Engineering Therapeutics With A Structural Approach

A conversation with Faraz A. Choudhury, cofounder and CEO, Immuto Scientific

Most drugs fail in the clinic due to toxicity and efficacy. Overlap between diseased and healthy targets often causes unwanted side effects, leading to failure in clinical trials. Even in approved treatments, many cancer therapies work by harming both healthy and diseased tissues, causing significant side effects and toxicity issues.

In this Q&A, Life Science Connect’s Morgan Kohler caught up with Faraz A. Choudhury, cofounder and CEO of Immuto Scientific, to discuss the company’s research with target proteins and their adoption of disease-specific conformations that could increase how therapeutics discriminate between diseased and healthy tissues.

How did your background in engineering translate to drug discovery and what insights does that background give Immuto Scientific in its preclinical work?

My training began in electrical engineering, where complex systems are studied through empirical measurement, rigorous modeling, and continuous iteration. Early in my career, I became interested in how similar principles could be applied to biology, particularly to protein structure. The scientific tools that existed at the time were excellent for characterizing purified proteins, yet far less suited for examining structure directly in native, disease-relevant environments. That gap became the foundation for my transition into structural proteomics.

Engineering taught me to treat biological systems as structured, information-rich environments rather than collections of isolated molecules. When I moved into the field of structural biology at the University of Wisconsin, I approached problems such as protein conformational changes with an engineering mindset: you probe it without disturbing it, gather high-fidelity data under true operating conditions, and then extract meaningful patterns that can guide design.

That mindset is central to how we work today. Drug discovery often struggles because early assumptions about target behavior do not hold when tested in more complex models. An engineering-based approach keeps us grounded in empirical measurement, particularly when studying structural biology in living systems. We focus not only on which proteins are present but on how their structures and interactions shift under disease pressure. This lens has helped us reveal forms of biology that do not appear in conventional genomic or proteomic snapshots.

In practice, this has led us to build platforms and design experiments that operate closer to the native state in the cell, rather than treating structure as something that can be studied only in isolation. It influences how we engineer our internal platforms with scalability, reproducibility, and systematization in mind. We use robotics and custom software to build workflows that are high-throughput, automated, and reproducible, which allows us to generate structural data sets on thousands of proteins in living disease models — something that wasn’t previously possible. We've built custom AI stacks and generative deep learning models to interpret the large, rich data sets; recognize patterns; and accurately model protein conformations. We apply engineering principles throughout the entire process because we see drug discovery as an information problem as much as a biology problem.

How does the team at Immuto Scientific address and overcome challenges faced in their R&D?

Drug discovery presents two persistent challenges: biological complexity and the difficulty of translating early insights into meaningful therapeutic possibilities. To move through these challenges, our team follows a scientific process that favors direct evidence from disease-relevant systems and cross-disciplinary perspective.

We start with patient-derived and clinically relevant models whenever possible. Working in these systems forces us to confront the true biology of disease rather than its simplified representation. For example, structural changes in proteins often arise from stress, microenvironmental shifts, and cellular signaling. These features are invisible in purified systems yet fundamental to understanding which targets truly distinguish diseased and healthy tissues. Beginning with native models reduces later attrition by providing early clarity about what is selective, accessible, and functionally relevant.

Another challenge is distinguishing promising biology from misleading artifacts. Our data sets are high-dimensional and too complex for human intuition alone. AI plays a central role in how we overcome this challenge. We use machine learning to extract patterns, model protein structures, prioritize targets, and guide antibody design — effectively turning R&D into a feedback-driven loop. As new data comes in, the models improve and our decisions become more precise over time. This learning-based approach helps us avoid common dead ends in preclinical development.

Our team is interdisciplinary in the truest sense: engineers, protein biochemists, data scientists, AI/ML experts, and translational researchers work together from the earliest stages. This diversity helps us challenge scientific assumptions and reach a shared understanding of how structural features emerge and how they should guide design choices. That collaborative process has proven essential for navigating the uncertainties that come with working at the edge of new biological territory.

How has Immuto Scientific chosen what disease areas to focus on?

Our initial focus grew organically from the biology. Some diseases rely heavily on genetic drivers, but many are driven by shifts in protein structure and activation states that are not well captured by mutation or expression profiling. Cancer is a clear example. Tumors often reshape their surface architectures in response to metabolic stress, signaling, or microenvironmental pressure, creating conformational features that differ fundamentally from those on healthy cells. These differences often determine how a tumor interacts with its surroundings and how it evades the immune system. This makes cancer one of the clearest and most urgent settings in which structural insights can guide new therapeutic opportunities.

Acute myeloid leukemia (AML) became our first internal program because of a strong observation that the disease presents a consistent, conformationally distinct structural feature on a surface protein across diverse patient samples. The target is present in a form that appears tied to the pathological state, not to any specific mutation pattern. Additionally, AML is an aggressive disease that affects over 100,000 patients worldwide with very limited therapeutic options.

More broadly, we choose disease areas based on where structural remodeling is central to pathology, where unmet need is high, and where existing approaches have struggled to achieve meaningful selectivity. Oncology meets all three criteria, and so do multiple immunological conditions where proteins shift into activated or dysregulated conformations that drive inflammatory signaling. Structural change is a unifying principle across these diverse indications.

Why those specific disease areas?

