Background: In the search of genetic determinants of Duchenne muscular dystrophy (DMD) severity, LTBP4,a member of the latent TGF-β binding protein family, emerged as an important predictor of functional outcome trajectories in mice and humans. Nonsynonymous single-nucleotide polymorphisms in LTBP4 gene associate with prolonged ambulation in DMD patients, whereas an in-frame insertion polymorphism in the mouse LTBP4 locus modulates disease severity in mice by altering proteolytic stability of the Ltbp4 protein and release of transforming growth factor-β (TGF-β). Givinostat, a pan-histone deacetylase inhibitor currently in phase III clinical trials for DMD treatment, significantly reduces fibrosis in muscle tissue and promotes the increase of the cross-sectional area (CSA) of muscles in mdx mice. In this study, we investigated the activity of Givinostat in mdx and in D2.B10 mice, two mouse models expressing different Ltbp4 variants and developing mild or more severe disease as a function of Ltbp4 polymorphism. Methods: Givinostat and steroids were administrated for 15 weeks in both DMD murine models and their efficacy was evaluated by grip strength and run to exhaustion functional tests. Histological examinations of skeletal muscles were also performed to assess the percentage of fibrotic area and CSA increase. Results: Givinostat treatment increased maximal normalized strength to levels that were comparable to those of healthy mice in both DMD models. The effect of Givinostat in both grip strength and exhaustion tests was dose- dependent in both strains, and in D2.B10 mice, Givinostat outperformed steroids at its highest dose. The in vivo treatment with Givinostat was effective in improving muscle morphology in both mdx and D2.B10 mice by reducing fibrosis. * Correspondence: firstname.lastname@example.org New Drug Incubator, Italfarmaco S.p.A., Milan, Italy University of Milano-Bicocca, Milan, Italy Full list of author information is available at the end of the article © The Author(s). 2021 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Licandro et al. Skeletal Muscle (2021) 11:19 Page 2 of 22 Conclusion: Our study provides evidence that Givinostat has a significant effect in ameliorating both muscle function and histological parameters in mdx and D2.B10 murine models suggesting a potential benefit also for patients with a poor prognosis LTBP4 genotype. Keywords: Duchenne nuscular dystrophy, Givinostat, mdx, D2.B10, LTBP4, HDAC inhibitor Background predictor of functional outcome trajectories. The LTBP4 Duchenne muscular dystrophy (DMD) is a lethal X- gene encodes the latent transforming growth factor-β linked disorder that leads to muscle wasting with an (TGF-β)-binding protein (Ltbp) 4 that binds TGF-β in average incidence of 1:3500 newborn males. DMD is due the extracellular matrix, sequestering this cytokine [16– to mutations in the gene encoding dystrophin, which 18]. During inflammatory processes, TGF-β is released cause the absence of the protein. The lack of dystrophin from the Ltbp4 complex by proteolysis of the proline- leads to muscle fiber degeneration, activation of chronic rich hinge domain. This cleavage leads to the activation inflammatory pathways and progressive muscle tissue re- of TGF-β . Once liberated from the complex, free placement by fibroblasts and adipocytes, triggering the TGF-β regulates collagen deposition and promotes fi- process leading to fibrosis development . In DMD brotic processes . During the physiological healing muscles, the normal regenerative pathways are subverted process, a transient release of TGF-β is necessary, and the abnormal substitution of damaged muscle fibers whereas sustained pro-inflammatory cytokine levels con- by fibrotic and adipose tissue leads to a severe reduction tribute to the pathological tissue degeneration processes, in muscle function. The main causes of the reduced life such as fibrosis . High TGF-β levels have been span of DMD patients are severe respiratory insuffi- shown to correlate with the severity of muscle fibrosis in ciency, due to the weakening of the diaphragm (DIA), DMD . and cardiac failure . In humans, single-nucleotide polymorphisms (SNPs) To date there is no standard therapy for DMD patients have been identified in the Ltbp4 protein, which origin- that leads to the healing of the disease; however, gluco- ate VTTT and IAAM haplotypes . DMD patients corticoid (GC) steroid treatment, corrective orthopedic homozygous for the IAAM Ltbp4 haplotype remained surgery, and assisted ventilation can contribute to im- ambulatory significantly longer (12.5 ± 3.3 years) than prove the quality of life of patients and to delay disease those homozygous for the VTTT haplotype (10.7 ± 2.1 progression . Attempts to specifically target individual years) . In addition, IAAM fibroblasts exposed to mutations have recently led to the approval of exon TGF-β show a reduction in phospho-SMAD levels when skipping oligonucleotides (Eteplirsen, Golodirsen, and compared to VTTT fibroblasts, in line with the concept Vitolarsen) by the FDA [4–6] and Ataluren for patients that LTBP4 regulates TGF-β activity [17, 22, 23]. with nonsense mutations by EMA [7, 8]. Only a minority There are two mouse strains that mimic the two dif- mdx of patients can benefit from these treatments, and the ferent human Ltbp4 haplotypes: C57BL10ScSn-Dmd / mainstay of disease management is GC steroids. GC J (hereafter referred to as mdx) have a 12-amino-acid in- steroid treatment (mainly prednisone and Deflazacort) sertion in the proline-rich region of Ltbp4. This contrib- was shown to beneficially influence all disease trajector- utes to a mild DMD phenotype due to a lower sensitivity ies, including survival and prolonged ambulation in to proteolysis and a reduced activation of TGF-β as it DMD patients [9–11], but this benefit comes at the cost occurs in the human IAAM haplotype ; D2.B10- mdx of significant side effects such as body weight (BW) in- Dmd /J (hereafter referred to as D2.B10) have a 12- crease, growth stunting, Cushing-like symptoms, mood amino-acid deletion in the same Ltbp4 region and func- changes, increased incidence of fractures, and suscepti- tionally resemble the human VTTT haplotype (severe bility to infections [12, 13]. Paradoxically, chronic GC DMD phenotype due to increased sensitivity to proteoly- steroid administration may promote muscle atrophy, sis and increased TGF-β activity) . Mdx mice harbor and mdx mice subjected to chronic GC steroid treat- a spontaneous point mutation in exon 23 of the dys- ment show a significant impairment of heart function trophin gene, leading to the loss of dystrophin  and and an increase of cardiac fibrosis, suggesting that pro- are routinely used as a rodent model of the DMD dis- longed steroid treatment may be detrimental to dys- ease even though it has a milder phenotype as compared trophic heart muscle [14, 15]. Thus, there is a clear need to DMD patients and a normal lifespan. The compara- for a broadly active and well-tolerated treatment for tively mild phenotype of mdx mice can also be attributed DMD patients. to the compensatory function of the dystrophin-related Recently, in the search of genetic determinants of protein utrophin, which is highly upregulated in regener- DMD severity, the LTBP4 gene emerged as an important ating muscle fibers in adult mdx mutants, except for the Licandro et al. Skeletal Muscle (2021) 11:19 Page 3 of 22 DIA muscle . The D2.B10 strain, created by back- Materials and methods crossing mdx mice onto the DBA/2J background, may Animal experiments be a superior DMD model as it better recapitulates sev- Study approval eral of the human characteristics of the DMD myopathy Procedures involving animals and their care were carried (reduced hindlimb muscle weight, fewer and atrophied out in conformity with institutional guidelines in com- myofibers, increased fibrosis, fat deposition and inflam- pliance with national and international laws and policies matory infiltrate, muscle weakness) when compared to (Italian Governing Law: D.lgs 26/2014 “Attuazione della strains with this mutant allele on other genetic back- direttiva 2010/63/UE sulla protezione degli animali uti- grounds (such as mdx mice) [24–26]. lizzati a fini scientifici”). The research project has been Pharmacological blockade of histone deacetylases authorized by the Italian Ministry of Health. (HDACs) decreases fibrosis and promotes compensa- tory regeneration in the mdx skeletal muscle [27–30]. Animals and study design The role of HDACs in the DMD muscle is not fully Mice were kept under pathogen-free conditions with a understood; however, if comparing the muscles of 12-h light/12-h dark cycle at a temperature of 22° ± 2° mdx mice to that of healthy C57BL/10 J mice, it can and 55% ± 10% humidity. Each cage was enriched with a be observed that the HDAC activity is high in mdx mouse house. Mice were regularly checked by a certified mice . Givinostat is a potent histone deacetylase veterinarian who was responsible for health monitoring, inhibitor (HDACi) currently in a phase III clinical animal welfare supervision, experimental protocols, and trial (www.Clinicaltrials.gov, clinical trial identifier: procedure revision. All mice were maintained under a NCT02851797) for DMD treatment. Givinostat treat- controlled diet (VRF1 diet, Charles River), with a daily ment has been effective in ameliorating morphology amount of chow of 4–5 g/mouse throughout the experi- and muscular function in mdx mice. In particular, ment  and received water ad libitum. Givinostat significantly reduces fibrosis in muscle tis- For the efficacy studies, C57BL/10 J (Stock No: sue and promotes the increase of cross-sectional area mdx 000665) wild type (wt) and C57BL/10ScSn-Dmd /J (CSA) of the muscles . In addition, immunohisto- (Stock No: 001801) 7–8-week-old male mice, DBA/2J wt chemical analysis performed on muscle biopsies of mdx (Stock No: 000671), and D2.B10-Dmd /J (Stock No: DMD patients demonstrates that Givinostat reduces 013141) 6–7-week-old male mice were purchased from inflammation, necrosis, and fibrosis in muscle tissue The Jackson Laboratories (Bar Harbor, Maine, USA). and promotes increase of the CSA of the myofibers After 5 days of acclimatization in the animal facility, that occupy a larger fraction of muscle tissue . mice were randomized into different treatment groups Theinfluenceof theLtbp4 polymorphism on theac- based on their BW. C57BL/10J wt mice were assigned to tivity of Givinostat is unknown. the naive wt group (healthy mice that received no treat- This study was therefore aimed at evaluating the ef- ment, but were subjected to functional tests), whereas fect of a long-term oral treatment with Givinostat mdx mice were assigned to the following groups (9 (15 weeks) in the two DMD mouse models. Moreover, mice/group; 4–5 mice/cage): naive mdx (mice that re- we also evaluated the effect of prednisone and Defla- ceived no treatment, but were subjected to functional zacort in D2.B10 mice, since this strain bears a closer tests), vehicle (0.5% methylcellulose, p.o.), and Givinostat resemblance to human DMD pathology [15, 32]. To (0.1, 0.3, 1, 5, 10, 25, and 37.5 mg/kg, p.o.). DBA/2J mice evaluate the efficacy of Givinostat and steroid treat- were assigned to the naive wt group (healthy mice that ment, we assessed muscle function and mice fatigabil- received no treatment, but were subjected to functional ity using in vivo behavioral tests (i.e., grip strength tests), whereas D2.B10 mice were assigned to the follow- and treadmill apparatus). Histopathological analyses ing groups (12–13 mice/group; 4–5 mice/gage): vehicle were also performed to assess the impact of Givino- (filtered tap water, p.o.), Givinostat (1, 5, 10, and 37.5 stat and steroid treatment on tissue morphology in mg/kg, p.o.), prednisone 1 mg/kg, and Deflazacort 1 mg/ terms of CSA, centralized nuclei, and fibrosis by kg (i.p.). histopathological analysis. We show that Givinostat mdx For the PK study, C57BL/10ScSn-Dmd /J (Stock has a significant and dose-dependent effect on muscle No: 001801) 9-week-old male mice were purchased from function in both models suggesting a potential benefit The Jackson Laboratories (Bar Harbor, Maine, USA). also on patients with the poor prognosis LTBP4 geno- After 