NEDD4L mediates intestinal epithelial cell ferroptosis to restrict inflammatory bowel diseases and colorectal tumorigenesis

Abstract

Various factors play key roles in maintaining intestine homeostasis. Disruption of the balance may lead to inflammatory bowel diseases and even colorectal cancer (CRC). Loss or gain of function of many key proteins can result in dysregulated intestinal homeostasis. Our research demonstrated that neural precursor cells expressed developmentally downregulated 4-like protein (NEDD4L, or NEDD4-2), a type of HECT family E3 ubiquitin ligase, played an important role in maintaining intestinal homeostasis. NEDD4L expression was significantly inhibited in intestinal epithelial cells (IECs) of patients with Crohn's disease, ulcerative colitis, and CRC. Global KO of NEDD4L or its deficiency in IECs exacerbated colitis induced by dextran sulfate sodium (DSS) and 2,4,6-trinitrobenzene sulfonic acid (TNBS) and CRC induced by azoxymethane and DSS. Mechanistically, NEDD4L deficiency in IECs inhibited expression of the key ferroptosis regulator glutathione peroxidase 4 (GPX4) by reducing the protein expression of solute carrier family 3 member 2 (SLC3A2) without affecting its gene expression, ultimately promoting DSS-induced IEC ferroptosis. Importantly, ferroptosis inhibitors reduced the susceptibility of NEDD4L-deficient mice to colitis and colitis-associated CRC. Thus, NEDD4L is an important regulator in IEC ferroptosis, maintaining intestinal homeostasis, making it a potential clinical target for diagnosing and treating IBDs.

Keywords: Cell biology; Epithelial transport of ions and water; Inflammation; Inflammatory bowel disease; Molecular biology.

Figure 1. NEDD4L expression is significantly downregulated in IECs of patients with IBDs.

(A and B) Statistical analysis of NEDD4L immunohistochemical (IHC) intensity in the biopsies from Xijing Hospital (cohort 1) (A) and representative IHC staining of sections traced with anti-NEDD4L antibody (B). Normal control (HC) n = 40 and UC n = 83. Scale bars: 50 μm. (C and D) Statistical analysis of NEDD4L IHC intensity in the biopsies from the First Affiliated Hospital of Zhejiang University School of Medicine (cohort 2) (C) and representative IHC staining of sections (D). Normal control (HC) n = 31, UC n = 36, and CD n = 41. Scale bars: 50 μm. (E and F) Statistical analysis of NEDD4L IHC intensity in the biopsies with disease status record from cohort 2 (E) and representative IHC staining of sections traced with anti-NEDD4L antibody (F). Mild n = 14 and moderate/severe n = 48. Scale bars: 50 μm. (G and H) Quantitative PCR (qPCR) analysis (G) and representative Western blotting (H) of NEDD4L in the mucosa from patients with IBDs and their corresponding normal tissues (n = 24 per group). (I and J) Western blotting analysis (I) and protein intensity analysis (J) according to I using ImageJ software (NIH) of NEDD4L from the IECs of the WT mice treated or not treated with DSS for 4 days (n = 5 per group). Red arrows indicate NEDD4L expression in IECs, and green arrows indicate NEDD4L expression in non-IECs. Data represent mean ± SEM. Each dot represents an independent sample. **P < 0.01; ***P < 0.001; ****P < 0.0001. Statistical analysis was performed using 1-way ANOVA with multiple comparisons in C, and a 2-tailed Student’s t test in A, E, G, and J.

Figure 2. Nedd4l deficiency in mice promotes DSS-induced experimental colitis in a non-hematopoietic cell–dependent manner.