These disease areas share a common challenge: traditional approaches do not provide a reliable path to high precision. Many targets that look promising at the genetic or expression level cannot be pursued safely because they remain present on critical healthy tissues. By focusing on diseases where pathological structure plays a defining role, we can detect features that would otherwise remain hidden.

These structural distinctions often arise from the disease microenvironment itself, rather than genetic alterations. In cancer, the shift toward hypoxia, altered metabolism, and immune evasion creates conditions that physically reshape proteins. In immune disorders, dysregulated signaling can tilt proteins toward active or inactive conformations that differ from their resting states. These structural signatures provide valuable selectivity precisely where traditional approaches struggle most.

We also prioritize conditions where patient heterogeneity has historically limited progress. Structural features tied to disease biology rather than genotype may provide opportunities that span diverse patient populations. The consistency of certain structural patterns across heterogeneous samples is one of the strongest signals we look for when selecting indications.

How is Immuto Scientific addressing the issue of translation from development to clinic?

Translation is one of the most difficult aspects of drug discovery. Many targets look promising in early studies yet fail to deliver selectivity, safety, or efficacy when tested in more complex biological systems. Our approach is to incorporate translational thinking at the earliest phase of discovery.

First, we prioritize observing protein features directly in living, disease-relevant systems. This avoids the common pitfall of basing target selection on purified or recombinant proteins that do not reflect their true structure in the tissue of interest. Structural biology in isolation often lacks the spatial and environmental context necessary for therapeutic design. By starting in native settings, we ensure that any structural feature we pursue is real in the context that matters.

Second, we evaluate disease specificity early rather than waiting until lead optimization. Our staged validation approach examines whether a structural feature persists in patient-derived samples across varying purity levels, tissue conditions, and clinical backgrounds. Only features that remain stable across these contexts advance.

Third, we incorporate functional considerations during target assessment. A structural distinction alone does not guarantee that a target internalizes, signals, or presents an accessible epitope. We integrate data from structural characterization, computational modeling, and cellular assays to understand whether a therapeutic could reliably engage the disease-specific form.

Finally, we maintain transparency about the limitations of each data type. Structural measurements need to be supported by biochemical, cellular, and translational experiments that collectively outline a realistic path toward clinical evaluation. The goal is not only to discover targets but to discover targets that stand a meaningful chance of succeeding in the clinic.

What is Immuto Scientific doing in its preclinical research that is different from what has traditionally been done?

Traditional approaches rely heavily on genomics or expression profiling to identify targets. These approaches reveal which genes or proteins are present or upregulated but offer limited visibility into the structural states that drive function and disease specificity. Yet structure governs how proteins interact, signal, and present epitopes. Structure also defines whether a therapeutic modality can discriminate between healthy and diseased tissues.

Our research focuses on measuring structural features of surface proteins in native biological settings, not after purification. This involves studying proteins as they exist in living cells, under conditions that preserve the microenvironmental pressures that shape their conformation. It also involves integrating empirical structural data with computational modeling to understand which features are present only in disease states.

We also align structural measurement with therapeutic design much earlier than is typical. Rather than treating structure and drug development as separate stages, we incorporate structural insights into binder generation and functional evaluation from the beginning. This coherence between discovery and design helps ensure that the earliest decisions support later therapeutic feasibility.

How has implementing a different approach shown up in your results?

This approach has changed the kinds of targets we discover, how we evaluate them, and the confidence with which we move forward.

One clear example is our work in AML. Instead of searching for differentially expressed proteins or mutation-driven targets, we examined structural differences directly in patient-derived samples. This revealed a conformationally distinct feature on a surface protein that appeared consistently across patients. It was not tied to a specific mutation or expression level. Instead, it reflected a fundamental aspect of AML biology. Because the structural feature was absent in matched healthy cells, it suggested that selective targeting may be achievable.

Early in vitro assays and preclinical studies reinforced this possibility. Bindings that recognized the disease-specific form did not show appreciable engagement of the normal form. In AML xenograft models, we observed strong in vivo efficacy of our drug, and it did not cause toxicity in human hematopoietic stem cells when administered to mice with a human immune system. These results provided early evidence that structural distinction could translate into actionable selectivity.

More broadly, across multiple programs, we have seen that starting with native-state structure allows us to identify targets that do not emerge from standard omics workflows. It also has allowed us to deprioritize targets that, although interesting initially, did not demonstrate structural stability or accessibility in more complex validation samples. This refinement prevents downstream investment in targets that are unlikely to succeed.

Our results demonstrate that studying structure in its true biological context can reveal new therapeutic possibilities that are both selective and widely present across heterogeneous patient populations. This is the most compelling indicator that a structural approach to target discovery can reshape what is possible in drug development.

About The Expert

Faraz A. Choudhury, Ph.D., is the cofounder and chief executive officer of Immuto Scientific, a biotechnology company pioneering structural surfaceomics to transform how new drug targets and therapeutics are discovered. He co-invented Immuto Scientific’s initial core technology and guided its development into a high-throughput structural proteomics platform capable of mapping protein conformations directly in living cells and patient samples.

Prior to founding Immuto Scientific, Choudhury served as a research scientist in the Department of Biochemistry at the University of Wisconsin-Madison, where he applied advanced mass spectrometry to study protein structures and interactions. He earned his Ph.D. in electrical engineering and a graduate certificate in entrepreneurship from UW–Madison, where his doctoral research focused on plasma technologies for both semiconductor and biological applications.