Targeted Drug Delivery in Inflammatory Bowel Disease Using a Novel Phospholipid-Based Prodrug Approach: Experimental and Computational Studies

Purpose: Current therapies of inflammatory bowel diseases (IBD) target a particular segment in the gastrointestinal tract regardless of where the inflammation is actually localized. We propose a novel prodrug approach that may address this issue by using phospholipase A2 (PLA2), an enzyme that hydrolyses the sn-2 position of phospholipids (PL) and is overexpressed in the inflamed tissues of IBD patients. This prodrug contains three main components: the PL, the drug moiety in the sn-2 position of the PL, and the carbonic linker connecting the two. Methods: PL-diclofenac prodrugs were carefully designed, and conjugates with 2, 4, 6 and 8-CH2 units were synthesized, purified and characterized by LC-MS and NMR. In vitro PLA2-mediated activation assay was performed with bee venom PLA2 for all 4 conjugates. Stability was tested on PL-diclofenac conjugate with 6 carbon linker in 3 different pHs, intestinal lumenal perfusate and plasma. Thermodynamic Integration and Umbrella Sampling/WHAM calculations of the PLA2-mediated activation of the different prodrugs were based on free energy changes in the rates of activation. The simulations also included comparison of bee venom and human PLA2. Results: The conjugates with linker length up to 6-carbon atoms, showed a linear correlation between the linker length and the degree and rate of PLA2-mediated hydrolysis. The most efficiently activated PL-diclofenac conjugate was the 6-carbon atom linker prodrug, while longer linker (8-carbon atoms) failed to further improve the prodrug activation, and was actually worse than the 6-carbon linker. The 6-carbon linker conjugate was found to be stable in intestinal perfusate, fresh plasma, and pH 4.0 and 6.8 buffers, but not at pH 1.0. Computational calculations were in excellent agreement with these in-vitro results. Moreover, bee venom and human PLA2 were found to exhibit excellent correlation between them. Conclusion: Based on experimental and computational studies, the optimal linker length for PLA2-mediated activation of PL-diclofenac prodrugs is 6-carbon atoms. Experimental results were in excellent correlation with our modern computational simulations; this combined in-vitro/in-silico approach has the potential to guide the design of the optimal PL-drug conjugate, and consequently to decrease the number of experimentations required. This work may become the basis of a novel drug targeting approach in the treatment of IBD.