Progressive transition of toxicological testing and decision making- Can we attain complete transition?

Toxicity Testing

Even though the application of toxicology for human benefit dated back to the Ebers Papyrus -circa 1500 BC, and Lex Cornelia (the first law against poisoning)-circa 82 BC, the involvement of animal testing for human safety from regulatory perspective, budded around 1920-1940, when J. W. Trevan proposed the use of the LD50 (lethal dose-50%) to determine the lethal dose of chemicals and the Sulfanilamide disaster that triggered Copeland Bill 1938, that in turn led to the formation of United States Food & Drug Administration (US FDA). The current toxicology testing, mostly referred as traditional toxicity testing for human health involves the myriad of animals in variety of complementary tests such as acute, repeated dose, genotoxicity, carcinogenicity, reproductive, developmental, irritation/corrosion, sensitization, phototoxicity, toxicokinetics etc.

Now, the critical challenge around us is, how accurate our fellow living beings, such as experimental animals help to predict safety in us! The answer is undoubtedly, not up to the mark because among 90% of candidate molecules, that fail in drug discovery, has the majority reasons due to toxicity (20%) and inefficacy (40%), both rely on animal testing. Does, that mean the cornerstone practices and decision making of regulatory authorities in past and now, to put drugs and industrial chemicals in market for humans, based on animal testing are meaningless? The answer is “No”, because we have been always applying the best science to our benefit and we should be thankful to all sacrificed animals who saved us from retinopathies of anti-malarial compounds, several carcinogens, and many more hazards. By posing above opinions, we don’t remain non-participating, rather in this article, we tried to evaluate the testing practices and identify the solution.

Shortcomings of Animal testing

Homogenous animals used: By nature, humans have different ethnicities, different genetic backgrounds, different lifestyle, different ages, different comorbidities etc. which are quite tough to mimic in animal studies, whereas animals belong to same in-bred strains and almost siblings which are young with no co-morbidities and no scope for heterogeneity in their biological make up. So, we fail to distinguish vulnerability accurately in single in-bred animal species.

High dose exposures: The design of animal studies usually employ high dose exposures and evaluate the range of adverse effects in definitive period, minimum of 24 hours to 2 years. However, in real case, human gets exposure to chemicals at much lower doses with greater duration, than used in animal studies and it ranges from couple of hours to their life time (60--–70 years). So, we try to extract safety information in different study design.

Rodents and Non-rodents are not 70 kg Human: The Absorption, Distribution, Metabolism, and Excretion (ADME) varies from species to species. We have no single species which helps to predict comprehensive safety accurately in humans. Rodents have good predictivity for hepatotoxicity, non-rodents such as dogs and pigs have good predictivity for cardiovascular toxicity, minipigs have good predictivity for renal toxicity, monkeys are more sensitive to neurodegenerative diseases, Guinea pigs are good models for respiratory chronic obstructive pulmonary disease (COPD) prediction when compared to rodents (1) etc. So, restricting to one/two species for predicting toxicity for humans seems futile. Moreover, the fact is that no species can completely represent another species in testing. It could be between rats and mice or rodents and humans, the biology varies with inter-species.  

DNA and tissue repair mechanisms not robust in animals: Humans have blessed with robust and higher repairing process at DNA and tissue levels when compared with laboratory rodents (2). In our routine genotoxicity testing, besides in vitro, we rely on rodent genotoxicity assays to arrive a conclusion. Do you know common table salt and sugar (sucrose) are genotoxic both in vivo and in vitro tests, where common salt at 2% concentration in diet caused clastogenicity in bladder epithelium of rats and sucrose induced 1.8–2 fold cII mutant frequency in colon of Big Blue rats when administered through diet at 1-0–30% (3). However, we all know from ages we are using above both ingredients. So, relying on animal genotoxicity results, sometimes not promising.

