| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Regular Articles |
McArdle Laboratory for Cancer Research, University of Wisconsin, Madison, Wisconsin 53706
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
60% compared with littermate controls. The identification of these independent loci within the Mcs1 COP allele provide a model of the genetic complexity of cancer. |
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
30%. Peto, however, points out that 30% is most likely an underestimate and defines the minimal percentage of the contribution of hereditary factors to the etiology of breast cancer (2)
. The inherited genetic component of breast cancer in a population consists of both highly penetrant genes at a low frequency in the population (e.g., BRCA1, BRCA2) and those that occur at a high frequency but have a low penetrance. The highly penetrant genes BRCA1 and BRCA2 account at most for 25% of the heritable genetic component of breast cancer, with the remainder likely attributable to the additive, dominant, and epistatic effects of low-penetrance genes. Breast cancer is thus a polygenic disease. Current risk models that include family history are limited in their predictive power. In a recent article, Pharoah et al. (3)
showed that in current predictive models, only 62% of cases could be predicted to occur in 50% of the population at highest risk, whereas 15% of breast cancer cases can be assigned to 10% of the population at highest risk. In contrast, it has been calculated that if we know half or all of the low- penetrance genes that control risk, we will be able to assign, respectively, 80 or 88% of all breast cancer cases to 50% of the population. Also, it is calculated if we have knowledge of only 50% of these genes, then we can assign 32% of cases to 10% of the women at highest risk. With knowledge of all low-penetrance genes, this estimate increases so that we could assign almost half of all cases to 10% of women at high risk (3)
. Thus, knowledge of such genes would vastly improve our ability to detect and prevent breast cancer.
Finding 50% of such low-penetrance genes using only existing means, which focus on human population studies using the approach of human genetic epidemiology, will be difficult. These methods generally focus on and test known genes that are involved in processes that might lead to breast cancer such as xenobiotic metabolizing enzymes (4)
and genes in DNA repair pathways (5)
. It would be difficult to directly identify unknown or unsuspected genes that lead to an altered level of susceptibility. This is especially true for genes that lead to breast cancer resistance. This is in part because of the fact that identifying a population as resistant to breast cancer cannot be readily accomplished for a disease such as breast cancer that occurs in 10% of a population. It is hard to distinguish between individuals with resistant genetics from those with good fortune.
One approach to identify low-penetrance genes that modulate breast cancer risk, especially those associated with resistance, is to initially use animal models to identify potential loci and genes and then translate this animal data to populations of women for comparative evaluation. In addition, the full characterization of animal models for breast cancer susceptibility will provide us with a better understanding of the degree of complexity of the polygenic etiology of breast cancer. We have begun the genetic dissection of the heritable susceptibility to breast cancer with two rat models that use the COP3 rat (6 , 7) and the Wistar-Kyoto rat (8) strains. Both rats are resistant to the induction of breast cancer. Using linkage-based mapping studies against a susceptible WF strain, we genetically identified four QTLs in each resistant rat strain and termed them Mcs1–8. Each strain has three QTLs that contribute to resistance and one that increases sensitivity to mammary cancer. Interestingly, only one QTL in each strain overlaps a QTL at the same chromosomal location in the other strain, i.e., only two total QTLs in common (COP Mcs2 and WKy Mcs6). There thus appear to be many independent rat QTLs that contribute to the sensitivity to mammary cancer. Here, we extend these studies by first physically confirming the prediction of a mammary cancer resistance QTL, Mcs1, in the COP rat by producing and characterizing congenic rats containing the Mcs1 COP allele. In addition, we extend our study of the complexity of the polygenic component of breast cancer by fine-mapping Mcs1 to determine that this locus contains more than one gene that contributes to mammary cancer susceptibility.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Genotype Analysis.
