Mathematical Models Including microRNA Levels of Mesenteric Adipose Tissue May Predict Postoperative Relapse in Crohn’s Disease Patients

Abstract

Recent evidence suggests that the mesenteric adipose tissue (MAT) near the affected intestine may play a role in Crohn's disease (CD) pathophysiology. Modulation of several transcripts has already been identified in the MAT of CD in the literature. Therefore, our aim was to validate the microRNA (miRNA) transcript levels and their target genes in the MAT of active CD patients and correlate them with clinical and epidemiological data. Samples from the MAT of surgical specimens from 25 active CD patients were obtained. The control group comprised fifteen patients who underwent surgery for other diseases, except inflammatory bowel diseases. Transcriptional levels of miRNA and their target genes were assessed by quantitative real-time polymerase chain reaction. The correlation between transcripts and clinical characteristics was obtained using multiple linear regression. The mathematical models (M) underwent a statistical filter to ensure robustness and reliability (P value < .05; adjusted R-squared (Rˆ2)> .99; correct predictions of more than 60%). miRNA-650 and miRNA-29c were upregulated in the MAT of CD compared to the control group (P< .0001 and P= .0032, respectively), besides presenting decreased levels of their target genes. Two were target genes of the miRNA-650: glutamine-fructose-6-phosphate transaminase 2 (P= .012) and aldehyde dehydrogenase 4 family (P= .0035); and 4 were targets of the miRNA-29c: cell death-inducing DFFA-like effector c (P= .001), E2F transcription factor-1 (P= .007), hypoxia-inducible factor 3 subunit alpha (P= .0029), and pyruvate dehydrogenase kinase 4 (P= .0054). We found 2 M with statistical strength and robustness. The performance test identified one model with 100% accuracy for predicting the month of recurrence and determining patients with less risk of early relapse after surgery. We demonstrate that miRNA-650 and miRNA-29c and some of their target genes, besides clinical and epidemiological variables, may be useful in a model to predict when disease relapse may occur in CD patients who underwent surgery. These findings constitute a potential tool to guide postoperative clinical management.