IKKα activation of NOTCH links tumorigenesis via FOXA2 suppression

Abstract

Proinflammatory cytokine TNFα plays critical roles in promoting malignant cell proliferation, angiogenesis, and tumor metastasis in many cancers. However, the mechanism of TNFα-mediated tumor development remains unclear. Here, we show that IKKα, an important downstream kinase of TNFα, interacts with and phosphorylates FOXA2 at S107/S111, thereby suppressing FOXA2 transactivation activity and leading to decreased NUMB expression, and further activates the downstream NOTCH pathway and promotes cell proliferation and tumorigenesis. Moreover, we found that levels of IKKα, pFOXA2 (S107/111), and activated NOTCH1 were significantly higher in hepatocellular carcinoma tumors than in normal liver tissues and that pFOXA2 (S107/111) expression was positively correlated with IKKα and activated NOTCH1 expression in tumor tissues. Therefore, dysregulation of NUMB-mediated suppression of NOTCH1 by TNFα/IKKα-associated FOXA2 inhibition likely contributes to inflammation-mediated cancer pathogenesis. Here, we report a TNFα/IKKα/FOXA2/NUMB/NOTCH1 pathway that is critical for inflammation-mediated tumorigenesis and may provide a target for clinical intervention in human cancer.

Figure 1. TNFα-Induced NOTCH1 Expression Requires IKKα

(A) Correlation between levels of p-IKKα/β and activated NOTCH1 (NICD) in 100 human primary HCC tumor specimens stained with antibodies specific to NICD and p-IKK (P = 0.003, Pearson’s chi-square test). (B) Two representative specimens from (A); arrows point to p-IKK and NICD. (C) Correlation between tumor grade and p-IKK expression. High p-IKK expression correlated significantly with higher tumor grade (P = 0.035, Pearson’s chi-square test). (D) Correlation between expression of p-IKK, NICD and tumor grade. High NICD expression correlated significantly with higher tumor grade (P = 0.01, Pearson’s chi-square test). (E) Top: WT MEFs were treated with or without TNFα (20 ng/ml) for 24 h, and cell lysates were subjected to western blot analysis for NICD expression. Bottom: IKKα and IKKβ were transfected into IKKα^−/− MEFs and IKKβ^−/− MEFs, respectively. After treatment with TNFα (20 ng/ml) for 24 h, cell lysates were subjected to western blotting. Lysates from MEFs without TNFα treatment served as control. α-Tubulin was used as loading control. (F) RBP-Jk Luc and IKKα siRNA were transfected into Hep3B cells for 48 h. After transfection, WT and IKKα-knockdown Hep3B cells were treated with TNFα for 24 h. Cell lysates were subjected to western blotting and luciferase reporter assay. Cells were transfected with nontargeting siRNA as control. Error bars represent SD (n = 3). * indicates, P < 0.05.

Figure 2. IKKα-Mediated NOTCH1 Activation Requires NUMB

(A) Hep3B and Huh-7 cells were treated with TNFα (20 ng/ml) for 24 h. Endogenous NICD and NUMB expression levels were examined by western blotting. α-Tubulin was used as loading control. (B) Knockdown of IKKα or double knockdown of IKKα and NUMB by siRNA in Hep3B cells. Nontargeting siRNA was used as control. After 24 h of TNFα treatment, cell lysates (30 µg) were subjected to western blotting to detect endogenous NICD and NUMB expression. α-Tubulin was used as loading control. (C) Western blot analysis of endogenous NICD and NUMB expression in WT MEFs and IKKα^−/− MEFs with or without reconstituted IKKα. α-Tubulin was used as loading control. (D) Hep3B cells were transfected with WT IKKα or KD IKKα. Endogenous NICD and NUMB expression was analyzed by western blotting. α-Tubulin was used as loading control. (E) RBP-Jk Luc and NUMB were transfected into Hep3B cells for 48 h. After transfection, Hep3B cells were treated with TNFα (20ng/ml) for 24 h. Cell lysates were subjected to the luciferase reporter assay. NUMB expression was analyzed by western blotting with 10 µg of total lysates. Error bars represent SD (n = 3). * indicates statistical significance (P < 0.05). (F) Hep3B cells were infected with retrovirus expressing IKKα or IKKα and NUMB. RNA extracts were purified from the cell lysates and quantitated with real-time PCR. The NOTCH1 target gene mRNA levels were normalized to the mRNA levels of target genes in vector-infected cells. Error bars represent SD (n = 3). * indicates statistical significance (P < 0.05). (G) Hep3B and Huh-7 cells were treated with TNFα (20 ng/ml) for 24 h. NUMB mRNA level was detected with qPCR. Error bars represent SD (n = 3). * indicates statistical significance (P < 0.05). (H) The TRANSFAC 7.0 and PROMO 8.3 programs were used to predict transcriptional factors that bind to the NUMB promoter. Five HNFs were identified from the prediction. (I) Each of the five indicated HNFs was transfected alone or together with IKKα into Hep3B cells. NUMB mRNA level was detected with real-time PCR. Error bars represent SD (n = 3). * indicates statistical significance (P < 0.05).

