Diabetic sensory neuropathies are a common, clinically observed sequelae of hyperglycemia and are characterized by a progressive degradation of primary afferent function 12. Functional and structural evidence suggest an early and frequent involvement of small diameter primary sensory neurons leading to nociceptive abnormalities 234. In order to examine basic mechanisms underlying this disorder, we utilized the OVE26 transgenic mouse model of diabetes mellitus 56 to examine the effects of long-standing hyperglycemia on enzyme histochemical indicators of sensory neuron metabolism and evaluate the potential utility of this model for future studies of diabetic neuropathy. The OVE26 mouse line uses cell-specific overexpression of calmodulin to destroy pancreatic β-cells and the result is a viable diabetic mouse (>1 year survival) that displays both early-onset (<1 week after birth) and chronic elevated blood glucose (>500 mg/dl) and decreased serum and pancreatic insulin (<50% of normal) 56. Enzyme histochemical techniques demonstrated to be sensitive to neuronal perturbation 7 were used to examine the impact of long-standing hyperglycemia and hypoinsulinemia on the distribution and activity of lysosomal acid β-glycerophosphatase (AP), cytochrome oxidase (CO), and NADPH-diaphorase (NADPH-d; a correlate of nitric oxide synthase in aldehyde fixed tissue 8) in both sensory ganglia and the spinal cord. It has been previously demonstrated 711 that the metabolic status of sensory neurons, as reflected by the endogenous activity of specific homeostatic enzymes, is sensitive to injury and perturbation. Therefore, these enzymes were selected as putatively reflective of mitochondrial function (CO), lysosomal or degradative activity (AP) and primary sensory neuron injury or repair (NADPH-diaphorase).

The OVE26 transgenic mouse line (characterized by the insulin promoter-linked overexpression of calmodulin in pancreatic β-cells) used in this study displays a well-characterized chronic hyperglycemia and hypoinsulinemia within days after birth. 56. Ten (five OVE26 transgenic and five age-matched, control FVB animals) aged (>365 days old) mice were anesthetized with pentobarbital, perfused with 4% paraformaldehyde and the lumbar spinal cord and sensory ganglia removed, sectioned and processed for AP, CO or NADPH-d enzyme histochemistry as previously described 79101112. Counts of primary sensory somata were conducted on toluidine blue counterstained sections of L5 spinal ganglia and quantified using previously published methodologies 712.

Quantitative analysis of CO, AP and NADPH-d staining was undertaken on both the dorsal horn of the L5 segment of the spinal cord and the large and small cells of L5 sensory ganglion using previously described densitometric analysis 712. The entire mediolateral extent of lamina I to III was selected for staining intensity measurement. Statistical analyses (t-test, Mann-Whitney Rank Sum test, one way analysis of variance, Kruskal-Wallis analysis of variance on ranks, z-test of proportions) were conducted using SigmaStat (Jandel). Controls for densitometric analysis consisted of: 1) simultaneous sectioning and mounting of diabetic and control tissue on the same slide to ensure identical histological processing; 2) statistical analysis to verify consistency of staining between animals within control and experimental groups; 3) correction for small fluctuations in tissue opacity/thickness by subtractive illumination whereby the density value of white matter was subtracted from the immediately adjacent ventral horn; and 4) manual adjustment and calibration of the video camera parameters and microscope illumination and acquisition of all images using identical settings. All experiments were conducted in accordance with the guidelines of our institutions and the National Institutes of Health regarding the care and use of animals for experimental procedures.

Prior to fixative perfusion, the phenotypic status of OVE26 diabetic mice were confirmed by their characteristic small eyes caused by the GR19 gene in their transgenic construct 5. All adult OVE26 mice maintained fed blood glucose levels of at least 400 mg/dl. At the histological level, a survey 13 of the ratio of small (50 and 500 μm² area) to large (500 and 1950 μm² area) primary sensory somata in the fifth lumbar spinal ganglia revealed a significant decrease in the proportion of small to large cells in diabetic (1.29:1 small:large perikarya) compared to control (1.94:1 small:large perikarya) mice (P < 0.05 by z-test of proportions; 417 cells measured). Quantitative densitometric analysis of the abundance and distribution of enzyme histochemical reaction product in dorsal root ganglia (DRG) revealed substantive differences between diabetic and control mice (720 cells were quantified for both densitometry and cell size; 240 cells for each enzyme). Small somata from the ganglia of diabetic mice exhibited lower levels of AP (13.4% decrease; P < 0.001) and CO (Fig 1A,B; 9% decrease; P < 0.001) reaction product and an increase in the density of the reaction product for NADPH-d (Fig. 1C,D; 13.2% increase; P < 0.001) in comparison to control animals. No differences were observed in large diameter neurons from diabetic as compared to control animals.

