Autoreactive T Cells and Chronic Fungal Infection Drive Esophageal Carcinogenesis

Abstract

Humans with autoimmune polyendocrinopathy-candidiasis-ectodermal dystrophy (APECED), a T cell-driven autoimmune disease caused by impaired central tolerance, are susceptible to chronic fungal infection and esophageal squamous cell carcinoma (ESCC). However, the relationship between autoreactive T cells and chronic fungal infection in ESCC development remains unclear. We find that kinase-dead Ikkα knockin mice develop APECED-like phenotypes, including impaired central tolerance, autoreactive T cells, chronic fungal infection, and ESCCs expressing specific human ESCC markers. Using this model, we investigated the link between ESCC and fungal infection. Autoreactive CD4 T cells permit fungal infection and incite tissue injury and inflammation. Antifungal treatment or autoreactive CD4 T cell depletion rescues, whereas oral fungal administration promotes, ESCC development. Inhibition of inflammation or epidermal growth factor receptor (EGFR) activity decreases fungal burden. Fungal infection is highly associated with ESCCs in non-autoimmune human patients. Therefore, autoreactive T cells and chronic fungal infection, fostered by inflammation and epithelial injury, promote ESCC development.

Keywords: EGFR; IKKalpha; autoimmune disease; autoreactive T cells; esophageal squamous cell carcinoma; fungal infection; inflammation; mucosal epithelium.

Figure 1. Esophageal SCCs and Fungal Infection in Ikkα^KA/KA Mice

(A) H&E– and immunohistochemistry (IHC)–K5-stained esophagi of 5-month-old WT and Ikkα^KA/KA mice. K5, keratin 5. Scale bar, 50 μm. (B) Immunoblotting (IB) shows (top) indicated protein levels in 4 WT esophagi and 5 Ikkα^KA/KA esophageal SCCs (ESCCs). β-Actin, protein-loading control. Intensities of IKKα, ΔNp63, p53, and p16 normalized by β-Actin and p-EGFR levels normalized by EGFR. Data represent means±SEM (3 repeats, bottom). *, p < 0.05; ***, p < 0.001, t test. (C) Gene-expression profiles (fold changes): Ikkα^KA/KA ESCCs versus (vs) WT esophagi using microarray (See Figure S1B and S1C). (D) H&E–stained esophagi of WT and Ikkα^KA/KA mice. Scale bar, 20 μm. Right panel: Grocott’s methenamine silver (GMS)–stained fungi. Positive, black spots. Scale bar, 5 μm. (E) Fungal genera identified by sequencing from cultured colonies (see Figure S1E). See also Figure S1 and Tables S1 and S2.

Figure 2. mTEC Is Essential for Preventing Chronic Fungal Infection and Esophageal Phenotypes

(A) H&E–stained thymuses of WT and Ikkα^KA/KA mice at 6 weeks of age. Lighter-stained areas are medullary regions. Boxes and lines in row one indicate amplified, lighter-stained medullary regions. Immunofluorescence (IF) staining for K5/UEA-1 in row three and UEA-1/Aire mTECs in row four. Scale bar, 50 μm. (B) Flow cytometry analyzes CD4, CD8, and CD4⁺CD25⁺Foxp3⁺ Treg cells from thymuses and spleens of Ikkα^KA/KA and WT mice at 6 weeks of age (top). Data represent means±SEM (3 repeats) (bottom). **, p < 0.01; n.s, not significant, t test; KA/KA, Ikkα^KA/KA. (C) Gene-expression profiles of mTEC mRNA of Ikkα^KA/KA (KA, n =3) and WT mice (n = 3). The heat map contains 600 genes. Spt1, salivary protein 1; Crp, C-reactive protein; Cldn, claudin; Krt, keratin. (D) IB shows indicated cytoplasmic and nuclear protein levels in mTECs of 4 WT and 8 Ikkα^KA/KA mice following anti-CD40 antibody treatment. α-Tubulin, cytoplasmic protein-loading control; Lamin B, nuclear protein-loading control. (E) IB shows nuclear p50 and p65 levels in mTECs from 4 WT, 8 Ikkα^KA/KA, and 4 Ikkα^KA/KA;K5.IKKα mice following anti-CD40 antibody treatment. Lamin B, nuclear protein-loading control. Loading controls see (D). (F) H&E–stained thymuses of WT, Ikkα^KA/KA, and Ikkα^KA/KA;K5.IKKα mice. Scale bar, 40 μm. (G) H&E–stained organs of Ikkα^KA/KA;K5.IKKα mice. Scale bar, 40 μm. See also Figure S2.

