ILAR Journal Online, Volume 38(1) 1997: Unusual Mammalian Models

Online Issues

<< All Back-issues

<< This Issue's Table of Contents

ILAR Journal V38(1) 1997
Unusual Mammalian Models

The Laboratory Opossum (Monodelphis domestica) in Laboratory Research
John L. VandeBerg and Edward S. Robinson

John L. VandeBerg, Ph.D., is Scientific Director and Edward S. Robinson, Ph.D., is a scientist at the Southwest Foundation for Biomedical Research, San Antonio, Texas.

The following article is a reprint of a manuscript from a book entitled Recent Advances in Marsupial Biology, edited by N.R. Saunders and L. Hinds, and published by the University of New South Wales Press, Sydney. ILAR Journal gratefully acknowledges permission from the editors and the publisher to reprint this manuscript.

INTRODUCTION

The South American gray short-tailed opossum, Monodelphis domestica, is a small (80-120 g), nocturnal marsupial native to Brazil and adjacent countries. It is docile, breeds readily in captivity, and produces large litters (typically 6-13). Since the first M. domestica were imported into the USA in 1978, efforts to develop this species as a laboratory marsupial have been highly successful (VandeBerg 1983, 1990). The standard laboratory diet is a commercial pelleted fox food provided ad libitum, and the animals are maintained in polycarbonate or polypropylene rodent cages under unremarkable conditions (VandeBerg 1990). This is the only marsupial species that has been produced in captivity in very large numbers, i.e., tens of thousands. Breeding colonies currently exist at academic and research institutions in Germany, the United Kingdom, Brazil, Australia, and at several locations in the USA. Because of its many advantageous characteristics as a laboratory animal and its economical production in captivity, M. domestica has become a prototype species for basic research on marsupial biology, in much the same way that the laboratory mouse is a prototype species for basic research on the biology of eutherian or placental mammals. Consequently, we now refer to M. domestica as the laboratory opossum.

The status of M. domestica as a laboratory animal was reviewed by VandeBerg (1990). Much progress has been made since then, both in basic biological research with this species and in its development as an animal model for research on human diseases. The present review focuses on recent progress.

MONODELPHIS Domestica AS A LABORATORY ANIMAL

Current Composition of Colony

Table 1 summarizes the origin of the current population of laboratory opossums. There were seven separate introductions of animals into captive breeding between 1978 and 1993; the animals came from four different localities in three states of Brazil, and a single locality in Bolivia. In total, 38 animals were imported, of which 28 have contributed to the current gene pool as founders. Only portions of the genomes of each of the 28 founders are still present, because some genes are lost in each generation by chance. For example, between 1978 and 1984, a computer simulation based on the pedigree structure of the colony estimated that 31% of the genetic material of the original nine animals imported in 1978 already had been lost (MacCluer and others 1986). Therefore, the genetic variability present in the colony is much less than that which was present in the 28 founders. van Oorschot and others (1992a) have described in detail the geographic origins of the founders obtained by 1988 and genetic differences between the groups descended from them.

Figure 1 illustrates the locations at which the founders were trapped. Some of these populations are separated by major rivers and mountain ranges which may substantially inhibit the transfer of genetic material between populations.

Founding animals from the five populations are designated in our laboratory as from Populations 1, 2, 3, 4, or 5, respectively, as are any animals descended entirely from founders of any one population. Population 1 is designated as the reference population, and interbreeding among populations has been restricted to Population 1 and any one other population. Animals descended from ancestors from Population 1 and Population 2, and having less than 30% of their genetic material derived from Population 1, are designated as members of Admixed Population 2. Admixed Population 3 has up to 37.5% of its genetic material from Population 1, and Admixed Population 4 has 50% of its genetic material from Population 4. The Bolivian animals (Population 5) have not been crossed with the Brazilian stocks.

Production Rate

More than 29,000 progeny have been weaned in our colony since 1980. Currently, the colony has approximately 2,400 individuals of which many are used in experimental research rather than in the breeding program. The breeding colony consists of about 700 adult females and 450 adult males, and it produces approximately 4,600 weanlings per year, in addition to sucklings used experimentally before weaning. Approximately 300 individuals are provided to other institutions each year as experimental subjects or breeding nuclei. The rest are used as research subjects, or as replacement breeders to perpetuate stocks or to continue the development of inbred strains. Detailed records have been kept on each animal in the colony, which is fully pedigreed, and the data base is computerized. Some pedigrees extend back as many as 22 generations.

Inbreeding

Because of the low number of founders and the large number of animals produced over many generations, a limited amount of inbreeding has been unavoidable in the colony at large. In addition, some groups have been purposely inbred over several generations to develop partially inbred strains. Brother-sister matings were used in initial attempts to develop inbred strains. One stock achieved an inbreeding coefficient of 0.911 before it was lost, but attempts to produce fully inbred strains (F = 0.986, equivalent to that achieved after 20 consecutive generations of full-sib matings) have not succeeded because of rapid decline in fecundity and infant survival after five or six generations. We hypothesize that the low recombination rate in females by comparison with eutherian mammals (van Oorschot and others 1992b, 1993) inhibits the formation of new gene combinations sufficiently quickly in brother-sister mating schemes to enable the acquisition of gene combinations compatible with an inbred state. However, we are succeeding in gradually increasing the inbreeding coefficients in partially inbred stocks by less severe inbreeding schemes, and are hopeful of achieving fully inbred strains in this manner. At the present time, the colony has 34 partially inbred stocks. Of those that continue to have high fecundity and infant survival rates, the most highly inbred has an average inbreeding coefficient of 0.674. Approximately 600 members of the colony have inbreeding coefficients in excess of 0.6. Since more than 60% of their gene loci have alleles that are identical by descent, these animals are valuable for research designed to detect recessive genes for susceptibility to experimentally induced diseases. They also are especially useful for intercrossing and backcrossing for linkage analysis and gene mapping purposes.

