Transplantation of Bone Marrow–Derived Mesenchymal Stem Cells Improves Diabetic Polyneuropathy in Rats

Six-week-old male Sprague-Dawley rats (Chubu Kagakushizai, Nagoya, Japan) with an initial body weight of 170–180 g were allowed to adapt to the experimental animal facility for 7 days. They were housed in an aseptic animal room at a temperature of 20–24°C and a humidity of 40–70%, with a 12-h light cycle and 12 fresh air changes per hour, and were allowed free access to rat diet and water. Diabetes was induced by intraperitoneal injection of STZ (60 mg/kg; Sigma Chemical, St. Louis, MO). Control rats received an equal volume of citric acid buffer. One week after STZ administration, rats with plasma glucose concentrations of >16 mmol/l were selected as the STZ-induced diabetic group. The Nagoya University Institutional Animal Care and Use Committee approved the protocols of this experiment.

Bone marrow was harvested by flushing the tibiae and femurs of adult male Sprague-Dawley rats (200–250 g body wt) with PBS. Mononuclear cells (MNCs) were isolated from the collected marrow using the Histopaque-density centrifugation method (14). Total MNCs were cultured in Dulbecco's modified Eagle's medium, with 10% fetal bovine serum (Sigma Chemical) on plastic dishes at 37°C in 5% humidified CO₂. Nonadherent cells were washed off, and adherent cells were expanded. When adherent cells were confluent (defined as passage 0), they were continuously cultured as MSCs until passage 3.

Cells were incubated with fluorescein isothiocyanate–conjugated mouse monoclonal antibodies against rat CD90 (Becton Dickinson, Franklin Lakes, NJ), fluorescein isothiocyanate–conjugated hamster anti-rat CD29 monoclonal antibody (Becton Dickinson), phycoerythrin-conjugated mouse monoclonal antibodies against rat CD34 (Santa Cruz Biotechnology, Santa Cruz, CA), and CD45 (Becton Dickinson) and were characterized as MSCs by fluorescence-activated cell sorting (FACSCalibur; Becton Dickinson). Isotype-identical antibodies served as controls.

MSCs were labeled with the red fluorescent dye PKH26 (Sigma Chemical) according to the manufacturer's protocol, with minor modifications. Briefly, 2 × 10⁷ MSCs were labeled by 2 × 10⁻⁵ to 2 × 10⁻⁶ mol/l PKH26 for 4 min at room temperature, and this reaction was terminated by the addition of 2 ml fetal bovine serum. Cells were washed twice with 5 ml Dulbecco's modified Eagle's medium and transplanted.

Eight weeks after the induction of diabetes, 1 × 10⁶ cells/limb of MSCs in 1.0 ml saline were injected into the right thigh and soleus muscles of normal and STZ-induced diabetic rats. Left hind limb muscles were treated with saline alone. Four weeks later, the following parameters were bilaterally measured.

To evaluate the nociceptive threshold, the current perception threshold was measured in 4, 8, and 12-week STZ-induced diabetic and age-matched normal rats using a CPT/LAB neurometer (Neurotron, Denver, CO) according to the method of Kiso et al. (15), with minor modifications. An electrode (SRE-0405-8; Neurotron) for stimulation was attached to plantar surfaces of the rats. The skin patch dispersion electrode (SDE-49-4; Neurotron) was applied to the back of the body, where the hair had been removed with a hair clipper. Each rat was kept in a Ballman cage (Natsume, Tokyo, Japan) suitable for light restraint in the awake state. Transcutaneous nerve stimuli of three sine-wave pulses (at 5, 250, and 2,000 Hz) were applied to the plantar surfaces of the rats. The intensity of each stimulation was gradually increased automatically (increments of 0.05 mA for 5 and 250 Hz, increments of 0.1 mA for 2,000 Hz). The minimum intensity at which each rat vocalized or was intensely startled with and without vocalization was defined as the current perception threshold. Five consecutive measurements were conducted at each frequency. The current perception thresholds were reported as the means of the values obtained at the fourth and fifth measurements because the threshold values obtained at the first, second, and third measurements slightly fluctuated.

