Effects of gonadal sex on lipid levels are abolished following statin treatment

To assess the effects of sex components on hypercholesterolemia, we made use of the apolipoprotein E (apoE)-deficient (Apoe–/–) mouse model, which is widely used in the study of hypercholesterolemia and atherosclerosis. Apoe–/– mice typically develop hypercholesterolemia (300–600 mg/dL) without dietary intervention, which is associated with elevated levels of LDL and VLDL, as well as reduced high-density lipoproteins (HDL) (reviewed in [21]). We generated Apoe–/– FCG mice on a C57BL/6 background (see Methods) to evaluate the role of gonadal and chromosomal sex in response to hypercholesterolemia and statin treatment (Fig. 1A, left panel). Throughout this study, statistical analyses of the FCG mice were performed by two-way ANOVA with gonadal sex and chromosomal sex as variables. The comparison between all mice with XX chromosomes (i.e., XX mice with ovaries plus XX mice with testes) and all mice with XY chromosomes (i.e., XY mice with ovaries plus XY mice with testes) allowed the detection of sex chromosome effects (Fig. 1A, middle panel). Similarly, comparison between all mice with ovaries (XX and XY with ovaries) and those with testes (XX and XY with testes) allowed the detection of gonadal effects. Since two genotypes are combined to generate groups for two-way ANOVA, the statistical power is also augmented. The analyses that we performed include plasma lipid levels and assessment of the liver transcriptome by RNA-seq and differential gene expression with pathway analyses (Fig. 1A, right panel).

Apoe–/– FCG mice were fed chow diet with or without simvastatin for 8 weeks. Prior to statin treatment, all four FCG genotypes were hypercholesterolemic (> 240 mg/dL) due to apo E deficiency. Mice with testes had higher cholesterol levels than those with ovaries, regardless of sex chromosome type (Fig. 1B, left). After statin treatment, there were no significant sex differences in the absolute cholesterol levels (Fig. 1B, middle). However, the sex genotype influenced the statin-induced change in cholesterol levels. Reductions in cholesterol levels occurred in a gonad-dependent manner with larger reductions in mice with testes compared to mice with ovaries (Fig. 1B, right). A significant interaction between gonad type and sex chromosome complement was identified by two-way ANOVA. In contrast to cholesterol levels, triglyceride levels did not vary across sex genotypes on chow or after statin treatment (Fig. 1C).

Free fatty acid levels in hypercholesterolemic mice were influenced by gonadal type, with higher levels in mice with testes than in mice with ovaries (Fig. 1D, left). After statin treatment, the effect of gonadal sex was not evident, but rather XX mice had higher fatty acid levels than XY mice (Fig. 1D, middle). The statin-induced changes in free fatty acid levels were affected by both gonad type and sex chromosome complement (Fig. 1D, right). With statin treatment, fatty acid levels were elevated in mice with ovaries and reduced in mice with testes; XX mice had higher levels than XY mice (Fig. 1D, right). Overall, statin altered plasma cholesterol and free fatty acid levels in a sex-dependent manner with the most robust effects in mice having XX chromosomes paired with ovaries or XY chromosomes paired with testes.

Gonadal and chromosomal sex independently impact the hepatic transcriptome in hypercholesterolemic mice

To identify genes and pathways that influence sex differences in hypercholesterolemia and statin response, we performed RNA-sequencing of liver from Apoe–/– FCG mice. To visualize the impact of gonads, sex chromosomes, and statin treatment on gene expression, we performed principal component analysis (PCA) of the RNA-seq data (Fig. 2A). The PCA generated separate clusters for gonadal type based on principal component 1, which accounted for 58% of the variation in the dataset. Within each gonadal type, the samples were separated into individual clusters for XX and XY sex chromosome complement based on principal component 2, which accounted for 23% of the variation within the dataset. Chow and statin treatments clustered together within each genotype indicating that statin treatment had a smaller impact on gene expression than sex genotype. We performed differential expression analysis with DESeq2 [15]; boxplots of the transformed counts show similar total counts across all 24 samples in the analysis (Fig. 2B).