Large animal size groups required: To have a successful marketing authorization for medicinal product and other chemicals such as pesticides etc., they need minimum 20–40 studies which involves large number of animals to characterize the safety and margin of exposures. However, sometimes toxicological studies flunk to yield meaningful interpretation due to juggle between biological significance and statistical significance that relies on group size of animals used. To yield meaningful scientific interpretation, it is often required to include sufficient group size and additionally other groups apart from treatment groups such as interim sacrifice and recovery groups (4).

Non-reproducible animal studies: Majority of animal studies failed to reproduce same results again, which rises question about the design of animal studies. Do you know, Bayer could not reproduce findings of 43 projects among 67 projects, that belongs to oncology and cardiology drug classes for women from their own in-house projects (5)

In addition to above shortcomings, the major drive for alternative animal tests can be formulated and represented as t × (E+E) =(2E)t* , where t stands for time taken for group of animal studies to yield meaningful interpretation on safety of humans and one E stands for factor of economic, where the animal testing package involves huge money, and another E stands for ethics, which is most fundamental backbone for humane science. The time taken for the animal experiments can be both independent (study duration) and dependent with regard to economical and ethical aspects because grant of large projects, procurement, and maintenance of animals require appreciable time. To illustrate, USA alone spends $16 billion annually for animal testing and NTP alone has share of $1.8 billion (6) and 100 million animals are being euthanized annually for training and testing in USA, which is alarming figure (7).

*- The formula is own individual thought. Don’t mean to deny other thoughts

Regulatory motives for alternatives to animal testing

The triggered initiative for alternatives to animal testing can be traced back to proposed 3R principle (Replace, Reduce, and Refine) by Russell and Burch in 1959 in their book “The Principles of Humane Experimental Technique” under Chapter-4.

Replacement: Prefer to use non-animal methods over animal methods, whenever it is possible to achieve same scientific objective

Reduction: Try to use fewer animals and obtain similar or more information when compared to required number of animals designated by protocol

Refine: Ameliorate the animal experimentation with respective to animal welfare by decreasing pain, suffering, and distress to animals

After the 3R principle proposal, several initiatives taken by different scientists in different corners of the World significantly to address animal welfare in testing, and in fact, some modified the original principles proposed by Russell and Burch and it is ambiguous to note “Alternatives to animal testing” address only Replacement (R) or all 3Rs (Replace, Reduce, and Refine)! (8). Regardless of that, alternatives to animal testing came into lime light globally after regulation passed by European Union, that is 7th amendment to Directive 2003/15/EC, which prohibited animal testing for cosmetic products since 2004 and cosmetic ingredients since 2009 and took an imperative call to restrict the marketing of cosmetic products and ingredients, which have been tested on animals. This pertinent call made cosmetic industries to rethink on testing strategies to launch their products and, it is appreciable step because animals need not suffer for enhancement of aesthetic appearance of humans. This might not hold true for life saving drugs. In addition, another legislation EU Directive 2010/63/EU “on the protection of animals used for scientific purposes” passed on January 2013 which fostered to develop specialized and trained research institutes all over Europe to design and validate alternatives to animal testing. Parallelly, European Chemicals Agency (ECHA) also stated clearly that animal testing should be last resort for generation of toxicity data.

On the other side of world in United States, the real motivation triggered by National Research Council (NRC) published report on “Toxicity Testing in the 21st Century: A Vision and a Strategy” in 2007, where it has suggested the path, to make use of advancements in other multi-disciplines of science for evaluation of health risks (9). It emphasized the need for reduction of animal testing and increase of human cell-based assays by application of more in-vitro and in-silico toxicology to assess toxicity pathways. 