A small tail section from each rat was removed at 1–4 weeks of age. DNA was extracted (Gentra, Minneapolis, MN), diluted, and PCR amplified using polymorphic microsatellite markers. Rats were genotyped using multiple microsatellite markers spanning 20–30 cM of the Mcs1 locus from D2Mit29 to D2Rat201 on the centromeric end of chromosome 2. In addition, rats at the N2 through N6 generations were also screened for the presence of the COP allele on chromosomes 1, 7, and 8, which were previously shown to contain the Mcs2, Mcs3, and Mcs4 QTLs affecting mammary tumor development (7)
. Heterozygous carriers were selected to specifically contain the chromosome 2 Mcs1 regions of interest while eliminating the COP alleles at the other QTLs. PCR reactions were performed in 96-well plates, combining 2.5 µl of diluted DNA (50–100 ng) with 2.5 µl of PCR master mix. The nonradioactive PCR master mix consisted of 1x PCR buffer [10 mM Tris-HCl (pH 8.3), 1.5 mM MgCl2, 50 mM KCl, and 0.001% gelatin], 250 µM deoxynucleotide triphosphates, 132 nM each primer, and 0.2 units of AmpliTaq polymerase (Applied Biosystems, Foster City, CA). The final concentrations in the radioactive PCR reaction mixture were 1x PCR buffer; 200 µM dCTP, dTTP, and dGTP, 20 µM dATP, 0.12 µM of each primer, 0.25 units of AmpliTaq DNA Polymerase, and 0.025 µl of [
-32P]dATP (NEN, Boston, MA, specific activity 3000 Ci/mmol, 10 mCi/ml). Each reaction underwent an initial denaturation of 94°C for 5 min followed by 35–40 cycles of 94°C for 1 min, 55°C for 1 min, 72°C for 30 s, and a final extension of 72°C for 5 min. PCR products were resolved on either Metaphor agarose (nonradioactive protocol) or acrylamide sequencing (32P protocol) gels and imaged using either a FluorImager or PhosphorImager (Amersham Biosciences, Piscataway, NJ), respectively.
Phenotype Analysis.
Female rats, at 50–55 days of age, were given a single oral gavage of 65 mg DMBA/kg rat body weight. The DMBA (Acros, Pittsburgh, PA) was suspended in sesame oil at a concentration of 20 mg/ml, heated in a boiling water bath until dissolved, and cooled to room temperature. At 15 weeks after DMBA administration, all rats were necropsied, and mammary carcinomas were identified by location and size. The total number of mammary carcinomas > 3 x 3 mm in diameter that developed in each rat was enumerated. Spleen sections were also removed and frozen for an additional source of DNA in case confirmation of genotype was needed. Statview (SAS Institute, Cary, NC) statistical analysis software was used and data regarding carcinoma development are presented as means ± SE. ANOVA was used when multiple generations for a congenic line existed to test for significant effects of generation on tumor multiplicity. Unpaired t tests were used to compare genotypes within congenic lines/strains.
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Congenic line B was shown to lack all other known COP mammary carcinoma susceptibility loci (Mcs2–4) by microsatellite marker genotyping within these loci (Table 1)
. This line was phenotyped for DMBA-induced mammary cancer induction at the N6 and N9 generations; at each generation, three genotypes at Mcs1 were tested: COP homozygous; WF/COP heterozygous; and WF homozygous. In addition, the parental WF and COP strains were also phenotyped. COP control rats (n = 16) developed an average of 0.25 ± 0.1 carcinomas/rat, whereas the WF control rats (n = 30) had 8.1 ± 0.7 carcinomas/rat at week 15 after DMBA treatment. The results of phenotyping are shown in Table 2
. For congenic line B, statistical analysis using ANOVA showed no significant difference between the N6F1 and N9F1 generations (P = 0.91), allowing us to pool these data for analysis. It can be seen that after averaging over both generations there was an effect of genotype (P < 0.0001). COP-homozygous Mcs1-congenic rats developed the least number of carcinomas, averaging 1.2 ± 0.6 carcinomas/rat (n = 5, P < 0.001), whereas WF/COP-heterozygous rats had an average of 3.0 ± 0.4 carcinomas/rat (n = 29, P < 0.0001). The WF-homozygous rats developed an average of 8.5 ± 1.1 carcinomas/rat (n = 8) that was not different from WF control rats. A congenic strain of line B, WF.COP-D2Mit29/D2Rat201, was made through brother-sister mating of COP-homozygous N10 rats. Phenotype analysis of this Mcs1-congenic strain revealed a strong resistance phenotype similar to that observed in the earlier congenic generations, with 1.5 ± 0.3 carcinomas developing/rat (n = 19) in the COP-homozygous rats and WF control rats developing an average of 6.3 ± 0.5 carcinomas/rat (n = 29; Table 2
).
