Benefits of Strategic Small-Scale Targeted Screening

Over the past decade, we have seen an accelerated integration of high-throughput screening (HTS) technologies into academic settings. Arguably, the main impetus for this “academic screening evolution” in the United States was the NIH Molecular Libraries Screening Network, which formalized broad-based access to HTS technologies, HTS (and assay development) expertise as well as the entrée of pharmaceutical and biotechnology industry-trained scientists into academic institutions. ¹ Moreover, with HTS in the NIH spotlight, many academic institutions initiated new or augmented existing in-house screening programs that complemented their basic research endeavors. These HTS programs are now integral components of many academic translational research, team science, and “bench to bedside” efforts, and ensure the nimbleness of the discovery-based academic science community. Since the network's decommissioning in 2011, other organized efforts have arisen (e.g., Academic Drug Discovery Consortium—addconsortium.org; Virginia Drug Discovery Consortium) to provide a forum for investigators involved with academic screening to network, collaborate, and to promote academic drug discovery efforts. ² Thus, it is clear that HTS is here to stay and, as a result, there is growing universal familiarity with the HTS-centric vernacular. Consequently, terms such as Z-factor, signal window, hit, and cherry-picking are becoming regular parlance for academic researchers. ^3,4 There is also a growing awareness of PAINS compounds, dark matter, diversity sets versus targeted libraries, and the strategic benefits of drug repurposing. ^{4 –7} In addition, a growing appreciation of critical paths (AKA screening paradigms, screening trees, or Go/No-Go strategies) has spawned extensive efforts in the establishment of secondary confirmation assays for use in lead optimization leading to preclinical and clinical development. ⁸ Critically, this burgeoning fluency with HTS-associated concepts may provide the impetus for a paradigm shift in HTS-based drug discovery and HTS-driven screening strategies.

In general, screening strategies are contingent on five major components: validated screening assays, compound libraries, automation, data capture instrumentation, and bioinformatics. Of these five main HTS components, the assay–compound library dyad is most crucial for a productive screening strategy (Fig. 1). Although automation, bioinformatics, and data capture instrumentation are critical tools that regulate throughput and assay formats, the assay–compound pairing determines the quality of the hit chemotypes that are channeled into cycles of secondary confirmation assays and medicinal chemistry optimization. Unfortunately, too often screening philosophies hinge upon “go big or go home” rather than thoughtfully pairing an assay with a compound library. Although large-scale screening (e.g., >100K compounds) has its utility and benefits, this strategy is resource intensive with often suboptimal returns on investment. Movement toward more physiologically representative assay systems, however, intimates that a “paradigm shift” in screening strategies may be forthcoming.

OPEN IN VIEWER

Fig. 1.

Assay–compound library dyad is critical for screening success. (1) The compound library and the screening assay determine the context, chemistry, and ultimately quality of the screening hits. (2) Bioinformatics, data capture instrumentation, and automation enhance throughput, accuracy, reproducibility, and allow different assay readouts effectively providing a solid foundation for screening activities. (3) Screening serves to identify chemotypes of interest and fulfill hit criteria. (4, 5) Successive rounds of confirmation assays and medicinal chemistry (i.e., SAR) refine selectivity and potency with subsequent impact on experimental efficacy and novelty of chemical structure. (6) The result is a chemical probe and, preferably, one that displays high-quality characteristics such as potency, selectivity, and novel chemistry, among others. High-quality chemical probes help validate molecular targets and may serve as a parental chemotype for drug development. Shown is MSP1-IN1 a small molecule inhibitor of MSP1, a dual-specificity kinase involved in spindle assembly checkpoint and chromosomal stability. ²¹ SAR, structure activity relationships; PAINS, pan-assay interference compounds.