Figure 3. IKKα Regulates mTEC and TCR Repertoire; T Cells are Autoreactive in Ikkα^KA/KA Esophagi

(A) Flow cytometry analyzes mTECs gated at CD45⁻Integrinα6⁺(Intgα6)UEA-1⁺ cells from thymuses of WT, Ikkα^KA/KA, and Ikkα^KA/KA;K5.IKKα mice at 8 to 10 weeks of age (left). Data represent mean±SEM (3 repeats, right). **, p < 0.01; ***, p < 0.001, one-way ANOVA. (B–D) Flow cytometry analyzes TCR Vβ5 repertoire in T cells of the thymus and spleen of WT, Ikkα^KA/KA, and Ikkα^KA/KA;K5.IKKα mice at 8 to 10 weeks of age (B–C). SP, single positive. Data represent mean±SEM (D, 3 repeats). *, p < 0.05; ***, p < 0.001, one-way ANOVA. (E) PCR for expression of Vβ5.1 and Vβ5.2 TCR repertoire in splenic T cells of mice. Gapdh is used to normalize Vβ5.1 and Vβ5.2 levels. Data represent mean±SEM (3 repeats). ***, p < 0.001, one-way ANOVA. (F–G) Flow cytometry analyzes CD3⁺CD44^Hi and CD3⁺CD44^Lo cells (F), and TCR Vβ5 positive CD4⁺ T cells (G) gated at CD45⁺ cells in esophagi of 8 WT and 4 Ikkα^KA/KA mice at 5 months of age. (H) Flow cytometry analyzes TCR Vβ5 repertoire positive cells gated at CD4 or CD8 from the thymus of chimeric mice. WT or Ikkα^KA/KA (KA/KA) BM was injected to KA/KA mice (left). WT BM or Ikkα^KA/KA (KA/KA) BM was injected to WT mice (right). Data represent mean±SEM (3 repeats). ***, p < 0.001; n.s, not significant, t test. (I) Flow cytometry analyzes CD3 T cells from spleens of chimeric mice. Data represent mean±SEM (3 repeats). n.s, not significant, t test. See also Figure S3.

Figure 4. T-Cell Depletion or Antifungal Drug Treatment Diminishes Inflammation and Malignancies in Ikkα^KA/KA Esophagi

(A) H&E–stained thymuses and esophagi of WT, Ikkα^KA/KA, antifungal-treated Ikkα^KA/KA, Ikkα^KA/KA;Rag1^−/−, and Ikkα^KA/KA;Cd4^−/− mice at 5 months of age (Scale bar, 50 μm) and IHC–stained esophagi with the anti-CD3 or anti-F4/80 antibody. AmpB, antifungal drug: amphotericin B; positive, brown color; Ikkα^KA/KA;Rag1^−/− and Ikkα^KA/KA;Cd4^−/− mice (n = 20/group). Scale bar, 30 μm. (B) A summary of esophagus phenotypes (severe inflammation and malignancy) from indicated mice. Fungus detection in oral swabs and esophagus extracts of indicated mice was shown; number, mouse numbers; M, treatment months. Fisher’s exact test was applied for statistical analyses (p value). (C) Fungal cultures from oral swabs of WT, Ikkα^KA/KA, and antifungal-treated Ikkα^KA/KA mice (see 4B). Data is representative of YPD agar plates containing chloramphenicol (10 μg/ml). (D) IB shows COX-2 levels in esophagi of 4 WT, 3 Ikkα^KA/KA, and 2 AmpB-treated Ikkα^KA/KA mice and their intensities analyzed by one-way ANOVA. ***, p < 0.001. β-Actin, protein-loading control. (E) PD-L1–IHC-stained esophagi of WT, Ikkα^KA/KA, and AmpB-treated Ikkα^KA/KA mice (n = 5/group). Scale bar, 30 μm. (F) IF staining detected 8-OHdG in esophageal squamous epithelial cells of WT, Ikkα^KA/KA, and antifungal-treated Ikkα^KA/KA mice (n = 3/group). White lines are located beneath esophageal basal cell layers. DAPI, the nucleus of cells. Scale bar, 30 μm. (G) Fungal cultures from the oral swabs of WT, Ighm^tm1Cgn (B-cell–lacking), and Ikkα^KA/KA;Ighm^tm1Cgn mice at 10 to 12 weeks of age. Data is representative of YPD agar plates (n = 5/group). See also Figure S4.

Figure 5. Ikkα^KA/KA Autoreactive T Cells and Fungal Infection Contribute to Esophageal Carcinogenesis

