ILAR Journal Online, Volume 41(2) 2000: Humane Endpoints for Animals Used in Biomedical Research and Testing

Online Issues

<< All Back-issues

<< This Issue's Table of Contents

ILAR Journal V41(2) 2000
Humane Endpoints for Animals Used in Biomedical Research and Testing

Humane Endpoints and Cancer Research
James Wallace

James Wallace, F.I.A.T., C.Biol., M.I.Biol., is a Fellow of the Institute of Animal Technology, Chartered Biologist, and Member of the Institute of Biology at The McElwain Laboratories Institute of Cancer Research, Surrey, United Kingdom.

Cancer is the term used to describe a number of diseases that are characterized by uncontrolled, abnormal growth of cells in tissues or the blood. Cancer development may be localized to specific tissues or cells or may spread from the primary tumor to other parts of the body to produce secondary tumor growth (metastases). Cancer is a serious disease common to humans and many species of animals. The worldwide incidence of human cancer is increasing (Boyle 1997), and the available therapies and outcome for certain malignancies, metastatic disease in particular, are less than satisfactory. Understanding the biologic processes that lead to the development of cancer, identifying potential carcinogens in the environment, and developing safe, effective treatments are major scientific and medical objectives.

Experimental animal models of cancer provide a link between how cancer develops at the molecular or cellular level and how the condition manifests and may be treated in complex, living organisms. In vivo experimental models of cancer allow the collection of scientific and preclinical data that is not available from cell cultures or human subjects. To this end, a variety of experimental tumor systems have been developed, mainly using laboratory rodents as hosts, to study cancer biology and treatment (Clarke 1996; Fidler 1986; Kallman 1987; Kerbel 1999; Welch 1997). The development of immunodeficient mouse models engrafted with human tumors has enabled anticancer drug/human tumor sensitivity and toxicity tests to be conducted. Genetically modified (transgenic) rodent models have been developed to promote the study of the molecular and genetic causes of the disease and to develop new therapeutic approaches (Amundadottir et al. 1996; Lipkin 1997; Rosenberg and Bortner 1999; Rovigatti et al. 1998).

Some experimental cancers can cause profound physiologic and metabolic changes that lead to an irreversible decline in the host. Laboratory animals may experience significant adverse effects as a result of experimentally induced cancers and the effects of investigative or treatment regimes (Montgomery 1990; UKCCCR 1997; van Loo et al. 1997). The routine, or intentional, use of either death of the animals or evidence of significant pain or distress as an experimental endpoint in scientific studies is to be discouraged (CCAC 1993; Hamm 1995; Olfert 1996; Tomasovic et al. 1998). In the Canadian Council on Animal Care document Ethics of Animal Investigation, it is stated, "toxicologic and biologic testing, cancer research and infectious disease investigation may, in the past, have required continuation until the death of animal. However, in the face of distinct signs that the processes are causing irreversible pain or distress, alternative endpoints should be sought that satisfy both the requirements of the study and the needs of the animal" (CCAC 1989, p. 2).

The United Kingdom Coordinating Committee also recognizes that animals used in cancer research may experience pain and/or distress and recommends the implementation of humane endpoints in cancer studies. The objectives they describe in Guidelines for the Welfare of Animals in Experimental Neoplasia (UKCCCR 1997) are quite clear; however, the practical methods to achieve them are less certain. Cancer is a complex condition, and the recognition and ranking of the severity of effects of experimental cancer on the animals can be problematic in practice. Development of universal scientific and humane endpoints for cancer studies is further frustrated by the number of scientific disciplines involved in cancer research, the range of scientific techniques and challenges employed, and the variety of tumor systems and animal models. However, a useful starting point is to encourage investigators to "seek the least traumatic techniques feasible that would allow the (scientific) question to be adequately evaluated" (Goy 1982, p. 346).

Scientific and Humane Endpoints

Scientific and humane endpoints are inextricably linked and should be developed concurrently. There is little purpose in terminating experiments prematurely if this action confounds the experimental objectives and potentially wastes animals. It is equally unacceptable to extend animal distress beyond the point required to meet the scientific objectives. Scientific endpoints are criteria that determine the outcome of an experiment. Humane endpoints are finite, compassionate limits placed on the amount of pain or distress any experimental animal will be allowed to experience. Scientific and humane endpoints utilize data derived from observation of changes in the condition of the animals and laboratory-based assays. Both scientific and humane endpoints should promote the objective of establishing a scientifically valid answer to the experimental question with a minimum of harm to the animals. This situation requires that specific scientific and humane endpoints are developed for each project and are subjected to continuous adjustment and review.

