This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Samuelson, D. J. Articles by Gould, M. N. Search for Related Content PubMed PubMed Citation Articles by Samuelson, D. J. Articles by Gould, M. N.

Fine Mapping Reveals Multiple Loci and a Possible Epistatic Interaction within the Mammary Carcinoma Susceptibility Quantitative Trait Locus, Mcs5

McArdle Laboratory for Cancer Research, University of Wisconsin-Madison, Madison, Wisconsin

Requests for reprints: Michael N. Gould, McArdle Laboratory for Cancer Research, University of Wisconsin-Madison, 1400 University Avenue, Madison, WI 53706-1599. Phone: 608-263-6026; Fax: 608-262-2824; E-mail: gould{at}oncology.wisc.edu.

	Abstract

Top Abstract Introduction Materials and Methods Results and Discussion References

 
To identify high-frequency, low-penetrance breast cancer modifier genes, we have developed a rat genetic model that uses the Wistar-Kyoto (WKy) inbred strain, resistant to developing 7,12-dimethylbenz[a]anthracene–induced mammary carcinogenesis, as a congenic donor and the susceptible Wistar-Furth (WF) strain as the recipient. Here, data from congenic rat lines containing smaller WKy genomic intervals of the Mcs5 quantitative trait locus region are presented to fine map three independently acting Mcs5 subloci. WKy-homozygous females from congenic lines defining Mcs5a, Mcs5b, and Mcs5c averaged, respectively, 4.0 ± 0.4, 11.6 ± 0.6, and 3.5 ± 0.4 mammary carcinomas per rat. These phenotypic values are statistically different from the WF-homozygous phenotype value of 8.0 ± 0.4, which is the baseline phenotype used for these experiments. We identified a likely Mcs5a x Mcs5b epistatic interaction that results in masking the increased susceptibility effect of the Mcs5b WKy allele by the Mcs5a WKy allele. We also provide evidence for a Mcs5a x Mcs5c interaction that is synergistic to decrease mammary carcinoma susceptibility below the additive effects of WKy alleles at each locus independently. The Mcs5 subloci are currently localized to 1.0, 7.5, and 4.5 Mb of rat chromosome 5, and the orthologous regions are on human chromosome 9 and mouse chromosome 4. These loci will provide unbiased candidate gene loci for evaluation in human case-control association studies.

	Introduction

Top Abstract Introduction Materials and Methods Results and Discussion References

 
Inherited risk of developing breast cancer for most individuals is controlled by an undetermined number of risk alleles. Many of these are expected to have higher population frequency compared with the rare risk alleles of BRCA1 or BRCA2. The collective effect of the low-penetrance breast cancer modifier genes on population risk likely exceeds that of the highly penetrant breast cancer susceptibility alleles, BRCA1 or BRCA2. Estimates indicate that if all the breast cancer risk alleles were known, almost 90% of breast cancer cases could be assigned to 50% of the population of women at highest risk (1). Understandably, these predictions have generated much interest in identifying breast cancer modifier gene loci. Historically, most work to identify modifier genes has focused on biased selection of candidate gene loci for human association studies. For example, many candidate genes that function in cancer-related pathways (e.g., DNA repair genes) have been investigated over the last two decades with little progress toward positive association results that were verifiable in large population studies (2). We are using rodent/human comparative genomics to incorporate an unbiased method of selecting modifier candidate gene loci for human breast cancer case-control association studies. Our study uses the rat mammary carcinogenesis model, which we view as a good model for human breast cancer. Many inbred rat strains are available that vary dramatically in mammary cancer susceptibility. Our studies use congenic rat lines to develop genetic models of breast cancer susceptibility. For the study here, the recipient rat strain providing the base susceptibility phenotype is the Wistar-Furth (WF), which is highly sensitive to 7,12-dimethylbenz[a]anthracene (DMBA)–induced mammary carcinogenesis. The donor strain, resistant to DMBA-induced mammary carcinogenesis, is the Wistar-Kyoto (WKy) strain. Classic quantitative trait linkage analyses using crosses between the WF and either of two resistant strains have identified a total of eight quantitative trait loci (QTL; refs. 3, 4). These loci are termed mammary carcinoma susceptibility loci 1 to 8 (Mcs1-Mcs8). We previously reported the confirmation of the QTL that had the highest LOD score in a backcross linkage mapping analysis (4), Mcs5, and defined the locus to a 115-Mb region of chromosome 5 using congenic rat lines (5). Here, we continue our comparative genomic search for breast cancer modifier genes by fine mapping the Mcs5 locus and identifying multiple subloci within it that modify mammary cancer risk.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results and Discussion References

