Adult (12 m) male and female C57/BL6 mice (9 males, five females) were used in this study for single hindlimb immobilization. Mice had ad libitum access to food and water and were on a normal light cycle (6am-6pm light/6pm-6am dark). All procedures were reviewed and approved by the Augusta University IACUC. The left hindlimb of each mouse was immobilized for 2 weeks using an adapted version of a Velcro cast (Aihara et al., 2017). Cast padding (McKesson, 16–010), connected to the pad of the foot by surgical tape, was wrapped from the pad of the foot around the hindlimb several times (Figure 1). This padding was secured in place with a layer of surgical tape. A drop of super glue was added at the top of the foot to complete the cast. Replacement of the cast was performed as needed throughout the 2 weeks, usually every 2–3 days. After 2 weeks of immobilization the animals were euthanized following IACUC approved procedures.

Immediately after euthanization, quadriceps muscles were dissected free, weighed, and placed in 10% buffered formalin. Muscles were removed from fixation after 24 h and stored in 70% ethanol. Samples were paraffin embedded, cut in cross section, and trichrome stained to facilitate measurement of fiber size. Fiber size was calculated by taking images from standardized areas of each cross section at ×20 magnification using a brightfield Leica DMCS microscope outfitted with a micropublisher six camera (QImaging). Ocular software (QImaging) was then used to take images of each sample which were then analyzed with ImageJ to calculate the area of individual muscle fibers. Thirty muscle fibers were measured per sample to calculate an average fiber size per muscle. Slides were stained for H&E using standard protocol and images were obtained using both 20x and 40x magnification.

A total of six 12-month-old C57/BL6 mice (three males and three females) mice were used for FACs sorting. The mice were immobilized for 2 weeks as described above. Promptly after euthanization quadriceps mass was measured, leg muscles were isolated, and muscles were then dissociated with forceps and placed into an enzymatic mixture (dispase II 2.5 U/mL, Roche, collagenase B 0.2% Roche, and MgCl2 5 mM) resting on ice. The mixture was placed in a water bath (37°C) for 45 min, briefly vortexing every 15 min to ensure even digestion. The samples were filtered using a 100 and 40 µm nylon cell strainer (BD Biosciences). The cells were then stained with the following antibodies: CD45 (Invitrogen, MCD4528), CD31 (Invitrogen, RM5228), CD11b (Invitrogen, RM2828), CD34 (BD Biosciences, 551,387), Ly-6A-Ly6E (BD Biosciences, 561,021), α7-integrin (R&D, FAB3518N), and for PDGFRa either CD140a (eBioscience, 17-1,401-81) or CD140a MicroBead Kit (Miltenyl Biotec, #130-101-547) was used. The cells were then sorted using BD Biosciences FACSCAria II SORP based on: CD31-/CD11b-/CD45-/Sca1+/CD34+/(CD140a) PDGFR-α+. The cells were sorted into 500 µL media, pelleted, and all but 50 µL of media was removed. 500µL of TRIzol was added to the samples, frozen on dry ice and then stored at −80. For those sorted with a bead kit, the cells were sorted according to: CD31-/CD11b-/CD45-/Sca1+/CD34+ and sorted into 500 µL media following manufacturer specifications and samples stored in TRIzol as described above (Supplementary Figure S1).

