PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of plosonePLoS OneView this ArticleSubmit to PLoSGet E-mail AlertsContact UsPublic Library of Science (PLoS)
 
PLoS ONE. 2009; 4(2): e4388.
Published online Feb 5, 2009. doi:  10.1371/journal.pone.0004388
PMCID: PMC2633613
Five Quantitative Trait Loci Control Radiation-Induced Adenoma Multiplicity in Mom1R ApcMin/+ Mice
Eiram Elahi,#1 Nirosha Suraweera,#1 Emmanouil Volikos,1 Jackie Haines,2 Natalie Brown,2 Gerovie Davidson,3 Mike Churchman,3 Mohammed Ilyas,4 Ian Tomlinson,5 and Andrew Silver1*
1Colorectal Cancer Genetics, Institute for Cell and Molecular Sciences, Barts and The London, Queen Mary's School of Medicine and Dentistry, London, United Kingdom
2HPA Radiation Protection Division, Chilton, Didcot, Oxon, United Kingdom
3Cancer Research UK Genotyping Facility, Clinical Pharmacology, University of Oxford, Radcliffe Infirmary, Oxford, United Kingdom
4School of Molecular Medical Sciences, Division of Pathology, Queen's Medical Centre, Nottingham, United Kingdom
5Molecular and Population Genetics Laboratory, London Research Institute, Cancer Research UK, London, United Kingdom
Sebastian D. Fugmann, Editor
National Institute on Aging, United States of America
#Contributed equally.
* E-mail: a.r.silver/at/qmul.ac.uk
Conceived and designed the experiments: EE NS MI IT AS. Performed the experiments: EE NS EV JH NB GD MC. Analyzed the data: EE NS EV. Contributed reagents/materials/analysis tools: EE NS. Wrote the paper: EE NS AS.
Received October 13, 2008; Accepted December 22, 2008.
Ionising radiation is a carcinogen capable of inducing tumours, including colorectal cancer, in both humans and animals. By backcrossing a recombinant line of ApcMin/+ mice to the inbred BALB/c mouse strain, which is unusually sensitive to radiation–induced tumour development, we obtained panels of 2Gy-irradiated and sham-irradiated N2 ApcMin/+ mice for genotyping with a genome-wide panel of microsatellites at ~15 cM density and phenotyping by counting adenomas in the small intestine. Interval and composite interval mapping along with permutation testing identified five significant susceptibility quantitative trait loci (QTLs) responsible for radiation induced tumour multiplicity in the small intestine. These were defined as Mom (Modifier of Min) radiation-induced polyposis (Mrip1-5) on chromosome 2 (log of odds, LOD 2.8, p = 0.0003), two regions within chromosome 5 (LOD 5.2, p<0.00001, 6.2, p<0.00001) and two regions within chromosome 16 respectively (LOD 4.1, p  = 4×10−5, 4.8, p<0.00001). Suggestive QTLs were found for sham-irradiated mice on chromosomes 3, 6 and 13 (LOD 1.7, 1.5 and 2.0 respectively; p<0.005). Genes containing BALB/c specific non-synonymous polymorphisms were identified within Mrip regions and prediction programming used to locate potentially functional polymorphisms. Our study locates the QTL regions responsible for increased radiation-induced intestinal tumorigenesis in ApcMin/+ mice and identifies candidate genes with predicted functional polymorphisms that are involved in spindle checkpoint and chromosomal stability (Bub1b, Casc5, and Bub1), DNA repair (Recc1 and Prkdc) or inflammation (Duox2, Itgb2l and Cxcl5). Our study demonstrates use of in silico analysis in candidate gene identification as a way of reducing large-scale backcross breeding programmes.
The genetic dissection of complex traits is a difficult and important challenge. A number of resistant and susceptibility quantitative trait loci (QTLs) for tumours arising in many tissues (lung, colon, skin and haemopoietic system) have been mapped using mice (1, 2). Unfortunately, relatively few QTLs have been translated into critical polymorphisms within genes that can be demonstrated to show a clear alteration in the risk of tumour development. The lack of a systematic approach to the selection of genes within QTLs coupled with extensive breeding programmes to define relatively small intervals without taking into account clusters of interacting genes has contributed to this disappointing return. The search of genes within QTLs has recently benefited from the advances in the mammalian genome research, particularly with regard to mouse strain-specific sequence databases allowing in silico discovery of gene-coding and gene-regulatory variants (3). This approach combined with bioinformatics, including software to interrogate polymorphisms for functional effects provide an effective method for discovering the genes behind multigenic diseases such as cancer.