Current stand of Regulatory authorities to phase out animal testing

US FDA is clear in regard to use of animal testing for cosmetics, where it encourages to use alternatives to animal testing on first hand before anyone decides to use animal data for justification of safety. In case, after considering available alternatives, if manufacturer decide to use animal testing, they can use. But the assurance of safety is primordial factor (10). US FDA has also identified the importance of animal testing in essential products such as drugs, biologics, and medical devices (11). Already FDA made an attempt and suggested a design to eliminate dogs in veterinary trails for an anti-parasitic drug (12). Generally, change always accompany resistance, however the responsible regulatory authority US FDA is not resistant to new ideas and still trying for humane science without compromising safety, that’s the reason they stated clearly that, they encourage alternatives and also where required, animal testing should not be ignored. Regardless of this, the dispute between Vanda Pharmaceuticals and US FDA on recommendation to carry out 9-month non-rodent dog study for neurokinin-1 receptor antagonist remains a point of interest to toxicology community (13). FDA in fact formed Toxicology Working Group with 6-part Framework to identify the emerging predictive toxicology methods and new technologies, which could be integrated and apply into regulatory safety and risk assessment. By this, FDA laying out a predictive toxicology roadmap that will have a great impact on future toxicology testing and decision making (14)

European Medicines Agency (EMA) also move in similar path, concurrently with the formation of working group on the application of 3Rs in regulatory testing of medicinal products (15), where it supports the alternatives to animal testing, but clearly stated that until enough confidence attained on new technology, animal testing is essential in some areas. EMA also published guideline on regulatory acceptance of 3Rs testing approaches, but clearly stated that they will not alone be taken a criterion for authorization of drug (16). EMA in its reflection paper has clearly identified the areas of toxicity, where alternatives still have not replaced completely animal testing with current knowledge on drugs. These areas include: qualification of impurities, safety pharmacology for central nervous system and respiratory, repeated dose/chronic toxicity, reproductive toxicity, and carcinogenicity (17).

The only slight difference between EMA and US FDA is, US FDA trying to focus on predictive toxicology methods which rely more on “replacement” and EMA trying to focus on “reduction” because EMA recommends including multiple goals in single in vivo toxicology study. For example, one should try to design chronic study for evaluating repeated dose effects, immunotoxicity, neurotoxicity, carcinogenicity, toxicokinetics, genotoxicity, and reproductive toxicity. However, by and large both global regulatory authorities are striving for effective implementation of 3Rs.

United States Environmental Protection Agency (US EPA) has passed memorandum to reduce animal testing and its related funding by 30% by 2025 and eliminate all mammalian studies by 2035. Additionally, announced $4.25 million to foster research on 3Rs by funding to 5 universities. By this, it is clear indication that, in coming 15 years, US EPA foresee complete transition could be achieved in toxicological testing for environmental chemicals to move away from traditional testing in animals (18).

In a recent publication by European Partnership for Alternative Approaches to Animal Testing (EPAA) which has a great expertise partnership of 7 federations and 36 global giant companies, clearly indicated that immediate replacement of repeated dose toxicity testing is not certain because of complexity of the knowledge, the repeated dose toxicity provide and the invaluable information derived from repeated dose toxicity studies, that helps in decision making and safety assessment. Nevertheless, they have not denied the scope of improvement in application of 3Rs in repeated dose toxicity (19).

Current Tools in the Transition

Toxicology being a borrowing science, can accommodate wide variety of technologies and strategies that facilitate to arrive a meaningful interpretation. Tried to mention here few technologies that could be fit into toxicological tool box for decision making and risk assessment.

In vitro models: These models are good choice to apply human cells for prediction of toxicity, so that at least species difference can be taken care. However, one should accept the fact that in vitro cannot represent in vivo for accurate prediction of adverse effects. In an attempt simulate in vivo conditions, scientists brought primary cell cultures, so that in vivo morphological and biochemical characteristics could be replicated. Further advancement made that is, 3D cell cultures, by adding third dimension to growing cells in suspensions, on artificial substrates, and matrices which helps to analyze cell-cell interaction, cell behavior, and signal transduction. Nevertheless, 3D cell cultures flagged issues such as accumulation of toxicity inside cell culture due to unavailability of oxygen and nutrition (22). Metabolism is a limitation, which is still not addressed in in vitro studies. To address this issue to some extent, this primary liver cell cultures derived from human could be useful than artificial addition of liver enzymes. Currently, extended versions such as 3D tissues are widely being used for regulatory testing, one among them is reconstructed human epidermis (Episkin) for skin irritation and corrosion (21).

