TAK1-mediated autophagy and fatty acid oxidation prevent hepatosteatosis and tumorigenesis

The MAP kinase kinase kinase TGFβ-activated kinase 1 (TAK1) is activated by TLRs, IL-1, TNF, and TGFβ and in turn activates IKK-NF-κB and JNK, which regulate cell survival, growth, tumorigenesis, and metabolism. TAK1 signaling also upregulates AMPK activity and autophagy. Here, we investigated TAK1-dependent regulation of autophagy, lipid metabolism, and tumorigenesis in the liver. Fasted mice with hepatocyte-specific deletion of Tak1 exhibited severe hepatosteatosis with increased mTORC1 activity and suppression of autophagy compared with their WT counterparts. TAK1-deficient hepatocytes exhibited suppressed AMPK activity and autophagy in response to starvation or metformin treatment; however, ectopic activation of AMPK restored autophagy in these cells. Peroxisome proliferator–activated receptor α (PPARα) target genes and β-oxidation, which regulate hepatic lipid degradation, were also suppressed in hepatocytes lacking TAK1. Due to suppression of autophagy and β-oxidation, a high-fat diet challenge aggravated steatohepatitis in mice with hepatocyte-specific deletion of Tak1. Notably, inhibition of mTORC1 restored autophagy and PPARα target gene expression in TAK1-deficient livers, indicating that TAK1 acts upstream of mTORC1. mTORC1 inhibition also suppressed spontaneous liver fibrosis and hepatocarcinogenesis in animals with hepatocyte-specific deletion of Tak1. These data indicate that TAK1 regulates hepatic lipid metabolism and tumorigenesis via the AMPK/mTORC1 axis, affecting both autophagy and PPARα activity.

TGFβ-activated kinase 1 (TAK1) is a MAP kinase kinase kinase (MAP3K) activated by TLRs, IL-1 receptor, TNF receptor 1, and TGFβ receptors (1, 2). In hepatocytes, TAK1 stimulates activation of IKK-NF-κB and JNK, which are crucial regulators of cell survival, proliferation, and tumorigenesis, as well as lipid metabolism and insulin sensitivity (3). IKK/NF-κB signaling prevents TNF- and ROS-mediated hepatocyte death, hepatic steatosis, and hepatocarcinogenesis (4). In contrast, sustained JNK activation promotes ROS generation, resulting in hepatocyte death, lipid accumulation, and development of hepatocellular carcinoma (HCC) (3). Since TAK1 controls activation of both IKK/NF-κB and JNK pathways, its role in liver pathophysiology has been hard to predict. We and others have previously demonstrated that ablation of Tak1 in hepatocytes results in spontaneous hepatocyte death, inflammation, fibrosis, and HCC development; these phenotypes depends on TNF and TGFβ receptor signaling (5, 6) and are more dramatic than the phenotypes observed in mice with hepatocyte-specific deletion of IKKβ or IKKγ/NEMO (5, 7–9).

TAK1 can also regulate AMPK activity through phosphorylation (10). Nutrient deprivation strongly activates AMPK, leading to the inhibition of mTOR complex 1 (mTORC1), a multifunctional protein kinase complex that regulates lipid biosynthesis, cellular proliferation, and autophagy (11, 12). AMPK can also stimulate autophagy through direct phosphorylation of ULK1 independently of mTORC1 (13). Under high nutrient conditions, when ATP levels are high, AMPK activity is inhibited, thereby activating mTORC1 and resulting in increased lipid synthesis in a SREBP-1c– and PPARγ-dependent manner (12). Moreover, mTORC1 inhibits PPARα activity, which regulates mitochondrial functions and fatty acid β-oxidation (FAO) (14). Hepatic FAO is impaired in mice with inactivated PPARα, SIRT1, or SIRT3 (15–17). These mice exhibit significant lipid deposition in the liver upon consuming a high-fat diet (HFD) or during fasting. AMPK activation and mTORC1 inhibition induce autophagy to remove and recycle cellular materials for biosynthesis or energy production when nutrients are limited. Autophagy promotes lipid degradation and prevents excessive lipid accumulation (18). It was suggested that TAK1 might contribute to the induction of autophagy through either the IKK complex or AMPK (10, 19). However, the physiological and pathophysiological significance of TAK1-dependent regulation of AMPK/mTORC1 signaling and autophagy and their involvement in lipid metabolism and HCC development in the liver remain elusive.

In the present study, we determined that TAK1 in hepatocytes prevents excessive lipid accumulation through AMPK activation, mTORC1 inhibition, and autophagy. TAK1 also favors PPARα-mediated FAO through the inhibition of mTORC1. Upon HFD feeding, TAK1 activity prevents excessive hepatic lipid accumulation, injury, and inflammation. Correspondingly, restoration of autophagy with the mTORC1 inhibitor rapamycin suppresses hepatic steatosis, fibrosis, and HCC progression in Alb-Cre Tak1^fl/fl (Tak1^Δhep) mice.

Nutrient deprivation causes severe lipid accumulation in Tak1-deleted livers. To determine the physiological role of TAK1 in the liver, we investigated its involvement in the response to fasting, which causes mobilization of lipids from peripheral depots into the liver (12, 20). Initially, livers appeared normal in 1-month-old Tak1^Δhep mice and their WT counterparts, but after 12 hours of fasting, livers of Tak1^Δhep mice became white. Tak1^Δhep mice showed increased lipid accumulation in hepatocytes compared with that seen in their WT counterparts, particularly after fasting (Figure 1, A and B). Hepatic triglycerides (TGs) in fasted Tak1^Δhep mice were 3-fold higher than those in fasted WT mice, whereas plasma FFAs were similarly elevated in both fasted Tak1^Δhep and WT mice (Figure 1, C and D). These data indicate that although FFA release from adipose tissue is not affected by loss of hepatic TAK1 expression, the deposition of TGs in the liver is TAK1 dependent. In WT livers, nutrient deprivation strongly inhibited S6 phosphorylation, an established marker of mTORC1 activity (Figure 1E). In contrast, Tak1^–/– livers showed marked increases in S6 phosphorylation before or after nutrient deprivation (Figure 1E), suggesting aberrant activation of mTORC1.

