The vision of the Institute of Cancer Research is that people may live their lives free from the threat of cancer as a life-threatening disease.
Cancer or malignant neoplasm refers to a class of diseases in which a group of cells display uncontrolled growth (in other words division beyond the normal limits), invasion (intrusion on and destruction of adjacent tissues), and sometimes metastasis (spreading to other locations in the body via lymph or blood). These three malignant properties of cancers differentiate them from benign tumors, which are self-limiting and do not invade or metastasize. Most cancers form a tumor but some, like leukemia, do not. Cancer may affect people at all ages, even fetuses, but the risk for most varieties increases with age; and cancers can affect all animals.
Cancer causes about 13% of all deaths and, according to the American Cancer Society, 7.6 million people died from cancer in the world during 2007. Where effective anticancer treatments do exist they can be very demanding on the patient.
The objective of using live animals in cancer research is to develop rapid diagnosis, better treatments for existing cancers, and an improved prognosis for patients. With this in mind, scientists engaged in experimental cancer research follow four main areas of investigation, some of which use laboratory animals. Cancer research scientists attempt to discern, detect, identify, and develop.
• To discern the biological mechanisms, scientists investigate different sites of origin in the body, why particular cancers are more prevalent in some tissues and not others, and the rate of growth and metastases of cancers.
• The detection of potential carcinogens is an important chain in the link to identify agents in the environment such as chemicals, potential carcinogenic materials, exhaust fumes from motor vehicles, and other agents that may be responsible for carcinogenesis.
• Identification of individuals at particular risk looks at epidemiological studies and historical data to determine who in the general population may be at greater risk of developing certain types of cancer. This particular area of investigation has taken an important step forward in recent years since the advent of genetic testing. Investigations may involve the examination of particular risk factors, including lifestyle (tobacco use, alcohol consumption, obesity, lack of physical activity) or genetic predisposition.
• Developing ways to cure or control clinical disease is usually achieved by improving the prognosis for patients through the use of drugs, chemotherapeutic agents, radiation therapy, and/or surgical intervention.
Laboratory rodents, usually mice and rats, have assisted scientists in the field of cancer research and it is clear that they will continue to do so. They are used as experimental models in cancer research studies only where there is a justified need and only if absolutely necessary. The Institute of Cancer Research does not use animals for research if nonanimal alternatives are available, and endeavors to set humane endpoints for all research involving laboratory animals.
Because we need to use live animals in some research programs, it is essential that these living creatures be afforded the best care at all times. Staff tasked with caring for animals in the laboratory are continually striving to improve and enhance animal husbandry and welfare. An important part of this process is the use of humane endpoints in our animal experiments.
[According to] the OECD, a humane endpoint can be defined as “the earliest indicator in an animal experiment of severe pain, severe distress, suffering, or impending death.” Investigators make use of different humane endpoints depending on the tumor model being studied in any particular animal, and try wherever possible to determine accurate, predictive, and reproducible humane endpoints.
Humane endpoints should be a consideration for all experiments involving animals, but are essential in situations that may involve suffering or death (e.g., acute toxicology, infection, cancer, or inflammatory disease). They are just one manifestation in the process of refinement of animal experiments. Humane endpoints are best used with prospective planning for their use, not ad hoc to address specific welfare concerns as they might arise.
There are several considerations in arriving at the objective assessment of pain and suffering and translating this into the appropriate endpoint in a given experiment. An important point is the requirement to continually improve our skills at observing the animals and assigning some objective values to the observations we make (usually these are based on animal behavior and physiology). We also need to know, in any given study, which observations are the most significant indicators of animal pain and suffering, and have scientific acceptance of these measurements; otherwise they become invalid and unworkable. It follows, therefore, that validation and monitoring of other study parameters are required to ensure robust predictability of the endpoint and minimal interference with the scientific objectives.
All personnel contribute to the care and welfare of the animals, and it’s important that these individuals be provided with the correct training and knowledge in order to develop their required skills, and for each to progress to a level of competency. It is impossible to recognize signs of pain, suffering, or distress in any animal if you do not, or are unable to, recognize (and have not received training to be able to do so) normal signs of good health in an animal. Training and competency are very important attributes, especially when dealing with humane endpoints. Staff development of skills is an evolving process, and a clear program of training and mentoring enhances animal welfare and staff morale.
Biomedical research encompasses all types of research including research into cancer. All research can be viewed as a giant puzzle. Humane endpoints are an important and essential part of the discovery process. Everyone in the worldwide research community has an individual role to play in creating parts of the puzzle in order to find new treatments, enhanced therapies, and ultimately attempts to cure some of our more difficult and challenging diseases, not just in the field of cancer research but in all areas.
It’s important that if we continue to use animals for experimental purposes we do this in the most humane manner at all times. All who are involved in animal research must have a clear sense of responsibility, but more importantly a strong sense of compassion for the animals in their care. By the correct use and validation of appropriate humane endpoints we will help to add important parts to the puzzle.
The topic of my presentation is humane endpoints in infectious disease. This is a very sensitive and difficult topic, which I deal with almost every day as the IACUC chair at the United States Army Medical Research Institute for Infectious Diseases (USAMRIID). The opinions I am expressing today are my own and not those of my employer, the US Army.
USAMRIID does infectious disease research on some of the most dangerous viruses and bacteria in the world, and we do this under biocontainment conditions, generally ABSL-3 and -4 conditions. USAMRIID is AAALAC accredited and, as IACUC chair I can avow emphatically, does everything under an approved animal protocol.
The study of infectious disease generally involves studies of disease pathogenesis, immune response to infection, and development of therapeutics and vaccines. Because it is ethically and morally wrong to perform clinical efficacy studies with humans, the FDA has developed the animal rule, which allows new drugs and biologic products to be tested in animals as a means to getting approval for human use. Safety testing must still occur in humans, and the animal model is critical to the success of the FDA approval process. It is necessary to understand that the animal study endpoints must be clearly related to the desired benefit in the human; generally these are related to enhancement of survival or prevention of major morbidity.
The animal welfare regulations require that procedures involving animals avoid or minimize discomfort, distress, and pain to the animals. The Public Health Service Policy states that animals undergoing chronic pain or distress should be euthanized as soon as feasible and appropriate, which leads to a discussion of humane endpoints. Death as an endpoint has always been a difficult issue in infectious disease research. Lack of reproducible animal models often leads to the use of death as an endpoint. The argument to support death as an endpoint is that euthanasia and termination of the study before scientific objectives are met compromise study results. On the other hand, the counterpoint is that progression of infectious diseases to death allows unnecessary suffering, which compromises research results.
Simply put, many animal models of infectious diseases are not clearly defined and it is difficult to reliably differentiate animals that will die from those that will recover despite showing severe clinical signs. There have been many instances in which a severely ill animal recovered from its experimentally induced infectious disease. Death of the animal is the final proof that the challenge was lethal and that the vaccine failed to protect. Therefore some investigators are reluctant to euthanize early or to use anything but death as an endpoint. An argument may be made, however, that actual physiological events are missed when death is the only criterion evaluated. The data gathered from monitoring these events can be used to develop early and humane endpoints. Additionally, for a variety of reasons including tissue autolysis, death of the animal diminishes sample collection. Therefore, it is important to consider requirements for the development of early and humane endpoints.
For an early endpoint to be acceptable, it must meet the following criteria: it must be indicative of inevitable progression to death; it must reliably differentiate the animals that will die from those that will recover despite showing severe signs of toxicity; and it must adequately mimic the death endpoint.
The benefits of humane endpoints are many and should be emphasized during the planning meetings with investigators. Specifically, the development of uniform methods to assess endpoint criteria contributes to the validity and the uniformity of the experimental data. Detailed observations of clinical signs may lead to increased discriminatory experimental power. Last but most importantly, use of humane endpoints avoids or terminates unnecessary pain and distress for the research animal.
It is very important to tailor the endpoints to each animal protocol. Different animal species react differently to the same viral or bacterial challenge. For instance, Ebola Zaire is lethal in five to seven days in cynomolgus monkeys, whereas it is lethal in seven to ten days in rhesus macaques. And in Mauritius-origin cynos, monkeypox is 100% fatal while in Chinese-origin cynos there may be only a 43% fatality. This emphasizes the importance of picking the right species and understanding the course of that disease in that species. Outcomes must also be defined; will morbidity suffice or must you go to the moribund condition?
