Journal of Neuropathology and Experimental Neurology
Copyright q 2004 by the American Association of Neuropathologists
Vol. 63, No. 6
June, 2004
pp. 561 573
Diabetes Mellitus and the Sensory Neuron
C. TOTH, MD, FRCPC, V. BRUSSEE, PHD, C. CHENG, MD,
AND
D. W. ZOCHODNE, MD, FRCPC
Abstract. Sensory neurons in diabetes may be primarily targeted by diabetes and their involvement may account for prominent sensory loss and pain in diabetic patients. Previous studies demonstrating evidence of excessive polyol flux, microangiopathy, and oxidative stress involving sensory axons and ganglia have been joined by more recent work demonstrating
altered neuron phenotype, mitochondrial dysfunction, ion channel alterations, and abnormal growth factor signaling. As such,
an interesting and unique panoply of molecular changes in primary sensory neurons has been identified in diabetic models.
Insulin deficiency and subsequent changes in second messenger signaling may also play an important role in how sensory
neurons respond to diabetes. Applying approaches to support sensory neurons in diabetes may be an important therapeutic
direction in diabetic patients.
Key Words:
Diabetes; Dorsal root ganglion; Neuron; Neuropathy.
INTRODUCTION
Human diabetes mellitus is associated with ‘‘endstage’’
complications involving the retina, kidney, and peripheral
nerve. Although exceptional control of diabetes-related
hyperglycemia may attenuate, delay, or prevent development of diabetic complications, reversal of complications does not occur. Among these complications, diabetes selectively targets the peripheral nervous system in
a widespread or diffuse fashion leading to polyneuropathy or selectively causing focal neuropathies. In patients
with polyneuropathy, sensory and perhaps autonomic
neurons appear to be involved early. Unfortunately, polyneuropathy is also not really an ‘‘endstage complication’’
but may, for example, develop in diabetic children.
A number of hypotheses have been proposed to explain the targeting of peripheral nerves by diabetes (Fig.
1). We will emphasize, however, those changes that are
particularly directed toward the cell body or perikaryon
of the sensory neuron. In the past, the direct targeting of
perikarya in ganglia has not been emphasized as much
as changes in distal axons. The perikaryon is an important target to consider, however, since loss of sensory
neurons may be an irretrievable consequence, limiting
options for repair. Alternatively, specific and early approaches that would support the sensory neuron may provide options to prevent such loss. This might be accomplished by better understanding of the potential role of
neuron growth factors, including insulin.
Why would disease of perikarya account for the features of diabetic polyneuropathy? The innervation of the
most distal target organs is the first to be disrupted by
early diabetic polyneuropathy. This causes clinical symptoms and signs of polyneuropathy, first in the toes and
From the Department of Clinical Neurosciences, University of Calgary, Calgary, Alberta, Canada.
Correspondence to: Dr. D.W. Zochodne, University of Calgary, Department of Clinical Neurosciences, Room 168, 3330 Hospital Drive,
N.W., Calgary, Alberta, Canada T2N 4N1. E-mail: dzochodn@
ucalgary.ca
then in the fingers. There may be at least 2 explanations
for why longer axons and their terminals are involved. It
may be that a longer ‘‘exposed’’ length of axon has a
higher likelihood of being targeted by blood-born or systemic damage. Another reason may be that impaired perikarya lose their ability to supply the molecules or other
structural components to their most distal terminals. Such
terminals consequently atrophy and then disappear. Examining the epidermal innervation of the skin in patients
or animals with diabetes does, in fact, identify such abnormal terminals with irregular contours and ‘‘retraction
bulbs.’’ In older parlance, this process was termed ‘‘dying
back,’’ although it was unclear whether the axon alone
or first its cell body was damaged.
Clinical Features of Diabetic Neuropathies
Diabetic sensory polyneuropathy usually precedes involvement of motor neurons in humans. In human studies, there is loss of axons in the sural nerve as well as in
the epidermis of the skin, and loss involves both myelinated and unmyelinated axons. Symptoms of sensory polyneuropathy generally begin with paresthesias, tingling,
prickling, and ‘‘pins and needles’’-like sensations that,
with progression, travel further up the foot and leg, later
involving fingers and hands. A characteristic feature of
diabetic polyneuropathy is a stocking and glove distribution of pain, paresthesiae, or loss of sensation. Like
other polyneuropathies, these features indicate susceptibility, as discussed above, of the longest axons (i.e. those
involving the toes) to be lost first. Pain descriptors often
used include burning, electrical-like sensations, aching,
tightening, and discomfort by light touch (tactile allodynia). While prominent pain has been thought to be due to
damage specifically to small myelinated and unmyelinated afferents subserving nociception, more recent studies
have failed to support this ‘‘small fiber painful neuropathy’’ phenotype. Sural nerve biopsy studies of patients
with diabetic painful neuropathy have identified loss of
large myelinated axons as well (1). It may be that more
561
562
TOTH ET AL
Fig. 1.
Potential insults and support mechanisms of diabetic sensory neurons.
relevant processes in pain development are the alterations
in ion channel distribution and properties with associated
impacts on nerve excitability, rather than simple loss of
a fiber class. Such changes are described below.
Frank sensory loss develops to pinprick perception,
temperature sensibility, light touch, vibration perception,
and proprioception. Such loss of protective sensation is
associated with the development of foot ulceration, joint
damage (Charcot joint), and sometimes amputation. Semmes-Weinstein monofilaments that bend when applied to
the foot at a defined force are sometimes used to screen
for such sensory loss. Distal to proximal progression of
sensory loss in the foot may be associated with Achilles
tendon areflexia. Loss of proprioception occurs when afferent loss has progressed more rostral with interruption
of more proximal tendon and muscle afferents sensitive
to distal joint movements. Similarly, marked gait unsteadiness develops later. The extent of the sensory deficit can
be more accurately measured using quantitative sensory
testing, which provides threshold values for vibration
sensitivity, cold perception, and heat sensation. Quantitative sensory testing has identified loss of vibration sensitivity threshold as a very sensitive early index of diabetic sensory polyneuropathy.
J Neuropathol Exp Neurol, Vol 63, June, 2004
Small (3 mm) punch biopsies of the skin can be used
to quantitatively measure numbers of epidermal and dermal axons labeled with the marker PGP 9.5. In diabetes,
both a loss of epidermal axons and altered appearances
with formation of ovoids occur. Sensory nerve conduction studies are the most widely used and validated tests
of axon integrity in diabetic patients. In diabetes, declines
in the amplitude of the sensory nerve action potential
occur first, followed by slowing of sensory conduction
velocity. Although sural nerve biopsies are not taken for
routine diagnostic purposes in diabetic patients, they
identify diffuse or multifocal loss of axons, axonal degeneration, regenerating axon clusters in a single basement membrane (‘‘pseudo-onion bulbs’’), segmental demyelination, and changes in microvessels, including
microthrombosis, basement membrane thickening, endothelial cell reduplication, and smooth muscle proliferation
(2, 3). Some features of microangiopathy may occur in
patients with early neuropathy, but whether these changes
occur in parallel with axon loss or induce this loss is
debated (1). Theriault et al (4) directly measured nerve
blood flow using laser Doppler flowmetry in diabetic
patients with early polyneuropathy who later underwent
sural nerve biopsies. Despite declines in sensory nerve
DIABETES MELLITUS AND THE SENSORY NEURON
action potential amplitudes and numbers of axons (correlated with one another), local blood flow measures
tended toward higher, rather than lower, values in diabetic
patients, particularly those with more significant neuropathy.