5 days of acclimatization in the animal facility, type. In addition, our dose-response studies shed light mdx mice were randomized into 4 treatment groups (28 on possible correlations between efficacy, histologic mice/group; 4 mice/cage) and 1 untreated group (4 parameters, and pharmacokinetic (PK) analysis that mice) based on their BW. Animals were administered could lead to further dose optimizations in the treat- with single oral dose of Givinostat at 5, 10, 25, and 37.5 ment of DMD patients. Licandro et al. Skeletal Muscle (2021) 11:19 Page 4 of 22 mg/kg and samples were collected at the following time NMD Neuromuscular Network) . The highest re- points: 0.5, 1, 1.5, 2, 3, 4, and 6 h after administration. corded value of maximal normalized strength (FNmax) obtained for each mouse at all the time points (from T0 Drug treatments to T15) was used for further analysis. In the efficacy studies, Givinostat powder (ITF2357) was suspended in 0.5% methylcellulose (Sigma-Aldrich) and Treadmill exercise administered p.o. (by gavage) at the doses of 0.1, 0.3, 1, Run to exhaustion performance of mdx and D2.B10 5, 10, 25, and 37.5 mg/kg qdx5x15weeks (administration mice was evaluated every 14 or 21 days, respectively, volume per mouse: 10 mL/kg) in mdx mice (see Add- using a treadmill apparatus (Ugo Basile SRL, Italy). The itional Figure 1A). Vehicle-treated mdx mice received exhaustion protocol consisted in an initially horizontal 0.5% methylcellulose suspension. Givinostat suspension running at 5 m/min for 5 min after which the speed was was stored at + 4 °C, and it was freshly prepared every 7 increased 1 m/min every minute until exhaustion (the days. Probably due to the impairment of their tongue procedure was compliant with the standard operating and masticatory muscles, the oral gavage procedure procedures of TREAT–NMD Neuromuscular Network) turned out to be difficult to perform, and chronic treat- . The test was concluded when mice remained for ment of D2.B10 mice by this route was considered to more than 10 s on the shocker plate (for both C57BL/10 bear a high risk of damaging the esophagus. For this rea- J and mdx mice) or when D2.B10 mice reached a max- son, Givinostat was administered in the drinking water. imum number of shocks. Due to the excessive stress Givinostat was dissolved in filtered tap water and was caused to D2.B10 mice by this functional test, a max- administered in drinking water (throughout 24 h) for imum number of 600 and 150 shocks was set. These ex- 105 days at the dose of 1, 5, 10, and 37.5 mg/kg/day, ac- perimental conditions were also applied to the DBA/2J cording to a daily estimate of water consumption of 4 wt healthy mice. ml/mouse/day . Water consumption was weekly After taking baseline readings for behavioral assays monitored throughout the entire duration of the study (FNmax and distance run by the animals), mice were re- weighing the bottle of each cage (maximum 4 mice per randomized based on these two parameters, also consid- cage). Prednisone and Deflazacort powder (Sigma-Al- ering the previous randomization performed on their drich) was dissolved in 2% DMSO (Sigma-Aldrich) and BW. diluted in sterile saline on injection days. Both drugs The grip strength and run to exhaustion tests were were administered weekly (for 15 weeks) by i.p. injection performed during the light-cycle phase, in the early at the dose of 1 mg/kg  in a volume of 10 mL/kg. hours of the morning (8:00–11:30 am). Both the func- Givinostat and steroid treatments started on days 7 and tional tests have been conducted after a training period 9 of the experimental plan, respectively (see Additional of 3 and 4 days, respectively, during which mice become Figure 1B). All mice were monitored daily and tolerabil- familiar with the procedures. The training period to the ity was evaluated on the basis of BW and clinical signs. functional tools started after the acclimatization period No treatment-related clinical signs were observed at any (see Additional Figure 1A and B). time point during the studies. In the PK study, formulations were prepared suspend- ing Givinostat in 0.5% methylcellulose at the concentra- Pharmacokinetic study tions of 0.5, 1, 2.5, and 3.75 mg/mL. The administered Blood, plasma, and muscle collection volume was 10 mL/kg to obtain 5, 10, 25, and 37.5 mg/ Mice were maintained under isoflurane anesthesia. kg dosages. Blood samples were collected in two aliquots from the retro-orbital plexus of mice and transferred into tubes Assessment of functional tests for the evaluation of containing heparin as anticoagulant (100 USPunits/mL) treatment efficacy and then mice were sacrificed by cervical dislocation. Maximal normalized strength One aliquot of 50 μL was diluted with the same volume In both efficacy studies, the effectiveness of treatments of water and immediately frozen in dry ice. This blood was evaluated every week by measuring the forelimb was analyzed in order to subtract the blood content of strength by a grip strength meter (Ugo Basile SRL, Italy). Givinostat from the muscle homogenate. The other ali- BW of mdx (see Additional Figure 2) and D2.B10 mice quot of blood was centrifuged for 10 min at 13,000 rpm (see Additional Figure 3) was also measured in order to at 4 °C and plasma was separated and frozen in dry ice. normalize the absolute grip strength of each mouse with The quadriceps of the right hindlimb was removed, respect to its BW. The protocol provides 5 measure- weighed, and then washed twice in 1 mL of PBS pH 7.4 ments for each mouse (the procedure was compliant and dried with paper, then it was frozen in dry ice. with the standard operating procedures of TREAT– Licandro et al. Skeletal Muscle (2021) 11:19 Page 5 of 22 Plasma analysis different time points to be examined by a pathologist Givinostat was determined in mouse plasma (50 μL) fol- (for more details see Additional Table 1). Due to the re- lowing a protein precipitation (200 μL of 1% HCOOH in sults obtained by the functional tests, muscles from only acetonitrile) and Ostro (Waters) plate filtration. The or- some treatment groups were analyzed in both DMD ganic phase was evaporated under a stream of nitrogen; models (see Additional Table 1). All the collected muscles the residue was re-dissolved in the reconstitution solvent were fixed in 10% buffered formalin solution (Bio-optica, (100 μL of 25% CH3CN-75% H O-0.05% TFA), and 5 μL Milan, Italy) at + 4 °C for at least 48 h. After fixation, mus- was analyzed. Givinostat concentrations were measured cular samples were transversely trimmed, caged, and par- by LC-MS/MS method, in the calibration range of 0.5– affin embedded overnight with an Automated Vacuum 400 ng/mL. Tissue Processor Floor (ETP, Histo-Line Laboratories) and included in paraffin blocks. Serial transverse cross- Blood analysis sections (4 μm thick) were cut with a microtome (Leica Givinostat was determined in diluted blood (100 μL, 1:2 RM 2255), collected onto uncoated glass slides, and in water) with a liquid-liquid extraction. Samples were stained using a standard protocol for hematoxylin and mixed and extracted with 2 mL of diethyl ether and cen- eosin (Mayer hematoxylin & Aqueous G Eosin 1%, Bio- trifuged, and the organic phase was separated and evap- Optica) and Sirius Red staining (Direct Red 80, Sigma- orated under a stream of nitrogen. The residue was re- Aldrich). For the image analysis and quantitation, dissolved in the reconstitution solvent (150 μL of 25% hematoxylin and eosin- and Sirius Red-stained slides were CH3CN-75% H O-0.05% TFA), sonicated, and filtered examined with an Olympus BX51 light microscope. From (regenerated cellulose syringe filters), and 5 μL was ana- each sample, 1 to 3 random microphotographs at the lyzed by LC-MS/MS method. The calibration range of magnification of × 4 and × 10 (TA, GAS and heart) or × the method was 1–200 ng/mL. 20 (DIA) were collected using Image-Pro Plus system. The digital images were processed by a pathologist using Quadriceps analysis ImageJ software (U.S. National Institutes of Health, Be- Mouse quadriceps were homogenized in PBS pH = 7.4 thesda, MD, USA). To quantify CSA and centralized nu- (1:10 w/v) with ultrasonic homogenizer (Sonoplus clei (%) on hematoxylin and eosin-stained sections, one Mini20 - Bandelin) maintaining the samples in an ice hundred myofibers were manually measured for each bath. Quadricep homogenate (100 μL) was added with muscle/mouse using a tablet pen (Wacom Intruos) and 200 μL of 0.5% TFA in CH3CN, vortexed, and centri- analyzed by the ImageJ software after a linear calibration fuged. Then, 100 μL of supernatant was diluted with of 150 μm. The fibrotic area, corresponding to the area H O (1:2). Samples were filtered (regenerated cellulose stained in red, was compared to the total area of the tissue syringe filters) and aliquots of 5 μL were analyzed. Givi- section, and the results were expressed as percentage fi- nostat concentrations were measured by LC-MS/MS brosis. The % of fibrosis of muscles in the different treat- method, in the calibration range of 1–40 ng/mL. ment groups was expressed by averaging the three values obtained from each muscle. Calculation In these experiments, the histopathological evaluation PK parameters were evaluated on mean curves using of muscle inflammatory infiltrate, adipose tissue depos- conventional non-compartmental methodology by the ition, regeneration, and degeneration of the muscular tis- software KineticaTM v. 5.1. Quadriceps concentrations sue was also performed. The severity of myodystrophy were expressed as nanograms per gram of muscle, this in the different muscles (TA, GAS and DIA) was quanti- value was obtained by multiplying the homogenate con- fied through a weighted histopathological method that centrations obtained for the homogenate volume, then considered some parameters scored by severity and ex- subtracting the residual blood content, as reported in lit- tension of the injury: muscle degeneration/necrosis, re- erature , considering the measured blood concentra- generation, inflammatory infiltrate, interstitial reaction, tions and dividing for the muscle weight. and adipose tissue deposition. Each parameter was clas- sified by severity (mild = 1, moderate = 2, and severe = Histological analysis 3) and extension (focal = 1, multifocal = 2, and diffuse = Tissue preparation, sectioning, and staining 3). The individual severity score was calculated for each Tibialis anterior (TA), gastrocnemius (GAS), DIA, and animal and an average score per group was calculated. heart were collected from mice maintained under iso- flurane anesthesia and sacrificed by cervical dislocation. Whole-genome miRNA expression profiling These muscles were collected from C57BL/10J wt and Plasma collection mdx (5 mice/group) at the end of treatments and from Blood samples were taken from the retro-orbital plexus DBA/2J wt and D2.B10 mice (5 mice/group) at two of mice maintained under isoflurane anesthesia. Blood Licandro et al. Skeletal Muscle (2021) 11:19 Page 6 of 22 was placed in tubes containing 50 μL (for 1 mL of total Statistical analysis blood) of EDTA (100 nM) and then it was centrifuged at Statistical analysis was performed using a GraphPad 13,300 rpm at + 4 °C for 10 min. Plasma was separated Prism version 8 software. (400–500 μL) and immediately frozen in dry ice and All experimental data were expressed as mean ± stored at − 80 °C until analysis. standard error (s.e.). Multiple statistical comparisons be- These analyses were carried out by Biogazelle (Tech- tween groups were performed by two-way ANOVA nologiepark 82, 9052 Zwijnaarde, Belgium). (mice BW, FNmax, and distance run) and by one-way Thirteen mouse plasma samples (naive wt, n = 3; naive ANOVA (centralized nuclei and fibrosis percentage), mdx, n = 3; vehicle, n = 4; Givinostat 37.5 mg/kg, n = 3) with Bonferroni’s multi-comparison test. P values ≤ 0.05 were delivered to Biogazelle in dry ice. were considered as statistically significant and are indi- cated in each figure legend; p values > 0.05 were consid- ered as statistically not significant (ns). RNA extraction RNA was isolated using the miRNeasy Serum/Plasma CSA statistical analysis Kit (Qiagen) following the manufacturer’s instructions. The deciles of the CSA of the animals treated with Givi- nostat