Perspectives On Animal Research
Volume 1, Supplement
An Evaluation of Ten Randomly-Chosen Animal Models
of Human Disease
Appendix C
Comments from Animal Model Researchers
I (SRK) requested comments from 25 researchers who had used animal models featured in this study, and I sent them our review of the animal model they had employed. I solicited comments for all ten animal models. I received six replies, and two granted permission to reproduce their responses. Relevant comments from the other four are quoted below, but their names are withheld. I chose to reply to some of these comments. We are grateful to those who took the time and effort to comment on our evaluation of their animal model.
3. Adenocarcinoma ol the Colon
"I must admit I have never heard of your Committee and I do not know in what context you wish to comment on the document.
"I have always held the view that DMH carcinogenesis in the rat is a good 'model' of human disease. If any 'model' of disease in animals is good at all!'"
5. Giardiasis
"Your report would be strengthened by mentioning that studies of Giardia muris infection have shed light on the mechanisms responsible for opportunistic infections in patients with iinmunodeficiency diseases, notably AIDS. Specifically, I am referring to work in which depletion of CD4+ or CD8+ lymphocytes, by treatment in mice with monoclonal antibodies against each of these T-cell subsets, provided direct evidence that CD4+ cells are necessary for clearance of G. muris at a normal rate."
I requested that this investigator expound on this point, and I received the following reply:
"...With reference to Giardia muris infection, it was shown some time ago that treatment of mice with anti-CD4 (anti-L3T4) monoclonal antibody induces susceptibility to chronic G. muris infection ... As you will appreciate, this work ... shows that clearance of G. muris depends on the CD4 (helper) 1-lymphocyte subset, but not on the CD8 (cytotoxic) subset. It was subsequently shown that G. muris-infected mice treated with anti-L3T4 monoclonal antibody, and consequently depleted of CD4-positive lymphocytes, had an impaired ability to generate an intestinal IgA response to G. muris trophorzoites. . -
"Lymphocyte depletion studies of the type outlined in the previous paragraph enable direct examination of the role of different lymphocyte subpopulations in the clearance of organisms that infect the gastrointestinal tract and other organs, such as the respiratory tract. Direct information of this type cannot be obtained from either human studies or in vitro studies with cultured cells. Another example of depletion studies of the sort that I have outlined above comes from the experience of investigators studying the role of different T-cell subsets in protective immunity against Leishmania parasites...
"Since the major clinical impact of AIDS consists in the development of opportunistic infections, direct in vivo analysis of the relative importance of different lymphocyte subsets in the development of protective immunity against various microorganisms sheds light on the immunopathology of AIDS, and of other immunodeficiency diseases such as common variable hypogammaglobulinemia..
"Studies of Giardia muris infection have had the unexpected bonus of advancing the basic understanding of intestinal immunophysiology. Such understanding is necessary for identifying 'missing links' that account for the development of particular types of opportunistic infection in immunodeflcient individuals; for example, deficiency of either helper I cells, certain immunoglobulin isotypes, or cytokines such as gamma-interferon. You will also, no doubt, be aware that depletion of CD4-positive lymphocytes, by treatment of mice with anti-L3T4 monoclonal antibody, has established the importance of the CD4 lymphocyte subset in the development of murine lupus, anti-DNA antibodies, and glomerulonephritis. This work has provided insight into the pathogenesis of systemic lupus erythematosus.
"Overall, I believe that it is unnecessarily restrictive to categorize Giardia muris infection simply as a 'mouse model of human giardiasis.' As I have indicated above, Giardia muris infection, which is normally cleared immunologically, is an excellent example of an infectious process that can be studied analytically so as to identify the crucial components of the immune response that are necessary for its elimination."
Reply:
This investigator contends that the murine G. muris model has been useful for understanding the role of the immune system in controlling parasitic infection. This information is most relevant for immune deficiency states, most notably AIDS. While a discussion of animal models for AIDS research is outside the scope of this study, it is useful to examine this investigator's points in greater detail.
