Int J Clin Exp Med 2019;12(6):7038-7046www.ijcem.com /ISSN:1940-5901/IJCEM0086000

Original Article Loss of cutaneous unmyelinated and myelinated fibers in streptozotocin-induced diabetic rats: a time course study

Qin Wang*, Shuai Wu*, Jing Ding, Ji-Hong Dong#, Xin Wang#

Department of Neurology, Zhongshan Hospital Fudan University, Shanghai, China. *Equal contributors. #Equal contributors.

Received September 26, 2018; Accepted April 8, 2019; Epub June 15, 2019; Published June 30, 2019

Abstract: Background: Quantification of density of intraepidermal nerve fibers (IENFD), targeting mainly unmyelin-ated C fibers using skin biopsies, has emerged as a promising method for assessment of diabetic peripheral neu-ropathy. However, detailed histological changes in unmyelinated fibers in combination with large myelinated fibers in rat models have not been adequately explored. Methods: The current study assessed both cutaneous unmyelinated and myelinated fiber changes, as well as pain-related behavior, in an 8-week streptozotocin (STZ)-induced rat model of diabetic mellitus (DM). Mechanical allodynia and thermal hyperalgesia were evaluated at baseline and weekly after injections with STZ. Quantification of IENFD, dermal nerve fiber (DNF) density, and density of intrapapillary myelinated endings (IMEs) was conducted at weeks 4, 6, and 8 post-injection. Results: Mechanical allodynia started at 1-week post-injection. It progressed through week 5, then plateaued at week 8. Thermal hyperalgesia occurred at week 4 and only lasted through week 5 in DM rats. IENFD reduction had already occurred at week 4. However, both DNFL and IME densities were still comparable at that time. IME density in the DM rats started to decrease right after week 4. It was significantly reduced at week 8. DNFL reduction in the DM rats did not occur until week 8. Conclusion: Present results show early persistent mechanical allodynia and delayed temporary hyperalgesia in this STZ-induced diabetic rat model. Significant IENF loss preceded changes in IMEs and DNFs, suggesting that IENFD quantification might be a potential marker for detection of early diabetic neuropathy.

Keywords: Diabetic neuropathy, mechanical allodynia, thermal hyperalgesia, intraepidermal nerve fiber density (IENFD), dermal nerve fibers (DNF), intrapapillary myelinated endings (IME)

Introduction

Diabetic peripheral neuropathy (DPN) is a com-mon complication of diabetes, affecting nearly half of all diabetic patients over the course of the disease [1]. Patients with DPN may have a variety of symptoms ranging from hyperalgesia, allodynia, and spontaneous pain to progressive hypoalgesia in their limbs [2]. However, the rea-sons why pain symptoms occur in certain gr- oups of patients and the relationships between pain generation and peripheral nerve damage remain unclear [3]. Due to length-dependent characteristics of diabetic neuropathy, the dis-tal parts of the limbs are first affected. This could affect both small-caliber fibers (such as unmyelinated C fibers and thinly myelinated Aδ fibers, which receive thermal and noxious stim-

uli [4]) and large myelinated Aβ fibers innervat-ing cutaneous mechanoreceptors, known as Meissner corpuscles (MCs) [5, 6]. Understanding the underlying pathological changes of DPN is critical in advancing optimal treatment.

Skin biopsies have become a commonly used alternative approach to assess pathophysiolog-ical changes in distal part of the limbs [7]. Qu- antification of density of intraepidermal nerve fibers (IENF) (the peripheral termini of nocicep-tive unmyelinated C fibers and thinly myelinated Aδ fibers) by skin biopsies has now become the gold standard method for assessment of dia-betic peripheral neuropathy [7, 8]. Reduction of IENF densities (IENFD) was found in both dia-betic patients and animal models, suggesting that loss of IENF could be an early indication of

A time course study of diabetic neuropathy in a rat model

7039 Int J Clin Exp Med 2019;12(6):7038-7046

diabetic neuropathy [4, 9-11]. Despite intensive efforts in studying IENF, few studies have evalu-ated large myelinated fibers. One study found a significant reduction of intrapapillary myelinat-ed endings (IMEs) (large myelinated fibers in- nervating MCs) in patients with diabetic neu-ropathy [6]. Dermal nerve fiber (DNF) quantifi-cation has been used to detect nerve fiber bundles in the superficial dermis [12]. However, there remains a lack of studies concerning par-allel histological changes of both myelinated and unmyelinated fibers in diabetic neuropa- thy.