While axonal atrophy contributes to conduction velocity slowing, such atrophy develops much later, at least in
experimental models, and atrophy in human diabetic sural
nerve biopsies has been difficult to demonstrate. Segmental demyelination, the hallmark of acquired demyelinating peripheral neuropathies, also occurs in diabetes
and can contribute to conduction velocity slowing, but is
generally not observed in experimental models where
slowing is present. Finally, in the BB Wistar model of
Type I diabetes, one group of investigators has suggested
that paranodal abnormalities known as axoglial dysjunction are important structural changes that may lead to
conduction slowing (5).
None of the therapies proposed to date truly reverse
diabetic sensory polyneuropathy, including trials of aldose reductase inhibitors, recombinant human nerve
growth factor, and many others not reviewed here. Even
pancreatic transplantation, rendering euglycemia has only
generated marginal improvements in neuropathy, particularly autonomic function, when patients are followed for
several years.
Potential Mechanisms of Diabetic Polyneuropathy
Convergent thinking is beginning to link established
derangements in diabetes into an overall scheme of the
pathogenesis of its complications, including neuropathy.
Briefly, these include excessive polyol (sugar alcohol)
flux into nerve and probably ganglia, oxidative stress,
glycosylation of nerve proteins, accelerated functional
and structural microangiopathy, and failure of neurotrophic support. Schwann cells accumulate excessive
polyols (6), particularly sorbitol, by accelerated flux
through the aldose reductase pathway. There are associated depletions of nerve myo-inositol, changes in protein
kinase C (PKC) subunits, and dysfunction of nerve Na/
K ATPase (7–9). Increased axonal Na content and later
Na channel migration out of nodes of Ranvier may account for early nerve conduction velocity slowing in diabetes (10, 11).
Oxidative stress is an early and critical pathogenic
event (12). Auto-oxidation of glucose and glycation products of glucose generate oxygen free radicals that include
the hydroxyl radical (OH●), superoxide anion (O2-), hydrogen peroxide, singlet oxygen, nitric oxide (NO), and
organic analogues. Moreover, generation of these species
occurs in the setting of lowered antioxidant defenses
characterized by lowered superoxide dismutase, catalase,
glutathione peroxidase, and glutathione reductase. The
consequences of oxidative stress include lipid peroxidation (elevated malondialdehyde, and conjugated dienes),
563
DNA damage (8-hydroxy-29-deoxyguanosine), and protein damage (protein carbonyls). Oxidative stress may
primarily target microvessels, in turn leading to vascular
damage and tissue ischemia. In peripheral nerve there is
an increase in malondialdehyde, conjugated dienes, and
lipid hydroperoxides with reduced levels of GSH (reduced glutathione) and glutathione peroxidase (12). Hyperglycemia and oxidative stress generate advanced glycosylation endproducts (AGEs) that are thought to bind
to receptors for AGEs (RAGEs), activating NF-kB, a potential mediator of cellular dysfunction (13).
Consumption of NO (nitric oxide) by AGEs in diabetes
may contribute to loss of vasodilation (functional microangiopathy) in diabetic microvessels. Alternatively, there
may be nitrergic stress, referring to damage of cellular
proteins by excessive NO or its metabolites acting as free
radical agents. NO combines with the superoxide radical
(O2-) to generate peroxynitrite (ONOO-), a highly potent
oxidizing agent that nitrates protein tyrosines, eventually
leading to cell death (14). NO itself targets thiol groups
on proteins, such as SNAP 25 and B50/GAP43, and damages DNA, in turn activating PARS (polyADP-ribose
synthase), an enzyme that depletes cellular ATP. Peripheral nerves and sensory ganglia in experimental diabetes
may have elevated NO synthesis, a feature that may contribute to local radical stress (15).
Microangiopathy, at first functional, then structural,
may arise from endothelial damage and NO quenching
during oxidative stress and polyol flux (2, 16–18). These
abnormalities are exacerbated by hyperviscosity, loss of
erythrocyte deformability, increased platelet aggregation,
and alterations in local oxygen release, all of which likely
contribute towards an eventual cascade of hypoxia and
ischemia (12, 19). Our lab, and others, have provided
extensive argument against the concept that early nerve
trunk ischemia ‘‘accounts’’ for diabetic neuropathy (see
review [3]). Instead, microvascular, axon and neuron
damage may develop in parallel.
Dorsal Root Ganglia and Diabetic Microangiopathy
Dorsal root ganglia (DRG) are invested with microvessels originating from segmental radicular arteries and
anastomoses from spinal arteries. Sensory ganglia have
interesting physiological features that may reflect greater
potential vulnerability to diabetes (20). DRG entrain
higher levels of local blood flow measuring 30 to 40 ml/
100g/min compared to 15 to 20 ml/100g/min in the sciatic nerve trunk. Mean ambient oxygen tensions measured using polarography within ganglia are lower than
those of the peripheral nerve trunk, indicating greater
metabolic demand and oxygen uptake (21, 22). Peripheral
nerves have a near linear relationship between mean arterial pressure and blood flow, whereas DRG have evidence of partial autoregulation. Combined, these physiological features of ganglia suggest heightened
J Neuropathol Exp Neurol, Vol 63, June, 2004
564
TOTH ET AL
Fig. 2. Artifactual subplasmalemmal vacuoles from lumbar DRG in a patient without known peripheral nerve or ganglion
disease. Such changes are a function of the postmortem interval. The white arrow indicates subplasmalemmal artifactual vacuoles.
flow-metabolic coupling in ganglia and may predispose
to more serious consequences of microangiopathy in diabetes (23, 24). DRG also have a more leaky blood-ganglia barrier than within nerve or brain.
We observed reductions in ganglia blood flow in experimental streptozotocin (STZ)-induced diabetes, with
measurements around 20 ml/100g/min (25, 26). These
changes were selective for ganglia, as the same cohort of
diabetic rats had normal concurrent nerve blood flow.
While such reductions might suggest that chronic ganglion ischemia damages diabetic sensory neurons primarily, it is also possible that such changes are secondary
as a result of decreased neuron metabolic demand.
Could episodic ischemia target ganglia? Endothelin, a
highly potent circulating endothelial elaborated vasoconstrictor has been reported to be elevated in some studies
of diabetic patients, but not all (27, 28). Exposure of ganglia to exogenous endothelin does induce local vasoconstriction. Like the nerve trunk, however, its actions on
diabetic microvessels are more intense and prolonged
than those on similar nondiabetics (29). This finding supports the concept that microangiopathy occurs early in
diabetes even if it is not a primary pathogenetic event.
