Potential tumor-specific antigens and immune landscapes identification for mRNA vaccine in thyroid cancer

Abstract

Background: Tumor mRNA vaccines have been identified as a promising technology for cancer therapy in multiple cancer types, while their efficacy in thyroid cancer (THCA) is unclear. Immunotyping is strongly associated with the immune microenvironment and immune status in cancer, thus it is important in vaccination and therapeutic response. This study is to identify potential valuable antigens and novel immune subtypes of THCA for immune landscape construction, thus screening patients suitable for mRNA vaccination.

Methods: The clinical information and gene expression files of 568 THCA cases were obtained from the TCGA dataset. The DNA copy number variation and the somatic mutation of THCA were visualized by the cBioPortal database. TIMER was used to investigate the immune infiltrating correlation with candidate antigens. Consensus clustering analysis was conducted to cluster data using the ConsensusClusterPlus package. The immune landscapes of THCA patients were visualized using the Monocle package. The critical hub genes for THCA mRNA vaccines were identified by WGCNA package. To validate, the immunohistochemistry and real-time quantitative PCR (RT-qPCR) were performed to detect the expression level of potential antigen for mRNA vaccine in tissue and cell lines in THCA.

Results: Thymidine kinase 1 (TK1) was identified as a potential biomarker of mRNA vaccine against THCA. It was confirmed to be significantly upregulated in THCA tissues and cells lines. Moreover, three novel immune subtypes of THCA were obtained based on the expression consistency of immune-associated genes. The S2 subtype was characterized as an immunological "cold" phenotype with a high expression of immunogenic cell death modulators. S1 and S3 subtypes were immunological "hot" phenotypes with immune checkpoints upregulation. Further, the immune landscape of THCA patients was visualized and ten hub genes for mRNA vaccines were identified.

Conclusion: TK1 was a tumor-specific antigen of mRNA vaccines. The patients belonging to the S2 subtype ("cold" tumor) were suitable for mRNA vaccine therapy in THCA. Notably, ten hub genes were conducted as potential biomarkers for identifying suitable patients for mRNA vaccination. These findings provided novel insights into mRNA vaccine development against THCA.

Keywords: TK1; immune landscape; immunotype; mRNA vaccine; thyroid cancer; tumor immune microenvironment.

Figure 1

Identifying the candidate antigens in thyroid cancer (THCA). (A) Chromosomal distribution of up-regulated genes in THCA. (B) The aberrant copy number genes, amplified (red) and deleted (blue) were shown in THCA. (C) The overlap count of samples in altered genome fraction. (A), <=0.0001; (B), 0.0001-0.001; (C), 0.001-0.01; (D), 0.01-0.10; (E), >0.10. (D) The overlap count of samples in mutation count. (E) The genes with the highest frequency in the altered genome alteration group. (F) The genes with the highest frequency in the mutation count group. *P < 0.05, **P < 0.01, ***P <0.001, and ‘ns’ represents insignificant.

Figure 2

Identifying the prognosis-related potential antigens in thyroid cancer (THCA). (A) Forest maps of single factor survival analysis of genes in THCA. (B) The upset diagram indicates the number of gene intersections in distinct groups. (C) Survival analysis of thymidine kinase 1 (TK1) in THCA performed by Kaplan-Meier (KM) plotter. (D) Correlation between TK1 expression with the purity of infiltrating cells and amount of B cells, macrophages, and dendritic cells. ‘ns’ represents insignificant.

Figure 3

Identifying the novel immune subtypes of thyroid cancer (THCA). (A) Venn diagram screening genes from significant disease-free survival (DFS) and immune infiltration in THCA. (B) Cumulative distribution function curve of immune-related genes in THCA cohort. (C) Sample clustering heatmap. (D) Survival analysis of three immune subtypes. (E-I) Distribution of S1-S3 across THCA (E) age, (F) gender, (G) DFS status, and (H, I) tumor stages. ‘ns’ represents insignificant.

Figure 4

A correlation between cellular and molecular characteristics with three immune subtypes of thyroid cancer (THCA). (A) Heatmap of differential enrichment scores of 28 immune cell signatures among THCA immune subtypes. (B) Boxplot of differential enrichment scores of 28 immune cell signatures among THCA immune subtypes. (C) Heatmap of differential enrichment scores of immune-related pathways among three immune subtypes. (D) Boxplot of differential enrichment scores of immune-related pathways among three immune subtypes. (E) Differential expression of immune checkpoint (ICP) genes among THCA immune subtypes. (F) Differential expression of immunogenic cell death (ICD) modulator genes among THCA immune subtypes. *P < 0.05, **P < 0.01, ***P <0.001, **** P <0.0001 and ‘ns’ represents insignificant.

Figure 5

Immune landscape of thyroid cancer (THCA). (A) The immune landscape of THCA. The location of individual patients in the immune landscape with the color corresponding to the immune subtypes. (B) Correlation between PCA1/2 and immune modules. (C) Patients are separated by the immune landscape based on their location. (D) Separated patients were associated with different clinical outcomes. **P < 0.01, and ‘ns’ represents insignificant.

Figure 6

Immune gene co-expression modules of thyroid cancer (THCA). (A) Sample clustering. (B) Scale-free fit index for various soft-thresholding powers and their mean connectivity. (C) Dendrogram of all differentially expressed genes clustered based on a dissimilarity measure (1-TOM). (D) Gene numbers in each module. (E) Differential distribution of feature vectors of each module in THCA subtypes. **** P<0.0001 and ‘ns’ represents insignificant.

Figure 7

Identifying the hub genes of mRNA vaccine for thyroid cancer (THCA). (A) Forest maps of single factor survival analysis of six modules of THCA. (B) Bubble plot showing top 10 kyoto encyclopedia of genes and genomes (KEGG) terms in the brown module. (C) The top 10 enriched terms in gene ontology (GO) analysis of brown module. (D) Correlation between the brown module feature vector and PCA1 in the immune landscape. (E) Differential prognosis in the brown module with high and low mean. (F) Protein-protein interaction (PPI) network of hub genes identified in the brown module. *P < 0.05, **P < 0.01, ***P <0.001, and ‘ns’ represents insignificant.