Experiment
Genetic background plays a dominant role in mammary gland development and breast cancer (BrCa). The primary goals of the current study were to describe the degree of phenotypic variation in mammary ductal development across 43 strains within the mouse diversity panel (MDP), to identify QTL associated with variations in mammary ductal development, and to determine whether any correlate with known BrCA loci in humans.
Female mice (N=4/strain except where otherwise noted) from 43 inbred mouse strains considered a part of the mouse diversity panel (MDP) were used. All animals were purchased from the Jackson Laboratory at 4 weeks of age. The strains studied included C57BL/6J, 129S1/SvImJ, A/J, AKR/J, BALB/cByJ, BPL/1J, BTBR_T+_tf/J, BUB/BnJ, C3H/HeJ, C57BLKS/J, C57BR/cdJ, C57L/J, C58/J, CAST/EiJ, CBA/J, CE/J, CZECHII/EiJ, DBA/2J, DDY/JclSidSeyFrkJ, FVB/NJ, I/LnJ, KK/HlJ, LG/J, LP/J, MA/MyJ, MOLF/EiJ, MRL/MpJ, MSM/Ms, NOD/ShiLtJ, NON/ShiLtJ, NZL/LtJ, NZO/HlLtJ, NZW/LacJ, PL/J, PWD/PhJ, PWK/PhJ, Qsi5, RIIIS/J, SEA/GnJ, SJL/J, SM/J, SWR/J, and WSB/EiJ.
Biopsies of the #4 mammary gland were collected at 6 and 12 weeks of age. Prior to each biopsy, the study animals were administered a subcutaneous injection of pregnant mare serum gonadotropin followed 45 to 47 hours later with a subcutaneous injection of human chorionic gonadotropin. Biopsies were collected at 24 hours following HCG-injection. These treatments are known to induce ovulation and synchronize all of the animals to the estrus phase of the cycle.
Traits were measured at 6 wks and 12 weeks of age and were referred to as area_6, perimeter_6, branches_6, length_6 and density_6 and area_12, perimeter_12, branches_12, length_12 and density_12, respectively. Both the 6 wk and 12 wk data were collected from wholemounts taken from the same animal and individual rates of change were calculated for all 5 traits as the difference between the 12wk measurement and the 6wk measurement. These traits were referred to as the D_traits: area_D, perimeter_D, length_D, branches_D, and density_D.
GWAS was conducted using the EMMA package in R in conjunction with a composite SNP dataset that was obtained from the Mouse Phenome Database (http://phenome.jax.org/). Coordinates for this dataset are based on GRCm38/mm10. Genome-wide significance thresholds were set by running EMMA on 1000 permutations of the data from each of the 15 traits. Twenty mammary ductal QTL (Mdq) containing associated SNP were detected on chromosomes 1, 2, 5, 6, 7, 9, 13, 15, 16, and 18, Table 1:
QTL Mdq1 (mammary ductal QTL 1) mapping to Chromosome 6 with SNP rs29972830, p=8.9 x 10(-12), was associated with the trait branches_D. Traits perimeter_D and length_D were also associated with the Mdq1 locus.
QTL Mdq2 (mammary ductal QTL 2) mapping to Chromosome 1 with SNP rs32398376, p=2.08 x 10(-9), was associated with the trait area_12. Traits perimeter_12, branches_12 and length_12 were also associated with the Mdq2 locus.
QTL Mdq3 (mammary ductal QTL 3) mapping to Chromosome 18 with SNP rs13483228, p=3.18 x 10(-9), was associated with the traits area_6, perimeter_6, branches_6, and length_6.
QTL Mdq4 (mammary ductal QTL 4) mapping to Chromosome 2 with SNP rs33051600, p=2.56 x 10(-8), was associated with the trait density_6.
QTL Mdq5 (mammary ductal QTL 5) mapping to Chromosome 13 with SNP rs30095733, p=6.33 x 10(-8), was associated with the trait branches_D.
