Should foetal tissue be used for medical research?

Planned Parenthood is an American non-profit women’s health organisation. They provide free education, healthcare and contraception to millions of women each year. They are also the largest abortion provider in the US, making them a prime target for anti-abortion campaigners such as The Center for Medical Progress (CMP).

Last year, CMP released secretly filmed videos of physicians at Planned Parenthood. The physicians spoke about conducting abortions in a way that keep foetuses intact for use in research. They also suggest that the Planned Parenthood are illegally profiting by ‘selling’ foetal tissue. The videos sparked a furore which resulted in the US Senate passing legislation to stop Government funding to Planned Parenthood. Although the legislation will not be passed, as it will certainly be vetoed by President Obama, it does prompt serious questions about the use of foetal tissue in medical research.

Millions of non-human animals are bred each year by profit-making businesses to sell to researchers. The research conducted on these animals is often completely irrelevant to humans. The foetal tissue will be produced regardless of the research. Furthermore, there is no evidence that Planned Parenthood or the researchers using foetal tissue are breaking any laws. If foetal tissue is legally obtained, and would otherwise be thrown away, why not use it to study diseases such as HIV/AIDS?

As long as women are not pressured into abortions for financial gain, isn’t human foetal tissue preferable to using specially-bred animals of a different species?

Missing Mice

There are a number of problems with scientific publications.  Some of which are avoidable, others are not.  Some are well-documented, others still remain veiled.  Some have the potential to significantly damage a study’s validity, while others may simply question it.  Sometimes a single flaw in the process can encompass many of these at once.

Last month Nature News posted an article on their website titled Missing Mice: Gaps in the data plague animal research.  The tagline, ‘reports of hundreds of biomedical experiments lack essential information’ outlines the increasingly evident failings in scientific publications of not documenting crucial study components.  While not the first to raise awareness of such short-comings, the Nature article focuses on two separate studies that reinforce this growing awareness.

The first study from the Charité Medical University in Berlin review 522 rodent-based experiments from 100 paper from 2000-2013, discovering that two thirds of them displayed a drop in the numbers of animals used between the methods and results sections.  Sometimes dropping one or several test subjects, for valid and noted reasons, is part of the scientific process.  However, in the experiments investigated, only 14 explained why.  This suggests potentially misleading results, or poor experimental protocol.  Additionally, it could also point to intentional or unintentional reporting biases, wherein individual results that would have confound the desired overall results were cast aside. Further analysis has showed that the relative statistical severity of removing selected data points from studies could affect the overall results by as much as 175%.

How many publications, especially those that are heavily cited, have discrepancies between the initial n value proposed and the actual number used in the final conclusions?  It would be seemingly straightforward to justify the use of few dozen or hundred mice for a study, then break that number up into treatment groups over several runs, subsequently removing data points here and there as and when they disagreed with the sought after result.  How many casual, or even invested readers for that matter, would go back and double check that all the numbers add up to the initial amount?

The second study examined whether 268 randomly selected biomedical papers provided full data as well as sufficient detail to replicate the work.  The Stanford-lead study discovered that none supplied full data and only one provided information adequate enough to reproduce the experiment.  Beyond this, it was stated that in 2014 33% of papers included conflict-of-interest statements, compared to 10% in 2000.

These findings are considerably discouraging.  Of course, they both reflect relatively small sample sizes, yet provide important insights into areas of scientific reporting that many take for granted.  The investigations carried out are part of a larger meta-analysis looking to address and identify the most common and problematic fallacies in scientific publications.  An immense task to uncover all the past inaccuracies, but incredibly valuable to avoid any future ones.  The ultimate risk to proper scientific output from practices like the ones discovered by these studies cannot be underestimated nor disregarded.  The immediate hope is that through widespread awareness across all areas of research of the potential complications and/or biases such activities can produce a fresh start, in sorts, can begin.  Naturally, not all of the reporting and methodological inaccuracies are intentional or malicious; yet increased universal understanding and appreciation of the implications of the issues related to them should hopefully lead to their reduction and a growth in more reliable and improved research.

Open Trials

Open Trials is a project led by Bad Science author Ben Goldacre that aims to form a complete collection of every clinical trial conducted around the world. A clinical trial is when a new medicine is tested in humans for the first time. The results from these trials are used to decide how well the medicine works and how safe it is. Currently, not all clinical trial results are published, especially when the findings are negative. This can have dangerous consequences on patient safety as medicine regulators and doctors may not be fully aware of a medicine’s effects. Therefore Open Trials is an important project.

However, before medicines are tested in humans for the first time, they are required to undergo pre-clinical tests in various species of non-human animals. These tests are required to show that the medicines are safe and effective in animal models before they are allowed to be tested in humans. However, the usefulness of these animal studies has been questioned because of the inherent biological differences between species. The results from animal studies are often very different from the results in humans, meaning that ineffective or unsafe medicines are given to humans in clinical trials. Conversely, potentially good medicines are rejected before they get to be trialed in humans because of poor results in animals.

What if we had an Open Trials–like project for pre-clinical animal studies, so that every animal study was recorded and all the results were openly available for others to review?

Here are some of the potential advantages of such a project:
– Overcoming issues such as publication bias. This occurs because studies with positive findings are more likely to be published leading to an unrepresentative and often misleading view of research that has been conducted.
– All data is given to regulators, so they can make a fully informed decision about whether a new medicine is allowed to be tested in humans.
– Finding out how useful and predictive the animal models are for human diseases and treatments.
– Reducing animal use by preventing duplication of animal studies, particularly ones with negative findings. If a study finds that a drug doesn’t work in mice, the results are unlikely to be published. Therefore other researchers might test the same drug again, without realising that it has already been proven ineffective.
Overall, more open reporting of all data can only be a good thing for science. Open trials is fantastic step forward, but more can be done for other types of experiments. This openness would be particularly valuable with animal studies due to the considerable ethical costs of the research.

Young Researchers — The Ethical Challenge

Rebecca Ram

Never before has there been a better opportunity for young researchers
to focus on replacement alternatives, not only to save animals
from unnecessary pain and suffering,
but also to pave the way for a career in cutting-edge innovation


The Lush Prize awards and encourages individuals or organisations who have contributed to the global initiative to end animal testing, in the fields of science, public awareness, lobbying and training, and also aims to support young researchers who wish to develop a career in animal-free toxicology. The total yearly prize fund is £250,000. The ‘Young Researcher’ category of the Lush Prize welcomes nominations from early-career scientists who are keen to progress in research without animal testing. The award offers four £12,500 bursaries, to reward research and development specifically in methods for the entire replacement of animals in toxicity testing.
In this article, Lush Prize refers to testing without the use of animals by using the term ‘non-animal’ methods. They are validated, scientifically robust methods of safety testing in their own right. The use of terms such as ‘alternatives’ or ‘replacement’ methods (while useful for clarity sometimes) may suggest that animal testing is the ‘gold standard’ of safety testing, when much of the scientific industry, along with a wealth of research evidence, confirms that, aside from the suffering involved, animal tests do not reliably predict human responses. In addition, the term ‘alternatives’ is also used to describe the Three Rs methods,1 previously summarised as follows:
Refinement: to minimise suffering and distress to animals;
Reduction: to minimise the number of animals used; and Replacement: to avoid the use of living animals. Whilst reduction or refinement methods are positive steps, they are not achievements according to the ethos of the Lush Prize. We consider only the final ‘R’ (Replacement) to be a genuine alternative, as the other two Rs still involve animal use.

A brief review of previous findings

To discuss the relevance of the Young Researcher prize, a number of key animal protection organisations were contacted and interviewed during previous research for the Lush Prize. Some of these contacts are quoted and provide useful updates throughout this article. The key messages are:

— Although the acceptance and recognition of new technologies is growing at an encouraging rate, animal-free toxicology is still the ‘less travelled’ path, and any early-career researcher trying to progress in this area is likely to meet at least some resistance or challenges along the way. That said, the environment for the discussion of alternatives to animal use is expanding, so it is vital to be persistent and remain true to one’s values in pursuing career and networking opportunities.
— A very proactive attitude is needed; ethical scientists must actively seek out their opportunities, but the rewards can be hugely successful.
— Continuing to promote the anti-animal testing message to all relevant individuals in the young/early-career researcher field, as well as communicating this message at an earlier stage of education, is vital.

One of the key positive findings from interviews with previous Young Researcher prize winners is that “Young scientists don’t always have the prejudices about animal testing being the ‘best’ way of doing things”.2 An unbiased view and fresh perspective on cutting-edge science is essential, combined with raising awareness earlier in the educational system. There is also a strong link between the Young Researcher and Training prizes, as the latter awards are relevant to those involved in the education of a number of audiences, from children in early-stage schooling, through GCSE/A-level, to the undergraduate and postgraduate levels and beyond. There is considerable scope for early-career researchers working in a broad range of scientific or technical fields to get involved in non-animal methods. This is coupled with the fact that the in vitro toxicity testing market is projected to be worth $17,227 million by 2018.3 The EU cosmetic testing ban has played a key part in driving this growth.

Mainstream funding and development of Replacement methods

The importance of funding offered by the Lush Prize continues to grow. This is especially relevant to the Young Researcher Prize, as the bursaries awarded will directly fund the development of methods to replace the use of animals in ‘frontline’ research. As highlighted in previous research conducted for the Lush Prize, lack of finance is a major obstacle to the availability of non-animal methods. The ongoing reliance on animal research means that it continues to receive the vast majority of the available funding. Furthermore, those interested in non-animal research, not only have to maintain the momentum on their specific ideas and methods, but also face a need to continually look for funding or sponsorship, which ultimately impacts on the amount of time they directly spend on their research, as acknowledged by PETA in an interview with Lush Prize in 2012: “…people may have good ideas about non-animal methods, but they’re continually going to be seeking support…and funding for those”.4
To provide some figures to illustrate the above points, in the UK in 2012–2013, over £300 million of public funding was spent on projects which  “include an element of animal use”.5In contrast, a sum of just under £9 million was awarded to Three R projects broadly termed as ‘alternatives’, with the NC3Rs awarding £7 million of this total.5 The NC3Rs state that, of the funding they provide, “around 55 per cent of research awards are directed primarily at replacement, 25 per cent for reduction and 20 per cent for refinement”.6            Therefore, within this £9 million, a much lower sum was awarded to genuine non-animal (Replacement) methods, as a significant amount of ‘alternatives’ funding is donated to the other two Rs, which still involve animal use. For example, previous NC3Rs funding includes projects which develop scales for recognising facial expressions of pain in monkeys7 or facial grimace scales in rabbits.8 To add further perspective, since it was established in 2004, the NC3Rs has awarded just over £37 million in project funding. Based on the above figures, this equates to 12% of just one year’s Government funding of projects which include animal research.
Research carried out by the BUAV in 2013 revealed the stark lack of funding devoted to alternatives to animal testing across the EU Member States. Just
€18.7 million were devoted to methods relating to  the Three Rs in 2013, by only seven countries, with most Member States failing to assign any funding at all, and half of them not responding to the survey. Given that the available figures for 2011 show the total combined annual science R&D (research and development) budget for the EU to be almost €257 billion, the amount spent on alternatives is wholly inadequate, equating to just 0.007% of the total expenditure.9
These disappointing figures demonstrate the importance of independent funding for non-animal research, such as that which the Lush Prize offers. At its launch in 2012, the £50,000 total prize money for the Young Researcher Prize was allocated to five potential winners. This has now changed to award four prizes of £12,500, in order to increase the funding awarded to each individual, whilst still recognising the work of several researchers. Feedback from previous prize winners has indicated that these bursaries provide a meaningful amount, so the slight increase in funding across four awards will be of even more benefit, to go toward both research expenses and the cost of consumables.

Current and ongoing opportunities for keen young researchers

Banning animal testing will stifle innovation?

The claim that a ban on animal testing would stifle innovation was regularly made by industry as the 2013 EU cosmetics testing ban came into effect.10 Far from impeding research, the ban (both the 2009 and 2013 phases) had the opposite effect, and was the direct driver for the launch of new research into non-animal methods through large-scale, multinational projects (e.g. ReProTect11) as these two critical deadlines approached. R&D on new methods is innovation in itself, and it provides the perfect opportunity for those who genuinely want to be involved in cutting-edge, next-generation science, without causing animal suffering. The EU has led the way in progress on the development of alternative methods of testing to animals, and is considered ‘a leader in innovation’, something which should be reflected in the opportunities it offers young researchers and emerging graduate scientists.