Therefore, the present study aimed to compare the progression of both cutaneous unmyelinat-ed and myelinated fiber changes (IENF, DNF, and IMEs densities) in an 8-week streptozoto-cin (STZ)-induced diabetic rat model. Addition- ally, pain-related behavioral changes (mechani-cal allodynia and thermal hyperalgesia) were assessed.

Materials and methods

Animals

Fifty 4-week old Sprague-Dawley rats were pur-chased from Shanghai Laboratory Animal Cen- ter of China Academy of Science. They were housed in a hooded cage at room temperature (RT) with a 12-hour light-dark cycle. They were provided food and water ad libitum. Experi- mental diabetes was induced in 28 rats by sin-gle intraperitoneal injections of 60 mg/kg STZ per rat (Sigma-Aldrich, St Louis, MO, USA) after 24-hour fasting. The STZ was dissolved in a 0.1 M sodium citrate buffer, with PH 4.5 (the DM group). The remaining 22 rats were injected with sodium citrate, without STZ, serving as normal controls (control group). Fasting blood glucose levels were measured 72 hours after injections and every two weeks thereafter, en- suring a constant hyperglycemic status during the study. Experimental diabetes was estab-lished for rats with a fasting blood glucose level >16.6 mmol/L. These rats were included in the DM group.

All experimental procedures were approved by the Animal Care and Use Committee of Fudan University and were in accordance with NIH Guidelines for the Care and Use of Laboratory Animals.

Behavioral testing

Mechanical allodynia and thermal hyperalgesia were assessed before STZ injections, estab-lishing the baseline level. They were assessed weekly, after injections, for 8 weeks. Mechanical allodynia was assessed by measuring the 50% hind paw withdrawal threshold using calibrated von Frey filaments, according to the up-down method of Dixon. Briefly, rats were placed in wire-mesh bottomed plastic chambers. They were acclimated for 30 minutes. A series of 8 calibrated von Frey filaments of increasing stiff-ness levels (0.4, 0.6, 1.4, 2, 4, 6, 8, and 15 g, Stoelting, Wood Dale, Illinois, US) were applied to the mid-plantar hind paw of each rat, perpen-dicularly, with sufficient force to cause a slight bending of the filament against the paw. It was held there for 4-6 seconds, with 10-minute intervals between applications of 2 filaments. The test started with the 2 g filament. A sharp withdrawal of the paw or immediate paw flinch-ing upon removal of the filament was consid-ered a positive response. Stimuli were applied consecutively, either in an ascending order in the absence of positive response or in a de- scending order if a positive response was ob- served. Testing consisted of five more stimuli after the first change in response occurred. The pattern of response was converted to a 50% paw withdrawal threshold (50% PWT) using the method of Dixon [13].

Paw withdrawal latency (PWL) to radiant heat was measured to assess thermal hyperalgesia, using the Model 390 Paw Stimulator Analge- sia Meter (IITC/Life Science Instruments, USA). The rats were placed on a glass platform in tr- ansparent plastic chambers and allowed to acclimate for at least 30 minutes. Radiant heat was applied to the plantar surface of the hind paws until the rat removed its paw from the glass. Both hind paws were tested, indepen-dently, with a 10-minute interval. PWL was de- fined as the time from application of the heat to paw removal. The heat had a constant intensity set to elicit a 10-12 second PWL in rats in the control group. The maximal time duration for heat application was 20 seconds, avoiding tis-sue damage. A total of 5 tests were conducted. The longest and shortest PWL scores for each hind paw were excluded. Final PWL was calcu-lated as the average of the remaining 3 PWLs [14].

