Myelinated Sensory Fibers
One of the earliest detectable structural changes in sensory myelinated fibers is the conspicuous enlargement or swelling of the nodal and paranodal axon, which correlates with the early Na+/K+-ATPase defect and increased axonal [Na+J (30,33) and is reversible after metabolic corrections. This abnormality is less frequent in early type 2 DPN in the BBZDR/Wor-rat, probably related to the milder defect in the Na+/K+-ATPase activity (31,56). Other early ultrastructural abnormalities observed in sensory nerves consist of glycogen accumulation in axonal mitochondria, so called glycogeno-somes, and misalignment of neuroskeletal structures (74,75). These structural changes are likely to reflect impaired axonal energy metabolism (74,76) and aberrant phosphorylation and nonalignment of particularly neurofilaments, respectively (75,77-79). Maligned neuroskeletal elements appear to induce phagocytotic activities by the Schwann cell that extends cytoplasmic loops engulfing abnormal axoplasm leading to so-called honeycombing of the axon (75). These early changes are associated with perturbed phosphorylation and synthesis of neurofilaments and decreased slow axonal transport leading to progressive axonal atrophy evident in the BB/Wor-rat already after 4 months of diabetes (Table 1) (45,80,81). Axonal atrophy leads to decreased circularity of axons associated with excessive myelin wrinkling best assessed in teased fiber preparations. The ultimate fate of these changes is axonal degeneration with secondary myelin breakdown and fiber loss. Significant fiber loss of 10% is already detectable in the sural nerve after 4 months of diabetes and increases to 33% after 11 months in the BB/Wor-rat (Table 1). Fiber loss in the tibial nerve at 11 months of diabetes is 12% and in the dorsal root 0% (81). Therefore, the degenerative processes progress in a proximal direction both in the peripheral as well as the central sensory axons, hence the characterization of DPN as a central-peripheral axonopathy of dying back type (81,82).
Primary demyelination is rare in type 1 DPN of the BB/Wor-rat, but somewhat more common in both human and experimental type 2 DPN (1,31). The reason for this is not well known, although more common comorbidities in both human and experimental type 2 diabetes such as hypercholesterolemia and triglyceridemia have to be considered.
In the BBZDR/Wor-rat the structural changes progress at a slower pace. So for instance, nodal/paranodal axonal swelling is only evident in 14-month diabetic rats, reflecting the milder Na+/K+-ATPase defect. Myelinated fiber atrophy of the sural nerve also progresses slower, becoming significant only after 6 months of diabetes and reaching a reduction of 11% at 14 months (cf. Table 1). At this time-point there is also a 10% loss of myelinated fibers in the sural nerve. On the other hand, segmental demyelination
Fig. 5. Axoglial dysjunction is a characteristic degenerative change of type 1 DPN. As the disease progresses (left to right, top panel), it results in a breach of the paranodal ion-channel barrier (red in the top panel), which connects the terminal myelin loops with the axolemma. This defect allows the nodal a-Na+-channels (yellow) to diffuse away from the node, resulting in severe nerve conduction deficits. Insulin receptors (IR, lower panel) are particularly clustered to the nodal axolemma and paranode. In type 1 DPN the expression of key nodal molecules such, as P-Na+-channel adhesive molecules (P), ankyrinG, contactin, and RPTPp is severely suppressed. Also, their ability to interact with neighboring proteins is hampered by impaired phos-phorylation. The expression of the pore forming a-Na+-channel (a) is not altered in type 1 DPN,
Fig. 5. Axoglial dysjunction is a characteristic degenerative change of type 1 DPN. As the disease progresses (left to right, top panel), it results in a breach of the paranodal ion-channel barrier (red in the top panel), which connects the terminal myelin loops with the axolemma. This defect allows the nodal a-Na+-channels (yellow) to diffuse away from the node, resulting in severe nerve conduction deficits. Insulin receptors (IR, lower panel) are particularly clustered to the nodal axolemma and paranode. In type 1 DPN the expression of key nodal molecules such, as P-Na+-channel adhesive molecules (P), ankyrinG, contactin, and RPTPp is severely suppressed. Also, their ability to interact with neighboring proteins is hampered by impaired phos-phorylation. The expression of the pore forming a-Na+-channel (a) is not altered in type 1 DPN, is three times as frequent in comparison with the type 1 BB/Wor-rat and affects 4.2% of all fibers in chronically diabetic BBZDR/Wor-rats (15,31). It is therefore obvious that the spectrum and progression of structural changes differ greatly between the two models despite the fact that cumulative hyperglycemic exposure is the same. These differences are not dissimilar from the differences previously described between human type 1 and type 2 DPN (1).