Toxicity testing is toxicity testing, regardless of purpose

The validated and accepted non-animal (replacement) toxicity testing methods that are now available have been developed largely due to the phased EU ban on the animal testing and marketing of cosmetic ingredients and finished products. As a result, discussions on the replacement of animals in toxicity testing are far more common and perhaps are considered more acceptable in the cosmetics field. However, young researchers working or studying in other areas of toxicity may feel less able to speak out about their research interests, especially if they involve replacement/non-animal methods, as these are seen as more controversial than ‘two Rs’ (reduction or refinement) approaches. It is therefore important to recognise that these methods are now of essential use in other chemical testing sectors, such as the food or pharmaceutical industries. This demonstrates that when a non-animal method is developed and accepted, it can potentially be applied to the testing of any substance, for any purpose. This may encourage young researchers to voice their interests in the development and use of non-animal methods.
This is especially relevant as, despite the development of alternatives for use in areas such as cosmetics, toxicity testing in animals continues in many other industries. For example, in the UK in 2013, over 375,000 toxicity tests (from a total 4.12 million procedures) were performed on animals (mice, rats, rabbits, guinea-pigs, dogs, cats, monkeys, birds and fish).12 Another important point with regard to the Young Researcher Lush Prize, is that almost half of all animal experiments in the UK are carried out at  universities. One of the most concerning findings is that the increase in the use of genetically-modified animals (mainly mice) and the increasing use of zebrafish are, in some contexts, being considered as ‘alternatives’. This was highlighted by FRAME in a report on the Home Office annual statistics on animal use in Great Britain in 2012.13
The 2011 EU figures14 showed that over 1 million animals (1,004,873, from a total of just under 11.5 million animals) were used in toxicity testing across the EU states in that particular year. Of these, 111,166 animals were used in tests that were not even required by law (categorised as ‘no regulatory requirements’). The  archaic and much criticised LD50/LC50 (lethal dose or lethal concentration test, which tests the amount of substance required to kill 50% of the animals tested) accounts for the majority of the animals used each year, along with other lethal tests (34%). The other main use is simply categorised as ‘other’ toxicology tests (22%), followed by chronic/sub-chronic toxicity and reproductive toxicity. There is no official figure for the number of toxicity tests still conducted on animals worldwide (from the estimated yearly total of 115 million animals15 used in all experiments), as many countries omit this information or do not even count the numbers of animals used. However, a revised  estimate by Lush Prize researchers puts the number of toxicity tests carried out on animals worldwide at almost 9.5 million (from a total 118 million animal experiments).16

After the 2013 marketing ban, much work is still to be done

Although the EU cosmetics legislation has been the major driver of the development of non-animal toxicity testing methods in recent years, there is still much more to be done. This is illustrated very clearly, given that “Over 80% of the world allows animals to be used in cruel and unnecessary cosmetics tests and these animal tested cosmetics can be purchased in every country across the globe.”17
The proposed 2013 EU marketing ban on animaltested  cosmetics did finally go ahead, though the European Commission had previously considered the possibility of delaying the deadline on the basis of recommendations that necessary but still missing’ alternative methods would take much longer to be developed. For example, estimates of another 5–9 years were proposed for methods for skin sensitisation and toxicokinetics to be developed, and possibly even longer for full replacement in these areas. No estimates were provided at all for when repeat-dose toxicity, reproductive toxicity or carcinogenicity tests on animals might be developed. These timelines were estimated in a report published by the Commission in 2011.18  In the three years since that time, aside from the introduction of the 2013 ban itself (which went ahead regardless of the lack of alternatives available, which was great news), further work had been ongoing in the areas of toxicity testing which still need development. For example, in 2013, the Joint Research Centre (JRC) published its EURL-ECVAM Strategy to Avoid and Reduce Animal Use in Genotoxicity Testing.19 Similarly, the five-year long NOTOX project,20 launched in 2011 and involving a network of scientific expertise from several countries, works “towards the replacement of current repeated dose systemic toxicity testing in human safety assessment”. NOTOX is part of a wider project , funded under the EU Seventh Framework Programme (FP7), known as SEURAT (Safety Evaluation Ultimately Replacing Animal Testing). This project combines the research efforts of over 70 European universities, public research institutes and companies, and regularly posts open vacancies and research opportunities.21 Of particular relevance is that SEURAT hosted a Young Scientists Summer School to discuss replacement of repeat-dose toxicity testing in animals.22

Focusing on key areas

As highlighted by EURL-ECVAM,23 a key factor in the development of non-animal methods is the integration of a number of alternative test methods into a ‘battery’ that successfully addresses a number of endpoints, especially those which are considered more complex or need to be considered in-depth. For example, several alternatives available for skin testing, examine how a substance may react in various stages of topical application, absorption, irritation or corrosion, and provide very targeted and quantitative results, especially when compared to a crude skin test in rabbits or guinea-pigs. Therefore, non-animal methods which are still under development or undergoing validation, or ‘gaps’ in the development of methods where the greatest use of animals still occurs, such as reproductive or chronic toxicity, could be helped by awarding the Lush Prize to young researchers to allow them to potentially channel their ideas or research themes for specific replacement projects and encourage them to specialise in key areas.
Never before has there been a better opportunity for young researchers to focus on replacement alternatives, not only to save animals from unnecessary pain and suffering, but also to pave the way for a career in cutting-edge innovation. As highlighted by the New England Anti-Vivisection Society (NEAVS) in a previous interview with Lush Prize, linking an early-career scientist’s research to increased income and sponsorship is key:
“I think the solution for graduate students who want  to do more progressive in vitro research is to find the granting agencies that will help bring money in [to an institution]. …The key to changing institutions is bringing in grant dollars. When someone who wants to develop in vitro alternatives can show that they can bring in million-dollar grants, then institutions are going to have to accept it. They’re not going to turn money away, even if they want to try to suppress a certain ideology”.24
What this means, in effect, is that, if a researcher has ideas, but can also say “if you fund me, I propose to cut your costs, save you time, increase income and improve your business”, whilst this might be viewed as a challenge, their proposals are much more likely  to be considered.

Challenges for ethical early-career scientists

As previously highlighted, there remain ongoing prejudices toward switching from animal to non-animal research. Resistance to change, combined with ‘comfort’ in repeating accepted, conventional  methods, allows the animal research industry to maintain the status quo, despite ever-increasing recognition that animal testing is a flawed, overrated and outdated system. It must also be noted that the industry has, for decades, consisted of a network, not only of researchers, but also breeders, suppliers and transporters of animals across the world, who rely on animal testing to continue. There are other factors to consider — for example, some scientists (especially senior-level researchers) have based their entire careers on the use of animals, and are unable or unwilling to consider anything else; they may view switching to non-animal research as  the daunting and unattractive option of ‘starting again’. This may also apply to earlier career individuals, who have followed the mainstream route into animal-based toxicology to progress their careers to date, for example, since leaving university. This is echoed by several previous prizewinners, who felt that the undergraduate level of their education was the most challenging arena in trying to avoid the use of animals. Nevertheless, one positive finding from interviews with previous Young Researcher Prize winners is that “Young scientists don’t always have the prejudices about animal testing being the ‘best’ way of doing things”.2
Finally, to provide some useful insight into the types of scholarships available to young researchers, Appendix 1 gives a summary of 15 PhD studentships recently funded by the UK NC3Rs. A full list is shown to illustrate the types of research being undertaken
— however, it must be noted that the studentships cover the broader remit of the Three Rs, rather than the ‘replacement only’ criterion that the Young Researcher Prize demands.

Download the full article here (including Appendix 1) 

Rebecca Ram
Lush Prize
41 Old Birley Street
Manchester M15 5RF

References and Notes

1 Russell, W.M.S. & Burch, R.L. (1959). The Principles of Humane Experimental Technique, 238pp. London, UK: Methuen.
2 Anon. (2012). Lush Young Researchers Prize 2012 — Research Paper, 22pp. Available at: http://www.
Young-Researchers-Prize-2012-Research-Paper.pdf (Accessed 05.11.15).
3 Anon. (2014). In-Vitro Toxicology Testing Market worth $17,227 Million by 2018. Available at: http://www. market-worth-17227-million-by-2018-25358636 1.html (Accessed 05.11.15).
4 Interview with PETA, 20 August 2012.
5 Willetts, D. (2014). Hansard Written Answers. Animal Experiments: Business, Innovation and Skills written question — answered on 11th March 2014. Available at: 2014-03-11a.188641.h (Accessed 05.11.15).
6 NC3Rs (2014). Funding schemes. Available at: http:// (Accessed 06. 06.14).
7 NC3Rs (2014). Quantifying the behavioural and facial correlates of pain in laboratory macaques. Available at:
(Accessed 05.11.15).
8 NC3Rs (2012). The Rabbit Grimace Scale — A new method for pain assessment in rabbits. Available at: E2%80%93-new-method-pain-assessment-rabbits
(Accessed 05.11.15).
9 Taylor, K. (2014). EU member state government contribution to alternative methods. ALTEX 31, 215– 218.
10 Anon. (2013). Europe Bans Marketing of Cosmetics Tested on Animals. Available at: http://ensnewswire. com/2013/03/11/europe-bans-marketing-o
f-cosmetics-tested-on-animals/ (Accessed 05.11.15). 11 Schwarz, M. (2011). Meta Analysis of a Battery Test of Reproductive Toxicity Assays: The ReProTect Experience. [Presentation given at Open Tox, Munich,
9–12 August, 2011.] Available at:
Talk-Schwarz.pdf (Accessed 05.11.15).
12 Home Office (2013). Annual Statistics of Scientific Procedures on Living Animals — Great Britain 2013, 59pp. London, UK: Her Majesty’s Stationery Office. Available at: uploads/system/uploads/attachment_data/file/3278 54/spanimals13.pdf (Accessed 05.11.15).
13 Hudson-Shore, M. (2013). Statistics of Scientific Procedures on Living Animals 2012: Another increase in experimentation — Genetically-altered animals dominate again. ATLA 41, 313–319.
14 Anon. (2013). Report from the Commission to the Council and the European Parliament. Seventh Report on the Statistics on the Number of Animals Used for Experimental and Other Scientific Purposes in the Member States of the European Union. COM (2013) 859 final, 14pp. Brussels, Belgium: European Commission. Available at:
9&from=EN (Accessed 05.11.15).
15 Taylor, K., Gordon, N., Langley, G. & Higgins, W. (2008). Estimates of worldwide laboratory animal use in 2005. ATLA 36, 327–342.
16 Anon. (2014). The 2014 Lush Prize: A Global View of
Animal Experiments 2014, 42pp. Available at: http://
View_of-Animal_Experiments_2014.pdf (Accessed 05.11.15).
17 Cruelty Free International (2012). Did you know animal tested cosmetics are for sale in every country in the world? Available at:  ttp://www.crueltyfree (Accessed 06.06.14).
18 Adler, S., Basketter, D., Creton, S., Pelkonen, O., van Benthem, J., Zuang, V., Andersen, K.E., Angers- Loustau, A., Aptula, A., Bal-Price, A., Benfenati, E.,
Bernauer, U., Bessems, J., Bois, F.Y., Boobis, A., Bran – don, E., Bremer, S., Broschard, T., Casati, S., Coecke, S., Corvi, R., Cronin, M., Daston, G., Dekant, W., Felter, S., Grignard, E., Gundert-Remy, U., Heinonen, T., Kimber, I., Kleinjans, J., Komulainen, H., Kreiling, R., Kreysa, J., Leite, S.B., Loizou, G., Maxwell, G., Mazzatorta, P., Munn, S., Pfuhler, S., Phrakonkham, P., Piersma, A., Poth, A., Prieto, P., Repetto, G., Rogiers, V., Schoeters, G., Schwarz, M., Serafimova, R., Tähti, H., Testai, E., van Delft, J., van Loveren, H., Vinken,
M., Worth, A. & Zaldivar, J.M. (2011). Alternative (nonanimal)
methods for cosmetics testing: Current status and future prospects — 2010. Archives of Toxicology 85, 367–485.
19 Corvi, R., Madia, F., Worth, A. & Whelan, M. (2013). EURL ECVAM Strategy to Avoid and Reduce Animal Use in Genotoxicity Testing, 48pp. Ispra, Italy: European Commission, Joint Research Centre, Institute for Health and Consumer Protection. Available at:
111111111/30088/1/jrc_report_en_34844_on line.pdf (Accessed 05.11.15).
20 NOTOX (undated). Welcome to NOTOX. Available at: (Accessed 05.11.15).
21 SEURAT (undated). Welcome to the SEURAT-1 website. Available at: (Accessed 05.11.15).
22 SEURAT (undated). SEURAT-1 & ESTIV Joint Summer School — 8–10 June 2014. Egmond aan Zee, Netherlands. Available at:
seurat-1/2014/summer-school/ (Accessed 05.11.15).
23 Anon. (2013). EURL ECVAM Progress Report on the Development, Validation and Regulatory Acceptance of Alternative Methods (2010–2013). Available at:
validation-regulatory-acceptancealternative- methods (Accessed 06.06.14).
24 Interview with NEAVS, 20 August 2012.

