Developing a new drug from scratch is an enormously costly and carbon-intensive process. On average, bringing a single drug to market generates 20,000 to 60,000 tons of CO2, with the pharmaceutical industry producing 48.55 tons of CO2 equivalents per $1 million spent — over 50 percent higher than the automotive industry.
These figures are driven not only by complex laboratory processes and clinical trials but also by the high failure rate. Approximately 75 percent of drug candidates never reach patients due to safety or efficacy issues. The result is a system that is financially, clinically, and environmentally unsustainable.
At the European Laboratory Research & Innovation Group (ELRIG) 2025 conference in Liverpool, DDN spoke with cofounders Jordan Lane and Layla Hosseini-Gerami of Ignota Labs to explore how their company is harnessing artificial intelligence (AI) to revive previously failed drugs, reduce the risk of failure, and significantly cut the pharmaceutical industry’s environmental footprint. By focusing on drugs that already show promise, Ignota Labs aims to accelerate development timelines, improve safety, and reduce carbon emissions by up to 80 percent compared with traditional drug discovery.
Understanding drug failure
To achieve their goal of reviving shelved drugs, Lane said that the team at Ignota Labs focuses on deeply understanding drug safety issues first. "We use our AI to map the mechanistic pathway of what's happened and why. If we can understand it, we can know what to change in the drug molecule. We can then buy or partner with the drug holders and develop a new clinical candidate, pushing it back into the clinic as quickly as possible.”
A prime example of this approach comes from Ignota Labs’ recent study on erlotinib and gefitinib, two cancer drugs used to treat non-small cell lung cancer (NSCLC). Both drugs are effective but known for their risk of liver toxicity.
While erlotinib’s toxicity mechanism was partially understood prior to the new study, gefitinib’s had remained unclear for over 20 years. Using their AI platform, SAFEPATH, Ignota Labs was able to confirm erlotinib’s mechanism, which involves inhibition of an enzyme called UGT-1A (uridine diphosphate-glucuronosyltransferase 1-1). This enzyme normally helps break down certain drugs in the liver, so its inhibition can lead to dangerous drug interactions and liver damage.
For gefitinib, SAFEPATH predicted a previously unknown mechanism of toxicity involving PRKD1 (Protein Kinase D1) and PRKD3, enzymes that regulate sphingolipid metabolism, a process critical for liver cell health. Experimental validation confirmed the computational findings, showing that inhibition of PRKD1 and PRKD3 leads to dose-dependent liver injury.
Further analysis revealed inter-patient variability in enzyme activity as a key factor explaining why some individuals develop hepatotoxicity while others tolerate treatment. These results provide a mechanistic foundation for tailoring doses based on molecular or metabolic profiling, enabling safer, more precise therapy.
Choosing appropriate candidates
Lane explained that Ignota looks for drug candidates that already have strong evidence backing their efficacy, and where the commercial opportunity is still viable. “If a drug hits a clinical issue in Phase 2, taking it back into development could leave us years behind other leads. And if we feel we can’t realistically change its profile, then it’s unfortunately not a candidate,” he said. Beyond those considerations, the platform is largely agnostic to drug type or disease area — the key requirement is that their technology can identify a credible mechanistic signal to investigate and act on.
To support their identification of the best candidates for drug revival, the company has developed an in-house system called RADAR, an agentic research tool that scans global datasets, scientific literature, and internal bioinformatics signals. Hosseini-Gerami explained that this tool “compiles a detailed report summarising why the drug failed, adverse events, mechanism of action, target engagement, and commercial potential.”
Once they find a suitable drug with a validated mechanistic hypothesis, Ignota’s strategy is to refine the molecule and return it to the clinic, often years faster than if a program began from scratch. Because earlier development work — such as pharmacology, formulation, and in some cases Phase 1 or Phase 2 data — already exists, financial and environmental costs are substantially lower than for a new discovery effort.
Reviving promising drug candidates
This strategy is already being applied to several of Ignota’s active programs. In October, Ignota Labs acquired several assets from Kronos Bio, including a CDK9 (cyclin-dependent kinase 9) inhibitor known as istisociclib and two SYK (spleen tyrosine kinase) inhibitors, entospletinib and lanraplenib.
Kronos stopped development of istisociclib after a Phase 1/2 trial in patients with platinum-resistant ovarian cancer was discontinued due to signs of neurotoxicity. However, CDK9 remains a highly promising target, with excellent kinase selectivity and oral bioavailability. Ignota’s approach focuses on reducing brain penetration and optimizing pharmacokinetics, aiming to preserve anti-cancer efficacy while avoiding central nervous system side effects.
These drugs would have had strong commercial potential to get to where they got … If we have now found safe and efficacious ways to revive them, it makes sense to invest further.
– Jordan Lane, Ignota Labs
The company is also advancing both entospletinib and lanraplenib, both of which disrupt immune cell signaling. These candidates were originally developed for B-cell malignancies and autoimmune diseases. Entospletinib progressed to Phase 2 trials for relapsed or refractory hematologic malignancies, including chronic lymphocytic leukemia (CLL). However, 74 percent of the patients experienced a grade 3 adverse event or worse, leading to discontinuation in nearly one-fifth of participants. Despite these setbacks, it still demonstrated that entospletinib could help overcome resistance in CLL to current treatments, potentially benefiting up to 30 percent of patients who no longer respond to standard care.
Lanraplenib is considered to be a second-generation SYK inhibitor, with enhanced pharmacokinetic properties. Lanraplenib has been in clinical trials for a range of autoimmune diseases, including lupus membranous nephropathy and Sjogren's syndrome. Ignota is now optimizing this inhibitor to achieve best-in-class efficacy in another autoimmune disease, immune thrombocytopenia, while avoiding the toxicity seen with current therapies. “These drugs would have had strong commercial potential to get to where they got,” Lane noted. “If we have now found safe and efficacious ways to revive them, it makes sense to invest further.”
A balanced view
While AI offers powerful tools, it has its own environmental footprint. Training large language models emits 100-500 tons of CO2, and everyday queries to systems like ChatGPT consume 30-50mL of water per query — enough to affect the annual drinking water supply of one million people at scale. Even so, Ignota Labs emphasizes that the carbon savings from reviving failed drugs far exceed the footprint of AI, reducing the need for repeated trials, wasted materials, and the extensive energy consumption of traditional drug discovery.
Most startups focus on creating entirely new drugs, however, Ignota has chosen to concentrate on refining existing candidates that others have left behind. By bringing failed drugs back to life, the company is not only exposing the untapped therapeutic potential of these molecules but also significantly reducing the time, cost, and environmental impact associated with traditional drug discovery.
