BCAA insufficiency leads to premature ovarian insufficiency via ceramide-induced elevation of ROS

Abstract

Premature ovarian insufficiency (POI) is a disease featured by early menopause before 40 years of age, accompanied by an elevation of follicle-stimulating hormone. Though POI affects many aspects of women's health, its major causes remain unknown. Many clinical studies have shown that POI patients are generally underweight, indicating a potential correlation between POI and metabolic disorders. To understand the pathogenesis of POI, we performed metabolomics analysis on serum and identified branch-chain amino acid (BCAA) insufficiency-related metabolic disorders in two independent cohorts from two clinics. A low BCAA diet phenotypically reproduced the metabolic, endocrine, ovarian, and reproductive changes of POI in young C57BL/6J mice. A mechanism study revealed that the BCAA insufficiency-induced POI is associated with abnormal activation of the ceramide-reactive oxygen species (ROS) axis and consequent impairment of ovarian granulosa cell function. Significantly, the dietary supplement of BCAA prevented the development of ROS-induced POI in female mice. The results of this pathogenic study will lead to the development of specific therapies for POI.

Keywords: ROS; ceramide; infertility; low BCAA diet; premature ovarian insufficiency.

Figure 1. Metabolic disorders in POI patients

1. Clinical information of the cohort.

2. PCA of serum metabolomics. Circles indicate 95% confidence intervals. N = 18.

3. Heatmap showing the relative abundance of the top 75 differential expressed metabolites. N = 18.

4. The relative abundance of valine (Val) and leucine‐isoleucine (Leu‐Ile). N = 18; Boxplot, central band stands for median, boxes stand for 50% of the data, and whiskers stand for min or max of the data.

5. The relative abundance of 3‐hydroxybutyrate (3‐HOB) and acetylcarnitine (AC). N = 18; Boxplot, central band stands for median, boxes stand for 50% of the data, and whiskers stand for min or max of the data.

6. The concentration of FGF21 in serum. Control, N = 18; POI, N = 20; Boxplot, central band stands for median, boxes stand for 50% of the data, whiskers stand for min or max of the data.

7. The concentration of GDF15 in serum. Control, N = 14; POI, N = 9; Boxplot, central band stands for median, boxes stand for 50% of the data, and whiskers stand for min or max of the data.

Data information: Error bars stand for SEM of biological repeats. The P‐value was calculated by a two‐tailed t‐test with 2‐way ANOVA correction. Source data are available online for this figure.

Figure EV1. Metabolic details of POI patients

1. A

   Volcano plot showing the fold‐change and P‐value of metabolites. N = 18.

2. B

   Enrichment score and P‐value of metabolites downregulated in POI patients. N = 18.

3. C

   Relative abundance of valine (Val), acetyl‐carnitine (AC) but not leucine‐isoleucine (Leu‐Ile) in the serum of the Shandong Cohort. N = 10; Truncated violin plot, central band stands for median, and dotted lines stand for the upper quartile or the lower quartile of the data.

4. D

   The correlations between metabolites/FGF21 and clinical parameters. The P and r were calculated by the nonparametric Spearman test. N = 56.

5. E, F

   The concentration of FGF21 in mouse serum. (E) N = 5; (F) Control N = 10; Ketogenic diet, N = 6.

6. G, H

   The concentration of FSH in mouse serum. (G) Control, N = 7; Fgf21 OE, N = 6; and (H) control, N = 10; Ketogenic diet, N = 6.

Data information: Error bars stand for SEM. The P‐value was calculated by a two‐tailed t‐test with 2‐way ANOVA correction. Source data are available online for this figure.

Figure 2. Low BCAA diet induces POI‐like phenotypes in young female mice

1. A, B

   The serum concentration of FSH and the changes of follicles in mice on a low BCAA diet for 1.5 months. (A) Control, N = 10; low BCAA, N = 8; (B) Control, N = 4; low BCAA, N = 4.

2. C

   The serum concentration of FSH in mice on a different batch of the low BCAA diet. N = 5.

3. D

   The serum concentration of FSH in mice on the low BCAA diet from Research Diet. N = 10.

4. E, F

   The serum concentration of AMH and E2 in mice on the low BCAA diet for 1.5 months. Control, N = 5; low BCAA, N = 3.

5. G, H

   The serum concentration of AMH and E2 in mice on the low BCAA diet for 3 months. N = 5.

6. I, J

   The pregnancy rate and number of pups per pregnancy on the low BCAA diet for 3 months. N = 4.

Data information: S1, Primordial; S2, Primary; S3, Secondary; S4, Antral; S5, Atretic. Error bars stand for SEM of biological repeats. The P‐value was calculated by two‐tailed t‐test with 2‐way ANOVA correction. Source data are available online for this figure.

Figure 3. Elevation of ceramide impaired granulosa cell function

1. A, B

   Elevation of serum ceramide in POI patients from the Fudan Cohort or the Shandong Cohort. Left, the relative abundance of total ceramide; right, the relative abundance of ceramide with specific acyl chain. (A) N = 18; (B) N = 10; Truncated violin plot, central band stands for median, and dotted lines stand for the upper quartile or the lower quartile of the data.

2. C

   Heatmap showing the top 75 changed lipids in the serum of mice on a low BCAA diet. N = 10.

3. D

   Elevation of ceramide in the ovaries of mice on a low BCAA diet. N = 10; Truncated violin plot, central band stands for median, and dotted lines stand for the upper quartile or the lower quartile of the data.

4. E

   Decreases of E2 secretions from KGN cells treated with ceramide. N = 3.

