Mitigating sarcoplasmic reticulum stress limits disuse-induced muscle loss in hindlimb unloaded mice.

Muscle disuse in the hindlimb unloaded (HU) mice causes significant atrophy and weakness. However, the cellular and molecular mechanisms driving disuse-muscle atrophy remain elusive. We investigated the potential contribution of proteins dysregulation by sarcoplasmic reticulum (SR), a condition called SR stress, to muscle loss during HU. Male, c57BL/6j mice were assigned to ground-based controls or HU groups treated with vehicle or 4-phenylbutyrate (4-PBA), a potent inhibitor of SR stress, once a day for three weeks. We report that the 4-PBA reduced the SR stress and partly reversed the muscle atrophy and weakness in the HU mice. Transcriptome analysis revealed that several genes were switched on (n = 3688) or differentially expressed (n = 1184) due to HU. GO, and KEGG term analysis revealed alterations in pathways associated with the assembly of cilia and microtubules, extracellular matrix proteins regulation, calcium homeostasis, and immune modulation during HU. The muscle restoration with 4-PBA partly reversed these changes along with differential and unique expression of several genes. The analysis of genes among the two comparisons (HU-v vs. control and HU-t vs. HU-v.) shows 841 genes were overlapped between the two comparisons and they may be regulated by 4-PBA. Altogether, our findings suggest that the pharmacological suppression of SR stress may be an effective strategy to prevent disuse-induced muscle weakness and atrophy.

Introduction

Mechanical unloading of skeletal muscle results in rapid loss of muscle mass and force with time[1]. This condition is relevant to a plethora of scenarios, from prolonged bed rest in patients with chronic pathologies[2,3] to microgravity during space flights[4]. Several interventions (dietary supplements, electrical stimulation, passive muscle contraction, and physical exercise) have been used to reduce disuse muscle atrophy[5]. However, technical challenges and poor compliance with these therapies warrant the necessity of a pharmacological intervention to boost muscle mass and force during prolonged inactivity.

The experimental rodent model of hindlimb unloading (HU) recapitulates several features of mechanical unloading, including muscle atrophy and weakness[6], and is an excellent model to test potential pharmacological interventions[7]. However, no drug therapy exists to effectively prevent muscle atrophy and weakness, partly because the exact molecular mechanism(s) coupling inactivity to muscle impediment are poorly understood. The emerging role(s) of endo/sarcoplasmic reticulum (ER/SR) in skeletal muscle diseases is only beginning to surface[8,9]. The SR plays a pivotal role in calcium homeostasis and protein folding in the mammalian skeletal muscle. However, overload of misfolded and/or unfolded proteins in the SR lumen, a condition called SR stress, can lead to pathological consequences in chronic diseases[10]. SR stress leads to the activation of a signaling network called the unfolded protein response (UPR) and its downstream consequences. Chronic elevation of UPR activates cell death pathways (apoptosis, inflammation, and autophagy) associated with degenerative muscle disorders, including aging, myopathies, and catabolic conditions[11]. However, establishing a direct causality between SR stress and muscle detriment requires assessing muscle mass and force by mitigating SR stress in disease models. To our knowledge, only a few studies have investigated the effects of inhibition of SR stress on skeletal muscle in disease conditions, none of them involve disuse atrophy. Thus, the treatment of mice with 4-phenyl butyrate (4-PBA), a pan-SR stress inhibitor, ameliorates muscle atrophy in burn injury[12] or genetic mutations associated with SR calcium dysregulation[13]. However, a recent study shows that short-term inhibition of SR stress exacerbates rather than protects against cancer cachexia[14]. These discrepancies warrant a thorough investigation of molecular alterations in mechanically unloaded skeletal muscle with 4-PBA treatment. However, such a characterization has not been performed before. Therefore, a careful dissection of global transcriptomic changes in skeletal muscle is required to characterize the contribution of SR stress to skeletal muscle impairment.

Here, we investigated the potential contribution of SR stress to muscle weakness and atrophy in HU mice. We show that pharmacological inhibition of SR stress in HU conditions partially restores muscle mass and strength. Partial muscle restoration is associated with alteration in signature transcriptomic changes in skeletal muscle. Using a non-biased approach, we identified several candidate genes and molecular pathways that dictate SR stress-induced muscle restoration in HU conditions. Our findings indicate that pharmacological inhibition of SR stress may be a promising therapy for disuse-induced muscle loss.