The route of the challenge is very important. The exposure route must be similar to that anticipated in humans per the FDA. This affects the time course and pathogenesis of the disease. It may be important to challenge at two or more doses, because this can help differentiate physiological changes between survivors and nonsurvivors. The viral and bacterial strain to be used should also be considered when developing endpoints. Ebola Zaire is uniformly lethal and has a shorter time course than Ebola Reston, which is also lethal but with a prolonged time course. Finally, it is important to consider human safety when dealing with infectious diseases.
When planning endpoints, one must consider observation frequency. It is critical to set reasonable observation frequencies to ensure human safety, the least stress to the animal, and investigator compliance. The frequency should be
set to minimize stress but allow for euthanasia, sample collection, and avoidance of progression to death. It is necessary to know whether the animal is nocturnal or diurnal and whether disruption of sleep will adversely affect the study. It is also necessary to determine when to increase your observation frequencies so as not to miss critical events.
As mentioned, human safety must always be considered when dealing with infectious diseases. Promoting animal welfare by increased monitoring of animals after exposure can jeopardize human safety. Therefore, investigators and the IACUC should be encouraged to look for other, less intrusive and safer methods of monitoring the animals, such as telemetry and in-room cameras.
Rodent species present their own challenges when developing humane endpoints. Rodents are generally group housed and they are not always individually identified, making their observation difficult. Additionally, clinical signs of illness in rodents can be subtle and nondiscriminatory in nature.
It may be necessary to consider objective versus subjective endpoint criteria. It is important to use a mixture of both, but when using subjective criteria with three different people observing the animals throughout the day, they must be very adequately trained on exactly what these criteria mean—e.g., “What is ruffled fur in a mouse and should it be added to my score sheet?” The IACUC must work with investigators in the development and use of humane endpoints. In many institutions, the IACUCs have developed strong policies stating that death as an endpoint is not acceptable. The IACUC should also require the use of intervention criteria or score sheets that clearly define when the animal is to be euthanized.
As I have already stated, the IACUC should work with the investigator to determine the best schedule of animal monitoring. Personnel that monitor the animals must understand normal species behavior as well as the clinical signs expected during the course of the disease. Observation frequencies should increase as the clinical signs become more severe and these observations need to be documented.
The IACUCs must ensure that there is an available point of contact for euthanasia so that when the time comes the animal will be euthanized promptly. In fact, it may be wise to have an alternate point of contact to ensure that when the score is met and it is time for euthanasia, this happens promptly.
When the clinical course of the infectious disease is not clearly defined for the animal species, the IACUC should consider the use of a pilot study to allow for criteria development. The IACUC should use subject matter experts to assist in developing the criteria and should consider the use of analgesics for each infectious disease, animal study, or protocol.
The IACUC should review the use of observation documentation as part of its postapproval compliance monitoring. Another issue that the IACUC must discuss is whether the humane endpoints should be the moribund or morbid condition. This is a difficult issue and there is no right answer. Each study must be considered separately. Often in assessing the effectiveness of the treatment or vaccine, the moribund state is used, while the morbid state would be used if it is
not necessary to know if the animal will die as a result of the treatment or vaccine failure.
In a score sheet that we use at USAMRIID with filovirus research done in macaques, if the score is equal to or greater than 10, the animal is administered pain alleviation. If it is greater than 20 the animal is considered terminally ill and is euthanized. Exceptions require consultations with the attending vet. The use of score sheets has progressed over the years and with each experiment refinements are made to improve them.
In conjunction with the investigator we have been able to add some objective criteria; e.g., if liver enzymes double, a score of 1 is assigned, and if they triple a score of 3 is given. The hope is to avoid the moribund end state and euthanize when we see liver enzymes increase.
So these are the kinds of things going on at USAMRIID in infectious disease research. Everyone has a score sheet, and every investigator is encouraged to define criteria or do a pilot study within that protocol so that future score sheets may be developed based on these criteria.
In summary, it must be the goal of all infectious disease researchers using animals and of the IACUCs that provide oversight for these animals to develop humane early endpoints. Good science and humane animal care require nothing less.
Technologies that enable the targeted manipulation of the genome have created new opportunities to study the role and interplay of specific genes in both the regulation and function of physiological and behavioral processes and the development of pathological conditions. Through the development of new or novel animal models, these techniques enable new insights into the molecular basis of disease processes and provide opportunities to develop targeted therapeutic approaches.
Despite the potential benefits from the use of these technologies, there are ethical issues in relation to their application, some of which relate to the impact on the welfare of the animals involved. The establishment of humane endpoints is a key strategy in achieving the goal of refinement; when the use of animals is scientifically justified but where there is a risk of those animals experiencing pain or distress, applying the process by which humane endpoints are implemented and reviewed underpins an informed and strategic approach to managing such risks.
Genetically modified (GM) animal models present particular challenges when developing criteria to set humane endpoints. I will provide an overview of the animal welfare issues presented in the application of GM technologies and discuss the opportunities and challenges to applying humane endpoints when GM animal models are developed.
The development of technologies that permit the targeted manipulation of genetic material—be that by transgenesis or targeted mutagenesis—has created opportunities to explore the organization, regulation, and function of molecular processes in both normal and pathological states in ways previously not possible. Further, the application of these methods has expanded the availability of
animal models that are more accurate analogues of the underlying disease processes and hence can be used to better understand disease processes and to develop new, targeted therapies.
While the potential benefits of the use of these technologies are recognized (Royal Society 2001; NRC 2002; Nuffield Council 2005), there is continuing public disquiet about their use (Einsiedel 2005). A range of issues are being raised, including fundamental ethical questions about the use of genetic modification (GM) technologies and notions of the sanctity of life and the autonomy of the individual as well as concerns about risks to human health and the environment. The welfare of the animals involved also has been a recurring issue and has been addressed in a number of reports and guidelines (for example, Royal Society 2001; Animal Procedures Committee 2001; Dennis 2002; Robinson et al. 2003; Brown and Murray 2006; Wells et al. 2006; NHMRC 2007; CCAC 2008).
The process by which humane endpoints are developed, validated, and reviewed is a key platform in making progress toward the goal of refinement when animals are used for scientific purposes (Morton 2000; Stokes 2000). Humane endpoints are used for two complementary purposes: identifying the onset of a disease process so that early intervention is possible either to initiate treatment or to enable an early, defined endpoint in a study; or, alternatively, to determine the point when an animal’s condition has deteriorated such that its involvement in the study should be terminated.
Setting humane endpoints involves identifying potential risks and validating criteria to, first, identify specific physiological or behavioral changes associated with the animal model and, second, assess the impact on the animal in relation to both the predicted effects of the experimental treatment and general criteria to assess the occurrence of pain and distress. Thus criteria are established upon which decisions can be based and outcomes reviewed. This is an iterative process that underpins informed decision making and validates the ongoing refinement of experimental procedures. Although the same processes apply to establish humane endpoints with GM animal models, as highlighted by Dennis (2000) there are particular difficulties in these circumstances brought about primarily by the unpredictability of the effects of GM technologies on phenotypic expression.
In recent years there has been a rapid escalation in the development of new GM models. In the biomedical sciences mice are by far the species most often used, but a range of species can be involved, including zebrafish, pigs, and nonhuman primates. Further, the pace and scope of the development of new GM animal models are likely to continue for the foreseeable future, which presents logistical challenges for the effective management of these animals, especially when this involves significant numbers of animals and many lines with differing phenotypes (Comber and Griffin 2007).
Welfare issues have been identified in relation to both the methods used to produce GM animals and the resulting phenotype.
Production of GM Animals
GM animal models are produced by a number of different methods that result in reduced or enhanced expression or inactivation of a gene. The most common methods used involve (1) transgenesis, where exogenous genetic material from either the same or another species is inserted in a fertilized blastocyst by microinjection, electroporation, or a nonpathogenic viral vector and then implanted in surrogate mothers; (2) targeted mutagenesis, which results in the presence or absence of a specific gene (“knock-in” or “knockout”), which is achieved by inserting modified genetic material in cultured embryonic stem cells that are injected into a blastocyst and implanted in surrogate mothers; or (3) random or chemical mutagenesis, where animals or their gametes are exposed to mutagens that increase the rate of mutations, resulting in the production of novel single gene mutations. Only a small percentage of animals produced will carry the modified genome and significant numbers of animals may be required to produce and maintain each GM line. Consequently, relative to the number of GM animals created, significantly more are produced and culled.
