BACH1 controls hepatic insulin signaling and glucose homeostasis in mice

Abstract

Hepatic insulin resistance is central to the metabolic syndrome. Here we investigate the role of BTB and CNC homology 1 (BACH1) in hepatic insulin signaling. BACH1 is elevated in the hepatocytes of individuals with obesity and patients with non-alcoholic fatty liver disease (NAFLD). Hepatocyte-specific Bach1 deletion in male mice on a high-fat diet (HFD) ameliorates hyperglycemia and insulin resistance, improves glucose homeostasis, and protects against steatosis, whereas hepatic overexpression of Bach1 in male mice leads to the opposite phenotype. BACH1 directly interacts with the protein-tyrosine phosphatase 1B (PTP1B) and the insulin receptor β (IR-β), and loss of BACH1 reduces the interaction between PTP1B and IR-β upon insulin stimulation and enhances insulin signaling in hepatocytes. Inhibition of PTP1B significantly attenuates BACH1-mediated suppression of insulin signaling in HFD-fed male mice. Hepatic BACH1 knockdown ameliorates hyperglycemia and improves insulin sensitivity in diabetic male mice. These results demonstrate a critical function for hepatic BACH1 in the regulation of insulin signaling and glucose homeostasis.

Fig. 1. BACH1 is elevated in the livers of individuals with obesity, patients with NAFLD, and obese mice.

a BTB and CNC homology 1 (BACH1) mRNA expression in the hepatocytes of lean (n = 10) individuals and individuals with obesity (n = 4) from the published single-cell transcriptome sequencing database (GEO accession no. GSE192742). b, c BACH1 mRNA expression (b) and protein expression (c) in the liver tissues from healthy subjects (Normal) and non-alcoholic fatty liver disease (NAFLD) patients (n = 6). d The representative images of hematoxylin and eosin (H&E), periodic acid-Schiff (PAS) staining, and immunohistochemistry assay of BACH1 from liver tissues of healthy subjects and NAFLD patients (scale bar = 100 μm, n = 6 individuals per group). e Quantitative data of immunohistochemistry assay of BACH1 in the liver tissues (n = 6). f Bach1 mRNA expression in liver cells from mice following a high-sugar diet (HSD) compared with chow diet (CD) from the published single-cell transcriptome sequencing database (GEO accession no. GSE182365). g, h Bach1 mRNA (n = 8) (g) and protein expression (n = 5) (h) in the liver tissues from male CD- and high-fat diet (HFD)-fed mice. i, j BACH1 protein expression in male ob/ob mice (n = 4) (i) and db/db mice (n = 3) (j). k Primary hepatocytes were treated with oleic acid (OA, 1 mM) for 12 h and then subjected to immunoblot analyses to determine the BACH1 protein expression (n = 3). Statistical analysis was performed by Wilcoxon Rank Sum test for (a) and (f), by two-tailed Mann–Whitney U test for (e), (g), and by unpaired two-tailed Student’s t-test for (b). Data are presented as mean values ± SD. Source data are provided as a Source Data file.

Fig. 2. Hepatic BACH1 deficiency improves obesity-induced insulin resistance and dysregulation of glucose homeostasis.