Clearly defined experimental objectives are essential to the development of both scientific and humane endpoints. The nature, intensity, and duration of the scientific challenge determine the potential consequences of the experiment for the animals. The experimental objective will determine the scientific approach, tumor system, animal model, and group sizes and will define the scientific and humane endpoints. For example, when the experimental objective is serial transplantation of tumors or determination of the oncogenicity of a cell line, the development of a single tumor of a predetermined size may act as both a scientific and humane endpoint. When the effects of therapy on the development of experimental tumors is to be evaluated, the adverse effects on the animals may be greater and a number of different scientific assays and humane endpoints may need to be employed. Whenever possible, scientific and humane endpoints should utilize criteria or assays that avoid significant animal distress or death.

Studies of tumor growth or therapy frequently utilize either in situ or excision assays to establish changes in the pattern of tumor growth or cell survival. Assays conducted in situ involve either various methods of determining changes in the pattern of growth of the tumors in the host or extension of the survival time of the host. Excision assays normally require removal of the tumor at the end of the experiment for additional laboratory evaluation.

Tumor growth or excision assays should replace "survival'' endpoints. Tumor growth delay, tumor doubling time (Td¹), tumor regression, and clonogenic survival of tumor cells are alternative assays for assessing the growth of solid tumors. When tumor growth delay, tumor regression, or Td is the scientific endpoint, the final size of the tumor(s) in both treated and untreated control animals should be the smallest practicable (e.g, 2 to 3 Tds of the size) at the beginning of treatment, rather than any fixed final tumor volume or weight. The potential problems with in situ assays are that (1) tumors may become too large, (2) tumor-associated disease may appear in the host, or (3) animals may die as a result of tumor development. Excision assays assess what fraction of the tumor cells are still clonogenic and can form colonies in vivo, in other animals, or in vitro ( Courtney and Mills 1978). Where either ascitic tumors or leukemias are involved, ex vivo clonogenic assays or the onset of clinical signs should replace moribundity or death of the host as an experimental endpoint.

Whenever possible, the potential consequences (in terms of pain, distress, or lasting harm) of the experimental challenge to the animals should be identified and controlled or eliminated at the outset (Wallace et al. 1990). A review of the literature relevant to the use of animal models in cancer research (Clarke 1997; Gart et al. 1986; Hanfelt 1997; Kallman 1987; UKCCR 1988; Welch 1997) will illustrate the practical and ethical issues involved in the design and conduct of cancer studies involving animals.

Animal welfare considerations have both qualitative and quantitative constituents. Experimental protocols should be subjected to statistical review to ensure that they employ refined experimental technique and use the minimum number of animals to produce valid scientific results (Gart et al.; 1986; Hanfelt 1997). Management of the experimental cancer program and animal care must also be considered (Schiffer 1997). Pilot experiments, dose range finding studies, or a staged experimental approach using a small number of animals can help determine the response to an unknown or novel experimental challenge and provide data to develop scientific and humane endpoints for later studies using larger groups of animals. These approaches are particularly encouraged when there is the potential for significant animal distress.

Tumor Characterization

Detailed knowledge of the growth characteristics and biology of the proposed tumor model and the onset and nature of any adverse effects on the animals is crucial to the establishment of both scientific and humane endpoints (Siemann 1987). Tumors in laboratory animals may be induced or transplanted or may occur spontaneously. Established methods of inducing experimental cancer in animals include exposure to carcinogenic substances, local or whole body irradiation, inoculation with viruses or malignant cells, genetic mutation, or (increasingly) the introduction of oncogenes by genetic modification of animals or transfection of cells. Experimentally induced tumors may develop at a predictable site and within a specific time frame, or their onset may be difficult to predict.

Tumors may be solid and localized or may be solid and metastasizing to produce secondary disease in other tissues. Solid tumors may develop either in the superficial tissues or internally. Certain types of malignancies may arise in the blood, bone marrow, or lymphatic system. In contrast to the development of clinical cancer, experimental tumor systems in animals are frequently characterized by a relatively short latent period and rapid tumor growth. The incidence, site of origin, and growth rate of an experimental tumor and the onset and nature of the adverse effects on the host will vary with (1) methods used to induce the tumors, (2) the biology of the tumor, (3) the site of development and the issues involved, (4) any associated experimental challenge, and (5) the response of the host.