A July 2003 report by a Joint Working Party on Refinement in the United Kingdom reviewed the relative advantages and disadvantages of the production of GM animals by either pronuclear injection or embryonic stem cell techniques and recommended strategies to promote both reduction in the numbers of animals involved and refinement of procedures to minimize impact (Robinson et al. 2003). The report recommended criteria to benchmark the efficacy of procedures so as to ensure production methods to maximize the potential to produce GM animals and management strategies to reduce surplus production. For each step in the process, the report recommends performance benchmarks (indicators when there is a need to review that process) and outlines possible causative factors that should be considered. Thus this report sets out current standards of good practice and provides a process to benchmark animal welfare outcomes in the context of the needs and justification for current methods.
With both these technologies, donor animals undergo various, and sometimes multiple, procedures with the risk of associated pain or distress. Strategies to manage and minimize the impact on the donors of surgical procedures, superovulation of females, and tissue biopsy for genotyping are discussed in this report. The recognition and uptake of opportunities to modify and refine these procedures will continue to play an important role in the future development and use of these methods.
While reports such as this highlight the need to be aware of the impact of these procedures on the animals involved in the production of GM animals, in
the one study to date these procedures were not shown to have a significant effect on the behavioral and physiological development of mouse progeny up to 30 weeks of age (Van der Meer et al. 2001).
GM Animal Models
GM animal models have been applied to the investigation of a range of human diseases such as diabetes, obesity, atherosclerosis, chronic heart failure, hypertension, cancer, autoimmune disease, and musculoskeletal and neurological disorders. However, not all GM animals are bred as disease models. GM animals may exhibit clinical disease but, given that the rationale behind the development of GM technologies is to tease out the role and function of individual genes or gene sequences, in many cases do not do so and that is not the intended outcome.
Wells and colleagues (2006) observed that in only a minority of GM animals are animal welfare problems evident and that, with transgenic animals where most often the purpose is to study the function of a DNA segment, adverse effects are uncommon and that for GM models developed using targeted mutagenesis (knock-in or knockout) where the purpose is to study the function of a single gene, either embryonic death or animals with no evidence of adverse effects are the most common outcomes. However, they noted that both targeted and random mutagenesis can lead to neonatal mortality or animals with compromised health or welfare.
When adverse effects do occur they either are predicted on the basis of the particular genetic modification or, notably, are not of a kind that was predicted to occur or are seen in circumstances where adverse effects were not anticipated. It is the uncertainty and low predictability of such events that present particular challenges when managing GM animal colonies. Such unpredicted adverse effects may arise for a variety of reasons including the overexpression or the absence of the specific gene, interactions with collocated genes, or the influence of the genetic background of donor animals or the background strain that may interact with the targeted modification. Furthermore, adverse effects may not be evident in the first generation and emerge only in subsequent generations (Dennis 2000).
Abnormalities in GM animals may affect the viability of offspring and their long-term survival and welfare and may be linked to the specific gene modification or reflect a peculiarity of the phenotype of the background strain. A diverse range of abnormalities have been reported, including hydrocephalus, cleft palate, malformed limbs, absence of teeth, poor mothering, absence of milk, poor thermoregulatory ability, increased aggression and cannibalism, clotting disorders, enhanced growth of tumors and development of metastases often at atypical sites, diabetes, osteoporosis, degenerative joint disease, respiratory disorders, inflammatory bowel disease, ulcerative colitis, liver and kidney
dysfunction, seizures, and sensory and locomotor abnormalities affecting sight, hearing, smell, balance, and social interactions. The occurrence of one or more of these abnormalities may necessitate the euthanasia of affected individuals but also may indicate the need to review the ongoing production of a particular line. In some of these conditions the impact can be alleviated by the implementation of treatment programs or changes to husbandry practices such as the provision of special diets, the placement of food and water on the bottom of the cage, and increased volume and changing of bedding (Brown and Murray 2006).
A higher than expected incidence of infectious disease has been observed in GM animals (Dennis 2002). As highlighted in the review by Franklin (2006), GM animals respond to infections in a similar way to immunodeficient animals: they develop clinical infections due to common opportunists or to agents that would normally result in asymptomatic infections. GM may affect host specificity of pathogens and infections may result in unusual or new phenotypes not necessarily due to immune defects.
When developing humane endpoints for GM animal models the uncertainty of the incidence, kind, and timing of adverse events presents a significant challenge (Dennis 2000). Furthermore, when animals develop concurrent diseases—for example, 26% of mice developed diabetes in a transgenic model (R6/2) of Huntington’s disease (Luesse et al. 2001)—the determination of an appropriate endpoint may be confounded. Unquestionably, the development and implementation of monitoring strategies to assess the impact of a specific genetic modification is essential to effectively manage the welfare of GM animals and to enable the development of effective humane endpoints.
When a new genetic line is created a detailed description of its phenotype must be undertaken. With the rapid increase in the number of new lines being created, especially in mice, reference databases have been established that document the methods used to create and maintain the GM line and its phenotype; details in relation to the onset of changes, disease progression, and suggested endpoints are included in some instances.
Although some concern has been expressed that the monitoring of GM animals could focus on a description of the phenotype with insufficient attention given to animal welfare indicators (Brown and Murray 2006), these processes can and should be complementary and there are important benefits in establishing effective humane endpoints when this occurs. A detailed phenotypic description, including animal welfare measures, will provide both a more accurate picture of the time course and characteristics of a phenotype and identify relevant indicators of negative effects on the animal’s welfare. Ideally, the quality of
these data will enable a more accurate determination of the specific settings for a humane endpoint by aligning phenotypic changes with animal welfare indicators and identification of special needs that can alleviate some effects.
Several protocols to monitor the welfare of GM animals have been developed (e.g., Dennis 2002; Wells et al. 2006) with many common elements.
Dennis (2002) emphasized the importance of at least daily monitoring when new lines are created to ensure that signs of illness, physical defects, injury, or abnormal behavior are detected and assessed, noting the importance of documenting what may seem to be unimportant changes—the frequency and specific elements of a monitoring program should detect both predicted and unforeseen changes. Dennis (2002) also stressed the need to include regular monitoring of the health status of GM mouse lines, including serological testing and postmortem examination. These measures also are an important component of developing a phenotypic description of a GM line.
Similarly, Wells and colleagues (2006) proposed a structural assessment of the welfare of new GM lines focusing on the initial phase in the creation and phenotypic assessment, the aim being to create a “welfare profile” so that, once a line is established, monitoring would focus on several welfare indicators specific for that line. However, as noted by other authors, there can be discrepancies in the phenotypic description of a given GM line between different institutions. Consequently, this kind of welfare assessment also should be undertaken when GM lines are newly introduced to an institution.
Wells and colleagues (2006) propose specific welfare assessments to be carried out in neonates and at weaning. In neonates, criteria such as skin color, surface temperature, activity, reflexes, response to touch, and evidence of a milk spot are proposed. At weaning, mice are assessed by appearance, coat condition, posture, gait, activity, clinical signs, and relative size; in addition, preweaning mortalities, evidence of aggression or stereotypies, and body weight are recorded, and more detailed behavioral assessments are recommended if behavioral problems are identified. If no animal welfare problems are identified in neonates or weanlings, animals are monitored during routine husbandry procedures. If animal welfare concerns either are identified in the assessment of neonates or weanlings or subsequently emerge, animals then undergo more detailed assessments to identify special needs and criteria for humane endpoints.
There is a convergence between protocols for monitoring animal welfare and for developing a phenotypic profile. As a minimum, Brown and Murray (2006) suggest that the following measures be included in phenotype screening: clinical chemistries, complete blood count, urinalysis, gross and histopathology of major organs, abnormal gross tissues and target organs, and an assessment of general health, sensory function, motor abilities, and behavioral tests as proposed by Crawley (1999). Proposals such as that developed by Rogers and col-
leagues (1997) and Crawley and Paylor (1997) to develop a comprehensive phenotypic profile have been widely adopted with various modifications and include measures relevant to animal welfare assessment. However, an important addition to these protocols is a comprehensive assessment of behavior that uses a range of laboratory-based tests to assess learning, memory, sensory motor activity, feeding behavior, pain, reproduction, and emotionality. These kinds of data may also assist in evaluating or predicting the impact of the GM on animal welfare.
Finally, when assessing phenotypic changes in GM animals, comparison with their wild-type, littermate controls is important.