Methods

Hindlimb unloading (HU) mice model

We randomly assigned 4-month-old, male c57BL/6j mice into ground-based controls (n = 10) and HU mice. HU mice were further divided into two groups and treated with either PBS as the vehicle (HU-v; n = 10) or 4-PBA (HU-t; 100 mg/kg/d for three weeks via intraperitoneal injections) (n = 8). The dose was chosen based on published literature[15]. Mice were kept under controlled environmental conditions (20 ± 1 °C, with light/dark periods of 12 h each) with food (standard chow diet for mice) and water provided ad-libitum. As previously described, the HU mice were suspended for 21 days in specially designed cages[16,17]. At the end of the experiments, mice were euthanized via cervical dislocation, and the gastrocnemius muscles were immediately excised, weighed, and snap-frozen in liquid nitrogen for further analysis, as described previously[18] (Fig. 1). The experimental protocol was approved by the University of Sharjah ACUC (Animal Care and Use Committee) in agreement with accepted international standards. All methods were carried out in accordance with the relevant guidelines and regulations. Additionally, all methods are reported in accordance with ARRIVE guidelines.

Fig. 1

Experimental design.

Experimental design of the study.

Grip strength measurements

We measured the grip strength using a Grip Strength Meter with a mesh grid pull bar (Columbus Instruments, Columbus, OH) specifically designed for mice. Mice were allowed to grip the metal grids of the grip meter with their paws and gently pulled backward in a horizontal plane until they could no longer hold the grip. The grip strength was obtained from the two forelimbs and all four limbs, including hindlimbs. The peak grip strength obtained in ten consecutive trials was designated as the mouse’s grip strength and was normalized to body weights, as described previously[17].

Sample collection and Library preparation

Total RNA was extracted from gastrocnemius muscles of the three groups of mice. Before fragmentation, the mRNAs were purified with magnetic beads tagged with Ploy-T tail. The first strand was prepared with the random hexamers primers, and then the second strand was also synthesized. The library was checked with Qubit and quantified with real-time PCR. Next, the library was analyzed with a bioanalyzer for size detection. The libraries were pooled, and then they were sequenced on Illumina platforms.

Clustering and sequencing

The samples were index-coded, and clusters were generated, followed by sequencing the library using an Illumina platform to generate paired-end reads.

Quality control

Raw reads in the FASTQ format were processed by fastp. To get the clean data, reads containing the adapter sequence and low quality were removed. Following the clean reads, Q20 and Q30, clean data were calculated. Further downstream analysis was calculated using clean data with good quality. Using the Spliced Transcripts software (STAR), paired-end clean reads were aligned to the reference genome. This alignment is based on the RNAseq alignment algorithm followed by seed clustering and stitching. Supplementary file 1 contains the QC statistics of the data, while Supplementary File 2 contains the mapping statistics.

Quantification

The numbers of reads that were mapped to each gene were counted using the FeatureCounts. To normalize the reads, Fragments Per Kilobase of transcript sequence per Million (FPKM) fragments mapped were calculated[19].

Differential expression analysis

Differential genes expression analysis was performed between two conditions/groups (three biological replicates per condition) using DESeq2 R package. The P values were calculated using Benjamini and Hochberg’s method for the false discovery rate (FDR). Genes with an adjusted P-value < 0.05 were assigned as significantly differentially expressed genes.

Gene ontology (GO), KEGG pathways, and reactome enrichment analysis

Biological ontologies is a common way for finding shared functions among genes. GO enrichment analysis of differentially expressed genes was performed by the clusterProfiler R package. GO terms with corrected P-value less than 0.05 (Fisher exact test) were considered significantly enriched in the differential expressed genes. For the KEGG enrichment, R package clusterProfiler was used to find KEGG pathways that were statistically significant among the differential expression genes, as described previously[20]. KEGG terms with adjusted P value less than 0.05 (Fisher exact test) were considered significant enrichment. For reactome enrichment analysis, the clusterProfiler was also used and terms with adjusted p-value less than 0.05 were considered significantly enriched.

Quantitative real-time PCR validation

A set of 13 randomly selected genes were chosen for RT-PCR to validate the transcript data. The procedure is described in detail elsewhere[21]. Briefly, total RNA was extracted using the RNase kit (Qiagen, Valencia, CA, USA) from 20 mg of frozen gastrocnemius tissues. cDNA was synthesized using QuantiTect Reverse Transcriptase kit (Qiagen 205,311, Germany), and quantitative real-time PCR was performed using SYBR Green PCR Master Mix with primers. The primer sequence is described in Table 1. Calculations were performed using delta-delta-ct method, as described previously[22].

Table 1

Primers’ sequence. Sequence of the primers used in the study.

