Mdx mice have a premature stop mutation in the exon 23 of the murine Dmd gene, which results in failure to translate dystrophin. These mice mimic various aspects of the human disease (Bulfield et al., 1984; Sicinski et al., 1989). C57BL/6-background mdx mice were developed by Dr. T. Sasaoka (National Institute for Basic Biology, Aichi, Japan) and were maintained in our animal facility. Age-matched and untreated male mdx littermates (P30 and P90, n = 3; P60, n = 6) and wild-type C57BL/6 mice (P30 and P90, n = 3; P60, n = 6) were used as controls in these metabolic studies. All mice remained healthy in appearance, activity, and body weight throughout the observation period. All experiments were conducted in accordance with the protocols described in the experimental protocols approved by the Ethics Committee for the Treatment of Laboratory Animals at the Nippon Medical School and Institute of Medical Science. The same male littermates were housed together in individually ventilated cages, with four to six mice per cage. Each group of mice was randomly assigned to a cage. All mice were maintained under a regular 12-h light/12-h dark diurnal lighting cycle with ad libitum access to food and water.

DPSCs were provided by JCR Pharmaceuticals (Hyogo, Japan). The cells were cultured in DMEM (Thermo Fisher Scientific) supplemented with 10% fetal bovine serum (Thermo Fisher Scientific) and 1% antibiotic-antimycotic solution (Wako Pure Chemical Industries) at 37°C in a 5% CO₂ atmosphere. The animals were randomly divided into two groups: control and MSC-treated groups. Systemic delivery of DPSCs (8.0 × 10⁵ cells/100 μL of PBS) into mdx mice via the tail vein was conducted using four injections (treated-mdx, n = 3, each) administered at an interval of 1 week, beginning at 4–5 weeks of age (body weight [BW] > 10 g), as previously reported (Nitahara-Kasahara et al., 2021). Age-matched male WT and mdx mice (n = 3 each) were used as controls. At the ages of 30-, 60-, and 90 days (P30, P60, and P90), the animals were euthanized by cervical cord dislocation, and tissues were excised for histological and molecular analysis.

Muscle samples were collected from the DPSC-treated mice and immediately frozen in liquid nitrogen-cooled isopentane. Subsequently, 8-μm-thick transverse cryosections were prepared from the skeletal muscles and stained with hematoxylin and eosin (H&E) using standard procedures. Tissue sections were visualized using the IX71 or IX83 microscope (Olympus, Tokyo, Japan). Quantification of nuclear infiltration and collagen-stained areas was performed using the HALO image analysis software (Indica Labs, Corrales, MN, USA). The nucleic area (%) was calculated by dividing the nuclear infiltration area by the total area.

Serum CK levels were determined using an ELISA mouse kit (Cloud-Clone, Corp., TX, USA) according to the manufacturer’s recommendations.

Plasma and skeletal muscle samples were collected from young (P30) and adult (P60 and P90) mdx mice and age-matched WT mice (three male mice each group). The collected tissue samples (30 mg) were stored at −80°C until the assay was performed. Blood collected from the heart was centrifuged at 1,500 g for 15 min at 4°C. Plasma was isolated from the resulting supernatant.

Mouse plasma (50 µL) was added to a 450 µL methanol solution prepared to achieve a concentration of 10 µM of the pretreatment internal standard for capillary electrophoresis-time-of-flight mass spectrometry (CE-TOFMS). This mixture was added to 500 μL of chloroform and 200 μL of Milli-Q water and centrifuged at 2,300 g and 4°C for 5 min. After centrifugation, 400 μL of the aqueous layer was transferred to an ultrafiltration tube (Ultrafree MC PLHCC, HMT, centrifugal filter unit 5 kDa) and centrifuged (9,100 ×g, 4°C, 120 min) followed by ultrafiltration. The filtrate was dried and dissolved again in 50 μL of Milli-Q water for measurement. For liquid chromatography-time-of-flight mass spectrometry (LC-TOFMS) analysis, 300 μL of mouse plasma was added to 900 μL of formic acid-acetonitrile solution (1%) prepared as the internal standard (6 μM) and was centrifuged (2,300 ×g, 4°C, 5 min). Phospholipids were removed from the supernatant by solid-phase extraction and dried. For measurement, the dried sample was dissolved in 100 μL of 50% isopropanol solution (v/v).