The ideal experimental tumor model system would be stable and exhibit reproducible growth, behavior, morphology, and sensitivity to the potential experimental challenge. Immunogenicity, cell kinetics, and heterogeneity should be known. It should be possible to predict the incidence, origin, and time of onset of tumor development and any associated disease. However, experimental tumor cell lines are frequently genetically unstable and subject to constant selection processes at the cellular level in the host and through laboratory processing. Continuous serial transplantation in vivo or transfer into a host of a different genotype may transform the growth characteristics of a tumor line (Rockwell 1987). Typical changes in tumor biology associated with serial transplantation include a reduced latent period, an increase in tumor incidence, reduced tumor doubling time, increased malignancy or ability to metastasize (Visonneau et al. 1998), and changes in the sensitivity to therapeutic agents (Clarke 1997). If cancer studies are to yield reliable and reproducible scientific results, then care must be exercised to minimize experimental variables by scrupulously following experimental tumor propagation protocols (Welch 1997). Changes in the immunogenicity or antigenicity of tumor lines may alter the growth characteristics of the tumor and modify responses to experimental therapy. The strain, sex, age, health status, and genetic background of the host animal are additional variables that can influence tumor development. The genetic and microbiologic status of the tumor cell lines and the animal recipients should be ascertained to avoid experimental artifacts (McKisic et al. 1996; Schiffer 1997). Transplantable tumors may be induced orthotopically, in the tissue or site of origin, or ectopically, usually subcutaneously in the flanks or by intravenous injection. Changing the inoculation site may change the growth characteristics of the tumors (Morikawa et al. 1988; Volpe and Milas, 1990). Knowledge of the origin, incidence, and time of onset of spontaneously developing tumors in the host animal is necessary if experiments or animal welfare are not to be compromised.

Preparatory Procedures and Experimental Challenge

To promote the engraftment of some experimental tumor lines, it may be necessary to modify the recipient's immunologic or physiologic status. Low-level whole body irradiation or immunosuppressive agents are frequently used to further suppress the immune response of immunodeficient rodents before inoculation with human tumor xenografts. Surgical ablation of the endocrine glands and subsequent hormone supplementation may be necessary to assist in the establishment of hormone-dependent tumors. Experimental challenges may include novel or established anticancer treatments or the administration of potential carcinogens. Experimental protocols should indicate whether the test substances are to be administered at potentially toxic doses. Invasive techniques may be needed to induce experimental cancer, administer an experimental challenge, or monitor an animal's response during an experiment. These procedures may present a greater challenge to an animal than that of a developing tumor. Humane endpoints should consider the cumulative effect of all the experimental challenges on the animals.

Monitoring Tumor Development

Unexpected or uncontrolled tumor development can result in unnecessary animal distress or mortality. Studies on experimental tumors frequently cannot be initiated until a detectable tumor of the appropriate size is present. During the experimental phase, the development of the tumor and condition of the animal should be monitored frequently. Longitudinal collections of data on tumor growth throughout the study may yield more powerful results than simple assessment of the final tumor burden. Repeated measurements will also assist in profiling tumor development and possible adverse effects on the animal. Experiments should be completed before tumor development or tumor-associated disease causes death or a significant deterioration in the host. Limits on tumor development will be determined by the experimental objectives, growth characteristics of the tumor, site of tumor development, and clinical condition of the animals. Systems must be in place to monitor the animals for tumor development, with established limits placed on the tumor burden or severity of tumor-associated disease. Frequency of monitoring will be determined by the growth characteristics of the tumor and the onset of critical phases in the experimental process or development of the tumor. Animals with newly transplanted tumors may require only routine observation during the early stages of tumor development; however, animals in the terminal stages of tumor-associated disease or drug toxicity may require monitoring several times a day. The incidence and growth rate of spontaneous tumors, some chemically induced tumors, or those arising in trans-genic animals may be difficult to predict. When grown as ascites, some tumors such as the mouse L1210 leukemia (Kline et al. 1972) and the Leydig cell rat tumor (Cooke et al. 1979) are rapidly lethal when growth is uncontrolled. Care must be taken to limit the distress of animals bearing potentially lethal tumors, particularly those used as untreated con-trois and not receiving therapy. When experiments involve the use of potentially lethal tumors, the experimental phase should be completed before the onset of the period of tumor-induced morbidity or death.

Monitoring the development of solid tumors in the superficial tissues and the condition of the animals is relatively straightforward. Rodents can sustain large, superficial, noninvasive tumors without any apparent adverse effects or restriction on their normal behavior. However, it is desirable to place limits on tumor burden, consistent with the objectives of the scientific inquiry. The UKCCCR guidelines (UKCCCR 1997) recommend limiting solid tumors to a maximum of 10% of the host's body weight. However, estimating the weight of a tumor in relation to the overall weight of the animal can be difficult, particularly if the tumor is growing while the host is losing weight. Measuring the tumor dimensions with calipers is a more reliable method of monitoring tumor growth. Endpoints requiring either tumors of a fixed size or treatment over a fixed time period may present difficulties, particularly if tumors grow very rapidly or slowly, ulcerate, or invade the underlying tissues. Large, superficial tumors may inhibit normal activity due to their bulk or anatomic location.