 
Inbred WF and WKy rats were obtained from Harlan Sprague-Dawley, Inc. (Indianapolis, IN) and fed Teklad lab blox chow and acidified water ad libitum. Rats were maintained in a 12-hour light/dark cycle in an Association for Assessment of Laboratory Animal Care–approved facility and all protocols were approved through the University of Wisconsin Medical School Animal Research Committee. Congenic lines were established and maintained as previously published (5). Congenics are defined as genetic lines developed on a WF genome and carrying the selected WKy alleles shown in Fig. 1. The congenic generation (number of backcrosses) used to determine each line phenotype and the approximate amount of unselected WKy sequence present at the respective line's generation are listed in Table 1. Primer information for the markers listed on the y axis of Fig. 1 is available from the Rat Genome Database,¹ except for the following markers with the respective primer pair sequences listed here:

• AU048474, 5'-GATCGCACAGCTGAAATTAAA and 5'-TCCTCCTGAGCAAAGAACCC
  
• gUwm47-10, 5'-GATAGGGACCATCCTAATCC and 5'-GCCAGAACAGAGTCTGAGTG
  
• gUwm41-6, 5'-AGCTCTTGAGCATTAGCAGT and 5'-CCTTTTCCCAAGAAAGAGAT
  
• gUwm40-18, 5'-GACTTAATGTGGGGAGTGAA and 5'-AGCACATATGGAGGTTTGAC
  
• gUwm45-5, 5'-CTAGAAAGGTGCTTTGGTTG and 5'-TCAGCTTCTCCTCCTTCC
  
• gUwm50-22, 5'-TGAGACCTGGAACTTATTGG and 5'-CAGAGCCTAAGGGATTCAG
  
• bUwm26-2, 5'-ACACTGAAGGTTGAGCTTGT and 5'-ACTCACCTCTGTTCATCCTG
  
• bUwm36-1, 5'-CCTGTGATTCCAACTCTAGG and 5'-GATTCCTCTGTTCTTCTGGA
  
• bUwm32-4, 5'-AGCAAGTTCTAGGACAACCA and 5'-CTGGGGAGTTTTTCCTTATT.
  

View larger version (20K): [in this window] [in a new window] [Download PPT slide]   Figure 1. Genetic map, QTL scan, and WF.WKy congenic line intervals defining the Mcs5 subloci on rat chromosome 5. Congenic lines containing specific WKy alleles spanning Mcs5 subloci and resistant to DMBA-induced mammary carcinogenesis (black vertical bars). Mammary carcinoma–sensitive congenic lines containing specific WKy alleles delimiting the Mcs5 subloci intervals are designated (white vertical bars). The superscript to the congenic line name (letter) is the immediate predecessor line the congenic line was derived from. Initially, there were three long lines (A, B, and C) established at the early generations of backcrossing to the WF strain. The maximum interval of each Mcs5 sublocus is marked by a gray vertical bar with genetic markers at the ends annotated by heavy dashed lines. Vertical axis, rat chromosome 5 genetic markers. The light dashed lines are markers at the ends of congenic lines that fall outside of Mcs5 subloci. Horizontal axis, LOD score from the original QTL analysis. The map scan associated with the corresponding marker intervals (solid wavy line) with the peak LOD score at 13. Approximately to scale relative to base pair distances.

