A Resurgence in Natural Product-Based Drug Discovery

Natural products and their substructures have long been valuable starting points for medicinal chemistry and drug discovery. Since the earliest days of medicine, we’ve turned to nature for our treatments. From the application of digitalis extract as a remedy for heart failure, to the use of Vitamin C to prevent scurvy, many of the first drug treatments were developed by studying the medicinal effects of plants, and isolating the specific compounds responsible for their therapeutic properties.    As the knowledge of medicinal chemistry and chemical synthesis advances, the pharmaceutical industry has become more adept at creating synthetic analogues of natural products to reduce the reliance on natural sources, or to improve drug properties such as therapeutic potency, bioavailability, or metabolism by carefully modifying a molecule’s structure (1). Indeed, it’s thought that approximately 40% of drugs available on the market today are derived from chemical structures found in nature (2, 3).   Yet over the past few decades, the influence of natural products on drug discovery has notably reduced, in part due to the perceived difficulties of isolating and synthesizing these complex molecules, as well as the challenges associated with screening them using high throughput assays, which are commonly used to identify potential lead compounds. The industry, however, could be at a turning point. In recent years, there has been a resurgence of interest in the inclusion of natural products and their substructures in compound screening collections. Here, the author considers how advances in technology and adoption of alternative screening strategies are playing a role in revitalizing natural product-based drug discovery.

The high-throughput screening revolution 

The arrival of combinatorial chemistry and high-throughput screening (HTS) three decades ago dramatically changed drug discovery. Through the use of simple chemical or cellular assays to quickly identify whether small molecules bind to a specific biological target, HTS approaches now allow potentially millions of compounds to be rapidly screened in the search for promising molecular structures that can be optimized into potent therapeutics. Coupled with strategies for the design of “drug-like” compound libraries and prioritization of hits-such as Lipinski’s “rule of five” amongst others-HTS has led to significant progress in the discovery of lead compounds and the reduction of attrition rates in pharmaceutical development.   Nonetheless, efforts to streamline drug discovery have also been confined to a relatively small area of chemical space. Combinatorial chemistry-based strategies have narrowed the search to small, flat, synthetically tractable compounds, with relatively few functional groups. With their large size and enhanced potential for hydrogen bond interaction, natural products can be the first to be deprioritized or removed from screening decks if a rigid interpretation of Lipinski’s rules is applied.    As a result, natural products populate areas of chemical space that have largely been ignored by modern drug discovery. In fact, it’s thought that only 20% of the ring systems found in natural products are present in drugs available on the market (4). But with their rich stereochemistry and wealth of pharmacophores, these compounds have the potential to be equally, if not more, effective than the structures prioritized by pharma in recent years. After all, natural products are already predisposed to interacting with biological systems, and the industry could be missing out on a wide range of therapeutically interesting structures.

A natural product renaissance

After decades of decline, natural products are now experiencing something of a renaissance, thanks to increased interest in their use for more challenging targets, such as those with antimicrobial potential or where protein–protein interaction mechanisms play a key role (5). On this first point, the wide variety of resistance mechanisms means that tweaking the structure of existing molecules is proving increasingly ineffective for the development of new antimicrobials. Looking to the compounds produced by plants, animals, and microorganisms in their own antimicrobial arms race could provide a better option for discovering the next generation of drugs. Additionally, the advent of genomic-based target discovery has revealed a wide range of targets, such as transcription factors and structural proteins, which do not fit the traditional “lock and key” model. One way to modulate these “undruggable” targets could be through the perturbation of protein–protein interactions, and it seems that larger molecules could stand the best chance of exerting an effect.   Increasingly, the industry is also realizing the potential gains that can be made by moving beyond oral administration-a drug delivery route that has traditionally focused drug design on structures favored by Lipinski’s rules. It is worth remembering that these rules don’t necessarily make a therapeutically effective drug molecule; they’re simply properties associated with a high probability of good bioavailability when absorbed through the gut. For localized delivery to the lungs, the factors prioritized by Lipinski’s rules can actually hinder a drug molecule’s therapeutic efficacy. By no longer restricting drug design to these narrow criteria, a much greater area of chemical space can be explored, including natural product frameworks. 

Fragment-based screening and technological advances

Advances in analytical technology and alternative assay strategies are also making natural product-based drug discovery more viable. Improvements in the analytical sensitivity and precision of assay detection techniques mean we can now probe more precisely how drug molecules interact with protein binding sites.   These advances have enabled fragment-based screening approaches to play an increasingly important role in hit discovery. Here, rather than searching for molecular matches using complete compounds, the interior structure of the binding pocket can be probed by introducing small chemical fragments based on specific functional groups. While individual fragments bind more weakly to the protein than larger molecules, these interactions can be studied in detail with highly sensitive techniques, such as nuclear magnetic resonance (NMR), surface plasmon resonance, and capillary electrophoresis. Supported by computational approaches, this structural information can then be pieced together to design a complete molecule.   Advances in computing power are also helping to make this process much faster and more effective. Today, protein binding sites can be routinely modeled to study how large compounds, such as natural products, may fit within this space. In this way, drug developers can better predict how functional groups in larger natural products might interact with binding pockets and how natural products might be structurally adapted to improve the fit within the binding site.   The reverse can also be done. Computational target prediction tools now allow the identification of potential targets and the prediction of off-target effects of lead compounds (6, 7). In combination with high-throughput phenotypic screening strategies, where a specific target may not yet be established, these methods can highlight potential avenues for further investigation.   Once suitable structures are identified, the latest biosynthetic strategies can make synthesizing these complex molecules much simpler. Recent advances in genomic sequencing have resulted in the identification of many metabolic pathways that can be exploited for the purposes of drug design. These approaches, based on modified enzymatic pathways within cells, can potentially be used to carry out complex chemical modifications to naturally-derived substrates that would otherwise involve resource-intensive synthetic chemistry. 

Conclusion

While nature has been an important source of medicines throughout human history, the value of natural products in drug discovery has been somewhat overlooked in recent times. There is, however, a resurgence of interest in their use, driven to a large extent by the recognition of their enormous potential in the search for new antimicrobials and their efficacy to challenging targets based on the disruption of protein–protein interactions. Alternative screening strategies, such as fragment-based and phenotypic approaches, as well as advances in assay detection technology, have the potential to open up unexplored areas of chemical space populated by these important structures in the search for new and effective treatments.  

References

1. Z. Guo, Acta Pharmaceutica Sinica B, 7 (2) 119-136 (2017).

2. E.J. Jacob, Current Science, 96 (6) 753-754 (2009).

3. D.J. Newman and G.M. Cragg, Journal of Natural Products, 79 (3) 629-661 (2016).

4. T. Rodrigues et al., Nature Chemistry, 8, 531-541 (2016).

5. A.L. Harvey, R. Edrada-Ebel, and R.J. Quinn, Nature Reviews Drug Discovery, 14 (2) 111-129 (2015).

6. T. Cheng et al., The AAPS Journal, 19 (5) 1264-1275 (2017).

7. A. Koutsoukas et al., Journal of Proteomics, 74 (12) 2554-2574 (2011).