On Replacing the Concept of Replacement

Michael Balls

Russell and Burch saw failure to accept the correlation
between humanity and efficacy as an example of rationalisation,
a psychological defence mechanism

While wondering what I could discuss in this column I looked, as I often do, in the abridged version1 of The Principles of Humane Experimental Technique,2 at Russell and Burch’s introduction of what I call the humanity criterion. It is part of their discussion of the sociological factors which are among the Factors Governing Progress. This is how part of page 101 of the abridged version reads:

In fact, really informative experiments must be as humane as would be conceivable possible, for science and exploration are indissolubly linked to the social activity of cooperation, which will find its expression in relation to other animals, no less than to our fellow humans. Conscious good will and the social operational method are useless as safeguards against the mechanism of rationalisation (in the pathological sense of the term – i.e. the mechanism of defence by which unacceptable thoughts or actions are given acceptable reasons to justify them to oneself and to others, while, at the same tie, unwittingly hiding the true, but unconscious, motives for them).

The bold type indicates my explanation, and I have to admit that, six years after preparing the abridged version of The Principles, I now found it difficult to see what Russell and Burch had intended to convey. I therefore looked back at the original book, and found this paragraph on pages 156−157:

In efficacy, or yield of information, the advantages of humane technique apply almost universally. The correlation between humanity and efficacy has appeared so often in this book that we need not labour the point. There is, however, a more fundamental aspect of this correlation, specially important in research. Science means the operational method — telling somebody else how to see what you saw. This method is one of the greatest of all human evolutionary innovations. It has, however, one drawback. It prevents permanent acceptance of false information, but it does not prevent wastage of time and effort. The activity of science is the supreme expression of the human exploratory drive, and as such it is the subject to the same pathology. The scientist is liable, like all other individuals, to block his exploration on some front where his reactions to childhood social experiences are impinged upon. When this happens to the experimental biologist, we can predict the consequence with certainty. Instead of really exploring, he will, in his experiments, act out on his animals, in a more or less symbolic and exaggerated way, some kind of treatment which he once experienced in social intercourse with his parents. He can rationalise this as exploration, and hence fail to notice the block. But in fact such acting out invariably occurs precisely when real exploration is blocked, and must be relinquished before real exploration can begin again. Hence, such experiments will be utterly wasteful, misleading, and uninformative. The treatment of the animals, for one thing, will inevitably be such as to impair their use as satisfactory models. The interpretation of the results will be vitiated by projection. Really informative experiments, must in fact be as humane as would be conceivably possible, for science and exploration are indissolubly linked to the social activity of cooperation, which will find its expression in relation to other animals no less than to our fellow humans. Conscious goodwill and the social operational method are useless as safeguards against the mechanism of rationalisation (in the pathological sense of the term).

Here, the underlining indicates what I omitted from the abridged version, and I now wonder why I did so. These words clearly reflect Russell’s interest in psychology — he later became a psychotherapist, and undoubtedly will have been influenced by discussions with his psychotherapist wife, Claire Russell. They could be seen as an explanation why some scientists did not appreciate the essential link between humanity and efficacy, and why Russell thought they needed what was offered by the Three Rs and the humanity criterion.

It is not clear what is meant by “the social operational method”, and consulting Google leads to only one hit — The Principles itself! “Conscious goodwill” is probably meant to contrast with unconscious rationalisation.  Perhaps what Russell meant is that, however sincere the intention may appear to be, support for the Three Rs is useless, unless it leads to active and practical commitment to their development and application.

We are often confronted with rationalisation, the pseudo-rational justification of irrational acts,3 and its relative, intellectualisation, a different defence mechanism (or way of making excuses), “where reasoning is used to block confrontation with an unconscious conflict and its associated emotional stress, where thinking is used to avoid feeling. It involves removing one’s self, emotionally, from a stressful event. Intellectualisation is one of Freud’s original defence mechanisms. Freud believed that memories have both conscious and unconscious aspects, and that intellectualisation allows for the conscious analysis of an event in a way that does not provoke anxiety.”4

I am not a psychoanalyst, and I think it would be unwise, even dangerous, were I to seek to delve into the underlying reasons why some scientists are so keen to run to animal experimentation as the first resort and to do so little to make possible its replacement. Nevertheless, I can say, without fear of contradiction, that this is another great example of how Russell and Burch’s wonderful book continues to give us food for thought and calls for action.

Professor Michael Balls

1 Balls, M. (2009). The Three Rs and the Humanity Criterion, 131pp. Nottingham, UK:  FRAME.
2 Russell, W.M.S. & Burch, R.L. (1959). The Principles of Humane Experimental Technique, xiv + 238pp. London, UK: Methuen.
3 Anon. (2015). Rationalization (psychology). San Francisco, CA, USA: Wikipedia Foundation, Inc. Available at: (Accessed 26.08.15).
4 Anon. (2015). Intellectualization. San Francisco, CA, USA: Wikipedia Foundation, Inc. Available at: (Accessed 26.08.15).

The Principles of Humane Experimental Technique is now out of print, but the full text can be found at The abridged version, The Three Rs and the Humanity Criterion, can be obtained from

Download a pdf copy of this post by clicking here.

Previous Wisdom of Russell & Burch posts from Michael Balls:

The Concept, Sources and Incidence of Inhumanity and its Diminution or Removal Through Implementation of the Three Rs.
The Wages of Inhumanity.
Fidelity and Discrimination.
The Factors Governing Progress. 
UFAW and Major Charles Hume. 
The Toxicity Testing Problem. 
The Use of Lower Organisms.
The Analysis of Direct Inhumanity.
William Russell: Polymath, Wordsmith, Classicist and Humourist .
Rex Leonard Burch: Humane Scientist and Gentle Man.
On the Proper Application of Appropriate Statistical Methods. 
Comparative Substitution. 
The Three Rs: The Way Forward .
The Choice of Procedures .
Rationalisation and Intellectualisation.


Read-across for Hazard Assessment: The Ugly Duckling is Growing Up

Wera Teubner and Robert Landsiedel

Increasing use of read-across in integrated approaches for the testing and assessment
of chemical hazards will ensure that it eventually matures into a beautiful swan

In the hazard assessment of chemicals, read-across describes a technique used to predict physicochemical, ecotoxicological and toxicological endpoints. If it is performed on several substances at a time, it is called ‘category formation’. Read-across is based on the experience that similar chemicals exhibit similar properties — with the crucial issue of knowing which properties determine similarity for a given endpoint. In this aspect, it is a relative of the quantitative/ qualitative structure–activity relationship (QSAR), and was sometimes simply termed ‘expert judgement’. The idea of the read-across concept being an ‘ugly duckling’ has mostly arisen from the difficulty in verifying the plausibility of its findings without actually performing the experimental studies. The Read-Across Assessment Framework (RAAF),1 published in May 2015, states that “Under REACH, any read-across approach must be based on structural similarity between the source and target substances”. However, the limited verification of readacross, and especially the limitations of the use of the read-across approach only to structural similarities, reflect a state of infancy that needs to be nurtured toward maturity in order to reap its maximum

When, in 1959, William Russell and Rex Burch published The Principles of Humane  Experimental Technique,2 calling for the replacement, refinement  and reduction of animal testing, a major focus was the quality of animal testing and the criticism that
poor planning and experimental techniques resulted in animal studies of limited value, and consequently in more testing than should have been needed. With the introduction of Good Laboratory Practice and of Organisation for Economic Co-operation and Development (OECD) test guidelines (TGs) and animal welfare policies, the quality of animal data has become much less of a problem, and refinement has considerably improved. The improvement of cell culture, tissue culture and molecular biology technology kindled the hope for replacement. Meanwhile, standalone in vitro methods (e.g. for skin and eye irritation) or batteries of tests (e.g. for skin sensitisation) can address local toxicity. Likewise, methods to address specific early effects or mechanisms, such as genotoxicity or oestrogenic activity, are available.

A major challenge today is the prediction of complex toxicological effects such as systemic  and developmental toxicity. Large research programmes, e.g. ToxCast or SEURAT, aim to meet this challenge.3,4 Any new approach to complex toxicological effects combines various methods (in silico, in vitro and in vivo) in a testing battery or strategy.5,6 These approaches use mechanistic information, and are constructed according to (putative) adverse-outcome pathways (AOPs).7 Such information is, of course, also useful in supporting the read-across of apical toxic effects of different chemicals. Read-across can actually become a successful part of many integrated approaches for testing and assessment (IATAs).

Traditionally, chemicals are considered candidates for read-across, if they share structural similarity or are metabolically or spontaneously transformed to common products. It is assumed that structural similarity will result in a common mode-of-action. When assessing wanted pharmacological activities or unwanted toxicological hazards in research and development, applying read-across is already possible when the substance in question still only exists on paper. High-quality predictions are valuable for success in product development. At some point, the predicted effects are determined experimentally for promising candidates, and it is at this point that the consequences of poor read-across hit back. Again, identifying the correct similarity between read-across source and target chemicals is crucial.
Figure 1
The ‘ugly duckling’ characteristics of read-across (Figure 1) originate from areas in which it is used as a quick (and cheap) means to generate hazard information, either to fulfil regulatory data requirements, or to identify and list substances allegedly of very high concern (no reference given here, since this PiLAS is not a pillory). It also may originate from the idea that any information is better than no information in situations where there is no budget, or when animal testing is simply out of the question. Global efforts to identify and substitute hazardous chemicals can only succeed, if so-called ‘regrettable  substitutions’ can be avoided. Neither overestimation nor  underestimation of hazards by read-across is helpful in this context. Actually, it takes a wide range of thorough considerations to perform a robust and meaningful read-across — and these need to be documented.  To toxicologists with long experience in their respective chemical space, similarity may seem so obvious that their read-across justifications are rather frustrating to comprehend.

The application of read-across and the related category approach received a boost when  the European Union (EU) introduced the REACH programme in 2006. The REACH legislation (EC Regulation 1907/2006)8 requires the hazard characterisation of all chemicals marketed in the EU, with actual data requirements dependent on the production and import tonnage and the use conditions. With the estimation that more than 20,000 chemicals would need to be assessed, the legislation needed to include provisions to use animal testing only as a last resort. The obligation of the European Chemicals Agency (ECHA) to report on the status of the implementation and use of non-animal test methods and testing strategies is actually laid down in Article 117(3) of the legislation. As of 1 October 2013, dossiers for 8,729 substances have been submitted to the ECHA. A readacross or category approach was used in up to 75% of analysed dossiers for at least one endpoint.9

Considering the huge number of chemicals that were to be registered within the short period of eight years, the REACH legislation introduced a previously mostly-unknown component to chemical legislation. It was proposed that acceptance of registration, if appropriate, would be granted after automated dossier screening. Any scientific review of toxicological data would then be performed at a later stage, and this review would have to be conducted for at least 5% of the registered substances. With this procedure, the opportunity for an upfront discussion on the data requirements and suitability to support a read-across approach is in no way considered. This registration strategy has the advantage of speed and a certainty of meeting submission deadlines, but  he disadvantage of uncertainty with regard to follow-up activities, the latter arising from the possibility that the read-across assessment might be judged to be deficient and the decision would then be made that the target substance must be tested.

Both challenges and improvements to read-across approaches have been triggered by cases where apparently small changes in structures resulted in vast changes of the hazard properties (so called ‘activity cliffs’). The most prominent examples originate from differences in the interactions of substances with enzymes and receptors. The substances 2-acetylamino fluorene (2-AAF) and 4-acetylaminofluorene (4-AAF) are structurally very similar. As well as being a bladder carcinogen, 2-AAF is a strong liver enzyme inducer, leading in long-term studies to liver tumours. However, 4-AAF only slightly induces liver enzymes and does not induce the formation of liver tumours.12 Enantiomers of 1-hydroxyethylpyrene are activated to mutagenic sulphates by different sulphotransferases, 13 and the enantiomers of Carvone smell of caraway or spearmint,14 to name but two examples. When looking at the two-dimensional description of a chemical only (e.g. SMARTS pattern or Tanimoto score), stereoisomers appear identical, but three-dimensional structure modelling for receptorbinding simulation can differentiate stereoisomers. Regardless, stereo-isomeric and regio-isomeric differences of molecules appear to be small alterations, as compared to the changes usually bridged by readacross (e.g. homologous series). It is important to know which aspect of similarity between two chemicals is governing their similar hazardous properties.

Structure–hazard relationships are a ‘long-shot’: In between the structure of a chemical and its apical toxic effect are its material properties (e.g. electrophilicity), system-dependent properties (e.g. ROS generation), molecular interactions (e.g. receptorbinding and DNA-binding) and early cellular responses (e.g. mutagenicity). It is crucial to know when structure information is sufficient, or when additional data, possibly closer to the apical effect, are needed, but this should not undervalue the research efforts undertaken to derive such properties from information on structure, nor does it mean that structure and material property are unrelated. This has been exemplified with skin sensitising chemicals of low molecular weight, where reaction classes identified from the chemical structure may be a more-instructive property to predict the protein-binding than general molecular descriptors.15–17 The reaction class is considering only the property that is essential to initiate the molecular initiating event of skin sensitisation, i.e. protein binding, whereas general molecular descriptors can ‘dilute’ this information with molecular
features of less relevance.