(A) Nedd4l-global-deficient mice (Nedd4l^+/–) and control littermates (Nedd4l^+/+) were given 4% DSS for 5 days followed by water to induce acute colitis. Mouse death was monitored until day 9. n = 20 per group. (B–F) Nedd4l^+/– and Nedd4l^+/+ mice were given 3% DSS for 5 days followed by water until day 9. n = 9 per group. Body weight change (B), bleeding scores (C), colon length (D), gross morphology images (E), and H&E staining of colons (F) from Nedd4l^+/+ and Nedd4l^+/– mice are shown. Red arrows point to epithelial degeneration and green arrows to inflammatory infiltrates. Scale bars: 200 μm or 50 μm (amplified sections). (G–J) Bone marrow from Nedd4l^+/+ (WT) and Nedd4l^–/– (KO) mice was transferred to WT (n = 7) and KO (n = 10) mice to generate bone marrow reconstitution mice. The bone marrow reconstitution mice were subjected to 3% DSS treatment for 5 days followed by water, and mouse death (G) and body weight changes (H) were monitored until day 9. (I and J) From a separate experiment, colon length (I) and gross morphology images (J) of colons from mice on day 6 after DSS treatment. n = 4 per group. Data represent mean ± SEM from at least 2 independent experiments. Each dot represents an independent sample. **P < 0.01; ****P < 0.0001. Statistical analysis was performed using a log-rank test in A and G, a 2-way ANOVA test in B, C, and H, and a 2-tailed Student’s t test in D and I.

Figure 3. Nedd4l deficiency in IECs promotes DSS-induced colitis in mice.

(A) Nedd4l IEC-deficient mice (Nedd4l^fl/fl Villin^Cre, n = 8) and control littermates (Nedd4l^fl/fl, n = 7) were given 2.5% DSS for 5 days followed by water to induce acute colitis. Mouse death was monitored until day 12. (B–F) In a separate experiment, Nedd4l^fl/fl Villin^Cre mice (n = 7) and control Nedd4l^fl/fl mice (n = 8) were given 2% DSS for 5 days followed by water until day 9 to induce colitis. Body weight change (B), bleeding scores (C), colon length (D), gross morphology images (E), and H&E staining of the colons (F) from Nedd4l^fl/fl Villin^Cre and Nedd4l^fl/fl mice are shown. Red arrows point to epithelial degeneration and green arrows to inflammatory infiltrates. Scale bars: 200 μm or 50 μm (amplified sections). (G and H) Colon-infiltrated immune cells of Nedd4l^fl/fl Villin^Cre and Nedd4l^fl/fl mice from B were analyzed by flow cytometry analysis (n = 3–4 per group). Data represent mean ± SEM from at least 2 independent experiments. Each dot represents an independent sample. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Statistical analysis was performed using a log-rank test in A, a 2-way ANOVA test in B and C, and a 2-tailed Student’s t test in D, G, and H.

Figure 4. Nedd4l deficiency in IECs promotes IEC ferroptosis, resulting in barrier integrity damage.

(A) KEGG analysis of colonic tissues on the 7th day from Nedd4l^fl/fl Villin^Cre and Nedd4l^fl/fl mice given 2% DSS. (B) The indicated mice were treated as in A and were orally fed with FITC-dextran (500 mg/kg) for 4 hours before sacrifice. The serum levels of FITC-dextran were detected by measurement of the mean fluorescence intensity (MFI) of FITC-dextran. (C) In a separate experiment, the indicated mice were treated as in A, and colon tissues were further subjected to ZO-1 immunofluorescence (IF) staining. Red IF indicates ZO-1, and blue (DAPI) indicates cell nucleus. Scale bars: 50 μm. (D) KEGG analysis of ubiquitylation mass spectrometry (MS) from IECs of the indicated mice treated as in A. (E–H) Colon tissues from DSS-treated Nedd4l^fl/fl Villin^Cre and Nedd4l^fl/fl mice were subjected to TUNEL (E and F) and 4-hydroxynonenal (4-HNE) (G and H) IHC staining. TUNEL (F) and 4-HNE (H) IHC staining was scored and analyzed. Scale bars: 50 μm. (I) In a separate experiment, IECs from DSS-treated Nedd4l^fl/fl Villin^Cre and Nedd4l^fl/fl mice were subjected to MDA analysis. (J) Representative transmission electron microscope images from colonic tissue sections of DSS-treated Nedd4l^fl/fl Villin^Cre and Nedd4l^fl/fl mice. Scale bars: 2 μm or 0.5 μm (amplified sections). (K and L) Representative microscope images (K) and flow cytometer analysis (L) of small-intestinal organoids isolated and cultured from crypts of Nedd4l^fl/fl Villin^Cre and Nedd4l^fl/fl mice treated with DMSO (Control), DSS (0.5% wt/vol), erastin (30 μM), erastin 2 (30 μM), and RSL3 (5 μM) for 24 hours, followed by DAPI and BODIPY C11 staining. n = 3 per group. Scale bars: 100 μm. Data represent mean ± SEM from at least 2 independent experiments. Each dot represents an independent sample. *P < 0.05; **P < 0.01; ***P < 0.001. Statistical analysis was performed using a 2-tailed Student’s t test in B, E, H, I, and L.