When initially described (1,55) these lesions were met with some controversy, because investigators failed to identify them in biopsy material from mainly type 2 diabetic patients or in acutely diabetic BB/Wor-rat (83-86). Following the recent elucidation of the longitudinal development of molecular changes underlying the nodal and paranodal abnormalities in type 1 and type 2 DPN (48,61), these changes are now firmly established.
The principal ultrastructural change is the breach of the junctional complexes adhering the terminal myelin loops to the paranodal axolemma, referred to as axoglial dysjunction (Fig. 5). This is followed by retraction of the myelin, paranodal demyelina-tion, a lesion which may be repaired by the lay down of small thinly myelinated intercalated internodes (55,61). Axoglial dysjunction is not specific for type 1 DPN, but occurs in a variety of neuropathies summarized by Yamamoto et al. (87).
The molecular compositions of the node of Ranvier and the paranodal apparatus are complex (Fig. 5) (88,89). Voltage-gated Na+ channels are concentrated to the nodal axolemma. They consist of the pore-forming Na+ channel a-subunit and two auxiliary Na+ channel subunits P1 and P2, which act as adhesive anchors (90) and interact with axonal ankyrinG (91). The P-subunits also interact with neurofascin, Nr-CAM, N-cadherin and L1 (89,92) mediating contacts with Schwann cell microvilli. Post-translational modification of ankyrinG by O-linked N-acetyl-glycosamine, inhibits phosphorylation of serin residues and prevents interaction with the P-subunits, and its interaction with receptor protein tyrosin phosphatase (RPTP)-P, which is mediated by tyrosine phosphorylation sites (61,93,94). It should be mentioned that the high affinity insulin receptor in peripheral nerve is concentrated to the nodal axolemma and colocalizes with axoglial junctions at the paranode (Fig. 5) (95).
At the paranode, myelin loops adhere to the axolemma through tight junctions constituting the paranodal ion-channel barrier separating nodal Na+ channels from juxtaparan-odal K+ channels. Caspr is the principal molecule of tight junctions and is coupled to contactin, serving as a receptor for RPTP-P (Fig. 5) (96,97). The cytoplasmic tail of caspr mediates protein-protein interactions by adducts of p85 to its protein 4.1 and SH3
Fig. 5. (Continued) although anchoring of the channel to the axolemma is compromised by the defective P-subunits. At the paranode, the backbone molecule of the tight junctions, caspr is responsible for the ion-channel barrier function and is compromised in its expression and post-translational activity by impaired binding of p85 to intracytoplasmic SH3 domains. Furthermore, caspr's interaction with contactin, RPTPp and P-Na+-channels is impaired. These molecular abnormalities lead to the progressive degeneration of the paranodal barrier allowing the now mobile a-Na+-channels to diffuse away from the nodal axolemma (upper panel). a: a-Na+-channel; P: P- Na+-channels; IR: insulin receptor. Reproduced by permission from ref. 48.
domains. p85 is the regulatory subunit of PI3-kinase, and is probably mediated by insulin signaling (Fig. 5) (61,97). Hence, the paranodal ion-channel barrier is not a static rigid structure but merely a complex of metabolically regulated protein-protein interactions.
