Autolysosomal β-catenin degradation regulates Wnt-autophagy-p62 crosstalk

Abstract

The Wnt/β-catenin signalling and autophagy pathways each play important roles during development, adult tissue homeostasis and tumorigenesis. Here we identify the Wnt/β-catenin signalling pathway as a negative regulator of both basal and stress-induced autophagy. Manipulation of β-catenin expression levels in vitro and in vivo revealed that β-catenin suppresses autophagosome formation and directly represses p62/SQSTM1 (encoding the autophagy adaptor p62) via TCF4. Furthermore, we show that during nutrient deprivation β-catenin is selectively degraded via the formation of a β-catenin-LC3 complex, attenuating β-catenin/TCF-driven transcription and proliferation to favour adaptation during metabolic stress. Formation of the β-catenin-LC3 complex is mediated by a W/YXXI/L motif and LC3-interacting region (LIR) in β-catenin, which is required for interaction with LC3 and non-proteasomal degradation of β-catenin. Thus, Wnt/β-catenin represses autophagy and p62 expression, while β-catenin is itself targeted for autophagic clearance in autolysosomes upon autophagy induction. These findings reveal a regulatory feedback mechanism that place β-catenin at a key cellular integration point coordinating proliferation with autophagy, with implications for targeting these pathways for cancer therapy.

Figure 1

Modulation of β-catenin expression regulates autophagosome number. (A–C) Measurement of autophagy during nutrient starvation in HT29 cells stably expressing YFP–LC3. (A) Increased LC3 puncta per cell following β-catenin knockdown with siRNA compared with a non-targeting (NT) control siRNA. Columns show autophagosome numbers (mean±s.e.m., n>200 cells in >20 fields of view per experiment of three independent treatments) assessed under normal and starved (2 h) conditions with β-catenin or non-targeting (NT) siRNA. Representative images (B) of YFP–LC3 puncta (green) and DAPI nuclei staining (blue) are shown. (C) Upper panel: western blotting showed increased LC3-II expression in HT29 cells with β-catenin knockdown compared to control cells after 2 h starvation. Lower panel: quantification of the LC3-II/β-actin ratio by densitometry (mean±s.e.m. of four independent treatments, *P=0.011 in nutrient conditions, *P=0.035 under starvation). (D and E) HCT116 β-catenin^WT/− cells overexpressing a control or β-catenin^S33Y plasmid nutrient starved for 24 h. (D) Immunostaining of LC3 (green) with DAPI nuclei staining (blue) showed decreased LC3 puncta with β-catenin^S33Y overexpression. (E) By western blotting, LC3-II decreased with β-catenin^S33Y overexpression (upper panel). Lower panel: quantification of the LC3-II/β-actin ratio by densitometry (mean±s.e.m. of four independent treatments, *P=0.015). (F) Comparison of isogenic β-catenin^WT/− and β-catenin^−/ΔS45 HCT116 cells. Upper panel: western blotting showed decreased LC3-II in HCT116 β-catenin^−/ΔS45 cells. Lower panel: western blotting confirmed by quantification using densitometry of the LC3-II/β-actin ratio (mean±s.e.m. of three independent treatments, *P=0.046). Source data for this figure is available on the online supplementary information page.

Figure 2

β-Catenin negatively regulates p62 expression but does not block autophagic flux. (A) Protein expression of p62 increased following β-catenin knockdown under normal and starved conditions in HT29 cells. (B) Comparison of p62 expression in isogenic HCT116 β-catenin^WT/− and HCT116 β-catenin^−/ΔS45 cells by western blotting showed decreased p62 expression in β-catenin^−/ΔS45 cells. (C) LC3 puncta number per cell during starvation and nutrient addback post starvation. Left panel: LC3 puncta number reduces in non-targeting (NT) siRNA control with nutrient addback. Right panel: LC3 puncta number reduces in β-catenin knockdown cells with nutrient addback (mean±s.e.m. of three independent treatments; n>200 cells in >20 fields of view per experiment). Western blotting for β-catenin confirmed knockdown. (D) Western blotting of LC3 and p62 in HT29 cells following β-catenin knockdown and starvation with lysosomal inhibitors. LC3-II and p62 protein expression showed an increase with β-catenin siRNA in the presence of autophagy flux inhibitors chloroquine (10 μM) or bafilomycin A1 (100 nM; applied for the final 30 min of the 8 h) compared to β-catenin knockdown alone. Quantification by densitometry of the LC3-II/β-actin ratio (mean±s.e.m. of four independent experiments). See also Supplementary Figure S1A. Source data for this figure is available on the online supplementary information page.