Evidently, properties and effects closer to the apical toxic effects are more predictive and less uncertain. Lately, the concept of applying ‘functionality’ rather than (or in addition to) material descriptors was proposed for nanomaterials.18–20 This can be taken a step further: Rather than using the molecular structure or the ‘functionality’, read-across can be based on the early biological effects or common modes-of-actions of two (or more) substances. Actually, such a concept is typically represented by the common classification of any chemical with a pH of > 11 as corrosive, but no one would consider calling it functionality-based or mode-of-action-based read-across. The concept of biologically-based activity relationship (QBAR, i.e. referring to QSAR, the structure-based activity relationship) has been discussed and exemplified by van Ravenzwaay et al.12 The example of different toxicities of the structurally-similar isomers, 2-AAF and 4-AAF, was given above. These differences are reflected in different metabolome-patterns induced by these two compounds. Another example are fibrates with structural similarity. Most of these fibrates also show toxicological and pharmacological similarity, based on the metabolome data. Gemfibrozil, however, does have different pharmacological and toxicological effects. The differences in the target organ (e.g. the kidney) for Gemfibrozil and its pharmacological effect (cholesterol lowering) can be identified, based on the metabolome data. This example shows that structurally similar chemicals need not necessarily have the same apical effects, and in this case biological data are needed to prove toxicological similarity.

The call for good science and documentation in hazard assessment, that was made by Russell and Burch,2 is as relevant now as it was in 1959. Indeed, guidance documents and reporting templates have undergone several refinements,21–23 strategies have been published,11,24,25 and recently, the ECHA has published the Read-Across Assessment Framework (RAAF).1 The latter aims at the quality control and transparency of read-across evaluations. It provides structure, and ensures that all relevant elements are addressed and will lead to a conclusion on whether or not a read-across is scientifically acceptable.

Documentation and justification for a read-across approach, in a form that it is sufficient and immediately understandable for an independent reviewer, is both challenging and time consuming. It is a considerable cost factor, which is easily underestimated in the preparation of registration dossiers. In addition, a letter of access, granting the rights to use the experimental data on read-across substances, must be available. In cases where more than one study are needed, the costs for getting the rights to refer to all read-across studies may match, or even exceed, the cost of a new study. In a favourable situation, the data on the read-across substances are already owned by one of the registrants, or they have been published in sufficient detail in a peer-review journal. In this case, refusal of a read-across assessment upon evaluation is much less costly, as compared to the situation where registrants have paid a competing company for a letter of access to now-useless
read-across studies.

Read-across approaches rely on existing experimental data on potential read-across source substances. Both the generation of new data and their dissemination via the ECHA website continue to provide opportunities for read-across. Most importantly, IT tools facilitate the identification of analogues and the easy display of existing data. The most sophisticated tool in this regard is the OECD QSAR toolbox,26 but already, simpler search tools such as eChemPortal27 permit a quick search for potential read-across candidates.

Read-across has found its way in other modern chemical legislation, such as the new chemical legislations in Korea (K-REACH) and China. It helps in the hazard  assessment of new cosmetic products that are banned from animal testing in the EU. Read-across case studies are discussed at the OECD level,28 illustrating the current worldwide interest in this approach.

One of the many important points made by Russell and Burch in their 1959 book,2 is the inappropriateness of blindly taking mammalian studies as the ‘gold standard’ for human health hazard assessment. It needs to be remembered that this can also be applied to the read-across approach, since most of the experimental data on the similar chemicals are animal data. Read-across assessments predicting the outcome of animal studies may be perfect with regard to fulfilling regulatory requirements, but the ultimate aim remains human health hazard assessment.

Developing sound and well-justified read-across and grouping will be neither quick nor easy (hence it should not be termed ‘non-testing’), and it will often require fortification by ‘mode-of-action-tailored’ experimental data, in order to cover chemicals with similar early interactions, but at first glance not necessarily closely-related structures. Newly generated ‘omics’ and in vitro data addressing early (biological) effects, as well as already-existing REACH dossiers,29 SEURAT30 and Toxcast31 data, offer tools to improve read-across, based on properties closer to the hazard (the apical effect) beyond the traditional concept based only on QSARs. Established AOPs and the identification of molecular initiating events (MIEs) facilitate this use of read-across (and were, on the other hand, often identified from a set of experimental data from structurally-related chemicals). The combination of different experimental data and their relation to apical toxic effects may indeed offer the most powerful tools to advance the Three Rs. Considerations of relevant data in creating a read-across case are also used to build IATAs. Both require a sound scientific case, relevant data to support them, and awareness (and acceptance) of their limitations.

Consensus on what an acceptable read-across looks like, is emerging whilst it is in the process of being used. For this, we have to nourish and nurture the duckling — and we have to recognise when it is no longer an ugly duckling, but has matured and become
a beautiful swan (Figure 2).
Figure 1
Author for correspondence:
Dr Robert Landsiedel
Experimental Toxicology and Ecology
Ludwigshafen am Rhein

Dr Wera Teubner
BASF Schweiz AG
Product Safety


1 ECHA (2015). Read-Across Assessment Framework (RAAF) [ECHA-15-R-07-EN], 38pp. Helsinki, Finland: European Chemicals Agency. Available at: http://
(Accessed 11.11.15).
2 Russell, W.M.S. & Burch, R.L. (1959). The Principles of Humane Experimental Technique, 238pp. London, UK: Methuen. Available online: http://altweb.jhsph.
3 Berggren, E., Amcoff, P., Benigni, R., Blackburn, K., Carney, E., Cronin, M., Deluyker, H., Gautier, F., Judson, R.S., Kass, G.E.N., Keller, D., Knight, D., Lilienblum,
W., Mahony, C., Rusyn, I., Schultz, T., Schwarz, M., Schüürmann, G., White, A., Burton, J., Lostia, A.M., Munn, S. & Worth, A. (2015). Chemical safety assessment
using read-across: Assessing the use of novel testing methods to strengthen the evidence base for decision making. Environmental Health Perspectives, E-pub ahead of print [doi: 10.1289/ ehp.1409342].
4 Worth, A., Barroso, J., Bremer, S., Burton, J., Casati, S., Coecke, S., Corvi, R., Desprez, B., Dumont, C., Goul iarmou, V., Goumenou, M., Gräpel, R., Griesinger,
C., Halder, M., Janusch Roi, A., Kienzler, A., Madia, F., Munn, S., Nepelska, M., Paini, A., Price, A., Prieto, P., Rolaki, A., Schäffer, M., Triebe, J., Whelan, M., Wittwehr, C. & Zuang, V. (2014). Alternative Methods for Regulatory Toxicology — A State-of-the-art review.
Report EUR 26797 EN, 461pp. Ispra, Italy: European Commission, Joint Research Centre, Institute for Health and Consumer Protection. Available at: http://echa.
report_en.pdf (Accessed 11.11.15).
5 Cronin, M.T., Madden, J., Enoch, S.J., & Roberts, D.W. (2013). Chemical Toxicity Prediction: Category Formation and Read-across, 191pp. London, UK: Royal
Society of Chemistry.
6 Rovida, C., Alépée, N., Api, A.M., Basketter, D.A., Bois, F.Y., Caloni, F., Corsini, E., Daneshian, M., Eskes, C.,Ezendam, J., Fuchs, H., Hayden, P., Hegele-Hartung,
C., Hoffmann, S., Hubesch, B., Jacobs, M.N., Jaworska,  J., Kleensang, A., Kleinstreuer, N., Lalko, J., Landsiedel, R., Lebreux, F., Luechtefeld, T., Locatelli, M.,  Mehling, A., Natsch, A., Pitchford, J.W., Prater, D., Prieto, P., Schepky, A., Schüürmann, G., Smirnova, L., Toole, C., van Vliet, E., Weisensee, D. & Hartung, T. (2015). Integrated Testing Strategies (ITS) for safety assessment. ALTEX 32, 25–40.
7 Willett, C., Caverly Rae, J., Goyak, K.O., Minsavage, G., Westmoreland, C., Andersen, M., Avigan, M., Duché, D., Harris, G., Hartung, T., Jaeschke, H., Kleensang, A.,
Landesmann, B., Martos, S., Matevia, M., Toole, C., Rowan, A., Schultz, T., Seed, J., Senior, J., Shah, I., Subramanian, K., Vinken, M. & Watkins, P. (2013). Building shared experience to advance practical application of pathway-based toxicology: Liver toxicity mode-of-action. ALTEX 31, 500–519.
8 Anon. (2007). Regulation (EC) No 1907/2006 of the European Parliament and of the Council of 18 December 2006 concerning the Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH), establishing a European Chemicals Agency,
amending Directive 1999/45/EC and repealing  Council Regulation (EEC) No 793/93 and Commission Regulation (EC) No 1488/94 as well as Council Directive 76/769/EEC and Commission Directives 91/155/EEC, 93/67/EEC, 93/105/EC and 2000/21/EC.
Official Journal of the European Union L136, 29.05.2007, 3–280.
9 ECHA (2014). The Use of Alternatives to Testing on Animals for the REACH Regulation: Second Report under Article 117(3) of the REACH Regulation [ECHA-
14-A-07-EN], 131pp. Helsinki, Finland: European Chemicals Agency. Available at: http://echa.europa. eu/documents/10162/13639/alternatives_test_ animals_2014_en.pdf (Accessed 11.11.15).
10 Ball, N., Bartels, M., Budinsky, R., Klapacz, J., Hays, S., Kirman, C. & Patlewicz, G. (2014). The challenge of using read-across within the EU REACH regulatory framework; how much uncertainty is too much? Dipropylene glycol methyl ether acetate, an exemplary case study. Regulatory Toxicolology & Pharmacology 68, 212–221.
11 Schultz, T.W., Amcoff, P., Berggren, E., Gautier, F., Klaric, M., Knight, D.J., Mahony, C., Schwarz, M., White, A. & Cronin, M.T.D. (2015). A strategy for structuring and reporting a read-across prediction of toxicity. Regulatory Toxicology & Pharmacology 72,
12 van Ravenzwaay, B., Herold, M., Kamp, H., Kapp,  M.D., Fabian, E., Looser, R., Krennrich, G., Mellert, W., Prokoudine, A., Strauss, V., Walk, T. & Wiemer, J.
(2012). Metabolomics: A tool for early detection of toxicological effects and an opportunity for biology based grouping of chemicals — from QSAR to QBAR.
Mutation Research 746, 144–150.
13 Hagen, M., Pabel, U., Landsiedel, R., Bartsch, I., Falany, C.N., & Glatt, H. (1998). Expression of human estrogen sulfotransferase in Salmonella typhimurium:
Differences between hHST and hEST in the enantioselective activation of 1-hydroxyethylpyrene to a mutagen. Chemico-Biological Interactions 109, 249–
14 Brenna, E., Fuganti, C. & Serra, S. (2003). Enantio-selective perception of chiral odorants. Tetrahedron: Asymmetry 14, 1–42.
15 Roberts, D.W., Aptula, A.O., Patlewicz, G., & Pease, C. (2008). Chemical reactivity indices and mechanism ‐based read‐across for non‐animal based assessment
of skin sensitisation potential. Journal of Applied Toxicology 28, 443–454.
16 Teubner, W., Mehling, A., Schuster, P.X., Guth, K., Worth, A., Burton, J., van Ravenzwaay, B. & Landsiedel, R. (2013). Computer models versus reality:
How well do in silico models currently predict the sensitization potential of a substance. Regulatory Toxicology & Pharmacology 67, 468–485.
17 Urbisch, D., Mehling, A., Guth, K., Ramirez, T., Honarvar, N., Kolle, S., Landsiedel, R., Jaworska, J., Kern, P.S., Gerberick, F., Natsch, A., Emter, R., Ashikaga, T., Miyazawa, M. & Sakaguchi, H. (2015). Assessing skin sensitization hazard in mice and men
using non-animal test methods. Regulatory Toxicology & Pharmacology 71, 337–351.
18 Hendren, C.O., Lowry, G.V., Unrine, J.M., & Wiesner, M.R. (2015). A functional assay-based strategy for nanomaterial risk forecasting. Science of the Total Environment 536, 1029–1037.
19 Arts, J.H., Hadi, M., Irfan, M.A., Keene, A.M., Kreiling, R., Lyon, D., Maier, M., Michel, K., Petry, T., Sauer, U.G., Warheit, D., Wiench, K., Wohlleben, W. & Landsiedel, R. (2015). A decision-making framework for the grouping and testing of nanomaterials
(DF4nanoGrouping). Regulatory Toxicology & Pharmacology 71, Suppl. 2, S1–S27.
20 Oomen, A.G., Bos, P.M., Fernandes, T.F., Hund-Rinke, K., Boraschi, D., Byrne, H.J., Aschberger, K., Gott ardo, S., von der Kammer, F., Kühnel, D., Hristozov, D., Marcomini,
A., Migliore, L., Scott-Fordsmand, J., Wick, P. & Landsiedel, R. (2014). Concern-driven integrated approaches to nanomaterial testing and assessment — Report of the
NanoSafety Cluster Working Group 10. Nanotoxicology 8, 334–348.
21 ECHA (2008). Guidance on Information Requirements and Chemical Safety Assessment, Chapter R.6: QSARs and grouping of chemicals, 134pp. Helsinki, Finland:
European Chemicals Agency. Available at: https://
requirements_r6_en.pdf (Accessed 12.11.15).
22 ECHA (2012). Practical Guide 6 — How to report readacross and categories [ECHA-10-B-11.1-EN], 32pp. Helsinki, Finland: European Chemicals Agency. Available
at: 10162/13655/pg_
report_readacross_en.pdf (Accessed 12.11.15).
23 ECHA (2013). Grouping of Substances and Read-across Approach — An illustrative example [ECHA-13-R-02-EN], 11pp. Helsinki, Finland: European Chemicals Agency.
Available at:
13628/read_across_introductory_note_en.pdf (Accessed 12.11.15).
24 Patlewicz, G., Ball, N., Becker, R.A., Booth, E.D., Cronin, M.T., Kroese, D., Steup, D., van Ravenzwaay, B. & Hartung, T. (2014). Food for Thought: Read-across approaches — misconceptions, promises and challenges ahead. ALTEX 31, 387–396.
25 Patlewicz, G., Ball, N., Boogaard, P.J., Becker, R.A. & Hubesch, B. (2015). Building scientific confidence in the development and evaluation of read-across. Regulatory
Toxicology & Pharmacology 72, 117–133.
26 OECD (undated). The OECD QSAR toolbox. Paris, France: Organisation for Economic Co-operation & Development. Available at:
theoecdqsartoolbox.htm (Accessed 12.11.15).
27 OECD (undated). The Global Portal to Information on Chemical Substances (eChemPortal). Paris, France: Organisation for Economic Co-operation & Development.   Available at:
index?pageID=0&request_locale=en (Accessed 12.11. 15).
28 OECD (2014). Series on Testing and Assessment No. 194: Guidance on Grouping of Chemicals, 2nd edition, ENV/JM/MONO(2014)4, 141pp. Paris, France: Organisation
for Economic Co-operation & Development.
29 ECHA (undated). Registered substances. Helsinki, Finland: European Chemicals Agency. Available at: http://echa. (Accessed 12.11.15).
30 SEURAT (undated). Welcome to the SEURAT-1 website. Available at: (Accessed 11.11.15).
31 EPA (2015). Toxicity Forecasting: Advancing the Next Generation of Chemical Evaluation. Durham, NC, USA: US EPA Research. Available at:
chemical-research/toxicity-forecasting (Accessed 12.11.