Figure 1

TAK1 deficiency causes severe fatty liver after fasting. One-month-old WT and Tak1^Δhep mice were fasted for 12 hours (n = 5 each). (A) Macroscopic appearance of livers of WT and Tak1^Δhep mice before and after fasting. (B) Oil Red O staining of lipid droplets. Original magnification, ×200. (C) Hepatic TG content and (D) serum levels of FFAs were measured. (E) Immunoblotting for total TAK1, p-S6, and total S6. Black bar, WT; gray bar, Tak1^Δhep mice. Data are presented as the means ± SEM. *P < 0.05; **P < 0.01.

AMPK activation and autophagy are inhibited in Tak1^–/– hepatocytes. Reduced p62/SQSTM1 expression, increased LC3B-II generation from LC3B-I, and formation of LC3B aggregates are markers of autophagy induction (21–23). We found that fasted WT livers exhibited lower p62 expression and higher LC3B-II amounts and accumulation of LC3B dots than did fed WT livers (Figure 2, A and B, and Supplemental Figure 1, A and B; supplemental material available online with this article; doi:10.1172/JCI74068DS1). In contrast, Tak1^–/– livers exhibited much higher p62 expression and no LC3B-II expression or less LC3B aggregates than did WT livers (Figure 2, A and B, and Supplemental Figure 1, A and B). Furthermore, fasting did not result in any increase in autophagy markers in Tak1^–/– livers, indicating that TAK1 is required for stimulation of liver autophagy during fasting. Chloroquine (CQ) treatment prevents autophagosome degradation, thereby inducing LC3B aggregation during fed states (23). In WT livers, CQ treatment increased LC3B aggregation, but LC3B aggregates were much less pronounced in Tak1^–/– livers before or after CQ treatment (Supplemental Figure 1F). In LC3B-GFP transgenic mice, the detection of free GFP released from an LC3B-GFP fusion protein by autophagosome degradation reflects autophagic flux (24). We found that free GFP was increased after fasting in WT LC3B-GFP transgenic livers; however, free GFP formation was insignificant in Tak1^Δhep-LC3B-GFP transgenic livers before and after fasting (Supplemental Figure 1G). EM showed that there were some lipid droplets with attached autophagic vesicles (AVs) in fed WT livers and that the number of such droplets was increased upon fasting (Figure 2C). However, lipid droplets with attached AVs were rare in Tak1^Δhep livers (Figure 2C). These results demonstrate that both basal and induced autophagy are suppressed in Tak1^–/– livers. In addition, expression of autophagy genes including Ulk1, Ulk2, Atg3, Atg5, and Atg8/Lc3b was suppressed in Tak1^Δhep mice, as revealed by microarray analysis and quantitative real-time PCR (Supplemental Table 1 and Supplemental Figure 2), further supporting the requirement of TAK1 for basal and induced autophagy in the liver.

Figure 2

Defective autophagy inTak1^Δhep mice. (A–C) One-month-old WT and Tak1^Δhep mice were fasted for 12 hours (n = 5 each). (A) Immunoblotting for hepatic expression of p62 and LC3B. (B) Immunohistochemistry for LC3B and quantification. LC3B aggregation was observed in fasted WT mice. Original magnification, ×1,000. (C) EM and its quantification. Autophagosomes associated with lipid droplets are indicated with arrows. Scale bars: 1 μm. Lipid droplets (LD); autophasic vacuoles (AV); ND, not detected. (D) Primary hepatocytes were isolated from WT mice. TAK1 kinase activity was analyzed after 1 or 2 hours of amino acid starvation. (E) Primary hepatocytes were isolated from WT and Tak1^Δhep mice. Expressions of TAK1, p-LKB1, LKB1, p-AMPK, AMPK, p-raptor, raptor, p-ULK1, and ULK1 after 2 and 6 hours of amino acid starvation were assessed by immunoblotting. (F) Primary hepatocytes were isolated from LC3B-GFP transgenic mice and Tak1^Δhep-LC3B-GFP transgenic mice. After 6 hours of amino acid starvation, LC3B-GFP accumulation in the cytoplasm was assessed (top). Original magnification, ×200. Black bar, WT; white bar, Tak1^Δhep mice. Immunoblotting for LC3B-GFP and p62 and quantification of the ratio of LC3B-II GFP to LC3B-I GFP (bottom). Similar results were obtained in 3 independent experiments. Data are presented as the means ± SEM. *P < 0.05; **P < 0.01.

Next, we examined TAK1 kinase activity upon starvation. TAK1 kinase activity was increased 1 hour after starvation (Figure 2D). Starvation also induced the phosphorylation of LKB1, AMPK, raptor, and ULK1 and increased the conversion of LC3B-I to LC3B-II and the accumulation of LC3B-GFP aggregates, while reducing the amount of p62 in WT hepatocytes (Figure 2, E and F). All of these responses were highly attenuated in Tak1^–/– hepatocytes (Figure 2, E and F), indicating that TAK1 is required for LKB1 and AMPK phosphorylation and induction of autophagy during starvation. Similarly, we observed that TAK1 kinase activity was increased upon metformin treatment (Supplemental Figure 3A), and metformin-induced AMPK activation and autophagy were suppressed in Tak1^–/– hepatocytes (Supplemental Figure 3, B–D). Accordingly, we examined whether ectopic AMPK activation can restore autophagy in Tak1^–/– hepatocytes. Overexpression of active AMPK increased LC3B-I-to-LC3B-II conversion and decreased p62 in Tak1^–/– hepatocytes, indicating that AMPK is an important downstream effector of TAK1 that mediates its effect on autophagy (Supplemental Figure 3E).

