Distinct blood protein profiles associated with ileal and colonic ulcers in Crohn’s disease

Background
Ileal and colonic Crohn’s disease (CD) are nowadays considered as separate entities. Studies are needed to better characterise the biological specificities of these subphenotypes. In fine, this research should offer opportunities for the development of personalised medicine.
Methods
By combining different technologies (proximity extension assay, selected reaction monitoring and high-sensitivity turbidimetric immunoassay (hsCRP)), 207 serum proteins were measured in CD patients presenting no endoscopic lesions (endoscopic remission) (n=23), isolated ileal ulcers (n=17) or isolated colonic ulcers (n=16) (Table 1). To compare the endoscopic activity between the ileum and colon, and in order to obtain an assessment of the overall lesion burden, we used the Crohn’s disease endoscopic index (CDEIS) without dividing its total score (sum of each gut segment) by the number of inspected segments. The data analysis was conducted using Wilcoxon-Mann-Whitney test and area under the receiver operating characteristics curve (AUROC). The AUROC were compared using the bootstrap test (2000 replications).
Results
The median of the total CDEIS was higher in CD patients with isolated colonic ulcers than those with isolated ileal ulcers (23.1 vs 8.0) (Table 1). When compared to endoscopic remission, the presence of isolated ileal ulcers and isolated colonic ulcers were specifically associated with the level of 6 and 18 serum proteins, respectively: (high: JUN, CNTNAP2; low: FCRL6, LTA, CLEC4A, NTF4) (Figure 1A); (high: hsCRP, IL6, APCS, CFB, MBL2, IL7, IL17A, CCL19, CXCL10, CSF3, IL10, CLEC4G, MMP12, VEGFA; low: CLEC3B, GSN, TNFSF12, TPSAB1) (Figure 1B). No protein was associated with both isolated ileal ulcers and isolated colonic ulcers (Figure 1). All CD patients in endoscopic remission showed a normal level of hsCRP (<5 mg.L-1). Compared to CD patients in endoscopic remission, CD patients with isolated colonic ulcers showed an increased median level of hsCRP (1.3 vs 5.9 mg.L-1, p-value=0.0045) and this was not the case for CD patients presenting isolated ileal ulcers (1.3 vs 2.2 mg.L-1, p-value=0.13). hsCRP detected ileal and colonic ulcers with an AUROC of 0.64 (p-value=0.07) and 0.77 (p-value=0.001), respectively. When compared to hsCRP (AUROC), the best marker pairs for detecting ileal or colonic ulcers were respectively CLEC4A×LTA (0.82 vs 0.64, p-value=0.06) or hsCRP×CSF3 (0.85 vs 0.77, p-value =0.02).
Conclusion
In CD patients, ileal and colonic ulcers were associated with distinct systemic responses. hsCRP showed a better capacity to detect patients with colonic than ileal ulcers. Some markers might help to detect ileal ulcers or to improve the ability of CRP for detecting colonic ulcers.