Mouse Models of Cancer

Identification of the first low penetrance tumor modifier genes

One of the major areas of research in the Mouse Models of Cancer Theme within the Genetics Program is the identification of low penetrance tumor modifier loci which can be identified using interspecific crossing between resistant and sensitive strains of mice. These mouse model systems provide a powerful tool to analyze these complex genetic traits that led to the formation of the Complex Trait Consortium that provides oversight for the use of mouse models in the analysis of cancer predisposition.

Drs Demant, Nagase and Elliott are key members of the internationally recognized group of scientists focusing on this approach (Abiola et al., Nat Rev Genet 2003; 4:911; Churchill et al., Nat Genet 2004; 36:1133). Dr Demant has been a pioneer in developing the systems to identify low penetrance modifier genes over the past decade (Demant, Nat Rev Genet 2003; 4: 721; Demant, Radiat Res 2005; 164:462). Work by Dr Demant’s group has localized > 50 novel susceptibility loci for lung or colon cancer using molecular mapping following interspecies breeding programs (Demant, Nat Rev Genet 2003; 4:721).

Using positional cloning strategies one of these genes, Scc1 (Susceptibility to colon cancer 1) has been identified as encoding the protein tyrosine phosphatase receptor-type J, (PTPRJ), (Ruivenkamp et al., Oncogene 2003; 22:3472). The PTPRJ gene has since been shown to be involved in LOH in human colon, lung, and breast cancers (Ruivenkamp et al., Oncogene 2003:22: 3472; Ruivenkamp et al., Oncogene 2003; 22:7258) and a specific haplotype of this gene has been shown in a large case/control study to be associated with altered risk to sporadic (common) breast cancer  (Lesueur et al., Am. J Human Genet 2005; 14:2349), thus providing valuable proof-of-principle validation of studies in animal models for identification of susceptibility loci in human cancers.

The successful cloning of low penetrance modifier cancer genes by the Demant group was soon followed by other examples (Ewart-Toland et al., Nat Genet 2003; 34: 403; Zhang et al., Nat Genet 2003; 33:145). Following the seminal work of Dr Nagase, in collaboration with Dr Alan Balmain (UCSF), defining low penetrance modifiers in a carcinogen-induced skin cancer model they had established, this group demonstrated that the aurora kinase A gene was responsible for increased susceptibility in well characterized strains of mice (Ewart-Toland et al., Nat Genet 2003; 33:145).

A functional amino acid change in an aurora box of AURKA reveals loss of the binding ability of E2 ubiquitin conjugating enzyme, UBE2N, (Ewart-Toland et al., Nat Genet 2003; 34:403).  It is predicted that this mutation stabilizes protein that, in turn, affects chromosome stability thereby increasing the risk of cancer development.

In human populations specific haplotypes of this gene have been shown to confer an increased risk to breast cancers. In collaboration between Dr Nagase and Dr Alan Balmain (UCSF), the major skin tumor susceptibility locus, Skts1, was mapped to within a 0.9cM genetic interval on proximal chromosome 7 using linkage and congenic mouse analyses. This single tumor modifier locus was shown to have a dramatic effect on the allelic preference for imbalance on chromosome 7, with at least 90% of tumors from the congenic mice showing preferential gain of markers on the chromosome carrying the susceptibility variant.  Importantly, these alterations allowed a higher resolution localization of Skts1 than was possible using congenic mouse strains alone (de Koning et al., Oncogene 2007; 26(28):4171). 

Similarly, Dr Elliott has identified loci involved in susceptibility to carcinogen-induced colon cancer. Earlier work indicated that the susceptibility found in ICR/Ha mice includes loci on Chrs 5, 12, 13 and 14. Congenic strains including loci on each of these chromosomes on a C57BL/6 background are susceptible and provide the means to identify the genes involved.

An important area of the analysis of the functional effects of tumor susceptibility genes involves the regulation of the impact of native and adaptive immunity on tumorigenesis. The study of numerous functional parameters of macrophage activation by Dr Demant’s group has revealed that many of these genes are genetically linked to lung cancer susceptibility genes on chromosomes 4, 8, and 19 (Fijneman et al., Cancer Res 2004; 64:3458). Studies of the genetics of T lymphocyte activation indicated that several aspects of this process are also linked genetically to the same lung cancer susceptibility genes (Lipoldova et al., Int J Cancer 2005; 114:394).

Lung inflammation is an important factor supporting development of lung cancer. Therefore, in order to complement the above assays of systemic immune response, Dr Demant is analyzing the susceptibility of the mouse lung to chemical and allergic inflammation (Piavaux et al., Genes Immun 2007; 8:28), which will lead to the identification of predisposing genes at the organ level. Interestingly, it has been shown (Horlings and Demant, Exp Lung Res 2005; 5:513) that the presence of infiltrating lymphocytes in lung tumors is genetically controlled, and early indications are that this process involves both of the major subclasses of T lymphocytes.

It also appears that it is the host genes that determine the propensity for lymphocyte infiltration. These findings have two potential translational implications: Firstly, since the absence of lymphocyte infiltration in human tumors and the inability to mount lymphocyte infiltration may be one of the mechanisms for failure of immunotherapy, the detection of host genes that support or hinder this process may help select those patients who are more likely to respond to immunotherapy.