Human stem cells: Embryonic and adult stem cells from human with the characteristics of pluripotency and multipotency can help us to understand tissue hemostasis and repair mechanisms well. They also offer advantages to study effects on cells that have capable to turn into different functional system. However, their quick change in genetics over time and the effects on culture of cells, that gets vary in differentiated cells when various functions gained over time resulted uncertainty in data with respect to real case exposure to humans.

Non-mammalian models: Models which employ Zebrafish, fruit fly, and roundworm found to be better options when we are interested to understand a greater number of signaling pathways relevant to human toxicity. Although, non-mammalians are different from mammalians in their anatomical structure, they still could throw light on physiology and pharmacology of chemicals when evaluating toxicity. As they do not contribute to kinetic similarities when compared to humans, they do not really fill the gap but offer better alternatives in terms of availability, cost, and screening. The fruit fly model already used in neurodegenerative disease for respective drug discovery (22). Zebra fish models are in fact good models for developmental toxicity because their embryos are transparent and easy to study developmental process.

High throughput testing: These are screening methods which include high automated systems which can test wide range of concentrations and wide variety of compounds. This will help to understand either toxicity pathways or mode of action based on design with number of chemicals and number of assays used. The EPA Toxcast program has applied successfully the High-throughput Screening (HTS) for identification of toxicity pathways and still in quest to discern more. Even though HTS has high potential to test, it accompanies shortcomings such as generation of big data and they are also expensive. Therefore, we need robust data analytical system in place to get toxicity conclusion. HTS could be coupled with imaging technologies to yield high content imaging technology to provide, temporal and spatial clues on operating biological process that could lead to toxicity when encountered with chemicals for cells and small mammals. The high content imaging technologies can be further explored well to understand the change in adaptive effects to adverse effect.

Omics Technologies: These include various technologies such as genomics, proteomics, transcriptomics, and metabolomics which help us to do efficient sequencing. The polymorphisms in genes can give best view of genetic toxicology. Another important application of genomics is the toxicogenomics which allows to study the adverse effects manifested when chemical has interacted at the genetic level. Proteomics play an important role in biomarkers for toxicity prediction. However, the shortcomings of proteomics such as stability of proteins and post-translational modifications create room of uncertainty with proteomics alone. Metabolomics is another omics technology which is often referred as holy grail, but it is the only technology which yields results on phenotype, because metabolomics gives final picture on toxic insult caused by toxic metabolites. However, metabolomics changes over time. The imperative point to note is any omics technology alone will not yield meaningful toxicity results, they always need to be integrated for better understanding of toxicity pathway. To illustrate, genomics alone will not give any clue to the root cause error, their integration with proteomics and metabolomics give correct idea about abnormal biological effects at organs system level. Similarly, toxicogenomics results alone force us to assume that mRNA levels reflect the final proteins levels, but when integrated with proteomics will allow us to understand the effects and mechanism. The integration of omics technology with several computational tools called as Systems biology approach is the best technology to rely on. The systems biology is making immense contribution for establishing toxicity pathways (20)

Microfluidic chips: These are new tools, where series of small chambers or organ compartments are created minutely on a chip with sample of tissue also available on chip, resulting finally the functional unit of specific organ on a chip. Instead of blood, alternate fluids circulated to mimic the in vivo organ condition and conduction. Need to mention one among them is human organs-on-chips project taken by Wyss Institute (23)

Integrated Testing strategies: This is the best strategy which can be used in the current scenario, where it uses both existing knowledge on animal experimentation and new knowledge derived from above technologies for quality human risk assessment. These are actually tiered testing approaches with the inbuilt decision points between sequential steps. This specific strategy will not allow to use blindly animals because they were used before, rather it recommends the toxicologist to move in each step with rationale and decision making. Generally, it recommends using cost and time efficient methods first and then if no conclusion arrived, then it triggers animal testing. Currently, most of the regulatory authorities adapt this strategy because it is systematic and transparent approach. The steps include: (i) identification and collation of already available data from various sources, if insufficient, then it triggers the (ii) application of computational toxicology tools such as Quantitative Structural activity relationship (QSAR) which predicts toxicological properties by structural knowledge. Generally, it is recommended to use expert based and knowledge based QSAR for better prediction. Parallelly, another method is usage of read-across (RA) approach, where one end-point is predicted from information of another substance on same end-point. ECHA already got best out of read-across approach, by which they avoided huge animal testing for generation of data in ECHA Dossiers. (iii) if no sufficient information obtained from above steps, animal testing has to be tried with sound scientific design. However, it is worthy to note, integrated testing strategies need high prospective validation before their application (20).

