Sirt2 is a nicotinamide adenine dinucleotide (NAD)-dependent protein lysine deacylase that can remove both acetyl group and long-chain fatty acyl groups from lysine residues of many proteins. It was reported to affect inflammatory bowel disease (IBD) symptoms in a mouse model. However, conflicting roles were reported, with genetic knockout aggravating while pharmacological inhibition alleviating IBD symptoms. These seemingly conflicting reports cause confusion and deter further efforts in developing Sirt2 inhibitors as a potential treatment strategy for IBD. We investigated these conflicting reports and elucidated the role of Sirt2 in the mouse model of IBD. We essentially replicated these conflicting results and confirmed that Sirt2 inhibitors' protective effect is not through off-targets as two very different Sirt2 inhibitors (TM and AGK2) showed similar protection in the IBD mouse model. We believe that the differential effects of inhibitors and knockout are due to the fact that the Sirt2 inhibitors only inhibit some but not all the activities of Sirt2. This hypothesis is confirmed by the observation that a PROTAC degrader of Sirt2 did not protect mice in the IBD model, similar to knockout. Our study provides an interesting example where genetic knockout and pharmacological inhibition do not align and emphasizes the importance of developing substrate-dependent inhibitors. Importantly, we showed that the effect of Sirt2 inhibition in IBD is through regulating the gut epithelium barrier by inhibiting Arf6-mediated endocytosis of E-cadherin, a protein important for the intestinal epithelial integrity. This mechanistic understanding further supports Sirt2 as a promising therapeutic target for treating IBD.