A Strong Genetic Determinant of Hyperoxia-Related Retinal Degeneration on Mouse Chromosome 6

Abstract

purpose. Hyperoxia-related retinal degeneration (HRRD) is a model system in the mouse in which elevated oxygen levels are used to induce retinal degeneration. The hypothesis for the present study was that strain differences in HRRD susceptibility are due to allelic variants of one or more genes in the mouse genome whose human orthologues should be important targets for research and drug development.

methods. C57BL/6J, A/J, or B.A-Chr6 mice were exposed to 75% oxygen (hyperoxia) or room air for 14 days. After death, one eye was fixed and processed for outer nuclear layer (ONL) thickness measurements. The retina and RPE/choroid were separately dissected from the fellow eye and processed for microarray analysis. Single nucleotide polymorphism (SNP) analysis for transcribed sequences from the C57BL/6J and A/J genomes was conducted using the NIH genome site.

results. C57BL/6J mice developed a significant retinal degeneration in the inferior hemisphere after 14 days of hyperoxia. Under identical conditions, A/J mice exhibited only minor changes. A significant genetic effect was located on chromosome 6. SNP analysis of known transcribed sequences on chromosome 6 combined with microarray expression analysis yielded 33 candidate genes.

conclusions. A significant genetic effect of susceptibility to HRRD is located on chromosome 6. In silico analysis of transcribed sequences results in a fairly small number of candidate genes.

A large number of monogenic retinal degenerations (RDs) have been reported in the mouse, and many of these are caused by mutations in genes with human orthologues that are related to known clinical entities. ¹ Retinal degenerations caused by multiple genes and/or environmental factors, however, are more complicated to study, both in humans and in mice. 

Although the complex retinal degenerations that are currently being studied in the mouse ^{2 3} may not precisely mimic human diseases, genetic studies in the mouse are made easier by exact environmental controls, the availability of many inbred strains with various physiological properties, and specialized recombinant inbred and chromosome substitution strains. ⁴ Recently developed informatics tools also take advantage of the published genome sequences of many of these strains to identify single nucleotide polymorphisms (SNPs), which may be the source of allelic variation leading to variable phenotypes. The rationale for studying complex retinal degenerations in the mouse is the expectation that at least some of the genes to be identified have human orthologues, which are also important in complex human retinal degenerations. 

Yamada et al. ^{5 6} first described hyperoxia-related retinal degeneration (HRRD). This degeneration was originally observed in the C57BL/6J mouse after a 14-day exposure to 75% oxygen (hyperoxia). ⁵ At this time point, animals exhibited a significant thinning of the outer nuclear layer (ONL) in the posterior retina of the eye. Recently, Walsh et al. ⁷ have identified BALB/cJ as a mouse strain that is resistant to HRRD when compared with C57BL/6J. 

The observation of a strain difference in susceptibility to HRRD suggests that this retinal degeneration can be studied by quantitative genetic approaches to identifying underlying genes. Our long-term goal is to identify one or more genes with alleles that are responsible for susceptibility to HRRD. We report the initial measurement of susceptibility to HRRD as a quantitative trait. We then present a new strain difference between C57BL/6J and A/J and the mapping of a significant genetic effect to chromosome 6 by the use of B.A chromosome substitution strains (CSSs). Finally, sequence and microarray expression analyses comparing transcribed sequences on chromosome 6 in the C57BL/6J and A/J mouse that yield an initial set of candidate genes are presented. 

Materials and Methods

Experimental Animals

C57BL/6J, A/J, and B.A-Chr6 (a CSS) mice at 4 to 6 weeks of age were purchased from The Jackson Laboratory (Bar Harbor, ME). Mice were housed under a 12-hr light–dark cycle with an illumination exposure of 2 to 7 foot candles. A standard rodent chow (Rodent Diet 5001; PMI Feeds, Inc.) and water (Napa Nectar; Systems Engineering, SE Laboratory Group Inc., Napa, CA) was provided ad libitum. All procedures conformed to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. These procedures were also authorized by the Internal Care and Use Committee for Laboratory Animals at the University of California, Davis. 

Hyperoxia Exposure

Mice were exposed to hyperoxia (75% oxygen) or room air (21% oxygen) for up to 14 days. The hyperoxia chamber was a standard 7-L induction chamber used for small animal anesthesia (VetEquip Inc., Pleasanton, CA). A mixture of medical grade oxygen and air was bubbled through water to humidify the air mixture entering the hyperoxia chamber. The hyperoxia chamber had 17 volume changes of exposure gas per hour. In addition, the percentage of oxygen exiting the chamber was continuously monitored (MiniOX I oxygen analyzer; MSA Instrument Division, Pittsburgh, PA). The chamber was opened daily to assess animal health and as necessary to feed and water the mice and clean the cages. 