If the mouse and human immune system are identical or nearly identical, then it is possible that manipulations of the immune system of mice may provide information that is relevant for people. If, however, there are significant differences between humans and mice in the functions of the various components of the immune system, then it will be impossible to translate information from the murine model to human clinical issues. Given that there are numerous interactions between the different parts of the immune system, each difference between humans and mice in the function of each component of the immune system results in broader differences in the function of the entire immune system.
Recent research has indicated that there are profound differences in the function of T lymphocyte subsets between humans and mice. In order to understand these points, a little background is helpful. T cells are defined by their surface antigens. Researchers have noted homologous function between mouse T cells that have an antigen called L3T4 and human T cells that have an antigen called CD4. Both have been called "helper" cells.
Similarly, the "suppressor" cells, murine T cells with Lyt-2 and Lyt-3 antigens and human T cells that express CD8 antigen, have similar functions. However, these homologous antigens are not the same. Sprent and Webb wrote, "Like CD8 molecules, CD4 (L3T4) molecules show considerable divergence from their homologues in other species, suggesting that these molecules have undergone rapid evolution...(11) Furthermore, "The Lyt-2 chain is the homologue of the CD8 chain in man, the CD8 molecule consisting of homodimers and multimers of a single chain; there is no apparent counterpart of the Lyt-3 chain in man."(1)
More importantly, there is good reason to believe that these analogous T lymphocytes in humans and mice have differences in function. Kaplan observed:
The experiments that established the prior paradigm utilized the differential expression of two cell surface molecules to prepare two subpopulations of T lymphocytes (CD4 and CD8). However, within these two major subpopulations, any potential heterogeneity was obscured by averaging of the results. Late in the last decade, a new technique was established that allowed for a much more precise dissection of cellular function; the ability to clone and maintain, via antigenic stimulation, individual non-transformed T cells provided a source of large numbers of homogeneous cells for investigation. The interpretation of experimental results has been an order of magnitude more clear-cut with the clones. Certain cloned cytotoxic effector cells have been shown to provide help for inimunoglobulin secretion and secrete lymphokines that enhance T and B cell proliferation. T cell clones have been described that mediate both help and suppression of immunoglobulin synthesis, Cones of lymphocytes have been grown for months and even years, and it has been shown that functional attributes can oscillate during this time according to rules that are not yet completely elucidated.(2)
He continued:
The T lymphocyte has the potential for inputs through many different molecules and the ability to respond with an array of multipotential lymphokines. This capability defines the plasticity of lymphocytes ... Since T lymphocyte function is not fixed in such a rigid way, the terms "helper," "cytotoxic, and "suppressor" T lymphocytes seemed inappropriate.(2)
In contrast, murine T-cells do not appear to have the same diversity as human T-cells. Sprent and Webb wrote:
...in the case of human T cells, a variety of mAb are being used to separate T cells into a bewildering complexity of subsets displaying different functions. We have elected not to discuss this phenomenology, largely because, as yet, there is little or no evidence that murine T cells — the main subject of this review — are divided into more than two phenotypically and functionally distinct subsets.(1)
Thus, I take exception to this investigator’s practice, common in immunology, of using the terms L3T4 lymphocyte and CD4 lymphocyte interchangeably. There is good reason to doubt the relevance for humans of studies of immune system dysfunction in mice.
1. Sprent J, Webb SR: Function and specificity of T cell subsets in the mouse. Adv Immunol 1987;41:39-132.
2. Kaplan DR: Lexicographer-scientists and the plasticity of lymphocytes. Perspec Biol Med 1988;32:31-37.
6. Crigler-Najjar Syndrome
"I have studied the document you sent to me, and while overall I agree that the Gunn rat is not a good model for understanding the molecular basis of the enzyme defect in bilirubin UDP-glucuronosyltransferase (UDPGT), I do feel that it has been of considerable value to investigators in the field --without it I doubt whether we would know as much about Crigler-Najjar syndrome as we do today. It has at least provided a starting point for the investigation of this intriguing disorder at the molecular level.
