Genome-Wide Modifier Screens:
Forward Genetic Approach to identify genes
involved in cancer development
Genetics 144,
Oncogenomics (Winter 2005, Dr. Charles Brenner)
By Michael Chen
Overview
Genetic, environmental, and stochastic factors are the trinity of cancer etiology. Thanks to advance in gene targeting technique in mice, reverse genetic approaches have been very successful in identifying genes involved in cancer development. However, as many cancer related genes are also involved in development. Embryonic lethality is common when genes are homozygously knocked-out in germline. Site-specific recombinase mediated conditional alleles are usually employed to bypass the embryonic lethality. Nevertheless, it is still relatively difficult to identify moderate or low penetrance alleles by reverse genetic approach because multiple hits would be required for tumor development. Forward genetic approach has been rapidly evolving to identify modifier loci involved in cancer development.
The complex traits are a result of multiple genetic determinants and each individual locus contributes a small effect to the overall phenotype. These quantitative trait loci (QTLs) may affect quantitative phenotype differently. QTL mapping provides the means to identify chromosome regions that contain genes work together to give a complex trait. Genetically identical inbred mice are valuable resource for identifying genetic factors influencing these traits. As discussed by Siracusa and co-workers (Ref 1), different inbred strains are first tested for susceptibility to certain cancer phenotype. The two strains that exhibit the extreme phenotype are usually subjected for further study. Intercrosses or backcrosses between these strains coupled with genotypic and phenotypic analysis of the offspring can identify the sites of chromosomal regions encoding genes responsible for differences in cancer susceptibility. Once the candidate genes are identified, reverse genetic approach could be used to confirm and identify the biochemical pathway. Another approach is by mutagenesis screen. Individual animals are subjected to physical or chemical mutagenesis and score for cancer incidence. Again, intercrosses or backcrosses are performed to determine the location of new mutations. Similar reverse genetic approaches are employed for further confirmation.
This report will take a paper published by Mao, et al., entitled: Fbxw7/Cdc4 is a p53-dependent, haploinsufficient tumor suppressor gene (Nature 432, p775, 2004, Ref 2), as an example to demonstrate how forward genetic approach could be utilized to identify cancer modifier genes.
Introduction
The p53 tumor suppressor protein has been found mutated or deleted in more than 50% of all human cancers. In cancers without p53 mutations, p53 function can still be attenuated by cytoplasmic sequestration or degradation by other oncoproteins. The major functions of p53 are summarized in Slide1. Following DNA damage, p53 can function as a transcriptional activator to result in cell cycle arrest, apoptosis and negative regulation of p53 itself. The p53 protein can also down-regulate the expression of growth promoting genes, such as Bcl-2 and cyclins. The p53 protein can also be a cofactor in a number of DNA repair pathways (Ref 3).
The p53 protein is a key regulator in cell cycle checkpoints. Ionization radiation (IR) induced DNA damage activates the PI3K-like kinases, ATM and ATR. ATM can phosphorylate p53 directly or via Chk2. ATR has also been implicated in p53 phosphorylation. ATM can also phosphorylate Mdm2, a p53 negative regulator, and inhibit the interaction between Mdm2 and p53, leading to an accumulation of p53 protein levels. One the major transcriptional targets of p53 is p21WAFgene, which encode an inhibitor of cyclin-dependent kinase. The p21WAF protein can bind to cdk4/CyclinD and cdk2/CyclinE complexes and result in a cycle block in late G1 and early S phase respectively, and maintain the Rb protein in a hypo-phosphorylated state bound to E2F. The p53 protein can also up-regulate 14-3-3s gene. 14-3-3 family members can bind and sequester cdc25 protein phosphotases, such as cdc25C, whose activity is required for the G2 to M phase transition. Chk1 is a downstream target of ATR. To bind 14-3-3s, cdc25C has to be phorphorylated first by Chk1. Chk1 also phosphorylate and activate Wee1 protein. Wee1 can maintain the cdc2/cyclin B complex in an inactive form, and prevents cell cycle into mitosis (Slide 2) (Ref 3).