After intraperitoneal injection of sodium pentobarbital (5 mg/100 g), rats were placed on a heated pad in a room maintained at 25°C to ensure a constant rectal temperature of 37°C. The sciatic-tibial MNCV between the ankle and sciatic notch was determined with a Neuropak NEM-3102 instrument (Nihon-Koden, Osaka, Japan), as previously described (5,16,17).

After evaluation of the MNCV, sciatic NBF (SNBF) was measured by the hydrogen clearance technique with an analog recorder (BW-4; Biochemical Science, Kanazawa, Japan) and an electrolysis tissue blood flow meter (RBA-2; Biochemical Science), as previously described (5,16,17), and calculated with the equation of Koshu et al. (18). During this measurement, rats were placed on a heated pad in a room maintained at 25°C to ensure a constant rectal temperature of 37°C.

Four weeks after the transplantation, rats were killed by an overdose of pentobarbital (15 mg/100 g). Thigh muscles for RNA extraction and the distal portions of sural nerves and soleus muscles for protein extraction were obtained from normal and STZ-induced diabetic rats. Tissues were snap-frozen in liquid nitrogen and kept at −80°C until use. Soleus muscles were embedded within OCT compound (Sakura Finetechnical, Tokyo, Japan), and cryostat sections were prepared for immunofluorescence analysis. The proximal portions of the sural nerves were fixed and immersed in 2.5% glutaraldehyde overnight. They were postfixed in osmium tetroxide, dehydrated, and embedded in Epon for morphometric analysis. Normal and STZ-induced diabetic rats were perfused with 500 ml of 4% paraformaldehyde fixative. After perfusion, the soleus muscles from both sides were immersed in 4% paraformaldehyde overnight, embedded in paraffin, and cut into 5-μm sections to evaluate the capillary number–to–muscle fiber ratio.

Four weeks after the transplantation, for tracking of transplanted MSCs labeled by PKH26, cryostat sections of the soleus muscles were observed using a fluorescence microscope. Then, to detect the colocalization of MSCs labeled by PKH26 and VEGF or bFGF protein, immunofluorescence staining was performed. Cryostat sections of the soleus muscles were fixed in 4% formaldehyde, blocked with 5% normal goat serum for 60 min, and incubated with rabbit anti-VEGF polyclonal antibody (1:100; Santa Cruz Biotechnology) or rabbit anti-bFGF polyclonal antibody (1:100; Santa Cruz Biotechnology) and then with Alexa Fluor 488–coupled goat anti-rabbit IgG antibody (1:250; Invitrogen, Carlsbad, CA). Finally, nucleus staining was performed using 4′,6-diamidino-2-phenylindole (DAPI; Wako, Osaka, Japan). Stained sections were observed using a fluorescence microscope, and merge images of PKH26 (red), growth factors (green), and DAPI (blue) were obtained.

Nerve growth factor (NGF) and neurotrophin 3 (NT-3) contents in the soleus muscles were determined with a quantitative two-site enzyme immunoassay using the Emax Immunoassay System (Promega, Madison, WI). The assays were performed according to the manufacturer's protocols. Samples were homogenized in lysis buffer (137 mmol/l NaCl; 20 mmol/l Tris HCl, pH 8.0; 1% NP40; 10% glycerol; 1 mmol/l phenylmethyl sulfonyl fluoride; 10 μg/ml aprotinin; 1 μg/ml leupeptin; and 0.5 mmol/l sodium vanadate) and centrifuged (14,000g for 30 min), and the supernatants were loaded on MaxiSorp 96-well plates (Nalge-Nunc International, Rochester, NY) precoated with anti-NGF or anti–NT-3 polyclonal antibodies. Each sample was assayed in duplicate. Absorbance from colorimetric reactions of horseradish peroxidase and TMB (3,3′,5,5′-tetramethyl benzidine) was determined by a Powerscan HT (Dainippon Pharmaceutical, Osaka, Japan) microplate reader (450 nm) and converted to NGF or NT-3 protein levels (pg/ml) by normalizing to the standard curve. The recoveries of exogenous NGF and NT-3 added to soleus muscle homogenates were measured, and the concentrations of NGF and NT-3 were calculated, taking into account the recoveries. Results were expressed as picograms per milligram protein.