A time course study of diabetic neuropathy in a rat model

7040 Int J Clin Exp Med 2019;12(6):7038-7046

Immunohistochemistry of skin biopsies

Biopsy samples were obtained at weeks 4, 6, and 8 after injections for immunohistochemis-try examinations. The rats were given an over-dose of chloral hydrate. Glabrous skin in the middle of their hind limbs was separated and dissected. The glabrous skin was chosen be- cause, unlike hairy skin, it not only has unmy-elinated C fibers and thinly myelinated Aδ fibers, but also large myelinated Aβ fibers innervating MCs or Merkel cells. Dissected glabrous skins were immediately fixed in Zomboni’s fixative (2% paraformaldehyde-picric acid) at 4°C for 24 hours. They were then cryoprotected in 20% sucrose -0.1 M phosphate buffer (PH 7.4) for 24-48 hours. Next, 40 μm-thick sections trans-verse to the epidermis were cut and stored at -20°C for further immunohistochemistry anal- ysis.

Six sections from a single footpad of a rat were randomly chosen for immunohistochemistry analysis using a free-floating protocol [15, 16]. Sections were blocked with 4% normal goat serum in 0.01 M PBS - 0.3% Triton X-100 for 2 hours at RT, then incubated at 4°C overnight in rabbit polyclonal anti-PGP 9.5 antibody (1:1000, Chemicon). This was conducted to stain all nerve fibers or mouse anti-myelin basic protein (MBP) primary antibody (1:200, Che- micon), observing the myelinated nerve fibers. After rinsing with PBS, the sections were incu-bated with biotinylated goat anti-rabbit IgG or biotinylated goat anti-mouse IgG (1:1000, Vector) at RT for 1 hour. They were quenched in 30% methanol/1% hydrogen peroxide for 30 minutes and incubated in avidin-biotin complex (Vector) for 1 hour. The final reaction product was revealed by the blue chromogen/peroxi-dase substrate (Vector SG substrate kit). Sec- tions were then mounted, air-dried, dehydrated in graded alcohol, cleared in xylene, and co- ver-slipped.

Quantification of nerve fiber loss

Quantification of IENFD, DNF, and IMEs densi-ties was conducted at weeks 4, 6, and 8 post-STZ injection.

IENFD is defined as the number of IENF per length of a section (IENF/mm). It was quantified and calculated by an observer blinded to the grouping of the rats using bright-field immuno-

histochemistry, according to the methods of Lauria et al. [16]. The number of PGP 9.5 posi-tive IENFs in each section was counted under a light microscope at high power magnification (400×). Individual IENFs crossing the dermal-epidermis junction were counted. Secondary branching within the epidermis was not includ-ed in the counting. Images were used to mea-sure the length of the epidermal surface. IENFD was calculated in at least 3 sections from 1 footpad, arriving at an average IENFD.

Quantification of DNF was also conducted, as described previously. Briefly, a dermal area of interest (AOI) 200 μm below the dermal-epider-mal junction was drawn and the length of each DNF was measured. DNF density is defined as the sum of the length of each DNF (DNFL) divid-ed by the area of AOI (mm2) [12].

Loss of myelinated fibers was evaluated by quantification of density of IMEs innervating MCs. The number of IMEs was counted using bright-field immunohistochemistry on at least 3 MBP-stained sections for each footpad of a rat. The density of IMEs is defined as the number of IMEs per length of the epidermis (IMEs/mm) [5].

Statistical analysis

SPSS 17.0 for Windows was used for data anal-ysis. Results are presented as mean ± stand-ard deviation (SD). For parameters with multi-ple time points, repeated measure analysis of variance (ANOVA) was applied. For inter-group comparisons, two-tailed student’s t tests were used for data with normal distribution and Mann-Whitney U-tests were used for data with non-normal distribution. P<0.05 indicates sta-tistical significance.

