Back to the Future: Predictive Toxicology
Safety Assessment
Alan Hoberman

Back to the Future: Predictive Toxicology

As we reduce our reliance on animal models, we need to look out for those unknown unknowns. Our final installment of the series.  

We’ve come a long way from the paradigm that risk assessment is simply based on animal safety studies. The proof is in the regulatory pudding. Authorities are increasingly supportive of integrated toxicology or weight-of-evidence approaches that employ a broad range of different models, and in some scientifically justified cases no animal studies at all. 

This week, Eureka examined two ways that predictive toxicology is being applied in testing drugs and chemicals. One story looked at how we can use human stem cells to de-risk T-cell therapies during the earliest stages of drug discovery, another looked at how Big Data can reduce animal testing. Moreover, the FDA and EPA is also receptive to using predictive toxicology in non-animal studies.

Which begs the questions, when will we be ready to use predictive toxicology without relying on animals? Many scientists argue that we will change our current testing paradigms once we have “complete” understanding of molecular and cellular biology. This knowledge along with an understanding of how a chemical, biologic, or gene therapy interacts with human cells should, theoretically, open the door to studies that rely entirely on non-animal methods, such as in vitro, in silico or in chemico tests.

However, I would caution that even when we have “complete” understanding of basic biology, we will also have to consider, what in the early 1960’s, was noted by Clarke Fraser and his group at McGill University about normal development of the fetus. Fraser’s lab found that injecting pregnant mice with cortisone caused the offspring to be born with a cleft palate, though the frequency of defects was higher in some strains than others. Fraser eventually determined that “an embryo’s genetically determined, normal, pattern of development could influence its susceptibility to a teratogen.”

Weighing Safety

Clearly, we will need to be cautious about how we test our drugs and chemicals. The public’s safety must be paramount. What we can do is to continue to improve our predictive toxicology tools. When the benefit of a new therapy outweighs the risk, such as the emerging CAR T-cell and T-cell receptor therapies for terminally ill patients with aggressive cancers, predictive toxicology using non-animal models is an effective way of weeding out the bad candidates while accelerating the good. On the other hand, it might not be so useful for vetting a new gene therapy that might cure sickle cell anaemia, but result in an increase in tumors years later in the cured patients.

This reminds me of a well-cited quote from our former US Secretary of Defense Donald Rumsfeld: “Reports that say that something hasn't happened are always interesting to me, because as we know, there are known knowns; there are things we know we know. We also know there are known unknowns; that is to say we know there are some things we do not know. But there are also unknown unknowns—the ones we don't know we don't know. And if one looks throughout the history of our country and other free countries, it is the latter category that tend to be the difficult ones.”

Rumsfeld was talking about the search in Iraq for weapons of mass destruction, but we can also apply his reasoning to toxicology, which at its core is all about determining risk.

The integrated approach

In conclusion, there are plenty of reasons why predictive toxicology should be included in the regulatory approval process for new drugs and environmental chemicals. We have more than 80,000 potentially hazardous chemicals that need to be evaluated, and not enough animal models that mimic exposure to chemicals in humans. The animal models we use to develop new drugs are imperfect, while our ability to use human-derived stem cells is advancing.

But does this mean that we will soon be running most preclinical safety programs without animals? No, because we still need to sort out those unknown unknowns. Fortunately, we are moving down the right path. The next steps will be a combination of in vitro and in vivo approaches or an integrative approach that will allow us to ensure safety, accounting for what we know and what we still have to learn.

You can find our entire series on trends in predictive toxicology here.