Lost In Translation
In drug development, several checkpoints must be successfully overcome before a drug can be used in humans. In the initial stages, basic scientific research is done to identify the mechanisms of disease and hypothesize potential drug candidates. Once drug candidates are selected, they are tested in vitro (in isolated cell systems). Drugs must then be filtered by two criteria: (1) efficacy (whether the drug works for its intended purpose) and (2) safety. This is done through preclinical trials, which are generally run in vivo (in animal models). If a drug proves to be effective and safe in animals during preclinical trials, it graduates to clinical testing in humans. Particularly in cancer drug development, the predictive value of animal models during preclinical trials is being called into question. If animals may not be the most predictive modalities for testing cancer drug safety and efficacy, is it worth subjecting animals to experimentation?
The paper “Lost in translation: animal models and clinical trials in cancer treatment” investigates the limitations of animal models, which are used to “translate” findings from a laboratory setting to clinical trials. A drug tested in animal models is successfully translated to clinical trials less than 8% of the time. Poor methodology may be responsible for the 92% failure rate. However, the inadequacy of animal models to accurately mimic human cancers also contributes greatly to the inability of drugs to graduate from success in animal models to success in clinical trials with humans. Rodents are the platform of choice for preclinical cancer research. Experimental tumors are grown in rodents and the drug is administered to test whether it improves the rodent’s outcome. However, human genetics, immunology, and cellular environments are very different from those of rodents. Even if rodents exhibit parallel pathogenesis of cancer, they are unlikely to replicate the exact physiological changes that humans experience when they have a tumor.
There are dire consequences to choosing an unrepresentative rodent model for testing cancer drugs that are intended for humans. The authors of this paper examined three notable instances when clinical cancer trials failed even when the drugs had succeeded in tests with animal models. One example is the TGN1412 drug, which was tested in a variety of animal platforms with great success at doses one hundred times higher than that administered to humans. However, when the drug was finally introduced to humans during clinical trials, it caused unprecedented and catastrophic organ failure in patients.
A class of drugs that is gaining traction with cancer researchers is therapeutic cancer vaccines. These stimulate the immune system to mobilize against tumors the same way a flu vaccine stimulates the immune system to recognize and marshal an attack against the flu virus if it were to infect the body. However, the authors warn that translating the success of cancer vaccines from animals to humans may not be successful. This is because the human immune system contains a complex assortment of immunological checkpoints and immunosuppressive elements that may prevent the immune system from responding to the vaccine the way it does in rodents. Considering the different physiology of animal models and their limited predictive value will be essential when bringing these drugs from the preclinical to the clinical testing phase.
Animal models are still the primary means of testing cancer drugs. However, if the safety of both the animal models and later human users of the drugs is compromised by animal experimentation, it is worth considering alternatives to animals for preclinical testing. A potential alternative discussed by the authors is the Phase 0 approach to preclinical testing. In phase 0 studies, drugs can be tested in extremely small doses on humans to determine how the drug is distributed and metabolized in the body, whether it targets the intended mechanisms, and whether it successfully overcomes pharmacological barriers such as absorption or penetration into target organs. This is a good strategy for testing the efficacy of a drug in the same system in which it will later be used. Aside from phase 0 studies, the authors also briefly mentioned other strategies for preclinical research including computer simulations, epidemiological studies, autopsies, and “organ-on-a-chip” studies.
Overall, cancer manifests differently in animals than in humans, which means animal testing for cancer drugs is not predictive of the drug’s efficacy and safety in humans. Cancer is a very active area of research, and if studies like this are emerging that demonstrate the potentially dangerous limitations of animal models, we need to take them seriously. For the sake of both animals and humans, we need to curb the use of animal models in preclinical research and commit to making better use of alternatives.