A new era for drug development technology

Eroom’s Law - the current trend in drug discovery

It has always intrigued me how Moore’s Law is so often discussed as though it is an immutable principle like the laws of physics. Moore’s Law, which suggests that the number of transistors in an integrated circuit doubles approximately every two years, to my mind appears more as a self-fulfilling prophecy. It seems to be more fuelled by the substantial amount of time and money invested in transistor development, rather than an inherent physical law. It aligns more with principles found in social sciences, such as economic laws like supply and demand, which demonstrate the robust yet adaptable nature of human behaviour.

It is with this mindset that I encountered the term “Eroom’s Law,” – effectively Moore’s Law backwards. Eroom’s Law observes that the number of new drugs approved per billion dollars spent on research and development (R&D) decreases by half approximately every nine years. This trend is attributed to factors including the escalating standards set by blockbuster pharmaceuticals, the increasingly risk-averse nature of regulatory bodies leading to increasing R&D costs, a propensity to throw additional resources at challenges and an overestimation of the productivity of traditional research methods.

While Eroom’s Law has held true in drug discovery for decades, it should not be perceived as an inevitability. Technological advancements in in vitro, ex vivo and in silico models now have the potential to drive down costs, increase accuracy and improve the speed of drug development. Technology advancements are also coinciding with a regulatory drive to minimise preclinical animal testing, which is helping to open the way for these alternative approaches.

The drive to minimise preclinical animal testing

The key regulating bodies are currently trending towards a desire to minimise the use of animals in preclinical studies. For example, the latest passing of the Food and Drug Administration’s (FDA) Modernization Act 2.0 removes the mandate for animal testing to assess the safety and efficacy of a drug. Similarly, BS EN ISO 10993-1:2020, the standard concerned with biocompatibility of medical devices, does not mandate animal testing. Instead, it states that in vivo animal testing is only justified when either existing scientific data or in vitro and ex vivo studies fail to “yield equally relevant” information to support a comprehensive assessment of the safety of a medical device.

A mere aspiration to reduce or eliminate animal testing is however insufficient; what is required are credible and validated alternatives. It is widely acknowledged that in vivo (inside a living body/organism) animal models exhibit limited predictive accuracy of the efficacy of medical interventions in humans. It is this issue that inspired Greg Tietjen to co-found Revalia Bio, a company working to change the landscape of pharmaceutical development. He recalled seeing “brilliant nanomedicine technologies that were engineered in a beautiful, artful way leading to hundreds of millions raised in funding, yet failed in clinical trials”. This is an incredibly common occurrence – 90% of drugs that make it to clinical trials fail to reach approval. Tietjen suggests this is likely due to a simple reason: “they were engineered in the wrong system”.

Nonetheless, it is crucial to adequately manage the risks associated with introducing novel medical technologies, prior to trialling them on living humans. To date, though imperfect, this requirement has made preclinical animal studies indispensable. However, recent technological advancements in biotechnology, computing power and artificial intelligence (AI) may offer viable alternatives to this conventional approach.

The current drug discovery technologies set to disrupt the development landscape, also referred to as new approach methodologies (NAMs), can be categorised into three groups, in vitro, ex vivo and in silico.

In vitro

Recent industry developments have begun to allow in vitro models of organs and diseases to be used more routinely in drug discovery and development. This is already providing highly valuable insights, such as predicted drug pharmacokinetics and toxicity, at a far lower cost than preclinical animal studies, and at a much earlier stage than with previous approaches.

In vitro, a term derived from Latin for ‘in glass’, is used in the scientific community to refer to the use of living cells outside of the living body/organism and in an artificial environment (e.g. a glass dish). This is typically conducted in a laboratory and referred to as ‘cell culture’. While 2D cell culture has been around for more than a century, the use of more advanced 3D cell culture techniques is relatively recent. In biology, cells are rarely in 2D – they are in complex 3D structures, involving multiple cell types. Cells take their cues from their surroundings, including topographic (the 3D extracellular microenvironmental structure), as well as other physical, mechanical, electrical and chemical cues. The more cell culture conditions can be tailored to mimic their native biological environments, the more relevant these in vitro models become.

This has been the scientific trend over several decades and in vitro models are now reaching a highly advanced stage. In vitro models take many forms, such as organoids, engineered tissues and organ-on-a-chip (OOC) technologies.

Organoids are 3D aggregates of thousands of cells, made from diverse cell types, with a size in the hundreds of micrometres (even reaching millimetre size scales). They can be tailored to effectively imitate the complexities of an organ’s in vivo physiology, resulting in ‘mini-organ’ constructs. As well as drug discovery and disease research, organoids have potential applications in therapeutic and diagnostic areas of personalised medicine.

