Rediscovering Covalent Drugs

Covalent drugs have been some of the most successful in the history of pharmaceuticals: few would disagree with the revolutions brought about through the discovery of aspirin or penicillin. However, the industry has shied away from the modality in recent decades – owing to concerns regarding irreversible off-site activity, safety and predictability. New developments in the field promise to bring covalent drugs back en vogue, aided by novel techniques and methods for discovering new candidates.

Covalent Bonds

What qualifies as a covalent drug is simple: a compound which forms covalent bonds with its target(s). Covalent bonds are at the core of chemistry: they are formed when an electron is shared between two atoms – and are some of the strongest chemical bonds in nature. Often, the reaction which forms the bond is irreversible, which can grant higher potency at lower doses to the pharmaceutical candidate.

But this also has a flip-side, which is one of the main reasons covalent drugs have been in the background of pharma. Should the compound have off-site activities beyond its intended target, side-effects caused through irreversible reactions will also be durable. Paracetamol’s ability to covalently bind to liver proteins, and hence its hepatotoxic properties, is emblematic of the anxiety over the potential for off-site covalent effects.  Covalent bonds require caution to get everything right – which is why, up until recently, most covalent drugs were discovered phenotypically rather than through target-based approaches. Their covalent mode of action was simply revealed afterwards. 

While designing traditional drugs with reversible interactions, weaker chemical forces such as van der Waals, hydrogen bonds, chemical polarity and allosteric interactions all contribute and can be exploited. However, non-covalent ligands such as this are limited in the number of interactions they can achieve while maintaining a low enough molecular mass to not jeopardize their pharmacological potency – a term known as ligand efficiency. Covalent drugs, because of their ability to form strong bonds, can exceed the efficiencies of such traditional drugs while still maintaining low masses. This presents the potential to fill unmet needs and intractable drug targets.

A Slow Resurgence

The design of targeted covalent inhibitors (TCI) – as most targeted covalent drugs are known due to their function – languished until the discovery of ibrutinib. Ibrutinib worked by irreversibly binding Bruton’s Tyrosine Kinase, inhibiting the B-cell receptor pathways that are often overactive in multiple cancers such as mantle cell lymphoma and chronic lymphocytic leukemia. The approval of ibrutinib sparked a revival of interest in developing novel TCIs – it was soon joined by covalent inhibitors of EGFR, another protein heavily implicated across multiple indications of cancer. Results have even been demonstrated in targeting KRAS mutants such as KRASG12C; KRAS has long been one of the most elusive targets in oncology. This resulted in one of the first approved drugs to target KRAS – sotorasib, although questions over toxicity remain – with in vivo rat models showing possible metabolites of the pharmaceutical as responsible for renal side effects.

The progress made by these drugs had shown that covalent drugs were a viable drug modality, even in target-based drug discovery approaches. But they remain limited to areas where the therapeutic potential outweighs the possibility of toxic off-site effects. Without a way to assess their full impact on the body and its inventory of proteins – the proteome, efforts to develop them are subject to chance. However, this is changing – advances in chemoproteomics and the advent of high-throughput proteomic screening aim to rapidly assess all potential protein targets of compound candidates and detect off-site activities early in development. 

Methodical Improvements

Chemoproteomics advances not only promise to improve safety profiles for TCIs, but they also hold potential in discovering new targets. This is particularly true of mass spectrometry methods, combined with fragment-based screening. Fragment-based screening takes advantage of low molecular weight, active compounds to identify proteins to which they bind without the potential for interference by irrelevant groups in the candidate compound. Successful hits can then be optimized for ideal ligand efficiencies. We talked to Dr. Gyorgy Keseru, who is intimately involved in fragment-based drug discovery for covalent drugs, to find out more on the subject.

Current limitations also include the kind of amino acid that covalent drugs can target. Most TCIs depend on binding to a cysteine amino acid – for which acrylamide groups are particularly useful. But finding ways to interact with other amino acids, such as aspartates, would open up a plethora of new, currently undruggable, targets. Some strides have been made in targeting weakly reactive amino acid residues such as tyrosine, lysine and serine – others such as aspartate, glutamate and histidine remain more challenging but have nevertheless seen some research work. 

Transformational Technologies 

We must also not underestimate the potential of predictive protein structure and binding models to further improve covalent drug discovery. After all, the biggest limitation in the development of covalent drugs remains the complete characterization of their effects on the proteome at costs that are not prohibitive. We have seen novel models employed in recent years, such as BIreactive, a machine-learning model, as well as databases such as CovalentInDB, which fully illustrate known covalent ligand-protein interactions. Other efforts include deep generative modeling for the design of protein kinase inhibitors – which are already a well-established class of covalent drugs, as illustrated by idrutinib.

We have even seen AI-assisted design of covalent inhibitors against SARS-CoV-2 – illustrating the versatility of the treatment modality. What these developments illustrate is that the increasing adoption of cutting-edge technologies can rapidly transform covalent drug discovery – and the standardization of current knowledge into shared databases can accelerate efforts worldwide. Covalent drugs faced a sluggish period where their potential for long-lasting, irreversible toxicity forced researchers to shy away from them – and for probably good reason. Our increased understanding of their functions has restored them to a mainstream place in the industry – and their potential to reach intractable targets while providing low-dose, potent solutions cannot be understated.

Nick Zoukas, Former Editor, PharmaFEATURES

Join Proventa International’s Medicinal Chemistry Strategy Meeting in London in 2022 to hear more on the developments in the covalent drug space from leading experts. Participate in closed roundtable format discussions on current topics from throughout the MedChem field, and forge new, productive collaborations.