Impaired ketogenesis in Leydig Cells drives testicular aging

Abstract

Testicular aging commonly leads to testosterone deficiency and impaired spermatogenesis, yet the underlying mechanisms remain elusive. Here, we show that Leydig cells are particularly vulnerable to aging processes in testis. Single-cell RNA sequencing identifies the expression of Hmgcs2, the gene encoding rate-limiting enzyme of ketogenesis, decreases significantly in Leydig cells from aged mice. Additionally, the concentrations of ketone bodies β-hydroxybutyric acid and acetoacetic acid in young testes are substantially higher than that in serum, but significantly diminish in aged testes. Silencing of Hmgcs2 in young Leydig cells drives cell senescence and accelerated testicular aging. Mechanistically, β-hydroxybutyric acid upregulates the expression of Foxo3a by facilitating histone acetylation, thereby mitigating Leydig cells senescence and promoting testosterone production. Consistently, enhanced ketogenesis by genetic manipulation or oral β-hydroxybutyric acid supplementation alleviates Leydig cells senescence and ameliorates testicular aging in aged mice. These findings highlight defective ketogenesis as a pivotal factor in testicular aging, suggesting potential therapeutic avenues for addressing age-related testicular dysfunction.

Fig. 1. Single-cell transcriptome profiling of young and aged mice testes.

a Schematic of the experimental workflow. b Left: tSNE (t-Distributed Stochastic Neighbor Embedding) plot showing the annotated testicular cell types of mice testes. Cells are colored and annotated by cell types. Right: tSNE plots showing distribution of different cell types in the young (top) and aged (bottom) testes. c tSNE plot showing the expression profiles of the indicated cell type-specific marker genes for the assessed cell types in mice testes. The color key, ranging from grey to blue, indicates low to high gene expression levels, respectively. d Left: heatmap showing the top 30 differentially expressed genes of each cell cluster. The scaled gene expression levels were colored according to Z-score at the top. Right: the corresponding GO terms enriched in each cell cluster with -log10 (adjusted P-value) colored according to the color key at the top. e tSNE plot showing young and aged LCs populations. Violin plots showing the expression of senescence markers in young and aged LCs. f, g Quantitative RT-PCR analysis of senescence markers (p21, Cxcl10) in LCs from young and aged mice. n = 3 per group. h Left: representative confocal images of testicular sections obtained from young and aged mice. NC: negative control. The sections were stained with senescence marker p21 (Cyclin-dependent kinase inhibitor 1 A), HSD3β (hydroxysteroid dehydrogenase-3β), and DAPI (4,6-diamino-2-phenyl indole). Right: quantitative analysis of the p21⁺ HSD3β⁺ cells. Scale bar: 50 μm. n = 3 per group. Data were presented as mean ± SEM. Significance was determined by Two-tailed t-test (f–h), or Two-sided Wilcoxon rank-sum test (e) or One-sided hypergeometric test with BH correction (d). Illustrations were created in BioRender. Xia, K. (2025) https://BioRender.com/2y29pus. Source data are provided as a Source Data file.

Fig. 2. LCs in the aged testis show impaired ketogenesis, which leads to a decrease in testicular ketone body concentration.

a Heatmap showing both the upregulated and downregulated DEGs in the aged group of LCs. The scaled gene expression levels are colored according to Z-score. Six upregulated or downregulated GO terms of DEGs were listed, with dot size indicating the range of −log10 (adjusted P-value). Color keys, ranging from grey to blue, indicate the absolute value of the NES. b Volcano plots of the upregulated and downregulated genes in GO terms listed in (a). c Diagram of ketogenesis. d Violin plots showing the expression of genes encoding enzymes for ketogenesis in LCs. e–g Quantitative RT-PCR analysis of genes encoding enzymes for ketogenesis in primary LCs. n = 3 per group. h Left: representative confocal images of testicular sections obtained from young and aged mice. NC: negative control. The sections were stained with HMGCS2 (3-hydroxy-3-methylglutaryl-CoA synthetase 2), HSD3β, and DAPI. Scale bar: 50 μm. Right: quantitative analysis of the HMGCS2⁺ HSD3β⁺ cells. n = 3 per group. i–l The concentration of ketone bodies in serum or testes. n = 4 per group in (i) and (k); n = 8 per group in (j), n = 10 per group in (l). Data were presented as mean ± SEM. Significance was determined by Two-tailed t-test (e–l), or Two-sided Wilcoxon rank-sum test with bonferroni correction (b, d) or Two-sided permutation test with BH correction (a). Illustrations were created with BioRender. Source data are provided as a Source Data file.