Another category of in vitro models is those which use scaffolds with the cells in culture, referred to as engineered tissues. The field of tissue engineering uses biomaterials to mimic the native extracellular matrix (ECM) to further support cell growth and function in cell culture, allowing above-millimetre-scale living tissue models to be fabricated. This field originally evolved with the aim of producing lab-grown tissues for patient transplantation, to trigger regeneration of diseased or damaged tissues. This same technology is now also being developed for drug discovery, lab-grown meat and other applications.

Organ-on-a-chip (OOC) technologies, also known as microphysiological systems (MPS), can incorporate various in vitro formats with other technologies, particularly microfluidics, to create complex multi-channel circuits. These can further simulate the activity of entire organ systems, such as blood flow. For example, CN Bio Ltd, a UK based start-up, launched its printer sized, OOC/MPS system in 2018, the PhysioMimix®. This device range, along with the corresponding consumables and assay protocols, has been shown to accurately predict human drug responses.

Emily Richardson, a lead scientist from CN Bio, recently discussed with Team the potential impact of this emerging drug development technology: “Ultimately, we expect OOC technology to become a pivotal tool in the drug development pipeline; to better predict drug properties and corresponding responses by human tissue. By filtering out non-efficacious or unsafe drugs earlier in the pipeline, or alternatively, advancing drugs that may have otherwise been disregarded at the preclinical stage, NAMs can save pharma companies billions of dollars in drug development costs.”

The supporting industry around in vitro models is also growing rapidly, with companies such as Crown Biosciences, a Contract Research Organisation (CRO), offering in vitro model studies as part of their services. Meanwhile, Molecular Devices, a supplier of bioanalytical measurement systems for drug discovery, is developing automated cell culture instrumentation, while there has also been an increasing number of biobanks storing organoids for purchase.

While the future of in vitro modelling looks bright, this drug development technology is still in its infancy and there is a long way to go before it is utilised widely by drug developers and accepted by regulators. As Richardson pointed out: “Trust and uptake of OOC is an interesting challenge for the whole sector. Whilst many pharmaceutical companies are already using the technology and regulators have expressed their openness to accept OOC data (as shown by the FDA Modernization Act 2.0), there is a hesitancy from some drug developers to fully implement these assays into their pipelines.” One route to pave the way for the commercialisation and universal adoption is through the standardisation of these platforms.

Ex vivo

Another emerging preclinical alternative in lab-based experiments is the use of living tissue or whole organs (as opposed to cells) outside of the body, known as ex vivo. There now exist a number of perfusion systems which provide whole organs with nutrients, oxygen and waste removal, so they can survive outside of the body for several days.

One example is a particularly complex liver perfusion system Team Consulting helped develop with our client OrganOx. The metra®, seen below, is a multi-award-winning device which can preserve liver transplants for up to 24 hours and provides real-time data which the surgeons can use to assess liver viability. The system not only provides blood to the liver, it also monitors and manages the temperature, oxygenation, nutrients, bile concentration and much more, through a custom control algorithm.

Perfusion systems like the metra® device are increasingly being utilised to support organ transplantation, however they are also being used to support the research and development of new drugs and other medical technologies. In the US, although approximately 170 million people are registered to be organ donors, only 3 in every 1,000 people die in a way that allows for the deceased’s organs to be donated for transplantation. Revalia Bio, mentioned previously, has a mission to utilise the data derived from ex vivo perfused human organ experiments to accelerate and de-risk the medical device and drug development processes prior to clinical trials on living humans. It was during Greg Tietjen’s work with organ transplantations that he realised the many organs being rejected “had all of these disease’s which were the very diseases we were trying to mimic in other systems”.

The use of whole, real human organs ex vivo presents unique experimental possibilities that cannot be replicated by ‘mini-organ’ models via in vitro cell cultures. This includes the examination of a drug’s impact on macroscale physiological processes and the study of novel medical devices, such as tools or surgical instruments. However, in vitro models have the potential to achieve significantly higher throughput than ex vivo models. Conducting an ex vivo experiment involves organ collection, setup on a perfusion system, running the experiment over several days and subsequent multi-parameter analysis, not to mention the sensitivity required for the organ donation process.

Use of these ex vivo methods are far more cost-effective than clinical trials and importantly, more representative than animal studies, yet are more resource constrained than in vitro experiments. In essence, both ex vivo and in vitro approaches are valuable pre-clinical tools and the choice between them should depend on selecting the most appropriate tool to address a specific research question.

In silico

Outside of the laboratory, drugs, biomarkers, organs and even whole bodies are also being simulated digitally, with remarkable accuracy. The virtual models of these systems are often referred to as ‘digital twins’ (take a look at Team’s recent article for an in-depth discussion on how digital biomarkers are impacting healthcare). Experiments which are conducted by means of computer modelling or simulation are referred to as in silico.