Figure 5. NEDD4L positively regulates SLC3A2 expression.

(A) Volcano plots of protein abundance fold change based on ubiquitination MS of Figure 4D. (B) Venn analysis shows the potential targets of NEDD4L based on interaction MS analysis in HCT116 cells stably expressing FLAG-tagged NEDD4L and ubiquitination MS analysis. The list shows the overlapped targets of NEDD4L in A and B. (C) Representative IHC staining of SLC3A2 from Nedd4l^fl/fl Villin^Cre and Nedd4l^fl/fl mice treated with DSS on day 5. Scale bars: 100 μm or 50 μm (amplified sections). (D and E) Western blotting analysis (D) and statistical analysis (E) of intensity of the indicated proteins in IECs from Nedd4l^fl/fl Villin (n = 4) and Nedd4^fl/fl (n = 7) mice treated as in Figure 3B. (F and G) Representative IHC staining (F) and correlative analysis (G) of SLC3A2 and NEDD4L from colonic sections from CD patients (n = 13). Scale bars: 50 μm. (H) Immunoblot analysis of the indicated proteins in small-intestinal organoids isolated and cultured from crypts of Nedd4l^fl/fl Villin^Cre and Nedd4l^fl/fl mice, with 0.5% DSS treatment for the indicated times. (I and J) NEDD4L-KO (sgNEDD4L) and negative control (sgNTC) HCT116 cell lines, or HCT116 cells transfected with Myc-tagged NEDD4L, Myc-tagged NEDD4L-C942A (Myc-NEDD4L-CA), or Myc-tagged null control plasmids (Ctrl), were treated with 2% DSS for the indicated times and then subjected to immunoblot analysis of the indicated proteins. (K–M) Immunoblot analysis of the indicated proteins in HCT116 cells (K), SW480 cells (L), and RKO cells (M) transfected with siRNA targeted to NEDD4L (siNEDD4L) or negative control (siNC) and treated as in I. Data represent mean ± SEM from at least 2 independent experiments. Each dot represents an independent sample. *P < 0.05; **P < 0.01; ***P < 0.001. Statistical analysis was performed using a 2-tailed Student’s t test in E, and a Pearson’s correlation test in G.

Figure 6. SLC3A2 negatively regulates ferroptosis.