Tissue Processing and Histology

The mice were killed with CO₂. Eyes were fixed and processed according to published procedures. ^{8 9} One eye from each mouse was fixed for 24 hours in 2% paraformaldehyde/2.5% glutaraldehyde in phosphate buffer. The eye was then bisected from the superior to inferior region through the optic disc, the lens was removed, and the tissue was embedded in a PolyBed 812/Araldite mixture. Single sections (1 μm) cut along the vertical meridian of the eye were then stained with 1% toluidine blue containing 1% sodium borate. 

The remaining eye was processed for RNA isolation as described previously with minor modifications. ¹⁰ The eye cup was immobilized to the dissecting Petri dish by a drop of Superglue (QuickTite; Loctite Corp. for Manco, Inc., Avon, OH), which allowed rapid dissection of the eye. Retinal tissue and the RPE/choroid were taken out sequentially, immediately immersed into 300 μL of RNA stabilizer (RNALater; Ambion, Austin, TX), and stored at −20°C. 

Quantification of ONL Thickness

The thickness of the ONL was quantified by an image processing method with a microscope and 40× objective (Nikon, Tokyo, Japan), in conjunction with image-analysis software (ImageProPlus; Media Cybernetics, Silver Spring, MD). ONL thickness was measured in nine superior and nine inferior fields, with 3 measurements taken in each field, for a total of 27 measurements in each of the superior and inferior hemispheres of the eye. ^{8 9} One section through the optic disc in a superior-to-inferior orientation was used for quantification of each eye. 

RNA Isolation

Samples for microarray analysis were collected from three animals from each of the treatments (normoxia and hyperoxia), from two eye tissues (retina and RPE/choroid), and from two strains (C57BL/6J and B.A-Chr6). Before RNA isolation, tissue was homogenized with a 22-gauge needle and a spin column (QIAshredder; Qiagen, Valencia, CA). Total RNA was isolated (RNeasy kit; Qiagen) according to a previously published protocol. ¹⁰ The purity of total RNA was initially determined from the A_260/280 ratio with a spectrophotometer (Beckman Coulter Inc., Fullerton, CA). Total RNA quality and concentration were determined using a microfluidics-chip–based assay that can detect nanogram quantities of RNA (RNA 6000 Nano Assay LabChip; Agilent Technologies, Inc., Palo Alto, CA) and run on microfluidics-based platform instrument (2100 BioAnalyzer; Agilent Technologies, Inc.). A typical yield from processing a single retina was 4 μg and 0.4 μg from a single RPE/choroid. 

Microarray Procedures and Data Analysis

Labeled RNA samples were prepared according to the manufacturer’s manual (Affymetrix, Santa Clara, CA). ¹¹ A labeled cRNA probe was prepared from the 16 highest-quality total RNA samples (two replicates per treatment/tissue/strain), and run on a total of 16 mouse genome microarrays (Mouse Genome 430A_v2.0 GeneChips; Affymetrix). We followed the manufacturer’s protocol for small sample which required only 0.100 μg of RNA/sample (GeneChip Eukaryotic Small Sample Target Labeling Protocol; Affymetrix). Hybridization, washing, and staining were performed on the fluidics station (GeneChip Fluidics Station 400; Affymetrix) according to the manufacturer’s standard protocols. Hybridized microarrays were then scanned using a microarray scanner (Agilent GeneArray Scanner; Agilent Technologies, Inc.). 

The acquisition, processing, and basic analysis of microarray-generated data were performed using the manufacturer’s gene expression analysis software (Affymetrix Gene Chip Operating Software [GCOS] and Gene Chip RNA Expression Analysis Software [GREX]), and raw data were normalized based on 100 housekeeping genes present on the microarray, and scaled to an average median signal value of 500. The GCOS algorithm was used to evaluate the abundance of each transcript represented on the array and labeled it as present (P), absent (A), or marginal (M). Comparison of normoxic and hyperoxic data was performed with GREX software (Affymetrix) generating signal log ratios (SigLogRatios) and significance estimates. SigLogRatios for biological replicates were averaged and SDs were calculated. Generally, SigLogRatio values greater or equal to 0.5 were considered to show upregulation, values smaller or equal to −0.5 downregulated, and values in between −0.5 and 0.5 as no change (NC), although GREX data analysis algorithm would make the final call taking into account the given probe’s signal/noise ratio. 