Fig. 3. Inhibition of Hmgcs2 accelerates senescence of Leydig cells.

a Schematic of the experimental workflow. b, c The level of ketone bodies in MLTC-1 cells. n = 3 per group. Cells were treated with Hymeglusin (3 μM) for 4 days before analysis of ketone bodies. d Left: representative images of SA-β-gal (senescence-associated β-galactosidase) staining. Scale bar: 75 μm. Right: quantitative analysis of the SA-β-gal⁺ cells. Cells were treated with Hymeglusin (3 μM) for 4 days before SA-β-gal staining. n = 3 per group. e Left: representative images of γ-H2AX (phosphorylated H2AX) staining. Scale bar: 25 μm. Right: quantitative analysis of the γ-H2AX⁺ cells. Cells were treated with Hymeglusin (3 μM) for 4 days before staining. n = 3 per group. f Left: representative images of H2AY (histone protein macroH2A.1) staining. Scale bar: 25 μm. Right: quantitative analysis of the H2AY⁺ cells. Cells were treated with Hymeglusin (3 μM) for 4 days before staining. n = 3 per group. g Left: representative images of Lamin B staining. Scale bar: 25 μm. Right: quantitative analysis of the Lamin B⁺ cells. Cells were treated with Hymeglusin (3 μM) for 4 days before staining. n = 3 per group. h, i Quantitative RT-PCR analysis of senescence markers (p21, Pai1). Cells were treated with Hymeglusin (3 μM) for 4 days before analysis. n = 3 per group. j Progesterone production of MLTC-1 cells. Cells were exposed to Hymeglusin (3 μM) for 4 days before analysis. n = 3 per group. k Heat map of differentially expressed genes in Hymeglusin-treated and control groups. l Upregulated or downregulated GO terms of Hymeglusin-treated MLTC-1 cells, with the dot size indicating the range of −log10 (adjusted P-value). Color keys, ranging from grey to blue, indicate the absolute value of the NES. Data were presented as mean ± SEM. Significance was determined by Two-tailed t-test (b–j) or Two-sided permutation test with BH correction (l). n represents the number of biological replicates in (b–j). Illustrations were created with BioRender. Source data are provided as a Source Data file.

Fig. 4. Conditional knockout of Hmgcs2 in LCs leads to testicular aging in young mice.