Secondly, the presence of infiltrating lymphocytes is an important positive prognostic factor in colon cancer and has been shown by others to be a better predictor than the standard pathological criteria. Similar findings were reported for melanomas and ovarian cancers. The definition of genes controlling the propensity of tumor infiltration by lymphocytes, therefore, will allow screening of individual patients to establish this constitutional propensity even before any tumor develops. 

In addition, Dr Demant is an active member of the Lung Disease Site Working Group (DSRG) where he is examining human bronchial epithelial tissues (dysplastic, metaplastic) as well as malignant lung tissue for a human gene counterpart to his murine models and collaborates with Drs Loewen and Reid (Cancer Prevention and Population Sciences Program).  This ability will potentially provide an important indication of the invasiveness of possible future cancers in an individual, thus providing important guidelines for determining the optimal individualized preventive and therapeutic strategies.

A novel resource for the generation of knockout mice

The NIH directed program to systematically generate knock out mice for all known genes has recently begun. Dr Yu, in collaboration with Dr Allan Bradley (Sanger Center, UK), pioneered the development of a mutagenic insertion and chromosome engineering resource (MICER) (Adams et al., Nat Genet 2004; 36:867) that will facilitate this effort. This resource includes ~100,000 ready-made targeting vectors which can be used to generate gene mutations as well as chromosomal rearrangements in mice, such as deletions, duplications, inversions, and translocations. These vectors have become a freely available public resource which has been provided to scientists world-wide for the functional analysis of mammalian genomes.

Recent work involving transgenic mice overexpressing the IL14a gene by Dr Cowell, in collaboration with Dr Ambrus (State University of New York at Buffalo) shows that these mice develop diffuse large cell lymphoma in aging animals (Shen et al., J Immunol 2006; 177:5676). However, when placed on a background of MYC oncogene over expression they develop lymphomas whose immunophenotype and cytology resemble the blastoid variant of mantle cell lymphoma (MCL). These tumors develop within 6 months and this system provides the only reported animal model for MCL (Ford et al., Blood 2007; 109(11):4899).

In addition, mouse models of hepatocellular carcinoma have been developed by Dr Demant following knockout of the Txnip gene (Sheth et al., Oncogene 2006; 25:3528).  Dr Cowell’s group have further shown that retroviral transfer of the ZNF198-FGFR1 fusion gene, which they discovered associated with myeloproliferative disease, results in the development of a myeloproliferative disease similar to that in humans when transplanted into sub-lethally irradiated host mice.  These mice also develop T-cell lymphomas, a phenotype consistent with the human disease as well. This mouse model provides another unique model for the study of leukemogenesis.

Chromosome engineering creates mice with megabase chromosome deletions/duplications

Gene knockouts and transgenic mice have greatly assisted the characterization of cancer genes where their involvement has been identified. In many cases, however, although the location of critical loci is known through the association with chromosome abnormalities, a specific candidate gene has not been identified. In some cases, increased dosage of specific genes is presumed to be involved rather than genetically determined by loss/gain of function.

To investigate this very important aspect of cancer development and predisposition, the science of ‘chromosome engineering’ has been developed, and one of the pioneers of this technically challenging and relatively recent science is Dr Yu, in collaboration with Dr Alan Bradley (Sanger Institute). In essence this technology can reproduce exact chromosome deletions and translocations in mice as found in the syntenic regions in humans. In this way, the role of structural chromosome changes in the development of cancer can be studied at physiological levels, and thus, chromosome engineering has become a powerful strategy to model human genomic rearrangements.

Work in Dr Yu’s laboratory has created a mouse model for the human chromosome deletion disorder involving 17q21.33-q23.2. Two rearrangements that represent large deletion and duplication events have been engineered in mice spanning 1.8-Mb and 6.9-Mb (Yu et al., Genetics 2006; 173: 297). The phenotypic analysis indicated that the 6.9-Mb deletion resulted in cardiovascular anomalies that are similar to those of human patients carrying deletions.

Annular pancreas in mouse carrying the 30 megabase duplication of human chromosome 21 created by chromosome engineering

In ongoing efforts using chromosome engineering technologies, (Li et al ., Hum Molec Genet 2007;16(11):1359-1366) Dr Yu, has now developed three individual duplications on mouse chromosomes 10, 11 and 16 which are syntenic to the entire human chromosome 21 and individually these mice have characteristic developmental defects of Down Syndrome patients. One of these duplications spans ~30 megabase pairs.

These mouse duplications have been sought after in the scientific community for more than a decade and no mouse lines with similar duplications has been reported to date. The breeding strategy to compound each of these duplications in a single mouse is underway. Down syndrome patients have a 500-fold increased risk of developing childhood leukemia which is related to the acquisition of mutations in the GATA1 transcription factor gene on the X chromosome. Traditional knockout strategies have been used to generate a mouse null for this gene in Dr Yu’s laboratory. In combination with the deletion strains, it will now be possible to characterize the genetic elements that are responsible for leukemia in these mice and provide a model for the specific disease to evaluate novel therapies.