SNP Analysis

Single nucleotide polymorphism (SNP) analysis in the expressed sequences was performed using the NCBI SNP database, Build 125 (www.ncbi.nlm.nih.gov/SNP/MouseSNP/ provided by the National Center for Biotechnology Information, Bethesda, MD). We compared C57BL/6J and A/J strains, limiting retrieval to chromosome 6. Only those genes affected by SNPs and expressed (as confirmed by microarray analysis) in the RPE/choroid or retina were retained in the analysis. 

Results

HRRD in the C57BL/6J Mouse

Yamada et al. ^{5 6} previously reported degeneration of the C57BL/6J mouse posterior retina after exposure to 75% oxygen for a period of 2 weeks. Our initial goal was to reproduce this result in our own laboratory. Figure 1shows a schematic of the hyperoxia exposure chamber that we used. The design utilizes constant flow to replace the entire chamber volume every 3.5 minutes (17 volume changes/h). This approach was taken to assure minimal accumulation of CO₂ in the chamber. Oxygen content was measured at the gas outlet of the chamber and adjusted on a regular basis to ensure a true level of 75% ± 1.5% oxygen. 

After 14 days of exposure, the control and experimental animals were killed and the eyes were prepared for quantification of the ONL thickness by the standard methods previously detailed by LaVail et al. ^{8 9}  

Figures 2A and 2Bshow photomicrographs of the C57BL/6J retina at 14 days after hyperoxia exposure as well as normoxic controls. These photomicrographs focus on the inferior hemisphere of the eye and illustrate a significant loss of ONL thickness in this region for hyperoxia versus control exposed eyes. At 10 days of hyperoxia, little or no thinning of the ONL was apparent in the inferior hemisphere (data not shown). 

Evaluation of the A/J Mouse Strain

We chose to evaluate the A/J strain for several reasons. It has been paired with C57BL/6J mice in many studies exploring physiology and oxidative stress. The entire sequence of the A/J genome has been determined, and several specialized substrains have been constructed for the purpose of quantitative genetics. ⁴  

Figure 2shows a comparison of the C57BL/6J (Fig. 2A 2B)and A/J (Figs. 2C 2D)strains after 14 days of hyperoxia versus control mice. It is clear from these micrographs that the A/J strain is much less susceptible to HRRD at 14 days when compared with the C57BL/6J strain. 

Quantification of ONL Thickness in the C57BL/6J and A/J Strains after HRRD

To put the observed degeneration of the inferior hemisphere of the retina on a quantitative basis, we used the standard procedures first described by LaVail et al. ^{8 9} Figure 3illustrates ONL thickness measurements for C57BL/6J (Fig. 3A)and A/J (Fig. 3B)mice under hyperoxic and normoxic conditions. This figure clearly demonstrates a region in the inferior hemisphere (enclosed by the box in both figures) in which a large degeneration occurred in the C57BL/6J mouse relative to the A/J mouse. Note that the data represent six individuals for each point, and that the variance of the data for these experiments gives the experiments sufficient power for this determination. 

Identifying a Chromosome with Significant QTL(s) for HRRD

Nadeau et al. ⁴ have devised a method to map significant QTLs to individual chromosomes, utilizing chromosome substitution strains (CSSs). These strains are bred between a specific pair of background and donor strains such as C57BL/6J and A/J. In each substrain, one specific chromosome from the background strain is replaced by the equivalent chromosome from the donor strain. 

We initially planned to survey the entire panel of B.A CSSs. To maximize the speed with which we might find a significant QTL, we prioritized CSSs by the chromosomal assignment of QTLs for oxidative stress–related phenotypes in other systems. Chromosome 2 was chosen first, as it contains the gene for Nfe2l2 (the mouse homolog of NRF2), a quantitative trait gene for other hyperoxia-related traits. ¹² An initial evaluation of C57BL/6J versus B.A-Chr2 indicated no significant difference between these strains with respect to HRRD susceptibility (data not shown). Our second choice was chromosome 6. Several QTLs for oxidative stress–induced phenotypes related to ozone and nickel treatment, as well as several QTLs being investigated in photoreceptor degeneration in the mouse are located on chromosome 6. ^{2 3 13 14}  

Our evaluation of B.A-Chr6 was striking (Figs. 2E 2F) . This CSS’s phenotype appeared to be much more like the A/J strain than the C57BL/6J. There was little to no apparent HRRD after 14 days. Figure 2shows photomicrographs for a standard vertical section through the optic disc for the C57BL/6J, A/J, and B.A-Chr6 strains. A large degeneration is apparent for the C57BL/6J strain but not the B.A-Chr6 CSS. Quantification of these results is given in Figure 3 . 

