A multi-cancer analysis unveils ITGBL1 as a cancer prognostic molecule and a novel immunotherapy target

Date source and analysis

After being uniformly processed by the Toil process, the RNAseq data in TPM format was downloaded from UCSC XENA (https://xenabrowser.net/datapages/) and the mRNA data for all 33 cancers was acquired from the TCGA (https://www.cancer.gov/ccg/research/genome-sequencing/tcga) and GTEx (https://www.gtexportal.org/home/) databases [13]. The RNAseq data in TPM (transcripts per million reads) format were log2-transformed and analysed by the Mann–Whitney U test to contrast ITGBL1 expression in paraneoplastic and tumour tissues. Differential expression of ITGBL1 with a p value analysis less than 0.05 was identified as differentially expressed. The 33 varieties of cancer included adrenocortical carcinoma (ACC), bladder urothelial carcinoma (BLCA), breast invasive carcinoma (BRCA), cervical squamous cell carcinoma and endocervical adenocarcinoma (CESC), cholangiocarcinoma (CHOL), colon adenocarcinoma (COAD), lymphoid neoplasm diffuse large B-cell lymphoma (DLBC), esophageal carcinoma (ESCA), glioblastoma multiforme (GBM), head and neck squamous cell carcinoma (HNSC), kidney chromophobe (KICH), kidney renal clear cell carcinoma (KIRC), kidney renal papillary cell carcinoma (KIRP), acute myeloid leukemia (LAML), brain lower grade glioma (LGG), liver hepatocellular carcinoma (LIHC), lung adenocarcinoma (LUAD), lung squamous cell carcinoma (LUSC), mesothelioma (MESO), ovarian serous cystadenocarcinoma (OV), pancreatic adenocarcinoma (PAAD), pheochromocytoma and paraganglioma (PCPG), prostate adenocarcinoma (PRAD), rectum adenocarcinoma (READ), sarcoma (SARC), skin cutaneous melanoma (SKCM), stomach adenocarcinoma (STAD), testicular germ cell tumors (TGCT), thyroid carcinoma (THCA), thymoma (THYM), uterine corpus endometrial carcinoma (UCEC), uterine carcinosarcoma (UCS), and uveal melanoma (UVM).

Colorectal and gastric cancer datasets were downloaded from the GEO database (https://www.ncbi.nlm.nih.gov/geo/). The dataset GSE9348 for colorectal cancer contains 12 normal tissues and 70 tumour tissues. The dataset GSE44706 for colorectal cancer contains 148 normal tissues and 98 tumour tissues. The dataset GSE19826 for gastric cancer contains 15 normal tissues and 12 tumour tissues. The dataset GSE54129 for gastric cancer contains 21 normal tissues and 111 tumour tissues.

Clinical sample sources

Six sets of tissues, one for each colorectal cancer and one for paracancerous tissue, were taken from patients receiving treatment for colorectal cancer at the Second Affiliated Hospital of Suzhou University’s Department of General Surgery. These samples were preserved using liquid nitrogen. Every hospitalised patient gave their informed consent for this study. Clinical excision sample use was authorised by the Ethics Committee of the Second Affiliated Hospital of Soochow University (approval number: JD-LK2023001-IR01). Clinical samples were collected and used in strict accordance with the guidelines.

Clinicopathological stage analysis

The link between ITGBL1 across cancer types and clinicopathological stage was examined using GEPIA (http://gepia.cancer-pku.cn/). Using the pathological stage as the variable, differential expression was computed using one-way analysis of variance (ANOVA), the method of differential gene expression analysis. Expression data were first converted to log₂(TPM+1) for differential analysis.

Cancer prognosis analysis

An analysis of survival performed on RNAseq data obtained from the TCGA database in HTSeq-FPKM format was statistically examined by log-rank and included OS, DSS, and PFI. Data on survival were visualised using the survminer package (version 0.4.9, https://cran.r-project.org/web/packages/survminer/) and statistical analysis was performed using the survival package (version 3.2-10, https://cran.r-project.org/web/packages/survival/). These packages come from the R project (version 4.1.3, https://www.r-project.org/). The prognostic supplemental data were from a Cell article [14].