Results

Establishment of an STZ-induced diabetic rat model

Male Sprague-Dawley rats were divided into 2 groups, the DM group (n=28) and control group (n=22) as described previously. In this study, a single injection of STZ induced persis-tent hyperglycermia (fasting blood glucose >16.6 mmol/L) in 82.1% of injected rats, indi-cating that the hyperglycemia model was suc-cessfully established in the DM group. Rats in

A time course study of diabetic neuropathy in a rat model

7041 Int J Clin Exp Med 2019;12(6):7038-7046

pared to the control gro- up (P<0.001). This variati- on changed with time (P= 0.039). Mechanical allody- nia occurred 1 week after STZ injections in the DM group (50% PWT: 9.15± 1.81 g vs 9.85±2.02 g, P=0.51 at baseline, and 6.81±2.54 vs 10.55±2.14 g, P=0.013 at week 1 for DM and controls, respec-tively). It progressed throu- gh week 5 (50% PWT: 3.7± 0.93 g vs 10.56±1.62 g, P<0.001, at week 5 for DM and controls, respectively), then plateaued afterward (50% PWT: 3.81±0.93 g vs 10.23±2.10 g, P<0.001, at week 8 for DM and con-trols, respectively) (Figure 2A).

Mean paw withdrawal late- ncy was not different bet- ween the DM group and control group (P=0.059). Thermal hyperalgesia was not obvious in the DM gr- oup, except for weeks 4 and 5, in which significantly decreased PWL was obser- ved. This indicated the pr- esence of thermal hyperal-gesia (PWL 13.06±1.96 s

the DM group grew substantially slower than those in the control group and weighed less (Figure 1A). Hyperglycemia in DM rats occurred around 1 week after injection. Fasting blood glucose levels increased progressively through-out the study period, while fasting blood glu-cose levels in control rats was maintained below 10 mmol/L throughout the study period (Figure 1B).

Mechanical allodynia and thermal hyperalge-sia in DM rats

Eight DM rats and 6 control rats were tested for mechanical allodynia and thermal hyperalge-sia. Mechanical allodynia, defined as signifi-cantly reduced 50% paw withdrawal threshold (50% PWT), was reduced in the DM group, com-

vs 13.52±1.30 s, P=0.51 at week 0, 10.44± 1.34 s vs 13.09±1.51 s, P=0.005 at week 4 and 10.01±1.66 s vs 12.53±0.57 s, P=0.009 at week 5 for DM and controls, respectively) (Figure 2B).

Progressive skin denervation in DM rats

Fifteen rats in each group (4 for week 4, 4 for week 6, and 7 for week 8) were used for quan-tification of IENFD, DNF, and IMEs densities.

Figure 3A and 3B show PGP 9.5-stained sec-tions, identifying IENF and DNF. Moderately to heavily stained nerve fibers were seen abun-dantly near the dermal-epidermal junction (Figure 3A). These epidermal nerve fibers trav-elled in bundles in the dermis layer, penetrated

Figure 1. Body weight and fasting blood glucose levels of DM and control rats. A. Body weight changes after streptozotocin (STZ) (DM) or sodium citrate (Control) injections in 4-week old Sprague-Dawley rats. B. Fasting blood glucose changes in the DM rats and control rats.

Figure 2. Behavioral findings of the STZ-induced diabetic rat model from week 0 to week 8. A. 50% paw withdrawal threshold (mechanical threshold) for the DM group (n=8) and the control group (n=6). B. Paw withdrawal latency (PWL) for the DM group and control group. Data are expressed as means ± standard deviation (SD), P represents differences between the DM and control groups, *P<0.05, **P<0.01, ***P<0.001.

A time course study of diabetic neuropathy in a rat model

7042 Int J Clin Exp Med 2019;12(6):7038-7046

the dermal-epidermal junction, and terminated at the surface of the epidermis in a perpendicu-lar fashion (Figure 3B). Compared to the con-trol group, significantly reduced IENFD was observed in the DM group at week 4 (30.23± 2.38 vs 35.19±2.82 fibers/mm for DM and controls, respectively, P<0.05). IENF loss pro-gressed through week 8 (27.39±2.25 vs 35.23±3.54 fibers/mm, P<0.01 at week 6, and 26.18±4.2 vs 34.84±4.31 fibers/mm, P<0.001 at week 8). At week 8, IENFD in the DM group was 24.9% less than that in the control group (Figure 4).

PGP 9.5 stained nerve bundles were also pre-sent in the subepidermal area. DNF densities (DNFL/mm2) were comparable between the two groups at weeks 4 and 6. At week 8, the DM group had a significantly reduced DNF den-sity, compared to the control group (7.27±0.92 vs 8.61±1.15 mm/mm2 for DM and controls respectively, P<0.05) (Figure 4), indicating da- mage and loss of superficial dermal nerve fib-ers at week 8 in DM rats.

