Oncogene (1998) 17, 3385 ± 3400 ã 1998 Stockton Press All rights reserved 0950 ± 9232/98 $12.00 http://www.stockton-press.co.uk/onc Mouse models in tumor suppression

Nader Ghebranious1 and Lawrence A Donehower*,1

Division of Molecular Virology and Department of Cell Biology, Baylor College of Medicine, Houston, Texas 77030, USA

Genetic lesions found in tumors are often targeted to the by loss or mutation of the second allele in sporadically negative growth regulatory tumor suppressor genes. arising tumors. Inherited cancer predisposition syn- Much of our understanding of tumor suppressor gene dromes are often characterized by the germ line function is derived from experimental manipulations in transmission of a single defective tumor suppressor cultured cells. Recently, however, the generation of mice allele in the a€ected families. Not surprisingly, the with germ line tumor suppressor gene mutations through likelihood of an a€ected individual incurring a gene targeting in embryonic stem cells has provided mutation in the remaining wild type tumor suppressor another dimension by allowing experimental studies of allele and developing a tumor is much higher than for tumor suppressor function in an organismal context. the normal individual with two intact alleles. Novel insights into the role of tumor suppressors in The use of small animals for modeling tumors in a development, di€erentiation, cell cycle control, and controlled experimental manner has been a mainstay of tumor suppression have been obtained from the studies cancer research for almost a century. However, the on these `knockout' mice. In addition, such mice may development of molecular genetic techniques to serve as disease models for humans with inherited cancer identify speci®c cancer-associated gene lesions, com- predisposition syndromes. Perhaps the greatest advan- bined with breakthroughs in the manipulation of the tage of many of the mouse tumor suppressor models is germ line of mice, have greatly facilitated the that they facilitate study of the roles of tumor suppressor generation of powerful new models for understanding gene loss in tumor initiation and progression in vivo. the mechanistic roles of speci®c genes in cancer Moreover, derivation of primary cells from tumor initiation and progression. The development of suppressor-de®cient mice has provided an important transgenic mice in the 1980s through zygote pro- resource for in vitro studies on the role of targeted nuclear injection methods has allowed the addition of genes in cell cycle regulation, DNA damage response, any cloned gene to the germ line of mice to determine regulation of apoptotic pathways, and preservation of the e€ects of the added gene on the resulting genomic stability. In this review, we discuss some of the phenotype (Palmiter and Brinster, 1986). Usually, the mechanistic insights provided by tumor suppressor- introduced exogenous gene is driven by a promoter de®cient mice and their utility as models for human which restricts expression of the transgene to a speci®c cancer syndromes. tissue compartment. This approach particularly lends itself to the analysis of the tumor promoting functions Keywords: tumor suppressors; knockout mice; cancer of dominant acting oncogenes, which can be expressed models; inherited cancer syndromes in speci®c tissues and might induce tumors within that tissue within a relatively short time frame (Hanahan, 1989). A large number of very useful mouse cancer models have been developed via this approach Introduction (Hanahan, 1989; Bedell et al., 1997). The recessive nature of tumor suppressors precluded Over the past twenty years, considerable progress has study of their in vivo e€ects through standard been achieved in understanding the molecular genetics transgenic approaches, however. Rather than constitu- of cancer formation. Genetic alterations in oncogenes tively expressing an activated oncogene, it would be and tumor suppressor genes occur consistently in many necessary to inactivate one or both endogenous tumor tumor types and experiments have shown that such suppressor alleles in order to study their role in cancer mutations can have a potent growth promoting e€ect. promotion. This was a more challenging task, but the Mutations in the oncogenes may constitutively activate development of mouse embryonic stem cell gene cellular proto-oncogenes so that they are refractory to targeting techniques has made such an approach regulation. The activated oncogene generally behaves feasible. Inactivation of endogenous tumor suppressor in a dominant manner in that its growth promoting genes in the germ line of mice has been achieved for e€ects can be maintained in the presence of a normal virtually every known tumor suppressor gene originally proto-oncogene allele. In contrast, tumor suppressor identi®ed in humans (Jacks, 1996; Bedell et al., 1997; mutations in tumors are recessive in nature. Inactiva- Venkatachalam and Donehower, 1998). tion or mutation of one allele is usually accompanied The advantages of such tumor suppressor `knock- out' mice have been multiple. First, the heterozygous version of a particular tumor suppressor knockout mouse may provide a useful animal model for the corresponding human cancer predisposition syndrome. While there have been a number of successes in this *Correspondence: LA Donehower, Division of Molecular Virology, One Baylor Plaza, Baylor College of Medicine, Houston, TX 77030, regard, the inherent di€erences between mouse and USA human have resulted in some incongruities. For Mouse tumor suppressor models N Ghebranious and LA Donehower 3386 example, mice heterozygous for the Rb gene develop genes) are thought to increase genetic instability in a pituitary adenomas rather than the retinoblastomas manner which indirectly promotes cell growth through typically observed in children with a defective germ line an increased mutation rate. Our focus will be on the Rb allele (Jacks et al., 1992; Hu et al., 1994). gatekeeper genes associated with well characterized Nevertheless, the Rb heterozygous mice do develop inherited cancer predisposition syndromes in humans. early tumors and thus have provided a number of We will describe the tumor phenotypes observed in the important insights into the role of Rb in tumorigenesis. tumor suppressor-de®cient mice and we will assess the A second advantage provided by the homozygous usefulness of some of these tumor suppressor knockout knockout mice (and not obtainable by cell culture or mice as models for the corresponding human cancer human tumor studies) has been the important insights syndrome. It is not our intention to provide an into the role of tumor suppressors in embryonic exhaustive review, but rather to highlight some of the development. The well characterized embryology of the contributions that such mice have made to our basic mouse has facilitated these studies. With a few notable understanding of tumor suppressor function. exceptions, mice null for a given tumor suppressor exhibit embryonic lethality and a distinct pattern of organ malformations. Analysis of the biological and p53-de®cient mice (Li-Fraumeni syndrome) molecular bases of the observed defects has facilitated our understanding of the role of particular tumor Discovered in 1979 as a cellular protein complexed suppressor genes in organogenesis, regulation of cell with a DNA tumor virus oncoprotein (Lane and cycle control, di€erentiation, and apoptosis. Crawford, 1979; Linzer and Levine, 1979), early The accumulation of numerous tumor-susceptible studies suggested that the p53 gene might encode a transgenic and knockout mice has aided investigators proto-oncogene, as mutant versions of p53 had in analysing genetic interactions among the oncogenes oncogenic activity in transformation assays (Jenkins and tumor suppressor genes. Intercrossing of two et al., 1984; Eliyahu et al., 1984; Parada et al., 1984). di€erent tumor suppressor knockout mice to generate However, by 1989, several groups had demonstrated doubly de®cient o€spring may reveal di€erent levels of that the wild type p53 gene encodes a tumor suppressor cooperativity between the two genes. In some cases protein (Finlay et al., 1989; Eliyahu et al., 1989; Baker embryonic lethality in doubly de®cient mice can be et al., 1990; Chen et al., 1990; Mercer et al., 1990). accelerated or delayed. In other situations, tumor Transfection of p53 into cultured cells inhibited incidence rates can be accelerated and tumor spectra transformation mediated by a variety of oncogenes, altered by the combined tumor suppressor de®ciencies. while Vogelstein and others had shown that a number Important and sometimes surprising new insights can of human tumors displayed mutation or loss of both be obtained from such combinations. Crossing the p53 alleles (Masuda et al., 1987; Nigro et al., 1989; tumor suppressor mutations into inbred strains of Baker et al., 1989). The subsequent demonstration that di€erent backgrounds has also revealed strain-speci®c a high fraction of individuals a€ected by a familial di€erences in tumor development and spectra, allowing cancer predisposition called Li-Fraumeni syndrome for the identi®cation of novel tumor modi®er genes. had germ line p53 mutations solidi®ed the status of The identi®cation of such modi®ers (as well as other p53 as a tumor suppressor (Malkin et al., 1990; cooperating oncogenes and tumor suppressor genes) Srivastava et al., 1990). Since these initial reports, will be greatly assisted by the availability of thousands thousands of papers have been published demonstrat- of strain-speci®c microsatellite markers spanning the ing that the p53 gene is the most frequently mutated mouse genome (Dietrich et al., 1995). gene in a wide variety of human cancers (for reviews, Another signi®cant bene®t is that primary cells see Gottlieb and Oren, 1996; Ko and Prives, 1996; lacking a speci®c tumor suppressor gene can often be Levine, 1997; Moll and Schramm, 1998). derived from the knockout mice embryos (either ES p53 may suppress tumorigenesis through multiple cells or embryonic ®broblasts). Biological or biochem- mechanisms. It has been shown that various cellular ical comparison of various attributes of the null cells stresses, including DNA damage, hypoxia, depletion of and wild type control cells in culture can impart ribonucleotide pools, serum starvation, and aberrantly important insights into the function of a given tumor activated oncogenes, may induce stabilization of p53 suppressor in preventing transformation and in through phosphorylation of speci®c sites on p53, regulating cell cycle control, growth signal transduc- preventing it from binding to the mdm2 protein which tion, di€erentiation, apoptosis, genomic stability, and facilitates its degradation (Prives, 1998). Stabilized p53 response to DNA damage. can induce either G1 cell cycle arrest or apoptosis The goal of this review is to outline some of the (Kastan et al., 1992; Yonish-Rouach et al., 1991; Lane, important new insights provided by experiments 1992). The decision by p53 about whether to intiate an directly utilizing the tumor suppressor-de®cient mice arrest program or an apoptotic program can be and cells derived from them. It is not our intention to in¯uenced by a number of variables, including cell describe every knockout mouse which exhibits an type, cytokine or growth factor concentrations, levels enhanced tumor predisposition. Rather, we will take of p53, or amount of damage (Ko and Prives, 1996). a more focused approach employing a useful classifica- Whether arrest or apoptosis is induced, the end result tion scheme recently proposed by Kinzler and is that the cell with DNA damage is prevented from Vogelstein (1998). They have subdivided the tumor replicating damaged DNA templates and incurring suppressors into two groups, the `gatekeepers' and the potentially oncogenic DNA lesions. `caretakers'. The gatekeepers directly regulate tumor The cell cycle arrest function of p53 is mediated cell growth (usually through cell cycle control largely by its ability to transcriptionally regulate an mechanisms), while the caretakers (often DNA repair extensive array of growth regulatory genes by both Mouse tumor suppressor models N Ghebranious and LA Donehower 3387 positive and negative mechanisms (Ko and Prives, 1992; Purdie et al., 1994; Jacks et al., 1994). The p53 1996; Levine, 1997). Positive transcriptional regulation heterozygous (p53+/7) mice are also susceptible to by p53 is through speci®c DNA binding to sites tumors, but these neoplasms arise later than those upstream or within the target genes. Important targets observed in the p537/7 mice. Tumors are rarely of p53 upregulation include p21WAF1/CIP1, a G1 cyclin- observed until nine months of age, but by 18 months, dependent kinase inhibitor, mdm2, a negative regulator roughly 50% have developed cancers, and by 24 of p53 function, and IGF-BP3, an anti-mitogenic months, over 95% of the animals have had tumors factor (el-Deiry et al., 1993; Barak et al., 1993; or died (Venkatachalam et al., 1998). This contrasts to Buckbinder et al., 1995). Important transcriptional a 20% death or tumor incidence in wild type animals targets implicated in apoptosis include bax, KILLER/ of the same strain (primarily C57BL/6) by 24 months. DR5, and FAS/Apo1 (Miyashita et al., 1995; Wu et The tumor spectrum in p53+/7 animals di€ers from al., 1997; Owen-Schaub et al., 1995). As might be that of the p537/7 mice, with osteosarcomas and soft expected, those genes repressed by p53 include positive tissue sarcomas being more prevalent than lymphomas growth regulators, including PCNA and bcl-2 (Deb et (Harvey et al., 1993; Jacks et al., 1994). Moreover, al., 1992; Miyashita et al., 1994). carcinomas occur more frequently in the p53+/7 In addition to its role in mediating G1 checkpoint animals. control, p53 has also been implicated in G2/M One issue with respect to the p53+/7 mice is checkpoints (Levine, 1997). Mouse ®broblasts and whether their observed tumor incidence and spectrum spleen cells de®cient in p53 undergo multiple rounds accurately models the incidence and spectrum of of DNA synthesis without completing chromosome tumors observed in Li-Fraumeni syndrome lineages. segregation, forming tetraploid and octaploid cells The correlation between mouse model and human (Cross et al., 1995). p53 seems to be an integral part syndrome in this case is reasonably close. On a life of the process that regulates the number of centro- span basis, a 