Increasing use of M. domestica

Animals superfluous to our own research needs have been available for distribution to other laboratories since 1984, and the species has already made important contributions in many areas of basic and biomedical research. The annual number of publications, which were indexed in BIOSIS Previews (all life science subjects), about M. domestica rose from 0 in 1978 to 14 in 1986 (VandeBerg 1990). The number has grown steadily since then by approximately five per year: 45 were indexed in 1992 and 54 in 1993, making a total of 276 indexed publications since 1978. The possibility that this increasing rate of expansion of the scientific literature on M. domestica is simply a reflection of increasing rate of publication on all marsupial species is not substantiated by a count of the numbers of publications concerning Macropus eugenii and Sminthopsis crassicaudata, two other commonly used laboratory marsupials; for both of those species the annual number of publications has remained approximately constant with means of 25.2 and 4.2, respectively, during the 5-year period from 1989 to 1993.

Some areas of investigation have developed because of the marsupial characteristics of this new laboratory animal; for instance, birth after a short gestation enables manipulation of young at an immature stage of development (Tyndale-Biscoe and Janssens 1988; Janssens and others 1995, Saunders and Hinds, in press, chapter 4). Others arose because of the value of an out-group of mammals for comparative research strategies; for example, marsupials exhibit paternal X-chromosome inactivation rather than random X-chromosome inactivation, as is characteristic of eutherian mammals. Laboratory opossums are used as a model for comparative research on these divergent strategies to accomplish balance between sexes in expression of X-linked genes (reviewed by VandeBerg and others 1987). Still other areas have arisen simply because preliminary studies on physiological systems of M. domestica revealed research opportunities that did not exist for the conventional laboratory animal species that were available. These other research opportunities are not related to the marsupial characteristics of laboratory opossums, but arose simply as a consequence of the existence of a new laboratory animal with its own species-specific characteristics.

Ultraviolet radiation induced skin and eye cancers and dietary induced hypercholesterolemia, discussed in this review, are diseases for which M. domestica has become an important model species for reasons unrelated to its marsupial characteristics.

M. Domestica AS A MODEL FOR MELANOMA

The incidence of human melanomas is increasing at an alarming rate, indeed at a faster rate than any other type of cancer except lung cancer in women. Ultraviolet radiation (UVR) is believed to be a major causative factor. Unfortunately, research on the role of UVR in inducing melanoma in humans has been hampered by the lack of a mammalian model in which melanoma can be induced by UVR alone. UVR is not a factor in the induction or promotion of melanoma in the Sinclair swine model (Greene and others 1994) and in mice, UVR must be combined with chemical carcinogens to induce melanomas (Kripke 1979). Although a newly developed transgenic mouse appears to be a promising model for research on UVR-induced melanomas (Mintz and Silvers 1993; Klein-Szanto and others 1994), the laboratory opossum has the distinction of being the only naturally existing mammal, other than humans, known to be susceptible to melanoma in response to UVR alone.

Melanocytic Nevi

In humans, some types of melanocytic nevi or moles are a risk factor for melanoma; occasionally a cell in these nevi can undergo transformation and become a melanoma cell. Conditions for the production of melanocytic nevi by repeated shaving of adult laboratory opossums and exposure to UVR were developed by Dr. R.D. Ley and his colleagues (see Kusewitt and others 1991, and references therein). The protocol in our laboratory involves exposure to a spectral peak of 302 nm (UVB) at a dose of 125 J/m²per exposure. This dose is equivalent to exposure at 40 cm from the lamp for 1 to 1.75 min, depending on the age of the lamp. A few animals exhibited slight erythema after the initial exposures, but none exhibited severe sunburn. In our experience, the incidence of melanocytic nevi after three exposures per week was 7% after 30 weeks and 14% after 45 weeks (VandeBerg and others 1994a). When the same exposure regimen was initiated in weanlings rather than young adults, the incidence was considerably lower: only 5% after 45 weeks (Robinson and others 1995).

Six cell lines were developed from melanocytic nevi induced in the earlier experiments with adults (Dooley and others 1993). Cytogenetic analyses revealed that three of them exhibited aneuploidy involving extra copies of chromosomes 3, 5, 7, and/or 8. This result suggested that aneuploidy of these particular chromosomes might be an early event in the induction of melanocytic nevi.

In an attempt to reduce the time and effort required to produce animals with nevi as potential precursors of melanoma, we devised a technology for irradiating laboratory opossums repeatedly between birth and 19 days of age, when they begin to become furred (Robinson and others 1994). This technology depends on the fact that M. domestica are pouchless, so the neonates are fully exposed on the ventral side of their mother. The protocol involves placing the UVR source beneath a wire grid, aimed up at the litter of a mother standing on the grid. With a mean litter size of seven, many individuals can be irradiated in a relatively short time, without the need for anesthesia or shaving.

Although the optimal exposure protocol remains to be determined, the combined results of all protocols attempted have yielded an incidence of 10% of nevus-beating animals at, or up to, 1 month after weaning (8 weeks) (Robinson and others 1994). Additional exposure protocols are currently being evaluated, and experiments are in progress to determine if susceptibility to nevus formation is under genetic control.