Some protocols may require inoculating tumors into multiple sites, usually subcutaneously. Estimating the number of tumors an animal can be expected to support is difficult, and no specific guidance is available. Whether it is preferable to use fewer animals with multiple tumors or a greater number of animals with only one tumor must be a balance between good science and animal welfare. The use of multiple tumor implants may bring into question the independence of the resulting scientific data with respect to possible tumor interaction. Multiple tumors may coalesce and compromise the animal's well-being; however, the use of multiple tumor implants in a single animal has been reported to reduce inter-animal variability and strengthen experimental data (Hanfelt 1997; Heitjan et al. 1992).

Tissue necrosis or ulceration of the skin overlying the developing tumor may occur. Ulceration can occur when tumors either develop subcutaneously or are inoculated into the derma of the host. Other causes of ulceration can include (1) situations in which large tumors develop on the ventral surface and are subjected to constant abrasion; and (2) certain tumor types (e.g., papillomas); and (3) cell lines that are predisposed to ulcerate. Ulceration may occur when the tumors are relatively small. When ulceration is characteristic of the tumor line, the aim should be to complete the experiment in the latent period before ulceration. As soon as a tumor has ulcerated, the growth pattern will alter, which may be sufficient grounds for terminating the experiment. Ulcerated or necrotic tissue may result in a continuous loss of body fluid and/or infection. When it is necessary to maintain an animal with an ulcerated tumor, both the status of the ulcerated tissues and the animal's overall condition must be assessed daily.

Internal Tumors

Tumors growing internally or tumor-associated metastatic disease may be difficult to monitor (Welch 1997). Uncontrolled development of tumors in internal organs may cause distress or death. There is increasing interest in growing experimental tumors orthotopically, for example, in the tissue or site of origin such as in breast, bladder, or prostate (Hoffman 1999; Kerbel et al. 1991; Killion et al. 1999; Stephenson et al. 1992), in preference to subcutaneous sites. Orthotopic culture or inoculation into internal tissues/organs may increase the risk of the animals' distress because the growth and metastis of these tumors may be difficult to monitor. Tumor development in sites where space for tumor expansion is restricted (e.g., the eye, footpad, brain, muscle, or tail) may cause distress and require specific justification and monitoring. Metastatic deposits in the lungs and other internal tissues may occur as a result of either inoculation of tumor cells into the vascular system, or cells escaping from a primary tumor, or from cancers of the blood or lymphatic systems. The removal of a primary tumor may stimulate the development of latent metastatic disease elsewhere.

The onset of clinical signs or endpoint assays conducted at autopsy are commonly used to confirm the presence of visible or microscopic tumor deposits. Alternative methods to detect or profile internal tumors in live animals may include serial termination of animals, investigative surgery, medical diagnostic imaging techniques, or palpation for either organ enlargement or tumor development. Surrogate markers of tumor development, such as consistent loss of body weight (Redgate et al. 1991) and behavioral or neurologic changes, may also be used as alternatives to either limiting clinical signs or death. The increasing availability of biomarkers such as bacterial lacZ (Kruger et al. 1998-99 ) or green fluorescence protein gene (Chalfie et al. 1994; Hoffman 1999) to label tumor cells can provide an opportunity to refine the detection of tumor deposits. Tagging metastatic cells with genes or stains can identify single cell micrometastatic deposits in tissues (Welch 1997). Video microscopy techniques can detect labeled micrometastases in live animals (Kan and Liu 1999). The presence of human DNA or tumor biomarkers can be used to detect tumor deposits or circulating cancer cells in an animal engrafted with human tumors.

Cancers of the Blood and Hemopoietic and Lymphatic Systems

Leukemias can be detected in the living animal by examining blood samples for the presence of circulating cancer cells or changes in the cellular constituents of the blood. Increases in circulating tumor cells or changes in blood constituents can forecast the onset of clinical symptoms. When human leukemia cells are engrafted into immunodeficient mice, the number of circulating human cancer cells is less predictive of either the onset or severity of clinical symptoms. In the absence of reliable laboratory-based assays, animals with leukemia or lymphomas should be observed for early clinical signs such as anemia, loss of condition or weight, and enlargement of the spleen and lymph nodes. Scientific end-points should precede limiting clinical signs such as consistent weight loss, clinical anemia, apathy, impaired respiration, or death.

Ascitic Tumors

Abdominal distension, anemia, solid tumor deposits, and loss of condition are associated with the development of ascites. When the L1210 mouse leukemia (REF) is grown as ascites, it is lethal approximately 8 days after inoculation. The terminal period is associated with abdominal ascites, dysponea, and piloerection. During the terminal phase, the animals should be inspected and continually assessed for early termination several times each day. Tumor cell survival assays may be used as surrogate endpoints in place of limiting clinical signs or death of the host.