at the different doses were plotted against the Small RNA sequencing and data processing corresponding deciles of the vehicle-treated mice and a Libraries for small RNA sequencing were prepared using linear regression analysis was performed on natural the NEBNext small RNA library prep kit (New England [CSAtreated = b CSAcontrol + a] and log_transformed Biolabs) according to the manufacturer’s instructions. data [log(CSAtreated) = log(a) + b log(CSAcontrol)]. If Briefly, 5 μl of RNA eluate was used as input for RNA the shape of the two distributions is not significantly al- adapter ligation (using 3′ and 5′ RNA adapters) followed tered by Givinostat treatment, the value of the slope (b) by reverse transcription and PCR amplification with bar- should not differ significantly from 1; in addition, if the coded primers. Size selection of individual libraries was value of the intercept (a) is significantly different from performed on a Pippin Prep system (Sage Science) to re- zero, the presence of a shift (S) between the two CSA- cover the ~ 147–157-nt fractions containing mature treated and CSAcontrol distributions may be considered. miRNAs. Next, qPCR-based normalization was used for In case of untransformed data, this means that the effect pooling of the individual small RNA libraries. Finally, of Givinostat treatment is additive and any decile of the the resulting small RNA pools were concentrated via CSAtreatment distribution may be derived from the con- ethanol precipitation and quantified using the Qubit 2.0 trol data summing up a constant quantity (S) to each fluorometer (Thermo Fisher Scientific) prior to sequen- fiber AREA of the control group (CSAtreated = CSA- cing with read length of 75 bp on a NextSeq 500 sequen- control + S). In case of log_transformed data, the effect cer (Illumina). of Givinostat has to be considered as a multiplicative Sequencing reads were filtered based on stringent read factor on any generic fiber of the muscle under evalu- quality control. After adapter trimming with Trimmo- ation, whose size, after the treatment, results K times matic, reads were collapsed and mapped to the reference larger in comparison to the control group value (CSA- genome (GRCm38) using Bowtie . Reads up to 25 treated the S or K values = K × CSAcontrol, with K = nucleotides are mapped with no mismatches, while lon- 10^a). The S or K parameter values were then deter- ger reads (mainly from RNA species other than miR- mined through an iterative estimation procedure, based NAs) are mapped allowing a maximum of 2 mismatches. on sum of squares (SSQ) minimization, directly compar- Using genome annotation data from Ensembl (release ing the CSA distribution observed in treated animals 84), UCSC (mm10), and miRbase (release 21), mapped with that obtained summing up the S value or multiply- reads were subsequently annotated to mature miRNAs ing by K the CSA values of the control group. and other small RNA species including tRNA fragments, Differential miRNA abundance was performed using rRNA, sn(o)RNAs, and piRNAs. Raw count data is re- the same DESeq2 package. A false discovery rate (FDR) ported for each isomiR and each mature miRNA (as the of 5% was applied on p values adjusted according to the sum of all isomiR reads from the canonical mature Benjamini and Hochberg method. miRNA locus). Raw miRNA reads were normalized DESeq2 applies a method for outlier detection . using the geometric mean-based method implemented These miRNAs have reads in at least one sample but are in DESeq2. Spectral maps were generated using mpm R ignored for the statistical test (hence no p value). Fur- package  based on DESeq2 normalized counts . thermore, DESeq2 applies a filtering on low counts; In this study, the contrasts analyzed were the following: these miRNAs have reads in at least one sample but naive mdx vs naive wt (T16) and Givinostat 37.5 mg/kg have a low average abundance. The statistical test is ap- vs vehicle (T16). plied for these miRNAs—they receive a p value, but they Licandro et al. Skeletal Muscle (2021) 11:19 Page 7 of 22 are excluded from multiple test correction—they receive Treadmill exhaustion test no adjusted p value. Givinostat treatment at the dose of 37.5 mg/kg signifi- cantly enhanced the performance of mice on day 31 and, between days 31 and 44, their performance (785 ± 97 Results and 693 ± 103 m, respectively) was superimposable on Givinostat improves muscle function in a dose-dependent that of wt mice (735 ± 66 and 728 ± 76 m on days 31 way in mdx mice and 44, respectively). On day 94, mice were able to run In order to start dissecting the influence of Ltbp4 poly- for 410 ± 40 m (the distance covered by wt mice was morphism on the efficacy of Givinostat in DMD mouse 657 ± 44 m), whereas the mdx vehicle mice ran for only models, we first set out to characterize the effect of our 92 ± 4 m (Fig. 2A). Lower doses did not reach signifi- HDAC inhibitor in mdx mice. Muscle function was in- cance, although a trend of a dose-response effect was vestigated using both the grip test and the treadmill ex- evident throughout the entire experimental period (Fig. haustion test. 2A). In addition to the distance, we also analyzed the time to exhaustion at the last time point (day 94). Givi- Grip test nostat, at the highest dose, significantly prolonged the Untreated mdx mice showed a small increase in max- time to exhaustion (Fig. 2B). The effect was also signifi- imal normalized strength up to day 21 from the baseline cant at lower doses, with 0.3 mg/kg being the lowest ef- value (day 0), followed by a progressive decline (Fig. 1). fective dose. Mdx vehicle-treated mice behaved as the mdx naive mice, clearly indicating that the vehicle did not exert any Pharmacokinetics of Givinostat in C57BL/10 mdx mice effect. In this setting, Givinostat treatment induced a re- The functional data reported thus far, point to a dose- markable, dose-dependent increase of the FNmax with dependent improvement of functional tests by Givino- the maximal efficacy observed at the dose of 37.5 mg/kg. stat. We next investigated the correlation with systemic On treatment day 49, this dose transiently improved and tissue-specific drug exposure by measuring blood, FNmax to a level that was comparable to that of wt plasma, and muscle Givinostat levels after single oral ad- healthy mice (Fig. 1). From this day on, the animals ministration of different doses of the drug. Mean con- started to diminish the developed strength but the bene- centration levels of Givinostat in plasma and quadriceps ficial effect of Givinostat remained significant up to day are shown in Fig. 3 (Fig. 3A,B, respectively), and mean 105 (see Additional Table 2). PK parameters in plasma, blood, and quadriceps are Fig. 1 Givinostat increases maximal normalized strength (FNmax) in mdx mice in a dose-response manner. Givinostat treatment started on day 7. Data are expressed as the mean ± standard error (n = 9; wt: wild type) Licandro et al. Skeletal Muscle (2021) 11:19 Page 8 of 22 Fig. 2 Givinostat increases distance run by mice (A) and time to exhaustion at day 94 (B)in mdx mice. Givinostat (Giv) treatment started on day 7. Data are expressed as the mean ± standard error (2-way ANOVA with Bonferroni’s multiple comparison test (A) and one-way ANOVA with Bonferroni’s multiple comparison test (B), *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001 vs Vehicle; n = 9; wt: wild type) reported in Table 1 (Table 1). Givinostat was detectable without a delay time for tissue distribution. The AUCs in all plasma, blood, and muscle samples up to 6 h after of Givinostat in the muscle were always higher than in the oral treatment at all doses investigated. plasma, indicating a preferential distribution of Givino- Blood quantification was mainly performed in order to stat into muscle tissue. subtract from muscle homogenate the amount of Givi- The results of the PK study show that, upon oral nostat in the residual blood contained in the muscle, ac- administration, Givinostat exposures increase in a cording to the physiological parameters reported in dose-dependent way both in terms of Cmax and AUC literature . The level of Givinostat in blood was al- and that the compound quickly distributes into mus- ways higher than in plasma indicating a distribution of cular tissue with an AUC about 4.5 times higher than the drug into blood cells. The AUC ratio was calculated in plasma. The efficacy of Givinostat may well be re- and a mean value of 1.6 was found. Similar PK profiles lated to these high exposure levels in the muscular both in plasma and in quadriceps were also found tissue. Licandro et al. Skeletal Muscle (2021) 11:19 Page 9 of 22 Fig. 3 Plasma and quadriceps Givinostat concentrations in mdx mice. Linear plot of plasma (A) and quadriceps (B) levels of different doses of Givinostat orally administrated to mdx mice (n = 4) Histopathological analysis fibrosis compared to that of naive wt mice. Givinostat Histopathology was performed to investigate whether reduced fibrosis at all the 3 doses analyzed compared to the observed functional improvements in Givinostat- that of vehicle-treated mice, but the effect was not sig- treated mdx mice correlate with better preservation of nificant (Fig. 4A). Figure. 4B shows an example of the muscle integrity at the histological level. The effect of reduction of fibrosis (visible through the reduction of Givinostat on fibrosis was analyzed. At the end of treat- red as the dose of Givinostat increases) in histological ment (T16), GAS, TA, DIA, and heart in naive mdx mice sections of GAS stained by Sirius Red staining. had a significantly higher fibrosis area compared to The effect of Givinostat treatment on myofiber cross- healthy muscles. Givinostat administered at 10 and 37.5 sectional areas was also evaluated. Determination of mg/kg significantly reduced fibrosis (by about 30%) in fiber CSA of GAS in mdx mice revealed an increase in GAS, whereas an effect on fibrosis in TA was observed fiber size after 105 days of Givinostat treatment at 37.5 only at the dose of 37.5 mg/kg (Fig. 4A). In the DIA, all mg/kg, when compared to vehicle-treated mice. GAS the 3 analyzed doses of Givinostat significantly reduced CSA of Givinostat-treated mice have doubled GAS CSA the amount of fibrosis by about 40% (Fig. 4A). In the values (peak at 1100 μm ) as compared to those of mdx heart, the amount of fibrosis seems to be more variable vehicle GAS (peak at 500 μm ) (Fig. 5A). According to than in other muscles. However, vehicle-treated mdx the statistical analysis, this difference could be explained mice seem to have a significantly higher amount of by assuming a multiplicative model effect with a 1.72 K, Licandro et al. Skeletal Muscle (2021) 11:19 Page 10 of 22 Table 1 Quantification of pharmacokinetic parameters ^Mean PK parameters for Givinostat in plasma, blood and muscle after oral administration of different doses to mdx mice which provides the multiplicative factor to be applied to normal pattern of myofiber regeneration, even though each fiber CSA value of the vehicle group to obtain the this parameter is difficult to interpret because central- expected CSA distribution after Givinostat at 37.5 mg/kg ized nuclei may persist even after muscle repair. Nuclear treatment. Interestingly, this K value is well in line with centralization was absent in GAS, TA, and DIA muscles that observed in a previously reported clinical study . from 24 to 25-week-old wt mice, while it was very dif- In TA, fiber CSA showed a limited increase after Givi- fuse in GAS (77 %), TA (66 %), and DIA (58 %) of nostat at 37.5 mg/kg treatment compared to vehicle- vehicle-treated mdx mice at the same age. No marked treated mice, with a shift toward wt CSA values treatment-related effects were observed: Givinostat at 1, (1400 μm ). TA CSA of Givinostat at 10 and 37.5 mg/kg 10, and 37.5 mg/kg treated mice showed a similar nu- treated mice have a peak value of 900 μm similar to that clear centralization to that of vehicle-treated mice (67.6 of mdx vehicle TA (700 μm ) (Fig. 5B). Statistical ana- ± 2.4%, 58 ± 3%, and 61.3 ± 3.8%, respectively, in GAS; lysis of the TA CSA distribution confirmed the results 71 ± 1.6%, 73.2 ± 3.7%, and 60.8 ± 2.5%, respectively, in observed using the additive model effect, with