RNA was extracted from frozen samples of thigh muscles using Isogen (Nippon Gene, Toyama, Japan), according to the manufacturer's instructions and quantified spectrophotometrically. Starting from 1 μg RNA, cDNA was synthesized using ReverTra Ace (Toyoba, Osaka, Japan), according to the manufacturer's descriptions. Primers and probes for VEGF, bFGF, and 18S rRNA for the endogenous control were purchased from TaqMan Gene Expression Assays (Applied Biosystems, Foster City, CA). The probe for 18S rRNA was fluorescently labeled with VIC and TAMRA, whereas probes for the genes of interest were labeled with FAM and TAMRA. Real-time quantitative PCR was performed and monitored using the Mx3000P QPCR System (Stratagene, La Jolla, CA). The PCR master mix was based on TaqMan Universal PCR Master Mix (Applied Biosystems). In the same reaction well, cDNA samples (5 μl, for a total volume of 25 μl per reaction) were analyzed both for the gene of interest and the reference gene (18S rRNA) using a multiplex approach. Relative quantity was calculated by the 2[DELTA]^[DELTA]Ct method (19).

The sections of soleus muscles fixed with paraformaldehyde were used for hematoxylin-eosin staining and immunostaining. These stainings were performed as previously described in detail (16). The vascular capillaries were stained by anti–von Willebrand factor polyclonal antibody (1:600; Dako, Tokyo, Japan) and were counted under light microscopy (×200) to determine the capillary number–to–muscle fiber ratio. Five fields from the each section were randomly selected for the capillary counts.

Proximal sural nerve sections were used for morphometric analysis. Semithin cross-sections of the sural nerve were stained with toluidine blue. Then, we assessed the total complement of sural nerve myelinated fibers and obtained the following parameters: fascicular area, density of myelinated fibers, occupancy rate, mean myelin area, mean axonal area, axon-to-myelin ratio, axonal diameter, myelin thickness, and axonal circularity (formula = 4π[area/perimeter {2}]) (20). These parameters were obtained using image processing and analysis software ImageJ (Research Services Branch of the National Institutes of Mental Health, Bethesda, MD).

The distal sural nerves were used for Western blotting. Samples were lysed in detergent lysis buffer (20 mmol/l Tris-HCl, pH 7.5; 150 mmol/l NaC1; 1 mmol/l EDTA; 1 mmol/l EGTA; 1% Triton; 2.5 mmol/l sodium pyrophosphate; 1 mmol/l β-glycerophosphate; 1 mmol/l Na₃VO₄; and 1 μg/ml leupeptin). The lysates were centrifuged at 14,000 rpm for 20 min at 4°C, and protein concentrations were measured using a bicinchoninic acid assay (Sigma Chemical). Ten micrograms of protein were separated by 7.5–10% SDS-PAGE and transferred to nitrocellulose membranes (Millipore, Billerica, MA). Membranes were blocked with Tween-20 Tris-buffered saline (10 mmol/l Tris-HCl, pH 7.5; 100 mmol/l NaCl; and 0.1% Tween 20) containing 5% nonfat dry milk before incubation with mouse monoclonal anti–neurofilament light-chain (NF-L) antibody (1:5,000; Chemicon, Temecula, CA), rabbit polyclonal anti–neurofilament medium-chain (NF-M) antibody (1:10,000; Chemicon), rabbit polyclonal anti–neurofilament heavy-chain (NF-H) antibody (1:10,000; Chemicon), and anti-actin antibody (1:10,000; Abcom, Cambridge, U.K.). Antigen detection was performed using enhanced chemiluminescence Western blotting detection reagents (Amersham Pharmacia Biotech, Piscataway, NJ) with horseradish peroxidase–conjugated secondary antibodies. Images were scanned and densities were determined by ImageJ. The expression of proteins was corrected for by actin density, and the expression in the limbs of normal rats treated with saline was arbitrarily set at 1.0.

All the group values were expressed as means ± SD. Statistical analyses were made by a one-way ANOVA, with the Bonferroni correction for multiple comparisons.