Melanomas

In some instances, we have further irradiated animals that had been exposed as sucklings. The second irradiation treatment schedule involved shaving and irradiating the weanlings three times per week, as in the earlier experiments. Of the animals that had a nevus at weaning 40% developed malignant melanoma after up to 45 weeks of post-weaning UVR. In each case, the original nevus continued to grow, underwent changes in appearance, and metastasized to one or more lymph nodes.

Cell lines clonally derived from pigmented cells from an affected lymph node provided definitive evidence of malignancy in this model (Robinson and others 1994). Not only were they highly dendritic and pigmented in vitro (properties of melanotic metastatic melanoma), but they were moderately tumorigenic and produced pigment after injection into immunoincompetent nude mice.

M. Domestica AS A MODEL FOR ANGlOGENESIS AND CORNEAL CANCER

Exposure to environmental or artificial UVR is widely believed to be a contributing factor to loss of transparency (opacification) of the cornea and to angiogenesis (the formation of blood vessels by neovascularization) (discussed in Applegate and Ley 1991). The laboratory opossum is particularly susceptible to the development of these abnormalities after repeated exposure to low doses of UVR, and provides an excellent model for investigation of the biological mechanisms involved. In many individuals of this species, opacification and neovascularization are followed by the development of corneal sarcomas. Although UVR-induced corneal cancers are uncommon and rarely metastasize in humans (Erie and others 1986), those that develop in M. domestica provide a valuable generalized sarcoma model.

Initiation and Progression

We developed a number-coded system for clinical assessment of histopathological changes in corneal phenotype caused by UVR exposure ranging from normal, through opacification and angiogenesis, to an advanced, highly invasive stage (VandeBerg and others 1994b). The development of most types of tumors is accompanied by angiogenesis. Angiogenesis is a clearly identifiable early event in corneas of UV-irradiated M. domestica during transformation of normal stromal (fibroblastic) cells into tumor (sarcoma) cells. Blood vessels originating in the limbal region around the periphery of the cornea colonize the stroma and become most pronounced at foci of neoplastic transformation. The resulting sarcoma cells soon spread through the cornea into adjacent ocular structures and at advanced stages they occupy much of the eyeball, eventually invading the orbit and migrating along the optic nerve to the brain. Little is known about the molecular mechanisms by which UVR induces these potentially life-threatening corneal sarcomas. Mutationally activated or amplified ras oncogenes (K-ras, N-ras, and H-ras) have been found in experimentally induced animal tumors and naturally occurring human tumors, and activation of K-ras may play a role in the development of UVR-induced corneal tumors in M. domestica (Sabourin and others 1992). Moreover, it appears that the opossum K-ras gene can be mutationally activated in a manner similar to the K-ras genes of eutherian mammals (Kusewitt and others 1993).

Genetics

One hundred fifty-one individuals, consisting of 33 sibships belonging to 9 families, completed at least 30 weeks of the UVR protocol, and 137 completed 45 weeks (VandeBerg and others 1994b). Although the sample size is small for genetic analysis, particularly because only a single generation was tested, estimates of heritability (h²⁾of corneal sarcoma were high. After 30 weeks of treatment, h²was 0.74 with a 95% confidence interval of 0.30 to 1.00; after 45 weeks, h²was 1.00 with the 95% confidence interval ranging down to 0.35. It will be important to refine the estimate of heritability with a reduced confidence interval, but already it is clear that susceptibility to UVR-induced corneal sarcoma is highly heritable in this population.

M. Domestica AS A MODEL FOR DIETARY I N DUCED HYPERCHOLESTEROLEMIA

Cardiovascular disease remains the greatest single cause of health problems and mortality in industrialized societies, and characteristics of blood cholesterol are among the most important risk factors. Intensive research on lipid metabolism since the mid- 1970s has greatly enhanced our understanding of disorders of human lipid metabolism. Some of that progress is attributable to the use of animal models, including mice, rabbits, swine, and nonhuman primates. Most of the emphasis and progress in research has related to constitutive lipid disorders, such as familial hypercholesterolemia, rather than the dietary-induced lipid disorders which are substantially responsible for the vast majority of cardiovascular disease. Animal models that exhibit genetic variation in lipemic responsiveness to dietary components are needed to dissect out the genes that are believed to contribute to individual variability in this physiological characteristic.

Research to date with the laboratory opossum indicates that this species provides a unique and valuable model for this purpose. It exhibits extensive individual variation in response to dietary fat and cholesterol, and that variation appears to be largely controlled by a single recessive gene. As in humans and other animal models, hyperlipidemia can lead to atherosclerotic lesions in M. domestica (HC McGill, Jr. and JL VandeBerg, unpublished observations). To our knowledge, no other animal model exists of a dietary induced hyperlipidemia that is both as pronounced as in M. domestica and as highly regulated by a single gene. Identification of this gene and characterization of its mechanism of action may lead to a better understanding of dietary-induced changes in human lipoprotein characteristics. Indeed, the omnivorous characteristic of M. domestica may make it a better model than conventional laboratory animals for research on physiological effects of dietary fat and cholesterol.