Fig. 2

Characterization of Apoe–/– FCG liver RNA-seq dataset. A Principal component analysis of liver RNA-seq data from Apoe–/– FCG mice fed chow diet without or with simvastatin for 8 wks. B Boxplots representing transformed RNA-seq counts across all 24 liver samples. C The genome-wide distribution of genes found to be differentially regulated by presence of ovaries vs. testes (gonadal DEG), or by presence of XX vs. XY chromosomes (Sex Chr DEG). DEG, differentially expressed genes

We assessed the influence of sex genotypes on gene expression in the hypercholesterolemic state without statin treatment. A genome-wide analysis of differential gene expression (≥ 1.25-fold difference, adj P < 0.05) identified 3223 genes with differential expression in mice with ovaries compared to testes, and 1390 genes with differential expression in XX compared to XY mice (Additional file 2: Table S2). We found that 5.9–12.7% of genes on each autosome and 5.5% of genes on the X chromosome had different expression levels in mice with ovaries compared to mice with testes (Fig. 2C, blue bars). The XX vs. XY chromosome complement conferred differential expression of 3.7–7.3% of genes/autosome, as well as 3.8% of genes on the X chromosome (Fig. 2C, red bars). The majority of Y chromosome genes are expressed at very low levels in liver and were therefore not included in this analysis. This analysis reveals that gonadal and chromosomal sex each influence expression of genes that map across all autosomes as well as on the X chromosome.

We determined the functional classification of genes that were differentially expressed based on gonadal or chromosomal sex in the liver of hypercholesterolemic mice by pathway enrichment analysis [16,17,18]. The 1972 DEG with elevated expression in mice with ovaries compared to testes were enriched for genes in the immune system, cell cycle, and cell signaling pathways of the 49 significant pathways identified (Fig. 3A, B, Additional file 2: Table S2, Additional file 3: Table S3). The 1251 DEG elevated in the presence of testes were enriched in functions that include the unfolded protein response (UPR) and endoplasmic reticulum (ER) stress pathways (Fig. 3A, B, Additional file 2: Table S2, Additional file 3: Table S3). These results demonstrate that distinct hepatic cellular processes are influenced at the gene expression level by presence of ovaries or testes.

Fig. 3

Gonadal and chromosomal sex impact gene expression in Apoe–/– FCG liver. MA plot shows mean expression and fold-change of genes with differential expression in A mice with ovaries compared to mice with testes (adjusted P < 0.05, fold-change > 1.25) and B the top 10 significant cellular pathways enriched in the differentially expressed genes (adjusted P < 0.05). MA plot shows mean expression and fold-change of genes with differential expression in C XX mice compared to XY mice (adjusted P < 0.05, fold-change > 1.25) and D the top 10 significant cellular pathways enriched in the differentially expressed genes (adjusted P < 0.05)

We also assessed the functional enrichment of genes with differential expression in mice with XX (with ovaries or testes) vs. XY chromosomes (with ovaries or testes). The majority of DEG due to sex chromosome complement (1242 of 1390 genes) had elevated expression in XY compared to XX mice (Fig. 3C, Additional file 2: Table S2). The XY DEG were enriched in immune system, Rho GTPase signaling, and other cell signaling pathways (Fig. 3D, Additional file 3: Table S3). Only 148 genes were expressed at higher levels in XX compared to XY liver, but these were enriched for genes associated with fatty acid oxidation in mitochondria and other aspects of fatty acid metabolism (Fig. 3D, Additional file 3: Table S3).

Gonadally regulated genes in FCG mice extend beyond sexually determined growth hormone-regulated genes

