Akt stimulates hepatic SREBP1c and lipogenesis through parallel mTORC1-dependent and independent pathways

Abstract

Through unknown mechanisms, insulin activates the sterol regulatory element-binding protein (SREBP1c) transcription factor to promote hepatic lipogenesis. We find that this induction is dependent on the mammalian target of rapamycin (mTOR) complex 1 (mTORC1). To further define the role of mTORC1 in the regulation of SREBP1c in the liver, we generated mice with liver-specific deletion of TSC1 (LTsc1KO), which results in insulin-independent activation of mTORC1. Surprisingly, the LTsc1KO mice are protected from age- and diet-induced hepatic steatosis and display hepatocyte-intrinsic defects in SREBP1c activation and de novo lipogenesis. These phenotypes result from attenuation of Akt signaling driven by mTORC1-dependent insulin resistance. Therefore, mTORC1 activation is not sufficient to stimulate hepatic SREBP1c in the absence of Akt signaling, revealing the existence of an additional downstream pathway also required for this induction. We provide evidence that this mTORC1-independent pathway involves Akt-mediated suppression of Insig2a, a liver-specific transcript encoding the SREBP1c inhibitor INSIG2.

Figure 1. mTORC1 activity is required for the insulin-stimulated induction of hepatic SREBP1c and de novo lipogenesis

A) Primary hepatocytes were serum starved overnight, pretreated for 30 min with vehicle or rapamycin (Rap; 20 nM), and stimulated for 6 h with insulin (Ins; 10 nM), as indicated. B) Hepatocytes were treated as in (A), and mRNA expression was measured by qRT-PCR. Transcript levels are shown relative to untreated controls. *P<0.006 (Srebp1c) and P<4×10⁻⁷ (Fasn) compared to unstimulated; **P<0.03 (Srebp1c) and P<0.002 (Fasn) compared to insulin-stimulated, vehicle-treated. C) Hepatocytes treated as in (A) were labeled with ¹⁴C-acetate and the levels of ¹⁴C incorporation into the lipid fraction are shown relative to untreated controls. *P<5×10⁻⁵ and **P<0.004 for the comparisons described in (B). Data are presented as the mean ± SEM from three (B) or four (C) independent experiments. (D,E) 8 week-old mice were fasted overnight (−) and then refed for 6 h (+) following a 30-min pre-treatment with vehicle or rapamycin (10mg/kg). N=4 mice per condition. (D) Liver lysates were immunoblotted for markers of mTORC1 signaling and full length (fl) and processed (p) SREBP1. (E) Gene expression was measured by qRT-PCR. Data are shown as mean ± SEM relative to untreated controls. *P<0.003 (Srebp1c), P<0.002 (Fasn) and P<0.0007 (Scd1) compared to fasted; **P<0.03 (Srebp1c), P<0.05 (Fasn), and P<0.0001 (Scd1) compared to fed, vehicle-treated.

Figure 2. Liver-specific Tsc1 deletion renders hepatic mTORC1 signaling constitutive and insulin independent

A) Schematic of the Tsc1^fl allele and resulting null allele following Cre-mediated excision. The position of the three PCR primers used for genotyping are shown with arrows. B) Representative PCR genotyping of liver tissue from the offspring of a cross between Tsc1^fl/fl and Albumin (Alb)-Cre Tsc1^fl/+ mice using the primers shown in (A). Genomic DNA from MEFs of the given genotype was used as controls for the three Tsc1 alleles. C) Immunoblots of liver, muscle, and fat tissue extracts from ad libitum-fed mice of the given genotypes are shown. D) Immunohistochemistry of phospho-S6 (S240/244) and staining with hematoxylin and eosin (H&E) were performed on adjacent liver sections from ad libitum-fed, 6 month-old mice of the given genotypes. E) 8 week-old Tsc1^fl/fl and LTsc1KO mice were fasted overnight (−) and then re-fed for 6 h (+), where indicated (N=3 mice per condition). Immunoblots of liver extracts from these mice are shown. F) Primary hepatocytes from Tsc1^fl/fl and LTsc1KO mice were serum-starved overnight and stimulated for 6 h with insulin (Ins; 10 nM) in the presence or absence of a 30-min pretreatment with rapamycin (Rap; 20 nM).