Plasma Cholesterol Levels

The fasting plasma cholesterol level in laboratory opossums fed a standard laboratory diet (National Complete Fox Food Pellets, Reproduction Diet, Milk Specialties Co., P.O. Box 119, New Holstein, Wisconsin) is 85 + 22 mg/100 ml (VandeBerg and Cheng 1985a). For experimental dietary challenges, the diet is enriched with lard from the basal level of 8.1 to 17.7% fat (dry weight) and with cholesterol from 0.04 to 0.61% dry weight. This level of dietary fat corresponds to about 40% of calories from fat, equivalent to the mean in many industrialized societies. The level of cholesterol is perhaps sixfold higher than the mean in North America, but not above the high end of the range of human cholesterol consumption. After laboratory opossums have eaten this challenge diet for 8 weeks, the population data for plasma cholesterol levels are distinctly bimodal. In two studies, the means of the low responding group were 109 + 13 mg/100 ml and 118 + 17 mg/100 ml (VandeBerg and Cheng 1985a; Rainwater and VandeBerg 1992). The means of the high responding groups in the two studies were 304 + 213 mg/100 ml and 593 -+ 74 mg/100 ml. Many animals attain plasma cholesterol levels in the range of 1000 to 1900 mg/ 100 ml after eating this diet for 8 weeks (JL VandeBerg, unpublished observations).

For the purpose of comparison, human plasma cholesterol levels in the range of 200-239 mg/100 ml are considered to be borderline-high, and levels in excess of 240 mg/ 100 ml confer high risk of cardiovascular disease. Individuals heterozygous for familial hypercholesterolemia typically have plasma cholesterol levels of 350-400 mg/100 ml.

Distribution of Cholesterol Among Lipoprotein Classes

A large compilation of data from animal models and human subjects indicates that the amount of cholesterol carried by very low and low density lipoproteins is a causative factor in cardiovascular disease, whereas that carried by high density lipoproteins is protective (see Gotto and others 1990, and references therein). A strong indicator of risk of cardiovascular disease is the ratio of low density lipoprotein cholesterol to high density lipoprotein cholesterol.

In M. domestica, which are low responders, this ratio is relatively constant regardless of diet: 0.40 + 0.06 mg/100 ml and 0.55 + 0.11 rog/100 ml for the standard and challenge diets, respectively. However, the ratio is dramatically increased in the high responders fed the challenge diet, from 0.49 + 0.05 to 6.48 + 1.28 mg/100 ml (Rainwater and VandeBerg 1992). Thus, responsiveness is primarily a characteristic of low density ]ipoprotein cholesterol.

The characteristics of M. domestica lipoproteins under these two dietary regimens have been investigated in detail (Rainwater and VandeBerg 1992). The close resemblance to human lipoproteins in physical characteristics and specific apolipoprotein contents, indicate that results for the laboratory opossum model can be appropriately interpreted in relation to human lipoprotein metabolism.

Genetics

The initial pedigree data were limited, but were consistent with the hypothesis that plasma cholesterol levels after dietary challenge were largely controlled by a single recessive gene (VandeBerg and Cheng 1985b).

More recently, data from 327 individuals have been analyzed by complex segregation analysis (Atwood and others 1994). This analysis focused on low density lipoprotein cholesterol, and concluded that plasma concentration in animals fed the challenge diet is inherited as a major gene. The results further indicated that 68% of the total variance in the population is attributed to variation at this gene locus. Such a high contribution by a single gene to any quantitative character is unusual, and emphasizes the potential value of the M. domestica model. One or more other genes appear also to influence low density lipoprotein cholesterol concentrations because the residual heritability accounted for 17% of the total variance. Thus, the model may enable an understanding of how a major gene interacts with modifying genes and environmental (dietary) factors in controlling a quantitative phenotype of medical importance.

GENE MAPPING

Considerable effort has been committed to marsupial gene mapping, and the results have been informative in relation to understanding mammalian karyotypic evolution. Most mapping data have been derived by in situ hybridization of tritium-labeled human gene probes to chromosomes of marsupial fibroblasts, and some were obtained by somatic cell genetic techniques. All published data derived for marsupials from these techniques are from macropodids (kangaroos), dasyurids (Australian "marsupial mice") and phalangerids (Australian possums) (O'Brien and others 1993). The laboratory opossum offers a good opportunity to expand the comparative approach to understanding mammalian genome evolution because American and Australian marsupials are only distantly related, having evolved as separate lineages for an estimated 70 million years (Maxson and others 1975; Lowenstein and others 1981). In contrast, it is estimated that 120 to 156 million years have elapsed since the separation of the marsupial and eutherian lineages (Hope and others 1990; Mark and Marotte 1992). Therefore, in instances where syntenies have been disrupted between orders of eutherian mammals or between Australian marsupials and eutherian mammals, the syntenic relationships in M. domestica can be used to determine which synteny represents the ancestral condition.

From the standpoint of biomedical research, the more important aspect of gene mapping is linkage relationships (compared with syntenic relationships). For that reason, we have focused our gene mapping strategies on linkage analysis. The intent is to construct a linkage map for localizing genes that affect normal physiological characteristics, physiological responses to environmental stimuli, and susceptibility to experimentally induced diseases. If such genes can be localized in the M. domestica genome, and if the genome is sufficiently well mapped to relate segments of chromosomes to the human genome, then it may be possible to clone those genes from human libraries and to identify variation in them in relation to human physiological conditions.

Karyotype

M. domestica, like all other marsupials, has a low number of chromosomes (2n = 18 in this species), which are large and ideally suited for cytogenetic analysis (Merry and others 1983). Techniques for high resolution G-banding and R-banding have been developed, and 223 bands can be resolved per haploid set of chromosomes (Pathak and others 1993).

