic treatment of painful neuropathy: a multicenter, double- therapeutic range, as it is excreted renally. Opioids blind, placebo-controlled trial in patients with diabetes mellitus. Epilepsia 1999;40(Suppl 6):S57–S59. including oxycontin have level 1 evidence to support 12. Barohn RJ, Sahenk Z, Warmolts JR, Mendell JR. The Bruns- their use in diabetic neuropathic pain.76,210 Exacer- Garland syndrome (diabetic amyotrophy). Revisited 100 bated constipation and cognitive dysfunction are years later. Arch Neurol 1991;48:1130–1135. 13. Baynes JW. Glycation and advanced glycation reactions. In: side effects that limit their use. The combination of Gries FA, Cameron NE, Low PA, Ziegler D, editors. Text- morphine and gabapentin had synergistic benefits in book of diabetic neuropathy. New York: Thieme; 2003. p 105–112. a group of patients that included those with painful 14. Beydoun A, Shaibani A, Hopwood M, Wan Y. Oxcarbazepine DPN.75 Tramadol is a weak opioid and SSRI (selec- in painful diabetic neuropathy: results of a dose-ranging tive serotonin reuptake inhibitor) that improves study. Acta Neurol Scand 2006;113:395–404. 85,89,210 15. Biessels GJ, Staekenborg S, Brunner E, Brayne C, Scheltens neuropathic pain. Newer agents also hold P. Risk of dementia in diabetes mellitus: a systematic review. some promise for diabetic neuropathic pain and Lancet Neurol 2006;5:64–74. include pregabalin, and duloxetine, a serotonin and 16. Bolton CF, Remtulla H, Toth B, Bernardi L, Lindsay RM, Maryniak O, et al. Distinctive electrophysiological features of norepinephrine uptake inhibitor. The role for can- denervated muscle in uremic patients. J Clin Neurophysiol nabanoids is as yet undetermined and local skin 1997;14:539–542.

Diabetes Mellitus and the PNS MUSCLE & NERVE August 2007 159 17. Borghini I, Ania-Lahuerta A, Regazzi R, Ferrari G, Gjinovci 37. Ceriello A, Giugliano D. Oxidative stress and diabetic com- A, Wollheim CB, et al. Alpha, beta I, beta II, delta, and plications. In: Alberti KGGM, Zimmet P, DeFronzo RA, Keen epsilon protein kinase C isoforms and compound activity in HH, editors. International textbook of diabetes mellitus. the sciatic nerve of normal and diabetic rats. J Neurochem Chichester, UK: John Wiley & Sons; 1997. p 1453–1461. 1994;62:686–696. 38. Chaudhry V, Stevens JC, Kincaid J, So YT. Practice advisory: 18. Boulton AJ, Malik RA, Arezzo JC, Sosenko JM. Diabetic utility of surgical decompression for treatment of diabetic somatic neuropathies. Diabetes Care 2004;27:1458–1486. neuropathy: report of the Therapeutics and Technology 19. Boulton AJ, Vinik AI, Arezzo JC, Bril V, Feldman EL, Free- Assessment Subcommittee of the American Academy of Neu- man R, et al. Diabetic neuropathies: a statement by the rology. Neurology 2006;66:1805–1808. American Diabetes Association. Diabetes Care 2005;28:956– 39. Chen YS, Chung SS, Chung SK. Noninvasive monitoring of 962. diabetes-induced cutaneous nerve fiber loss and hypoalgesia 20. Brett FM, Kalichman MW, Calcutt NA, Mizisin AP. Effects of in thy1-YFP transgenic mice. Diabetes 2005;54:3112–3118. seven days of galactose feeding and aldose reductase inhibi- 40. Cheng C, Zochodne DW. Sensory neurons with activated tion on mast cells and vessel morphometry in rat sciatic caspase-3 survive long-term experimental diabetes. Diabetes nerve. J Neurol Sci 1996;141:6–12. 2003;52:2363–2371. 21. Brewster WJ, Fernyhough P, Diemel LT, Mohiuddin L, Tom- 41. Cheng HL, Randolph A, Yee D, Delafontaine P, Tennekoon linson DR. Diabetic neuropathy, nerve growth factor and G, Feldman EL. Characterization of insulin-like growth fac- other neurotrophic factors. Trends Neurosci 1994;17:321– tor-I and its receptor and binding proteins in transected 325. nerves and cultured Schwann cells. J Neurochem 1996;66: 22. Bril V. Electrophysiological testing. In: Gries FA, Cameron 525–536. NE, Low PA, Ziegler D, editors. Textbook of diabetic neu- 42. Cheng HL, Russell JW, Feldman EL. IGF-I promotes periph- ropathy. New York: Thieme; 2003. p 177–184. eral nervous system myelination. Ann N Y Acad Sci 1999;883: 23. Bril V, Buchanan RA. Long-term effects of ranirestat (AS- 124–130. 3201) on peripheral nerve function in patients with diabetic 43. Cherian PV, Kamijo M, Angelides KJ, Sima AA. Nodal Naϩ- sensorimotor polyneuropathy. Diabetes Care 2006;29:68–72. channel displacement is associated with nerve-conduction 24. Bril V, Ellison R, Ngo M, Bergstrom B, Raynard D, Gin H. slowing in the chronically diabetic BB/W rat: prevention by Electrophysiological monitoring in clinical trials. Roche aldose reductase inhibition. J Diabetes Complications 1996; Neuropathy Study Group. Muscle Nerve 1998;21:1368–1373. 10:192–200. 25. Bril V, Werb MR, Greene DA, Sima AAF. Single-fiber elec- 44. Clifft JK, Kasser RJ, Newton TS, Bush AJ. The effect of tromyography in diabetic peripheral polyneuropathy. Mus- monochromatic infrared energy on sensation in patients cle Nerve 1996;19:2–9. with diabetic peripheral neuropathy: a double-blind, place- 26. Brismar T, Sima AA. Changes in nodal function in nerve bo-controlled study. Diabetes Care 2005;28:2896–2900. fibres of the spontaneously diabetic BB-Wistar rat: potential 45. Daube JR. Electrophysiological testing in diabetic neuropa- clamp analysis. Acta Physiol Scand 1981;113:499–506. thy. In: Dyck PJ, Thomas PK, editors. Diabetic neuropathy. 27. Brown MJ, Asbury AK. Diabetic neuropathy. Ann Neurol Philadelphia: WB Saunders; 1999. p 222–238. 1984;15:2–12. 46. Diabetes Control & Complications Trial Research Group. 