In silico modeling is emerging as an important partial alternative to animal testing. 

Animals, in particular genetically-engineered mice, have long been used to test medicines before they go into humans. Regulators require animal toxicology data, and in vivo studies are also used to inform clinical development

Yet animal models are very poor predictors of drugs’ safety and efficacy in people. Size, behavior and diet are among the most glaring – but far from the only – reasons for that: a mouse brain, heart or liver is just a tiny fraction of a human brain, heart or liver.  Widespread animal use also raises ethical concerns. Meanwhile, alternative tools for predicting drugs’ toxicity and efficacy are gaining traction. 

That’s why regulators and policy-makers are calling for change. In late September, the US Senate passed a bill to remove the requirement for animal testing for new drugs. It doesn’t ban animal tests, but it opens the door to alternatives.

FDA’s Alternative Methods group and the European Medicines Agency are already investigating non-animal tools to help predict the effects of new drugs, with the goal of reducing, refining or replacing some animal testing. In silico modeling – computer-based simulations of physiological systems – is one such tool. Others are improved cell-based testing methods such as organoids (stem-cell derived cell clusters that behave like particular tissues) and organs-on-chips (chips lined with living cells and tiny fluid channels, designed to capture organ-level physiology).  

These efforts matter not just for animal welfare, but in order to improve the safety and efficiency of drug R&D. Animal use is a very early filter in drug discovery. Its dismal predictive power means that some toxic drugs slip through into human trials, and that some effective treatments are pre-emptively culled.

In silico testing and other emerging alternatives are not just cheaper, faster and kinder than using caged animals for pre-clinical testing. They are  also in some cases likely to be better than animal models at predicting efficacy and toxicity in humans, particularly given the rise of precision medicine.

Many new therapies are tailored to particular human genetic or molecular sequences and may show little or no activity in animals. Moreover, animals are often bred or cloned to reduce variability, yet precision medicine is all about understanding and exploiting those differences. Examining and monitoring variability is precisely where in silico models can be helpful.

In silico and other new R&D methods are becoming more sophisticated and powerful. The disease and treatment models developed at nova, for example, are built, tested and refined using an ever-increasing variety and volume of data and insights. They can integrate information from multiple sources, including different animal species – rather than the single one provided by any given animal model. Organ chips are also becoming more intricate and better at mimicking physiological systems. They’re also being linked up to simulate whole-body responses – even the most accurate heart-on-a-chip can’t flag up what effects a new molecule might have in the kidneys or brain. 

Ultimately, in silico and other technologies will do more than mimic animal studies. They will bring greater precision to R&D, and may even become central to the development process. One day these models may even determine which (minimal) real-life animal and human studies are necessary, rather than simply being fed with existing animal and human data.

Yet these systems won’t replace animals until and unless they are convincingly shown to be better at predicting drugs’ behavior. For now, a demonstrably predictive in silico (virtual) patient or “body-on-a-chip” remains elusive. 

So animal testing will remain a core part of drug R&D for the foreseeable future. But as medicines become more complex – including antibody, cellular and gene-based therapies – in silico simulations and other predictive methods will take on greater importance. They will gradually replace some animal models, arming drug developers with insights that allow them to move confidently into more targeted and  less costly development studies.

Accelerating uptake of these alternatives is an ethical and economic imperative.