the shift TA; 57 ± 2.9%, 61.4 ± 3.5%, and 41.4 ± 4.5%, respect- effect parameter S = 140 μm . ively, in DIA). In DIA, no major change in CSA distribution was evi- As reported in the literature, the high percentage in dent in mice treated with Givinostat 1, 10, and 37.5 mg/ centralized nuclei in mdx mice is due to their character- kg compared to vehicle-treated mice (CSA peak value in istic mechanism of continuous regeneration after muscle vehicle-treated mice was 550 μm , the same observed in degeneration  and could also explained a lack of Givinostat 37.5 mg/kg treated mice). marked effect in treated mice. The percentage of centralized nuclei on muscle tissue The histological evaluation of muscle inflammatory in- was also analyzed. Centralized nuclei are considered a filtrate, adipose tissue deposition, regeneration, and Licandro et al. Skeletal Muscle (2021) 11:19 Page 11 of 22 Fig. 4 Effect of Givinostat treatment on muscle fibrosis in mdx mice. Givinostat significantly reduces fibrosis in gastrocnemius (GAS), tibialis anterior (TA), and diaphragm (DIA) muscles of mdx mice. In the heart, the effect of Givinostat was not significant. Analysis was performed at T16 time point. Data are expressed as the mean ± standard error (1-way ANOVA with Bonferroni’s multiple comparison test, *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001 vs vehicle; n = 5; wt: wild type) (A). Representative images of Sirius Red staining of GAS muscle transverse sections of naive wild type (wt), naive mdx, vehicle, Givinostat 1, 10, and 37.5 mg/kg treated mice (B) degeneration of the muscular tissue was not significantly enrichment of specific miRNAs in FAP-derived exo- different among the treatment groups. In particular, the somes and increased levels of exosomal miR-206 were degree of inflammatory infiltrate and adipose tissue de- required to promote MuSC activation and expansion position was very limited in all samples and, therefore, ex vivo and to stimulate regeneration of dystrophic mus- not useful for discriminating the efficacy of the different cles in vivo. These data point to the importance of inves- doses of Givinostat (Additional Table 3). tigating miRNA levels in response to Givinostat Functional and histological improvements as a func- treatment in vivo. To investigate whether the observed tion of administered doses are summarized in Additional effects of Givinostat on muscle function and histology Table 4 (see Additional Table 4). correlate with changes in plasma miRNA levels, we used the RNAseq approach to evaluate the whole-genome Whole-genome miRNA expression profiling miRNA expression profile in plasma of mdx mice MicroRNAs (miRNAs) are small, approximately 22 nt, treated with Givinostat at the dose of 37.5 mg/kg. In this non-coding RNA molecules that post-translationally study we evaluated the differential expression of miRNA regulate gene expression. Several miRNAs were de- between selected groups of samples collected at T16. scribed to be involved in skeletal muscle proliferation, The number of replicates per group were compared in a differentiation, and regeneration . In addition, miR- pairwise manner. The compared conditions (i.e., con- NAs were proposed to be useful serum biomarkers for trasts) and the number of up- and downregulated miR- muscular dystrophy . Previous publications have im- NAs are listed in Additional Table 5 (see Additional plicated miRNAs as mediators of exosome-regulated Table 5). biological processes [44–46]. Exosomes, or extracellular We focused on the most highly modulated miRNAs vesicles, are gaining a lot of attention in the scientific with known association with dystrophic disease at T16 community due to their role in cell-cell communication. for the naive mdx vs naive wt contrast. The analysis of Fibro-adipogenic progenitor (FAP)-derived exosomes ac- these two groups is instructive to comprehend if the epi- cumulate in the interstitial space of regenerating muscles genetic treatment ameliorates the mdx phenotype by re- and can mediate miRNA transfer to MuSCs. In vivo ex- storing the altered expression of certain miRNAs. To posure to HDAC inhibitors was reported to lead to an evaluate the effect of Givinostat on the modulation of Licandro et al. Skeletal Muscle (2021) 11:19 Page 12 of 22 Fig. 5 Graph of the analysis of myofiber cross-sectional area (CSA) of muscles in mdx mice. Frequency graphs showing fiber CSA distributions in gastrocnemius (GAS) (A) and tibialis anterior (TA) (B) (5 mice/group, except for Naive wild type (wt) and Naive mdx groups, 4 mice/group). Each distribution was calculated as percentage of the total number of cases to account for the different number of cases within the groups. In Fig. 5A are also reported three examples of hematoxylin and eosin-stained transverse sections of myofibers areas in GAS of Naive WT, vehicle, and Givinostat 37.5 mg/kg treated mice miRNAs, we chose the last treatment (T16) and the myofibers [52, 53]. The abundance of these three miR- most efficacious dose (37.5 mg/kg) vs vehicle contrast. NAs correlates with the progression of dystrophic path- In the T16 naive mdx vs naive wt contrast, 120 miR- ology . NAs were significantly upregulated and 66 miRNAs were The most significantly downregulated miRNAs in T16 significantly downregulated in plasma of mdx mice. naive mdx vs naive wt contrast were miR-122-3p and Among the upregulated miRNAs we found miR-133a-5p miR-122-5p that are probably involved in the regulation (FC = 5.973; the most upregulated miRNA), miR-133b- of muscle growth and lipid deposition , but, to our 3p (FC = 5.441), miR-133a-3p (FC = 4.693), miR-378b knowledge, these miRNAs were so far not described to (FC = 4.109), miR-206-3p (FC = 4.065), miR-1a-3p (FC be dystrophy-related. = 3.423), miR-22-5p (FC = 2.991), and miR-22-3p (FC = Among the Givinostat-induced miRNAs (T16 Givi- 2.550) (Fig. 6). All of these miRNAs have been reported nostat 37.5 mg/kg vs vehicle contrast), the only two to be elevated in the dystrophic mice [47, 48]. Moreover, that were significantly upregulated were miR-449a-5p miR-133, miR-1 and miR-206 levels were found to be el- (FC = 4.388) and miR-92b-3p. The upregulation of evated also in the plasma of DMD patients . these two miRNAs in mdx muscle after HDACi treat- In mice, miR-133a-3p, miR-1a-3p, and miR-206-3p are ment (trichostatin A—TSA) was already reported in specifically muscle-enriched miRNAs that regulate both the literature [56, 57]. the proliferation of myofibers and the myogenic differen- miR-449a-5p is involved in different pathways, such as tiation during muscle growth [50, 51]. In particular, TGF-β, TNF, MAPK, Wnt, FoxO, PI3K-Akt, and Hippo miR-206-3p plays a key role in dystrophic muscle regen- signalling pathway and was proposed to be linked to eration and is known to be enriched in regenerating stem cell pluripotency . Licandro et al. Skeletal Muscle (2021) 11:19 Page 13 of 22 Fig. 6 Volcano plots of differentially expressed miRNA in wt and mdx mice. The analysis was performed for Naive mdx vs Naive wild type (wt) (A) and Givinostat 37.5 mg/kg vs vehicle (B) contrasts after 15 weeks of treatment (T16). Log2 fold-changes estimated by DESeq2 versus –log10 adjusted p values (FDR: false discovery rate) are indicated in the plot. Black dots indicate miRNA with non-significant fold change (FDR > 0.05). Pink dots indicate significantly differentially expressed miRNAs at FDR < 0.05 miR-92b-3p, along with miR-133a-3p, is one of the Treatment efficacy evaluation in D2.B10 mouse model cardiac specific miRNAs involved in heart failure (i.e., Having characterized the functional, histological, and cardiomyocyte apoptosis, hypertrophy, and inflamma- molecular effects of Givinostat on mdx mice with a mild tion) and was reduced in the coronary sinus of pa- dystrophic phenotype, we next evaluated the drug in the tients with heart failure . In particular, miR-92b- more severe D2.B10 model. We also included GC steroid 3p was observed to be decreased in mouse hyper- treatment groups in this experiment to evaluate the effi- trophic cardiomyocytes upon angiotensin II antagonist cacy of Givinostat in comparison to the clinically rele- treatment and its overexpression can attenuate the vant standard of care treatment. Both prednisone and hypertrophic phenotype by downregulating the hyper- Deflazacort were used for this purpose. trophy related genes [58–60]. Mdx mice show a mild Unfortunately, we observed that repeated, chronic cardiomyopathy and, according to the literature, we oral administration of Givinostat by gavage turned found that at T16, miR-92b-3p in naive mdx plasma out to be difficult and unsafe for these mice. This dif- was downregulated compared to naive wt mice (FC = ficulty may be related to a more pronounced oropha- − 0.896, not significant). After chronic Givinostat ryngeal or esophageal dysfunction in D2.B10 mice as treatment (T16 Givinostat 37.5 mg/kg vs vehicle), our compared to mdx mice. We noticed that feeding diffi- analysis showed a significant upregulation (FC = culties and oropharyngeal or esophageal dysfunctions 2.130) of miR-92b-3p, suggesting a possible beneficial were also reported to occur in Duchenne patients effect on cardiac function of mdx mice upon treat- . As a consequence, the drug was given in drink- ment with Givinostat. Since we focused exclusively on ing water (see “Drug treatments” paragraph in “Mate- plasma miRNA levels, it will be interesting to investi- rials and methods” section). To estimate the daily gate the effects of Givinostat on miRNA levels in dose administered to the animals using this proced- skeletal or cardiac muscle in future studies. ure, water consumption was measured every week The results of each contrast are displayed as volcano during the study. We calculated the actual doses of plots including all miRNAs in Fig. 6. Our results are in Givinostat ingested by mice, to be in the following agreement with literature data that indicate a profound ranges: 0.8–1.1 mg/kg, 3.6–6mg/kg, 7.2–11.6 mg/kg, dysregulation of miRNA levels in skeletal muscle from and 27.5–41.5 mg/kg. We noticed that these data were mdx mice . Givinostat treatment does not revert the in line with the expected doses of 1, 5, 10, and 37.5 overall pattern toward normalcy but specifically acts on mg/kg, respectively. An increase in the water intake two miRNAs that have been reported to be functionally by the animals, and therefore in the dose of Givino- linked to muscle physiology and more specifically to car- stat is probably due to the growth, and therefore, the diac pathologies. weight gain of the animals over time. Licandro et al. Skeletal Muscle (2021) 11:19 Page 14 of 22 Functional activity measurement significantly different from prednisone. Significant differ- In an analogy to the characterization done on the mdx ences in FNmax values between Givinostat at 37.5 mg/ mice, muscle function of D2.B10 mice was investigated kg versus prednisone were observed only at days 90 and using both the grip test and the treadmill exhaustion 97 (p < 0.05) and, at the same experimental time points, test. a significant difference was also seen between Givinostat at 37.5 mg/kg vs Deflazacort (p < 0.05 and p < 0.01, Grip test respectively). The developed FNmax in DBA/2J wt mice increased throughout the entire experimental period (Fig. 7), while Treadmill exhaustion test FNmax in D2.B10 vehicle-treated mice tended to de- During the exhaustion test evaluation, we had to set a crease starting from day 14 (Fig. 7), indicating a progres- maximum number of shocks that mice received until the sive impairment of muscle function. The effect of end of each test with respect to the protocol applied to Givinostat on FNmax was dose-dependent, with max- mdx mice in order to avoid unnecessary stress to the an- imal efficacy observed at the dose of 37.5 mg/kg. Givino- imals (see “Assessment of functional tests for the evalu- stat at 10 and 37.5 mg/kg as well as prednisone ation of treatment efficacy” paragraph in “Materials and treatment increased the FNmax of D2.B10-treated mice methods” section). This adjustment led to an apparent to that of wt healthy mice starting from day 21 until worsening of the performance of healthy DBA/2J which days 49–62 (Fig. 7). From day 62 onwards, D2.B10- we therefore ascribe to a change in experimental settings treated mice started to diminish the FNmax, but the ef- and not to an increased exhaustion. Despite the reduced fect of Givinostat (administered at 5, 10 and 37.5 mg/kg) number of shocks, Givinostat at its top dose improved and prednisone remained significantly elevated until day running performance of D2.B10 mice with a clear in- 104 (see Additional Table 6). The lowest dose that crease at day 63 (Fig. 8A). This effect lasted until the exerted a significant effect from day 21 to day 104 on end of treatment period where the improvement was FNmax was 10 mg/kg. Deflazacort and Givinostat at 1 statistically significant. In fact, on day 105, Givinostat- mg/kg significantly improved the FNmax for a limited treated mice were able to run for 473 ± 22 meters, time interval, i.e., from day 28 to day 90 and from day whereas the D2.B10 vehicle-treated mice ran for only 42 to days 90/104, respectively (see Additional Table 6). 