Expression Profile of Serum CircFUNDC1 and CircUHRF1 Can Differentiate Between Colorectal Cancer and Inflammatory Bowel Diseases (Ulcerative Colitis and Crohn's Disease)

Abstract

Background: Colorectal cancer (CRC) is a worldwide burden. Circular RNAs are promising biomarkers for diagnosing and prognosis of CRC.

Objective: To investigate the possible association of sera levels of CircFUNDC1 and CircUHRF1 expression with predisposition and clinicopathological findings in CRC, ulcerative colitis (UC), and Crohn's disease (CD) in Egyptian patients.

Methods: The serum levels of CircFUNDC1 and CircUHRF1 were evaluated in 113 Egyptian subjects divided into four groups; CRC (31), UC (26), and CD (25) and compared to healthy controls (31) using quantitative polymerase chain reaction.

Results: The median values of log2 serum fold change (FC) of CircFUNDC1 in CRC, UC, and CD patients were 9.11, 6.58, and 6.17, respectively. It was upregulated in all case groups. CRC, UC, and CD patients had significantly higher serum CircFUNDC1 levels than controls (p < 0.001). However, there were no significant differences among patient groups (CRC, UC, and CD). The medians of log 2 of serum FC CircUHRF1 in patients with CRC, UC, and CD were -2.00, 3.33, and 3.12, respectively. The CircUHRF1 serum level was lower in the CRC group of patients, with no significant difference between the CRC group and the controls. Serum CircUHRF1 was significantly overexpressed in patients with UC and CD compared to the CRC groups or controls (p < 0.001). By Roc curve analysis, both genes can differentiate CRC patients from inflammatory bowel disease (IBD) patients or healthy controls with p < 0.05.

Conclusion: Serum CircFUNDC1 is a biomarker for CRC, while CircUHRF1 is a biomarker of IBD.

Keywords: CircFUNDC1; CircUHRF1; Colorectal cancer; Crohn's disease; Ulcerative colitis.

FIGURE 1

(A) Post hoc pairwise comparisons between the four studied groups regarding log2 of serum fold change of CircFUNDC1. There were significant differences between CRC, UC, or CD and controls. p values (Controls vs. UC = < 0.001, Controls vs. CD = < 0.001, Controls vs. CRC = < 0.001, UC vs. CD = 1.000, UC vs. CRC = 0.194, CD vs. CRC = 0.255). Log2 of the fold change of target gene values was obtained to represent the large values. Controls were considered 1 by the equation 2^−ΔΔCt, and log2(1) equals 0. (B) Post hoc pairwise comparisons between the four studied groups regarding log2 of serum fold change of CircUHRF1. There were significant differences between UC or CD and controls. Also, there was a significant difference between CRC and UC or CD. There were no differences between CRC and controls or between UC and CD. p values (Controls vs. CRC = 1.000, Controls vs. UC < 0.001, Controls vs. CD < 0.001, CRC vs. UC < 0.001, CRC vs. CD < 0.001, UC vs. CD = 1.000). Log2 of the fold change of target gene values was obtained to represent the large values. Controls were considered 1 by the equation 2^−ΔΔCt, and log2 of (1) equals 0.

FIGURE 2

(A) Correlation between CircUHRF1 and ALT in CRC. (B) Correlation between two genes in CRC. (C) Correlation between two genes in UC. (D) Correlation between two genes in CD. (E) Correlation between CircUHRF1 and duration of illness in CD. (F) Correlation between CircFUNDC1 and HB in CD. (G) Correlation between CircFUNDC1 and HTC in CD.

FIGURE 3

ROC curve analysis to show how well serum CircFUNDC1 and CircUHRF1 work as diagnostic tools across the various study groups. (A) Serum CircFUNDC1 was shown to differ CRC from healthy controls with a best cut‐off value of 7.88, an AUC (95% CI) of 0.941 (0.877–1.00), p < 0.001, a sensitivity of 98.58, specificity of 91.65, and overall precision of 95.11%. Also, serum CircUHRF1 could differ CRC from healthy controls with a best cut‐off value of 2.01, an AUC (95% CI) of 0.863 (0.768–0.957), p < 0.001, sensitivity of 90.29, specificity of 88.74, and overall accuracy of 89.51%. (B) Serum CircFUNDC1 was shown to differ UC from healthy controls with a best cut‐off value of 4.08, an AUC (95% CI) of 0.863 (0.868–957), p < 0.001, a sensitivity of 92%, specificity of 90%, and overall precision of 90%. Also, serum CircUHRF1 could differ UC from healthy controls with a best cut‐off value of 2.14, an AUC (95% CI) of 0.824 (0.719–0.928), p < 0.001, sensitivity of 90%, specificity of 85.6%, and overall accuracy of 88%. (C) Serum CircFUNDC1 was shown to differ CD from healthy controls with a best cut‐off value of 3.11, an AUC (95% CI) of 0.902 (0.820–984), p < 0.001, a sensitivity of 94%, specificity of 89%, and overall precision of 92%. Also, serum CircUHRF1 could differ UC from healthy controls with a best cut‐off value of 1.95, an AUC (95% CI) of 0.843 (0.743–0.943), p < 0.001, sensitivity of 91%, specificity of 80%, and overall accuracy of 86%. (D, E) Serum CircFUNDC1 was shown to differ CRC from UC OR CD with the best cut‐off value of 3.09 and 2.87, respectively, with a total accuracy of around 93%. Serum CircUHRF1 was shown to differ CRC from UC OR CD with the best cut‐off value of −2.13 and −1.98, respectively, with a total accuracy of around 90%. (F) Serum CircFUNDC1 and CircUHRF1 could not differentiate between UC and CD.