There are a number of ways to reduce the impact of GM on the welfare of a particular line (NHMRC 2007). The rapid development and refinement of GM technologies that limit temporal or spatial gene expression has resulted in refinements to the way in which the expression of phenotypes can be targeted and benefits the welfare of the GM animal by limiting or negating the expression of negative effects. Two common strategies used in the production and maintenance of GM animals are, when there is an unacceptable level of morbidity, mortality, chronic disease, or abnormal behavior in homozygote animals, to maintain the GM line in heterozygous animals and, when the GM line is no longer needed for current research programs using cryopreservation, to store embryos, sperm, and ovaries.
GM Models in the Neurosciences
There has been a significant increase in the number of GM animal models in the neurosciences used in the study of neurodegenerative diseases such as Alzheimer’s, Huntington’s, or Parkinson’s disease, and psychiatric illness such as schizophrenia, depression, and anxiety, obsessive compulsive disorders, and pain and stress. In some circumstances, for example Huntington’s disease, a single gene may be involved, but many of these conditions involve complex gene interactions and the use of transgenic or knockout models provides new opportunities to study the function and interplay of individual genes to elucidate factors that influence the regulation and modulation of neural substrates (see for example the discussion by Mogil and Grisel 1998 in relation to pain studies, and Muller and Keck 2002 in relation to stress). Furthermore, the development of knockout lines has created the possibility of studying the role and function of a single gene in relation to behavior (Nelson and Young 1998; Anagnostopoulos et al. 2001).
An overview of the scope of GM animal models in the neurosciences provides some insight into the opportunities and challenges that GM animals present.
One of the drivers for the development of a battery of behavioral tests to be used in the development of phenotype profiles for new GM lines has been the potential to use these animal models in the neurosciences. Consequently, one of the defining characteristics of these animal models will be changes to one or more behavioral tasks indicative of cognitive, emotional, sensory, or motor function. Changes in an animal’s ability to perform such tasks may relate to the experience of pain, stress, anxiety, fear, or depression. In further study of these models, a suite of specific behavioral tasks will be selected relevant to the hypothesis being tested (Crawley 1999).
In many GM lines animals do not show any evidence of clinical disease or abnormal behaviors but demonstrate a change in one or more tasks. For example, in a study designed to look at dysfunction in the serotonergic system, which is implicated in psychiatric conditions such as anxiety and depression, compared to their wild-type controls 5-HT1A knockout mice showed increased anxiety in the elevated-plus maze test and decreased reactivity in the open-field test, whereas 5-HT1B knockouts showed the reverse, but neither line showed any difference in development, feeding behaviors, reproductive performance, or any other evidence of abnormalities (Zhuang et al. 1999). Changes in behavior such as increased aggression, altered maternal care, seizures, and impaired motor coordination and sensory abilities are seen in knockout mice where such changes are linked to the targeted gene (Nelson and Young 1998; Anagnostopoulos et al. 2001). Furthermore, transgenic and knockout mice with these kinds of modifications may develop changes that affect their ability to interact with their physical and social environment. For example, changes to genetic components of the dopaminergic system in mice are associated with changes to their olfactory ability (McGrath et al. 1999), resulting in increased aggression, changes in their social interaction (Rodriguiz et al. 2004), and increased fear response (El-Ghundi et al. 2001).
In these kinds of models, setting criteria for humane endpoints presents particular challenges. This is not an issue when animals display signs of clinical disease or abnormal behaviors, such as seizures, but when the only evidence of behavioral change is in the performance of a behavioral task during a brief exposure to an artificial environment and there is no evidence of change in any other measures, the decision is not so clear. Evidence of altered emotionality or cognitive ability in a behavior test does not indicate that such experiences are part of an animal’s day-to-day condition. The animal’s negative experiences may be limited to the brief test period and in these circumstances the frequency of testing should be considered in limiting impact. However, the occurrence of these kinds of behavioral changes concurrent with evidence of changes under normal living conditions shifts the weight of evidence and may indicate animal welfare concerns. For example, mice deficient in the extracellular matrix glycoprotein tenascin-R (TN-R) showed increased anxiety when tested in the open-
field and elevated-plus maze tests and decreased locomotor activity, but also showed significant changes in circadian activity in their home cage (Freitag et al. 2003).
There are differing views as to the interpretation of stereotypic behavior in relation to animal welfare and, as shown in a review by Mason and Latham (2004), although in most circumstances where this occurs it is likely to be linked to poor welfare, there are exceptions. There are a number of reports where transgenic or knockout mice display stereotypic behavior with a range of genetic modifications (for example, Ambree et al. 2006; Berger et al. 2006; Chartoff et al. 2001; Hines et al. 2008; Mohn et al. 1999; Rodriguiz et al. 2004). A recent report on the recognition and alleviation of distress prepared by an ILAR committee (NRC 2008) concluded that stereotypies are undesirable and their prevention is likely to improve animal welfare. The weight that is given to the presence of stereotypies in setting humane endpoints may be contentious. The context in which these occur may be relevant, but a special case would need to be made to maintain animals exhibiting stereotypies that cannot be alleviated under home cage conditions.
In the context of neurodegenerative models where there is progressive deterioration of an animal’s condition over a prolonged time, there are examples of where the early detection of the onset of the disease identifies early intervention points and provides an opportunity to test therapeutic efficacy. For example, a transgenic model of Huntington’s disease measuring behavioral changes in open-field and elevated-plus maze tests detected the onset of a deterioration in motor activity prior to evidence of changes in anxiety levels (Klivenyi et al. 2006), and the development of a transgenic model of Alzheimer’s disease showed cognitive and neurophysiological defects before the development of overt neuropathology (Gimenez-Llort et al. 2007). Further, a report by Drage and Heinrichs (2005) shows not only how the husbandry of a seizure-prone E1 mouse can be modified to eliminate the onset of seizures when the mice are held by the tail but also evidence of behavioral and cardiovascular changes that can be used as predictors before the onset of seizures.
There are ongoing challenges in the interpretation of behavioral phenotypic changes in relation to the fidelity to specific gene effects, including the confounding influences of husbandry and housing conditions, which are relevant to the setting of humane endpoints. Changes that result from compensation or developmental effects of the mutation, the influence of the genetic background strain, the influence of maternal behavior on adult phenotype, or pleiotropy can all confound interpretation of a phenotype (Gingrich and Hen 2000); differences in the background strain can result in significant phenotypic differences in pain-related measures (Lariviere et al. 2001); and rearing conditions and neonatal handling can affect behavioral responses to pain and stress in adult mice (Sternberg et al. 2003; Parfitt et al. 2004).
A number of studies have demonstrated the influences of laboratory conditions on phenotypic expression (Crabbe et al. 1999; Wahlsten et al. 2003). Würbel (2001) has argued that more attention should be given to the animal’s living
conditions and hypothesized that animals that experience an enriched environment (EE) would be less susceptible to minor environmental changes and therefore provide a more “standardized” response to test conditions. Although Wolfer and colleagues (2004) showed that EE did not result in differences in behavioral tests when applied to two inbred mouse strains and their F1 hybrids, studies in transgenic models of Alzheimer’s disease indicate the need to carefully evaluate the influence of EE in specific GM animals (Jankowsky et al. 2003; Richter et al. 2008). Recent studies showing that even subtle changes in EE or cage design are associated with significant changes in tests used for behavioral phenotyping (Tucci et al. 2006; Kallnik et al. 2007; Pietropaolo et al. 2007) highlight the urgent need to further investigate these kind of effects.
The establishment of humane endpoints in GM animal models presents particular challenges due to the unpredictable nature and occurrence of adverse events. However, the scope and depth of monitoring required to accurately describe a phenotype, together with careful monitoring to assess animal welfare, provide a comprehensive framework to establish humane endpoints with a high level of accuracy as well as informing the development of effective strategies to reduce the impact of a specific genetic modification. Studies that can identify and demonstrate ways to modify confounding influences on the phenotypic expression of a specific gene will enable refinement of the setting of humane endpoints that will benefit both scientific and animal welfare outcomes.
Ambree O, Touma C, Bortz N, Keyvani K, Paulus W, Palme R, Westaway D, Sachser N. 2006. Activity changes and marked stereotypic behaviour precede AB pathology in TgCRND8 Alzheimer mice. Neurobiol Aging 27:955-964.
Anagnostopoulos AV, Mobraaten LE, Sharp JJ, Davisson MT. 2001. Transgenic and knockout databases: Behavioral profiles of mouse mutants. Physiol Beh 73:675-689.
Animal Procedures Committee. 2001. Rep Biotech. www.apc.uk.