Toxicology testing- Transition and Decision

Toxicological science in past, present, and future has only one goal, to provide knowledge on safety, that could enable us to survive in the evolving world. Safety being primordial importance, we need to do the best in identification and addressal of problems. Our forerunners provided fundamentals of toxicological testing and set a path to move forward, we need to use together the fundamentals and our technological advancements to get best out of science. The major tribulation currently in the girth of toxicology is application of animal testing data in decision making and risk assessment. It is obvious that, to know risk of hazard in humans, we cannot subject whole humans for testing at initial stages and in large scale. So, we tried to check in alternate mammals and started reproducing them for laboratory testing. Those countless animals scarified their lives and gave information and still giving information about safety in humans. On the other hand, as we cannot trial on ourselves firstly, does not mean we can subject several rodents and non-rodents to unnecessary testing because every living organism on the earth have the best opportunity to reach and maintain their full genetic potential. Therefore, we need to have rationale based, evidence based, and strategic based toxicological testing in place with the consistent efforts to reduce, refine, and replace animals in toxicological testing. Toxicological decision making in current scenario should not tilt on one side of beam balance such as decision based on completely alternatives to animal testing or solely on animal testing.  

From the entire discussion, we clearly realized and wanted to communicate that transition is in progress and it is essential, however the transition might not be 100%. Alternatives to animal testing are really a boon to toxicology because, it allows to do humane science because after all we are humans. Safety testing in animals for cosmetics is not required because majority of humans could still survive without cosmetics. Similarly, the prioritization of chemicals for hazard assessment and screening flock of suitable drug molecules to identify lead molecule may not need large scale animal testing because prioritization and screening are not the end stages of decision making, rather they are initial stages of decision making. On the other side, in vitro cannot reflect an integrated biological system (in vivo) to understand the comprehensive biology and neither adverse nor adaptive pathways because human exposure to chemicals do not occur in isolated cells and tissues. So, the intelligent and strategic cocktail of alternatives to animal testing and animal experiments need to be employed for decision making. It is important to note that to gain confidence on alternatives to animal testing, we need robust validation and also effective communication among developers, regulators and public. In addition to these, extensive focus on Physiologically based Pharmacokinetic models (PBPK) and robust methods to derive First in Human dose (FHD) can really make us closer to real human safety. The current cynosure Artificial Intelligence (AI) has already entered into drug toxicity and safety domain, where Machine learning and Deep Learning will act as curators to manage huge data. It is interesting to note that US FDA released plans for new framework on application of AI algorithms for safe medical devices (24). We realized that, in future the management of big data generated by alternatives to animal testing would be challenging. To overcome we may need to adapt few steps such as data transparency, data availability (to avoid maximum extent of confidentiality without affecting their business), data management (by AI algorithms), and harmonization with the combination of consistent communication, collaboration and commitment.

By and large, we would like to conclude that “Biology always remains variable”, Therefore to understand the abnormal biological effects (adverse effects), we need to use set of variety tools (alternatives + animal testing) but not fixed tools.

 “It is the mark of an instructed mind to rest assured with that degree of precision that the nature of the subject admits, and not to seek exactness when only an approximation of the truth is possible”                                                                                                                                              -Aristotle

Authors Note: By posing above opinions and thoughts, my intention is not to oppose the great work happening on both the ends alternatives to animals testing and animal testing, because we feel both have great value and importance. In fact, there should be two parties who work consistently to favor the goal at their ends in a beam balance, then only we can strike the perfect balance and evaluate the toxicity.  