Tak1 deficiency causes defects in PPARα-mediated FAO in hepatocytes. To investigate the role of TAK1 in hepatic lipid metabolism, we assessed expression of genes related to FAO and lipid metabolism. The mRNA levels of Ppara, Hmgcs2, Cpt1a, Acox1, Cyp4a10, and Sirt3 in fed WT livers were higher than in fed Tak1^–/– livers (Figure 3A). Furthermore, these mRNAs were upregulated after 12 hours of fasting in WT livers, but their expression in Tak1^–/– livers remained lower than in WT livers, even after fasting (Figure 3A). Hepatic FAO was also elevated after fasting in WT mice, and it was suppressed in Tak1^–/– livers both before and after fasting (Figure 3B). Likewise, WT livers expressed higher levels of ApoB and Mttp mRNA than did Tak1^–/– livers both before and after fasting and had higher TG secretion rates (Figure 3, C and D). These results suggest that reduced FAO and impaired TG secretion contribute to increased hepatic steatosis in Tak1^Δhep mice after fasting.

Figure 3

Gene expression involved in FAO and induction of FAO are suppressed in Tak1 ^Δhep mice. One-month-old WT and Tak1^Δhep mice were fasted for 12 hours (n = 5 each). (A) Hepatic expression of FAO-related genes (Ppara, Hmgcs2, Cpt1a, Acox1, Cyp4a10, and Sirt3) was measured by quantitative real-time PCR. (B) FAO was measured in WT and Tak1^Δhep mice in vivo. 5 μCi [¹⁴C] palmitate was injected into mice with or without fasting, and hepatic FAO was measured. FAO is shown as the fold change compared with WT fed controls. (C) mRNA expression of ApoB and Mttp was measured by quantitative real-time PCR. (D) VLDL secretion rate was measured. Black bar, WT; gray bar, Tak1^Δhep mice. Data are presented as the means ± SEM. *P < 0.05; **P < 0.01.

Because Tak1 deficiency caused overactivation of mTORC1 (Figure 1E) and mTORC1 negatively regulates PPARα signaling (14), we expected that PPARα activity would be suppressed in Tak1^Δhep mice. Microarray transcriptome profiling including gene ontology and KEGG pathway analysis revealed that the PPAR signaling and fatty acid metabolism pathways were in fact the 2 major pathways that were suppressed in the livers of Tak1^Δhep mice (P = 3.53 × 10^–6 and P = 2.4 × 10^–7, respectively) (Supplemental Tables 2 and 3). Forty-eight PPARα target genes were reported to be involved in FAO and fatty acid metabolism (16, 25, 26). We found that expression of 28 (after Student’s t test) and 14 (after Bonferroni correction) of these genes was significantly decreased in Tak1^–/– livers compared with that in WT livers (Supplemental Figure 4 and Supplemental Table 4). This prompted us to further examine PPARα-regulated genes using quantitative real-time PCR. Ppara mRNA expression itself was reduced in Tak1^–/– hepatocytes (Figure 4A). Treatment with the PPARα agonist WY14643 induced Hmgcs2, Acox1, Ehhadh, Acsl1, Peci, Cyp4a10, Cyp4a12, and Cpt1a mRNA in WT hepatocytes, but with the exception of Cpt1a, none of these genes were induced in Tak1^–/– hepatocytes (Figure 4A). We confirmed that the effect of WY14643 on PPARα target gene expression was mediated through PPARα and that WY14643 had no significant effect on PPARγ target genes (Supplemental Figure 5). Congruently, we found that WY14643-induced FAO was markedly reduced in Tak1^–/– hepatocytes (Figure 4B). We also examined the effect of WY14643 on palmitate-induced lipid accumulation in hepatocytes. Treatment with palmitate resulted in the accumulation of lipid droplets in WT hepatocytes, and this was strongly enhanced in Tak1^–/– hepatocytes (Figure 4C). Interestingly, WY14643 treatment markedly inhibited palmitate-induced lipid accumulation in WT hepatocytes but had no effect in Tak1^–/– hepatocytes (Figure 4C). These results demonstrate that Tak1 deficiency suppresses PPARα expression and PPARα-induced gene expression, as well as FAO, resulting in enhanced lipid deposition in Tak1^–/– hepatocytes.

Figure 4

PPARα-mediated FAO is impaired in Tak1^–/– hepatocytes. (A) Primary hepatocytes were isolated and incubated with palmitate (125 μM) and PPARα agonist WY14643 (30 μM) for 6 hours. mRNA expression of Ppara, Hmgcs2, Acox1, Ehhadh, Acsl1, Peci, Cyp4a10, Cyp4a12, and Cpt1a was measured by quantitative real-time PCR. (B) Hepatocytes were preincubated with palmitate (250 μM) and WY-14643 (30 μM) for 16 hours followed by incubation with [¹⁴C]-palmitate (125 μM) and carnitine (1 mM) for 2 hours. FAO was measured by scintillation counter as acid-soluble metabolites and captured CO₂. Black bar, WT; gray bar, Tak1^–/– hepatocytes. (C) Hepatocytes were preincubated with palmitate (250 μM) with or without WY-14643 (30 μM) for 16 hours. Lipid accumulation was assessed by Oil Red O staining. Original magnification, ×200. Similar results were obtained in 3 independent experiments. Data are presented as the means ± SEM. *P < 0.05; **P < 0.01.

Rapamycin restores autophagy and lipid degradation in Tak1^–/– hepatocytes. We examined whether inhibition of mTORC1 with rapamycin can restore autophagy in Tak1^–/– hepatocytes. Autophagy, evidenced by increased LC3B-II GFP generation, accumulation of LC3B GFP aggregates, and reduced p62 accumulation, was induced in both WT and Tak1^–/– hepatocytes treated with rapamycin (Figure 5, A and B). Additionally, rapamycin treatment suppressed palmitate-induced lipid accumulation in both WT and Tak1^–/– hepatocytes (Supplemental Figure 6). We examined whether the rapamycin-induced reduction in lipid accumulation is dependent on autophagy. Autophagy inhibition by silencing of Atg5 increased lipid accumulation in both WT and Tak1^–/– hepatocytes and attenuated the response to rapamycin in both cell types (Supplemental Figure 6). These results suggest that the ability of rapamycin to reduce lipid accumulation in Tak1^–/– hepatocytes is largely dependent on autophagy.

