Dental pulp stromal cells can improve muscle dysfunction in animal models of 1 Duchenne muscular dystrophy 2

36 Background: Duchenne muscular dystrophy (DMD) is an inherited progressive disorder 37 that causes skeletal and cardiac muscle deterioration with chronic inflammation. Dental 38 pulp stromal cells (DPSCs) are attractive candidates for cell-based strategies for DMD 39 because of their immunosuppressive properties. Therefore, we hypothesized that systemic 40 treatment with DPSCs might show therapeutic benefits as an anti-inflammatory therapy. 41 Methods: To investigate the potential benefits of DPSC transplantation for DMD, we 42 examined the disease progression in DMD animal model, mdx mice, by comparing them 43 with different systemic treatment conditions. The DPSC-treated model, a canine X-linked 44 muscular dystrophy model in Japan (CXMD J ), which has a severe phenotype similar to 45 that of DMD patients, also underwent comprehensive analysis, including 46 histopathological findings, muscle function, and locomotor activity. 47 Results: We demonstrated a therapeutic strategy for the long-term functional recovery in 48 DMD using repeated DPSC administration. DPSC-treated mdx mice and CXMD J showed 49 no serious adverse events. MRI findings and muscle histology suggested that DPSC 50 treatment downregulated severe inflammation in DMD muscles and demonstrated a 51 milder phenotype after DPSC treatment. DPSC-treated models showed an increased 52 recovery in grip-hand strength and improved tetanic force and home-cage activity. Interestingly, maintenance of long-term running capability and stabilized cardiac function was also observed in 1-year-old DPSC-treated CXMD J . Conclusions: We developed a novel strategy for the safe and effective transplantation of DPSCs for DMD recovery, including regulate inflammation at a young age. This is the first report on the efficacy of systemic DPSC treatment, from which we can propose that DPSCs may play an important role in delaying the DMD disease phenotype. MSCs in DMD for

Medical). CRP levels were measured using a colorimetric assay with an FDC3500 clinical 148 biochemistry analyzer. 149

ELISA 150
The serum levels of IL-6 were determined using the Quantikine ELISA mouse (R&D 151 Systems). A canine IL-6 immunoassay (R&D Systems) was carried out according to the 152 manufacturer's recommendations. 153

Grip strength 154
Forelimb grip strength was measured using a grip strength meter (MK-380M; Muromachi 155 Kikai) as previously described [26]. Three trials of five measurements per trial were 156 performed with a resting period of five seconds between trials. The average tension force 157 (grams) was calculated for each group of mice. 158

Analysis of locomotor activity 159
Physiological mouse activity was analyzed in each cage with a computerized wheel 160 system (dual activity monitor system, SHINFACTORY) by counting the number of 161 wheel revolutions during each 5 min interval using ACTIMO-DATA II software [27]. 162 The activity of dogs was monitored and counted using an infrared sensor system 163 (Supermex, Muromachi Kikai) as previously described [28]. The average daily locomotor 164 activity shown by the dogs over five days and nights (12 h light/dark cycles) was 165 calculated. We also compared the 15-meter running time of CXMD J during the 166 experimental period. The running speed was averaged from four measurements. To 167 determine the acceleration parameter, we used portable wireless hybrid 168 sensors (TSND121; ATR-Promotions) on the thoracic and lumbar regions of dogs, as 169 described previously [29]. The acceleration magnitude (AM) was calculated from the 170 three acceleration vectors (Ax, Ay, Az) as the square root of the sum of the three-axial 171 values (AM = √Ax 2 + Ay 2 + Az 2 ) [30], and was averaged for each trial. The relative 172 components of the AM along the three axes (%) were calculated by dividing the absolute 173 values of each axis by the AM [31], and these components that were averaged in each trial 174 were calculated as acceleration ratios (Ax ratio, AY ratio, AZ ratio). 175

Magnetic resonance imaging 176
Images of the T2-weighted and fat-saturated T2-weighted series were obtained in CXMD J 177 anesthetized with an inhalational mixture of 2% -3% isofluorane and oxygen according 178 to the same method as described in previously [32] with constant monitoring of heart rate 179 and oxygen saturation. We examined the crus muscles of the lower limbs using a 180  The average SNR (Ave SNR) was calculated with equation described as previous our 188 report [29]. 189

Hindlimb extensor strength test 190
The two hindlimbs in CXMD J were evaluated by measuring the wrist flexion and 191 extension strength using a custom-made torque measurement device. Stimulation 192 frequencies of 60 Hz activate muscles that extend or push the hind paw against the ground. 193 A transducer captures the torque generated when the paw pushes against the force plate. 194 The maximal torque was expressed as a percentage of predicted values computed using a 195 model based on control values [33]. 196

