Targeting cellular senescence in dystrophin^−/−/utrophin^−/−double knockout mice improves musculoskeletal health and increases lifespan

Abstract

Duchenne muscular dystrophy (DMD) is a severe genetic muscle disease caused by mutations of the dystrophin gene. Previous studies have detected senescent cells in the skeletal muscle of human DMD, dystrophin-deficient mice (Mdx), and rats. This study aimed to use a more severe dystrophin^−/−/utrophin^−/− (dKO-Hom) mouse model to identify which cells become senescent and if targeting cellular senescence can improve bone quality and muscle pathology in dKO-Hom mice. Immunohistochemistry of P21 and GLB1 revealed significantly more senescent cells in the skeletal muscle tissues of 4-week-old Mdx and dKO-Hom mice compared to WT mice, but not in the bone tissue. The senescent cells were predominantly macrophages (GLB1⁺/CD68⁺). Treatment of dKO-Hom mice with ruxolitinib improved spine L5 trabecular bone microarchitecture and ameliorated skeletal muscle histopathology by decreasing senescent macrophages (GLB1⁺CD68⁺, FUCA1⁺/CD68⁺ or P21⁺/CD68⁺) and senescent-associated phenotypes (SASP) such as macrophage migration inhibitory factor (MIF) in skeletal muscle. Ruxolitinib treatment also improved heart muscle pathology by decreasing senescent macrophages. Additionally, ruxolitinib treatment increased muscle grip strength and treadmill endurance of Mdx mice. Moreover, ruxolitinib significantly extended the lifespan of dKO-Hom mice after 12 days of treatment. Furthermore, treatment of dKO-Hom mice with ruxolitinib and deflazacort synergistically improved bone microarchitecture of the spine L5 vertebrate and the proximal tibia trabecular bone (BV/TV, Tb.N, Tb.Th) by increasing osteoblast cells and decreasing osteoclasts. Co-administration of ruxolitinib and deflazacort also synergistically ameliorated skeletal muscle and heart pathology. Therefore, targeting senescent cells with ruxolitinib represents a promising approach for treating DMD patients but warrants further studies in humans.

1. Introduction

Duchenne muscular dystrophy (DMD) is a severe genetic muscle disease that occurs due to the mutations of the dystrophin gene, resulting in loss of functional dystrophin expression. It affects 1 in 3000 boys [1,2]. The manifestations of DMD are severe with patients succumbing to the disease in their third or fourth decade of life, as there is currently no cure. In addition to their muscular manifestations, DMD patients are also at an increased risk of skeletal fractures [3,4]. Currently, glucocorticoid treatment is the standard of care for muscular dystrophy, which improves the overall health of DMD patients and extends ambulatory time [5–7]. Deflazacort-treated boys experience a longer duration of ambulation than prednisone-treated patients (15.6 vs 13.5 years). Furthermore, deflazacort is also associated with a lower risk of scoliosis, improved ambulatory function, a greater percentage of lean body mass, a shorter stature, and a lower body weight, after adjusting for age and steroid treatment duration [8]. However, glucocorticoid treatment, including deflazacort, increases vertebrate fracture risk in DMD patients (with some reporting a 16-fold increase compared to untreated), and, unfortunately, boys with a history of fracture(s) had a rapid decline of function [9–12].

Restoring functional dystrophin is the ultimate solution to eventually cure DMD patients, but few gene therapy approaches have been translated into clinical treatments due to a myriad of different mutations of the dystrophin gene. Early studies using AAV-vectors to deliver micro-, mini-, or full-length dystrophin genes have been shown to express truncated or full-length functional dystrophin proteins, and consequently ameliorate skeletal muscle pathology, improve growth, inhibit spine deformation, and extend lifespans using different mouse models, including dystrophin^−/−/utrophin^−/− (dKO-Hom) mice [13–16]. The combination of AAV-VEGFa and AAV-mini-dystrophin further improved skeletal muscle pathology as demonstrated by decreased inflammatory cell infiltration, central nuclear fibers, fibrosis, and increased dystrophin and nNOS expression [17]. More recently, ELEVIDYS, an AAV-based gene therapy (micro-dystrophin) has been approved by the FDA for the treatment of ambulatory pediatric patients aged 4 through 5 years with DMD and a confirmed mutation in the DMD gene [18]. CRISPR/-Cas9 technology is another gene therapy approach aimed at restoring functional endogenous dystrophin by guided RNA-mediated excision of mutant dystrophin genes such as its exon 23. Using AAV9 to deliver CRISPR/cas 9 and gene-editing components (gRNA) to postnatal mdx mice by administration at different modes and times restored dystrophin protein expression in cardiac and skeletal muscle to varying degrees, and its expression increased from 3 to 12 weeks after the injections, leading to enhanced skeletal muscle function [19,20]. More recently, base and prime editing have also been used to edit the mutant dystrophin genes to restore functional dystrophin protein, and have demonstrated high efficacy [21,22].

The dKO-Hom mouse is a mouse model that recapitulates the disease’s clinical manifestations more closely than dystrophin^−/−mice (Mdx) with more severe muscle histopathology including muscle necrosis, fibrosis, fat infiltration, heterotopic ossification (HO), kyphosis, and a shortened life span [23–27]. Previously, we have shown that dKO-Hom mice also exhibit a spectrum of musculoskeletal abnormalities, including bone osteopenia [28]. Bone osteopenia was found to be a secondary consequence of the muscle disease, with mice demonstrating declines in osteoblasts, osteoclasts, and osteocytes in bone [29]. Furthermore, we also found impaired fracture healing in dKO-Hom mice [30]. Targeting the prostaglandin E2 (PGE2)/PGE2 receptor 2 (EP2) signaling pathway with EP2 antagonist PF04418948 improved muscle pathology and long bone microarchitecture [31].

Cellular senescence has been found to be the fundamental mechanism of aging. Targeting cellular senescence with different senolytic drugs has been shown to significantly improve aged-related fragility, disease, and bone loss [32–35]. Cellular senescence also plays an important role in other diseases [36,37]. Previous studies have also detected senescent cells in both mouse and rat models of DMD, as well as the muscle tissues of human DMD patients [38,39]. Ablation of P16 or using the senolytic drug ABT263 has been shown to improve muscle pathology and function in dystrophic muscle [39]. No study has investigated if senolytic drugs also improve the microarchitecture of the bones in DMD. Therefore, the aim of this study was to identify which cells and tissues become senescent in dKO-Hom mice and if targeting cell senescence can improve bone phenotypes and muscle pathology, as well as the general health of dKO-Hom mice.

2. Materials and methods

2.1. Animal breeding

All experiments were approved by The Institutional Animal Care and Use Committee (IACUC) animal protocol of Colorado State University (#1234). This study followed the Animal Research: Reporting of In Vivo Experiments guidelines (ARRIVE). DKO-Hom mice and Mdx mice at different ages were generated using dystrophin^−/−/Utrophin^+/− (dKO-Het) mice bred at the Animal Facility of Colorado State University. DKO-Het mice were provided by Dr. Bing Wang from the University of Pittsburgh created by Dr. Grady et al. [23]. DKO-Hom mice at 4-week--old were used for different experiments after genotyping. 4-week-old dKO-Hom mice were chosen due to the findings of a previous study which showed bone loss at this age [29]. C57BL/10ScSnJ (WT) mice were purchased from The Jackson Laboratory and bred at the animal facility of Colorado State University to generate WT mice at the desired age.

2.2. Detection of cellular senescence and senescence-associated phenotype (SASP)

4-week-old WT, Mdx, and dKO-Hom mice (n = 5/group) were sacrificed, and the gastrocnemius muscles were dissected and placed onto a cork surface with freezing medium. This was then snap frozen in 2-methylbutane in liquid nitrogen and stored at −80°C freezer until sectioning. Thigh muscles were collected and immediately frozen on dry ice and stored at −80°C for later RNA extraction and quantitative polymerase chain reaction (Q-PCR) analysis. Tibia and femur bone tissues were dissected and fixed in neutral buffered formalin (NBF, Sigma-Millipore) for 4 days, decalcified, processed, and then paraffin-embedded for histology.