The skeletal muscle sample (30 mg) added to 750 μL of 50% acetonitrile solution (v/v) was crushed using a crusher under cooling (1,500 rpm, 120 s × 3 times). The same volume of 50% acetonitrile solution (v/v) was added, and the sample was centrifuged at 2,300 g at 4°C for 5 min. The upper layer (400 μL) was transferred to an ultrafiltration tube (Ultra-free MC PLHCC, HMT, centrifugal filter unit 5 kDa) and was centrifuged (9,100 ×g, 4°C, 120 min), followed by ultrafiltration. The filtrated sample was dried and dissolved again in 50 μL of Milli-Q water for CE-TOFMS. For LC-TOFMS analysis, the skeletal muscle sample from the lower legs (30 mg) added in 500 μL of 1% formic acid-acetonitrile solution was crushed (1,500 rpm, 120 s × three times, and 1,500 rpm, 120 s × one time after addition of 167 μL Milli-Q water) using a crusher under cooling conditions and then centrifuged (2,300 × g, 4°C, 5 min). The supernatant was added to 500 μL of 1% formic acid-acetonitrile AMEOR-HMT-0096 and 167 μL of Milli-Q water to precipitate and stirred. Three tubes with a volume of 350 µL (NANOSEP 3 K Ω, PALL) were transferred to an ultrafiltration unit. The samples in the tubes were subjected to centrifugation (9,100 ×g, 4°C, 120 min) and ultrafiltration. After phospholipids in the mixture were removed using solid-phase extraction, the sample was dried and dissolved in 100 μL of 50% isopropanol solution (v/v) for analysis.

The mass spectrometers of the CE-TOFMS system (Agilent; Santa Clara, CA, USA) (Capillary: Fused silica capillary i.d. 50 μm × 80 cm, CE voltage: Positive, 27 kV; MS ionization: ESI Positive or ESI Negative; MS capillary voltage: 4,000 V, and MS scan range: m/z 50–1,000) and Agilent 1,200 series RRLC system SL (Column: ODS column, 2 × 50 mm, 2 μm; Column temp.: 40°C, Flow rate: 0.3 mL/min, Run time: 20 min, Post time: 7.5 min, Gradient condition: 0–0.5 min: B 1%, 0.5–13.5 min: B 1%–100%, 13.5–20 min: B 100%, MS ionization mode: ESI Positive, MS Nebulizer pressure: 40 psi, MS dry gas flow: 10 L/min, MS dry gas temp: 350°C, MS capillary voltage: 3500 V, MS scan range: m/z 100–1700) were operated in positive and negative electrospray ionization conditions. Three independent measurements were performed for each group.

Candidate compounds were assigned to 451 peaks in the plasma sample and 336 peaks in the skeletal muscle using the mass-to-charge ratio (m/z), migration time (MT), and retention time (RT) values of the substances in the HMT metabolite library (Human Metabolome Technologies Inc.) and the Kyoto Encyclopedia of Genes and Genomes (KEGG). For group comparisons, relative area ratios were calculated for each of the peaks corresponding to skeletal muscles and plasma, facilitating the identification of candidate compounds. Quantitative analysis was performed for the candidate compounds. To this end, the calibration curves incorporated peak areas corrected by internal standards, and concentrations were calculated via single-point calibration of 100 μM of substance (internal standard of 200 μM). Ingenuity pathway analysis (IPA; QIAGEN, CA, USA) was performed based on the quantified data from the plasma of each mouse group.