MBP-stained sections identified abundant MBP-positive myelinated fibers in the upper dermis of the footpad. Some ascended into the dermis papillary and innervated the Meissner corpuscles (Figure 3C-E). IMEs densities and integrity levels were comparable between the two groups at week 4 (8.85±1.52 vs 10.45± 1.57 fibers/mm, P=0.195 for DM and controls, respectively). However, IME densities in DM rats decreased progressively over the next sev-eral weeks. At week 8, significantly reduced IMEs densities were observed in the DM group (7.1±1.38 vs 11.51±0.89 fibers/mm, P<0.001 for DM and controls, respectively) (Figure 5).

IENFD reduced prior to DNFL and IMEs density

At week 4, significant reduction of IENFD had already taken place in DM rats and progressed through week 8. In contrast, DM rats and con-trol rats had comparable DNFL and IMEs densi-ties at week 4. IMEs densities in DM rats start-ed to decrease after week 4. At week 8, IMEs densities in DM rats were significantly reduced.

Figure 3. Normal appearance of epidermal and dermal nerve fiber profiles in glabrous skin from hind paw of the rats. (A, B) were PGP-9.5 immuno-stained sections. (A) In low magnification field (×100), moderately to heavily stained nerve fiber bundles were seen abundantly near the dermal-epidermal junctions and run through the sub-papillary dermis. (B) In high magnification field (×400), IENF can be seen cross the dermal-epidermal junctions and run towards the skin surface with multiple branches. Arrow heads-nerve bundles in the dermis; Arrows-IENF. (C-E) were MBP immune-stained sections. (C) In low magnification field (×100), MBP stained myelinated nerve fibers can be seen abundantly in the upper dermis of the footpad. (D, E) In high magnification field (×400), myelinated fibers can be seen running in the upper dermis (D), ascending into the dermis papillary and innervated the Meissner corpuscles (E). Arrow heads-dermis myelinated fibers, Arrows-intrapapillary myelinated endings. The bar represents 50 um.

A time course study of diabetic neuropathy in a rat model

7043 Int J Clin Exp Med 2019;12(6):7038-7046

to generalized nerve fiber denervation in this painful diabetic neuropathy model.

Discussion

The present study exam-ined pain-related behavio-ral and cutaneous patho-logical changes in a STZ-induced diabetic rat model during an 8-week study pe- riod. Two key results were found: (1) STZ-induced DM rats exhibited early persis-tent mechanical allodynia and delayed temporary hy- peralgesia; and (2) Signifi- cant IENF loss preceded changes in IMEs and DNF, indicating that unmyelinat-ed fibers were damaged earlier in diabetes and IE- NFD might be a potential marker for detection of ea- rly diabetic neuropathy.

Behavioral results in mec- hanical sensitivity profiles were consistent with most previous studies, showing significant mechanical all- odynia early in the disea- se course [17-19]. However, thermal hyperalgesia only appeared for a short period in the current study. Beha- vioral changes of diabet- ic rodents are complicated and vary in different rat or mice models. It begins with an acute metabolic pha- se, featuring the slowing of nerve conduction and hy- peralgesia. These are usu-ally reversible. With the pro-gression of the disease co-

Significant DFNL reduction in DM rats also did not occur until week 8. Epidermal unmyelinated fibers were damaged prior to myelinated fibers in the dermis papillary and nerve fiber bundles in the subepidermal area. In week 8, all kinds of cutaneous nerve fibers were reduced. These results suggest that cutaneous nerve fiber loss began with unmyelinated fibers and progressed

urse, more severe functional abnormalities develop in concert with onset of progressive structural changes in the nerves [20]. Different pathogenic mechanisms may produce multiple manifestations of neuropathy. The current DM rat model showed definite mechanical allodynia and thermal hyperalgesia in week 4 to week 6, making it more useful to simulate the clinical

Figure 4. Changes of IENFD and DNF densities in DM and control groups at different time points. (A-F) PGP9.5-immunostained skin sections of control and diabetic rats 4 weeks (A, B), 6 weeks (C, D), 8 weeks (E, F) after modeling. (G, H) Bar graphs show differences of IENFD (G) and DNFL (H) between DM and con-trol group in 4 weeks, 6 weeks, and 8 weeks after modeling. Data are expressed as means ± standard deviation (SD), P represents differences between the DM and control groups, *P<0.05, **P<0.01, ***P<0.001. The bar represents 50 μm.