Berger S, Wolfer DP, Selbach O, Alter H, Erdmann G, Reichardt H, Chepkova A, Welzl H, Hass H, Lipp H, Schutz G. 2006. Loss of the limbic mineralocorticoid receptor impairs behavioral plasticity. Proc Nat Acad Sci U S A 103:1995-2000.
Brown MJ, Murray KA. 2006. Phenotyping of genetically engineered mice: Humane, ethical, environmental and husbandry issues. ILAR J 47:118-123.
CCAC [Canadian Council on Animal Care]. 2008. CCAC Guidelines on Genetically Engineered Animals. Ottawa. Draft document.
Chartoff EH, Marck BT, Matsumoto AM, Corsa DM, Palmiter RD. 2001. Induction of stereotypy in dopamine-deficient mice requires striatal D1 activation. Proc Nat Acad Sci U S A 98:10451-10456.
Comber J, Griffin G. 2007. Genetic engineering and other factors that might affect human-animal interactions in the research setting. J Appl Anim Welf 10:267-277.
Crabbe JC, Wahlsten D, Dudek BC. 1999. Genetics of mouse behavior: Interactions with laboratory environment. Science 284:1670-1672.
Crawley JN. 1999. Behavioral phenotyping of transgenic and knockout mice: Experimental design, and evaluation of general health, sensory functions, motor abilities and specific behavioral tests. Brain Res 835:18-26.
Crawley JN, Paylor R. 1997. A proposed test battery and constellations of specific behavioral paradigms to investigate the behavioral phenotypes of transgenic and knockout mice. Horm Beh 31:197-211.
Dennis MB. 2000. Humane endpoints for genetically engineered animal models. ILAR J 41:94-98.
Dennis MB. 2002. Welfare issues of genetically modified animals. ILAR J 43:100-109
Drage MG, Heinrichs SC. 2005. Phenotyping the untouchables: Environmental enhancement of behavioral and physiological activation in seizure-prone E1 mice. Epilepsy Behav 6:35-42.
Einsiedel EF. 2005. Public perceptions of transgenic animals. Rev OIE 24:149-157.
El-Ghundi M, O’Dowd BF, George SR. 2001. Prolonged fear responses in mice lacking dopamine D1 receptor. Brain Res 892:86-93.
Franklin DL. 2006. Microbial considerations in genetically engineered mouse research. ILAR J 47:141-155.
Freitag S, Schachner M, Morellini F. 2003. Behavioral alterations in mice deficient for extracellular matrix glycoprotein tenascin-R. Behav Brain Res 145:189-207.
Gimenez-Llort L, Blazquez G, Canete T, Johansson B, Oddo S, Tobena A, LaFerla FM, Fernandez-Teruel A. 2007. Modeling behavioral and neuronal symptoms of Alzheimer’s disease in mice: A role for intraneuronal amyloid. Neurosci Biobehav Rev 31:125-147.
Gingrich JA, Hen R. 2000. The broken mouse: The role of development, plasticity and environment in the interpretation of phenotypic changes in knockout mice. Curr Opin Neurobiol 10:146-152.
Hines RH, Wu L, Hines DJ, Steenland H, Mansour S, Dahlhaus R, Singaraja RR, Cao X, Sammler E, Hormuzdi SG, Zhuo M, Husseini A-E. 2008. Synaptic imbalance, stereotypies and impaired social interactions in mice with altered Neuroligin 2 expression. J Neurosci 28:6055-6067.
Jankowsky JL, Xu G, Fromholt D, Gonzales V, Borchelt DR. 2003. Environmental enrichment exacerbates amyloid plaque formation in a transgenic mouse model of Alzheimer disease. J Neuropath Exp Neurol 62:1220-1227.
Kallnik M, Elvert R, Ehrhardt N, Kissling D, Mahabir E, Welzl G, Faus-Kessler T, Angelis M, Wurst W, Schmidt J, Holter S. 2007. Impact of IVC housing on emotionality and fear learning in male C3HeB/FeJ and C57BL/6J mice. Mammal Gen 18:173-186.
Klivenyi P, Bende Z, Hartai Z, Penke Z, Nemeth H, Toldi J, Vecsei L. 2006. Behaviour changes in a transgenic model of Huntington's disease. Behav Brain Res 169:137-141.
Lariviere WR, Chesler EL, Mogil JS. 2001. Transgenic studies of pain and analgesia: Mutation or background genotype? J Pharmacol Exp Therap 297:476-473.
Luesse H-G, Schiefer J, Spruenken A, Puls C, Block F, Kosinski CM. 2001. Evaluation of R6/2 HD transgenic mice for therapeutic studies in Huntington’s disease: Behavioral testing and impact of diabetes mellitus. Behav Brain Res 126:185-195.
Mason GJ, Latham N. 2004. Can’t stop, won’t stop: Is stereotypy a reliable animal welfare indicator? Anim Welf 13:S57-S69.
McGrath MJ, Campbell KM, Burton FH. 1999. The role of cognitive and affective processing in a transgenic mouse model of cortical-limbic neuropotentiated compulsive behavior. Behav Neurosci 113:1249-1256.
Mogil JS, Grisel JE. 1998. Transgenic studies of pain. Pain 77:107-128.
Mohn AR, Gainetdinov RR, Caron MG, Koller BH. 1999. Mice with reduced NMDA receptor expression display behaviors related to schizophrenia. Cell 98:427-436.
Morton DB. 2000. A systematic approach to establishing humane endpoints. ILAR J 41:80-86.
Muller MB, Keck ME. 2002. Genetically engineered mice for studies of stress-related clinical conditions. J Psychiatr Res 36:53-76.
Nelson RJ, Young KA. 1998. Behavior in mice with targeted disruption of single genes. Neurosci Biobehav Rev 22:453-462.
NHMRC [National Health and Medical Research Council]. 2007. Guidelines for the generation, breeding, care and use of genetically modified and cloned animals for scientific purposes. Canberra: Australian Government.
NRC [National Research Council]. 2002. Animal Biotechnology: Science-based Concerns. Washington: National Academies Press.
NRC. 2008. Recognition and Alleviation of Distress in Laboratory Animals. Washington: National Academies Press.
Nuffield Council. 2005. The Ethics of Research Involving Animals. London: Nuffield Council on Bioethics.
Parfitt DB, Levin JK, Saltstein KP, Klayman AS, Greer LM, Helmreich DL. 2004. Differential early rearing environments can accentuate or attenuate the responses to stress in male C57BL/6 mice. Brain Res 1016:111-118.
Pietropaolo S, Mintz M, Feldon J, Yee BK. 2007. The behavioral sequelae following the prevention of home-cage grid climbing activity in C57BL/6 mice. Behav Neurosci 121:345-355.
Richter H, Ambree O, Lewejohann L, Herring A, Keyvani K, Paulus W, Palme R, Touma C, Schabitz WR, Sachser N. 2008. Wheel-running in a transgenic mouse model of Alzheimer’s disease. Behav Brain Res 190:74-84.
Robinson V, Morton DB, Anderson D, Carver JFA, Francis RJ, Hubrecht R, Jenkins E, Mathers KE, Raymond R, Rosewell I, Wallace J, Wells DJ. 2003. Refinement and reduction in production of genetically modified mice. Lab Anim 37(Suppl 1).
Rodriguiz RM, Chu R, Caron MG, Wetsel WC. 2004. Aberrant responses in social interaction of dopamine transporter knockout mice. Behav Brain Res 148:185-198.
Rogers DC, Fisher EMC, Brown SDM, Peters J, Hunter AJ, Martin JE. 1997. Behavioral and functional analysis of mouse phenotype: SHIRPA, a proposed protocol for comprehensive phenotype assessment. Mammal Gen 8:711-713.
Royal Society UK. 2001. The Use of Genetically Modified Animals. http://royalsociety.org/.
Sternberg WF, Ridgway CG. 2003. Effects of gestational stress and neonatal handling on pain, analgesia and stress behavior of adult mice. Physiol Beh 78:375-383.
Stokes WS. 2000. Reducing unrelieved pain and distress in laboratory animals using humane endpoints. ILAR J 41:69-61.
Tucci V, Lad HV, Parker A, Polley S, Brown DM, Nolan PM. 2006. Gene-environment interactions differentially affect mouse strain behavioral parameters. Mammal Gen 17:1113-1120.