For each group, data were excluded when data acquisition was difficult due to weight loss, debilitation, or death or when there were concerns regarding peculiar values, mechanical errors in measurement, or external environmental influences because motor function and tissue structure could not be correctly assessed. Data are presented as the mean ± standard deviation (SD). Differences between the two groups were assessed using unpaired two-tailed t-tests. Multiple comparisons between three or more groups were performed using analysis of variance (ANOVA, n = 3, Tukey’s post hoc test). Statistical significance was defined as ^* p < 0.05, ^** p < 0.01, ^*** p < 0.001, and ^**** p < 0.0001 and was calculated using GraphPad Prism eight or 9 (GraphPad, La Jolla, CA, USA).

Mdx mice were repeatedly administered DPSCs via the tail vein (Figure 1A). The cross-section of the tibialis anterior (TA) muscle of the untreated mdx mice showed fibers with smaller (regenerating) and larger (hypertrophic) diameters, centrally nucleated fibers (CNFs), spread muscle interstitium, and cell infiltration interspersed in the muscle interstitium (Figure 1B). Histopathological findings observed in the repeatedly DPSC treated mdx mice included limited interstitial muscle area and nuclear infiltration. Quantitative analysis revealed a reduction in the nuclear inflammation area in the DPSC-treated TA muscle compared with that in untreated mdx mice (Figure 1C). In addition, high concentrations of circulating CK decreased temporarily after DPSC treatment in mdx mice (Figure 1D).

We investigated the therapeutic effects of DPSC-treatment on the DMD pathology-specific metabolic abnormalities. To understand the biological responses to DPSC treatment or environmental changes in disease-specific features, we analyzed metabolic disturbances in plasma and TA muscles from WT (healthy), dystrophic mdx, and DPSC-treated mdx mice. Metabolic analysis by CE-TOFMS and LC-TOFMS revealed several substances in the plasma samples (451 metabolites; CE-TOFMS, cation: 195, anion: 95; LC-TOFMS, positive: 72, negative: 89, Supplementary Table S1) and skeletal muscle samples (460 metabolites; CE-TOFMS, cation: 212, anion: 126; LC-TOF-MS, positive: 85, negative: 37, Supplementary Table S2).

In the case of plasma samples, we used metabolic profiling to identify and quantitate a total of 165 metabolites, which revealed that changes in a total of 85 metabolites were statistically significant (p ≤ 0.05). The levels were either elevated or decreased in the comparison of untreated mdx mice versus WT (P30, 21 metabolites; P60, 26 metabolites; P90, 37 metabolites). Significant alterations were observed in 25 metabolites upon comparison of DPSC-treated mdx mice versus untreated mdx mice (P60, 11 metabolites; P90, 22 metabolites), and 31 metabolites were significantly changed upon the comparison of DPSC-treated mdx versus WT (P60, 27 metabolites; P90, 9 metabolites) (Supplementary Table S3).

Using global metabolic profiling, we identified and quantitated metabolites in the skeletal muscles. We found that changes in 62 metabolites were statistically significant (p ≤ 0.05), which either increased or reduced in the comparison of untreated mdx mice versus WT (P30, 23 metabolites; P60, 30 metabolites; P90, 37 metabolites). A total of 33 metabolites were significantly altered in the comparison of DPSC-treated mdx mice versus untreated mdx mice (P60, 11 metabolites; P90, 25 metabolites), and 50 metabolites were significantly changed in the comparison of DPSC-treated mdx mice versus WT (P60, 21 metabolites; P90, 41 metabolites) (Supplementary Table S4).

Based on quantitative metabolite differences, principal component analysis (PCA) showed a clear separation between WT and dystrophic muscles (Figures 2A, B). The first and second PCA components explained 17.2% and 14.7% of the variation in plasma (Figure 2A) and 22.4% and 12.5% of the variation in the TA muscles (Figure 2B), respectively. In addition, DPSC-treated mdx mice showed an intermediate distribution between these groups in the PCA graph. While plasma samples showed negligible change with respect to the time course among mouse groups in the PCA results, we observed time-series variation in the analysis of data derived using the partial least-squares method (PLS) (Supplementary Figure S1). In the skeletal muscles, PCA analysis described clear time-series changes between 30 and 60–90 days for both WT and mdx mice. These data suggest that DPSC-treated mdx mice demonstrated a PCA distribution closer to that of WT mice than to that of untreated mdx mice in both plasma and skeletal muscle samples.