Figure 5

Rapamycin restored autophagy and FAO-related genes in Tak1^–/– livers. (A and B) Primary hepatocytes were isolated from LC3B-GFP and Tak1^Δhep-GFP-LC3B transgenic mice. Cells were treated with rapamycin (5 μM) for 2 hours. (A) Immunoblotting for LC3B GFP and p62. (B) LC3B GFP accumulation in cytoplasm was assessed. Original magnification, ×200. White bar, vehicle; black bar, rapamycin. Similar results were obtained in 3 independent experiments. (C and D) One-month-old WT and Tak1^Δhep mice were treated with rapamycin (5 mg/kg) for 3 days (n = 5 each). (C) Immunoblotting for hepatic p62 and LC3B expression. (D) Immunohistochemistry for LC3B and its quantification. LC3B aggregation in WT and Tak1^Δhep mice treated with rapamycin. Original magnification, ×1,000. (E–H) Mice treated with rapamycin (5 mg/kg) for 3 days followed by fasting for 12 hours (n = 5 each). (E) Macroscopic appearance of livers from WT and Tak1^Δhep mice with fasting and rapamycin treatment. (F) Hepatic TG levels were measured, and (G) hepatic lipid deposition was assessed by quantification of Oil Red O staining. (H) Hepatic expression of Ppara, Hmgcs2, Cpt1a, and Sirt3 was measured by quantitative real-time PCR. Black bar, WT; gray bar, Tak1^Δhep mice. Data are presented as the means ± SEM. *P < 0.05; **P < 0.01.

Subsequently, we tested the effect of rapamycin in vivo. Rapamaycin suppressed mTORC1 activity in WT and Tak1^–/– livers, as shown by decreased S6 phosphorylation (Figure 5C). Consistent with the results of the in vitro experiments, rapamycin treatment induced autophagy in both WT and Tak1^–/– livers, as demonstrated by decreased p62 and increased LC3B-II and LC3B aggregates (Figure 5, C and D, and Supplemental Figure 1C). Moreover, rapamycin treatment reduced hepatic TG and lipid accumulation in fasted Tak1^Δhep mice (Figure 5, E–G). Rapamycin treatment also increased expression of FAO-related genes, such as Ppara, Hmgcs2, Cpt1a, and Sirt3 in Tak1^Δhep mice to levels comparable to those in WT mice (Figure 5H). These findings suggest that rapamycin restored autophagy and PPARα function by inhibiting excessive mTORC1 activity in Tak1^–/– livers.

Loss of Tak1 aggravates steatohepatitis. Since advanced experimental fatty liver disease showed decreased expression of TAK1 (Supplemental Figure 7) and Tak1^–/– hepatocytes stored more lipid than WT hepatocytes after treatment with palmitate (Figure 4C and Supplemental Figure 6), we investigated the effect of Tak1 deficiency on the response to an HFD. Four-week-old WT and Tak1^Δhep mice were placed on an HFD for 12 weeks. We observed that hepatic lipid deposition and hepatic TGs, FFAs, and cholesterol were markedly increased in Tak1^Δhep mice relative to WT mice (Figure 6, A–C, and Supplemental Figure 8D). Hepatic expression of Srebp1c, Dgat1, Pparg, and Fasn was significantly upregulated in Tak1^Δhep mice compared with that in WT mice fed an HFD (Figure 6D). Moreover, expression of inflammatory and fibrogenic genes including Tnf, Il6, Ccl2, Col1a1, and Tgfb1 and serum ALT levels were also markedly elevated in HFD-fed Tak1^Δhep mice (Figure 6, E and F). Hepatic expression of Afp, an HCC marker, was significantly increased in the HFD-fed Tak1^Δhep mice compared with that observed in Tak1^Δhep mice fed a normal chow diet (Figure 6G), suggesting that HFD feeding accelerates tumorigenesis in Tak1^Δhep mice. Additionally, p62 accumulation was significantly enhanced in HFD-fed Tak1^Δhep mice (Figure 6H and Supplemental Figure 1D), suggesting that HFD feeding further reduced autophagy in these mice. These findings collectively suggest that the impaired autophagy and increased lipogenesis caused by TAK1 deficiency render Tak1^Δhep mice highly susceptible to the adverse effects of an HFD. Interestingly, we observed no significant differences in body weight, food intake, or blood lipid and glucose levels between WT and Tak1^Δhep mice fed an HFD (Supplemental Figure 8).

Figure 6

Loss of TAK1 enhances liver injury, lipid accumulation, inflammation, and fibrosis. One-month-old WT and Tak1^Δhep mice were fed an HFD for an additional 12 weeks (n = 5 each). (A) Macroscopic appearance of livers. (B) Oil Red O staining and its quantification. Original magnification, ×200. (C) Hepatic TG levels. (D) Hepatic mRNA expression of Srebp1c, Dgat1, Pparg, and Fasn was measured by quantitative real-time PCR. (E) Hepatic expression of Tnf, Il6, Ccl2, Col1a1, and Tgfb1 mRNA. (F) Serum ALT levels. (G) mRNA expression of Afp. (H) Immunoblotting for p62. Black bar, WT; gray bar, Tak1^Δhep mice. Data are presented as the means ± SEM. *P < 0.05; **P < 0.01.

mTORC1 inhibition blocks HCC development in Tak1^Δhep mice. We also examined whether excessive mTORC1 activity and defective autophagy in Tak1^Δhep mice contribute to spontaneous hepatocarcinogenesis. Tak1^Δhep mice develop spontaneous liver cancer whose histology shows a trabecular or steatotic type of HCC, with expression of typical HCC markers, including α fetoprotein, glypican 3, glutamine synthetase, and heat shock protein 70 (Supplemental Figure 10 and refs. 5, 8, 27, 28). Inhibition of mTORC1 with rapamycin suppressed hepatocyte death and reduced serum ALT in 1-month-old Tak1^Δhep mice (Figure 7, A and B). Given that autophagy is strongly suppressed in HCC in Tak1^Δhep mice, as demonstrated by the accumulation of p62 aggregates (Figure 7F and Supplemental Figure 1E), we examined whether inhibition of mTORC1 activity and restoration of autophagy by rapamycin prevent HCC development. We treated Tak1^Δhep mice with rapamycin during early (from 2 to 10 weeks after birth) or late (from 7 to 9 months of age) stages of HCC development. The early treatment significantly reduced the number of liver tumors (Figure 7C), but did not change their maximal size compared with those in the vehicle-treated Tak1^Δhep mice (S. Inokuchi-Shimizu and E. Seki, unpublished observations). Moreover, late-stage rapamycin treatment significantly reduced the number and size of tumors and the extent of liver fibrosis in Tak1^Δhep mice (Figure 7, D and E, and Supplemental Figure 9).