Van der Meer M, Baumans V, Olivier B, Kruitwagen CLJJ, Van Dijk JE, Van Zutphen LFM. 2001. Behavioral and physiological effects of biotechnology procedures used for gene targeting in mice. Physiol Behav 73:719-730.
Wahlsten D, Metten P, Phillips TJ, Boehm II SL, Burkhart-Kasch S, Dorow J, Doerksen S, Downing C, Fogarty J, Rodd-Henricks K, Hen R, McKinnon CS, Merrill CM, Nolte C, Schalomon M, Schlumbohm JP, Sibert JR, Wenger CD, Dudek BC, Crabbe JC. 2003. Different data from different labs: Lessons from studies of gene-environment interaction. J Neurobiol 54:283-311.
Wells DJ, Playle LC, Enser WEJ Flecknell PA, Gardiner MA, Holland J, Howard BR, Hubrecht R, Humphreys KR, Jackson IJ, Lane N, Maconochie M, Mason G, Morton DB, Raymond R, Robinson V, Smith JA, Watt N. 2006. Assessing the welfare of genetically altered mice. Lab Anim 40:111-114.
Wolfer DP, Litvin O, Morf S, Nitsch RM, Lipp HP, Würbel H. 2004. Cage enrichment and mouse behavior. Nature 432:821-822.
Würbel H. 2001. Ideal homes? Housing effects on rodent brain and behavior. Trends Neurosci 24:207-211.
Zhuang X, Gross C, Santarelli L, Compan V, Trillat AC, Hen R. 1999. Altered emotional states in knockout mice lacking 5-HT1A or 5-HT1B receptors. Neuropsychopharmacology 21(2 Suppl):52S-60S.
The concept of animal welfare is inextricably bound up with ethics, with an ethics component. Animal welfare is in essence what we owe an animal and to what extent. This is not very well understood by animal users, particularly by the agricultural community. I was on the Pew Commission where we frequently heard that animal welfare is simply a matter of sound science. It is not. It is a combination of sound science and ethics.
The relevant ethics that figures in the animal welfare equation comes from the societal ethic for animals, and there is in fact a new societal ethic emerging over the past 40 years that will likely dominate not only the West but, insofar as the West is the source of science in the East and elsewhere, the East as well.
In recent years ethnocentrism has become a dirty word and a variety of factors have converged to create a bias against the bias in favor of our own culture. Postmodernism, feminism, atonement for past imperialistic sins, political correctness have all converged in favor of a putative neorelativistic tolerance for multiculturalism that we would have historically dismissed offhandedly.
In fact, we do not accept many principles from other cultures. We don’t accept tribal butchery, we don’t accept clitoridectomies in young women, we don’t accept the Taliban repression of women. But multiculturalism has certainly exacted some costs. Consider for example the extraordinary proliferation of evidentiary baseless alternative medicine, some of which is purportedly borrowed from the traditions of “other cultures” and on which the US public spent no less than $40 billion in 2005. This is not based in evidence and not based in science.
Hence too our current concern: How do we arrive at a conception of animal welfare that does justice to the bewildering array of views of this concept across different cultures? Part of the reason this issue creates perturbation among scientists is their historical disavowal of ethics being integral to science. The mantra is: “Science is value free in general and ethics free in particular.”
When I was trained in science in the ’60s I got that mantra. My students are still getting it, although it is not quite as widespread as it was, fortunately, because it really is a detriment to the thriving of science in our society, which already is facing formidable obstacles.
Thus it is widely believed that animal welfare can be explicated without reference to values, simply on the basis of objective biological fact. In reality, the variation across cultures in views of animal welfare can be found historically intraculturally. It is simply magnified by considerations of cultural variability.
Consider the following: In 1981 in response to burgeoning societal concerns about animal welfare, the US agricultural community, represented by the Council for Agricultural Science and Technology (CAST), published Scientific Aspects of the Welfare of Food Animals. Reflecting a ubiquitous view among producers, the CAST report spoke of welfare as follows: The principal criteria used thus far as indices of the welfare of animals in production systems have been rate of growth or production, efficiency of feed use, efficiency of reproduction, mortality, and morbidity. In other words, the welfare of an animal is defined and determined by how well it fulfills the human purpose to which it is put, with no reference to how it feels, whether or not it suffers pain, distress, anxiety, boredom, loneliness, frustration, inability to move or be with con-specifics, and so on.
Implicit in this definition are a set of values and a set of moral obligations that are easily extracted: Humans are morally obliged to provide animals only with a set of conditions that allows the animal to fulfill the purpose for which it is kept by humans. In Kant’s terminology, then, animals are in no way ends in themselves, they are strictly means to an end, a human end. Animal welfare is based solely on these human ends. In metaphorical terms, welfare is to animals as sharpness is to a saw, what is needed for both to be functional tools.
At roughly the same historical moment, the early 1980s, other definitions of animal welfare were promulgated. In the writings of Marian Dawkins, Ian Duncan, and myself, one essential feature of welfare was argued to rest in what the animal experienced, its subjective states. The moral position implicit in such views was that animals ought to be at least in some measure in Kantian terms “ends in themselves” because they were conscious, and what they experienced mattered to them. By the way, at that time much of the scientific community was agnostic about the concept of animal consciousness. They overtly denied either the existence or the knowability of animal consciousness; therefore what we did to animals and how we forced them to live didn’t matter.
In my view of welfare, animals have intrinsic value rather than merely instrument value—that is, value merely as tools—because they are capable of valuing in their subjective life what happens to them. There were other definitions of welfare—e.g., the Farm Animal Welfare Council (FAWC) notion of the
Five Freedoms1 that grew out of the Brambell Commission, and so forth. The point here is that even in British and American culture, one could find numerous very obviously different and incompatible definitions of animal welfare, based in radically differing views of the moral status of animals, separated irreconcilably by disparate ethical values underlying them. Thus, the existence of divergent views of animal welfare, differing across cultures, does not raise any new conceptual problems that were not already present by virtue of the intracultural value-based differences in views of what constitutes animal welfare.
It is in no way surprising that animal welfare should have emerged as a moral issue in the latter part of the 20th century, because of the precipitous changes in the nature of animal use that transpired in the mid-20th century. For the entire history of civilization, the overwhelming use of animals in society was in agriculture—food, fiber, locomotion, and power—and the key to success in agriculture was good husbandry. Husbandry meant putting your animals in the optimal conditions dictated by the animal’s biological natures and needs, and augmenting their native ability to survive and thrive by provision of food during famine, water during drought, help in birthing, medical attention, and protection from predation.
This has been called the ancient contract. It was based on the insight that producers did well if and only if animals did well, and vice versa. Thus, husbandry was, in equal measure, a prudential and a moral imperative, sanctioned by what is after all the ultimate motivation for human beings, self-interest. Thus defining animal welfare and animal ethics was a nonissue.
In fact, the only animal ethic that was needed and can be found as early as the Bible arose from the need to cover the case of the small number of psychopaths and sadists who were unmoved by self-interest. In other words, if you were using animals in an agricultural way, which was the primary use of animals, you needed to put them in decent conditions and provide for their needs during famine and drought and so forth and so on in order to be financially successful.
Defining animal ethics and animal welfare became an issue when the nature of agriculture changed from husbandry to industry. The values changed as well. Primacy was now given to the values of efficiency and productivity. Whereas one can characterize husbandry as putting square pegs in square holes, round pegs in round holes, and creating as little friction as possible doing so, the Industrial Revolution provided us with “technological sanders,” as it were, that allowed us to force square pegs into round holes, round pegs into triangular holes, while still keeping animals productive—things like air-handling systems, antibiotics, vaccines, etc. If one had tried to develop these systems during the era of animal husbandry, the animals would be dead in weeks of disease spreading like wildfire, but with these sanders we can force square pegs into round holes.
1These are freedom from thirst, hunger, and malnutrition; freedom from discomfort; freedom from pain, injury, and disease; freedom to express normal behavior; and freedom from fear and distress; available online (www.fawc.org.uk/freedoms.htm).
What was lost was the isomorphism between the animal well-being and productivity that characterized husbandry, and thus animal welfare and animal ethics became an issue instead of a presupposition of animal use. This was potentiated by the advent at roughly the same time, the 1940s, of large amounts of animal research and testing, again representing significant animal use that violated the symbiosis inherent in husbandry.
The research community typically deflected this issue by being agnostic both about the relevance of ethics to science and about animal consciousness. I was a principal architect of the 1985 federal laboratory animal laws in the United States. In 1982 when I went before Congress to defend them, I was asked to prove that there was a need for such a law. The research community claimed they were already controlling pain in research animals.
