May the assessment of baseline mucosal molecular pattern predict the development of gluten related disorders among microscopic enteritis?

AIM To evaluate mucosal baseline mRNA expression of tissue transglutaminase 2 (tTG2), interferon gamma (IFNγ), toll-like receptor 2 (TLR2) and Myeloid Differentiation factor 88 (MyD88) in patients with microscopic enteritis (ME). METHODS We retrospectively enrolled 89 patients with ME of different etiology, which was defined within a 2-year mean period of follow-up. Baseline histological examination was performed on Hematoxylin-Eosin stained sections and CD3 lymphocyte immunohistochemistry was used for intraepithelial lymphocyte count (IELs). ME was defined according to the criteria of Bucharest Consensus Conference. For each patient, formalin embedded biopsy samples of the duodenum referred to the period of ME diagnosis were retrieved. Real-time polymerase chain reaction (RT-PCR) was used to detect the amount of mRNA coding for tTG2, IFNγ, TLR2 and MyD88, and the quantity was expressed as fold change compared to controls. Control group was represented by duodenal normal specimens from 15 healthy subjects undergoing endoscopy for functional symptoms. Comparisons among continuous variables were performed by One way analysis of variance (ANOVA) and Bonferroni's test. The χ(2) test was used for categorical variables. Pearson's test was used to evaluate correlations. Receiver operating curves were drawn for all four markers to estimate sensitivity and specificity in discriminating the development of CD and GS. RESULTS After a period of follow up of 21.7 ± 11.7 mo, the following diagnoses were achieved: gluten related disorders in 48 subjects (31 CD; 17 GS) and non-gluten related ones in 41 (29 Irritable Bowel Syndrome - IBS; 12 Others). CD patients had the highest tTG2 levels (8.3 ± 4.5). The ANOVA plus Bonferroni analysis showed that CD > Other ME > GS = IBS > negative controls. A cut off value of 2.258 was able to discriminate between CD and GS with a sensitivity of 52.94% and a specificity of 87.1%. Additionally, CD patients had the highest IFNγ levels (8.5 ± 4.1). ANOVA plus Bonferroni demonstrated CD > Other ME > GS = IBS > negative controls. A cut off of 1.853 was able to differentiate CD and GS with a sensitivity of 47.06% and a specificity of 96.77%. Patients with non gluten-related causes of ME exhibited the highest TLR2 levels (6.1 ± 1.9) as follows: Other ME > CD = GS = IBS > negative controls. TLR2 was unable to discriminate CD from GS. Patients with CD overexpressed MyD88 levels similarly to non gluten-related causes of DL (7.8 ± 4.9 and 6.7 ± 2.9), thus CD = Other ME > GS = IBS > negative controls. A cut off of 3.722 was able to differentiate CD from GS with a sensitivity of 52.94% and a specificity of 74.19%. IELs count (15-25 and more than 25/100 enterocytes) strongly correlated with mRNA levels of all tested molecules (P < 0.0001). CONCLUSION Our results confirm that a single marker is unable to predict a discrimination among ME underlying conditions as well as between CD and GS. Mucosal high levels of tTG and IFNγ mRNA may predict the development of CD more than GS with high specificity, despite an expected low sensitivity. TLR2 does not discriminate the development of CD from GS. MyD88 levels indicate that intestinal permeability is more increased when a severe intestinal damage underlies ME in both gluten related and unrelated conditions. Therefore, the results of the present paper do not seem to show a clear translational value.