The genetic control of avascular area in mouse oxygen-induced retinopathy

Abstract

Purpose: The C57BL/6ByJ and BALB/cByJ inbred strains of mice are, respectively, susceptible and resistant to oxygen-induced retinopathy (OIR). The purpose of this work was to investigate the genetic control of the retinal avascular area in mouse OIR using a mapping cross.

Methods: The central retinal avascular area was measured on postnatal day 16 (P16) in C57BL/6ByJ, BALB/cByJ, 101 (C57BL/6ByJ x BALB/cByJ)F₂, and 116 (BALB/cByJ x C57BL/6ByJ)F₂ mice that had been subjected to the OIR protocol. A genome-wide scan was performed of selected albino and non-albino mice to determine quantitative trait loci associated with weight and avascular area.

Results: C57BL/6ByJ mice had significantly larger avascular areas than BALB/cByJ ones. Albino mice of the F₂ generation had smaller avascular areas than the non-albino mice. Genotyping was performed at 856 informative single nucleotide polymorphisms approximately evenly distributed across the genome from each of 85 selected F₂ mice. Weight, sex, and the paternal grandmother were found to act as additive covariates associated with the avascular area on P16; mapping analyses that used a model incorporating these covariates found a quantitative trait locus on chromosome 7 related to avascular area. Mapping analyses that used a model that did not incorporate covariates found a quantitative trait locus on chromosome 9 related to avascular area. A quantitative trait locus for bodyweight on P16 was mapped to chromosome 5.

Conclusions: The retinal avascular area in the mouse OIR model is under genetic control. Revascularization in OIR is related to the weight, strain of paternal grandmother, sex, and albinism. Our data support the existence of a quantitative trait locus on chromosome 5 that influences weight after exposure to hyperoxia, as well as quantitative trait loci on chromosomes 7 and 9 that modify susceptibility to OIR.

Figure 1

Illustration of the breeding scheme used in the mapping cross. The BALB/cByJ and C57BL/6ByJ parental strains are crossed in both directions. The resulting F₁ mice are inbred through brother–sister mating to produce the F₂ generation. This figure shows the manner in which the Y chromosomes and recombined (black and white-striped) and unrecombined (solid color) X chromosomes are passed through the generations of the cross. White chromosomes are inherited from the BALB/cByJ albino parent, and black chromosomes are inherited from the C57BL/6ByJ pigmented parent. The different coat colors shown in the F₂ generation are for illustrative purposes only. The color of the F₂ animals is unrelated to either gender or the direction of the cross.

Figure 2

Retinas after oxygen induced retinopathy (OIR) protocol. A: Flatmounted retina of a BALB/cByJ mouse on postnatal day 16 after exposure to the OIR protocol. The retina is relatively well vascularized at this time point. B: A retina from a C57BL/6ByJ mouse at the same time point after exposure to the OIR protocol. The central retina has large areas of avascularity.

Figure 3

Mean avascular areas of parental strains after adjusting for influence of weight. This shows that the BALB/cByJ strain has smaller areas of avascular retina in response to hyperoxia than the C57BL/6ByJ strain (ANOVA, p=0.0002).

Figure 4

The inverse association between avascular area and weight of F₂ mice at postnatal day 16 after exposure to oxygen-induced retinopathy. The avascular area was shown to correlate with weight with a coefficient of −0.1743. This association was shown to be statistically significant (Welch’s two-sample t-test, p<10⁻⁸).

Figure 5

Comparison of avascular areas for albino and non-albino F₂ mice. The mean avascular area of albino mice was found to be 0.67 mm² versus 1.20 mm² for non-albino mice, a statistically significant difference (ANOVA, p<0.0001).

Figure 6

The results of genetic linkage mapping using central retinal avascular area as the phenotype. A: Linkage map of avascular area across the whole genome created using no covariates (blue) and using weight, paternal grandmother, and sex as additive covariates in the analysis (red). Empirically derived genome-wide (alpha=0.05) thresholds of significance are shown as horizontal lines. For mapping without covariates the threshold of significance was a logarithm of odds (LOD) score of 3.75. For mapping with covariates the threshold of significance LOD score of 3.81. Mapping without covariates identified peaks on chromosomes 7 (LOD=8.90) and 9 (LOD=3.75). Mapping with covariates identifed a peak on chromosome 7 (LOD=9.83). B: Linkage map of avascular area across chromosome 9 created without inclusion of covariates, showing a peak at SNP rs4135590 (LOD=3.75, p=0.05). C: Effect plots of primary peak on chromosome 9. The BALB/cByJ allelotype was found to be additively protective against the development of OIR. D: Linkage map of avascular area across chromosome 7 created by including weight, paternal grandmother, and sex as additive covariates, showing a peak between SNPs rs13479506 and CEL-7_115892950 (LOD=9.83, p<0.004). E: Effect plot of the peak on chromosome 7 at the nearest marker, CEL-7_115892950. The BALB/cByJ allelotype at this locus is shown to have a protective effect in a recessive manner.

Figure 7

The results of genetic linkage mapping using bodyweight 96 h after exposure to hyperoxia as the phenotype. A: A genome-wide linkage map of weight on postnatal day 16 after exposure to the oxygen-induced retinopathy model. The threshold of significance was found to be 3.86 (α=0.05), as determined by permutation tests. B: Shows linkage across chromosome 5, with significant linkage at marker CEl-5_93945748 (LOD=3.91, p=0.024). C: Effect plot of genotypes at this marker on postnatal day 16; homozygosity for the BALB/cBJ allelotype at this locus is associated with reduced weight in a recessive manner (ANOVA, p<0.001). The BALB/cByJ allele is designated A^B/c, and the C57BL/6ByJ alleles is designated A^C57.