296 ± 17 m (the distance covered by wt mice was 648 ± The effect of prednisone on FNmax was comparable to 31 m). On the contrary, the effect of Deflazacort, pred- that of Givinostat 5 e 10 mg/kg for all the treatment nisone, and of the lower doses of Givinostat was not sta- period, while the effect of Deflazacort was lower and not tistically significant (Fig. 8A). Fig. 7 Effect of Givinostat and steroids on maximal normalized strength (FNmax) in D2.B10 mice. Givinostat and steroid treatment started on days 7 and 9, respectively (wt: wild type) Licandro et al. Skeletal Muscle (2021) 11:19 Page 15 of 22 Fig. 8 Effect of Givinostat and steroids on both the distance run (A) and time to exhaustion at day 105 (B) in D2.B10 mice. From day 44, a maximum number of shock (150) was set. Givinostat and steroid treatment started on days 7 and 9, respectively (2-way ANOVA with Bonferroni’s multiple comparison test, *p < 0.05; ****p < 0.0001 vs vehicle; wt: wild type) In addition to the distance, we also analyzed the time vehicle-treated mice (by about 26%) (Fig. 9). No signifi- to exhaustion at the last time point (day 105). Givinostat, cant effect was observed at time point T16 upon either at the highest dose, significantly prolonged the time to Givinostat or steroid treatment. In TA (T8), both ste- exhaustion (Fig. 8B). roids and Givinostat (administered at 37.5 mg/kg) sig- nificantly decreased fibrosis compared to vehicle-treated Histopathological analysis mice (by about 22%) (Fig. 9). No significant effects were In DBA/2J wt mice, GAS, TA, and DIA had a low per- observed on fibrosis at time point T16. At T8, DIA of centage of fibrotic area when compared to dystrophic animals treated with Givinostat, as well as with prednis- muscles at both time points T8 (3.0 ± 0.5 %, 3.0 ± 0.4 % one, showed a significant reduction of the amount of fi- and 3.9 ± 0.3 %, respectively) and T16 (1.8 ± 0.5 %, 2.3 ± brosis by about 33 and 45%, respectively. Deflazacort 0.3 % and 4.7 ± 0.8 %, respectively). In GAS (T8), pred- had no significant effect (Fig. 9). At T16, the efficacy of nisone significantly reduced fibrosis compared to both Givinostat at 37.5 mg/kg and prednisone was lost, Licandro et al. Skeletal Muscle (2021) 11:19 Page 16 of 22 Fig. 9 Givinostat and steroid treatment ameliorate muscle fibrosis in D2.B10 mice. Effect of Givinostat and steroids on the percentage of fibrosis in gastrocnemius (GAS), tibialis anterior (TA), and diaphragm (DIA) of D2.B10 mice after 8 weeks of treatment (T8) (1-way ANOVA with Bonferroni’s multiple comparison test, *p < 0.05; **p < 0.01; ****p < 0.0001 vs vehicle; n = 5; wt: wild type) possibly due to the exhaustion of regenerative potential doses are summarized in Additional Table 8 (see Add- with age . itional Table 8). The CSA of GAS and TA of DBA/2J wt mice were higher compared to that of dystrophic mice (1500 μm Discussion 2 2 at T8 and 900 μm at T16 in GAS and 1200 μm at T8 DMD patients homozygous for the IAAM Ltbp4 haplo- and 1050 μm at T16 in TA). Moreover, at both T8 and type remained ambulatory significantly longer than those T16, no major changes in CSA distribution were ob- homozygous for the VTTT haplotype . This import- served in GAS and TA of mice treated with Givinostat ant correlation was confirmed in other studies [22, 23] 2 2 administered at 37.5 mg/kg (600 μm at T8 and 450 μm where a delay in the loss of ambulation in patients with 2 2 at T16 in GAS and 300 μm at T8 and 300 μm at T16 the IAAM phenotype treated with corticosteroids was in TA) and prednisone at 1 mg/kg (600 μm at T8 and also observed. 2 2 2 300 μm at T16 in GAS and 300 μm at T8 and 300 μm In our study, we observed that Givinostat was effective at T16 in TA) compared to D2.B10 vehicle-treated mice in both mdx and D2.B10 murine models which function- 2 2 (1050 μm at T8 and 300 μm at T16 in GAS and ally represent the mild and more severe phenotypes of 2 2 600 μm at T8 and 300 μm at T16 in TA). DMD, respectively. Nuclear centralization was absent or very low in GAS It is further interesting to compare the effects of Givi- and TA muscles from wt mice of 24–25 weeks old nostat on the two mouse strains used in this project. (around 1%), while it was very diffuse in GAS (25.1 ± 1.5 D2.B10 mice could not be chronically dosed by daily %) and TA (21.2 ± 1.0%) of D2.B10 vehicle-treated mice gavage and received the drug in their drinking water. of the same age. No marked treatment-related effects Even though water consumption measurements have were observed: Givinostat at 37.5 mg/kg and prednisone- shown that the total doses that D2.B10 mice received treated mice sowed a similar nuclear centralization to were comparable to those of the mdx mice, an accurate that of the vehicle group (22.4 ± 2.1 % and 21.6 ± 1.7 %, PK monitoring upon administration in drinking water is respectively, in GAS and 19.8 ± 2.4 % and 22.0 ± 2.4 %, essentially impossible. It is likely however, that the respectively in TA). plasma concentration curve of Givinostat will be much The histological evaluation of muscle inflammatory flatter with this mode of administration and that lower infiltrate, adipose tissue deposition, regeneration, and circulating concentrations will be reached, even though degeneration of the muscular tissue was not signifi- similar AUC values are expected. We notice that in this cantly different among the treatment groups. In par- model, even though strong functional benefit could be ticular, the degree of inflammatory infiltrate and shown, only a transient effect on fibrosis and no signifi- adipose tissue deposition was very limited in all sam- cant effect on CSAs was observed. It is possible that this ples and, therefore, not useful for discriminating the lack of effect could relate to the severeness of the disease efficacy of the different doses of Givinostat or steroids developed by these animals that may not be efficiently (Additional Table 7). countered by an HDACi. Like for the pharmacologic ef- The statistical analyses of both functional and histo- fect in mdx mice, an alternative explanation could be logical improvements as a function of administered that a robust effect on fibrosis and CSA requires higher Licandro et al. Skeletal Muscle (2021) 11:19 Page 17 of 22 circulating concentrations (Cmax-driven effect) that are monitoring forelimb force with this assay, were robust not attained when the compound is administered with and reproducible and allowed us to establish clean dose- drinking water. In line with this reasoning, Givinostat response relationships. Using this readout, Givinostat was inactive in an in vivo pulmonary fibrosis model showed a remarkable, dose-dependent improvement in when administered with the diet (McKinsey TA, unpub- FNmax with top doses transiently approaching values lished results). This mode of administration is likely to seen in wt animals. In mdx mice, the treadmill distance give similar low-level exposures as our administration and time to exhaustion also transiently approached wt with drinking water in D2.B10 mice. values, whereas the analogous data obtained with Although doubts remain about their ability to prolong D2.B10 mice are more difficult to interpret due to proto- ambulation in DMD patients, the pros and cons of their col changes that were necessary during the experiment long-term use and the choice of which corticosteroid/ to safeguard animal well-being. Still, also in this case, regimen to use, GC are routinely used for DMD treat- the top dose showed a statistically significant improve- ment. We evaluated both prednisone and Deflazacort in ment over vehicle-treated animals in both distance and the more severe DMD mouse model and we found that time to exhaustion evaluations. We conclude that Givi- there were small differences between the two GC ste- nostat improves muscle function, irrespective of the roids in terms of their ability to improve muscle per- Ltbp4 status. Functional benefits were transient in both formance. Prednisone and Deflazacort were unable to mouse strains but even in the decline phase a statistically affect cross-sectional area or the number of centralized significant difference with respect to vehicle-treated ani- nuclei in D2.B10 muscles. In addition, prednisone sig- mals was observed. nificantly reduced fibrosis in TA, GAS, and DIA, while The mechanism of action of HDACis in muscle re- Deflazacort was effective only in TA. generation is complex, multifaceted, and not com- Recently, Givinostat was shown to be highly efficacious pletely understood. Fibro-adipogenic progenitor cells in two distinct murine models of diastolic dysfunction were proposed to be key actors in these processes with preserved ejection fraction . In these experi- [56, 57, 62]. In young animals, these cells have been ments, the drug was admixed to the diet. In these car- shown to be reprogrammed by HDACis to deliver diac models, Givinostat blocked left ventricular diastolic pro-regenerative factors and to mobilize muscle- dysfunction due to hypertension and suppressed aging- resident stem cells to engage in tissue repair . induced diastolic dysfunction in normotensive mice. No This effect was no longer observable in aged animals, effect on fibrosis was observed by the authors, but an suggesting that exhaustion of regenerative capacity impairment of cardiac myofibril relaxation, as a previ- may occur over time . Whether this is the basis ously unrecognized, myocyte-autonomous mechanism for the transient functional improvements observed in for diastolic dysfunction was discovered that was pro- our experiments needs to be further explored. Also, posed to be regulated by myofibrillar protein acetylation. the relevance of this exhaustion process for the treat- We can therefore assert that Givinostat is effective in ment of human disease with HDACis needs to be counteracting fibrosis in the DMD mouse models and established. The outcome of the ongoing clinical trial not in pulmonary fibrosis and diastolic dysfunction mur- with Givinostat in adult Becker dystrophy patients ine models. (ClinicalTrials.gov identifier: NCT03238235) will be A specific role for HDAC2 in the control of myofibrils revealing in this respect. was suggested that apparently was affected already by Givinostat, at the top dose, induced significant in- very low circulating HDACi concentrations. It would be creases of the CSAs of TA and GAS muscles in mdx interesting to investigate if myofiber protein acetylation mice, while no effect on nuclear centralization was seen. also occurs in skeletal muscles and to what extent it af- Whether the increased CSA is a result of regeneration, fects the function of dystrophic muscles. of a countering of atrophic processes or due to protec- Givinostat has previously been shown to be effective in tion of muscle fibers from damage is presently unknown. the mdx mouse model and is presently being studied in We note that Givinostat potently inhibits muscle atro- phase III clinical trials in DMD patients. phy pathways in vitro (manuscript in preparation) and In this preclinical work, we wanted to further dissect has a known anti-inflammatory activity that could miti- the pharmacological activity of the molecule with a spe- gate ROS-mediated tissue damage in an inflammatory cial focus on functional readouts, dose dependence, and environment. Both effects could contribute