Gene set enrichment analysis

The Molecular Signature Database (MSigDB, https://www.gseamsigdb.org/gsea/index.jsp) website provided a “GMT” file with 50 hallmark gene sets. This file was used to compute the normalised enrichment score (NES) and false discovery rate (FDR) for each biological process for every type of cancer. The R package “clusterProfiler” (https://bioconductor.org/) was used to carry out the gene set enrichment analysis [15], and the outcomes are summarised in the bubble plots described in the R package “ggplot2” (https://cran.r-project.org/web/packages/ggplot2/). These packages come from the R software (https://www.r-project.org/).

Immunological correlation analysis

The TIMER2.0 database (http://timer.cistrome.org/) was utilised to examine the relationship between ITGBL1 and 19 distinct categories of immune cells that infiltrate, such as B cells, CAFs, CD4⁺ T cells, CD8⁺ T cells, progenitor cells, dendritic cells, endothelial cells (Endos), eosinophils (Eos), γδ T cells, stem cells, macrophages, mast cells, monocytes, neutrophils, NK cells, NK T cells, T-cell follicular helper (TfH) cells and regulatory T cells (Tregs). TIMER2.0 has the ability to use algorithms such as EPIC, xCell, TIMER, CIBERSORT and others to assess the significance of ITGBL1 and other immune cells that infiltrate cancerous tissue.

ITGBL1 and immune subtypes as well as molecular subtypes in pancancer were found to be related using the TISIDB database (http://cis.hku.hk/TISIDB/). The main immune and molecular subtypes are the following: C1 (wound healing), C2 (IFN-gamma dominant), C3 (inflammatory), C4 (lymphocyte depleted), C5 (immunologically quiet), and C6 (TGF-β dominant).

Drug sensitivity of ITGBL1

Data processed for drug sensitivity was uploaded from the CellMiner™ database (https://discover.nci.nih.gov/cellminer/home.do) [16]. For 60 cancer cell lines, drug sensitivity data is available in the CellMiner™ database. The R packages “impute” (https://bioconductor.org/), “limma” (https://bioconductor.org/), “ggplot2” (https://cran.r-project.org/web/packages/ggplot2/), and “ggpubr” (https://cran.r-project.org/web/packages/ggpubr/index.html) were employed for analysis and visualisation of all the data.

Cell culture

Human gastric cancer cell line MGC803 and the human colorectal cancer cell line HCT-8 were used in the study. The HCT8 cells were purchased from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). The MGC803 cells were purchased from Beyotime Biotechnology (Shanghai, China). All cells were tested for mycoplasma. RPIM (cat No.: SH30809.01, HyClone, USA) medium containing 10 % foetal bovine serum (FBS, #A5669701, Gibco, Thermo Fisher Scientific, USA) and 1 % penicillin streptomycin (#15140122, Gibco, Thermo Fisher Scientific, USA) was employed to cultivate HCT-8 and MGC803 cells at 37 °C and 5 % CO₂.

Western blot

We lysed the cells and tissues with radio-immunoprecipitation assay (RIPA) buffer (Cat# R0020, Solarbio, Beijing, China). 5× SDS-PAGE Protein loading buffer (Cat: 20315ES20, YEASEN, Shanghai, China) was used to prepare the proteins. The primary antibodies included ITGBL1 antibody (Absin, Shanghai, China), horseradish peroxidase (HRP)-labelled goat anti-rabbit IgG (H+L) (Beyotime, Shanghai, China) and horseradish peroxidase (HRP)-labelled goat anti-mouse IgG (H+L) (Beyotime, Shanghai, China). We used polyvinylidene difluoride (PVDF) film (IPVH00010, Merck Millipore, USA) for transfer and 5 % skimmed milk powder for closure and then applied the primary antibody at 4 °C for the entire night and the secondary antibody for 1 h the following day at room temperature. An ultrasensitive electrochemiluminescence (ECL) chemiluminescence kit (Cat. No: P10100, NCM Biotech, Suzhou, China) was used to detect protein expression levels. Amersham ImageQuant 800 (cytiva, USA) was used for the detection and analysis of proteins in membranes. ImageJ (v1.54a) was used for quantitative analysis.