Echocardiography 208
Echocardiographic images of unanesthetized dogs were obtained using a Vivid S6 209 Dimensions (GE Healthcare) probe equipped with a 96-element linear array transducer 210 (i13L) transmitting at 10 MHz as described previously [34]. The EF (%) was calculated 211 using M-mode parameters based on multiple measurements. 212

Statistical analysis 213
Data are presented as mean ± S.D. Differences between two groups were assessed using 214 unpaired two-tailed t-tests. Multiple comparisons between three or more groups were 215 performed using one-way or two-way ANOVA. Statistical significance is defined as * P < 216 0.05, ** P < 0.01, *** P < 0.001, and **** P < 0.0001. Statistical significance was calculated 217 using Excel (Microsoft) and GraphPad Prism 8. 218

Systemic injection of human DPSCs (hDPSCs) into dystrophic mice 221
Mdx mice received a single dose of hDPSCs (high dose, 1.0 × 10 6 cells, or low dose, 5.0 222 × 10 5 cells) or repeated administration of high-and low-dose hDPSCs via the tail vein 223 ( Figure 1A). None of the hDPSC-treated mdx mice showed any significant effect on body 224 weight (BW) during the experiments ( Figure 1B). Grip strength in mdx mice showed 225 significant restoration after repeated administration of high-dose hDPSC (Fig. 1C, 226 Supporting Information Table S1, mdx vs. repeated high dose of hDPSC-mdx, P = 0.0002; 227 WT vs. repeated high dose of hDPSC-mdx, P = 0.995). However, the grip strength did 228 not improve in mice administered a single high-dose nor repeat low-dose injection. The 229 grip strength of high-dose hDPSC-treated mdx mice was not significantly different from 230 that of 1-year-old mdx mice or WT mice (Supporting Information Figure S1A). 231 We examined the progressive resistance at wheel running in mdx mice. The 232 hDPSCs-treated mdx mice had improved running speed compared to mdx mice ( Figure  233 1D, Supporting Information Table S2, mdx vs. repeated high dose of hDPSC-mdx, P = 234 0.007; WT vs. repeated high dose of hDPSC-mdx, P = 0.642; mdx vs. single high dose of 235 hDPSC-mdx, P = 0.664; mdx vs. low dose of hDPSC-mdx, P = 0.019) and had a daily 236 running distance similar to WT mice ( Figure 1E, Supporting Information Table S2, mdx 237 vs. repeated high dose of hDPSC-mdx, P = 0.0069; WT vs. high dose of hDPSC-mdx, P 238 = 0.214; mdx vs. single high dose of hDPSC-mdx, P = 0.07; mdx vs. low dose of hDPSC-239 mdx, P = 0.015). Surprisingly, there was a difference in running speed between the 240 repeated treatment and untreated groups at 1 year of age ( Figure 1F, mdx vs. repeated 241 high dose of hDPSC-mdx, P = 0.022), although their daily running distance was not 242 significantly different (Figure. S1B and Table S2, P = 0.24). 243 The cross-section of the tibialis anterior (TA) muscle of mdx mice showed 244 smaller (regenerating fibers) and larger (hypertrophic fibers) fiber diameter in the 245 dystrophic muscles, centrally-nucleated-fibers (CNFs), spread muscle interstitium, and 246 cell infiltration interspersed in the muscle interstitium (Figure 2 A-D). The 247 limited muscle interstitium, nuclear infiltration ( Figure 2A, B, D), and the reduced 249 frequency of larger fiber areas ( Figure 2C), but not in a dose-dependent manner. We also 250 observed that the CNFs in dystrophic muscle, which are indicative of regenerated 251 myofibers following degeneration, were reduced in the hDPSC-treated mdx mice with a 252 repeated high dose ( Figure 2E), suggesting that degeneration was regulated in the 253 hDPSC-treated muscle. When we examined the distribution of the hPDSCs by human-254 specific Alu-PCR, one week after the transplantation, many cells accumulated in the lung, 255 and some survived in the skeletal muscle ( Figure 2F), but these were not detected for a 256 long period of time. 257 Altogether, our data supported the conclusion that short-term amelioration of 258 the DMD phenotype was observed in all groups of hDPSC-treated mdx mice. Among 259 them, the mice repeatedly treated with high-dose of hDPSCs showed long-term and 260 remarkable beneficial effects on the DMD phenotype. 261