A time course study of diabetic neuropathy in a rat model

7044 Int J Clin Exp Med 2019;12(6):7038-7046

a reduction in epidermal in- nervations in rat models of diabetic neuropathy [9, 17, 21]. The current study, however, is the first to in- vestigate concurrent chan- ges of myelinated fibers. One study of diabetic mice showed that densities of myelinated fibers were also reduced with the behavior of mechanical sensory lo- ss. Pathological changes could be improved with ne- urotrophin treatment [22]. The current study used IM- Es as a target to evaluate the myelinated Aβ fibers in- nervating MCs. It was found that this specific kind of fi- ber was preserved in the early stage of diabetes. It began to show a tendency toward decline in week 6. DNFL reflects damage of nerve bundles, which inclu- des both unmyelinated and myelinated fibers in the su- bepidermal area [12]. DNFL reduction appeared even later in week 8. Data sug-gests an early apparent de- generation of unmyelinated nerve axon in the epider-mis. This is consistent with the ‘dying-back phenome-non’ of diabetic neuropa-thy. The early preservation of myelinated fibers indicat-ed that myelin may have the function of protecting nerve axons against meta-bolic injuries.

Functional consequences of cutaneous nerve fiber loss are complicated. The

manifestation of patients with painful diabetic neuropathy instead of insensate neuropathy.

Present results showed that significant IENF loss preceded changes in IMEs and DNF, indi-cating that intraepidermal unmyelinated fiber loss happened before the impairment of myeli-nated fibers. Previous studies have also found

current study observed an early occurrence of mechanical allodynia in diabetic neuropathy, in accord with previous reports. These studies placed the onset of mechanical allodynia at 1 week after STZ-injections and fully-developed allodynia by 2-8 weeks [17-19]. Some studies have suggested that mechanical allodynia co- uld possibly result from a direct effect of hyper-

Figure 5. Changes in IME densities in DM and control groups at different time points. (A-F) MBP-immuno-stained skin sections of control and diabetic rats 4 weeks (A, B), 6 weeks (C, D), 8 weeks (E, F) after modeling. (G) Bar graph shows differences of IME between DM and control group in 4 weeks, 6 weeks, and 8 weeks after modeling. Data are expressed as means ± standard deviation (SD), P represents differences between the DM and control groups, *P<0.05, **P<0.01, ***P<0.001. The bar represents 50 μm.

A time course study of diabetic neuropathy in a rat model

7045 Int J Clin Exp Med 2019;12(6):7038-7046

glycemia on the peripheral nervous system, rather than from a structural deficit [18]. In the current study, mechanical allodynia progressed through week 5, then plateaued until the end of the 8-week study. Thermal hyperalgesia app- eared in week 4 and week 5. Pathological ex- aminations also showed IENF reduction in week 4, indicating that IENF loss might be relevant to the maintenance of mechanical allodynia and generation of thermal hyperalgesia. However, the progression of IENF loss and IMEs loss from week 5 to week 8 was not paralleled by the behavioral manifestation. With the progres-sive loss of cutaneous nerve fibers, mechanical allodynia plateaued and thermal hyperalgesia behavior even disappeared. This indicates a dissociation between fiber loss and behavioral manifestation. Previous studies have also re- ported an onset of behavioral deficits prior to quantifiable intraepidermal fiber loss [9, 23]. In addition, a consensus has not yet be reached concerning whether IENFD is different between patients with painful and painless diabetic neu-ropathy and whether IENF loss could possibly increase the risk of neuropathic pain [3, 24]. This dissociation indicates that structure ch- anges may not be functionally relevant. In addi-tion to overt fiber loss, it is likely that metabolic damage leading to dysfunction of intact fibers may contribute to pain behavior. This dysfunc-tion could include electrophysiological or neu-rochemical changes [21, 25]. Future studies should investigate the expression of functional proteins related to pain on the nerve terminals, such as voltage-gated sodium channels or TRPVs.