Sex differences in hepatic gene expression have previously been attributed to growth hormone signaling, as gonad type drives sex differences in circulating levels of growth hormone through action on the pituitary. Male mice (XY testes) have low levels of circulating growth hormone with spikes at irregular frequencies while female mice (XX ovaries) have more frequent release of growth hormone leading to more consistent growth hormone-dependent signaling in the liver [22, 23]. In a study that assessed male/female differences in hepatic gene expression in CD-1 mice, 951 protein-coding genes were differentially expressed (adj P < 0.01). To identify genes that differ between the sexes due to the differential pattern of growth hormone release, growth hormone levels in male mice in that study were made to more closely resemble females by persistent growth hormone treatment for 4 days [19]. Of the original 951 genes with expression differences between males and females, the expression levels of 427 (44.9%) of these genes were altered after growth hormone treatment in male mice, suggesting that sex differences in growth hormone dynamics underlies about half of the observed male–female differences in hepatic gene expression in that model [19]. To identify growth hormone-dependent genes in our hypercholesterolemic FCG mice, we compared the 427 growth hormone-responsive genes identified by Lau-Corona et al. to the genes with differential expression in FCG mice with ovaries compared to mice with testes. Of the 427 genes, 309 genes (72.4%) were regulated by gonad type in FCG mice (adj P < 0.05) (Additional file 4: Table S4). Notably, however, our studies in hypercholesterolemic FCG mice identified an additional 2,554 genes with hepatic expression levels influenced by gonad type. 633 of the genes showing differential expression in FCG mice with ovaries compared to testes were previously found to be differentially expressed in mice with genetic deletion of the growth hormone receptor in liver (Additional file 4: Table S4) [20]. Thus, in our model, although sex differences in growth hormone may contribute to differential gene expression, additional mechanisms likely contribute to a majority of the gonad-type specific gene expression changes in our hypercholesterolemic mice.

ER stress response gene expression is elevated in mice with testes

Since ER stress-related pathways were the predominant pathways upregulated in mice with testes, we further evaluated the sex determinants of expression levels for specific genes within these pathways. The ER plays key roles in protein synthesis and trafficking as well as the synthesis of lipids, including sterols, phospholipids, and triglycerides. Disruption of ER homeostasis induces stress responses that are associated with the development and progression of metabolic diseases including obesity, fatty liver, type 2 diabetes and atherosclerosis [24, 25]. All protein-encoding genes in the unfolded protein response pathway (Reactome R-HSA-381119) [26] were evaluated by two-way ANOVA on CPM values from our RNA-seq data to identify genes impacted by gonadal effects and/or sex chromosome effects. Of the 92 protein-encoding genes, 36 genes were differentially expressed in mice with testes (XX testes plus XY testes) vs. ovaries (XX ovaries plus XY ovaries), and 7 genes were influenced by sex chromosome complement (XX ovaries plus XX testes compared to XY ovaries plus XY testes) (two-way ANOVA, P < 0.05) (Fig. 4A). We confirmed the expression patterns of representative genes via real-time PCR in additional mouse liver samples beyond those used in the RNA-seq. Consistent with RNA-seq data, the expression levels of Psmd14, Hyou1, Dnjab11, and Sec61a1 were upregulated in the livers of mice with testes compared to mice with ovaries (Fig. 4B). These data suggest that gonadal sex is a key determinant of sex differences in gene expression of ER stress-associated pathways in hypercholesterolemic liver, while chromosomal sex has only a minor impact.

Fig. 4

Genes in ER stress and UPR pathways show increased expression in mice with testes. A Heatmap displays relative hepatic expression levels of genes related to ER stress and unfolded protein response in Apoe–/– FCG genotypes (N = 3). Expression of each genotype was relative to that in XX mice with ovaries. B Real-time PCR quantification of representative ER stress genes from A (Psmd14, Hyou1, Dnjab11, and Sec61a1) in liver of Apoe–/– FCG mice. * Indicates gonadal sex effect and † indicates chromosomal sex effect. * or †P < 0.05 and **P < 0.01 by two-way ANOVA. N = 4–6. In two-way ANOVA analyses, two groups of mice were combined such that the number of mice in each group used for statistical analysis is 7–12

Fatty acid metabolism genes are independently regulated by gonadal and chromosomal sex