(A–C) The multiple cell lines, including HCT116 cells (A), SW480 cells (B), and RKO cells (C), were transfected with siRNA targeted to SLC3A2 (siSLC3A2) or negative control (siNC). The cells were treated with 2% DSS for the indicated times and then subjected to CCK8 assay. (D–F) The multiple cell lines were treated as in A–C with or without Fer-1 (2 μM) treatment. The cells were then subjected to flow cytometry analysis of BODIPY C11 staining to measure lipid peroxidation production. (G–I) The multiple cell lines were treated as in A–C for the indicated times and then subjected to immunoblot analysis of the indicated proteins. (J–M) FLAG-tagged SLC3A2 or FLAG-tagged null control plasmids were overexpressed in HCT116 cells. The cells were treated with 2% DSS or indicated inducers for the stated times, and then subjected to CCK8 assay (J), MDA assay (K), flow cytometry analysis of BODIPY C11 staining (L), and immunoblot analysis of the indicated proteins (M). (N–P) HCT116 cells were transfected with oligonucleotides specific for siRNA negative control (siNC) or NEDD4L (siNEDD4L), and then FLAG-tagged SLC3A2 or FLAG-tagged null control plasmid was overexpressed in the cells. The cells were treated with 2% DSS for the indicated times and then subjected to CCK8 assay (N) and lipid peroxidation (O). (P) Immunoblot analysis of the indicated proteins. Data represent mean ± SEM from at least 2 independent experiments. Each dot represents an independent sample. *P < 0.05; **P < 0.01; ***P < 0.001. Statistical analysis was performed using a 2-tailed Student’s t test in A–F, J–L, N, and O.

Figure 7. NEDD4L ubiquitinates SLC3A2.

(A) Immunoblot analysis of NEDD4L and SLC3A2 coimmunoprecipitated with anti-SLC3A2 antibody from lysates of HCT116 cells treated with 2% DSS for the indicated times. (B and C) Immunoblot analysis of Myc-tagged proteins and FLAG-tagged SLC3A2 coimmunoprecipitated with anti-Myc antibody from lysates of HEK293T cells cotransfected with indicated plasmids. WCL, whole-cell lysate. (D) Immunoblot analysis of NEDD4L, SLC3A2, and Ub, which were coimmunoprecipitated with anti-Ub antibody from lysates of NEDD4L (sgNEDD4L) or negative control (sgNTC) KO HCT116 cells treated with 2% DSS for the indicated times. (E) Immunoblot analysis of total ubiquitination of FLAG-tagged SLC3A2 following coimmunoprecipitation with anti-FLAG antibody from lysates of HEK293T cells cotransfected with indicated plasmids. (F) Immunoblot analysis of Ub-linked FLAG-tagged EGFP or SLC3A2 incubated with Myc-tagged NEDD4L, Myc-tagged NEDD4L-C942A (CA), or Myc-tagged EGFP recombinant protein in the presence of the full complement of ubiquitination reaction components, including E1, E2, Ub, and ATP, in vitro. (G and H) Immunoblot analysis of ubiquitination of FLAG-tagged SLC3A2 following coimmunoprecipitation of SLC3A2 with anti-FLAG antibody from lysates of HEK293T cells cotransfected with indicated plasmids. (I) Immunoblot analysis of K63Ub, K48Ub, Ub, GPX4, TFRC, SLC3A2, NEDD4L, and actin, which were coimmunoprecipitated with anti-SLC3A2 antibody from lysates of NEDD4L (sgNEDD4L) or negative control (sgNTC) KO HCT116 cells treated with 2% DSS for the indicated times and pretreated with 20 μM MG132 for 6 hours. (J) Immunoblot analysis of total ubiquitination of FLAG-tagged SLC3A2 following coimmunoprecipitation of SLC3A2 with anti-FLAG antibody from lysates of HEK293T cells cotransfected with indicated plasmids.

Figure 8. NEDD4L regulates DSS-induced colitis through ferroptosis.