Primer name	Forward	Reverse	Reference
s-XBP1	CTGAGTCCGAATCAGGTGCAG	GTCCATGGGAAGATGTTCTGG	[44]
u-XBP1	CAGCACTCAGACTATGTGCA	GTCCATGGGAAGATGTTCTGG	[44]
ATF4	GGGTTCTGTCTTCCACTCCA	AAGCAGCAGAGTCAGGCTTTC	[45]
CHOP	CCACCACACCTGAAAGCAGAA	AGGTGAAAGGCAGGGACTCA	[45]
ADAM1B	TTCCCTCCATGAGGAATACG	GTGCCTTCCTCTTTGCAGTC	[46]
PEG3	GGTTCAGTGTGGGTGCACTAGACT	GCTCACACCCAAGGGCTTGAGCGT	[47]
H2Q7	CGGGCCAACACTCGCTGCAA	GTATCTGCGGAGCGACTGCAT	[48]
NLE1	TATCAAGCTGTGGGATGGC	AGCATATACCTCATCGGCGT	[49]
MEG3	GGACTTCACGCACAACACGTT	GTCCCACGCAGGATTCCA	[50]
ELN	GGAGTTCCCGGTGGAGTCTATT	ACCAGGAATGCCACCAACACCTG	[51]
SCARF1	CTCTCCAGAGGTGCTCAACC	ATGCCTCCATCAGTGGTCTC	[52]
UBXN10	GAGTCTGTGCAACGGTCT CA	TCCTGGCTTGAATCCTCTTG	Self-designed primer
HSF5	GCTGTAGGACAATTTCACCGG	TTCCAAGGGAGTTCTGCCAC	[53]
EID3	AGGAGGAGGAAGGCTCAGAC	GCCTCTCTGGTTCTGCTCAC	[54]
NME8	GACGATGCGGTTAAGGTCTC	TTGCCTCTGCATCAGTATGG	Self-designed primer
ZDHHC19	TTGCTGCCTTCAATGTAACG	TGAGAAGTTGAGCGAGACGA	Self-designed primer
HSPA1B	CAAGATCACCATCACCAACG	ATGACCTCCTGGCACTTGTC	[55]
GAPDH	CATCACTGCCACCCAGAAGACTG	ATGCCAGTGAGCTTCCCGTTCAG	[56]

(s-XBP1; spliced x-box-binding protein1, u-XBP1; unspliced x-box-binding protein1, ATF4; activating transcription factor4; CHOP; C/EBP homologous protein, ADAM1B; ADAM Metallopeptidase Domain 1B, PEG 3; paternally expressed 3, H2Q7; histocompatibility 2, Q region locus 7, NLE1; notchless protein homolog 1, MEG3; maternally expressed 3, ELN; elastin, SCARF1; Scavenger Receptor Class F Member 1, UBXN10; UBX domain protein 19, HSF5; heat shock transcription factor 5, EID3; EP300 Interacting Inhibitor Of Differentiation 3, NME8; NME family member 8, ZDHHC19; Zinc Finger DHHC-Type Palmitoyltransferase 19, HSPA1B; heat shock protein family A member 1B).

Western blot

Muscle tissues were homogenized in RIPA buffer containing 50 mM Tris (pH = 7.4), 150 mM NaCl and protease inhibitors. Proteins were quantified using the Bio-Rad kit (Sigma-Aldrich, Poole, UK) and transferred to a nitrocellulose membrane after electrophoresis using 8–15% polyacrylamide gels, as described before[23]. The western blot was derived from the same experiment and the tissues were processed in parallel. The western blot was derived from the same experiment and the tissues were processed in parallel. Membranes were probed using anti-GRP78 primary antibody (Cat # 3177, Cell Signaling Technology) at 1:1000 dilution and a secondary antibody at 1:10,000 (HRP-linked anti-rabbit IgG; Cat # 7074 S, cell signaling, Danvers, MA 01923, USA) as described previously[22].

Histology

Gastrocnemius muscles were immediately embedded in optimal cutting temperature (OCT) compound and snap frozen. Sections were cut at 10 µm with a Leica 3050 cryotome and stained for hematoxylin and Eosin, as previously described[18]. Zeiss LSM 510 Meta confocal microscope (Melville, NY, USA) was used for imaging and the Image J software (National Institute of Health, Bethesda, MD, USA) was used for image analysis, as described previously[18].

Raman spectroscopy

The experimental Raman spectra were obtained by using Renishaw inVia confocal Raman microscope. Three specimens of gastrocnemius muscles were selected from each of the three groups of mice. We collected ten spectra from randomly selected locations of each sample to obtain the average. Each recording involved exposure of a 50 µm muscle segment for 10 s to a 785 nm laser of 1% intensity (14 mW). The spectral range was kept between 100 and 1200 cm−1, representing the peaks for biological molecules. All spectra were collected at the intervals of 500 cm−1 and then stitched together to obtain the full spectrum.

Statistical analysis

All numerical values are presented as mean ± SEM, and the comparisons among the groups were performed by one-way analysis of variance (ANOVA) and Turkey’s multiple comparison tests, with a single pooled variance. For the real time PCR, unpaired student’s t-test was performed. Data were analyzed using GraphPad Prism 9 (GraphPad Software, La Jolla, CA), and p < 0.05 was considered statistically significant.