a Schematic of AAV (adeno-associated virus) mediated Hmgcs2 conditional knockout. b Left: 4 weeks post-injection, representative testicular sections from AAV-GFP and AAV-Cre infected mice, stained with LCs marker HSD3β, HMGCS2, and DAPI. NC: negative control. Scale bar: 50μm. Right: quantitative analysis of HMGCS2⁺ HSD3β⁺ cells. n = 3 per group. c, d 4 weeks post-injection, testicular AcAc and BHB concentration in AAV-GFP and AAV-Cre infected mice. n = 4 per group in (c), n = 7 per group in (d). e Left: 2 months post-injection, representative testicular sections from AAV-GFP and AAV-Cre infected mice, stained with LCs marker CYP17A1, and SA-β-gal. Scale bar: 75 μm. Right: quantitative analysis of CYP17A1⁺ SA-β-gal⁺ cells. n = 3 per group. f Left: 2 months post-injection, representative testicular sections from AAV-GFP and AAV-Cre infected mice, stained with senescence marker p21, HSD3β, and DAPI. NC: negative control. Scale bar: 50 μm. Right: quantitative analysis of p21⁺ HSD3β⁺ cells. n = 3 per group. g, h 2 months post-injection, quantitative RT-PCR analysis of senescence markers (p21, Cxcl10) in AAV infected LCs. n = 3 per group. i–l Serum testosterone (i), intratesticular testosterone (j), serum Insl3 (Insulin-like peptide 3) (k), serum LH (Luteinizing Hormone) (l) level of the indicated groups. n = 10 per group in (i), n = 6 per group in (j), n = 13 per group in (k), n = 13 per group in (l). m–o 2 months post-injection, representative HE (hematoxylin-eosin) stained testicular sections obtained from AAV-GFP and AAV-Cre infected mice. Scale bar: 500 μm (m). The percentages of seminiferous tubules with varying diameters (n) and epithelial thickness (o). n = 3 per group. p Representative light micrographs of sperm acquired from indicated groups. Scale bar: 100 μm. q–s Sperm concentration (q), proportion of sperm with motility (r), proportion of sperm with progressive motility (s). n = 20 per group. Data were presented as mean ± SEM. Significance was determined by Two-tailed t-test. Illustrations were created with BioRender. Source data are provided as a Source Data file.

Fig. 5. BHB alleviates cellular senescence by promoting histone acetylation via HDAC inhibition.

a Experimental scheme of H₂O₂ (hydrogen peroxide) induced senescence, AcAc (5 mM) or BHB (5 mM) treatment and analysis. b Left: representative images of SA-β-gal staining in MLTC-1 cells of the indicated groups. Scale bar: 75 μm. Right: quantitative analysis of the SA-β-gal⁺ cells. n = 4 per group. c Left: representative images of senescence marker p21 in MLTC-1 cells of the indicated groups. Scale bar: 25μm. Right: quantitative analysis of the p21⁺ cells. n = 4 per group. d, e Quantitative RT-PCR analysis of senescence marker (p21, Cxcl10) in the indicated groups. n = 3 per group. f Progesterone production of MLTC-1 cells. n = 3 per group. g Right: Representative western blots for H3K9bhb and H3K9ac and Histone H3. Left: quantitative analysis of H3K9bhb and H3K9ac protein levels. Relative to Histone H3. MLTC-1 cells were treated with BHB (0, 1, 2, 3, 4, 5 mM) for 24 h before analysis. n = 3 per group. h Left: representative images of H3H9ac staining in MLTC-1 cells treated with BHB for 24 h. Scale bar: 20μm. Right: quantitative analysis of fluorescence intensity. n = 3 per group. i HDAC activity in MLTC-1 cells treated with BHB (5 mM) or vorinostat (1 μM). n = 3 per group. j Left: representative images of SA-β-gal staining in MLTC-1 cells of the indicated groups. Senescent cells were treated with BHB (5 mM) or vorinostat (1 μM) for 48 h before staining. Scale bar: 75μm. Right: quantitative analysis of the SA-β-gal⁺ cells. n = 3 per group. k–m Quantitative RT-PCR analysis of senescence markers (p16, p21, Cxcl10) in the indicated groups. Senescence cells were treated with BHB (5 mM) or vorinostat (1 μM) for 48 h before analysis. n = 3 per group. n Progesterone production of MLTC-1 cells. Senescent cells were treated with BHB (5 mM) or vorinostat (1 μM) for 48 h before analysis. n = 3 per group. Data were presented as mean ± SEM. Significance was determined by one-way ANOVA (b–g, j–n) or Kruskal-Wallis test (h). n represents the number of biological replicates in (b–g, j–n). Illustrations were created with BioRender. Source data are provided as a Source Data file.