So I went out and, with a colleague at the Library of Congress, did a literature search on laboratory animal analgesia. How many papers did I find? I found none on laboratory animal analgesia. When I broadened it to animal analgesia, I found two, one of which said there ought to be papers, and the other said here is what we know: it was a one-pager in the Journal of the American Veterinary Medical Association and it was very honest about not knowing anything and how there was a moral imperative to know.
As public cognizance of the radical changes in animal use grew beginning in the 1960s and 1970s, efforts in favor of restoring fairness to animal use began to pervade Western society, beginning in Britain in the 1960s and resulting in the view that animals were entitled to the famous five basic freedoms. The ensuing years saw the emergence in Western society of “the new social ethic for animals.”
As anyone attending to cultural history can easily determine, the issue of animal treatment assumed major social prominence beginning about 1970. Whereas 30 years ago in the United States one would have found no federal bills pending in Congress pertaining to animal welfare, the last decade has witnessed up to 50 and 60 per year. On a state level in 2004, there were well over 2,100 bills proposed in state legislatures pertaining to animal welfare; there were over 200 in California alone.
Most Western countries have recently adopted new laws protecting laboratory animals and ensuring control of their pain, often despite opposition from the research community, which preferred a laissez-faire approach. Britain is of course a notable exception, given the act of 1876.
Much of Northern Europe and the European Union have introduced major restrictions on confinement agriculture, probably the most dramatic being the Swedish law of 1988 that abolished confinement agriculture as taken for granted in the US, and created what the New York Times very presciently called in 1988 a “bill of rights for farm animals.”
Although the US has been slow in developing its concern for agricultural animals, in the last few years it has tended to accelerate, partially due to referenda, legislative initiatives to abolish the most egregious of practices. The Pew
Commission report (PCIFAP.org) also educated a myriad of people who didn’t really know about agriculture before.
Well, we can proffer evidence indefinitely, but I think enough has been said and placed in evidence to buttress my claim regarding social concern. So the question that arises is, if there is that much social concern, how is it expressing itself ethically?
Historically, both the laws protecting animals and the social ethic informing them were extremely minimalist, in essence forbidding—and this is language from the legal system, from the cruelty laws as well as from judicial interpretations of those laws—deliberate, willful, sadistic, deviant, extraordinary, unnecessary cruelty not essential, as one judge put it, to ministering to the necessities of man, or completely outrageous neglect.
Those of you involved in animal welfare may well be aware that early efforts to regulate animal research invoked the cruelty laws and tried to present in evidence certain research that was “cruel,” and the universal judicial assessment was that research is not the sort of thing that can be cruel. It is not deviant, it is not sadistic, it does not betoken psychopathic behavior, etc. That is why it was essential to develop, as one judge put it, a new ethic for animals, by going not to the judiciary but to the legislature.
The ethic of anticruelty is found in the Bible and in the Middle Ages. St. Thomas Aquinas, while affirming that although animals were not direct objects of moral concern, nonetheless forbade cruelty to them on the grounds that those who would be cruel to animals will inexorably graduate to people. This is an insight that has subsequently been buttressed by decades of social scientific research—our last dozen serial killers all had early histories of animal abuse. Those involved with battered women’s shelters know that provisions must be made for the woman’s animal or the husband who is a batterer will go back and hurt the animal to get back at the woman. Psychiatrists acknowledge animal abuse…as sentinel behavior for subsequent psychopathology.
Roughly beginning in 1800, anticruelty laws, reflecting the anticruelty ethic, were codified in the legal systems of most Western societies. The key notion explaining why there was a demand for a new ethic can be found in the fact that the old ethic was so restricted in scope. If I draw a pie chart representing all the suffering that animals experience at human hands, how much would you say was the result of deliberate cruelty, a lot or a little? Every audience says a little. One week I spoke to PETA at San Francisco State and the Billings Rodeo Association of Montana, and they both said 1%. Well, if only 1% is covered by the cruelty ethic, then 99% is not. What that means is, as society has begun to concern itself with the other 99%, it has sought a vocabulary, an ethic, of expressing that concern in a manner that doesn’t invoke cruelty, which is essentially irrelevant.
Most animal suffering results from putatively reasonable and defensible uses—industrial agriculture, which is meant to provide cheap and plentiful food; scientific research, which advances knowledge, cures disease, and provides medicaments.
In the 1970s when I was hired by a veterinary school to develop the field of veterinary ethics, I realized pretty early that the moral status of animals was a fundamental question in veterinary ethics. That led me to think about what sort of ethics society was likely to develop if indeed it was to continue to be concerned about animals. It dawned on me after about two years that ethics does not come ex nihilo—it doesn’t come out of nothing. As Plato said, all ethics builds on preestablished ethics. I surmised that society would look to the ethic we have for people and modify it, change it—mutatis mutandis, as philosophers say—appropriately change it to fit the animal situation.
What aspect of our ethics is in fact being extended? One that is applicable to animal use is the fundamental problem of weighing the interests of the individual human against the general public. Different societies have provided different answers to this problem. Totalitarian societies opt to devote little concern to the individual, favoring instead the state or the Reich or the Volk or whatever their version of the general welfare may be. At the other extreme, anarchical groups such as communes give primacy to the individual and very little concern to the group; hence they tend to enjoy a very transient existence, such as the communes of the 1960s did.
In Western society, however, a balance is struck. Although most of our decisions are made to benefit the general welfare, fences are built around individuals to protect the individual’s basic human interests from being sacrificed for the majority. Thus we protect individuals from being silenced even if the majority disapproves of what they say. We protect individuals from having their property seized without compensation, even if such seizure benefits the general welfare. We protect individuals from torture, even if they planted a bomb in an elementary school and refuse to divulge its location.
What are these interests that we protect? We protect the interests of the individual that we consider essential to being human, to human nature, from being submerged for the sake of the common good.
What are these fences around human individuals called? They are rights. I’m not obviously going to be using the animal rights locution in the same way as the animal rights people do. What they really mean is animal liberation. All the legislative flurry of activity, the 2,400 bills proposed and similar acts, is an attempt to provide societal guarantees of proper animal treatment since husbandry no longer ensures it. It is absurd to suggest that these are the same rights that humans have, because animals do not have the same natures that humans have. I thought about not using the locution of rights, but I knew you would realize that the concept was being invoked. However, it is the concept being invoked by the general public.
If you look at surveys (which I don’t really tend to trust but they are indicators), close to 90% of the public will affirm that animals have rights. I have lectured to 15,000 Western ranchers in every ranching state. What percentage of them would say animals have rights? In my experience, well over 90% would.
An example of that occurred when the governor called a conference on the effects of animal welfare and animal rights on Colorado agriculture about 18
years ago. The opening speaker was the head of the Colorado Cattlemen’s Association. He said, “If I had to raise animals like the chicken people do, I’d get the hell out of the business.” I work very closely with these people. I just brokered a deal between the Humane Society of the United States and Colorado agriculture to avoid the costly referendum that took place in California, Proposition 2, banning veal crates and battery cages and gestation crates. It would have cost Colorado agriculture $12 million to fight it and lose two to one, and they didn’t have $12 million, so we were able to broker a compromise.
People are seeking to build fences around animals. There were two surveys, one done by Gallup, one by Oklahoma State University, both of which had almost identical results, although you would expect a discrepancy because Oklahoma State is a very strong agricultural school and the poll was not particularly agriculturally oriented. They both found that 80% of the general public wants to see proper treatment of farm animals ensured by legislation.
People in society are seeking to build fences around animals to protect the animals and their interests and their natures, which following Aristotle I have called telos. Those of you who studied Aristotle know what he means by telos: the biological and behavioral and psychological needs and interests that are constitutive of a given type of animal—e.g., the pigness of the pig, as one of my book reviewers once wrote, or the cowness of the cow. They are trying to protect that from being totally submerged in the quest for human general welfare, and are trying to accomplish it by going to the legislature.
With good husbandry, respect for telos occurred automatically. In industrial agriculture where it is no longer automatic, and also in animal testing, people wish to see it legislated.
Very simply, the new ethic recognizes that fish must swim, birds must fly. Clearly, then, the notion that animals ought to have legal protection for fundamental aspects of their natures, a notion actualized in the Swedish agricultural law of 1988 and implicit in the Brambell Commission, is a mainstream phenomenon.