a Liver weight (left) and liver/body weight ratio(right) of male Bach1^LKO and Bach1^fl/fl mice fed a CD or HFD for 12 weeks were examined (n = 8 mice per group). Bach1^LKO: hepatocyte-deficient Bach1 mice, Bach1^fl/fl: the controls to Bach1^LKO mice. b Fasting blood glucose levels (left), fasting insulin levels (middle), and corresponding homeostasis model assessment of insulin resistance (HOMA-IR) index (right) of Bach1^LKO and Bach1^fl/fl mice after 12 weeks of CD or HFD treatment (n = 8 mice per group). c Glucose tolerance tests (GTT) were measured at week 12 of CD or HFD feeding (n = 8 mice per group). The area under the curve (AUC) was used to quantify the GTT results. d Insulin tolerance tests (ITT) were measured at week 12 of CD or HFD feeding (n = 8 mice per group). The AUC was used to quantify the ITT results. e Periodic acid-Schiff (PAS) staining of liver sections from Bach1^LKO and Bach1^fl/fl mice fed a CD or HFD for 12 weeks (scale bar = 100 μm, n = 8 mice per group). f A glycogen assay was carried out to determine the glycogen content in the livers of Bach1^LKO and Bach1^fl/fl mice (n = 8 mice per group). g mRNA expression of phosphoenolpyruvate carboxykinase1 (Pck1) and glucose-6-phosphatase catalytic subunit (G6pc) was measured in the livers of Bach1^LKO and Bach1^fl/fl mice with HFD feeding for 12 weeks by real-time quantitative PCR (qRT-PCR) (n = 6 mice per group). h Left: Western blot analysis of essential markers of insulin signaling (insulin receptor β (IR-β), AKT, forkhead box protein O1 (FOXO1), and glycogen synthase kinase-3β (GSK-3β)) in the liver tissues from Bach1^LKO and Bach1^fl/fl mice fed with HFD after injected i.p. with insulin after overnight fasting (n = 6 mice per group). Right: Phosphorylated protein levels were normalized to total protein. Statistical analysis was performed by two-way ANOVA followed by Tukey’s test for (a–d), and (h), by two-way ANOVA followed by Kruskal–Wallis test with Dunn multiple comparisons test for (f), by two-tailed Welch test for G6pc in (g), by unpaired two-tailed Student’s t-test for Pck1 in (g). Data are presented as mean values ± SD. Source data are provided as a Source Data file.

Fig. 3. Hepatic overexpression of BACH1 facilitates HFD-induced insulin resistance.

a Liver weight (left) and liver/body weight ratio (right) of male Bach1^LTG and N^TG mice fed a CD or HFD for 12 weeks (n = 8 mice per group). Bach1^LTG: hepatocyte-specific Bach1 overexpressed mice, N^TG: the controls to Bach1^LTG mice. b Fasting blood glucose levels (left), fasting insulin levels (middle), and HOMA-IR index (right) of Bach1^LTG and N^TG mice after 12 weeks of CD or HFD treatment (n = 8 mice per group). c GTTs were measured at week 12 of CD or HFD feeding (n = 8 mice per group). The area under the curve (AUC) was used to quantify the GTT results. d ITTs were measured at week 12 of CD or HFD feeding (n = 8 mice per group). The AUC was used to quantify the ITT results. e PAS staining of liver sections from Bach1^LTG and N^TG mice fed a CD or HFD for 12 weeks (scale bar = 100 μm, n = 8 mice per group). f A glycogen assay was carried out to determine the glycogen content in the liver tissues of Bach1^LTG and N^TG mice (n = 8 mice per group). g mRNA expression of Pck1 and G6pc was measured in the liver tissues of Bach1^LTG and N^TG mice with HFD feeding for 12 weeks by qRT-PCR (n = 6 mice per group). h Left: Western blot analysis of essential markers of insulin signaling (IR-β, AKT, FOXO1, and GSK-3β) in the liver samples from Bach1^LTG and N^TG mice fed with HFD after injected i.p. with insulin after overnight fasting (n = 6 mice per group). Right: Phosphorylated protein levels were normalized to total protein. Statistical analysis was performed by two-way ANOVA followed by Kruskal–Wallis test with Dunn multiple comparisons test for a (left), by two-way ANOVA followed by Tukey post hoc tests for (a) (right), (b–d), (f), and (h), by two-tailed Student’s t-test for (g). Data are presented as mean values ± SD. Source data are provided as a Source Data file.