CCK8 proliferation assay

Cell Counting Kit-8 (Cat. No: C6005, NCM Biotech, Suzhou, China) was employed in order to identify the proliferation of cancer cells. We spread cancer cells from the negative control and experimental groups in 96-well plates at 2000 cells per well and set up five replicate wells. At 0 h, 24 h, 48 h and 72 h, 10 μL of CCK8 solution was added to each well and incubated at 37 °C for 1 h. The optical density (OD) values at 450 nm were measured using a microplate spectrophotometert (EPOCH-SN, BioTek Epoch, USA) under light-avoiding conditions. GraphPad Prism software (v9.0.0) was used to assess the proliferative capacity of the different groups.

Migration assay

We first spread the negative control and experimental group cells in six-well plates, and after the cells reached 90 % confluence, a straight line was gently scratched in the centre from top to bottom with the tip of the pipette. Finally, the plate was washed three times with phosphate buffer saline (1× PBS, pH 7.4, #C10010500BT, gibco, USA), an appropriate amount of culture medium was added to each well, and microscopic photographs were captured at 0 and 48 h to evaluate the capacity to heal scratches. A fluorescence microscope (IX73, Olympus Corporation, Tokyo, Japan) was used for image acquisition. GraphPad Prism software (v9.0.0) was used to analyse and compare migration capabilities.

Statistical analysis

An unpaired t-test was used to compare the two groups, and the data were presented as the mean±standard deviation. The Kaplan–Meier method was employed to assess the relationship between prognosis and ITGBL1 expression levels in patients. p <0.05 was the threshold for statistical significance in this case. The data analysis and statistics were performed using GraphPad Prism software (v9.0.0).

Analysis of ITGBL1 expression in normal and tumour tissues

By means of RNAseq data analysis from the GTEx and TCGA databases, we ascertained the variation of ITGBL1 expression in 33 different types of cancer. According to the results, ITGBL1 was expressed differently in 27 cancers, and a statistically significant difference was evident with a p-value <0.05. Among them, the expression of ITGBL1 was greater in tumour tissues compared to normal tissues in 13 cancers (Figure 1A). To increase the reliability of the data, we also performed ITGBL1 validation through the GEO database. We examined the expression of ITGBL1 in normal and tumour tissues using the colorectal cancer datasets GSE9348 and GSE44706, and the results showed that the expression of ITGBL1 in tumour tissues was elevated (Figure 1B). Comparably, ITGBL1 expression in the gastric cancer datasets GSE19826 and GSE54129 showed similar final results (Figure 1C).

Figure 1:

Expression of ITGBL1 in pancancerous tissues. (A) Evaluation of transcript level expression of ITGBL1 in tumour and normal tissues. (B, C) Variations in ITGBL1 expression between tumour and healthy tissues in the GEO dataset. (D) ITGBL1 expression in colorectal cancer samples and surrounding tissues was confirmed by western blot analysis. *p<0.05, **p<0.01, ***p<0.001. ns, not statistically significant.

Furthermore, we confirmed ITGBL1’s protein expression in six pairs of colorectal and paracancerous tissues obtained from clinical samples, and the findings demonstrated that colorectal cancer tissues had elevated levels of ITGBL1 (Figure 1D). Fascinatingly, this outcome is in line with the GEO database analysis, opening the door for a thorough investigation into ITGBL1 as a potential high-risk factor in cancer.

Clinicopathological stages

The influence of ITGBL1 on tumour pathological stages was examined using the GEPIA database. We screened and analysed tumours with statistically valid results and found that in ACC, ESCA, HNSC, BLCA, KIRP, STAD, THCA, and OV, ITGBL1 was substantially linked to the formation of tumours (Figure 2). In ACC, ESCA, HNSC, BLCA, KIRP, and THCA, the expression of ITGBL1 tended to increase with tumour progression, indicating a positive correlation between ITGBL1 and the progression of these tumours. However, in stage IV, ESCA and OV, ITGBL1 was expressed at low levels (Figure 2B and H). Figure 2F reveals an opposite trend of ITGBL1 in STAD progression, indicating a negative correlation between ITGBL1 and STAD staging (Figure 2F). In summary, ITGBL1 may have important diagnostic value in the progressive stage of tumours.