Figure 3. IKKα Interacts with and Phosphorylates FOXA2 In Vivo and In Vitro

(A) IKKα and FOXA2 were transfected into 293T cells and analyzed by reciprocal co-immunoprecipitation and immunoblotting using indicated antibodies. (B) Lysates of Hep3B cells were analyzed by reciprocal co-immunoprecipitation and immunoblotting using indicated antibodies. (C) In vitro transcribed FOXA2 proteins (labeled with S35) were incubated with commercially available recombinant His-IKKβ and GST-IKKα and then pulled down with His/GST beads. IVT: in vitro transcription and translation. S35 concentration in cell lysates was determined by film exposure. (D) FOXA2 phosphorylation was identified by in vitro kinase assay. GST-FOXA2 was pulled down by GST beads from Hep3B cells and incubated with commercially available IKKα for 30 min. (E) Mass spectrum of FOXA2 phosphorylation sites by IKKα using samples (without isotope) from (D). (F) Hep3B cells were co-transfected with FOXA2 and WT IKKα or KD IKKα. pFOXA2 expression was detected using an antibody specific to pFOXA2 (S107/111). Cell lysates (10 µg) wereloaded for western blotting. (G) IKKα was co-transfected with WT FOXA2 or FOXA2-SSAA. pFOXA2 (S107/111) was detected using an antibody specific to pFOXA2 (S107/111). Total FOXA2 expression and IKKα expression are also shown. (H) Western blot analysis of pFOXA2 with an antibody specific to pFOXA2 (S107/111) using samples (without isotope) from (D). Total FOXA2 expression and IKKα expression are also shown. (I) Hep3B cells were treated with TNFα and harvested at 10, 30, 60, and 120 min after treatment. Cell lysates (40 µg) were used for western blot analysis of pFOXA2 with an antibody specific to pFOXA2 (S107/111). Total FOXA2 expression and IKKα expression are also shown. (J) IKKβ^−/− MEFs were treated with TNFα for 45 min. Cell lysates were subjected to cell fractionation. After immunoprecipitation by the IKKα antibody, the nuclear and cytoplasmic fractions were subjected to western blot analysis with indicated antibodies. Lamin and tubulin were used as cell fractionation controls.

Figure 4. FOXA2 Phosphorylation by IKKα Suppresses FOXA2 Transactivation Activity

(A) NUMB promoter luciferase reporter and FOXA2 in the presence of different doses of WT IKKα or KD IKKα were transfected into 293T cells for 48 h. Cell lysates were subjected to luciferase reporter assays after transfection. Expression levels of WT IKKα, KD IKKα, and FOXA2 were analyzed by western blotting. Error bars represent SD (n = 3). * indicates statistical significance (P < 0.05). (B) Co-transfection of NUMB-promoter Luc and WT FOXA2, FOXA2-SSAA, or FOXA2-SSEE plus the indicated Flag-tagged IKKα in 293T cells for 48 h. Lysates of 293T cells were subjected to luciferase assays. Error bars represent SD (n = 3). * indicates statistical significance (P < 0.05). Expression levels of IKKα, WT FOXA2, and FOXA2 mutants were detected by western blot analysis. (C) Hep3B cells were infected with retrovirus expressing FOXA2-SSAA or FOXA2-SSEE. RNA extracts from the cells were isolated and subjected to real-time RT-PCR for NUMB mRNA expression. Error bars represent SD (n = 3). * indicates statistical significance (P < 0.05). (D) Hep3B cells were treated with TNFα for 24 h, and cell lysates were harvested and subjected to primary ChIP assay using the FOXA2 antibody. The purified DNA from the ChIP assay was subjected to real-time PCR using NUMB primers (to left of arrow). The immunoprecipitated products of Hep3B cells (with TNFα treatment) were desalted and re-ChIP-ed by IKKα and NOTCH1 antibody. Real-time PCR using NUMB primers was repeated using the purified DNA from the ChIP assay (to right of arrow). The products from the second immunoprecipitation were desalted again and re-subjected to re-ChIP assay by NOTCH1 and IKKα (as a positive control). The purified DNA from ChIP assay was subjected to real-time PCR using NUMB primers (below downward arrow). Error bars represent SD (n = 3). * indicates statistical significance (P < 0.05). (E) Endogenous FOXA2 in Hep3B cells was knocked down by shRNA against the 5′-UTR. After shRNA knockdown of FOXA2, we transfected WT FOXA2, FOXA2-SSAA, or FOXA2-SSEE into FOXA2^−/− Hep3B cells, followed by TNFα treatment for 24 h. Cell lysates were analyzed by western blotting using indicated antibodies. (F) RBP-Jk Luc and WT FOXA2, FOXA2-SSAA, or FOXA2-SSEE were co-transfected into Hep3B cells for 24 h, followed by TNFα treatment for 24 h. Cell lysates were analyzed by luciferase reporter assay. Error bars represent SD (n = 3). * indicates statistical significance (P < 0.05). (G) NOTCH1 expression was knocked down by NOTCH1 siRNA in Hep3B cells, which were treated with TNFα for 24 h, and lysates were analyzed by western blotting.