Figure 3

β-Catenin deletion increases LC3 puncta and p62 expression in mouse intestinal epithelium in vivo. (A–D) LC3 staining (red) in mouse intestinal crypt and villus epithelium increased (white arrowheads) 2 days post β-catenin deletion in β-catenin^−/lox-villin-creERT2 mice (C and D) compared to control β-catenin^+/lox-villin-creERT2 mice (A and B). Further magnification of the areas marked by white squares is shown (lower panels). (E–H) p62 staining (green) increased in the crypt and villus epithelium 2 days after β-catenin deletion in β-catenin^−/lox-villin-creERT2 mice (G and H) compared to control in β-catenin^+/lox-villin-creERT2 mice (E and F). DAPI staining (blue) identifies nuclei. β-Catenin deletion was confirmed by immunofluorescence (Supplementary Figure S1B). Red and green channel levels were adjusted post acquisition for clarity (equal level changes applied across the entire figure). Further magnification of the areas marked by white squares is shown (lower panels). (I and J) Western blotting on extracts from mouse intestinal epithelial tissue demonstrated increased p62 and LC3-II protein expression 2 (I) and 4 (J) days post β-catenin deletion in β-catenin^−/lox-villin-creERT2 mice compared to control β-catenin^+/lox-villin-creERT2 mice. Source data for this figure is available on the online supplementary information page.

Figure 4

The Wnt/β-catenin/TCF pathway controls p62 expression. (A) Relative p62/SQSTM1 mRNA levels by quantitative real-time-PCR (qRT–PCR) increased following β-catenin knockdown in HT29 cells (mean±s.e.m., three independent experiments performed in triplicate, ***P<0.001). β-Catenin knockdown was confirmed by western blotting. (B) Relative p62 mRNA levels by qRT–PCR decreased with β-catenin^S33Y overexpression after 24 h of starvation in HCT116 β-catenin^WT/− cells (mean±s.e.m., three independent experiments performed in triplicate, ***P<0.001). β-Catenin^S33Y overexpression was confirmed by western blotting. (C) The increase in p62 protein expression induced by β-catenin siRNA was attenuated in HT29 cells treated with 10 μg/μl cycloheximide and starvation for 8 and 24 h compared to vehicle control. A complementary experiment using the transcription inhibitor actinomycin D is shown in Supplementary Figure S1C. (D) Overexpression of Fzd6 (green, left panel) increased p62 (red, right panel) protein expression in HEK293T cells. As an internal control, cells not overexpressing Fzd6 are shown in the field of view. (E) Doxycycline induction (1 μg/ml) of DNTCF4 increased p62 and LC3-II protein expression in doxycycline-inducible DNTCF4 LS174T-L8 cells. LGR5 downregulation by DNTCF4 confirmed inhibition of β-catenin/TCF4 signalling. (F) Relative p62 mRNA levels (48 h) by qRT–PCR increased following doxycycline induction of DNTCF4 in LS174T-L8 cells (mean±s.e.m., three independent experiments performed in triplicate, ***P<0.001). (G) Upper panel: Wnt3a (200 ng/ml) treatment of HCT116 β-catenin^WT/− cells (24 h) decreased LC3-II and p62 protein expression by western blotting. Lower panel: quantification by densitometry of the LC3-II/β-actin ratio (mean±s.e.m. of three independent experiments). Source data for this figure is available on the online supplementary information page.