Endothelin vasoconstriction-rendered ischemic damage
and DNA fragmentation was largely selective to diabetic
ganglia damaging perikarya (29). There was also downstream axonal degeneration of sensory axons in the sural
nerve not directly exposed to ischemia. Intraganglionic
J Neuropathol Exp Neurol, Vol 63, June, 2004
axonal damage was observed early with axotomy cell
body reactions, including eccentric nuclei and central
chromatolysis.
Diabetic Sensory Neurons: Survival and Phenotype
In human diabetic subjects, biopsies of the sural nerve
identify loss of axons, segmental demyelination, and
structural changes in microvessels. While axon loss might
arise from retraction of the axon terminals despite preservation of ganglion neurons, it might also arise from
irretrievable loss of the entire neuron, including perikaryon and axon. Surprisingly, few pathological studies of
human sensory ganglia in diabetes mellitus have been
published. Greenbaum et al described loss of neurons,
replacement with nests of Nageotte, axon retraction
bulbs, and vacuolation of sensory neurons in 3 patients
with clinical diabetic polyneuropathy (30). In retrospect,
vacuolar changes in ganglia sensory neurons, including
subplasmalemmal scalloping, may not represent genuine
disease, but instead an artifact of tissue handling. Some
vacuolar changes observed in neurons represent swollen
double membrane mitochondria; their prevalence directly
related to postmortem interval (31) (Fig. 2). As such,
reports of microvacuolar neuronopathy in experimental
diabetes may similarly not reflect disease (32). Schmidt
et al, however, have identified a unique dystrophic structural change in proximal axonal segments of autonomic
DIABETES MELLITUS AND THE SENSORY NEURON
565
Fig. 3. Proximal axonal dystrophic changes (arrows) in human DRG observed in humans with advanced age or diabetes.
Loss of a neuron replaced by a nest of Nageotte (A, arrow) and proximal axon dystrophic swellings from sensory neurons (arrows,
B, C are H & E stain, D–F are silver stain). Reproduced with permission (78).
ganglia of long-term diabetic rats (8- to 12-month duration) and in human sensory ganglia from diabetic subjects
or older persons (Fig. 3) (33). Dystrophic proximal axons
appear strictly localized to ganglia, sometimes indenting
perikarya and containing prominent accumulations of
neurofilament. The significance of dystrophic axons is
uncertain, possibly related to arrested axoplasmic transport or a unique vulnerability of proximal axonal segments within ganglia. In separate work, similar proximal
axon segments in a long-term diabetic model had prominent expression of activated caspase-3 and PARP (poly
[ADP-ribose] polymerase) (34). In long-term models, lipofuscin pigment granules are more prominent in diabetic
neurons (32, 35). Beyond these changes, and overall perikaryal atrophy, there appear to be few other obvious morphological distinctions between diabetic and nondiabetic
neurons.
Is there actual loss of diabetic sensory neurons in experimental models? This appears to depend on the model
chosen. The absence of overt loss on direct neuronal
counts, for example, in the STZ rat model of Type I diabetes probably indicates that it is a milder model of human disease. Some studies have reported evidence of
DNA fragmentation and presumed apoptosis in large proportions of ganglion neurons in early STZ rat diabetes
(36, 37). Such findings of severe neuronal loss, however,
are not matched by direct counts either in the ganglia or
downstream sural nerve. It may be that some DNA fragmentation identified by TUNEL labeling or caspase activation provides evidence that sensory neurons suffer
‘‘apoptotic stress,’’ but nonetheless survive. We studied
a long-term (12 month) model of experimental rat STZinduced diabetes using a rigorous 3-dimensional physical
dissector counting approach. Overall numbers were
J Neuropathol Exp Neurol, Vol 63, June, 2004
566
TOTH ET AL
Fig. 4. Atrophic unmyelinated (top panels) and small myelinated axons (lower panels) in rat sural nerve. After 6 months of
diabetes, axons atrophy and demonstrate reduced numbers of neurofilaments (left panels) compared to nondiabetic littermates
(right panels) of similar age. Reproduced with permission from Scott et al (74).
preserved, corresponding with intact downstream sural
axon populations (35). Both perikarya and axons, however, undergo atrophy with an apparent shift of sizes to
smaller size categories. Overall, rats appear relatively resistant to diabetic neuropathy, while yet exhibiting measurable changes such as reductions in conduction velocity, axonal atrophy, and a series of molecular alterations
(see below). The importance, then, of this model may be
in its reflection of early changes of neuropathy preceding
degeneration and loss.
In mice rendered diabetic with STZ, the situation appears somewhat different and this model may better parallel human disease. While 12 months of diabetes in a
rat represents a substantial proportion of its lifespan, 6
months of diabetes in mice may provide relatively similar
diabetic exposure. Both may better represent decades of
human disease than the short-term models extensively reported in the literature. By 4 months of diabetes, STZ
J Neuropathol Exp Neurol, Vol 63, June, 2004
mice have evidence of active sensory axonal degeneration. By 8 months of diabetes there was evidence of loss
of epidermal innervation of the hindpaw footpads but
also substantial (;30%) loss of sensory neurons counted
using the physical dissector approach (Kennedy and Zochodne, unpublished data). Rigorous counting of sensory
neurons using dissector approaches have not been carried
out in Type II models or in the BB Wistar rat model of
Type I diabetes. Loss of perikarya requires that remaining
intact neurons collaterally reinnervate denervated target
organs. At this time, it is uncertain what barriers diabetes
imposes upon collateral sprouting of a residual population of sensory neurons.
What happens to sensory neurons prior to dropout? In
long-term model of STZ rat diabetes, axonal atrophy develops by 2 to 3 months, preceding later perikaryal atrophy (Fig. 4). Hyperosmolar shrinkage from hyperglycemia probably does not account for such changes
567
DIABETES MELLITUS AND THE SENSORY NEURON
TABLE
Molecular Changes Described in Experimental Diabetic Sensory Neurons in Ganglia
Molecule
Neurofilament H
Neurofilament M
Neurofilament L
ta1-tubulin
aCGRP
bCGRP
TrkA
Trk C
p75
PACAP
SP
Somatostatin
GAP43/B50
Nav1.8
cGMP
PKCb-II
NFkB
CREB
CCK
Galanin
BCl-2
ERK
JNK
P38
Activated Caspase-3
PARP
HSP27
IR
Calcium
Nav1.3,1.6,1.9
b3
Role
Change
Heavy subunit of neurofilament cytoskeleton
Medium subunit of neurofilament cytoskeleton
Light subunit of neurofilament cytoskeleton
Microtubule constituent
Calcitonin gene-related peptide, neuropeptide
Calcitonin gene-related peptide, neuropeptide
NGF receptor
NT-3 receptor
Low affinity neurotrophin receptor
Pituitary adenylate cyclase activating peptide
Substance P, neuropeptide
Neuropeptide
Growth associated protein
Sodium channel
Cyclic nucleotide messenger
Protein kinase C isoform
Second messenger
Cyclic AMP response element binding protein
Cholecystokinin neuropeptide
Neuropeptide
Anti-apoptotic protein
MAPK second messenger
MAPK second messenger
MAPK second messenger
Cysteine proteinase
Poly (ADP- ribose) polymerase
Heat shock protein
Insulin receptor
Sodium channel
Sodium channel subunit
because of the long delays in the appearance of atrophy
after hyperglycemia appears. Progressive dysfunction of
neurons implies a failure to elaborate critical structural
nerve proteins, of which neurofilament and tubulin are
key members. Both perikaryal and axonal atrophy can be
directly related to a loss of neurofilaments. Distal diabetic
axons appear to suffer a double hit, with failed synthesis
of neurofilament and slowed axonal transport (38).