Digital modelling has come a long way in a short space of time and the in silico field has emerged thanks to advances in computing power. Scientists measure computing performance in FLOPS (floating point operations per second). For comparison, humans can typically solve simple addition problems at a rate of 1 FLOPS. Or, put another way, 1 second of exascale computing power equates to every human on the planet solving addition problems continuously for 4 years.

As a result, exascale computing has unlocked the ability to accurately model systems and organs down to the atomic scale. ELEM Biotech SL, a spin-out company from the Barcelona Supercomputing Centre, is one notable example. It has developed cardiovascular and respiratory systems which show how organs work in different healthy and diseased states, as well as predicting how effective treatments will be at restoring vital functions.

The in silico field is not only investigating the performance of new drug candidates either. It is also designing them, from scratch, through the use of artificial intelligence (AI). InSilico Medicine made headlines in 2022 when it became the first company to reach Phase I clinical trials with a novel drug for treating idiopathic pulmonary fibrosis, in which both the novel drug target and the novel molecule were discovered using the company’s proprietary AI platforms. It took them just 18 months from starting the project to having both the novel drug target and candidate identified, with a further 12 months to commence Phase I clinical trials. This 30-month period cost just £2.6m, a miniscule figure compared to the billions of pounds and decades it often takes for traditional pharmaceutical R&D to reach a similar stage.

However, a significant number of drug candidates encounter setbacks in Phase II and III clinical trials, meaning the ultimate challenge for AI-developed drugs is still to be met. As such, the integration of AI into pharmaceutical pipelines as standard practice may not be imminent. Despite this, the inevitability of AI’s capacity to assimilate knowledge and evolve suggests it is only a matter of time until we see widespread adoption.

Recent research suggests a significant surge in investments for AI-driven drug discovery technology companies, which reached $24.6 billion in 2022 – triple the amount from just four years ago. Notably, major pharmaceutical companies are the ones taking the lead in this increased investment, engaging in acquisitions and signing licensing deals. This indicates a clear strategic focus from these traditional players on the potential of AI in drug discovery.

The future outlook for drug discovery

In vitro models, including organoids, engineered tissues and organ-on-a-chip technologies, have benefitted from decades of advances in cell bioengineering, biomaterials engineering and microfluidics, to offer complex and relevant insights into human physiology at a fraction of the cost of traditional animal studies. Research estimates that if organ-on-a-chip platforms were to be fully integrated into pharmaceutical workflows, R&D costs would fall by 26%, amounting to $49bn a year.

Ex vivo methodologies, facilitated through perfusion systems such as OrganOx, which leverages advanced sensing and control technologies, have the capability to repurpose donated organs initially deemed unsuitable for transplantation. This process extends the viability of these organs outside of the body, thereby contributing a highly valuable resource to accelerate research and development for new drugs and medical technologies.

The rise of in silico modelling, powered by exascale computing and artificial intelligence, is transforming drug discovery and design. Companies like InSilico Medicine have demonstrated the ability to fast-track drug development at a fraction of the time and cost traditionally incurred.

These are not merely fringe endeavours, undertaken by a handful of small spin-outs from Universities either. There is a significant and growing swell of interest and investment from the big players. Collaborative publications between commercial organ-on-a-chip developers and industry partners have doubled from four years ago and are now numbering in the hundreds. Roche and AstraZeneca are emerging as global leaders working to validate organ-on-a-chip models, with 13 and 8 collaborative publications, respectively. Others, including Sanofi, Merck, GSK, Pfizer, J&J and many more have also been active, each co-authoring 2 or more collaborative publications in this field.

Similarly, traditional pharmaceutical companies are eager not to miss the opportunities presented by the innovators in AI-driven drug discovery. Companies such as J&J, GSK, AstraZeneca, Novartis, Pfizer, Sanofi and Eli Lilly, among others, have made substantial investments in AI for drug discovery. The private equity landscape for the sector has seen several companies raising hundreds of millions in venture capital, such as Beijing-based MegaRobo Technologies ($300 million Series C) and Massachusetts-based ConcertAI ($150 million Series C), along with many more.

This signifies more than just a departure from outdated methodologies; it could be signalling a paradigm shift impactful enough to break Eroom’s Law. While challenges remain, the combination of regulatory changes and technological innovation paints a promising picture for the future – where drug development technology is not only more efficient and affordable, but also minimises, or possibly eradicates, the need for animal testing. The pharmaceutical landscape is evolving and this may indeed be the beginning of a transformative era that redefines the future of drug discovery technology.