Fig. 4. BACH1 inhibits insulin signaling in hepatocytes.

a Left: Western blot analysis of the phosphorylation and total protein levels of IR-β, AKT, FOXO1, and GSK-3β in the primary hepatocytes isolated from male Bach1^fl/fl and Bach1^LKO mice with or without insulin (100 nM) stimulation. Right: The phosphorylated protein levels normalized to total protein (n = 3 mice per group). b Left: Western blot analysis of the phosphorylation and total protein levels of IR-β, AKT, FOXO1, and GSK-3β in the primary hepatocytes isolated from N^TG and Bach1^LTG mice with or without insulin (100 nM) stimulation. Right: The phosphorylated protein levels normalized to total protein (n = 3 mice per group). c Left: Phosphorylation of key molecules of the insulin pathway was determined in the HepG2 cells infected with adenoviruses coding for BACH1 (AdBACH1) or GFP (AdGFP) and then stimulated with different concentrations of insulin for 10 min. Right: The phosphorylated protein levels normalized to total protein (n = 3 biological replicates). Statistical analysis was performed by two-way ANOVA followed by Tukey post hoc tests for (a–c). Data are presented as mean values ± SD. Source data are provided as a Source Data file.

Fig. 5. BACH1 interacts with protein tyrosine phosphatase 1B (PTP1B) and IR-β by the BTB domain.

a Left: Western blot analysis of the phosphorylation and total protein levels of IR-β, AKT, and GSK-3β in primary hepatocytes isolated from Bach1^LKO mice. Hepatocytes were transfected with vectors coding for HA-tagged versions of the full-length mouse Bach1 sequence (BACH1-HA) or mutant sequences lacking the Bzip domain (BACH1^ΔBzip-HA) or BTB domain (BACH1^ΔBTB-HA) with or without insulin (100 nM) stimulation for 10 min. Right: Phosphorylated protein levels were normalized to total protein (n = 3 biological replicates). b The coimmunoprecipitation (co-IP) assay of BACH1 and IR-β, PTP1B, phosphatase and tensin homolog (PTEN), and serine/threonine protein phosphatase 2 A (PP2A) in mouse primary hepatocytes. Cell lysates were immunoprecipitated with control mouse IgG or BACH1, and immunoblotting was used to detect IR-β, PTP1B, PTEN, PP2A, and BACH1 (n = 3 biological replicates). c Representative immunostaining of BACH1-GFP (green), IR-β-cherry (red), and PTP1B-D181A-BFP (purple) in HepG2 cells. Nuclei were stained with DAPI (blue) (Scale bar = 20 μm, n = 3 biological replicates). d, e Glutathione s-transferase (GST) pull-down assay. The lysate from HEK293T cells expressed IR-β-GFP (d) or expressed PTP1B-HA (e) were incubated with GST-tagged BACH1 protein (GST-BACH1). Immunoblotting was used to detect IR-β-GFP and GST-BACH1 (d) or PTP1B-HA and GST-BACH1 (e) (n = 3 biological replicates). f, g The co-IP analyses in HEK293T cells. HEK293T cells were transfected with two vectors, one coding for PTP1B-HA (f) or IR-β-GFP (g), and the other coding for Flag-tagged versions of the full Bach1 sequence or deletion sequences lacking the BTB domain. The BACH1 was immunoprecipitated with an anti-Flag antibody. PTP1B was detected in the precipitate with an anti-HA antibody (f), and IR-β was detected in the precipitate with an anti-GFP antibody (g) (n = 3 biological replicates). Statistical analysis was performed by one-way ANOVA followed by Tukey post hoc tests for a. Data are presented as mean values ± SD. Source data are provided as a Source Data file.

Fig. 6. BACH1 enhances the interaction between PTP1B and IR-β in response to insulin.

a The co-IP analyses of the interaction of IR-β and PTP1B in the liver lysates from HFD-fed male Bach1^LKO and Bach1^fl/fl mice. Mice were injected i.p. with insulin (100 nM) after overnight fasting. Liver lysates from HFD-fed Bach1^LKO and Bach1^fl/fl mice were immunoprecipitated with anti-PTP1B antibodies and subjected to immunoblot analysis with antibodies against IR-β (upper). The quantification data of IR-β protein levels was shown (lower) (n = 6 biological replicates). b Representative immunostaining of IR-β-cherry (red), PTP1B-D181A-BFP (purple), and GFP (green) in HepG2 cells with or without BACH1-GFP overexpression after being treated with 100 nM insulin for 10 min. Nuclei were stained with DAPI (blue) (upper). Colocalization rates were calculated and showed (lower) (Scale bar = 20 μm, n = 3 biological replicates). c Left: Western blot analysis of the phosphorylation and total protein levels of IR-β, AKT, GSK-3β, and FOXO1 in the presence or absence of insulin (100 nM, 10 min) in primary hepatocytes infected with AdBACH1 or AdGFP and then transfected with PTP1B-siRNAs. Right: Phosphorylated protein levels were normalized to total protein (n = 3 biological replicates). Statistical analysis was performed by two-way ANOVA followed by Tukey post hoc tests for (a) and (c). Data are presented as mean values ± SD. Source data are provided as a Source Data file.