50% tumor incidence in mice by 18 somes in a cell. Mouse ®broblasts from p53 null mice months of age (a mouse life span is 30 ± 36 months) is produce abnormal numbers of centrosomes after a few not very dissimilar from a 50% tumor incidence in doublings in cell culture and initiate spindles with three a€ected Li-Fraumeni patients by 30 years of age or four poles (Fukasawa et al., 1996). Thus, (Malkin et al., 1990). Moreover, the tumor types preservation of global genomic stability has been observed in mice, soft tissue sarcomas, osteosarcomas, shown to be an important function of p53, as multiple and lymphomas, are also signature tumors often seen assays have shown that cells missing it display more in Li-Fraumeni families (Malkin, 1994). However, the frequent genome aberrations of various types (Bischo€ spectrum concordance is not complete, as the frequent et al., 1992 Yin et al., 1992; Livingstone et al., 1992). breast and brain tumors seen in Li-Fraumeni patients p53 may also control a third cell cycle checkpoint are infrequently observed in p53+/7 mice. (G0/G1) mediated by the growth arrest speci®c (Gas-1) A central feature of a tumor suppressor is its protein which encodes a membrane protein that is recessiveness. Thus, inactivation of both copies of a highly expressed in growth restricted conditions (G0). tumor suppressor is considered to be a prerequisite for Gas-1 can restrict ®broblasts to G0 only when wild cancer formation, as originally postulated in Knud- type p53 is present (Del Sal et al., 1995). Consistent son's `two hit' model for tumor suppressors (Knudson, with this observation, is that hepatocytes null for p53 1971, 1985). Indeed, this rule is observed for p53 in also show an increase in G1 cells compared to that most sporadically arising tumors, where point mutation with wild type (Yin et al., 1998). This data suggests in one allele is often accompanied by the loss of the that p53 also function to maintain cells in G0, either by second allele (Baker et al., 1990). Surprisingly, facilitating their exiting from M phase or by preventing however, when the fate of the remaining wild type G0/G1 transition. p53 allele in tumors arising in p53+/7 mice was Given the importance of the p53 gene in cell cycle investigated, we found that this allele is retained control and cancer prevention, development of a structurally intact in about 50% of the tumors as mouse model was an obvious priority. Several groups measured by Southern blot hybridization analysis and have been successful in developing p53-de®cient mice DNA sequencing (Venkatachalam et al., 1998). Tests using gene targeting in embryonic stem cells of p53 function in these tumors with retained p53 also (Donehower et al., 1992; Gondo et al., 1994; Jacks et indicated that the wild type p53 retained normal al., 1994; Purdie et al., 1994; Tsukada et al., 1993). The transcriptional regulation and apoptosis induction majority of p53 null mice develop normally. However, activities. Interestingly, an extensive survey of Li- a fraction of the female null embryos display defects in Fraumeni tumors has revealed that roughly half of neural tube closure resulting in an overgrowth of these appear to retain an intact wild type p53 allele neural tissue in the region of the mid-brain, a condition (Varley et al., 1997). Thus, p53 may be an exception to known as exencephaly (Armstrong et al., 1995; Sah et the two hit model in that the mere reduction of p53 al., 1995). dosage level may be sucient to promote cancer Not surprisingly, p53 null mice develop tumors at a formation (Venkatachalam et al., 1998). very young age. By 6 months of age, 75% of p537/7 Another interesting issue with respect to such mouse mice have developed tumors of various types and all models is the issue of strain-dependent di€erences in have succumbed to tumors by 10 months of age tumorigenesis. The p537/7 and p53+/7 mice of (Donehower et al., 1992). Thymic T-cell lymphomas 129/Sv background develop tumors sooner on average are the most frequently arising tumor type in the null than do their counterparts of predominantly C57BL/6 mice, but lymphomas of B-cell origin, soft tissue background (Donehower et al., 1995a). Moreover, in sarcomas, osteosarcomas, testicular teratomas, and the C57BL/6 p537/7 mice, the predominant tumor other types are also observed (Donehower et al., type is lymphoma, whereas in the null 129/Sv strain, Mouse tumor suppressor models N Ghebranious and LA Donehower 3388 testicular teratomas are as common as lymphomas. not a€ected by p53 status, even in the early The dramatic susceptibility of 129/Sv null mice to preneoplastic stages. In contrast, cell proliferation testicular tumors may be attributed to the fact that rates as measured by percentages of tumor cells in S wild type mice of this strain have a modest phase and mitosis were signi®cantly increased in the susceptibility to this tumor (Stevens and Little, 1954). p537/7 Wnt-1 transgenic mammary tumors com- Thus, absence of p53 may exacerbate a previous tumor pared to their p53+/+ counterparts. Similar results predisposition. have been obtained by Hundley et al. (1997) for an Despite its usefulness, the p53-de®cient mouse as a MMTV-c-ras/p53 mammary tumor model. model to obtain mechanistic insights into the role of Another contributing factor to the rapid appearance p53 loss in tumorigenesis has certain limitations. and growth of tumors missing p53 could be through Among them are the diverse array of tumors observed increased levels of genomic instability in these tumors. and the paucity of control tumors which arise in the Cell culture experiments using both human and murine p53+/+ littermates in order to form a reliable cells have shown that the absence of p53 consistently comparison of p53-dependent and p53-independent confers increased aneuploidy, increased rates of drug tumorigenesis pathways. To circumvent these inherent induced gene ampli®cation, and abnormal centrosome limitations, a number of investigators have crossed the duplication (Yin et al., 1992, Livingstone et al., 1992, p53-de®cient mice to transgenic mice that are Fukasawa et al., 1996). In the p53-de®cient Wnt-1 susceptible to a single tumor type. The o€spring of transgenic model cited above, we found that that the such crosses often develop tumors in a single tissue in a absence of p53 in a mammary tumor conferred relatively short amount of time. Moreover, the same signi®cantly higher rates of aneuploidy and chromoso- tumor types with normal (p53+/+) and reduced mal copy number aberrations than the mammary (p53+/7) complements of p53 can be compared to tumors which retained wild type p53 (Donehower et tumors missing p53 (p537/7) for an array of al., 1995b). This correlation of p53 loss with biological, biochemical, and genetic properties. chromosomal instability has also been con®rmed in Since p53 in cell culture studies a€ects apoptosis, cell sarcomas and lymphomas from the p53-de®cient mice proliferation, and genomic stability, it is obvious that (Venkatachalam et al., 1998). such processes could also be a€ected by p53 in a The various p53-associated tumor models described tumorigenic context. The bitransgenic models cited above are not in complete agreement. It appears in above have been particularly useful for assessing how some models that p53 loss results in abrogated tumor the loss of p53 a€ects tumor initiation and progression. cell apoptosis and that this may be a key rate limiting Early experiments with the p53-de®cient mice have step in tumor progession. In other models, tumor cell shown that normal tissues in the p53 null mice are apoptosis rates are low and relatively unin¯uenced by particularly refractory to radiation-induced apoptosis p53 status, while cell proliferation rates are dramati- (Lowe et al., 1993; Clarke et al., 1993). With respect to cally a€ected by p53 loss. In each model, the results tumors, several studies have shown that absence or loss which are observed are probably a€ected by the of p53 in a tumor model is accompanied by reduced cooperating oncogene/tumor suppressor or the tissue levels of tumor cell apoptosis (Symonds et al., 1994; type expressing it. The Wnt-1 and MMTV-c-ras Howes et al., 1994; Pan and Griep, 1994). Symonds et transgenes may abrogate apoptosis independently of al. (1994) have demonstrated that a transgenic mouse p53 while inactivated Rb clearly does not. The Wnt-1 expressing high levels of a SV40 large T antigen and MMTV c-ras models were thus useful in fragment that inactivates Rb in the choroid plexus eliminating apoptosis as a tumor variable and develop slow growing choroid plexus with high allowing the more direct comparison of cell prolifera- numbers of apoptotic cells. When these mice were tion rates in the presence and absence of p53. Thus, it made null for p53 through crossing to p53-de®cient is our hypothesis that p53 loss contributes to tumor mice, the choroid plexus tumors developed sooner and progression through a variety of mechanisms including grew faster, largely due to greatly reduced tumor cell the ones described here and those that are yet to be apoptosis. Similar e€ects of p53 loss on tumor discovered. It is expected that the current p53-de®cient progression have been observed in other mouse mice and new mutant versions (e.g. p53 alleles with models with Rb de®ciency (Howes et al., 1994; Pan point mutations or tissue-speci®c p53 deletions) will and Griep, 1994). play an important role in elucidating some of these Given the involvement of p53 in multiple phases of novel tumorigenic mechanisms. cell cycle control, it could be expected that tumors missing p53 would show higher rates of tumor cell proliferation than those tumors with intact p53. In RB-de®cient mice (inherited childhood retinoblastoma) fact, at least two tumor models utilizing p53 knockout mice have shown this to be the case (Hundley et al., Childhood retinoblastoma has both an inherited and 1997; Jones et al., 1997). In one of these models, we sporadic component. About 60% of cases are crossed mammary tumor susceptible Wnt-1 transgenic nonhereditary and 40% of cases are inherited (News- mice to p53-de®cient mice (Jones et al., 1997). Not ham et al., 1998). Positional cloning was used to clone unexpectedly, mammary tumors from the p537/7 the gene based on linkage analysis of familial Wnt-1 transgenic female mice arose sooner and grew retinoblastoma cases. The retinoblastoma susceptibil- much faster than those from p53+/+ Wnt-1 trans- ity gene, RB, was cloned and characterized by a genic females (Donehower et al., 1995b; Jones et al., number of groups (Friend et al., 1986; Fung et al., 1997). However, when tumors were assessed for 1987; Lee et al., 1987). Both copies of the RB gene are apoptosis and cell proliferation rates, apoptosis rates inactivated by somatic mutations during the develop- in the mammary tumors were found to be modest and ment of sporadic retinoblastomas, a malignant tumor Mouse tumor suppressor models N Ghebranious and LA Donehower 3389 arising from the nuclear layer of the retina (Bookstein The absence of retinoblastoma in these mice has et al., 1988; Horowitz et al., 1989; Dunn et al., 1989). been attributed to species di€erences, such as differ- This gene is also mutated in many other sporadically ences in life span and number of cells in the target arising tumor types, including osteosarcomas, soft tissue. Interestingly, retinoblastoma has been observed tissue sarcomas and carcinomas of the breast, lung in mice expressing SV40 T-antigen (which inactivates and bladder (Newsham et al., 1998). Individuals who Rb) in the retina (Windle et al., 1990). These results inherit a nonfunctional copy of RB have an estimated suggest that the development of retinoblastoma in mice 90% probability of developing retinoblastoma at an requires inactivation of additional genes, possibly early age (Knudson et al., 1991). The RB gene product including p53 (Windle et al., 1990). Indeed, p53 loss is a nuclear phosphoprotein, which shares homology may be involved in retinoblastoma in mice, as with two additional family members, p107 and p130, expression of human papilloma virus E7 oncoprotein neither of which seem to play a role in tumorigenesis in the retinal cells, which binds and inactivates Rb (Weinberg, 1995). protein only, results in retinal degeneration with high The RB protein is an important regulator of the G1 level of apoptosis, with the cells restricted from checkpoint. It binds to members of the E2F family, undergoing terminal di€erentiation. However, expres- transcription factors needed for transcription of S sion of E7 in a null p53 background leads to phase early genes, and prevents S phase entry retinoblastoma (Howes et al., 1994). (Weinberg, 1995). Hyper-phosphorylation of RB by Is there a synergistic interaction between p53 and G1/S cyclin-dependent kinases releases E2F from RB- Rb in other contexts of tumorigenesis? Doubly mediated repression and enables progression of the cell heterozygous p53 and Rb mice exhibit shorter life into S phase. span and develop endocrine tumor not seen in the Rb-de®cient mice have been generated by at least parental animals (Williams et al., 1994; Harvey et al., three groups (Clarke et al., 1992; Lee et al., 1992; Jacks 1995). These in vivo demonstrations of Rb and p53 et al., 1992). Mice that are homozygous for Rb cooperativity con®rm the multilevel interactions of de®ciency die at about day 14.5 of gestation. The these two proteins observed in molecular studies. It hematopoietic system is abnormal in these mice, has recently been demonstrated that E2F-1 upregula- exhibiting a signi®cant increase in the number of tion (occurring during Rb de®ciency) stimulates immature nucleated erythrocytes. In addition, neuro- p19ARF expression which in turn results in inhibition genesis is defective as cells in the central and peripheral of mdm2 and stabilization of p53 (Bates et al., 1998). nervous system undergo extensive apoptosis and Moreover, a recent report revealed an apparent cross aberrant S phase entry. Interestingly, these apoptotic talk between p53 and RB during induction of and proliferative e€ects were largely abrogated by apoptosis by p53 (Gottlieb and Oren, 1998). Deple- introducting E2F-1 nullizygosity into the Rb7/7 tion of IL-3 from the medium of some tumor cell lines embryos (Tsai et al., 1998). In addition, the embryos results in caspase-mediated RB cleavage, occurring had an extended viability, exhibiting embryonic preferentially within cells which express functional