Dr. Peter Tummino, president of R&D at Nimbus, spoke with Pharmaceutical Executive about how new technologies are changing the drug development process. It isn’t just adding new tools to researchers’ toolbox, its changing the foundational approach to discovery.
Key Takeaways
- Computational chemistry combined with AI technologies has become a powerful tool for assessing small molecule properties, high-resolution protein structures, and molecular mechanism of action, as well as optimizing target selection and small molecule design.
- One major challenge in small molecule drug discovery today is how the goalposts for achievement are being defined for AI.
- The real advantage of AI is that it can empower scientists and make them much more capable.
Pharmaceutical Executive: What role will computational chemistry play in drug discovery and development?
Dr. Peter Tummino: Computational chemistry combined with AI technologies has become a powerful tool for assessing small molecule properties, high-resolution protein structures, and molecular mechanism of action, as well as optimizing target selection and small molecule design. At Nimbus, it's fully embedded in how we innovate. It’s not just a tool we use, but fundamental to our approach.
We apply our physics-based drug discovery engine to visualize small molecule-protein interactions with extraordinary detail and predict potency, selectivity, and other key properties. Combined with AI and machine learning tools, we can design molecules with optimal druglike properties with greater precision.
Computational chemistry tools will increasingly improve medicinal chemists' decision-making in drug discovery. Similar to the genomics revolution 20 years ago, genomics didn't replace traditional approaches, but it provided powerful tools for biologists. The industry is moving toward a hybrid model where computational chemists work alongside medicinal chemists. Over time, the model will evolve to one where chemists possess computational expertise and are adept at applying both disciplines to designing molecules.
PE: What is Nimbus' target selection and drug discovery process?
Tummino: We take a disciplined approach to target selection. We focus on areas of high unmet medical need that impact large patient populations. We select well-validated targets where we are not taking biological validation risk, but instead focus on making better molecules than what other companies may be able to achieve. These are often targets where others can't achieve the potency or selectivity needed for efficacy with an appropriate therapeutic window. We're therapy area agnostic, working across multiple areas to find the best opportunities that fit our drug discovery approach, always looking for substantial competitive advantages in our molecules rather than incremental improvements. From the start, we design to achieve improved small molecule attributes that will benefit patients.
This approach has enabled success with difficult targets like AMPK, where we recently achieved a significant preclinical milestone in our Eli Lilly collaboration. This is a target that has long been recognized as a promising therapeutic target for multiple clinical indications in metabolic disease, but previous attempts to effectively drug the target were unsuccessful due to the technical challenges of developing selective activators of specific AMPK isoforms. Nimbus has leveraged its computational approaches and structure-based drug design capabilities to overcome these historical obstacles. We have applied our approach to other challenging targets including ACC, TYK2, HPK1, and WRN, with strong cases for best-in-class molecules in each area. Best-in-class is only meaningful if it creates an opportunity to improve patient treatment.
Part of the success of our approach as a small biotech is the willingness to stop programs and redeploy resources. We set high standards and make objective decisions to focus our resources on programs with the most promising potential for patients with significant unmet needs.
PE: What does the company have planned for the next five years?
Tummino: Nimbus is advancing a diverse, multi-therapeutic pipeline. We've had three programs enter the clinic in our first 15 years, and we're poised to have three more first-in-human programs in just three years, a significant acceleration that reflects the maturation of our discovery engine.
I'm especially excited about two programs. First is our novel non-covalent Werner syndrome helicase (WRN) inhibitor, which entered Phase 1/2 trials earlier this year. WRN is a well-validated target for microsatellite instability (MSI) solid tumors. Our preclinical data show the compound is potent, selective, and produces significant tumor regression and sustained complete responses at low doses in MSI-high models refractory to immunotherapy and chemotherapy. We believe it has best-in-class potential for patients with limited treatment options.
Second is our salt-inducible kinase (SIK) inhibitor for immunological and inflammatory diseases. SIK2 controls both inflammation and its resolution. Our compelling preclinical data indicate that potent selective SIK2 inhibition dials down pro-inflammation cytokines in overactive immune cells while boosting IL-10, a powerful anti-inflammatory cytokine that initiates tissue healing. This combination of anti-inflammatory activity and tissue healing may be particularly valuable for inflammatory bowel disease.
Each of the four programs we’ve put into the clinic are currently still in the clinic, and we are working to expand that success, focusing on bringing truly novel, best-in-class medicines to patients who need them most.
PE: What does Nimbus consider to be the biggest challenges impacting the industry?
Tummino: One major challenge in small molecule drug discovery today is how the goalposts for achievement are being defined for AI. The current focus has been primarily on speed, particularly reducing timelines from program initiation to clinical proof-of-concept (POC). While faster timelines are important, I believe this misses the larger opportunity. The most critical goal in small molecule drug discovery should be effectively drugging the most important therapeutic targets for a particular disease indication. When speed becomes the primary metric, competitive biotechs tend to pursue targets that are easier to drug rather than the most impactful ones. This approach risks higher failure rates in Phase 2 trials.
In contrast, consider the successes we’ve seen with companies like Amgen and Mirati, who have successfully drugged KRAS for oncology, and Kymera, who has made breakthroughs with STAT6 for autoimmune diseases. These are major advances in treatment options with significant potential to benefit patients. These programs likely did not fit the aspirational timelines that the industry talks about, but they represent the kind of meaningful progress our field should prioritize when employing AI to improve drug development.
The biggest challenges I see is the industry losing sight of what should be at the center of drug discovery: outstanding scientists. You really need pure, outstanding scientists to succeed. Average science is not going to be able to tackle these very difficult clinical indications and develop breakthrough medicines.
The real advantage of AI is that it can empower scientists and make them much more capable, but it's just a tool. There is a growing narrative that AI will replace human judgment, but it takes human expertise to make sound drug development decisions. Companies succeed by approaching AI with healthy skepticism, integrating it thoughtfully as a force multiplier in the hands of experienced scientists, while keeping the focus on the fundamentals: recruiting and rewarding outstanding scientists as the most important people in the organization.