Figure 7

Induction of autophagy inhibits HCC growth in Tak1 ^Δhep mice. (A and B) One-month-old WT and Tak1^Δhep mice were treated with or without rapamycin (5 mg/kg) for 3 days (n = 5 each). (A) Hepatocyte apoptosis was assessed by TUNEL staining. (B) Serum ALT levels. (C) Tak1^Δhep mice were treated with rapamycin (5 mg/kg) twice weekly from 2 to 10 weeks of age, and the livers were harvested from the mice at 36 weeks of age (n = 12 each). Macroscopically visible tumors were counted. (D and E) Tak1^Δhep mice were treated with rapamycin (5 mg/kg) twice weekly from 28 to 36 weeks of age, and then the livers were harvested (n = 12). (D) Macroscopically visible tumors were counted, and maximal size was measured. (E) Macroscopic appearance of the liver. (F) Expression of p62 in HCC in Tak1^Δhep mice as assessed by immunoblotting for p62. (A–D) Black bar, vehicle; white bar, rapamycin. Data are presented as the means ± SEM. *P < 0.05; **P < 0.01.

Two well-established functions of TAK1 are the regulation of IKK/NF-κB signaling and JNK/AP-1 signaling downstream of IL-1, TNF, TGFβ, and TLRs. Through these pathways, TAK1 was suggested to modulate cell survival, proliferation, inflammation, fibrosis, and tumorigenesis, as well as lipid metabolism and insulin resistance (2). Disruption of the Ikbkg gene in hepatocytes, which encodes the IKK regulatory/scaffold subunit, combined with HFD feeding, results in exacerbated steatohepatitis, a phenotype similar to that of Tak1 ablation (29). Remarkably, TAK1 ablation in liver has a stronger tumorigenic effect than does Ikbkg ablation (5, 8, 9), suggesting that enhanced susceptibility to hepatocyte death caused by the loss of IKK-NF-κB activity accounts for only part of the adverse consequences of Tak1 ablation. Furthermore, Ikbkg ablation in Tak1^–/– livers attenuates HCC development (5, 8, 9), suggesting that TAK1 suppresses tumorigenesis through an IKK-independent mechanism. Since TAK1 was shown to control AMPK activity in cells (10, 30), we postulated that decreased AMPK activity and enhanced mTORC1 activation might contribute to some of the pathologies associated with loss of hepatic TAK1, including hepatosteatosis and HCC development. The results shown above provide strong support to this hypothesis and demonstrate that TAK1 is a physiologically relevant activator of AMPK and autophagy and a negative regulator of mTORC1, which contribute to lipid metabolism during feeding and fasting. Furthermore, we show that excessive mTORC1 activity is an important contributor to the spontaneous hepatic tumorigenesis seen in Tak1^Δhep mice. In this respect, the liver phenotype in Tak1^Δhep mice resembles that of Tsc1^Δhep mice, whose hepatocytes lack TSC1, a negative regulator of mTORC1 (31). In both strains of mice, rapamycin inhibits HCC development and hepatic steatosis (14, 31)

Our data demonstrate that starvation-induced LKB1 and AMPK phosphorylation requires TAK1 in hepatocytes, although the mechanism by which TAK1 controls LKB1 and AMPK activities remains elusive (10, 30). Activated AMPK then phosphorylates raptor, which inhibits mTORC1 activity, resulting in stimulation of autophagy (Supplemental Figure 11 and refs. 12, 13). The contribution of TAK1 to starvation-induced autophagy, however, has been controversial. One study reported that TAK1 is required for starvation-induced autophagy (10, 19), while another study showed that TAK1 is dispensable for starvation-induced autophagy but is required for TRAIL-induced autophagy (10, 19). Given the different cell types examined in the previous studies and in our current study, it is conceivable that the requirement of TAK1 in starvation-induced autophagy is largely cell-type specific and may reflect differential effects on AMPK. Although our results show that starvation- and metformin-induced AMPK activation and autophagy are dependent on TAK1 and that ectopic expression of activated AMPK reverses the autophagy defect in Tak1^–/– hepatocytes, suggesting TAK1-mediated autophagy through AMPK. However, a recent report demonstrated that TAK1 also participates in autophagy through the inhibition of p70S6K phosphorylation by interacting with raptor independently of AMPK (32). This suggests that the increased basal activity of mTORC1 and suppressed basal autophagy in Tak1^–/– hepatocytes may be mediated by a direct effect of TAK1 on mTORC1-p70S6K activity. It should be noted that TAK1–binding protein 2 (TAB2) and TAB3 also regulate autophagy through their effects on beclin 1 and may therefore be involved in TAK1-regulated autophagy in the liver (33, 34). In addition, TAK1 deficiency results in reduced expression of several autophagy genes, including Atg3, Atg5, Atg8, and Ulk1, providing another mechanism by which TAK1 can influence autophagy.

In addition to autophagy, mTORC1 inhibits FAO by suppressing PPARα expression and function (14). PPARα is a powerful FAO inducer involved in lipid breakdown and energy production through induction of FAO-related genes, including Hmgcs2, Acox1, and Cpt1a. Absence of Tsc1 causes a strong mTORC1 activation that decreases expression of PPARα and its target genes, thereby demonstrating another similarity between Tak1^Δhep and Tsc1^Δhep mice (14). Restoration of PPARα activity by ectopic PPARα introduction or treatment with a PPARα agonist, however, does not restore PPARα function in Tsc1^Δhep mice, suggesting that strong mTORC1 overactivation may also affect functions of cofactors (e.g., NCoR1) that require PPARα-mediated functions (14). Due to overactivation of mTORC1 in Tak1^–/– livers, expression of PPARα and its target genes, as well as FAO, are suppressed, which may contribute to elevated hepatic steatosis in Tak1^Δhep mice upon fasting or after HFD feeding. Moreover, TAK1 deficiency inhibited expression of SIRT3, which is required for deacetylation and activation of LCAD and HMGCS2 (17, 35), suggesting that FAO may also be suppressed through inhibition of SIRT3 in Tak1^–/– hepatocytes.

