Muscle biopsy RNA-seq from 5 week bedrest study volunteers

Sedentary lifestyle, chronic disease or microgravity can cause muscle deconditioning that then has an impact on other physiological systems. An example is the nervous system, which is adversely affected by decreased physical activity resulting in increased incidence of neurological problems such as chronic pain. We sought to better understand how this might occur by conducting RNA sequencing experiments on muscle biopsies from human volunteers in a 5-week bed-rest study with an exercise intervention arm. We also used a computational method for examining ligand-receptor interactions between muscle and human dorsal root ganglion (DRG) neurons, the latter of which play a key role in nociception and are generators of signals responsible for chronic pain. We identified 1352 differentially expressed genes (DEGs) in bed rest subjects without an exercise intervention but only 132 DEGs in subjects with the intervention. Thirty-six genes were shared between the exercise and no intervention groups. Among 591 upregulated muscle genes in the no intervention arm, 26 of these were ligands that have receptors that are expressed by human DRG neurons. Overall design: Skeletal muscle biopsies were collected at the Institute for Exercise and Environmental Medicine. Methods and subject characteristics for this study have been previously described (Krainski et al., 2014). Subjects were healthy, nonsmoking, adults aged 20-54 years. Subjects were randomly assigned to one of two groups: bed rest only (BR-CON; n=9, 1 female), or bed rest with exercise countermeasure (BR-EX; n=16, 2 female). All subjects underwent-6-degree head down tilt bed rest for 35 consecutive days. BR-EX subjects received an exercise countermeasure consisting of rowing ergometer training on 6 days/week and biweekly resistance training for the duration of the bed rest period. A needle muscle biopsy of 300 mg tissue on average was obtained from the midbelly of the right (pre) and left (post) vastus lateralis muscle before and after the five week bed rest period (Krainski et al., 2014). Part of each biopsy was used for a previously published study (Krainski et al., 2014); at least 50 mg per sample remained for use in the present study. Library Generation and Sequencing Based on available muscle following initial ultrastructural and histochemical analysis (Krainski et al., 2014) biopsy samples from 13 subjects underwent RNA sequencing, consisting of all three female samples (1 BR-CON, 2 BR-EX), and ten male samples (5 BR-CON, 5 BR-EX; matched for age and exercise tolerance at baseline). Researchers performing sequencing and subsequent analyses were informed of subject pair matches and whether samples were from baseline or post-intervention, but were blinded to intervention received by individual subjects (i.e. the researchers did not know whether subjects were sedentary or exercised during bedrest until the first round of analysis was complete). Following RNA extraction (RNeasy Plus Universal Mini Kit, Qiagen), RNA yield was quantified using a Nanodrop system (ThermoFisher Scientific), and RNA quality assessed by fragment analyzer (Advanced Analytical Technologies). Stranded mRNA library preparation and sequencing were completed at the University of Texas at Dallas (UT Dallas), TX, USA. 76-bp, single-end sequencing of RNA-seq libraries was performed on the Illumina Hi-Seq sequencing platform. Mapping and Quantification of RNA-seq data: RNA-seq read files (fastq files) were checked for quality by FastQC (Babraham Bioinformatics) and read trimming was done based on Phred score and per-base nucleotide content, with trimmed, 60 bp reads from the 13 – 72 bp positions being used for downstream analysis. Trimmed reads were mapped to the reference transcriptome GENCODE v27 (Frankish et al., 2019), and its underlying reference genome GRCh38.p10 retaining only uniquely mapped reads by using the STAR (v2.7.3a) aligner (Dobin et al., 2013) with the following command line mapping parameters: STAR --runThreadN 20 --runMode alignReads --genomeLoad NoSharedMemory -- outMultimapperOrder Random --outSAMtype BAM SortedByCoordinate --outSAMattributes All --outSAMattrIHstart 0 --outSAMprimaryFlag AllBestScore --outBAMcompression-1 -- outBAMsortingThreadN 20 --outFilterType BySJout --outFilterMultimapNmax 10 -- outFilterMismatchNoverReadLmax 0.06 --twopassMode Basic --quantMode GeneCounts -- alignSJoverhangMin 5 --alignSJDBoverhangMin 3 --alignIntronMax 1000000 -- alignMatesGapMax 1000000 --outFilterIntronMotifs RemoveNoncanonical -- outSAMstrandField intronMotif --outFilterScoreMinOverLread 0.3 -- outFilterMatchNminOverLread 0.3 --readFilesCommand zcat --alignSoftClipAtReferenceEnds No --alignEndsType EndToEnd Relative abundance per sample (in Transcripts per Million or TPM) was calculated from the mapped reads using Stringtie (Pertea et al., 2015). Genes were filtered to only retain protein-coding genes and TPMs were re-normalized per sample to a million based only on the set of coding genes A fold-change ratio (post intervention / baseline) was calculated for each coding gene in every subject.