Phenotypic values for Mcs5 subloci combinations were predicted using the applicable additive, dominance, and interaction effects that were determined empirically using heterozygous and homozygous females from congenic lines and F₁ offspring of congenic line crosses. The mean mammary carcinomas per rat of the WF-homozygous genotype for each experiment was used as the baseline phenotype; thus, the genetic components are expressed relative to the WF phenotype (i.e., no WKy alleles at any of the Mcs loci). The one-sample t test function of StatView was used to compare mammary carcinoma multiplicity statistics to the predicted phenotypic values. The final predictive model for the WKy Mcs5 QTL mammary carcinoma multiplicity phenotype was determined to be:

where Mcs5 is the mammary carcinoma susceptibility/rat phenotype, µ is the mean effect with no WKy alleles present at any of the Mcs loci (i.e., WF homozygous), A is the additive effect of the WKy allele relative to the WF-homozygous phenotype, D is the dominance deviation of the WKy allele relative to the additive genotypic value of the heterozygote, and I is the interaction between WKy alleles at Mcs5 subloci. The effects of WKy alleles on mammary carcinoma multiplicity were also estimated using the general linear model function with a Poisson regression in the R statistical computing environment (The R Foundation for Statistical Computing). Data used were mammary carcinoma multiplicities from the congenic lines that contained Mcs5 loci and crosses of these lines (lines C, F, O, Q, Y, and OxY).

	Results and Discussion

Top Abstract Introduction Materials and Methods Results and Discussion References

 
Figure 1 illustrates the Mcs5 QTL region of rat chromosome 5 that was defined by statistical analysis of backcross data and further characterized by the congenic rat lines WF.WKy-D5Wox7/D5Uwm37 (line C) and WF.WKy-D5Rat26/D5Uwm42 (line D; refs. 4, 5). Line C females with two WKy alleles at the Mcs5 locus (WKy homozygous) developed an average of 1.3 ± 0.2 mammary carcinomas per rat (n = 43), which was less (P < 0.0001) than females that inherited two WF alleles at the Mcs5 locus (WF homozygous) and averaged 7.7 ± 0.8 mammary carcinomas per rat (n = 28; Table 1). The objective of the work here was to fine map the Mcs5 locus in line C using recombinant congenic lines that carry shorter segments of the WKy chromosome 5 Mcs5 locus. Several WF.WKy-Mcs5-region recombinant congenic rat lines were bred that contained different segments of the WKy chromosome 5 candidate region (Fig. 1). The recombinant congenic lines and the genetic markers that define their WKy intervals are defined as follows: line E, WF.WKy-gUwm47-10/D5Uwm37; line F, WF.WKy-gUwm40-18/AU048474; line G, WF.WKy-D5Wox8/gUwm41-6; line O, WF.WKy-gUwm40-18/gUwm45-5; line Q, WF.WKy-gUwm50-22/AU048474; line T, WF.WKy-bUwm26-2/bUwm36-1; and line Y, WF.WKy-bUwm32-4/D5Mit4.

As shown in Table 1, WKy-homozygous females from line O averaged 4.0 ± 0.4 mammary carcinomas per rat (n = 49), which was statistically different from the WF-homozygous littermates (P < 0.0001) and the WF parental strain (P < 0.0001). The WKy mammary cancer resistance allele in congenic line O spans a 4.5-Mb region from genetic markers gUwm40-18 to gUwm45-5. However, the candidate region defining the Mcs5a locus in this line is further delimited by the overlap of the proximal end of line O with the distal end of mammary carcinoma susceptible congenic line G. Line G females that are WKy homozygous averaged 8.9 ± 0.5 mammary carcinomas per rat (n = 53), which was not statistically different from the WF parent strain or the line G WF-homozygous females that average 7.5 ± 0.8 carcinomas per rat (n = 10; Table 1). Congenic line G is defined proximally by marker D5Wox8 and distally by gUwm41-6 (Fig. 1). Thus, the region of overlap with line O is between the distal and proximal ends of lines G and O, respectively. This overlap region is defined by the markers gUwm40-18 and gUwm41-6 and is not expected to contain Mcs5a susceptibility genes or elements. Taken together, Mcs5a is thus contained in the 1-Mb region between markers gUwm41-6 and gUwm45-5. The negative result obtained with line G also rules out the possibility that the proximal end of positive line C from markers D5Wox7 to gUwm41-6 contains independently acting Mcs5 subloci.