Spontaneous Tumors

Spontaneous tumors can arise in most tissues, and their development most closely resembles the clinical course. Typically, spontaneous tumors are characterized by a variety of histologic types, slow development, and cell turnover; if malignant, they may be poorly differentiated. The onset of spontaneous neoplasia is frequently unpredictable, and cancer cases rarely arise in sufficient numbers or at the appropriate time to provide sufficient subjects for experimental purposes. The type of spontaneous tumor, incidence, and clinical onset of tumor-associated disease will vary with the species, strain, age, and sex of the host. Diet or other nonexperimental variables may also influence tumor incidence and time of onset. Knowledge of the background incidence and biology of spontaneous tumors in the host animal is essential if their development is not to conflict with experimental cancer studies or cause unnecessary animal distress or mortality. Some strains of laboratory mice such as AKR and SJL/J have a predictable incidence of spontaneous lethal tumors, which may limit their use in some scientific studies.

Paraneoplastic Syndromes

Tumor development may be associated with significant disturbance of the host's normal physiologic or metabolic processes. Cachexia, one of the most serious and obvious conditions, is characterized by an energy imbalance leading to consistent weight loss, wasting of muscles, and eventual death. Enhanced nutritional intake does not usually improve the condition. Cachexia may be associated with the use of specific tumor lines such as the MAC 16 mouse colonic adenocarcinoma (Bibby et al. 1987) or Leydig cell rat tumor (Cooke et al. 1979); however, all tumor-bearing animals should be monitored for signs of cachexia. The onset of intentionally induced cachexia in individual animals may be uncertain. A significant and sustained loss of weight may be necessary before cachexia is confirmed. The scientific objective should be to complete the experimental phase as soon as possible after the onset of cachexic symptoms but before the appearance of limiting clinical signs. Humane endpoints for cachexic animals should limit the loss of weight (maximum of 25% less than starting body weight) and emaciation, as indicated by condition scoring.

Cancer in Genetically Modified Animals

Transgenic and knockout animal models with a genetically engineered predisposition to develop cancer are being used increasingly in cancer research (Amundadottir et al. 1996; Donehower et al. 1992; Lipkin 1997; Rovigatti et al. 1998; Rosenberg and Bortner 1999). Cancer development is a multistage process, and there is the expectation that transgenic animal models will provide information on the role of specific genes in tumorigenesis and in the development of novel anticancer therapies. The p53 knockout mouse is predisposed to develop a variety of different neoplasms relatively early in life (Donehower et al. 1992). The APC^minmouse (Moser et al. 1990) is an ethylnitrosurea-induced mutation adenomatous polyposis coli (APC) gene. APC^minmice mimic the inherited colon cancer syndrome in humans. The gene is lethal in the homozygous state, but min/+ mice will develop tumors of the small intestine that result in intestinal obstruction and anemia. A major welfare concern with the genetic manipulation of animals is the unpredictability of the adverse effects, particularly cancer, affecting large numbers of animals. When double or even triple transgenes are created, these changes may act synergistically, increasing the incidence and speed of onset of experimental cancer and other effects (Amundadottir et al. 1996; Rosenberg and Bortner 1999). Tumorigenic variance may be compounded by differences in the genetic background of the genetically modified animals (Noguchi and Noguchi 1985). Continuous monitoring of genetically modified animals for tumor development or phe-notypic abnormalities may be necessary from birth onward.

Humane Endpoint Considerations

Protocols to monitor the condition of the animals and respond to any deterioration in their condition is an essential part of implementing scientific and humane endpoints. All those concerned with the use and care of the animals should be familiar with the monitoring systems and how to respond. Observation frequency will be determined by the animals' status and progress of the experimental challenge. Useful monitoring systems may include a list of observations of general appearance and behavior, measurable clinical signs, signs of pain or distress, and specific observations directed to changes arising from the scientific challenge (Irwin 1968). Experimental protocols should include a full description of the proposed scientific challenge and a description of any tumor models used.

The objective assessment of pain or distress in animals arising as a result of scientific procedures is fraught with difficulties. A number of systems have described specific and general clinical signs of pain and distress in laboratory animals (Montgomery 1990; Morton and Griffiths 1985; UFAW 1989; Wallace et al. 1990). Categorizing adverse signs and applying a numerical value to each sign permit evaluation of the increasing severity of the symptoms and the cumulative effect on the animal (Morton and Griffiths 1985). Limiting clinical signs for tumor-bearing animals are described in the UKCCCR guidelines (UKCCCR 1997). Developing distress scoring systems for cancer studies focuses attention on the animal's condition and can lead to refinements of humane end points. However, although these systems can reliably identify animals that are unwell, they do not automatically provide a reliable prognosis of either long-term survival or time of death. Not all adverse clinical signs are equally indicative of the seriousness of an animal's condition. Continuing body weight loss, progressive dehydration, anorexia, hypothermia, and lethargy--if present to-gether-indicate that an animal is in a life-threatening situation. Acute body weight loss, piloerection, or hunched posture indicates that an animal is not well but does not confirm the seriousness of the underlying condition. When cancer studies are involved, emaciation may be a more reliable indicator of a serious condition than loss of body weight. There is some concern that terminating an experiment early on the basis of a series of clinical observations alone could lead to a loss of essential data and a waste of animals. An association between the observed changes in the animal's condition, the underlying cause, and the potential outcomes should be established. Continued progress in refining experimental and humane endpoints may require rigorous validation of distress scoring systems and consideration of alternative strategies.

