Dr. Isaac Bentwich- Founder and CEO, Quris-AI.
Scientific experimentation on live animals has always been a faulty and highly controversial practice, yet it has played a pivotal role in pre-clinical drug testing methods for more than 80 years. Recent regulatory moves are changing the game by allowing researchers to embrace new non-animal testing models, unleashing a wave of innovation fueled by artificial intelligence (AI) that will transform drug development and benefit society in the years ahead.
Eighty-four years of Apples-to-Oranges Animal Testing
In the fall of 1937, more than 100 people died after ingesting an improperly prepared antibiotic known as Elixir Sulfanilamid. Congress quickly responded to the tragedy by passing the Federal Food, Drug, and Cosmetic Act of 1938, which mandated animal testing for every drug.
Decades later, novel drugs still go through a time- and cost-intensive pre-clinical process involving animal testing to measure quality, efficacy, and safety for humans. Industry bodies such as the International Council for Harmonization of Technical Requirements for Pharmaceuticals provide the frameworks for these tests, based on feedback from regulatory authorities, scientific experts, and existing research.
These frameworks rely on animal models to try and predict how a drug will work across the human body’s various facets: how it will be metabolized by the liver, excreted by the kidney, etc. Yet a large body of research demonstrates that a drug’s effect on humans cannot be accurately predicted by animal studies. There are simply too many critical differences in the organ systems, as well as the composition, expression, and catalytic activities of drug-metabolizing enzymes, to effectively extrapolate data from animal models to humans. Various statistics and tools are used to try to bridge the animal-human gap and accurately determine drug safety. Some predictability approaches, for instance, compare variations of animal species, such as rats, dogs, monkeys, and mice, to predict compounds’ metabolic behavior in humans. Yet, a staggering 92% of drug candidates that successfully pass animal testing later fail in clinical trials in humans! This means that animal testing is a terrible predictor for whether a drug candidate will work safely and efficaciously in the human body, a predictor that is in fact wrong nine out of ten times.
Furthermore, we humans are very different from each other. Our bodies are incredibly nuanced – shaped by everything from a unique genetic makeup to weight, hormones, and metabolic behavior, to lifestyle choices and environmental influences.
Daunting Pre-Clinical Requirements Drive Misconduct and Animal-Welfare Concerns
Documented cases of research misconduct reveal just how far some will go to circumvent pre-clinical hurdles and advance to human trials. For instance, this year federal investigators charged a former University of Pennsylvania professor for manipulating results of federally funded studies involving grisly animal-testing practices for a novel drug designed to treat human brain injuries. Some of the researcher’s reported actions – including relabeling results from old studies and falsifying critical data, among other things – made the drug appear more effective than it was. Not only does this extreme example illustrate the breakdown between animal and human testing, but it also highlights the major ethical dilemma associated with the practice.
In 1959, William Russell and Rex Burch from the Universities Federation for Animal Welfare (UFAW) tackled this thorny issue by establishing a set of principles for humane animal testing called the 3Rs: replacement, reduction, and refinement. Decades later, researchers around the world still adhere to these principles, toeing the line between minimizing animal suffering and advancing the human condition. Yet reports of excessively inhumane practices continue to surface with concerning frequency. A notable example is Elon Musk’s medical device company, Neuralink, which tests brain implants on primates and is currently under federal investigation for potential animal welfare violations.
The Deeply Flawed Drug Discovery and Development System Demands an Overhaul
The traditional pre-clinical drug development process is inherently flawed, taking critical time away from drug manufacture and keeping life-changing drugs out of reach from those who need them most. The staggering 90 percent failure rate for novel drugs in clinical trials underscores this truth. For every breakthrough drug to hit the market, there are nine others that die somewhere along the discovery and development lifecycle – often before they even make it to clinical trials.
Drug candidates that successfully reach the pre-market evaluation step face a host of new challenges – chiefly, a widespread lack of patient diversity. Racial/ethnic minorities are 1.5 to 2.0 times more likely than whites to have chronic diseases. Yet a global study of 150,000 patients in 29 countries at five different periods between 1997 to 2014 revealed that nearly 90% of trial participants are white. Unfortunately, this is not a new phenomenon, nor has it improved over time. Inadequate representation can significantly impact clinical trial data, translating into sub-par treatment strategies, increased safety issues, and deteriorated outcomes for both patients and researchers.
Pharmaceutical R&D also comes at an astronomical cost, whether a novel drug finally reaches patients or not. A 2023 Deloitte study of 20 pharmaceutical companies shows that the estimated average cost of developing a drug, including the cost of failure, increased from $2 billion in 2021 to $2.3 billion in 2022. Though the analyzed companies spent a collective $139 billion on R&D in 2022, projected returns on investment in pharmaceutical R&D fell to 1.2% – the lowest ROI observed in 13 years. Outmoded pre-clinical practices, rampant patient underrepresentation, time- and cost-intensive development cycles and decreased ROI are driving pharmaceutical companies to challenge conventional thinking and start taking the steps needed to rethink the drug development process to make it more efficient.