QTL Mdq6 (mammary ductal QTL 6) mapping to Chromosome 16 with SNP rs46013670, p=7.65 x 10(-8), was associated with the trait density_12.
QTL Mdq7 (mammary ductal QTL 7) mapping to Chromosome 1 with SNP rs50444438, p=1.05 x 10(-7), was associated with the traits area_12, perimeter_12, and length_12.
QTL Mdq8 (mammary ductal QTL 8) mapping to Chromosome 2 with SNP rs27176282, p=1.24 x 10(-7), was associated with the trait density_D.
QTL Mdq9 (mammary ductal QTL 9) mapping to Chromosome 9 with SNP rs36588132, p=1.40 x 10(-7), was associated with the traits perimeter_D, and branches_D.
QTL Mdq10 (mammary ductal QTL 10) mapping to Chromosome 6 with SNP rs29830755, p=2.06 x 10(-7), was associated with the trait branches_D.
QTL Mdq11 (mammary ductal QTL 11) mapping to Chromosome 7 with SNP rs3696903, p=3.81 x 10(-7), was associated with the trait density_12.
QTL Mdq12 (mammary ductal QTL 12) mapping to Chromosome 5 with SNP rs33753649, p=5.27 x 10(-7), was associated with the trait area_D.
QTL Mdq13 (mammary ductal QTL 13) mapping to Chromosome 6 with SNP rs30305626, p=6.76 x 10(-7), was associated with the traits branches_12 and length_12.
QTL Mdq14 (mammary ductal QTL 14) mapping to Chromosome 2 with SNP rs27275156, p=8.29 x 10(-7), was associated with the trait branches_D.
QTL Mdq15 (mammary ductal QTL 15) mapping to Chromosome 5 with SNP rs29778211, p=8.38 x 10(-7), was associated with the trait perimeter_6.
QTL Mdq16 (mammary ductal QTL 16) mapping to Chromosome 5 with SNP rs31558147, p=8.43 x 10(-7), was associated with the trait density_D.
QTL Mdq17 (mammary ductal QTL 17) mapping to Chromosome 15 with SNP rs46415121, p=1.06 x 10(-6), was associated with the traits perimeter_12 and length_12.
QTL Mdq18 (mammary ductal QTL 18) mapping to Chromosome 2 with SNP rs27336088, p=1.21 x 10(-6), was associated with the trait length_D.
QTL Mdq19 (mammary ductal QTL 19) mapping to Chromosome 9 with SNP rs29782631, p=1.52 x 10(-6), was associated with the trait perimeter_D.
QTL Mdq20 (mammary ductal QTL 20) mapping to Chromosome 1 with SNP rs36274122, p=1.52 x 10(-6), was associated with the traits perimeter_12 and length_12.
To determine if any of the Mdq QTL detected in the current study were syntenic with previously described human BrCa loci, the UCSC genome browser was used to perform a liftover of the murine coordinates for each locus from mm10 to hg19. Combining the results with a previously published summary of human BrCa risk loci (Ghoussaini et al. 2013) revealed that 5 (Mdq5, Mdq8, Mdq11, Mdq12 and Mdq15) of the 20 Mdq were syntenic with previously detected BrCa risk loci. [Table 1].
To identify potential candidate genes with the Mdq loci, additional genotype calls were obtained from 9 MPD databases and included the calls for each of the 20 Mdq lead SNP. For each of the lead SNP within a locus, both D and r2 were calculated for surrounding SNPs at up to 10 Mbp on either side. Nineteen of the 20 identified loci had detected LD blocks. Intersection of these blocks with genomic data from the Mouse Phenome Database identified 157 potential candidate genes [Table 1].
The most highly significant (P=110(-11)) association of the study was with QTL Mdq1 (MMU6), whose LD block contained the genes Plxna4, Plxna4os1, and Chchd3. On MMU5, QTL Mdq15 was syntenic to a human BrCa locus on h12q24.5 and contained the genes Tbx3 and Tbx5.
Analysis Tools