Figure 5. FOXA2 Blocks IKKα-Induced Cell Proliferation and Tumor Growth In Vitro

(A, B, and C). Hep3B cells were infected with retrovirus expressing WT FOXA2, FOXA2-SSAA, or FOXA2-SSEE in the presence of (A) control shRNA, (B) NUMB shRNA, or (C) NOTCH1 shRNA. Cells (2.5 × 10³) were plated in 96-well plates, and cell growth was determined by MTT assay each day for 4 days. (D. and E). Anchorage-independent growth assay of Hep3B cells infected with retrovirus expressing WT FOXA2, FOXA2-SSAA, or FOXA2-SSEE in the presence of (D) NUMB shRNA or (E) NOTCH1 shRNA, compared with control shRNA. Colony numbers (mean ± SD) in week 3 are shown (n = 6). * indicates statistical significance (P < 0.05, Student’s t test). Representative images from the anchorage-independent growth assay are shown below. (F) Hep3B cells were infected with retrovirus expressing WT FOXA2, FOXA2-SSAA, or FOXA2-SSEE either without (top) or with (bottom) NUMB shRNA. Infected cells were subjected to BrdU assay and analyzed by flow cytometry.

Figure 6. FOXA2 Blocks IKKα-Induced Cell Proliferation and Tumor Growth In Vivo

(A) Hep3B cells were infected with retrovirus expressing control or IKKα shRNA. Mice were injected subcutaneously with 1 × 10⁶ cells. Error bars represent SD (n = 5). Please see the Experimental Procedures section for more detailed information. (B) Tumor tissues from mice that received transplants were analyzed 10 days after transplantation and subjected to western blot analysis for indicated antibodies. Right: Expression levels of pFOXA2, NUMB, and NOTCH1 quantified with ImageJ software. (C) Hep3B cells were infected with retrovirus expressing WT FOXA2, FOXA2-SSAA, or FOXA2-SSEE in the presence of (top) control shRNA or (bottom) NUMB shRNA. Error bars represent SD (n = 5). Please see the Experimental Procedures section for more detailed information. (D) Tumor issues from (C) were harvested 10 days after transplantation and subjected to western blot analysis for NICD, FOXA2, and NUMB. α-Tubulin was included as control. (E) Comparison of pFOXA2 (S107/111), IKKα, and NICD expression in tumors and their adjacent normal tissues. Protein expression was examined by immunoblotting 30 pairs (tumor and normal) of liver tissue samples and quantified with ImageJ software. Right: Four representative pairs (N, normal; T, tumor). Protein expression levels in tumor samples were normalized to those in paired normal tissues. Protein expression levels higher (or lower) than those in normal tissues were defined as “high” (or “low”).(P < 0.01, Pearson’s chi-square test.) (F) Correlation analysis of pFOXA2 and IKKα/NICD expression levels in tumors (R² > 0.5, Spearman’s rank correlation test).

Figure 7. Proposed Model of FOXA2 Phosphorylation by IKKα

FOXA2 phosphorylation by IKKα leads to activation of NOTCH1 signaling through downregulation of NUMB and thereby induces tumorigenesis.