Butyrate reverses ferroptosis resistance in colorectal cancer by inducing c-Fos-dependent xCT suppression

Abstract

Ferroptosis has emerged to be a promising approach in cancer therapies; however, colorectal cancer (CRC) is relatively insensitive to ferroptosis. Exactly how the gut microenvironment impacts the ferroptotic sensitivity of CRC remains unknown. Herein, by performing metabolomics, we discovered that butyrate concentrations were significantly decreased in CRC patients. Butyrate supplementation sensitized CRC mice to ferroptosis induction, showing great in vivo translatability. Particularly, butyrate treatment reduced ferroptotic resistance of cancer stem cells. Mechanistically, butyrate inhibited xCT expression and xCT-dependent glutathione synthesis. Moreover, we identified c-Fos as a novel xCT suppressor, and further elucidated that butyrate induced c-Fos expression via disrupting class I HDAC activity. In CRC patients, butyrate negatively correlated with tumor xCT expression and positively correlated with c-Fos expression. Finally, butyrate was found to boost the pro-ferroptotic function of oxaliplatin (OXA). Immunohistochemistry data showed that OXA non-responders exhibited higher xCT expression compared to OXA responders. Hence, butyrate supplementation is a promising approach to break the ferroptosis resistance in CRC.

Keywords: Butyrate; Colorectal cancer; Ferroptosis; c-Fos; xCT.

Graphical abstract

Fig. 1

Butyrate enhances ferroptosis sensitivity in CRC cells. a, Fecal levels of SCFAs in CRC patients (n = 32) and healthy controls (n = 32) were evaluated by SCFA metabolomics (for isobutyrate or caproate detection, n = 31 or 25 respectively, since some patients had low levels of isobutyrate or caproate which were below the detection threshold). b, HCT116 cells were pretreated with different SCFAs (1 mM) for 8 h, followed by treatment of erastin (30 μM), RSL3 (15 μM), FIN56 (5 μM) or DMSO for 24 h. Cell viability was evaluated by CCK8 assay. c, The synergistic effects of butyrate and erastin were assessed by the SynergyFinder tool. d, HCT116 cells were treated with butyrate, erastin, or combination and cell viability was evaluated by a RTCA system. Normalized Cell Index means that in each group, the cell indexes were normalized to the corresponding cell index of the time point of adding erastin. e, HCT116 cells were pretreated with butyrate (1 mM) for 8 h in the presence of Z-VAD-FMK (10 μM), ferrostatin-1 (Fer-1, 2 μM), or necrostatin-1 (Nec-1, 2 μM), followed by the treatment of erastin for 24 h. Cell viability was evaluated by CCK8 assay. f-i HCT116 cells were pretreated with butyrate (1 mM) followed by erastin (20 μM) treatment. Mitochondrial morphology was evaluated by transmission electron microscope (f). The intracellular ROS production (g), GSH levels (h) and lipid peroxidation (i) were measured. MFI = mean fluorescence intensity. j, k CRC organoids from AOM/DSS mice were treated with butyrate, erastin, or combination. Organoid viability was measured by MTT assay (j) and PI staining (k). Representative images (left) and quantification data (right) are shown. Data are represented as the mean ± SD. **p < 0.01; ***p < 0.001, two-tailed unpaired Student's t-test.

Fig. 2

Butyrate supplementation enhances tumor susceptibility to ferroptosis in CRC mice. a, HCT116-bearing mice were administered butyrate, erastin, or both (n = 5/group). Tumor volume was monitored. b, Tumors were photographed and weighed at the experimental endpoint (Day 23). c, Tumor GSH levels were evaluated. d, Lipid peroxidation was evaluated by 4-HNE staining. e-g AOM/DSS-CRC mice were administered with butyrate, erastin, or both (n = 5/group). The size and number of tumor nodules were measured (e). (f) The levels of GSH in tumor tissues were evaluated. (g) The lipid peroxidation levels in tumor tissues were evaluated by 4-HNE staining. Data are represented as the mean ± SD. *p < 0.05; **p < 0.01; ***p < 0.001, two-tailed unpaired Student's t-test.