The current study was limited by the short ob- servational duration. Additionally, the first me- asured time point for skin pathological exami-nations was set at week 4 post-STZ injections, in accord with most previous reports. Therefore, whether IENFD changes happen before week 4 deserves further investigation. More behavioral phenotype and skin pathological studies con-cerning structural and pain-related functional proteins are warranted in both DPN patients and animal models.

Conclusion

In present study, STZ-induced diabetic rats de- veloped persistent mechanical allodynia thr- oughout the 8-week test period. However, ther-mal hyperalgesia developed much later and

had a shorter duration. Free intraepidermal nerve fibers were significantly reduced in week 4, while dermal nerve fibers and intrapapillary myelinated fibers were damaged later in week 8. Significant IENF loss preceded changes in IMEs and DNF, suggesting that IENFD quantifi-cation might be a potential marker for detec-tion of early diabetic neuropathy.

Acknowledgements

The authors would like to thank Prof Wenli Mi and Yanqing Wang in the Department of In- tegrative Medicine and Neurobiology, Fudan University, for kindly providing the place, equip-ment, and technique support for the experi-ment. The work was supported by funding from the National Science Foundation of China (81- 671597) and the Youth Grant from Zhongshan Hospital, Fudan University, China (2014ZSQN- 41).

Disclosure of conflict of interest

None.

Address correspondence to: Ji-Hong Dong and Xin Wang, Department of Neurology, Zhongshan Hos- pital Fudan University, Shanghai 200032, China. Tel: 86-21-64041990-2973; Fax: 86-21-64041990; E- mail: dong.jihong@zs-hospital.sh.cn (JHD); Tel: 86- 21-64041990-2976; Fax: 86-21-64041990; E-mail: wang.xin@zs-hospital.sh.cn (XW)

References

[1] Callaghan BC, Cheng HT, Stables CL, Smith AL and Feldman EL. Diabetic neuropathy: clinical manifestations and current treatments. Lancet Neurol 2012; 11: 521-534.

[2] Peltier A, Goutman SA and Callaghan BC. Painful diabetic neuropathy. BMJ 2014; 348: g1799.

[3] Sorensen L, Molyneaux L and Yue DK. The re-lationship among pain, sensory loss, and small nerve fibers in diabetes. Diabetes Care 2006; 29: 883-887.

[4] Shun CT, Chang YC, Wu HP, Hsieh SC, Lin WM, Lin YH, Tai TY and Hsieh ST. Skin denervation in type 2 diabetes: correlations with diabetic duration and functional impairments. Brain 2004; 127: 1593-1605.

[5] Myers MI, Peltier AC and Li J. Evaluating der-mal myelinated nerve fibers in skin biopsy. Muscle Nerve 2013; 47: 1-11.

[6] Peltier AC, Myers MI, Artibee KJ, Hamilton AD, Yan Q, Guo J, Shi Y, Wang L and Li J. Evaluation

A time course study of diabetic neuropathy in a rat model

7046 Int J Clin Exp Med 2019;12(6):7038-7046

of dermal myelinated nerve fibers in diabetes mellitus. J Peripher Nerv Syst 2013; 18: 162-167.

[7] Sommer C and Lauria G. Skin biopsy in the management of peripheral neuropathy. Lancet Neurol 2007; 6: 632-642.

[8] Joint Task Force of the EFNS and the PNS. European federation of neurological societies/peripheral nerve society guideline on the use of skin biopsy in the diagnosis of small fiber neuropathy. Report of a joint task force of the European federation of neurological societies and the peripheral nerve society. J Peripher Nerv Syst 2010; 15: 79-92.

[9] Beiswenger KK, Calcutt NA and Mizisin AP. Dissociation of thermal hypoalgesia and epi-dermal denervation in streptozotocin-diabetic mice. Neurosci Lett 2008; 442: 267-272.

[10] Pittenger GL, Ray M, Burcus NI, McNulty P, Basta B and Vinik AI. Intraepidermal nerve fi-bers are indicators of small-fiber neuropathy in both diabetic and nondiabetic patients. Diabe- tes Care 2004; 27: 1974-1979.