Protein Polymorphisms

Initial efforts to identify genetic markers focused on blood protein polymorphisms in descendants from Population I and Population 2. Fifty-four blood protein markers have been surveyed by electrophoretic- and isoelectric-focusing techniques and polymorphism was detected in 13 of them (ACP1, ACP2, ADH3, AK1, ALDH3, APRT, AT3, C6, C7, DIAl, GAPD, GPT, PI; abbreviations from Human Gene Mapping 11, 1991) (Sevilir and others 1989; van Oorschot and VandeBerg 1989; VandeBerg 1990; Holmes and others 1990, 1991, 1992; Bell and others 1992, van Oorschot and others 1992a,b, 1993). The mean proportion of heterozygous loci for these markers is approximately 5% per individual. This level of polymorphism is consistent with that in most eutherian species and slightly higher than that in macropodids (16.7% of loci heterozygous, 4% mean heterozygosity per individual; Cooper and others 1979).

Numerous additional protein markers and three more populations (3, 4, and 5) remain to be surveyed. The founders of these three populations were trapped long distances from the founders of Populations 1 and 2 (Figure 1), so we anticipate the discovery of considerably more protein polymorphisms in the laboratory population of M. domestica.

DNA Polymorphisms

Efforts to identify restriction length polymorphisms in genomic DNA of M. domestica have been initiated. Preliminary results from several human probes used in Southern blot analyses are promising (PB Samollow, personal communication), and one polymorphic system detected by a probe for the N-ras oncogene (ATCC #41030, 1.5 Kb genomic DNA fragment) has been unequivocally established as consistent with Mendelian inheritance. This probe detects two HindIII polymorphisms in M. domestica. One consists of two alleles represented by 11.5 and 8.5 Kb fragments. Animals of mixed ancestry from Populations 1 and 2 exhibited frequencies of 0.32 and 0.68 for the two alleles (van Oorschot and others 1991). The other is a two-allele system involving 7.0 and 4.5 Kb fragments (Perelygin and others 1994). All animals typed from Populations 1 and 2 were homozygous for the 4.5 Kb fragment whereas the two founders from Joaima (Population 3) were homozygous for the 7.0 Kb fragment. Although the data are too limited to be interpreted as indicative of fixed allelic differences, they do support the notion of substantial differences in gene frequencies between the populations (Perelygin and others 1994).

To increase the number of markers available for linkage analysis more quickly, and to increase the efficiency of typing animals for the purpose of linkage analysis, 10-mer primers are being used to generate random amplified polymorphic DNA markers for typing by the polymerase chain reaction. Initial results have established 9 restriction length polymorphisms that segregate as Mendelian markers, indicating that these will be a useful adjunct to protein polymorphisms and random amplified polymorphic DNA markers in developing a linkage map (PB Samollow, personal communication). Another strategy that is planned for development is the isolation of M. domestica microsatellite sequences for developing polymerase chain reaction primers specific for this species.

Recombination and Linkage

Two linkage groups have been identified in M. domestica, and both exhibit much lower recombination rates in females than in males (van Oorschot and others 1992b, 1993). AK1 and PI exhibit a recombination rate of 0.087 + 0.23 (standard error) in females and are unlinked in males, and GPT and the C6-C7 haplotype exhibit a recombination rate of 0.172 + 0.056 in females and are unlinked in males. C6 and C7 have not yielded recombinants in either sex, and also are very tightly linked in eutherian mammals (Coto and others 1991; Eldridge and others 1983; Orren and others 1985).

Because of the estimated times of divergence of marsupials and eutherian mammals, these 2 linkage groups are informative in regard to mammalian genome evolution. AK1 and PI are not syntenic in mice or humans; C6 and C7 are tightly linked in mice and humans and presumably arose by gene duplication; and GPT is syntenic with C6-C7 in mice (chromosome 15) but not in humans (C6-C7 on chromosome 5, GPT on chromosome 8) (O'Brien and others 1993). Based on these comparisons, it seems likely that the C6-C7-GPT synteny is ancestral to the split between marsupials and eutherian mammals and was broken during evolution of the human genome.

The sex difference in recombination rate is consistent with results of cytogenetic analyses, which revealed that chiasmata are clustered near the telomeres in female meiosis, but occur more uniformly along the length of the chromosomes in male meiosis (Hayman and others 1988). These recombination and cytogenetic characteristics parallel those previously observed in an Australian dasyurid, Sminthopsis crassicaudata (Bennett and others 1986).

The sex-specific difference in recombination frequencies will be valuable for detecting syntenies and doing long-range mapping via matings established with multiply heterozygous females, whereas multiply heterozygous males can be used in crosses designed for shorter range mapping.

FUTURE DIRECTIONS

Colony Maintenance and Development

We plan to maintain random breeding stocks of animals belonging to Populations 1, 2, and 5, although Population 5 will probably become inbred too quickly to enable its perpetuation without admixing with another population. In addition, random breeding stocks of Admixed Populations 2, 3, and 4 will be maintained. These populations will enable the preservation of as much genetic variability as is practical, while simultaneously providing genetically distinct stocks that may serve different purposes.

Although some of the 34 partially inbred stocks will probably be lost, we will pursue the goal of developing fully inbred strains by gradually increasing the mean inbreeding coefficients of the existing stocks.

Melanoma

The capability of screening large numbers of animals for susceptibility to UVR-induced nevi has led us to embark on a genetic analysis of this characteristic. Because the nevus phenotype is scored prior to sexual maturity in animals irradiated as sucklings, prospective matings based on phenotype can be implemented. Whether the results establish a genetic basis to susceptibility or not, the model makes feasible research on the molecular mechanisms by which UVR can cause melanocytes to become nevi and eventually metastatic melanoma. The whole animal model also may be useful for testing chemopreventative and chemotherapeutic agents for melanoma. An initial investigation of this type supported the feasibility of the model for this purpose (Maenpaa and others 1993).

