Sex difference in BAT thermogenesis depends on PGC-1α-mediated phospholipid synthesis in mice

Abstract

Brown adipose tissue (BAT), a thermogenic tissue that plays an important role in systemic energy expenditure, has histological and functional sex differences. BAT thermogenic activity is higher in female mice than in male mice. However, the molecular mechanism underlying this functional sex difference has not been fully elucidated. Herein, we demonstrate the role and mechanism of PGC-1α in this sex difference. Inducible adipocyte-specific PGC-1α knockout (KO) mice display mitochondrial morphological defects and decreased BAT thermogenesis only in females. Expression of carbohydrate response-element binding protein beta (Chrebpβ) and its downstream de novo lipogenesis (DNL)-related genes are both reduced only in female KO mice. BAT-specific knockdown of ChREBPβ displays decreased DNL-related gene expression and mitochondrial morphological defects followed by reduced BAT thermogenesis in female wild-type mice. Lipidomics reveals that, PGC-1α increases ether-linked phosphatidylethanolamine (PE) and cardiolipin(18:2)₄ levels through Chrebpβ-dependent and -independent mechanisms in female BAT. Furthermore, PGC-1α enhances the sensitivity of female BAT estrogen signaling, thereby increasing Chrebpβ and its downstream DNL-related gene expression. These findings demonstrate that PGC-1α-mediated phospholipid synthesis plays a pivotal role in BAT thermogenesis in a sex-dependent manner.

Fig. 1. PGC-1α is highly expressed in BAT and is essential for maintaining cold tolerance in female mice.

a Pgc1a mRNA levels in BAT (n = 7 Male and n = 8 Female, p = 0.0001), inguinal WAT (iWAT) (n = 8 Male and Female), and gonadal WAT (gWAT) (n = 8 Male and n = 7 Female). b Western blots for PGC-1α proteins in BAT (left) and protein levels normalized to α-tubulin (right). n = 3 per group. p = 0.0124. c Pgc1a mRNA levels in BAT at 30 °C (n = 8 Male and n = 7 Female), 25 °C (n = 10 Male and Female, p = 0.0025), and 10 °C (n = 8 Male and Female, p = 0.0003). d Rectal temperature in Male Control (n = 7), Male KO (n = 8), Female Control (n = 5), and Female KO (n = 8) mice under acute cold exposure. p = 0.4146 (Male Control vs. Male KO), p = 0.0080 (Female Control vs. Female KO). Representative thermal images of interscapular surface (e) and interscapular surface temperature (f) in Male Control (n = 7), Male KO (n = 8), Female Control (n = 5), and Female KO (n = 8) mice under acute cold exposure. p = 0.5386 (Male Control vs. Male KO), p = 0.0103 (Female Control vs. Female KO). g Oxygen consumption (VO₂) recordings in response to NE in Male Control (n = 8), Male KO (n = 8), Female Control (n = 7), and Female KO (n = 8) mice. p = 0.2275 (Male Control vs. Male KO), p = 0.0214 (Female Control vs. Female KO). Data are expressed as the mean ± SEM. Data were analyzed using unpaired two-sided t-test (a–c) and two-way repeated measures ANOVA (d, f, g). Significance is indicated (*p < 0.05; **p < 0.01; ***p < 0.001). ns not significant. Source data are provided as a Source Data file.

Fig. 2. PGC-1α modulates mitochondrial membrane structure in female BAT cells.

a Representative electron micrographs of mitochondria from the BAT of Control and KO mice. The areas outlined in red boxes indicate enlarged images of mitochondria. Scale bars = 1 μm (n = 3 biologically independent experiments). b Histograms showing the distribution frequency (%) of mitochondrial section areas (0–3 µm²). c Total cristae length per mitochondrion. n = 30 per group. p = 0.2604 (Male Control vs. Male KO), p < 0.0001 (Male Control vs. Female Control), p < 0.0001 (Female Control vs. Female KO). d Representative blots for electron transport chain complexes (upper panel) and protein levels normalized to α-tubulin (lower panel); Complex I (n = 6 per group), Complex II (n = 6 Male Control, n = 6 Male KO, n = 5 Female Control, and n = 6 Female KO), Complex III (n = 6 per group), Complex IV (n = 4 Male Control, n = 6 Male KO, n = 5 Female Control, and n = 6 Female KO) and Complex V (n = 6 per group). p = 0.0014 (Complex I, Male Control vs. Female Control), p = 0.0157 (Complex I, Female Control vs. Female KO), p = 0.0275 (Complex IV, Male Control vs. Female Control). Data are expressed as the mean ± SEM. Data were analyzed by one-way ANOVA with Tukey’s post hoc test (c, d). Significance is indicated (*p < 0.05; **p < 0.01; ****p < 0.0001). ns not significant. Source data are provided as a Source Data file.