Distal to the Mcs5a locus on rat chromosome 5 is Mcs5b, defined by congenic line Q, which has an unexpected phenotype of increased mammary carcinoma susceptibility compared with WF-homozygous rats. As shown in Table 1, line Q WKy-homozygous females averaged 11.6 ± 0.6 mammary carcinomas per rat (n = 59). This was significantly higher (P < 0.01) than the WF-homozygous group that averaged 7.6 ± 0.6 mammary carcinomas per rat (n = 30) and the WF parental strain (P < 0.0001). Distally, the Mcs5b locus is delimited by the overlap of line Q and congenic line T (Fig. 1), which has a mammary carcinoma susceptibility phenotype similar to the WF inbred strain. Line T females that were WKy homozygous averaged 9.7 ± 0.5 mammary carcinomas per rat (n = 34), which was not statistically different (P > 0.05) from the WF-homozygous females that averaged 8.5 ± 0.7 carcinomas per rat (n = 21; Table 1) but was statistically different from the WKy-homozygous line Q females (P < 0.01). Congenic line T overlaps the distal end of line Q (Fig. 1), further narrowing the Mcs5b locus to the region spanned, minimally, by proximal marker gUwm50-22 and distal marker bUwm26-2. Thus, the Mcs5b locus maps to the 7.5-Mb region of rat chromosome 5.

Distally, the susceptible line T overlaps congenic line Y that contains the Mcs5c locus. In DMBA-induced mammary carcinogenesis experiments, WKy-homozygous females from line Y averaged 3.5 ± 0.4 mammary carcinomas per rat (n = 30), which was a significant reduction in tumor multiplicity compared with the WF-homozygous females that averaged 9.6 ± 1.1 mammary carcinomas per rat (n = 29; P < 0.0001) and the WF parental strain (P < 0.0001; Table 1). Congenic line E overlaps the distal end of line Y and was used to further define the Mcs5c locus because of its mammary carcinoma susceptibility phenotype. Line E females that were WKy homozygous for this region of rat chromosome 5 averaged 6.0 ± 0.5 mammary carcinomas per rat (n = 47), which was not statistically different from the WF-homozygous or the inbred WF females that, respectively, averaged 7.7 ± 0.8 and 6.4 ± 0.6 mammary carcinomas per rat (n = 20 and n = 27). Considering the mammary carcinoma–resistant phenotype of line Y and the susceptible phenotypes of overlapping lines E and T, the Mcs5c locus is defined by a 4.5-Mb region of rat chromosome 5 from bUwm36-1 to gUwm47-10.

To determine the effect of one WKy allele at each of the Mcs5 subloci, the phenotypes of heterozygous and homozygous rats were compared from each congenic line for which the WKy-homozygous phenotype deviated significantly (P < 0.05) from the WF susceptibility phenotype (Table 1). Line O (Mcs5a) females with one WKy allele (WKy/WF) developed an average of 5.9 ± 0.4 mammary carcinomas per rat (Table 1), which was 66% of the average number of mammary carcinomas per rat that developed in the line-control littermates with two WF alleles (WF homozygous; Fig. 2) that averaged 8.9 ± 0.8 mammary carcinomas per rat (P < 0.01). Line O heterozygous rats were also significantly different from the WKy-homozygous group that averaged 4.0 ± 0.4 mammary carcinomas per rat (n = 49; P < 0.05). The observed mammary tumor multiplicity phenotype for line O heterozygous females, which was approximately intermediate (5.9 observed versus 6.5 midpoint) between the two homozygous phenotypic values (4.0 and 8.9), was not statistically different from the midpoint (P = 0.08), suggesting that no dominance exists with respect to the WKy and WF Mcs5a alleles.

View larger version (24K): [in this window] [in a new window] [Download PPT slide]   Figure 2. Relative number of mammary carcinomas per rat induced by DMBA in WF.WKy congenic lines harboring chromosome 5 Mcs5 loci. Columns, percentages of mammary carcinomas per rat relative to the average number of mammary carcinomas per rat induced in the WF-homozygous genotype (set at 100%) as calculated using the values from Table 1; bars, ±SE. WF-homozygous (white columns), WKy/WF-heterozygous (gray columns), and WKy-homozygous (black columns). , estimated Mcs5a + Mcs5b heterozygote phenotype based on the combined additive effects of Mcs5a and Mcs5b WKy alleles at each locus and the dominance effect of the WKy allele at the Mcs5b locus (i.e., assuming no epistatic deviation between the Mcs5a and Mcs5b loci). , estimated Mcs5a + Mcs5c heterozygote phenotype based on the combined additive effects at the Mcs5a and Mcs5c loci (i.e., assuming no Mcs5a X Mcs5c interaction). , estimated Mcs5 (Mcs5a, Mcs5b, and Mcs5c) heterozygous phenotype is based on the combined additive effects of Mcs5a + Mcs5b + Mcs5c, the dominance effect of the WKy allele relative to the WF allele at the Mcs5b locus, the epistatic interaction of Mcs5a x Mcs5b, and the Mcs5a x Mcs5c synergistic interaction.