By far the majority of tumours used for QTL mapping purposes have been induced using high doses of chemical carcinogens and mutagens (2). This raises the possibility that a number of the QTLs will reflect genes involved in xenobiotic metabolism and clearance from the body rather than bona fide modifiers of carcinogenesis. Furthermore, the majority of the general human population are unlikely to be exposed to the chemicals used in these studies outside the context of occupational exposure. Consequently, it remains to be seen whether any QTLs identified through chemical tumorigenesis studies have homologues relevant to the appearance of cancer in humans. A complementary approach is to use rodent models where the primary genetic event leading to tumorigenesis is already known, and then to map modifiers of disease severity through selective breeding. The best example of this rationale involves the use of the multiple intestinal neoplasia (Min) mice, the first mouse model system involving the adenomatous polyposis coli (Apc) gene. ApcMin/+ mice develop numerous intestinal adenomas and are a good model of human familial adenomatous polyposis (FAP). The model has provided examples of modifying loci (called modifiers of Min, Mom) in mice, demonstrating the principle of genetic modulation of disease severity represented by, in these cases, intestinal adenoma multiplicity. To date, the chromosomal positions of Mom1-3, Mom6 and Mom7 have been reported (4) and functional polymorphisms identified for Mom1 and 2 (5).
Ionizing radiation is a well-characterized carcinogen capable of inducing tumours, including colorectal cancer (CRC), in both humans and animals (6). All humans are exposed to radiation as a consequence of environmental background; occupational and medical radiation sources provide additional hazards. For example, some 6,000 patients undergo pelvic radiotherapy each year and many show inflammatory responses in the gut, a known risk factor for CRC (7). Using a recombinant line of ApcMin/+ mice and the BALB/c mouse strain, which is unusually sensitive to ionising radiation–induced tissue damage and tumour development, we have previously shown that adenoma multiplicity in irradiated ApcMin/+ mice can be modified by two QTLs, which map to chromosome 16 segments from BALB/c (8). We have now undertaken the first genome-wide scan for additional loci modifying radiation-induced tumour multiplicity in the intestine of N2 backcross Min mice. The successful identification of candidate genes for QTLs using a combination of in silico and bioinformatics demonstrates the effectiveness of our approach.
Identification of main effect QTLs controlling adenoma multiplicity
We have conducted a genome-wide study involving microsatellite genotyping of constitutional DNA from panels of 2 Gy x-irradiated (84 males and 58 females) and sham-irradiated (60 males and 51 females) N2 ApcMin/+ mice. Each N2 mouse was genotyped using a total of 112 microsatellites with an average interval space between markers of 15 cM based on Kosambi's distances created by the MapManager QTX programme. Following our earlier observations on ApcMin/+ mice that the upper part of the small intestine (duodenum and jejunum) was more sensitive than the lower (ileum) region to radiation-induced adenoma formation (8) and that there may be modifiers that influence tumour multiplicity in the different parts, animals were assessed for adenoma multiplicity in the upper (USI) and lower small intestine (LSI) at time of sacrifice (all mice >80 days; Mean: 0 Gy, 241 days; 2 Gy, 180 days). The data for USI and LSI were analysed separately. Interval mapping (IM) and permutation testing (PT) were used to determine threshold values for each trait and an overall threshold value of 12.6 likelihood ratio statistic (LRS) was calculated from over 18 separate PTs (P<0.05); empirical p values were calculated for the observed LRS values (Table 1).
Table 1
Table 1
Main effect QTLs for tumour multiplicity in the small intestine of mice identified using a panel of 142 N2 irradiated (2Gy) mice.
