The middle path
August 16, 2010 § Leave a comment
Drug discovery tends to happen via one of two main ways. Either you look for a phenotypic effect — as in, chewing that willow bark eased my headache, or this fungal extract stopped my rat liver membrane prep from making cholesterol — or you look for a specific effect on a purified target. Both have disadvantages [and advantages, of course]: an effect found in a phenotypic screen may be very hard to track down to a particular small molecule, and a specific mechanism, while a purified-target assay may give you many hits that don’t turn out to be useful in the complex environment of a whole cell, let alone the human body. Now Riki Eggert’s lab has shown that there is a middle path: targeting a whole pathway (Castoreno et al. 2010 Small molecules discovered in a pathway screen target the Rho pathway in cytokinesis. Nat Chem Biol. 6 457-63 PMID: 20436488).
Castoreno et al. were interested in finding inhibitors of the Rho pathway, a complex and many-branched pathway that is at the heart of several key cellular functions such as migration, adhesion, and cytokinesis, the process by which two daughter cells separate from each other once their DNA has doubled and divided. Rho itself is a GTPase that switches between two states, GTP-bound (active) and GDP-bound (inactive). Quite a few regulatory proteins are involved in this cycle, including the Guanine nucleotide Exchange Factors (GEFs), which encourage the exchange of bound GDP for GTP, and the GTPase Activating Proteins (GAPs), which encourage cleavage of the GTP to form GDP, inactivating Rho. Active Rho has several targets, and at least three of the pathways activated by Rho converge on cytokinesis.
Why not simply screen for small molecules that inhibit Rho itself? After all, a protein that binds to a small molecule like GTP clearly has a small-molecule binding pocket, a place where a small molecule can fit and displace water molecules, increasing the entropy of the overall system. So GTPases might be thought to be druggable, as the ATP-binding kinases are. Unfortunately, it has been hard to find inhibitors of small GTPases such as Rho and its cousin Ras, though many have tried, probably because their affinity for GTP is higher than that of a kinase for ATP. And it’s also hard to look for inhibitors of the downstream pathways, since many of the enzymes in those pathways are difficult to assay. Before this work there were only three classes of small molecules known to inhibit this pathway: Rho kinase inhibitors, used in the clinic to treat cardiovascular diseases; the statins, which indirectly inhibit Rho activity by preventing the synthesis of isoprenyls, which are needed to localize Rho to the membrane; and a recently reported inhibitor of formin, one of Rho’s downstream targets.
Castoreno et al. therefore took a different approach, inspired by classical genetic modifier screens. Modifier screens start with a mutant that has a defined, non-lethal phenotype, and use this background to screen for additional mutations that either exacerbate or relieve the phenotype. In the pathway screen performed here, the initial mutation is replaced with an RNAi treatment that knocks down part, but not all, of Rho’s activity. This treatment partly blocks the process of cytokinesis, so that some of the cells that try to divide can’t; cells that fail cytokinesis end up as an over-large cell with two nuclei [like a double-yolked egg]. The screen is then for small molecules, not mutants, that either increase the number of binucleate cells or reduce it. The key here is to find conditions that partially block the activity of Rho. If you block Rho activity too thoroughly, you won’t find enhancers because it will be hard to see the difference between almost completely blocked and completely blocked. On the other hand, blocking Rho activity is what makes the screen (mostly) pathway specific. Anything that perturbs the activity of the Rho pathway will be easier to see in the context of partial inhibition.
The authors ran the screen on 23,000 compounds, including some known bioactives and fungal extracts, using automated image analysis to count the number of binucleate cells after treatment, and came up with 9 candidate enhancers of Rho pathway activity in cytokinesis, which they call Rhodblocks. (They found suppressors too, but many of them were cell cycle inhibitors, stopping the cell before it gets to the division stage, and so they’re still sorting out which suppressors are interesting). The activity of the compounds was modest — they’re active at 10-100 µM — but convincing.
Next, they tried knocking down other parts of the Rho pathway, such as the GAP, the GEF, or the proteins responsible for the next steps in the downstream pathways, and tested how much the different Rhodblock compounds synergized with each RNAi treatment. Synergy was evaluated relative to a multiplicative model that assumed the effect of the RNAi was independent of the effect of the small molecule. The different Rhodblocks showed different synergy/antagonism patterns with various parts of the pathway, indicating that each compound could act by a distinct mechanism. Rhodblock 6 was shown to inhibit Rho kinase, one of the key downstream targets of activated Rho, in a pure-protein assay. The targets of the other compounds are yet to be identified.
Some might ask, why do I need a small molecule when I have RNAi? Well, apart from the potential for pointing the way towards drug discovery — which is a long road, but an important one — the big advantage of small molecules over RNAi is that they can be used acutely. With RNAi, you need to wait hours or days for the lack of mRNA to translate into a lack of protein, depending on how fast the turnover of your protein is. With small molecules, you start to see the effect just a few minutes after you add them, and what’s more you can wash them out again and see the process you just inhibited re-start. (The disadvantage, of course, is that you have to worry even more about specificity than you do with RNAi.) Castoreno et al. also make the good point that in a situation like cytokinesis, where scaffolding is important — at least two large multiprotein complexes need to be assembled to allow furrow ingression and the completion of cytokinesis — absence of a given protein may have a very different effect from reduction in that protein’s activity.
One reason I like this work is that by screening for inhibitors of all parts of the pathway at once, we may turn up unexpected inhibitors of proteins that were not thought to be “druggable”; yet, the approach is more targeted than an ordinary phenotypic screen, and so you have some idea of where to start when you go looking for the mechanism. I’ll be watching the next installments of the story with interest.
Castoreno AB, Smurnyy Y, Torres AD, Vokes MS, Jones TR, Carpenter AE, & Eggert US (2010). Small molecules discovered in a pathway screen target the Rho pathway in cytokinesis. Nature Chemical Biology, 6 (6), 457-63 PMID: 20436488