lethality at 17.5 days gestation, rather than 14.5 days p53. Thus, p53 may modulate apoptosis both by for the Rb7/7 embryos. These results indicate that diminishing the activity of Rb (which inhibits the apoptosis induction occurring in Rb null cells is apoptosis), and by positive regulation of the mediated largely by E2F-1 overexpression. Recent apoptotic machinery itself through transcription- studies in an Rb-de®cient tumor model by Pan et al. dependent and independent mechanisms. (1998) have con®rmed that the absence of E2F-1 reduces apoptosis in a p53-dependent manner. Given the signi®cant level of embryonic develop- p16 and p19ARF-de®cient mice (dysplastic nevus ment and organogenesis observed in the Rb-de®cient syndrome) embryos, despite the critical importance of Rb in mediating the G1/S transition, it could be argued that The incidence of melanoma has been increasing rapidly other Rb-related molecules might functionally com- and 37 000 cases per year have recently been reported pensate for Rb loss. In fact, two Rb family members in the US (Kamb and Herlin, 1998). Melanomas arise which exhibit structural and biochemical similarities to somatically from melanocyte precursor cells, but an Rb, p107 and p130, have been inactivated in the germ inherited susceptibility to melanoma, called dysplastic line of mice (Cobrinik et al., 1996; Lee et al., 1996). nevus syndrome, has also been observed. The a€ected p107 and p130 null mice develop normally, but individuals in the melanoma-prone families have an doubly de®cient p1077/7; p1307/7 mice show estimated risk of developing a malignant melanoma of neonatal lethality and bide®cient p1077/7; Rb7/7 53% by age 80 (Kamb and Herlin, 1998). Genetic mice show earlier embryonic lethality (11.5 days) than analysis of somatic and inherited melanomas showed monode®cient Rb7/7 mice (Lin et al., 1996). Thus, frequent loss of heterozygosity at chromosome 9p21 it appears likely that the Rb family members do have (Fountain et al., 1992). Studies of melanoma cell lines some overlapping functions as well as distinct localized the deletions to a site containing two genes functions. encoding the cyclin-dependent kinase inhibitors p16 Mice that are heterozygous for the Rb gene show no and p15 (Kamb et al., 1994; Nobori et al., 1994; Jen et signs of retinoblastoma or precursor lesions. However, al., 1994). Fine mapping studies showed that inactivat- virtually all of the Rb heterozygotes develop pituitary ing mutations were usually in the p16 locus, identifying adenomas (Jacks et al., 1992; Hu et al., 1994). These it as a tumor suppressor and responsible for the tumors consistently show loss of heterozygosity of the majority of the melanoma-susceptible kindreds (Kamb wild-type Rb allele, consistent with Knudson's `two hit' et al., 1994; Liu et al., 1995; Stone et al., 1995). The hypothesis (Hu et al., 1994). signi®cance of the p16 locus in tumor suppression was Mouse tumor suppressor models N Ghebranious and LA Donehower 3390 underscored by the observation that it was homo- gene but not p16. This was accomplished by deleting zygously deleted at high frequency in cell lines derived exon E1b, while leaving all of the p16INK4a coding from tumors of lung, breast, brain, bone, skin, bladder, sequence intact. Surprisingly, p19ARF-de®cient mice kidney, ovary and lymphocyte (Kamb and Herlin, develop tumors early in life in a manner similar to 1998). In addition, point mutations and small deletions INK4a-de®cient mice. One third of the p19ARF7/7 are common in pancreatic adenocarcinomas, esopha- null mice exhibited malignant tumors by 6 months of geal carcinomas, biliary tract cancers, and in families age (mostly, ®brosarcomas and lymphomas), but no with hereditary susceptibility to melanoma and tumors were observed in the p19ARF+/+ or pancreatic cancer (Goldstein et al., 1995; Hussussian p19ARF+/7 mice. Following dimethylbenzanthra- et al., 1994). cene (DMBA) treatment, skin tumors develop by The p16 (or INK4a) locus encodes two di€erent 13 ± 15 weeks of age in p19ARF null mice but not in transcripts each initiated by a distinct promoter (Duro control groups (Kamijo et al., 1997). et al., 1995; Stone et al., 1995; Mao et al., 1995). Each Interestingly, MEFs derived form p19ARF7/7 transcript is composed of two common exons, E2 and embryos, unlike p19ARF+/7 and p19ARF+/+ E3, and two di€erentially expressed ®rst exons, E1a or MEFs, proliferate faster and do not undergo any E1b.E1a is the exon encoded in the p16 transcript, detectable senescence even up to ten passages. whereas in p19ARF (alternative reading frame), it is However, they promptly stop proliferating when E1b (Quelle et al., 1995). Initiation from E1b results in infected with a retrovirus encoding p19ARF. It is a transcript of a di€erent reading frame. Thus, noteworthy, that p537/7 MEFs also do not undergo p19ARF protein shares no homology with p16. p16 any senescence (Harvey et al., 1993). Thus, the e€ect was ®rst found to be associated with CDK4 in human observed of p19ARF on cell growth could be mediated cells (Xiong et al., 1993), and was cloned thereafter and by p53 either directly or through p21WAF1/CIP1. Con- characterized as a CDK4/6 inhibitor (Serrano et al., sistent with this hypothesis is the observation that 1993). basal p21WAF1/CIP1 levels are reduced in p19ARF7/7 The p16 protein induces G1 arrest through its direct MEFs compared to p19ARF+/+ MEFs (Kamijo et binding to CDK4/6 molecules. This growth suppres- al., 1997), which is further supported by the recent sion is mediated by RB, since overexpression of p16 ®nding that p19ARF physically interacts with mdm2 results in G1 arrest only in cells with functional RB and prevents it from complexing with p53 and (Guan et al., 1994; Koh et al., 1995; Lukas et al., 1995; inactivating it (Pomerantz et al., 1998; Kamijo et al., Medema et al., 1995; Serrano et al., 1995). 1998; Stott et al., 1998; Zhang et al., 1998). p19ARF, the alternative product of the p16INK4a The similarity between the INK4a and the p19ARF locus, has the ability to arrest cell proliferation both at mouse models raises an important question about the G1 and G2 phases of the cell cycle. However, the relative contributions of p16 and p19ARF in tumor- mechanism for such suppression does not involve direct igenesis. In human tumors, mutations a€ecting both inhibition of any of the known CDKs (Quelle et al., alleles are often observed, as well as mutations in only 1995). More recently, it was shown that p19ARF the p16 allele. However, p14ARF (the human version potently suppresses oncogenic transformation in of mouse p19ARF) mutations are infrequently ob- primary cells and that this function is abrogated when served in human tumors (Haber, 1997). Thus, the p53 is neutralized by viral oncoprotein and dominant- generation and characterization of a mouse with a p16- negative mutants, but not by the p53 antagonist mdm2 speci®c mutation may be important in further (Pomerantz et al., 1998; Kamijo et al., 1998; Stott et clarifying some of these issues. al., 1998; Zhang et al., 1998). p19ARF appears to block the mdm2-induced degradation of p53 through physical interaction with Mdm2 (Prives, 1998). Apc-de®cient mice (familial adenomatous polyposis) INK4a-de®cient mice have been generated using gene targeting techniques in ES cells. The targeting Colon cancer is one of the most frequently arising resulted in the deletion of exons 2 and 3 of the INK4a malignancies in the western world, with an estimated locus, thus disrupting both p16 as well as p19ARF 60 000 deaths per year in the United States alone protein function (Serrano et al., 1996). The homo- (Kinzler and Vogelstein, 1998). In one form of colon zygous null INK4a7/7 mice develop normally and cancer, familial adenomatous polyposis (FAP), the are fertile. Spontaneous tumors, mostly sarcomas and individual inherits a single mutant allele of the lymphomas, develop by approximately 30 weeks of age adenomatous polyposis coli gene (APC) and develops in about 68% of the null mice. The INK4a7/7 mice hundreds of polyps in the colon (Bulow, 1987, Groden display features consistent with abnormal extramedul- et al., 1991; Joslyn et al., 1991; Powell et al., 1992). lary hematopoisis, suggesting that p16 or p19ARF may Some of these polyps may eventually progress to a normally regulate the proliferation of some hemato- malignant carcinoma form (Bulow, 1987). APC poietic lineages or their progenitors. Unlike humans, mutations also frequently occur in sporadically arising none of the null mice developed skin melanoma, which colon carcinomas (Kinzler and Vogelstein, 1998). could be due to the fact that mice are relatively Mutation and loss of both APC alleles are often a resistant to the development of cutaneous melanomas. relatively early event in colon polyps (Kinzler and Analysis of mouse embryonic ®broblasts (MEFs) Vogelstein, 1998). obtained from INK4a7/7 mice established that these The human APC gene was identi®ed by positional cells have a high proliferation index as well as colony cloning (Groden et al., 1991; Kinzler et al., 1991; formation eciency (Serrano et al., 1996). Nishisho et al., 1991). The gene encodes a large 2844 More recently, Kamijo et al. (1997) targeted the amino acid multi-domain polypeptide, which is highly INK4a locus in a di€erent way, disrupting the p19ARF conserved between human and mouse (Groden et al., Mouse tumor suppressor models N Ghebranious and LA Donehower 3391 1991; Kinzler et al., 1991; Su et al., 1992). The APC an adenovirus encoding the Cre recombinase, the mice protein was shown to bind to beta-catenin, micro- developed adenomas within 4 weeks. The adenomas tubules and an uncharacterized protein named EB1 showed deletion of Apc exon 14, indicating that the loss (Joslyn et al., 1993; Munemitsu et al., 1994, 1995; of Apc function was caused by Cre-loxP-mediated Rubin®eld et al., 1995, 1996; Su et al., 1993). recombination (Shibata et al., 1997). Accordingly, the APC protein may be involved in Intercrossing of C57BL/6 Min mice with other cellular adhesion, migration, signaling and prolifera- strains has revealed modi®er loci which speci®cally tion (Hoscheutzky et al., 1994; Nathke et al., 1996; a€ected the polyp numbers and Min-induced morbidity Rubinfeld et al., 1996; Wong et al., 1996). (Bilger et al., 1996). Sensitive strains for Mom1 (e.g. The ®rst Apc-de®cient mice were not generated by C57BL/6) developed higher numbers of polyps than gene targeting methods, but by a mouse germ line were observed for resistant strains (e.g. AKR). One of random mutagenesis protocol emplying ethylnitrosur- the major tumor modi®ers identi®ed was the Mom1 ea, a point mutagen. One line of mutagenized mice, (Modi®er of Min 1) locus. Mapping studies showed called Min (Multiple intestinal neoplasms) displayed an that Mom1 segregated with markers on the distal end anemic pheotype, which was traced to bleeding from of mouse chromosome 4 (Dietrich et al., 1993). Recent multiple polyps in the intestine (Moser et al., 1990). positional cloning experiments have shown that a The Min phenotype resembled that of FAP. This secretory phospholipase 2 group 2a (Pla2g2a) is the resemblance led to the characterization of the Min Mom1 gene. Pla2g2a function in cleavage of fatty acids mutation as a nonsense mutation in the mouse Apc from lipids and is expressed in the small intestinal homologue. (Su et al., 1992). The heterozygous mice crypts (MacPhee et al., 1995). Con®rmation of the rarely live beyond 5 months of age, due to secondary Pla2g2a as the Mom1 gene was provided by the e€ects to the tumors including anemia and intestinal demonstration that all sensitive strains had a blockage (Moser et al., 1990). The intestinal tumors premature stop codon in the Pla2g2a gene while the invariably show loss of the remaining wild type allele gene was intact in resistant strains. Moreover, Cormier (Luongo et al., 1994; Levy et al., 1994). In addition, et al. (1997) crossed a Pla2g2a transgene into a the Apc protein has been shown to be important in sensitive strain of Min mice and demonstrated reduced development as homozygosity of the Min mutation intestinal polyp formation typical of a resistant strain. leads in early embryonic lethality (Moser et al., 1993). Recent analyses indicate that there are other modi®ers The phenotype of these mice di€ers from the human of the Min phenotype in addition to Mom1 (Bilger et FAP phenotype in a few aspects, among which are the al., 1996; Shoemaker et al., 1998). origin of the tumor. In the Min mice, tumors appear in the small intestine, whereas in human FAP patients, the majority of tumors develop in the large intestine Nf1 and Nf2-de®cient mice (neuro®bromatosis type 1 (Bilger et al., 1996). In addition, about 10% of female and type 2) Min mice develop mammary tumors, which has not been seen in human FAP patients (Moser et al., 1993). The neuro®bromatoses are genetic disorders that It should be noted that a large majority of the FAP primarily a€ect cell growth of neural tissues. The patients develop additional complications not seen in most common type, neuro®bromatosis type 1 (NF1), is Min mice, such as bone and body wall tumors, skin an autosomal dominant disorder a€ecting about 1 in cysts and enlarged pigmented retinal cells (Bullow, 3000 individuals (Gutmann and Collins, 1998). These 1987). The lack of extra-intestinal lesions, such as that disorders can cause benign tumors to grow from nerves seen in FAP patients, in the Min mouse model is at any location, most commonly in Schwann cells. A attributed to the involvement of genes other than Apc subset of these tumors may also progress into in the process of tumorigenesis. In favor of this malignant neuro®brosarcomas. Moreover, multiple argument is the observation that Min/+, p537/7 hyperpigmented areas, also called `cafe au lait mice develop pancreatic cancer and some of the body macules', are a common characteristic in these wall lesions seen in FAP patients (Clarke et al., 1995; patients. NF-1 patients are also at risk for developing W Dove, personal communication). malignant myeloid disorders as well as optic gliomas Apc-de®cient mice have also been generated using and pheochromocytomas (Hope and Mulvihill, 1981). gene targeting (Oshima et al., 1995). The targeting Likewise, infants and children with NF1 are at high resulted in the truncation of the Apc gene product at risk of developing malignant myeloid disorders (Bader amino acid 716. The heterozygous mice develop and Miller, 1978). multiple polyps throughout