Tumor-antigens and immune landscapes identification for prostate adenocarcinoma mRNA vaccine

Abstract

Prostate adenocarcinoma (PRAD) is a leading cause of death among men. Messenger ribonucleic acid (mRNA) vaccine presents an attractive approach to achieve satisfactory outcomes; however, tumor antigen screening and vaccination candidates show a bottleneck in this field. We aimed to investigate the tumor antigens for mRNA vaccine development and immune subtypes for choosing appropriate patients for vaccination. We identified eight overexpressed and mutated tumor antigens with poor prognostic value of PRAD, including KLHL17, CPT1B, IQGAP3, LIME1, YJEFN3, KIAA1529, MSH5 and CELSR3. The correlation of those genes with antigen-presenting immune cells were assessed. We further identified three immune subtypes of PRAD (PRAD immune subtype [PIS] 1-3) with distinct clinical, molecular, and cellular characteristics. PIS1 showed better survival and immune cell infiltration, nevertheless, PIS2 and PIS3 showed cold tumor features with poorer prognosis and higher tumor genomic instability. Moreover, these immune subtypes presented distinguished association with immune checkpoints, immunogenic cell death modulators, and prognostic factors of PRAD. Furthermore, immune landscape characterization unraveled the immune heterogeneity among patients with PRAD. To summarize, our study suggests KLHL17, CPT1B, IQGAP3, LIME1, YJEFN3, KIAA1529, MSH5 and CELSR3 are potential antigens for PRAD mRNA vaccine development, and patients in the PIS2 and PIS3 groups are more suitable for vaccination.

Keywords: Immune landscape; Immune subtypes; Immunotherapy; Prostate adenocarcinoma; Tumor antigens; mRNA vaccine.

Fig. 1

Identification of potential PRAD-specific antigens for mRNA vaccine development. A Distribution of the upregulated and downregulated genes across the chromosomes. B Mutation status of the top 20 mostly mutated genes of each PRAD sample. C Samples overlapping in altered genome fraction and mutation count groups. D Potential tumor antigens associated with the overall survival (OS) of PRAD (13 genes). E Potential tumor antigens associated with disease-free PRAD interval (70 genes). F Identification of tumor antigens associated with both the OS of disease-free PRAD interval (8 genes). G Kaplan–Meier survival curves of patients with PRAD according to the expression of eight potential tumor PRAD antigens. H Correlation of IQGAP3, CELSR3, and KIAA1529 expression with the infiltration of B cells, macrophages, and dendritic cells

Fig. 2

Identification of immune subtypes and immune landscape of PRAD. A Identification of the clusters of TCGA-PRAD cohort using partition around medoids algorithm. B-C Survival comparison among the PRAD immune subtypes in the training cohort and validation cohort. D The prediction of the response to anti-PD-L1 immunotherapy for PRAD immune subtypes. E The distribution of PIS1, PIS2, and PIS3 in the groups with or without biochemical recurrence. F Association between PRAD immune subtypes and existing pan-cancer immune subtypes. G Copy number variation (CNV) across chromosomes across the PRAD immune subtypes. H-I Mutation counts and tumor mutation burden across PIS1, PIS2, and PIS3. J Homologous recombination deficiency score for each PRAD immune subtype. K The comparison of mRNA stemness index across PIS1 to PIS3. L Immune landscape of PRAD. Each point represents a patient and the immune subtypes are color-coded. M Association between two principal components and immune cells. N Immune landscape of the PIS1 and PIS3 subsets. O Immune landscape of four subsets of samples from extreme locations. P Survival curves of four subsets of samples from extreme locations. * P < 0.01, ** P < 0.001, and *** P < 0.0001