In-silico QTL mapping of postpubertal mammary ductal development in the mouse uncovers potential human breast cancer risk loci - PubMed (original) (raw)
In-silico QTL mapping of postpubertal mammary ductal development in the mouse uncovers potential human breast cancer risk loci
Darryl L Hadsell et al. Mamm Genome. 2015 Feb.
Abstract
Genetic background plays a dominant role in mammary gland development and breast cancer (BrCa). Despite this, the role of genetics is only partially understood. This study used strain-dependent variation in an inbred mouse mapping panel, to identify quantitative trait loci (QTL) underlying structural variation in mammary ductal development, and determined if these QTL correlated with genomic intervals conferring BrCa susceptibility in humans. For about half of the traits, developmental variation among the complete set of strains in this study was greater (P < 0.05) than that of previously studied strains, or strains in current common use for mammary gland biology. Correlations were also detected with previously reported variation in mammary tumor latency and metastasis. In-silico genome-wide association identified 20 mammary development QTL (Mdq). Of these, five were syntenic with previously reported human BrCa loci. The most significant (P = 1 × 10(-11)) association of the study was on MMU6 and contained the genes Plxna4, Plxna4os1, and Chchd3. On MMU5, a QTL was detected (P = 8 × 10(-7)) that was syntenic to a human BrCa locus on h12q24.5 containing the genes Tbx3 and Tbx5. Intersection of linked SNP (r(2) > 0.8) with genomic and epigenomic features, and intersection of candidate genes with gene expression and survival data from human BrCa highlighted several for further study. These results support the conclusion that mammary tumorigenesis and normal ductal development are influenced by common genetic factors and that further studies of genetically diverse mice can improve our understanding of BrCa in humans.
Figures
Figure 1. Lattice plot showing scatter, histograms, and Pearson’s correlations of the strain means for all traits measured
A total of 172 virgin female mice across 43 different inbred strains (4/strain) were imported to our animal facility from the Jackson Laboratory at 5 wks of age. Samples were collected at from 171 females at 6 weeks of age (A) and 158 females at 12 weeks (B) of age. Because sequential biopsies were taken from the same animal, the difference (C) between 6 and 12 week biopsies were used as an indicator of rate for each of the traits. Scatter plots in the lower left, histograms along the diagonal and Pearson’s correlations in the upper right.
Figure 2. Distributions for quantitative ductal development traits among virgin female mice from inbred strains within the MDP
Ductal development traits were measured in inbred mouse strains (n 2–5/strain) from the MDP at either 6 (n=171) or 12 (n=158) wks of age. The trait distributions at 6 week (pink) or 12 weeks (blue) of age are presented as histograms for ductal area (A), ductal perimeter (B), total branch count (C), total duct length (D), and branch density (E). All five traits were increased (P<0.05) in 12 week samples in comparison to 6 week samples.
Figure 3. Comparison of mammary ductal development traits among inbred strains from within the MDP show 4 development clusters
A heatmap displays the mean normal scores for six quantitative traits describing mammary ductal morphology at either 6 wks, 12 wks, or as the difference between 6 and 12 wks of age (A). All samples were collected from estrus-synchronized females. Strains were hierarchically clustered based the means for each trait-time point combination. Cells colored white have not data for the WSB/EiJ strain. A genetic distance matrix is shown (B) to indicate clustering of closely related strains. Strain names in blue indicate “classical” strains and strains names underlined indicate current common use strains. Detailed descriptive statistics for each of the strains along with ANOVA results and heritability estimates are provided in Table S1.
Figure 4. Images of mammary wholemounts prepared from strains representing each of the development clusters
The images shown are of the left #4 gland collected at 6wks (A,C,E,G,I, K, M, and O) of age and the right #4 gland collected at 12 wks (B,D,F,H,J, L, N, and P) of age from the same female within each strain. At the upper extreme of development in cluster 1 were the BUB/BnJ (C and D) and KK/HlJ (A and B) strains. In cluster 2 were the AKR/J (E and F) and PL/J (G and H) Strains. In cluster 3 were the FVB/NJ (I and J) and C57BL/6J (K and L) srains. At the lower extreme of ductal development were the BALB/cByJ (M and N), and CZECHII/EiJ (O and P) strains.
Figure 5. Imaging of the mammary ductal tree in 3D emphasizes the presence of local differences in mammary ductal topology
Whole mammary glands from females at 6 wks of age were immunostained with an antibody to e-cadherin, imaged by optical projection tomography and then reconstructed in 3D for the CZECHII/EiJ (A; supplemental video 1) and KK/HlJ (D; supplemental video 2) strains. Tree Surveyor allowed for the production of skeletons (B and E) and phylograms (C and F) for the visualization of local patterning variations.