the small intestine, similar The NF1 gene was identi®ed in 1990 by positional to that seen in Min mice. Tumor analysis indicated that cloning (Viskochil et al., 1990; Wallace et al., 1990). the wild-type Apc gene is lost during tumorigenesis The gene is more than 300 kb long, composed of about (Oshima et al., 1995). The homozygous mice, as with 50 exons, and encodes a protein, neuro®bromin, of the Min mice, die in utero before day 8 of gestation. To 2818 amino acids. Neuro®bromin has extensive circumvent embryonic lethality due to the Apc homology with the catalytic region of the RAS de®ciency in mice, Shibata et al. (1997) used a GTPase-activating protein, GAP, and the IRA1 and conditional gene targeting system. Using the Cre-loxP IRA2 gene products of the yeast S. cerevisiae (Ballester system, the Apc gene was inactivated speci®cally in the et al., 1990). This GAP domain functions to negatively colorectal epithelium by inserting loxP sites in the regulate Ras by catalyzing the conversion of the active introns around exon 14 of the Apc gene, and GTP-bound form of Ras to the inactive GDP-bound introducing the mutant allele into the mouse germline. form (Hall, 1990). The NF1 gene is presumed to follow Mice homozygous for Apc-loxP insertion were normal. Knudson's two hit rule, since the majority of germ line However, upon infection of the colorectal region with NF1 mutations produce no protein, and clonally Mouse tumor suppressor models N Ghebranious and LA Donehower 3392 derived malignant neuro®brosarcomas express no and cytoskeleton interface and was named merlin detectable NF1 protein (Basu et al., 1992; DeClue et (meosin, ezrin, radixin like protein). The NF2 protein al., 1992). is thought to be involved in cytoskeletal reorganiza- Two research groups have reported the targeted tion events. disruption of the Nf1 gene in the mouse germ line More recently, Jacks and colleagues used gene (Brannan et al., 1994; Jacks et al., 1994). Mice targeting techniques to disrupt the Nf2 gene in mouse homozygous for the Nf1 mutation die during develop- ES cells (McClatchey et al., 1997). The homozygous ment between days 12.5 and 14 of gestation. The cause mutation at the Nf2 locus leads to embryonic failure at of early lethality has been attributed to a cardiac embryonic day (E) 6.5 ± 7.0, with the embryos dysfunction, in which several abnormalities in the heart displaying poorly developed extraembryonic structures exist, including disoriented and poorly developed and a lack of organized extraembryonic ectoderm myocardial ®bres, overall hypoplasticity, and ventri- (McClatchey et al., 1997). cular septal defects caused by insucient ventricular The Nf2 heterozygous mice are predisposed to a growth and abnormalities in valve lea¯et formation. broad spectrum of malignant tumors, though these These defects indicate that neuro®bromin has an tumors develop relatively late in the mouse life span important role in heart development. This idea is (mean incidence at 22 months of age) (McClatchey et further supported by the observation of high levels of al., 1998). The tumor spectrum was quite varied, though neuro®bromin in the heart during during develop- osteosarcomas were the most frequently observed and mental periods that coincide with the development of arose in 63% of the Nf2+/7 mice. Surprisingly, none the abnormalities (Huynh et al., 1994; Brannan et al., of the tumor types observed in the heterozygous mice 1994). Brannan et al. (1994) also reported a delay in resembled the tumor spectrum observed in NF2 renal, hepatic as well as skeletal muscle development in patients, and were much more malignant in nature the Nf1 null mice. than the relatively benign human neoplasms. Interest- Mice heterozygous for the Nf1 mutation have do not ingly, unlike tumors in p53-de®cient mice (Taverna et develop neuro®bromas or pigmentation defects such as al., 1998) and wild type mice (Frith et al., 1981), the that seen in humans (Brannan et al., 1994; Jacks et al., Nf2+/7 tumors were highly metastatic, with meta- 1994). Nevertheless, the loss of one Nf1 allele seems to static lesions frequently found in the lung and liver as increase the overall rate of tumor development in the well as other sites (McClatchey et al., 1998). mouse. About 19% of the heterozygotes develop McClatchey et al. (1998) also demonstrated linkage- pheochromocytoma, which a€ects about 1% of the speci®c e€ects of Nf2 and p53 cooperativity in NF1 patients, as well as myeloid leukemias. These tumorigenesis. In the mouse, the Nf2 gene and p53 tumors displayed loss of heterozygosity for the wild- gene are both located on chromosome 11 about 40 cm type Nf1 allele, which is also common among human apart. Doubly de®cient Nf2+/7; p53+/7 mice with NF1 patients (Jacks et al., 1994). However, despite the the two mutant alleles in cis (on the same chromosome) lack of neuro®bromas in the Nf1+/7 animals, or in trans (on separate chromosomes) were monitored chimeric mice containing a mixture of normal cells for tumors. Heterozygotes with the mutant alleles in cis and Nf17/7 cells did develop neuro®bromas similar demonstrated dramatically accelerated tumor forma- to that seen in human NF1 patients, but not of tion compared to the mice with trans con®guration of Schwann cell origin, as is the case in humans mutant alleles. However, even the trans double (Cichowski et al., 1996). heterozygotes showed accelerated tumor formation Even though the Nf17/7 mice die early during compared to the parental p53+/7 and Nf2+/7 development, transplantation of hematopoietic cells mice, indicating a marked cooperativity between these from the Nf17/7 embryos into lethally irradiated two tumor suppressor genes. mice resulted in chronic myeloid leukemias derived from the donor Nf1-de®cient cells (Largaespada et al., 1996). Hematopoietic cells derived from fetal liver of BRCA1 and BRCA2-de®cient mice (inherited breast Nf17/7 mouse embryos were shown to have a cancer susceptibility) selective hypersensitivity to GM-CSF, a feature shared with leukemic cells derived from NF1 patients (Bollag The tumor suppressor genes BRCA1 and BRCA2 have et al., 1996; Largaespada et al., 1996). The elevated been implicated in inherited predisposition to breast levels of Ras-GTP and the reduction in Nf1-GAP cancer (Miki et al., 1994; Cannon-Albreight and activity are likely to be the cause of GM-CSF Skolnick, 1996; Claus et al., 1991; Easton et al., 1993; hypersensitivity, as well as for the development of Marcus et al., 1996; Stratton and Wooster, 1996; Szabo myeloid neoplasms. and King, 1995; Couch and Weber, 1998). Mutations of Neuro®bromatosis type 2 (NF-2), is an autosomal BRCA1 also predispose to ovarian cancer (Gayther et dominant disorder a€ecting about 1 in 50 000 al., 1995; Johannsson et al., 1996; Serova et al., 1996), individuals. It causes benign tumors in the nervous whereas, mutations in BRCA2 predispose to male breast system including Schwanomas, meningiomas, and cancer as well as pancreatic cancer (Goggins et al., 1996; ependymonas (MacCollin and Gusella, 1998). Muta- Schutte et al., 1995; Teng et al., 1996). tions and loss of heterozygosity of the NF2 locus are Nearly 45% of all familial breast cancer cases have common in NF2 patients with sporadically occurring mutations in BRCA1 or BRCA2, while mutations of tumors, suggesting that NF2 behaves as a classical these genes are infrequent in sporadic non-familial tumor suppressor gene (MacCollin and Gusella, 1998). breast tumors (Lancaster et al., 1996; Miki et al., 1994; The NF2 gene was cloned using positional cloning Teng et al., 1996). Both BRCA1 and BRCA2 are (Rouleau et al., 1993). The protein product has considered tumor suppressor genes due to the frequent homology with proteins at the plasma membrane loss of heterozygosity of the wild-type allele in tumors Mouse tumor suppressor models N Ghebranious and LA Donehower 3393 from patients with germ line mutations in BRCA1 or protein have resulted in homozygous mice that are BRCA2 (Smith et al., 1992; Collins et al., 1995). viable (Connor et al., 1997; Friedman et al., 1998). Interestingly, there is some similarity in the These mice are approximately one-third normal size biochemical properties in the protein product of these and are susceptible to thymic lymphomas rather than two genes, both have transcription activation function mammary tumors. Embryo ®broblasts derived from the (Chapman et al., 1996; Milner et al., 1997) and can homozygous mice are defective in proliferation and colocalize with Rad51, a protein involved in the display a progressively reduced growth rate with regulation of recombination and double stranded passaging. Patel et al. (1998) have shown that the DNA repair, suggesting that these two genes may Brca2 homozygous cells exhibit spontaneous chromo- function in DNA repair (Sharan et al., 1997; Scully et somal abnormalities, with frequent formation of al., 1997). Moreover, BRCA2 appears to directly triradial and quadriradial chromosomes. These struc- complex with Rad51 (Sharan et al., 1997). In tures are thought to arise through aberrant chromo- addition, both genes are coordinately regulated during some exchanges and are also observed in Rad51- development (Blackshear et al., 1998), and both gene de®cient cells. Such results indicate that Brca2 and products form a stable interaction in mitotic and Rad51 are important in maintaining correct chromo- meiotic cells (Chen et al., 1998). some segregation and resolution of recombination Given their potential role in DNA repair, both intermediates during meiosis and mitosis. Interest- BRCA1 and BRCA2 proteins have been referred to as ingly, the increased appearance of these aberrant `caretakers' rather than `gatekeepers', which tend to be structures was also correlated with high levels of p53 associated with direct cell cycle control (Kinzler and and p21WAF1/CIP1, indicating a stimulated DNA damage Vogelstein, 1997). In this respect, inactivation of response. BRCA1 or BRCA2 does not promote tumor initiation Two groups have demonstrated a partial rescue of directly, but rather facilitates genetic instability and the Brca1 and Brca2 null embryonic lethality through mutations of genes, speci®cally, the gatekeeper genes. simultaneous mutation of p53 (Hakem et al., 1996; This interpretation is consistent with the ®ndings that Ludwig et al., 1997). Double mutant embryos in Brca1 BRCA1 and BRCA2 interact with Rad51, and the and p53 or Brca2 and p53 survive an additional one to similarities in the sensitivity of their null mice to two days of gestation, suggesting perhaps that radiation. Moreover, mutations in `caretaker' genes activation of p53 activity by defective DNA repair in occur primarily as germ line mutations in inherited the Brca-de®cient embryo may inhibit proliferation cancer syndromes and are less frequent in somatically during development and induce earlier abnormalities. arising tumors (Couch and Weber, 1998; Kinzler and In support of this hypothesis is the observation that Vogelstein, 1997). Thus, the primary role of BRCA1 mutation of p21WAF1/CIP1 has been shown to be capable and BRCA2 in tumor suppression may be through of delaying the Brca1 null mouse embryonic lethality promotion of DNA repair, rather than through arrest (Hakem et al., 1997). of cell proliferation. Nevertheless, recent data impli- Surprisingly, both Brca1 and Brca2 heterozygous cates BRCA1 in cell cycle control as transfection of mice do not exhibit any cancer predisposition (Hakem BRCA1 into cancer cells inhibits S phase entry through et al., 1996; Lim et al., 1996; Liu et al., 1996; Ludwig et induction of p21WAF1/CIP1 (Somasundaram et al., 1997). al., 1997; Sharan et al., 1997; Suzuki et al., 1997). This In addition, BRCA1 has been reported to associate di€erence between mouse and man may relate a with p53 and stimulate its transcriptional activity number of potential explanations, including the (Zhang et al., 1998). di€ering functional activities of the murine and human Multiple groups have reported the disruption of the Brca proteins, to the di€erences in life span between Brca1 gene in the mouse (Hakem et al., 1996; Liu et mouse and human, or to a greater resistance to Brca al., 1996; Ludwig et al., 1997). Homozygous Brca1 LOH in murine tissues. Carcinogen treatment or mutant mice die before day 7.5 of embryogenesis. crosses of the Brca1 and Brca2 heterozygous mice to Mutant mice are poorly developed, with no evidence of other tumor-susceptible mice may yet reveal some mesoderm formation. An interesting observation in tumorigenic predisposition. these mice is that they show reduced cell proliferation together with high expression of the CDK inhibitor, p21WAF1/CP1. This observation suggests that the death of WT-1 de®cient mice (Wilms' tumor) the mutant embryos are due to failure in the proliferation needed for the establishment of the Urogenital malformations as well as the appearance of di€erent germ layers (Hakem et al., 1996). childhood tumors of the kidney are caused by Gene targeting which leads to early truncation of mutations in the Wilms' tumor-associated gene, WT-1 The Brca2 gene leads to embryonic lethality in mouse (Baird et al., 1992; Bruening et al., 1992; Cowell et al., (Sharan et al., 1997; Suzuki et al., 1997). The embryos 1991; Davis et al., 1991; Haber et al., 1990; Hu€ et al., arrest at the same time as the onset of detectable Brca2 1991; Little et al., 1992; Pelletier et al., 1991a; Ton et expression after day 6.5 of gestation. Like Brca1- al., 1991; van Heyningen et al., 1990). Consistent with de®cient mice, apoptosis is normal in Brca2 null mice, this phenotype, the WT-1 gene has been shown to be whereas cellular proliferation is impaired (Sharan et al., involved in the urogenital development (Pelletier et al., 1997; Suzuki et al., 1997). Interestingly, Brca2-de®cient 1991b, Pritchard-Jones et al., 1990). mice exhibit increased sensitivity to radiation (Sharan The WT-1 gene maps to the short arm of et al., 1997), which is a feature shared with Rad51- chromosome 11. Wilms' tumors show cytogenetically de®cient mice (Lim et al., 1996). visible deletions as well as loss of heterozygosity at the More recent targeting experiments in the Brca2 gene WT-1 locus, suggesting that these tumor suppressor utilizing smaller C-terminal truncations in the encoded gene follows Knudson's two hit rule (Knudson, 1985; Mouse tumor suppressor models N Ghebranious and LA Donehower 3394 Haber and Housman, 1992). The gene product of WT-1 PTEN-de®cient mice (Cowden syndrome) is