28. Brownlee M, Cerami A, Vlassara H. Advanced glycosylation The effect of intensive treatment of diabetes on the devel- end products in tissue and the biochemical basis of diabetic. opment and progression of long-term complications in in- N Engl J Med 1988;318:1315–1321. sulin-dependent diabetes mellitus. N Engl J Med 1993;329: 29. Brownlee M, Vlassara H, Kooney A, Ulrich P, Cerami A. 977–986. Aminoguanidine prevents diabetes-induced arterial wall pro- 47. Diederich D, Skopec J, Diederich A, Dai FX. Endothelial tein cross-linking. Science 1986;232:1629–1632. dysfunction in mesenteric resistance arteries of diabetic rats: 30. Brunet A, Datta SR, Greenberg ME. Transcription-depen- role of free radicals. Am J Physiol 1994;266:H1153–H1161. dent and -independent control of neuronal survival by the PI3K-Akt signaling pathway. Curr Opin Neurobiol 2001;11: 48. Dogra S, Beydoun S, Mazzola J, Hopwood M, Wan Y. Oxcar- 297–305. bazepine in painful diabetic neuropathy: a randomized, pla- 31. Brussee V, Cunningham FA, Zochodne DW. Direct insulin cebo-controlled study. Eur J Pain 2005;9:543–554. signaling of neurons reverses diabetic neuropathy. Diabetes 49. Durante W, Sen AK, Sunahara FA. Impairment of endothe- 2004;53:1824–1830. lium-dependent relaxation in aortae from spontaneously 32. Bucala R, Tracey KJ, Cerami A. Advanced glycosylation prod- diabetic rats. Br J Pharmacol 1988;94:463–468. ucts quench nitric oxide and mediate defective endotheli- 50. Dyck JB, Dyck PJ. Diabetic polyneuropathy. In: Dyck PJ, um-dependent vasodilatation in experimental diabetes. Thomas PK, editors. Diabetic neuropathy. Toronto: WB J Clin Invest 1991;87:432–438. Saunders; 1998. p 255–278. 33. Calver A, Collier J, Vallance P. Inhibition and stimulation of 51. Dyck JB, O’Brien PC, Bosch EP, Grant I, Burns TM, Wind- nitric oxide synthesis in the human forearm arterial bed of ebank AJ, et al. Results of a controlled trial of IV methyl- patients with insulin-dependent diabetes. J Clin Invest 1992; prednisolone in diabetic lumbosacral radiculoplexus neu- 90:2548–2554. ropathy (DLRPN): a preliminary indication of efficacy. J 34. Cameron NE, Cotter MA, Dines KC, Maxfield EK. Pharma- Peripher Nerv Syst 2005:10(Suppl 1). cological manipulation of vascular endothelium function in 52. Dyck PJ. Hypoxic neuropathy: does hypoxia play a role in non-diabetic and streptozotocin-diabetic rats: effects on diabetic neuropathy? The 1988 Robert Wartenberg lecture. nerve conduction, hypoxic resistance and endoneurial cap- Neurology 1989;39:111–118. illarization. Diabetologia 1993;36:516–522. 53. Dyck PJ, Davies JL, Wilson DM, Service FJ, Melton LJ, 35. Cameron NE, Cotter MA, Low PA. Nerve blood flow in early O’Brien PC. Risk factors for severity of diabetic polyneurop- experimental diabetes in rats: relation to conduction defi- athy: intensive longitudinal assessment of the Rochester Di- cits. Am J Physiol 1991;261:E1–E8. abetic Neuropathy Study cohort. Diabetes Care 1999;22: 36. Carroll SL, Byer SJ, Dorsey DA, Watson MA, Schmidt RE. 1479–1486. Ganglion-specific patterns of diabetes-modulated gene ex- 54. Dyck PJ, Dyck PJ, Larson TS, O’Brien PC, Velosa JA. Patterns pression are established in prevertebral and paravertebral of quantitative sensation testing of hypoesthesia and hyper- sympathetic ganglia prior to the development of neuroax- algesia are predictive of diabetic polyneuropathy: a study of onal dystrophy. J Neuropathol Exp Neurol 2004;63:1144– three cohorts. Nerve Growth Factor Study Group. Diabetes 1154. Care 2000;23:510–517.

160 Diabetes Mellitus and the PNS MUSCLE & NERVE August 2007 55. Dyck PJ, Giannini C. Pathologic alterations in the diabetic 74. Giannini C, Dyck PJ. Pathological alterations in human dia- neuropathies of humans: a review. J Neuropathol Exp Neu- betic polyneuropathy. In: Dyck PJ, Thomas PK, editors. Di- rol 1996;55:1181–1193. abetic neuropathy. Philadelphia: WB Saunders; 1999. p 279– 56. Dyck PJ, Hansen S, Karnes J, O’Brien P, Yasuda H, Wind- 296. ebank A, et al. Capillary number and percentage closed in 75. Gilron I, Bailey JM, Tu D, Holden RR, Weaver DF, Houlden human diabetic sural nerve. Proc Natl Acad SciUSA RL. Morphine, gabapentin, or their combination for neuro- 1985;82:2513–2517. pathic pain. N Engl J Med 2005;352:1324–1334. 57. Dyck PJ, Karnes JL, O’Brien P, Okazaki H, Lais A, Engelstad 76. Gimbel JS, Richards P, Portenoy RK. Controlled-release oxy- J. The spatial distribution of fiber loss in diabetic polyneu- codone for pain in diabetic neuropathy: a randomized con- ropathy suggests ischemia. Ann Neurol 1986;19:440–449. trolled trial. Neurology 2003;60:927–934. 58. Dyck PJ, Kratz KM, Karnes JL, Litchy WJ, Klein R, Pach JM, 77. Goldstein DJ, Lu Y, Detke MJ, Lee TC, Iyengar S. Duloxetine et al. The prevalence by staged severity of various types of vs. placebo in patients with painful diabetic neuropathy. Pain diabetic neuropathy, retinopathy, and nephropathy in a 2005;116:109–118. population-based cohort: The Rochester Diabetic Neuropa- 78. Goss JR, Goins WF, Lacomis D, Mata M, Glorioso JC, Fink thy Study. Neurology 1993;43:817–824. DJ. Herpes simplex-mediated gene transfer of nerve growth 59. Dyck PJ, Lais A, Karnes JL, O’Brien P, Rizza R. Fiber loss is factor protects against peripheral neuropathy in streptozo- primary and multifocal in sural nerves in diabetic polyneu- tocin-induced diabetes in the mouse. Diabetes ropathy. Ann Neurol 1986;19:425–439. 2002;51:2227–2232. 60. Dyck PJ, Norell JE, Dyck PJ. Methylprednisolone may im- 79. Greenbaum D, Richardson PC, Salmon MV, Urich H. Patho- prove lumbosacral radiculoplexus neuropathy. Can J Neurol logical observations on six cases of diabetic neuropathy. Sci 2001;28:224–227. Brain 1964;87:201–214. 61. Dyck PJ, Windebank AJ. Diabetic and nondiabetic lumbosa- 80. Greene DA, Lattimer SA, Sima AA. Sorbitol, phosphoinositi- cral radiculoplexus neuropathies: new insights into patho- des, and sodium-potassium-ATPase in the pathogenesis of physiology and treatment. Muscle Nerve 2002;25:477–491. diabetic complications. N Engl J Med 1987;316:599–606. 62. Edstrom A, Ekstrom PA. Role of phosphatidylinositol 3-ki- 81. Gries FA, Cameron NE, Low PA, Ziegler D. Textbook of nase in neuronal survival and axonal outgrowth of adult diabetic neuropathy. New York: Thieme; 2003. mouse dorsal root ganglia explants. J Neurosci Res 2003;74: 82. Gupta S, Sussman I, McArthur CS, Tornheim K, Cohen RA, 726–735. Ruderman NB. Endothelium-dependent inhibition of Na(ϩ)-Kϩ ATPase activity in rabbit aorta by hyperglycemia. 63. Eisenberg E, Lurie Y, Braker C, Daoud D, Ishay A. Lam- Possible role of endothelium-derived nitric oxide. J Clin otrigine reduces painful diabetic neuropathy: a randomized, Invest 1992;90:727–732. controlled study. Neurology 2001;57:505–509. 