Fig. 3. PGC-1α regulates Chrebpβ gene expression by modulating chromatin accessibility in female BAT cells.

a Venn diagram showing the number of downregulated genes in Male and Female KO mice compared with Controls. b Gene Ontology (GO) analysis of genes downregulated only in Female KO mice. The top 10 enriched terms showing Metascape-generated enrichment p-values using cumulative hypergeometric distributions. c Gene expression of Chrebpβ (n = 10 per group), Acly (n = 10 per group), Acss2 (n = 10 per group), Fasn (n = 10 per group), Acaca (n = 8 Male Control, n = 9 Male KO, n = 8 Female Control, and n = 10 Female KO) and Elovl6 (n = 10 per group). p = 0.0400 (Chrebpβ), p = 0.0144 (Acly), p = 0.0405 (Acss2), p = 0.0297 (Fasn) for comparison between Female Control and Female KO. d Proportion of assay for transposase-accessible chromatin (ATAC) peaks unique to each group of Control and KO mice. e Venn diagram showing the number of genes with reduced peak counts on ATAC-seq and reduced gene expression on RNA-seq in Female KO mice compared with Control mice. f Gene Ontology analysis performed on the 93 genes commonly downregulated in both ATAC-seq and RNA-seq. g Chromatin accessibility near the transcription start site (TSS) of Chrebpβ in the BAT of Control and KO mice. Data are expressed as the mean ± SEM. Data were analyzed by one-way ANOVA with Tukey’s post hoc test. Significance is indicated (*p < 0.05). Source data are provided as a Source Data file.

Fig. 4. PGC-1α regulates TCA cycle metabolism in female BAT.

a Heat map illustrating differential metabolites based on metabolomic analysis in the BAT of Control and KO mice of both sexes. n = 5 per group. b Enrichment analysis of metabolic pathways using MetaboAnalyst. The x-axis indicates the pathway impact values derived from the pathway topology analysis, and the y-axis indicates the –log10(p-values) obtained from the pathway enrichment analysis using the hypergeometric test (one-sided). P-values were adjusted for multiple comparisons using the false discovery rate (FDR) method. c Score plot of principal component analysis (PCA) of all metabolites. Each point represents an individual sample. d Top 5 metabolites contributing to PC2 loading. e Concentration of TCA cycle metabolites in BAT of each group. n = 5 per group. p = 0.0082 and 0.0040 for acetyl-CoA; p = 0.6022 and 0.0389 for succinic acid; p = 0.1974 and 0.0445 for fumaric acid; p = 0.2551 and 0.0575 for malic acid; and p = 0.0497 and 0.0021 for NADH, comparing Male Control vs. Female Control and Female Control vs. Female KO, respectively. f Correlation between the levels of TCA cycle metabolites and the peak of VO₂ upon NE administration. Data are expressed as the mean ± SEM. Data were analyzed by one-way analysis of variance with Tukey’s post hoc test (e) or linear regression (f). Significance is indicated (*p < 0.05; **p < 0.01). Source data are provided as a Source Data file.

Fig. 5. PGC-1α changes the membrane phospholipid profiles of female BAT.

a Heat map illustrating differential lipid species based on lipidomic analysis in the BAT of Control and KO mice of both sexes. n = 5 per group. b Score plot of partial least squares (PLS) discriminant analysis of all lipids. Each point represents an individual sample. c Top 5 lipids contributing to PLS1 loading. d Relative amounts of total cardiolipin (CL). n = 5 per group. e Various molecular species of cardiolipin. Numbers above the bar graph represent the molecular species composition. n = 5 per group. f Percentage of CL(18:2)₄ in total CL. n = 5 per group. p = 0.0005 and 0.0129 for comparisons between Male Control vs. Female Control and Female Control vs. Female KO, respectively. Relative amounts of various molecular species of ether-linked phosphatidylethanolamine (PE) (g) and phosphatidylcholine (PC) (h). n = 5 per group. i Relative amounts of coenzyme Q (CoQ). n = 5 per group. p = 0.0002 and 0.0747 for PE-O(18:1/22:6), p = 0.0059 and 0.0123 for PE-O(18:2/18:1), p = 0.0008 and 0.0477 for PE-O(18:2/22:6), p = 0.0051 and 0.0064 for PC-O(16:1/18:2), p < 0.0001 and 0.0085 for PC-O(16:1/20:4), p = 0.0022 and 0.0149 for PC-O(16:1/22:6), and p = 0.0020 and 0.2054 for CoQ9, comparing Male Control vs. Female Control and Female Control vs. Female KO, respectively. Data are expressed as the mean ± SEM. Data were analyzed by one-way ANOVA with Tukey’s post hoc test (d–i). Significance is indicated (*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001). ns not significant. Source data are provided as a Source Data file.