The malignant melanoma cell lines are attractive as experimental model systems for melanoma research because they are derived from tumors that were initiated and promoted by UVR alone and they retain the ability to produce pigment (unlike most human metastatic melanoma cell lines currently available). These cell lines should be useful reagents for the discovery and development of anti-melanoma therapies, including conventional chemotherapeutic studies of cytotoxic and cytostatic activity in vitro. There is also the prospect of successful allogeneic grafting of metastatic cell lines in this species, for use in in vivo chemotherapy trials and for testing immunomodulatory therapies (Robinson and others 1994).

Corneal Cancer

UVR-induced sarcomas in the cornea of M. domestica provide a valuable generalized sarcoma model, particularly for studies on angiogenesis and oncogenic transformation in this translucent, accessible, and immunologically privileged site. In normal cells, angiogenesis is controlled by a balance between endogenous factors that stimulate new blood vessel growth and factors that inhibit it. In contrast, tumor cells are potently angiogenic as a result of decreased production of inhibitors and increased secretion of inducers (Dameron and others 1994). The anterior chamber of the eye of M. domestica is a favorable and readily available site for in vivo studies on the effects of reagents that are known or likely to promote or inhibit angiogenesis. For example, it will be possible to examine the effects of known inhibitory factors of angiogenesis (Fidler and Ellis 1994), including the recently identified and characterized antiangiogenesis factor, angiostatin (O'Reilly and others 1994). M. domestica corneal sarcoma cells also grow well in culture. It will therefore be possible to combine in vitro and in vivo strategies to test the efficacy of established or candidate chemotherapy and immunotherapy reagents in this new sarcoma model.

Hypercholesterolemia

In addition to providing a model for genetic analysis of dietary responsiveness, the laboratory opossum also has clear potential for research on other dietary factors that affect plasma cholesterol levels. For example, the model can be used to examine the lipemic effects of different types of dietary fat alone, in combination with one another, and in combination with dietary cholesterol. As the major gene and modifying gene(s) become better understood, and perhaps even identified, this system could serve as a model not only of biomedical importance, but also of importance to understanding complex genotype by genotype and genotype by diet interactions. In addition, the model has clear potential for testing the efficacy of cholesterol-lowering drugs.

Gene Mapping

Although few linkage groups have been identified in all marsupial species combined, M. domestica is well suited for linkage mapping because of its high reproductive capacity. Preliminary data suggest that Population 5 (from Bolivia) may be quite different from the Brazilian populations of M. domestica in gene frequencies, and that the Bolivian population may be monomorphic for some alleles not present in the Brazilian population (PB Samollow, personal communication). We plan to produce hybrids and backcross progeny from matings between Populations 1 and 5 and to use this panel of animals as the focus of future linkage mapping efforts. We anticipate that the backcross involving Populations 1 and 5 together with the use of polymerase chain reaction technologies to identify random amplified polymorphic DNA markers and microsatellite polymorphisms will rapidly increase the pace of linkage mapping in the species.

CONCLUSION

Although the laboratory opossum appeared on the research scene less than two decades ago, its value to existing research programs in comparative biology and medicine is evident. Given the wealth of baseline data that now exists for this species, its ready availability, and the substantial number of investigators using it in experimental research, the years ahead are bound to offer exciting new opportunities to exploit this prototype laboratory marsupial.

ACKNOWLEDGMENTS

We thank Dr. Paul B. Samollow for permission to cite unpublished data. Acquisition of data and preparation of the manuscript were supported in part by NIH grant POI RR09919 and a grant from the Robert J. Kleberg, Jr. and Helen C. Kleberg Foundation.

REFERENCES

Atwood LD, Kammerer CM, VandeBerg JF, Hackelman SM, VandeBerg .IL. 1994. A major gene affects V+LDL cholesterol concentrations in Monodelphis domestica. Texas Genetics Society, 21st Annual Mtg., Houston, Texas. Program and Abstracts, p 21(14).

Applegate LA, Ley RD. 1991. DNA damage is involved in the induction of opacification and neovascularization of the cornea by ultraviolet radiation. Exp Eye Res 52:493-497.

Bell K, Arthur H, van Oorschot RAH, VandeBerg .IL. 1992. Antithrombin III (AT3) polymorphism in the marsupial Monodelphis domestica: identification and genetics. Biochem Genet 30:591-601.

Bennett JH, Hayman DL, Hope RM. 1986. Novel sex differences in linkage values and meiotic chromosome behaviour in a marsupial. Nature 323: 59-60.

Cooper DW, Johnston PG, VandeBerg .IL, Maynes GM, Chew GW. 1979. A comparison of genetic variability of X linked and autosomal loci in kangaroos, man, and Drosophila. Genet Res 33:243-252.

Coto E, Martinez-Naves E, Dominguez O, DiScipio RG, Urra ,IM, Lopez-Larrea C. 1991. DNA polymorphisms and linkage relationship of the human complement component C6, C7, and C9 genes. Immunogenetics 33:184-187.

Dameron KM, Volpert OV, Tainsky MA, Bouck N. 1994. Control of angiogenesis in fibroblasts by p53 regulation of thrombospondin-1. Science 265:1582-1584.

Dooley TP, Mattern VL, Moore CM, Porter PA, Robinson ES, VandeBerg JL. 1993. Cell lines derived from untraviolet radiation-induced benign melanocytic nevi in Monodelphis domestica exhibit cytogenetic aneuploidy. Cancer Gen Cytogen 71:55-66.

Eldridge PR, Hobart M J, Lachmann PJ. 1983. The genetics of the sixth and seventh components of complement in the dog: polymorphism, linkage, locus duplication, and silent alleles. Biochem Gen 21:81-91.