Download the pdf here

A new approach to optimise the use of animal models in drug discovery through Big Data Sharing

Jiaqi Lu and Jianfei Wang

Sharing full details of animal models would enhance
consistency of models between establishments and
reduce the numbers of animals used for set-up and validation

With the continued rapid growth of Information technology (IT) and Cloud technology, the  concept of so called ‘Big Data’ has emerged over the last decade. Right now, this buzzword-phrase is being heard and talked about much more in the pharmaceutical industry and health care sectors.1–3 As a result of increased investment and better supply of information, pharmaceutical companies have recently been collating years of research and development data, including in vitro and animal preclinical test information, into medical databases. Meanwhile, the governments and other public stakeholders have been opening their vast stores of health-care knowledge, including data from clinical trials and information on patients acquired through public medical insurance programmes. 4,5 In parallel, some cross-boundary large IT companies, such as Google and Apple, have also jumped into this hot-tub and expect financial returns from the accelerating value and innovation in the health care and drug discovery industries.6,7

Historically, drug discovery and development has been a relatively isolated endeavour, with little information sharing evident among different pharmaceutical companies and academic researchers. In recent years, however, it has become apparent that pharmaceutical R &D is suffering from declining success rates and stagnant pipelines. This provides the impetus and an opportunity to change the landscape of drug discovery by utilising Big Data information. The current ability to generate and store vast amounts of information has led to an abundance of data and a growth in the discipline of Systems Pharmacology. Data are generated from several stages in the drug discovery process, including pre-clinical animal tests and Phase I, II and III trials, as well as post-marketing monitoring.8 Effectively utilising these data will help pharmaceutical companies to better identify new potential drug candidates and to develop safe, effective, approved medicines more quickly. In this article, a new approach to optimising the use of animal models in drug discovery through Big Data is explored.

Challenges and opportunities

Economic pressures, perhaps more than any other factor, are driving the demand for Big Data analysis and applications in drug discovery; this is appropriate, as it costs more than one billion US dollars to test and develop one new drug, and it often takes far more than ten years. In the early pre-clinical stage, the increasing costs of high-throughput screening of compound candidates, and of safety and efficacy studies in animal models, are major financial challenges to all pharmaceutical companies.

Animal model data are an important part of drug discovery Big Data — however, primary data generated from drug discovery animal research are infrequently accessed or re-used remotely from where they were generated. There is limited access to detailed drug discovery animal model data, since full information about the protocols has not always been published in research papers. This has led to a drive by many journals to enhance the depth of details given in the Methods sections  of manuscripts. Without this attention to detail, there have been missed opportunities for the continuous optimisation and  improvement of scientific methods and enhancement of innovation.

On the other hand, with the strengthening social pressures to avoid the use of laboratory animals in drug discovery, pharmaceutical companies and academia are finding it hard to demonstrate the application of the Three Rs principles to the satisfaction of the public. The sharing of full details of the animal models used in drug discovery would enhance the consistency of models between establishments and reduce the numbers of animals used to set up and validate the models.

Last, but not least, changing the mind set about confidentiality is a big challenge for all public and private organisations. Pharmaceutical R&D has always been a ‘secretive’ activity, conducted within the confines of the R&D department, with little external collaboration. Unless it is possible to identify an ideal future state with non-competitive aims, there is little value to investing in improving Big Data sharing capabilities. Today, public–private partnerships still represent a concept to be tested — therefore, if a new approach of sharing the full details of animal models used in drug discovery demonstrated its value and reduced attrition, then there would be an enlarged space for future developments in sharing information.

Landscape and strategy

The sharing and cross-analysis of pre-competitive drug discovery animal model information across public research organisations, pharmaceutical companies, Three Rs organisations, biotech companies and contract research organisations (CROs), through a one-stop sharepoint, would contribute to the simplification of the partners’ operating systems, would facilitate the delivery of more products of value through reduced attrition, and would enable all the partners to build trust by demonstrating their commitment to the Three Rs.

This one-stop sharepoint would allow data circulation within and beyond the original partnership. By enhancing interdisciplinary scientific reviews, animal studies could be optimised. Raw data, including cross-therapeutic animal models and protocols, drugvehicle effects, and positive and negative study results, would permit the assessment of the positive predictive value of each animal model. In addition, new information and hypotheses generated from data cross-analysis could be available to all partners, which would maximise the value of the animal research data. This non-competitive animal model information, such as guidelines on contemporary best practice and innovative alternative approaches to animal research, could also contribute to public knowledge and enhance animal welfare. In the end, by exchanging this information, all the partners would bolster external collaborations within and beyond the original partnership (Figure 1).
Figure 1

The importance of partnerships

The key component to achieving this goal is through partnership. No matter how public  research organisations, pharmaceutical companies, Three Rs organisations, CROs and biotech companies break the silos by enhancing collaboration with external partners,
all stakeholders can extend their knowledge and data networks through partnership.

Ideally, the following objectives could be achieved: identifying and discontinuing the use of animal models that are not sufficiently robust or fail to translate in the clinic; optimisation of the design and validation of animal models and protocols (e.g. to improve translation of animal models between laboratories or decrease model severity, to facilitate the informed choice of animal models, to share best practices and Three Rs advances, and to reduce duplication of efforts).

Academic partners could share insights from the latest scientific breakthroughs in cross-therapeutic animal models and make a wealth of innovation available. Normally, academia is willing to help improve the transfer of animal models between laboratories and the intra-laboratory reproducibility. Collaborations between the pharmaceutical companies could quickly identify and discontinue the use of animal models that identify treatments that fail to translate into efficacious medicines in the clinic. This could then lead to optimising the design/validation of animal models and protocols as a next step. This partnership could reduce clinical attrition, which would, in turn, reduce the financial cost of whole drug discovery process. Through collaboration with Three Rs organisations in order to learn best practices and Three Rs advances, stakeholders would enhance animal welfare by direct innovation and implementation of non-animal alternatives where they are the most needed, or by refining animal models to decrease severity and variability. Maximising external collaborations with CROs and new biotech companies could quickly add to or scale up internal capabilities and provide access to expertise in advanced technologies and animal models which would otherwise require establishment in-house.

Although this pattern of collaboration appears to be a win–win situation for all the partners, belief in the benefits of the data sharing culture and active participation still needs to be inspired. All stakeholders would have to recognise the value of Big Data analysis and sharing and be willing to act on its insights; a fundamental mind-set shift for many and one that may prove difficult to achieve. Confidentiality issues would also continue to be a major concern, although new IT technology can readily enhance private information protection in the databases. However, stakeholders would still have to be vigilant and watch for potential problems, as the increasing amount of information on animal models that is becoming openly available has the potential to the misunderstood by the general public.


Big Data sharing is a new approach to optimising the use of animal models in drug discovery. Sharing animal model information, such as protocols, study results (including drug-vehicle effects and positive and negative data) and translational outcomes, in a single cross-therapeutic platform that uses a standard data capture and common ontology framework, would permit the secondary analysis of multiple datasets.

This would lead to higher efficiency for the assessment of preclinical drug efficacy and pharmacokinetics, and would also reduce the welfare impact on animals. Ultimately, it will also facilitate the re-use of animal data in the wider and more complex scenario of drug R&D, by facilitating linkage with other datasets, such as safety assessment datasets, chemistry and pharmacokinetics and pharmacodynamics (PK/PD). Crosspharma
collaborations in animal research are identified as a high-impact opportunity for accelerating scientific innovation and improving scientific output in animal model research for drug discovery, and for more tangible contributions to the Three Rs ethical principles across all the pharmaceutical industry. Subsequently, animal test dataset sharing across multiple pharmas and some prominent CROs, would further permit the appropriate assessment of the value of animal use in drug discovery and would lead to a reduction in the numbers of animals used in this work. Finally, if the pool of animal datasets generated by pharmas/CROs were ultimately augmented by the experimental data from many prominent academic institutions, it would be possible to generate an animal test search engine similar to Google Scholar. In an ideal situation, any proposal to carry out an animal test or use a particular animal model, should start with a search to identify suitable assays and an assessment of the potential utility of the model. This could be done by viewing existing positive and negative animal test data, as well as easily contacting partners experienced in these assays for advice. In addition, further animal model optimisation could be performed or unnecessary animal tests could be prevented. This would ultimately reduce animal use and reduce drug discovery costs and would speed up the drug discovery process by affording a greater chance of successful translation of efficacy to the clinic.


This work was supported by Key Projects in the National Science & Technology Pillar Program (No.  2011BAI15B03). The authors wish to thank Dr David Tattersall and Cheng Gao for their valuable suggestions.
Dr Jiaqi Lu
Department of Laboratory Animal Sciences
GlaxoSmithKline, R&D China

Author for correspondence:
Dr JianFei Wang
Head, Laboratory Animal Sciences
GlaxoSmithKline, R&D China
2F, Building 3
898 Halei Road
Zhangjiang Hi-Tech Park
Shanghai 201203


1 Fabricio, F.C. (2014). Big data in biomedicine. Drug Discovery Today 19, 433–440.
2 Marx, V. (2013). The big challenges of big data. Nature, London 498, 225–260.
3 Groves, P., Kayyali, B., Knott, D. & Kuiken, V.S. (2013).   The ‘big data’ revolution in healthcare: Accelerating value and innovation, 19pp. Center for US Health System Reform, Business Technology Office, McKinsey & Company.
4 SOTP (2012). Obama Administration Unveils “Big Data” Initiative: Announces $200 Million in New R&D Investments, 4pp. Washington, DC, USA: Office of Science & Technology Policy, Executive Office of the President, White House. Available at: https://www. _ data_press_release_final_2.pdf (Accessed 19.09.15).
5 NIH (2015). NIH-led effort launches Big Data portal for Alzheimer’s drug discovery. Bethesda, Maryland, USA: National Institutes of Health. Available at: http:// (Accessed 19.09.15).
6 Harris, D. (2015). Google, Stanford say big data is key to deep learning for drug discovery. Houston, TX, USA: Knowingly, Corp. Available at: 2015/03/02/google-stanford-say-big-data-is-key-todeep-learning-for-drug-discovery/ (Accessed 19.09.
7 Anon. (2015). Apple HealthKit. Cupertino, CA, USA: Apple Inc. Available at: healthkit/ (Accessed 19.09.15).
8 Zhang, J., Hsieh, J.H. & Zhu,  H. (2014). Profiling animal toxicants by automatically mining public bioassay data: A Big Data approach for computational toxicology.
PLoS One 9, e99863.