In addition to the previously reported QTLs on chromosome 16 (p<0.0001; 8), we found two highly significant QTLs (p<0.0001; Table 1; Fig. 1) on chromosome 5, one significant QTL on chromosome 2, (p<0.0001; Table 1; Fig. 1) and one suggestive QTL on chromosome 17 for USI trait (p<0.001; Table 1). The two QTLs on chromosome 5 explained 15% and 19% of the total trait variance, the two QTLs on chromosome 16 accounted for 11% and 14% and chromosome 2 explained 9%. Overall, these significant QTLs of the USI account for 68% trait variation and requires further investigation. Only suggestive QTLs were identified for the LSI on chromosomes 1, 6, 7, 8 and 11 (p<0.01; Table 1). The 0 Gy dataset yielded only suggestive QTLs on chromosomes 3 and 13 for the USI trait and on chromosome 6 for the LSI (Table 2). We define the most significant QTLs (Table 1 and Fig. 1) as Mom radiation-induced polyposis (Mrip) loci. To confirm the existence of two independent USI QTLs on chromosome 5, mice were genotyped using an additional marker D5Mit139 (69.0 cM) and the IM and PT re-run. This revealed a substantial QTL effect (LRS of 28.8; P<0.0001) at the beginning of the interval defined by D5mit139 and the next proximal marker, D5Mit188 (64 cM). This locus retained its strong association (highly significant at p<0.0001) following further PT using 10,000 scans (chromosomes 1–19). To confirm results from MapManager, we repeated IM analysis for the irradiated cohort of mice using a compatible format Cartographer QTL Version 1.17 (9,10). PT (1,000 tests; 2cM window) with all markers for each of the four traits for 2Gy dataset provided similar LRS threshold values as MapManager. The only exception was D2Mit395 (LOD 2.7), which showed borderline significance when analysed using MapManager, but with Cartographer the marker fell just below the significance threshold of LOD 2.8.
Figure 1
Figure 1
A genome wide scan to show the LOD profiles of chromosomes 1–19 with adenoma multiplicity in the USI of the 2 Gy cohort.
Table 2
Table 2
Main effect QTLs for tumour multiplicity in the small intestine of mice identified using a panel of 112 N2 sham-irradiated mice.
Composite Interval Mapping (CIM; (11,12)) was used to assess any background or ‘ghost’ effects associated with the multiple QTLs found on both chromosome 5 and 16. A test window of 30 cM was selected based on genetic distance between the two QTLs on chromosome 5 (D5Mit201 and D5Mit139; 30 cM) and on 16 (Prkdc and D16Mit189; 20 cM). A maximum of five significant markers outside of the test window, were selected as background controls and fitted to a forward and reverse regression models as coefficients to control for genetic variation (13). The number of markers selected was sufficient to account for background QTL effects, and both IM and CIM were equally efficient in detecting QTLs with large effects on chromosomes 5 and 16. When compared with CIM, IM tended to have the highest power of detecting QTLs at D5Mit201, D5Mit139, and at Prkdc; there was no difference between mapping methods for D16Mit189. Importantly, no false QTLs were detected under the threshold values, and both IM and CIM gave similar likelihood findings. These analyses confirmed that the two peaks observed on chromosomes 5 and on 16 are independent QTLs with strong effects (Fig. 2 and and33 respectively). We were now able to confirm Mrips as five independent loci (Mrip1–5; Table 1; Fig. 13)3) influencing adenoma multiplicity in the USI of Min recombinant mice.
Figure 2
Figure 2
The LRS profile of Mrip2 and Mrip3 against markers distributed along chromosome 5 is determined using CIM (composite interval mapping) and compared with IM (interval mapping).
Figure 3
Figure 3
The LRS profile of Mrip4 and Mrip5 against markers distributed along chromosome 16 is determined using CIM (composite interval mapping) and compared with IM (interval mapping).
Selection of genes within QTL regions
We focussed on the five most significant main effect QTLs, Mrip1–5. Gene lists were drawn up for the chromosomal segment up and down stream (±15 cM) of the most likely position for each QTL (Table 3). The markers for each peak were: for chromosome 2, D2Mit395 (Mrip1); for chromosome 5, D5Mit139 (Mrip2) and D5Mit201 (Mrip3); for chromosome 16, D16Mit189 (Mrip4) and Prkdc (Mrip5). Interestingly, Mrip5 contained the Ritls loci associated with radiation induced thymic lymphoma (14) and the Mrip1 region encompasses the Scc (Colon tumour susceptibility) 1 and 2 loci associated with colon cancer susceptibility (15). For all QTLs combined, a total of 373 genes containing coding non-synonymous single nucleotide polymorphism (CnSNPs; n = 678) were identified between the two inbred strains using the MPD database. Olfactory genes (n = 166) containing 138 CnSNPs were found within the Mrips, but were excluded from further analysis because of an assumption of lack of functional relevance. Genes from this list was further selected based on their involvement in tumorigenesis, inflammatory response and radiation effects (Table 3). By using these processes we selected a number of genes for sequence verification in the BALB and B6Min/+ line. In total, 14 genes from Mrip1, 14 genes from both Mrip2 and Mrip3, and 11 genes from Mrip4 and Mrip5 were chosen for verification by sequencing (Table 3).