References

1.      Wright, J. L., Cosio, M., & Churg, A. (2008). Animal models of chronic obstructive pulmonary disease. american journal of physiology-lung cellular and molecular physiology, 295(1), L1-L15.

2.      MacRae, S. L., Croken, M. M., Calder, R. B., Aliper, A., Milholland, B., White, R. R., ... & Zhang, Z. D. (2015). DNA repair in species with extreme lifespan differences. Aging (Albany NY), 7(12), 1171.

3.      Pottenger, L. H., Bus, J. S., & Gollapudi, B. B. (2007). Genetic toxicity assessment: employing the best science for human safety evaluation part VI: when salt and sugar and vegetables are positive, how can genotoxicity data serve to inform risk assessment?. Toxicological sciences, 98(2), 327-331.

4.      EMA/CHMP- Guideline on Repeated Dose Toxicity. Published on 21 February 2008

5.      Prinz, F., Schlange, T., & Asadullah, K. (2011). Believe it or not: how much can we rely on published data on potential drug targets?. Nature reviews Drug discovery, 10(9), 712.

6.      PETA Petition: Reduction of animal testing to reduce government waste at tax payers’ expense.

7.      PETA: Experiments on Animals-Overview

8.      Tannenbaum, J., & Bennett, B. T. (2015). Russell and Burch's 3Rs then and now: the need for clarity in definition and purpose. Journal of the American Association for Laboratory Animal Science, 54(2), 120-132.

9.      National Research Council. (2007). Toxicity testing in the 21st century: a vision and a strategy. National Academies Press.

10.  Animal Testing and Cosmetics, United States Food and Drug Administration, 2017

11.  Why are animals used for testing medical products? United States Food and Drug Administration, 2019

12.  Statement by FDA Commissioner Scott Gottlieb, M.D., on efforts to reduce animal testing through a study aimed at eliminating the use of dogs in certain trials, Published on November 16, 2018.

13.  A Doggone Shame for Vanda: DC District Court Grants FDA’s Remand Motion on Dog Study/IND Clinical Hold Challenge- by Sara W. Koblitz on March 19, 2019.

14.  FDA’s Predictive Toxicology Roadmap- Published on December 2017.

15.  EMA Joint CVMP/CHMP Working group on the Application of the 3Rs in Regulatory Testing of Medical Products. Biennial report 2016/2017

16.  EMA Guideline on the principles of regulatory acceptance of 3Rs (replacement, reduction, refinement) testing approaches, 2016

17.  EMA, Reflection paper providing an overview of the current regulatory testing requirements for medicinal products for human use and opportunities for implementation of the 3Rs, 2018.

18.  Efforts to Reduce Animal Testing at EPA. Directive signed on September 10, 2019

19.  Laroche, C., Annys, E., Bender, H., Botelho, D., Botham, P., Brendler-Schwaab, S., ... & Dent, M. (2019). Finding synergies for the 3Rs–Repeated dose toxicity testing: Report from an EPAA partners' forum. Regulatory Toxicology and Pharmacology, 108, 104470.

20.  van Vliet, E. (2011). Current standing and future prospects for the technologies proposed to transform toxicity testing in the 21st century. ALTEX-Alternatives to animal experimentation, 28(1), 17-44.

21.  Alépée N, Grandidier MH, Cotovio J. Sub-categorisation of skin corrosive chemicals by the EpiSkin? reconstructed human epidermis skin corrosion test method according to UN GHS: Revision of OECD Test Guideline 431. Toxicology in Vitro. 2014 Mar 1;28(2):131-45.

22.  Bilen, J., & Bonini, N. M. (2005). Drosophila as a model for human neurodegenerative disease. Annu. Rev. Genet., 39, 153-171.

23.  Human Organs-on-chips, Microfluidic devices lined with living human cells for drug development, disease modelling, and personalized medicine. Wyss Institute.

24.  Basile, A. O., Yahi, A., & Tatonetti, N. P. (2019). Artificial intelligence for drug toxicity and safety. Trends in pharmacological sciences.