Fig. 7. BACH1 aggravates insulin resistance through a PTP1B-dependent manner.

a Liver weights (left) and liver/body weight ratio (right) of male N^TG and Bach1^LTG mice injected with AAV-TBG-GFP or AAV-shPtpn1 and subjected to HFD challenge for 12 weeks (n = 8 mice per group). b Fasting blood glucose levels (left), fasting insulin levels (middle), and corresponding HOMA-IR index (right) of the above 4 groups of mice (n = 8 mice per group). c Left: GTTs were measured at week 12 of HFD feeding (n = 8 mice per group). Right: The AUC was used to quantify the GTT results. d ITTs were measured at week 12 of HFD feeding (n = 8 mice per group). Right: the AUC for ITTs. e, f PAS staining (e) and glycogen assay (f) were carried out to determine the glycogen content in the mouse livers. Representative PAS staining was shown in (e) (n = 8 mice per group, Scale bar = 100 μm). g Left: Western blot analysis of essential markers of the insulin pathway (IR-β, AKT, FOXO1, and GSK-3β) in the livers from the above 4 groups of mice after insulin administration. Right: Phosphorylated protein levels were normalized to total protein (n = 6 mice per group). Statistical analysis was performed by two-way ANOVA followed by Tukey post hoc tests for a, b, 0 min, and 90 min in (c), (d), (f), and (g), and by two-way ANOVA followed by Kruskal–Wallis test with Dunn multiple comparisons test for 120 min in (c). Data are presented as mean values ± SD. Source data are provided as a Source Data file.

Fig. 8. Knockdown of BACH1 ameliorates hyperglycemia and insulin resistance in db/db diabetic mice.

a Male db/db mice were injected with AAV control or AAV-shBach1 for four weeks, and the serum fasting blood glucose levels of db/db mice were examined (n = 8 mice per group). b GTTs were measured (n = 8 mice per group) (left). The AUC was used to quantify the GTT results (right). c ITTs were measured (n = 8 mice per group) (left). The AUC was used to quantify the ITT results (right). d The glucose infusion rate (GIR) was calculated for the last 40 min of insulin infusion (n = 5 mice per group). e Basal and clamped rates of hepatic glucose production (HGP) were calculated (n = 5 mice per group). f Suppression of HGP in db/db mice receiving either AAV control or AAV-shBach1 was determined (n = 5 mice per group). g A glycogen assay was carried out to determine the glycogen content in the livers of db/db mice receiving either AAV control or AAV-sh Bach1 (n = 5 mice per group). h mRNA expression of Pck1 and G6pc was measured in the livers of db/db mice receiving either AAV control or AAV-shBach1 by qRT-PCR (n = 6 per group). i Western blot analysis of essential markers of insulin signaling (IR-β, AKT, and GSK-3β) in the liver tissues from db/db mice receiving either AAV control or AAV-shBach1 after injected i.p. with insulin after overnight fasting (n = 3 per group). j Working model of the role of hepatic BACH1 in insulin signaling regulation. The increased BACH1 expression induced by excessive nutrient intake impaired insulin signaling and aggravated insulin resistance by facilitating the binding of PTP1B and IR-β. Statistical analysis was performed by two-tailed Student’s t-test for a, b, 0 min, and 90 min in (c), (d–f), and (h), by two-tailed Mann–Whitney U test for 120 min and 150 min in (c), and for (g). Data are presented as mean values ± SD. Source data are provided as a Source Data file. Figure 8j was created with Biorender under a paid subscription (agreement number: IZ263QDPCW).