Figure 2:

Relationships between ITGBL1 expression and major pathological stages. Clinicopathological stages included stage I, stage II, stage III, and stage IV of ACC (p=0.003), ESCA (p<0.05), HSNC (p<0.05), BLCA (p<0.001), KIRP (p<0.001), STAD (p=0.001), THCA (p<0.001) and OV (p<0.05).

Analysis of the prognostic significance of ITGBL1 in pancancer

To investigate more thoroughly the necessity of ITGBL1 in the progression of tumours, we used univariate Cox regression analysis and three types of survival analysis, including OS, DSS, and PFI, to investigate the prognostic significance of ITGBL1. The Cox risk proportional model revealed that ITGBL1 had a connection with the OS of ACC (p=0.036), BLCA (p=0.005), CESC (p=0.02), COAD (p=0.007), DLBC (p=0.041), GBM (p<0.001), KIRP (p=0.001), LAML (p=0.025), LGG (p<0.001), MESO (p=0.048), OV (p=0.03) and STAD (p=0.018) (Figure 3A). Except in DLBC and LAML, where ITGBL1 is a low-risk factor, ITGBL1 is a high-risk factor for every other type of cancer. Kaplan–Meier survival analysis was also consistent, with high expression of ITGBL1 in ACC, BLCA, CESC, COAD, GBM, KIRP, LGG, MESO, OV, and STAD predicting a worse outcome. In contrast, patients with high ITGBL1 expression in LAML and DLBC had a longer OS (Figure 3B–M). In addition, high expression of ITGBL1 in patients with ACC, BLCA, CESC, COAD, GBM, KIRP, LGG, MESO, and OV was also observed in DSS, predicting shorter survival (Figure S1). Similar findings were found by the PFI analysis, which indicated that high ITGBL1 expression and poor prognosis were associated with these cancers (Figure S2). A combination of the three survival analyses suggests that ITGBL1 is likely to be a good marker of survival prognosis.

Figure 3:

Examination of the connection between ITGBL1 expression and overall survival (OS). (A) Forest plot for the association between OS and ITGBL1 in 33 different disease categories. (B–M) The Kaplan–Meier survival analysis examines the relationship between OS and ITGBL1 expression.

GSEA and immunological correlation

To further investigate the possible functions of ITGBL1 in tumour mechanisms, we searched for ITGBL1-associated cancer hallmarks by analysing differentially expressed genes (DEGs) between high- and low-expressing ITGBL1 subgroups in 33 cancers and performing GSEA. According to the GSEA results, ITGBL1 was extensively enriched in immune-related pathways across cancers, including the inflammatory response, IL-6 JAK-STAT3 signalling, IL-2 JAK-STAT3 signalling, KRAS signalling pathway and allograft rejection pathway (Figure 4A). Janus kinase (JAK) signal transducer and activator of transcription (STAT) pathways perform a crucial function in immune system modulation, especially cytokine receptors, which regulate the polarization of T helper cells [17, 18]. KRAS mutations are associated with tumour proinflammatory responses and KRAS-induced immune regulation, leading to immune escape in the tumour microenvironment [19]. Allograft rejection, which T cells are involved in suppressing, is inextricably linked to the immune system [20]. These results indicated that ITGBL1 was likely to be involved in the tumour microenvironment and related to tumour infiltrating immune cells, paving the way for further studies on immune function.

Figure 4:

Related functional pathways and immunophenotyping of ITGBL1. (A) The hallmark gene set enrichment analysis and immune subtypes of ITGBL1 across cancers. The FDR value is indicated by the size of the circles, and the NES score is denoted by the colour. (B) Correlation of ITGBL1 expression with pancancer immune subtypes investigated by TISIDB in ACC, BLCA, CESC, COAD, LGG, KIRP, MESO and OV. C1 (wound healing); C2 (IFN-gamma dominant); C3 (inflammatory); C4 (lymphocyte depleted); C5 (immunologically quiet); C6 (TGF-β dominant). (C) ITGBL1 expression correlates with TISIDB-investigated molecular subtypes in COAD, KIRP, LGG, and OV.