We found that inhibition of mTORC1 with rapamycin successfully restored autophagy and PPARα function and thereby reduced lipid accumulation in the Tak1^–/– livers. This beneficial effect of mTORC1 inhibition was largely dependent on restoration of autophagy, further suggesting that the autophagic defect is responsible for the increase in hepatic steatosis caused by the TAK1 deficiency. As previously observed in Tsc1^Δhep mice (31), we found that rapamycin treatment also blocked fibrosis and HCC development in Tak1^Δhep mice.

Hypernutrition causes an mTORC1 activation that contributes to lipogenesis through activation of SREBP-1c and PPARγ and inhibition of lipid breakdown through suppression of autophagy (12, 18). Furthermore, hepatic TAK1 expression is decreased during dietary and genetic obesity (Supplemental Figure 7), allowing for an additional enhancement of mTORC1 activity that increases lipogenesis and suppresses autophagy and FAO. Degradation of p62-bound polyubiquitinated proteins within autophagosomes is mediated by autophagy (22). The decrease in autophagy in Tak1^–/– hepatocytes results in accumulation of p62, which leads to further activation of mTORC1, as recently described (36, 37). The decrease in autophagy and the increase in mTORC1 activity resulted in enhanced development of hepatosteatosis in HFD-fed Tak1^Δhep mice. In addition, we found that liver inflammation and fibrosis were substantially increased in Tak1^Δhep mice after HFD feeding. Despite extensive investigations, the trigger(s) or factor(s) that cause progression from simple steatosis to steatohepatitis remain unknown. Given the fact that hepatic TAK1 expression is suppressed in advanced fatty liver disease, factors that lead to a further decrease in TAK1 activity or increase mTORC1 activation may contribute to the progression from simple steatosis to steatohepatitis.

Since the IKK/NF-κB pathway is known to prevent HCC formation and TAK1 acts upstream of the IKK/NF-κB pathway, it was suggested that spontaneous HCC development in Tak1^Δhep mice is caused by the absence of IKK/NF-κB function (5, 7–9). However, the present results indicate that excessive mTORC1 activity is equally important or even more important for spontaneous HCC development in Tak1^Δhep mice. It has been reported that the complete absence of autophagy causes the development of benign liver tumors with p62 accumulation rather than the development of HCC (38, 39). As recently shown in the pancreas (40), malignant progression seems to require a certain level of autophagy, thus explaining why Tak1^Δhep mice, in which autophagy is reduced but not completely absent, develop malignant HCC, whereas Alb-Cre Atg5^fl/fl or Alb-Cre Atg7^fl/fl mice only develop benign adenomas.

Despite aggravated hepatosteatosis, body weight and insulin resistance were not altered in HFD-fed Tak1^Δhep mice compared with WT mice fed an HFD, which is similar to the phenotype of Ikbkg^Δhep mice (29). This indicates a disassociation between hepatosteatosis and insulin resistance. Hepatic TAK1 and IKK-NF-κB prevent hepatosteatosis, but do not affect systemic insulin sensitivity.

Remarkably, in 30% of human prostate cancers, chromosome 6q15.1 containing the Tak1 locus is deleted (41). Although clinical evidence for TAK1 deficiency in human HCC is still lacking, deficiency in TAK1 may promote tumor development due to overactivation of mTORC1 and suppression of autophagy. Since rapamycin treatment of HCC patients after liver transplantation results in reduced HCC recurrence (42), it is plausible that inhibition of mTORC1 activity may become a useful therapy for HCC (43). In summary, TAK1 is a physiological regulator of AMPK and autophagy and a negative regulator of mTORC1. TAK1 deficiency causes excessive mTORC1 activation and autophagy defects, which contribute to enhancement of hepatosteatosis and spontaneous hepatic tumorigenesis.

View Supplemental data

LC3B-GFP transgenic mice were provided by Noboru Mizushima (Tokyo Medical and Dental University, Tokyo, Japan). Pingping Li (UCSD, Department of Medicine) provided liver samples from ob/ob mice. This study was supported by NIH grants R01AA02172 (to E. Seki), R01DK085252 (to E. Seki), R01CA155120 (to M. Karin), P42 ES010337 (to E. Seki, D.A. Brenner, M. Karin, and M.H. Ellisman), and R01AA020753 (to K. Asahina); by a pilot grant from the UCSD Digestive Diseases Research Development Center (DK080506, to E. Seki); by the National Natural Science Foundation of China (30500658 and 81370550, to L. Yang); a Postdoctoral Fellowship (to Y.S. Roh) and a Liver Scholar Award (to E.J. Park) from the American Liver Foundation; a postdoctoral fellowship from Susan G. Komen for the Cure (KG111506, to X. Wu); and a Postdoctoral Fellowship for Research Abroad, a Research Fellowship for Young Scientists from the Japan Society for the Promotion of Science, and a Uehara Memorial Foundation Fellowship (to K. Taniguchi).

Address correspondence to: Ekihiro Seki, UCSD, 9500 Gilman Drive, MC# 0063, Biomedical Research Facility 2, Rm 4A21, La Jolla, California 92093-0063, USA. Phone: 858.822.3347; E-mail: ekseki@ucsd.edu.

Conflict of interest: The authors have declared that no conflict of interest exists.

Reference information: J Clin Invest. 2014;124(8):3566–3578. doi:10.1172/JCI74068.

