Screening of immunosuppressive cells from colorectal adenocarcinoma and identification of prognostic markers

Abstract

Background: Colorectal cancer (CRC) is the most common type of gastrointestinal malignant tumour. Colorectal adenocarcinoma (COAD) - the most common type of CRC - is particularly dangerous. The role of the immune system in the development of tumour-associated inflammation and cancer has received increasing attention recently.

Methods: In the present study, we compiled the expression profiles of 262 patients with complete follow-up data from The Cancer Genome Atlas (TCGA) database as an experimental group and selected 65 samples from the Gene Expression Omnibus (GEO) dataset (of which 46 samples were with M0) as a verification group. First, we screened the immune T helper 17 (Th17) cells related to the prognosis of COAD. Subsequently, we identified Th17 cells-related hub genes by utilising Weighted Gene Co-expression Network Analysis (WGCNA) and Least Absolute Shrinkage and Selector Operation (LASSO) regression analysis. Six genes associated with the prognosis in patients with COAD were identified, including: KRT23, ULBP2, ASRGL1, SERPINA1, SCIN, and SLC28A2. We constructed a clinical prediction model and analysed its predictive power.

Results: The identified hub genes are involved in developing many diseases and closely linked to digestive disorders. Our results suggested that the hub genes could influence the prognosis of COAD by regulating Th17 cells' infiltration.

Conclusions: These newly discovered hub genes contribute to clarifying the mechanisms of COAD development and metastasis. Given that they promote COAD development, they may become new therapeutic targets and biomarkers of COAD.

Keywords: COAD; Immunosuppressive cells; Th-17 cells.

Figure 1. Flow chart

Figure 2. Quantifying immune cell infiltration in TCGA and GEO samples

(A) Kaplan–Meier curves of Th17 cells in immune cells. Genetic analysis of the TCGA and GEO databases showed that the high-invasion group had a significant inhibitory effect on the prognosis of COAD, and Th17 cells were associated with a better prognosis (GEO samples: P=0.009, HR = 5.79, 95% CI: 1.55−21.6, TCGA samples: P=0.001, HR = 3.11, 95% CI: 1.57−6.15). (B) Forrest plot of univariate Cox regression analysis in COAD (Th17 cells were associated with prognosis and HR < 1, both as protective factors in both TCGA and GEO).

Figure 3. Systematic clustering of 262 COAD tumour samples and clinical information

Figure 4. Scale-free conformance index and average connectivity calculated at different β values (the numbers in the figure represent the corresponding soft threshold power. An approximate scale-free topology can be achieved at a soft threshold power of 4)

Figure 5. Gene-cluster tree diagram

Based on consensus topological overlap, each colour module represents a colour-coding module containing a set of highly connected genes. Each module includes at least 50 genes (larger modules are relatively more meaningful).

Figure 6. Heat map of different modules with different clinical features

Each row corresponds to a consistency module, and each column corresponds to one type of clinical information (screening based on the residual tumour, pathological stage, overall survival (OS), OS events, and Th17 cells). The absolute value of the P-value indicates the correlation size (the module name is displayed on the left side of each cell, and the associated intensity and direction are shown on the right side of the heat map).

Figure 7. LASSO regression analysis

(A) Distribution of LASSO coefficients for eight related genes. (B) Partial likelihood bias of the LASSO coefficient distribution. The vertical dashed line indicates the minimum partial likelihood deviation.

Figure 8. Relationship between risk scores and the expression levels of the six hub genes

Figure 9. Time-dependent ROC curve analysis for clinical prediction models

Figure 10. Relationship between high- and low-risk scores and OS

The OS of the low-risk score group was significantly higher compared with the high-risk score group.