Figure 3. LTsc1KO mice are protected from diet-induced hepatic steatosis

A) A cohort of male Tsc1^fl/fl (n=3) and LTsc1KO (n=4) mice were weighed between the ages of 5 and 25 weeks, while being fed a high-fat diet (HFD) for the final 16 weeks. B) HFD-fed male Tsc1^fl/fl (n=6) and LTsc1KO (n=6) mice were subjected to DEXA scanning. C,D) Liver sections from 6-month old Tsc1^fl/fl (n=7) and LTsc1KO (n=7) mice fed a HFD, as in (A), were stained with H&E. (C) Representative sections from these mice are shown (scale bars: 500 µM (40X), 50 µM (200X). D) The results from blinded scoring of the H&E-stained liver sections for negative/mild, moderate, or severe steatosis are provided. E) Liver triglycerides (TG) were measured in ad libitum-fed Tsc1^fl/fl (n=5) and LTsc1KO (n=5) mice on a HFD, as in (A). TG levels are shown normalized to protein concentration and are presented as the mean ± SEM. *P<0.05 F,G) Expression of mRNA in the livers of ad libitum-fed Tsc1^fl/fl (n=4) and LTsc1KO (n=4) mice on a HFD, as in (A), was measured by qRT-PCR for the given genes, and levels are presented as mean ± SEM relative to Tsc1^fl/fl. *P<0.05 (Mcad).

Figure 4. LTsc1KO livers and hepatocytes have defects in the induction of SREBP1c expression and de novo lipogenesis

A) Expression of mRNA in the livers of male ad libitum-fed Tsc1^fl/fl (n=4) and LTsc1KO (n=4) mice on a HFD was measured by qRT-PCR for the given genes, and levels are presented as mean ± SEM relative to Tsc1^fl/fl. *P<0.005 (Srebp1c), P<0.05 (Fasn), and P<0.004 (Scd1). B) Immunoblots of full-length (fl) and processed (p) SREBP1, FASN, and SCD1 are shown in liver lysates from mice described in (A). (C, D) 8 week-old Tsc1^fl/fl and LTsc1KO mice were fasted overnight (−) with or without refeeding for 6 h (+), as indicated (N=4 per condition). (C) Liver lysates from Tsc1^fl/fl and LTsc1KO mice were immunoblotted for full length (fl) and processed (p) SREBP1 and its targets. The ratio of SREBP1(p) to SREBP1(fl) in the fasted and refed state was quantified for both genotypes and is presented as mean ± SEM relative to fasted wild-type samples. *P<0.01 for the fold induction by feeding compared to Tsc1^fl/fl livers. (D) RNA was isolated from the livers of these mice and gene expression was measured by qRT-PCR and is shown as mean ± SEM relative to fasted samples for each genotype. *P<0.0005 (Scd1) and P<0.003 (Fasn) for the fold induction by feeding compared to Tsc1^fl/fl livers. Note: There are no significant differences in fasted expression levels between Tsc1^fl/fl and LTsc1KO livers. E) Primary hepatocytes from Tsc1^fl/fl and LTsc1KO mice were serum starved overnight followed by insulin stimulation (Ins; 10 nM) for 6 h. Cells were labeled with ¹⁴C-acetate for the final 4 h, and its incorporation into the lipid fraction was measured. Data are presented as mean ± SEM relative to unstimulated wild-type samples. *P<0.01 for the fold induction by insulin compared to Tsc1^fl/fl hepatocytes. F) Expression of mRNA from primary hepatocytes treated as described in (E) was analyzed by qRT-PCR for the given genes, and levels are presented as mean ± SEM relative to the unstimulated samples for each genotype. *P<0.013 (Srebp1c) and P<0.023 (Fasn) for the fold induction by insulin compared to Tsc1^fl/fl hepatocytes. Data in (E) and (F) are representative of three independent experiments.

Figure 5. Attenuation of Akt signaling in LTsc1KO livers and hepatocytes is responsible for the defect in SREBP1c induction

