Design and R&D of covalent drugs

  • For a long time, electrophilic groups have been a minefield in drug development. Therefore, in the classic medicinal chemistry textbooks of the past, it is always recommended to avoid introducing functional groups such as epoxide, acridine and Michael receptor into the structure of drug molecules, because these functional groups are highly reactive and may interact with a wide range of biological macromolecules and cause serious toxic side effects.


    Aspirin, Lansoprazole and Clopidogrel have all been found in recent studies to act as covalent bonds with their targets. Inspired by this, the research and development of covalent drugs are slowly entered people’s field of vision. It has shown advantages that non-covalent combination drugs are difficult to achieve, such as longer-lasting efficacy, lower therapeutic doses and less resistance to drugs.


    Due to the different mechanisms of covalent and non-covalent binding, it is difficult to truly reflect their efficacy and safety by using traditional evaluation indicators such as dissociation constants, IC50, EC50 during the research and development process. The formation of covalent bonds between drug molecules and their targets is influenced by the reaction rate, whereas non-covalent is only a thermodynamic equilibrium process that shows the activity of a compound within a short period of time. Therefore, the use of traditional evaluation indicators often leads to misjudgments.


    Covalently drugs and natural products

    Discovering new drugs from natural products is a common strategy in new drug research and development projects. Many drugs are discovered based on this approach, such as paclitaxel, xylophone, morphine, etc. Another approach is to structurally modify natural products to obtain compounds that are more effective, such as salicylic acid to aspirin, morphine to methadone. According to statistics, about 60% of the drugs in clinical practice are obtained based on the above strategy.


    Although covalently drugs have only become popular in recent years, the concept is not uncommon in nature. Many antibiotics, such as penicillin, showdomycin and fosfomycin, interact with their targets in bacteria in the form of covalent bonds. Lipstatin is an irreversible inhibitor of pancreatolipase isolated from streptomyces.


    The above examples of natural products prove that covalent binding interactions are an overlooked treasure trove in drug design. Most of the drug targets are proteins. Residues such as serine, lysine, cysteine and histidine contain nucleophilic active functional groups (hydroxyl, sulfhydryl, amino, etc.), so the protein can act as an excellent nucleophile and can interact with electrophilic active groups to form covalent bonds. The greatest difficulty in the design of such compounds lies in the selectivity, which is prone to serious side reactions, resulting in failure of research and development.


    Currently, around 30% of drugs targeting enzymes are in the form of covalent binding, mainly because this design concept has only been accepted in recent years. Prior to this, active reactive groups were structures that were avoided wherever possible in drug design. Telaprevir, developed by Merck and approved for marketing by the FDA in 2011, achieves its antiviral effect by forming a hemiacetal with the catalytic serine residue (hydroxyl group) in the HCV protease and inhibiting its activity. Initially, the compound showed poor activity in the standard IC50 test and was nearly abandoned, however, in a specially designed activity test, it performed well. The development of Telaprevir is a profound example, which highlights that the development of covalent drugs requires a different set of evaluation methods that are different from traditional non-covalently bound drugs.


    Another classic example is the development of Afatinib. It is an irreversible inhibitor of the EGFR receptor (launched in 2013). The electrophilic active group acrylamide in its structure forms a covalent bond with a cysteine residue (sulfhydryl group) in the active site of the EGFR receptor, which overcomes the resistance problems of the first generation of EGFR (gefitinib, erlotinib, etc.) receptor inhibitors, and shows good activity against non-resistant EGFR receptors. These two successful cases demonstrate the potential of covalent drugs. On the other hand, their development process provides valuable lessons for the development of other covalent drugs.


    Drug-target binding process

    Similar to non-covalent drugs, the covalent drug first interacts with the target to form a drug-target conjugate. Since this process is thermodynamic, equilibrium can be reached very quickly and its affinity can be described by parameters such as dissociation constant Ki or IC50. The difference lies in the formation of covalent bonds in the second step, which is formed at a slower rate compared to the first step, and exists a reaction equilibrium constant Ki*. Therefore, the overall binding of the covalently bound drug requires consideration of two parameters: Ki and Ki*. When Krea is much larger than Krev-rea, the reaction equilibrium constant tends to infinity, and the binding between the drug and the target can be regarded as irreversible covalent binding. When the difference between Krea and Krev-rea is not very large, i.e. Ki* is within a reasonable value range, the binding between the drug and the target can be regarded as reversible covalent binding.


