Non-coding RNA-mediated high expression of SFXN3 as a prognostic biomarker associated with paclitaxel resistance and immunosuppressive microenvironment in head and neck cancer

Abstract

Chemoresistance is the leading cause of poor prognosis in head and neck squamous cell carcinoma (HNSC); however, promising biomarkers to identify patients for stratified chemotherapy are lacking. Sideroflexin 3 (SFXN3) is an important mitochondrial serine transporter during one-carbon metabolism, which is involved in the proliferation of cancer cells. However, the specific role of SFXN3 in HNSC remains unknown. In this study, we performed expression and survival analysis for SFXN3 in pan-cancer using data from The Cancer Genome Atlas (TCGA) and Genotype-Tissue Expression (GTEx) and found that SFXN3 served as a potential oncogene in HNSC. Notably, SFXN3 expression was found to be positively associated with enriched tumor-infiltrating macrophages, other immune suppressive cells, and immune checkpoint expression and resistance to paclitaxel. Gene, clinical, and immune variables included in the univariate and multivariate analyses showed that SFXN3 expression was an independent risk factor. Moreover, the LINC01270/hsa-miR-29c-3p/SFXN3 axis was identified as the most likely upstream non-coding RNA-related pathway of SFXN3 in HNSC using bioinformatic analysis, expression analysis, correlation analysis, and survival analysis. Taken together, our findings demonstrated that a non-coding RNA-mediated high expression of SFXN3 is a prognostic biomarker and is associated with the immunosuppressive microenvironment in HNSC.

Keywords: head and neck cancer; immune infiltration; paclitaxel resistance; sideroflexin 3; targeted therapy.

Figure 1

Expression analysis and survival analysis for SFXN3 in pan-cancer. (A) The expression of SFXN3 in pan-cancer based on TCGA cancer and normal data analyzed. (B) Cox regression analysis on overall survival was performed on SFXN3 expression in pan-cancer. (C) The overall survival (OS) analysis in patients with HNSC. (D) Cox regression analysis on disease-specific survival (DSS) was performed on SFXN3 expression in pan-cancer. (E) The DSS analysis for SFXN3 expression in patients with HNSC determined. -: no significant difference, **p value < 0.01; ***p value < 0.001; ****p value < 0.0001.

Figure 2

Correlation between SFXN3 expression with tumor immune cell infiltration in patients with head and neck squamous cell carcinoma (HNSC). (A) Tumor Immune Estimation Resource (TIMER) was used to analyze the correlation between SFXN3 expression and immune effector cells. (B) Tumor-infiltrating immune cells in HNSC samples were estimated using the CIBERSORT algorithm. (C) Correlation analysis of SFXN3 expression with the markers of M1 or M2 polarized phenotypes of macrophages. *p value < 0.05; **p value < 0.01; ***p value < 0.001.

Figure 3

SFXN3 expression correlates significantly with markers of immunosuppressive molecules and cells. (A) Correlation between SFXN3 expression and markers of immunosuppressive cells (MDSCs, TAMs, and Tregs). (B) Correlation between SFXN3 expression with molecules involved in inhibition of T-cell activation (CD28, CTLA4, CD274, PDCD1LG2, VSIR, and TIGIT).

Figure 4

Univariate and multivariate Cox regression analyses on survival in patients with head and neck squamous cell carcinoma (HNSC). (A) Univariate Cox regression analysis on survival in HNSC. (B) Multivariate Cox regression analysis on survival in HNSC.

Figure 5

SFXN3 expression as a potential predictor for drug sensitivity. (A) The compound activity z scores of chosen compounds in relation to SFXN3 expression, as shown by correlation analysis. (B) Correlation analysis using the Spearman correlation test for SFXN3 expression and activity z scores of paclitaxel, docetaxel, paclitaxel, methotrexate, carboplatin, fluorouracil, cisplatin, 5-fluoro deoxy uridine 10mer, and gemcitabine. (C) IC50 values of chemotherapeutics, including paclitaxel, docetaxel, methotrexate, 5-fluorouracil, cisplatin, and gemcitabine in the high SFXN3 expression HNSC cell lines compared with those in the low SFXN3 expression HNSC cell lines. HNSC cell lines with high SFXN3 expression were found to possess higher IC50 for docetaxel. The p-values were calculated using the Wilcoxon rank-sum test. *p value < 0.05.

Figure 6

Identification of hsa-miR-29c-3p as a potential upstream miRNA of SFXN3 in HNSC. (A) Venn diagram of predicted miRNAs, downregulated miRNAs, and miRNAs related to better survival in HNSC. (B) The expression of hsa-miR-29c-3p in HNSC and control normal samples as determined using the CancerMIRNome database. (C) The prognostic value of hsa‐miR‐29c‐3p in HNSC assessed using the CancerMIRNome. (D) The expression correlation between hsa-miR-29c-3p and SFXN3 in HNSC.

Figure 7

Correlation analysis, expression analysis, and survival analysis for upstream lncRNAs of hsa-miR-29c-3p in HNSC. (A) Correlation analysis between hsa-miR-29c-3p and LINC01270, NOP14-AS1, LINC01907, and MIR193BHG, respectively. (B) Correlation analysis between SFXN3 and LINC01270, NOP14-AS1, LINC01907, and MIR193BHG. (C) The expression of LINC01270, NOP14-AS1, LINC01907, and MIR193BHG in TCGA HNSC compared with that in normal tissues using ENCORI. (D) The expression of LINC01270, NOP14-AS1, LINC01907, and MIR193BHG in TCGA HCC compared with “TCGA normal” or “TCGA and GTEx normal” data. (E) The OS analysis for LINC01270, NOP14-AS1, LINC01907, and MIR193BHG in HNSC. *p value < 0.05; ***p value < 0.01.

Figure 8

Schematic model of the LINC01270/hsa-miR-29c-3p/SFXN3 axis as a potential regulatory pathway in HNSC.