Our gene enrichment analyses demonstrated that gonadal and chromosomal sex each impact expression of lipid metabolism genes. In particular, XX vs. XY chromosomes influenced fatty acid metabolism and beta oxidation genes, and testes vs. ovaries influenced peroxisomal lipid metabolism (Fig. 3B, D; Additional file 3: Table S3). We further investigated the sex component regulation of genes involved in fatty acid synthesis and oxidation across the FCG genotypes. Of the 37 genes in the fatty acid synthesis pathway (Reactome R-HSA-75105.7), 21 genes were regulated by gonadal sex, sex chromosome complement, or a combination of both sex factors in livers of hypercholesterolemic mice (two-way ANOVA, *P < 0.05) (Fig. 5A). Gonadal sex influenced expression of 10 fatty acid synthesis genes. Elovl3 was strongly increased by the presence of testes, but the majority of the other 10 gonadally regulated fatty acid synthesis genes were down-regulated in mice with testes. The sex chromosome complement influenced expression of 7 fatty acid synthesis genes, with the majority upregulated in XY compared to XX mice. Four fatty acid synthesis genes showed more complex regulation, with significant effects of both gonadal and chromosomal sex.

Fig. 5

Gonadal and chromosomal sex regulate fatty acid metabolism. A Heatmaps display relative hepatic expression levels of genes related to fatty acid synthesis and peroxisomal beta oxidation in Apoe–/– FCG genotypes (N = 3). Expression of each genotype was relative to that in XX mice with ovaries. *P < 0.05 for gonadal or chromosomal sex by two-way ANOVA. Real-time PCR quantification of representative genes from A showing B chromosomal sex effects (Acsl1 and Acsl4), C gonadal sex effects (Scd2 and Phyh) and D both gonadal and chromosomal sex effects (Slc27a2). Two-way ANOVA with * donating gonadal sex effect and † denoting chromosomal sex effect. * or †P < 0.05 and ** or ††P < 0.01. N = 4–6. In two-way ANOVA analyses, two groups of mice were combined such that the number of mice in each group used for statistical analysis is 7–12

Peroxisomal fatty acid oxidation genes were also largely influenced by gonadal sex, but in a pattern distinct from the fatty acid synthesis genes. Of 29 peroxisomal lipid metabolism-related genes (Reactome R-HSA-390918.5), 12 were regulated by gonadal sex, primarily with increased expression due to testes compared to ovaries (Fig. 5A). Four of the 29 peroxisomal oxidation genes were expressed at lower levels in mice with XY compared to XX chromosomes.

We confirmed patterns of representative fatty acid metabolism genes identified by RNA-seq in a larger number of mice by real-time PCR. The acyl-CoA synthetase genes Acsl1 and Acsl4 were both regulated by sex chromosome complement, but in opposing directions: XX chromosomes promoted higher expression of Acsl1, while XY chromosomes promoted higher expression of Acsl4. (Fig. 5B). Scd2 and Phyh were oppositely regulated by gonadal sex, with ovaries promoting higher Scd2 expression, and testes enhancing Phyh expression (Fig. 5C). Slc27a2 exhibited combined gonadal and sex chromosome regulation with higher expression in XX mice compared to XY mice, but also higher expression in mice with testes compared to mice with ovaries (Fig. 5D). Overall, fatty acid metabolism exhibits complex sex-dependent regulation with fatty acid biosynthesis and peroxisomal degradation impacted in opposite directions by gonadal and chromosomal sex.

Sex-dependent transcriptional response to statin treatment is influenced by XY sex chromosome complement and testes

To identify sex-dependent effects of statin, Apoe–/– FCG mice received simvastatin in chow diet for 8 weeks. The role of sex components on statin-induced alterations in the hepatic transcriptome were identified by comparing statin-treated mice to chow diet controls within each genotype. Statin treatment influenced a similar number of genes in mice with testes (949 DEG between statin and chow) and ovaries (939 DEG between statin and chow) (Fig. 6A). However, only ~ 5% (94) of the genes altered by statin treatment were shared by mice with ovaries and testes, indicating distinct target genes for statin-associated regulation depending on gonadal type. We analyzed genes with up and down regulation in response to statin (Fig. 6B, C) for functional enrichment. In mice with testes, statin down-regulated genes were enriched for fatty acid metabolism pathways, and statin upregulated genes were enriched for cholesterol biosynthesis pathways (Fig. 6D). In mice with ovaries, statin-regulated genes did not show enrichment for specific biological pathways (Fig. 6E).