The acute physiologic and chronic health evaluation (APACHE) system is a severity of disease classification system used by physicians for at-risk human patients (Knaus et al. 1985). The system uses a combination of patient health evaluation, physiologic, biochemical, and hematologic indicators (e.g., blood pressure, heart rate, serum sodium, and serum creatinine) to assess the condition of seriously ill human patients and provide a prognosis of the outcome. Increasing deviation (above and below normal values) is scored using a scale from 0 to 4. Overall assessment of the subject's condition can then be made using a combination of specific physiologic values, expert consensus, or, less satisfactorily, mathematical score alone. A similar approach using specific batteries of tests could be adapted for use in laboratory animals under a variety of circumstances.

Perturbed hematologic, physiologic, metabolic, or biochemical values may precede the onset of clinical signs. Uti-nary, hematologic, or other biomarkers of organ dysfunction may provide more objective and reproducible experimental or humane endpoints than clinical observations alone (Poon and Chu 1998, 1999). White blood cell and hematocrit values can indicate hematologic toxicity (Clarke 1997), infection, anemia, and other conditions and provide a surrogate end-point in place of limiting clinical signs. In addition to tissue changes, clinical and experimental cancer can produce general or specific physiologic, biochemical, and metabolic disturbances in the host. Plasma lactic dehydrogenase activity is increased in many malignancies, particularly lymphoma (Pannell and Kotasek 1997). The presence of some malignancies and response to treatment can be detected by the presence of specific tumor markers, such as carcinoembry-onic antigen (CEA) for colorectal cancer and alphetoprotein in hepatocellular cancer. Organ dysfunction, tissue damage, or repair associated with tumor development can also be determined from altered physiologic, hematoligic, or biochemical values.

Laboratory assays utilizing small quantities of body fluids or tissues and telemetry now make it possible to monitor metabolic or physiologic values in small animals. The i-STAT is a hand-held, point-of-care human blood analyzer that utilizes single-use multisensor cartridges (Bingham et al. 1999). Thirteen hematologic and biochemical values including urea, glucose, sodium, potassium, chloride, hemat-ocrit, pH, PCO_2,and PO₂can be assessed simultaneously from a single small (50- to 100-~L) blood sample. This analyzer and similar systems may have a role in monitoring the changing physiologic status of small laboratory animals.

New technological developments in electronics, molecular biology, and diagnostic imaging enable scientists to monitor subtle physiologic and metabolic changes taking place in the experimental animal. These developments, together with an approach of regarding experimental animals as "patients," can yield improved scientific data from fewer animals and lead to more refined scientific and humane endpoints. The use of objective measurements of dysfunction in animals will also assist in the process of developing uniform experimental and humane endpoints in cancer studies.

References

Amundadottir LT, Merlini G, Dickson RB. 1996. Transgenic models of breast cancer. Breast Cancer Res Treat 39:119-135.

Bibby MC, Double JA, Ali SA, Fearon KC, Brennan RA, Tisdale MJ. 1987. Characterisation of transplantable adenocarcinoma of the mouse colon producing cachexia in recipient animal. J Natl Cancer lnst 78:539-546.

Bingham D, Kendall J, Clancy M. 1999. The portable laboratory: An evalu-

ation of the accuracy and reproducibility of i-STAT. Ann Clin Blochem 36(Pt 1):66-71. Boyle P. 1997. Global burden of cancer. Lancet 349:23-26.

CCAC [Canadian Council for Animal Care]. 1989. Policy Statement, Ethics of Animal Investigation. Ottawa, Canada: CCAC.

CCAC [Canadian Council for Animal Care]. 1993. Ethics of animal investigation. In: Olfert ED, Cross BM, McWilliam AA, eds. Guide to the Care and Use of Laboratory Animals. Vol 1, 2nd ed. Ottawa, Canada: CCAC. p 315-350.

Chalfie M, Tu Y, Euskirschen G, Ward WW, Prasher DC. 1994. Green fluorescent protein as a marker for gene expression. Science 263:802-805.

Clarke R. 1996. Human breast cancer cell line xenografts as models of breast cancer--The immunobiologies of recipient mice and characteristics of several tumorigenic cell lines. Breast Cancer Res Treat 39:69-86.