In type 1 diabetic BB/Wor-rat the expression of ankyrinG, contactin, RPTPP, and P-Na+-channel subunits are unaltered after 2 months of diabetes, a time-point at which axoglial dysjunction is undetectable. However, at 8 months of diabetes, the expression of these molecules is significantly decreased. In addition, the glycation of ankyrinG is increased coupled with decreased phosphorylation (61), the socalled "yin-yang" relationship, hence compromising its protein-protein interaction. Interestingly the pore-forming a-Na+ channel expression is not altered. Of the paranodal constituents, caspr is unaltered at 2 months, but shows a 25% reduction in expression at 8 months at which time there is also marked suppression of contactin and RPTPP (61). Similar changes are not detectable in 8-month BBZDR/Wor-rats (61), suggesting that these aberrations in type 1 diabetes are associated with impaired insulin signaling. Indirect evidence for this construct is provided by the beneficial effects of insulinomimetic C-peptide (48,61), which prevents both the impaired expression of these molecular elements and maximizes their perturbed post-translational modifications, necessary for protein-protein interaction.
However, more chronic 14-month diabetic BBZDR/Wor-rats, do show evidence of nodal and paranodal degenerative changes, such as significant paranodal demyelination and increased frequencies of intercalated internodes (31).
Therefore, the progressive axoglial dysjunction first evident after 4 months of diabetes in the BB/Wor-rats has a molecular underpinning that appears to be caused by impaired insulin signaling, thereby, explaining the differences between DPN in type 1 and type 2 diabetes with respect to nodal and paranodal pathology.
The sensory C-fiber population is probably the most vulnerable anatomical component in DPN. This is evidenced by the occurrence of C-fiber related symptoms already during prediabetic conditions (72,73) and C-fiber atrophy in the glucose-intolerant and insulin deprived GK-rat (70).
In the type 1 BB/Wor-rat, sural nerve C-fibers show profound axonal atrophy and a 50% fiber loss at 8 months of diabetes (Fig. 6), leaving behind increased frequencies of denervated Schwann cell profiles and collagen pockets. The degeneration of C-fibers is associated with loss of mesaxonal junctional complexes of supporting Schwann cells, exposing C-fibers directly to the endoneurial environment, so-called type 2 axon/Schwann cell relationship (46). Interestingly, as with the paranodal ion-channel barrier, the insulin receptor colocalizes with C-fiber mesaxons (95,98). Whether insulin signaling abnormalities are of pathogenetic significance here as demonstrated for axoglial dysjunction has not been explored.
The C-fiber pathologies are associated with decreased sciatic nerve content of NGF and systemic insulin with consequent effects on the expression of SP and CGRP in dorsal root ganglion (DRG) cells.
In the type 2 BBZDR/Wor-rat the C-fiber population in sensory nerve is substantially less affected with normal axonal size and number (Fig. 6) even after 8 months exposure to severe hyperglycemia. Consequently the frequencies of denervated Schwann cell profiles and type 2 axon/Schwann cell relationships are not different from nondiabetic
control rats (46). The relative absence of significant C-fiber pathology is associated with normal expressions of neurotrophic factors and their receptors as well as normal content of SP and CGRP in dorsal root ganglia (46), and substantially milder functional deficits.
Therefore, the significant differences between C-fiber involvement between the two models necessitated investigations concerning whether insulin action may have an overriding impact. Type 1 BB/Wor-rat replenished with proinsulin C-peptide (71) for 8 months demonstrated corrections of the DRG insulin receptor, IGF-1R, TrkA and Trk C-receptors, and normalization of peripheral NGF and NT-3. These effects resulted in normal SP and CGRP contents in DRG's and prevention of C-fiber loss and partial prevention of C-fiber atrophy (Fig. 6) (47), resulting in a functionally mild C-fiber neuropathy similar to that seen in the type 2 BBZDR/Wor-rat. Hence, these data strongly suggest that C-fiber neuropathy, resulting in painful diabetic neuropathy is caused mainly by impaired insulin and C-peptide signaling mediated through its effect on the expression of neurotrophic factors with secondary effects on SP and CGRP positive neurons.
DRG apoptosis has been implied as a potential component in sensory neuropathies. In the BB/Wor-rat these data has not been confirmed. Rather apoptotic stresses occurring in DRG neurons in the BB/Wor-rat appears to be completely counter balanced by various survival factors (99).
Was this article helpful?
This guide will help millions of people understand this condition so that they can take control of their lives and make informed decisions. The ebook covers information on a vast number of different types of neuropathy. In addition, it will be a useful resource for their families, caregivers, and health care providers.