Table 3
Table 3
Summary of the process used in selecting genes from QTL regions.
Predicting functional mutations in selected genes using in silico analysis
Mrip1
A combination of bioinformatics programmes was used to identify possible functional mutations for the CnSNPs (Table 3). We identified possibly damaging mutations in six genes within the Mrip1 region: Duox2, Bubr1, Mertk, Dll4, Casc5 and Bub1 (Table 4). Three BALB specific single nucleotide polymorphisms (SNPs) were found in the Duox2 gene: H627R has only partial conservation among species (36%), but was predicted to be a damaging mutation according to Polyphen with a position specific independent count (PSIC) of 1.524. The polymorphisms, S553G and N378S were located within the extra cellular, N terminal heam peroxidase domain, although both were not predicted to be damaging substitutions.
Table 4
Table 4
Predictive assessment of BALB specific amino acid polymorphisms for potentially damaging alterations.
Two CnSNPs within Dll4 were observed, including a substitution at G627S predicted to be damaging according to Polyphen analysis due to a significant hydrophobic substitution effect (PSIC score = 1.516). This substitution causes an uncharged polar molecule (S) to change into a non-polar molecule (G) at position 627 according to SIFT. This amino acid was not evolutionarily conserved (6/16) and did not belong to any known functional domain. Although the Q419L alteration was within the 3ed EGF-like domain, the mutation itself was predicted to not cause any functional alterations to the protein's activity (Table 4).
Five CnSNPs were detected within Bub1b, three of which (S240G, P278T and V312A) are within a Mad2 dependent, Cdc20 binding domain necessary for effective cell cycle checkpoint functions (16). With regard to the other two CnSNPs, both SIFT and Polyphen analysis indicated that all these substitutions, although within functional domains, were tolerated. Casc5 and Bub1 polymorphisms (T1285I and P152S, respectively) were predicted to be deleterious (PSIC>1.5), although none were within any known functional domains.
Mrip2 and Mrip3
Possible functional CnSNPs were identified for Centd1, Recc1 and Cxcl5. The Centd1, and Cxcl5 polymorphisms (R278 and R10H, respectively) were found in regions of the proteins that were highly conserved between different vertebrates whilst the N66Y polymorphism from Recc1 was only partially conserved at 79%. Although these regions in which these polymorphisms were found did not form part of any known functional domains, the R278P, R10H and Y66N were not tolerated substitutions according to Polyphen predictions (PISC = 1.582). The change from a basic molecule, R, to the non-polar molecule, P, in the R278P Centd1 polymorphism was predicted ‘probably damaging’ (PSIC = 2.270) and ‘intolerant’ using Polyphen and SIFT respectively. The N66Y polymorphism of Recc1 was also predicted ‘probably damaging’ (PSIC = 2.44) using Polyphen, but ‘tolerant’ according to SIFT. All the protein prediction analyses indicated that these substitutions could be functionally important.
Mrip4 and Mrip5
Prkdc is the most likely candidate for Mrip4 (8,14,17,18,19). Of the two identified Prkdc polymorphisms, R2140C involves the substitution of a basic molecule (R) to an uncharged polar molecule (C) which was predicted as damaging (PSIC = 2.495) using Polyphen and intolerant according to SIFT analysis. The amino acid sequence is highly conserved with invertebrates and forms part of the Nuc194 domain involved in DNA repair (Table 4). Within the Mrip5 region the Itgb2l gene on chromosome 16 has two substitutions, D509N and G539S, and both polymorphisms were considered to be deleterious according to Polyphen (PSIC = 2.193 and 1.628, respectively). We also found that the D509 amino acid residue was conserved in all vertebrates (100%; 11/11) for which sequence data was available (Table 4).
