CRYAB predicts clinical prognosis and is associated with immunocyte infiltration in colorectal cancer

Data source

The Cancer Genome Atlas (TCGA) (https://genome-cancer.ucsc.edu/) provides scholars and researchers with clinical and pathological information on 33 types of cancer. Data of colorectal cancer (CRC) patients with RNA-Seq expression and matching clinicopathological information was obtained through the TCGA tool cancer browser. Because the database is publicly available and accessible, approval from the local ethics committee was not necessary.

GEO database

The GEO database is a comprehensive gene expression library in the National Center for Biotechnology Information (NCBI) (https://www.ncbi.nlm.nih.gov/geo/) and is one of the largest collections of gene chips in the world.

Immunochemistry

Fresh colorectal cancer tissues and normal tissues were collected, immediately immersed in 4% paraformaldehyde overnight, dehydrated by an ethanol series (70%, 80%, 90%, and 100%), clarified in xylene, and paraffin-embedded. They were subjected to a temperature of 60 °C for 2 h, deparaffinized, hydrated with xylene and ethanol, and then the recovered nuclear antigen was washed with PBS and hydrogen peroxide solution. Additional slices were randomly selected and incubated with rabbit anti-CRYAB antibody (#15808-1-AP; Proteintech, Wuhan, China) at 4 °C overnight, followed by HRP incubated goat anti-rabbit mouse universal antibody (1: 3000, K5007, DAKO, Denmark) for 60 min at room temperature. After the slides were placed in PBS and decolorized, they were developed with DAB chromogenic solution. In addition, all slides (nuclei) were counterstained with 5 µg/mL Harris at room temperature for 3 min. Finally, the image was captured under a microscope (Nikon Eclipse E200; Tokyo, Japan).

Statistical analysis of survival

Patients in the experimental group and the validation groups were divided into two subgroups according to the median expression of the CRYAB gene: a high CRYAB expression group and a low CRYAB expression group. The effect of CRYAB expression level on the clinical outcomes of CRC patients was investigated using Kaplan–Meier (KM) survival curves, and a prognostic classifier was constructed to compare survival differences. The KM survival curve was implemented using the “survminer” package for analysis and visualization.

Analysis of CRYAB expression levels in colorectal cancer and normal colorectal samples

To compare CRYAB expression patterns between tumor and normal tissues, the differential expression of CRYAB was examined in TCGA datasets using the Tumor Immunity Estimation Resource (TIMER) 2.0, a comprehensive online resource for the analysis of immune infiltrates and gene expression in different cancer types (Li et al., 2017a; Li et al., 2017b; Li et al., 2020). In TIMER2.0, we selected “Exploration,” then “Gene_De,” followed by entering the gene name “CRYAB” in the text box and clicking “submit.” Oncomine is a publicly accessible cancer gene chip database and web-based data mining platform containing 715 data sets and 86,733 samples (Rhodes et al., 2004; Rhodes et al., 2007). We searched the Oncomine server for human CRC, and chose the differential gene analysis segment (Normal Analysis vs. Cancer) to retrieve the results.

TIMER database analysis

The association between the expression of CRYAB and the presence of five infiltrating immune cells (neutrophils, macrophages, CD4 + T cells, CD8 + T cells) and three immune-related genes (CD2, DC3D and CD3E) in CRC patients was evaluated using the TIMER database (http://timer.cistrome.org/).

Meta-analysis

A meta-analysis of data from the TCGA, GEO, and ICGC databases was performed to evaluate the significance of CRYAB expression in CRC prognosis. Heterogeneity among the included studies was determined by the I²-value obtained from the Cochrane Q test and the P-value obtained from the chi-square test. In cases of heterogeneity (I² ≥ 50% or P < 0.05), the results were summarized using a random-effects model. Otherwise, a fixed-effect model was used for analysis. The “meta” R package (R version 4.0.0) was used to perform the meta-analysis.

GO and KEGG enrichment analyses

GO and KEGG enrichment analyses were performed using R 4.0.2 and the R packages “org. Hs.eg.db,” “ggplot2,” “Cluster Profiler,” and “enrich plot.” Only terms with P-value < 0.05 were considered significantly enriched.

Patient characteristics

The RNA sequencing data of 306 samples from the TCGA database and detailed clinical prognostic information were included in the analysis. Patients were divided into a low expression group (n = 153) and a high expression group (n = 153). Age, gender, and tumor size did not differ significantly between the high and low expression groups (P = 0.1695, 0.6467, 0.3418), however rates of lymphatic metastasis, metastasis and stage were significantly different between the high and low expression groups (P = 8.00E−04, 0.0372, 0.0101) (Table 1).

CRYAB expression is higher in normal tissues than in tumor samples

The CRYAB gene is significantly downregulated in CRC (Fig. 1A). TIMER analysis of CRYAB expression in various cancer types showed that CRYAB expression was significantly lower in CRC tumor tissues than in normal tissues (Fig. 1B).

An analysis of the mRNA expression levels of CRYAB in TCGA samples showed that CRYAB expression was lower in tumor samples than in normal tissues (P <  2.22e−16) (Fig. 1C). This was verified in the GEO database (P = 0.045) (Fig. 1F). A meta-analysis of the above three datasets was performed to evaluate the correlation between overall survival (OS) and CRYAB gene expression and to obtain more objective conclusions. Because there was no statistically significant difference between the three datasets (P = 0.55, I² = 0%), a fixed effects model was used to evaluate the combined hazard ratio (HR) and 95% confidence interval (CI). A relatively high expression of the CRYAB gene was significantly correlated with poor OS (HR = 1.23, 95% CI [1.11–1.35], P <  0.0001; Fig. 1G), indicating that CRYAB may be a predictor of poor OS.