In clinics, the loss of p53 function due to mutation or degradation can potentially alter the response of tumor to radiotherapy. The impact could be summarized as the "five R" of radiotherapy: radiosensitivity, redistribution, repair, reoxygenation and repopuplation (Slide 3) (Ref 3). For radiosensity, cells lost p53 will result in decrease in apoptosis and terminal arrest. For redistribution, p53 deficient cells will show loss of checkpoint control and increase of S phase population. For repair, p53 deficient cells will show decreased in genetic stability. For Reoxygenation, p53 deficient cells show higher resistance to hypoxial condition. For repopulation, p53 deficient cells show increase in proliferation and decrease in growth factor dependency.
Experimental Rational
Homozygous p53 null mice usually develop
lymphomas and sarcomas but rarely epithelial tumors. Heterozygous mice with one
functional p53 allele develop tumors slower than their homozygous null
counterparts. One possible reason is the second hit is required for tumor to
develop. If the second hit is another p53 in the heterozygous mice, the overall
pattern of genetic changes should be similar to p53 homozygous null mice.
However, if the primary second hit is on another gene rather than p53, we would
expect to find certain genetic changes in tumors from p53 heterozygous mice
only. To test this, mice were bred as shown in the following breeding scheme
and subjected to ionizing radiation for tumor induction.
Mutagenesis
Screens and determine chromosomal and sub-chromosomal locations of mutations
The spectrum of genetic alterations in lymphomas from F1 p53 +/-, F1 backcross p53+/-, and F1 backcross p53-/- were compared base on frequency of allelic deletion or imbalance. Chromosome 3, 16, and 18 showed p53-dependent LOH, whereas no difference was seen in LOH frequency on chromosome 12 and 19 (Ref 2, Fig 1a). For example, almost all tumors from F1 p53 +/- showed LOH, whereas only about 50% tumors from F1 backcross p53+/- showed of LOH and only 10% of tumors from F1 backcross p53-/- showed LOH. Sub-chromosomal losses were detected in the interval between D3Mit139 and D3MA38 on chromosome 3. Several candidate genes implicated in tumorigenesis including Mts1, TrkA, and Fbxw7/Cdc4. However, no inactivating mutations found in the coding region of Mts1 from tumors showed LOH. Analysis of 85 additional radiation-induced lymphomas from (C57BL/6J X 129/Sv) F1 (p53+/- or p53-/-) defined the LOH breakpoint at the distal region of chromosome 3 between D3MA44 and D3MA57. Only Fbxw7 gene is in this region (Ref2, Fig. 1b).
Identification of modifier genes
The mouse Fbxw7 coding region
sequence shows three spliced forms (a, b, and g) (Ref
2, S Fig 2a). Fbxw7 protein is a member of the F-box protein family (Ref
4, Fig 1). Mammalian Fbxw7 protein contains one F-box motif (F) and seven
WD40 repeats (WD). The interrelationships among the F-box protein family
members are shown in the phylogenetic tree (Ref
4, Fig 2). Fbxw7 protein is an ubiquitin ligase implicated in the control
of chromosome stability (Ref
5). Fbxw7
(hCDC4) mutations have been found in colorectal cancers and adenomas (Ref
4, Fig 1). Two tumors from the p53+/- mice showed mutations in WD40 domains (Ref
2, Fig. 1c).
Confirmation by
transgenesis:
The relationship
between p53 and Fbxw7
A 10-kilobase promoter region of Fbxw7gene was searched for p53 DNA-response elements. Total nine putative
p53 DNA-responsive elements were identified. To test if the activation of p53
by radiation can induce the expression of Fbxw7. Mouse embryonic fibroblasts (MEFs) are
subjected to g-radiation then assay for Fbxw7expression. As shown in the figure, Fbxw7 is
expressed in MEFs derived from p53+/+ and p53+/- but p53-/- mice after g-radiation
(Ref
2, Fig 2a). Down-regulation of Fbxw7 by siRNA results in growth
advantage in p53+/- MEFs, but not in p53-/- MEFs (Ref
2, Fig 2cde).