Figure 5

β-Catenin and TCF4 associate with the p62/SQSTM1 gene promoter. (A) Chromatin immunoprecipitation (ChIP) demonstrates β-catenin and TCF4 binding to the p62 promoter. HT29 cells were grown in normal or starvation conditions for 2 h and subjected to ChIP analysis with the indicated antibodies (IgG was used as a negative control). Binding of RNA Pol II, β-catenin and TCF4 to the p62 promoter region was measured by quantitative PCR and expressed as percent enrichment relative to the input chromatin. During nutrient deprivation, binding of RNA Pol II to the p62 promoter increased; binding of β-catenin to the p62 promoter decreased; TCF4 binding did not change (data are from one representative experiment of at least three independent experiments performed in triplicate). PCR products subjected to agarose gel electrophoresis are shown in Supplementary Figure S2A. (B) Relative p62 mRNA expression by qRT–PCR increases ∼14-fold after 24 h starvation in HT29 cells (mean±s.e.m., three independent experiments performed in triplicate, ***P<0.001). (C) Binding of acetyl-Histone H3 to the p62 promoter increased under starvation conditions, suggesting p62 gene derepression (data are from one representative experiment of at least three independent experiments performed in triplicate). (D) Binding of β-catenin to the p62 promoter (relative to β-catenin binding to a non-target gene promoter, GAPDH) significantly decreased under starvation conditions (mean±s.e.m., three independent experiments performed in triplicate, **P<0.01).

Figure 6

Autophagy induction reduces Wnt/β-catenin signalling. (A) Western blotting showing that autophagy was induced after 2 h starvation, as evidenced by decreased p62 protein and increased LC3-II protein expression. β-Catenin protein expression decreased over a 24-h period of starvation in HT29 cells. (B) Autophagy decreased following Atg7 knockdown as shown by increased p62 and decreased LC3-II protein expression. β-Catenin protein expression increased following Atg7 siRNA. Reduction of LC3 puncta with Atg7 siRNA was confirmed by immunofluorescence and is shown in Supplementary Figure S2B. (C, D) Reduction of TopFlash activity in HT29 cells after 24 h of autophagy induction using (C) starvation or (D) 100 nM mTOR inhibitor PP242 (C and D, mean±s.e.m., three independent experiments performed in triplicate, ***P<0.001). Autophagy induction was confirmed by western blotting and is shown in Supplementary Figures S2C and S2D). (E, F) qRT–PCR shows reduction of Wnt-target gene expression (E) Axin2 (*P=0.017) and (F) Cyclin D1 (**P=0.0036) after 8 h starvation (mean±s.e.m., three (Axin2) or six (Cyclin D1) independent experiments performed in triplicate). (G–I) Inhibition of Wnt-induced TopFlash activity by autophagy induction in (G) HCT116 β-catenin^WT/− cells (mean±s.e.m., three independent experiments performed in triplicate, **P<0.01) and (H) RKO cells (mean±s.e.m., three independent experiments performed in triplicate, *P<0.05). (I) Reduction of Wnt3a-induced Cyclin D1 gene expression RKO cells after 12 h treatment with autophagy induction using starvation or PP242 (mean±s.e.m., two independent experiments). (J, K) TopFlash activity following Atg7 knockdown in (J) HT29 cells (mean±s.e.m., four independent experiments performed in triplicate, *P=0.0488) and (K) RKO cells with 24 h Wnt3a treatment (mean±s.e.m., three independent experiments performed in triplicate, *P=0.0282). (L) A representative western blot of RKO cells following Wnt3a treatment and Atg7 siRNA. (M) Relative β-catenin mRNA levels did not change after 8 h starvation by qRT–PCR (mean±s.e.m., three independent experiments performed in triplicate). (N) Western blot showing prevented β-catenin protein degradation during starvation in the presence of lysosomal autophagy inhibitors chloroquine (10 μM) and bafilomycin A1 (100 nM) compared to starvation alone. (O) Inhibition of autophagy using wortmannin (50 nM) prevented β-catenin protein degradation during starvation. (P) Western blot analysis of RKO cells expressing a myc–tagged β-catenin mutant (S33A, S37A, T41A, S45A) resistant to proteasomal degradation (myc–β-catenin^AAAA). RKO cells were subject to nutrient starvation for 2, 8 and 24 h, and autophagy induction was confirmed by increased LC3-II and decreased p62 protein expression. Both endogenous β-catenin and myc–β-catenin^AAAA protein expression decreased during starvation. (Q) Proteasome inhibition with MG132 (10 μM) did not prevent starvation-induced decrease in β-catenin protein expression. Source data for this figure is available on the online supplementary information page.