"You may want to include some information we published very recently on Crigler-Najjar syndrome and the Gunn rat, and I have enclosed a reprint of the relevant article for your perusal. In this work we have analysed hepatic microsomes from a patient with Crigler-Najjar syndrome, type I by immunoblotting with an antibody which recognizes several rat and human UDPGT isoenzymes, including the bilirubin and phenol forms. We have previously published such an analysis of these microsomes, using the peroxidase system of detecting antigen/antibody interactions. [Burchell et al, Mol. Aspects Med. 9, 429-455 (1987)], which suggested that an isoenzyme of UDPGT was absent in this sample. In the latest paper, however, we have used the much more sensitive (up to 20-fold) alkaline phosphatase linked system to visualise the enzyme proteins on the Western blot, and this shows quite clearly that there are no major differences in the pattern of immunoreactive polypeptides in the liver microsomes for the Crigler-Najjar patient and in an age-matched normal infant. The original blots were difficult to interpret as a result of the induction of several UDPGT isoenzymes following treatment of the patient with phenobarbitone. Use of the alkaline phosphatase system means that much smaller amounts of protein can be loaded onto the gel, thereby avoiding this problem. This data strongly suggests that the molecular origins of the defect in the bilirubin UDPGT are fundamentally different in the Gunn rat and Crig!er-Najjar syndrome. This is the best evidence to date in this direction."
Mike Coughtrie, Ph.D.
8. Benign Monoclonal Gammopathy
"..The statement that BMG appears to represent a premalignant process in humans, or that 'there is evidence that BMG may be a precursor to MM in human beings' is an old-fashioned speculation ... the best answer is to quote Dr. Waldenstrom himself: After reviewing the studies of Axelsson and his own -- 'This must prove that late development of malignancy is quite rare.' Indeed, the most important study on this subject in humans is the nearly 25 years lasting follow-up study in unselected population (7000) in Sweden by Axelsson. Of the original 64 subjects with MG, only 4 were found to have or were suspected to have B-cell malignancy within the period of 20 years and no new malignancies had appeared 4 years after that. If BMG would represent a premyeloma stage, many more persons should have developed an overt disease. There is no hard evidence in the literature on MG individuals who later developed a malignancy that the MM developed from the same clone; it may be quite possible that the MG was MM from the beginning, however under successful control of the host defense mechanisms which eventually failed with aging and allowed progression of the disease. A smoldering MM with similar history has also been observed in the CS7BL mice. The problem is that B-cell malignancies show a very large scale of biological activity and clinical pattern, from a long lasting and relatively 'benign' form to a progressive malignant form; but that does not mean that another condition does not exist -- a benign proliferative disorder with a different biology and pathogenesis and in far higher frequency ... (many investigators) agree with the view that in both human and mouse situations, if BMG would be a premalignant condition, the development of malignancy from that BMG-cell clone should take place in much higher frequency than from an unaffected B. cell clone. This is obviously not the case. Nobody said that a transition from BMG to MM is impossible, but if it happens, it is too rare an event to mark BMG as a premalignant condition. Clinical studies in this respect have a great disadvantage, the selection of subjects. If not properly diagnosed (without a sufficiently long follow-up), a group denominated as BMG may consist of MG of different origin. In mice, the diagnosis of a given MG can be confirmed by transplantation experiments: Malignant MG are continuously transplantable for unlimited number of generations with high take frequencies, the cell is immortal; BMG is not immortal, it is transplantable, however for a maximum of 4 generations with decreasing take frequency; MG due to T<B cell deficiency are not transplantable. These facts are an enormous asset of studies in animal models, because they help to better understand the processes and conditions evolving in the proliferative disorders; they can not directly be confirmed in the human situation, one can not on purpose transplant a proliferating clone in humans! However, further research may lead to other evidence of differences between BMG and MM which could be verified in humans. The most likely level of such differences seems to me subcellular, differences in activations or disturbances of different oncogenes. Whether this will first be found in human or animal studies is of little importance to me, whatever way that would lead to our understanding of the processes should be used."