to preserve the influence of Ltbp4 polymorphism. The treadmill larger muscle fibers. We also observed a dose-dependent assay is known to be associated with large experimental effect on fibrosis in TA, GAS, and DIA muscles from variability. Therefore, the grip test was used for the first mdx mice. The antifibrotic effect was evident at the top time to monitor the functional benefit of Givinostat doses and this contrasts with the dose dependence of treatment. In our hands, the data obtained by functional improvements that were highly significant Licandro et al. Skeletal Muscle (2021) 11:19 Page 18 of 22 also at lower doses. These findings will be further dis- different administration route of the two compounds: cussed below. Vorinostat has very poor oral bioavailability in mice and Multiple dose PK experiments were done in mdx mice Colussi and colleagues administered this compound i.p., to establish efficacy/PK correlations. To put these data  whereas in our experiments Givinostat was given into context, we need to consider that Givinostat is a p.o. I.p. dosing is expected to give much higher peak relatively non-specific inhibitor of all 11 zinc-dependent plasma concentrations (Cmax value in CD1 mice was HDAC subtypes. Minetti and colleagues have shown that 123 ng/ml at 0.25 h after i.p. administration of 5 mg/kg Entinostat (MS-275), an HDAC inhibitor that selectively dose) than p.o. administration. Remarkably, also TSA, inhibits HDACs 1, 2, and 3 is as active as the pan- when given i.p., is active in the mdx model, even though HDACi TSA , while Colussi and colleagues have this molecule is extremely short-lived in vivo and was demonstrated that ablation of HDAC2 expression reca- reported to have a half-life of only 6.3 min upon i.p. dos- pitulates many features of small molecule inhibitors, ing in mice . However, i.p. dosing of TSA lead to mi- confirming a key role of the inhibition of this enzyme in cromolar peak plasma concentrations that are well mediating the pharmacologic activity of HDACis . above the enzyme inhibition IC50 values. Taking these Mechanistically, the impaired nNOS activity in DMD data together, at least some of the pharmacologic effects muscles was proposed to lead to a decrease in HDAC2 of HDACis in this model seem to be driven by transient nitrosylation, causing its retention on miRNA promoters peak plasma concentrations (Cmax values in mdx mice and leading to an increase in the expression of fibrosis/ treated with Givinostat: 24 ng/ml for the 5 mg/kg p.o. oxidative stress-related genes . miR-206 was sug- dose; 807 ng/ml for the 37.5 mg/kg p.o. dose). gested to be under the control of HDAC2 and expressed The question then becomes: how transient target in- in activated satellite cells where it represses the Pax7 fac- hibition may still lead to long-lived pharmacodynamic tor, thus allowing differentiation to proceed . Finally, effects? In this respect, an intriguing observation was re- HDAC3 was shown to be an obligate mediator of the in- cently published , showing that the pharmacologic flammatory gene expression program of macrophages effects of HDACi in mdx mice are, to a large extent, me- . In the light of these findings, it is likely that inhib- diated by extracellular vesicles that are produced by ition of HDACs 1, 2, and 3 is a major driver of the activ- mesenchymal cells such as fibro-adipogenic progenitors ity of Givinostat in DMD. (FAPs). These extracellular vesicles are involved in From our PK experiments, we conclude that Givino- microRNA transfer to muscle stem cells and exposure of stat rapidly accumulates in muscle tissue, where it dystrophic FAPs to HDACis increases the intra- reaches a concentration 4.5 times higher than in plasma extracellular vesicles levels of a subset of miRNAs that and that HDAC inhibition is probably going to be transi- regulate biological processes such as regeneration, fibro- ent, since concentration curves rapidly declined after a sis, and inflammation. Significantly, extracellular vesicles few hours. Therefore, the improvement of functional derived from HDACi-exposed FAPs were able to pheno- and histological parameters obtained in mdx mice can copy the pharmacologic effects of HDACis upon a single be also explained by the high distribution levels of Givi- injection in dystrophic mice, indicating long-lasting nostat into the muscular tissue. pharmacodynamic effects. Since the targeted HDACs 1, 2, and 3 are essential en- It is well known that the activity of HDACi on differ- zymes in most cells, their continuous, sustained inhib- ent biological pathways can greatly vary as a function of ition would probably not be tolerated. In fact, attempts their concentrations. As an example, we discovered that to deliver HDACis by continuous infusion, leading to the anti-inflammatory activity of HDACi is exerted at their sustained inhibition, were poorly tolerated in our much lower concentrations compared to those needed hands (C.S., unpublished observations). for an antitumor activity [68, 69]. Thus, a possible model It is interesting to compare our dose-response curves that reconciles all of these findings could be the follow- to those reported by Colussi and colleagues using Vori- ing: HDACis may trigger myofiber-autonomous effects nostat in the mdx mouse model . In that report, a leading to functional improvements at relatively low bell-shaped dose-response curve of Vorinostat with a concentrations that can be reached by low-dose adminis- peak for both functional and histologic effects at the low tration through oral gavage or upon admixture of the dose of 1.2 mg/kg was shown. We notice that this behav- drug to the diet or to drinking water. However, transient ior differs from our observations with Givinostat, for exposure to higher concentrations of HDACis, well which both functional and histologic parameters im- above the Ki values of the main targets HDACs 1, 2, and proved steadily with increasing dose up to 37.5 mg/kg, 3, may be needed to trigger pro-regenerative and antifi- the highest dose tested. Vorinostat and Givinostat have brotic effects. These drug levels can be reached by high- similar enzyme inhibition profiles . A possible ex- dose p.o. administration or by lower dose i.p. dosing. planation for this discrepancy could be linked to the Transient HDAC inhibition may result in durable Licandro et al. Skeletal Muscle (2021) 11:19 Page 19 of 22 pharmacodynamic outcomes that are mediated by extra- Abbreviations BW: Body weight; CSA: Cross-sectional area; DMD: Duchenne muscular cellular vesicles. This picture needs further confirmation dystrophy; DIA: Diaphragm; FNmax: Maximal normalized strength; from additional preclinical experiments and quantitative GAS: Gastrocnemius; GC: Glucorticoids; HDAC: Histone deacetylase; PK/PD correlations need to be established using relevant HDACi: Histone deacetylase inhibitor; LTBP4 gene: Latent transforming growth factor-β binding protein 4; miRNAs: MicroRNAs; ns: Not significant; biomarkers. If confirmed, these data could serve as the PK: Pharmacokinetic; TA: Tibialis anterior; TGF-β: Transforming growth factor- basis for optimizing the dosing scheme of HDACis in β; wt: Wild type the clinic. While the lack of dystrophin that leads to chronic Supplementary Information muscle damage is the causal factor of Duchenne dys- The online version contains supplementary material available at https://doi. trophy, the pathologic repair process that is triggered org/10.1186/s13395-021-00273-6. under these conditions and that leads to a progressive substitution of muscle tissue by fibrotic and adipose Additional file 1: Table 1. Muscle sampling in both mdx and D2.B10 tissue is a key pathogenic process. Restoring dys- mice. Summary of muscle sampling at the two different time points (T8 = after 8 weeks of treatment; T16 = after 15 weeks of treatment) in both trophin expression in all muscle fibers would in the- mdx and D2.B10 studies (n = 5; GAS = gastrocnemius; TA = tibialis orycurethediseaseand attempts to at least partially anterior; DIA = diaphragm). do so using exon skipping oligos or a drug inducing Additional file 2: Table 2. Statistical analysis of maximal normalized point-mutation read-through have conducted to first strength in mdx mice. Summary of statistical analysis of maximal normalized strength in wt and mdx mice (wt: wild type). 2-way ANOVA approvals of targeted therapies in Duchenne patients with Bonferroni’s multiple comparison test was performed (*p < 0.05; **p . Promising clinical data have been recently pub- < 0.01; ***p < 0.001; ****p < 0.0001 vs Vehicle). lished on AAV-based gene therapy, suggesting that Additional file 3: Table 3. Histopathological evaluation of the severity further, urgently needed therapy improvements could of myodystrophy in different muscles of mdx mice at T16. The histopathological method considered some parameters scored by be in sight for those patients that cannot benefit from severity and extension of the injury: muscle degeneration/necrosis, presently available therapeutic options . Still, all regeneration, inflammatory infiltrate, interstitial reaction and adipose of these therapeutic approaches have their limitations tissue deposition. Each parameter was classified by severity (mild = 1, moderate = 2 and severe = 3) and extension (focal = 1, multifocal = 2 and even if the present gene therapy approaches were and diffuse = 3). The individual severity score was calculated for each 100% efficacious, they would lead to the expression of animal and an average score per group was determined (group mean a truncated version of dystrophin, at best turning fatal total score) (statistical analysis: 1-way ANOVA with Bonferroni’s multiple comparison test. Mean values ± SD vs Vehicle; n = 5; T16 = sampling Duchenne into a milder Becker dystrophy type of dis- after 15 weeks of treatment; DIA = diaphragm; GAS = gastrocnemius; ease. This implies that a two-pronged approach focus- MTS = mean total score; TA = tibialis anterior; wt = wild type;). ing on both dystrophin restoration and a Additional file 4: Table 4. Summary of the statistical analysis results of normalization of the repair process is likely to provide functional and histological parameters in mdx mice. Givinostat administered at the dose of 37.5 mg/kg led to significant improvements the most benefit to patients. in both functional tests (T8 and T16) and histological evaluations (except HDACis are ideally suited for this purpose since they for heart) (T16). Statistical analysis: functional tests, 2-way ANOVA with act on multiple aspects of muscle repair through modu- Bonferroni’s multiple comparison test; histological parameters, 1-way ANOVA with Bonferroni’s multiple comparison test. Mean values ± s.e. (*p lation of inflammation, collagen deposition, adipocyte < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001 vs Vehicle; ns = not differentiation, and muscle fiber regeneration . In significant; s = significant based on multiplicative model effect in addition, they may have direct effects on muscle con- gastrocnemius and additive model effect in tibialis anterior, as described in Statistical analysis section in Materials and Methods paragraph; T8 = traction. These desirable activities are mediated by dif- sampling after 8 weeks of treatment; T16 = sampling after 15 weeks of ferent HDAC subtypes that are expressed in different treatment; CSA = cross sectional area). cell types and are likely to have different requirements in Additional file 5: Table 5. Summary of differentially expressed miRNA terms of HDACi concentration and duration of inhib- in Naive wt, Naive mdx, Givinostat 37.5 mg/kg and vehicle mdx mice. A large number of statistically significant, differentially expressed miRNAs ition. Additional preclinical experiments will be needed could be identified in all contrasts. Included: number of miRNAs used in to dissect the role of individual HDAC subtypes in these the analysis with non-zero total read count; up: number of miRNAs processes and to define the optimal profile and dosing upregulated at FDR < 0.05; down: number of miRNAs downregulated at FDR < 0.05. FDR: false discovery rate. schedule of HDAC inhibitors. Additional file 6: Table 6. Statistical analysis of maximal normalized strength in D2.B10 mice. Summary of statistical analysis of maximal normalized strength in wt and D2.B10 mice (wt: wild type). 