Safe systemic transplantation of hDPSCs into CXMD J 263
We next investigated the possibility of long-term benefit in hDPSC-treated animals using 264 dog models. We started with the administration of hDPSCs in CXMD J with the DMD 265 phenotype in the acute phase, when the disease signs were already observable (n = 3 per 266 group, Figure 3A, Table 1). Eight systemic injections of 4 × 10 6 cells/kg were performed 267 on three CXMD J dogs (12205MA, 13201MA, 13304MA) with two courses of weekly 268 injections for 4 weeks (Table 1). After each injection, we carefully monitored activity, 269 heart rate, respiratory rate, and appearance of any abnormal signs. During development, hDPSC-treated CXMD J showed well growth, and no severe weight loss due to continuous 271 administration ( Figure 3B). No obvious abnormalities related to hepato-renal damage or 272 anemia due to systemic administration in all hDPSC-treated CXMD J were noted in blood 273 tests, which included the determination of alkaline phosphatase (ALP), aspartate 274 transferase (AST), blood urea nitrogen (BUN) levels, or also C-reactive protein (CRP) 275 levels ( Figure 3C, Figure S2A). 276 Spontaneous locomotor activity measured using an infrared sensor system 277 showed a largely reduced mobility of CXMD J with aging [28]. In contrast, hDPSC-treated 278 CXMD J maintained activity for longer periods compared with untreated dogs until they 279 turned 1-year-old ( Figure 3D), suggesting that no serious adverse effects on hDPSC-280 treated CXMD J were identified. 281

Regulatory effects of hDPSC treatment on inflammation in CXMD J 282
During the experiments, serum IL-6 levels in CXMD J did not increase over the normal 283 range after hDPSC-treatment, whereas an increase was transiently detected in the 284 untreated CXMD J ( Figure 4A). To address the regulation of progressive inflammation, 285 the intensity of T2-signals on MRI was measured, which is characteristics in the 286 necrosis/edema and inflammatory lesions on CXMD J . When comparing the quantitative 287 changes of higher T2-signals (4-6 sites) in hindlimb muscles between 2 and 7 months of 288 age, these signals were significantly reduced in the hDPSC-treated CXMD J ( Figure 4B To investigate the pathological changes in hDPSC-treated muscle, we next examined 295 cross-sections of DMD muscles. Dystrophic phenotypes, including nuclear infiltration 296 and spread of muscle fiber interstitium were downregulated in the skeletal muscle of 297 hDPSC-treated CXMD J ( Figure 5A, Figure S4). Although dystrophic muscles also 298 displayed a high myofiber size variability due to a higher number of smaller fibers and 299 the occurrence of hypertrophic fibers, the fiber size distribution in the TA muscles shifted 300 toward a lower number of both smaller and hypertrophic fibers in the case of hDPSC-301 DMD ( Figure 5B, P = 0.0425, Figure S5). Immunostaining analysis showed significantly 302 suggesting that systemic hDPSC treatment can improve the dystrophic phenotype. By 308 Alu-PCR analysis, we also confirmed that circulating transplanted hDPSCs were not 309 detectable in blood within 48 h after injection ( Figure S2B). Seven weeks after treatment, 310 the retention of hDPSCs were confirmed in part of the skeletal muscle, such as TA and 311 extensor digitorum longus (EDL) muscle, and cardiac muscle (left ventricular, LV), but 312 not detectable in the lung and diaphragm of recipient dogs ( Figure 5D).

Improved locomotor activity in hDPSC-treated CXMD J 314
The CXMD J model displays progressive clinical impairment with a rapid decline in the 315 walking ability of dogs with progressive weakness, abnormal stiff limbs, and short strides 316  was higher than that of control DMD (4645 ± 183.9 counts, P = 0.0075) in 12-month-old 322 dogs ( Figure 3D), even if still was largely different from that of normal dogs (41746 ± 323 6241 counts, P < 0.0001). Video data showed an increased mobility of hDPSC-CXMD J 324 compared to untreated CXMD J in the cage based on jumping and playfulness (Supporting 325 Information, movie S1). These observations encouraged us to investigate whether hDPSC 326 could increase the locomotor activity of CXMD J . We monitored the 15 meters running 327 speed of CXMD J to determine motor function and confirmed that CXMD J had a slower 328 speed according to their progressive phenotype ( Figure 6A, Supporting Information, 329 movie S2, 3). Meanwhile, hDPSC-treated CXMD J maintained their running speed and 330 were active for more than 12 months (vs. control DMD, P < 0.00001; vs. normal, P < 331 0.0005). We also measured the multiple acceleration parameters, which severely decrease 332 with age in dystrophic dogs compared to normal dogs, as we have reported [29]. When 333 using the acceleration parameter to evaluate motor function, acceleration magnitudes 334 (AM) were not significantly different between untreated and hDPSC-treated CXMD J in 335 either the thoracic or lumbar region ( Figure. S7 and Table S4). But interestingly, the 336 higher AM (>10000 mG) maintenance ratio was rarely reached in CXMD J , but was 337 observed more frequently in hDPSC-treated CXMD J ( Figure 6B). 338

