p62/Sqstm1 promotes malignancy of HCV-positive hepatocellular carcinoma through Nrf2-dependent metabolic reprogramming

Abstract

p62/Sqstm1 is a multifunctional protein involved in cell survival, growth and death, that is degraded by autophagy. Amplification of the p62/Sqstm1 gene, and aberrant accumulation and phosphorylation of p62/Sqstm1, have been implicated in tumour development. Herein, we reveal the molecular mechanism of p62/Sqstm1-dependent malignant progression, and suggest that molecular targeting of p62/Sqstm1 represents a potential chemotherapeutic approach against hepatocellular carcinoma (HCC). Phosphorylation of p62/Sqstm1 at Ser349 directs glucose to the glucuronate pathway, and glutamine towards glutathione synthesis through activation of the transcription factor Nrf2. These changes provide HCC cells with tolerance to anti-cancer drugs and proliferation potency. Phosphorylated p62/Sqstm1 accumulates in tumour regions positive for hepatitis C virus (HCV). An inhibitor of phosphorylated p62-dependent Nrf2 activation suppresses the proliferation and anticancer agent tolerance of HCC. Our data indicate that this Nrf2 inhibitor could be used to make cancer cells less resistant to anticancer drugs, especially in HCV-positive HCC patients.

Figure 1. Activation of Nrf2 in HCC cells expressing phospho-mimetic p62.

(a) Enzymes involved in glucose and glutamine metabolism regulated by Nrf2. (b) Immunoblot analysis. FLAG-tagged p62 and its mutants were overproduced in the indicated HCC cells using an adenovirus vector. At 48 h after infection, cytosolic and nuclear fractions were prepared and subjected to immunoblot analysis with the specified antibodies. Data were obtained from three independent experiments.

Figure 2. Gene expression of Nrf2-targets in HCC cells harbouring phospho-mimetic p62.

Quantification of mRNA levels of Nrf2 target genes in the indicated HCC cells expressing GFP, FLAG-p62 or its mutants. Values were normalized to the amount of mRNA in HCC cells expressing GFP. The experiments were performed three times. Data are presented as means±s.e. *P<0.05, **P<0.01, ***P<0.001 as determined by the Welch t-test. We used mouse p62 constructs in these experiments.

Figure 3. Metabolic intermediates in HCC cells harbouring phospho-mimetic p62.

Quantification of metabolic intermediates in Huh7 cells expressing GFP, wild-type, S351E or S351A. The experiments were performed three times. Data are presented as means±s.e. *P<0.05, **P<0.01 as determined by the Welch t-test.

Figure 4. Glucose and glutamine metabolism in HCC cells harbouring phospho-mimetic p62.

(a) Tracer study using [U-¹³C₆] glucose. Huh7 cells expressing GFP, S351E or S351A were incubated with [U¹³C₆] glucose for 1 h and analysed. The experiments were performed ten times. Data are presented as means±s.e. *P<0.05, **P<0.01, ***P<0.001 as determined by the Welch t-test. (b) Tracer study using [U-¹³C₅] glutamine. Huh7 cells expressing GFP, S351E or S351A were incubated with [U-¹³C₅] glutamine for 6 h and analysed. The experiments were performed three times. Data are presented as means±s.e. *P<0.05, **P<0.01, ***P<0.001 as determined by the Welch t-test. We used mouse p62 constructs in these experiments.

Figure 5. Tolerance to anti-cancer drugs and proliferation potency of HCC cells expressing phospho-mimetic p62.