2.3. Treatment of mice with ruxolitinib and other senolytic drugs

4-week-old dKO-Hom mice were divided into two groups (n = 8/group including both males and females) and treated with ruxolitinib (60 mg/kg/d)(CT-INCB, Chemietek), a JAK1/2 inhibitor, and vehicle (5 % dimethyl sulfoxide (DMSO) in phosphate-buffered saline (PBS)) by oral gavage daily. Mice were then sacrificed 12 days after treatment. Gastrocnemius muscle tissues were collected for histology and immunofluorescent staining. Thigh muscle tissues were collected for RNA isolation and Q-PCR as stated above. The lumbar spine, right tibia, and femur bones were collected for micro-CT and histology analysis. The left side femur and tibia were frozen at −80°C for biomechanical testing. To test if other senolytic drugs were also effective, 4-week-old dKO-Hom mice were also treated with fisetin (Item No. 15246, Cayman Chemical Inc, 20 mg/kg, N = 8) or vehicle (0.5 % sodium carboxylmethyl cellulose (C5013–500G,Sigma-Aldrich) (Suspension, N = 9) for 2 weeks. Furthermore, 4-week-old dKO-Hom mice were also treated with dasatinib (D,Cat#:SML2589–50MG, MilliporeSigma)+quercetin (Q, Cat#:PHR1488–1G, 50 mg/kg, MilliporeSigma) (D+Q) or vehicle (60 % Phosal 50PG, 10 % ethanol and 30 % PEG-400) once a week for 2 weeks. Mice were sacrificed 2 weeks after treatment, and the gastrocnemius and thigh muscles were harvested for histology, immunofluorescent staining, and Q-PCR, respectively. The lumbar spine, right tibia, and femur were harvested for micro-CT and histology analysis as stated below.

2.4. Lifespan determination

To detect whether targeting cell senescence could extend lifespan, 4-week-old dKO-Hom mice were treated with either ruxolitinib or vehicle (control) for 12 days (n = 8–10 including both males and females) and allowed to live until natural death. The lifespan was compared between the ruxolitinib-treated and vehicle-treated mice.

2.5. Treatment of Mdx mice with ruxolitinib and muscle grip strength as well as endurance test

To test if targeting cellular senescence with ruxolitinib will also improve muscle function in Mdx mice, 6-week-old Mdx mice (N = 8 males and females) were purchased from The Jackson Laboratory. These mice were treated with 60 mg/Kg ruxolitinib and vehicle (5 % DMSO in PBS) for two weeks. Mice were then subjected to grip strength testing using a DS2–50N grip strength device (QZSD-02) provided by Dr. Tom LaRocca’s laboratory from Colorado State University. Five trials were conducted, the highest and lowest were removed, and the data from the remaining three trials were averaged for final analysis. After overnight rest, the two groups of mice were also subjected to treadmill endurance testing using the PanLab H. Harvard Apparatus Treadmill device kindly provided by Dr. Kelly Santangalo’s laboratory from Colorado State University. Mice were first trained for 10 min to acclimate to treadmill running, then rested for 5 min, and subsequently subjected to treadmill running for 1 h using 0.4 intensity and 20 cm/second speed. 4 mice were tested in a group each time. The travel distance, time of shock, and number of shocks were automatically documented by the device. After the treadmill test, mice were allowed to rest for approximately 4 h and then sacrificed. Blood was collected for the isolation of serum using a retro-orbital method. Gastrocnemius muscles were harvested and snap frozen on cork using 2-methylbutane in liquid nitrogen. Bone tissues, including spine tissues, were collected for micro-CT analysis using a Viva CT 80 and analyzed using built-in software. H&E staining was performed on the gastrocnemius muscle as stated below. Serum creatine kinase (CK) (Cat#:MA-CK, RayBiotech) and alanine transaminase (ALT) (Cat#MA-ALT, RayBiotech) were measured according to the manufacturer’s manual.

2.6. Determination of the synergistic effect of ruxolitinib and deflazacort on skeletal muscle and bone health in dKO-Hom mice

Since deflazacort is a commonly used palliative medication to treat DMD patients, treatment with ruxolitinib and deflazacort was further investigated to determine if they had synergistic effects on both muscle and bone in dKO-Hom mice. 4-week-old dKO-Hom mice (both males and females) were treated with vehicle (5 % DMSO in PBS) (N = 8), deflazacort (1 mg/kg/day) (Item #20386, Cayman Chemical) (N = 10), and deflazacort (1 mg/kg/d) plus ruxolitinib (CT-INCB, Chemietek, 60 mg/kg/day) (D+R) (N = 8) for 12 days. For the D+R group, deflazacort was gavaged in the morning while ruxolitinib was gavaged later in the afternoon based on body weight. Mice were then sacrificed 12 days after treatment. The gastrocnemius muscle, the thigh muscles of the left side, the lumbar spine, the right side tibia and femur were harvested and processed as in 2.1. Bone tissues were then fixed with NBF for 4 days, and micro-CT was performed to investigate bone microarchitecture. Bone tissues were decalcified,processed and embedded in paraffin.

2.7. Bone biomechanical testing using three-point bending assessment of tibia

All bone specimens were harvested and stored in a −80°C freezer until testing. Following thawing at room temperature for 24 h in PBS, tibial length was measured using digital calipers. Mechanical properties of the tibia were then evaluated through three-point bending at the mid-diaphysis using an Instron 34SC-1 universal testing machine equipped with a 100 N load cell (Instron, Norwood, MA, USA). The tibia was positioned horizontally on two lower supports spaced 6 mm apart [40,41], with the anterior surface facing downward [42]. This orientation ensured maximal stability during testing. Quasi-static loading was applied to the medial surface at the midshaft using a displacement rate of 2 mm/min. Load and displacement data were recorded at a frequency of 50 Hz. The resulting load-displacement curves were analyzed using a custom MATLAB script. Functional parameters of the tibia were quantified, including ultimate load, stiffness, deflection, and energy absorbed up to ultimate load and yield force. Ultimate load was defined as the peak load before failure. Once this point was established, corresponding deflection and energy absorption were calculated. Stiffness was determined from the steepest slope of the linear portion of the curve preceding ultimate load, with the analysis requiring a minimum of 50 data points and a load displacement correlation exceeding 0.99 [43,44].

2.8. Skeletal muscle immunofluorescence staining

The gastrocnemius muscle was cut into 8 μm sections using a Leica Cryostat and stored at −80°C until staining later. Slides were dried at room temperature for 10 min and then fixed with 4 % paraformaldehyde in PBS for 8 min. Section slides were then washed 3 times with PBS and blocked with 5 % donkey serum in PBS for 1 h at room temperature, and then incubated with primary antibodies diluted in 5 % donkey serum overnight. Primary antibodies and their dilutions were as listed below: rat anti-CD68 (ab53444, Abcam, 1:200 dilution), rabbit anti-β-galactosidase (GLB1) Cat#:15518–1-AP, Proteintech, 1:200 dilution), rabbit anti-alpha-L-fucosidase (FUCA1) (Catalog# PA5–115256, ThermoFisher Scientific, 1:200 dilution), rabbit anti-P21 (ab188224, Abcam, 1:400 dilution), rat anti-CD31(BD553370, BD Biosciences, 1:300 dilution), mouse anti-embryonic myosin heavy chain (eMHC) (F1.652, DSHB, 1:50 dilution) and paired box transcription factor 7 (PAX7) (DSHB, 1:100 dilution). Double immunofluorescence staining of GLB1/CD68, FUCA1/CD68, P21/CD68, FUCA1/CD31, eMHC/GLB1, PAX7/GLB1 was performed. After overnight incubation with primary antibody, slides were then washed three times in PBS in a Coplin jar and subsequently incubated in donkey anti-rat 488 and donkey anti-rabbit-594 (Jackson ImmunoResearch) at 1:200 dilution for 2 h at room temperature. Slides were then washed and nuclear counterstaining was performed using 4′,6-diamidino-2-phenylindole (DAPI) at 1 μg/ml in PBS for 10 min. Slides were then washed 3 times with PBS, then 3 times with deionized water, and coverslipped with an aqueous mounting medium (Vector Laboratories). Fluorescent images were captured using the Nikon Ti microscope. Images were taken at the highest cellularity area of each muscle section. At least 8 images from different section levels of muscle were taken for quantification. Positive cells were counted using Image J.

2.9. Micro-CT analysis

Bone specimens were fixed for 4 days in NBF and then scanned with the Viva-CT 80 using 70kVP and 114 μA (Scanco Medial LLC). The micro-CT scan and definition of VOI followed a protocol as previously described [29]. A Gause Sigma= 0.8, Support= 1, and threshold = 163 was used to define trabecular bone. A Gauss Sigma= 0.8, Support= 1, and threshold 200 was used to define cortical bone. The spine L5 and proximal tibia trabecular bone microarchitecture quantification parameters include bone volume/total volume (BV/TV), trabecular number (Tb.N), trabecular thickness (Tb.Th), trabecular separation (Tb.Sp) and bone volume density (BV density). For midshaft femur cortical bone, the cortical bone thickness (Ct. Th) and cortical bone volume density were analyzed.

