BNIP3-mediated mitophagy boosts the competitive growth of Lenvatinib-resistant cells via energy metabolism reprogramming in HCC

Abstract

An increasing evidence supports that cell competition, a vital selection and quality control mechanism in multicellular organisms, is involved in tumorigenesis and development; however, the mechanistic contributions to the association between cell competition and tumor drug resistance remain ill-defined. In our study, based on a contructed lenvitinib-resistant hepatocellular carcinoma (HCC) cells display obvious competitive growth dominance over sensitive cells through reprogramming energy metabolism. Mechanistically, the hyperactivation of BCL2 interacting protein3 (BNIP3) -mediated mitophagy in lenvatinib-resistant HCC cells promotes glycolytic flux via shifting energy production from mitochondrial oxidative phosphorylation to glycolysis, by regulating AMP-activated protein kinase (AMPK) -enolase 2 (ENO2) signaling, which perpetually maintaining lenvatinib-resistant HCC cells' competitive advantage over sensitive HCC cells. Of note, BNIP3 inhibition significantly sensitized the anti-tumor efficacy of lenvatinib in HCC. Our findings emphasize a vital role for BNIP3-AMPK-ENO2 signaling in maintaining the competitive outcome of lenvitinib-resistant HCC cells via regulating energy metabolism reprogramming; meanwhile, this work recognizes BNIP3 as a promising target to overcome HCC drug resistance.

Fig. 1. Cell competition drives the dominant growth of lenvatinib-resistant cells in HCC.

A Schematic workflow of our experimental strategy. Huh7 or Huh7R was cocultured with Huh7m at 1:1 ratio or cultured alone respectively. Cells were cultured for 48 h after cell attachment and then performed by high content imaging analysis and flow cytometry analysis. B High content imaging of cell growth over a time course of coculture after cell attachment in CC group (CCHuh7R+CCHuh7m) using 20X objective lens (Details see Video 5). C (Left) Representative H&E stain and fluorescence imaging of cells in NCC group-and CC group-mice. (Right) Mcherry positive cells in NCC group and CC group were counted with ImageJ. N = 5. D High content imaging of cell growth at 0 h, 24 h, 48 h after cell attachment in Alone group (Huh7, Huh7m, Huh7R), NCC group (NCCHuh7+NCCHuh7m) and CC group (CCHuh7R+CCHuh7m) using 10X objective lens (Details see Video 6–10). E Quantitative statistics of cell proliferation rate in (D). F, G Schematic workflow (F), tumor photos and statistic analysis of tumor volume (G) in our HCC orthotopic model. N = 5. H (Left) Representative H&E stain and fluorescence imaging of cells in Huh7R group-and CC group-mice. (Right) Mcherry positive cells in CC group were counted with ImageJ. N = 5. I Flow cytometry analysis of cell proportion in NCC group, CC group and single Huh7m mixed with single Huh7R at 24 h and 48 h after cell attachment. mcherry+ (m+): Huh7m, NCCHuh7m, CCHuh7m; mcherry- (m-): Huh7R, NCCHuh7, CCHuh7R. J Quantitative statiststics of ratio (m+/m-) in (F). Three independent experiments were conducted, and the values are represented by means ± SEM using t-test (C, G) or two-way ANOVA with Tukey’s multiple comparisons test (E, J). *p < 0.05, ns non-statistically significant. See also Figs. S1 and S2.

Fig. 2. Lenvatinib-resistant cells display upregulated glycolytic metabolism in cell competition at transcriptomic level.

