Challenges and Limitations in Rational Drug Design – a Case Study on the cAMP-receptor Epac2

Let’s assume a cellular signalling cascade controlling a medical relevant process (as for example insulin secretion). Let’s further assume that the cascade contains one step depending on the interaction between a small molecule and a protein. The design of a drug mimicking that small molecules would be logical and consequent. Many signalling cascades are unrevealed these days but still the number of rational designed drugs is limited. With the cAMP-receptor Epac2 as an example I will present successful steps but also limitations of such a design process.

cAMP (cyclic adenosine-3’,5’monophosphat) is a second messenger produced by many cell types. cAMP has a regulatory function in a wide range of biological processes, including olfaction, pacemaker activity, regulation of gene expression, insulin secretion, and many others. Cellular cAMP levels are sensed by proteins containing cyclic-nucleotide binding (CNB) domains, such as protein kinase A (PKA) or the guanine nucleotide exchange factors (GEFs) Epac1 and Epac2. In pancreatic β-cells cAMP augments glucose induced insulin secretion and Epac2 is assumed to be involved in this process.

We have developed chemical modified cAMP analogues that act as Epac2-selective agonist both in vitro and in vivo. This process had required the biochemical characterisation of the activation process, the determination of x-ray structures of the Epac2 protein and the systematic testing of about 100 cAMP analogues. The analogues were design by a structure-guided approach. In this process we encounter many unexpected but intriguing features of the Epac2-ligand interaction. We obtained a cAMP analogue (S-220) that activates selectively Epac2. Furthermore, S-220 activated Epac2 with much higher potency than cAMP itself. The selectivity as well as the high potency is explained on the structural level. S-220 is able to potentiate insulin secretion in isolates islets.