Total RNA was extracted from sorted cells using Single Cell RNA Purification Kit (NORGEN) according to the recommended procedure. RNA quality was assessed using Agilent 2,100 Bioanalyzer with RNA 6000 Pico kit (Agilent Technologies) and assured of RNA Integrity Number (RIN) greater than 5.0. Total RNA from 12 samples were processed for cDNA library preparations using SMARTer Stranded RNA-seq kit v2 Pico Input (Clontech_TaKaRa). Briefly, 1 ng of total RNA was fragmented based on the quality of the RNA input and converted to cDNA fragments. These cDNA fragments then had the addition of a single “A” base and subsequent ligation of the adapter. Following purification, the cDNA libraries were treated with ZapR and R-Probes v2 (Clontech_TaKaRa) to deplete rRNA (18S and 28S). The rRNA depleted libraries were enriched with PCR and purified to create the final libraries. Each library was examined by Bioanalyzer and Qubit HS DNA kit (Thermofisher) to evaluate library quality and quantity, respectively. A total of 12 libraries were pooled and run on an Illumina NextSeq500 sequencer with high output using 150-cycle for 75bp paired-end at the Integrated Genomics Shared Resource in the Georgia Cancer Center of Augusta University. BCL files generated from the NextSeq500 are converted to FASTQ files for downstream analysis. The pair-end raw sequence reads from 12 RNA samples were transferred to Partek Flow server via SFTP producing 12 paired RNA-seq samples. The analysis was performed using Partek Flow RNA-Seq tool. After pre-alignment QA/QC check and base trimming the sequencing reads were aligned to mm10 genome using the Bowtie2 aligner. These aligned reads were then quantified to annotation model mm10-emsembl-release-97 using Partek E/M algorithm, followed by data normalization (offset, scaling, and log base two transformation) to generate sequence counts of 23,416 unique mouse genes for down-stream differential expression analysis. Pre-alignment, the total reads ranges from 33 + million to 44 + million reads per sample with QA/QC score of at least 33.83 per sample. Through pre-alignment QA/QC check and base trimming the sequencing reads were aligned to mm10 genome using the Bowtie2 aligner. The alignment rate ranges from 31 to 61% per sample with post-alignment QA/QC score of at least 34.14.

Droplet digital PCR (ddPCR) was performed using approximately 10 ng total RNA from isolated FAPs transcribed into cDNA using the iScript cDNA Synthesis Kit (Biorad) according to the manufacturer’s instructions. The ddPCR assay utilized a QX200 ddPCR Supermix for Probes (no dUTP) from Bio-Rad. The PCR products were quantified using the Bio-Rad QX200 droplet reader and subsequently analyzed with the QuantaSoft software as previously described (Helwa et al., 2017). On average 15,000 droplets were obtained during the process for each sample, with a requirement of 10,000 droplets needed for a sample to be analyzed. There also had to be a clear separation of background (noise) droplets and the signal droplets analyzed. Probe assays for IL-1β (dMmuEG5065947, Biorad), Actb (dMmuEG5193531, Biorad) and Cdkn2a (dMmuEG5193677) were used to detect gene expression.

Gastrocnemius muscles from twelve immobilized and twelve nonimmobilized samples were sonicated, and RNA was isolated using the TRIzol reagent (Invitrogen) according to the manufacturer’s instructions. Total RNA was isolated using a RNeasy Mini Kit (Qiagen). 1µg of total RNA was then reverse transcribed using the iScript cDNA Synthesis Kit (BioRad). PCR was then run using primers for UBC (Forward: 5′-AGC​CCA​GTG​TTA​CCA​CCA​AG-3′, R: 5′-ACC​CAA​GAA​CAA​GCA​CAA​GG-3′), ActB (F: 5′-AGC​CAT​GTA​CGT​AGC​CAT​CC-3′, R: 5′-GCT​GTG​GTG​GTG​AAG​CTG​TA-3′) and IL1β (F: 5′-GCA​ACT​GTT​CCT​GAA​CTC​AAC​T-3′, R: 5′-ATC​TTT​TGG​GGT​CCG​TCA​ACT-3′).

Slides were submerged in xylene and 100% EtOH then treated with hydrogen peroxide for 10 min at room temperature. Slides were placed in pretreatment (ACD Bio) and boiled at 99°C for 30 min using a rice cooker. The slides were then incubated with protease plus (ACD Bio) for 30 min at 40°C in the EZ Hybrid Oven (ACD Bio). RNAscope multiplex assay protocol was then performed according to manufacturer instructions. In brief, the probe mixtures (ACD Bio) were placed on the slides for a 2-h incubation at 40°C. Probes used in this study were for IL-1β (NM_008,361.3, 2–950), Cdkn2a (NM_009877.2, 35-886) and PDGFRα (NM_011058.2, 223-1161). Positive and negative control probes were provided by ACD Bio. Afterwards, the slides were incubated overnight at 4°C in 5xSSC buffer (Millipore Sigma). The slides were then incubated with AMP1, AMP2 and AMP3 (ACD Bio) at 40C. This was followed by incubation of HRP-C1 (ACD Bio), followed by the corresponding opal dye, followed by HRP block (ACD Bio). This was repeated for C2 and C3. Opal dyes 570, 620 and 690 (Akoya Biosciences) were used and diluted in TSA dilution buffer (ACD Biosciences). Once complete, mounting media with Dapi (Vector Labs) was added and the slide covered with a cover slip. Slides were then imaged using a Leica Stellaris confocal microscope at the Cell and Tissue Imaging Core Laboratory of Augusta University.