Higher CRYAB mRNA expression in CRC is associated with shorter OS

According to the KM chart, CRC cases with higher CRYAB mRNA expression had a shorter OS (P = 0.027) (Fig. 1D) and progression-free survival (P = 0.027) in the test cohort (Fig. 1E). The results showed that the expression of CRYAB was lower in colorectal tumors than in normal colorectal tissues.

Colorectal cancer CRYAB expression analysis

The expression of CRYAB in CRC was confirmed by immunohistochemistry (Figs. 2A–2F). The results confirmed that CRYAB expression was lower in CRC tissues than in normal colorectal tissues. We cut three sets of immunohistochemical slices, measured the positive rate of CRYAB among them, and drew the corresponding histogram. the positive rate of CRYAB was higher in tissue adjacent to tumors than in actual cancer tissue (p < 0.0001) (Fig. 2G)

Correlation between CRYAB expression and clinical characteristics

We analyzed the association between CRYAB mRNA expression and clinicopathological parameters in CRC patients and found that CRYAB level was not correlated with gender (P = 0.63) (Fig. 3B). We also found that CRYAB expression decreased with age (P = 0.026) (Fig. 3A), and that increased CRYAB expression was associated with distant metastasis from CRC (Fig. 3C). Low expression of CRYAB was associated with colorectal cancer progression from N0 to N1 and from N0 to N2 (P < 0.001, 0.01) (Fig. 3D). Low expression of CRYAB was also associated with CRC tumor size progression from T2 to T3, with an even stronger association of low CRYAB expression with the progression from T2 to T4 (P = 0.07, 0.02). In terms of lymph node metastasis, the expression level of CRYAB did not change after N1 (Fig. 3E). High CRYAB expression was associated with tumor progression from Stage I to III, from Stage I to IV, from Stage II to II, and from Stage II to IV (P < 0.001, 0.001, 0.05, 0.01). The level of CRYAB gene methylation was not associated with age, sex, TNM stage, or tumor stage (Figs. 3G–3L).

Correlation analysis between infiltrating immune cells and CRYAB expression

Tumor infiltrating lymphocytes impact cancer survival (Mou et al., 2021). Therefore, we analyzed the correlation between CRYAB expression and four infiltrating immune cells (neutrophils, macrophages, CD8 + T cells, and CD4 + T cells) and three immune-related genes (CD2, CD3D, and CD3E). The results showed that the expression level of CRYAB is comparable to neutrophils (r = 0.364, P = 4.6e−10), macrophages (r = 0.515, P = 4.23e−20), CD4+ T cells (r = 0.321, P = 4.99e−08), CD8 + T cells (r = 0.134, P = 2.6e−02), CD2 (r = 0.176, P = 1.49e−04), CD3D (r0.176, r = P = 1.12e−02) and CD3E (r = 0.209, P = 6.45e−06) (2021.5.9) and that infiltration levels were significantly positively correlated with CRYAB expression with significance defined as P < 0.05 (Fig. 4). In addition, CRYAB expression was significantly correlated with markers of M2-like macrophages in pan-carcinoma and CRC, including TGFB1 (Fig. 5A and 5B), MRC1 (Fig. 5C and 5D), and CD163 (Fig. 5E and 5F). These results indicate that in addition to the immunosuppressive microenvironment of colorectal cancer, high CRYAB expression was related to macrophage infiltration and polarization.

DNA methylation analysis

The degree of methylation of cg12598198 was the highest, followed from high to low by cg14276286, cg15227610, cg1158277, cg15204861, cg12947833, cg15318568, cg00514609, cg07476508, cg10048349, cg13210534, cg15545878, and cg13084335 (Fig. 6A). Among them, methylation sites such as cg13084335, cg15545878, cg13210534, and cg15318568 were positively correlated with low expression of CRYAB. The sites cg12598198, cg14276286, cg15227610, cg1158277, cg15204861, cg12947833, cg00514609, cg07476508, and cg10048349 and other methylation sites were negatively correlated with low expression of CRYAB (Figs. 6B–6O). The methylation level of CRYAB was lower in CRC than in the normal control group, indicating that changes in methylation are related to the abnormal expression of CRYAB. In addition, DNA methylation may be related to the molecular mechanism underlying the low expression level of CRYAB in tumor tissues and also to the pathogenesis of CRC.

GO and KEGG enrichment analyses

We performed a GO enrichment analysis and a KEGG pathway analysis. Fig. 3A shows the top 30 significantly enriched upregulated pathways. The GO pathway analysis showed that relatively high expression of CRYAB was associated with extracellular structure organization (GO:0043062, P = 2.13E−64), ECM organization (GO:0030198, P = 1.76E−64), collagen-containing ECM (GO:0062023, P = 3.59E−85), and ECM structural constituent (GO:0005201, P = 3.86E−70) (Fig. 7A). The KEGG pathway analysis identified many enriched pathways, including PI3K-Akt signaling pathway, focal adhesion, human papillomavirus infection, ECM-receptor interaction, and protein digestion and absorption, and that CRYAB genes were strongly linked to them (Fig. 7B)