In this Q&A, VoxCell’s Graham Craig explores why human-relevant, vascularized tissue models are reshaping preclinical development — and how they could redefine the future of drug testing.
In recent years, we’ve seen a significant move toward human-relevant models. What do you believe is the primary driver behind this shift in drug development?
Graham Craig (GC): I believe there are two main reasons for this shift. In recent years, we have seen the emergence of innovative new drug modalities, including biologics, antibody-drug conjugates (ADCs) and immunomodulators. According to the 2024 Drug Modalities Report, nine of the top 10 drugs may well be new modalities by 2029. This is potentially exciting, but it comes with its own challenges.
These drug modalities promise the opportunity to address more complex diseases. However, they present unique toxicology profiles and effects that must be thoroughly assessed to ensure patient safety. As drugs become ever more specific to humans, it is increasingly unsatisfactory to rely on legacy preclinical models to predict human outcomes. We are effectively relying on preclinical systems to predict human delivery, target engagement, efficacy and safety in settings where species biology diverges.
The second factor is economics. For drug developers, late-stage failure is costly. They cannot afford to encounter obstacles late in development because drugs exhibit unexpected effects that are not highlighted in preclinical nonhuman testing. Teams are increasingly being measured on factors such as capital efficiency, cycle time and probability of technical success.
These two drivers are elevating human-relevant models from “nice to have” to “need to have” by enabling a more seamless transition from non-clinical signals to human responses.
With the introduction of new guidelines, such as the FDA Modernization Act 3.0, how is the regulatory landscape changing the way drug developers approach their non-clinical data?
GC: Recent legislation in the USA and Europe reflects growing support for non-animal testing methodologies. In the US, animal testing was previously regarded as the only acceptable route for demonstrating non-clinical safety before moving on to in-human studies. The FDA Modernization Act 3.0 removed this statutory requirement, opening the door to a range of new approach methodologies (NAMs). The European Medicines Agency (EMA) has also championed the use of NAMs to replace, reduce and refine the use of animal models in medicine testing across the EU.
The FDA itself has highlighted the potential of approaches such as human organ model-based lab testing and AI-based computational modeling in producing treatments faster and more reliably. As a result, the landscape is shifting from a culture of “animal by default” to one that prioritizes “fit-for-purpose evidence,” where developers must produce rigorous, well-structured data that demonstrates efficacy and regulatory compliance.
In practice, this means developers must now design non-clinical packages around the credibility of their predictions, rather than tradition. While the statutory requirements have loosened to accommodate more modern approaches, companies are under even greater pressure to demonstrate a stronger mechanistic rationale, greater human relevance and clearer context-of-use across all datasets.
How is this increasing alignment on NAMs helping to accelerate the use of human-relevant models in drug development?
GC: To a great extent, this is a case of legislation catching up to the science. There was already significant belief in the efficacy of human-relevant models. However, uncertainty around regulatory acceptance and standards had an understandable effect on innovation and adoption.
The introduction of more definitive guidance around NAMs and human-centric approaches will almost certainly accelerate that process. Legislation such as the Modernization Act 3.0 encourages drug developers to invest in NAMs with greater confidence. That confidence will only increase as regulators, contract research organizations (CROs), pharma companies and technology providers collaborate and converge on key issues such as performance expectations, definitions and context-of-use. The FDA has an ambitious roadmap to make animal studies the exception rather than the norm within the next three to five years, and is already working to highlight acceptable NAMs and streamline nonclinical programs. It is also collaborating with agencies such as the National Institutes of Health (NIH) to establish pathways for validating NAMs.
This increasing alignment will lead to greater standardization. We are now entering a phase of agreeing on shared benchmarks, reproducibility criteria and clearer regulatory guidance on how NAM data should be presented in submissions. Once those norms are in place, adoption becomes less about advocacy and more about workflow integration.
Why is the presence of a functional vascular network so critical for accurately predicting how a drug will behave in a human patient?
GC: A functional vascular network is crucial to determining the efficacy of a drug, particularly drugs that are highly complex and human-specific. If preclinical testing is carried out using methodologies that do not adequately predict human outcomes, the risk of misleading results increases.
For example, in humans, most drugs do not simply “diffuse into tissue.” Instead, they are delivered. This means that exposure is determined by several factors, including perfusion, vascular permeability and endothelial barrier function. If a model lacks functional vasculature, it cannot capture real-world penetration limits because the biology is too dominated by diffusion and exposure is non-physiologic.
Furthermore, our blood vessels are an intricate part of our complex biology, not just “plumbing.” For many therapeutics, the endothelium is the first meaningful interface that determines whether a drug reaches its target and what its off-target liabilities may be. Our endothelial cells actively regulate inflammation, leakage, immune trafficking, and cross-talk with surrounding tissue. All of these elements are fundamental to obtaining physiologically relevant, reproducible results that reflect how a drug would behave in humans.
Drug development is increasingly about failing fast and failing early to save resources. How do vascularized 3D human tissue models help developers prioritize their most promising programs with greater confidence?
GC: One of the most compelling benefits of vascularized 3D tissue models is that they enable teams to test a broader range of considerations in a single experiment. Instead of simply investigating whether a drug effectively binds the target, you can explore whether it arrives in the right place, at the right concentration, for the right duration and whether it generates the intended response. By testing both delivery and biology, you can measure functional endpoints and mechanistic readouts, providing the opportunity to explore the full biological story.
This improves program prioritization in two important ways. Firstly, it allows teams to rank candidates based on true tissue exposure and target engagement, rather than simply potency in an oversimplified system. Secondly, it means you can uncover liabilities such as poor penetration, vascular disruption, inflammatory activation or lack of efficacy at an earlier stage. This allows you to make decisions faster, with fewer expensive surprises down the line.
Late-stage failure is one of the most significant risks in this industry. In what ways can human-relevant models help identify safety or efficacy issues before a drug ever enters a first-in-human trial?
GC: Late-stage failure is often caused by one of two major factors. Either the drug does not work in humans in precisely the way it does in animals, or an adverse effect emerges that was not predicted in testing.
Human-relevant models respond to these issues by enabling teams to anchor decisions in human biology earlier. This is particularly vital in immune signaling, receptor pharmacology and some vascular liabilities, where differences between species can prove problematic in testing.
These models can identify mismatches at an early stage. If you discover that you cannot achieve the penetration necessary for drug efficacy, or you encounter safety signals under human-relevant exposure conditions, you can course-correct at that point, rather than discovering these issues after investing in clinical readiness and execution.
Looking five years ahead, how do you expect the “standard” non-clinical testing process to evolve, and what role will bio-fabricated tissues play in that new ecosystem?
GC: In five years’ time, I expect we will see a standard that is more modular and evidence-based. Targeted animal work may still be employed where informative, but it will be supported by a larger, more credible NAM layer that is mechanistic, human-relevant and increasingly standardized. The FDA has already signalled its openness to this shift.
Biofabricated, vascularized tissues have the potential to assist in several important ways. They can help teams to deliver evidence on delivery and penetration risk, target engagement, human-relevant efficacy and early-safety derisking.
As performance expectations and context-of-use become clearer, tissues such as these will develop from “supporting data” into “key decision data.” At that point, they will become essential to developing the next generation.
Newsletters
Receive the latest pharmaceutical news, personalities, education, and career development – weekly to your inbox.