Based on these comments, as well as comments on other aspects of the paper, we made some revisions and sent this to the above investigator. The second reply included the following comments:
"(An) objection concerns (the) statement that 'BMG itself raises issues more related to basic science than to human patients, because BMG is asymptomatic.' First of all, any clinical evaluation of a disease or condition which does not take into account the results of a relevant 'basic' or applied research is worthless. This would degrade the medical science to the medieval age. What is your definition of basic and clinical research? Was the development of labeling indices for the differential diagnosis between MM and MGUS not a basic research? Further, BMG is not always asymptomatic, the paraprotein may possess an activity which may cause even serious problems (e.g., some neurological disorders). Isn't your approach a bit artificial?
"In my opinion, a good clinician should try to improve his understanding of all B-cell proliferative disorders, because MG of the malignant category is not the only one which is clinically 'relevant' (e.g., MG due to T<B immunodeficiency). For this purpose, studies in the animal models of these disorders are of enormous value. Specifically, studies of the animal model of BMG demonstrated (among others) that there exists a B-cell proliferative disorder in which the clonal cell is autonomous but not immortal. This is in contrast to MM, where the clonal cell is autonomous and 'immortal'. Is this impossible in humans? Isn't the mere fact that people start thinking, when reading the results of animal studies or discussing them, enough to give some benefit to the animal model of BMG? Of course, the human disease remains of primary importance and human 'material' should be investigated whenever possible. But you simply cannot do all experiments in humans...
"...Conceming ... an increased number of plasma cells in lymph nodes of C57BL mice, this was written only in the preliminary short communication and not in the two other definitive papers. The presence of an increased number of plasma cells in lymph nodes is a phenomenon related to aging (both in man and mice) but not to the presence of BMG."
Reply:
I appreciate the detailed review of this document. This investigator questions whether BMG in humans is a premalignant process, and there is evidence to suggest that it is not. The essay (see chapter 8) addresses this issue also, and I agree that 1) this issue is not settled and 2) there is reason to suspect that BMG and MM may be distinct pathological processes. This investigator also noted recent evidence that MM may occur in C57BL mice.(1) This paper did not mention whether the MM clone was identical to a previous BMG clone, which would have demonstrated conversion of BMG to MM in mice.
This investigator strongly encourages the study of all B-cell proliferative disorders, and I agree. It is possible that a better understanding of benign processes will yield insights to our understanding of pathologic conditions. I do not know whether the mouse model of BMG will be helpful in this regard, and therefore I do not advocate abandonment of this model. The opinion of this investigator, as well as Waldenstrom, is that the mouse model can serve important functions. However, it should be kept in mind that our study of animal models had a limited scope, and our findings should not be misconstrued as an indictment of all animal research.
Regarding basic research, I agree with this investigator that the definition of "basic" versus "applied" research is not always clear. We considered "basic" research to include studies designed to understand fundamental physiological processes. We considered "applied" research to be directed to the understanding or treatment of specific disease states. Certainly, many research projects do not fall neatly into one category or the other. We tended to avoid basic science issues in this study because it was necessary to limit the scope of the project. One reason we chose not to address basic science questions is that a cost-benefit analysis of basic research is most difficult. Because the benefits are, almost by definition, unforeseen and most likely in the distant future, the value of the research is almost impossible to ascertain. However, I believe that basic science research can have important clinical applications and can be worthy of support. Differences in anatomy and physiology between humans and animals may limit the utility of basic science research with animals. Whether such research is an efficient use of resources has not, to my knowledge, been fully addressed.
