Use of Laboratory Animals in the Postgenome Era
Robert R. Rich
Emory University School of Medicine
I am pleased to have the opportunity to bring the somewhat removed but, I hope, useful perspective of a dean to the important questions being addressed today. I would like to begin with several general observations on the interface between new trends in biomedical research and their impact on the issues of pain and distress in laboratory animals. These observations require the caveat that my perspective is from a school of medicine. Explicitly, some of my observations and generalizations do not apply to other entities, such as schools of veterinary medicine, because I will discuss biomedical research as it relates to human disease.
PREDICTABLE TRENDS IN RESEARCH
Consider with me how completion of genome sequencing in humans and numerous model species will likely change biomedical research over the next 20 years, a time frame I will refer to as postgenome medicine.
I assert that we will see the ascendancy of what I would call systems biology. We have heard several times today that in the past decade, the interest in experimental animals has increased as molecular biologists, utilizing the power of transgenic and gene knockout animals, have moved their focus from the molecular laboratory to the intact animal. I am confident that over the next 20 years, this trend will accelerate. Additionally, this trend will require scientists with molecular training to focus increasingly on the complex interactions between cells and
tissues and whole organ systems that can be accomplished only in living animals. Consequently, there will be a marked increase in use of and dependence on research involving animal models. Thus, we should recognize that there will not be a reduction in the total number of laboratory animals used. On the contrary, to take full advantage of the postgenome revolution in biomedical research, we will increasingly call on the use of animal models to solve problems of human illness.
Molecular Genetic Technology
Nevertheless, I also believe there will be a considerable reduction in the heterogeneity of animal models and species used. We will focus instead on those animal models with which we can most effectively exploit the power of molecular genetic technologies. We will observe a particularly dramatic increase in the use of rodents as novel animal models of human diseases. In contrast, at schools of medicine (and I explicitly exclude schools of veterinary medicine) I believe the use of most large animal species will decline substantially. Of course, there will be some important exceptions to this latter generalization. For example, new advances in xenotransplantation will require certain large animal models that will be uniquely useful because they have been genetically engineered for experimental, and possibly within 20 years clinical, use. Similarly, nonhuman primates will remain uniquely useful for certain purposes, such as studies in cognitive neuroscience.
I believe one of the more important consequences of the availability of new molecular technologies is that they will change the very nature of the use of laboratory animals. This use will result in new experimental capabilities. Increasingly, the basis of pain and distress in laboratory animals will not be a reflection of something done to them other than the alteration of their genes. Laboratory animals, predominantly mice, fish, and invertebrates, will be genetically manipulated in ways that result in development of disease or functional disorders. In that sense, they will resemble human beings who are genetically predisposed to different diseases. The animals will develop these disorders or diseases as a consequence of genetic alterations of the germ line.
PREDICTABLE CONSEQUENCES OF GENETIC BREEDING
I believe that our understanding of novel approaches to the treatment and prevention of human (and animal) diseases will be greater than we can even imagine today. As a corollary, however, I also believe that it is unrealistic to have as a foreseeable goal the elimination of pain and distress in experimental animal models that replicate or mimic human diseases. I submit that as long as we can develop animal models by genetic manipulation and gain novel insights into cures, preventions, and treatments of diseases, we should focus on the management rather than the elimination of pain and distress.
Definitions and Categories
We have heard considerable discussion today about the current USDA definition and protocol categorization based on anticipated pain and distress. I agree with speakers who have argued that the current definitions are flawed. I believe part of the flaw is the focus on process rather than outcome. For example, the categorization includes such variables as the use of analgesia or anesthesia, rather than simply asking the more simple (and important) question of whether a proposed protocol would involve an element of animal pain and distress. In this respect, I believe the approach of the Humane Society of the United States, simply to classify pain and distress according to three categories of outcome, represents an improvement. It is important to have a process whereby investigators, veterinarians, and IACUCs assess protocols and monitor research in an effort to minimize pain and distress.
Dr. Haywood admonished us earlier today to keep it simple. I agree completely. Laboratory animals are best served by a simple classification that speaks exclusively to the outcome. We need to use simple words that all of us generally understand and can apply in global assessment, as opposed to overdefining or overquantifying such variables as temporal and intensity issues. I believe a more simple clinical judgment would focus more appropriately and consistently on qualitatively minimizing animal pain and distress and would use the practical experience and intuitive judgment of trained veterinarians and investigators. In fact, I am not truly convinced that it is useful to have distinct definitions of pain and distress. We know what those words mean. The issue is simply whether they are absent (or minimal), moderate, or severe.
I believe that exemplary procedures generally associated with moderate or severe pain are useful as guides to investigators and to IACUCs. I would argue, however, that such exemplars should not be codified in regulation or policy, which risks replacing good judgment with a “check-off” mentality. I believe the more simple notion of global assessment, as I interpret the Humane Society categorization, serves animals well and provides a regulatory framework that can be applied and enforced with greater consistency. Simplifying the process of protocol preparation and review, and elimination (to the extent possible) of the “hassle factors” that promote investigator cynicism and reluctant cooperation, would also contribute significantly to realization of our shared goal of minimizing pain and distress.
Replication of Human Diseases
Nevertheless, I want to reemphasize that just as human disease inherently involves pain and distress, it will not be possible to eliminate these conditions in
laboratory animal research as we attempt to replicate human diseases. However, one of the attractive aspects of the Humane Society's approach for simplifying the categorization of pain and distress is to facilitate work among investigators, veterinary staff, and IACUC members to try to “down-categorize” otherwise painful or distressful protocols with appropriate use of drugs and other modes of pain and distress relief. We should focus on the overall pain or distress experienced and incorporate modalities of pain or distress relief into study designs. Perhaps then we will be increasingly able to classify protocols not as Category E but instead as Category D or C because the appropriate interventions have been applied.
I also advocate increased attention to and definition of specific experimental endpoints that can minimize pain and distress. This would be another advantage of a more simple categorization that, instead of emphasizing the processes involved, requires an answer to the following questions:
“Did the animal utilized in this study experience moderate or severe pain or distress?”
“If so, could we have down-categorized an otherwise painful procedure by choosing an earlier experimental endpoint?”
Obviously, earlier experimental endpoints might be as effective as interventions with drugs.
Finally, I want to echo the other speakers who have pointed out that a cooperative venture is required. No one involved in the responsible conduct of science does not want to minimize pain and distress of laboratory animals. High-quality science requires that we do this. The investigators who design and execute a study, the IACUC members who review the protocols, and the veterinary staff who monitor the process all share the goal of humane care and treatment of our incredibly valuable laboratory animals. By working together, I believe we can develop policies and procedures that are relatively simple, can be understood, can be consistently applied, and are, in fact, good for the animals.
QUESTIONS AND ANSWERS
DR. DE HAVEN (Ron DeHaven, USDA): Dr. Rich, you suggested that we not codify examples of the categories (minor, moderate, and severe) in regulation
or policy. How then, from a regulator's standpoint, would you ensure uniform interpretation of categories?
DR. RICH: Your question is excellent. I do believe that exemplars are helpful. I suggest that we, as a community, can agree on some exemplars but still avoid encouraging checklist mentality inspections of university animal usage, which lack a thoughtful approach to what was actually done. I believe we are ill-served when checklists and clipboards replace good judgment. However, having some community consensus regarding certain procedures would ensure that appropriate corrective measures would be taken if an inspector found substantial discordance from this consensus in a particular institution's categories of procedures or protocols. However, to the extent that codification of exemplars would tend to focus on process rather than outcomes, I believe this approach could actually make inspection and enforcement less, rather than more, fair and effective.