1. Shim JH, et al. TAK1, but not TAB1 or TAB2, plays an essential role in multiple signaling pathways in vivo. Genes Dev. 2005;19(22):2668–2681.
   View this article via: PubMed CrossRef Google Scholar
2. Sakurai H. Targeting of TAK1 in inflammatory disorders and cancer. Trends Pharmacol Sci. 2012;33(10):522–530.
   View this article via: PubMed CrossRef Google Scholar
3. Seki E, Brenner DA, Karin M. A liver full of JNK: signaling in regulation of cell function and disease pathogenesis, and clinical approaches. Gastroenterology. 2012;143(2):307–320.
   View this article via: PubMed CrossRef Google Scholar
4. Karin M. Nuclear factor-kappaB in cancer development and progression. Nature. 2006;441(7092):431–436.
   View this article via: PubMed CrossRef Google Scholar
5. Inokuchi S, et al. Disruption of TAK1 in hepatocytes causes hepatic injury, inflammation, fibrosis, and carcinogenesis. Proc Natl Acad Sci U S A. 2010;107(2):844–849.
   View this article via: PubMed CrossRef Google Scholar
6. Yang L, et al. Transforming growth factor-beta signaling in hepatocytes promotes hepatic fibrosis and carcinogenesis in mice with hepatocyte-specific deletion of TAK1. Gastroenterology. 2013;144(5):1042–1054 e1044.
   View this article via: PubMed CrossRef Google Scholar
7. Maeda S, Kamata H, Luo JL, Leffert H, Karin M. IKKbeta couples hepatocyte death to cytokine-driven compensatory proliferation that promotes chemical hepatocarcinogenesis. Cell. 2005;121(7):977–990.
   View this article via: PubMed CrossRef Google Scholar
8. Bettermann K, et al. TAK1 suppresses a NEMO-dependent but NF-κB-independent pathway to liver cancer. Cancer Cell. 2010;17(5):481–496.
   View this article via: PubMed CrossRef Google Scholar
9. Luedde T, et al. Deletion of NEMO/IKKγ in liver parenchymal cells causes steatohepatitis and hepatocellular carcinoma. Cancer Cell. 2007;11(2):119–132.
   View this article via: PubMed CrossRef Google Scholar
10. Herrero-Martin G, et al. TAK1 activates AMPK-dependent cytoprotective autophagy in TRAIL-treated epithelial cells. EMBO J. 2009;28(6):677–685.
    View this article via: PubMed CrossRef Google Scholar
11. Russo GL, Russo M, Ungaro P. AMP-activated protein kinase: A target for old drugs against diabetes and cancer. Biochem Pharmacol. 2013;86(3):339–350.
    View this article via: PubMed CrossRef Google Scholar
12. Zoncu R, Efeyan A, Sabatini DM. mTOR: from growth signal integration to cancer, diabetes and ageing. Nat Rev Mol Cell Biol. 2011;12(1):21–35.
    View this article via: PubMed CrossRef Google Scholar
13. Egan DF, et al. Phosphorylation of ULK1 (hATG1) by AMP-activated protein kinase connects energy sensing to mitophagy. Science. 2011;331(6016):456–461.
    View this article via: PubMed CrossRef Google Scholar
14. Sengupta S, Peterson TR, Laplante M, Oh S, Sabatini DM. mTORC1 controls fasting-induced ketogenesis and its modulation by ageing. Nature. 2010;468(7327):1100–1104.
    View this article via: PubMed CrossRef Google Scholar
15. Djouadi F, Brandt JM, Weinheimer CJ, Leone TC, Gonzalez FJ, Kelly DP. The role of the peroxisome proliferator-activated receptor alpha (PPAR α) in the control of cardiac lipid metabolism. Prostaglandins Leukot Essent Fatty Acids. 1999;60(5–6):339–343.
    View this article via: PubMed Google Scholar
16. Purushotham A, Schug TT, Xu Q, Surapureddi S, Guo X, Li X. Hepatocyte-specific deletion of SIRT1 alters fatty acid metabolism and results in hepatic steatosis and inflammation. Cell Metab. 2009;9(4):327–338.
    View this article via: PubMed CrossRef Google Scholar
17. Hirschey MD, et al. SIRT3 regulates mitochondrial fatty-acid oxidation by reversible enzyme deacetylation. Nature. 2010;464(7285):121–125.
    View this article via: PubMed CrossRef Google Scholar
18. Singh R, et al. Autophagy regulates lipid metabolism. Nature. 2009;458(7242):1131–1135.
    View this article via: PubMed CrossRef Google Scholar
19. Criollo A, et al. The IKK complex contributes to the induction of autophagy. EMBO J. 2010;29(3):619–631.
    View this article via: PubMed CrossRef Google Scholar
20. McCue MD. Starvation physiology: reviewing the different strategies animals use to survive a common challenge. Comp Biochem Physiol A Mol Integr Physiol. 2010;156(1):1–18.
    View this article via: PubMed CrossRef Google Scholar
21. Kroemer G, Marino G, Levine B. Autophagy and the integrated stress response. Mol Cell. 2010;40(2):280–293.
    View this article via: PubMed CrossRef Google Scholar
22. Komatsu M, Kageyama S, Ichimura Y. p62/SQSTM1/A170: physiology and pathology. Pharmacol Res. 2012;66(6):457–462.
    View this article via: PubMed CrossRef Google Scholar
23. Klionsky DJ, et al. Guidelines for the use and interpretation of assays for monitoring autophagy. Autophagy. 2012;8(4):445–544.
    View this article via: PubMed CrossRef Google Scholar
24. Mizushima N, Yamamoto A, Matsui M, Yoshimori T, Ohsumi Y. In vivo analysis of autophagy in response to nutrient starvation using transgenic mice expressing a fluorescent autophagosome marker. Mol Biol Cell. 2004;15(3):1101–1111.
    View this article via: PubMed Google Scholar
25. Rakhshandehroo M, et al. Comprehensive analysis of PPARα-dependent regulation of hepatic lipid metabolism by expression profiling. PPAR Res. 2007;2007:26839.
    View this article via: PubMed Google Scholar
26. Rakhshandehroo M, Knoch B, Müller M, Kersten S. Peroxisome proliferator-activated receptor α target genes. PPAR Res. 2010;2010:612089.
    View this article via: PubMed Google Scholar
