Tumor Antigens

Gene expression profiles were utilized to decipher the TIME characterization of CRC samples with multiple bioinformatics tools. The immunophenoscore (IPS) was applied to assess the immune state of each sample. IPS is a scoring scheme that quantifies the immunogenicity of tumor samples using a variety of markers of immune response or immune tolerance. The higher the IPS z-score, the stronger the immunogenicity of the sample (31 (link)). To measure the infiltration abundance of immune cell populations in tumor tissues, two different tools were applied. First, we leveraged the R package ESTIMATE to infer the fraction of stromal and immune cells in CRC samples, and generated two scores including the immune and stromal scores (32 (link)). Next, to describe a more detailed landscape of immune cell types infiltration, we employed the single sample gene set enrichment analysis (ssGSEA) implemented in R package GSVA. The gene sets for marking each cell were obtained from the research of Charoentong, which stored 28 human immune cell subtypes (31 (link)).
In addition, to predict their putative response to immune checkpoint blockade (ICB), CRC samples were scored using T-cell inflammatory signature (TIS) and Tumor Immune Dysfunction and Exclusion (TIDE) approaches. TIS proposed by Ayers et al. was used to predict clinical response to PD-1 blockade. The signature was composed of 18 inflammatory genes associated with antigen presentation, chemokine expression, cytotoxic activity, and adaptive immune resistance (33 (link)). The TIDE algorithm (http://tide.dfci.harvard.edu/) integrates the expression signature of two primary mechanisms of immune evasions: T cell dysfunction and T cell exclusion, to model tumor immune evasion. Patients with higher TIDE score suggest the greater potential of tumor immune evasion; thus, these patients would derive worse immunotherapy response (34 (link)).

Immunophenoscore (IPS) is a superior predictor of response to anti-CTLA-4 and anti-PD-1 regimens, which quantify the determinants of tumor immunogenicity and characterize the intratumoral immune landscapes and cancer antigenomes 42 (link). The scoring scheme developed from a panel of immune-related genes belonging to the four classes: MHC-related molecules (MHC), checkpoints or immunomodulators (CP), effector cells (EC) and suppressor cells (SC). The weighted averaged Z score was calculated by averaging the samplewise Z scores of the four classes within the respective category and the sum of the weighted averaged Z score was calculated as the IPS. The Tumor Immune Dysfunction and Exclusion (TIDE) algorithm proposed by Jiang et al. was utilized to model distinct tumor immune evasion mechanisms 45 (link), including dysfunction of tumor infiltration cytotoxic T lymphocytes (CTLs) and exclusion of CTLs by immunosuppressive factors. A higher TIDE score indicated tumor cells more likely to induce immune escape, thus indicating a lower response rate to ICI treatment. The Estimation of Stromal and Immune Cells in Malignant Tumors using Expression Data (ESTIMATE) algorithm 46 (link), which takes advantage of the unique properties of the transcriptional profiles to infer the tumor cellularity as well as the tumor purity. By using the ESTIMATE algorithm, we calculated the immune and stromal scores to predict the level of infiltrating immune and stromal cells and these form the basis to infer tumor purity. Tumor tissues with abundant immune cell infiltration represented a higher immune score and lower level of tumor purity.

Gene ontology (GO) and Kyoto encyclopaedia of genes and genomes (KEGG) analyses were performed to view the differential potential bio function pathways between cluster 1 and 2. GO, and KEGG analysis was conducted by employing gene set variation analysis (GSVA) and gene set enrichment analysis (GSEA) with the R package ‘GSVA’²⁹ and ‘clusterProfiler’,³⁰ respectively.
Two types of immunogram were introduced to evaluate tumour immunogenicity, as previously reported.³¹, ³² Pathways including T cell immunity, absence of inhibitory molecules, absence of checkpoint expression, absence of inhibitory cells, recognition of tumour cells, trafficking and infiltration, priming and activation, innate immunity, interferon‐gamma response, proliferation, and glycolysis were evaluated by using ssGSEA with R package ‘GSVA’.
The gene set of immune escape‐related genes was obtained from the previous study.³³ As for single‐cell RNA‐seq analysis, genes that were not detected in over 80% of all samples were excluded.
The tumour immune landscape was depicted by the ESTIMATE algorithm, CIBERSORT algorithm, and xCELL algorithm. The ESTIMATE algorithm and xCELL algorithm were performed with R package ‘ESTIMATE’³⁴ and ‘xCell’,³⁵ respectively. The CIBEERSORT algorithm was performed as guided in https://cibersortx.stanford.edu/.³⁶, ³⁷R package ‘CellChat’ was introduced to predict cell–cell communication based on ligand‐receptor pairs between tumour cells and immunocytes.³⁸ Significant differential ligand‐receptors pairs between cluster 1 and cluster 2 were selected.

The online cBioPortal tool (version 5.0.2, http://www.cbioportal.org/, accessed on 7 August 2022) was inaugurated for worldwide researchers especially those interested in cancer genomics, gathering manifold data mainly from tissue samples (TCGA, International Cancer Genome Consortium) and malignant cells (Cancer Cell Line Encyclopedia), and others [19 (link)]. Herein, relying on a total of 293 samples from the TCGA cohort, the genome alteration status of PRCC was entirely detected and visualized by right of the cBioPortal for the dissection of potential tumor antigens.