a proline/glutamine rich amino terminal domain protein, which contains four zinc ®nger domains at the PTEN (for `phosphatase and tensin homologue deleted carboxy terminal (Buckler et al., 1991; Call et al., 1990; from chromosome 10') has recently been identi®ed as a Morris et al., 1991). In addition, the WT-1 protein has candidate tumor suppressor gene and has been found homology to the EGR family of transcription factors, to be mutated in gliomas (Rasheed et al., 1997), and has been shown to transcriptionally repress genes endometrial tumors (Tashiro et al., 1997), prostate with the EGR-1 recognition sites (Drummond et al., cancers (Cairns et al., 1997), and breast cancers (Rhei 1992; Call et al., 1990). et al., 1997). Germ line mutations in the PTEN gene In 1993, Kreidberg et al. (1993) successfully have been associated with Cowden disease, an inherited disrupted the Wt-1 gene in mice using gene targeting cancer predisposition syndrome characterized by early technique in ES cells. Homozygous mice for the Wt-1 development of breast, thyroid, and brain tumors (Eng mutation die at day 11 of gestation of kidney and and Parsons, 1998). Two related familial syndromes, gonad developmental failure. In addition, the mutation Lhermitte-Duclos disease and Bannayan-Zonana syn- revealed developmental abnormalities in the mesothe- drome, show similar predispositions to tumors as lium, heart, and lungs. Despite the clear role of Wt-1 in Cowden disease and have germ line PTEN mutations, development, mice heterozygous for the Wt-1 mutation but have some developmental di€erences which do not develop tumors. distinguish them from one another (Zonana et al., 1976; Eng et al., 1994). The pTEN gene encodes a protein with homology to VHL-de®cient mice (von Hippel-Lindau syndrome) dual speci®city protein phosphatases and tensin, a protein associated with the cytoskeleton at focal Inheritance of an inactivated form of the VHL tumor adhesions (Eng and Parsons, 1998). PTEN interacts suppressor gene predisposes patients to develop von with focal adhesion kinase and reduces its phosphor- Hippel-Lindau disease. The disease is characterized by ylation. It was also shown to reduce cell migration, development of tumors at multiple sites, including formation of focal adhesions, and extracellular matrix retinal angiomas, hemangioblastomas of the central interactions (Tamura et al., 1998). It is the ®rst nervous system, pheochromocytomas, renal cell carci- example of a classical tumor suppressor that is a nomas, and pancreatic cancers (Gnarra et al., 1996; phosphatase. Linehan and Klausner, 1998). Moreover, somatic Recently, PTEN-de®cient mice have been generated mutations in one allele of the VHL gene and by Di Cristofano et al. (1998) and Stambolic et al. inactivation of the other through hypermethylation or (1998). PTEN null embryos displayed an early mutation, occurs in approximately 80% of sporadic embryonic lethal phenotype. Interestingly, the null clear cell renal carcinoma (Gnarra et al., 1996; Linehan embryos exhibited a dramatic overgrowth in the and Klausner, 1998). cephalic and caudal regions. Moreover, BrdU labeling Like most tumor suppressor genes, the human VHL assays indicated a signi®cantly higher proliferative gene has been identi®ed by positional cloning (Latif et index in the cells of the developing null embryo al., 1993). The VHL protein structure has been compared to its wild type siblings (Stambolic et al., deduced and was shown to physically interact with 1998). Despite the enhanced proliferation of PTEN elongin B and elongin C, RNA polymerase II null cells in vivo, ES cells and ®broblasts derived from elongation factors. Loss of VHL appears to result in these embryos showed no enhanced proliferation. ES an inappropriate increase in hypoxia-inducible cells, however, did exhibit enhanced anchorage- mRNAs, which may account for the high degree of independent cell growth (Di Cristofano et al., 1998). vascularization of VHL-associated cancers (Kaelin et Moreover, ®broblasts showed greatly decreased sus- al., 1998). In addition, the elongin-VHL complex was ceptibility to apoptosis following treatment with a shown to interact with a human homologue of a variety of apoptosis-inducing agents. This reduction in member of a conserved gene family, CUL-2, involved apoptosis was traced to reduced levels of depho- in cell cycle and growth control (Aso et al., 1995; Duan sphorylation and increased kinase activity of protein et al., 1995; Kibel et al., 1995; Pause et al., 1997). kinase B/Akt, a kinase which mediates cell survival Introduction of wild type VHL into VHL-negative (Stambolic et al., 1998). Increased kinase activity of renal carcinoma cells suppresses their growth rate by protein kinase B protects cells from apoptosis induced causing them to exit the cell cycle into G0/quiescence by withdrawal of survival signals. Thus, absence of (Pause et al., 1998). PTEN may contribute to tumorigenesis either through Gene targeting techniques have been employed to increasing cell proliferation rates (as observed in the inactivate the VHL gene in the germ line of mice (Gnarra early embryos), by protection against apoptosis (as et al., 1997). Heterozygous mice developed normally and observed in the MEF studies), or by increased showed no phenotype, whereas mice that are homozygous invasiveness consistent with the enhanced anchorage for Vhl deletion die between days 10.5-12.5 in utero. The independence observed in PTEN7/7 ES cells. likely cause of death in these mice was attributed to Con®rmation that PTEN mutations can predispose inadequate vasculogenesis in the placenta, a phenotype to cancer was demonstrated by the fact that PTEN+/ consistent with the involvement of VHL in regulating 7 mice are susceptible to early tumors. Stambolic et levels of angiogenesis-associated mRNAs (Gnarra et al., al. (1998) noted a frequent development of T cell 1997). One interesting aspect of ®broblasts derived from lymphomas, while Di Cristofano et al. (1998) described VHL null embryos was a gross defect in ®bronectin matrix a broader array of tumor types, including apparent assembly, indicating that VHL may also have an colon adenocarcinomas, leukemia, and germ cell extracellular role (Ohh et al., 1998). tumors. However, they also observed hyperplasias Mouse tumor suppressor models N Ghebranious and LA Donehower 3395 and dysplasias of the prostate, skin, and colon, which have grouped the mouse models according to the are characteristic of the three syndromes with PTEN similarity of the tumor types observed in the germ line mutations. The PTEN+/7 T-cell lympho- heterozygous mice with those in the corresponding mas of Stambolic et al. (1998) were associated with human syndrome. As indicated, Apc, p53, Nf1, and PTEN loss of heterozygosity, whereas the various PTEN heterozygotes (Group I) exhibit some degree of PTEN+/7 tumors described by Di Cristofano et al. correlation. Li-Fraumeni syndrome patients and (1998) appeared to retain the wild type PTEN allele. p53+/7 mice both show frequent development of However, the structural integrity of the PTEN allele at osteosarcomas, soft tissue sarcomas, and hematopoietic the nucleotide level and its expression in these tumors neoplasms (Malkin, 1998). However, the correlation is remain to be determined. Stambolic et al. (1998) imperfect, since Li-Fraumeni patients exhibit frequent concluded that PTEN belongs to the `gatekeeper' breast tumors and brain tumors, which are relatively category of tumor suppressor genes as it directly rare in the p53+/7 mice (Malkin et al., 1990; a€ects cell growth regulatory pathways. Donehower, 1996). Familial adenomatous polyposis patients develop carcinomas in the large intestine, while Min mice develop tumors exhibit tumors in the small Conclusions intestine, which should be considered a relatively minor di€erence with respect to the Min mice serving as a The mouse tumor suppressor models described here useful model (Bilger et al., 1996). Nf1 heterozygotes have an array of interesting phenotypes that have shed and NF1 patients both show a susceptibility to new light on the role of these genes in embryonic pheochromocytomas and myeloid leukemia, though development, cell cycle regulation, and tumor suppres- Nf1+/7 mice fail to develop neuro®bromas, the sion. In addition, the crossing of these models to each signature tumor of NF1 patients (Cichowski et al., other and to other informative models has facilitated 1996). However, chimeric mice composed of Nf17/7 the study of the cooperativity of di€erent combinations cells have been developed and these have displayed of tumor suppressors and oncogenes in embryogenesis neuro®broma-like lesions along various nerves of the and tumorigenesis. peripheral nervous system (Cichowski et al., 1996). One of the rationales for generating these mice is Finally, PTEN heterozygotes exhibited thyroid and that they may serve as a useful model for the colon tumors and prostate, skin and colon dysplasias, corresponding human inherited cancer syndrome. As which are characteristic of Cowden disease patients (Di can be seen in Table 1 showing the tumor spectra Cristofano et al., 1998; Eng and Parsons, 1998). observed in the mouse models, the mice have achieved Group II mice represent those heterozygous models varying levels of success in emulating the cognate that develop tumors, but not of the same type observed human inherited cancer predisposition. In Table 1 we in the cognate human syndrome. This category

Table 1 Classi®cation of tumor suppressor genes based on human and mouse phenotype di€erences Group Gene Tumors in humans Tumors in +/7 mice Phenotype in 7/7 mice Similarity to humans I APC Colon carcinomas Intestinal carcinomas Embryonic lethality Some similarity p53 Breast cancer Mammary tumors (rare) Lymphomas Some similarity Osteosarcomas Osteosarcomas Osteosarcomas Soft tissue sarcomas Soft tissue sarcomas Soft tissue sarcomas Leukemias/Lymphomas Lymphomas Brain tumors Brain tumors (rare) NF-1 Neuro®bromas Pheochromocytomas Embryonic lethality Some similarity Pheochromocytomas Optic gliomas Schwannomas Myeloid leukemia Myeloid leukemia PTEN Hamartomas Lymphomas Embryonic lethality Some similarity Breast cancer Gonado-stromal tumors Thyroid tumors Thyroid tumors Colon tumors Germ cell tumors II RB Retinoblastoma Pituitary adenomas Embryonic lethality No similarity NF2 Schwannomas Osteosarcomas Embryonic lethality No similarity Meningiomas Lymphomas Ependymomas Lung adenocarcinomas Fibrosarcomas Liver tumors III p16INK4 Melanomas No tumors Lymphomas No similarity Pancreatic adenocarcinomas Fibrosarcomas p19ARF Melanomas No tumors Lymphomas No similarity Pancreatic adenocarcinomas Fibrosarcomas IV BRCA1 Breast cancer No tumors Embryonic lethality No similarity Ovarian cancer BRCA2 Breast cancer No tumors Embryonic lethality No similarity WT-1 Kidney tumors No tumors Embryonic lethality No similarity VHL Kidney tumors No tumors Embryonic lethality No similarity Other tumor types Mouse tumor suppressor models N Ghebranious and LA Donehower 3396 includes the Rb and Nf2 heterozygous mice. The Rb di€erences could be explained by species-dependent heterozygotes develop only pituitary adenomas, unlike di€erences in susceptibility to loss of the remaining susceptible children with retinoblastoma susceptibility wild type tumor suppressor allele. Or perhaps the (Newsham et al., 1998). The Nf2 heterozygotes reduced target cell number or shorter life span of mice succumb to malignant and metastatic osteosarcomas, a€ect susceptibility in certain tissues. Moreover, the lymphomas, lung adenocarcinomas, and liver tumors propensity of many of the tumor suppressor-de®cient rather than to the relatively benign tumors of the mouse models to develop lymphomas and sarcomas central nervous system seen in NF2 a€ected humans may merely re¯ect the fact that the C57BL/6 strain into (McClatchey et al., 1998; MacCollin and Gusella, which the mutated alleles are usually crossed are highly 1998). Thus, while the group II mice may not be susceptible to these tumor types. Thus, the tumor useful in modeling speci®c tumors for the human suppressor lesions in the mice may greatly exacerbate syndrome, the very fact that they develop tumors an earlier predisposition. The crossing of these lesions makes them quite useful for investigating the under- into di€erent strains may alter the tumor spectra and lying molecular mechanisms by which their tumor provide the foundation for identifying novel tumor suppressor lesions predispose to early cancers. modi®er loci, as has been demonstrated by Dove p16 and p19ARF-de®cient mice could be considered (Bilger et al., 1996). to be members of Category III. Cancers have not been In summary, we have described the developmental reported for the heterozygous p16 and p19ARF mice, and tumorigenic consequences of mouse germ line but the nullizygous animals do develop tumors mutations in the `gatekeeper' category of tumor (sarcomas and lymphomas) unlike those observed in suppressors for which there is a corresponding defect the melanoma prone human syndrome (Kamb and in a human inherited cancer syndrome. As we have Herlyn, 1998). seen, the tumor suppressor-de®cient mice and the cells The ®nal category, Group IV, is composed of derived from them have provided excellent tools for the models in which no tumors above background levels development of critical new insights into tumor have been observed in the heterozygotes. This group suppressor function. As new human tumor suppres- includes the Brca1, Brca2, Vhl, and Wt-1 hetero- sors are identi®ed, mouse knockout models for these zygotes. It may be possible to show accelerated tumors new genes will be generated and further conceptual in mice of this category through carcinogen protocols advances will be made. Moreover, the development of or through crosses to other tumor-suceptible models. second generation gene targeting technologies have A number of explanations have been proposed for made it possible to make more subtle gene alterations the lack of correlation in tumor phenotypes between or conditional lesions that can provide more functional mouse and human (Jacks, 1996). The species-speci®c information beyond that provided by the mice with di€erences could be a result of relatively subtle simple null alleles. The next several years promise to be di€erences in growth control pathways in mouse and very exciting ones in the development of mouse models human cells in particular tissues. Alternatively, the for tumor suppressors.