Fig. 3

Butyrate inhibits xCT expression via HDAC inhibition. a, HCT116 cells were treated with DMSO or erastin (20 μM) for 10 h in the presence or absence of butyrate (1 mM), then subjected to RNA-Seq. The expression levels of ferroptosis–related genes were shown. b,c HCT116 cells were treated with butyrate as indicated, xCT expression was evaluated by qPCR (b) and immunoblotting (c). d, Mouse CRC organoids were treated with butyrate; xCT expression was evaluated by qPCR. e, Organoids from CRC patients were stimulated with 1 mM butyrate for 12 h, xCT expression was evaluated by qPCR. f, g xCT expression in HCT116 tumors (f) and AOM/DSS tumors (g) were evaluated by qPCR. h, HCT116 cells treated with butyrate were cultured in cystine-sufficient or cystine-low conditions. Cell viability was evaluated by CCK8. i, HCT116 cells were treated with erastin (20 μM) or RSL3 (10 μM), either alone or in combination with butyrate in the absence or presence of NAC. Cell viability was evaluated by CCK8. j, xCT^OE and control HCT116 cells were treated with a combination of butyrate + erastin. Cell viability was evaluated by CCK8. k, xCT^OE or control HCT116 cells were inoculated s.c into nude mice, which were then treated with DMSO or butyrate + erastin (n = 5–6/group). Tumor growth was monitored. l, On day 23 mice were sacrificed and tumors were weighed. m, Tumor GSH levels were measured. n, HCT116 cells were pretreated with pertussis toxin (PT) for 2 h, followed by butyrate treatment for 12 h xCT expression was evaluated by qPCR. o, HCT116 cells were treated with TSA (10 mM). xCT expression was evaluated by qPCR. Data are represented as the mean ± SD. *p < 0.05; **p < 0.01; ***p < 0.001, two-tailed unpaired Student's t-test.

Fig. 4

Low butyrate concentration is correlated with high xCT expression in CRC patients. a, b The levels of xCT protein (a) or mRNA (b) in CRC tissues and adjacent-normal tissues were evaluated by tissue microarrays (n = 47 for COAD patients, n = 18 for READ patients) or qPCR (n = 32/group), respectively. COAD = colon adenocarcinoma, READ = rectum adenocarcinoma. c, Correlations between tumor xCT mRNA expression and fecal SCFA concentration were analyzed by Spearman's rank correlation test (n = 32. For isobutyrate or caproate detection, n = 31 or 25 respectively. Since some patients have low levels of isobutyrate or caproate which were below the detection threshold). Data are represented as the mean ± SD. *p < 0.05; **p < 0.01; ***p < 0.001, two-tailed unpaired (a) or paired (b) Student's t-test.

Fig. 5

Butyrate induces c-Fos expression in CRC cells. a, HCT116 cells were treated with PBS or butyrate (1 mM) for 12 h in the presence of CHX (5 μg/ml), xCT expression was evaluated by qPCR. b, HCT116 cells were treated with 1 mM butyrate for 1 h. The genes having increased chromatin accessibilities (fold change >1.5) in their promoter region were overlapped with the butyrate-upregulated genes from RNA-Seq data, and with the genes whose products were predicted to bind to SLC7A11 promoter using hTFtarget tool (http://bioinfo.life.hust.edu.cn/hTFtarget#!/). c, Differential peaks of ATAC-seq signal in PBS- or butyrate-treated HCT116 cells at FOS and SLC7A11 promoter region. d, e HCT116 cells were treated with butyrate as indicated and c-Fos expression was evaluated by qPCR (d) and immunoblotting (e). f, Mouse CRC organoids were treated with butyrate, erastin, or both and c-Fos expression was evaluated by qPCR. g, Organoids from CRC patients were stimulated with 1 mM butyrate, c-Fos expression was evaluated by qPCR. h, HCT116 cells were treated with TSA and c-Fos expression was evaluated by qPCR. i, C-Fos was overexpressed in HCT116 cells, the activity of the SLC7A11 promoter was evaluated by a dual-luciferase reporter assay. j, The binding of c-Fos to the SLC7A11 promoter was examined by a ChIP assay. k, The expression of xCT was evaluated by qPCR after c-Fos overexpression. l, Different c-Fos protein domains were ectopically expressed in HCT116 cells. The expression of xCT was evaluated by qPCR. m, n C-Fos expression in HCT116 tumors (m) and AOM/DSS tumors (n) were evaluated by qPCR. o, The mRNA expression of c-Fos in CRC tissues and adjacent-normal tissues were evaluated by qPCR. p, The correlation between tumor c-Fos mRNA expression and fecal butyrate concentrations was analyzed by Spearman's rank correlation test (n = 32). q, The correlation between tumor c-Fos expression and xCT expression was analyzed by Spearman's rank correlation test (n = 32). Data are represented as the mean ± SD. *p < 0.05; **p < 0.01; ***p < 0.001, two-tailed unpaired (a-n) or paired (o) Student's t-test.