At 8 weeks, STZ-induced diabetic rats showed severe hyperglycemia (normal rats [n = 10]: 5.6 ± 0.7 mmol/l, STZ-induced diabetic rats [n = 11]: 24.4 ± 1.4 mmol/l; P = 0.0002) and significant reductions in body weight (normal rats [n = 10]: 516 ± 26 g, STZ-induced diabetic rats [n = 11]: 294 ± 32 g; P = 0.0004).

Most cultured adherent cells showed the fibroblastic morphology that is characteristic to MSCs. Although they expressed CD29 and CD90, a majority of adherent cells were negative for CD34 and CD45, which is consistent with MSCs previously reported (8,21).

Four weeks after the transplantation, MSCs labeled with red fluorescent dye PKH26 resided in the gaps between muscle fibers (Fig. 1). As far as we could observe, they kept their features and did not participate in the tissue structures, such as blood vessels, muscles, and nerve fibers.

bFGF and VEGF mRNA expressions in MSC-injected thigh muscles of 12-week STZ-induced diabetic rats significantly increased by 1.79- and 1.62-fold, respectively, compared with those in saline-injected thigh muscles (Fig. 2). Immunohistochemically, red-colored cells indicating PKH26-positive MSCs were surrounded by a green color suggesting local secretion of bFGF and VEGF from the MSCs (Fig. 3A and B).

NGF contents in the soleus muscles of the saline-injected side of STZ-induced diabetic rats (497 ± 76 pg/mg protein, P = 0.0070) were significantly reduced compared with those of normal rats (1,084 ± 465 pg/mg protein). MSC transplantation did not improve the reduced NGF contents in STZ-induced diabetic rats (Fig. 4A). Similarly, NT-3 contents were significantly decreased in the saline-injected side of STZ-induced diabetic rats (271 ± 46 pg/mg protein, P = 0.0020) compared with those of normal rats (555 ± 161 pg/mg protein), and transplantation of MSCs did not show any effects on the NT-3 contents (Fig. 4B).

The current perception threshold at 5 and 250 Hz tended to be decreased in 4-week STZ-induced diabetic rats compared with that in age-matched normal rats, but the differences were not significant (Fig. 5A). At 8 weeks, there were no significant differences in the current perception threshold at 5, 250, or 2,000 Hz between STZ-induced diabetic rats and age-matched normal rats (Fig. 5B). In 12-week STZ-induced diabetic rats, the current perception threshold at 5 Hz in the saline-injected limb was significantly increased compared with that in normal rats (normal: 0.22 ± 0.06 mA, diabetic-saline: 0.31 ± 0.06 mA; P = 0.0016), representing hypoalgesia in the STZ-induced diabetic rats. In the MSC-transplanted limb, this deterioration in sensation was normalized (diabetic-MSCs: 0.22 ± 0.03 mA, P = 0.0022). Similar results were obtained with a pulse at 250 Hz. The current perception threshold at 250 Hz in the saline-injected limb was significantly increased compared with that in normal rats (normal: 0.21 ± 0.05 mA, diabetic-saline: 0.45 ± 0.19 mA; P < 0.0001). In the MSC-transplanted limb, this deficit was significantly prevented (diabetic-MSCs: 0.25 ± 0.06 mA; P = 0.0003). On the other hand, 12-week STZ-induced diabetic rats showed no alteration in the current perception threshold at 2,000 Hz. The injection of MSCs into normal rats did not induce significant changes in the current perception threshold at 5, 250, or 2,000 Hz (Fig. 5C).

The MNCV in the sciatic-tibial nerves of 8-week STZ-induced diabetic rats was significantly delayed compared with that of normal rats (normal: 54.6 ± 2.5 m/s, diabetic: 38.1 ± 3.3; P = 0.0080) (Fig. 6A). The delayed MNCV in the saline-injected limb (37.6 ± 3.4 m/s) of 12-week STZ-induced diabetic rats was significantly ameliorated by MSC transplantation (46.8 ± 5.5 m/s, P = 0.0042) (Fig. 6B). However, MSC transplantation did not alter the MNCV of the sciatic-tibial nerves in the normal rats.

