Stress increases sperm respiration and motility in mice and men

Abstract

Semen quality and fertility has declined over the last 50 years, corresponding to ever-increasing environmental stressors. However, the cellular mechanisms involved and their impact on sperm functions remain unknown. In a repeated sampling human cohort study, we identify a significant effect of prior perceived stress to increase sperm motility 2-3 months following stress, timing that expands upon our previous studies revealing significant stress-associated changes in sperm RNA important for fertility. We mechanistically examine this post-stress timing in mice using an in vitro stress model in the epididymal epithelial cells responsible for sperm maturation and find 7282 differentially H3K27me3 bound DNA regions involving genes critical for mitochondrial and metabolic pathways. Further, prior stress exposure significantly changes the composition and size of epithelial cell-secreted extracellular vesicles that when incubated with mouse sperm, increase mitochondrial respiration and sperm motility, adding to our prior work showing impacts on embryo development. Together, these studies identify a time-dependent, translational signaling pathway that communicates stress experience to sperm, ultimately affecting reproductive functions.

Fig. 1. Sperm motility increased 2-3 months following elevated perceived stress in healthy men.

a Schematic of the human cohort study. Briefly, whole ejaculate was collected for one to six months, and Perceived Stress Score (PSS) was reported for the month of each collection (Time 0), as well as one month (Time −1), two months (Time −2), and three months (Time −3) prior to each collection. At each collection, sperm motility was measured by computer automated semen analysis (CASA) in fresh whole ejaculate. The association between prior PSS and motility was modeled using mixed effects modeling by the nlme package. b–f Whole ejaculate volume, concentration, % motile sperm, total sperm count, and total motile sperm count (TMC) across all visits. N = 34 participants, each providing one monthly semen sample for up to six months. Data are mean ± SEM; one-way analysis of variance demonstrates no significant changes in volume, concentration or % motile sperm. Kruskal-Wallis test demonstrates no significant changes in (e) total sperm count or (f) total motile sperm count. g Models of the association between PSS at Time 0 and VSL (β = 0.001, t(108) = 0.25, p = 0.81) at collection. h Models of the association between PSS at Time −1 and VSL (β = −0.0003, t(107) = −0.10, p = 0.92) at collection. i Models of the association between PSS at Time −2 and VSL (β = 0.005, t(107) = 1.67, p = 0.09) at collection. j Models of the association between PSS at Time −3 and VSL (β = 0.008, t(107) = 3.09, p < 0.01) at collection. For g, N = 34 across 146 observations, and for h–j, N = 34 across 145 observations due to one missed visit by a participant. For all models, slope significance was tested by a two-sided t-test per the standard nlme package, shaded gray (g), blue (h), red (i), and purple (j) regions reflect the SEM and tick marks on the x-axis denote individual PSS for each observation. Source data are provided as a Source Data file. Figure 1a created with BioRender.com released under a Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International license (https://creativecommons.org/licenses/by-nc-nd/4.0/deed.en).

Fig. 2. Prior stress dramatically increases H3K27me3 binding distribution using CUT&RUN sequencing analysis.

a Schematic of our previously established EEC in vitro post-corticosterone paradigm with treatment beginning at confluency (day 0). Colors indicate: The treatment period (red), the wash out period (light-blue), and the post-corticosterone timepoint (dark-blue) when CUT&RUN was performed. Black circles indicate media changes. b The CUT&RUN mechanism where protein A/G-MNase cleaves DNA bound by the ubiquitous transcriptional repressor, H3K27me3, and DNA is extracted for sequencing. c Percent of differentially bound H3K27me3 loci detected in all samples within specific components of the nearest gene and (d) the distribution of loci relative to the transcription start site (TSS). Color represents the distance from the TSS. Panels c and d summarize all 7282 peaks identified as differentially bound and do not depend on the magnitude of binding in any individual sample. e Representative summary plots of H3K27me3 binding in representative samples on day 9. Blue lines represent the H3K27me3-enriched cluster peaks. The green line represents the H3K27me3-depleted peaks. Shading represents the SEM. f Heatmap of 7282 differentially bound loci post-treatment in one representative sample from each group. Rows reflect individual peaks, and columns reflects a base pair up to 5 Kb from the beginning of the peak. Dark blue indicates enrichment relative to yellow. For the differential binding analysis represented in e, f, N = 4 in post-vehicle and 5 in post-corticosterone treatment group; FDR < 0.05. g Gene set enrichment analysis of the genes aligning with H3K27me3-associated features. For the top 30 pathways, the node size corresponds to count while node color corresponds with statistical significance (adjusted p-value, utilizing the “BH” method). N = 4 in post-vehicle and 5 in post-corticosterone treatment groups. Figure 2a,b created with BioRender.com released under a Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International license (https://creativecommons.org/licenses/by-nc-nd/4.0/deed.en).