|
|
|
. When these five recombinant lines were phenotyped, it became apparent that the Mcs1 locus in itself was complex. COP-homozygous rats from three of the nonoverlapping lines (Q, QQ, and T) showed resistance to DMBA-induced mammary tumorigenesis, whereas congenic rats from the other two lines (W and K) developed as many tumors as WF-homozygous littermate rats (Table 3)
. By analyzing these data, we identified three subloci that each independently contributed resistance to the induction of mammary carcinomas within the Mcs1 region as defined by congenic strain line B. They are termed in order of identification as Mcs1a, Mcs1b, and Mcs1c (Fig. 1)
.
|
2 cM), was phenotyped in the N10 generation, COP-homozygous rats developed an average of 3.3 ± 0.4 carcinomas/rat (n = 38, P < 0.0001), whereas WF/COP heterozygous rats had an average of 5.0 ± 0.4 carcinomas (n = 33, P = 0.0018), both statistically significant from WF-homozygous littermate rats (7.6 ± 0.7, n = 20). This genomic region of Mcs1a was additionally reduced to the distal marker D2Uwm14 after phenotyping congenic line W. Congenic line W did not confer a resistance phenotype with these COP-homozygous and WF/COP-heterozygous rats developing an average of 7.2 ± 1.0 and 6.8 ± 0.8 carcinomas/rat, respectively. The production of a congenic strain for line Q, WF.COP-D2Mit29/D2Uwm13 at the N12 generation verified the Mcs1a resistance phenotype with these COP-homozygous rats developing 3.6 ± 0.4 carcinomas/rat (n = 37, P < 0.0001) compared with WF controls with 6.3 ± 0.5 (n = 29; Table 3
).
The region of Mcs1c (6 cM) was defined by congenic line QQ with the proximal marker D2Rat2 and distal marker D2M13Mit286. Rats produced from the N10 carriers that had a COP-homozygous genotype at Mcs1c developed 3.0 ± 0.6 carcinomas/rat (n = 11), which was significantly less than WF-homozygous littermate females (7.1 ± 0.7, n = 19, P = 0.0003). The WF/COP-heterozygous females yielded an average carcinoma/rat of 5.6 ± 1.0 (n = 17) that, although intermediate between the COP- and WF-homozygous littermates, was not significantly different from either of these groups.
Mcs1b was defined proximally by marker D2Uwm17 and distally by D2Rat16, a region of 13 cM. Line K, which overlaps line T minimally at the genetic marker D2Uwm17, did not show a resistance phenotype because COP-homozygous rats developed 8.8 ± 1.3 carcinomas/rat (n = 8). In contrast, COP-homozygous rats of the N9 and N10 generations for line T developed 3.5 ± 0.5 carcinomas/rat (n = 21, P < 0.0001). The heterozygous females in this region averaged 7.6 ± 0.8 mammary carcinomas/rat (n = 18), which was not statistically different from the WF-homozygous littermate data (8.3 ± 0.8, n = 18) but was significantly different (P < 0.0001) than the line T COP-homozygous rats.
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
47% reduction in the number of DMBA-induced mammary carcinomas developing while having two COP alleles at Mcs1 would result in a 74% reduction in the absence of the Mcs-2, Mcs-3, Mcs-4 QTLs (7)
. If the congenic results presented in Table 2
are used to provide an estimate of the effects of the Mcs1 COP allele in its heterozygous and homozygous forms versus littermate WF-homozygous rats, one observes 1.2, 3.0, and 8.5 carcinomas/rat for rats, which carry two, one, and zero copies of the COP allele at Mcs1, respectively. Thus, rats with two copies show a statistically significant reduction of 85% compared with rats with zero copies while having one copy also produces a significant reduction of 65% compared with having no copies. The actual reductions because of having one or two copies of a COP allele at the Mcs1 QTL approximates the predicted degree of resistance from the Poisson regression model (7)
. The congenic strain for line B data showed a 76% reduction compared with WF control rats, additionally confirming the predicted value of 74% tumor reduction in rats with two copies of the Mcs1 COP allele.