References

Armstrong JF, Kaufman MH, Harrison DJ and Clarke AR. Bollag G, Clapp DW, Shih S, Adler F, Zhang YY, (1995). Curr. Biol., 5, 931 ± 936. Thompson P, Lange BJ, Freedman MH, McCormick F, Aso T, Lane WS, Conaway JW and Conaway RC. (1995). Jacks T and Shannon K. (1996). Nat. Genet., 12, 144 ± 148. Science, 269, 1439 ± 1443. BooksteinR,LeeEY,ToH,YoungLJ,SeryTW,HayesRC, Bader JL and Miller RW. (1978). J. Pediatr., 92, 925 ± 929. Friedmann T and Lee WH. (1988). Proc. Natl. Acad. Sci. Baird PN, Groves N, Haber DA, Housman DE and Cowell USA., 85, 2210 ± 2214. JK. (1992). Oncogene, 7, 2141 ± 2149. Brannan CI, Perkins AS, Vogel KS, Ratner N, Nordlund Baker SJ, Fearon ER, Nigro JM, Hamilton SR, Preisinger ML, Reid SW, Buchberg AM, Jenkins NA, Parada LF and AC, Jessup JM, vanTuinen P, Ledbetter DH, Barker DF Copeland NG. (1994). Genes Dev., 8, 1019 ± 1029. and Nakamura Y. (1989). Science, 244, 217 ± 221. Bruening W, Bardeesy N, Silverman BL, Cohn RA, Machin Baker SJ, Markowitz S, Fearon ER, Willson JK and GA, Aronson AJ, Housman D and Pelletier J. (1992). Nat. Vogelstein B. (1990). Science, 249, 912 ± 915. Genet., 1, 144 ± 148. Ballester R, Marchuk D, Boguski M, Saulino A, Letcher R, Buckbinder L, Talbott R, Velasco-Miguel S, Takenaka I, Wigler M and Collins F. (1990). Cell, 63, 851 ± 859. Faha B, Seizinger BR and Kley N (1995). Nature, 377, Barak Y, Juven T, Ha€ner R and Oren M (1993). EMBO J., 646 ± 649. 12, 461 ± 468. Buckler AJ, Pelletier J, Haber DA, Glaser T and Housman Basu TN, Gutmann DH, Fletcher JA, Glover TW, Collins DE. (1991). Mol. Cell. Biol., 11, 1707 ± 1712. FS and Downward J. (1992). Nature, 356, 713 ± 715. Bulow S. (1987). Dan. Med. Bull., 34, 1 ± 15. Bates S, Phillips AC, Clark PA, Stott F, Peters G, Ludwig CairnsP,OkamiK,HalachmiS,HalachmiN,EstellerM, RL and Vousden KH. (1998). Nature, 395, 124 ± 125. Herman JG, Jen J, Isaacs WB, Bova GS and Sidransky D. Bedell MA, Largaespada DA, Jenkins NA and Copeland (1997). Cancer Res., 57, 4997 ± 5000. NG. (1997). Genes Dev., 11, 11 ± 43. Call KM, Glaser T, Ito CY, Buckler AJ, Pelletier J, Haber Bilger A, Shoemaker AR, Gould KA and Dove WF. (1996). DA, Rose EA, Kral A, Yeger H, Lewis WH. et al. (1990). Sem. Cancer Biol., 7, 249 ± 260. Cell, 60, 509 ± 520. Bischo€JR,CassoDandBeachD.(1992).Mol. Cell. Biol., Cannon-Albright LA and Skolnick MH. (1996). Sem. 12, 1405 ± 1411. Oncol., 23, 1±5. Blackshear PE, Goldsworthy SM, Foley JF, McAllister KA, Chapman MS and Verma IM. (1996). Nature, 382, 678 ± 679. Bennett LM, Collins NK, Bunch DO, Brown P, Wiseman Chen PL, Chen YM, Bookstein R and Lee WH. (1990). RW and Davis BJ. (1998). Oncogene, 16, 61 ± 68. Science, 250, 1576 ± 1580. Mouse tumor suppressor models N Ghebranious and LA Donehower 3397 Chen J, Silver DP, Walpita D, Cantor SB, Gazdar AF, Easton D, Ford D and Peto J. (1993). Cancer Surv., 18, 95 ± Tomlinson G, Couch FJ, Weber BL, Ashley T, Livingston 113. DM and Scully R. (1998). Mol. Cell, 2, 317 ± 328. el-Deiry WS, Tokino T, Velculescu VE, Levy DB, Parsons R, Cichowski K, Shih TS and Jacks T. (1996). Sem. Cancer Trent JM, Lin D, Mercer WE, Kinzler KW and Vogelstein Biol., 7, 291 ± 298. B. (1993). Cell, 75, 817 ± 825. Clarke AR, Maandag ER, Roon MV, Van Lught NMT, van Eliyahu D, Raz A, Gruss P, Givol D and Oren M (1984). der Valk M, Hooper ML, Berns A and Riele H. (1992). Nature, 312, 646 ± 649. Nature, 359, 328 ± 330. Eliyahu D, Michalovitz D, Eliyahu S, Pinhasi-Kimhi O and Clarke AR, Purdie CA, Harrison DJ, Morris RG, Bird CC, Oren M. (1989). Proc. Natl. Acad. Sci. USA, 86, 8763 ± Hooper ML and Wyllie AH. (1993). Nature, 362, 849 ± 8767. 852. Eng C, Murday V, Seal S, Mohammed S, Hodgson SV, Clarke AR, Cummings MC and Harrison DJ. (1995). Chaudary MA, Fentiman IS, Ponder BA and Eeles RA. Oncogene, 11, 1913 ± 1920. (1994). J. Med Genet., 31, 458 ± 461. Claus EB, Risch N and Thompson WD. (1991). Am.J.Hum. Eng C and Parsons R. (1998). The genetic basis of human Genet., 48, 232 ± 242. cancer. Vogelstein B & Kinzler K. (eds). McGraw-Hill Cobrinik D, Lee MH, Hannon G, Bronson RT, Dyson N, Co., New York, pp. 519 ± 526. Harlow E, Beach D, Weinberg RA and Jacks T. (1996). Finlay CA, Hinds PW and Levine AJ. (1989). Cell, 57, 1083 ± Genes Dev., 10, 1633 ± 1644. 1093. Collins N, McManus R, Wooster R, Mangion J, Seal S, Fountain JW, Karayiorgou M, Ernsto€ MS, Kirkwood JM, LakhaniSR,OrmistonW,DalyPA,FordD,EastonDF. Vlock DR, Titus-Ernsto€ L, Bouchard B, Vijayasaradhi et al. (1995). Oncogene, 10, 1673 ± 1675. S, Houghton AN, Lahti J. et al. (1992). Proc. Natl. Acad. Connor F, Bertwistle D, Mee PJ, Ross GM, Swift S, Sci. USA, 89, 10557 ± 10561. Grigorieva E, Tybulewicz VL and Ashworth A. (1997). Friedman LS, Thistlethwaite FC, Patel KJ, Yu VP, Lee H, Nat. Genet., 17, 423 ± 430. Venkitaraman AR, Abel KJ, Carlton MB, Hunter SM, Cormier RT, Hong KH, Halberg RB, Hawkins TL, Colledge WH, Evans MJ and Ponder BA. (1998). Cancer Richardson P, Mulherkar R, Dove WF and Lander ES. Res., 58, 1338 ± 1343. (1997). Nat. Genet., 17, 88 ± 91. Friend SH, Bernards R, Rogelj S, Weinberg RA, Rapaport Couch, F & Weber, BL. (1998). The genetic basis of human JM, Albert DM and Dryja TP. (1986). Nature, 323, 643 ± cancer. Vogelstein B & Kinzler K. (eds). McGraw-Hill 646. Co., New York, pp. 537 ± 564. Frith CH, Little®eld NA and Umholtz R. (1981). J. Natl. CowellJK,WadeyRB,HaberDA,CallKM,HousmanDE Cancer Inst., 66, 703 ± 712. and Pritchard J. (1991). Oncogene, 6, 595 ± 599. Fukasawa K, Choi T, Kuriyama R, Rulong S and Vande Cross SM, Sanchez CA, Morgan CA, Schimke MK, Ramel Woude GF. (1996). Science, 271, 1744 ± 1747. S, Idzerda RL, Raskind WH and Reid BJ. (1995). Science, Fung YK, Murphree AL, T'Ang A, Qian J, Hinrichs SH and 267, 1353 ± 1356. Benedict WF. (1987). Science, 236, 1657 ± 1661. DavisLM,ZabelB,SengerG,LudeckeHJ,MetzrothB,Call Gayther SA, Warren W, Mazoyer S, Russell PA, Harrington K, Housman D, Claussen U, Horsthemke B and Shows PA,ChianoM,SealS,HamoudiR,vanRensburgEJ, TB. (1991). Genomics, 10, 588 ± 592. Dunning AM. et al. (1995). Nat. Genet., 11, 428 ± 433. DebS,JacksonCT,SublerMAandMartinDW.(1992).J. Gnarra JR, Duan DR, Weng Y, Humprey JS, Chen DY, Lee Virol., 66, 6164 ± 6170. S, Pause A, Dudley CF, Latif F, Kuzmin I, Schmidt L, DeClue JE, Papageorge AG, Fletcher JA, Diehl SR, Ratner Duh FM, Stackhouse T, Chen F, Kishida T, Wei MH, N, Vass WC and Lowy DR. (1992). Cell, 69, 265 ± 273. Lerman MI, Abar Z, Klausner RD and Linehan WM. Del Sal G, Ruaro EM, Utrera R, Cole CN, Levine AJ and (1996). Biochim. Biophys. Acta, 1242, 201 ± 210. Schneider C. (1995). Mol. Cell Biol., 15, 7152 ± 7160. Gnarra JR, Ward JM, Porter FD, Wagner JR, Devor DE, Di Cristofano A, Pesce B, Cordon-Cardo C and Pandol® PP. Grinberg A, Emmert-Buck M, Westphal H, Klausner RD (1998). Nat. Genet., 19, 348 ± 355. and Linehan WM. (1997). Proc. Natl. Acad. Sci. USA, 94, Dietrich WF, Copeland NG, Gilbert DJ, Miller JC, Jenkins 9102 ± 9107. NA and Lander ES. (1995). Proc. Natl. Acad. Sci. USA, Goggins M, Schutte M, Lu J, Moskaluk CA, Weinstein CL, 92, 10849 ± 10853. Petersen GM, Yeo CJ, Jackson CE, Lynch HT, Hruban Dietrich WF, Lander ES, Smith JS, Moser AR, Gould KS, RH and Kern SE. (1996). Cancer Res., 56, 5360 ± 5364. Luongo C, Borenstein N and Dove W. (1993). Cell, 75, Goldstein AM, Fraser MC, Struewing JP, Hussussian CJ, 631 ± 639. Ranade K, Zametkin DP, Fontaine LS, Organic SM, Donehower LA, Harvey M, Slagle BL, McArthur MJ, Dracopoli NC, Clark Jr. WH. et al. (1995). N. Engl. J. Montgomery CA, Butel JS and Bradley A. (1992). Med., 333, 970 ± 974. Nature, 356, 215 ± 221. Gondo Y, Nakamura K, Sasaoka T, Ito K, Kimura M and DonehowerLA,GodleyLA,AldazCM,PyleR,ShiYP, Katsuki M. (1994). Biochem. Biophys. Res. Commmun., Pinkel D, Gray J, Bradley A, Medina D and Varmus H. 202, 830 ± 837. (1995b). Genes Dev., 9, 882 ± 895. Gottlieb E and Oren M. (1998). EMBO J., 17, 3587 ± 3596. Donehower LA, McArthur MJ, Montgomery SAJ, Park SH, Gottlieb TM and Oren M. (1996). Biochim. Biophys. Acta, Thompson T, Ford RJ and Bradley A. (1995a). Mol. 1287, 77 ± 102. Carcinog., 14, 16 ± 22. Groden J, Thliveris A, Samowitz W, Carlson M, Gelbert L, Donehower LA. (1996). Sem. Cancer Biol., 7, 269 ± 278. Albertsen H, Joslyn G, Stevens J, Spirio L, Robertson M, Drummond IA, Madden SL, Rohwer NP, Bell GI, Sukhatme Sargeant L, Krapcho K, Wol€ E, Burt R, Hughes JP, VP and Rauscher FI. (1992). Science, 257, 674 ± 678. Warrington J, McPherson J, Wasmuth J, Le Paslier D, Duan DR, Pause A, Burgess WH, Aso T, Chen DY, Garrett Abderrahim H, Cohen D, Leppert M and White R. (1991). KP, Conaway RC, Conaway JW, Linehan WM and Cell, 66, 589 ± 600. Klausner RD. (1995). Science, 269, 1402 ± 1406. GuanKL,JenkinsCW,LiY,NicholsMA,WuX,O'Keefe Dunn JM, Phillips RA, Zhu X, Becker A and Gallie BL. CL, Matera AG and Xiong Y. (1994). Genes Dev., 8, (1989). Mol. Cell. Biol., 9, 4596 ± 4604. 