In the spinal cord, all observed differences were confined to lamina I to III. Motoneuron somata in the ventral horn appeared both qualitatively and quantitatively similar in diabetic and control animals. In diabetic animals, there was an observable loss of AP reaction product in lamina I and II of the dorsal horn (Fig. 2A) as compared to control mice (Fig. 2B). Similarly, these laminae appeared to have qualitatively fewer NADPH-d labeled fibers and neuronal somata in diabetic (Fig. 2C) as compared to control animals. (Fig. 2D). The decrease of both AP and NADPH-d labeling was most profound in the medial portion of lamina I and II. Quantitative densitometric analysis supported the qualitative observations and revealed significantly reduced levels of AP (P = 0.026; 27 sections quantified) and NADPH-d (P < 0.001; 27 sections quantified) reaction product in lamina I and II of the dorsal horn of control and diabetic mice (Table 1). No significant differences were observed in qualitative staining appearance or intensity of CO reaction product labeling in the dorsal horn of diabetic, compared to non-diabetic animals (31 sections quantified).

Here we have demonstrated that chronic hyperglycemia has an impact on both the survival and metabolic profile of primary sensory neurons. The observed decrease in the ratio of small to large diameter primary sensory somata in diabetic animals most likely represents a loss of unmyelinated or small myelinated primary sensory neurons although a relative increase in the number of large myelinated neurons, however, unlikely, cannot be discounted. Nonetheless, the former interpretation is supported by the observed decrease in AP labeling in the dorsal horn of the spinal cord. The observed decrease in AP intensity in the surviving small neuronal somata from the DRG of the OVE26 animals, as compared to small neurons from control DRG, suggests that acid phosphatase activity in those cells is depressed. As FRAP containing sensory neurons represent a subpopulation of unmyelinated C-fibers 14, our results suggest that there is a loss, or at least a metabolic disruption of, unmyelinated neurons in the dorsal horn and in DRG. These results are consistent with the possibility of apoptosis and cell loss in the DRG of rodent models of diabetes 1516 although differences in model, species and duration of hyperglycemia must be considered 1718 along with the likelihood that there are a spectrum of different diabetic neuropathies including painful (small fiber involvement) and non-painful (large-fiber involvement) syndromes 23. In light of this, it should not be entirely unexpected that our observations of a putative small fiber disorder complements findings of large fiber disorders in alternate animal models of diabetes 19.

Cytochrome oxidase is the terminal enzyme in the electron transport chain, and is therefore considered to be a strong indicator of somatic mitochondrial activity. The decrease in CO staining within small DRG neurons, as compared to a similar cohort of small neurons from control animal ganglia, suggests a disruption of oxidative metabolism which corresponds well with results from other animal models of diabetes that demonstrate diminished CO activity or disruptions in mitochondrial morphology or function 16202122. Alternatively, our observed decrease in CO staining may reflect a simple decrease in mitochondrial number, as DRG neurons exposed to high glucose in vitro exposure contain fewer mitochondria 23. The lack of an observed change in CO activity in the dorsal horn is not unexpected as both physical (axotomy) and functional (tetrodotoxin) disconnection have previously been shown to leave CO activity in the dorsal horn unaltered 7.

In DRG and the dorsal horn of the spinal cord, NADPH-d activity levels have been previously shown to be responsive to peripheral neuronal injury or attenuation of electrical activity 7. Our results suggest that in addition to the pathological state that led to our observed loss of sensory neurons (and diminished NADPH-d labeling in the dorsal horn), there is a ongoing perturbation resulting in increased NADPH-d labeling in small DRG neurons from hyperglycemic animals as compared to the primary sensory somata from normoglycemic animals. As utilized here, NADPH-d enzyme histochemical reaction product represents nitric oxide synthase activity 8. The elevated ganglionic NADPH-diaphorase and diminished CO labeling in the DRG of hyperglycemic mice is consistent with previously proposed inhibition of CO activity 24 by the product of nitric oxide synthase, nitric oxide. Although the reported statistically significant changes may appear to be quantitatively modest, these percent changes represent group averages. Qualitatively and quantitatively, the changes are more pronounced in some animals and tissues sections, and less obvious in others. This is not unexpected as chronic diseases processes impact individuals with profound variability in both severity and temporal progress.

Our results suggest that the OVE26 model of chronic hyperglycemia does alter the overall neurochemical profile of the sensory nervous system through cell loss and/or altered enzyme activity and that this pathology seems to specifically impact unmyelinated and/or small myelinated primary sensory neurons.

The author(s) declare that they have no competing interests.

RZ completed this work as part of his doctoral dissertation and was involved in the writing of this manuscript and contributed both intellectually and practically to the content. PE created, characterized and supplied the transgenic mice and was also involved in the writing of this manuscript and contributed both intellectually and practically to the content. PC provided the lab, supervision, and support for this work, exclusive of that associated with generation and characterization of the mouse model. PC was also involved in the design and coordination of this study and participated in the writing of this manuscript and contributed both intellectually and practically to the content. All authors read and approved the final manuscript.