Molecular networks of human muscle adaptation to exercise and age

Abstract

Physical activity and molecular ageing presumably interact to precipitate musculoskeletal decline in humans with age. Herein, we have delineated molecular networks for these two major components of sarcopenic risk using multiple independent clinical cohorts. We generated genome-wide transcript profiles from individuals (n = 44) who then undertook 20 weeks of supervised resistance-exercise training (RET). Expectedly, our subjects exhibited a marked range of hypertrophic responses (3% to +28%), and when applying Ingenuity Pathway Analysis (IPA) up-stream analysis to ~580 genes that co-varied with gain in lean mass, we identified rapamycin (mTOR) signaling associating with growth (P = 1.4 × 10(-30)). Paradoxically, those displaying most hypertrophy exhibited an inhibited mTOR activation signature, including the striking down-regulation of 70 rRNAs. Differential analysis found networks mimicking developmental processes (activated all-trans-retinoic acid (ATRA, Z-score = 4.5; P = 6 × 10(-13)) and inhibited aryl-hydrocarbon receptor signaling (AhR, Z-score = -2.3; P = 3 × 10(-7))) with RET. Intriguingly, as ATRA and AhR gene-sets were also a feature of endurance exercise training (EET), they appear to represent "generic" physical activity responsive gene-networks. For age, we found that differential gene-expression methods do not produce consistent molecular differences between young versus old individuals. Instead, utilizing two independent cohorts (n = 45 and n = 52), with a continuum of subject ages (18-78 y), the first reproducible set of age-related transcripts in human muscle was identified. This analysis identified ~500 genes highly enriched in post-transcriptional processes (P = 1 × 10(-6)) and with negligible links to the aforementioned generic exercise regulated gene-sets and some overlap with ribosomal genes. The RNA signatures from multiple compounds all targeting serotonin, DNA topoisomerase antagonism, and RXR activation were significantly related to the muscle age-related genes. Finally, a number of specific chromosomal loci, including 1q12 and 13q21, contributed by more than chance to the age-related gene list (P = 0.01-0.005), implying possible epigenetic events. We conclude that human muscle age-related molecular processes appear distinct from the processes regulated by those of physical activity.

Figure 1. Resistance exercise training induces an all-trans retinoic acid differential gene expression signature that is common with endurance exercise training.

A) Forty-four subjects completed a 20 wk supervised resistance exercise training program (RET) and biopsy RNA was profiled before and 72 hr after training. Following SAM analysis of the 38 subjects that demonstrated a clear physiological gain, the gene list was uploaded to the Ingenuity Pathway Analysis database (IPA) and the up-stream regulators were identified using IPA's new up-stream tool. An All-trans-retinoic acid (Tretinoin) gene expression literature network was found to have a significant overlap with the RET dataset (p = 6×10⁻¹³). Furthermore, the direction changes of the common transcripts were sufficiently similar enough to conclude that ATRA like activity was increased (Z-score = 4.5). B) We re-evaluated our earlier RNA responses to endurance exercise training (EET) study which involved biopsy profiles before and after 6 weeks endurance training in 24 subjects , using the IPA tool and while the ATRA was not independently significant the vast majority of the genes within the RET ATRA signature were regulated in an identical manner by EET.

Figure 2. Inhibition of the mTOR-related expression network is correlated with gains in lean mass following RET.

A) Quantitative SAM analysis was used to relate the change in RNA expression in response to 10 wk RET in 44 subjects. The change in gene expression was related to the change in lean mass (%) and a false discovery rate calculated based on permutation of the subject labels. Data were imported into IPA and 384 genes (FDR<5%) could be mapped to the data-base for up-stream analysis. An active rapamycin signature, equating to inhibition of mTOR signaling was discovered (Z-score = 2.8 for directional consistency; P-value for transcript overlap p = 1.4×10⁻³⁰). B) Given the strength of the negative statistical association between the rapamycin signature, we then plotted the data to establish the precise nature of the relationship. We presented the mean gains in lean mass by quartiles establishing that 25% of the subject demonstrated negligible changes in lean mass. C) We selected a representative subset of the genes from Figure 2A and plotted the mean changes with respect to lean mass changes. This established that those with the greatest lean mass actually had a reduction in mTOR related genes with RET and not simple a lesser increase as one might have expected from first inspection of Figure 2A.