The ‘Genomic Revolution’ and its Impact on Medical Research

Rehma Chandaria

The sequencing of the human genome held great promise for better
disease treatment and a concomitant reduction in animal experiments.
Neither promise has been adequately fulfilled.

In 1953, four scientists — James Watson, Francis Crick, Maurice Wilkins and Rosalind Franklin — uncovered the double helix structure of DNA, the molecule that carries genetic information in all living beings. In the 60 years since then, the ways in which we study and manipulate genes has changed considerably. In 2001, the human genome was sequenced, providing hope that common and serious conditions, such as Alzheimer’s disease and heart disease, could be predicted before they manifested clinically, and that personalised treatments for many other human conditions could be found. However, a decade on from this historic scientific achievement,  questions were raised about the lack of real clinical progress made from this vast genetic knowledge.1,2 In part, this is due to unrealistic expectations and the difficulties involved in translating discoveries made in the laboratory to clinical products that can benefit patients.

The sequencing of the human genome and better technologies for studying human genetics came with the expectation that we would no longer have to rely on animals to learn about human diseases and to find treatments. However, in reality, there has been a sharp increase in the use of genetically-modified animals, which has been responsible for the overall increase in animal procedures undertaken in the UK3 and elsewhere, in recent years.

So what actually has been achieved as a result of the ‘genomic revolution’, and what future progress can we realistically expect to see? Furthermore, is the use of animals still justified in this genomics era?

Some examples of successes arising from the genomic revolution

New research technologies

Since 1953, and particularly accelerated after the launch of the Human Genome Project in 1990, there have been significant advances in the research techniques used to study genes. Some particular highlights include the development of the Polymerase Chain Reaction (PCR) in 1985, and progress made in sequencing methods which means that the precise order of bases in DNA can now be obtained quickly and efficiently. PCR is still routinely used by researchers across the globe, to amplify a small amount of DNA to  generate millions of copies of a particular gene.

Understanding and treating genetically inherited diseases

The individual genes responsible for causing many inheritable diseases have been identified as a result of the genomic revolution. Consequently, gene therapy is now a possibility, where a faulty gene is corrected by inactivation or replacement with a functional version of the gene. In 2012,  Glybera became the first gene therapy treatment to be approved  in Europe.4 The treatment involves the delivery of an intact copy of the faulty gene in the rare inherited genetic disorder, lipoprotein lipase deficiency. Many other gene therapy treatments are in human clinical trials, although it is still an experimental
technique with several hurdles to be overcome before it can be considered as a widespread treatment option.


Cancer usually occurs as a result of mutations in the genetic sequence,5 and therefore diagnosis and personalised treatment of cancer is one of the primary successes of the genomic revolution. The genome sequencing of different cancers has revealed common mutations and patterns which determine how cancers develop.6 Mutations in genes such as BRCA1 have been identified as increasing the risk of developing breast cancer, meaning that preventative action, such as surgery to remove the breasts, can be offered to carriers of these mutations. Treatments have been developed against cancers with particular genetic profiles. For example, trastuzumab (Herceptin) is a drug mainly used to treat breast cancer, but it is only effective against cancers which overexpress the HER2 protein.7 Another example of this selective treatment is the drug gefitinib, which is only active against the 10–15% of lung cancers that carry EGFR mutations.8

Why have more treatments not been developed since the human genome was sequenced?

Of the 14,000 genes sequenced to date, only around 3800 are associated with a genetic condition that arises when the gene is faulty.9 The most common human diseases are not caused by a single gene defect, but are associated with multiple genes, each one having a small contributing effect. Other important contributing factors include influences such as diet, gender, ethnicity, age and environment, and epigenetics. Epigenetics refers to changes to the DNA molecules and the protein with which the DNA is associated to form chromatin. These epigenetic changes affect whether particular genes are switched on or off. Although every cell in the body of an organism has identical DNA, due to epigenetic modifications the characteristics of a liver cell, for example, are very different to those of a nerve cell. In addition, DNA interacts with other molecules in cells, such as RNA, signalling molecules and receptors. These complicating factors mean that someone can appear to have the genetic features of a disease such as diabetes mellitus, but not express it clinically. Therefore the idea of predicting and preventing diseases based on a person’s genetic make-up continues to be much further away than initially was hoped.

Genetically-altered animals

Since the human genome was sequenced, there has been a large increase in the use of genetically altered animals, and in particular mice, with the aim of learning about the roles of individual genes in whole animals. A genetically-altered animal has been bred or engineered to mutate, remove or insert a gene of interest. However, as the complexities already described above suggest, it is not always possible to find out the role of a gene in humans just by altering it in a mouse. Interactions with the environment and other components in a cell mean that the effects can vary greatly between species.10 In  fact, there are even examples of gene knockouts having different effects in different strains of mice.11 Therefore, we have to question whether data from genetically-altered mice can really provide insights into complex genetic interactions in humans. This is particularly relevant because of the inefficient process by which the desired genetic alterations are achieved, which results in large numbers of animals being bred but not used in experiments. There are several steps in the breeding procedure at which surplus animals are produced. The gene of interest (transgene) is constructed by using recombinant DNA technology, and then inserted into embryos which are implanted into female mice with the hope that they become pregnant and give birth to mice carrying the transgene. However, only 25% of the animals will contain the transgene, and even fewer will prove to be satisfactory for further study.12 Furthermore, extensive cross-breeding and back-crossing is required to produce animals with a homozygous genetic background in which both copies of the gene are correctly modified. The majority of the animals produced in these steps will not contain the correct genotype, are therefore considered surplus to requirements and are usually killed.13 Data from The Netherlands indicate that the number of animals that are bred but not used is almost equal to the number of animals used in experiments.14 This number of surplus animals continues to increase, due to the ongoing rise in the use of genetically-altered animals.

There are different levels of understanding human biology: the whole body, which has historically been studied by using non-human animals; tissue and organs; the cellular level; and the molecular level — which is where genomics is focused. To fully grasp how diseases develop, and to find new treatments, we need to understand the processes that happen at every level, and not just molecular processes. Despite the great successes of the genomic revolution, it is important to realise that work on genes is not the be all and end-all of biomedical research. There are complex systems and processes that occur in the body which are not related to genetics. Examples of this at the cellular and tissue levels are post-translational modifications, which occur to proteins after  they have been produced, and interactions with the extracellular matrix that surrounds cells, providing mechanical and biochemical cues that influence how cells behave. With this in mind, it is crucial to maintain efforts in all areas and at all levels of biology, rather than looking at genes and genomics in isolation. It is also vital that this is done by using scientifically- valid approaches, which provide results that are relevant to humans. This will not be achieved by using animals, and especially not genetically-modified animals. Very exciting progress has been made by biologists who have successfully used mathematical approaches to integrate data provided by in vitro techniques and thereby usefully simulate and predict physiological responses in vivo.15 This suggests that a more-complete understanding of human disease can be obtained by using human tissues and by investing in research across different levels.


From improvements in laboratory research technologies, to increased knowledge of genetic contributions to disease and genetic therapies, much has been achieved as the result of the genomic revolution. However, the resulting clinical outcome seen by patients has been much smaller than was anticipated, due to difficulties in translating research into actual treatments. This is due to the complex nature of how genes interact with other factors contributing to disease, as well as an over-reliance on animal models that are inherently different to the humans we are trying to treat. However, if the scientific community is able to turn its focus to human-based research, then the future prospects for new drugs and therapies to improve human health will be much greater.

Rehma Chandaria
Russell & Burch House
96–98 North Sherwood Street
Nottingham NG1 4EE


1 Evans, J.P., Meslin, E.M., Marteau, T.M. & Caulfield, T. (2011). Genomics. Deflating the genomic bubble. Science, New York 331, 861–862.
2 Marshall, E. (2011). Human genome 10th anniversary. Waiting for the revolution. Science, New York 331, 526–529.
3 Hudson-Shore, M. (2014). Statistics of Scientific Procedures on Living Animals 2013: Experimentation continues to rise — the reliance on genetically altered animals must be addressed. ATLA 42, 261–266.
4 Ylä-Herttuala, S. (2012). Endgame: Glybera finally recommended for approval as the first gene therapy drug in the European Union. Molecular Therapy 20, 1831–1832.
5 Loeb, K.R. (2000). Significance of multiple mutations in cancer. Carcinogenesis 21, 379–385.
6 Feng, H., Wang, X., Zhang, Z., Tang, C., Ye, H., Jones, L., Lou, F., Zhang, D., Jiang, S., Sun, H., Dong, H., Zhang, G., Liu, Z., Dong, Z., Guo, B., Yan, H., Yan, C., Wang, L., Su, Z., Li, Y., Nandakumar, V., Huang, X.F., Chen, S.Y. & Liu, D. (2015). Identification of genetic mutations in human lung cancer by targeted sequencing. Cancer Informatics 14, 83–93.
7 Yu, D. & Hung, M.C. (2000). Overexpression of ErbB2  in cancer and ErbB2-targeting strategies. Oncogene 19, 6115–6121.
8 Paez, J.G., Jänne, P.A., Lee, J.C., Tracy, S., Greulich, H., Gabriel, S., Herman, P., Kaye, F.J., Lindeman, N., Boggon, T.J., Naoki, K., Sasaki, H., Fujii, Y., Eck, M.J., Sellers, W.R., Johnson, B.E. & Meyerson, M. (2004). EGFR mutations in lung cancer: Correlation with clinical
response to gefitinib therapy. Science, New York 304, 1497–1500.
9 Barlow-Stewart, K. (2012). The human genetic code — the human genome project and beyond. Fact Sheet 24, 6pp. St Leonards, NSW, Australia: Centre for Genetics Education. Available at: http:// www. and Resources/
Genetics-Fact-Sheets/TheHumanGeneticCodeThe HumanGenomeCodeandBeyondFS24 (Accessed 30.09. 15).
10 Van Zutphen, L.F. (2000). Is there a need for animal models of human genetic disorders in the postgenome era? Comparative Medicine 50, 10–11.
11 Pearson, H. (2002). Surviving a knockout blow. Nature, London 415, 8–9.
12 Smith, K.R. (2002). Animal genetic manipulation — a utilitarian response. Bioethics 16, 55–71.
13 Combes, R.D. & Balls, M. (2014). Every silver lining has a cloud: The scientific and animal welfare issues surrounding a new approach to the production of transgenic animals. ATLA 42, 137–145.
14 Hendriksen, C. & Spielmann, H. (2014). New techniques for producing transgenic animals — a mixed blessing from both the scientific and animal welfare perspectives. ATLA 42, 93–94.
15 Butcher, E.C., Berg, E.L. & Kunkel, E.J. (2004). Systems biology in drug discovery. Nature Biotechnology 22, 1253–1259.

The Use of 3-D Models as Alternatives to Animal Testing

Hajime Kojima

A number of three-dimensional in vitro models are now available,
but significant further developments are needed before their routine
and widespread use as alternatives to animal testing will be possible

Download a pdf of this article

The development and validation of new ex vivo and in vitro test methods are urgently needed, in order to expand the use of alternatives to animal testing worldwide.  A number of such tests are already used for screening in a wide range of pharmaceutical developments, as well as in toxicological testing for regulatory purposes. These in vitro models are not commonly used, however, except to evaluate local toxic and genotoxic effects. Other toxicological fields currently utilise fish and other animals for testing, rather than in vitro or other non-animal alternatives.

I personally am hoping for the development of new ex vivo and in vitro test methods, because they are correlated with the successful development and application of regenerative medicine and tissue engineering. One important element of this research that has made significant progress is the development of novel cell types, such as cell lines, primary cultured cells, embryonic stem (ES) cells, induced pluripotent stem (iPS) cells, and mesenchymal stem cells (MSCs). There are, however, various limitations inherent in the use of cultured monolayer cells, which is why much work is currently under way in the development of three-dimensional (3-D) cell culture models. The 3-D models are superior to monolayer culture models in promoting higher levels of cell differentiation and tissue organisation, and being more appropriate because of the flexibility of the ECM (extracellular matrix) gels used, which can accommodate shape changes and intracellular connections. Rigid monolayer culture substrates are not capable of this, which is why they are not suitable for properly assessing the modes of action of medicines, toxicants and other substances.

