Bridging Academia and Industry: The Evolution of 3D Models in Drug Development

The landscape of drug development is rapidly evolving, with a notable shift towards human-relevant, predictive, animal-free models and in vitro innovations. This represents a critical step forward in efforts to enhance the efficacy and safety of new medicines. The industry's search for alternatives to traditional preclinical drug testing methods has led to a growing interest in the potential of three-dimensional (3D) in vitro models. These advanced models offer a more accurate reflection of human biology, potentially revolutionizing how drugs are developed and tested.

To delve deeper into this subject, Technology Networks spoke with Madhu Lal-Nag, chief scientific officer at InSphero, a 3D in vitro model company with a mission to “modernize drug discovery in ways that inspire researchers everywhere to reach their full potential and fuel a new era of breakthrough therapies.”

Our discussion explored the current state of in vitro model development, the importance of addressing disease-specific challenges and representing ethnic and genetic diversity in research.

Laura Lansdowne (LL): Could you share details about your current position at InSphero, your professional history and your expertise in drug development?

Madhu Lal-Nag (ML-N): I've been at InSphero for just over a year now, but my relationship with the company actually started many years ago.

I'm trained as a molecular oncologist. Shortly after I graduated, in 2009, I started working on the development of 3D models ‒ I was one of the first researchers to test InSphero’s plates.

I was at the National Institute of Health (NIH) while completing my postdoc and I went from using ex vivo models at NIA/NIH to developing them for high-throughput screening at the National Center for Advancing Translational Sciences (NCATS), to test small molecule drugs.

After approximately two years I moved on to become the head of the Trans NIH RNAi Facility. Here, we used cell-based 3D models for whole-genome RNAi and CRISPR screens. In that role, my laboratory developed and implemented many imaging assays for oncology using 3D models, as well as running infectivity assays and biochemical assays, etc.

I then joined the US Food and Drug Administration (FDA), which was an incredible experience. the longest time, we'd been developing models with a focus on efficacy. And when you are creating these models, you ask questions like, “How well does the drug engage my target? Are there any off-target effects?”

It may seem promising, but in reality, until you administer that drug to a human being, you just don’t know what you are dealing with. The moment it reaches safety testing, there's a 50:50 chance it's not going to cut it.

During my time at the FDA, I gained a huge respect for drug safety. It doesn't matter how good your drug is, if it's not safe, it doesn’t matter. That realization was really very humbling.

I then moved into drug development in pediatrics for a little while, with a focus on oncology, which was a great learning experience. I would have continued that for longer, but when this opportunity presented itself, I had to go for it – this area is my first love (other than my children of course!).

When I was working with these models 13‒14 years ago, it was a very academic exercise. But now the industry is on the brink of adopting this approach; the regulators are endorsing the use of these models and validating them. It’s a very exciting time.

Currently, there is no defined roadmap that states “by XYZ date we expect you to transition to non-animal models.” But I don’t think it will be long before we see this shift, because guidance is now available, and we are being encouraged to use these in vitro models.

Nowadays the regulatory authorities want take part in these conversations. Historically there has always been a “wall” between the regulators and those they regulate, but there's a lot more interaction now, and that's very heartening to see.

LL: In the interview I conducted with InSphero in 2021, automating body-on-a-chip and microfluidics processes for a high-throughput setting was flagged as a key challenge. Can you elaborate on current progress in this area?

ML-N: I think we've definitely made a lot of progress. But it comes back to another question ‒ what are you looking to address using your model? There are many different complex in vitro models available, but you have to determine which is best to use to get your answer.

InSphero’s models recapitulate liver complexity and can be used in a high-throughput setting. Well-for- well, our models give their users information about human relevant physiology and insights into mechanism of action for their investigational therapeutic, but they will not give you answers to questions related to perfusion of flow, for example. Only a liver-on-a-chip would be able to do that, but you never get the same scalability.

InSphero twin platform with liver and cancer microtissues. Credit: InSphero.