83. Halestrap AP. Calcium, mitochondria and reperfusion in- 64. Ekberg K, Brismar T, Johansson BL, Jonsson B, Lindstrom P, jury: a pore way to die. Biochem Soc Trans 2006;34:232–237. Wahren J. Amelioration of sensory nerve dysfunction by 84. Hanada M, Feng J, Hemmings BA. Structure, regulation and C-peptide in patients with type 1 diabetes. Diabetes 2003;52: function of PKB/AKT—a major therapeutic target. Biochim 536–541. Biophys Acta 2004;1697:3–16. 65. Ewing DJ, Boland O, Neilson JM, Cho CG, Clarke BF. Auto- 85. Harati Y, Gooch C, Swenson M, Edelman S, Greene D, nomic neuropathy, QT interval lengthening, and unex- Raskin P, et al. Double-blind randomized trial of tramadol pected deaths in male diabetic patients. Diabetologia 1991; for the treatment of the pain of diabetic neuropathy. Neu- 34:182–185. rology 1998;50:1842–1846. 66. Fadool DA, Tucker K, Phillips JJ, Simmen JA. Brain insulin 86. Heidenreich KA. Insulin and IGF-I receptor signaling in receptor causes activity-dependent current suppression in cultured neurons. Ann N Y Acad Sci 1993;692:72–88. the olfactory bulb through multiple phosphorylation of 87. Hill MA, Meininger GA, Larkins RG. Alterations in micro- Kv1.3. J Neurophysiol 2000;83:2332–2348. vascular reactivity in experimental diabetes mellitus: contri- 67. Fernyhough P, Gallagher A, Averill SA, Priestley JV, Houn- bution of the endothelium? In: Molinatti GM, Bar RS, Bel- som L, Patel J, et al. Aberrant neurofilament phosphoryla- fiore F, Porta M, editors. Endothelial cell function in tion in sensory neurons of rats with diabetic neuropathy. diabetic microangiopathy: problems in methodology and Diabetes 1999;48:881–889. clinical aspects. Basel: Karger; 1990. p 118–126. 68. Fernyhough P, Smith DR, Schapansky J, Van Der Ploeg R, 88. Holland NR, Crawford TO, Hauer P, Cornblath DR, Griffin Gardiner NJ, Tweed CW, et al. Activation of nuclear factor- JW, McArthur JC. Small-fiber sensory neuropathies: clinical kappaB via endogenous tumor necrosis factor alpha regu- course and neuropathology of idiopathic cases. Ann Neurol lates survival of axotomized adult sensory neurons. J Neuro- 1998;44:47–59. sci 2005;25:1682–1690. 89. Hollingshead J, Duhmke RM, Cornblath DR. Tramadol for 69. Folli F, Bonfanti L, Renard E, Kahn CR, Merighi A. Insulin neuropathic pain. Cochrane Database Syst Rev 2006;3: receptor substrate-1 (IRS-1) distribution in the rat central CD003726. nervous system. J Neurosci 1994;14:6412–6422. 90. Hotta N, Toyota T, Matsuoka K, Shigeta Y, Kikkawa R, 70. Franke TF, Hornik CP, Segev L, Shostak GA, Sugimoto C. Kaneko T, et al. Clinical efficacy of fidarestat, a novel aldose PI3K/Akt and apoptosis: size matters. Oncogene 2003;22: reductase inhibitor, for diabetic peripheral neuropathy: a 8983–8998. 52-week multicenter placebo-controlled double-blind paral- 71. Freeman R. Autonomic testing. In: Bolton CF, Brown WF, lel group study. Diabetes Care 2001;24:1776–1782. Aminoff MJ, editors. Neuromuscular function and disease. 91. Huang TJ, Price SA, Chilton L, Calcutt NA, Tomlinson DR, Toronto: WB Saunders; 2002. p 483–500. Verkhratsky A, et al. Insulin prevents depolarization of the 72. Freynhagen R, Strojek K, Griesing T, Whalen E, Balkenohl mitochondrial inner membrane in sensory neurons of type 1 M. Efficacy of pregabalin in neuropathic pain evaluated in a diabetic rats in the presence of sustained hyperglycemia. 12-week, randomised, double-blind, multicentre, placebo- Diabetes 2003;52:2129–2136. controlled trial of flexible- and fixed-dose regimens. Pain 92. Huang TJ, Sayers NM, Fernyhough P, Verkhratsky A. Diabe- 2005;115:254–263. tes-induced alterations in calcium homeostasis in sensory 73. Gao WQ, Shinsky N, Ingle G, Beck K, Elias KA, Powell- neurones of streptozotocin-diabetic rats are restricted to Braxton L. IGF-1 deficient mice show reduced peripheral lumbar ganglia and are prevented by neurotrophin-3. Dia- nerve conduction velocities and decreased axonal diameters betologia 2002;45:560–570. and respond to exogenous IGF-1 treatment. J Neurobiol 93. Ido Y, Chang K, LeJeune W, Tilton RG, Monafo WW, Wil- 1999;39:142–152. liamson JR. Diabetes impairs sciatic nerve hyperemia in-

Diabetes Mellitus and the PNS MUSCLE & NERVE August 2007 161 duced by surgical trauma: implications for diabetic neurop- 115. Krendel DA, Costigan DA, Hopkins LC. Successful treatment athy. Am J Physiol 1997;273:E174–E184. of neuropathies in patients with diabetes mellitus. Arch Neu- 94. Ishii H, Jirousek MR, Koya D, Takagi C, Xia P, Clermont A, rol 1995;52:1053–1061. et al. Amelioration of vascular dysfunctions in diabetic rats 116. Lauria G, Cornblath DR, Johansson O, McArthur JC, Mell- by an oral PKC beta inhibitor. Science 1996;272:728–731. gren SI, Nolano M, et al. EFNS guidelines on the use of skin 95. Johnson PC, Doll SC, Cromey DW. Pathogenesis of diabetic biopsy in the diagnosis of peripheral neuropathy. Eur J Neu- neuropathy. Ann Neurol 1986;19:450–457. rol 2005;12:747–758. 96. Judzewitsch RG, Jaspan JB, Polonsky KS, Weinberg CR, Hal- 117. Lawrence E, Brain SD. Altered microvascular reactivity to ter JB, Halar E, et al. Aldose reductase inhibition improves endothelin-1, endothelin-3 and NG-nitro-L-arginine methyl nerve conduction velocity in diabetic patients. N Engl J Med ester in streptozotocin-induced diabetes mellitus. Br J Phar- 1983;308:119–125. macol 1992;106:1035–1040. 97. Kamiya H, Zhang W, Sima AA. C-peptide prevents nocicep- 118. Lee GH, Proenca R, Montez JM, Carroll KM, Darvishzadeh tive sensory neuropathy in type 1 diabetes. Ann Neurol JG, Lee JI, et al. Abnormal splicing of the leptin receptor in 2004;56:827–835. diabetic mice. Nature 1996;379:632–635. 98. Kamiya H, Zhangm W, Sima AA. Apoptotic stress is counter- 119. Lesser H, Sharma U, LaMoreaux L, Poole RM. Pregabalin balanced by survival elements preventing programmed cell relieves symptoms of painful diabetic neuropathy: a random- death of dorsal root ganglions in subacute type 1 diabetic ized controlled trial. Neurology 2004;63:2104–2110. BB/Wor rats. Diabetes 2005;54:3288–3295. 120. Li F, Abatan OI, Kim H, Burnett D, Larkin D, Obrosova IG, 99. Karagiannis SN, King RH, Thomas PK. Colocalisation of et al. Taurine reverses neurological and neurovascular defi- insulin and IGF-1 receptors in cultured rat sensory and cits in Zucker diabetic fatty rats. Neurobiol Dis 2006;22:669– sympathetic ganglion cells. J Anat 1997;191:431–440. 676. 