Introduction Colitis in homozygous mK8-/- mice as a result of a primary epithelial defect and heterogeneous missense mutations in keratin (K) 8 and K18 in IBD suggest a possible association between simple epithelial keratins and IBD. Our previous work using mass spectrometry (MS) and western immunoblotting suggests alterations in the insoluble forms of K8, K18 and K19 in intestinal epithelial cells in colitis phenotypes compared to controls. We have previously shown an increase in insoluble keratin in epithelium from patients with longstanding pancolitis (LSPC) and decreased levels in epithelium from patients with dysplasia (in both the lesion and rectal mucosa), recent onset colitis (ROUC) and PSC associated colitis. There is also a reduction in insoluble keratins in active areas of inflammation compared to inactive areas.
Methods Paired biopsies were taken from patients with active colitis (one from the actively inflamed area and another from a proximal inactive area, n = 10) and patients with dysplasia (biopsies taken from the dysplastic area and an area of rectal mucosa, n = 3). Single rectal biopsies were taken from patients with ROUC (n = 8), LSPC (n = 10) and PSC colitis (n = 7). These tissue sections underwent IHC staining for K8, K18 and K19. Semi-quantitative assessment of keratin expression was performed using a pixel counting algorithm and a manual scoring system, which scored staining intensity at the surface and base of the crypts and extent of crypt staining.
Results There was no difference in K8 expression measured by manual scoring between the groups. K18 showed a significant difference in pixel count scores between the active and inactive inflammation (0.078 p = 0.022) reflecting a shift towards more extensive, moderate and weak staining in the active group. There was also a significant change in the manual scores for the ratio of surface intensity to crypt intensity (p = 0.028) representing a greater surface compared to crypt K18 expression. All other groups were similar in K18 expression and pattern. There was no difference in K19 expression measured by manual and pixel count scoring between the groups.
Conclusion The results suggest a difference between MS and IHC approaches. If the IHC data hold true they may suggest that the changes in keratin expression profile that occur with duration of disease, inflammation and dysplasia may be a result of changes in keratin solubility rather than a change in the total expression of keratins. However, possible explanations include too small sample size for IHC and lack of sensitivity of IHC by comparison with MS.
Disclosure of Interest None Declared