Finally, in Figure 4 , the quantitative HRRD data are presented in a normalized fashion for comparison among strains and F₁ animals from reciprocal C57BL/6J X B.A-Chr6 crosses. 

All pair-wise relationships were tested for significant differences. The phenotype of the B.A-Chr6 strain was not significantly different from the donor A/J strain, indicating that a substantial genetic effect was located on chromosome 6. Hyperoxia exposure of F₁ animals from reciprocal C57BL/6J X B.A-Chr6 crosses revealed a dominant/additive mode of inheritance of C57BL/6J susceptibility to HRRD. To frame the scope of the differences between C57BL/6J and A/J on chromosome 6, we next pursued an in silico analysis. 

SNPs on Chromosome 6 between the C57BL/6J and A/J Strains for Transcribed Sequences

We analyzed the C57BL/6J and A/J sequences using the NIH site Entrez SNP for RefSNPs on chromosome 6 (http://www.ncbi.nlm.nih.gov/SNP/MouseSNP.cgi/ Provided by the National Center for Biotechnology Information, Bethesda, MD) for the transcribed sequences for all known genes and found a total of 911 in db SNP Mouse Build 125 (Table 1) . 

Of the total number of SNPs found, 49 were located in the coding region of known genes. Nonsynonymous changes in the coding sequence were the result of 24 of these SNPs, whereas the remainder represented silent or synonymous changes. The analysis further yielded the presence of 382 SNPs in the 5′ and 3′ UTRs of mRNAs and 412 SNPs within introns in known genes. 

Expression Analysis for Genes Containing SNPs

Individual genes can have more than one SNP. These SNPs can be grouped together within identified genes. Not all identified genes will be expressed in any given tissue, however. To discover the total number of genes with SNPs that are actually expressed, either in the retina or the RPE/choroid, microarray analysis was performed. The results of these two analyses are presented in Table 2 . This table gives information only on genes that were found to be expressed and for which in silico analysis indicated the presence of SNPs in the coding region. 

Ten genes were identified that had both evidence of expression and nonsynonymous change in the coding region between the C57BL/6J and A/J strains. These 10 genes are detailed in Table 3 . 

The analysis of SNPs in the 5′ and 3′ untranslated region (UTR) of genes is less straightforward. For this analysis, SNPs were again grouped into genes, and the expression of these genes was examined by microarray analysis. For SNPs in the 5′ and 3′ UTR, however, the alteration of phenotype may lie in the relative levels of normoxic expression, in the location where the gene is expressed (RPE/choroid versus retina), or in the relative change of expression between normoxic and hyperoxic conditions. The details of these analyses are given in Appendix I and Appendix II. Table 4summarizes the data resulting in the addition of 23 new genes to our list of candidates. 

Discussion

Hyperoxia has long been known to result in photoreceptor cell death, ¹⁵ and it occurs in the outer retina of degenerating Royal College of Surgeons (RCS) rat models of retinal degeneration ¹⁶ and P23H mutant rhodopsin transgenic ¹⁷ rats, leading to the idea that the loss of some photoreceptors produces a more toxic environment in the outer retina. Hyperoxia also has been suggested to cause cone cell death late in retinal degenerations that result from rod photoreceptor mutations, ¹⁸ as well as to be involved with the thinning of inner retinal vessels after photoreceptor cell loss in human and animal retinal degenerations. ^{17 19}  

The use of hyperoxia in retinal research is not new. This type of exposure for newborn mouse pups produces what is now considered to be a standard model for retinal neovascularization. ²⁰ Yamada et al. ⁵ first described the progressive degeneration of the posterior retina in the C57BL/6J mouse after 14 days of hyperoxia. Recently, Walsh et al. ⁷ have also described HRRD in the C57BL/6J mouse and identified BALB/cJ as a resistant strain when compared with the C57BL/6J strain. 

Yu et al. ²¹ previously measured the intraretinal Po ₂ in the rat retina. These experiments demonstrated that the Po ₂ at the level of the RPE and photoreceptors was a linear function of inspired Po ₂. This is undoubtedly related to the high Po ₂ in the choroid and the volume of the choroidal circulation. The inner retina, however, displays significant autoregulation as does much of the brain, making the relationship between inspired Po ₂ and tissue Po ₂ a nonlinear function. 