100. Karasik A, Hattori M. Use of animal models in the study of 121. Li F, Drel VR, Szabo C, Stevens MJ, Obrosova IG. Low-dose diabetes. In: Kahn CR, Weir GC, editors. Joslin’s diabetes poly(ADP-ribose) polymerase inhibitor-containing combina- mellitus. Malvern, PA: Lea & Febiger; 1994. p 317–350. tion therapies reverse early peripheral diabetic neuropathy. 101. Keen H, Payan J, Allawi J, Walker J, Jamal GA, Weir AI, et al. Diabetes 2005;54:1514–1522. Treatment of diabetic neuropathy with gamma-linolenic 122. Li X-G, Zochodne DW. Microvacuolar neuronopathy is a acid. The gamma-Linolenic Acid Multicenter Trial Group. post-mortem artifact of sensory neurons. J Neurocytol 2003; Diabetes Care 1993;16:8–15. 32:393–398. 102. Kennedy JM, Rubin I, Lauritzen M, Zochodne DW. Injury- 123. Llewelyn JG, Thomas PK, Gilbey SG, Watkins PJ, Muddle JR. induced nitric oxide synthase activity in regenerating dia- Pattern of myelinated fibre loss in the sural nerve in neu- betic peripheral nerves. Soc Neurosci Abstr 2002;31:449. ropathy related to type 1 (insulin-dependent) diabetes. Dia- 103. Kennedy JM, Zochodne DW. The regenerative deficit of betologia 1988;31:162–167. peripheral nerves in experimental diabetes: its extent, tim- 124. Low PA, Benrud-Larson LM, Sletten DM, Opfer-Gehrking ing and possible mechanisms. Brain 2000;123:2118–2129. TL, Weigand SD, O’Brien PC, et al. Autonomic symptoms 104. Kennedy JM, Zochodne DW. Influence of experimental di- and diabetic neuropathy: a population-based study. Diabetes abetes on the microcirculation of injured peripheral nerve. Care 2004;27:2942–2947. Functional and morphological aspects. Diabetes 2002;51: 125. Low PA, Lagerlund TD, McManis PG. Nerve blood flow and 2233–2240. oxygen delivery in normal, diabetic, and ischemic neuropa- 105. Kennedy JM, Zochodne DW. Experimental diabetic neurop- thy. Int Rev Neurobiol 1989;31:355–438. athy and spontaneous recovery: is there irreparable damage? 126. Low PA, Nickander KK. Oxygen free radical effects in sciatic Diabetes 2005;54:830–837. nerve in experimental diabetes. Diabetes 1991;40:873–877. 106. Kennedy JM, Zochodne DW. Impaired peripheral nerve re- 127. Low PA, Nickander KK, Tritschler HJ. The roles of oxidative generation in diabetes mellitus. J Peripher Nerv Syst 2005; stress and antioxidant treatment in experimental diabetic 10:144–157. neuropathy. Diabetes 1997;46:S38–S42. 107. Kennedy WR. Opportunities afforded by the study of unmy- 128. Lyons TJ, Jenkins AJ. Glycation, oxidation, and lipoxidation elinated nerves in skin and other organs. Muscle Nerve in the development of the complications of diabetes: a car- 2004;29:756–767. bonyl stress hypothesis. Diabetes Rev 1997;5:365–391. 108. Kennedy WR, Navarro X, Goetz FC, Sutherland DE, Najarian 129. Malik RA. The pathology of human diabetic neuropathy. JS. Effects of pancreatic transplantation on diabetic neurop- Diabetes 1997;46:S50–S53. athy. N Engl J Med 1990;322:1031–1037. 130. Malik RA, Kallinikos P, Abbott CA, van Schie CH, Morgan P, 109. Kennedy WR, Nolano M, Wendelschafer-Crabb G, Johnson Efron N, et al. Corneal confocal microscopy: a non-invasive TL, Tamura E. A skin blister method to study epidermal surrogate of nerve fibre damage and repair in diabetic pa- nerves in peripheral nerve disease. Muscle Nerve 1999;22: tients. Diabetologia 2003;46:683–688. 360–371. 131. Malik RA, Newrick PG, Sharma AK, Jennings A, Ah-See AK, 110. Kennedy WR, Wendelschafer-Crabb G, Johnson T. Quanti- Mayhew TM, et al. Microangiopathy in human diabetic neu- tation of epidermal nerves in diabetic neuropathy. Neurol- ropathy: relationship between capillary abnormalities and ogy 1996;47:1042–1048. the severity of neuropathy. Diabetologia 1989;32:92–102. 111. Kihara M, Low PA. Impaired vasoreactivity to nitric oxide in 132. Malik RA, Tesfaye S, Newrick PG, Walker D, Rajbhandari experimental diabetic neuropathy. Exp Neurol 1995;132: SM, Siddique I, et al. Sural nerve pathology in diabetic 180–185. patients with minimal but progressive neuropathy. Diabeto- 112. Kochar DK, Rawat N, Agrawal RP, Vyas A, Beniwal R, Kochar logia 2005;48:578–585. SK, et al. Sodium valproate for painful diabetic neuropathy: 133. Malik RA, Tesfaye S, Thompson SD, Veves A, Sharma AK, a randomized double-blind placebo-controlled study. QJM Boulton AJM, et al. Endoneurial localisation of microvascu- 2004;97:33–38. lar damage in human diabetic neuropathy. Diabetologia 113. Korthals JK, Gieron MA, Dyck PJ. Intima of epineurial arte- 1993;36:454–459. rioles is increased in diabetic polyneuropathy. Neurology 134. Malik RA, Veves A, Masson EA, Sharma AK, Ah-See AK, 1988;38:1582–1586. Schady W, et al. Endoneurial capillary abnormalities in mild 114. Kowluru R, Bitensky MW, Kowluru A, Dembo M, Keaton PA, human diabetic neuropathy. J Neurol Neurosurg Psychiatry Buican T. Reversible sodium pump defect and swelling in 1992;55:557–561. the diabetic rat erythrocyte: effects on filterability and im- 135. Malik RA, Veves A, Walker D, Siddique I, Lye RH, Schady W, plications for microangiopathy. Proc Natl Acad SciUSA et al. Sural nerve fibre pathology in diabetic patients with 1989;86:3327–3331. mild neuropathy: relationship to pain, quantitative sensory

162 Diabetes Mellitus and the PNS MUSCLE & NERVE August 2007 testing and peripheral nerve electrophysiology. Acta Neuro- dation, and energy metabolism in diabetic precataractous pathol 2001;101:367–374. lens. Invest Ophthalmol Vis Sci 1999;40:680–688. 136. Mallozzi C, Di Stasi AM, Minetti M. Peroxynitrite modulates 157. Obrosova IG, Van Huysen C, Fathallah L, Cao XC, Greene tyrosine-dependent signal transduction pathway of human DA, Stevens MJ. An aldose reductase inhibitor reverses early erythrocyte band 3. FASEB J 1997;11:1281–1290. diabetes-induced changes in peripheral nerve function, me- 137. Max MB, Lynch SA, Muir J, Shoaf SE, Smoller B, Dubner R. tabolism, and antioxidative defense. FASEB J 2002;16:123– Effects of desipramine, amitriptyline, and fluoxetine on pain 125. in diabetic neuropathy. N Engl J Med 1992;326:1250–1256. 