27. Di Tommaso L, et al. Diagnostic value of HSP70, glypican 3, and glutamine synthetase in hepatocellular nodules in cirrhosis. Hepatology. 2007;45(3):725–734.
    View this article via: PubMed CrossRef Google Scholar
28. International Consensus Group for Hepatocellular Neoplasia. . Pathologic diagnosis of early hepatocellular carcinoma: a report of the international consensus group for hepatocellular neoplasia. Hepatology. 2009;49(2):658–664.
    View this article via: PubMed CrossRef Google Scholar
29. Wunderlich FT, et al. Hepatic NF-κB essential modulator deficiency prevents obesity-induced insulin resistance but synergizes with high-fat feeding in tumorigenesis. Proc Natl Acad Sci U S A. 2008;105(4):1297–1302.
    View this article via: PubMed CrossRef Google Scholar
30. Xie M, et al. A pivotal role for endogenous TGF-β-activated kinase-1 in the LKB1/AMP-activated protein kinase energy-sensor pathway. Proc Natl Acad Sci U S A. 2006;103(46):17378–17383.
    View this article via: PubMed CrossRef Google Scholar
31. Menon S, et al. Chronic activation of mTOR complex 1 is sufficient to cause hepatocellular carcinoma in mice. Sci Signal. 2012;5(217):ra24.
    View this article via: PubMed Google Scholar
32. Shin JH, et al. TAK1 regulates autophagic cell death by suppressing the phosphorylation of p70 S6 kinase 1. Sci Rep. 2013;3:1561.
    View this article via: PubMed Google Scholar
33. Criollo A, et al. Inhibition of autophagy by TAB2 and TAB3. EMBO J. 2011;30(24):4908–4920.
    View this article via: PubMed CrossRef Google Scholar
34. Takaesu G, Kobayashi T, Yoshimura A. TGFβ-activated kinase 1 (TAK1)-binding proteins (TAB) 2 and 3 negatively regulate autophagy. J Biochem. 2012;151(2):157–166.
    View this article via: PubMed CrossRef Google Scholar
35. Shimazu T, et al. SIRT3 deacetylates mitochondrial 3-hydroxy-3-methylglutaryl CoA synthase 2 and regulates ketone body production. Cell Metab. 2010;12(6):654–661.
    View this article via: PubMed CrossRef Google Scholar
36. Duran A, et al. p62 is a key regulator of nutrient sensing in the mTORC1 pathway. Mol Cell. 2011;44(1):134–146.
    View this article via: PubMed CrossRef Google Scholar
37. Linares JF, Duran A, Yajima T, Pasparakis M, Moscat J, Diaz-Meco MT. K63 polyubiquitination and activation of mTOR by the p62-TRAF6 complex in nutrient-activated cells. Mol Cell. 2013;51(3):283–296.
    View this article via: PubMed CrossRef Google Scholar
38. Takamura A, et al. Autophagy-deficient mice develop multiple liver tumors. Genes Dev. 2011;25(8):795–800.
    View this article via: PubMed CrossRef Google Scholar
39. Inami Y, et al. Persistent activation of Nrf2 through p62 in hepatocellular carcinoma cells. J Cell Biol. 2011;193(2):275–284.
    View this article via: PubMed CrossRef Google Scholar
40. Rosenfeldt MT, et al. p53 status determines the role of autophagy in pancreatic tumour development. Nature. 2013;504(7479):296–300.
    View this article via: PubMed CrossRef Google Scholar
41. Wu M, et al. Suppression of Tak1 promotes prostate tumorigenesis. Cancer Res. 2012;72(11):2833–2843.
    View this article via: PubMed CrossRef Google Scholar
42. Kawahara T, Asthana S, Kneteman NM. m-TOR inhibitors: what role in liver transplantation? J Hepatol. 2011;55(6):1441–1451.
    View this article via: PubMed CrossRef Google Scholar
43. LoPiccolo J, Blumenthal GM, Bernstein WB, Dennis PA. Targeting the PI3K/Akt/mTOR pathway: effective combinations and clinical considerations. Drug Resist Updat. 2008;11(1–2):32–50.
    View this article via: PubMed Google Scholar
44. Sato S, et al. Essential function for the kinase TAK1 in innate and adaptive immune responses. Nat Immunol. 2005;6(11):1087–1095.
    View this article via: PubMed CrossRef Google Scholar
45. Wu X, Sun SC. Retroviral oncoprotein Tax deregulates NF-κB by activating Tak1 and mediating the physical association of Tak1-IKK. EMBO Rep. 2007;8(5):510–515.
    View this article via: PubMed CrossRef Google Scholar
46. Storch J, Zhou YX, Lagakos WS. Metabolism of apical versus basolateral sn-2-monoacylglycerol and fatty acids in rodent small intestine. J Lipid Res. 2008;49(8):1762–1769.
    View this article via: PubMed CrossRef Google Scholar
47. Liu Y, et al. Knockdown of acyl-CoA:diacylglycerol acyltransferase 2 with antisense oligonucleotide reduces VLDL TG and ApoB secretion in mice. Biochim Biophys Acta. 2008;1781(3):97–104.
    View this article via: PubMed Google Scholar
48. Adachi M, Brenner DA. High molecular weight adiponectin inhibits proliferation of hepatic stellate cells via activation of adenosine monophosphate-activated protein kinase. Hepatology. 2008;47(2):677–685.
    View this article via: PubMed CrossRef Google Scholar
49. Lee JH, et al. Maintenance of metabolic homeostasis by Sestrin2 and Sestrin3. Cell Metab. 2012;16(3):311–321.
    View this article via: PubMed CrossRef Google Scholar
50. Sears DD, et al. Mechanisms of human insulin resistance and thiazolidinedione-mediated insulin sensitization. Proc Natl Acad Sci U S A. 2009;106(44):18745–18750.
    View this article via: PubMed CrossRef Google Scholar
51. Hsiao A, Worrall DS, Olefsky JM, Subramaniam S. Variance-modeled posterior inference of microarray data: detecting gene-expression changes in 3T3-L1 adipocytes. Bioinformatics. 2004;20(17):3108–3127.
    View this article via: PubMed CrossRef Google Scholar
52. Hsiao A, Ideker T, Olefsky JM, Subramaniam S. VAMPIRE microarray suite: a web-based platform for the interpretation of gene expression data. Nucleic Acids Res. 2005;33(Web Server issue):W627–W632.
    View this article via: PubMed Google Scholar