2.10. Histology

After micro-CT scan, the bone specimens were decalcified with 10 % ethylenediaminetetraacetic acid disodium salt dihydrate (EDTA) plus 1 % sodium hydroxide (pH=7.2) for 4 weeks, and the tissues were then processed, dehydrated, and embedded in paraffin. Paraffin sections were cut at 5 μm for different staining. H&E staining was performed using AnaTech Hematoxylin Extra Strength and Eosin Y reagents following the manufacturer’s protocol. For heart H&E quantification, 8–14 sections for each animal were imaged for both lower and higher magnifications, and each section was examined for the damage extent of inflammation (mild: less than 20 inflammatory cells, medium: 50–100 inflammatory cells, and severe with more than 100 inflammatory cells per foci of muscle). The percentage of sections with inflammation was evaluated and quantified. Herovici’s staining was performed using a protocol as previously described [45,46]. TRAP staining was performed using an Acid Phosphatase, Leukocyte (TRAP) Kit (Sigma, 387A-1KT) following the manufacturer’s protocol.

2.11. Immunohistochemistry

Immunohistochemistry for bone and muscle tissues was performed according to a previous protocol [47]. Paraffin sections of bone tissues were cut into 5 μm thickness, deparaffinized using xylene, and rehydrated in serial concentrations of alcohol to water. Then antigen retrieval was performed using a citrate buffer (pH=6, 10 mM) for P21 and pH = 9 Tris-EDTA buffer for osteocalcin and GLB1 in Coplin jars in a 92°C water bath for 20 min. Then, the slides were kept in Coplin jars and cooled for 10 min. The slides were then washed 3 times in PBS and blocked with 5 % donkey serum in PBS for 1 h at room temperature. The primary antibody was added onto the slides and incubated overnight. The primary antibody was diluted in 5 % donkey serum. The primary antibody dilutions were: rabbit anti-P21 (ab188224, 1:400 dilution), rabbit anti-osterix (ab22552, Abcam, 1:1000 dilution) and rabbit anti-osteocalcin (23418–1-AP, Proteintech, 1:200 dilution) and GLB1 (Cat#:15518–1-AP1, Proteintech: 1:400 dilution). For osterix, no antigen retrieval was needed. After overnight incubation, slides were washed three times with PBS, and endogenous peroxidase was inactivated with 0.5 % hydrogen peroxide (H₂O₂) in PBS for 30 min at room temperature. After another 3 washes in PBS, slides were incubated in biotinylated horse anti-rabbit secondary antibody (BA-1100–1.5, Vector Laboratory, 1:300). After secondary antibody incubation and washing, slides were incubated with ABC reagents (VECTASTAIN^® Elite^® ABC-HRP Kit, Peroxidase (Standard),PK-6100, Vector Laboratories)for 2 h at room temperature. After another 3 times of washing with PBS, DAB Substrate Kit, Peroxidase (HRP), with Nickel, (3,3′-diaminobenzidine) (SK-4100,Vector Laboratories) was used to reveal positive cells with brown color. Slides were washed with tap water and then counterstained with Hematoxylin QS (H-3404–100, Vector Laboratories) for 30 s, rinsed with tap water, and blue nuclei with running tap water for 10 min. Slides were then dehydrated in gradient alcohol, cleared with xylene, and mounted with Cytoseal (Fisher Scientific). All histology images were captured using NIS-Elements software equipped with the Nikon microscope. The positive cells were quantified using Image J and normalized to the bone surface (mm) for bone tissues.

2.12. RNA extraction, cDNA synthesis and Q-PCR analysis

For all experiments, thigh muscle tissues were dissected and immediately placed in the −80°C freezer for later RNA extraction using Trizol reagent (Invitrogen) following the manufacturer’s protocol. Total RNA was dissolved in DNAse/RNAase free water and concentration was measured with a nanoplate equipped on a Tecan Infinite 2000 plate reader. cDNA synthesis was performed using Iscript Reverse Transcription Supermix (1708841, BioRad) using 1 μg total RNA. cDNA was diluted 1:5 with DNAase/RNAase free H2O. Q-PCR was performed using the SsoAdvanced Universal SYBR^® Green Supermix Q-PCR kit (1725271, BioRad) using 10 μl reaction with two replicates. The targets analyzed included GLB1, FUCA1, p16^Ink4A (P16), P21^CIP/KIP (p21), plasminogen activator inhibitor-1(PAI-1), macrophage migration inhibitory factor (MIF), interleukin1β (Il1β), interleukin 6 (IL6), inducible nitric oxide synthethase (iNOS), CD86, CD206, CD68, paired box transcriptor 7 (PAX7), myogenin (MYOG), myogenic differentiation D (MYOD) while glyceraldehyd-3-phosphate dehydrogenase (GAPDH) was used as a housekeeping gene. The mRNA expression level was expressed as 2^{-Delta-Delta CT} using the WT or vehicle treated mice as the control group depending on the experiments. All primers were designed using Primer 3 Input [48–50]. Primers used in this study are listed in Table 1.

Table 1.

Primer information.

Gene name	Forward primers (5’-3’)	Reverse primers (5’-3’)	Product size(bp)
GLB1	AACACTGAGGCCATGGTACG	AGGTGGGGAGTATGAGGTCC	109
FUCA1	TGGTCTGATCGTCCCCATCT	TCTTTTCCGACTGCACCCTC	113
P16	GTCGCAGGTTCTTGGTCACT	CGAATCTGCACCGTAGTTGA	246
P21	CGGTGGAACTTTGACTTCGT	CAGGGCAGAGGAAGTACTGG	159
PAI–1	AGTCTTTCCGACCAAGAGCA	ATCACTTGCCCCATGAAGAG	209
MIF	CCGGACCAGCTCATGACTTT	ACTGTAGTTGCGGTTCTGGG	102
Il1β	ACTCATTGTGGCTGTGGAGA	TTGTTCATCTCGGAGCCTGT	199
Il–6	CCGGAGAGGAGACTTCACAG	CAGAATTGCCATTGCACAAC	134
GAPDH	CCGGGGCTGGCATTGCTCTC	GTGTTGGGGGCCGAGTTGGG	190
iNOS	GCAGAGATTGGAGGCCTTGT	CCTGATCCAAGTGCTGCAGA	280
CD86	CCGGATGGTGTGTGGCATAT	TGAGCAGCATCACAAGGAGG	166
CD206	AACCAGTTCCTTGAGCTCGG	CTGATTAGGGCAGCCGGTAG	289
MyoD	TGAATGAGGCCTTCGAGACG	GCCTGCAGACCTTCGATGTA	112
MyoG	GTGAATGCAACTCCCACAGC	AGTTGGGCATGGTTTCGTCT	195
PAX7	GACTCCGGATGTGGAGAAAA	GAGCACTCGGCTAATCGAAC	145
CD68	TTCTGCTGTGGAAATGCAAG	AGAGGGGCTGGTAGGTTGAT	241

2.13. Statistical analysis

All data were analyzed using Graphpad Prism 10. A two-tailed T-test was used for two group comparisons, while analysis of variance (ANOVA) was used for three group comparisons followed by Tukey’s post-hoc for multiple comparisons or non-parameter test. A P value of ≤ 0.05 was considered statistically significant.

3. Results

3.1. dKO-Hom mice demonstrated high level of cellular senescence and SASP gene expression in the skeletal muscle at 4-week-old

In order to detect the presence of cellular senescence in the dKO-Hom mice, we collected thigh muscle tissues from 4-week-old WT, Mdx and dKO-Hom mice, extracted RNA and performed Q-PCR analysis. Both Mdx and dKO-Hom mice were found to have significantly higher levels of GLB1, the gene that encodes senescent β-galactosidase. The expression of GLB1 in dKO-Hom mice was more than 10-fold of the WT mice and also relatively higher than the Mdx mice (Fig. 1A). FUCA1 was also significantly elevated in both the Mdx and dKO-Hom mice compared to the WT mice with more profound elevations in the dKO-Hom mice (Fig. 1B). Furthermore, P16 was significantly increased in both Mdx and dKO-Hom mice with dKO-Hom mice demonstrating increases of more than 20-fold compared to WT mice (Fig. 1C). P21 mRNA expression was also significantly higher in both Mdx and dKO-Hom mice with dKO-Hom mice displaying a 15-fold increase when compared to the WT mice. In addition, SASP molecules, IL-6, PAI-1, MIF and IL1β were all significantly increased in Mdx and dKO-Hom mice with approximately 20-, 5-, 3- and 50-fold increases, respectively, compared to WT mice (Fig. 1 E–H). These results taken together indicate elevated cellular senescence and SASP in the skeletal muscles of dystrophic mice (higher in dKO-Hom mice) compared to WT mice.

