O

Higher testosterone levels in female sprinters can contribute to athletic success, allowing them to reach a higher competitive level (Ahmetov et al. 2020). Similarly, in an assessment of a large cohort of elite female athletes, sprinters showed higher testosterone levels than long-distance runners (Bermon et al. 2014). In an assessment of a large cohort of elite male athletes, sprinters showed higher free testosterone levels than athletes in other sports (e.g., long-distance runners) (Bermon and Garnier 2017). Given that muscle hypertrophy (and the increase in muscle function it brings) has a performance-enhancing effect in sports that depend on strength and power, higher levels of testosterone create an advantage (Wood and Stanton 2012). One hundred and forty-eight physically active individuals (47 females, 101 males) were assessed for cross-sectional area (CSA) of fast-twitch muscle fibers. The aim of the present study was to test the association of 822 testosterone-increasing SNPs with muscle-related traits (muscle fiber size, fat-free mass and handgrip strength) and to validate the identified SNPs in independent cohorts of strength and power athletes. Circulating testosterone levels are a heritable trait with anabolic properties in various tissues, including skeletal muscle.
Relatively few functional pathways were enriched for the DEGs identified as responsive to testosterone. (G) Separation of individuals along the first two principal components describing variation in transcriptome-wide expression, alongside the mean (±SE) values for (H) PC1, and (I) PC2 for each group. (C) Frequency distributions of log2 fold change (FC) values from (A,B), with gray indicating overlap between juvenile and subadult distributions and colors indicating an excess of low sex bias in juveniles and high sex bias in subadults. Similarly, we used Wilcoxon and Bartlett’s tests to assess whether the sexes differed in the median absolute value and variance in log2 (fold change between juveniles and subadults). Summary statistics for each gene in each experimental contrast (e.g., log2 FC, F statistics, and p values) are available as Supplementary Material. Prior to analysis, we filtered libraries to remove low-expression transcripts using edgeR function "FilterByExpr," which retained 13,464 transcripts for analysis.
Among these genes, 4 were higher (Table S21 in file S1) and 4 were lower (Table S22 in file S1) in T-treated than control males in both tissues. In males, 8 genes were differentially expressed between T-treated and control individuals in both medial amygdala and hypothalamus (Figure 2c). Heat maps show scaled individual expression scores for just the genes that were significantly differentially expressed between T-treated and control individuals in each sex (a,b,d,e). In both brain regions, there were significant differences in expression between T-treated and control females.
Many of the genes we identified as differentially expressed by sex were also related to cell growth and cytoskeleton structure. If these transcription factors have a subtle effect on expression of other genes, then perhaps they affected the expression of many other genes, but below the level of detection of this experiment. Follistatin plays a wide-range of roles throughout the body ; however, in the brain it suppresses the release of follicle stimulating hormone, a key regulator of gametogenesis , from the pituitary . Combined, these findings suggest that males may be producing more steroids (or their precursors), but also more rapidly turning them over, compared to females. Others have noted that pooling tissues comprised of discrete regions can reduce ability to detect differences , likely because changes in gene expression in particular regions may be masked if the expression level was modified in only a subset of tissues that were analyzed collectively. One gene (Cytochrome P450 19A1) was significantly differentially expressed between T-treated and controls of both sexes in the hypothalamus, and that gene was more highly expressed in T-treated than control individuals in both sexes. In the medial amygdala, 36 isogroups (of 15,279 expressed) were significantly differentially expressed between T-treated and control males (Figure 2d) with 15 higher in T-treated than control males (Table S17 in file S1) and 21 lower in T-treated than control males (Table S18 in file S1).
This is one of the genes annotated with the GO term response to nutrient levels, and the role of CB1 in signaling hunger may be involved in the reduced body mass and fattening induced by T-treatment , . Cannabinoid receptor 1 (CB1) was more highly expressed in the hypothalamus of T-treated than control females. Another set of differentially expressed genes appears to be related to the metabolic and activity effects of T-treatment , , . Similarly, monoamine oxidase A (MAO-A) is less expressed in the medial amygdala of T-treated than control females.
Correlation between genetically predicted testosterone (PGST) and gene expression values for the NUPR1L, PSPH1, and PTPRD transcripts in the arterial tibial (left panel), skeletal muscle (middle panel), and pancreas (right panel) tissue in males (A) and females (B). However, the genetic correlation for testosterone levels between males and females is close to zero, suggesting distinct biology of testosterone between the two sexes3,18. In this study, we correlated predicted testosterone and within-sex gene expression measures across 40 human tissues to identify genes that show sex-differential control of gene expression and examine how this varies across tissues. Since there are variable outcomes in gene expression after TP exposure depending on sex chromosome complement, inherent epigenetic programming may be involved which has set the stage for differential response to such hormones in each sex. These feminizing and masculinizing events as a result of exposing embryonic neural stem cells to a male sex hormone are perhaps identifying key genes responsible for establishing typical male gene expression within the cells of the developing central nervous system.
For example, some genes that were significantly upregulated with age in one sex were nonetheless sex-biased toward greater expression in subadults of the opposite sex (Figures 2E,F). In the middle column, female age-biased DEGs from (B) are mapped onto plots describing (E) sex-biased expression, and (H) age-biased expression in males. Although these associations are expected in part because the same subadult transcriptomes that we used to identify sex-biased DEGs were also used to characterize age-biased expression in each sex, these statistical associations nonetheless confirm an important assumption of our approach (i.e., that patterns of age bias within each sex capture features of the development of sex-biased gene expression). Approximately half (233 of 478; 49%) of the significantly sex-biased DEGs in subadults were also identified as significantly age-biased DEGs in either males or females (Supplementary Figure S2A).