Fig. 6. BAT-specific ChREBPβ knockdown in female mice recapitulates the metabolic phenotype of female PGC-1α knockout mice.

a Gene expression of Chrebpβ and DNL-related genes in female BAT injected with adeno-associated viruses (AAV) -shScramble or AAV-shChrebp. n = 6 per group. p = 0.0014 for Chrebpβ, p = 0.0001 for Acly, p = 0.0013 for Acss2, p < 0.0001 for Fasn, p = 0.0005 for Acaca, and p = 0.1527 for Elovl6. b Gene expression of Pgc1a. n = 6 per group. p = 0.0474. c Oxygen consumption (VO₂) recordings in response to NE. n = 7 per group. p = 0.0122. d Representative electron micrographs of mitochondria from BAT of AAV-Scramble and AAV-shChrebp mice. Scale bar = 1 μm (n = 3 biologically independent experiments). e Total cristae length per mitochondrion. n = 30 per group. p = 0.0271. f Percentage of CL(18:2)₄ in total CL. n = 5 per group. g Percentage of ether-linked PEs in total lipids. n = 5 per group. p = 0.0236, 0.0358, and 0.0395 for PE-O(16:1/16:1), PE-O(16:1/18:1), and PE-O(18:1/16:0), respectively. h D₂O-labeled components of ether-linked PEs. n = 3 per group. p = 0.3849, 0.0376, and 0.9923 for PE-O(16:1/16:1), PE-O(16:1/18:1), and PE-O(18:1/16:0), respectively. Data are expressed as the mean ± SEM. Data were analyzed using unpaired two-sided t-tests (a, b, e), paired one-sided t-tests (f–h), and two-way repeated measures ANOVA (c). Significance is indicated (*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001). Source data are provided as a Source Data file.

Fig. 7. Female BAT PGC-1α regulates systemic energy metabolism in coordination with estrogen signaling.

a Oxygen consumption (VO₂) recordings in response to NE in vehicle and tamoxifen (TMX) -treated mice. n = 4 per group. p = 0.4129 (Male Veh vs. Male TMX), p = 0.0168 (Female Veh vs. Female TMX). b Representative electron micrographs of mitochondria from BAT of control and TMX-treated mice. Scale bar = 1 μm. c Total cristae length per mitochondrion. n = 30 per group. p < 0.0001 for Male Veh vs. Female Veh, and p < 0.0001 for Female Veh vs. Female TMX. d Venn diagram showing the number of downregulated genes in male and female TMX-treated mice compared with controls. e Gene Ontology (GO) analysis of genes downregulated only in Female KO mice. The top 10 enriched terms are shown with enrichment p-values generated by Metascape using cumulative hypergeometric tests (one-sided). P-values were adjusted for multiple comparisons using the Benjamini-Hochberg procedure (FDR). f Gene expression of Chrebpβ and DNL-related genes in BAT. n = 5 per group. p = 0.0454 and 0.0225 for Acss2 (Male Veh vs. Female Veh and Female Veh vs. Female TMX, respectively); p = 0.0447 for Elovl6 (Female Veh vs. Female TMX). g Pgc1a mRNA levels in BAT. n = 5 per group. p = 0.0025 (Male Veh vs. Female Veh). h Gene expression in BAT explants from the BAT of Control and KO mice of both sexes treated with vehicle (n = 8 Male Control, n = 7 Male KO, n = 6 Female Control, and n = 8 Female KO) or 17β-estradiol (E2) (n = 8 Male Control, n = 7 Male KO, n = 7 Female Control, and n = 8 Female KO). p = 0.0465 for Female Control Veh vs. E2, and p = 0.0128 for Female Control E2 vs. Female KO E2 for Chrebpβ; p = 0.0339 for Female Control E2 vs. Female KO E2 for Acly; p = 0.0329 for Female Control E2 vs. Female KO E2 for Fasn. Data are expressed as the mean ± SEM. Data were analyzed by two-way repeated measures ANOVA (a) and one-way ANOVA with Tukey’s post hoc test (c, f–h). Significance is indicated (*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001). ns not significant. Source data are provided as a Source Data file.