Erie JC, Campbell RJ, Liesegang TJ. 1986. Conjunctival and corneal intra-epithelial and invasive neoplasia. Ophthalmology 93:176-183.

Fidler I J, Ellis LM. 1994. The implications of angiogenesis for the biology and therapy of cancer metastasis. Cell 79:185-188.

Greene JF Jr, Townsend JS, Amoss MS Jr. 1994. Histopathology of regression in Sinclair swine model of melanoma. Lab Invest 71:17-24.

Gotto AM Jr, LaRosa JC, Hunnunghake D, Grundy SM, Wilson PW, Clarkson TB, Goodman DS, Hay JW. 1990. The cholesterol facts. A summary of the evidence relating dietary fats, serum cholesterol and coronary heart disease. A joint statement by the American Heart Assocation and the National Heart, Lung and Blood Institute. Circulation 81:1721-1733.

Hayman DL, Moore HDM, Evans EP. 1988. Further evidence of novel sex differences in chiasma distribution in marsupials. Heredity 61:455-458.

Holmes RS, van Oorschot RAH, VandeBerg JL. 1990. Genetics of alcohol dehydrogenase and aldehyde dehydrogenase from Monodelphis domes-tica cornea: further evidence for identity of corneal aldehyde dehydrogenase with a major soluble protein. Genet Res 56:259-265.

Holmes RS, van Oorschot RAH, VandeBerg JL. 1991. Aldehyde dehydrogenase (ALDH) isozymes in the gray short-tailed opossum (Monodelphis domestica): tissue and subcellular distribution and biochemical genetics of ALDH3. Biochem Genet 29:163-175.

Holmes RS, vail Oorschot RAH, VandeBerg JL. 1992. Biochemical genetics of alcohol dehydrogenase isozymes in the gray short-tailed opossum (Monodelphis domestica). Biochem Genet 30:215-231.

Hope R, Cooper S, Wainwright B. 1990. Globin macromolecular sequences in marsupials and monotremes. Australian J Zool 37:289-313.

Klein-Szanto AJP, Silvers WK, Mintz B. 1994. Ultraviolet radiation-induced malignant skin melanoma in melanoma-susceptible transgenic mice. Cancer Res 54:4569-4572.

Kripke ML. 1979. Speculations on the role of ultraviolet radiation in the development of malignant melanoma. J Natl Cancer Inst 63:541-548.

Kusewitt DF, Applegate LA, Ley RD. 1991. Ultraviolet radiation-induced skin tumors in a South American opossum (Monodelphis domestica). Vet Pathol 28:55-65.

Kusewitt D F, Kelly G, Sabourin CLK, Ley RD. 1993. Characterization of the K ras gene of the marsupial Monodelphis domestica. DNA Sequence 4:37-42.

Lowenstein JM, Sarich VM, Richardson BJ. 1981. Albumin systematics of the extinct mammoth and Tasmanian wolf. Nature 291:409-411.
MacCluer JW, VandeBerg JL, Read B, Ryder OA. 1986. Pedigree analysis by computer simulation. Zoo Biol 5:147-160.

Maenpaa J, Dooley T, Wurz G, VandeBerg JL, Robinson ES, Emshoff V, Sipila P, Wiebe V, Day C, DeGregorio M. 1993. Topical toremifene: a new approach for cutaneous melanoma. Cancer Chemother Pharmacol 32:392-395.

Mark RF, Marotte LR. 1992. Australian marsupials as models for the developing mammalian visual system. Trends Neurosci 15:51-57.

Maxson LR, Sarich VM, Wilson AC. 1975. Continental drift and the use of albumin as an evolutionary clock. Nature 255:397-400.

Merry DE, Pathak S, VandeBerg JL. 1983. Differential NOR activities in somatic and germ cells of Monodelphis domestica (Marsupialia, Mammalia). Cytogenet Cell Genet 35:244-251.

Mintz B, Silvers WK. 1993. Transgenic mouse model of malignant skin melanoma. Proc Natl Acad Sci USA 90:8817-8821.

O'Brien SJ, Peters J, Searle AG, Womack JE, Johnson PA, Graves JAM. 1993. Report of the committee on comparative gene mapping. In: Cuticchia AJ, Pearson PL, editors. Human Gene Mapping 1993. A compendium. Baltimore: John Hopkins University Press. p 846-892.

O'Reilly MS, Holmgren L, Shing Y, Chen C, Rosenthal RA, Moses M, Lane WS, Cao Y, Sage EH, Folkman J. 1994. Angiostatin: a novel angiogenesis inhibitor that mediates the suppression of metastases by a Lewis lung carcinoma. Cell 79:315-328.

Orren A, Hobart M J, Nash HR, Lachmann PJ. 1985. Close linkage between mouse genes determining the two forms of complement component C6 and component C7, and Cis action of a C6 regulatory gene. Immunogenet 21:591-599.

Pathak S, Ronne M, Brown NM, Furlong CL, VandeBerg JL. 1993. A high-resolution banding pattern ideogram of Monodelphis domestica chromosomes (Marsupialia, Mammalia). Cytogenet Cell Genet 63:181-184.

Perelygin AA, Samollow PB, VandeBerg JL. 1994. A new HindIII RFLP of NRAS in the laboratory opossum, Monodelphis domestica. Animal Genet.

Rainwater DL, VandeBerg JL. 1992. Dramatic differences in lipoprotein composition among gray short-tailed opossums (Monodelphis domestica) fed a high cholesterol/saturated fat diet. Biochimica Biophysica Acta 1126:159-166.