Improvement in skeletal muscle and cardiac dysfunction 339
Finally, we investigated whether repeated systemic administration of hDPSCs would lead 340 to long-term improvement of dystrophic muscle function. An instantaneous force by 341 torque evaluation was used to assess the skeletal muscle function. The tetanic force on 342 the CXMD J hindlimbs was 51.0 ± 12.3 % (3.12 ± 1.0 N·m/s; P < 0.0001) compared to 343 normal dogs (6.12 ± 0.49 N·m/s), while all hDPSC-treated CXMD J (4.96 ± 1.24 N·m/s) 344 showed significantly stronger torque values (81.0 ± 12.8 % of normal dogs, P = 0.042) 345 compared to untreated CXMD J (P = 0.0039) as described in Figure 6b. In regards to 346 cardiac function, CXMD J shows progressive cardiac dysfunction, which is similar to 347 DMD patients present with dilated cardiomyopathy [22,38]. Echocardiography showed 348 that left ventricular (LV) function was maintained, with higher levels of ejection fraction 349 (EF) in hDPSC-treated CXMD J (mean ± SD, 67.3 ± 0.53%) than that in control DMD 350 (60.5 ± 3.2%, P = 0.001), and comparable to normal dogs (69.6 ± 4.7%, Figure 6C). treated with high-dose of hDPSCs in the acute DMD phase also supports the conclusion 367 that effects are long-term. Since transplanted cells are temporary but accumulate in 368 muscle tissues, it is generally considered that hDPSCs play a protective role against 369 inflammation in the dystrophic muscle. In our study, this was supported by the 370 histopathological appearance of the hDPSC-treated muscle, with findings such as reduced 371 areas of nucleic infiltration (Figure 2). 372 In our experiments using dog models, repeated systemic hDPSC injections into 373 the CXMD J were safe and caused no severe side effects without the need for 374 immunosuppression (Figure 3). Since hDPSCs share the characteristics of clinically-used 375 BM-MSCs that lack HLA-DR expression, these cells are not likely to be subjected to 376 immunological attack in the recipient body. 377 The repeated use of hDPSCs in the CXMD J prevented severe inflammation 378 with an IL-6 surge, as validated by localized cell infiltration that was much more localized, 379 and attenuation of the T2 signals in muscles on MRI (Figure 4). These facts indicate that 380 hDPSCs have an immune-modulatory effect in DMD and may attenuate the 381 histopathological changes that leading to dysfunction in dystrophic muscles. 382 Histopathological appearance improvements after hDPSC administration evidence the 383 functional recovery of dystrophic muscle ( Figure 5). Importantly, the home cage activity 384 and running function of hDPSC-treated CXMD J were maintained until they reached 1 385 year of age ( Figure 3D and 6A). These therapeutic effects of hDPSCs are considered to 386 be more effective in the long-term maintenance of running function, a capacity that 387 diminishes with age in the disease, rather than contributing to its recovery. It appears that 388 hDPSC treatment at a young age could alleviate the DMD phenotype preserving the 389 whole-body muscle function. 390 Our results showed the acceleration parameter to determine the instantaneous 391 running ability of CXMD J ( Figure 6B). Since there is a large difference in the evaluation 392 of running speed amount individual and there is some problem even for 6 minutes walking 393 in DMD patients, it can be applied to assess outcomes in clinical trials for hereditary 394 neuromuscular disorders including DMD by introducing the acceleration parameter into 395 the evaluation of running ability. 396 Significantly stronger isometric torque values in hDPSC-treated CXMD J 397 clearly demonstrate that the progressive loss in limb muscle strength is ameliorated by 398 repeated hDPSC treatment ( Figure 6C). Echocardiography showed that decreased EF in 399 CXMD J due to progressive cardiac dysfunction [22,38] was rescued in the hDPSC- do not exhibit such ability, and can be expected to be novel.
We previously provided evidence that severe phenotypes in IL-10 knockout 419 mdx mice, such as increased M1-macrophage infiltration, high inflammatory factor levels, is upregulated in response to TNF-a stimulation ( Figure S9). The binding of SDF-1 to 443 both CXCR4 and CXCR7 is responsible for the production of paracrine mediators, 444 including VEGF, β-FGF-1, and HGF, which exert mitogenic, anti-apoptotic, pro-445 angiogenic, and anti-inflammatory effects [49]. Based on this, we consider that DPSC-446 specific tissue repair mechanisms may be associated with DMD treatment.