158. Orchard TJ, LLoyd CE, Maser RE, Kuller LH. Why does 138. McQuay H, Carroll D, Jadad AR, Wiffen P, Moore A. Anti- diabetic autonomic neuropathy predict IDDM mortality? An convulsant drugs for management of pain: a systematic re- analysis from the Pittsburgh Epidemiology of Diabetes Com- view. BMJ 1995;311:1047–1052. plications Study. Diabetes Res Clin Pract 1996;34(Suppl): 139. Merlot S, Firtel RA. Leading the way: directional sensing S165–S171. through phosphatidylinositol 3-kinase and other signaling 159. Oskarsson P, Ljunggren JG, Lins PE. Efficacy and safety of pathways. J Cell Sci 2003;116:3471–3478. mexiletine in the treatment of painful diabetic neuropathy. 140. Middlemas A, Delcroix JD, Sayers NM, Tomlinson DR, The Mexiletine Study Group. Diabetes Care 1997;20:1594– Fernyhough P. Enhanced activation of axonally transported 1597. stress-activated protein kinases in peripheral nerve in dia- 160. Otto M, Bach FW, Jensen TS, Sindrup SH. Valproic acid has betic neuropathy is prevented by neurotrophin-3. Brain no effect on pain in polyneuropathy: a randomized, con- 2003;126:1671–1682. trolled trial. Neurology 2004;62:285–288. 141. Mizisin AP, Li L, Calcutt NA. Sorbitol accumulation and 161. Perkins BA, Olaleye D, Zinman B, Bril V. Simple screening transmembrane efflux in osmotically stressed JS1 schwan- tests for peripheral neuropathy in the diabetes clinic. Dia- noma cells. Neurosci Lett 1997;229:53–56. betes Care 2001;24:250–256. 142. Mizisin AP, Powell HC. Schwann cell changes induced as 162. Pirart J. Diabetes mellitus and its degenerative complica- early as one week after galactose intoxication. Acta Neuro- tions. A prospective study of 4,400 patients observed between pathol 1997;93:611–618. 1947 and 1973. Part 1. Diabetes Care 1978;1:168–188. 143. Mizisin AP, Powell HC, Myers RR. Edema and increased 163. Pirart J. Diabetes mellitus and its degenerative complica- endoneurial sodium in galactose neuropathy. Reversal with tions. A prospective study of 4,400 patients observed between an aldose reductase inhibitor. J Neurol Sci 1986;74:35–43. 1947 and 1973. Part 2. Diabetes Care 1978;1:252–263. 144. Mohamed AK, Bierhaus A, Schiekofer S, Tritschler H, 164. Polydefkis M, Hauer P, Griffin JW, McArthur JC. Skin biopsy Ziegler R, Nawroth PP. The role of oxidative stress and as a tool to assess distal small fiber innervation in diabetic NF-kappaB activation in late diabetic complications. BioFac- neuropathy. Diabetes Technol Ther 2001;3:23–28. tors 1999;10:157–167. 165. Pop-Busui R, Marinescu V, Van Huysen C, Li F, Sullivan K, 145. Mulder DW, Lambert EH, Bastron JA, Sprague RG. The Greene DA, et al. Dissection of metabolic, vascular, and neuropathies associated with diabetes mellitus. Neurology nerve conduction interrelationships in experimental dia- 1961;11:275–284. betic neuropathy by cyclooxygenase inhibition and acetyl-L- 146. Namikawa K, Honma M, Abe K, Takeda M, Mansur K, Obata carnitine administration. Diabetes 2002;51:2619–2628. T, et al. Akt/protein kinase B prevents injury-induced mo- 166. Price SA, Zeef LA, Wardleworth L, Hayes A, Tomlinson DR. toneuron death and accelerates axonal regeneration. J Neu- Identification of changes in gene expression in dorsal root rosci 2000;20:2875–2886. ganglia in diabetic neuropathy: correlation with functional 147. Navarro X, Kennedy WR, Aeppli D, Sutherland DE. Neurop- deficits. J Neuropathol Exp Neurol 2006;65:722–732. athy and mortality in diabetes: influence of pancreas trans- 167. Pugliese G, Tilton RG, Speedy A, Chang K, Santarelli E, plantation. Muscle Nerve 1996;19:1009–1016. Province MA, et al. Effects of very mild versus overt diabetes 148. Navarro X, Sutherland DE, Kennedy WR. Long-term effects on vascular haemodynamics and barrier function in rats. of pancreatic transplantation on diabetic neuropathy. Ann Diabetologia 1989;32:845–857. Neurol 1997;42:727–736. 168. Quattrini C, Jeziorska M, Malik R. Small fiber neuropathy in 149. Newrick PG, Wilson AJ, Jakubowski J, Boulton AJ, Ward JD. diabetes: clinical consequence and assessment. Int J Low Sural nerve oxygen tension in diabetes. Br Med J 1986;293: Extrem Wounds 2004;3:16–21. 1053–1054. 169. Raskin J, Pritchett YL, Wang F, D’Souza DN, Waninger AL, 150. Nukada H. Mild ischemia causes severe pathological Iyengar S, et al. A double-blind, randomized multicenter changes in experimental diabetic nerve. Muscle Nerve 1992; trial comparing duloxetine with placebo in the management 15:1116–1122. of diabetic peripheral neuropathic pain. Pain Med 2005;6: 151. Oates PJ. Polyol pathway and diabetic peripheral neuropa- 346–356. thy. Int Rev Neurobiol 2002;50:325–392. 170. Raskin P, Donofrio PD, Rosenthal NR, Hewitt DJ, Jordan 152. Obrosova IG, Drel VR, Pacher P, Ilnytska O, Wang ZQ, DM, Xiang J, et al. Topiramate vs placebo in painful diabetic Stevens MJ, et al. Oxidative-nitrosative stress and poly(ADP- neuropathy: analgesic and metabolic effects. Neurology ribose) polymerase (PARP) activation in experimental dia- 2004;63:865–873. betic neuropathy: the relation is revisited. Diabetes 2005;54: 171. Richter RW, Portenoy R, Sharma U, LaMoreaux L, Bock- 3435–3441. brader H, Knapp LE. Relief of painful diabetic peripheral 153. Obrosova IG, Fathallah L. Evaluation of an aldose reductase neuropathy with pregabalin: a randomized, placebo-con- inhibitor on lens metabolism, ATPases and antioxidative trolled trial. J Pain 2005;6:253–260. defense in streptozotocin-diabetic rats: an intervention 172. Roberts RE, McLean WG. Protein kinase C isozyme expres- study. Diabetologia 2000;43:1048–1055. sion in sciatic nerves and spinal cords of experimentally 154. Obrosova IG, Fathallah L, Stevens MJ. Taurine counteracts diabetic rats. Brain Res 1997;754:147–156. oxidative stress and nerve growth factor deficit in early ex- 173. Rosenstock J, Tuchman M, LaMoreaux L, Sharma U. Pre- perimental diabetic neuropathy. Exp Neurol 2001;172:211– gabalin for the treatment of painful diabetic peripheral 219. neuropathy: a double-blind, placebo-controlled trial. Pain 155. Obrosova IG, Li F, Abatan OI, Forsell MA, Komjati K, Pacher 2004;110:628–638. P, et al. Role of poly(ADP-ribose) polymerase activation in 174. Rowbotham MC, Goli V, Kunz NR, Lei D. Venlafaxine ex- diabetic neuropathy. Diabetes 2004;53:711–720. tended release in the treatment of painful diabetic neurop- 156. Obrosova IG, Stevens MJ. Effect of dietary taurine supple- athy: a double-blind, placebo-controlled study. Pain 2004; mentation on GSH and NAD(P)-redox status, lipid peroxi- 110:697–706.

