Transcriptional Levels of PBX in colorectal cancer patients
So far, researchers have identified four PBX family members in mammalian cells. However, their role in colorectal cancer progression is poorly understood. To investigate whether PBX family members are implicated in the progression of colorectal cancer, we identified the expression of PBX family members in colorectal cancer tumor tissue and normal tissue using the Oncomine database, which contains 465, 449, 443, and 258 unique datasets for PBX1, PBX2, PBX3, and PBX4. respectively. According to the Oncomine data, we found that the transcript levels of PBX1 and PBX3 were significantly reduced in colorectal cancer tissues compared to normal tissues, while the transcript levels of PBX4 were remarkably elevated (Fig. 1A). In addition, we further explored the relevant expression levels of PBX transcripts in colorectal cancer cell lines through the CCLE database, and the results are displayed in Fig. 1B.
To further examine the differential expression of PBX family members in patients with colorectal cancer, we used GEPIA dataset to compare the mRNA expression of PBX in colon and rectal cancer and their corresponding normal tissues, respectively. The results showed that PBX1, PBX2 and PBX3 were expressed at lower levels in colon and rectal cancer tumor tissues than in normal tissues, whereas PBX4 was expressed at significantly higher levels in colon and rectal cancer tumor tissues (Fig. 2). These results were generally in accordance with that of Oncomine.
Correlation between the mRNA Levels of PBX and the Clinicopathological Parameters of colorectal cancer patients
We investigated whether the expression of PBX was correlated with the pathological stages in colorectal cancer patients. The correlation between differentially expressed PBX and the pathological stages of colorectal cancer patients was evaluated using the GEPIA database. The results showed that the PBX1 and PBX2 groups were highly variable, whereas the PBX3 and PBX4 groups did not exhibit significant differences (Fig. 3). These data suggest that PBX1 and PBX2 may function essentially in the pathogenesis and progression of colorectal cancer.
To validate the transcriptional profiles of PBX family members, we analyzed the expression of PBX1, PBX2, PBX3 and PBX4 proteins in colon and rectal cancers and their corresponding normal tissues through the Human Protein Atlas database. Representative images of immunohistochemical staining showed that PBX1 and PBX3 proteins were barely detectable in colon or rectal cancer tissues, but were expressed in the corresponding normal tissues (Fig. 4). PBX2 protein was expressed in both colorectal cancer tumors and their normal tissues, but at much lower levels in tumor sections (Fig. 4). The expression level of PBX4 protein was increased in colon cancer than in normal colon tissues. Unfortunately, there was no remarkable difference in the expression of PBX4 protein between rectal cancer tissues and their corresponding normal tissues (Fig. 4).
Association of PBX mRNA expression with prognostic value of colorectal cancer patients
The present study next examined the relevance of PBX family members to clinical outcomes in colorectal cancer patients, using an online analysis tool (GEPIA) to analyze the expression profiles and clinical information of colorectal cancer patients in the TCGA cohort. As shown in Fig. 5, we present the overall survival (OS) and disease-free survival (DFS) curves for patients with colon and rectal cancer. The high expression level of PBX4 was negatively correlated with the OS of colon cancer patients. Besides, in rectal cancer patients, the expression of PBX2 was tightly correlated with shorter DFS, with higher expression levels resulting in shorter DFS in patients. However, PBX1 and PBX3 appeared to have no significant impact on OS and DFS in colon and rectal cancer patients.
Furthermore, to investigate the prognostic value of PBX family members in colorectal cancer, we used the Kaplan-Meier plotter database to assess the prognostic value of PBX family members. Interestingly, Kaplan-Meier curve and log rank test analysis revealed that elevated levels of PBX3 (HR = 6.27, p = 0.039) mRNA expression were significantly associated with longer OS in colorectal cancer patients (Fig. 6).