Another important element is the development of new biomaterials for use as scaffolds for effecting proper intercellular connections. These take the form of collagen gels, spheroids and fibres, and they are fundamental for good 3-D models, which not only rely on the cells, but also on the use of the proper biomaterials. Also, at present it is difficult, if not impossible, to effect the adequate exposure of monolayer cells to substances that are not readily soluble in culture medium. Many researchers expect that 3-D models will provide a solution to such issues.

In this report, I would like to outline the current status of this research, together with both the limitations and the future potential that 3-D models represent for the development of non-animal test methods.

Recent trends in 3-D models


As early as 1970, Thomas et al. reported on the modelling of organs by using animal cells.1 Since then, many researchers have attempted to culture the liver, kidney, heart, blood vessels and various other organs, by using animal or human cells.2 Most of these
models were surrogates for external organs — including human dermis, epidermis, full-thickness and pigmented epidermis models — and a number of them are now commercially-available worldwide3, 4 for use in safety assessment and efficacy testing. These models are useful both for dermal research and for the safety assessment of skin corrosion, skin irritation and dermal absorption. The human pigmented epidermis model is used extensively in the cosmetics industry, to evaluate the whitening efficacy of new cosmetic ingredients.

Other models include the human ocular or corneal epithelium, oral epithelium, conjunctival epithelium, gingival epithelium, vaginal epithelium, bladder epithelium, intestinal epithelium, colon epithelium, alveolar epithelium,  vasculogenesis/angiogenesis5 and cardiovascular models,5 several of which are also commercially available,3,4 and are used worldwide in research and for toxicological safety assessments. The alveolar epithelium model,6 in particular, is used to assess the effects of nanoparticles, which increasingly appear in industrial products and are considered a potential cause of respiratory toxicity in humans.

There is also a significant amount of research on 3-D models of hepatocytes, based on biomaterials such as collagen gels, spheroids and fibres. Primary hepatocytes or cell lines derived from the liver are useful for studying long-term culture effects, the maintenance of functional structure, and the functional expression of the human liver. Similar liver models from a variety of animal species are being considered for use in pharmaceutical screening.

The regulatory use of 3-D models

The current Organisation for Economic Co-operation and Development (OECD) Test Guidelines (TGs) address human health hazard endpoints for skin corrosion, skin irritation, and eye irritation following exposure to a test chemical. These TGs describe in vitro procedures for identifying chemicals (substances and mixtures) not requiring classification and labelling for local toxicological damage, in accordance with the UN Globally Harmonised System of Classification and Labelling of Chemicals (GHS):7
— TG428: Skin Absorption: In Vitro Method8 This TG describes an in vitro procedure that has been designed to provide information on absorption of a test substance, ideally radio-labelled, that has been applied to the surface of a skin sample separating the donor chamber and receptor chamber of a diffusion cell. Static and flow-through diffusion cells are both acceptable for use in this assay. Skin from human or animal sources can be used. Although viable skin is preferred, non-viable skin can also be used. The absorption of a test substance during a given time period (normally 24 hours) is measured by analysis of the receptor fluid and the distribution of the test chemical in the test system; the absorption profile over time should be presented.
— TG430: In Vitro Skin Corrosion: Transcutaneous Electrical Resistance Test Method (TER)9 This TG describes an in vitro procedure that is useful for identifying non-corrosive and corrosive substances and mixtures, based on the rat skin transcutaneous electrical resistance (TER) test method. The test chemical is applied to three skin discs for a duration not exceeding 24 hours. Corrosive substances are identified by their ability to produce a loss of normal stratum corneum integrity and barrier function, which is measured as a reduction in the TER below a threshold level (5kΩ for rats). A dye-binding step incorporated into the test procedure permits the determination of whether or not increases in ionic permeability are due to physical destruction of the stratum corneum.
— TG431: In Vitro Skin Corrosion: Reconstructed Human Epidermis (RhE) Test Method10 This TG describes an in vitro procedure that is useful for identifying non-corrosive and corrosive substances and mixtures, based on a 3-D human skin model which reliably reproduces the histological, morphological, biochemical, and physiological properties of the upper layers of human skin, including a functional stratum corneum. The procedure with reconstituted human epidermis is based on the principle that corrosive chemicals are able to penetrate the stratum corneum by diffusion or erosion, and are cytotoxic to the underlying cell layers. Cell viability is measured by enzymatic conversion of the vital dye MTT (3-[4,5-dimethylthiazol- 2-yl]-2,5-diphenyltetrazolium bromide; yellow tetrazole) into a blue formazan salt that is quantitatively measured after extraction from the tissues (the MTT assay). Corrosive substances are identified by their capacity to reduce cell viability below the defined threshold.
— TG439: In Vitro Skin Irritation — Reconstructed Human Epidermis Test Method11 This TG describes an in vitro procedure that is useful for hazard identification of irritant chemicals (substances and mixtures) in accordance with GHS Category 2. It is based on reconstructed human epidermis (RhE), which in its overall design closely mimics the biochemical and physiological properties of the upper parts of the human skin. Cell viability  is measured by using the MTT assay. Irritant test chemicals are identified by their ability to decrease cell viability below defined threshold levels (below or equal to 50% for GHS Category 2). There are four validated test methods that conform to this TG. The use of this model in phototoxicity testing is described in the ICH (International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use) Test Guideline S10.12
— TG492: Reconstructed Human Cornea-like Epithelium (RhCE) Test Method for Identifying
Chemicals Not Requiring Classification and Labelling for Eye Irritation or Serious Eye
Damage13 This TG describes a test method for identifying chemicals that do not require classification and labelling for eye irritation or serious eye damage, by using a reconstructed human cornea-like epithelium (RhCE). This tissue construct closely mimics the histological, morphological, biochemical and physiological properties of the human corneal epithelium. The purpose of this TG is to describe the procedures used to evaluate the eye hazard potential of a test chemical, based on its ability to induce cytotoxicity in the RhCE tissue construct, as measured by using the MTT assay.

Future potential and limitations of the 3-D models

Future potential

A range of TGs describing test methods that use epidermal and/or ocular models are already available worldwide for regulatory use. The quality of the procedures that use these models is maintained by the suppliers. TG428 includes the use of an ex vivo skin model for assessing the effects of exposure to chemicals. In the future, in vitro full-thickness skin, intestine and alveolar models are expected to be used for assessing the effects of exposure to chemicals. It is absolutely necessary for these models to evaluate absorption at the threshold of the  physiologically based toxicokinetic (PBTK) model. On the other hand, I expect new developments for the hepatocyte model. Since 1997, the European Medicines Agency (EMA) and US Food and Drug Administration (FDA) Guidelines14,15 have required a CYP (cytochrome P450) induction assessment for new pharmaceuticals. However, human CYP induction for the safety assessment of a broad spectrum of test chemicals (e.g. cosmetics, food additives, pesticides, mixtures) is currently not systematically addressed by any OECD TG. Despite this shortcoming, the induction of CYP enzymes in monolayer hepatocytes by drugs, and the potential of 3-D models for use in this type of study, are receiving attention from researchers.

Furthermore, ‘human-on-a-chip’ and ‘organ-on-a-chip’ research focuses on in vitro human organ constructs for the heart, liver, lung and the circulatory system in communication with each other. The goal is to assess effectiveness and/or toxicity of drugs in a way that is relevant to humans and their ability to  process these pharmaceuticals. The 3-D culture models fail to mimic the cellular properties of organs in many aspects, including cell-to-cell interfaces or the complete organ as a whole. The application of microfluidics in organ-on-a-chip methodologies provides
for the efficient transport and distribution of  nutrients and other soluble items throughout the viable 3-D tissue constructs. Organs-on-chips are referred to as the ‘next wave’ of 3-D cell culture models, that mimic the whole living biological activities of organs, and their dynamic mechanical properties and biochemical functions.


Unfortunately, the current models need significant further developments, and most of them are constructed with only one cell type. Therefore, their construction and functions are not comparable to ex vivo models. I hope for further advances in these areas, particularly because 3-D epithelium models have advanced very little over the past decade. I expect the development of 3-D models of a wide variety of cell types to be achieved, and that a model constructed with differentiated cells (including different types of stem cells) will be produced in the near future — for example, a full-thickness skin model that includes melanocytes, Langerhans cells and hair follicle cells derived from stem cells. In addition, the toxicological biomarker for all of the current 3-D models, and the one that is accepted in the OECD TGs, is cytotoxicity. Actually, cytotoxicity is one biomarker, but I do not consider this to be a specific biomarker based on mode of action. Like specific CYP enzymes, specific toxicological biomarkers for each developed organ should be used. The economic viability of developing a wide variety of small-scale 3-D models remains precarious. A production platform that enhances efficiency is needed. The long lead-time required to prepare 3-D models is another factor that drives up costs. Further study of 3-D modelling by using cells differentiated from ES or iPS cells is impossible without the development of quality control criteria for the system used to differentiate the cells. It is difficult to coordinate the longterm maintenance of 3-D models with combinations of cells, and it will be necessary to co-culture with organ-derived substances and reconstructed blood vessels, in order to promote the development of humans-on-chips or organs-on-chips.

Dr Hajime Kojima
National Institute of Health Sciences
1-18-1 Kamiyoga
Tokyo 158-8501


1 Thomas, J.A. (1970). Organ Culture, 512pp. New York, NY, USA: Academic Press.
2 Antoni, D., Burckel, H., Josset, E. & Noel, G. (2015).Three-dimensional cell culture: A breakthrough in vivo. International Journal of Molecular Sciences 16, 5517–5527.
3 MatTek (2015). In Vitro Tissue Models. Ashland, MA, USA: MatTek Corporation. Available at: (Accessed 30.07.15).
4 Xenometrix (2015). Homepage. Allschwil, Switzerland: Xenometrix AG. Available at: (Accessed 30.07.15).
5 Heinonen, T. (2015). Better science with human cellbased organ and tissue models. ATLA 43, 29–38.
6 Jamin, A., Sr (2015). Predicting respiratory toxicity using a human 3D airway (EpiAirway™) model combined with multiple parametric analysis. Applied In Vitro Toxicology 1, 55–65.
7 Anon. (2015). Globally Harmonised System of Classif ication and Labelling of Chemicals. [In Japanese.] Tokyo, Japan: Ministry of Health, Labour & Welfare. Available at: (Accessed 30.07.15).
8 OECD (2004). Test Guideline No. 428: Skin Absorption: In Vitro Method, 8pp. Paris, France: Organisation for Economic Co-operation and Development. Available at:
health-effects_20745788 (Accessed 30.07.15).
9 OECD (2015). Test Guideline No. 430: In Vitro Skin Corrosion: Transcutaneous Electrical Resistance Test Method (TER), 20pp. Paris, France: Organisation for Economic Co-operation and Development. Available at: 739-en;jsessionid=4dc4tau8sk8k1.x-oecd-live-03 (Accessed 27.08.15).
10 OECD (2015). Test Guideline No. 431: In Vitro Skin Corrosion: Reconstructed Human Epidermis (Rhe) Test Method, 33pp. Paris, France: Organisation for Economic Co-operation and Development. Available at:; jsessionid=4dc4tau8sk8k1.x-oecd-live-03 (Accessed 27.08.15).
11 OECD (2015). Test Guideline No. 439: In Vitro Skin Irritation: Reconstructed Human Epidermis Test Method, 21pp. Paris, France: Organisation for Economic Co-operation and Development. Available at:
439-in-vitro-skin-irritation-reconstructed-humanepidermis- test-method_9789264242845-en;jsessionid =4dc4tau8sk8k1.x-oecd-live-03 (Accessed 27.08.15).
12 US FDA (2014). S10 Photosafety Evaluation of Pharmaceuticals: Guidance for Industry, 21pp. Silver Spring, MD, USA: US Department of Health and Human Services, Food and Drug Administration. Available at: complianceregulatoryinformation/guidances/ucm337572.pdf#search=’ICH+S10 (Accessed 30.07.15).
13 OECD (2015). Test Guideline No. 492: Reconstructed Human Cornea-like Epithelium (RhCE) Test Method for Identifying Chemicals Not Requiring Classification
and Labelling for Eye Irritation or Serious Eye Damage, 25pp. Paris, France: Organisation for Economic Co-operation and Development. Available at:;jsessionid =4dc4tau8sk8k1.x-oecd-live-03 (Accessed 27.08.15).
14 US FDA (2012). Guidance for Industry Drug Interaction Studies — Study Design, Data Analysis, Implications for Dosing, and Labeling Recommendations: Draft Guidance, 79pp. Silver Spring, MD, USA: US Department of Health and Human Services, Food and Drug Administration. Available at: downloads/drugs/guidancecomplianceregulatory
information/guidances/ucm292362.pdf (Accessed 30. 07.15).
15 EMA (2012). Guideline on the Investigation of Drug Interactions, 59pp. London, UK: European Medicines Agency.