The mechanism by which hyperoxia stimulates oxidative stress has been discussed in the literature. ²² Oxygen is presumed to diffuse into the cell and from the cytoplasm into the mitochondria. At higher than normal Po ₂ inside the mitochondria, various elements of the electron transport chain can directly reduce molecular oxygen to the superoxide anion. This species can then be dismutated into hydrogen peroxide and water among other reactions leading to reactive oxygen intermediates. 

We submitted animals after 14 days of hyperoxia for routine pathology. We were surprised to learn that no other tissues in the body exhibited significant pathology. In particular, the lungs appeared normal. This tissue is clearly massively affected when mice are exposed to 95% oxygen for up to 72 hours. ²³  

Our original intent was to survey the entire B.A CSS panel. We chose a specific order of initial strains hoping to find a significant QTL early in our survey. Chromosome 2 was selected first, as Cho et al. ¹² have demonstrated that a polymorphism in the NRF2 gene is responsible for different susceptibilities to pulmonary damage induced by hyperoxia. NRF2 is a gene that codes for a key transcription factor in the oxidative stress response. We found that B.A-Chr2 had the same phenotype as the C57BL/6J (sensitive). Chromosome 6 was our second choice for a variety of reasons. Previous QTLs on chromosome 6 have been identified for retinal degenerations in the mouse as well as nickel, ozone, and hyperoxia-induced pulmonary damage. ^{3 14 24 25} Our findings on the phenotype of B.A-Chr6 indicate that a significant genetic effect in HRRD susceptibility is accounted for by genes on this chromosome. 

The identification of a single chromosome which has one or more QTLs allows us to focus our attention on only 5% of the genome. Chromosome 6 in the mouse has 150 Mb containing 1822 identified genes according to the current NCBI Mus musculus chromosome 6 map Build 35.1 (NIH Map Viewer http://www.ncbi.nlm.nih.gov/mapview/). It is likely that this represents only 50% of the transcribed sequences, when undiscovered novel genes, new exons, antisense RNAs, and various types of noncoding but functional RNA species are considered. ²⁶  

Our analysis was limited to transcribed sequences, and we therefore did not examine polymorphisms in promoter sequences. This is a serious limitation of our study, and a SNP between the C57BL/6J and A/J found in any nontranscribed part of the genome could be responsible for the difference in phenotype. 

Given this note of caution, however, mouse geneticists now typically use in silico sequence analysis combined with expression analysis to shorten the list of candidate genes that may be studied in advance of the completion of fine mapping. ²⁷ We performed our SNP analysis using mouse data on the NIH site mentioned earlier. 

Individual candidates can be selected from our list of candidate genes for which enough information exists to make direct follow-up studies appropriate. The gene for paraoxonase 2 (PON2), for example, is identified as a candidate that has a SNP in the 5′ UTR region of the mRNA at the SNP flank position 208 (in the NT_039340.4). This gene is a member of the paraoxonases (PON) gene cluster, which includes Pon1, -2, and -3. Polymorphisms in the human PON1 gene have previously been associated with susceptibility to age-related macular degeneration (AMD). ^{28 29} The product of PON1 is a protein bound to lipoprotein complexes in the serum; its function is to reduce oxidized lipid species. PON1 polymorphisms have been associated with other diseases such as atherosclerosis, ³⁰ and the common role of oxidative stress in all these phenotypes suggests that this may be a susceptibility gene with influence on a variety of diseases. 

Our next studies will therefore concentrate on the genetic fine mapping of chromosome 6. Ultimately, genetic proof from knock-in or complementation testing by transgenesis with appropriate BAC clones will be necessary to prove that any set of SNPs or genes is responsible for phenotypic variation in HRRD. 

 Supported by an unrestricted grant from Research to Prevent Blindness (Department of Ophthalmology, University of California, Davis), National Eye Institute (NEI) Grants 5R01EY006473-17 (LMH) and R01EY01919 (MML), NEI Core Grants P30EY12576 and P30EY02162, Research to Prevent Blindness, and the Foundation Fighting Blindness.

 Submitted for publication July 24, 2006; accepted November 27, 2006.

 Disclosure: Z. Smit-McBride, None; S . L. Oltjen, None; M . M. LaVail, None; L . M. Hjelmeland, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Schematic diagram of the hyperoxia chamber.

Figure 2.