A statistically significant difference also exists between the average number of carcinomas developing in the Mcs1 COP-homozygous and the WF/COP-heterozygous rats. This observation supports the hypothesis of the presence of a semidominant Mcs1 gene that yields a phenotype of mammary cancer resistance. Alternatively, the Mcs1 locus, which covers 30 cM, might contain multiple genes that together produce a high degree of resistance to breast cancer. We addressed this alternative hypothesis by collecting recombinant lines within the Mcs1 interval. These recombinant rats served as founders of new congenic lines. Rats in these lines were phenotyped for their sensitivity to DMBA-induced mammary cancer. Interestingly, we found several subloci, each defined by nonoverlapping genetic markers that acted independently to confer resistance to DMBA-induced mammary carcinogenesis. These subloci were termed Mcs1a, Mcs1b, and Mcs1c.
Mcs1a was defined by congenic lines Q and W. COP-homozygous rats at the Mcs1a locus (line Q) had a 57% reduction in the development of mammary carcinomas/rat as littermates that were homozygous for the WF allele. An intermediate number of carcinomas were found in the heterozygotes, which was significantly different from rats homozygous for the COP allele at this Mcs1a locus, making Mcs1a a semidominant locus with respect to the resistance phenotype and additive with respect to tumor multiplicity.
Mcs1c defined by congenic line QQ had a 58% reduction in the development of carcinomas/rat in rats homozygous for the COP allele compared with those littermates homozygous for the WF allele. There was no significant difference in carcinoma development between Mcs1c COP-homozygous and -heterozygous rats, nor was the heterozygous group different from the WF-homozygous littermates. Although it is not possible to make a definitive conclusion regarding the heterozygous female phenotype in this Mcs1c region, this is most likely because of insufficient power. It should be noted that a strong trend (P = 0.059) suggests the possibility that the Mcs1c COP allele acts in a semidominant manner with respect to the resistance phenotype.
Mcs1b defined by line T had a 58% reduction in carcinomas/rat in rats homozygous for the COP allele at this sublocus compared with littermates that were homozygous for the WF allele and 54% reduction compared with females that were heterozygous. Thus, one copy of the COP allele has no effect on tumor development, suggesting that the Msc1b is a recessive locus or the WF allele is completely dominant to the COP allele at this locus.
These data suggest that genetically identified QTLs can be complex, harboring multiple genes. It is likely that these genes act in the same direction to confer resistance and not an increase in sensitivity. If they did not act in the same direction, their combined effects and close proximity would likely have prevented the Mcs1 locus from being identified initially by linkage analysis. As it is, all three of the Mcs1 subloci/genes contribute to mammary cancer resistance. One can speculate that it is likely that many of the QTLs identified as components in multigenic disease may contain several genes acting in the same quantitative direction. A corollary to this speculation is that it would be difficult to identify QTLs in regions of the genome in which two or more genes are present that contribute similarly to the genetic etiology of a disease but act in opposing directions. The closer these genes are linked, the more difficult is their identification using genetic linkage analysis. The genetic identification of such individual genes in such a complex genetic environment may require moving from a linkage/microsatellite approach to an association/SNP3 methodology.
These data together with other published data regarding rat breast cancer susceptibility models help to delineate the complexity of the multigenic model of breast cancer. Using two rat strains resistant to mammary cancer, we have identified six loci that contribute to mammary cancer resistance and two to increased sensitivity (6, 7, 8)
. Using a new statistical approach, we are also beginning to identify loci that, while having no significant main effect on mammary cancer susceptibility (e.g., Mcsm1), do act to modify the actions of loci with main effects (8)
. The results presented here, together with our previously published results (8)
, begin to model the high level of genetic complexity underlying the multigenic disease of breast cancer. The Mcs1 locus was chosen for our first detailed characterization based on its small 1-LOD interval; however, it is interesting to note that only Mcs1a falls within this interval. This observation supports using a wide genetic interval (20–30 cm) surrounding the 1-LOD interval in producing initial congenic animals for any QTL under study.