Considering the observed changes in metabolite levels, we next investigated how mdx and DPSC-treated mice exhibited metabolite imbalances compared to WT mice in the plasma and skeletal muscle. The heat map in Figures 2B, D shows the results of hierarchical clustering analysis (HCA). In HCA of skeletal muscle, WT and mdx mice showed notable differentiation consistent with PCA findings (Figure 2C), especially evident between the 30- and 60–90-day groups (Figure 2D). While less prominent than the differentiation between WT and mdx mice, we also detected some alterations in HCA patterns in DPSC-treated mdx mice compared to those in the untreated mdx mice, particularly at 60–90 days of age.

The top and bottom 30 factor loadings annotated in the Human Metabolome Database (HMBD) containing PC1 and PC2 using PCA for the plasma samples (Table 1) as well as PC1 and PC2 for the skeletal muscle samples are listed in Table 2.

When we focused on the regions of the HCA map where a marked difference between WT and mdx mice was observed, we noticed that several amino acid metabolites were present in addition to HMBD. In terms of concentration, amino acid metabolites, such as Asn (asparagine), Asp, Lys, and Met, were accumulated in plasma samples (Supplementary Table S3), and Asn and Gln were abundant in the skeletal muscles of the mdx mice compared to those in the WT mice (Supplementary Table S4). Ser in plasma and Ile and Tyr in the skeletal muscle were lower than those in the WT. Accumulated amino acid metabolites in the mdx mice were downregulated in the DPSC-treated group, including Asn (ratio of mdx ver. WT P60, 1.5, p = 0.021; treated-mdx ver. mdx, 0.7; p = 0.046) and Met (ratio of mdx ver. WT P60, 1.27, p = 0.015; treated-mdx ver. mdx, 0.83, p = 0.035) in the plasma samples and Asn (ratio of mdx ver. WT P90, 1.49, p = 0.012; treated-mdx ver. mdx, 0.72; p = 0.023) in the skeletal muscles.

To compare the exact metabolite amounts, we focused on the quantification of their areas, as reported by the peaks of the mass spectrum. We found that some metabolites, such as phosphocreatine, homocarnosine, S-methylcysteine, guanidinosuccinic acid, and thiamine, were altered in the plasma samples of the 60–90-day-old mdx mice compared to those in the WT. Among these metabolites, statistical analysis confirmed that S-methylcysteine in the DPSC-treated mice at P90 showed significant differences compared to that in the mdx mice (Figure 3A). Analysis of plasma samples from 30 to 90-day-old mice among the three groups showed that the metabolites creatine, inosine monophosphate (IMP), and carnosine accumulated in the early phase of the mdx mice at the age of 30–90 days (Figure 3B). Additionally, glucose 6-phosphate (G-6-P), fructose 6-phosphate, and thymidine levels were higher in the mdx mice than in WT mice after the age of 60 days. In contrast, fumaric acid was reduced in the mdx mice after the age of 60 days than in WT mice. Creatine, IMP, G-6-P, and fumaric acid levels in DPSC-treated mdx mice were not significantly different from those in WT mice at 90 days of age. Furthermore, in the quantification data of the skeletal muscle, the metabolites, threonic acid, pantothenic acid, O-acetylhomoserine, sphinganine, sphingosine, ethanolamine phosphate, and uric acid accumulated in the mdx mice at the age of 60–90 days, whereas homocarnosine levels decreased compared to those in WT mice (Figure 4A). Especially, the elevation of Ans and uric acid in the mdx mice after DPSC treatment showed a reduction at P90, indicating significant differences when compared to the untreated mdx mice (DMD vs. treated-med, Ans, p = 0.008; uric acid, p = 0.006). Next, we examined the changes in metabolites in the skeletal muscles of mice at the ages of 30–90 days by comparing the three groups of mice (Figure 4B). Choline, creatine, and putrescine levels in the mdx mice increased at 60 and 90 days of age. The metabolites choline and Asn were downregulated in DPSC-treated mdx mice (Figures 4A, B). While the levels of carnosine in the mdx mice were lower than those in WT mice, the differences were not significant when compared with the levels in WT mice after DPSC treatment. In addition, we found that cis-5,8,11,14,17-eicosapentaenoic acid, nicotinamide, and ornitin were transiently elevated in the skeletal muscles of DPSC-treated mice compared with those in the WT and mdx mice (Supplementary Table S4; Supplementary Figure S4).