In addition, ITGBL1 was significantly characterized in EMT in all 33 cancers, which was related to the function of ITGBL1 itself (Figure 4A). EMT enables tumour cells to acquire mesenchymal cell properties, increasing the invasiveness and drug resistance of cancer cells while suppressing antitumour immune responses [21]. ITGBL1 also participated in the ultraviolet light (UV) response, myogenesis pathway, etc.

Based on the results of GSEA across cancers, ITGBL1 is probably involved in the immune regulation of tumours, so it is necessary to investigate the immune correlation between ITGBL1 and tumours. We first used TISIDB to analyse whether ITGBL1 correlates with immune subtypes and molecular subtypes. The results indicated a relationship between ITGBL1 and immune subtypes in ACC (p=1.81e-02), BLCA (p=3.56e-05), CESC (p=1.94e-02), COAD (p=8.27e-03), LGG (p=9.71e-03), KIRP (p=9.55e-04), MESO (p=7.18e-02) and OV (p=1.57e-05). Among the eight kinds of cancers, the C4 type showed low expression of ITGBL1 (Figure 4B). The expression of ITGBL1 was statistically significant in the molecular subtypes COAD (p=4.16e-02), KIRP (p=1.95e-04), LGG (p=1.7e-21), and OV (p=7.29e-24) (Figure 4C). In the HM-indel subtype of COAD, ITGBL1 expression was low, whereas in the CIN subtype, it was highly expressed. ITGBL1 is expressed at low levels in the C1 isoform of KIRP and highly expressed in the C2c-CIMP isoform. There was relatively low expression of ITGBL1 in the G-CIMP-LOW isoform of LGG and relatively high expression of ITGBL1 in the PA-like isoform. ITGBL1 is upregulated in the mesenchymal isoform of OV, while it is downregulated in the proliferative isoform.

Analysis of infiltrating immune cells in pan carcinoma

Together, various immune cell subsets impact the tumour microenvironment and further impact the effectiveness of tumour immunotherapy [22], so we made use of the TIMER2.0 database to investigate whether the expression level of ITGBL1 could be corroborated with infiltrating immune cells in a range of cancers. The results indicated that the relevance of ITGBL1 and infiltrating immune cells was remarkable in most cancer species, especially B cells, CAFs, dendritic cells, Endos, stem cells, macrophages, monocytes, neutrophils, NK cells, and Tregs (Figure 5). We screened seven cancers and found that the expression of ITGBL1 had a favourable association with B cells and CD8⁺ T cells in COAD, ESCA, HNSC, LGG, OV, STAD, and THCA (Figure S3). However, the expression trend of ITGBL1 in LGG was opposite to that in CD4⁺ T cells, neutrophils, macrophages and myeloid dendritic cells (Figure S3d). In the ESCA, there was a negative correlation between neutrophil expression and ITGBL1, while the rest of the infiltrating immune cells showed a positive tendency (Figure S3b).

Figure 5:

The relevance of B cells and infiltrating immune cells in pancytosis. Infiltrating immune cells include B cells, CAFs, CD4⁺ T cells, CD8⁺ T cells, progenitor cells, dendritic cells, Endos, EOSs, γδ T cells, stem cells, macrophages, mast cells, monocytes, neutrophils, NK cells, NKT cells, TFH cells and regulatory T cells. Negative correlations are shown in blue, and positive correlations are shown in red.

Immunotherapy related to ITGBL1

The advent of immunotherapy has revolutionised the way that tumours are discovered and given advanced cancer patients new options for treatment [23]. In immunotherapy, immunostimulatory substances prime the body’s defences to identify and combat cancerous cells, while immunosuppressive factors cause cancer cells to undergo immune escape. According to the heatmap, positive correlations were observed between ITGBL1 and the majority of immunostimulatory and inhibitory factors (Figure 6A and B). Vital immune system mediators are chemokines and chemokine receptors, and the heatmap demonstrates that ITGBL1 is positively correlated with both in the majority of cancers (Figure 6C and D) [24]. As an illustration, CXCL12 and its receptor CXCR4, which are extremely related to ITGBL1, can promote tumour growth, angiogenesis and metastasis and are an excellent immunotherapeutic strategy [25].