The investigator commented that transplantation studies have demonstrated that MM cell lines are immortal in mice, as opposed to BMG, which can be successively transplanted four times at most. While I agree that such studies cannot be done in humans, it may be possible to obtain analogous information using in vitro techniques. The advantage of such in vitro research, of course, is that human cell lines would be used. Given that we do not know the etiology of BMG or MM in humans or mice, it is possible that the diseases have fundamental differences. While there are similarities in clinical presentation, the diseases could differ in important respects. Because the B-cell proliferative disorders have a wide range of clinical expression in both mice and humans, it may be inaccurate to compare one B-cell disorder in mice to another in humans. Consequently, we cannot be sure that the transplantation studies in mice are relevant to human monoclonal proliferative disorders. On the other hand, in vitro work with human tissues, when possible, would seem to be more directly applicable to humans.
1. Radl J, Croese JW, Zurcher C, Vanden Enden-Vieveen MHM, de Leeuw AM: Multiple Myeloma, Model No. 360, in, Capen CC, Jones TC, Migaki G (eds): Handbook Animal Models of Human Disease Fasc. 17. Washington DC, Registry of Comparative Pathology, Armed Forces Institute of Pathology, 1979.
9. Cloboid Cell Leukodystrophy
"Thank you for sending me your article on the evaluation of the twitcher mouse as an animal model of human disease. My work solely concerned the therapeutic possibilities by bone marrow transplantation. I feel you have slightly underplayed this aspect as bone marrow transplantation has long been considered as the only option for otherwise untreatable cases of many different inborn errors of metabolism in man. Despite the findings to which you refer in this field, Yeager et al. markedly lengthened the survival of twitcher mice, and I and my colleagues found increased galactosylcerebrosidase activity in the bone marrow."
Mary J. Seller BSc PhD DSc
Reply
Hoogerbrugge et al. reported in 1988:
Transplantation of enzymatically normal congenic bone marrow was earlier found to result in prolonged survival and increased levels of galactosylceramidase in the visceral organs of twitcher mice. It is now reported that bone marrow transplantation results in increased galactosylceramidase levels in the central nervous system (CNS) ... the infiltration of enzymatically competent, donor-derived macrophages was accompanied by extensive remyelination in the CNS.(1)
These exciting results suggest that bone marrow transplantation does offer some hope for human patients with GLD. However, Suzuki et al., doing similar studies, cautioned, "...the presence of inclusions in oligodendrocytes in the twitcher with BMT, even after 100 days of age, indicates that the basic enzymatic defect involving oligodendrocytes could not be completely corrected with BMT."(2)
I was unable to find any evidence of BMT therapy for human GLD patients. I do not know if it is reasonable to attempt BMT on these children, given the considerable risks associated with the procedure, but the value of the animal research to resolve this issue is debatable. Given that murine GLD differs from the human condition in isoenzyme activity, onset of symptoms, clinical manifestations, and in other respects, the success or failure of BMT should have little influence on whether BMT is indicated for human babies with GLD. Fortunately, there are tests to assess whether BMT is likely to help alleviate a metabolic disease. In 1981, Hobbes et al. reported improvement of a patient with Hurler's syndrome after BMT, and they wrote:
These promising results merit cautious consideration of treatment by BMT of patients in the early stages of some types of MPS (mucopolysaccharidoses) and other lysosomal enzyme disorders, especially when correction of the defect in cultured flbroblasts by added enzyme can be demonstrated in vitro.(3)
1. Hoogerbrugge PM, Suzuki K, Suzuki K, et al.: Donor-derived cells in the central nervous system of twitcher mice after bone marrow transplantation. Science 1988;239:1035-1038.
2. Suzuki K, Hoogerbrugge PM, Poorthuis BJHM, Van Bekkum DW, Suzuki K: The twitcher mouse: Central nervous system pathology after bone marrow transplantation. Lab Invest 1988;58:302-308.