Fig. 6

Butyrate treatment breaks xCT-mediated ferroptosis resistance in a c-Fos-dependent manner. a, Control or c-Fos^OE HCT116 cells were treated with erastin for 24 h. Cell viability was evaluated by CCK8. b, Different c-Fos protein domains were ectopically expressed in HCT116 cells followed by erastin treatment for 24 h. Cell viability was evaluated by CCK8. c, Control, c-Fos^KD, or c-Fos^KDxCT^KD HCT116 cells were treated with erastin, butyrate, or both for 24 h. Cell viability was evaluated by CCK8. d, Mouse CRC organoids were treated with butyrate + erastin in the presence of a c-Fos inhibitor (T5224, 20 μM) for 24 h. Organoid viability was evaluated by MTT. e, Control, c-Fos^KD, or c-Fos^KDxCT^KD HCT116 cells were inoculated s.c. into nude mice, which were then treated with butyrate + erastin. Tumor growth was monitored. f, Mice were sacrificed on day 24 and tumors were weighed. g, Tumor GSH levels were evaluated. Data are represented as the mean ± SD. *p < 0.05; **p < 0.01; ***p < 0.001, two-tailed unpaired Student's t-test.

Fig. 7

Butyrate reduces the ferroptosis resistance of CSCs. a, The luminal-to-crypt gradient of butyrate in the large intestine. b, The expression of xCT in LGR5⁺ and LGR5^- non-CSCs were evaluated by analyzing a GEO dataset (GSE92961). c, The expression of xCT in CD133⁺ CSCs and CD133^- non-CSCs in CRC patients were evaluated by analyzing a single-cell sequencing dataset (GSE146771). d, The expression of LGR5 was evaluated in AOM/DSS CRC tissues by qPCR. e, The viability of colon organoids isolated from AOM/DSS CRC mice administered with butyrate, erastin or in combination were evaluated by MTT assay. f, CRC organoids were treated with erastin (10 μM), butyrate (0.5 mM), or both for 24 h. The expression of LGR5 was evaluated by qPCR. g, HCT116 CSC spheres were treated with erastin (20 μM), butyrate (0.5 mM), or both for 24 h. The viability was examined by PI staining. h, The expression of CD133 was evaluated by qPCR. Data are represented as the mean ± SD. *p < 0.05; **p < 0.01; ***p < 0.001, two-tailed unpaired Student's t-test.

Fig. 8

Butyrate potentiates the therapeutic sensitivity of OXA. a, HCT116 cells were treated with OXA (20 μM), butyrate (1 mM), or both for 36 h, in the presence or absence of Z-VAD, nec-1, or fer-1. Cell viability was evaluated by CCK8. b, The synergistic effects of butyrate and OXA were assessed by SynergyFinder. c-e HCT116 cells were treated as indicated for 36 h, GSH levels (c), lipid peroxidation (d), and mitochondria morphology (e) were evaluated. f, HCT116 cells were treated as indicated for 36 h, cell viability was evaluated by CCK8. g, h CRC organoids were treated with OXA, butyrate, or both. Organoid viability was determined by MTT (g) and PI staining (h). i, The expression of LGR5 was evaluated by qPCR. j-l HCT116 cells were inoculated s.c. into nude mice, which were then treated with OXA or a combination of OXA plus butyrate (n = 6/group). Tumor growth was monitored (j), tumor weight was evaluated on day 24 (k), 4-HNE staining was performed to identify lipid peroxidation (l). m, Hep3b cells were inoculated s.c. into nude mice, which were then treated with OXA or OXA plus butyrate (n = 5/group). Tumor growth was monitored. n, The expression of xCT and c-Fos in HCT116 tumor tissues were evaluated by qPCR. o, The expression of xCT in OXA responders (R, n = 19) and non-responders (NR, n = 19) was examined by IHC. Data are represented as the mean ± SD. *p < 0.05; **p < 0.01; ***p < 0.001, two-tailed unpaired Student's t-test.