Fig. 1.

Q-PCR analysis of senescence and SASP molecules in thigh muscles of 4-week-old mice. A. GLB1 mRNA. B. FUCA1 mRNA. C. P16 mRNA. D. P21 mRNA. E. IL-6 mRNA. F. PAI-1 mRNA. G. MIF mRNA. H. IL1β mRNA. N = 5 for each group, including both males and females. Exact P values are indicated between group bars.

3.2. The senescent cells (mainly macrophages) in dKO-Hom mice are predominantly located in the skeletal muscle tissues

Next, the goal was to determine which cells become senescent in the Mdx and dKO-Hom mice at 4 weeks of age. Immunohistochemistry was performed for P21, one of the senescent markers. P21⁺ cells were found to be absent in the cortical bone tissues of mice but were highly expressed in the damaged skeletal muscle area (Fig. 2A). P21⁺ cells in both Mdx and dKO-Hom mice were significantly higher than in the WT mice (Fig. 2B). Immunohistochemistry staining of GLB1 was also performed, which revealed that GLB1 was mainly expressed in the damaged skeletal muscle area, but not in the femur cortical bone. GLB1⁺ cells were significantly higher in the skeletal muscle of Mdx and dKO-Hom mice than in the WT mice (Fig. 2C–D). Immunofluorescence staining was further performed to determine if macrophages become senescent by colocalizing GLB1 with CD68, a pan-macrophage marker. GLB1⁺CD68⁺ cells (senescent macrophages) were found to be significantly higher in the muscle tissues of both Mdx and dKO-Hom mice when compared to the WT muscle tissues (Fig. 2E,F). Furthermore, GLB1⁺CD68⁺/Total CD68⁺ cells (senescent macrophage percentage) also significantly increased in both Mdx and dKO-Hom mice, with dKO-Hom mice also demonstrating significantly higher levels than the Mdx mice (Fig. 2G). The GLB1⁻CD68⁺ cells (non-senescent macrophages) were also found to be significantly increased in both Mdx mice and dKO-Hom mice compared to WT mice (Fig. 2H). Total CD68⁺ macrophages were also significantly higher in the Mdx and dKO-Hom mice compared to the WT mice (Fig. 2I). However, GLB1⁺CD68⁻ cells (non-macrophage senescent cells) were minimal, and no significant changes were found between Mdx and dKO-Hom mice and WT mice (Fig. 2J).

Fig. 2.

Senescent cells are predominantly present in skeletal muscle tissues. A. Immunohistochemistry staining of P21 of bone and skeletal muscle tissues. P21⁺ cells were not found in the femur cortical bone tissues, but highly expressed in the skeletal muscle tissue. P21⁺ cells stained brown and nuclei stained blue. Scale bars = 100 μm. B. P21⁺ cells quantification. C. Immunohistochemistry of GLB1. GLB1⁺ cells stained brown in the cytoplasm. GLB1⁺ cells were highly expressed in the skeletal muscle tissues, but not the femoral cortical bone. D. GLB1⁺ cells quantification. E. Double immunofluorescence staining of GLB1 and CD68. CD68⁺ cells are stained green, GLB1⁺ cells are stained red. Cells stained in yellow or orange are GLB1⁺/CD68⁺ cells. Cells stained only green are GLB1⁻CD68⁺cells. Cells that are only stained red are GLB1⁺CD68⁻ cells. Scale bar = 100 μm. F-J. Quantification of GLB1⁺CD68⁺cells, GLB1⁺CD68⁺/Total CD68⁺ percentage, GLB1⁻CD68⁺ cells and GLB1⁺CD68⁻cells. N = 5–8 for each group, including males and females. Exact P values are indicated between group bars.

3.3. Ruxolitinib treatment improved bone microarchitecture of dKO-Hom mice

Treatment of dKO-Hom mice with the senotylic drug (ruxolitinib) was further investigated to determine if it improved bone microarchitecture and skeletal muscle pathology. DKO-Hom mice were treated at 4- week-old with ruxolitinib (60 mg/Kg/day) for 12 days. BV/TV and Tb.Th of spine L5 trabecular bone was significantly increased in the ruxolitinib-treated mice compared to the vehicle-treated mice (Fig. 3A–C). No significant changes of Tb.N, Tb.Sp, and BV density were observed (data not shown). For the proximal tibia trabecular bone, a trend of increased BV/TV, Tb.Th was observed in the ruxolitinib-treated mice compared to vehicle-treated mice (Fig. 3D–F). No significant changes of Tb.N, Tb.Sp and BV density were revealed (data not shown). No significant changes were observed for the femur cortical bone thickness (Ct.Th) or bone volume density between the two groups (Fig. 3G–I). Ruxolitinib treatment did not change the body weight of dKO-Hom mice compared to the vehicle-treated mice (Fig. 3J). Three-point bending bone biomechanical testing of the tibia was performed and revealed that ruxolitinib treatment did not significantly affect the tibia length, ultimate load, stiffness, yield force or energy absorbed, but significantly increased deflection (Fig. 3K–P).

Fig. 3.

Ruxolitinib treatment improved bone microarchitecture. A-C. Micro-CT 3D images of spine L5 and quantifications of BV/TV and Tb.Th. D-F, Micro-CT 3D images of proximal tibia and quantifications of BV/TV and Tb.Th. G-H. Micro-CT 3D images of femur cortical and quantification of Ct.Th, BV density of femur cortical bone. J. Body weight across different time points. K. Tibia length. L. Tibia ultimate load. M. Tibia stiffness. N. Tibia yield force. O. Tibia deflection. P. Tibia energy absorbed. N = 8 for each group, including males and females. Exact P values are indicated between group bars. Scale bars= 100 μm.

In addition to ruxolitinib, dKO-Hom mice were also treated with D+Q once a week for two weeks and micro-CT analysis of the spine and long bones was performed. D+Q did not significantly improve spine L5 trabecular bone microarchitecture including BV/TV, Tb.Th (Supplemental figure 1A–C) or Tb.N, Tb.Sp and BV density (data not shown). D+Q treatment also did not significantly change the bone microarchitecture parameters of proximal tibia trabecular bone (Supplemental Figure 1D–F). Furthermore, D+Q did not significantly change midshaft femur cortical bone parameters, including Ct.Th and BV density (Supplemental Figure 1G–I). D+Q treatment also did not change skeletal muscle pathology as revealed by H&E staining and Von Kossa staining (Supplemental Fig.2A–B).

dKO-Hom mice were also treated with 20 mg/kg/d fisetin for 2 weeks, followed by micro-CT for the spine trabecular bone, proximal tibia trabecular bone, and femur cortical bone. No significant improvement of fisetin-treated dKO-Hom mice was found when compared to vehicle-treated dKO-Hom mice for the bone microarchitecture parameters of the spine L5 trabecular bone, proximal tibia trabecular bone, and femur cortical bone (Supplemental Figure 3A–I).

3.4. Ruxolitinib treatment improved bone quality by decreasing osteoclasts

Herovicis’ staining showed pink (collagen 1) thicker trabecular bone in spine L5 trabecular bone in ruxolitinib-treated dKO-Hom mice when compared to vehicle-treated control mice (Fig. 4A). H&E staining also showed thicker bone trabeculae in the ruxolitinib group compared to the vehicle-treated dKO-Hom mice in spine L5 (Fig. 4B). Osterix (OSX)⁺osteoprogenitor cells in the spine L5 trabecular bone were not significantly affected by ruxolitinib treatment (Fig. 4C,E). However, ruxolitinib treatment significantly decreased TRAP⁺ osteoclasts when compared to the vehicle-treated mice (Fig. 4D, F).

Fig. 4.

Bone histology of spine L5 and lifespan. A. Herovici’s staining of spine L5. Collagen type 1 stained pink-red while collagen type 3 stained dark blue. B. H&E Staining of the spine L5 vertebra. C,E. Immunohistochemistry of OSX for osteogenic progenitor cells and quantification. OSX⁺ cells stained in brown in the nuclei on bone surface. D,F, TRAP staining for osteoclasts and quantification. G. Lifespan. Insets highlight positive cells in C and D. All scale bars = 100 μm. N = 8 for each group, including males and females. Exact P values are indicated between group bars.

An additional experiment was performed to determine if ruxolitinib treatment can improve the lifespan of dKO-Hom mice. Ruxolitinib-treated dKO-Hom mice demonstrated significantly increased average lifespans when compared to the vehicle-treated mice after being treated for 12 days (Fig. 4G).