2939 ± 2952. Duro D, Bernard O, Della Valle V, Berger R and Larsen CJ. (1995). Oncogene, 11, 21 ± 29. Mouse tumor suppressor models N Ghebranious and LA Donehower 3398 Gutman, D & Collins, F. (1998). The genetic basis of human Kamb A, Gruis NA, Weaver-Feldhaus J, Liu Q, Harshman cancers. Vogelstein B & Kinzler K. (eds). McGraw-Hill K, Tavtigian SV, Stockert E, Day RS, Johnson BE and Co., New York, pp. 423 ± 442. Skolnick MH. (1994). Science, 264, 436 ± 440. Haber DA. (1997). Cell, 91, 555 ± 558. Kamb, A & Heryln, M. (1998). The genetic basis of human Haber DA, Buckler AJ, Glaser T, Call KM, Pelletier J, Sohn cancer. Vogelstein B & Kinzler K. (eds). McGraw-Hill RL, Douglass EC and Housman DE. (1990). Cell, 61, Co,. New York, pp. 507 ± 518. 1257 ± 1269. Kamijo T, Zindy F, Roussel MF, Quelle DE, Downing JR, Haber DA and Housman DE. (1992). Adv. Cancer Res., 59, Ashmun RA, Grosveld G and Sherr CJ. (1997). Cell, 91, 41 ± 68. 649 ± 659. Hakem R, de la Pompa JL, Elia A, Potter J and Mak TW. Kamijo T, Weber JD, Zambetti G, Zindy F, Roussel MF and (1997). Nat. Genet., 16, 298 ± 302. Sherr CJ. (1998). Proc. Natl. Acad. Sci. USA, 95, 8292 ± HakemR,delaPompaJL,SirardC,MoR,WooM,Hakem 8297. A, Wakeham A, Potter J, Reitmair A, Billia F, Firpo E, Kastan MB, Zhan Q, El-Deiry WS, Carrier F, Jacks T, Hui CC, Roberts J, Rossant J and Mak TW. (1996). Cell, Walsh WV, Plunkett BS, Vogelstein B and Fornace AJ. 85, 1009 ± 1023. (1992). Cell, 71, 587 ± 597. Hall A. (1990). Science, 249, 635 ± 640. Kibel A, Iliopoulos O, DeCaprio JA and Kaelin Jr. WG. Hanahan, D. (1989). Science, 246, 1265 ± 1275. (1995). Science, 269, 1444 ± 1446. Harvey M, McArthur MJ, Montgomery CA, Butel JS, Kinzler KW, Nilbert MC, Su LK, Vogelstein B, Bryan TM, Bradley A and Donehower LA. (1993). Nat. Genet, 5, LevyDB,SmithKJ,PreisingerAC,HedgeP,McKechnie 225 ± 229. D. et al. (1991). Science, 253, 661 ± 665. HarveyM,VogelH,LeeEYHP,BradleyAandDonehower Kinzler KW and Vogelstein B. (1997). Nature, 386, 761 ± 763. LA. (1995). Cancer Res., 55 , 1146 ± 1151. Kinzler KW and Vogelstein B. (1998). Science, 280, 1036 ± Hope DG and Mulvihill JJ. (1991). Adv. Neurol., 29, 33 ± 56. 1037. Horowitz JM, Yandell DW, Park SH, Canning S, Whyte P, Kirby ML, Kumiski DH, Myers T, Cerjan C and Mishima N. Buchkovich K, Harlow E, Weinberg RA and Dryja TP. (1993). Dev. Dyn., 198, 296 ± 311. (1989). Science, 243, 937 ± 940. Knudson A, Vande Woude GF, Friend SH, Cavenee WK Hoschuetzky H, Aberle H and Kemler R. (1994). J. Cell and Brodeur GM. (1991). Cancer Res., 51, 5435 ± 5439. Biol., 127, 1375 ± 1380. Knudson AG. (1971). Proc. Natl. Acad. Sci. USA, 68, 820 ± Howes KA, Ransom N, Papermaster DS, Lasudry JGH, 823. Albert DM and Windle JJ. (1994). Genes Dev., 8, 1300 ± Knudson AG. (1985). Cancer Res., 45, 1437 ± 1443. 1310. Ko LJ and Prives C. (1996). Genes Dev., 10, 1054 ± 1072. Hu N, Gutsmann A, Herbert DC, Bradley A, Lee WH and Koh J, Enders GH, Dynlacht BD and Harlow E. (1995). Lee EY. (1994). Oncogene, 9, 1021 ± 1027. Nature, 375, 506 ± 510. Hu€V,MiwaH,HaberDA,CallKM,HousmanD,Strong Kreidberg JA, Sariola H, Loring JM, Maeda M, Pelletier J, LC and Saunders GF. (1991). Am.J.Hum.Genet.,48, HousmanDandJaenischR.(1993).Cell, 74, 679 ± 691. 997 ± 1003. Lancaster JM, Wooster R, Mangion J, Phelan CM, Cochran Hundley JE, Koester SK, Troyer DA, Hilsenbeck SG, Subler C, Gumbs C, Seal S, Barfoot R, Collins N and Bingell G. MA and Windle JJ. (1997). Mol. Cell. Biol., 17, 723 ± 731. (1996). Nat. Genet., 13, 238 ± 240. Hussussian CJ, Struewing JP, Goldstein AM, Higgins PA, Lane DP and Crawford LV. (1979). Nature, 278, 261 ± 263. Ally DS, Sheahan MD, Clark WHJ, Tucker MA and Lane DP. (1992). Nature, 358, 15 ± 16. Dracopoli NC. (1994). Nat. Genet., 8, 15 ± 21. Largaespada DA, Brannan CI, Jenkins NA and Copeland Huynh DP, Nechiporuk T and Pulst SM. (1994). Dev. Biol., NG. (1996). Nat. Genet., 12, 137 ± 143. 161, 538 ± 551. LatifF,ToryK,GnarraJ,YaoM,DuhFM,OrcuttML, JacksT,FazeliA,SchmittEM,BronsonRT,DoodellMA Stackhouse T, Kuzmin I, Modi W, Geil L. et al. (1993). and Weinberg RA. (1992). Nature, 359, 295 ± 300. Science, 260, 1317 ± 1320. Jacks T, Shih TS, Schmitt EM, Bronson RT, Bernards A and LeeEYHP,ChangC-Y,HuN,WangY-CJ,LaiC-C,Herrup Weinberg RA. (1994). Nat. Genet., 7, 353 ± 361. K, Lee W-H and Bradley A. (1992). Nature, 359, 288 ± 294. Jacks T, Remington L, Williams BO, Schmitt EM, Halachmi Lee MH, Williams BO, Mulligan G, Mukai S, Bronson RT, S, Bronson RT and Weinberg RA. (1994). Curr. Biol., 4, Dyson N, Harlow E and Jacks T. (1996). Genes Dev., 10, 1±7. 1621 ± 1632. Jacks T. (1996). Annu.Rev.Genet.,30, 603 ± 636. Lee WH, Bookstein R, Hong F, Young LJ, Shew JY and Lee James CD and Olson JJ. (1996). Curr. Opin. Oncol, 8, 188 ± EY. (1987). Science, 235 , 1394 ± 1399. 195. Levine AJ. (1997). Cell, 88, 323 ± 331. Jen J, Harper JW, Bigner SH, Bigner DD, Papadopoulos N, LevyDB,SmithKJ,Beazer-BarclayY,HamiltonSR, Markowitz S, Willson JK, Kinzler KW and Vogelstein B. Vogelstein B and Kinzler KW. (1994). Cancer Res., 54, (1994). Cancer Res., 54, 6353 ± 6358. 5953 ± 5958. Jenkins JR, Rudge K and Currie GA. (1984). Nature, 312, Lim DS and Hasty P. (1996). Mol. Cell. Biol., 16, 7133 ± 651 ± 654. 7143. Johannsson O, Ostermeyer EA, Hakansson S, Friedman LS, Lin SJ, Skapek S and Lee EYHP. (1996). Sem. Cancer Biol., Johansson U, Sellberg G, Brondum-Nielsen K, Sele V, 7, 279 ± 289. Olsson H, King MC and Borg A. (1996). Am.J.Hum. Linehan, WM & Klausner, RD. (1998). The genetic basis of Genet., 58, 441 ± 450. human cancer. Vogelstein B & Kinzler K. (eds). McGraw- JonesJM,AttardiL,GodleyLA,LauciricaR,MedinaD, Hill Co., New York, pp. 455 ± 474. Jacks T, Varmus H and Donehower LA. (1997). Cell Linzer DI and Levine AJ. (1979). Cell, 17, 43 ± 52. Growth & Di€erent., 8, 829 ± 838. Little MH, Prosser J, Condie A, Smith PJ, Van Heyningen V Joslyn G, Carlson M, Thliveris A, Albertsen H, Gelbert L, and Hastie ND. (1992). Proc. Natl. Acad. Sci. USA, 89, Samowitz W, Groden J, Stevens J, Spirio L, Robertson M. 4791 ± 4795. et al. (1991). Cell, 66, 601 ± 613. Liu CY, Flesken-Nikitin A, Li S, Zeng Y and Lee WH. Joslyn G, Richardson DS, White R and Alber T. (1993). (1996). Genes Dev., 10, 1835 ± 1843. Proc. Natl. Acad. Sci. USA, 90, 11109 ± 11113. Liu L, Lassam NJ, Slingerland JM, Bailey D, Cole D, Kaelin WG, Iliopoulos O, Lonergan KM and Ohh M. Jenkins R and Hogg D. (1995). Oncogene, 11, 405 ± 412. (1998). J Intern. Med, 243, 535 ± 539. Mouse tumor suppressor models N Ghebranious and LA Donehower 3399 Livingstone LR, White A, Sprouse J, Livanos E, Jacks T and Newsham I, Hadjistilianou T & Cavenee WK. (1998). The Tlsty TD. (1992). Cell, 70, 605 ± 620. genetic basis of human cancer. Vogelstein B & Kinzler K. Lowe SW, Schmitt EM, Smith SW, Osborne BA and Jacks T. (eds). McGraw-Hill Co., New York, pp. 392 ± 363. (1993). Nature, 362, 847 ± 849. Nigro JM, Baker SJ, Preisinger AC, Jessup JM, Hostetter R, Ludwig T, Chapman DL, Papaioannou VE and Efstratiadis Cleary K, Bigner SH, Davidson N, Baylin S and Devilee P. A. (1997). Genes Dev., 11, 1226 ± 1241. (1989). Nature, 342, 705 ± 708. Lukas J, Parry D, Aagaard L, Mann DJ, Bartkova J, Strauss NishishoI,NakamuraY,MiyoshiY,MikiY,AndoH,Horii M, Peters G and Bartek J. (1995). Nature, 375, 503 ± 506. A, Koyama K, Utsunomiya J, Baba S and Hedge P. Luongo C, Moser AR, Gledhill S and Dove WF. (1994). (1991). Science, 253, 665 ± 669. Cancer Res., 54, 5947 ± 5952. Nobori T, Miura K, Wu DJ, Lois A, Takabayashi K and MacCollin, M & Gusella, JF. (1998). The genetic basis of Carson DA. (1994). Nature, 368, 753 ± 756. human cancer. Vogelstein B & Kinzler K. (eds). McGraw- Ohh M, Yauch RL, Lonergan KM, Whaley JM, Stemmer- Hill Co., New York, pp. 492 ± 454. Rachamimov AO, Louis DN, Gavin BJ, Kley N, Kaelin MacPhee M, Chepenik KP, Liddell RA, Nelson KK, WGJ and Iliopoulos O. (1998). Mol. Cell, 1, 959 ± 968. Siracusa LD and Buchberg AM. (1995). Cell, 81, 957 ± Oshima M, Oshima H, Kitigawa K, Kobayashi M, Itakura C 966. and Taketo MM. (1995). Proc. Natl. Acad. Sci. USA, 92, Malkin D, Li FP, Strong LC, Fraumeni JFJ, Nelson CE, 4482 ± 4486. KimDH,KasselJ,GrykaMA,Bischo€FZandTainsky Owen-Schaub LB, Zhang W, Cusack JC, Angelo LS, Santee MA. (1990). Science, 250, 1233 ± 1238. SM, Fujiwara T, Roth JA, Deisseroth AB, Zhang WW Malkin D. (1994). Biochim. Biophys. Acta, 1198, 197 ± 213. and Kruzel E. (1995). Mol. Cell Biol., 15, 3032 ± 3040. Malkin, D. (1998). The genetic basis of human cancer. Palmiter RD and Brinster RL. (1986). Annu.Rev.Genet.,20, Vogelstein B & Kinzler K. (eds). McGraw-Hill Co., New 465 ± 499. York, pp. 393 ± 408. Pan H and Griep AE. (1994). Genes Dev., 8, 1285 ± 1299. Mao L, Merlo A, Bedi G, Shapiro GI, Edwards CD, Rollins Pan W, Sun T, Hoess R and Grafstrom R. (1998). BJ and Sidransky D. (1995). Cancer Res., 55, 2995 ± 2997. Carcinogenesis, 19, 765 ± 769. Marcus JN, Watson P, Page DL, Narod SA, Lenoir GM, Parada LF, Land H, Weinberg RA, Wolf D and Rotter V. Tonin P, Linder-Stephenson L, Salerno G, Conway TA (1984). Nature, 312, 649 ± 651. and Lynch HT. (1996). Cancer, 77, 697 ± 709. Patel KJ, Vu VP, Lee H, Corcoran A, Thistlethwaite FC, Masuda H, Miller C, Koe‚er HP, Battifora H and Cline MJ. Evans MJ, Colledge WH, Friedman LS, Ponder BA and (1987). Proc. Natl. Acad. Sci. USA, 84, 7716 ± 7719. Venkitaraman AR. (1998). Mol. Cell, 1, 347 ± 357. McClatchey AI, Saotome I, Ramesh V, Gusella JF and Jacks PauseA,LeeS,WorrellRA,ChenDY,BurgessWH, T. (1997). Genes Dev., 11, 1250 ± 1260. Linehan WM and Klausner RD. (1997). Proc. Natl. Acad. McClatchey AI, Saotome I, Mercer K, Crowley D, Gusella Sci. USA, 94, 2156 ± 2161. JF, Bronson RT and Jacks T. (1998). Genes Dev., 12, Pause A, Lee S, Lonergan KM and Klausner RD. (1998). 