You need to understand where in the drug discovery continuum your model fits and what answers it can give you, if you know what questions to ask. So, for now, as leaders in the in vitro toxicology space, we're able to model mechanism of action, run assays for predictive toxicology and can compare across preclinical species in a 384-well plate format, which is tremendous progress. We are one of the very few model providers that can provide sufficient physiological relevance in a scalable and reproducible system, which greatly increases the value proposition for anyone who uses it.

LL: How do you balance the need for innovation in safety testing with the commercial realities of drug development, and how does this balance impact the adoption of new technologies?

ML-N: That's an excellent question. What we've been able to do successfully over a period of time, is build a very valuable network of our customers.

While innovation is key, innovation in silo doesn't make a whole lot of sense.

We make sure to engage with key stakeholders, including our customers, the regulators and others such as the Critical Path Institute. We comprised a think tank, so that any products we develop that will be adopted by customers, has already been vetted. Our customers know what the gaps are, and regulators know what they want and need to see ‒ it’s making sure the pieces all fit together.

LL: What are the specific challenges you face when developing disease-specific in vitro models?

ML-N: Patient heterogeneity is a key challenge. In the metabolic space (e.g., when developing our liver disease and diabetes models) we use different donors, which enables us to identify donor-specific differences. This is especially important when you're thinking about the impact of single nucleotide polymorphisms.

In the oncology space, we're also focusing a lot more on patient heterogeneity, and the involvement of the tumor microenvironment with the addition of patient, disease and niche-relevant cell types, or we use patient-derived cells to be able to model that heterogeneity.

There are some aspects of certain disease physiologies that even induced pluripotent stem cells (IPSCs) cannot recapitulate. For example, if there is an epigenetic effect, an IPSC model wouldn't be able to recapitulate that. It’s really important to maximize what information we can glean from patient-derived material.

LL: How does InSphero address the representation of ethnic and genetic diversity in its 3D microtissues?

ML-N: InSphero is part of consortium with many public-private partnerships both in the US and in Europe, that aim to address patient diversity.

This is something that should be considered throughout the drug development process, at both the preclinical and clinical stages. When I was at the FDA, Dr. Janet Woodcock, who was the director of the Center for Drug Evaluation and Research (CDER) at the time, really pushed for the inclusion of different demographics in clinical trial design. To ensure the drug is going to work in the patients it’s intended to treat, it is imperative to have that representation not only in the clinical trial setting but in the preclinical models as well, to accurately capture those responses.

LL: Looking ahead, what are the major milestones you anticipate in the next 5‒10 years for animal-free testing in drug discovery and development?

ML-N: I really do believe adoption of these models will happen and regulators are going to play a huge role in making this come together. From the perspective of the vendors, method-providers and pharmaceutical companies, what is incumbent upon us, is that we have to show these models are first and foremost human-relevant, which I think we're already doing.

However, we must also prove that they are at least as predictive ‒ if not more predictive ‒ than the models they're used to seeing.

We need to generate enough data to demonstrate that you can achieve robust and reproducible results at a higher throughput, in a shorter time and at an acceptable cost, for them to endorse the use of these models as alternatives to preclinical animal models.

In the space of complex in vitro models, I don't think we've done enough to push working in a collaborative manner.

Every organ system or complex in vitro model has its own niche where it provides a lot of value.

What we need to do as a community is understand how different companies can work together to provide the ultimate benefit to researchers who are focused on therapeutic development.

There needs to be mutual trust. I think it will happen, but the question is: how quickly?

Dr. Madhu Lal-Nag was speaking to Laura Elizabeth Lansdowne, Managing Editor for Technology Networks.

About the Interviewee

Before joining InSphero as chief scientific officer in 2023, Madhu Lal-Nag served as the director of the Trans NIH RNAi Facility at the National Center for Advancing Translational Sciences and as a Program Lead at the US Food and Drug Administration. She completed her PhD in molecular and cellular oncology at George Washington University and holds a MSc degree in bioscience business from Keck Graduate Institute.