In this study of a recombinant ApcMin/+ BALB/cByJ mouse model we have identified five significant QTLs associated with intestinal tumour multiplicity in the USI. The sham irradiated cohort of mice yielded only suggestive QTLs for the USI and the LSI trait. Of the three suggestive spontaneous QTLs none map to chromosome 18 suggesting that Mom2, 3 and 7 are not involved in determining adenoma multiplicity in this genetic background or only have very weak influences that would require substantially more mice to investigate. Also, these loci may not have a role in this study because of the genetic background used. For instance, Mom2 is a spontaneous mutation not carried by the mice used here and the recessive allele conferring susceptibility to severe disease (Mom3) is not present. There are a number of alleles of Mom7 segregating amongst mouse strains, but unfortunately BALB has not been investigated to date (4); our data would suggest that this loci has no, or a very weak, influence on adenoma multiplicity in the mice used here. Of the main effects radiation associated QTLs investigated in this paper, Mrip2 and Mrip3 showed the most significant association with tumour multiplicity followed by Mrip1, Mrip4 and Mrip5. In silico and predictive methodologies were used to identify possible genes underlying these QTLs. With advances in whole genome databases, and the availability of near complete sequence data on different strains of mice, it is now possible to use in silico data analysis to identify the genes that map within delineated chromosomal intervals and to find any variations in amino acids sequence in candidate genes (20). Hence, QTL mapping in conjunction with in silico tools, including subsequent protein predictive programming to assign a functional role, can provide a powerful and relatively robust means of identifying susceptibility genes in multigene diseases such as tumorigenesis (Tables 4 and and55).
Table 5
Table 5
The function of candidate genes and their role in radiation response and tumorigenesis.
Our approach has identified ten candidate genes, most of which are involved in spindle checkpoint and chromosomal stability (Bub1b, Casc5, and Bub1), DNA repair (Recc1 and Prkdc) or inflammation (Duox2, Itgb2l and Cxcl5; Table 5). Dll4 is a key player in the Notch signalling pathway known to regulate tumour angiogenesis, as does Cxcl5, and growth (Table 5). The role of Centd1 is not understood, although the protein functions down stream of RhoA to regulate focal adhesion dynamics suggesting an involvement in invasion and metastasis in cancer (21). Our results might suggest other roles for this protein in tumorigenesis and further functional tests.
From our study it is evident that QTLs can be systematically searched to identify possible candidate genes responsible for disease by using a combination of in silico methods, although it should be cautioned that gene selection at any given time would be based on genes already known or suspected to be involved in tumorigenesis. In addition to CnSNPs and SSNPs studied in this paper there are many other BALB specific SNPs within these QTL regions. Some of these SNPs are situated within regulatory domains outside the gene coding regions such as promoters, enhances and untranslated regions (UTR)s. These SNPs may ultimately influence the regulation of gene expression. ExpressionQTL (eQTL) analysis, an alternative in silico approach, provides a means of identifying regulatory SNPs associated with the expression of genes located within QTLs (22). We will employ eQTL analysis in the future to identify BALB specific regulatory SNPs associated with radiation induced tumorigenesis in the ApcMin/+ BALB/cByJ mouse model.
In summary, we have identified five novel QTLs controlling susceptibility to radiation-induced polyposis in ApcMin/+ mice. By using an effective in silico approach, we were able to identify potential candidate genes within these QTLs. The functional influence of these polymorphisms within our candidate genes will require substantial in vivo and in vitro investigations.
Animals, adenoma induction and Scoring
A recombinant inbred line (Line I) of ApcMin/+ mice on a C57BL/6 background (B6Min/+) that showed limited intraline variation in adenoma numbers was established and maintained as detailed previously (23). Female BALB/cByJ (BALB) mice were crossed with male B6Min/+ to yield F1 (N1) progeny. N1 male ApcMin/+ were then backcrossed to female BALB to produce N2 offspring that carry at least one copy of the BALB-derived modifier of Min (Mom1) dominant resistance allele (Mom1R). Irradiations and sham-irradiations were performed as given previously on N2 ApcMin/+ mice (8). Animals were housed in conventional cages with water, and standard maintenance diet, provided ad libitum. Mice were killed by CO2 asphyxiation when quality of life was compromised. Intestinal tracts were prepared for adenoma counting as detailed earlier (8,23). The small intestine was divided into two equal portions with the upper segment containing all of the duodenum and jejunum. The number of adenomas arising in the upper (USI) and lower small intestine (LSI) was determined at time of sacrifice. Given that numbers are unchanged All procedures involving mice were carried out in accordance with the United Kingdom Animals (Scientific Procedures) Act 1986 and with guidance from local ethics committees on animal experimentation.
