Sideroflexin family genes were dysregulated and associated with tumor progression in prostate cancers

Abstract

Sideroflexin (SFXN) family genes encode for a group of mitochondrial proteins involved in cellular processes such as iron homeostasis, amino acid metabolism, and energy production. Recent studies showed that they were aberrantly expressed in certain human cancers. However, there is a paucity of information about their expression in prostate cancer. In this study, we took a comprehensive approach to investigate their expression profiles in benign prostate tissue, prostate-derived cell lines, and prostate cancer tissues using multiple transcriptome datasets. Our results showed that SFXN1/3/4 genes were predominantly expressed in prostate tissue and cell lines. SFXN2/4 genes were significantly upregulated while the SFXN3 expression was significantly downregulated in malignant tissues compared to benign tissues. SFXN4 expression was identified as a diagnostic biomarker and prognostic factor for unfavorite survival outcomes. In advanced prostate cancers, SFXN2/4 expressions were positively correlated with the androgen receptor signaling activity but negatively correlated with the neuroendocrinal features. Further analysis discovered that SFXN5 expression was significantly elevated in neuroendocrinal prostate cancers. In conclusion, SFXN2/4 expressions are novel biomarkers in prostate cancer diagnosis and prognosis.

Keywords: Androgen receptor modulation; Castration-resistance; Neuroendocrinal progression; Prostate cancer; Sideroflexin family genes.

Fig. 1

Expression of SFXN family genes in benign prostate tissue and prostate cancer cell lines. A–C The expression levels of SFXN family genes were downloaded from the cBioportal platform and RNA-seq RPKM values were utilized for the analysis. ANOVA test, * p < 0.05; **** p < 0.0001. D Real-time qPCR experiments were conducted using the cellular RNAs extracted from prostate cell lines as indicated. E DepMap data gene knockout effect was downloaded from the UALCAN platform

Fig. 2

Expression of SFXN family genes in primary prostate cancers. Case-matched pairs (A) or group cohort (B) comparisons were conducted between benign and malignant tissues using the TCGA-PRAD dataset (log₂ [TPM + 1]). The asterisks indicate a significant difference compared to the control group. * p < 0.05; ***, p < 0.001. C ROC analysis was conducted to determine the potential of prediction in distinguishing benign and malignant prostate tissues

Fig. 3

Expression of SFXN family genes in subgroups with specific molecular signatures or different Gleason scores. A–E SFXN expresion in different molecular signature groups. F–G SFXN2/3/4 expression in different Gleason score groups. Gene expression data was downloaded from the TCGA-PRAD dataset on the UALCAN platform. The asterisks indicate a significant difference compared to the benign group. * p < 0.05; ** p < 0.01; ***, p < 0.001; ****p < 0.0001. The case numbers in each group: normal (52), ERG (152), ETV1 (28), ETV4 (14), FLI (4), FOXA1 (9), IDH1 (3), and SPOP (35)

Fig. 4

Correlations of SFXN family genes with tumor immune infiltrations in prostate cancers. A–E The TCGA-PRAD RNS-seq dataset was utilized for the correlation analysis with tumor immune infiltrations on the XIANTAO platform. The red line squares indicated special factors with SFXN gene expression. F–J Single-cell sequencing data of gene expression in the immune infiltrations were downloaded and visualized using the online HuPSA database

Fig. 5

Kaplan–Meier survival analysis. SFXN4 expression as prognostic factor for overall survival (A), diseasespecific survival (B) and progression-free interval outcomes (C). Gene expression data for the SFXN4 expression were extracted from the TCGA-PRAD RNA-seq dataset. Kaplan-Meier survival curves were generated with a minimum p-value approach on the XIANTAO platform

Fig. 6

Correlation of the SFXN family gene expression with the AR score (A–C) or NEPC score (D–E) in CRPC patients. Gene expression at the mRNA levels (log₂ [TPM + 1]) was extracted from the SU2C/PCF dataset on the cBioportal platform. The asterisk indicates a significant difference between the two groups

Fig. 7

Expression of the SFXN family genes in CRPC patients. A–E CRPC tissues without NE or with NE features were divided into two groups for comparison of SFXN expressions. Data at the mRNA levels were shown as log₂[FPKM + 1] values extracted from the SU2C/PCF dataset on the cBioportal platform. The p-values shown in each panel were derived from Student’s t-test. F–J The mRNA expression levels of SFXN family genes were compared in subgroups based on tumor pathological features, as indicated. The asterisks indicate a significant difference from ANOVA analysis. Red asterisks, upregulation vs the benign control; green asterisks, downregulation vs the benign control. **** p < 0.001, ** p < 0.01

Fig. 8

Expression of SFXN2/4 genes in mouse prostate or human prostate xenografts in nude mice after castration or testosterone replacement. Microarray data were extracted from the NCBI GEO profiles of GDS2562 (A), GDS4107 (B), and GDS4120 (C). The relative values were calculated against the lowest value in the control group before statistical analysis. The asterisk indicates a significant difference. Student’s t-test, ** p < 0.01; **** p < 0.0001