Development and Validation of a Low-fidelity Simulator to Suture a Laparotomy in Rabbits

Juan J. Pérez-Rivero, Tonantzin Batalla-Vera and Emilio Rendón-Franco

An easily constructed, low-cost simulator
is assessed for its efficacy in the surgical training
of veterinary science undergraduates

Download a pdf of this article


There is a growing need for the development of alternatives to reduce, replace and refine  the use of animals for surgical training in contemporary veterinary education at the undergraduate level. In the present study, a simulator to suture a midline laparotomy in the rabbit was designed, that could be constructed from widely-available and low-cost materials. The simulator was used to develop surgical skills in students at the undergraduate level of veterinary medicine. Thirty-five, third-year veterinary students, with no previous surgical experience, were divided into two groups: a control group that did not use the simulator (n = 19), and an experimental group that used the simulator three times to practise the suturing of a laparotomy (n = 16). Later, both groups performed
a midline laparotomy in an anaesthetised rabbit, and the rate of closure of each anatomical plane (peritoneum, additional reinforcement, and skin) was measured.

The usefulness of simulators

The surgical training of undergraduate students by using live animals provides few opportunities for real training and is applicable only to certain surgical techniques. In addition, it also raises serious ethical and animal welfare considerations. The students themselves are also subjected to a level of stress, this being, in most cases, a cause of errors. Consequently, they do not adequately benefit from the training provided.1,2

In veterinary medicine and animal sciences, the Three Rs principles are being implemented as widely as possible. This involves the reduction, replacement and refinement of animal use, both in experiments and in teaching.3 One way to accomplish this is through the use of various simulators in their different forms, such as synthetic simulators, multimedia simulations, virtual reality, carcasses, and ethically sourced animal tissues.4,5 These provide training alternatives, which permit the acquisition of skills to successfully meet the needs of future clinical and surgical experiences with live patients, and to ensure that maximum educational value is achieved during practical training.6

The fidelity of a simulator is determined by how much realism is provided through characteristics such as visual cues, touch, the ability to feedback, and interaction with the student. In general, simulators can be divided into two groups: high-fidelity simulators, which are usually highly technical, detailed and realistic; and low-fidelity simulators, which have a low level of realism, are usually made with widely available and low-cost materials, are often portable, and can be used on a table. Despite their simplicity, the latter group of simulators assist the development of psychomotor skills.7 Some limitations of the use of simulators are related to their cost or difficulty in sourcing spare parts. Moreover, despite the large number of simulators that have been developed, few studies have been conducted to evaluate their effectiveness, leaving the concept of teaching through simulators at an empirical stage.8 Therefore, it is necessary to develop inexpensive, easy-to-construct simulators that support the process of surgical teaching, and also to quantitatively assess the effectiveness of their use. Therefore, the aim of this work was to develop and validate a low-fidelity simulator to assist in the teaching of the correct technique for closing a rabbit midline laparotomy.
Simulator assembly

Development of the simulator

A 10cm long and 4cm wide opening was made in an empty plastic 500ml solution bottle (Pisa Agropecuaria, Guadalajara, Jalisco, Mexico), leaving protruding areas to represent both the xiphoid process and the pubic symphysis (Figure 1a). To give support to the bottle, an internal  cardboard lining was added, as well as three 3ml syringes widthwise (Figure 1b). Two 3mm thick silicone sheets were made by pouring 270ml of PE53® silicone rubber (Poliformas Plasticas, Mexico City, Mexico) into a mould, 23cm long by 13cm wide, which was allowed to set at room temperature for 24 hours.

The back-board from a standard paper clipboard was used as the simulator base, with the plastic bottle placed onto the board and the first sheet of silicone overlaid, in order to simulate the peritoneum (Figure 1c). Subsequently, the rectus abdominis muscles were simulated by placing two sheets of 3mm thick × 28cm long × 21cm wide polypropylene around the bottle (Foamy; Mylin, Mexico City, Mexico), leaving a gap of 3cm in width along the entire midline. Finally, this layer was covered with the second sheet of silicone to simulate skin, and both sheets of silicone were tightened onto the clipboard base with paper clips (Figure 2a). The appropriate size head, thorax and abdominal organs were fashioned from cotton fabric and added to the simulator prior to use (Figure 2a).
General view

Simulator validation

Thirty-five students in the third year of a veterinary medicine and zootechnics course at the Universidad Autónoma Metropolitana, Unidad Xochimilco, with no previous experience in surgery, received a 120-minute theory session, supported with slides, on the midline laparotomy technique and suture in rabbits.9 This was part of the Surgical-Veterinary Therapeutic Bases module. Later, the students were divided into two groups: the experimental group (n = 16), which  was organised in four surgical teams of four participants each, and the control group (n = 19), which was divided in four groups of four participants and one three-participant group. Each student was assigned his/her rotation within the group, in such a way that they all covered all the roles once (surgeon, first assistant, scrub nurse, and anaesthesiologist). Each surgeon/first assistant team (according to the assigned rotation) of the experimental group used the laparotomy simulator two days prior to the practice on the live animals. They were asked to repeat three times the following procedure: put the surgical drapes in place (Figure 2b); perform a 7cm incision, including all the layers of the simulator; suture, with continuous stitches, the first silicone layer (peritoneum), which was reinforced with inverted ‘U’ stitches; and suture, with Sarnoff stitches, the second silicone layer (skin). The first assistant was only allowed to help the surgeon in handling the surgical instruments that were used. The closure of planes was performed by using nylon 2-0 suture (Figure 3).
Simulator in use

Subsequently, the participants of both groups performed midline laparotomies on 35 clinically healthy New Zealand rabbits (Oryctolagus cuniculus), suturing midline (peritoneum) with continuous stitching, reinforcing (muscular fascia) with inverted ‘U’ stitches, and suturing the skin with Sarnoff stitches, all performed under general anesthesia, according to the method previously described by Perez-Rivero and Rendón-Franco.10 Both the control group and the experimental group performed one surgery weekly. In total, evaluations were completed in 4 weeks (i.e. one week for each participant from each team).

Since the lengths of the incisions were different in all the cases, the rate of closure of each anatomic plane and all planes in total, was calculated as follows: the length of each incision was measured (in centimetres), and this was divided by the time (in minutes) taken to complete the suturing. The result was expressed in minutes per linear centimetre of incision (MLCI). During the whole process, each group was supervised by two professors and five assistant instructors.
Table 1

Statistical Analysis

Students having the prior role of first assistant, scrub nurse, and/or anaesthesiologist, would have previously observed and/or helped in the performance of the laparotomy. This could have resulted in an improvement in their performance when participating as actual surgeons. To rule out these effects, total MLCI values were compared among the members of each group, according to whether they acted as the surgeon in week 1, 2, 3 or 4, to ascertain whether there was a significant difference in their surgical proficiency, by using the one-way ANOVA with a significant value p < 0.05.

Once the effects of previous observation and/or assistance were ruled out, the MLCI values of each individual anatomical plane and the totals were compared between the control and the experimental groups, by means of the ANOVA test (significant value p < 0.05). All tests were performed by using the PAST® program.11

Ethical and animal welfare considerations

The present protocol was approved by the Comité Interno para el Cuidado y Uso de los Animales de Laboratorio (Internal Committee for the Welfare and Use of Laboratory Animals) from the Universidad Autónoma Metropolitana Unidad Xochimilco, with
reference number DCBS.CICUAL.02.10.

Results of simulator use

Comparisons of the proficiency of group members according to the week in which they acted the role of surgeon did not show a difference (p > 0.05), supporting the idea that observation and/or assistance did not improve technique. When comparing the MLCI values of each plane as well as total MLCI values between the control group and the experimental group, all were different (p < 0.05) with a higher rate of closure for the experimental group. The MLCI values of each group, as well as their comparisons, are shown in Table 1.

The experimental group performed the three planes of laparotomy suture in 5.34 ± 1.63 minutes per linear centimetre of incision (MLCI), compared to the control group that performed it in 7.03 ± 1.77 MLCI. This difference was significant (one-way ANOVA; p < 0.05) and showed that repeating the procedure three times with the simulator improved
suturing skills in a laparotomy.



When comparing MLCI values among the participants of each group independently, and not presenting differences, it is evident that observing and/or helping during the procedure did not render psychomotor skills or abilities in the participants. The use of complementary strategies, such as the use of the simulator, is necessary for a student to develop manual dexterity and the instrument skills required for the successful application of sutures.1,6

On the other hand, the experimental group demonstrated better suture skills for the laparotomy in rabbits after performing three repetitions of the procedure in the simulator. These findings agree with those reported by Aggarwal,12 who found in his study that laparoscopic surgeons required two repetitions of a particular procedure in a simulator, in order to learn it. The simulator required them to hold the tissue, lift it up, place a clip, and then cut; for trainees, seven repetitions were required to learn to perform the same procedure. However, we have to take into consideration that this particular procedure would have a longer learning curve than performing a suture.


The results make evident the advantages of the use of simulators, when recommended as training devices for undergraduate students. However, these models should be considered as complementary tools in the teaching of surgical procedures, for they help in the acquisition of skills and abilities that lead to better performance in real patients, and eventually reduce the number of training events that require the use of live animals.13

More studies are required to determine the time and number of necessary repetitions in training with these bench simulators, in order to reach an adequate level of proficiency. Further work will also be needed to make the simulators more realistic, and to investigate ways in which to take maximum advantage of this training tool.

Author for correspondence:
Dr Juan J. Pérez-Rivero
Departamento de Producción Agrícola y Animal
Universidad Autónoma Metropolitana Unidad
Calzada del Hueso 1100
Colonia Villa Quietud
Delegación Coyoacán 04960
Mexico City


1 Langebæk, R., Eika, B., Jensen, A.L., Tanggaard, L., Toft, N. & Berendt, M. (2012). Anxiety  in veterinary surgical students: A quantitative study. Journal of Veterinary Medical Education 39, 331–340.
2 Smeak, D.D. (2007). Teaching surgery to the veterinary novice: The Ohio State University experience. Journal of Veterinary Medical Education 34, 620–627.
3 Russell, W.M.S. & Burch, R.L. (1959). The Principles of Humane Experimental Technique, 238pp. London, UK: Methuen.
4 Martinsen, S. & Jukes, N. (2005). Towards a humane veterinary education. Journal of Veterinary Medical Education 32, 454–460.
5 Kumar, A.M., Murtaugh, R., Brown, D., Ballas, T., Clancy, E. & Patronek, G. (2001). Client donation program for acquiring dogs and cats to teach veterinary gross anatomy. Journal of Veterinary Medical Education 28, 73–77.
6 Valliyate, M., Robinson, N.G. & Goodman, J.R. (2012). Current concepts in simulation and other alternatives for veterinary education: A review. Veterinarni Medicina 57, 325–337.
7 Perez-Rivero, J.J. & Rendón-Franco, E. (2012). Experience of the use of table-top simulators as alternatives in the primary surgical training of veterinary undergraduate students. ATLA 40, P10–P11.
8 Schout, B.M.A., Hendrickx, A.J.M., Scheele, F., Bemel mans, B.L.H. & Scherpbier, A.J. (2010). Validation and implementation of surgical simulators: A critical
review of present, past, and future. Surgical Endoscopy 24, 536–546.
9 Griffon, D.J., Cronin, P., Kirby, B. & Cottrell, D.F. (2000). Evaluation of a hemostasis model for teaching ovariohysterectomy in veterinary surgery. Veterinary Surgery 29, 309–316.
10 Perez-Rivero, J.J. & Rendón-Franco, E. (2014). Cardiorespiratory evaluation of rabbits (Oryctolagus cuniculus) anesthetized with a combination of tramadol, acepromazine, xylazine and ketamin3. Archivos de Medicina Veterinaria 46, 145–149.
11 Hammer, Ø., Harper, D.A.T. & Ryan, P.D. (2001). PAST: Paleontological statistics software package for education and data analysis. Paleontología Electrónica 4, 1–9.
12 Aggarwal, R., Grantcharov, T.P., Eriksen, J.R., Blirup, D., Kristiansen, V.B., Funch-Jensen, P. & Darzi, A. (2006). An evidence-based virtual reality training program for novice laparoscopic surgeons. Annals of Surgery 244, 310–314.
13 Denadai, R., Oshiiwua, M. & Saad-Hossne, R. (2014). Teaching elliptical excision skills to novice medical students: A randomized controlled study comparing low- and high-fidelity bench models. Indian Journal of Dermatology 59, 169–175.