2-way ANOVA Conclusion with Bonferroni’s multiple comparison test was performed (*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001 vs Vehicle). Our data show that Givinostat improves the functional Additional file 7: Table 7. Histopathological evaluation of the severity activities and modifies the histological parameters in a of myodystrophy in different muscles of D2.B10 mice at T8 and T16. The dose-dependent way regardless of the Ltbp4 haplotype; histopathological method considered some parameters scored by indeed, Givinostat is effective in both the mdx (DMD severity and extension of the injury: muscle degeneration/necrosis, regeneration, inflammatory infiltrate, interstitial reaction and adipose mild phenotype) and in D2.B10 (DMD severe pheno- tissue deposition. Each parameter was classified by severity (mild = 1, type) murine models. Licandro et al. Skeletal Muscle (2021) 11:19 Page 20 of 22 manuscript. LC and RPo performed the analysis of the samples and analyzed moderate = 2 and severe = 3) and extension (focal = 1, multifocal = 2 the data. GF contributed to the critical reading of the manuscript. All authors and diffuse = 3). The individual severity score was calculated for each read and approved the final manuscript. animal and an average score per group was determined (group mean total score) (statistical analysis: 1-way ANOVA with Bonferroni’s multiple Author’s information comparison test. Mean values ± SD vs Vehicle; n = 5; T8 = sampling after Not applicable. 8 weeks of treatment; T16 = sampling after 15 weeks of treatment; DIA = diaphragm; GAS = gastrocnemius; MTS = mean total score; TA = tibialis anterior; wt = wild type;). Funding This work was supported by Regione Lombardia (grant “Accordi per la Additional file 8: Table 8. Summary of the statistical analysis results of Ricerca e l’Innovazione Cofinanziata dal POR FESR 2014-2020”, CUP: functional and histological parameters in D2.B10 mice. Givinostat E76D17000080009). administered at the doses of 5, 10 or 37.5 mg/kg and steroids (except for Deflazacort at T16) led to significant improvements in grip strength test, Availability of data and materials whereas, we observed a statistically significant improvement in the The datasets used and/or analyzed during the current study are available exhaustion test only with the highest dose of Givinostat at T16. For the from the corresponding author on reasonable request. histological analysis we observed a statistically significant result only in fibrosis at T8 in diaphragm (DIA) for all the doses of Givinostat administered and for Prednisone treatment. Givinostat at 37.5 mg/kg also Declarations significantly counteracted fibrosis in tibialis anterior (TA); Prednisone was able to diminish fibrosis in TA and gastrocnemius (GAS), whereas the Ethics approval and consent to participate antifibrotic effect of Deflazacort was observed only in TA. Statistical All mice were handled according to our Institutional Guidelines in analysis: functional tests, 2-way ANOVA with Bonferroni’s multiple compliance with national and international laws and policies (Italian comparison test; histological parameters, 1-way ANOVA with Bonferroni’s Governing Law: D.lgs 26/2014 “Attuazione della direttiva 2010/63/UE sulla multiple comparison test. Mean values ± s.e. (*p < 0.05; **p < 0.01; ***p < protezione degli animali utilizzati a fini scientifici”). The research project has 0.001; ****p < 0.0001 vs Vehicle; ns = not significant; T8 = sampling after been authorized by the Italian Ministry of Health. 8 weeks of treatment; T16 = sampling after 15 weeks of treatment; CSA = cross sectional area). Consent for publication Additional file 9: Figure 1. Experimental plan of treatments and Not applicable. functional tests in mdx (A) and D2.B10 (B) studies. Givinostat was orally administered by daily gavage in mdx mice, whereas it was dissolved in Competing interests drinking water for D2.B10 mice starting from day 7. Deflazacort and SAL, FL, CS, R Pomarico, R Perego, and GF declare that they have financial Prednisone were weekly administered by i.p. injection at the dose of 1 competing interest as they are Italfarmaco’s employees. LC declares that he mg/kg starting from day 9 only in D2.B10 mice. The grip strength and has no competing interests. run to exhaustion tests have been conducted after a training period of 3 and 4 days, respectively, during which mice become familiar with the Author details 1 2 procedures. Grip strength test was performed every week in both the New Drug Incubator, Italfarmaco S.p.A., Milan, Italy. University of studies, instead, run to exhaustion performance of mdx and D2.B10 mice Milano-Bicocca, Milan, Italy. School of Medicine and Surgery, University of was evaluated every 14 or 21 days, respectively, using a treadmill 4 Milano-Bicocca, Milan, Italy. Preclinical Development, Italfarmaco S.p.A., apparatus. Milan, Italy. Additional file 10: Figure 2. Effect of Givinostat on mdx mice body weight. Baseline BW values of 8 weeks old C57BL/10J wt and Naive mdx Received: 16 February 2021 Accepted: 27 May 2021 mice were 24.8 ± 0.8 and 27.7 ± 0.8 grams, respectively. At the end of the treatment period (day 105), the BW of mice treated with 25 and 37.5 mg/kg of Givinostat was 32.9 ± 0.9 and 33.1 ± 0.9 g respectively, i.e. References similar to that of the wt mice (33.3 ± 0.9 g) but significantly different (p < 1. Flanigan KM. Duchenne and Becker muscular dystrophies. Neurol Clin. 2014; 0.01 and p < 0.05, respectively) from that of the vehicle-treated mdx mice 32(3):671–88. https://doi.org/10.1016/j.ncl.2014.05.002. (37.2 ± 0.8 g), suggesting that Givinostat counteracted the pathologic BW 2. Koeks Z, Bladen CL, Salgado D, van Zwet E, Pogoryelova O, McMacken G, gain of mdx mice (BW = body weight; wt = wild type). 2-way ANOVA et al. Clinical outcomes in Duchenne muscular dystrophy: a study of 5345 with Bonferroni’s multiple comparison test was performed. patients from the TREAT-NMD DMD Global Database. J Neuromuscul Dis. 2017;4(4):293–306. https://doi.org/10.3233/JND-170280. Additional file 11: Figure 3. Effect of Givinostat, Prednisone and 3. Goemans N, Buyse G. Current treatment and management of Deflazacort on D2.B10 mice body weight. The baseline BW mean value of D2.B10 mice was 18.2 ± 0.6 grams, whereas the weight of the wt mice dystrophinopathies. Curr Treat Options Neurol. 2014;16(5):287. https://doi. org/10.1007/s11940-014-0287-4. was 24.8 ± 0.5 grams (p < 0.0001) and they still remained heavier than 4. Syed YY. Eteplirsen: first global approval. Drugs. 2016;76(17):1699–704. the dystrophic mice for all the duration of the experiment: at day 105, Review. https://doi.org/10.1007/s40265-016-0657-1. the mean BW of D2.B10 mice was 23.3 ± 0.6 g, whereas the BW of wt mice was 29.4 ± 0.5 g (p < 0.0001). There were no significant differences 5. Al Shaer D, Al Musaimi O, Albericio F, de la Torre BG. 2019 FDA TIDES (peptides and oligonucleotides) harvest. Pharmaceuticals (Basel). 2020;13(3): in BW of D2.B10 vehicle-treated mice compared to that of D2.B10 mice 40. Review. https://doi.org/10.3390/ph13030040. treated with either Givinostat or steroids (BW = body weight; wt = wild 6. Sheikh O, Yokota T. Advances in genetic characterization and genotype- type). 2-way ANOVA with Bonferroni’s multiple comparison test was performed. phenotype correlation of Duchenne and Becker muscular dystrophy in the personalized medicine era. J Pers Med. 2020;10(3):111. https://doi.org/10.33 90/jpm10030111. Acknowledgements 7. Bushby K, Finkel R, Wong B, Barohn R, Campbell C, Comi GP, et al. Ataluren The authors are grateful to M. Rochetti for the statistical analysis performed treatment of patients with nonsense mutation dystrophinopathy. Muscle on cross-sectional areas of TA, GAS, and DIA muscles, and to PL Puri and P Nerve. 2014;50(4):477–87. https://doi.org/10.1002/mus.24332. Bettica for critical reading of the manuscript and for the helpful discussions. 8. Ryan NJ. Ataluren: first global approval. Drugs. 2014;74(14) Review:1709–14. https://doi.org/10.1007/s40265-014-0287-4. Authors’ contributions 9. Bello L, Gordish-Dressman H, Morgenroth LP, Henricson EK, Duong T, SAL, FL, and CS conceived the project idea. SAL performed the experiments, Hoffman EP, et al. Prednisone/prednisolone and deflazacort regimens in the analyzed the data, and wrote much of the manuscript. FL contributed to CINRG Duchenne Natural History Study. Neurology. 2015;85(12):1048–55. analyze the data. FL, CS, RPo, RPe, and LC contributed to the writing of the https://doi.org/10.1212/WNL.0000000000001950. Licandro et al. Skeletal Muscle (2021) 11:19 Page 21 of 22 10. McDonald CM, Henricson EK, Abresch RT, Duong T, Joyce NC, Hu F, et al. 29. Colussi C, Brioschi M, Tremoli E, Straino S, Spallotta F, Mai A, et al. Proteomic Long-term effects of glucocorticoids on function, quality of life, and survival profile of differentially expressed plasma proteins from dystrophic mice and in patients with Duchenne Muscular Dystrophy: a prospective cohort study. following suberoylanilide hydroxamic acid treatment. Proteomics Clin Appl. Lancet. 2018;391(10119):451–61. https://doi.org/10.1016/S0140-6736(17)321 2010;4(1):71–83. https://doi.org/10.1002/prca.200900116. 60-8. 30. Consalvi S, Mozzetta C, Bettica P, Germani M, Fiorentini F, Del Bene F, et al. 11. Bylo M, Farewell R, Coppenrath VA, Yogaratnam D. A review of Deflazacort Preclinical studies in the mdx mouse model of Duchenne muscular for patients with Duchenne muscular dystrophy. Ann Pharmacother. 2020; dystrophy with the histone deacetylase inhibitor Givinostat. Mol Med. 2013; 54(8):788–94. Review. https://doi.org/10.1177/1060028019900500. 19(1):79–87. https://doi.org/10.2119/molmed.2013.00011. 12. Matthews E, Brassington R, Kuntzer T, Jichi F, Manzur AY. Corticosteroids for 31. Bettica P, Petrini S, D'Oria V, D'Amico A, Catteruccia M, Pane M, et al. the treatment of Duchenne muscular dystrophy. Cochrane Database Syst Histological effects of Givinostat in boys with Duchenne muscular Rev. 2016;5:CD003725. dystrophy. Neuromuscul Disord. 2016;26(10):643–9. https://doi.org/10.1016/j. 13. Wong BL, Rybalsky I, Shellenbarger KC, Tian C, McMahon MA, Rutter MM, nmd.2016.07.002. et al. Long-term outcome of interdisciplinary management of patients with 32. Coley WD, Bogdanik L, Vila MC, Yu Q, Van Der Meulen JH, Rayavarapu S, Duchenne muscular dystrophy receiving daily glucocorticoid treatment. J et al. Effect of genetic background on the dystrophic phenotype in mdx Pediatr. 2017;182:296–303 e1. mice. Hum Mol Genet. 2016;25(1):130–45. https://doi.org/10.1093/hmg/ddv4 14. Sali A, Guerron AD, Gordish-Dressman H, Spurney CF, Iantorno M, Hoffman 60. EP, et al. Glucocorticoid-treated mice are an inappropriate positive control 33. Mantuano P, Sanarica F, Conte E, Morgese MG, Capogrosso RF, Cozzoli A, for long-term preclinical studies in the mdx mice. Plos One. 2012;7(4): et al. Effect of a long-term treatment with metformin in dystrophic mdx e34204. https://doi.org/10.1371/journal.pone.0034204. mice: a reconsideration of its potential clinical interest in Duchenne 15. Quattrocelli M, Barefield DY, Warner JL, Vo AH, Hadhazy M, Earley JU, et al. muscular dystrophy. Biochem Pharmacol. 2018;154:89–103. https://doi.org/1 Intermittent glucocorticoid steroid dosing enhances muscle repair without 0.1016/j.bcp.2018.04.022. eliciting muscle atrophy. J Clin Invest. 2017;127(6):2418–32. https://doi.org/1 34. Bachmanov AA, Reed DR, Beauchamp GK, Tordoff MG. Food intake, water 0.1172/JCI91445. intake, and drinking spout side preference of 28 mouse strains. Behav Genet. 2002;32(6):435–43. https://doi.org/10.1023/A:1020884312053. 16. Heydemann A, Ceco E, Lim JE, Hadhazy M, Ryder P, Moran JL, et al. Latent TGF-beta-binding protein 4 modifies muscular dystrophy in mice. J Clin 35. De Luca A. Use of grip strength meter to assess the limb strength of mdx Invest. 2009;119(12):3703–12. https://doi.org/10.1172/JCI39845. mice. In: Experimental protocols for DMD animal models. Treat-NMD 17. Flanigan KM, Ceco E, Lamar KM, Kaminoh Y, Dunn DM, Mendell JR, et al. Neuromuscular Network. 2019. https://treat-nmd.org/wp-content/uploads/2 LTBP4 genotype predicts age of ambulatory loss in Duchenne muscular 016/08/MDX-DMD_M.2.2.001.pdf. Accessed 3 September 2008. dystrophy. Ann Neurol. 