Unique Molecular Changes in Diabetic Neurons
While axonal and neuronal losses are absolute measurable endpoints in the progression of a progressive neurological disorder, sublethal neuronal dysfunction may be
more difficult to identify. For example, declines in conduction velocity reflect an alteration in axon excitability
but do not imply potential demise. While the mouse model does illustrate that neuron loss is an eventual consequence of diabetes, it may be that there is chronic adaptation of neurons to limit apoptosis. Pro-apoptotic
signals generated from oxidative stress, chronic ischemia,
polyol flux, and loss of trophic support may be accompanied by rises in neuronal survival signals. The overall
balance and signaling in such pathways and their relation
↓ Hyperphosphorylation
↓ Hyperphosphorylation
↓ Hyperphosphorylation
↓
↓
↓
↓
↓
↓
↓
↓
↓
↓
↓
↓
↓
↓
↓
↔
↔
↔
↑
↑
↑
↑
↑
↑
↑
↑
↑
↑
Reference
(49, 74)
(35, 43, 74)
(49, 74)
(35, 49, 74)
(35, 75)
(35)
(35)
(35)
(35)
(35)
(35)
(76)
(49, 74)
(48, 56)
(48)
(48)
(52)
(52)
(35)
(35)
(34)
(42)
(42)
(42)
(34)
(34)
(35)
(77)
(46)
(56)
(57)
to disease duration are largely unknown. What is apparent is that chronically impaired diabetic neurons develop
some molecular changes known to occur after injury.
These changes are critical for neuron survival and regrowth and change of the primary phenotype from transmission to regeneration. Tomlinson, Fernyhough and colleagues pointed out the resemblance in alterations of
neuron mRNAs between diabetes and injury. These alterations include a downregulation of mRNA levels, neurofilament H and L subunits (NFH, NFL), CGRP, and
substance P among others. Reductions in peptide mRNAs
were corrected by nerve growth factor (NGF) (39). Similarly, NGF levels in tissue targets and nerves of diabetic
rats are reduced, prompting the hypothesis that NGF depletion accounts for molecular changes in diabetic sensory neurons (40).
The full repertoire of molecular changes in diabetic
sensory neurons, however, differs from that of injury (Table). For example, progressive long-term diabetes was
not only associated with declines in mRNA levels of neurofilament subunits (light, medium, and heavy), but also
with declines in ta-1 tubulin and GAP43/B50, molecules
that provide critical components for regenerating growth
J Neuropathol Exp Neurol, Vol 63, June, 2004
568
TOTH ET AL
Fig. 5. In situ hybridization images of L5 DRG from diabetic (12-month duration) rats (right) or littermate controls (left)
probed to label the medium subunit of the neurofilament triplet protein (NfM). Diabetics had a progressive decline in the neuronal
expression of mRNA for NfM, illustrated as less intense labeling on the right panels. Studies completed in the Verge laboratory,
University of Saskatchewan and reproduced with permission from Zochodne et al (35). Stars indicate a DRG neuron with intense
staining (left) or diminished staining (right).
cones (Fig. 5). Both ta-1 tubulin and GAP43/B50 are
upregulated within injured neurons. Other changes in the
long-term experimental diabetic model appearing in later,
but not earlier, time points include mRNAs coding for
important structural and transmitter proteins or peptides:
aCGRP, bCGRP, substance P, Trks A, B, and C, and PACAP (35). The full panoply of progressive mRNA changes has suggested that the diabetic phenotype reflects a
degenerative rather than an injury response of the neuron.
The distinction is important because it may alter how we
view potential rescue approaches clinically.
Nerve injury upregulates mRNA and protein levels of
the constitutive neuronal nitric oxide synthase (nNOS) in
sensory neurons (41). Neither short- nor long-term rat
diabetes was associated with significant changes in
mRNA or protein levels of nNOS, eNOS (constitutive
endothelial NOS), or iNOS (inducible inflammatory
NOS). Overall NOS activity, however, measured by conversion of labeled arginine to citrulline, was elevated in
both the ganglia and sciatic nerves of long-term diabetic
rats (15). Moreover, sensory neuron perikarya and proximal axonal segments had immunoreactivity for nitrotyrosine, a footprint of NO presence (34). Overall, these
J Neuropathol Exp Neurol, Vol 63, June, 2004
findings have suggested the possibility that chronic rises
in NOS activity, generating elevated local NO levels,
might create long-term nitrergic stress in diabetic ganglia.
Heightened activity of mitogen-activated protein kinases (MAPK) in sensory neurons may be a critical signaling alteration in diabetes (42). Of the MAPK types,
ERK and p38 were both activated in adult rat sensory
neurons in culture by oxidative stress, and were increased
in sensory neurons from 8-week STZ-induced diabetic
rats. JNK was activated by high glucose and appeared in
12-week STZ rat nerve (42, 43). Chronic rises in MAPK
activity may be pro-apoptotic and may influence other
aspects of neuronal function. One consequence may be
excessive neurofilament phosphorylation, a post-translational modification of the axon latticework that determines their spacing and perhaps longevity (43). Such
changes appear to be reversible, as demonstrated by the
impact of neurotrophin-3 (44) or insulin (45).
Two additional features indicate profound alterations
in how sensory neurons function and handle their metabolism in diabetes. Huang et al used fluorescent video imaging of lumbar DRGs to calculate cytosolic calcium dynamics (46). Both large and small neuron populations had
DIABETES MELLITUS AND THE SENSORY NEURON
569
Fig. 6. Prominent expression of activated caspase-3, poly (ADP-ribose) polymerase (PARP), and nitrotyrosine in L5 DRG
sensory neurons from long-term (12 month) diabetic rats using immunohistochemistry. Expression was not associated with active
neuron dropout or apoptosis, indicating neurons ‘‘under siege’’ but with survival. A: Illustrates labeling for activated caspase-3
(green). Note its presence in nuclei, cytoplasm, and in proximal axon segments. Some satellite cells also appear to have caspase
labeling. B: The identical section labeled for PARP (red) illustrating its overlap with caspase and yet more prominent labeling
in proximal axonal segments. C: The overlay image demonstrating colocalization of caspase and PARP. D: A higher power view
of the overlay illustrating patchy cytoplasmic and nuclear labeling for caspase-3 and prominent expression in proximal axons
surrounding the neuron. E: Illustrates a small neuron (arrowhead) and a proximal axon segment of a neighboring neuron (arrow)
labeled with an antibody directed against nitrotyrosine, a footprint of excessive nitrergic stress. F: The light microscopy image
of the section shown in panel E, demonstrating the histological features of the DRG neurons. Bar: A–C, E, F 5 50 mm; D 5 20
mm. Panels E and F: Copyright q 2003 American Diabetes Association. From Diabetes, Vol. 52, 2003:2363–2371. Reprinted
with permission from the American Diabetes Association.
cytoplasmic calcium levels twice that of controls, reversible by neurotrophin-3. The same group was imaged using rhodamine 123 to calculate mitochondrial inner membrane potential. Diabetes depolarized the potential, but
this was restored by low doses of insulin (47). Kim et al
reported that assays from ganglia of rats with diabetes of
4 to 6 weeks had declines in PKC isoenzyme protein
levels, particularly PKCb-II, but a higher relative membrane fraction of PKCb-II (48). cGMP levels were also
reduced in diabetics. It is, however, uncertain if the
changes were specific for neurons since whole ganglion
assays were used.