Diabetes Mellitus and the PNS MUSCLE & NERVE August 2007 163 175. Rundles RW. Diabetic neuropathy—general review with re- 195. Schmidt RE, Yip HK. Retrograde axonal transport in rat ileal port of 125 cases. Medicine 1945;24:111–160. mesenteric nerves. Characterization using intravenously ad- 176. Russell JW, Sullivan KA, Windebank AJ, Herrmann DN, ministered 125I-nerve growth factor and effect of chemical Feldman EL. Neurons undergo apoptosis in animal and cell sympathectomy. Diabetes 1985;34:1222–1229. culture models of diabetes. Neurobiol Dis 1999;6:347–363. 196. Schratzberger P, Walter DH, Rittig K, Bahlmann FH, Pola R, 177. Saarto T, Wiffen PJ. Antidepressants for neuropathic pain. Curry C, et al. Reversal of experimental diabetic neuropathy Cochrane Database Syst Rev 2005;CD005454. by VEGF gene transfer. J Clin Invest 2001;107:1083–1092. 178. Said G, Slama G, Selva J. Progressive centripetal degenera- 197. Shaw JE, Zimmet PZ, Gries FA, Ziegler D. Epidemiology of tion of axons in small fibre diabetic polyneuropathy. Brain diabetic neuropathy. In: Gries FA, Cameron NE, Low PA, 1983;106:791–807. Ziegler D, editors. Textbook of diabetic neuropathy. New 179. Schmidt AM, Hori O, Cao R, Yan SD, Brett J, Wautier JL, et York: Thieme; 2003. p 64–82. al. RAGE: a novel cellular receptor for advanced glycation 198. Shy ME, Frohman EM, So YT, Arezzo JC, Cornblath DR, end products. Diabetes 1996;45:S77–S80. Giuliani MJ, et al. Quantitative sensory testing: report of the 180. Schmidt AM, Hori O, Chen JX, Li JF, Crandall J, Zhang J, et Therapeutics and Technology Assessment Subcommittee of al. Advanced glycation endproducts interacting with their the American Academy of Neurology. Neurology 2003;60: endothelial receptor induce expression of vascular cell ad- 898–904. hesion molecule-1 (VCAM-1) in cultured human endothelial 199. Sima AA. Peripheral neuropathy in the spontaneously dia- cells and in mice. A potential mechanism for the accelerated betic BB-Wistar-rat. An ultrastructural study. Acta Neuro- vasculopathy of diabetes. J Clin Invest 1995;96:1395–1403. pathol (Berl) 1980;51:223–227. 181. Schmidt RE. The role of nerve growth factor in the patho- 200. Sima AA. The development and structural characterization genesis and therapy of diabetic neuropathy. Diabet Med of the neuropathies in the spontaneously diabetic BB Wistar 1993;10:10S–13S. rat. Metabolism 1983;32:106–111. 182. Schmidt RE, Beaudet LN, Plurad SB, Dorsey DA. Axonal 201. Sima AA, Bril V, Nathaniel V, McEwen TA, Brown MB, cytoskeletal pathology in aged and diabetic human sympa- Lattimer SA, et al. Regeneration and repair of myelinated thetic autonomic ganglia. Brain Res 1997;769:375–383. fibers in sural-nerve biopsy specimens from patients with 183. Schmidt RE, Cogswell BE. Tyrosine hydroxylase activity in diabetic neuropathy treated with sorbinil. N Engl J Med sympathetic nervous system of rats with streptozocin-in- 1988;319:548–555. duced diabetes. Diabetes 1989;38:959–968. 202. Sima AA, Calvani M, Mehra M, Amato A. Acetyl-L-carnitine improves pain, nerve regeneration, and vibratory perception 184. Schmidt RE, Dorsey D, Parvin CA, Beaudet LN, Plurad SB, in patients with chronic diabetic neuropathy: an analysis of Roth KA. Dystrophic axonal swellings develop as a function two randomized placebo-controlled trials. Diabetes Care of age and diabetes in human dorsal root ganglia. J Neuro- 2005;28:89–94. pathol Exp Neurol 1997;56:1028–1043. 203. Sima AA, Hay K. Functional aspects and pathogenetic con- 185. Schmidt RE, Dorsey DA, Beaudet LN, Reiser KM, Williamson siderations of the neuropathy in the spontaneously diabetic JR, Tilton RG. Effect of aminoguanidine on the frequency of BB-Wistar rat. Neuropathol Appl Neurobiol 1981;7:341–350. neuroaxonal dystrophy in the superior mesenteric sympa- 204. Sima AA, Lattimer SA, Yagihashi S, Greene DA. Axo-glial thetic autonomic ganglia of rats with streptozocin-induced dysjunction. A novel structural lesion that accounts for diabetes. Diabetes 1996;45:284–290. poorly reversible slowing of nerve conduction in the spon- 186. Schmidt RE, Dorsey DA, McDaniel ML, Corbett JA. Charac- taneously diabetic bio-breeding rat. J Clin Invest 1986;77: terization of NADPH diaphorase activity in rat sympathetic 474–484. autonomic ganglia — effect of diabetes and aging. Brain Res 205. Sima AA, Lorusso AC, Thibert P. Distal symmetric polyneu- 1993;617:343–348. ropathy in the spontaneously diabetic BB-Wistar rat. An 187. Schmidt RE, Dorsey DA, Roth KA. Immunohistochemical ultrastructural and teased fiber study. Acta Neuropathol characterization of NPY and substance P containing nerve (Berl) 1982;58:39–47. terminals in aged and diabetic human sympathetic ganglia. 206. Sima AA, Zhang W, Li ZG, Murakawa Y, Pierson CR. Molec- Brain Res 1992;583:320–326. ular alterations underlie nodal and paranodal degeneration 188. Schmidt RE, Grabau GG, Yip HK. Retrograde axonal trans- [125] in type 1 diabetic neuropathy and are prevented by C-pep- port of Nerve Growth Factor in ileal mesenteric nerves tide. Diabetes 2004;53:1556–1563. in vitro: effect of streptozotocin diabetes. Brain Res 1986; 207. Sima AAF, Brismar T, Yagihashi S. Neuropathies encoun- 378:325–336. tered in the spontaneously diabetic BB Wistar rat. In: Dyck 189. Schmidt RE, Plurad DA, Plurad SB, Cogswell BE, Diani AR, PJ, Thomas PK, Asbury AK, Winegrad AI, Porte D Jr, editors. Roth KA. Ultrastructural and immunohistochemical charac- Diabetic neuropathy. Toronto: WB Saunders; 1987. terization of autonomic neuropathy in genetically diabetic 208. Sima AAF, Pierson CR. Metabolic alterations in experimen- Chinese hamsters. Lab Invest 1989;61:77–92. tal models. In: Gries FA, Cameron NE, Low PA, Ziegler D, 190. Schmidt RE, Plurad DA, Roth KA. Effects of chronic exper- editors. Textbook of diabetic neuropathy. New York: Thi- imental streptozotocin induced diabetes on the noradrener- eme; 2003. p 96–105. gic and peptidergic innervation of the rat alimentary tract. 209. Simpson LO. Altered blood rheology in the pathogenesis of Brain Res 1988;458:353–360. diabetic and other neuropathies. Muscle Nerve 1988;11:725– 191. Schmidt RE, Plurad SB. Ultrastructural and biochemical 744. characterization of autonomic neuropathy in rats with 210. Sindrup SH, Andersen G, Madsen C, Smith T, Brosen K, chronic streptozotocin diabetes. J Neuropathol Exp Neurol Jensen TS. Tramadol relieves pain and allodynia in polyneu- 1986;45:525–544. ropathy: a randomised, double controlled trail. Pain 1999; 192. Schmidt RE, Plurad SB, Modert CW. Experimental diabetic 83:85–90. autonomic neuropathy characterization in streptozotocin di- 211. Sindrup SH, Gram LF, Brosen K, Eshoj O, Mogensen EF. abetic Sprague Dawley rats. Lab Invest 1983;49:538–552. The selective serotonin reuptake inhibitor paroxetine is ef- 193. Schmidt RE, Plurad SB, Sherman WR, Williamson JR, Tilton fective in the treatment of diabetic neuropathy symptoms. RG. Effects of aldose reductase inhibitor sorbinil on neuro- Pain 1990;42:135–144. axonal dystrophy and levels of myo-inositol and sorbital in 212. Singhal A, Cheng C, Sun H, Zochodne DW. Near nerve local sympathetic autonomic ganglia of streptozotocin-induced insulin prevents conduction slowing in experimental diabe- diabetic rats. Diabetes 1989;38:569–579. tes. Brain Res 1997;763:209–214. 194. Schmidt RE, Scharp DW. Axonal dystrophy in experimental 213. Stern D, Yan SD, Yan SF, Schmidt AM. Receptor for ad- diabetic autonomic neuropathy. Diabetes 1982;31:761–770. vanced glycation endproducts: a multiligand receptor mag-