For differential analysis of RNA-seq data normalized gene counts were analyzed through a 2 × 2 ANOVA model (with leg and gender as the two factors) in Partek Flow. A p-value of 0.05 was used as the cut-off value for the factor leg with absolute fold change between left and right legs being at least 2-fold without interaction effect between leg and gender factors. Functional enrichment analysis of the RNA-seq data was performed using ToppFun analysis (Chen et al., 2009) of the top 25 differentially (increased) expressed genes. ddPCR results were analyzed using Exact Wilcoxon rank sum two sample test statistic with normalization of the gene of interest using the mean expression of ActB for each group. The normalized expression values were used in the statistical analysis. Paired t-tests were used to compare quadriceps weight, relative percentage weights and FACs sorted FAP cell numbers. Unpaired t-tests were used for the PCR analysis of IL-1β in whole muscle lysate. GraphPad Prism (GraphPad Software, San Diego, California, United States, version 8.0.0) was used for analysis unless otherwise indicated.

The 2-week single hindlimb immobilization period resulted in a significant loss of quadriceps muscle mass (p < 0.0001) and quadriceps fiber size (p < 0.0001) in both sexes (Figures 2A,B). There was not a significant difference detected in overall body weight among all mice (pooled data), although a loss of weight was seen in male mice (p < 0.05). These results suggest that the single hindlimb immobilization model used in this study can induce significant muscle atrophy over a two-week timeframe. This muscle atrophy can be visualized with smaller fibers in the H&E images in the immobilized muscle compared to the nonimmobilized muscle (Figure 2C).

Total counts of FAP cell number were calculated during FACS sorting. Analysis of the counts showed that both FAP cell number and total cell number tended to be lower in the immobilized limb (Supplementary Table S2), so that FAP cell number normalized to total number of cells did not differ significantly between the immobilized and nonimmobilized limbs (p = 0.12).

There were a total of 919 genes differentially expressed with IL-1β identified as the gene whose fold-change (positive) was the greatest between muscle from the left (immobilized) and right limbs (Table 1). Functional enrichment analysis of the RNA-seq data also revealed changes in IL-1 signaling for both molecular functions (Figure 3A) and pathways (Figure 3B) with hindlimb immobilization.

The qRT-PCR analysis of cDNA from whole immobilized and nonimmobilized muscle showed a significant increase in IL1β compared to the nonimmobilized muscle (p < 0.05, Figure 4A). The ddPCR analysis of the samples used for RNAseq confirmed the increase seen in IL1β expression among FAPs (p = 0.0051, Figure 4B). ddPCR was also performed for Cdkn2a expression as a marker of senescence in FAPs isolated from immobilized and non-immobilized quadriceps. Results showed a significant increase in Cdkn2a in FAPs isolated from the immobilized limb (Figure 4C). A full list of the 919 differentially expressed genes can be found in Supplementary Table S1.

RNAscope images show co-localization of IL-1β and PDGFRα in sections of immobilized quadriceps muscle (Figure 5 and Supplementary Figure S2). Co-localization of PDGFRα and Cdkn2a was also seen in immobilized muscle (Figure 5 and Supplementary Figure S2). In the non-immobilized contralateral control limbs, there was little to no co-localization seen, confirming that the immobilization treatment is inducing the expression of IL-1β and Cdkn2a in PDGFRα+ cells (Figure 6 and Supplementary Figure S3). These results support the ddPCR, PCR and RNAseq data showing that PDGFRα+ FAP cells increase the expression of these two genes after 2 weeks of immobilization. Negative control staining can be found in Supplementary Figure S4 (Supplementary Figure S4).