3.5. Treatment with ruxolitinib decreased senescent macrophages and SASP in skeletal muscle

CD68/GLB1 double immunofluorescence staining was performed to detect if ruxolitinib treatment decreased senescent cells in the gastrocnemius muscle tissue. GLB1⁺CD68⁺ cells (senescent macrophages) were found to be significantly reduced in the ruxolitinib-treated dKO-Hom mice muscle compared to the vehicle-treated muscle (Fig. 5A–B). The GLB1⁺CD68⁺ cell percentage (senescent macrophage percentage) also decreased (P = 0.073)(Fig. 5C). No significant differences were found for GLB1⁺CD68⁻ cells (non-macrophage senescent cells) between the ruxolitinib-treated and vehicle-treated groups (Fig. 5D). After performing FUCA1 and CD68 double staining, a significant decrease of FUCA1⁺/CD68⁺ cells/200X field (senescent macrophages) (Fig. 5E–F) was observed with no significant changes of the FUCA1⁺/CD68⁺ cell percentage (senescent macrophage percentage) or FUCA1⁺/CD68⁻ cells (non-macrophage senescent cells) after ruxolitinib treatment (Fig. 5G–H). Moreover, FUCA1/CD31 double immunofluorescent staining was performed, and the results indicated that very few CD31⁺cells expressed FUCA1(Fig. 5I). Further, no significant differences of FUCA1⁺/CD31⁺ cells (senescent endothelial cells) between the ruxolitinib-treated and vehicle-treated groups were observed (Fig. 5J). However, FUCA1⁺/CD31⁻ cells (non-endothelial senescent cells) were significantly decreased by ruxolitinib treatment compared to the vehicle-treated group (Fig. 5K). Interestingly, FUCA1⁻CD31⁺ cells (non-senescent endothelial cells) were also decreased by ruxolitinib treatment compared with the vehicle-treated group (Fig. 5L). Addtionally, CD68/P21 double staining was performed and revealed that CD68⁺/P21⁺ cells (senescent macrophage) were significantly decreased in the ruxolitinib-treated group compared to vehicle-treated group (Fig. 5M,N). The CD68⁺/P21⁺ cell percentage (senescent macrophage) was also significantly decreased in the ruxolitinib-treated group compared to the vehicle-treated group (Fig. 5O). Furthermore, CD68⁻/P21⁺ cells (non-macrophage senescent cells) were also significantly decreased (Fig. 5P). It was also noted that newly regenerated multinuleated myofibers also expressed P21 indicating that not all P21⁺ cells were senescent cells. eMHC/GLB1 double immunofluorescence staining revealed that eMHC⁺ cells were almost undetectable in the control group and very few were detected in the muscle tissue area of the ruxolitinib-treated group. eMHC did not colocalize with GLB1 (Fig. 5Q). PAX7/GLB1 double immunofluorescence staining demonstrated that PAX7 was mainly detected in the non-inflammatory area and did not colocalize with GLB1. Very few PAX7⁺ cells were detected in the control group, but relatively more were detected in the ruxolitinib-treated group (Fig. 5R). Quantification showed that the ruxolitinib-treated group had a relatively greater number of PAX7⁺/GLB1⁻ cells compared to the control group (Fig. 5S, P = 0.1264). H&E staining was performed to investigate if ruxolitinib treatment also improved muscle pathology. It was revealed that inflammation was decreased in the muscles of mice treated with ruxolitinib compared to the vehicle-treated group. More regenerated muscle fibers were also observed in the skeletal muscle of the ruxolitinib-treated mice compared to the vehicle-treated group (Fig. 5T). To further investigate if ruxolitinib treatment affects thigh muscle senescence gene expression, Q-PCR analysis of the thigh muscle was performed and revealed that ruxolitinib treatment significantly decreased the senescent genes GLB1 and MIF (Fig. 5U–V). Ruxolitinib treatment also trended towards a decrease of TRAP (P = 0.1026) and BMP4 mRNA (P = 0.057) (Fig. 5W–X). Furthermore, ruxolitinib treatment did not significantly change M1 macrophage (iNOS and CD86), M2 macrophage (CD206), pan macrophage markers (CD68), or myogenic differentiation transcription factors PAX7, MYOG or MYOD at mRNA level (Fig. 5Y).

Fig. 5.

Ruxolitinib treatment decreased senescent macrophages and SASP in the skeletal muscle. A. GLB1/CD68 double immunofluorescent staining. GLB1 stained in red color, CD68 stained in green color. Double positive cells stained in yellow or orange color in merged images. B-D quantification of GLB1⁺/CD68⁺/200X field, GLB1⁺/CD68⁺ cell percentage and GLB1⁺/CD68⁻ cells/200X field. E. FUCA1/CD68 double immunofluorescent staining. FUCA1 stained in red color and CD68 stained in green color. Double positive cells showed in yellow or orange in merged images. F-H. Quantification of FUCA1⁺/CD68⁺, FUCA1⁺/CD68⁺ cell percentage and FUCA⁺CD68⁻ cells. I. FUCA1/CD31 double positive cells to detect senescent endothelial cells. FUCA1 stained in red and CD31 stained in green. Double positive cells stained in orange or yellow. J-L. Quantification of FUCA1⁺CD31⁺ cells/200X field, FUCA1⁺CD31⁺cells percentage, FUCA1⁻CD31⁺ cells/200X field (non-senescent endothelial cells). M. CD68/P21 double staining. P21 stained red in the nuclei, CD68 stained green in the cytoplasm and membrane. Double positive cells stained in orange or yellow in merged images. N-P. Quantification of CD68⁺/P21⁺cells/200X field, CD68⁺/P21⁺ cell percentage and CD68⁻P21⁺ cells/200X field (non-macrophage senescent cells). Q. eMHC/GLB1 double immunofluorescence staining. eMHC⁺ cells stained bright green, GLB1⁺ cells stained red. eMHC⁺ cells were almost undetectable in the control group and very few cells were detected in the muscle tissue of the ruxolitinib-treated group. eMHC did not colocalize with GLB1. R-S. PAX7/GLB1 double immunofluorescence staining and quantification. PAX7 did not colocalize with GLB1. Very few PAX⁺/GLB1⁻ cells were detected in the control group, but relatively more were detected in the ruxolitinib-treated group. T. H&E staining showed the most severe damage area of each group in the gastrocnemius muscle. U-X. Q-PCR analysis of GLB1, MIF, TRAP and BMP4 mRNA expression in the thigh muscle. Y. Q-PCR analysis of M1 and M2 macrophage markers as well as PAX7, MYOG and MYOD mRNA expression. N = 8 for each group, including males and females. Exact P values are indicated between group bars. Scale bars= 100 μm for 200X and 200 μm for 400X (PAX7).

Histology analysis of D+Q-treated mouse gastrocnemius muscle was further performed. H&E staining demonstrated no significant improvement of the general muscle pathology at low (large image, 20X) and higher magnification (100X) Supplemental Fig.3A). Von Kossa staining revealed no significant decrease in mineralization of the gastrocnemius muscle (Supplemental Fig.3B–C).

3.6. Ruxolitinib treatment improved heart histology

H&E staining revealed that heart muscle damage and inflammation were relatively mild compared to those in skeletal muscle. Ruxolitinib treatment decreased heart muscle damage (Fig. 6A). Qualitive analysis of heart pathological changes demonstrated that ruxolitinib treatment increased the percentage of sections without any inflammation, decreased the percentage of mild inflammation as well as the total percentage sections with inflammation, while also demonstrating a trend of decreased medium and severe inflammation (Fig. 6B). CD68/GLB1 double immunofluorescence staining was also performed. Severe inflammation was detected in the control group with many CD68⁺GLB1⁺ cells present in this area (stained in yellow or orange). Fewer macrophages were detected in the ruxolinib-treated heart muscle (Fig. 6C). In addition, ruxolitinib treatment significantly decreased CD68⁺GLB1⁺ (senescent macrophage), CD68⁻GLB1⁺(non-macrophage senescent cells), CD68⁺GLB1⁻(non-senescent macrophage) and total CD68⁺ macrophages in the heart (Fig. 6D–G).

Fig. 6.

Histology and immunofluorescence staining of ruxolitinib-treated dKO-Hom mice heart tissues. A. H&E staining showed the best and worst changes of the heart tissues in each group. Best histology sections had minimal damage, whereas the worst histology demonstrated the most severe muscle damage of each group. B. Quantification of inflammation changes between the two groups. C. CD68/GLB1 double immunofluorescence staining. CD68 stained in green and GLB1 stained in red. CD68⁺GLB1⁺ cells stained in orange or yellow indicate senescent macrophages. D-G. Quantification of CD68⁺GLB⁺, CD68⁻GLB1⁺, CD68⁺GLB1⁻, and total CD68⁺ cells in the 200X field. N = 8 for each group, including males and females. Exact P values are indicated between group bars. Scale bars = 250 μm for 40X and 100 μm for H&E staining or 100 μm for CD68/GLB1 immmunofluorescence staining.