164 Diabetes Mellitus and the PNS MUSCLE & NERVE August 2007 nifying cell stress in diverse pathologic settings. Adv Drug 234. Toth C, Schmidt AM, Tuor UI, Francis G, Foniok T, Brussee Deliv Rev 2002;54:1615–1625. V, et al. Diabetes, leukoencephalopathy and rage. Neurobiol 214. Stevens EJ, Carrington AL, Tomlinson DR. Nerve ischemia Dis 2006;23:445–461 in diabetic rats: time-course of development, effect of insulin 235. Tuck RR, Schmelzer JD, Low PA. Endoneurial blood flow treatment plus comparison of streptozotocin and BB mod- and oxygen tension in the sciatic nerves of rats with exper- els. Diabetologia 1994;37:43–48. imental diabetic neuropathy. Brain 1984;107:935–950. 215. Stevens MJ, Dananberg J, Feldman E, Lattimer SA, Kamijo 236. Valensi P, Le DC, Richard JL, Farez C, Khodabandehlou T, M, Thomas TP, et al. The linked roles of nitric oxide, aldose Rosenbloom RA, et al. A multicenter, double-blind, safety reductase and, (Naϩ,Kϩ)-ATPase in the slowing of nerve study of QR-333 for the treatment of symptomatic diabetic conduction in the streptozotocin diabetic rat. J Clin Invest peripheral neuropathy. A preliminary report. J Diabetes 1994;94:853–859. Complications 2005;19:247–253. 216. Stewart J. Diabetic truncal neuropathy: topography of the 237. Valk GD, de Sonnaville JJ, van Houtum WH, Heine RJ, van sensory deficit. Ann Neurol 1989;25:233. Eijk JT, Bouter LM, et al. The assessment of diabetic poly- 217. Stewart JD, McKelvey R, Durcan L, Carpenter S, Karpati G. neuropathy in daily clinical practice: reproducibility and Chronic inflammatory demyelinating polyneuropathy validity of Semmes Weinstein monofilaments examination (CIDP) in diabetics. J Neurol Sci 1996;142:59–64. and clinical neurological examination. Muscle Nerve 1997; 218. Suarez GA, Dyck PJ. Quantitative sensory assessment. In: 20:116–118. Dyck PJ, Thomas PK, editors. Diabetic neuropathy. Philadel- 238. Veglio M, Sivieri R, Chinaglia A, Scaglione L, Cavallo-Perin phia: WB Saunders; 2006. p 151–170. P. QT interval prolongation and mortality in type 1 diabetic 219. Sugimoto K, Murakawa Y, Sima AA. Expression and localiza- patients: a 5-year cohort prospective study. Neuropathy tion of insulin receptor in rat dorsal root ganglion and Study Group of the Italian Society of the Study of Diabetes, spinal cord. J Peripher Nerv Syst 2002;7:44–53. Piedmonte Affiliate. Diabetes Care 2000;23:1381–1383. 220. Sugimoto K, Murakawa Y, Zhang W, Xu G, Sima AA. Insulin 239. Vincent AM, Russell JW, Low P, Feldman EL. Oxidative receptor in rat peripheral nerve: its localization and alterna- stress in the pathogenesis of diabetic neuropathy. Endocr tively spliced isoforms. Diabetes Metab Res Rev 2000;16:354– Rev 2004;25:612–628. 363. 240. Vinik A. Clinical review: use of antiepileptic drugs in the 221. Sumner CJ, Sheth S, Griffin JW, Cornblath DR, Polydefkis M. treatment of chronic painful diabetic neuropathy. J Clin The spectrum of neuropathy in diabetes and impaired glu- Endocrinol Metab 2005;90:4936–4945. cose tolerance. Neurology 2003;60:108–111. 241. Vinik A, Ullal J, Parson HK, Casellini CM. Diabetic neurop- 222. Tesfamariam B, Palacino JJ, Weisbrod RM, Cohen RA. Al- athies: clinical manifestations and current treatment op- dose reductase inhibition restores endothelial cell function tions. Nat Clin Pract Endocrinol Metab 2006;2:269–281. in diabetic rabbit aorta. J Cardiovasc Pharmacol 1993;21: 242. Vinik AI, Bril V, Kempler P, Litchy WJ, Tesfaye S, Price KL, 205–211. et al. Treatment of symptomatic diabetic peripheral neurop- 223. Tesfaye S, Chaturvedi N, Eaton SE, Ward JD, Manes C, athy with the protein kinase C beta-inhibitor ruboxistaurin Ionescu-Tirgoviste C, et al. Vascular risk factors and diabetic mesylate during a 1-year, randomized, placebo-controlled, neuropathy. N Engl J Med 2005;352:341–350. double-blind clinical trial. Clin Ther 2005;27:1164–1180. 224. Theriault M, Dort J, Sutherland G, Zochodne DW. Local 243. Vinik AI, Pittenger GL, Burcus NI, Anderson SA, Stansberry human sural nerve blood flow in diabetic and other poly- KB, McNear E, et al. Neuroprotective effects of novel neu- neuropathies. Brain 1997;120:1131–1138. rotherapeutic topiramate in diabetic patients. Diabetologia 225. Thienel U, Neto W, Schwabe SK, Vijapurkar U. Topiramate 2003;46:A317. in painful diabetic polyneuropathy: findings from three dou- 244. Wang C, Li Y, Wible B, Angelides KJ, Ishii DN. Effects of ble-blind placebo-controlled trials. Acta Neurol Scand 2004; insulin and insulin-like growth factors on neurofilament 110:221–231. mRNA and tubulin mRNA content in human neuroblastoma 226. Thrainsdottir S, Malik RA, Dahlin LB, Wiksell P, Eriksson SH-SY5Y cells. Brain Res Mol Brain Res 1992;13:289–300. KF, Rosen I, et al. Endoneurial capillary abnormalities 245. Ward KK, Low PA, Schmelzer JD, Zochodne DW. Prostacy- presage deterioration of glucose tolerance and accompany clin and noradrenaline in peripheral nerve of chronic ex- peripheral neuropathy in man. Diabetes 2003;52:2615–2622. perimental diabetes in rats. Brain 1989;112:197–208. 227. Tilton RG, Chang K, Hasan KS, Smith SR, Petrash JM, Misko 246. Watkins PJ. Facial sweating after food: a new sign of diabetic TP, et al. Prevention of diabetic vascular dysfunction by autonomic neuropathy. Br Med J 1973;1:583–587. guanidines. Inhibition of nitric oxide synthase versus ad- 247. Wellmer A, Misra VP, Sharief MK, Kopelman PG, Anand P. vanced glycation end-product formation. Diabetes 1993;42: A double-blind placebo-controlled clinical trial of recombi- 221–232. nant human brain-derived neurotrophic factor (rhBDNF) in 228. Timperley WR, Boulton AJ, Davies-Jones GA, Jarratt JA, diabetic polyneuropathy. J Peripher Nerv Syst 2001;6:204– Ward JD. Small vessel disease in progressive diabetic neurop- 210. athy associated with good metabolic control. J Clin Pathol 248. Wendt T, Tanji N, Guo J, Hudson BI, Bierhaus A, Ramasamy 1985;38:1030–1038. R et al. Glucose, glycation, and RAGE: implications for am- 229. Tolins JP, Schultz PJ, Raij L, Brown DM, Mauer SM. Abnor- plification of cellular dysfunction in diabetic nephropathy. mal renal hemodynamic response to reduced renal perfu- J Am Soc Nephrol 2003;14:1383–1395. sion pressure in diabetic rats: role of NO. Am J Physiol 249. West IC. Radicals and oxidative stress in diabetes. Diabet 1993;265:F886–F895. Med 2000;17:171–180. 230. Tomlinson DR. Neurobiology of diabetic neuropathy. San 250. White MF. The insulin signalling system and the IRS pro- Diego: Academic Press; 2002. teins. Diabetologia 1997;40:S2–17. 231. Toth C, Brussee V, Cheng C, Zochodne DW. Diabetes mel- 251. Wilbourn AJ. Diabetic entrapment and compression neurop- litus and the sensory neuron. J Neuropathol Exp Neurol athies. In: Dyck PJ, Thomas PK, editors. Diabetic neuropa- 2004;63:561–573. thy. Toronto: WB Saunders; 1999. p 481–508. 232. Toth C, Brussee V, Martinez JA, McDonald D, Cunningham 252. Williams E, Timperley WR, Ward JD, Duckworth T. Electron FA, Zochodne DW. Rescue and regeneration of injured microscopical studies of vessels in diabetic peripheral neu- peripheral nerve axons by intrathecal insulin. Neuroscience ropathy. J Clin Pathol 1980;33:462–470. 2006;139:429–449. 253. Williamson JR, Ostrow V, Eades D, Chang K, Allison W, Kilo 233. Toth C, Brussee V, Zochodne DW. Remote neurotrophic C, et al. Glucose-induced microvascular functional changes support of epidermal nerve fibres in experimental diabetes. in nondiabetic rats are stereospecific and are prevents by an Diabetologia 2006;49:1081–1088. aldose reductase inhibitor. J Clin Invest 1990;85:1167–1172.