The COP rat is resistant to spontaneous (9)
, hormonally induced (10, 11, 12)
, and directly acting (9
, 13)
carcinogen-induced cancers of the mammary gland. Our model of chemically induced mammary cancer uses the indirectly acting carcinogen DMBA, a synthetic polycyclic aromatic hydrocarbon. It was selected for both pragmatic and theoretical reasons over other possible mammary carcinogens, including the directly acting carcinogen N-nitroso-N-methylurea, ionizing radiation, as well as hormones such as estrogens or prolactin. Although all of these various carcinogens are indeed important, the main reason we chose DMBA is that it is a very efficient mammary carcinogen that produces 6–10 carcinomas/susceptible (WF) rat in these studies. In contrast, COP rats developed 0.25 carcinomas/rat. This large ratio of developing carcinomas not only allows us to distinguish between WF and COP rats but also allows us the ability to phenotype many loci and subloci that control susceptibility to mammary cancer in the COP rat. It has been shown that the COP and WF rat strains were equally able to activate DMBA and showed a similar mammary spectrum of DNA adducts (14
, 15)
. The pleiotropic effects of DMBA as a carcinogen make it more likely to identify many susceptibility genes than, for example, mammary cancer induced by the activation of a specific oncogene. In this context, it must be stressed that DMBA is a very poor activator of ras as compared with N-nitroso-N-methylurea (16, 17, 18, 19)
. Finally, DMBA-induced rat mammary carcinomas share morphological similarities with most common human breast carcinomas (20)
, and this model system is widely used for preclinical evaluations.
Comparative analysis of rat, human, and mouse bacterial artificial chromosome sequences has shown that all three genomes are collinear in the Mcs1 region. The region of rat chromosome 2 encompassing Mcs1a, Mcs1c, and Mcs1b is collinear with human chromosome 5q14-q12 (inverted orientation) and mouse chromosome 13C1 (same orientation). The genome sequence of these three species is near completion as the individual genome project groups continue to fill gaps and finalize the assemblies. Although there is no evidence, to date, that the human region homologous to Mcs1 is implicated in breast cancer, this could be attributable to the high complexity of this multigenic disease and the difficulty in identifying low-penetrance genes in the human population. The identification of the Mcs1 loci in the rat is thus likely to lead to the identification of novel and important genes involved in human breast cancer.
In conclusion, we have physically verified the existence and effect of the Mcs1 QTL. The identification of three independent resistance subloci within the Mcs1 allele demonstrates an additional layer of genetic complexity underlying mammary cancer that will likely extrapolate to breast cancer. These data will provide important mapping information that will be critical in positionally cloning multiple Mcs1 cancer susceptibility genes. Delineation of the function of such genes will hopefully translate into prevention and/or treatment targets for human breast cancers.
|
FOOTNOTES |
|---|
1 Supported by the NIH Grant CA28954. 
2 To whom requests for reprints should be addressed, at McArdle Laboratory for Cancer Research, University of Wisconsin, 1400 University Avenue, Madison, WI 53706.
3 The abbreviations used are: COP, Copenhagen; WF, Wistar Furth; QTL, quantitative trait locus; DMBA, 7,12-dimethylbenz(a)anthracene; LOD, logarithm of odds ratio; SNP, single nucleotide polymorphism.
Received 4/28/03. Revised 6/20/03. Accepted 6/30/03.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
)anthracene in the rat: a comparative analysis with human breast tumors. Arch. Pathol. Lab. Med., 126: 915-927, 2002.[Medline]This article has been cited by other articles:
|
|
 |
|
  D. J. Samuelson, B. A. Aperavich, J. D. Haag, and M. N. Gould Fine Mapping Reveals Multiple Loci and a Possible Epistatic Interaction within the Mammary Carcinoma Susceptibility Quantitative Trait Locus, Mcs5 Cancer Res., November 1, 2005; 65(21): 9637 - 9642. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
, indicates the presence of two COP alleles for congenic lines resulting in reduced mammary tumor development (one to four carcinomas/rat); 
indicates the presence of two COP alleles for congenic lines incapable of conferring resistance to tumor development (six to eight carcinomas/rat).
,indicate areas of unknown genotype because of recombination. Three independent regions of chromosome 2 were capable of conferring resistance to mammary cancer development in COP-homozygous congenic rats. These genomic regions are shown as Mcs1a, Mcs1b, and Mcs1c. The chromosome markers used to identify the congenic rats are listed to the far right of the figure with genetic distances in cM to the left of the marker names. The Mcs1 QTL 1-LOD interval is also shown between D2Mit29 and D2Uwm13.