Given a superimposed axonal injury in diabetes, is the
sensory neuron capable of mounting an appropriate ‘‘injury phenotype’’? This does not appear to be the case.
Tomlinson et al noted, using Northern analysis, that upregulation of mRNA of GAP43/B50 and Ta-1 tubulin occurring after axotomy in diabetic rats (10 weeks) was
reduced compared to nondiabetic littermates (49). Neurofilament H and L and TrkA mRNA levels declined to
lower levels in diabetics than controls. Retrograde loss of
sensory neurons following nerve transection, prominent
in neonates, is relatively mild in adults. Unlike the relatively massive loss of neurons that develops retrogradely
in neonatal ganglia after axotomy, resistance to such loss
in adult rodents may be due to other sources of trophic
support besides that of target tissues. In diabetic rats, retrograde sensory neuron loss is much greater than in nondiabetic littermates, suggesting impairment of trophic
support (Kennedy and Zochodne, unpublished data).
Overall, the molecular phenotype of the diabetic sensory neuron does suggest a cell ‘‘under siege’’ from a
variety of stressors, at least for a period of time. In longterm (12 month) STZ diabetic rats, we found evidence of
increased activated caspase-3 expression, a feature usually considered to be pre-mortem (34) (Fig. 6). Despite
the presence of caspase-3, three-dimensional counting
(discussed above), light electron microscopic appraisal of
nuclear morphology, and TUNEL labeling failed to identify concurrent active apoptosis and loss (34, 35). Recent
J Neuropathol Exp Neurol, Vol 63, June, 2004
570
TOTH ET AL
work has clarified that caspase-3 expression need not necessarily specify imminent cellular demise (50), but what
prevents its full action in diabetic neurons is uncertain.
An interesting related finding in this cohort of rats was
the presence of heightened PARP (poly ADP-ribose polymerase) expression, an enzyme induced by DNA damage.
Might PARP, hitherto considered a mediator of cell death,
instead help to protect neurons? Another interesting finding was that there were rises in HSP 27 mRNA in sensory
neurons, in contrast to the nearly universal downregulation of most mRNA species examined (35). HSP27 is a
chaperone protein linked to neuron injury and survival
(51). We did not observe alterations in BCl-2 expression,
but the full repertoire of pro- and anti-apoptotic signals
in a relevant model of diabetes has not been examined.
Tomlinson also reported declines in the binding of NFkB
and CREB transcription factors, both of importance in
survival of neurons, from 12-week diabetic rat ganglia
(52).
How diabetes alters ion channel expression in sensory
neurons may be relevant in considering neuropathic pain,
a major clinical feature of diabetic neuropathies. While
changes of channel protein synthesis have a bearing on
their downstream investment into axons, their insertion
into perikarya membranes may have an equally important
role. Wall and Devor described the origin of ectopic discharges related to pain perception from ganglia neurons
(53). Behavioral evidence of pain, mostly mechanical allodynia, does develop in diabetic models (54, 55), although only relatively short duration models have been
studied. Craner et al described rises of both mRNA and
protein levels of the sodium channels Nav 1.3, 1.6, and
1.9 coinciding with mechanical allodynia, but downregulation of Nav 1.8 (56). Kim et al similarly reported decreased protein levels of Nav 1.8 (48). Nav 1.8 and 1.9
tetrodotoxin-resistant channels are expressed in small
sensory neurons, whereas Nav 1.3 is tetrodotoxin-sensitive and Nav 1.6 is expressed by myelinated axons at
nodes of Ranvier. Again, as with the molecular changes
described earlier, changes in sodium channel expression
differed from those expected of axon transection (axotomy), where declines of both Nav 1.8 and 1.9 would be
expected. The mRNA for the b3 auxiliary subunit of the
sodium channel may be upregulated selectively in medium sized Ad diabetic sensory neurons, whereas the b1
subunit was unchanged (57).
Hirade et al investigated the physiological consequences of 8 months of diabetes on isolated diabetic sensory
neurons (58). Overall, a rise in sodium current density
through tetrodotoxin-resistant channels and a leftward
shift in the voltage dependence of this conductance (to
more negative values) were noted. Both changes may increase the excitability of sensory neurons, but it is uncertain whether these physiological changes occur strictly
due to the rise in Nav1.9.
J Neuropathol Exp Neurol, Vol 63, June, 2004
Neurotrophic Growth Factors and Diabetic Sensory
Neurons
The potential role of the neurotrophin family (NGF,
BDNF, NT-3, NT-4/5) in supporting diabetic sensory neurons has been considered for some time (40) and has been
reviewed elsewhere. BDNF, a molecule that supports medium sized sensory neurons and motor neurons, has had
increases in its mRNA observed in DRGs and sciatic
nerves of experimental diabetic rats; this may add a protective or reparative mechanism for neurons (59). NT-3
supports large sensory neurons associated with myelinated proprioceptive afferents. Exogenous delivery has altered some features of experimental diabetes, including
conduction abnormalities, neurofilament accumulation,
and levels of MAPKs (44, 60, 61). Target tissue levels
and retrograde axonal transport of NT-3 appear to be reduced by diabetes (62).
Two non-neurotrophin family growth factors not considered in further detail here are CNTF and GDNF. CNTF
synthesis by diabetic Schwann cells may be impaired
(63). In response to nerve transection, GDNF mRNA expression in Schwann cells and DRG satellite cells increases (64).
Perineuronal satellite cells are an interesting population
of cells closely related to Schwann cells that surround,
support, and possibly protect neurons in sensory ganglia
(65). Features include an intimate association with neurons, scant cytoplasm, a high surface-volume laminar
structure, and a basement membrane (65). NGF and NT3 have both been observed to be upregulated in perineuronal satellite cells after injury, implying an important
local paracrine supportive role for sensory neurons (66).
In addition, such cells may scavenge free radicals and
diminish oxidative stress. It is not known whether diabetes targets these important cells.
Role of Insulin and Insulin-Like Growth Factor-1 in
Sensory Neurons
Hyperglycemia has long been considered the critical
insult to peripheral nerves that accounts for neuronal
damage. However, a loss of insulin signaling in Type I
diabetics or resistance to its action in Type II diabetes
might yet be relevant in the development of nerve disease. Insulin and closely related and regulated insulinlike growth factors belong to a family of growth factors
mediating metabolism, growth, cell differentiation, and
survival of most tissues in mammals. In the peripheral
nervous system, insulin and insulin-like growth factor-1
(IGF-1) are neurotrophic factors important for neuronal
survival and they influence phenotypic expression in
DRG neurons. Insulin and IGF-1 stimulate survival, neuritic outgrowth, and other cellular responses within the
adult sensory neuron (67). At low physiological insulin
571
DIABETES MELLITUS AND THE SENSORY NEURON
concentrations, insulin appears to stimulate neuritic outgrowth using insulin receptors, while at supraphysiological doses, insulin binds to type I IGF receptors (68).