Fig. 3. Co-regulation of transcriptional networks preceded transcriptional downregulation, indicating that prior corticosterone treatment influenced the trajectory of a cellular allostatic set point.

a Confluent EEC cultures were exposed to our in vitro post-corticosterone paradigm. RNA was isolated before treatment at day 0 (confluency), at the end of treatment (day 3), and at subsequent media exchanges during the wash-out period (days 6 and 9), denoted by white arrows. Circles indicate media exchanges. b Heatmap of all genes from RNA-sequencing of EEC RNA isolated at the denoted timepoints. Unbiased hierarchical clustering of samples is visualized by dendrogram. c Module-trait associations between module eigengenes calculated from gene expression patterns in the EEC dataset and time or treatment. Rows correspond to a module eigengene while columns align with time or treatment. Color designates the correlation according to the legend. Pearson correlation and p value for each module are presented. d The top ten significant biological processes identified by functional analysis of genes in the MEpink module. FDR is presented in line with each bar. e RNA-seq expression data were assessed by principal component analysis which revealed that component 1 accounts for 42.7% of the variance and aligns with treatment while component 2 accounts for 22.8% of variance and aligns with maturation (N = 4 plates). Color represents treatment group. f Differential expression analyses revealed 272 DEGs between day 9 post-corticosterone and post-vehicle samples visualized by heatmap. Hierarchical clustering of samples visualized by dendrogram (N = 4 plates, adjusted p < 0.05, p value adjustment method = “BH”). Source data are provided as a Source Data file. Figure 3a created with BioRender.com released under a Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International license (https://creativecommons.org/licenses/by-nc-nd/4.0/deed.en).

Fig. 4. Cellular allostatic set point maintained decreased basal energy requirements and ATP production rates driven by decreased mitochondrial complex I function.

a Schematic of the timeline for our EEC in vitro corticosterone treatment paradigm. b Representative Mito Stress Test data where cultures were exposed to oligomycin, FCCP, and rotenone and antimycin A. N = 12 in post-vehicle and 11 in post-corticosterone treatment group. c Baseline mitochondrial oxygen consumption rate (OCR). N = 12 in the post-vehicle and 11 in the post-corticosterone treatment group, two-sided t-test, ****p < 0.0001. d, e Total and oxidative ATP production rates. N = 12 in post-vehicle and 11 in post-corticosterone treatment group; two-sided t-test, ****p < 0.0001. f Schematic of the citric acid cycle and electron transport chain contributions to oxidative respiration. Red arrows emphasize how substrates contribute NADH for complex I activity. Letters annotate the corresponding panel. g Representative Seahorse experiment in which EEC cultures were permeabilized with recombinant perfringolysin O, followed by one of the following: glutamate and malate, pyruvate and malate, alpha-ketoglutarate and malate, or succinate. N = 9 in the post vehicle and 8 in the post-corticosterone treatment group. The change in OCR (dOCR) was significantly less in post-corticosterone EECs following injection of glutamate and malate (h) and pyruvate and malate (i), but not alpha ketoglutarate and malate (j) or succinate (k), a complex II substrate. N = 9 in the post vehicle and 8 in the post-corticosterone treatment group; two-sided t-test, **p = 0.0013, ****p < 0.0001. l Complex I oxidized NADH at a significantly reduced rate in post-corticosterone EECs, indicating decreased complex I activity. N = 5; two-sided t-test. **p = 0.0087. m Ndufa1 protein band volume relative to alpha-tubulin loading control band volume from EEC whole cell lysate. Processing of the blot is described in detail in. N = 3 in the post-vehicle and 4 in the post-corticosterone treatment group; two-sided t-test, *p = 0.0364. For all, data are mean ± SEM. For all appropriate, respirometry data is normalized by average ug protein/well from post-corticosterone or post-vehicle treatment wells without Mito Stress Test injections. The post-vehicle treatment group is represented by black or gray points or bars, post-corticosterone treatment group is represented by light blue. Source data are provided as a Source Data file. Figure 4a, f created with BioRender.com released under a Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International license (https://creativecommons.org/licenses/by-nc-nd/4.0/deed.en).

Fig. 5. Prior stress exposed EEC secreted EVs increased sperm respiration and motility.

a Schematic of the timeline for the in vitro corticosterone treatment paradigm. The white arrow indicates when media was collected for ultracentrifugation and EV isolation. b Plot of the size distribution of EVs isolated from day 9 EEC conditioned media. N = 8. c Area under the curve analysis (AUC) of the distribution confirmed there was no difference in EV concentration. N = 8. Student’s t-test, p > 0.05. d Nanoparticle tracking analysis revealed post-corticosterone EEC EVs (cort-EEC EVs) had a decreased median size. N = 8. Two-sided t-test, *p = 0.0106. e Schematic detailing the protocol of conditioned media collection to EV isolation to sperm incubation and whole cell respirometry. f Baseline mitochondrial oxygen consumption rate. N = 7 wells. Two-sided t-test, *p = 0.0334. g Extracellular acidification rate. N = 7 wells. Two-sided t-test, ****p < 0.0001. h ATP production rate. N = 7 wells. Two-sided t-test, **p < 0.0081. i Glycolytic ATP production rate. N = 7 wells. Two-sided t-test, ****p < 0.0001. j Schematic detailing the CASA velocity outcomes following the EV incubation detailed in e. Curve velocity is red, average velocity is blue, and linear velocity is black. Increases in k curve velocity (**p < 0.0037), l average velocity (**p < 0.0098), and m linear velocity (*p < 0.0165) as determined by CASA of all sperm analyzed. For k–m N = 8 for vehicle and 7 for corticosterone EEC EV incubation groups; Two-sided t-test. For all panels, data are mean ± SEM. For sperm respiratory data, average number of cells/well was used for normalization. The post-vehicle treatment group is represented by black or gray points or bars, post-corticosterone treatment group is represented by light blue. Source data are provided as a Source Data file. Figure 5a, e, j created with BioRender.com released under a Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International license (https://creativecommons.org/licenses/by-nc-nd/4.0/deed.en).