Figure 6. Mammary branch density at 12 wks of age is positively correlated to mammary tumor metastasis
Strain means from the current study were combined with previously published observations on mammary tumor latency and the incidence of lung metastases in F1 offspring from inbred strains crossed with transgenic mice overexpressing polyomavirus middle T (FVB/N-Tg(MMTVPyVT)634Mul/J). Scatterplots, histograms, and Pearson’s r are shown for those traits most highly correlated to tumor latency and metastatic index (A). Traits with a Pearson’s r >0.3 were significantly correlated (p < 0.05). Visualization of branch density at 12 wks of age suggests the presence of two groups of strains each with ductal density that is highly correlated to lung metastasic index (B). Tumor data are from (Lifstedt et al. 1998).
Figure 7. Significant genome-wide associations for mammary development traits
Genome-wide association analysis was conducted using EMMA. Manhattan plots are shown for ductal area (A, F, K) perimeter (B, G, L) branches (C, H, M), length (D, I, N), and density (E, J, O) at 6wk (A–E), 12wk (F–J) and for the delta between 6 and 12 wk (K–O). Regions with significantly associated SNPs (cyan) were identified using a Bonferroni-adjusted threshold of 2×10−6. Significant associations were detected on MMU 1, 2, 5, 6, 7, 13, 15, 16, and 18. Regions of synteny between mouse and human were identified using the UCSC genome browser. These regions were then cross referenced with a list of recently summarized human BrCa risk loci. Chromosomal locations of ductal development loci are shown as squares. Each square is labeled with the corresponding location in the human genome. Mouse loci overlapping human BrCa loci are colored red. Locations of the human BrCa loci were obtained from (Ghoussaini et al 2013).
Figure 8. Visualization of local LD in the top 3 candidate regions
LD was assessed by calculating r2 for SNP within a region that was 1 MBP on either side of the lead SNP for the top 3 associations based on P-value. These associations were detected on MMU 6 (A), 1 (B), and 18 (C). The associated SNP are plotted along with their respective P-values, as well as recently derived estimates of local recombination rates (Brunschwig et al 2012). Potential candidate genes are also shown as well as potential causal SNP contained within and around genes. For each plot, the lead SNP is shown as the largest symbol, colored blue along with it respective id and p-value. Symbols coded red are SNP in high LD (r2 >0.8) with the lead SNP, Symbols colored orange are in moderate LD (0.6 < r2 < 0.8). For green symbols, r2 is between 0.4 and 0.6. Cyan symbols have low LD (0.2 < r2 < 0.4). Blue symbols have no linkage (r2<0.2).
Figure 9. Visualization of local LD in candidate regions with overlap to human BrCa risk alleles or containing known mammary ductal development genes
LD was assessed by calculating r2 for SNP within a region that was 1 MBP on either side of the lead SNP for the top QTL that overlapped with known human BrCa loci (Mdq11 and 15) or contained genes with strong evidence supporting a direct role in mammary ductal development (Mdq19) based on p-value. These QTL were detected on MMU5 (A), 6, (B), and 9 (C). The associated SNP are plotted along with their respective P-values, as well as recently derived estimates of local recombination rates (Brunschwig et al 2012). Potential candidate genes are also shown as well as potential causal SNP contained within and around genes. For each plot, the lead SNP is shown as the largest symbol, colored blue along with it respective id and p-value. Symbols coded red are SNP in high LD (r2 >0.8) with the lead SNP, Symbols colored orange are in moderate LD (0.6 < r2 < 0.8). For green symbols, r2 is between 0.4 and 0.6. Cyan symbols have low LD (0.2 < r2 < 0.4). Blue symbols have no linkage (r2<0.2).
Figure 10. Select GWAS candidate genes correlate with breast cancer aggressiveness and distant metastasis-free survival
Kaplan-Meier plots in breast cancer patients from the Kessler et al. dataset, demonstrating six of the 43 candidate genes underlying QTL associated with mammary ductal development in the mouse were also linked to alterations in distant metastasis-free survival (DMFS). Kaplan-Meier plots are shown for Pcna (A), Pgr (B), Rgl1 (C), Tmem230 (D) (human ortholog: C20orf30), Lrp2 (E), and Rassf2 (F). Results are from analysis of public gene expression data from 1340 primary breast cancers, compiled from published studies (Kessler et al. 2012).
References
- Ambros V. The functions of animal microRNAs. Nature. 2004;431:350–355. - PubMed
- Andres AC, Schonenberger CA, Groner B, Hennighausen L, LeMeur M, Gerlinger P. Ha-ras oncogene expression directed by a milk protein gene promoter: tissue specificity, hormonal regulation, and tumor induction in transgenic mice. Proceedings. of. the. National. Academy. of. Sciences. of. the. United. States. of. America. 1987;84:1299–1303. - PMC - PubMed
- Apolant H. Die epithelian geschwulste des maus. Arbeiten Koniglchn Ins Exp The Zu Frankfurt. 1906;1:61.
- Atwood CS, Hovey RC, Glover JP, Chepko G, Ginsburg E, Robison WG, Vonderhaar BK. Progesterone induces side-branching of the ductal epithelium in the mammary glands of peripubertal mice. J. Endocrinol. 2000;167:39–52. - PubMed
Publication types
MeSH terms
LinkOut - more resources
Full Text Sources
Other Literature Sources
Medical
Molecular Biology Databases