SNBF in 8-week STZ-induced diabetic rats was significantly decreased compared with that in normal rats (normal: 21.7 ± 3.2 ml · min⁻¹ · 100 g⁻¹, diabetic: 8.3 ± 3.9 ml · min⁻¹ · 100 g⁻¹; P < 0.0001) (Fig. 7A). The decrease in the SNBF in 12-week STZ-induced diabetic rats (9.2 ± 5.3 ml · min⁻¹ · 100 g⁻¹) was significantly ameliorated (18.2 ± 5.3 ml · min⁻¹ · 100 g⁻¹, P = 0.0081) by MSC transplantation (Fig. 7B). However, MSC transplantation induced no alteration of the SNBF in normal rats.

The vasculatures were visualized by von Willebrand factor immunostaining, a specific marker for endothelial cells. Quantitative analyses revealed that the capillary number–to–muscle fiber ratio in the saline-injected side of STZ-induced diabetic rats (0.38 ± 0.04, P = 0.0011) was significantly reduced compared with that of normal rats (0.53 ± 0.05). Transplantation of MSCs significantly augmented the ratio (0.55 ± 0.08, P < 0.0001) compared with the saline-injected side in STZ-induced diabetic rats (Fig. 7C). Transplantation of MSCs into normal rats showed no significant differences.

The results of the myelinated fiber morphometry of sural nerves are shown in Table 1. Twelve-week STZ-induced diabetic rats showed no changes in fiber number, axonal size, myelin area, or fiber size compared with normal rats. The axonal circularity in the saline-injected limbs of STZ-induced diabetic rats (0.771 ± 0.011, P = 0.0047) was significantly decreased compared with that of normal rats (0.794 ± 0.008). In the MSC-transplanted limb, this decrement was normalized (0.796 ± 0.009, P = 0.0003). MSC transplantation did not alter the axonal circularity in normal rats.

The protein contents of NF-H, NF-M, and NF-L in distal sural nerves were similarly decreased in the saline-injected limbs of 12-week STZ-induced diabetic rats (P < 0.0001). This decrease was fully prevented by the transplantation of MSCs (P < 0.0001) (Fig. 8). In normal rats, the transplantation of MSCs induced no alterations in the neurofilament contents.

The present study demonstrated that transplantation of MSCs improved hypoalgesia, delayed MNCV, reduced SNBF, and decreased axonal circularity in diabetic nerves of the treated limbs. The immunohistological study revealed that the capillary number–to–muscle fiber ratio were increased in the MSC-injected hind limb skeletal muscles. Four weeks after the transplantation, MSCs maintained their viabilities and secreted angiogenic factors such as VEGF and bFGF at the injected sites.

In previous studies, MSCs have been shown to differentiate into a number of mesenchymal cell types, including osteoblasts, chondrocytes, and adipocytes (8). Additionally, MSCs have been reported to have the potential to differentiate into neural cells, such as astrocytes (22), oligodendrocytes (23), and Schwann cells (24,25). Moreover, several subpopulations of bone marrow, including MSCs, are speculated to differentiate into the cellular components of vascular structures (14,26–29). In this study, MSCs stayed in the gap between muscle fibers at transplanted sites and were not incorporated into tissue structures of recipient animals. These results indicate that the therapeutic effects obtained in this study were not mediated through the differentiation or fusion of MSCs.

