Top 5 AIdesigned drugs in trials

As of late, the convergence of AI and pharmaceutical research has provided groundbreaking discoveries that are rebuilding the drug discovery and development field. In this listicle, we share some of the top most remarkable drug developments that have been made possible by advancements in machine learning algorithms and computational techniques. Besides illustrating the potential of AI to revolutionize medicine, these examples also highlight the valuable collaboration of technological advancements combined with human ingenuity and expertise.

1. INS018_055

You might have heard about INS018_055 through the grapevine or in the news, and with good reason: it is actually the first fully AI-generated drug to enter human clinical trials. While there are alternative AI-designed drugs in trials, INS018_055 is unique in the fact that it is both AI-designed and its novel target was also AI-discovered. The mystery drug target has not been revealed by Insilico Medicine (NY, USA) on competitor grounds, however, we do know that the target regulates at least three pathways involved in fibrosis, the Wnt pathway, the YAP/TAZ pathway and the TGF- pathway. The drug was developed using Insilco’s Pharma.AI platform, which includes two AI-driven drug discovery and development tools, PandaOmics and Chemistry42 . While PandaOmics allows the multi-omics identification and analysis of novel targets involved in particular disease pathways using AI algorithms, Chemistry42 combines machine learning with medicinal chemistry and computational technologies to design novel molecules with favorable properties for the targets that were discovered using PandaOmics . Its creator? Insilico Medicine. What is it for? INS018_055 was created to treat idiopathic pulmonary fibrosis, a chronic progressive disease that causes fibrosis (scar tissue) to accumulate in the lungs, making breathing increasingly difficult. The condition affects approximately 3 million people globally . What stage is it at now? Phase II trials.

2. EXS-21546

EXS-21546 is an AI-designed A2A receptor antagonist, created by Exscientia (Oxford, UK) to treat patients with advanced solid tumors. Exscientia considers themselves pioneers in AI drug discovery, and rightfully so. They produced the first AI-designed molecules to reach clinical trials. Unfortunately, this drug didn’t make it to approval, but the company is still a huge contender in the AI drug game. EXS-21546 was generated by the company’s Centaur ChemistTM platform, which uses AI to design and prioritize novel compounds for synthesis whilst also enabling the platform to learn strategy from human experts. As of November 2022, the drug received approval to progress to Phase I/II trials, where it will be tested alongside anti-PD-1 therapy in immunotherapy relapsed or refractory renal cell carcinoma (RCC) and non-small cell lung cancer (NSCLC) patients. Exscientia hopes to expand testing the drug on other tumor types in future trials . Its creator? Exscientia. What is it for? EXS-21546 was designed to treat solid tumors exhibiting high adenosine signatures. Adenosine is a well-known immunosuppressant and is often synthesized at elevated levels in tumor microenvironments compared to healthy tissues, therefore is an attractive target for therapeutic interventions. Certain tumors produce high levels of adenosine, which binds to and activates A2A receptors present on immune cells. This in turn dampens the immune system’s natural anti-tumor responses. EXS-21546 is currently under investigation for its ability to inhibit the activation of A2A receptors by high adenosine concentrations. In doing so, this drug has the potential to enhance the immune system’s anti-tumor activities and avoid adenosine-mediated immunosuppression . What stage is it at now? Phase I/II trials.

3. Baricitinib

Within 48 hours, BenevolentAI (London, UK) used their AI-driven drug discovery platform to explore and analyze more than 85 data sources of biomedical information for mechanisms related to viral infection and inflammatory responses in COVID-19, including data from academic literature, clinical trials and multiomics data. From this huge mass of data, deep learning algorithms were used to extract novel hypotheses based on more than a billion relationships between proteins, targets, genes, diseases and drugs. These connections are stored in BenevolentAI’s Knowledge Graph, which was then further analyzed with AI tools to guide drug discovery and potential applications from pre-existing data. Their analysis pointed to baricitinib as the strongest candidate . Its creator? Originally owned by Eli Lilly (IN, USA) but identified for repurposing by BenevolentAI. What is it for? Baricitinib was actually a drug originally approved for the treatment of rheumatoid arthritis. However, in February 2020, BenevolentAI discovered that this same drug could be used to treat COVID-19 owing to its anti-viral and anti-inflammatory effects. Both the WHO and the FDA advocate the use of baricitinib in treating hospitalized COVID-19 patients and the COV-BARRIER trial has particularly promising results; there was a 38% reduction in fatalities among hospitalized adult COVID-19 patients who received standard of care alongside baricitinib. This is just one example of how combining machine learning with human expertise can unveil fresh insights concealed within extensive volumes of published data and research . What stage is it at now? FDA approved.

4. ISM3091

You might have seen this name in the news recently, likely because it was granted US FDA initial investigational new drug (IND) approval for the treatment of patients with solid tumors. Like INS018_055, ISM3091 was designed with Insilico’s machine learning tool Chemistry42 . Its creator? Insilico Medicine. What is it for? The treatment of solid tumors, which account for approximately 90% of adult cancers. ISM3091 is a highly selective, potent small molecule inhibitor of USP1, a gene encoding a deubiquitinase that is known to play a significant role in the DNA damage response and therefore cancer progression. The drug has exhibited strong anti-proliferative activity against BRCA1 mutant cells, as well as in HRR-proficient models, Interestingly, ISM3091 also possesses the exciting potential to combat PARP inhibitor resistance - an increasing issue in cancer therapy . What stage is it at now? Phase I trials (as of July this year).

5. REC-2282

Last, but not least, we have REC-2282. Having been granted Fast Track designation by the US FDA, as well as Orphan Drug designation by both the European Commission and US FDA in October of 2021, REC-2282 has now successfully progressed through to Phase II/III trials. The drug was identified as a potential candidate to treat disease resulting from mutation in the NF2 gene using the company’s AI-driven drug discovery platform: Recursion OS. This platform uses machine learning to identify relationships between biological contexts and chemical entities from huge datasets . Its creator? Recursion Pharmaceuticals (UT, USA). What is it for? The treatment of NF2-mutated meningiomas. Meningiomas are primary tumors of the meninges of the central nervous system. More than 34,000 patients are diagnosed with sporadic meningiomas in the US each year. In patients with progressive meningiomas, more than a third are driven by mutations in the gene NF2. NF2-mutated meningiomas are typically treated with surgery and radiotherapy or with off-label therapies of limited or unproven efficacy. The lack of approved therapies to prevent the progression of NF2-mutated meningiomas, which are frequently an aggressive tumor type, represents a significant unmet medical need. In addition, the syndrome neurofibromatosis type-2, which is associated with both sporadic and autosomal dominant inherited mutations in the NF2 gene, gives rise to meningiomas in approximately half of neurofibromatosis type-2 patients. In the US and EU5, neurofibromatosis type-2 affects approximately 33,000 patients and there are no approved therapies. REC-2282 is an HDAC inhibitor - an enzyme involved in tumor cell growth, including NF2-mutated mengiomas . What stage is it at now? Phase II/III trials. Be sure to get in touch if you have any queries about Future Medicine AI or are interested in publishing in the journal, please contact Commissioning Editor, Emma Hall (e.hall@future-science-group.com). Disclaimers:The opinions expressed in this feature are those of the author and do not necessarily reflect the views of Future Medicine AI Hub or Future Science Group.