One step closer to targeting RAS

The RAS family of small GTPases play critical roles in many types of human cancer. Activating RAS mutations are the most frequent type of oncogene mutations in human cancers, and are especially common in pancreatic, lung, and colorectal cancer. However, failure to obtain clinically useful inhibitors for RAS or any other GTPases suggests this target family is a therapeutic challenge. Consequently, significant efforts have been shifted toward targeting of downstream effectors of RAS including the Raf-MEK-ERK kinase pathway and the PI3K-AKT-mTOR kinase pathway. A third arm of RAS effector signaling, RAL (Ras-like) has not been targeted until recently and like RAS, is activated, but unlike RAS is not commonly mutated in cancer. RAL also shares a high structure similarity with RAS and other GTPase of the RAS superfamily. The two isoforms RALA and RALB are both important drivers of the proliferation, survival and metastasis of multiple human cancers. In our recent publication we reported the discovery of small molecule inhibitors of RAL.¹

GTPases including RAS and RAL exists in two interchangeable forms in cells, the GDP-bound inactive state and the GTP-bound active state. However, attempts to target the nucleotide binding site of small GTPases have been unsuccessful because of their high affinity for the guanine nucleotides GDP/GTP and the millimolar concentration of these nucleotides in cells. Furthermore, targeting this site would raise issues of specificity and subsequent toxicity in view of its high conservation across GTPases. One of the lessons learned from the history of kinase inhibitors is that some of the most effective kinase inhibitors proved to not be competitive with ATP, but were allosteric inhibitors that modify the conformation changes of kinases such as MEK.² Applying this idea to RAL, we analyzed the protein structure of RAL using computational modeling and identified a potential binding pocket on the surface of inactive RAL, which is absent from the active form. Importantly, this site does not overlap the nucleotide binding site and accommodates the independent binding of small molecules. We hypothesized that compounds that bind to this site would function as non-competitive inhibitors that can prevent the conformation change induced by nucleotide exchange and subsequent effector binding. We then performed computational screening of a 500,000 compound virtual library and identified 88 hits which were further characterized in cellular and biochemical assays, and then tested in an animal model. Eventually we identified RBC8 as a lead compound and synthesized a more potent analog BQU57 which was shown to be potent inhibitors of RAL activity and tumor growth both in vitro and in vivo.

The compounds we developed are the first of their kind given the concept of targeting the inactive form of a GTPase. Our approach also has the potential to be applied to all other small GTPases of the RAS family as long as these are not mutated in cancer. However, there are 2 major challenges associated with this structure-based approach. First, success of structure-based drug design relies heavily on the quality and accuracy of the structure information we have. The fine structures of many proteins have not been solved yet; not to mention that the protein structures we currently know either represents one of the possible conformations of the protein (X-RAY) or the average of all the possible confirmations (NMR). Secondly, not all GTPases will have a druggable allosteric site. As noted above, this is likely the case for GTPases that bear activating mutations. RALA or RALB mutations are rare (<1%) in human cancer or cancer cell lines (https://tcga-ata.nci.nih.gov/tcga and http://cancer.sanger.ac.uk/cosmic) making the targeting of RAL using our approach viable. An added complexity for RAS is that each different mutation type (G12C, G12D, G12V, or G13D) could have a different impact on the protein structure. Searching for targetable binding pockets on the active protein surface will be an option in these cases. Lodging small molecules in such pockets could inhibit the protein-protein interactions between RAS and its activators or effectors thus inhibiting RAS signaling. New crystal structures of RAS bound with different effectors and activators will help identify such novel binding interfaces that can be disrupted with small molecules. Recently, 3 studies have identified such binding pockets on the surface of RAS that are targetable with small molecules to prevent either activator or effector binding. Sun et al.³ and Maurer et al.⁴ both discovered compounds that bind to pockets on the KRAS surface and prevent SOS-mediated activation; while Shima et al.⁵ identified small molecules that bind to the surface of HRAS and inhibit its binding to the effector C-RAF.

In conclusion, although the RBC8/BQU57 compounds require further optimization for eventual human use, they offer a first “proof of principle” that targeting RAL GTPases is feasible and has effects in vivo. Our success in targeting RAL also demonstrated that structure-based screening and design of small molecule inhibitors for RAS is promising irrespective of whether they are directed at the inactive or active form of the molecule. Together, these studies are getting us closer to targeting RAS in human cancer.