Gene/protein expression profiling and targeting biochemical pathways:
Down-stream targets of Fbxw7
Notch4 levels were increased when Fbxw7 was downregulated
by siRNA in the presence of wild-type p53, but not cyclin E, Norch1 and
Presenilin1 (Ref
2, Fig 2f). Wild-type MEFs treated with Fbxw7 siRNA showed increase
in the percentage of aneuploid cells by FACS analysis, indicating increased
genomic instability (Ref
2, S Fig3a). Aurora-A, a
mitotic control proteins, and c-Jun, are both up-regulated when MEFs were
treated with Fbxw7 siRNA in the presence of wild-type p53 (Ref
2, Fig 2gh).
Haploinsufficiency
MEFs from Fbxw7+/- and Fbxw7-/- both showed increase in
levels of Aurora-A and Notch4, indicating that Fbxw7 has the properties
of a haploinsufficient tumor suppressor gene (Ref
2, Fig 2f). Gama-irradiated Fbxw7+/- mice showed accelerated
tumorgenesis when compared with wild-type mice (Ref
2, Fig 3).
Tissue specificity
Irradiated p53 +/- mice develop primarily lymphomas, but almost
never develop epithelial tumors. The double heterozygous p53+/- Fbxw7+/-
mice developed epithelial lung and liver tumors and a range of lesions in the
ovary (Ref
2, Table1, Fig
4). Other mouse models of cancer that result in genetic instability also
have given rise to epithelial tumors. It is possible that development of
epithelial tumor might require certain level of genetic instability. It also
suggested that Fbxw7 might involve in the protection of epithelial cell in the
consequence of DNA damage leading to genetic instability. The role of Fbxw7 in
mediating p53 dependent DNA damage responses is shown in the figure.
Conclusion
Activation of p53 by DNA damage might induce
the expression of Fbxw7. Upregulation of Fbxw7 levels might cause downregulation of
Aurora-A, Cyclin E and c-Jun, as well as other targets, which will lead to
growth arrest at various points in the cell cycle. Loss of p53 causes failure
to activate certain genes, including Fbxw7, in response to DNA damage.
Genetic alterations involving Fbxw7are not observed in p53deficient
mouse tumors. (Ref
2, S Fig 5b)
.
Forward genetics in mammalian is relatively difficult when compare with other genetic model organism, such as Drosophila or zebrafish, because of the number of offspring and generation time. However, genome-wide microsatellite markers have accelerated genomic mapping of novel genes involved in cancer development. We can foresee that more candidate genes could be identified by this methodology.
Reference:
1.
Siracusa L.D., Silverman, K.A., Koratkar, R., Markova, M., Buchberg, A.M. Ed.
by Brenner, C and Duggan, and D. Oncogenomics:
Molecular Approach to Cancer. Ch12,
255-90 John Wiley & Sons, Inc (2004)
2.
Mao, J.H., Perez-Losada, J., Wu, D., Delrosario, R., Tsunematsu, R., Nakayama,
K.I., Brown, K., Bryson, S., & Balmain, A. Fbxw7/Cdc4 is a p53-dependent,
haploinsufficient tumour suppressor gene.
Nature
432,775-9 (2004).
3.
Cuddihy, A.R. & Bristow, R.G. The p53 protein family and radiation
sensitivity: Yes or no? Cancer Metastasis Rev. 23, 237-57 (2004).
4.
Jin, J., Cardozo, T., Lovering, R.C., Elledge, S.J., Pagano, M., & Harper,
J.W. Systematic analysis and nomenclature of mammalian F-box proteins. Genes Dev.18, 2573-80 (2004).
5.
Rajagopalan, H., Jallepalli, P.V., Rago, C., Velculescu, V.E., Kinzler, K.W.,
Vogelstein, B., & Lengauer, C. Inactivation of hCDC4 can cause chromosomal
instability. Nature 428,77-81 (2004).