Clarke R. 1997. Issues in experimental design and endpoint analysis in the study of experimental cytotoxic agents in-vivo in breast cancer and other models. Breast Cancer Res Treat 46:255-278.

Cooke BA, Lindh LM, Janszen FHA, van Driel M J, Bakker CP, van der Plank MP, van der Molen HL. 1979. A Leydig cell tumour. A model for the study of lutropin action. Biochim Biophys Acta 583:320-331.

Courtney DV, Mills J. 1978. An in vitro colony assay for human tumors grown in immune suppressed mice and treated in vivo with cytotoxic agents. Br J Cancer 37:261-268.

Donehower LA, Harvey M, Slagle BL, McArthur M J, Montgomery CA Jr, Butel JS, Bradley A. 1992. Mice deficient for p53 are developmentally normal but susceptible to spontaneous tumors. Nature 19:215-221.

Fidler IJ. 1986. Rationale and methods for the use of nude mice to study the biology and therapy of human cancer metatastsis. Cancer Metast Rev 5:29-49.

Gart JJ, Krewski D, Lee PN, Tarone RE, Wahrendorf J. 1986. In: Gart JJ, Krewski D, Lee PN, Tarone RE, Wahrendorf J, eds. Statistical Methods in Cancer Research. Vol IlL The Design and Analysis of Long Term Animal Experiments. Lyon, France: IARC Scientific Publication No. 79:1-219.

Goy RW. 1982. Policy statement on principles of the ethical uses of animals at the Wisconsin Regional Primate Center. Am J Primatol 3:345-347.

Hamm TE. 1995. Proposed institutional animal care and use committee guidelines for death as an endpoint in rodent studies. Contemp Top 34:69-71.

Hanfelt JJ. 1997. Statistical approaches to experimental design and date analysis of in-vivo studies. Breast Cancer Res Treat 46:279-302.

Heitjan DF, Derr JA, Satyaswaroop PG. 1992. The multi-site tumor transplantation model for human endometrial carcinoma: A statistical evaluation. Cell Prolif 25:193-203.

Hoffman RM. 1999. Orthotopic transplant mouse models with green fluorescent protein-expressing cancer cells to visualise metastasis and an-giogenesis. Cancer Metast Rev 17:271-272.

Irwin S. 1968. Comprehensive observational assessment: In: A systematic, quantitative procedure for assessing the behavioural and physiologic state of the mouse. Psychopharmacology (Berlin) 13:222-257.

Kallman RF, ed. 1987. Rodent Tumor Models in Experimental Cancer Therapy. New York: Pergamon Press.

Kan Z, Liu, T-J. 1999. Video microscopy of tumor metastasis: Using the green fluorescent protein (GFP) gene as a cancer-cell-labeling system. Clin Exp Metast 17:49-55.

Kerbel RS. 1999. What is the optimal rodent model for anti-tumor drug testing? Cancer Metast Rev 17:301-304.

Kerbel RS, Cornil I, Theodorescu D. 1991. Importance of orthotopic transplantation procedures in assessing the effects of transfected genes on human tumor growth and metastasis. Cancer Metast Rev 10:201-215.

Killion J J, Radinsky R, Fidler IJ. 1999. Orthotopic models are necessary to predict therapy of transplantable tumors in mice. Cancer Metast Rev 17:279-284.

Kline I, Gang M, Tyrer DD, Venditti JM, Artis EW, Goldin A. 1972. Evaluation of antileukemic agents in advanced leukemia LI210 in mice. X. Cancer Chemother Rep 3:1-69.

Knaus WA, Zimmerman JE, Wagner DP, Draper EA, Lawrence DE. 1985. APACHE II--A severity of disease classification system. Crit Care Med 13:818-829.

Kruger A, Schirrmacher V, Khokha R. 1998-99. The bacterial lacZ gene: An important tool for metastasis research and evaluation of new cancer therapies. Cancer Mestast Rev 17:285-294.

Lipkin M. 1997. New rodent models for studies of chemopreventive agents. J Biochem 28/29(Suppi): 144-147.

McKisic MD, Patzurzo FZ, Smith AL. 1996. Mouse parvovirus infection potentiates rejection of tumor allografts and modulates T cell affector functions. Transplantation 61:292-299.

Montgomery CA. 1990. Oncologic and toxicologic research: Alleviation and control of pain and distress in laboratory animals. Cancer Bull 42:230-237.

Morikawa K, Walker SM, Nakajima M, Pathak S, Jessup MJ, Fidler IJ. 1988. Influence of organ environment on the growth, selection and metastasis of human colon carcinoma: Cells in nude mice. Cancer Res 48:6863-687 I.