Robinson ES, VandeBerg JL, Hubbard GB, Dooley TP. 1994. Malignant melanoma in UV-irradiated laboratory opossums: initiation in suckling young, metastasis in adults, and xenograft behavior in nude mice. Cancer Res 54:5986-5991.

Robinson ES, Hubbard GB, VandeBerg JL. 1995. Ultraviolet radiation-induced skin lesions in laboratory opossums (Monodelphis domestica) exposed from the weanling stage. Arch Dermatol Res 287:333-337.

Sabourin CLK, Freeman AG, Kusewitt DF, Ley RD. 1992. Identification of a transforming ras oncogene in an ultraviolet radiation-induced corneal tumor of Monodelphis domestica. Photochem Photobiol 55:417-424.

Saunders NR, Hinds L. In Press. Recent Advances in Marsupial Biology. Sydney: University of New South Wales Press.

Sevilir NS, Torgerson E, Guajardo AC, VandeBerg JL, Stone WH. 1989. Acid phosphatase polymorphism in a marsupial, Monodelphis domestica. Biochem Genet 27:755-759.

Tyndale-Biscoe CH, Janssens PA. 1988. The Developing Marsupial. Models for Biomedical Research. Berlin: Springer-Verlag.

VandeBerg JL. 1983. The gray short-tailed opossum: a new laboratory animal. ILAR News 26(3):9-12.

VandeBerg JL. 1990. The gray short-tailed opossum (Monodelphis domestica) as a model didelphid species for genetic research. Australian J Zool 37:235-247.

VandeBerg JL, Cbeng M-L. 1985a. A heritable diet-induced hypercholesterolemia in a laboratory marsupial. 7th International Symposium on Atherosclerosis, Melbourne, Australia, Program and Abstracts.

VandeBerg JL, Cheng M-L. 1985b. Dyslipoproteinemia in a laboratory marsupial, Monodelphis domestica. Isozyme Bull 18:66.

VandeBerg JL, Robinson ES, Samollow PB, Johnston PG. 1987. X-linked gene expression and X-chromosome inactivation: marsupials, mouse and man compared. Isozymes. Curt Top Biol Med Res 15:225-253.

VandeBerg JL, Williams-Blangero S, Hubbard GB, Ley RD, Robinson ES. 1994a. Genetic analysis of ultraviolet radiation-induced skin hyperpla-sia and neoplasia in a laboratory marsupial model (Monodelphis domestica). Arch Dermatol Res 268:12-17.

VandeBerg JL, Williams-Blangero S, Hubbard GB, Robinson ES. 1994b. Susceptibility to ultraviolet-induced corneal sarcomas is highly heritable in a laboratory opossum model. Intl J Cancer 56:119-123.

van Oorschot RAH, VandeBerg JL. 1989. Genetic variation in adenylate kinase 1, glyceraldehyde-3-phosphate dehydrogenase, and glutamic-pyruvate transaminase in the marsupial Monodelphis domestica. Biochem Genet 27:761-765.

van Oorschot RAH, Birmingham V, VandeBerg JL. 1991. HINDIII RFLP for NRAS in Monodelphis domestica. Nucleic Acids Res 19:6976.

van Oorschot RAH, Williams-Blangero S, VandeBerg JL. 1992a. Genetic diversity of laboratory gray short-tailed opossums (Monodelphis domes-tica): effect of newly introduced wild-caught animals. Lab Anim Sci 42:255-260.

van Oorschot RAH, Porter PA, Kammerer CM, VandeBerg JL. 1992b. Severely reduced recombination in females of the South American marsupial Monodelphis domestica. Cytogenet Cell Genet 60:64-67.

van Oorschot RAH, Birmingham V, Porter PA. Kammerer CM, VandeBerg JL. 1993. Linkage between C6, C7, and GPT in the marsupial Monodelphis domestica. Biochem Genet 31:215-222.

TABLE 1 Founders of the laboratory population of M. domestica

Month/ year pool	Locality of collection (state/town)¹	Latitude, longitude	Laboratory code for the population	Number sent to U.S.	Number that contributed to existing gene pool	Proportionate contribution to existing gene pool
05/78	Pernambuco/Exu²	7.28S, 39.45W	1	9	9	0.472
07/84	Paraiba/Piraua³	7.29S, 35.31W	2	2	2	0.481
12/84	Paraiba/Piraua	7.29S, 35.31W	2	11	7	0.481
10/85	Paraiba/Piraua	7.29S, 35.31W	2	8	3	0.481
9/88	Paraiba/Pirava	7.29S, 35.31W	2	3	2	0.481
10/90	Minas Gerais/Joaima	16.39S, 41.05W	3	2	2	0.013
11/92	Minas Gerais/ Conselheiro Mota⁴	18.17S, 44.00W	4	1	1	0.011
06/93	Santa Cruz/Brecha 3⁵	17.58S, 63.03W	5	2	2	0.023
			TOTAL	38	28	1.000

¹ All localities are in Brazil, except Santa Cruz, which is in Bolivia.
² Exu is 45 km southwest of the city of Crato.
³ Piraua is 25 km west of the city of Timbafiba.
⁴ Conselheiro Mota is 37 km west of the city of Diamantina.
⁵ Bracha 3 is 27 km southeast of the city of Santa Cruz.

FIGURE 1 Map of South America showing the sites at which each of the 5 sets of founders were trapped. The numbers 1 through 5 refer to the laboratory codes for the population sampled at each of the 5 sites (see Table 1). Approximate distances between adjacent sites are indicated in kilometers.