Reference
The Diversity Outbred heterogeneous stock (J:DO) is a developing mouse population derived from progenitor lines of the Collaborative Cross (CC). The CC is a panel of recombinant inbred (RI) mouse strains that combines the genomes of eight genetically diverse founder strains - A/J, C57BL/6J, 129S1/SvImJ, NOD/ShiLtJ, NZO/HlLtJ, CAST/EiJ, PWK/PhJ, and WSB/EiJ - to capture nearly 90% of the known variation present in laboratory mice (Churchill et al. 2004). Animals from 160 incipient CC lines at early stages of inbreeding were used to establish the DO population, which is maintained by a randomized outbreeding strategy that avoids brother-sister matings. The DO and CC populations thus capture the same set of natural allelic variants derived from a common set of eight founder strains, with DO mice being outbred and the CC population being inbred.
CTC (2004), Churchill, G. A., et al. The Collaborative Cross, a community resource for the genetic analysis of complex traits. Nat Genet. 36, 1133-7.
Experiment
Pancreatic cancer (PC) is one of the leading causes of cancer mortality. To understand the genetic causes of pancreatic cancer (PC), the authors conducted a genome-wide association study (GWAS) using the diversity outbred (DO) mice population to identify susceptibility genes underlying 7,12-dimethylbenzanthraene (DMBA) induced PC.
Two hundred and seventy 5th-generation DO mice were purchased from the Jackson Laboratory at the age of about 12 weeks. 136 mice were male and 134 were female.
For each mouse, serial tissue sections (5 m each) were cut from formalin-fixed Paraffin-Embedded (FFPE) pancreatic tissues and 40 separate sections were picked and stained with H&E and examined histologically under a light microscope. The pancreatic lesions were counted from the H&E-stained sections of each pancreatic sample. The percent lesion area of PC in DO mice was calculated as 100 x (the pancreatic lesion area/the whole pancreata area) and used as the phenotype data in the study.
Whole genome genotyping of all the DNA samples was outsourced to GeneSeek (http://www.neogen.com/GeneSeek), which completed the work using the Mouse Universal Genotyping Array (MUGA), a 7,851 SNP-based genotyping array built on the Illumina Infinium platform (Collaborative Cross Consortium, 2012). Markers on the MUGA are distributed genomewide with average spacing of 325 Kb and standard deviation of 191 Kb. After removing SNPs with low call rates (> 10% genotyping missing rate in mice), the authors successfully genotyped 6,485 SNPs distributed across the whole genome for 228 DO mice from generations G5 (n = 251 for total available mice for sacrificing, while 23 were excluded from genotyping due to the lack of obvious PC phenotype).
The authors used a robust statistical method, efficient mixed model association (EMMA) approach (Kang, 2008) to perform GWAS in the DO mice population. EMMA can effectively correct for population structure and genetic relatedness among DO mice so the authors utilized this method to assess SNP association with percent PC lesion area.
Four SNPs (QTL) reached genome-wide significance (genome coordinates relative to GRCm38/mm10):
Pancq1 (pancreatic cancer QTL 1) maps to Chr 14: 116,143,616 bp (UNC140360310 [rs51650625]) with a p-value of 5.81E-07. The associated gene is Gpc5 (MGI:1194894).
Pancq2 (pancreatic cancer QTL 2) maps to Chr 17: 4,383,171 bp (backupJAX00429894) with a p-value of 8.08E-07. The associated gene is Arid1b (MGI:1926129).
Pancq3 (pancreatic cancer QTL 3) maps to Chr 16: 95,300,644 bp (UNC160242626) with a p-value of 1.17E-06. The associated gene is Kcnj6 (MGI:104781).
Pancq4 (pancreatic cancer QTL 4) maps to Chr 1: 77,706,916 bp (UNC010148938) with a p-value of 1.60E-06. The associated gene is Epha4 (MGI:98277).
Inbred mice microarray expression data of the four genes reaching genomewide significance level were used for eQTL (expression QTL) analysis. The authors found that the expression level of one gene, Gpc5, was strongly and consistently associated with the same SNP identified in the initial GWAS of PC lesion size (rs51650625, p = 2.94 x 108). The allele G of rs51650625 is associated with significant decreased expression in Gpc5 as compared with the allele A. Specifically, the 'AG' genotype associated with a 42% reduced Gpc5 expression compared to the 'AA' genotype and the 'GG' genotype associated with a 63% reduced Gpc5 expression compared to the 'AA' genotype. In addition, the G allele of rs51650625 is also significantly associated with PC lesion size in DO mice revealed by the initial GWAS. These indicated that the rs51650625 G allele down-regulate the expression of Gpc5 gene via cis-regulation and the corresponding reduced Gpc5 expression leads to the increased PC lesion size in mice.
Together, the data support that Gpc5 as a tumor suppressor gene involved in the etiology of PC.
Analysis Tools