3.7. Ruxolitinib treatment improved skeletal muscle strength and endurance of Mdx mice

To further investigate if ruxolitinib treatment improved muscle function in Mdx mice, additional experiments were performed, including grip strength and treadmill endurance tests. Ruxolitinib treatment did not significantly change body weight (Fig. 7A). Ruxolitinib treatment showed a trend of increasing grip strength (P = 0.0837) compared to the control group. Treadmill results revealed that ruxolitinib treatment slightly but statistically significantly increased treadmill travel distance compared to the control group (Fig. 7B). Ruxolitinib also significantly decreased the shock time and number of shocks on the treadmill compared to the control group (Fig. 7C–D). Micro-CT results demonstrated that ruxolitinib treatment did not significantly change heterotopic bone in the muscle surrounding spine and did not change spine L5 vertebrate trabecular bone micro-architecture parameters, including bone volume (BV), BV/TV, Tb.N, Tb.Th, Tb.Sp and BV density (Fig. 7F–L). Although the serum CK activity of the ruxolitinib-treated Mdx mice showed an increased trend (P = 0.0734) (Fig. 7M), ruxolitinib treatment did not significantly change ALT serum activity (Fig. 7N). Despite the H&E staining of the Mdx gastrocnemius muscle revealing overall less severe muscle damage than dKO mice, the ruxolitinib-treated group improved muscle pathology in Mdx mice (Fig. 7O).

Fig. 7.

Effect of ruxolitinib treatment on muscle strength, bone parameters and histology of Mdx mice. A. Body weight of mice at different time points. B. Grip strength. C. Travel distance on treadmill in 1 h. D. Time of shock on treadmill. E. Number of shocks on treadmill. F. Micro-CT overview of lumbar spine and spine L5 vertebrate microarchitecture. G-L. Spine L5 trabecular bone BV, BV/TV, Tb.N, Tb.Th, Tb.Sp and BV density. M. Serum CK activity. N. Serum ALT activity. O. H&E staining of gastrocnemius skeletal muscle. N = 8 for each group, including males and females. Exact P values are indicated between group bars. Scale bars = 100 μm for 40X and 50 μm for 200X.

3.8. Ruxolitinib synergistically improved bone quality when used simultaneously with deflazacort

Since steroids are a commonly used palliative therapy that benefits DMD patients while extending the patient’s life span, additional experiments were performed by combining deflazacort with ruxolitinib. The results indicated that treatment with deflazacort for 2 weeks did not impair the bone quality of dKO-Hom mice, but significantly increased spine L5 vertebral trabecular bone BV/TV and Tb.N compared to the vehicle-treated group, and decreased Tb.Sp as well as trending towards increased Tb.Th (P = 0.0616). No significant changes of BV density of spine L5 vertebrae trabecular bone were detected in deflazacort group compared to the control group. Importantly, D+R further significantly increased BV/TV compared to the vehicle and deflazacort groups by markedly increasing Tb.Th compared to the deflazacort group with no significant changes of Tb.N, Tb.Sp and BV density. These results demonstrate the synergistic effects of ruxolitinib and deflazacort in improving spine trabecular bone (Fig. 8A–F). For the proximal tibia trabecular bone, deflazacort alone was also found to significantly increase BV/TV, Tb.N, significantly decrease Tb.Sp while increasing Tb.N (P = 0.0878) compared to the vehicle group. No significant effects on the BV density were observed in the deflazacort-treated group compared to vehicle-treated group. More importantly, D+R further significantly increased BV/TV, Tb.N, Tb.Th while decreasing Tb.Sp compared to the vehicle-treated group, and significantly increased BV/TV, Tb.N, Tb.Th while decreasing Tb.Sp compared to the deflazacort group (Fig. 8.G–L). No significant changes in BV density were observed compared with the vehicle and deflazacort groups. For the femoral cortical bone, no significant effects of deflazacort or D+R were observed for femur Ct.Th, but both deflazacort and D+R significantly increased femur BV density compared to the vehicle group. D+R treatment did not further increase BV density compared to the deflazacort group (Fig. 8M–O). Furthermore, the body weights of the deflazacort and D+R groups were found to be low at the beginning (day 1) and at day 6 after treatment. D+R group also had significantly lower body weight at day 12 (Supplemental Fig.4 A, please note two mice’s body weights in the control group were missing at day 1). The tibia length was found to be shorter in deflazacort group and significantly shorter in the D+R group compared to the control group which likely attributed to the smaller size of the mice (Supplemental data Fig. 4B). Tibia biomechanical testing results demonstrated significantly lower ultimate load in the deflazacort group and a trend of lower stiffness in this group compared to the control group (Supplemental figure 4C–D). No significant changes were observed in the yield force, deflection, and energy absorbed in the deflazcort and D+R groups compared to the control group (Supplemental Fig.4E–G).

Fig. 8.

Ruxolitinib and deflazacort synergistically improved bone microarchitecture. A. Micro-CT 3D images of spine L5 trabecular bone. B-F. BV/TV, Tb.N, Tb.Th, Tb.Sp and BV density of spine L5 trabecular bone. G. Micro-CT 3D images of proximal tibia trabecular bone. H-L. BV/TV, Tb.N, Tb.Th, Tb.Sp and BV density of proximal tibia trabecular bone. M. Micro-CT 3D images of femur midshaft cortical bone. N-O. Ct.Th and BV density of midshaft femur cortical bone. N = 8–11 for different groups, including males and females. Exact P values are indicated between group bars. Scale bars= 100 μm.

3.9. Ruxolitinib and deflazacort synergistically improve skeletal muscle pathology in dKO-Hom mice

We further investigated if ruxolitinib and deflazacort have synergistic effects on muscle pathology. H&E staining was performed on the gastrocnemius muscle, and revealed that deflazacort alone can decrease inflammation,but it did not improve heterotopic bone formation. Importantly, D+R is more effective in improving muscle pathology by decreasing inflammation, heterotopic bone formation and enhancing muscle regeneration (Fig. 9A.B). GLB1/CD68 double staining was subsequently performed and revealed that deflazacort alone significantly decreased GLB1⁺/CD68⁺ cells/200X field compared to the vehicle-treated control group, while D+R more dramatically decreased GLB1⁺/CD68⁺ cells than deflazacort alone (Fig. 9C–D). Deflazacort alone did not significantly decrease GLB1⁺CD68⁺/total CD68⁺ cells (senescent macrophage percentage), while D+R dramatically decreased the GLB1⁺/CD68⁺/total CD68⁺ cell percentage (Fig. 9E). Neither deflazacort nor D+R significantly decreased CD68⁺/GLB1⁻ cells (non-senescent macrophage) or total CD68⁺ cells (Fig. 9F–G). eMHC/GLB1 double immunofluorescence staining was further performed and revealed that eMHC⁺ cells were barely identified in the skeletal muscle of the control group, while only a few eMHC positive cells were detected in the deflazacort group and relatively more in the D+R group, but many central nuclear myofibers were detected (Fig. 9H). eMHC positive cells did not colocalize with GLB1. PAX/GLB1 double staining indicated that PAX⁺ cells were mainly located in the non-inflammatory area, where GLB1-positive cells were not seen and did not colocalize with GLB1. PAX7⁺ cells were barely present in the gastrocnemius muscle in the control group, and more were present in the deflazacort group, but the highest number was present in the D+R group

Fig. 9.

Synergistic beneficial effect of ruxolitinib and deflazacort on skeletal muscle tissues. A-B. H&E staining at 40X and 100X magnification. Large amounts of inflammatory cells are present in the gastrocnemius muscle tissues in the vehicle-treated group. Deflazacort decreased inflammatory cells, but had no effect on heterotopic bone formation (dark red patch-like staining), while the D+R group obviously decreased inflammatory cells and HO formation and improved muscle pathology. C. GLB1/CD68 double immunofluorescence staining of gastrocnemius muscle. GLB1 stained red, CD68 stained green. Double positive cells stained orange or yellow in merged images. Insets highlight double positive cells. D-G. Quantification of CD68⁺/GLB1⁺/200X field, CD68⁺/GLB1⁺/total CD68⁺ percentage, CD68⁺GLB1⁻ cells/200X field, Total CD68⁺cells. H. Double immunofluorescence staining of eMHC and GLB1. eMHC stained green and GLB1 stained red. eMHC does not colocalize with GLB1. I. PAX7/GLB1 double immunofluorescence staining. PAX7 stained green in the nuclei and GLB1 stained red in cytoplasm. J. Quantification of PAX7⁺/GLB1⁻ positive cells. K-L. Q-PCR analysis of M1 macrophage marker iNOS and CD86. M. Q-PCR analysis of M2 macrophage marker CD206. N-P. Q-PCR analysis of myogenic transcription factors PAX7, MYOG and MYOD. Q. Q-PCR analysis of MIF. Scale bars = 250 μm for 40X, 100 μm for 100X and 200X magnification and 50 μm for 400X magnification. N = 8–11 for different groups. Exact P values are indicated between group bars.