Figure 6:

Correlation of ITGBL1 and immune regulation. (A) The heatmap depicts the correlation between ITGBL1 and immunostimulatory factors. (B) The heatmap indicates the correlation between ITGBL1 and immunosuppressive factors. (C) The heatmap depicts the correlation between ITGBL1 and chemokines. (D) The heatmap represents the correlation between ITGBL1 and chemokine receptors. (E) Correlation of ITGBL1 expression and TMB. (F) Link between MSI and the expression of ITGBL1. *p<0.05, **p<0.01, ***p<0.001.

TMB and MSI are promising biomarkers for prediction in tumour immunotherapy [26]. According to Figure 6E, the expression of ITGBL1 was remarkably associated with TMB in 11 out of 33 cancers, including PRAD, THYM, LAML, LIHC, STAD, KIRP, BRCA, KIRC, LUSC, SKCM, and LUAD (Figure 6E). ITGBL1 was positively correlated with the TMB of PRAD, THYM and LAML but negatively correlated with the other eight cancers. Therefore, high expression of ITGBL1 predicts a higher number of tumour mutations in PRAD, THYM and LAML, which is more beneficial for improving the efficacy of tumour immunotherapy. High expression of ITGBL1 indicated a high incidence of gene mismatches in PRAD and SARC, whereas low expression of ITGBL1 indicated a high incidence of gene mismatches in STAD, LUSC, HNSC, and BRCA (Figure 6F). Overall, ITGBL1 can play an effective guiding role in tumour immunotherapy.

The drug sensitivity of ITGBL1

CellMiner™ is an online tool designed to mine publicly accessible datasets of human cancer cell lines related to genetics, pharmacology, and genomics, and we used this database to explore the relationship between ITGBL1 and drug sensitivity [27]. The analysis results showed that ITGBL1 was positively correlated with SGX-523, JNJ-3887618, AZD-1390, JNJ-38877605, dimethylfasudil, PF-04217903, olaparib, P-529, IDN-C227, and staurosporine. Conversely, ITGBL1 increased the sensitivity to APR-246, HYPOTHEMYCIN, bosutinib, BP-1-102, BGB-283, and LY-2606368 (Figure 7A–P). The results suggest that ITGBL1 can influence the drug sensitivity of these inhibitors, which will also provide more research directions for ITGBL1 in tumour mechanisms and improving prognosis.

Figure 7:

Analysis of drug sensitivity towards ITGBL1 according to the CellMiner™ database.

Elevated ITGBL1 induces cancer cell multiplication and migration

The database results indicated that ITGBL1 was highly expressed in colorectal and gastric cancers, which also predicted a poor prognosis, so we decided to validate the functional experiments in MGC830 and HCT8 cells. We transfected cancer cells with three different ITGBL1 siRNAs and validated them using Western blotting. The results showed that the knockdown effect of si-ITGBL1-1 was most significant in HCT8 cells, and the knockdown effect of si-ITGBL1-2 in MGC803 cells was the best (Figure 8A and B). Therefore, we decided to perform relevant experiments after the knockdown of HCT8 and MGC803 with these two siRNAs. The CCK8 assay can reflect the proliferation ability of cancer cells, which was used to demonstrate that the knockdown of ITGBL1 reduced the capacity of colorectal cancer cells and gastric cancer to proliferate over time (Figure 8C and D). Using scratch assays, we compared the migratory capacities of the cells in the negative control and knockdown groups to examine the impact of ITGBL1 on tumour cell migration. The results of our analysis showed that downregulating ITGBL1 can decrease the migratory capacity of cancer cells (Figure 8E and F). Consequently, we hypothesised that ITGBL1 is a tumour-causing oncogene that may contribute to the progression of cancer; therefore, targeting it with medication might enhance patient survival.

Figure 8:

Functional implications of ITGBL1 on proliferation and migration of cancer cells. (A, B) Western blot analysis of the transfection efficiency of three siRNAs after transfection of HCT8 and MGC803 cells. (C, D) Effect of ITGBL1 knockdown on the proliferation of colorectal cancer cells and gastric carcinoma cells. (E, F) Impact of ITGBL1 suppression on gastric and colorectal cancer cells’ ability to migrate. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.