The line Y (Mcs5c) heterozygous group developed 64% of the number of mammary carcinomas per rat of the WF-homozygous littermate females (Fig. 2), whereas the WKy-homozygous line Y (Mcs5c) females averaged 36% the number of mammary carcinomas relative to the WF-homozygous phenotype. Line Y heterozygous females averaged 6.1 ± 0.6 mammary carcinomas per rat, which was significantly different from the WF-homozygous genotype (P < 0.01; Table 1). The WKy-homozygous and WKy/WF-heterozygous females from line Y were also significantly different (P < 0.05) from each other. The mammary carcinoma multiplicity phenotype observed for line Y heterozygous females (6.1 mammary carcinomas per rat) was not statistically different (P = 0.20) from the midpoint value (6.6 mammary carcinomas per rat) calculated from the two homozygous phenotypic values (3.5 and 9.6), indicating that no dominance exists with respect to the WKy and WF Mcs5c alleles.

Congenic line C heterozygous females, which had one WKy allele of each of the identified Mcs5 subloci, were tested for this study to determine if the resistance phenotype of the heterozygote was different from the WKy-homozygous phenotype. Line C heterozygous rats averaged 2.1 ± 0.3 mammary carcinomas per rat (n = 43; Table 1). This was significantly different from the line C females that inherited only WF Mcs5 alleles (WF homozygous; P < 0.0001) but not different from the line C WKy-homozygous females (1.3 ± 0.2 mammary carcinomas per rat; n = 43). Thus, the combined haploid effects of one copy of each of the Mcs5 subloci contribute to a Mcs5 WKy allele that is completely dominant to the WF allele. The congenic line experiments presented here indicate that Mcs5a, Mcs5b, and Mcs5c are currently the only identified independently acting Mcs5 susceptibility subloci; a distinct possibility exists that additional subloci within Mcs5a, Mcs5b, or Mcs5c may be defined in the future. The combined effects of the three currently defined Mcs5 subloci could be hypothesized to account for the strong resistance phenotypes of the line C WKy-homozygous and WKy/WF-heterozygous genotypes. However, although the additive effects on mammary carcinoma multiplicity of the WKy alleles at the Mcs5a and Mcs5c loci are to decrease susceptibility, the WKy Mcs5b allele acts in the opposite direction to increase susceptibility. In other words, the additive effects of Mcs5a, Mcs5b, and Mcs5c together do not account for the strong phenotype exhibited by line C WKy-homozygous and WKy/WF-heterozygous females, which would contain WKy alleles at all three currently identified Mcs5 subloci. For example, considering the additive and dominance effects of the WKy alleles at all three loci together (A_Mcs5a + A_Mcs5b + A_Mcs5c + D_Mcs5b), the estimated additive genotypic value for the line C heterozygote relative to the WF-homozygous phenotype (7.7 mammary carcinomas per rat) is 5.9 mammary carcinomas per rat. This value is calculated using the statistics from Table 1 {[7.7 + (7.7 x –27%) + (7.7 x 26%) + (7.7 x –31%)] + (5.2 x 14%)}. However, the observed genotypic value for the heterozygote (line C) is 2.1 ± 0.3 mammary carcinomas per rat (Table 1), which is statistically different from the 5.9 estimate (P < 0.0001).