A Schematic workflow of our experimental strategy by Figdraw. Huh7, Huh7m and Huh7R cultured alone; Huh7 or Huh7R cocultured with Huh7m at 1:1 ratio respectively. Cells were cultured for 48 h after cell attachment and then coculture groups were separated using flow cytometry sorting for RNA-seq, Western blot and other purposes. B Volcano plot showing differentially expressed genes in CCHuh7R vs CCHuh7R group. C As in (B) but in CCHuh7R vs NCCHuh7 group. D As in (B) but in CCHuh7R vs CCHuh7m group. E As in (B) but in Huh7R vs Huh7 group. F GSEA analysis of oxidative phosphorylation enrichment in CCHuh7R vs CCHuh7m group. G GSEA analysis of glycolysis enrichment in CCHuh7R vs CCHuh7m group. H As in (G) but in CCHuh7R vs NCCHuh7 group. I As in (G) but in CCHuh7R vs Huh7R group. J Flow cytometry analysis of 2-NBDG-FITC-A and cell proportion at 48 h after cell attachment in Alone group, NCC group and CC group. K Mountain map of 2-NBDG-FITC-A in (J). * CCHuh7m vs NCCHuh7m (p < 0.05); $ CCHuh7m vs Huh7m (p < 0.05); ¥ CCHuh7R vs NCCHuh7 (p < 0.05); # CCHuh7R vs Huh7R (p < 0.05); & Huh7R vs Huh7 (p < 0.05). L Cellular lactic acid production levels at 48 h after cell attachment in Alone group, NCC group and CC group by flow cytometry cell sorting. M Protein levels of ENO2, GLUT1, HK1, HK2, MCT1, MCT4, PFKP, PKM2, LDHA, LDHB and GAPDH at 48 h after cell attachment in Alone group, NCC group and CC group detected by western blot. N Quantitative statistics of protein levels in (M). Three independent experiments were conducted, and the values are represented by means ± SEM using two-way ANOVA with Tukey’s multiple comparisons test. *p < 0.05, #p < 0.05, $p < 0.05, ¥p < 0.05, &p < 0.05, ns non-statistically significant. See also Fig. S3.

Fig. 3. Enhanced glycolysis supports the winner status of lenvatinib-resistant cells.

A Flow cytometry analysis of 2-NBDG-FITC-A and cell proportion at 48 h after cell attachment in CC group with low glucose or Calcitriol treatment. B Quantitative statistics of 2-NBDG-FITC-A intensity in (A). * CCHuh7R (Low glucose) vs CCHuh7R (Control) (p < 0.05); # CCHuh7m (Low glucose) vs CCHuh7m (Control) (p < 0.05); ns CCHuh7R (Calcitriol) vs CCHuh7R (Control) (p > 0.05); NS CCHuh7m (Calcitriol) vs CCHuh7m (Control) (p > 0.05). C Quantitative statistics of ratio of m + /m- in (A). D Flow cytometry analysis of mitochondrial membrane potential (Rhodamine123-FITC-A) and cell proportion at 48 h after cell attachment in CC group with low glucose or Calcitriol treatment. E Quantitative statistics of 2-NBDG-FITC-A in (D). ns CCHuh7R (Low glucose) vs CCHuh7R (Control) (p > 0.05); NS CCHuh7m (Low glucose) vs CCHuh7m (Control) (p > 0.05); * CCHuh7R (Calcitriol) vs CCHuh7R (Control) (p < 0.05); # CCHuh7m (Calcitriol) vs CCHuh7m (Control) (p < 0.05). F Quantitative statistics of ratio of m + /m- in (D). G Flow cytometry analysis of cell death (SYTOX Green-FITC-A) and cell proportion at 48 h after cell attachment in CC group with low glucose or Calcitriol treatment. H Quantitative statistics of Dead m + /m+ ratio in (G). I Quantitative statistics of m + /m+ ratio in (G). J High content imaging of cell growth at 0, 24 h, 48 h after cell attachment in CC group with low glucose or Calcitriol treatment using 10X objective lens. K Quantitative statistics of cell number in (J) (Details see Video 11–13). Three independent experiments were conducted, and the values are represented by means ± SEM using two-way ANOVA with Tukey’s multiple comparisons test (B, E and K) or Ordinary one-way ANOVA with Sidak’s multiple comparisons test (C, F, H and I). *p < 0.05, #p < 0.05, ns/NS non-statistically significant.