Initial HTS strategies clearly reflected available assay technologies and, hence, there was a significant emphasis on isolated recombinant proteins or two-dimensional cell-based viability assays. HTS is now poised to enter a new phase of innovation. Three-dimensional cell-based screening systems comprising spheroids or multiple cell types, including primary cells, with inclusion of extracellular matrix, are rapidly supplanting “flat biology” (i.e., two-dimensional cell culture). ^{9 –12} The rationale is that these cell-based model systems more accurately replicate what occurs in vivo and would enhance the translation of experimental discoveries. ^10,13 Likewise, in vitro protein–protein interaction and multiprotein in vitro biochemical assays are increasing in frequency and offer more physiological context and, to a certain extent, target specificity than traditional activity-based assays. ^14,15 Equally critical are whole organism-based screening models (e.g., Caenorhabditis elegans and Danio rerio), which are essentially self-contained, complete biological systems with their experimental significance centering upon the conservation of human disease genes and disease pathways. ¹⁶ Using these “high-physiocontextual” cell-based, whole organism, and in vitro biochemical assays in conjunction with large diversity compound sets to screen for primary actives (i.e., hits) is obviously an option. However, exploitation of focused (also referred to as targeted) libraries could be a more compelling alternative as they may better harmonize with screening assay endpoints/targets and may be more likely, in certain experimental scenarios, to yield high-confidence hits (Table 1).

Classic molecular target class-specific libraries (i.e., kinase, phosphatase, nuclear receptor, and G protein-coupled receptor) are perhaps the best recognized, but as our understanding of specific molecular targets or molecular target classes expands, our potential types of targeted libraries also increase (Table 1). Thus, there are commercially available libraries expressly directed toward identification of bromodomain, histone deacetylase, and PDZ domain small molecule inhibitors, for example. Most often, these newer library assemblies reflect topical research trends such as interest in metabolism, epigenetics, autophagy, or inflammation. Alternatively, unique targeted libraries can be created using computational algorithms de novo according to individual investigator criteria. ¹⁷ The latter strategy would alleviate concerns regarding the novelty of targeted chemotypes and should offer a better standing with respect to intellectual property. Targeted libraries may be especially attractive to smaller screening centers as they are usually more manageable in size, which conserves resources allowing for more comprehensive screening and subsequent characterization, and, perhaps more importantly, their selective acquisition or creation can complement specific institutional areas of research excellence. Collectively, the pairing of assays with high potential for physiological translation and select targeted libraries coupled with a de-emphasis of throughput (as a key screening metric for as a screening strategy) may be strategically more sound. Nonetheless, it is crucial that the reproducibility and rigor of HTS are maintained in the lower throughput scenario to ensure the identification high-quality lead chemotypes.

As with all things scientific, screening has been subject to transformation, hence, the development of HTS and the subsequent derivation of high-content screening (i.e., HCS). ¹⁸ Emphasis on the “high-physiocontextual” screening assay–targeted library dyad may portend the next logical HTS progression—a smaller scale inclusive screen that yields high-confidence hits. This screening strategy may help alleviate a key research bottleneck, which is the paucity of chemical probes (and ensuing drug leads) for the vast majority of molecular targets. ¹⁹ Critically, when this “high-physiocontextual” screening assay-targeted library pairing is implemented with chemical probe “fitness factors” in mind (i.e., potency, selectivity, context, and chemistry), there would be a pronounced potential for the discovery and development of high-quality chemical research tools. ²⁰ High-quality chemical probes, in turn, would have a profound “trickledown effect” through the attenuation of irreproducible and misleading data. This strategy would also allow for the concentration of “discovery-based” resources on more sophisticated smaller scale screening strategies and release investigators from the “go big or go home” screening mindset. Moreover, it would embolden investigators to design, optimize, as well as validate more physiologically relevant screening platforms and thoughtfully pair a screening assay with particular compound libraries rather than using large diversity sets by default. Undeniably, the “high-physiocontextual” screening assay strategy might require more in terms of effort and expense, but the returns on investment would potentially be greater. Meaning, the identification and development of a high-quality chemical probe with bona fide credibility as a drug lead.

The possible advantages of the “high-physiocontextual” screening assay–targeted library dyad are largely waiting for exploitation by the research community. Logic suggests that smaller scale academic screening centers can be early adopters of this avant garde screening strategy, capitalizing on their specific areas of research expertise, leading the way toward the discovery of high-quality chemical probes. Integration and implementation of this “high-physiocontextual” screening strategy may also render obsolete the concept of the “undruggability” of select molecular target classes. Thus, unleashing HTS to its next developmental level may radically change chemical probe and drug discovery.