Assuming that Mcs5a, Mcs5b, and Mcs5c are distinct individual loci, we evaluate the potential nonadditive interactions among them. To determine if one of the Mcs5 subloci conferring resistance to the development of mammary carcinomas showed an epistatic interaction with the Mcs5b locus, we tested line F (Fig. 1). Congenic line F contains WKy alleles at the Mcs5a and Mcs5b loci but not at the Mcs5c locus. Therefore, line F heterozygous females, based upon the additive effects of Mcs5a and Mcs5b and the dominance deviation of Mcs5b, would be expected to have a 15% increase in mammary carcinomas per rat relative to the WF genotype. This difference as illustrated in Fig. 2 may be experimentally indistinguishable from the WF phenotype. However, the mammary carcinoma multiplicity of line F heterozygous females was statistically different (P < 0.05) from that of the WF-homozygous females (Table 1); but, contrary to the predicted increase based on the additive and dominance components of Mcs5a + Mcs5b, the WKy allele in line F conferred a decreased susceptibility compared with the WF genotype. WKy-homozygous females from line F also had a decreased susceptibility phenotype. They developed 4.9 ± 0.3 mammary carcinomas per rat (n = 89), which was significantly different (P < 0.001) from the WF-homozygous females from line F that developed 7.5 ± 0.5 mammary carcinomas per rat (n = 50; Table 1). Unlike the line O (Mcs5a only) WKy-homozygous versus heterozygous comparison, the line F WKy-homozygous genotype was not statistically different from the line F heterozygous genotype. These results indicate that the Mcs5a WKy allele, as defined by congenic lines phenotyped in this study, masks the effects of the Mcs5b WKy allele; that is, Mcs5a has a potential epistatic effect. In support of this, when the reductions in mammary carcinomas per rat relative to the WF-homozygous genotypes for lines F and O were compared, the line F heterozygous females (A_Mcs5a + A_Mcs5b + D_Mcs5b + I_Mcs5aMcs5b) were not statistically different from the line O heterozygous females (A_Mcs5a; Fig. 2).

To determine whether the Mcs5a and Mcs5c WKy alleles in the absence of the Mcs5b WKy allele could account for the strong phenotype conferred by the Mcs5 WKy allele, which is currently defined as A_Mcs5a + A_Mcs5b + A_Mcs5c + D_Mcs5b + I_Mcs5aMcs5b (where I_Mcs5aMcs5b represents a term for an epistatic interaction between Mcs5a and Mcs5b), we crossed the WF.WKy lines O and Y to determine the Mcs5a x Mcs5c mammary carcinoma susceptibility phenotype. The resulting heterozygous females from this cross had an average of 3.4 ± 0.4 mammary carcinomas per rat (n = 36), significantly less (P < 0.0001) than the average of 9.7 ± 0.8 mammary carcinomas per rat (n = 13) for the WF-homozygous congenic females (Table 1). Compared with the WF-homozygous genotype, the Mcs5a x Mcs5c resulted in 35% the number of mammary carcinomas per rat (3.4 versus 9.7; Fig. 2). The expected additive value of one WKy allele at each of the Mcs5a and Mcs5c loci, based on the results from congenic lines O and Y, would be estimated to transmit a reduction in mammary carcinoma multiplicity that would correspond to 4.1 mammary carcinomas per rat (9.7-2.6-3.0; Fig. 2). This would be 42% of the number of mammary carcinomas per rat compared with the WF-homozygous genotype (4.1 versus 9.7; Fig. 2). The estimated phenotypic value of line O x line Y (4.1 mammary carcinomas per rat) is statistically different (P < 0.05) from the 3.4 ± 0.4 mammary carcinomas per rat observed, indicating that there is a potential interaction between the Mcs5a and Mcs5c WKy alleles.

We were also able to account for the strong phenotype of line C (Mcs5) by crossing the congenic lines that independently contained Mcs5a or Mcs5c fixed loci. Rats from the Mcs5 congenic line C and the Mcs5a x Mcs5c cross had similar mammary carcinoma multiplicities relative to the WF-homozygous averages for each group, 27 ± 3.4% and 35 ± 4.1%, respectively (Fig. 2). These relative mammary carcinoma multiplicity phenotypes for the two groups were not statistically different (P = 0.09). Therefore, the strong phenotype of line C (Mcs5c), where the heterozygote developed 27% of the number of mammary carcinomas per rat compared with the WF-homozygous genotype, may be explained by the additive, dominance, and interaction effects of the Mcs5a, Mcs5b, and Mcs5c loci (Mcs5 = µ + A_Mcs5a + A_Mcs5b + A_Mcs5c + D_Mcs5b + I_Mcs5aMcs5b + I_Mcs5aMcs5c). Using this formula, the estimated phenotype for the line C (Mcs5) heterozygote relative to the WF-homozygous phenotypic value is 2.3 mammary carcinomas per rat [7.7 + (–2.1) + 2.0 + (–2.4) + 0.7 + (–3.1) + (–0.5)]. This estimate is not statistically different from the 2.1 mammary carcinomas per rat observed for the line C heterozygote (Table 1). The estimate for the Mcs5 heterozygote without the I_Mcs5aMcs5c component is 2.8 mammary carcinomas per rat, which in a one-sample t test is significantly different (P < 0.01) from the observed phenotype (2.1 mammary carcinomas per rat), providing additional support for an interaction between Mcs5a and Mcs5c WKy alleles. It should be noted that we cannot reject the hypothesis the Mcs5b and Mcs5c loci interact.