    Binding Mechanisms

    Most of the drug targets are proteins, which can essentially act as nucleophilic reagents because their structures are rich in functional groups such as hydroxyl, sulfhydryl and amino groups. Covalent binding compounds usually contain electrophilic functional groups in their structures, such as Michael receptors, epoxy, halogen, carbonyl, isocyanine and other structures (Figure 5), which can act as electrophiles, and the two react with each other to form new covalent bonds. The main types of reactions involved are: acylation reactions, alkylation reactions, Michael addition, disulfide bonding, Pinner reactions, etc. The choice of covalent bond formation method depends on the nature of the target binding site.


    In 2014, one of the top 10 best-selling drugs in the US was esomeprazole (the racemic form is omeprazole). Omeprazole was invented in the 1970s and went on the market in 1988 after several optimizations and clinical studies. Two years later, it was discovered that its proton pump inhibition was based on a covalent bond form. Under acidic conditions, omeprazole is activated near the target site to form an active sulfenamide derivative, and then forms a covalent bond with a cysteine residue (sulfhydryl group) in the target site to inhibit gastric acid secretion.


    Another top-selling covalent drug is clopidogrel, which needs to be activated by P450 oxidase in the liver to produce a sulphur-containing compound that then forms a disulfide bond with a cysteine residue in the P2Y12 receptor to inhibit platelet coagulation. The above-mentioned two drugs are now very widely used and were not originally designed to act in a covalent bonding form. It was not until later in the mechanism research that they were found to produce irreversible inhibition by covalent bonding with the target.


    Compounds containing cyano groups usually undergo Pinner reactions with tnucleophilic functional groups in proteins such as hydroxyl and sulfhydryl groups to form generally reversible imine ester bonds. A successful example is the DPP4 inhibitors (for example saxagliptin and vildagliptin) for the treatment of diabetes mellitus type 2. In addition, there is Odancate, which is used for the treatment of osteoporosis in postmenopausal women. However, due to its potential risk of stroke, Merck announced in 2016 that it would abandon this compound.


    There are several other covalent binding mechanisms, such as alkylation, acylation and Michael addition. They are all essentially the interactions between nucleophilic groups in biomolecules and electrophilic groups in drug molecules.


    The control of selectivity

    The selectivity of covalently bound drugs is crucial. In terms of selectivity, it is now a common strategy to first design a highly selective non-covalent binding precursor to the target and then, on the basis of this compound, to optimize the design of the active functional group, i.e. to find a suitable "shell" and add a suitable "warhead", and then install the "warhead".


    A more typical example is the research and development of Osimertinib. The initially obtained lead compound 1 showed excellent binding in the in vitro target binding assays, but its affinity plummeted by nearly 70 times in cellular assays, which may be caused by the high concentration of ATP in the cell binding to its competing target. Therefore, the third-generation lung cancer target drug Osimertinib was successfully developed by using an irreversible covalent binding action strategy. This drug was approved by the FDA and the EU in 2017, and was certified by the CFDA in the same year.


    In addition to the strategies mentioned above, selectivity can also be achieved through the physicochemical properties of different tissues in the human body. For example, Omeprazole can act selectively on the proton pump as it requires strongly acidic conditions to activate.


    Alternatively, selectivity can be achieved through the careful design of electrophilic groups. As shown in Figure 10, the Michael addition of α-cyano acrylamide to a sulfhydryl group is a reversible process, and the addition product can be stabilized in a specific target due to the non-covalent effect of other amino acid residues, while other non-target targets cannot achieve this effect, thus achieving selectivity. Of course, this strategy needs to be backed by very powerful bioinformatics, otherwise, it is difficult to achieve.


    Toxicity issues

    Covalent drug design strategies are now becoming more and more sophisticated, and the risks should not be underestimated, especially the irreversible covalent binding effects. The formation of a new covalent bond between a small molecule and a target can cause changes in the structure of the protein and produce an immune response, but such reports are relatively rare. Off-target effects are also an important cause of toxicity, and improving the selectivity of compounds and lowering therapeutic doses are effective solutions to this problem.


    The majority of covalent binding drugs are currently focused on the anti-cancer field, but with the continuous development of technology and theories, such drugs will become available in more and more disease areas.