Fig. 6

Statin effects on hepatic gene expression are influenced in distinct ways by gonadal and chromosomal sex. A Venn diagram shows number of genes altered by statin treatment depending on presence of ovaries or testes, and the minimal overlap between the two. MA plots display mean expression levels and fold-change of differentially expressed genes (DESeq2, > 1.25-fold altered expression, P < 0.05) in B statin-treated mice with testes compared to chow diet controls and C statin-treated mice with ovaries compared to chow diet controls. Significantly enriched cellular pathways for statin-induced gene expression changes in D mice with testes and E mice with ovaries (no enriched pathways). F Venn diagram shows number of genes altered by statin treatment depending on presence of XX or XY chromosomes, and the minimal overlap between the two. MA plots display mean expression levels and fold-change of differentially expressed genes (DESeq2) in G statin-treated XY mice compared to chow diet controls and H statin-treated XX mice compared to chow diet controls. Significantly enriched cellular pathways for statin-induced gene expression changes in I XY mice and J XX mice (no enriched pathways)

Statin-induced gene regulation was also influenced by chromosomal sex. Compared to mice fed a chow diet, statin treatment altered expression of 605 genes in XY mice and 612 genes in XX mice, but only 39 (~ 6%) of the dysregulated genes were shared between the two sex chromosome complement types (Fig. 6F). In mice with XY sex chromosomes, statin treatment upregulated 162 genes and down-regulated 442 genes. In mice with XX sex chromosomes, statin treatment upregulated 391 genes and down-regulated 221 genes (Fig. 6G,H). For XY mice, statin upregulated genes were enriched for cholesterol biosynthesis pathways, while XX mice had no significant pathway enrichment for the statin-induced DEG (Fig. 6I,J).

Statin treatment induces cholesterol biosynthesis gene expression only in mice with XY sex chromosomes or testes

Statin drugs bind to the rate-limiting enzyme in cholesterol biosynthesis (HMG CoA reductase) to inhibit its activity. The complex feedback mechanisms that control HMGCoA reductase levels can lead to enhanced transcription of the corresponding gene, as well as other genes within the pathway, in response to statin [27, 28]. The subsequent upregulation of HMG CoA reductase protein levels may reduce the effectiveness of statin drugs. Our data above suggested that statin-induced compensatory upregulation of cholesterol synthetic gene expression is driven by testes and XY chromosome complement. We investigated further by assessing the effect of gonadal and chromosomal sex on expression of genes in the cholesterol biosynthetic pathway in response to statin. Statin upregulated 18 of 25 cholesterol biosynthetic genes in the liver of XY mice, with little or no increase in XX liver (Fig. 7A). The increased expression in XY mice was most pronounced in XY mice with testes, suggesting an additional impact from gonadal sex.

Fig. 7

Cholesterol biosynthesis pathway gene expression is upregulated in XY mice in response to statin. A Heatmap displays relative hepatic expression levels of genes altered by sex within the cholesterol biosynthesis pathway as fold-change of statin treated compared to chow diet for each of the Four Core genotypes (N = 3 per genotype and diet). * indicates gonadal sex effect and † indicates chromosomal sex effect. * or †P < 0.05 by two-way ANOVA. B–E Real-time PCR quantification of representative genes from A illustrates the effects of gonadal and/or chromosomal sex on statin-induced alterations in gene expression in the cholesterol biosynthesis pathway. Two-way ANOVA with * denoting gonadal sex effect and # denoting effect due to statin treatment. * or #P < 0.05 and ** or ##P < 0.01. N = 4–6. In two-way ANOVA analyses, two groups of mice were combined such that the number of mice in each group used for statistical analysis is 7–12

We confirmed the sex chromosome differences in statin regulation of representative cholesterol synthetic genes by real-time PCR. Mice with XY sex chromosomes upregulated Hmgcs, Hmgcr, Mvk, and Sqle with statin treatment compared to chow diet controls, while mice with XX sex chromosomes did not (Fig. 7B–E). Hmgcr expression also showed a significant gonad effect with higher expression in mice with ovaries compared to testes, irrespective of sex chromosome type and statin treatment (Fig. 7C). Our results reveal that sex differences in statin-regulated gene expression depend on the individual gene, and specific genes may be influenced by gonadal sex, chromosomal sex, or a combination of the two.