To further test the hypothesis that Mcs5a interacts with Mcs5b and Mcs5c, we used a general linear model that incorporated a Poisson regression analysis to estimate the effect of WKy alleles on mammary carcinoma multiplicity. The estimated fold change in mammary carcinoma multiplicity for the main and interaction effects are shown in Table 2. As expected, the main effects for each Mcs5 sublocus were significant (P < 0.0001). The magnitude of the fold changes in mammary carcinoma susceptibilities with one versus two WKy alleles at Mcs5a and Mcs5c loci were additive, whereas the WKy allele at the Mcs5b locus exhibited a degree of dominance over the WF allele. Furthermore, the fold changes in mammary carcinoma multiplicity phenotype estimated by the model for one WKy allele at each of the Mcs5 subloci were similar to the additive effects calculated in Table 1 for each individual congenic line. The model predicted a significant (P < 0.0001) interaction between Mcs5a and Mcs5b (I_Mcs5aMcs5b) when one WKy allele was present at each locus. The presence of additional WKy alleles at the Mcs5a and Mcs5b loci did not significantly increase the effect of this potential epistatic interaction. The model also predicted an interaction between Mcs5a and Mcs5c. The model indicated (P = 0.06) there may be an interaction between WKy alleles at the Mcs5a and Mcs5c loci when one WKy allele was present at each locus. The interaction between Mcs5a and Mcs5c was significant (P < 0.01) when two WKy alleles were present at each locus. The model also predicted an epistatic interaction between WKy alleles at the Mcs5b and Mcs5c loci (P < 0.01).

The current data support the existence of at least three distinct subloci within the Mcs5 QTL. It is possible, however, that one or more of these subloci may harbor multiple genetic elements that modulate susceptibility to mammary cancer through independent or interactive activities. In spite of this, the three currently identified subloci and their likely epistatic interactions clearly classify the Mcs5 QTL as complex. Interestingly, the only other rat mammary carcinoma susceptibility QTL that has been fine mapped, Mcs1, has also been shown to have at least three subloci that modulate risk (7). In addressing the complexity of the Mcs1 locus, we postulated that because many mammary cancer modifiers are likely to have relatively low penetrance, especially as heterozygous alleles, QTL mapping in a backcross study might most readily detect areas of the genome that harbor multiple risk modifiers in a limited genomic interval. Furthermore, we surmised that these loci in close proximity to each other should all act in a similar direction to increase or decrease risk. Mcs5 seems more complex than Mcs1 in that, contrary to the above assumption, the subloci in Mcs5 do not act in the same direction. Mcs5a and Mcs5c act to decrease risk, whereas Mcs5b acts to increase risk. If these three subloci acted solely in an independent manner (i.e., additive), it is unlikely that the Mcs5 QTL would have been detected in our backcross linkage study. The Mcs5 QTL was likely discovered, because Mcs5a in its WKy haplotype epistatically minimized the risk enhancement activity of Mcs5b. Thus, the linkage study identified the Mcs5 QTL as its effect was similar to the sum of Mcs5a and Mcs5c. Most linkage studies would lack the resolution to individually identify these three Mcs5 subloci due to their close proximity. The proximity of subloci within both Mcs5 and Mcs1 raise the question of why this clustering occurs. Two extreme possibilities exist. The first would suggest that there are thousands of mammary cancer modifier genetic elements randomly dispersed in the genome, and the current experimental power in most linkage modifier mapping experiments define a subset of these, which lie in close proximity. The second is compatible with a more limited number of modifier genes that are located in proximity due to functional constraints or their evolutionary origins from a single ancestral gene. As we identify and functionally characterize distinct genetic elements within these loci, we will be better able to address these and other possibilities.