5. F

   The serum concentration of FSH in mice with ceramide treatment. N = 7.

6. G

   Number of pups from mice with ceramide treatment. Control, N = 4; ceramide treatment, N = 5.

7. H

   The changes in follicle count from mice with ceramide treatment. N = 7.

Data information: S1, Primordial; S2, Primary; S3, Secondary; S4, Antral; S5, Atretic. Error bars stand for SEM of biological repeats. The P‐value was calculated by a two‐tailed t‐test with 2‐way ANOVA correction. Source data are available online for this figure.

Figure EV2. Lipid metabolism features of POI

1. A

   Heatmap showing the top 25 differentially abundant metabolites in the ovaries of mice on a low BCAA diet. Control, N = 10; low BCAA, N = 8.

2. B

   The relative abundance of BCAAs in the ovaries of mice on a low BCAA diet. Control, N = 10; low BCAA, N = 8; Boxplot, central band stands for median, boxes stand for 50% of the data, and whiskers stand for min or max of the data.

3. C

   GSEA of the proteomics data in the serum of the Fudan Cohort. N = 11.

4. D

   GSEA of RNA‐seq data from the liver in the mice on a low BCAA diet or control diet. N = 6.

5. E

   The PCA of serum from mice on a low BCAA diet or control diet. N = 10.

6. F

   The relative cell viability of KGN cells with ceramide treatment. N = 6.

7. G

   The percentage of Annexin V positive cells of KGN cells with ceramide treatment. N = 3.

8. H–J

   The pregnancy rate, body weight, and food intake from mice with ceramide treatment. N = 10.

9. K

   The relative abundance of ceramide in the serum of mice with myriocin treatment. N = 3; Truncated violin plot, central band stands for median, and dotted lines stand for the upper quartile or the lower quartile of the data.

10. L

    The serum concentration of FSH in mice with ceramide treatment. N = 8.

11. M

    The changes in follicle count from mice with ceramide treatment. N = 8.

Data information: S1, Primordial; S2, Primary; S3, Secondary; S4, Antral; S5, Atretic. Error bars stand for SEM. The P‐value was calculated by a two‐tailed t‐test with 2‐way ANOVA correction. Source data are available online for this figure.

Figure 4. BCAA‐induced POI via ceramide‐ROS axis

1. A

   GSEA results of RNA‐seq data. N = 5.

2. B

   sNuc‐seq data of ovarian nonimmune cells. Granulosa cells (cluster 0), mesothelial cells (cluster 1), thecal cells (cluster 2), and endothelial cells (cluster 3).

3. C, D

   (C) Relative expression of genes related to ROS in granulosa cells and (D) the concentration of E2 secreted by KGN cells with H₂O₂ treatment. N = 3.

4. E, F

   Relative abundance of GSH and GSSG. N = 3.

5. G

   The concentration of E2 secreted by KGN cells with ceramide treatment w/o NAC. N = 4.

6. H–J

   The serum concentration of FSH, AMH, and E2. (H) Control, N = 10; BCAA, N = 5; TG = 9; TG + BCAA = 8; (I) Control, N = 5; BCAA, N = 5; TG = 6; TG + BCAA = 6; (J) N = 5.

7. K

   The changes in follicle count. Control, N = 5; BCAA, N = 4; TG = 5; TG + BCAA = 5.

8. L

   Body weight. Control, N = 5; BCAA, N = 5; TG = 5; TG + BCAA = 4.

9. M

   Pregnancy rate. N = 4.

10. N

    Number of pups from mice with TG or TG + BCAA supplement treatment. N = 4.

Data information: Error bars stand for SEM of biological repeats. The P‐value was calculated by a two‐tailed t‐test with 2‐way ANOVA correction. Source data are available online for this figure.

Figure EV3. Low BCAA induces POI via elevation of ROS

1. A

   Genes with two‐fold changes and P‐value < 0.05 are highlighted in the volcano plot. N = 5.

2. B

   Upregulation of ROS‐related genes in KGN cells treated with ceramide. N = 5.

3. C

   Violin plots showing the expression of classical markers of nonimmune cells in ovaries from sNuc‐seq data.

4. D

   Relative activity of ROS‐related factors in patients. Control, N = 30; POI, N = 60.

Data information: Error bars stand for SEM. The P‐value was calculated by a two‐tailed t‐test with 2‐way ANOVA correction. Source data are available online for this figure.

Figure EV4. BCAA supplement protected the granulosa cells from ROS inducer

1. A

   Heatmap showing the relative abundance of metabolites in KGN cells treated with ceramides. N = 3.

2. B

   The concentration of E2 secreted by KGN cells with H₂O₂ treatment w/o BCAT2 inhibitor. N = 5.

3. C

   The concentration of E2 secreted by KGN cells with ceramide treatment w/o BCAT2 inhibitor. N = 5.

4. D

   Elevation of ceramide in the serum of TG‐treated mice. N = 5.

5. E

   BCAA supplement prevented the elevation of FSH in mice with 3‐NPA treatment. Control, N = 7; BCAA, N = 5; 3‐NPA, N = 9; 3‐NPA + BCAA, N = 16.

6. F

   BCAA supplement rescued the decrease of primordial follicles in mice with 3‐NPA treatment. Control, N = 3; BCAA, N = 3; 3‐NPA, N = 5; 3‐NPA + BCAA, N = 4.

Data information: S1, Primordial; S2, Primary; S3, Secondary; S4, Antral; S5, Atretic. Error bars stand for SEM. The P‐value was calculated by a two‐tailed t‐test with 2‐way ANOVA correction. Source data are available online for this figure.