Insulin binds to the a subunit of insulin receptors (IRs)
where tyrosine autophosphorylation of the b subunit is
promoted, subsequently leading to activation of the catalytic domain and phosphorylation of cellular substrates,
including the insulin receptor substrate (IRS) proteins and
Shc (69). The phosphorylation of IRS-1 or IRS-2 (also
called 4PS1 [70]), on multiple tyrosine residues creates
an active signaling complex via recruitment of various
proteins, including phosphatidylinositol 3 (PI3) kinase,
Grb2, Nck, Crk, Fyn, and SHP2. Insulin activity thus
provides signal transduction through PI3K activation as
well as Grb 2/Sos association with IRS-2/4PSI (70). The
Ca21-activated phospholipid-dependent PKC is stimulated by insulin and IGF-1.
IGF-1 is a polypeptide growth factor that is produced
and released from target tissues. IGF receptors are present
on neurons and assist in internalization and retrograde
transport of IGF (71). Binding of IGFs to a subunits of
IRs or type 1 IGF receptors occurs extracellularly and
activates the kinase within the b subunit situated intracellularly. The major substrates for the IGF-1 receptor are
Shc and IRS proteins. IGF-1 induces a sustained tyrosine
phosphorylation of Shc and also stimulates association
with Grb2 (72). Tyrosine phosphorylation of Shc occurs
in parallel with IGF-1-stimulated activation of extracellular signal-regulated kinase (ERK) (72). Alternatively,
IGF-1 induces a transient tyrosine phosphorylation of
IRS-2 and stimulates an association of IRS-2 with Grb2.
As well, immunolocalization studies show IRS-2 and
Grb2, but not Shc, to be concentrated at the tip of the
extending growth cone where membrane ruffling is most
active. This suggests that association of Shc with Grb2
is essential for IGF-1-mediated neurite outgrowth, while
the IRS-2-Grb2-PI3K complex may regulate growth cone
extension and membrane ruffling (72).
IGF-1 has been demonstrated to have anti-apoptotic
effects and anti-oxidant effects upon neuronal cells. Apoptotic effects of excessive glucose in cultured DRG neurons can be controlled with supplementation of IGF-1,
which does not occur with supplementation of NGF (37).
It is postulated that IGF-1 mediates its anti-apoptotic effect through the PI3K pathway, although this pathway
must have differences from pathways stimulated by NGF,
as these also act through the PI3K pathway (37). IGF-1
also upregulates uncoupling protein 3 (UCP3), a member
of the mitochondrial transporter superfamily found within
the inner membrane of mitochondria (73). UCP3 appears
to be important in the regulation of production of reactive
oxygen species in mitochondria, giving it a role as an
anti-oxidant in diabetes.
Within the DRG in experimental diabetic neuropathy,
changes occur in cellular expression of IGF-1; a preferential expression of IGF-1 and IGF receptors occurs in
small (,25 mm) diameter DRG diabetic neurons (71).
Loss of IGF neurotrophic support to DRG neurons may
possibly relate to perturbed regulation of sodium channel
expression and neuropathic pain observed in experimental models (56, 71).
CONCLUSION
Knowledge of the complexity of changes within the
diabetic DRG and their role in diabetic neuropathy is ever
increasing. Along with theories such as excessive polyol
flux, oxidative stress, protein glycosylation, and accelerated microangiopathy, there are additional pathophysiological explanations such as failed neurotrophic support,
downregulation and alterations of mRNA, and changes
in signaling pathways involving MAPK. The DRG may
be the most important site for continuing investigations
into the etiologies of diabetic neuropathy and is likely the
most important location for examining the prevention of
neuropathic disease in diabetics. It is probable that a convergence of separate pathophysiological mechanisms
leads to DRG dysfunction in diabetic neuropathy. Future
research into the importance of the DRG in diabetic neuropathy may need to examine the importance of neuroglial interactions and their role in the support of the DRG
neuron.
ACKNOWLEDGMENTS
CT is a Clinical Fellow of the Alberta Heritage Foundation for Medical Research (AHFMR). DZ is a Senior Medical Scholar of AHFMR.
Brenda Boake provided editing assistance for the manuscript.
REFERENCES
1. Malik RA, Veves A, Walker D, Siddique I, Lye RH, Schady W,
Boulton AJ. Sural nerve fibre pathology in diabetic patients with
mild neuropathy: Relationship to pain, quantitative sensory testing
and peripheral nerve electrophysiology. Acta Neuropathol 2001;
101:367–74
2. Malik RA, Newrick PG, Sharma AK, et al. Microangiopathy in
human diabetic neuropathy: Relationship between capillary abnormalities and the severity of neuropathy. Diabetologia 1989;32:
92–102
3. Zochodne DW. Nerve and ganglion blood flow in diabetes: An
appraisal. In: Tomlinson D, eds. Neurobilogy of diabetic neuropathy. San Diego: Academic Press, 2002:161–202
4. Theriault M, Dort J, Sutherland G, Zochodne DW. Local human
sural nerve blood flow in diabetic and other polyneuropathies. Brain
1997;120:1131–38
5. Sima AAF, Sugimoto K. Experimental diabetic neuropathy. Diabetologia 1999;42:773–88
6. Mizisin AP, Shelton GD, Wagner S, Rusbridge C, Powell HC. Myelin splitting, Schwann cell injury and demyelination in feline diabetic neuropathy. Acta Neuropathol 1998;95:171–74
7. Borghini I, Ania-Lahuerta A, Regazzi R, et al. Alpha, beta I, beta
II, delta, and epsilon protein kinase C isoforms and compound activity in the sciatic nerve of normal and diabetic rats. J Neurochem
1994;62:686–96
J Neuropathol Exp Neurol, Vol 63, June, 2004
572
TOTH ET AL
8. Greene DA, Lattimer SA, Sima AA. Sorbitol, phosphoinositides,
and sodium-potassium-ATPase in the pathogenesis of diabetic complications. N Engl J Med 1987;316:599–606
9. Sima AA, Sugimoto K. Experimental diabetic neuropathy: An update. Diabetologia 1999;42:773–88
10. Sima AAF, Brismar T, Yagihashi S. Neuropathies encountered in
the spontaneously diabetic BB Wistar rat. In: Dyck PJ, Thomas PK,
Asbury AK, et al, eds. Diabetic neuropathy. Toronto: W.B. Saunders, 1987
11. Cherian PV, Kamijo M, Angelides KJ, Sima AA. Nodal Na(1)channel displacement is associated with nerve- conduction slowing