Fig. 6. Enhancing ketogenesis in LCs of aged mice alleviates testicular aging.

a Schematic of AAV-mediated Hmgcs2 overexpression. b Left: 2 weeks post-injection, representative testicular sections from AAV-Control and AAV-O.E. infected mice, stained with LCs marker HSD3β, HMGCS2, and DAPI. NC: negative control. Scale bar: 50μm. Right: quantitative analysis of HMGCS2⁺ HSD3β⁺ cells. n = 3 per group. c Testicular BHB concentration 2 weeks post-injection. n = 5 per group. d Left: 2 weeks post-injection, representative western blots for H3K9ac and Histone H3 in testes of different mice. Right: quantitative analysis of H3K9ac protein level, relative to Histone H3. n = 3 per group. e Left: 2 weeks post-injection, representative western blots for FOXO3a and GAPDH in testes of different mice. Right: quantitative analysis of FOXO3a protein level, relative to GAPDH. n = 3 per group. f Left: 2 months post-injection, representative testicular sections from different mice stained with LC marker CYP17A1, and SA-β-gal. Scale bar: 75μm. Right: quantitative analysis of CYP17A1⁺ SA-β-gal⁺ cells. n = 3 per group. g Left: 2 months post-injection, representative testicular sections from different mice stained with senescence marker p21, HSD3β, and DAPI. Scale bar: 50μm. Right: quantitative analysis of p21⁺ HSD3β⁺ cells. n = 3 per group. h, i 2 months post-injection, quantitative RT-PCR analysis of senescence markers (p21, Cxcl10) in LCs. n = 3 per group. j–m Serum testosterone (j), intratesticular testosterone (k), serum Insl3 (l), serum LH (m) level of the indicated groups. n = 12 per group in (j), n = 4 per group in (k), n = 12 per group in (l), n = 7 per group in (m). n–p 2 months post-injection, representative HE stained testicular sections. Scale bar: 500 μm (n). The percentages of seminiferous tubules with varying diameters (o) and epithelial thickness (p). n = 3 per group. q–t Representative light micrographs of sperm from the indicated groups. Scale bar: 100 μm (q). Sperm concentration (r), proportion of sperm with motility (s), proportion of sperm with progressive motility (t). n = 10 per group. Data were presented as mean ± SEM. Significance was determined by one-way ANOVA (b–i, k, l, o–t) or Kruskal-Wallis test (j, m). Illustrations were created with BioRender. Source data are provided as a Source Data file.

Fig. 7. Oral administration of BHB partially alleviates testicular aging in aged mice.

a Schematic of oral administration of BHB and subsequent analysis. b Water and food consumption between salt and BHB-treated mice. n = 10 per group. c The concentration of BHB in serum and testes. n = 5 per group. d Left: the representative western blots for H3K9ac and Histone H3 in testes of different mice. Right: quantitative analysis of the H3K9ac protein level, relative to Histone H3. n = 3 per group. e Representative western blots for FOXO3a and GAPDH in testes of different mice. Right: quantitative analysis of the FOXO3a protein level, relative to GAPDH. n = 3 per group. f Left: representative images of testicular sections from indicated groups. The sections were stained with LCs marker CYP17A1, and SA-β-gal. Scale bar: 75 μm. Right: quantitative analysis of the CYP17A1⁺ SA-β-gal⁺ cells. n = 3 per group. g Left: representative confocal images of testicular sections from indicated groups. The sections were stained with senescence marker p21, HSD3β, and DAPI. Scale bar: 50 μm. Right: quantitative analysis of the p21⁺ HSD3β⁺ cells. n = 3 per group. h, i Quantitative RT-PCR analysis of senescence markers (p21, Cxcl10). n = 3 per group. j–m Serum testosterone (j), intratesticular testosterone (k), serum Insl3 (l), serum LH (m) level of the indicated groups. n = 20 per group in (j), n = 5 per group in (k), n = 12 per group in (l), n = 12 per group in (m). n–q Representative light micrographs of sperm acquired from indicated groups (n). Scale bar: 100 μm. Sperm concentration (o), proportion of sperm with motility (p), proportion of sperm with progressive motility (q). n = 10 per group. Data were presented as mean ± SEM. Significance was determined by one-way ANOVA (b-water consumed, c–h, k, o–q) or Kruskal-Wallis test (b-food consumed, i, j, l, m). Illustrations were created with BioRender. Source data are provided as a Source Data file.