Genotyping
Genotyping of ApcMin/+ was performed as given earlier (24). Genome-wide microsatellite analysis was conducted using standard methods and involved the following markers: chromosome (chr) 1, D1Mit430, 169, 132, 495, 17 and 292; chr2, D2Mit327, 395, 411, 229, 148, and 230; chr3, D3Mit178, 51, 320 and 352; chr 4, D4Mit227, 18, 193, 348, 308 and 256; chr5, D5Mit387, 352, 201, 157, 188, 168 and 409; chr6, D6Mit1, 83, 123, 284, 287, 194, 15 and 198; chr7, D7Mit178, 117, 228, 276, 323, 101 and 362; chr8, D8Mit155, 124, 289, 292, 45, 211, 213 and 49; chr9, D9Mit250, 90, 285, 336, 355, 347, 201 and 151; chr10, D10Mit189, 106, 194, 31, 42, 95, 233 and 103; chr11, D11Mit216, 339, 177 and 179 ; chr 12, D12Mit182, 91, 143 and 17; chr13, D13Mit115, 179, 142, 107, 260 and 78; chr14, D14Mit126, 127, 174, 60, 102 and 107; chr15, D15Mit100, 70 and 159; chr16, D16Mit182, 4, 5, 189 and 106; chr17, D17Mit143, 51, 180, 20 and 142; chr18, D18Mit222, 208, 49 and 4; chr19, D19Mit28, 106, 90, 103 and 6. Prkdc was genotyped as given in (19). All available irradiated and un-irradiated mice were genotyped.
Statistical analysis
The MapManager QTX program (25) was used to perform interval mapping (IM) and permutation testing (PT). Mapping of likeliest position of a QTL on a chromosome and the association with a given trait is determined by the statistic value described as the Likelihood Ratio Statistic (LRS) or the Log of Odds (LOD). The observed LRS was tested against the null hypothesis that the QTL did not occur by chance in the genome. Empirical significance thresholds (suggestive, significant and highly significant) were computed for each individual trait (USI and LSI) and empirical p values calculated for each observed LRS. IM analysis was also performed using Cartographer QTL (9). Composite interval mapping (CIM; (11,13,14) to investigate multiple significant peaks observed from IM was conducted in Cartographer.
Selection and interrogation of target genes
Following the identification of the physical positions of each QTL, a comprehensive list of genes mapping to a region ±15 cM from the most likely position of the QTLs was collated using the Ensemble databases. These lists were checked for known tumour susceptibility loci, and searched for coding non-synonymous single nucleotide polymorphisms (CnSNP) between BALB and C57BL/6J using MPD database (26). All selected polymorphisms were verified by re-sequencing DNA taken from the mice used in the backcross. Sequencing was carried out on a 3100 ABI prism sequence analyser; primers and protocols are available on request.
Bioinformatic tools used to predict functional influence of non-synonymous polymorphisms
The functional influences of CnSNPs were studied using bioinformatics software, including Polyphen (27) and SIFT (28). Sequence alignment and conservation of the SNPs were analysed by EMBL-EBI and Polyphen. A Search Tool for the Retrieval of Interacting Genes or Proteins (STRING) was used to identify whether the CnSNPs belonged to known or predicted functional domains (29). These domains were also confirmed using scientific literature searches.
Footnotes
Competing Interests: The authors have declared that no competing interests exist.
Funding: Cancer Research UK funded this research. The Funder (Cancer Research UK) had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.