in the chronically diabetic BB/W rat: Prevention by aldose reductase inhibition. J Diabetes Complications 1996;10:192–200
12. Low PA, Nickander KK, Tritschler HJ. The roles of oxidative stress
and antioxidant treatment in experimental diabetic neuropathy. Diabetes 1997;46:S38–S42
13. Hofmann MA, Drury S, Fu C, et al. RAGE mediates a novel proinflammatory axis: A central cell surface receptor for S100/calgranulin polypeptides. Cell 1999;97:889–901
14. Mallozzi C, Di Stasi AM, Minetti M. Peroxynitrite modulates tyrosine-dependent signal transduction pathway of human erythrocyte
band 3. FASEB J 1997;11:1281–90
15. Zochodne DW, Verge VM, Cheng C, et al. Nitric oxide synthase
activity and expression in experimental diabetic neuropathy. J Neuropathol Exp Neurol 2000;59:798–807
16. Bucala R, Tracey KJ, Cerami A. Advanced glycosylation products
quench nitric oxide and mediate defective endothelium-dependent
vasodilatation in experimental diabetes. J Clin Invest 1991;87:
432–38
17. Durante W, Sen AK, Sunahara FA. Impairment of endotheliumdependent relaxation in aortae from spontaneously diabetic rats. Br
J Pharmacol 1988;94:463–68
18. Kihara M, Low PA. Impaired vasoreactivity to nitric oxide in experimental diabetic neuropathy. Exp Neurol 1995;132:180–85
19. Simpson LO. Altered blood rheology in the pathogenesis of diabetic
and other neuropathies. Muscle Nerve 1988;11:725–44
20. Zochodne DW. Is early diabetic neuropathy a disorder of the dorsal
root ganglion? A hypothesis and critique of some current ideas on
the etiology of diabetic neuropathy. J Peripher Nerv Syst 1996;1:
119–30
21. Zochodne DW, Ho LT. Unique microvascular characteristics of the
dorsal root ganglion in the rat. Brain Res 1991;559:89–93
22. Zochodne DW, Ho LT. Neonatal guanethidine treatment alters endoneurial but not dorsal root ganglion perfusion in the rat. Brain
Res 1994;649:147–50
23. Greene DA, Winegrad AI, Carpentier JL, Brown MJ, Fukuma M,
Orci L. Rabbit sciatic nerve fascicle and ‘endoneurial’ preparations
for in vitro studies of peripheral nerve glucose metabolism. J Neurochem 1979;33:1007–18
24. Kadekaro M, Crane AM, Sokoloff L. Differential effects of electrical stimulation of sciatic nerve on metabolic activity in spinal
cord and dorsal root ganglion in the rat. Proc Natl Acad Sci USA
1985;82:6010–13
25. Zochodne DW, Ho LT. The influence of sulindac on experimental
streptozotocin-induced diabetic neuropathy. Can J Neurol Sci 1994;
21:194–202
26. Zochodne DW, Ho LT, Allison JA. Dorsal root ganglia microenvironment of female BB Wistar diabetic rats with mild neuropathy. J
Neurol Sci 1994;127:36–42
27. Takahashi K, Ghatei MA, Lam H-C, O’Halloran DJ, Bloom SR.
Elevated plasma endothelin in patients with diabetes mellitus. Diabetologia 1990;33:306–10
28. Bertello P, Veglio F, Pinna G, et al. Plasma endothelin in NIDDM
patients with and without complications. Diabetes Care 1994;17:
574–77
J Neuropathol Exp Neurol, Vol 63, June, 2004
29. Xu Q-G, Cheng C, Sun H, Thomsen K, Zochodne DW. Local sensory ganglion ischemia induced by endothelin vasoconstriction.
Neuroscience 2003;122:897–905
30. Greenbaum D, Richardson PC, Salmon MV, Urich H. Pathological
observations on six cases of diabetic neuropathy. Brain 1964;87:
201–14
31. Li X-G, Zochodne DW. Microvacuolar neuronopathy is a post-mortem artifact of sensory neurons. J Neurocytol 2003;32:393–98
32. Sasaki H, Schmelzer JD, Zollman PJ, Low PA. Neuropathology and
blood flow of nerve, spinal roots and dorsal root ganglia in longstanding diabetic rats. Acta Neuropathol 1997;93:118–28
33. Schmidt RE, Beaudet LN, Plurad SB, Dorsey DA. Axonal cytoskeletal pathology in aged and diabetic human sympathetic autonomic ganglia. Brain Res 1997;769:375–83
34. Cheng C, Zochodne DW. Sensory neurons with activated caspase3 survive long-term experimental diabetes. Diabetes 2003;52:
2363–71
35. Zochodne DW, Verge VMK, Cheng C, Sun H, Johnston J. Does
diabetes target ganglion neurons? Progressive sensory neuron involvement in long term experimental diabetes. Brain 2001;124:
2319–34
36. Kishi M, Tanabe J, Schmelzer JD, Low PA. Morphometry of dorsal
root ganglion in chronic experimental diabetic neuropathy. Diabetes
2002;51:819–24
37. Russell JW, Sullivan KA, Windebank AJ, Herrmann DN, Feldman
EL. Neurons undergo apoptosis in animal and cell culture models
of diabetes. Neurobiol Dis 1999;6:347–63
38. Medori R, Autilio-Gambetti L, Jenich H, Gambetti P. Changes in
axon size and slow axonal transport are related in experimental
diabetic neuropathy. Neurology 1988;38:597–601
39. Diemel LT, Brewster WJ, Fernyhough P, Tomlinson DR. Expression
of neuropeptides in experimental diabetes: Effects of treatment with
nerve growth factor or brain-derived neurotrophic factor. Brain Res
Mol Brain Res 1994;21:171–75
40. Brewster WJ, Fernyhough P, Diemel LT, Mohiuddin L, Tomlinson
DR. Diabetic neuropathy, nerve growth factor and other neurotrophic factors. Trends Neurosci 1994;17:321–25
41. Verge VMK, Xu Z, Xu XJ, Wiesenfeld-Hallin Z, Hokfelt T. Marked
increase in nitric oxide synthase mRNA in rat dorsal root ganglia
after peripheral axotomy: In situ hybridization and functional studies. Proc Natl Acad Sci USA 1992;89:11617–21
42. Purves T, Middlemas A, Agthong S, et al. A role for mitogenactivated protein kinases in the etiology of diabetic neuropathy.
FASEB J 2001;15:2508–14
43. Fernyhough P, Gallagher A, Averill SA, et al. Aberrant neurofilament phosphorylation in sensory neurons of rats with diabetic neuropathy. Diabetes 1999;48:881–89
44. Sayers NM, Beswick LJ, Middlemas A, et al. Neurotrophin-3 prevents the proximal accumulation of neurofilament proteins in sensory neurons of streptozocin-induced diabetic rats. Diabetes 2003;
52:2372–80
45. Brussee V, Cunningham FA, Zochodne DW. Direct insulin signaling
of neurons reverses diabetic neuropathy. Diabetes (in press)
46. Huang TJ, Sayers NM, Fernyhough P, Verkhratsky A. Diabetesinduced alterations in calcium homeostasis in sensory neurones of
streptozotocin-diabetic rats are restricted to lumbar ganglia and are
prevented by neurotrophin-3. Diabetologia 2002;45:560–70
47. Huang TJ, Price SA, Chilton L, et al. Insulin prevents depolarization of the mitochondrial inner membrane in sensory neurons of
type 1 diabetic rats in the presence of sustained hyperglycemia.