To understand the significance of the metabolites that underwent disease-specific changes or that varied after DPSC administration, we carried out “diseases and biological functions” analysis using IPA. To detect early changes, we performed the IPA analysis at P60, immediately following DPSC administration. The results revealed canonical pathways that were altered in the plasma samples of mice in the DPSC-treatment group. A total of 21 pathways with predicted upper or lower activation states were detected when untreated mdx mice and DPSC-treated mdx mice were compared (Table 3). As predicted by the activation z-score, the top pathways activated in mdx mice included “transport of amino acids, mobilization of Ca²⁺, uptake of D-glucose, release of nitric oxide, synthesis of prostaglandin E2”. Hence, the relevant factors were reduced in DPSC-treated mice. The top pathways in the mdx mice included “uptake of amino acids, storage or concentration of triacylglycerol, uptake of glutamine family amino acid, and conversion of lipid.” Next, we confirmed the differences in these parameters between WT and mdx mice along with variability ratios in the untreated and DPSC-treated mdx mice (Table 4). We found that the “transport of L-amino acid” and “uptake of L-amino acid” were enriched in the mdx mice compared with those in the WT. In contrast, “conversion of lipid” and “uptake of amino acids” were decreased in mdx mice. The pathways that were altered in the untreated mdx mice compared with those in the WT were increased by approximately two-fold in mdx mice compared to those in the DPSC-treated group, except for the “uptake of L-amino acid.” These results imply that the enriched pathways in the mdx mice were downregulated in DPSC-treated mdx mice, resembling WT mice closely. In contrast, pathways such as the “uptake of amino acids,” “concentration of triacylglycerol,” “uptake of glutamine family amino acid,” and “conversion of lipid,” which were downregulated in the mdx mice compared with those in the WT, were upregulated in the DPSC-treated group, indicating a trend toward similarity with WT mice in the DPSC-treated group.

IPA was used to identify the upstream regulators that explained the observed changes in the metabolites because upstream regulators were activated early and subsequently contributed to downstream metabolic changes (Figure 5; Supplementary Table S5). This suggests that several upstream regulators are critical for DMD and MSC treatment. The most activated upstream regulators in WT vs. mdx mice included disease-specific factors, LEP, NOS3, and 3-nitropropionic acid, while the inhibited state targeted 6–8 molecules such as creatinine, D-sphingosine, glutathione, and L-arginine. LDL and BHMT in the activated state targeted 5–6 molecules such as cholesterol, D-sphingosine, and glutathione.

In the case of DPSC-treated mdx mice, the top activated upstream regulators included methamphetamine in an inhibitory state targeting molecules such as 4-hydroxy-3-methoxyphenylacetic acid, 5-hydroxytryptamine, creatinine, GABA, and glutathione as predicted, and upstream regulators BHMT, UCP1, NOS1, and GNRH1 in the activated state, targeting molecules such as creatinine, GABA, L-arginine, L-glutamic acid, oleic acid, phosphorylcholine, and glycero-3-phosphocholine. These results demonstrated that disease-specific or DPSC-treated DMD-specific pathways, including upstream regulators, could be identified using IPA.