Fig. 4. Enhanced glycolytic flux of lenvatinib-resistant cells (winner) requires BNIP3-mediated mitophagy.

A UpSet plot showing Top 50 DEGs’ overlapping genes in CCHuh7R vs Huh7R group and CCHuh7R vs NCCHuh7 group. B (Left) High content immunofluorescence imaging of colocalization of autophagosomes (Cy5.5-LC3: purple) and mitochondria (TOMM20: green) at 48 h after cell attachment in Alone group, NCC group and CC group using 63X water immersion objective lens. (Right) Quantification of colocalization between LC3 (purple peak) and TOMM20 (green peak) in the above groups. Purple/green peak height represents fluorescence intensity of LC3B/TOMM20; overlapping peaks indicate colocalization numbers between LC3B and Tomm20. C Transmission electron microscope imaging of Huh7R and CCHuh7R, the arrow (orange) indicates mitochondria of cell, the arrow (red) represents mitophagy activity in cell. The enlargement picture of the dashed boxes in CCHuh7R shows the details of specific mitophagy activity. D Protein levels of LC3B, TOMM20 and BNIP3 at 48 h after cell attachment in Alone group, NCC group and CC group detected by western blot. E Quantitative statistics of protein levels in (C). F (Left) High content immunofluorescence imaging of colocalization of autophagosomes (Cy5.5-LC3: purple) and mitochondria (TOMM20: green) at 48 h after cell attachment in Huh7R (Alone group) /CCHuh7R (CC group) transfected with shRNA against BNIP3 (shBNIP3), BNIP3 overexpression plasmid (oeBNIP3) or oeBNIP3+shBNIP3 using 63X water immersion objective lens. (Right) Quantification of colocalization between LC3 (purple peak) and TOMM20 (green peak) in the above groups. Purple/green peak height represents fluorescence intensity of LC3B/TOMM20; overlapping peaks indicate colocalization numbers between LC3B and Tomm20. G Protein levels of LC3B, TOMM20, BNIP3, ENO2, GLUT1, HK1, HK2, MCT1, MCT4 and PFKP at 48 h after cell attachment in Huh7R (Alone group) /CCHuh7R (CC group) transfected with shBNIP3, oeBNIP3 or shBNIP3+oeBNIP3 detected by western blot. H Quantitative statistics of protein levels in (H). I Flow cytometry analysis of 2-NBDG-FITC-A and cell proportion at 48 h after cell attachment in Huh7R (Alone group) /CCHuh7R (CC group) transfected with shBNIP3, oeBNIP3 or shBNIP3+oeBNIP3. J Mountain map of 2-NBDG-FITC-A in (J). * CCHuh7R-oeBNIP3 vs CCHuh7R-Control (p < 0.05); # CCHuh7R-shBNIP3 vs CCHuh7R-Control (p < 0.05); & CCHuh7R-shBNIP3+oeBNIP3 vs CCHuh7R-shBNIP3 (p < 0.05); $ Huh7R-oeBNIP3 vs Huh7R-Control (p < 0.05); ¥ Huh7R-shBNIP3 vs Huh7R-Control (p < 0.05); % Huh7R-shBNIP3+oeBNIP3 vs Huh7R-shBNIP3 (p < 0.05). K Quantitative statistics of ratio m + /m- in (J). L Cellular lactic acid production levels at 48 h after cell attachment in Huh7R (Alone group) /CCHuh7R (CC group) transfected with shBNIP3, oeBNIP3 or shBNIP3+oeBNIP3 by flow cytometry cell sorting. (M-N) Schematic workflow (M), tumor photos and statistic analysis of tumor volume (N) in our HCC subcutaneous model under lenvatinib treatment with or without olomouine. N = 5. O (Left) Representative H&E stain and fluorescence imaging of cells in CC group-mice with or without lenvatinib or lenvatinib+olomouine. (Right) Mcherry positive cells in corrresponding groups were counted with Image (J). N = 5. Three independent experiments were conducted, and the values are represented by means ± SEM using two-way ANOVA with multiple comparisons test (E, H, L, J and M) or Ordinary one-way ANOVA with Sidak’s multiple comparisons test (K, N, O). *p < 0.05, #p < 0.05, $p < 0.05, ¥p < 0.05, &p < 0.05, %p < 0.05, ns non-statistically significant. See also Figs. S4 and S5.