Diabetes 2003;52:2129–36
48. Kim H, Sasaki T, Maeda K, Koya D, Kashiwagi A, Yasuda H.
Protein kinase Cbeta selective inhibitor LY333531 attenuates diabetic hyperalgesia through ameliorating cGMP level of dorsal root
ganglion neurons. Diabetes 2003;52:2102–9
DIABETES MELLITUS AND THE SENSORY NEURON
49. Mohiuddin L, Tomlinson DR. Impaired molecular regenerative responses in sensory neurones of diabetic rats: Gene expression
changes in dorsal root ganglia after sciatic nerve crush. Diabetes
1997;46:2057–62
50. Hoeppner DJ, Hengartner MO, Schnabel R. Engulfment genes cooperate with ced-3 to promote cell death in Caenorhabditis elegans.
Nature 2001;412:202–6
51. Lewis SE, Mannion RJ, White FA, et al. A role for HSP27 in
sensory neuron survival. J Neurosci 1999;19:8945–53
52. Purves TD, Tomlinson DR. Diminished transcription factor survival
signals in dorsal root ganglia in rats with streptozotocin-induced
diabetes. Ann N Y Acad Sci 2002;973:472–76
53. Wall PD, Devor M. Sensory afferent impulses originate from dorsal
root ganglia as well as from the periphery in normal and nerve
injured rats. Pain 1983;17:321–39
54. Calcutt NA, Jorge MC, Yaksh TL, Chaplan SR. Tactile allodynia
and formalin hyperalgesia in streptozotocin- diabetic rats: Effects
of insulin, aldose reductase inhibition and lidocaine. Pain 1996;68:
293–99
55. Ahlgren SC, Levine JD. Mechanical hyperalgesia in streptozotocindiabetic rats. Neuroscience 1993;52:1049–55
56. Craner MJ, Klein JP, Renganathan M, Black JA, Waxman SG.
Changes of sodium channel expression in experimental painful diabetic neuropathy. Ann Neurol 2002;52:786–92
57. Shah BS, Gonzalez MI, Bramwell S, Pinnock RD, Lee K, Dixon
AK. Beta3, a novel auxiliary subunit for the voltage gated sodium
channel is upregulated in sensory neurones following streptozocin
induced diabetic neuropathy in rat. Neurosci Lett 2001;309:1–4
58. Hirade M, Yasuda H, Omatsu-Kanbe M, Kikkawa R, Kitasato H.
Tetrodotoxin-resistant sodium channels of dorsal root ganglion neurons are readily activated in diabetic rats. Neuroscience 1999;90:
933–39
59. Fernyhough P, Diemel LT, Brewster WJ, Tomlinson DR. Altered
neurotrophin mRNA levels in peripheral nerve and skeletal muscle
of experimentally diabetic rats. J Neurochem 1995;64:1231–37
60. Mizisin AP, Calcutt NA, Tomlinson DR, Gallagher A, Fernyhough
P. Neurotrophin-3 reverses nerve conduction velocity deficits in
streptozotocin-diabetic rats. J Peripher Nerv Syst 1999;4:211–21
61. Middlemas A, Delcroix JD, Sayers NM, Tomlinson DR, Fernyhough P. Enhanced activation of axonally transported stress-activated protein kinases in peripheral nerve in diabetic neuropathy is
prevented by neurotrophin-3. Brain 2003;126:1671–82
62. Fernyhough P, Diemel LT, Tomlinson DR. Target tissue production
and axonal transport of neurotrophin-3 are reduced in streptozotocin-diabetic rats. Diabetologia 1998;41:300–306
63. Calcutt NA, Muir D, Powell HC, Mizisin AP. Reduced ciliary neurontrophic factor-like activity in nerves from diabetic or galactosefed rats. Brain Res 1992;575:320–24
573
64. Hammarberg H, Piehl F, Cullheim S, Fjell J, Hokfelt T, Fried K.
GDNF mRNA in Schwann cells and DRG satellite cells after chronic sciatic nerve injury. NeuroReport 1996;7:857–60
65. Pannese E. The satellite cells of the sensory ganglia. Adv Anat
Embryol Cell Biol 1981;65:1–111
66. Zhou XF, Deng YS, Chie E, et al. Satellite-cell-derived nerve
growth factor and neurotrophin-3 are involved in noradrenergic
sprouting in the dorsal root ganglia following peripheral nerve injury in the rat. Eur J Neurosci 1999;11:1711–22
67. Recio-Pinto E, Rechler MM, Ishii DN. Effects of insulin, insulinlike growth factor-II, and nerve growth factor on neurite formation
and survival in cultured sympathetic and sensory neurons. J Neurosci 1986;6:1211–19
68. Recio-Pinto E, Ishii DN. Effects of insulin, insulin-like growth factor-II and nerve growth factor on neurite outgrowth in cultured
human neuroblastoma cells. Brain Res 1984;302:323–34
69. White MF, Yenush L. The IRS-signaling system: A network of
docking proteins that mediate insulin and cytokine action. Curr Top
Microbiol Immunol 1998;228:179–208
70. Patti ME, Sun XJ, Bruening JC, et al. 4PS/insulin receptor substrate
(IRS)-2 is the alternative substrate of the insulin receptor in IRS1-deficient mice. J Biol Chem 1995;270:24670–73
71. Craner MJ, Klein JP, Black JA, Waxman SG. Preferential expression of IGF-I in small DRG neurons and down-regulation following
injury. NeuroReport 2002;13:1649–52
72. Kim B, Cheng HL, Margolis B, Feldman EL. Insulin receptor substrate 2 and Shc play different roles in insulin-like growth factor I
signaling. J Biol Chem 1998;273:34543–50
73. Gustafsson H, Tamm C, Forsby A. Signalling pathways for insulinlike growth factor type 1-mediated expression of uncoupling protein
3. J Neurochem 2004;88:462–68
74. Scott JN, Clark AW, Zochodne DW. Neurofilament and tubulin
gene expression in progressive experimental diabetes. Failure of
synthesis and export by sensory neurons. Brain 1999;122:2109–17
75. Rittenhouse PA, Markle RA, Marchand JE, Kream RM, Leeman
SE. Streptozotocin-induced diabetes produces a decrease in pituitary substance P content and preprotachykinin mRNA. Neurosci
Lett 1996;211:77–80
76. Rittenhouse PA, Marchand JE, Chen J, Kream RM, Leeman SE.
Streptozotocin-induced diabetes is associated with altered expression of peptide-encoding mRNAs in rat sensory neurons. Peptides
1996;17:1017–22
77. Sugimoto K, Murakawa Y, Sima AA. Expression and localization
of insulin receptor in rat dorsal root ganglion and spinal cord. J
Peripher Nerv Syst 2002;7:44–53
78. Schmidt RE, Dorsey D, Parvin CA, Beaudet LN, Plurad SB, Roth
KA. Dystrophic axonal swellings develop as a function of age and
diabetes in human dorsal root ganglia. J Neuropathol Exp Neurol
1997;56:1028–43
J Neuropathol Exp Neurol, Vol 63, June, 2004