Morton DB, Griffiths PHM. 1985. Guidelines on the recognition of pain, distress and discomfort in experimental animals and an hypothesis for assessment. Vet Rec 116:431-436.

Moser AR, Pitot HC, Dove WF. 1990. A dominant mutation that predisposes to multiple intestinal neoplasia in the mouse. Science 247:322-324.

Noguchi T, Noguchi M. 1985. A recessive mutation (ter) causing germ cell deficiency and a high incidence of congenital testicular teratomas in 129/Sv-ter mice. J Natl Cancer Inst 75:385-392.

Olfert ED. 1996. Considerations for defining an acceptable endpoint in toxicological experiments. Lab Anim 25:28-33.

Pannell P, Kotasek D. 1997. Cancer and Clinical Biochemistry. Clinical Biochemistry in Medicine Series. London: Association of Clinical Biochemists, ACB Venture Publications.

Poon R, Chu I. 1998. Urinary biomarkers as humane endpoints in toxicological research. In: Hendriksen CFM, Morton DB, eds. Humane End-

points in Animal Experiments for Biomedical Research. Proceedings of the International Conference, Zeist, Netherlands. London: Royal Society of Medicine Press.

Poon R, Chu I. 1999. The role of blood tests in the refinement of human endpoints for animal testing. Proceedings of the Third World Congress on Alternatives and Animal Use in the Life Sciences. Bologna, Italy. (Forthcoming).

Redgate ES, Deutsch M, Boggs SS. 1991. Time of death of CNS tumor-beating rats can be reliably predicted by body weight-loss patterns. Lab Anim Sci 41:269-273.

Rockwell S. 1987. Maintenance of tumor systems and appropriate treatment for experimental tumors. In: Kallman RF, ed. Rodent Tumor Models in Experimental Cancer Therapy. New York: Pergamon Press. p 29-36.

Rosenberg MP, Bortner D. 1999. Why transgenic mice and knockout animals models should be used (for drug efficacy studies in cancer). Cancer Metast 17:295-299.

Rovigatti U, Afanasyeva T, Brander S, Hainfellner JA, Kiess M, Maddalena A, Malin G, Rtilicke T, Steinbach J, Weissenberger J, Aguzzi A. 1998. Transgenic mice as research tools in neurocarcinogenesis. J Neurovirol 4:159-174.

Schiffer SP. 1997. Animal welfare and colony management in cancer research. Breast Cancer Res Treat 46:313-331.

Siemann DW. 1987. Satisfactory and unsatisfactory tumor models: Factors influencing the selection of a tumor model for experimental evaluation. In: Kallman RF, ed. Rodent Tumor Models in Experimental Cancer Therapy. New York: Pergamon Press. p 12-15.

Stephenson RA, Dinny CPN, Gohki K, Ordonez NG, Killion J J, Fidler Ii. 1992. Metastatic model for human prostate cancer using orthotopic implantation in nude mice. J Natl Cancer Ins 84:951-957.

Tomasovic SP, Coghlan LG, Gray KN, Mastromarino AJ, Travis EL. 1998. IACUC evaluation of experiments requiring death as an endpoint: A cancer centre's recommendations. Lab Anim 17:31-34.

UFAW [Universities Federation for Animal Welfare]. 1989. Guidelines for the Recognition and Assessment of Pain in Animals. Association of Veterinary Teachers and Research Workers, eds. Potters Bar: UFAW.

UKCCCR [United Kingdom Coordinating Committee on Cancer Research]. 1988. Guidelines for the Welfare of Animals in Experimental Neoplasia. London: UKCCCR.

UKCCCR [United Kingdom Coordinating Committee on Cancer Research]. 1997. Guidelines for the Welfare of Animals in Experimental Neoplasia London: UKCCCR.

van Loo PLPE, Everse LA, Bernsen MR, Baumans V, Hellebrekers LJ, Kruitwagen CLJJ, den Otter W. 1997. Analgesics in mice used in cancer research: Reduction of discomfort? Lab Anim 31:318-325.

Visonneau S, Cesano A, Torosian MH, Miller EJ, Santoli D. 1998. Growth characteristics and metastatic properties of human breast cancer xenografts in immunodeficient mice. Am J Pathol 152:1299-1311.

Volpe JPG, Milas L. 1990. Influence of tumor transplantation methods on tumor growth rate and metastatic potential of solitary tumors derived from metastases. Clin Exp Metast 8:381-389.

Wallace J, Sanford J, Smith MW, Spencer KV. 1990. The assessment and control of the severity of scientific procedures on laboratory animals. Report of the Working Party of the Laboratory Animal Science Association. Lab Anim 24:97-103.

Welch DG. 1997. Technical considerations for studying cancer metastases in-vivo. Clin Exp Metast 15:272-306. ¹Abbreviation used in this article: Td, tumor doubling time.