We’re told phenotypic screening is the answer to the lack of drug discovery productivity, ‘Return to the old ways and old-style productivity will return’. Well, a paper has just been published in ACS Chemical Biology which may throw some interesting light on this approach.
Brian Shoichet et al have had a long standing interest in the physico-chemical properties of screening molecules. Their first paper in this series, published in Journal of Medicinal Chemistry in 2002, arrived just in time to salvage some of my tarnished reputation with the screening group at a previous company. We were working on inhibitors of a novel antibacterial target and for some time had been following a set of compounds with flat mM SAR. All active molecules were highly lipophilic. No matter which solubilising groups you added to improve physico-chemical properties, activity always disappeared. Only as we added successive fluorines and iodines did activity increase. The observation that lipophilic molecules (such as the ones we were making) tended to form aggregates in solution and sequester protein was a ‘eureka moment’ for us. It explained our SAR (or lack of it) and allowed both chemists and biochemists to climb down from the increasingly confrontation positions we’d been adopting at project meetings! Our follow-up antibacterial projects had a strong focus on physico-chemical properties and bacterial membrane penetration.
At the same company we also had a ‘skunkworks’ project, run in our spare time, in which we used an oncology cell line, developed by a friend during his PhD, to screen some focused sets of compounds. We were looking for phenotypic responses to compound. Once again, screening initially at 20mM single concentration, we discovered some mM hit compounds but none of the nM activity we’d been hoping for.
Perhaps throwing some light on our results, Shoichet et al have used transmission electron microscopy (TEM) to show that some oncology drugs have a propensity to form colloidal aggregates inside cells (not just in biochemical assays). The aggregates act as reservoirs of compound within cells effectively reducing the cellular concentration of the compound. Thus, in contrast to the effect observed for biochemical assays, this can be expected to lead to a raised false negative hit rate. Active molecules will look less active because their effective cellular concentration is lower. The obvious answer is to add detergent to the system but of course most detergents are toxic to cells. However, they discovered that 0.025% Tween-80 was non-toxic in their cell lines. In the presence of compound the detergent was sufficient to break up the colloidal aggregates almost completely. They comment that ‘the monomeric drug forms were substantially more toxic than the colloidal forms, which consistently showed no significant anti-proliferative effects.’ So, returning to our ‘skunkworks’ project, perhaps some of our mM hits would have been considerably more potent had we developed the assay with a small amount to detergent to break up aggregates…we’ll never know!
The authors then go on discuss the observation that Evan’s Blue, a dye frequently used to measure vasculature leakage in tumours, also readily forms aggregates. In the past the concentration of Evans Blue dye within tumours (“enhanced permeability and retention (EPR) effect”) had been ascribed to the binding of Evans Blue to albumin and then transport of albumin across leaky membranes. The authors suggest this is not the case but that Evans Blue colloidal aggregates may cross cellular membranes anyway, bound or unbound to albumin. They conclude with the intriguing suggestion that an ability to form aggregates may be required to enable the concentration of certain anti-cancer drugs in tumours.
‘Evans Blue colloids are 120 nm in radius and will likely, themselves, permeate tumor tissue via the EPR effect, absent any protein. Indeed, such behaviour might, under the right circumstances, be true of other molecules, including colloid forming drugs. This would have a profound effect on their distribution and efficacy in vivo and may merit further study.