Identification and Validation of Biomarkers to Predict Early Diagnosis of Inflammatory Bowel Disease and Its Progression to Colorectal Cancer

Abstract

Inflammatory bowel disease (IBD) has become a common global health problem as prevalence continues to rise. It is often associated with increased risk of colorectal cancer (CRC) development. Limitations in current IBD biomarker-based diagnosis hinder the accuracy of early detection of CRC progression. Therefore, in this study, we proposed the use of transcription factor (TF)-based biomarkers that can potentially detect the transition of IBD to CRC. Various bioinformatic analysis and online database validations, and RT-qPCR validations were performed to identify possible diagnostic TFs. RUNX1 was identified as a promising TF that regulates 106 IBD/CRC-related genes. The incorporation of RUNX1 in combination with currently known IBD biomarkers, FEV + NFKB1 + RELA, achieved a comparable sensitivity and specificity scores of 99% and 87%, respectively, while RUNX1 in combination with known CRC markers, CEA + TIMP1 + CA724 + CA199, achieved a sensitivity and specificity score of 97% and 99%, respectively. Furthermore, a small pilot RT-qPCR-based analysis confirmed a demarcated shift in expression profiles in CA724, CEA, RUNX1 and TIMP1 in IBD patients compared to CRC patients' tissue samples. Specifically, CA724 is noticeably elevated in IBD, while the levels of CEA, RUNX1 with TIMP1 are probable genes that may be employed in discerning IBD progression to CRC. Therefore, these preliminary results once validated in large patient cohorts could potentially have a significant impact on CRC disease stratification, resulting in a more precise prediction for treatment and treatment outcomes, especially in South African patients.

Keywords: Colorectal cancer; Diagnostic biomarkers; In silico and wet-lab validation; Inflammatory bowel disease; Predictive markers; Transcription factors.

Fig. 1

GO annotations and KEGG enrichment Pathway analysis of 289 IBD-related genes Biological processes,Cellular components,Molecular functions and KEGG Pathways found to be enriched. The size of the black circles represents the relative number of genes showing enrichment of a process/pathway. Red/blue bars represent the statistical significance at p < 0.01 for KEGG pathways, whereas p < 005 represents the significance cut-off for GO annotations

Fig. 2

Kaplan–Meier Plots depicting OS and DSF for the 5 biomarkers gene signature set and 4 biomarkers gene signature set The combination of 5 biomarkers (CEA + TIMP1 + CA724 + RUNX1 + CA199) expression was associated with a worse OS, hazard ratio (HR) = 1.3, log rank P = 0.2 and b worse DFS, HR = 1.6, log rank P = 0.028. The combination of 4 biomarkers gene set (CEA + TIMP1 + CA724 + RUNX1) expression was associated with c worse OS, hazard ratio (HR) = 1.8, log rank P = 0.2 and d worse DFS, HR = 2.1, log rank P = 0.014

Fig. 3

In vitro validation of in silico markers identified using South African Patient samples Relative gene expression profile of CA724, CEA, RUNX1 and TIMP1 in IBD (n = 3) and CRC patients (n = 3) obtained by RT-qPCR analysis. Endogenous expression of these genes was employed to normalise the data to the healthy patient group (n = 3). Data represent the mean ± standard deviation (S.D) (n = 6). The Kruskal–Wallis test was employed for statistical analysis to assess significance across all three patient groups (healthy, IBD and CRC) which proved a significant difference does exist between these groups, where p < 0.05 (*) and ns is non-significant as represented above

Fig. 4

Validation of biomarkers in IBD (UC and CD) samples using IBD TaMMa database and in CRC samples using Kaplan–Meier Plotter. Human IBD data was queried in the IBD TaMMa database for disease groups (CD and UC) vs. control normal samples, and a Log₂ expression values were plotted for genes RUNX1 and TIMP1 using FDR<1^-10 as a default cut-off. b represents the expression profiles of RUNX1 and TIMP1 in normal and CRC (tumour and metastatic) samples, whereas c showcases the signature expression analysis (RUNX1+TIMP1) which calculates the means of the selected gene signature across each patient sample one by one in the database, d shows the comparison of Log₂FC in IBD and CRC samples using two databases. FDR, false discovery rate, CD Crohn’s disease, UC ulcerative colitis, IBD inflammatory bowel disease, CRC colorectal cancer, FC fold change