Analysis of gene mutations and interactions of PBX family members in colorectal cancer patients
Once the prognostic value of wild-type PBX family members was determined, we next investigated the genetic changes of PBX genes in colorectal cancer. Analysis of the TCGA dataset from an online tool, cBioPortal, showed that two or more alterations were detected in different subtypes of colorectal cancer. Additionally, the total mutation rate (missense mutations, truncations, amplifications and deletions) of PBX family members was 4% in a sample of 1949 colorectal cancer patients, and the individual mutation rates were 1.7%, 1.2%, 1% and 1.5% for PBX1, PBX 2, PBX3 and PBX4, respectively (Fig. 7A and 7B).
In addition, we analyzed the correlation between the members of the PBX family through the GEPIA database. The results showed that there was a significant positive correlation between both PBX1, PBX2 and PBX3, while a significant negative correlation existed between PBX1 and PBX4 (Fig. 7C). Subsequently, we constructed the protein-protein interaction network of PBX family members and their related genes using GeneMANIA. The results showed that PBX family members are closely implicated in sensory organ development, myeloid cell differentiation, and regulation of hematopoietic function (Fig. 7D).
Association of expression levels of PBX family members with immune cell infiltration in colon or rectal cancer
Events of immune infiltration in the tumor microenvironment (TME) function critically in the progression of tumors and influencing the patient's clinical outcome in cancer[28]. To investigate whether the expression of PBX family members was associated with the level of tumor immune infiltration, TIMER database was used to analyze the correlation between PBX family members and immune cell infiltration. The results showed that the expression of PBX1 mRNA was significantly correlated with the infiltration levels of B cells, CD4 + T cells and macrophages, but not with CD8 + T cells (COAD), neutrophils (READ) and dendritic cells (READ) in colon and rectal cancers (Fig. 8A). Additionally, PBX2 mRNA expression showed a remarkable positively correlation with the infiltration levels of CD4 + T cells, CD8 + T cells and macrophages in colon cancer, but not with neutrophils and dendritic cells in rectal cancer (Fig. 8B). Interestingly, PBX3 mRNA expression showed a significant positive correlation with the infiltration levels of all 6 immune cells in colon and rectal cancers (Fig. 8C). As for PBX4, it was observed that the expression level of its transcript dramatically correlated with the degree of infiltration of CD4 + T cells, CD8 + T cells and macrophages in colon cancer. However, it was only closely correlated with the infiltration levels of neutrophils in rectal cancer (Fig. 8D). Since tumor purity is an essential element influencing immuno-infiltration analysis of clinically tumor specimens, we further analyzed the correlation between PBX family members and tumor purity, and the results showed that only PBX3 mRNA expression was significantly correlated with tumor purity in colon and rectal cancers (Fig. 8).
Subsequently, we further analyzed the correlation of mRNA expression levels of PBX family members in colon and rectal cancers with markers of different types of immune cells, including typical markers of different types of immune cells such as CD8 + T cells, T cells, B cells, monocytes, tumor-associated macrophages, neutrophils, macrophages, DC cells, NK cells, etc., and functional T cell markers of different types such as Th1 cells, Th2 cells and Treg cells, using the TIMER database (Supplementary table 1–4). Adjusted for correlation by purity, we found that PBX1 and PBX3 mRNA expression levels significantly correlated with most immune marker genes of multiple immune cells and functional T cells in colon and rectal cancers, whereas PBX2 and PBX4 mRNA showed correlation with only a limited set of immune cell markers (Supplementary table 1–4).
Interestingly, we found a robust correlation between PBX1 mRNA expression and the expression levels of markers in Monocyte, TAM, M2 Macrophage, Neutrophils, and Treg cells. For example, marker genes CD86 and CSF1R for Monocyte, CCL2 and CD68 for TAM, and CD163 and MS4A4A for M2 Macrophage (Supplementary table 1). In addition, except for markers of B cell, Natural killer cell and T cell exhaustion in rectal cancer, PBX3 was significantly correlated with virtually all other types of immune markers in colon or rectal cancer (Supplementary table 3). The above results suggest that PBX1 and PBX3 mRNA expression levels are relevant to the degree of tumor immune infiltration in colon or rectal cancer.