Figure 7

LC3 and β-catenin co-localise in the mouse intestinal epithelium. (A and B) Immunofluorescence of LC3 (red) and β-catenin (green) expression in the intestinal epithelium following 2 days tamoxifen treatment in control (A: β-catenin^+/lox-villin-creERT2) and β-catenin deleted (B: β-catenin^−/lox-villin-creERT2) mice. (C and D) Magnified areas from B revealing co-localisation of LC3 (red) and β-catenin (green). Arrowheads indicate co-localisation of LC3 and β-catenin. (E) Linescan analyses from C and D showing staining intensity of indicated co-localised puncta. Red and green channel levels were adjusted post acquisition (equal changes applied across the entire figure) and the blue channel was removed for clarity.

Figure 8

β-Catenin directly interacts with the autophagy protein LC3. (A) Co-immunoprecipitation of YFP–LC3 or negative control YFP in HT29 cells. Binding of endogenous β-catenin to YFP–LC3 was detected after 8 h of autophagy induction by starvation and starvation in the presence of lysosomal autophagy flux inhibitor chloroquine (10 μM). Input and immunodepleted lysates are shown. (B) Immunoprecipitation of endogenous LC3 in HT29 cells. IgG was used as a negative control. (C) β-Catenin contains a W/YXXI/L motif (putative LC3-interacting region) at amino-acid positions 504–507. β-Catenin wild-type (green highlight) and W504A/I507A double point mutant (blue highlight) sequences used in D–F are shown. (D) Pulldown assays using recombinant GST or GST–LC3B and lysates from HEK293 cells expressing HA–tagged wild-type β-catenin (HA–β-catenin^WT) or W504A/I507A β-catenin (HA–β-catenin^W504A/I507A). HA–β-catenin^W504A/I507A exhibited reduced GST–LC3B binding compared to HA–β-catenin^WT. (E) Recombinant myc–tagged His–LC3 (His–LC3–myc) interacted with recombinant GST–β-catenin^WT in vitro, but binding to GST–β-catenin^W504A/I507A was reduced (upper panel). Input recombinant proteins visualised with Coomassie blue are shown (lower panels). (F) Co-immunoprecipitation experiments using lysates from HEK293 cells transiently expressing YFP–LC3 with β-catenin^WT or β-catenin^W504A/I507A starved for 8 h with chloroquine (10 μM). YFP–LC3 immunoprecipitated β-catenin^WT but not β-catenin^W504A/I507A. Input lysates are shown and immunoprecipitated p62 and YFP–LC3 served as positive controls. (G) HA–β-catenin^{S33A/S37A/T41A/S45A/W504A/I507A} (HA–β-catenin^6A) was more resistant to the starvation-induced reduction in β-catenin protein expression than HA–β-catenin^S33Y. (H) Relative cell death in HT29 cells subject to nutrient starvation for 24 h. Atg7 knockdown increased cell survival. The increase in cell survival following Atg7 knockdown was significantly reversed by simultaneously depleting β-catenin (double knockdown of Atg7 and β-catenin). Data are the mean±s.e.m. of three independent experiments performed in triplicate, ***P<0.001; **P<0.01. Source data for this figure is available on the online supplementary information page.

Figure 9

Working model summarising the crosstalk between Wnt/β-catenin signalling and autophagy described in this study. β-Catenin is a cellular integration point coordinating proliferative signalling with autophagy. Under normal physiological conditions (nutrient rich) when autophagy is required at basal levels only, β-catenin limits autophagy and functions as a transcriptional co-repressor of p62. During nutrient deprivation (starvation), the inhibitory function of β-catenin on autophagy is reduced, p62 becomes derepressed, and β-catenin is targeted for autophagic degradation.