Animal-Free Methods for Drug Safety Testing

Animal-free testing methods are being increasingly explored and implemented as an alternative to animal models, in efforts to provide a more accurate prediction of a drug candidate’s efficacy and safety. InSphero recently hosted a virtual roundtable discussion on the topic of animal-free testing. Expert panelists included Dr Stefan Platz, senior vice president of clinical pharmacology and safety sciences within AstraZeneca's biopharmaceuticals R&D unit, Prof. Thomas Hartung, professor of evidence-based toxicology and director of the Center for Alternatives to Animal Testing at Johns Hopkins Bloomberg School of Public Health and Prof. Armin Wolf, chief scientific officer at InSphero.       

Technology Networks had the pleasure of speaking with Armin Wolf, to learn more about the topics discussed, including the limitations related to using animal disease models in preclinical drug development, advances in in vitro three-dimensional (3D) models and their use as tools to assess drug toxicity. 

Laura Lansdowne (LL): There continues to be much discussion among the scientific community when it comes to ethical and scientific considerations regarding animal testing. What would you consider to be the key challenges related to the use of animals in drug safety testing studies?

Armin Wolf (AW): The key challenge with animal testing is that since the 1960s healthy animals are used to predict drug safety in diseased patients by setting safety margins by the so-called therapeutic index. Although animals can be regarded as “complete systems” responding to direct and indirect drug effects and thus may serve as a surrogate for healthy humans, they cannot fully serve as surrogates for patients. Healthy humans and patients may respond differently to the same drug. Patients are more fragile than healthy subjects and thus are more susceptible to the toxicity of a test item at lower dosages.

Another key challenge is that the immune system of animals and human is different. Since many toxic reactions in humans are mediated by the immune system, these reactions cannot be predicted by animals. One example is drug-induced liver injury (DILI) which can cause fatal liver damage in humans.

LL: In the panel discussion, it was noted that approximately 90% of investigational new drug candidates that were progressed into clinical phases failed. Do you believe that the use of animal models is a contributing factor to this high drug attrition rate, are there other factors that should be considered?

AW: Correct – that was in a recent paper published by the European Medicines Agency (EMA). They were looking at the failure rates of all submitted investigational new drug (IND) candidates during a recent 10-year period. It turned out that 90% of clinically tested candidates failed because of a lack of efficacy, safety issues and other reasons. In the discovery and development phases, preclinical animal species are used as well as cell cultures of various origins. This suggests that animal testing contributes to this high failure rate but also the low predictivity of 2D cell cultures might be involved and play a role for the failure.

LL: In the panel discussion Thomas Hartung noted that creating disease models in animals and then testing toxicity of the substance is not a good approach, when attempting to improving predictivity of drug safety for patients. Could you comment on this and perhaps talk about the role advanced in vitro models could play?

AW: Animal disease models demonstrated strong limitations in terms of translatability to human and particularly to patients. There is a definite need for more human-specific models. 3D liver microtissue (= spheroids) in vitro cultures are an example of these models showing clear advantages in comparison to 2D liver cell cultures. It was demonstrated in a recent study evaluating 108 DILI-annotated clinically tested drugs for its liver cytotoxicity. It turned out that 3D models demonstrated higher sensitivity and specificity compared to 2D models in terms of DILI prediction.

Advanced therapeutic modalities are amplifying the need for better, more sophisticated predictive in vitro human 3D microtissue models. In vitro models with immune competence would be such an advanced model for the testing of biologics which are different from the testing of small molecules. Normally, biologics are tested in vivo in monkeys which are more like humans compared to rodents or non-rodents. However, in some cases, for example, when testing bispecific antibodies, monkeys cannot be used, due to the high human specificity of these applied molecules.

Other advanced models would be patient-centric models. These are in vitro models including aspects of the underlying disease and which can enhance the toxicity of a test item.

LL: In recent years there has been tremendous effort focused towards advancing in vitro 3D models to study drug toxicity – are sex-specific, age and ethnicity differences considered at this stage of development, and if so, how does this help to identify potential differences in drug response among these populations?

AW: This is an important question since we know that drug toxicity may vary between different ethnic groups. It may also vary because of the underlying diseases, body mass index, age and gender which are risk factors, particularly for DILI. These risk factors increase the toxic potential of drugs and thus increase the risk for fatal liver failure in patients. For the testing of risk factors, animals cannot be used.

3D microtissues have demonstrated superior performances in drug safety testing compared to 2D cell cultures in general. To eliminate specific differences between donors, InSphero engineered 3D microtissues from 10 pooled donors, 5 females and 5 males. This is a very good representation of the average Caucasian population. Similar types of microtissues can also be engineered for other ethnic groups with underlying specific polymorphisms like Asian or Indian. Testing new drugs ahead of clinical trials with use of “clinical studies on chip” would allow better risk assessments for the patients.

Improved risk assessment for patients can also be obtained from liver tissues from patients with a specific underlying disease. For instance, more than 20% of people in Western countries are suffering from steatosis and from non-alcoholic steatohepatitis (NASH). Anti-steatotic or anti-NASH drugs tested in healthy animals could strongly deviate from the toxic effects in patients. Toxicity testing in 3D microtissues in the respective in vitro disease model would allow a much better assessment for this fragile patient population.

LL: In the panel discussion Stephan Platz commented that “we are in a precompetitive phase where I believe the safety of patients is dominating above any kind of commercial interest”. Could you comment on this further?

AW: I completely agree. That’s why several public-funded consortia working on these topics. There are also areas that are not public-funded and can only be explored by joint forces. I can provide an example for such a privately initiated precompetitive consortium in the field of translational toxicology. InSphero recently established a consortium with several pharmaceutical partners (EMD Serono, Genentech, Pfizer and Sanofi), and we are exploring liver toxicity in preclinical animal species for in vitro DILI validation. By testing drug candidates, which are well characterized in vivo in preclinical species and in vitro in 3D microtissues from the respective animal species, we can potentially create a link between the in vivo and in vitro effects. These activities will help provide more insight into test systems and will set the ground for increased regulatory acceptance and the formation of alternative animal tests. Once better test systems are used for safety assessment, patients will benefit from safer drugs.

LL: Do you envision a day when animals are not used in any way in drug discovery and development? Is that possible?

AW: Yes, but it will take many years – but the good news is that we can already reduce animal testing substantially by making good use of available in vitro models. There is also a lot of progress in the field of body-on-a-chip and microfluidics. In so called “proof of concept” studies, we are learning how to bring different organ systems together which are interacting in a microfluidic, physiological way. One of the next hurdles to overcome is the automation of these processes in a high throughput setting. This will have an additional impact on the drug discovery process. InSphero is well advanced in this respect, providing customized platforms to enable the production of robust and reproducible data for pragmatic applications which can easily be implemented into industrial-level workflows. Having concrete cases whereby drugs are registered without animal testing will increase interest from the pharmaceutical industry and motivate the scientific community. All these developments will inspire additional advances, to move beyond animal testing.

Armin Wolf, PhD., was speaking with Laura Lansdowne, Managing Editor for Technology Networks.