Figure 3. Using principal component analysis to evaluate the relationship between physiological and acute protein signaling events to RET induced gains in lean mass.

A) Change in lean mass following 20 wk RET and a number of physiological parameters which demonstrated the most variance were scaled to a common value and plotted using principal component analysis in R. Principal component (PC) 1 captured the major variance in lean mass gains across subjects however none of the commonly postulated physiological parameters varied with lean mass (linear regression analysis demonstrated no significant association also). PC2, the second largest proportion of independent variance also demonstrated no association between factors such as fiber type or age and gains in lean mass. B) Phospho-protein signaling 2 hr after a combined exercise and nutrition acute intervention (to promote anabolic signaling) were scaled and plotted with change in lean mass following 20 wk RET. Again these acute signaling events shared little common variance in either PC1 or PC2 with changes in lean mass with 20 wk RET.

Figure 4. Differential gene expression analysis, contrasting young and old subjects, does not produce a reliable biomarker signature of age.

Several attempts have been to define a set of genes that differ in skeletal muscle between young and old human subjects. We re-analysed three of the most robust and largest human studies with common methods, including our new study, and contrasted the genes identified to be differentially regulated using SAM analysis and Gene Ontology analysis. No common pattern of differential gene expression could be found using this analysis method indicating that no prior gene signature for muscle ageing can be considered a reliable marker of muscle age in humans. Gene ontology analysis found that both the Trappe and Melov data sets had modest enrichment in mitochondrial genes, which were down-regulated with age however this was not true for the DRET study and both Melov and Trappe data-sets had elderly with much lower physical fitness levels making it impossible to attribute these changes to age per se with differential expression analysis.

Figure 5. Quantitative SAM analysis using a continuum of age versus gene expression produces network hubs that are activated with human muscle age.

Using a total of 97 U133+2 Affymetrix gene-chips newly produced from two independent studies, the DRET study and the HERITAGE family study we produced a novel analysis that relied on the full age-range present in these data sets. A) We first found a set of genes that co-varied with age in the DRET study and then confirmed that 580 of these were also related to age in the HERITAGE study. Mitochondrial genes were not a feature of this linear age vs gene analysis. We then mapped the Affymetrix probe-sets to the IPA database and examined the up-stream analysis output. We found in IPA that the age-related dataset was consistent with the activation of the PGR (z-score = 2.6 and p-value = 0.001) and RXR (z-score = 2.0 and p-value = 0.0001) proteins and 5-fluorouracil agonism (Z-score = 2.2 and p-value = 0.0005). B) We noted that some members of these age-related networks were also associated with lean mass gains in humans. However about 50% of the common genes were positively associated with lean mass gain and age; and 50% were regulated in a discordant manner. Clearly some responses can be causal, some may be purely correlative and some may represent compensatory events.

Figure 6. Quantitative SAM analysis using a continuum of age versus gene expression produces network hubs that are inhibited with human muscle age.

Using a total of 97 U133+2 Affymetrix gene-chips newly produced from two independent studies, the DRET study and the HERITAGE family study we produced a novel analysis that relied on the full age-range present in these data sets. A) We first found a set of genes that co-varied with age in the DRET study and then confirmed that 580 of these were also related to age in the HERITAGE study. Mitochondrial genes were not a feature of this linear age vs gene analysis. We then mapped the Affymetrix probe-sets to the IPA database and examined the up-stream analysis output. We found in IPA that the age-related dataset was consistent with including c-MYC (z-score = −2.8 and p-value<0.0001) and CLDN7 (z-score = −2.6 and p-value = 0.05). B) A few members of these age-related networks were also associated with lean mass gains in humans and this included mTOR regulated genes, which were negatively associated with increasing age and thus in contrast to the lean-mass association.

Figure 7. Positional enrichment analysis identified age-related genes that were over-represented on 4 chromosomes including Chromosome 1 and 13.

A plot of selected genes found to be over-represented at 1q12 and 13q21. A) Positional enrichment analysis identified 3 genes at each loci and B) the relationship with age of each gene was plotted with the linear correlation coefficient provided and the FDR derived from the qSAM analysis (FDR<5%). Genes at these loci included proteins that are known to influence mTOR related signalling.