(a) Effect of phospho-mimetic p62 on anti-cancer drug resistance. The indicated HCC cells were infected with adenovirus for GFP, wild-type p62 or its mutants for 60 h. Thereafter, the cells were treated with sorafenib or cisplatin at the indicated concentration for 48 h, and the survival ratio was determined. The experiments were performed three times. Data represent means±s.e. *P<0.05, **P<0.01, ***P<0.001 as determined by the Welch t-test. (b) Effect of phospho-mimetic p62 on proliferation. The indicated HCC cells were infected with adenovirus for GFP, wild-type p62 or its mutants. Proliferation was measured from 60 h after infection. Initial cell numbers were normalized to 1. The experiments were performed three times. Data represent means±s.e. *P<0.05, **P<0.01, ***P<0.001 as determined by the Welch t-test. (c) Glutathione levels in the presence or absence of BSO. The indicated HCC cells were infected with adenovirus expressing S351E. At 60 h after infection, the cells were cultured in the presence or absence of BSO for 72 h, and then glutathione was quantitated. The experiments were performed three times. Data represent means±s.e. *P<0.05, **P<0.01, ***P<0.001 as determined by the Welch t-test. (d) Immunoblot analysis. Indicated HCC cell lines were cultured as described in c, and their cytosolic and nuclear fractions were prepared and subjected to immunoblot analysis with the specified antibodies. Data were obtained from three independent experiments. (e) Proliferation in the presence or absence of BSO. The indicated HCC cells were infected with adenovirus for S351E. At 60 h after infection, the cells were cultured in the presence or absence of BSO for the indicated time, and proliferation was measured. Initial cell numbers were normalized to 1. The experiments were performed three times. Data represent means±s.e. *P<0.05, **P<0.01, ***P<0.001 as determined by the Welch t-test. We used mouse p62 constructs in these experiments.

Figure 6. Persistent activation of Nrf2 in autophagy-deficient mouse livers.

(a) Immunoblot analysis. Liver homogenates and nuclear fractions of female Atg7^f/f, Atg7^f/f;Alb-Cre, Atg7^f/f;Nrf2^f/f and Atg7^f/f;Nrf2^f/f;Alb-Cre mice aged at 5 weeks were prepared, and were subjected to immunoblotting with the indicated antibodies. (b) Immunofluorescence analysis. Liver sections described in a were double-immunostained with a combination of anti-phosphorylated p62 (green) and anti-p62 (red) antibodies, or of anti-Keap1 (green) and anti-p62 (red) antibodies. Scale bars, 20 μm. (c) Nrf2-dependent gene expressions in autophagy-deficient livers. Total RNAs were prepared from livers of female Atg7^f/f (n=4), Atg7^f/f;Alb-Cre (n=4), Atg7^f/f;Nrf2^f/f (n=4) and Atg7^f/f;Nrf2^f/f;Alb-Cre mice (n=4) aged at 5 weeks. Values were normalized to the amount of mRNA in the livers of Atg7^f/f or Atg7^f/f;Nrf2^f/f mice. The experiments were performed three times. Data are means±s.e. *P<0.05, **P<0.01 and ***P<0.001 as determined by the Welch t-test.

Figure 7. Metabolic intermediates in autophagy-deficient mouse livers.

Quantification of metabolic intermediates in glucose and glutamine metabolism in livers of female Atg7^f/f (n=5), Atg7^f/f;Alb-Cre (n=5), Atg7^f/f;Nrf2^f/f (n=10) and Atg7^f/f;Nrf2^f/f;Alb-Cre mice (n=13) aged at 5 weeks. The amount of glycogen was also quantified. Data are means±s.e. *P<0.05, **P<0.01 and ***P<0.001 as determined by the Welch t-test.

Figure 8. Dynamics of S349-phosphorylated p62 in human HCC.

(a) Immunoblot analysis of human HCC lysates with the indicated antibodies. Types of HCC are indicated; the other patients are four cases of non-alcoholic HCC (No. 5, 8, 10 and 15), and one case of combined hepatocellular and cholangiocarcinoma (No. 14). N: non-tumour region; T: tumour region. (b) Immunohistochemical images. Paraffin sections of indicated human HCC patients were stained with anti-p62, anti-S349-phosphorylated p62 or anti-Keap1 antibody. Scale bar, 100 μm. (c) Double immunofluorescence microscopy. Three cases of HCV-positive human HCC containing typical aggregates were double-immunostained with anti-p62 and anti-S349-phosphorylated p62 (left) or anti-p62 and Keap1 (right) antibodies. Merged images of p62 (red) and phosphorylated p62 (green) or p62 (red) and Keap1 (green) are shown on the right. Scale bar, 50 μm.

Figure 9. K67, a chemical compound that inhibits the interaction between Keap1 and S349-phosphorylated p62.