(A) Left: H&E–stained esophagi of Ikkα^KA/KA;Rag1^−/− mice receiving WT or Ikkα^KA/KA T-cell transfer (n = 5/group, 2 repeats). Arrows indicate fungi. Right: IHC-stained esophagi of Ikkα^KA/KA;Rag1^−/− mice receiving WT or Ikkα^KA/KA T cells with anti-CD3 and F4/80 antibodies. Scale bar, 30 μm. (B) Fungal colony numbers in YPD agar plates from oral swabs of Ikkα^KA/KA mice and Ikkα^KA/KA mice receiving WT T-cell injection at day 3 and day 20. M, months of age. Data represent mean±SEM (n = 3/group). *, p < 0.05; **, p < 0.01, one-way ANOVA. (C) Relative mRNA levels of cytokines in T cells from the spleens of WT, Ikkα^KA/KA, and Ikkα^KA/KA;K5.IKKα mice following treatment with an anti-CD3 antibody (CD3) only or anti-CD3 antibody and zymosan A (ZY, 2 μg/ml). Data represent mean±SEM (n = 3/group). *, p < 0.05; **, p < 0.01; n.s, not significant, t test. (D) Fungal colony numbers in YPD agar plates from oral swabs of Ikkα^KA/KA mice and Ikkα^KA/KA mice treated with aspirin for 20 days. M, months of age. Data are analyzed by t test (n = 3 for 5- to 8-month-old mouse group; n = 5 for 3-month-old mouse group). *, p < 0.05; **, p < 0.01. (E) Fungal colony numbers in YPD agar plates from oral swabs of Ikkα^KA/KA mice and Ikkα^KA/KA mice treated with GW2974 (0.6 mg/mouse/day), 5 times per week for 30 days. M, months of age. Data are analyzed by t-test (n = 4/group). *, p < 0.05. See also Figure S5.

Figure 6. Fungal Infection Promotes Pathogenesis of Mouse Oral and Esophageal Cancers and Fungi are Detected in HESCCs

(A–B) Oral fungal infection and tumors detected in WT, Ikkα^KA/+ (KA/+), and Ikkα^KA/KA (KA/KA) mice orally infected with fungi (A). Oral-Fungi, monitoring fungi in oral swabs; the first (1st) infection; Oral-Eso-Tum, oral or esophageal tumors; 4th, the fourth infection. Data are analyzed by Fisher’s exact test. **, p < 0.01; ***, p < 0.001; ****, p < 0.0001; n.s, not significant (B). (C) Representative H&E–stained esophagus and oral cavity of WT mice, and esophageal SCC in situ and oral SCC from Ikkα^KA/KA mice after 4 fungal infections. Scale bar, 50 μm. (D) A working model of esophageal SCC development promoted by impaired central tolerance mediated autoimmunity, T–cell defects, inflammation, fungal infection, and IKKα reduction. mTECs, medullary thymic epithelial cells; SP4, CD4 single-positive T cell; MΦ, macrophage; lines with circles on mTEC surfaces are tissue-restricted antigens; Aire, Aire expression in the nucleus of mTECs; multiple crossed lines in the nuclei of tumor cells, DNA damage. Down arrow, decrease; up arrow, increase; cross lines, inactivation. (E) Left panels (photos): Human ESCC sections stained with Periodic Acid-Schiff (PAS) staining or GMS for fungal staining. Left panel 1, Scale bar 40 μm; Left panels 2 and 3, Scale bar 5 μm. An arrow indicates the area that contains PAS-stained fungi. Right panel: Identified fungal species from paraffin sections of HESCC cases by DNA sequencing. (F) H&E–stained human normal esophageal tissue sections. Scale bar, 40 μm. (G) Data in Figures 6E and 6F are statistically analyzed by Fisher’s exact test. ***, p < 0.001. HNETs, human normal esophageal tissues = 50 and HESCCs = 80. See also Figure S6.

Figure 7. Increased Macrophage and T Cells and PD-L1 Levels, and Reduced IKKα in HESCCs

(A) Macrophage (CD68) numbers in the human tissue array containing 60 HESCCs and 60 adjacent tissues (Adjacent-E), as analyzed by t-test. ****, p < 0.0001. (B) T-cell (CD3) numbers in the human tissue array containing 60 HESCCs and 60 adjacent tissues, as analyzed by t test. ****, p < 0.0001. (C) Staining intensities of IKKα using IHC with an anti-IKKα antibody in the human tissue array containing 60 HESCCs and 60 adjacent tissues. Staining intensities show SCCs vs adjacent tissues: strong staining (+++), 5% vs 38.3%; medium staining (++), 25% vs. 56.7%; and weak staining (+), 70% vs. 5%. Data are statistically analyzed by Chi-square test for each staining group between SCCs and adjacent tissues. ****, p < 0.0001. (D) IB shows IKKα levels in human SCCA431 SCC cell line following TNFα (TNF, 10 ng/ml) and IFNγ (IFN, 10 ng/ml) treatment at 48 and 72 hr. Cont, Control. (E) CHUK mutations in HESCCs obtained from two studies (cbioportal.org). IF, in-frame; Del, deletion. (F) Left panel: Staining intensities of PD-L1 using IHC with an anti-PD-L1 antibody in the human tissue array containing 60 HESCCs and 60 adjacent tissues. Staining intensities show SCCs vs adjacent tissues. Strong staining (+++) to negative staining (−). Data are analyzed by Chi-square test for each staining group between SCCs and adjacent tissues. ****, p < 0.0001. Right panel: IHC-stained adjacent tissue and HESCC. Scale bar, 40 μm. See also Figure S7.