Diabetes Mellitus and the PNS MUSCLE & NERVE August 2007 165 254. Xu Q-G, Cheng C, Sun H, Thomsen K, Zochodne DW. Local 267. Zochodne DW, Cheng C. Diabetic peripheral nerves are sensory ganglion ischemia induced by endothelin vasocon- susceptible to multifocal ischemic damage from endothelin. striction. Neuroscience 2003;122:897–905. Brain Res 1999;838:11–17. 255. Xu QG, Li X-Q, Kotecha SA, Cheng C, Sun HS, Zochodne 268. Zochodne DW, Cheng C. Neurotrophins and other growth DW. Insulin as an in vivo growth factor. Exp Neurol 2004; factors in the regenerative milieu of proximal nerve stump 188:43–51. tips. J Anat 2000;196:279–283. 256. Yamagishi, S, Masuta, N, Okamoto, K, Yagihashi, S. Alter- 269. Zochodne DW, Cheng C, Sun H. Diabetes increases sciatic ations of protein kinase C activity in the peripheral nerve of nerve susceptibility to endothelin-induced ischemia. Diabe- STZ-induced diabetic mice overexpressing human aldose tes 1996;45:627–632. reductase. Diabetes 2001;50;773–P. 270. Zochodne DW, Ho LT. Normal blood flow but lower oxygen 257. Yamagishi S, Uehara K, Otsuki S, Yagihashi S. Differential tension in diabetes of young rats: microenvironment and the influence of increased polyol pathway on protein kinase C influence of sympathectomy. Can J Physiol Pharmacol 1992; expressions between endoneurial and epineurial tissues in 70:651–659. diabetic mice. J Neurochem 2003;87:497–507. 271. Zochodne DW, Ho LT. The influence of indomethacin and 258. Yasuda H, Dyck PJ. Abnormalities of endoneurial microves- guanethidine on experimental streptozotocin diabetic neu- sels and sural nerve pathology in diabetic neuropathy. Neu- ropathy. Can J Neurol Sci 1992;19:433–441. rology 1987;37:20–28. 272. Zochodne DW, Ho LT. Diabetes mellitus prevents capsaicin 259. Yasuda H, Sonobe M, Yamashita M, Terada M, Hatanaka I, from inducing hyperaemia in the rat sciatic nerve. Diabeto- Huitian Z et al. Effect of prostaglandin E1 analogue TFC 612 logia 1993;36:493–496. on diabetic neuropathy in streptozocin-induced diabetic rats 273. Zochodne DW, Ho LT. The influence of sulindac on exper- Comparison with aldose reductase inhibitor ONO 2235. Di- imental streptozotocin-induced diabetic neuropathy. Can abetes 1989;38:832–838. J Neurol Sci 1994;21:194–202. 260. Yuen KC, Baker NR, Rayman G. Treatment of chronic pain- 274. Zochodne DW, Ho LT, Allison JA. Dorsal root ganglia mi- ful diabetic neuropathy with isosorbide dinitrate spray: a croenvironment of female BB Wistar diabetic rats with mild double-blind placebo-controlled cross-over study. Diabetes neuropathy. J Neurol Sci 1994;127:36–42. Care 2002;25:1699–1703. 261. Zalba G, Beaumont J, San Jose G, Fortuno A, Fortuno MA, 275. Zochodne DW, Isaac D, Jones C. Failure of immunotherapy Diez J. Vascular oxidant stress: molecular mechanisms and to prevent, arrest or reverse diabetic lumbosacral plexopa- pathophysiological implications. J Physiol Biochem 2000;56: thy. Acta Neurol Scand 2003;107:299–301. 57–64. 276. Zochodne DW, Murray MM, van der Sloot P, Riopelle RJ. 262. Zenon GJ, Abobo CV, Carter BL, Ball DW. Potential use of Distal tibial mononeuropathy in diabetic and nondiabetic aldose reductase inhibitors to prevent diabetic complica- rats reared on wire cages: an experimental entrapment neu- tions. Clin Pharm 1990;9:446–457. ropathy. Brain Res 1995;698:130–136. 263. Zhang W, Kamiya H, Ekberg K, Wahren J, Sima AA. C- 277. Zochodne DW, Nguyen C. Increased peripheral nerve mi- peptide improves neuropathy in type 1 diabetic BB/Wor- crovessels in early experimental diabetic neuropathy: quan- rats. Diabetes Metab Res Rev 2007;23:63–70. titative studies of nerve and dorsal root ganglia. J Neurol Sci 264. Zhang W, Yorek M, Pierson CR, Murakawa Y, Breidenbach 1999;166:40–46. A, Sima AA. Human C-peptide dose dependently prevents 278. Zochodne DW, Sun H-S, Cheng C, Eyer J. Accelerated dia- early neuropathy in the BB/Wor-rat. Int J Exp Diabetes Res betic neuropathy in axons without neurofilaments. Brain 2001;2:187–193. 2004;127:2193–2200. 265. Ziegler D, Nowak H, Kempler P, Vargha P, Low PA. Treat- 279. Zochodne DW, Verge VM, Cheng C, Hoke A, Jolley C, ment of symptomatic diabetic polyneuropathy with the an- Thomsen K, et al. Nitric oxide synthase activity and expres- tioxidant alpha-lipoic acid: a meta-analysis. Diabet Med sion in experimental diabetic neuropathy. J Neuropathol 2004;21:114–121. Exp Neurol 2000;59:798–807. 266. Zochodne DW. Nerve and ganglion blood flow in diabetes: 280. Zochodne DW, Verge VMK, Cheng C, Sun H, Johnston J. an appraisal. In: Tomlinson D, editor. Neurobiology of dia- Does diabetes target ganglion neurons? Progressive sensory betic neuropathy. San Diego: Academic Press; 2002. Vol. 50, neuron involvement in long term experimental diabetes. p 161–202. Brain 2001;124:2319–2334.

166 Diabetes Mellitus and the PNS MUSCLE & NERVE August 2007