1. Demant P. Cancer susceptibility in the mouse: genetics, biology and implications for human cancer. Nat Rev Genet. 2003;4:721–734. [PubMed]
2. Dragani TA. 10 years of mouse cancer modifier loci: human relevance. Cancer Res. 2003;63:3011–3018. [PubMed]
3. Marshall KE, et al. In silico discovery of gene-coding variants in murine quantitative trait loci using strain-specific genome sequence databases. Genome Biol. 2002;3:RESEARCH0078. [PMC free article] [PubMed]
4. Kwong LN, et al. Identification of Mom7, a novel modifier of Apc(Min/+) on mouse chromosome 18. Genetics. 2007;176:1237–1244. [PubMed]
5. Silverman K, Korarkor R, Siracusa LD, Buchberg AM. Identification of the modifier of Min 2 (Mom2) locus, a new mutation that influences Apc-induced intestinal neoplasia. Genome Res. 2002;12:88–97. [PubMed]
6. United Nations Scientific Committee on Effects of Atomic Radiation (UN-SCEAR) Sources and Effects of Ionizing Radiation. Report to the General Assembly with Scientific Annexes. Official Record: Fifty-fifth session, Supplement No. 46 (A/55/46) New York: United Nations; 2000.
7. Turina M, Mulhall AM, Mahid SS, Yashar C, Galandiuk S. Frequency and surgical management of chronic complications related to pelvic radiation. Arch Surg. 2008;143:46–52. [PubMed]
8. Degg NL, et al. Adenoma multiplicity in irradiated ApcMin mice is modified by chromosome 16 segments from Balb/c. Cancer Res. 2003;63:2361–2363. [PubMed]
9. Basten CJ, Weir BS, Zeng ZB. Smith C, Gavora JS, Benkel B, Chesnais J, Fairfull W, Gibson JP, Kennedy BW, Burnside EB, editors. Zmap–a QTL cartographer. In Proceedings of the 5th World Congress on Genetics Applied to Livestock Production: Computing Strategies and Software, 1994. Volume 22, pages 65-66. Published by the Organizing Committee, 5th World Congress on Genetics Applied to Livestock Production, Guelph, Ontario, Canada.
10. Basten CJ, Weir BS, Zeng ZB. 2004. QTL Cartographer Version 1.17. Department of Statistics, North Carolina State University, Raleigh, NC.
11. Jansen RC, Stam P. High resolution of quantitative traits into multiple loci via interval mapping. Genetics. 1994;136:1447–1455. [PubMed]
12. Zeng ZB. Precision mapping of quantitative trait loci. Genetics. 1994;36:1457–1468. [PubMed]
13. Zeng ZB. Theoretical basis for separation of multiple linked gene effects in mapping quantitative trait loci. Proc Natl Acad Sci USA. 1993;90:10972–10976. [PubMed]
14. Mori N, Okumoto M, Yamate JA. Susceptibility locus for radiation lymphomagenesis on mouse chromosome 16. J Radiat Res. 2000;41:367–372. [PubMed]
15. Moen CJ, Groot PC, Hart AA, Snoek M, Demant P. Fine mapping of colon tumor susceptibility (Scc) genes in the mouse, different from the genes known to be somatically mutated in colon cancer. Proc Natl Acad Sci USA. 1996;93:1082–1086. [PubMed]
16. Davenport J, Harris LD, Goorha R. Spindle checkpoint function requires Mad2-dependent Cdc20 binding to the Mad3 homology domain of BubR1. Exp Cell Res. 2006;312:1831–1842. [PubMed]
17. Mori N, Matsumoto Y, Okumoto M, Suzuki N, Yamate J. Variations in Prkdc encoding the catalytic subunit of DNA-dependent protein kinase (DNA-PKcs) and susceptibility to radiation-induced apoptosis and lymphomagenesis. Oncogene. 2001;20:3609–3619. [PubMed]
18. Sipley JD, et al. Gene for the catalytic subunit of the human DNA-activated protein kinase maps to the site of the XRCC7 gene on chromosome 8. Proc Natl Acad Sci USA. 1995;92:7515–7519. [PubMed]
19. Yu K, et al. Elevated breast cancer in irradiated BALB/c associates with unique functional polymorphism of the Prkdc (DNA-dependent protein kinase catalytic subunit gene. Cancer Res. 2001;61:1820–1824. [PubMed]