Fig. 5. The effect of glycolysis of winner is ENO2 dependent.

A Heatmap exhibiting expression levels of glycolysis-related genes between CCHuh7R and Huh7R. B As in (A) but between CCHuh7R and NCCHuh7. C UpSet plot displaying overlapping glycolysis-related genes based on (A) and (B). D Flow cytometry analysis of 2-NBDG-FITC-A and cell proportion at 48 h after cell attachment in Huh7R (Alone group) /CCHuh7R (CC group) transfected with shRNA against ENO2 (shENO2), ENO2 overexpression plasmid (oeENO2) or oeENO2+shENO2. E Mountain map of 2-NBDG-FITC-A in (D). * CCHuh7R-oeENO2 vs CCHuh7R-Control (p < 0.05); # CCHuh7R-shENO2 vs CCHuh7R-Control (p < 0.05); & CCHuh7R-shENO2+oeENO2 vs CCHuh7R-shENO2 (p < 0.05); $ Huh7R-oeENO2 vs Huh7R-Control (p < 0.05); ¥ Huh7R-shENO2 vs Huh7R-Control (p < 0.05); % Huh7R-shENO2+oeENO2 vs Huh7R-shENO2 (p < 0.05). F Quantitative statistics of ratio m + /m- in (D). G Cellular lactic acid production levels at 48 h after cell attachment in Huh7R (Alone group) /CCHuh7R (CC group) transfected with shENO2, oeENO2 or shENO2+oeENO2 by flow cytometry cell sorting. H Protein levels of LC3B, TOMM20, BNIP3, ENO2, GLUT1, HK1, HK2, MCT1, MCT4 and PFKP at 48 h after cell attachment in Huh7R (Alone group) /CCHuh7R (CC group) transfected with shENO2, oeENO2 or shENO2+oeENO2 detected by western blot. I Quantitative statistics of protein levels in (H). J (Left) High content immunofluorescence imaging of colocalization of autophagosomes (Cy5.5-LC3: purple) and mitochondria (TOMM20: green) at 48 h after cell attachment in Huh7R (Alone group) /CCHuh7R (CC group) transfected with shENO2, oeENO2 or shENO2+oeENO2 using 63X water immersion objective lens. (Right) Quantification of colocalization between LC3 (purple peak) and TOMM20 (green peak) in the above groups. Purple/green peak height represents fluorescence intensity of LC3B/TOMM20; overlapping peaks indicate colocalization numbers between LC3B and Tomm20. Three independent experiments were conducted, and the values are represented by means ± SEM using a two-way ANOVA with Tukey’s multiple comparisons test (E, G and I) or Ordinary one-way ANOVA with Sidak’s multiple comparisons test (F). *p < 0.05, #p < 0.05, $p < 0.05, ¥p < 0.05, &p < 0.05, %p < 0.05, ns = non-statistically significant. See also Fig. S5.

Fig. 6. AMPK functions as a signaling bridge in BNIP3-ENO2 crosstalk in lenvatinib-resistant cells.