3. Hobbs JR, Barrett AJ, Chambers D, et al: Reversal of clinical features of Hurler's disease and biochemical improvement after treatment by bone-marrow transplant. Lancet 1981;ii:709-712.
10. Large Granular Lymphocytic Leukemia
"...Your review is well written and usually accurate in its interpretation. Specific comments ... (regarding the contention that) in rats, it (LGL leukemia) is an acute disease while in humans it is chronic. This statement is not quite correct. Remember that rats live a mean of 2.5 years while humans live a mean of 75 years. The leukemia in rats generally is clinically apparent for 2-6 weeks, This time period in humans may be equivalent to 60-180 weeks (1-3 years). Also, the rat disease is more uniform than in humans and represents the more malignant cases in the human LGL proliferative disease spectrum.
"In all five papers ... none of the authors credited the animal model with any important insights into the [human disease]." While this statement is correct, it is not definitive nor accurate. See my comments below.
"Review of Animal Models: Animal models can serve many purposes.
They can represent identical diseases in humans or serve as models for
studying certain aspects of human diseases. In the case of LGL leukemia,
(rats) were the first to document in any species, including humans, the
occurrence of LGL leukemia as a neoplastic disease entity ... Dr. Stromberg's
animal model paper subsequently stimulated interest in human LGL leukemia.
As you can see from the literature, in the past 5 years, several large
case series have been reported for the human disease spectrum. LGL leukemia
in rats can also serve as a model to study the mechanisms of cell killing
by lymphocytes. Several recent papers have appeared using transplanted
rat tumors or tumor cell cultures to isolate LGL granules and characterize
the cytolysins present in the granules. LGL leukemia is an important
endpoint for carcinogenesis testing (safety assessment) of chemicals
including food additives, pesticides and drugs. Recently ethylene oxide
was found to increase the incidence of LGL leukemia in F344 rats. A few
years later, epidemiological studies provided evidence for the leukemogenesis
of ethylene oxide in humans occupationally exposed ... Thus, animal models
can serve in many ways."
Reply
This investigator raised several important points about animal models in general, but I will first address points relevant to the animal model of human LGL leukemia, While it is possible that this model "represents the more malignant cases in the human LGL proliferative disease spectrum", it is also possible that this animal model represents a fundamentally different disease process from human LGL leukemia. As noted to chapter 10, LGL leukemia tends to have a benign course and generally does not require therapy. Second, I question whether "Dr. Stromberg's animal model paper subsequently stimulated interest in human LGL leukemia." Loughran and Starkebaum noted that McKenna et al. had reported in 1977 a blood dyscrasia associated with increased circulating LGL before this model was identified.
This investigator noted that in vitro studies with LGL tumor cells have enhanced our understanding of LGL function. Given that it is possible to use human LGL to develop these cell lines, it would appear that use of human tissues would be preferable for the understanding of human LGL properties. Whether F344 rats prove useful for carcinogenicity screening remains to be seen. I am unaware of data regarding sensitivity and specificity, and one example does not demonstrate that use of this model would have high predictive value.
Finally, the view that animal models may "serve as models for studying certain aspects of human disease" is widely held. Rarely, if ever, are animal model diseases identical to human diseases. If the animal and human diseases are not identical, insights to the disease process in the animal model need to be verified in human patients before they may be accepted as valid for people. Furthermore, as long as the animal model and human disease differ, discoveries about the animal disease would seem to have a low chance of being relevant to humans. Indeed, using an animal model that is not identical to a human disease involves a process that may be logically flawed. When the animal model differs from the human disease, investigators tend to focus attention on specific attribute(s) of the model. The choice of attributes necessarily involves finding those aspects of the animal disease that are similar to the human disease. Consequently, identification of "relevant" aspects of the animal model requires precedent knowledge of the human disease. Thus, insights to the animal model follow insights to the human disease, and it is difficult, if not impossible, for the animal model to lead the process of medical discovery.