On the other hand, MSCs have been reported to secrete a wide array of cytokines, such as bFGF, VEGF, monocyte chemoattractant protein-1, and stem cell–derived factor-1 (12). In fact, MSC transplantation has been performed in the field of ischemic diseases, and it has been suggested that its plausible effects would be mediated largely through paracrine actions of locally released arteriogenic cytokines, including bFGF and VEGF (11,13,30). Although bFGF or VEGF has the independent ability to induce angiogenesis, the use of bFGF and VEGF in a coordinated manner has been shown to induce functional neovascularization (31). Furthermore, bFGF and VEGF have been reported to have neurosupportive effects (5,6). In our study, bFGF and VEGF mRNA expressions in STZ-induced diabetic rats were upregulated in MSC-injected hind limb muscles, and MSCs maintained their viabilities and abilities to secrete bFGF and VEGF even 4 weeks after the transplantation. These observations suggest that the interventional effects of MSC transplantation on DPN demonstrated here would be achieved, for the most part, by the paracrine actions of bFGF and VEGF, and these cytokines would work as both angiogenic and neurotrophic factors. We also evaluated the protein levels of VEGF in the soleus muscles using a commercially available enzyme-linked immunosorbent assay kit. The VEGF content was decreased in the saline-treated limbs of 12-week STZ-induced diabetic rats (normal-saline: 68.4 ± 13.8 pg/mg protein, normal-MSCs: 62.8 ± 13.7 pg/mg protein, diabetic-saline: 28.0 ± 16.3 pg/mg protein), which is consistent with a previous report by Schratzberger et al. (6). Although this decrease was not significantly ameliorated by the transplantation of MSCs, the protein content of VEGF tended to increase (diabetic-MSCs: 48.0 ± 13.3 pg/mg protein) and was accompanied by the increased mRNA expression of VEGF. The protein contents of bFGF could not be measured, but it can be speculated that MSC transplantation would have effects on the protein content of bFGF similar to those of VEGF. Therefore, VEGF and bFGF secreted by the transplanted MSCs would additively exert angiogenic and neurotrophic actions.

Longitudinal and further studies are required to assess how long MSCs work in the injected sites and whether they differentiate and incorporate into several tissues or not. The contributions of other potential cytokines, such as monocyte chemoattractant protein-1 or stem cell–derived factor-1, on DPN should be also evaluated.

Our study revealed that the capillary number–to–muscle fiber ratio in the hind limb skeletal muscles were significantly lower in STZ-induced diabetic rats than in normal rats, which is consistent with human studies (32), and MSC transplantation induced neovascularization. The level of therapeutic vasculogenesis induced by MSC transplantation was similar to that by endothelial progenitor cell transplantation, as we previously reported (16), and the differentiation or incorporation of MSCs to vessels was not observed, suggesting that neovasculalization induced by transplantation would be mediated through the actions of angiogenic factors such as VEGF and bFGF secreted from MSCs, which would explain the improvement of NBF. In addition, both VEGF and bFGF have been reported to induce endothelial nitric oxide synthase (eNOS) production and to increase local blood flow through vasodilatation. Although we did not assess the eNOS levels in vasa nervorum, mRNA expressions of bFGF and VEGF were measured in the soleus muscles. Both expressions were significantly elevated in the MSC-transplanted limbs of STZ-induced diabetic rats but not in those of normal rats. Therefore, eNOS production could be induced in the MSC-transplanted limbs of STZ-induced diabetic rats. Thus, both the induction of eNOS production and angiogenesis by MSCs would ameliorate impaired NBF of sciatic nerves in STZ-induced diabetic rats. However, it remains unknown why MSC transplantation did not have any effects on mRNA expressions of bFGF and VEGF or NBF in normal rats. It can be speculated that diabetic conditions such as hyperglycemia and impaired local microcirculation might cause or require the transplanted MSCs to more strongly produce and secrete bFGF and VEGF. Further studies are in progress in our laboratory.

The impairment of NBF is one of the major pathogenic factors in the development of DPN. This notion is clinically and experimentally supported by various studies, which reported that the amelioration of NBF by various treatments improved impaired nerve functions (16,17,33–40). The beneficial effects of MSC transplantation on NCV and NBF in this study could be comparable with those of the therapeutic modalities reported previously. However, there is disagreement as to whether reduced NBF can account for diabetic neuropathy (41,42). Other factors including neurotrophic factors such as NGF and NT-3 could be involved in the improvement of NCV. In the present study, NGF and NT-3 protein contents were reduced in hind limb muscles of STZ-induced diabetic rats, consistent with previous works that demonstrated the role of these neurotrophines in DPN (43,44). However, MSC transplantation improved NCV without affecting the protein levels of NGF and NT-3 in hind limb muscles of STZ-induced diabetic rats. Therefore, the therapeutic effects of MSC transplantation on NCV would not be mediated through NGF or NT-3. As mentioned above, it can be speculated that VEGF and bFGF supplied by transplanted MSCs contributed to the improvement of NCV through the amelioration of NBF or direct actions to axons and Schwann cells.