A GSEA analysis of AMPK signaling pathway enrichment in BNIP3 high expression vs BNIP3 low expression group of TCGA-LIHC cohort. B As in (A) but in ICGC-LIHC cohort. C Protein levels of AMPK and p-AMPK at 48 h after cell attachment in Alone group, NCC group and CC group detected by western blot. D Quantitative statistics of protein levels in (C). E Correlation analysis of BNIP3 expression and AMPK signaling pathway in TCGA-LIHC cohort. F As in (E) but in ICGC-LIHC cohort. G Flow cytometry analysis of 2-NBDG-FITC-A and cell proportion at 48 h after cell attachment in Huh7R (Alone group) /CCHuh7R (CC group) treated with AMPK-activator-2, AMPK-IN-3 or AMPK-IN-3 + AMPK-activator-2. H Mountain map of 2-NBDG-FITC-A in (G). * CCHuh7R+AMPK-activator-2 vs CCHuh7R-Control (p < 0.05); # CCHuh7R-AMPK-IN-3 vs CCHuh7R-Control (p < 0.05); & CCHuh7R+AMPK-IN-3 + AMPK-activator-2 vs CCHuh7R+AMPK-IN-3 (p < 0.05); $ Huh7R-AMPK-activator-2 vs Huh7R-Control (p < 0.05); ¥ Huh7R+AMPK-IN-3 vs Huh7R-Control (p < 0.05); % Huh7R+AMPK-IN-3 + AMPK-activator-2 vs Huh7R+AMPK-IN-3 (p < 0.05). I Quantitative statistics of ratio m + /m- in (G). J Cellular lactic acid production levels at 48 h after cell attachment in Huh7R (Alone group) /CCHuh7R (CC group) treated with AMPK-activator-2, AMPK-IN-3 or AMPK-IN-3 + AMPK-activator-2 by flow cytometry cell sorting. K (Top) High content immunofluorescence imaging of colocalization of autophagosomes (Cy5.5-LC3: purple) and mitochondria (TOMM20: green) at 48 h after cell attachment in Huh7R (Alone group) /CCHuh7R (CC group) treated with AMPK-activator-2 using 63X water immersion objective lens. (Bottom) Quantification of colocalization between LC3 (purple peak) and TOMM20 (green peak) in the above groups. Purple/green peak height represents fluorescence intensity of LC3B/TOMM20; overlapping peaks indicate colocalization numbers between LC3B and Tomm20. L Protein levels of AMPK, p-AMPK and ENO2 at 48 h after cell attachment in Huh7R (Alone group) /CCHuh7R (CC group) treated with AMPK-activator-2, shENO2 or shENO2 + AMPK-activator-2 detected by western blot. M Quantitative statistics of protein levels in (L). N Protein levels of AMPK, p-AMPK and BNIP3 at 48 h after cell attachment in Huh7R (Alone group) /CCHuh7R (CC group) treated with oeBNIP3, AMPK-IN-3 or oeBNIP3 + AMPK-IN-3 detected by western blot. O Quantitative statistics of protein levels in (N). Three independent experiments were conducted, and the values are represented by means ± SEM using a two-way ANOVA with Tukey’s multiple comparisons test (D, H, J, M and O) or Ordinary one-way ANOVA with Sidak’s multiple comparisons test (I). *p < 0.05, #p < 0.05, $p < 0.05, ¥p < 0.05, &p < 0.05, %p < 0.05, ns = non-statistically significant. See also Fig. S6.

Fig. 7. Cell competition mechanism between lenvatinib-resistant cell (Winner) and lenvatinib-sensitive cell (Loser) in HCC.

In HCC coculture system, overexpressed BNIP3 activates mitophagy activity through binding with LC3, which eliminates damaged mitochondria of lenvatinib-resistant HCC cell; mitochondria reduction results in the decrease of oxidative phosphorylation (OXPHOS) levels; subsequently, energy imbalance signal (ADP/ATP) caused by weaken OXPHOS triggers the phosphorylation of AMPK sensor; enabled AMPK specifically targets ENO2 and therefore endows resistant cell with enhanced glycolysis, which facilitates the competitive and grabbing abilities for glucose from environment and opponent, eventually boosting cell proliferation. In lenvatinib-sensitive HCC cell, this failed mechanism by which BNIP3 modulates energy metabolism shifting continually activates and overuses mitochondrial OXPHOS for energy compensation, which aggravates damaged mitochondria accumulation and ROS production; limited metabolism function and oxidative pressure disqualify lenvatinib-sensitive HCC cell’s competition for resource, which eventually induce cell apoptosis.