2013;73(4):481–8. https://doi.org/10.1002/ana.23819. 36. Grange RW. Use of treadmill and wheel exercise to assess dystrophic state. 18. Hammers DW, Hart CC, Matheny MK, Wright LA, Armellini M, Barton ER, In: Experimental protocols for DMD animal models. Treat-NMD et al. The D2.mdx mouse as a preclinical model of the skeletal muscle Neuromuscular Network. 2015. https://treat-nmd.org/wp-content/uploads/2 pathology associated with Duchenne muscular dystrophy. Sci Rep. 2020; 016/08/MDX-DMD_M.2.1.003-34.pdf. Accessed 12 November 2008. 10(1):14070. 37. Brown RP, Delp MD, Lindstedt SL, Rhomberg LR, Beliles RP. Physiological parameter values for physiologically based pharmacokinetic models. Toxicol 19. Ceco E, Bogdanovich S, Gardner B, Miller T, DeJesus A, Earley JU, et al. Ind Health. 1997;13(4):407–84. https://doi.org/10.1177/074823379701300401. Targeting latent TGFβ release in muscular dystrophy. Sci Transl Med. 2014; 6(259):259ra144. 38. Langmead B, Trapnell C, Pop M, Salzberg SL. Ultrafast and memory-efficient 20. Lu J, Liu Q, Wang L, Tu W, Chu H, Ding W, et al. Increased expression of alignment of short DNA sequences to the human genome. Genome Biol. latent TGF-beta-binding protein 4 affects the fibrotic process in scleroderma 2009;10(3):R25. https://doi.org/10.1186/gb-2009-10-3-r25. by TGF-beta/SMAD signaling. Lab Invest. 2017;97(5):591–601. https://doi. 39. Wouters L, Göhlmann HW, Bijnens L, Kass SU, Molenberghs G, Lewi PJ. org/10.1038/labinvest.2017.20. Graphical exploration of gene expression data: a comparative study of three 21. Bernasconi P, Torchiana E, Confalonieri P, Brugnoni R, Barresi R, Mora M, multivariate methods. Biometrics. 2003;59(4):1131–9. https://doi.org/10.1111/ et al. Expression of transforming growth factor-beta 1 in dystrophic patient j.0006-341X.2003.00130.x. muscles correlates with fibrosis. Pathogenetic role of a fibrogenic cytokine. 40. Love MI, Huber W, Anders S. Moderated estimation of fold change and J. Clin. Invest. 1995;96(2):1137–44. https://doi.org/10.1172/JCI118101. dispersion for RNA-seq data with DESeq2. Genome Biol. 2014;15(12):550. https://doi.org/10.1186/s13059-014-0550-8. 22. Bello L, Kesari A, Gordish-Dressman H, Cnaan A, Morgenroth LP, Punetha J, et al. Genetic modifiers of ambulation in the Cooperative International 41. Bourgon R, Gentleman R, Huber W. Independent filtering increases Neuromuscular Research Group Duchenne Natural History Study. Ann detection power for high-throughput experiments. Proc Natl Acad Sci U S Neurol. 2015;77(4):684–96. https://doi.org/10.1002/ana.24370. A. 2010;107(21):9546–51. https://doi.org/10.1073/pnas.0914005107. 23. Van den Bergen JC, Hiller M, Böhringer S, Vijfhuizen L, Ginjaar HB, Chaouch 42. Giordani L, Sandoná M, Rotini A, Puri PL, Consalvi S, Saccone V. Muscle- A, et al. Validation of genetic modifiers for Duchenne muscular dystrophy: a specific microRNAs as biomarkers of Duchenne muscular dystrophy multicentre study assessing SPP1 and LTBP4 variants. J Neurol Neurosurg progression and response to therapies. Rare Dis. 2014;2(1):e974969. https:// Psychiatry. 2015;86(10):1060–5. https://doi.org/10.1136/jnnp-2014-308409. doi.org/10.4161/21675511.2014.974969. 24. Rodrigues M, Echigoya Y, Fukada SI, Yokota T. Current translational research 43. Matsuzaka Y, Kishi S, Aoki Y, Komaki H, Oya Y, Takeda S, et al. Three novel and murine models for Duchenne muscular dystrophy. J Neuromuscul Dis. serum biomarkers, miR-1, miR-133a, and miR-206 for Limb-girdle muscular 2016;3(1):29–48. Review. https://doi.org/10.3233/JND-150113. dystrophy, Facioscapulohumeral muscular dystrophy, and Becker muscular 25. Fukada S, Morikawa D, Yamamoto Y, Yoshida T, Sumie N, Yamaguchi M, dystrophy. Environ Health Prev Med. 2014;19(6):452–8. https://doi.org/10.1 et al. Genetic background affects properties of satellite cells and mdx 007/s12199-014-0405-7. phenotypes. Am J Pathol. 2010;176(5):2414–24. https://doi.org/10.2353/ajpa 44. Lee Y, Andaloussi SE, Wood MJA. Exosomes and microvesicles: extracellular th.2010.090887. vesicles for genetic information transfer and gene therapy. Hum Mol Genet. 2012;21(R1):R125–34. https://doi.org/10.1093/hmg/dds317. 26. Rodrigues M, Echigoya Y, Maruyama R, Lim KRQ, Fukada SI, Yokota T. Impaired regenerative capacity and lower revertant fibre expansion in 45. Braicu C, Tomuleasa C, Monroig P, Cucuianu A, Berindan-Neagoe I, Calin GA. dystrophin-deficient Mdx muscles on DBA/2 background. Sci Rep. 2016;6(1): Exosomes as divine messengers: are they the Hermes of modern molecular 38371. https://doi.org/10.1038/srep38371. oncology? Cell Death Differ. 2015;22(1):34–45. https://doi.org/10.1038/cdd.2 27. Minetti GC, Colussi C, Adami R, Serra C, Mozzetta C, Parente V, et al. 014.130. Functional and morphological recovery of dystrophic muscles in mice 46. Iraci N, Leonardi T, Gessler F, Vega B, Pluchino S. Focus on extracellular treated with deacetylase inhibitors. Nat Med. 2006;12(10):1147–50. https:// vesicles: physiological role and signalling properties of extracellular doi.org/10.1038/nm1479. membrane vesicles. Int J Mol Sci. 2016;17(2):171. https://doi.org/10.3390/ 28. Colussi C, Mozzetta C, Gurtner A, Illi B, Rosati J, Straino S, et al. HDAC2 ijms17020171. blockade by nitric oxide and histone deacetylase inhibitors reveals a 47. Mizuno H, Nakamura A, Aoki Y, Ito N, Kishi S, Yamamoto K, et al. common target in Duchenne muscular dystrophy treatment. Proc Natl Acad Identification of muscle-specific microRNAs in serum of muscular dystrophy Sci USA. 2008;105(49):19183–7. https://doi.org/10.1073/pnas.0805514105. animal models: promising novel blood-based markers for muscular Licandro et al. Skeletal Muscle (2021) 11:19 Page 22 of 22 dystrophy. PLoS One. 2011;6(3):e18388. https://doi.org/10.1371/journal.pone. 65. Chen X, Barozzi I, Termanini A, Prosperini E, Recchiuti A, Dalli J, et al. 0018388. Requirement for the histone deacetylase Hdac3 for the inflammatory gene 48. Coenen-Stass AML, Sork H, Gatto S, Godfrey C, Bhomra A, Krjutškov K, et al. expression program in macrophages. Proc Natl Acad Sci U S A. 2012; Comprehensive RNA-sequencing analysis in serum and muscle reveals 109(42):E2865–74. https://doi.org/10.1073/pnas.1121131109. novel small RNA signatures with biomarker potential for DMD. Mol Ther 66. Bradner JE, West N, Grachan ML, Greenberg EF, Haggarty SJ, Warnow T, Nucleic Acids. 2018;13:1–15. https://doi.org/10.1016/j.omtn.2018.08.005. et al. Chemical phylogenetics of histone deacetylases. Nat Chem Biol. 2010; 6(3):238–43. https://doi.org/10.1038/nchembio.313. 49. Cacchiarelli D, Legnini I, Martone J, Cazzella V, D’Amico A, Bertini E, et al. 67. Sanderson L, Taylor GW, Aboagye EO, Alao JP, Latigo JR, Coombes RC, et al. miRNAs as serum biomarkers for Duchenne muscular dystrophy. EMBO Mol Plasma pharmacokinetics and metabolism of the histone deacetylase Med. 2011;3(5):258–65. https://doi.org/10.1002/emmm.201100133. inhibitor trichostatin a after intraperitoneal administration to mice. Drug 50. Chen JF, Mandel EM, Thomson JM, Wu Q, Callis TE, Hammond SM, Metab Dispos. 2004;32(10):1132–8. https://doi.org/10.1124/dmd.104.000638. et al. The role of microRNA-1 and microRNA-133 in skeletal muscle 68. Leoni F, Zaliani A, Bertolini G, Porro G, Pagani P, Pozzi P, et al. The antitumor proliferation and differentiation. Nat Genet. 2006;38(2):228–33. https:// histone deacetylase inhibitor suberoylanilide hydroxamic acid exhibits doi.org/10.1038/ng1725. antiinflammatory properties via suppression of cytokines. Proc Natl Acad Sci 51. Koutsoulidou A, Mastroyiannopoulos NP, Furling D, Uney JB, Phylactou LA. U S A. 2002;99(5):2995–3000. https://doi.org/10.1073/pnas.052702999. Expression of miR-1, miR-133a, miR-133b and miR-206 increases during 69. Leoni F, Fossati G, Lewis EC, Lee JK, Porro G, Pagani P, et al. The histone development of human skeletal muscle. BMC Dev Biol. 2011;11(1):34. deacetylase inhibitor ITF2357 reduces production of pro-inflammatory https://doi.org/10.1186/1471-213X-11-34. cytokines in vitro and systemic inflammation in vivo. Mol Med. 2005;11(1- 52. Yuasa K, Hagiwara Y, Ando M, Nakamura A, Takeda S, Hijikata T. MicroRNA- 12):1–15. https://doi.org/10.2119/2006-00005.Dinarello. 206 is highly expressed in newly formed muscle fibers: implications 70. Mendell JR, Sahenk Z, Lehman K, Nease C, Lowes LP, Miller NF, et al. regarding potential for muscle regeneration and maturation in muscular Assessment of systemic delivery of rAAVrh74.MHCK7.micro-dystrophin in dystrophy. Cell Struct Funct. 2008;33(2):163–9. https://doi.org/10.1247/csf. children with Duchenne muscular dystrophy: a nonrandomized controlled trial. JAMA Neurol. 2020;77(9):1–10. 53. Liu N, Williams AH, Maxeiner JM, Bezprozvannaya S, Shelton JM, Richardson 71. Consalvi S, Saccone V, Giordani L, Minetti G, Mozzetta C, Puri PL. Histone JA, et al. microRNA-206 promotes skeletal muscle regeneration and delays deacetylase inhibitors in the treatment of muscular dystrophies: epigenetic progression of Duchenne muscular dystrophy in mice. J Clin Invest. 2012; drugs for genetic diseases. Mol Med. 2011;17(5-6):457–65. https://doi.org/1 122(6):2054–65. https://doi.org/10.1172/JCI62656. 0.2119/molmed.2011.00049. 54. Roberts TC, Godfrey C, McClorey G, Vader P, Briggs D, Gardiner C, et al. Extracellular microRNAs are dynamic non-vesicular biomarkers of muscle turnover. Nucleic Acids Res. 2013;41(20):9500–13. https://doi.org/10.1093/na Publisher’sNote r/gkt724. Springer Nature remains neutral with regard to jurisdictional claims in 55. Wang Z, Li Q, Chamba Y, Zhang B, Shang P, Zhang H, et al. Identification of published maps and institutional affiliations. genes related to growth and lipid deposition from transcriptome profiles of pig muscle tissue. PLoS One. 2015;10(10):e0141138. https://doi.org/10.1371/ journal.pone.0141138. 56. Saccone V, Consalvi S, Giordani L, Mozzetta C, Barozzi I, Sandoná M, et al. HDAC-regulated myomiRs control BAF60 variant exchange and direct the functional phenotype of fibro-adipogenic progenitors in dystrophic muscles. Genes Dev. 2014;28(8):841–57. https://doi.org/10.1101/gad.2344 68.113. 57. Sandonà M, Consalvi S, Tucciarone L, De Bardi M, Scimeca M, Angelini D, et al. HDAC inhibitors tune miRNAs in extracellular vesicles of dystrophic muscle-resident mesenchymal cells. EMBO Rep. 2020;21(9):e50863. https:// doi.org/10.15252/embr.202050863. 58. Ben-Zvi I, Volinsky N, Grosman-Rimon L, Haviv I, Rozen G, Andria N, et al. Cardiac-peripheral transvenous gradients of microRNA expression in systolic heart failure patients. ESC Heart Fail. 2020;7(3):835–43. https://doi.org/10.1 002/ehf2.12597. 59. Hu ZQ, Luo JF, Yu XJ, Zhu JN, Huang L, Yang J, et al. Targeting myocyte- specific enhancer factor 2D contributes to the suppression of cardiac hypertrophic growth by miR-92b-3p in mice. Oncotarget. 2017;8(54):92079– 89. https://doi.org/10.18632/oncotarget.20759. 60. Yu XJ, Huang YQ, Shan ZX, Zhu JN, Hu ZQ, Huang L, et al. MicroRNA-92b-3p suppresses angiotensin II-induced cardiomyocyte hypertrophy via targeting HAND2. Life Sci. 2019;232:116635. https://doi.org/10.1016/j.lfs.2019.116635. 61. Jaffe KM, McDonald CM, Ingman E, Haas J. Symptoms of upper gastrointestinal dysfunction in Duchenne muscular dystrophy: case-control study. Arch Phys Med Rehabil. 1990;71(10):742–4. 62. Mozzetta C, Consalvi S, Saccone V, Tierney M, Diamantini A, Mitchell KJ, et al. Fibroadipogenic progenitors mediate the ability of HDAC inhibitors to promote regeneration in dystrophic muscles of young, but not old Mdx mice. EMBO Mol Med. 2013;5(4):626–39. https://doi.org/10.1002/emmm.2012 63. Jeong MY, Lin YH, Wennersten SA, Demos-Davies KM, Cavasin MA, Mahaffey JH, et al. Histone deacetylase activity governs diastolic dysfunction through a nongenomic mechanism. Sci Transl Med. 2018;10(427):eaao0144. 64. Cacchiarelli D, Martone J, Girardi E, Cesana M, Incitti T, Morlando M, et al. MicroRNAs involved in molecular circuitries relevant for the Duchenne muscular dystrophy pathogenesis are controlled by the dystrophin/nNOS pathway. Cell Metab. 2010;12(4):341–51. https://doi.org/10.1016/j.cmet.2010. 07.008.
Skeletal Muscle – Springer Journals
Published: Jul 22, 2021
Keywords: Duchenne nuscular dystrophy; Givinostat; mdx; D2.B10; LTBP4; HDAC inhibitor