1121 ± 1133. Proc. Natl. Acad. Sci. USA, 95, 993 ± 998. Medema RH, Herrera RE, Lam F and Weinberg RA. (1995). Pelletier J, Bruening W, Li FP, Haber DA, Glaser T and Proc. Natl. Acad. Sci. USA, 92, 6289 ± 6293. Housman DE. (1991a). Nature, 353, 431 ± 434. MercerWE,ShieldsMT,AminM,SauveGJ,AppellaE, Pelletier J, Schalling M, Buckler AJ, Rogers A, Haber DA Romano JW and Ullrich SJ. (1990). Proc. Natl. Acad. Sci. and Housman D. (1991b). Genes Dev., 5, 1345 ± 1356. USA, 87, 6166 ± 6170. Pomerantz J, Schreiber-Agus N, Liegeois NJ, Silverman A, Miki Y, Swenson J, Shatuck-Eidens D, Futreal PA, Harsh- Alland L, Chin L, Potes J, Chen K, Orlow I, Lee HW, manK,TavtigianS,LiuQ,CochranC,BennettLM,Ding Cordon-Cardo C and DePinho RA. (1998). Cell, 92, 713 ± W, Bell R, Rosenthal J, Hussey C, Tran T, McClure M, 723. Frye C, Hattier T, Phelps R, Haugen-Strano A, Katcher H, Powell SM, Zilz N, Beazer-Barclay Y, Bryan TM, Hamilton Yakumo K, Gholami Z, Sha€er D, Stone S, Bayer S, Wray SR, Thibodeau SN, Vogelstein B and Kinzler KW. (1992). C, Bogden R, Dyananth P, Ward J, Tonin P, Narod S, Nature, 359, 235 ± 237. Bristow PK, Norris FH, Helvering L, Morrison P, Rosteck Pritchard-Jones K, Fleming S, Davidson D, Bickmore W, P, Lai M, Barrett JC, Lewis C, Neuhausen S, Cannon- Porteous D, Gosden C, Bard J, Buckler A, Pelletier J, AlbrightL,GodgarD,WisemanR,KambAandSkolnick Housman D. et al. (1990). Nature, 346, 194 ± 197. MH. (1994). Science, 266, 66 ± 71. Prives C. (1998). Cell, 95, 5±8. Milner J, Ponder B, Hughes-Davies L, Seltmann M and Purdie CA, Harrison DJ, Peter A, Dobbie L, White S, Howie Kouzarides T. (1997). Nature, 386, 772 ± 773. SEM, Salter DM, Bird CC, Wyllie AH, Hooper ML and MiyashitaT,KrajewskiS,KrajewskaM,WangHG,Lin Clarke AR. (1994). Oncogene, 9, 603 ± 609. HK, Liebermann DA, Ho€man B and Reed JC. (1994). Quelle DE, Zindy F, Ashmun RA and Sherr CJ. (1995). Cell, Oncogene, 9, 1799 ± 1805. 83, 993 ± 1000. Miyashita T and Reed JC. (1995). Cell, 80, 293 ± 299. Rasheed BK, Stenzel TT, McLendon RE, Parsons R, Moll UM and Schramm LM. (1998). Crit.Rev.OralBiol. Friedman AH, Friedman HS, Bigner DD and Bigner Med., 9, 23 ± 37. SH. (1997). Cancer Res., 57, 4187 ± 4190. Morris JF, Madden SL, Tournay OE, Cook DM, Sukhatme Rhei E, Kang L, Bogomolniy F, Federici MG, Borgen PI and VP and Rauscher FJ. (1991). Oncogene, 6, 2339 ± 2348. Boyd J. (1997). Cancer Res., 57, 3657 ± 3659. Moser AR, Pitot HC and Dove WF. (1990). Science, 247, RouleauGA,MerelP,LutchmanM,SansonM,ZucmanJ, 322 ± 324. Marineau C, Hoang-Xuan K, Demczuk S, Desmaze C, Moser AR, Mattes EM, Dove WF, Lindstrom MJ, Haag JD Plougastel B. et al. (1993). Nature, 363, 515 ± 521. and Gould MN. (1993). Proc. Natl. Acad. Sci. USA, 90, Rubinfeld B, Souza B, Albert I, Munemitsu S and Polakis P. 8977 ± 8981. (1995). J. Biol. Chem., 270, 5549 ± 5555. Munemitsu S, Souza B, Muller O, Albert I, Rubinfeld B and Rubinfeld B, Albert I, Por®ri E, Fiol C, Munemitsu S and Polakis P. (1994). Cancer Res., 54, 3676 ± 3681. Polakis P. (1996). Science, 272, 1023 ± 1026. Munemitsu S, Albert I, Souza B, Rubinfeld B and Polakis P. Sah VP, Attardi L, Mulligan GJ, Williams BO, Bronson RT (1995). Proc. Natl. Acad. Sci. USA, 92, 3046 ± 3050. and Jacks T. (1995). Nat. Genet., 10, 175 ± 180. Nathke IS, Adams CL, Polakis P, Sellin JH and Nelson WJ. (1996). J. Cell Biol., 134, 165 ± 179. Mouse tumor suppressor models N Ghebranious and LA Donehower 3400 Schutte M, da Costa LT, Hahn SA, Moskaluk C, Hoque AT, Tashiro H, Blazes MS, Wu R, Cho KR, Bose S, Wang SI, Li Rozenblum E, Weinstein CL, Bittner M, Meltzer PS, J, Parsons R and Ellenson LH. (1997). Cancer Res., 57, Trent JM. et al. (1995). Proc. Natl. Acad. Sci. USA, 92, 3935 ± 3940. 5950 ± 5954. Taverna D, Ullman-Cullere M, Rayburn H, Bronson RT and Scully R, Chen J, Plug A, Xiao Y, Weaver D, Feunteum J, Hynes RO. (1998). Cancer Res., 58, 848 ± 853. Ashley T and Livingston DM. (1997). Cell, 88, 265 ± 275. TengDH,BogdenR,MitchellJ,BaumgardM,BellR, Serova O, Montagna M, Torchard D, Narod SA, Tonin P, Berrys,DavisT,HaPC,KehrerR,JammulapatiS,Chen Sylla B, Lynch HT, Feunteun J and Lenoir GM. (1996). Q, Ot K, Skolnick MH, Tavtigian SV, Jhanwar S, Am.J.Hum.Genet.,58, 42 ± 51. Swedlund B, Wong AK and Kamb A. (1996). Nat. Genet., Serrano M, Hannon GJ and Beach D. (1993). Nature, 366, 13, 241 ± 244. 704 ± 707. Ton CC, Hu€ V, Call KM, Cohn S, Strong LC, Housman Serrano M, Gomez-Lahoz E, DePinho RA, Beach D and DE and Saunders GF. (1991). Genomics, 10, 293 ± 297. Bar-Sagi D. (1995). Science, 267, 249 ± 252. Tsai KY, Hu Y, Macleod KF, Crowley D, Yamasaki L and Serrano M, Lee HW, Chin L, Cordon-Cardo C, Beach D and Jacks T. (1998). Mol. Cell, 2, 293 ± 304. DePinho RA. (1996). Cell, 85, 27 ± 37. Tsukada T, Tomooka Y, Takai S, Ueda Y, Nishikawa S, Sharan SK, Morimatsu M, Albrecht U, Lim DS, Regel E, Yagi T, Tokunaga T, Takeda N, Suda Y, Abe S, Matsuo I, Dinh C, Sands A, Eichele G, Hasty P and Bradley A. Ikawa I and Aizawa S. (1993). Oncogene, 8, 3313 ± 3322. (1997). Nature, 386, 804 ± 810. van Heyningen V, Bickmore WA, Seawright A, Fletcher JM, ShibataH,ToyamaK,ShioyaH,ItoM,HirotaM, Maule J, Fekete G, Gessler M, Bruns GA, Huerre- Hasegawa S, Matsumoto H, Takano H, Akiyama T, Jeanpierre C, Junien C. et al. (1990). Proc. Natl. Acad. Toyoshima K, Kanamaru R, Kanegae Y, Saito I, Sci. USA, 87, 5383 ± 5386. Nakamura Y, Shiba K and Noda T. (1997). Science, Varley JM, McGown G, Thorncroft M, Santibanez-Koref 278, 120 ± 123. MF, Kelsey AM, Tricker KJ, Evans DGR and Birch JM. Shoemaker AR, Moser AR, Midgley CA, Clipson L, Newton (1997). Cancer Res., 57, 3245 ± 3252. MA and Dove WF. (1998). Proc. Natl. Acad. Sci. USA, 95, Venkatachalam S and Donehower LA. (1998). Mutat. Res., 10826 ± 10831. 400, 391 ± 407. Smith SA, Easton DF, Evans DG and Ponder BA. (1992). Venkatachalam S, Shi YP, Jones SN, Vogel H, Bradley A, Nat. Genet., 2, 128 ± 131. Pinkel D and Donehower LA. (1998). EMBO J., 17, Somasudaram K, Zhang H, Zheng Y-X, Houvras Y, Peng Y, 4657 ± 4667. Wu GS, Licht JD, Weber BL and El-Deiry WS. (1997). Viskochil D, Buchberg AM, Xu G, Cawthon RM, Stevens J, Nature, 389, 187 ± 190. Wol€ RK, Culver M, Carey JC, Copeland NG, Jenkins Srivastava S, Zou Z, Pirollo K, Blattner W and Chang EH. NA. et al. (1990). Cell, 62, 187 ± 192. (1990). Nature, 348, 747 ± 749. Wallace MR, Marchuk DA, Andersen LB, Letcher R, Odeh Stambolic V, Suzuki A, de la Pompa JL, Brothers GM, HM, Saulino AM, Fountain JW, Brereton A, Nicholson J, MirtsosC,SasakiT,RulandJ,PenningerJM,Siderovski Mitchell AL. et al. (1990). Science, 249, 181 ± 186. DP and Mak TW. (1998). Cell, 95, 29 ± 39. Weinberg RA. (1995). Cell, 81, 323 ± 330. Stevens LC and Little CC. (1954) Proc. Natl. Acad. Sci. USA, Williams BO, Remington L, Albert Daniel M., Mukai S, 40, 1080 ± 1087. Bronson RT and Jacks T. (1994). Nat. Genet., 7, 480 ± 484. Stone S, Jiang P, Dayananth P, Tavtigian SV, Katcher H, Windle JJ, Albert DA, O'Brien JM, Marcus DM, Disteche Parry D, Peters G and Kamb A. (1995). Cancer Res., 55, CM, Bernards R and Mellon PL. (1990). Nature, 343, 2988 ± 2994. 665 ± 668. StottFJ,BatesS,JamesMC,McConnellBB,StarborgM, Wong MH, Hermiston ML, Syder AJ and Gordon JI. (1996). Brookes S, Palmero I, Ryan K, Hara E, Vousden KH and Proc. Natl. Acad. Sci. USA, 93, 9588 ± 9593. Peters G. (1998). EMBO J., 17, 5001 ± 5014. WuGS,BurnsTF,McDonaldER3,JiangW,MengR, Stratton MR and Wooster R. (1996). Curr. Opin. Genet. Krantz ID, Kao G, Gan DD, Zhou JY, Muschel R, Dev., 6, 93 ± 97. Hamilton SR, Spinner NB, Markowitz S, Wu G and El- Su LK, Kinzler KW, Vogelstein B, Preisinger AC, Moser Deiry WS. (1997). Nat. Genet., 17, 141 ± 143. AR, Luongo C, Gould KA and Dove WF. (1992). Science, Xiong Y, Zhang H and Beach D. (1993). Genes Dev., 7, 256, 668 ± 670. 1572 ± 1583. Su LK, Vogelstein B and Kinzler KW. (1993). Science, 262, Yin L, Ghebranious N, Chakraborty S, Sheehan CE, Ilic Z 1734 ± 1737. and Sell S. (1998). Hepatol., 27, 73 ± 80. Suzuki A, De la Pompa J, Hakem R, Elia A, Yoshida R, Mo Yin Y, Tainsky MA, Bischo€ FZ, Strong LC and Wahl GM. R, Nishina H, Chuang T, Wakeham A, Itie A, Koo W, (1992). Cell, 70, 937 ± 948. Billia P, Ho A, Fukumoto M, Hui CC and Mak TW. Yonish-Rouach E, Resnitzky D, Lotem J, Sachs L, Kimchi A (1997). Genes Dev., 11, 1242 ± 1252. and Oren M. (1991). Nature, 352, 345 ± 347. Symonds H, Krall L, Remington L, Saenz-Robles M, Lowe Zhang Y, Xiong Y and Yarbrough WG. (1998). Cell, 92, S, Jacks T and Van Dyke T. (1994). Cell, 78, 703 ± 711. 725 ± 734. Szabo CI and King MC. (1995). Hum. Mol. Genet., 4, 1811 ± Zonana J, Rimoin DL and Davis DC. (1976). J. Paediatr., 1817. 89, 600 ± 603. Tamura M, Matsumoto K, Aota S, Parsons R and Yamada KM. (1998). Science, 280, 1614 ± 1617.