A Novel Ferroptosis-Based Molecular Signature Associated with Biochemical Recurrence-Free Survival and Tumor Immune Microenvironment of Prostate Cancer

Abstract

Objective: To identify ferroptosis-related molecular clusters, and to develop and validate a ferroptosis-based molecular signature for predicting biochemical recurrence-free survival (BCRFS) and tumor immune microenvironment of prostate cancer (PCa).Materials and Methods: The clinical data and transcriptome data of PCa were downloaded from TCGA and GEO database. Ferroptosis-related genes (FRGs) were obtained from FerrDb database. We performed consensus clustering analysis to identify ferroptosis-related molecular subtypes for PCa. Univariate and multivariate Cox regression analysis were used to establish a ferroptosis-based signature for predicting BCRFS. Internal verification, external verification and subgroup survival analysis were then successfully performed.Results: There was a total of 40 differentially expressed FRGs in PCa. We then identified three ferroptosis-related molecular clusters of PCa, which have significantly different immune infiltrating cells, tumor immune microenvironment and PD-L1 expression level. More importantly, a novel ferroptosis-based signature for predicting BCRFS of PCa based on four FRGs (including ASNS, GPT2, NFE2L2, RRM2) was developed. Internal and external verifications were then successfully performed. Patients with high-risk score were associated with significant poor BCRFS compared with those with low-risk score in training cohort, testing cohort and validating cohort, respectively. The area under time-dependent Receiver Operating Characteristic (ROC) curve were 0.755, 0.705 and 0.726 in training cohort, testing cohort and validating cohort, respectively, indicating the great performance of this signature. Independent prognostic analysis indicated that this signature was an independent predictor for BCRFS of PCa. Subgroup analysis revealed that this signature was particularly suitable for younger or stage T III-IV or stage N0 or cluster 1-2 PCa patients. Patients with high-risk score have significantly different tumor immune microenvironment in comparison with those with low-risk score. The results of qRT-PCR successfully verified the mRNA expression levels of ASNS, GPT2, RRM2 and NFE2L2 in DU-145 and RWPE-1 cells while the results of IHC staining exactly verified the relative protein expression levels of ASNS, GPT2, RRM2 and NFE2L2 between PCa and BPH tissues.Conclusions: This study successfully identified three ferroptosis-related molecular clusters. Besides, we developed and validated a novel ferroptosis-based molecular signature, which performed well in predicting BCRFS and tumor immune microenvironment of PCa.

A novel ferroptosis-related gene prognostic index for prognosis and response to immunotherapy in patients with prostate cancer

“…A few previous studies have examined the role of FRGs in predicting BCR for PCa patients undergoing radical prostatectomy in the TCGA cohort ( 13 – 15 ). Lv et al.…”

“…Ke et al. also developed a 4-gene (ASNS, GPT2, NFE2L2 and RRM2) ferroptosis-based signature for predicting the BCR of PCa ( 15 ). However, its AUC values were relatively lower than those of our FRGPI.…”

“…A few previous studies have examined the role of FRGs in predicting BCR for PCa patients undergoing radical prostatectomy in the TCGA cohort (13)(14)(15). Lv et al and Liu et al included nine FRGs (AIFM2, AKR1C1, AKR1C2, CBS, FANCD2, FTH1, G6PD, NFS1, and SLC1A5) and seven FRGs (AKR1C3, ALOXE3, ATP5MC3, CARS1, MT1G, PTGS2, and TFRC) into their risk models, respectively (13,14).…”

“…The predictive performance of ferroptosis-related biomarkers on long-term prognosis in PCa patients has not been well-characterized. A few ferroptosisrelated gene (FRG) signatures have been developed to predict the prognoses for PCa patients after radical prostatectomy, but these signatures or risk models were established almost based on BCRFS that was, as mentioned above, not a good surrogate for OS (13)(14)(15).…”

Identification of Ferroptosis-Related Prognostic Signature and Subtypes Related to the Immune Microenvironment for Breast Cancer Patients Receiving Neoadjuvant Chemotherapy

“…FRGs including 150 drivers, 109 suppressors, and 123 markers were collected from FerrDb ( 19 , 23 , 24 ). A univariate Cox proportional hazards regression analysis was conducted to filtrate prognostic FRGs in the GSE25055 cohort with p < 0.05 considered to be statistically significant using the “coxph” function.…”