(Fig. 9I). Quantification of PAX7⁺/GLB1⁻ cells revealed a non-significant increase in deflazacort group and significant increase in the D+R group compared to the control group (Fig. 9J). Q-PCR results revealed that both the deflazacort and D+R groups significantly decreased iNOS mRNA expression, a marker of M1 macrophages, compared to the control group (Fig. 9K). The D+R group also showed a trend of decreased CD86 compared to the deflazacort group (P = 0.0907). Neither group had a significant change to the M2 macrophage marker CD206 mRNA or the myogenic transcription markers PAX7, MYOG and MYOD at the mRNA level (Fig. 9M–P). D+R also showed a trend of decreased MIF mRNA compared to control group (P = 0.1046) (Fig. 9Q).

In addition, heart histology was also performed to evaluate if deflazacort and ruxolitinib had beneficial effects on the heart tissue. H&E staining revealed that overall, the inflammation and damage were relatively less severe in the heart than in the gastrocnemius muscle (Supplemental Fig.5A). Quantification of the heart sections for each mouse (8–14 sections/animal for both 40X and 100X) demonstrated deflazacort showed a greater percentage of sections with no inflammation compared to the control group (P = 0.081) (Supplemental Fig.5B). Significantly lower percentages of sections with mild inflammation were found for both the deflazacort and D+R treated groups (Supplemental Fig.5C). No significant differences were found for the percentage of medium and severe inflammation sections which were very low between all groups (Supplement Fig.5D–E). A trend of decreased total inflammation section percentage was found between the deflazacort and control groups. The D+R group also showed a relatively lower percentage of total inflammation sections (Supplemental Fig.5F). Furthermore, serum CK activity was measured and revealed that the deflazacort group had significantly lower CK activity than both control and D+R groups while no significant differences were observed between the D+R group and the control group (Supplemental Fig.5G). No significant differences were found between any groups for serum ALT activity (Supplemental Fig.5H).

3.10. Ruxolitinib and deflazacort synergistically improve bone pathology in dKO-Hom mice

Herovici’s staining was performed for the spine and long bone tissues. Deflazacort alone increased pink-stained (collagen 1) trabecular bone quantity compared to the vehicle-treated control group, while D+R increased both pink-stained trabecular bone quantity and thickness for the spine L5 trabecular bone (Fig. 10A). Significant differences were not observed for the pink-stained femur cortical bone (Fig. 10B). For the proximal tibia, both deflazacort and D+R increased pink-stained trabecular quantity compared to the vehicle-treated control (Fig. 10C). H&E staining revealed more pink-stained trabecular bone in the deflazacort-treated group compared to the vehicle-treated group. D+R appeared to increase both trabecular quantity and thickness as compared to the vehicle-treated group (Fig. 10D). For the femur cortical bone, no significant differences were observed between the deflazacort group or the D+R group compared to the control group (Fig. 10E). Moreover, both the deflazacort and the D+R groups had significantly higher pink-stained (collagen 1) trabecular bone than the control group. Notably, more bubble-like adipose cells were found in the bone marrow of the proximal tibia in the deflazacort group compared to the control group. D+R also demonstrated more bubble-like adipose cells compared to the control group, but not more than the deflazacort group, which indicated this effect is likely caused by deflazacort rather than ruxolitinib (Fig. 10F).

Fig. 10.

Herovici’s and H&E staining of bone tissues treated with deflazacort and D+R.A. Herovici’s staining of spine L5 trabecular bone. Collagen 1 stained in pink-red, collagen 3 stained dark blue, bone marrow stained light blue. B. Herovici’s staining of femur cortical bone. Cortical bone stained in dense pink-red. C. Herovici’s staining of proximal tibia trabecular bone. Trabecular bone stained in pink red. D. H&E staining of spine L5 trabecular bone. E. H&E staining of femur cortical bone. F. H&E staining of proximal tibia trabecular bone. More trabecular bones were observed in the deflazacort and D+R groups. Many bubble-like adipose cells were present in deflazacort and D+R groups, but no differences between deflazacort and D+R groups were observed. N = 8–11 for different groups, including males and females. Scale bar = 200 μm.

3.11. Ruxolitinib and deflazacort synergistically affect osteoblasts and osteoclasts

Immunohistochemistry staining was used to detect osteocalcin-positive (OCN⁺) osteoblasts for spine trabecular bone. Deflazacort significantly increased OCN⁺ osteoblasts in the spine L5 trabecular bone compared to the control group. Furthermore, D+R also significantly increased OCN⁺ osteoblasts compared to the control group, but not significantly more than the deflazacort group (Fig. 11A–B). Interestingly, D+R also significantly increased the TRAP⁺ osteoclasts of the spine L5 trabecular bone compared to the control and deflazacort groups (Fig. 11C–D). For the proximal trabecular bone, deflazacort significantly increased the OCN⁺ osteoblasts in the proximal tibia trabecular bone, while D+R did not significantly increase OCN⁺ osteoblasts compared to the control or deflazacort groups (Fig. 11E–F). Furthermore, TRAP⁺ osteoclasts in the proximal tibia were significantly increased in the deflazacort group compared to the control group, while D+R significantly decreased TRAP⁺ osteoclasts compared to the deflazacort group (Fig. 11G–H).

Fig. 11.

Synergistic effect of ruxolitinib and deflazacort on bone cells. A. Immunohistochemistry staining of osteocalcin to detect osteoblasts. Osteoblasts are stained brown on the trabecular bone surface. B. Quantification of OCN⁺ cells/bone surface. C. TRAP staining for osteoclasts of spine L5. Osteoclasts stained violet-red on the bone surface with multiple nuclei. D. Quantification of TRAP⁺osteoclasts on the bone surface of spine L5 vertebra trabecular bone. E. Immunohistochemistry staining of OCN of the proximal tibia. Osteoblasts stained brown and were located on the bone surface. F. Quantification of OCN⁺ osteoblasts on bone surface. G. TRAP staining for the proximal tibia trabecular bone. TRAP⁺ osteoclasts stained violet red on the bone surface. H. Quantification of TRAP⁺osteoclasts. Scale bar = 100 μm. N = 8–11 for different groups, including both males and females. Exact P values are indicated between group bars.

4. Discussion

The most important findings of this study are that both Mdx and dKO-Hom mice exhibit severe cellular senescence and SASP in skeletal muscle but not in bone. The GLB1/CD68 double immunofluorescence staining demonstrated that the senescent cells are predominantly macrophages. Targeting senescent cells with ruxolitinib decreased senescent macrophages, ameliorated muscle pathology, and improved trabecular bone microarchitecture of dKO-Hom mice. More strikingly, ruxolitinib treatment also increased the lifespan of the dKO-Home mice. Furthermore, ruxolitinib treatment improved the grip strength and treadmill endurance of Mdx mice. Finally, treatment of dKO-Hom mice with ruxolitinib and deflazacort synergistically improved the bone microarchitecture of the spine L5 vertebrae and the proximal tibia trabecular bone, and reduced muscle pathology despite not improving cortical bone biomechanical properties.