20. Belknap JK, et al. QTL analysis and genomewide mutagenesis in mice: complementary genetic approaches to the dissection of complex traits. Behav Genet. 2001;31(1):5–15. [PubMed]
21. Geho DH, Bandle RW, Clair T, Liotta LA. Physiological mechanisms of tumour-cell invasion and migration. Physiology. 2005;20:194–200. [PubMed]
22. MacLaren EJ, Sikela JM. Cerebellar gene expression profiling and eQTL analysis in inbred mouse strains selected for ethanol sensitivity. Alcohol Clin Exp Res. 2005;29:1568–1579. [PubMed]
23. Haines J, et al. Genetic basis of variation in adenoma multiplicity in ApcMin/+ Mom1s mice. PNAS. 2005;102:2868–2873. [PubMed]
24. Luongo C, Moser AR, Gledhill S, Dove WF. Loss of Apc in intestinal adenomas from Min mice. Cancer Res. 1994;54:5947–5952. [PubMed]
25. Manly KF, Jr, Cudmore RH, Meer JM. MapManager QTX, cross-platform software for genetic mapping. Mamm Genome. 2001;12:930–932. [PubMed]
26. Bogue MA, Grubb SC, Maddatu TP, Bult CJ. Mouse Phenome Database (MPD). Nucleic Acids Res. 2007;35(Database issue):D643–649. [PMC free article] [PubMed]
27. Sunyaev S, Ramensky V, Bork P. Towards a structural basis of human non-synymous single nucleotide polymorphisms. Trends in Genet. 2000;16:198–200. [PubMed]
28. Ng GC, Henikoff S. Predicting deleterious amino acid substitutions. Genome Res. 2001;11:863–874. [PubMed]
29. von Mering C, et al. STRING 7 - recent developments in the integration and prediction of protein interactions. Nucleic Acids Res. 2007;35 (Database issue):D358–623. [PubMed]
30. Bedard K, Krause KH. The NOX family of ROS-generating NADPH oxidases: physiology and pathophysiology. Physiol Rev. 2007;87:245–313. [PubMed]
31. Luxen S, Belinsky SA, Knaus UG. Silencing of DUOX NADPH oxidases by promoter hypermethylation in lung cancer. Cancer Res. 2008;68:1037–1045. [PubMed]
32. Abal M, et al. APC inactivation associates with abnormal mitosis completion and concomitant BUB1B/MAD2L1 up-regulation. Gastroenterology. 2007;132:2448–2458. [PubMed]
33. Katoh M, Katoh M. Notch signaling in gastrointestinal tract. Int J Oncol. 2007;30:247–251. [PubMed]
34. Kiyomitsu T, Obuse C, Yanagida M. Human Blinkin/AF15q14 is required for chromosome alignment and the mitotic checkpoint through direct interaction with Bub1 and BubR1. Dev Cell. 2007;13:663–676. [PubMed]
35. Grady WM. Genomic instability and colon cancer. Cancer Metastasis Rev. 2004;23:11–27. [PubMed]
36. Yoon HY, et al. ARAP2 effects on the actin cytoskeleton are dependent on Arf6-specific GTPase-activating-protein activity and binding to RhoA-GTP. J Cell Sci. 2006;119:4650–4666. [PubMed]
37. Hashiguchi K, Matsumoto Y, Yasui A. Recruitment of DNA repair synthesis machinery to sites of DNA damage/repair in living human cells. Nucleic Acids Res. 2007;35:2913–2923. [PMC free article] [PubMed]
38. Baier PK, Eggstein S, Wolff-Vorbeck G, Baumgartner U, Hopt UT. Chemokines in human colorectal carcinoma. Anticancer Res. 2005;25:3581–3584. [PubMed]
39. Dimberg J, Dienus O, Löfgren S, Hugander A, Wågsäter D. Expression and gene polymorphisms of the chemokine CXCL5 in colorectal cancer patients. Int J Oncol. 2007;31:97–102. [PubMed]
40. Weterings E, Chen DJ. DNA-dependent protein kinase in nonhomologous end joining: a lock with multiple keys? J Cell Biol. 2007;179:183–186. [PMC free article] [PubMed]
41. Hojgaard A, et al. Altered localization of CXCL13 expressing cells in mice deficient in Pactolus following an inflammatory stimulus. Immunology. 2006;119:212–223. [PubMed]
Articles from PLoS ONE are provided here courtesy of
Public Library of Science