(a) Working hypothesis about an inhibitors of the interaction between Keap1 and S349-phosphorylated p62. (b) Structural formulas of K67 and Cpd16. (c) Crystal structure of Keap1-DC (white) in complex with K67 (cyan). Overall structure of Keap1-DC, shown as a ribbon model. K67 (cyan) and some of the potential interacting residues of Keap1-DC are shown in stick representation. A close-up view of K67 bound to Keap1-DC is shown. Intermolecular electrostatic interactions are depicted as broken red lines. (d) In vitro pull-down assay. Full-length Keap1 was allowed to form a complex with K67 (left panel) or Cdp16 (right panel) at the indicated concentration. Upper panel: Keap1 binding to His-GST-tagged Neh2(Δ17-51) was estimated by Oriole fluorescent gel staining. Middle and bottom panels: Keap1 binding to OSF-p-p62 (middle panel) or OSF-Nrf2 (bottom panel) was estimated by immunoblot analysis with anti-Keap1 antibody. Data are representative of three independent experiments. (e) Modelling of Keap1 (white)–K67 (cyan)–DLGex (green; upper left panel) and Keap1–Cpd16 (orange)–DLGex (lower left panel). Upper right: schematic representation of the interface between DLGex and K67. Lower right: schematic representation of the interface between DLGex and Cpd16. Potential hydrogen bonds are indicated by dashed lines between the atoms involved, whereas hydrophobic contacts are represented by an arc with spokes radiating towards the ligand atoms they contact. (f) In vitro pull-down assay. Full-length Keap1 was allowed to form a complex with K67 (left panel) or Cdp16 (right panel) at the indicated concentration. Keap1 binding to His-GST-tagged Neh2(1-56) was estimated by Oriole fluorescent gel staining. Data are representative of three independent experiments. (g) In vitro pull-down assay with biotinylated compounds. His-GST-tagged Neh2(1-56) binding to the Keap1-DC complexed with biotinylated K67 or Cpd16 was estimated by Coomassie brilliant blue staining. Data are representative of three independent experiments.

Figure 10. Inhibitory effects of K67 on tumour growth and tolerance to anti-cancer agents.

(a) Immunoprecipitation assays. Huh1 cells expressing FLAG-Keap1 were cultured in the presence or absence of indicated compound (50 μM) for 12 h. The immunoprecipitant with anti-FLAG antibody was subjected to immunoblot analysis. Experiments were performed three times. (b) Ubiquitination of Nrf2. Huh7 cells expressing S351E were cultured in the presence or absence of indicated compound (50 μM). After 4 h of treatment, the cells were treated with or without lactacystin (10 μM) for 8 h. The immunoprecipitant with anti-Nrf2 antibody was subjected to immunoblot analysis. Experiments were performed three times. (c) Immunoblot analysis. Huh1 cells were cultured in the presence of 50 μM K67 or Cpd16 for the indicated times. Cytosolic and nuclear fractions were prepared and subjected to immunoblot analysis. Experiments were performed three times. (d) Quantification of mRNA levels of Nrf2 target genes in Huh1 cells treated with 50 μM K67 for the indicated times. Values were normalized to the amount of mRNA in non-treated Huh1 cells. Experiments were performed three times. Data represent means±s.e. *P<0.05, **P<0.01, ***P<0.001 as determined by the Welch t-test. (e) Effect of K67 on cell proliferation. Huh1 and Huh7 cells were pre-cultured in the presence of DMSO, K67 or Cpd16. Proliferation was measured starting at 72 h after pre-culture (n=4). Initial cell numbers were normalized to 1. Experiments were performed three times. Data represent means±s.e. *P<0.05, **P<0.01, ***P<0.001 as determined by the Welch t-test. (f) Effect of K67 on resistance to anti-cancer drugs. Huh1 and Huh7 cells were pre-cultured in the presence of DMSO or K67 for 96 h. Thereafter, the cells were treated with sorafenib, cisplatin or either drug in combination with K67 for 48 h, and survival ratio was determined (n=4). Experiments were performed three times. Data represent means±s.e. *P<0.05, **P<0.01, ***P<0.001 as determined by the Welch t-test. (g) Schematic diagram of cancer malignancy mediated by the p62–Keap1–Nrf2 axis.