One of the most extraordinary things about writing the laboratory animal laws was the fact that the public did not divide on party lines. Support for controlling pain and suffering in animals, for example, was invariant across Democrats and Republicans.
People were not saying do not use animals in research. What they are saying is, if you use animals in research, control the pain, control the distress. Distress is demarcated from pain in the US laws.
Conceptually speaking in terms of legal theory, animals cannot have rights because they are property. The old slave decisions established that anything that is property cannot have rights. This is a solecism, a legal oxymoron. However, this could not be changed without a Constitutional amendment although a lot of legal scholars are trying to do precisely that.
There was an animal law conference at Harvard Law School two years ago where 350 people filled every space and 300 were turned away. Over 100 law schools have courses in animal law, and a big thrust of most of those law profes-
sors is enfranchising animals and abolishing invasive animal use. But the same functional goal can be accomplished by restricting how animal property is used, which is exactly what the proliferation of laws attempts to do, including the laboratory animal laws. The day they passed I was sitting with Tom Wolfle from NIH and Dale Schwindaman from USDA, and they both shook my hand saying, “Congratulations, you have established certain minimal rights for animals.” These men were hardly radicals and essentially what they were saying was that an animal now had the right to have its pain controlled if pain is inflicted in the course of research, unless you were studying pain.
The good news is that we have gone from two published papers on analgesia to more than 11,000, with a correlative increase in its use, however deficient that use may still be.
So with this analysis in mind, we can begin to answer the question of cultural relativity of concepts of animal welfare. If our account is correct, there is not great disparity across at least different Western societies: all believe morally that animals should legally have their natures and interests protected and this should be accomplished by the legal/regulatory system. This is perhaps truer in Europe than in America.
Insofar as this notion seems to pervade Western democratic societies, which dominate the world politically and economically, it appears that this notion will dominate as the key notion of animal welfare, even as Western democratic notions of human rights have dominated discourse regarding human ethics.
People in other countries are beginning to realize that this notion will dominate. For example, two years ago I addressed 300 Southeast Asian agriculture animal producers who were greatly interested in what is happening in the West because they knew that they would have to abide by those standards if they were going to trade with the West. Recent announcements by Chinese government officials explicitly state that pressures of globalization are forcing China to consciously consider animal welfare and animal welfare legislation for the first time in its history.
As more and more US research is being shipped to other countries for economic reasons, we can be morally certain that public opinion will demand that it be accompanied by the new ethic. Judy MacArthur Clark has a project to try to bring Western ethical standards to these countries where the research will be exported.
As we argued earlier, the concept of animal welfare is deeply value-laden, both in what we choose to consider ingredient in an animal’s welfare and to what extent we are willing to satisfy those welfare concerns. This in turn first led to producers saying that welfare is what the animal requires to do the job we expect of them. That has been turned around by the new ethic and placed the locus of welfare in the animal, not in our generosity or largesse. That is the source of the notion of rights, that they are entitled rather than simply being a matter of our will.
We have argued that the new ethic is intended to restore the fair contract inherent in husbandry, and it is primarily agricultural. It happened with research first in the US because we are removed from agriculture. My average Columbia PhD friend still thinks farms are Old McDonald’s farms. We have argued that the new ethic is intended to restore the fair contract inherent in husbandry and to ensure that animals lead decent lives. We have further argued that the source of our primary obligation to animals is derived from attending to the animals’ natures, even as the rights of humans are based in respecting the essentials of human nature. How does this notion transfer to animals?
In the US Constitution and in the foundational documents of other democratic societies, the relevant concept of human nature was derived from people’s reaction to being denied certain things, from oppression. Having been denied freedom of religion or belief, people demanded that such belief be protected from governmental intrusion. Similarly, this is true of the seizure of property. Philosophically, the notion of human nature is of course very problematic, with many theories abounding about what human nature is and with some philosophers, notably existentialists and Marxists, affirming that there is no such thing. Interestingly enough, I would argue that whatever position you take on human nature, the notion of animal nature is far less problematic than the notion of human nature.
Animal life is far less plastic than human nature and is far less influenced by culture, and is thus far easier to define. It is more obvious, for example, that lions are predators than that humans are. Determining what animals are evolved for is simpler than answering the same question about humans. So obvious is it that animals have a nature that Aristotle, the greatest philosopher of common sense in antiquity, made it the cornerstone of his biology, and correlatively made biology based in telos the root metaphor for explaining everything in the universe. Whereas for Cartesian and post-Cartesian modern biology, biology is best expressed in terms of physics and chemistry, for Aristotle physics and chemistry were to be explained using functional biological categories. Physics was for Aristotle the biology of dead matter, to put it paradoxically. Biological categories, functional categories, are the most appropriate categories for explanation when it comes to living things.
So in De Anima, which is his biology, Aristotle lays out a functional template for biology that still serves as the framework for the way biology is taught in high school. Any living thing, says Aristotle, is a constellation of functions constitutive of its nature, and all living things are to be described in terms of how they fulfill these functions—locomotion, reproduction, nutrition, excretion, sensation, and so on. We characterize living things in terms of how they fulfill these functions. These functions, then, I would argue, constitute the telos of any type of animal—the pigness of the pig, the dogness of the dog. Aristotle says, tellingly, this nature is knowable simply by intelligent and repeated observation. Respect for the animal’s nature was essential for traditional agriculture: the greater the respect, the better the husbandry, the more productive was the ani-
mal. The fact of agricultural success attested to knowledge of animal telos. Under extensive conditions, productivity did betoken good welfare.
Modern agricultural use circumvents respect for telos and forces square pegs into round holes. Other animal uses ignore telos—for example, zoos and maintenance of animals in research settings, where animals are housed in conditions developed largely out of convenience for us but in violation of the needs flowing from their natures, as when nocturnal burrowing creatures are kept in polycarbonate cages under bright illumination or when social animals are kept in isolation in zoos, or the most egregious example I ever experienced in my youth, a giraffe cage in which the giraffe could not stand up. Such a situation would not occur today, which in a weak way attests evidentially to the claim that society is worried about animal telos.
Both Tom Wolfle and I in the early 1980s and David Morton today have pointed out that the conditions under which we keep animals are probably more invasive and more harmful to the animal than the number of overt invasive protocols. My understanding is that maybe 10-15% of protocols are seriously invasive in research. But 100% of animals are kept under conditions that are inimical to their basic natures.
I would thus argue that in today’s world, animal welfare is being defined in terms of animal telos—that is, meeting the needs and interests that matter to the animal by virtue of its biological and psychological nature. According to my analysis, complete satisfaction of the animal’s telos would constitute what could be legitimately called happiness for the animal. Thus, a happy lion would be a lion kept under extensive conditions with other lions, allowing the full range of lion behavior, including predatory behavior. A miserable lion would be one kept alone in a small cage. The relevant ethical judgment for lions in captivity would be to create a space that functionally approximates the ideal, as the research community has done with primates. Thus, pigs in a straw-based pen system would be happier than a sow in a sow stall, but not as happy as a sow with free access to foraging and shelter from inclement weather. There is a huge body of empirical data from Edinburgh on natural behavior in pigs, particularly sows. In my view, part of the job of what is called animal welfare science is getting as close as possible to happiness for the captive animal.
So there is more to being ethical to the research animals than simply minimizing pain. There are all these other parameters. I find personally talking of animal happiness unproblematic. Indeed, I would argue that animal happiness is far clearer than human happiness, given the curse of human reflective consciousness. A person may have every wish he or she ever wanted fulfilled, yet not be happy for a multitude of reasons. Everyone has friends like this—for example, people possessed of neurotic worry about losing it, neurotic worry about whether they deserve it or not. We have no reason to believe that animals are capable of such nonproductive navel-gazing. There are few human cases of happiness as paradigmatic as the horse let out of the small corral after winter into a large green pasture, after being fenced for months: the kicking up of the heels,
the breaking of wind, the exuberance of the gallop, and the whinny express joy more clearly than any human affirmations. Typically, animals don’t lie.
In sum, we have argued that emerging social ethics for animals in democratic societies will largely dictate the form animal welfare takes, and particularly in the research area, since social and economic pressure will help impose it on other societies. This emerging ethic emphasizes the rights animals should have based on their biological and psychological natures or teloi. The extent to which such telos can be accommodated will vary with circumstance, but the ideal remains clearly demarcated. This idea was necessary to counter the 20th century tendency to see animal welfare as strictly determined by the human purposes to which the animal is put.