AI-Fueled Innovation Introduces Powerful Animal-Testing Alternatives
The FDA Modernization Act 2.0 marked a huge step forward in the historically sluggish, process-heavy realm of pharmaceutical regulatory reform. Signed into law on December 29, 2022, the act refuted the Federal Food, Drug, and Cosmetics Act of 1938 and opened the door to animal-testing alternatives. On the heels of this landmark act came the Food and Drug Administration’s Innovative Science and Technology Approaches for New Drugs (ISTAND), a pilot program that aids in evaluating novel drug development tools (DDTs) for specific regulatory uses.
Other countries have taken similar steps to encourage alternative models, and in some cases moving towards transitioning away from animal testing completely in certain situations. For instance, this year UK banned animal testing for ingredients used exclusively in cosmetics and Australia rolled out a detailed plan for shifting to non-animal testing models. These recent regulatory moves propel the industry forward, ending decades of unchallenged legislation that hampers drug development and relies on outdated processes that no longer serve patients’ (or researchers’) needs.
While novel DDTs span the technological spectrum, AI-based innovations are driving unprecedented disruption and delivering new opportunities in non-animal science. After achieving the first critical step, learning how to effectively develop organ-on-chip – miniaturized 3D versions of human organs created either from primary cells or from stem cells – to model diseases and predict human response to drugs outside of the body, researchers quickly adopted the technology. While the initial waves of experimentation could be considered relatively simple use cases such as replicating bone marrow, at the time they were groundbreaking.
As the rate of innovation accelerated, researchers began developing more extensive applications and shifted focus to creating patient[1]specific stem cells and organoids en masse. Achieving this goal allowed them to model more complex diseases and test new treatments in more reliable and affordable ways. However, there were still too many limitations.
The next step in this scientific journey was integrating advanced AI together with organ-on-chip biology and doing so with increasingly interconnected organ-on-chip systems, with integrated nano-sensing capabilities. The resulting AI-enabled patient-on-a-chip technology is now taking previous platforms to the next level by simulating actual patients instead of isolated organs. With this capability, scientists can begin to better understand how a drug impacts the entire human body. This now opens the door to more accurately gauging drug toxicity and efficacy at the individual level, while connecting the dots across large and diverse patient profile groups to optimize study design and costs, eliminate non-viable drug candidates earlier in the pre-clinical phase, and, ultimately, get safer drugs to patients faster.
This new phase is extremely powerful, because it unleashes, and indeed harnesses, the power of AI and the unique synergy between AI and patient-on-chip biology. As patient-on-a-chip platforms tapping vast stem cell systems continue to evolve, becoming more sophisticated, automated, and tightly integrated with AI – they allow for the generation of a massive amount of data that better recapitulates the interactions of known drugs with these miniaturized human tissues, and then this highly predictive data is used to train the machine-learning platform, delivering unmatched drug safety and efficacy prediction capabilities. With these breakthroughs, drug developers can better vet novel compounds across a diverse patient population before ever introducing them to a human patient. It is hoped that these powerful capabilities will allow R&D teams to significantly improve and scale drug testing and will make the investigational new drug (IND) application process dramatically more efficient over time.
As AI continues to transform more areas of drug research and development, the pace of progress is increasing in lockstep. AI is rapidly becoming an integral part of every step in the drug discovery and development process – putting patient efficacy and wellness fully back into focus. AI is already very successful in improving target discovery, molecule design, drug-target fit, mode-of-action analysis, and many other aspects of drug discovery and development. AI can also ease resource-intensive processes for recruiting and retaining trial patients to overcome pervasive enrollment challenges and keep trial timelines on track. Finally, throughout trials, AI can enhance medical wearables, sensors, and/or video monitoring solutions to automate laborious tasks and improve patient and clinician experiences.
While conventional drug development practices will not disappear overnight, AI advancements offer a glimpse into a near future where novel drug data and discovery and development processes are optimized at every phase of the lifecycle. And thanks in large part to more modern legislation, AI-powered platforms will gradually replace and greatly reduce the reliance on antiquated, ineffective animal testing.
Publication Details
This article appeared in Pharmaceutical Outsourcing:
Vol. 4, No. 24
Oct/Nov/Dec 2023
Pages: 31-33
Subscribe to our e-Newsletters
Stay up to date with the latest news, articles, and events. Plus, get special
offers from Pharmaceutical Outsourcing – all delivered right to your inbox! Sign up now!