A) Immunoblots of liver lysates from ad libitum-fed male Tsc1^fl/fl and LTsc1KO mice following a 16-week HFD are shown for the indicated phospho-(P) and total proteins. (B) 8 week-old Tsc1^fl/fl and LTsc1KO mice were fasted overnight (−) with or without refeeding for 6 h (+), as indicated, prior to immunoblot analysis of liver lysates. C) Primary hepatocytes from Tsc1^fl/fl (+) and LTsc1KO (−) mice were serum starved overnight and then stimulated with insulin (Ins; 10 nM) for 1 h, where indicated. D) Following overnight serum starvation, LTsc1KO hepatocytes were treated with rapamycin (Rap; 20 nM) for 30 min, where indicated, prior to insulin stimulation (10 nM) for 6 h. E) Primary hepatocytes from Tsc1^fl/fl and LTsc1KO mice were infected with adenovirus expressing either empty vector (Vec) or myr-Akt2 (Akt2), and 6 h post-infection, cells were serum starved overnight before lysis and immunoblotting. F) Following overnight serum starvation, uninfected cells were stimulated with insulin (10 nM) for 6 h, whereas cells infected as in (E) were left unstimulated. Cells were labeled with ¹⁴C-acetate for the final 4 h, and its incorporation into the lipid fraction was measured. Data are presented as mean ± SEM relative to unstimulated Tsc1^fl/fl samples for the uninfected group or to the vector-expressing Tsc1^fl/fl samples for the infected group for at least two independent experiments. *P<0.05 for the difference in fold induction by insulin compared to Tsc1^fl/fl hepatocytes. G) Primary hepatocytes were infected as described in (E), with vector-expressing cells being treated with insulin (10 nM) for 6 h, where indicated. Expression of mRNA was measured by qRT-PCR for the given genes, and levels are presented as mean ± SEM relative to the unstimulated vector-expressing group for each genotype. *P<0.027 (Srebp1c) and P<0.024 (Fasn) for the fold induction by insulin compared to Tsc1^fl/fl hepatocytes.

Figure 6. The mTORC1-independent pathway to SREBP1c involves Akt-mediated suppression of Insig2a

A) Expression of Insig2a in the livers of ad libitum-fed male Tsc1^fl/fl (n=4) and LTsc1KO (n=4) mice on a HFD was measured by qRT-PCR and levels are presented as mean ± SEM relative to Tsc1^fl/fl. *P<0.0002. B) 8 week-old mice were fasted overnight (−) and then refed for 6 h (+) following a 30-min pretreatment with vehicle, rapamycin (Rap; 10mg/kg) or Aktviii (Akti; 50mg/kg) (N=4 per condition). Hepatic RNA was isolated and gene expression was measured by qRT-PCR. Data are shown as mean ± SEM relative to fasted controls. *P<3×10⁻⁶ (Insig2a) and P<0.0002 (Srebp1c) compared to fasted; **P<0.01 and ***P<0.0008 compared to vehicle-treated fed mice. Note: There is no significant difference between the fold suppression of Insig2a by refeeding in the presence or absence of rapamycin or between Insig2a levels in fasted mice versus those fed in the presence of Aktviii. C) Tsc1^fl/fl primary hepatocytes were serum starved overnight and pretreated for 30 min with vehicle, rapamycin (20 nM), or Aktviii (10 µM) prior to 6 h stimulation with insulin (Ins; 10 nM), gene expression was measured by qRT-PCR. Data are presented as the mean ± SEM relative to unstimulated cells and are representative of two-independent experiments. Insig2a: *P<0.003 for insulin-stimulated suppression and P<0.01 for insulin-stimulated suppression in the presence of rapamycin. Note: There is no significant difference between the fold suppression of Insig2a by insulin in the presence or absence of rapamycin. Srebp1c: *P<0.0003 compared to unstimulated; **P<0.003 and ***P< 0.03 compared to vehicle-treated, insulin-stimulated samples. D) Primary hepatocytes from Tsc1^fl/fl and LTsc1KO mice were left uninfected or were infected with adenovirus expressing either empty vector (Vec) or myr-Akt2 (Akt2), and 6 h post-infection, cells were serum starved overnight. Uninfected cells were stimulated with insulin (10 nM) for 6 h, whereas infected cells were left unstimulated. Insig2a expression was measured by qRT-PCR. Data are presented as mean ± SEM relative to the unstimulated samples for the uninfected group or the vector-expressing samples for the infected group for each genotype and are representative of at least two independent experiments. *P<0.05 compared to the fold suppression by insulin in Tsc1^fl/fl cells. The fold suppression of Insig2a induced by Myr-Akt2 in Tsc1^fl/fl and LTsc1KO hepatocytes is not significantly different (NS). E) Tsc1^fl/fl (Tsc1 +) and LTsc1KO (Tsc1 −) primary hepatocytes were transiently transfected with control or Insig2 siRNA. 24-h post-transfection, cells were serum starved overnight followed by a 6 h stimulation with insulin (10nM). Gene expression was measured by qRT-PCR and levels are presented as mean ± SEM relative to unstimulated controls for each genotype. *P<0.04 (Insig2a) and *P<0.01 (Srebp1c) compared to the fold-change in response to insulin in Tsc1^fl/fl hepatocytes treated with control siRNAs. The insulin-stimulated induction of Srebp1c in control siRNA-treated Tsc1^fl/fl and Insig2 siRNA-treated LTsc1KO hepatocytes is not significantly different (NS). F) Model of two parallel pathways downstream of insulin and Akt2 required for the induction of hepatic SREBP1c and lipogenesis.