This study demonstrated that both Mdx and dKO-Hom mice expressed significantly higher levels of GLB1, FUCA1, P16, and P21 in their muscle tissues. SASP genes IL6, PAI-1, MIF, and IL1-β were all significantly increased in the Mdx and dKO-Hom mice. Furthermore, increased senescent cells were predominantly located in the skeletal muscle tissues, but not in the bone tissues, as revealed by P21 and GLB1 immunohistochemistry staining. The senescent cells were mainly macrophages as demonstrated by CD68/GLB1 double staining. A previous study using dystrophic rats demonstrated that the senescent cells are satellite cells and mesenchymal progenitor cells in the muscle tissue, and that P21, P16, and P19 were significantly increased in DMD rats. P16, P14, and P21 were also increased in the muscle tissues of DMD patients compared to non-DMD patients [39]. Others using a D2 Mdx mouse model showed that senescent cells are CDKN1 (P21)-positive macrophages, while in the C57-Mdx mice, the senescent cells are endothelial cells [51]. In our study using Mdx and dKO-Hom mice, we found CD68⁺ macrophages are predominantly the cells that become senescent in the muscle tissues using colocalization of GLB1/CD68 staining, and not GLB1⁺/CD68⁻ cells (non-macrophage senescent cells) because there were no differences between the Mdx, dKO-Hom, and WT mice. Further, immunohistochemistry staining for P21 and GLB1 revealed a large number of P21⁺ cells or GLB1⁺ cells in the damaged inflammatory area of the skeletal muscle of both the Mdx and dKO-Hom mice, but not in their bone tissues. CD31⁺ endothelial cells also did not become senescent, as the untreated dKO-Hom mice had very few FUCA1⁺/CD31⁺ cells, while senolytic drug treatment did not change this pattern. It is therefore believed that macrophages are the predominant cells that become senescent in both the Mdx and dKO-Hom mice. Indeed, a previous study in dKO-Hom mice also demonstrated that mesenchymal progenitor cells proliferated more in the dKO-Hom mice’s muscle tissues than in the WT mice’s muscle tissues [26].

Our results also demonstrated that targeting cellular senescence with ruxolitinib improved muscle histopathology in the dKO-Hom mice by decreasing inflammation and senescent macrophages (GLB1⁺/CD68⁺, FUCA1⁺/CD68⁺, and P21⁺/CD68⁺ cells) in the skeletal muscle tissues but not GLB1⁺CD68⁻ or FUCA1⁺CD68⁻ or P21⁺/CD68⁻ cells, as all of these are non-macrophage senescent cells. However, ruxolitinib treatment did not have any effect on FUCA1⁺CD31⁺ (senescent endothelial cells) which are very rare (about 1 in 200X field), while the FUCA1⁺/CD31⁻ cells (non-endothelial senescent cells) were significantly decreased by ruxolitinib treatment. Interestingly, the number of FUCA1⁻/CD31⁺ (non-senescent endothelial cells) decreased after ruxolitinib treatment which likely occurred due to reduced inflammation and decreased cellularity in the damaged muscle tissue area. Ruxolitinib treatment also improved heart pathology by decreasing GLB1⁺/CD68⁺ cells.

Several previous studies demonstrated that DMD patients are at increased risk of fracture and often become wheelchair-bound due to fractures [11,12,52]. Our previous study also demonstrated that dKO-Hom mice developed bone osteopenia 4 weeks after birth and had delayed fracture healing [29,30]. Therefore, it was further investigated if treatment with ruxolitinib improved bone microarchitecture in dKO-Hom mice. Our results revealed that dKO-Hom mice treated with ruxolitinib for 12 days had significantly increased bone mass as demonstrated by increases in both BV/TV and Tb.Th of spine L5 trabecular bone. This is especially encouraging because DMD patients also suffer from osteopenia and vertebral fractures. A previous study has demonstrated that ruxolitinib increased the BV/TV and Tb.Th of aged mice, while ruxolitinib treatment also suppressed osteoclasts without affecting osteoblasts [34]. Ruxolitinib treatment also decreased the number of spontaneous fractures and increased hind limb and whole body mineral contents in a mouse model of Hutchinson–Gilford progeria syndrome (HGPS), increased grip strength, and improved other aging phenotypes [53]. In this study, ruxolitinb treatment also improved grip strength and treadmill endurance of C57 Mdx mice. Ruxolitinib treatment also extended the lifespan of dKO-Hom mice and heart histology, indicating improvement of general health. A previous study demonstrated that the ablation of P16 in DMD rats increased PAX7⁺ and MyoD⁺ cells, while decreasing some SASP such as TGFβ1, CTGF, and MMP2, but the rats developed rhabdomyosarcoma at 9-months-old. The senolytic drug ABT263 increased muscle regeneration, maintained muscle strength and decreased SASP such as IL6, TGFβ1 and IL1β [39]. The authors, however, did not investigate if the treatment of the senolytic drug improved bone quality in the DMD rats. The current study further demonstrated ruxolitinib treatment decreased osteoclast numbers in spine trabecular bone and trended towards an increase in osteoblast number on the bone surface. These findings are consistent with Farr J et al.’s results in aged mice [34]. Ruxolitinib treatment was also found to decrease GLB1 and MIF expression in the dystrophic muscle tissue. TRAP and BMP4 expression was further decreased in ruxolitinib-treated mice compared to vehicle-treated mice.

However, dKO-Hom mice treated with D+Q did not improve bone microarchitecture or muscle pathology in the current study. A recent clinical trial also demonstrated that D+Q did not significantly reduce bone resorption in post-menopausal women despite its effectiveness in decreasing osteoclasts in aged mice [34,54]. Fisetin at 20 mg/Kg also did not improve bone microarchitecture in dKO-Hom mice. These results imply that the choice of senolytic drug (dosing regimen, timing after disease progression) in each specific disease should be mechanistically and disease specific as cellular senescence can be induced by many adverse factors. In the Mdx and dKO-Hom mice, both senescent and SASP genes were dramatically elevated in the skeletal muscle tissues, and the macrophages were also dramatically increased, therefore, targeting Jak1/2 with ruxolitinib is more effective than both D+Q and fisetin. The duration of senolytic drug treatment also plays a role. A previous study using dKO-Hom mice has shown that fisetin treatment at 20 mg/Kg for 4 weeks decreased macrophage number and increased PAX7⁺ satellite cells [38]. Additionally, the timing of the treatment may also play a role. In the current study, the mice were treated at 4 weeks of age while the literature [38] did not specify the age of mice.

To facilitate translation, ruxolitinib was further tested to see if it had a synergistic effect when combined with deflazacort, a commonly used steroid for the treatment of DMD patients. Remarkably, deflazacort treatment alone for 12 days also significantly increased BV/TV and Tb. N, while decreasing Tb.Sp, and trended towards an increase in Tb.Th for the spine L5 vertebrae and the proximal tibia trabecular bone compared to vehicle-treated dKO-Hom mice. Even more encouraging, simultaneous treatment of deflazacort and ruxolitinib more dramatically increased BV/TV than deflazacort alone by increasing Tb.Th and decreasing Tb.Sp, a finding most obvious in proximal tibia trabecular bone. Although both the deflazacort and D+R groups did not increase femoral cortical bone Ct.Th, both treatments significantly increased BV density in the femur cortical bone. The finding that tibia biomechanical properties did not change is likely due to the fact that prior to treatment, the cortical bone was thinner and the mice were smaller in size in the deflazacort or D+R groups relative to control group. These results demonstrate the enhanced benefits of combining deflazacort and ruxolitinib. Furthermore, immunohistochemistry staining demonstrated that deflazacort increased osteoblasts in the bone tissues while ruxolitinib counteracted the increase of osteoclasts by deflazacort at least in the proximal tibia trabecular bone. The finding that combining deflazacort and ruxolitinib increased both Tb.N and Tb.Th indicates new bone formation both on the existing trabecular bone (Tb.Th) as well as the generation of new trabecular bone (increasing in Tb.N). Our data also showed that ruxolitinib further improved the benefit of deflazacort on muscle pathology by decreasing the number of senescent macrophages (GLB1⁺/CD68⁺), while the combined use of deflazacort and ruxolitinib also decreased the senescent macrophage percentage despite not significantly decreasing non-senescent macrophages (GLB⁻/CD68⁺).

A limitation to the generalizability of the study is that it did not consider gender/sex issues even if we utilized both males and females throughout the study. Furthermore, to translate the results into clinical application for the treatment of patients with DMD requires further longterm evaluation as well as monitoring of potential side effects.

5. Conclusion

In summary, this study demonstrated severe cellular senescence and SASP in the muscle tissues of both Mdx and dKO-Hom mice, but not in the bone tissues. The senescent cells were mainly macrophages. Targeting cellular senescence in the muscle tissue of dKO-Hom mice improved muscle and heart histopathology as well as trabecular bone microstructure while extending the lifespan of dKO-Hom mice. Ruxolitinib also improved the grip strength and treadmill endurance of Mdx mice. Furthermore, combined ruxolitinib and deflazacort treatment synergistically enhanced the beneficial effects of deflazacort and ruxolitinib on bone microarchitecture, muscle, and heart pathology. This study provides new insights on the application of senotylic pharmaceuticals to treat DMD patients by protecting both bone microarchitecture and improving muscle pathology.

Appendix A. Supporting information

Supplementary data associated with this article can be found in the online version at doi:10.1016/j.phrs.2025.108016.

Data availability

Data will be made available on request.