Nedd4l^fl/fl Villin^Cre and Nedd4l^fl/fl mice pretreated with Fer-1 (5 μM/kg) or DMSO were given 2% DSS for 5 days, and on the 9th day the mice were sacrificed for collection of colonic tissues and IECs. Nedd4l^fl/fl + DMSO n = 3, Nedd4l^fl/fl Villin^Cre + DMSO n = 4, Nedd4l^fl/fl + Fer-1 n = 6, Nedd4l^fl/fl Villin^Cre + Fer-1 n = 4. (A–F) Body weight change (A), colon length (B), gross morphology images (C), histological score (D), representative H&E staining (E), and TUNEL staining of the colon sections (F) from the indicated mice. (G–J) In a separate experiment, IECs and colon tissues from mice treated as in A were subjected to flow cytometry analysis of EpCAM, CD45, and propidium iodide (PI) staining (G and H), 4-HNE IHC staining (I), and ZO-1 IF staining (J). (K–M) qPCR analysis (K), Western blotting analysis (L), and protein intensity analysis (M) of the indicated proteins of IECs treated as in A. Nedd4l^fl/fl + DMSO n = 3–5, Nedd4l^fl/fl Villin^Cre + DMSO n = 3, Nedd4l^fl/fl + Fer-1 n = 4–6, Nedd4l^fl/fl Villin^Cre + Fer-1 n = 3–5, as indicated in the figure. Scale bars: 50 μm. Data represent mean ± SEM from at least 2 independent experiments. Each dot represents an independent sample. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Statistical analysis was performed using a 2-way ANOVA test in A, and 1-way ANOVA with multiple comparisons B, D, G, H, K, and M.

Figure 9. Nedd4l deficiency in IECs promotes AOM/DSS–induced CRC in mice.

(A) MRI images of Nedd4l^fl/fl Villin^Cre and Nedd4l^fl/fl mice treated with AOM/DSS for 90 days. (B–D) Tumor numbers (Nedd4l^fl/fl n = 15, Nedd4l^fl/fl Villin^Cre n = 21) (B), tumor size (n = 6 per group) (C), and representative morphology images of colons (D) from the AOM/DSS–treated mice on day 90. (E–G) Representative IHC staining of sections from the tumor, adjacent tumor, and distal normal tissues of AOM/DSS–treated Nedd4l^fl/fl Villin^Cre and Nedd4l^fl/fl mice with anti-Ki67 antibody (E) and anti–4-HNE antibody (G), and statistical analysis of Ki67⁺ cells (F) according to E (n = 4 per group). (H) Schematic diagram of the treatment plan for AOM/DSS–treated Nedd4l^fl/fl Villin^Cre and Nedd4l^fl/fl mice with ddH₂O or DFOM. (I–L) Representative morphology images of colons (I), tumor numbers (J), statistical analysis of 4-HNE IHC staining score (K), and representative images of 4-HNE IHC staining (L) from the treated mice as in I. Nedd4l^fl/fl + ddH₂O n = 5, Nedd4l^fl/fl Villin^Cre + ddH₂O n = 5, Nedd4l^fl/fl + DFOM n = 8, Nedd4l^fl/fl Villin^Cre + DFOM n = 8. Scale bars: 50 μm. Data represent mean ± SEM from at least 2 independent experiments. Each dot represents an independent sample. *P < 0.05; **P < 0.01; ***P < 0.001. Statistical analysis was performed using a 2-tailed Student’s t test in B, C, and F, and 1-way ANOVA with multiple comparisons in J and K.

Figure 10. Expression of NEDD4L is significantly downregulated in IECs of patients and mice with CRC.

(A–D) WT mice were treated with AOM/DSS, and IECs (on day 0, day 15, and day 60) and tumor nodes (on day 90) were collected for immunoblot analysis (A), protein intensity analysis (B), qPCR analysis (C), and correlative analysis of the indicated proteins (D). n = 3 per group. (E and F) Representative NEDD4L and 4-HNE IHC staining of sections from the tumor, adjacent tumor, and distal normal tissues of patients with CRC (E), and statistical analysis of NEDD4L and 4-HNE IHC staining intensity (F) according to E. n = 55. (G and H) Representative SLC3A2, GPX4, and NEDD4L IHC staining sections from the tumor tissues of patients with CRC (G), and correlative analysis between SLC3A2, GPX4, and NEDD4L IHC staining intensity score (n = 55) (H). Scale bars: 50 μm. Data represent mean ± SEM from at least 2 independent experiments. Each dot represents an independent sample. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Statistical analysis was performed using 1-way ANOVA with multiple comparisons in B, C, and F, and a Pearson’s correlation test in D and H.