In the present study, we evaluated sensory nerve functions using a CPT/LAB neurometer. The neurometer is a device that measures selectively the current perception thresholds in three classes of afferent fibers by applying transcutanous sine-wave stimuli at three frequencies of 2,000, 250, and 5 Hz via surface electrodes at a current intensity in the range of 0.01–9.9 mA to the skin. The pulses at 2,000, 250, and 5 Hz mainly stimulate large myelinated (Aβ-), small myelinated (Aδ-), and small unmyelinated (C-) fibers, respectively. The neurometer is now widely used clinically to evaluate the effects of analgesic drugs and peripheral nerve function in various painful neuropathies, including DPN (45–48). Although previous studies (15,49–51) reported methods for measurement of the current perception threshold with the neurometer in rats and sensory nerve functions in animals, no studies have evaluated STZ-induced diabetic animals. This is the first report that evaluates DPN in rats using the neurometer. Twelve weeks after the induction of diabetes, STZ-induced diabetic rats showed no abnormalities in the current perception threshold at 2,000 Hz. However, hypoalgesia at 5 and 250 Hz were observed in the STZ-induced diabetic rats, and transplantation of MSCs improved this abnormality. These results are consistent with clinical findings that an impairment of small fibers occurs as an earlier event than that of large fibers in the natural history of DPN (52–54). However, STZ-induced diabetic rats did not show hyperalgesia at 4 weeks of diabetes, in contrast with previous reports that evaluated small fiber functions of STZ-induced diabetic rats using conventional testing methods, such as a thermal plantar test. Calcutt et al. (55) reported that STZ-induced diabetic rats showed thermal hyperalgesia at 4 weeks of diabetes. In this study, there was a tendency of hyperalgesia for 5- and 250-Hz stimulations at 4 weeks of diabetes, but the differences were not significant. The animals were restrained in Ballman cages and kept under unusual stress during the measurements. This factor could affect the sensory function. Thus, we can speculate that the current perception thresholds measured by the neurometer might not be suitable for evaluating hyperalgesia in STZ-induced diabetic animal models. In the morphometry of sural nerves, myelinated fiber loss or atrophy was not observed in the saline-injected limbs of 12-week STZ-induced diabetic rats. However, the axonal circularity was significantly decreased compared with that in normal rats. Additionally, we demonstrated that the protein contents of NF-H, NF-M, and NF-L in distal sural nerves were decreased in the saline-injected limbs of STZ-induced diabetic rats. These results are in agreement with previous studies that showed a relationship between neurofilament deficits and distorted circularity (56). Therefore, we suggest that the axonal degeneration in STZ-induced diabetic rats was reflected by the distorted circularity and decreased NF-H, NF-M, and NF-L in the distal sural nerves of STZ-induced diabetic rats. In the MSC-transplanted limbs of STZ-induced diabetic rats, both the impaired circularity and reduced protein contents of the neurofilaments were almost normalized by MSC transplantation and were accompanied by ameliorated nerve functions. However, the precise relationships between neurofilament deficits, distorted circularity, and functional measurements remain unknown.

Transplantation of MSCs did not have any effects on the nerve functions or SNBF on the opposite side, and transplanted MSCs remained at the transplanted sites. These results suggest that local transplantation of MSCs has no systemic effects. In the transplanted sites, MSCs did not construct ectopic abnormal tissues, including tumorous growths.

In conclusion, we have demonstrated the beneficial effects of transplantation of MSCs on DPN. MSCs are relatively easy to obtain and can be expanded to sufficient numbers for cell therapy. Although further studies designed to reveal other useful aspects of MSC transplantation on DPN might be required, transplantation of MSCs appears promising as a therapeutic strategy for DPN.

This research was supported in part by Grant-in-Aid for Scientific Research 18613019 from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) and Aichi-Gakuin University High-Tech Research Center Project for Private Universities: matching fund subsidy from MEXT, Japan.

The authors thank Yuko Maehata, Michiko Yamada, and Keiko Shimamoto for technical assistance.