The work reported here furthers our investigations into the comparative genetics of breast cancer and shows the complexity of the multigenic etiologic components of breast cancer. This approach identifies novel susceptibility loci that, when translated to humans, may provide an unbiased selection method of novel candidate loci for evaluation in breast cancer case-control association studies.

	Acknowledgments

 
Grant support: NIH grant CA77494 and U.S. Army Medical Research and Materiel Command postdoctoral fellowship grant DAMD17-03-0280 (D.J. Samuelson).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Dr. Laurie A. Shepel for her critical evaluation of this article and suggestions that enhanced it and Dr. Robert Mau for his assistance with the general linear model and the R computing environment.

	Footnotes

 
¹ http://rgd.mcw.edu/.

² UCSC Genome Browser, http://genome.ucsc.edu/.

Received 5/ 3/05. Revised 8/ 5/05. Accepted 9/ 2/05.

	References

Top Abstract Introduction Materials and Methods Results and Discussion References

 

1. Pharoah PD, Antoniou A, Bobrow M, Zimmern RL, Easton DF, Ponder BA. Polygenic susceptibility to breast cancer and implications for prevention. Nat Genet 2002;31:33–6.[CrossRef][Medline]
2. Wooster R, Weber BL. Breast and ovarian cancer [review]. N Engl J Med 2003;348:2339–47.[Free Full Text]
3. Shepel LA, Lan H, Haag JD, et al. Genetic identification of multiple loci that control breast cancer susceptibility in the rat. Genetics 1998;149:289–99. Erratum in: Genetics 1998;149:1627.[Abstract/Free Full Text]
4. Lan H, Kendziorski CM, Haag JD, Shepel LA, Newton MA, Gould MN. Genetic loci controlling breast cancer susceptibility in the Wistar-Kyoto rat. Genetics 2001;157:331–9.[Abstract/Free Full Text]
5. Samuelson DJ, Haag JD, Lan H, et al. Physical evidence of Mcs5, a QTL controlling mammary carcinoma susceptibility, in congenic rats. Carcinogenesis 2003;24:1455–60.[Abstract/Free Full Text]
6. Wheeler DL, Barrett T, Benson DA, et al. Database resources of the National Center for Biotechnology Information. Nucleic Acids Res 2005;33:D39–45.[Abstract/Free Full Text]
7. Haag JD, Shepel LA, Kolman BD, et al. Congenic rats reveal three independent Copenhagen alleles within the Mcs1 quantitative trait locus that confer resistance to mammary cancer. Cancer Res 2003;63:5808–12.[Abstract/Free Full Text]

This article has been cited by other articles:

				M. N. Gould The Utility of Comparative Genetics to Inform Breast Cancer Prevention Strategies Genetics, October 1, 2009; 183(2): 409 - 412. [Abstract] [Full Text] [PDF]

				D. J. Samuelson, S. E. Hesselson, B. A. Aperavich, Y. Zan, J. D. Haag, A. Trentham-Dietz, J. M. Hampton, B. Mau, K.-S. Chen, C. Baynes, et al. Rat Mcs5a is a compound quantitative trait locus with orthologous human loci that associate with breast cancer risk PNAS, April 10, 2007; 104(15): 6299 - 6304. [Abstract] [Full Text] [PDF]

				B. S. Schaffer, C. M. Lachel, K. L. Pennington, C. R. Murrin, T. E. Strecker, M. Tochacek, K. A. Gould, J. L. Meza, R. D. McComb, and J. D. Shull Genetic Bases of Estrogen-Induced Tumorigenesis in the Rat: Mapping of Loci Controlling Susceptibility to Mammary Cancer in a Brown Norway x ACI Intercross. Cancer Res., August 1, 2006; 66(15): 7793 - 7800. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Samuelson, D. J. Articles by Gould, M. N. Search for Related Content PubMed PubMed Citation Articles by Samuelson, D. J. Articles by Gould, M. N.