[11] Lauria G and Lombardi R. Small fiber neuropa-thy: is skin biopsy the holy grail? Curr Diab Rep 2012; 12: 384-392.

[12] Lauria G, Cazzato D, Porretta-Serapiglia C, Ca- sanova-Molla J, Taiana M, Penza P, Lombardi R, Faber CG and Merkies IS. Morphometry of dermal nerve fibers in human skin. Neurology 2011; 77: 242-249.

[13] Dixon WJ. Efficient analysis of experimental ob-servations. Annu Rev Pharmacol Toxicol 1980; 20: 441-462.

[14] Inoue T, Takenoshita M, Shibata M, Nishimura M, Sakaue G, Shibata SC and Mashimo T. Long-lasting effect of transcutaneous electri-cal nerve stimulation on the thermal hyperal-gesia in the rat model of peripheral neuropa-thy. J Neurol Sci 2003; 211: 43-47.

[15] McCarthy BG, Hsieh ST, Stocks A, Hauer P, Macko C, Cornblath DR, Griffin JW and Mc- Arthur JC. Cutaneous innervation in sensory neuropathies: evaluation by skin biopsy. Ne- urology 1995; 45: 1848-1855.

[16] Lauria G, Lombardi R, Borgna M, Penza P, Bi- anchi R, Savino C, Canta A, Nicolini G, Marmiroli P and Cavaletti G. Intraepidermal nerve fiber density in rat foot pad: neuropathologic-neuro-physiologic correlation. J Peripher Nerv Syst 2005; 10: 202-208.

[17] Lee YF, Lin CC and Chen GS. Temporal course of streptozotocin-induced diabetic polyneurop-athy in rats. Neurol Sci 2014; 35: 1813-1820.

[18] Dobretsov M, Hastings SL, Romanovsky D, Stimers JR and Zhang JM. Mechanical hyperal-gesia in rat models of systemic and local hy-perglycemia. Brain Res 2003; 960: 174-183.

[19] Courteix C, Eschalier A and Lavarenne J. Strep- tozocin-induced diabetic rats: behavioural evi-dence for a model of chronic pain. Pain 1993; 53: 81-88.

[20] Biessels GJ, Bril V, Calcutt NA, Cameron NE, Cotter MA, Dobrowsky R, Feldman EL, Ferny- hough P, Jakobsen J, Malik RA, Mizisin AP, Oates PJ, Obrosova IG, Pop-Busui R, Russell JW, Sima AA, Stevens MJ, Schmidt RE, Tesfaye S, Veves A, Vinik AI, Wright DE, Yagihashi S, Yorek MA, Ziegler D and Zochodne DW. Pheno- typing animal models of diabetic neuropathy: a consensus statement of the diabetic neuropa-thy study group of the EASD (Neurodiab). J Peripher Nerv Syst 2014; 19: 77-87.

[21] Johnson MS, Ryals JM and Wright DE. Early loss of peptidergic intraepidermal nerve fibers in an STZ-induced mouse model of insensate diabetic neuropathy. Pain 2008; 140: 35-47.

[22] Christianson JA, Ryals JM, Johnson MS, Dob- rowsky RT and Wright DE. Neurotrophic modu-lation of myelinated cutaneous innervation and mechanical sensory loss in diabetic mice. Neuroscience 2007; 145: 303-313.

[23] Wright DE, Johnson MS, Arnett MG, Smittkamp SE and Ryals JM. Selective changes in nocifen-sive behavior despite normal cutaneous axon innervation in leptin receptor-null mutant (db/db) mice. J Peripher Nerv Syst 2007; 12: 250-261.

[24] Krishnan ST, Quattrini C, Jeziorska M, Malik RA and Rayman G. Abnormal LDIflare but normal quantitative sensory testing and dermal nerve fiber density in patients with painful diabetic neuropathy. Diabetes Care 2009; 32: 451-455.

[25] Ringkamp M and Meyer RA. Injured versus un-injured afferents: who is to blame for neuro-pathic pain? Anesthesiology 2005; 103: 221-223.