It’s becoming increasingly apparent that treatments for complex diseases such as cancer, neurodegenerative diseases and depression are unlikely to target a single molecule in the pathological pathway. Drug developers are turning to new ways to exploit more than one drug target at a time. Here we look at some of the approaches being taken to achieve therapeutic responses that are greater than a single molecule can achieve alone.
Bispecific antibody approaches
Antibody based drugs have been making a major impact on health for several decades. But for some time scientists have been looking at different antibody formats to potentiate their activity. An area that’s creating a buzz right now is the development of bispecific antibodies. These are antibodies that recognize two different epitopes either on the same or on different antigens. This gives them two main advantages: they can target two different biological mechanisms simultaneously and can link targets that are distal to each other.
It’s an approach being embraced in cancer immunotherapy, with two bispecific drugs (emicizumab and blinatumomab) currently approved. More than 200 others are in or about to enter clinical trials.
So why this sudden spark of activity in this area? According to this opinion piece, it’s the realization that monospecific targeting through classical IgG formats is not going to be enough to successfully treat diseases, and ultimately, to help patients. Cancer is one of the leading indications that has seen a sudden explosion in the use of bispecifics and is focused on two approaches: either engaging the immune system or targeting dual targets on tumor cells. Bispecifics are not just effective at targeting cancer, drug developers working on other diseases such as Alzheimer’s are also getting in on the act.
According to Liusong Yin, Senior Director of Antibody Services at GenScript, there are two major challenges in developing bispecific antibodies. “First, by over-engineering naturally generated antibodies this can cause an antibody therapeutic to produce its own immune response, making the drug ineffective,” he explains. “The non-natural format of bi-specific antibodies, such as product instability, low expression level, and complex purification processes can also lead to manufacturing issues.”
Genscript aims to avoid this by engineering antibodies that have naturally generated monoclonal antibody backbones, providing several advantages over mono-specific therapeutics. “They show bio-superiority and a better safety profile over monotherapy or combinatorial therapeutics, and what we’re finding with our platform is that it can generate high yields and concentration in formulation with desirable stability,” explains Yin.
An alternative bispecific strategy is currently being pursued by F-star Biotechnology. They have designed an antibody which incorporates the conventional features of an IgG including the Fab antigen-binding region but has an additional binding site in the Fc region. Specially designed antibody fragments called Fcabs can then be plugged into the Fc region meaning it is possible to construct versatile molecules capable of associating with different targets. They’ve used this approach to design a dual checkpoint inhibitor, currently in preclinical development.
This move towards dual-targeting therapeutics is also starting to spill into small-molecule drug discovery, where computational chemistry could speed up the search for molecules with multiple activity.
Jens Carlsson, Associate Professor at Uppsala University and the Science for Life Laboratory, is using virtual screening to identify potential drug candidates that have dual activity against G protein-coupled receptor targets. “We have been working a lot on finding selective molecules that bind to one target and not another. But now we are taking the opposite approach and finding molecules that act on proteins that are very different but relevant for the same disease.”
Carlsson and colleagues used structure-based virtual screening to find compounds that would simultaneously target and suppress two unrelated molecules associated with Parkinson’s disease – the A2A adenosine receptor (A2AAR) and monoamine oxidase B enzyme (MAO-B).1 After screening 5.4 million compounds they identified 24 that had elevated binding potential to both, and experiments showed that four compounds targeted both proteins. Two of these will be tested further in preclinical models.
Carlsson was surprised to find these potential dual-action drugs. “I thought it would be more difficult. But the advantage of our approach is that we now have access to large libraries of millions of molecules that would not be accessible to experimental approaches. We have a high failure rate but it’s cheap failure as only a small set of compounds need to be tested experimentally.”
They are now extending the approach to other complex diseases including cancer, where the adenosine receptor is increasingly thought to play a role. “A multi-targeted approach will be efficient in treatment of cancer, and in other difficult diseases in neuroscience like Alzheimer’s disease and depression, where the classical approach of finding a single target and hitting it just won’t work. For HIV, yes, once you hit the virus you can take it out, but there’s no single receptor for depression.”
Carlsson hopes that the drug developers will recognize the potential that computational chemistry has in finding dual-action molecules. “When I started in this research field it was considered to be very difficult to find any active molecules with virtual screening and we couldn’t screen large libraries because we didn’t have the computational power. Now when we have tons of computational power, it turns out that we already have methods that can be useful for finding molecules with very complex properties. Our approach shows a lot of promise.”
Smarter drug combinations
Perhaps the most well-established approach to exploiting multiple targets is the use of combination therapies – where treatments are expected to act synergistically, providing an additive response over either drug alone. But although this approach is used often, it still presents significant challenges when it comes to clinical development.
More than 10,000 ongoing clinical trials are currently registered in the US alone investigating combination therapies for complex diseases, and there are even more preclinical-research articles on drug combinations.2 Yet, this is a small proportion of all the potential combinations that could be tested. So how do you select which drugs to combine?
Moreover, the multiple comparisons required to get a clear signal on whether a drug combination has an adequate degree of synergy is a challenge. This is before you consider looking at secondary PK/PD or toxicology outcomes which can make such studies even more complex.
Systems biology and artificial intelligence may well hold the answer. Conducting better preclinical studies based on predictive modeling which selects drugs that, based on their mechanism of action, are most likely to show synergistic effects could help to select the most promising combinations to take into larger clinical studies.
To this end, researchers are being called on to develop a set of defined mechanistic “rules” for combination treatments in cancer, which would allow treatment modalities to be rationally and effectively combined based on the biology of resistance and potentially the evolutionary projection of the treated tumor.
A further issue is the availability of proprietary drugs for combination studies. To address this, the National Cancer Institute has launched the NCI Formulary, a public-private partnership that aims to give investigators who would like to perform combination studies rapid access to drugs from multiple collaborating pharmaceutical companies.
The hope is that more rational selection of combinations combined with intelligent study design and defined criteria of success will make for more successful clinical trials.
1. Jaiteh, M. et al. Docking screens for dual inhibitors of disparate drug targets for Parkinson’s disease. J. Med. Chem. 2018;61:5269–78
2. Editorial. Rationalizing combination therapies. Nat. Med. 2017;3:1113