 

Photomicrographs of the inferior posterior region of C57BL/6J, A/J, and B.A-Chr6 mouse retinas after 14 days of hyperoxia and control sections. Control C57BL/6J mouse retina in room air (A). C57BL/6J mouse retina after 14 days of 75% oxygen exposure (B). Control A/J mouse retina in room air (C). A/J mouse retina after 14 days of 75% oxygen exposure (D). Control B.A-Chr6 mouse retina in room air (E). B.A-Chr6 retina after 14 days of 75% oxygen exposure (F).

Figure 3.

 

Retinal topography of HRRD in the C57BL/6J, A/J and B.A-Chr6. ONL thickness measurements for the C57BL/6J (A), A/J (B), and B.A-Chr6 (C) strains are plotted. These representations of retinal topography are referred to as spider graphs. Box: three points from the inferior retina that showed the greatest differences and were used to compare the control and hyperoxia groups.

Figure 4.

 

Graphic representation of the comparison of the posterior/inferior ONL thickness measurements for the C57BL/6J, A/J, and B.A-Chr6 strains and F₁ animals from the reciprocal C57BL/6J × B.A-Chr6 cross. Individual values for each animal represent the average ONL thickness of three fields as indicated in the box in Figure 3 . The values for each strain were normalized to 100% for mice exposed to normoxia (control). Each treatment had an n= 6. The C57BL/6J control and hyperoxia groups were significantly different (P < 0.001). There was no significant difference between treatments for the A/J or B.A-Chr6 strains.

Table 1.

 

Summary of RefSNPs between C57BL/6J and A/J Strains on Chromosome 6

SNPs	Total (Build 125)
Total on chromosome 6	911
Reference (mainly coding region)	49
Nonsynonymous amino acid changes	24
Synonymous amino acid changes	24
mRNA UTR	382
Introns	412

Table 2.

 

Summary of Genes with RefSNPs between C57BL/6J and A/J Strains on Chromosome 6

Genes	Total (Build 125)
Genes with SNPs causing nonsynonymous amino acid change	13
Expression confirmed	10
Genes with 5′ and 3′ UTR mRNA SNPs	43
Expression confirmed	31
Expression change confirmed	23

Table 3.

 

Genes on Chromosome 6 with Nonsynonymous Amino Acid Sequence Changes

Gene Symbol	Chromosome Location	Entrez Gene ID	Gene Name
Cntn6	6 E2	53870	Contactin 6
Htra2	6 C3|6 34.75 cM	64704	HtrA serine peptidase 2
Loxl3	6 C3|6 34.76 cM	16950	Lysyl oxidase-like 3
Lrmp	6 G3|6 71.0 cM	16970	Lymphoid-restricted membrane protein
Mansc1	6 G1	67729	MANSC domain containing 1
Vwf	6 F3|6 60.8 cM	22371	Von Willebrand factor homologue
A930040G15Rik	6 B2.3	77827	RIKEN cDNA A930040G15 gene
A030007L17Rik	6 B3	68252	Cation transport conserved domain
4930469P12Rik	6 G3	67636	Growth-hormone inducible soluble protein
1100001H23Rik	6 G1	66857	Laminin A conserved domain

 Shown are genes with nonsynonymous amino acid sequence changes between the C57BL/6J and A/J strains (NIH Build 125) expressed in the retina or RPE/choroid.

Table 4.

 

Genes on Chromosome 6 with SNPs in 5′ or 3′ mRNA UTRs

Gene Symbol	Chromosome Location	Entrez Gene	Gene Title
Acrbp	6 F2	54137	Proacrosin binding protein
Anxa4	6D1	11746	Annexin A4
Asb4	6A1|6 0.6 cM	65255	Ankyrin repeat and SOCS box-containing protein 4
C1s	6 F2	50908	Complement component 1, s subcomponent
Cntn4	6E2	269784	Contactin 4
Cntn6	6E2	53870	Contactin 6; NB-3
Dusp11	6D1	72102	Dual sp. phosphatase 11 (RNA/RNP complex 1)
Emp1	6G1	13730	Epithelial membrane protein 1
Eps8	6G1	13860	EGFR pathway substrate 8
Frmd4b	6D3	232288	FERM domain containing 4B
Glcc1	6A1	170772	Glucocorticoid-induced transcript 1
Hirip5	6D2	56748	Histone cell cycle reg. defective interact. protein 5
M6pr	6F2	17113	Mannose-6-phosphate receptor, cation dependent
Mgl1	6D1	23945	Monoglyceride lipase
Mrpl19	6 C3|6 34.55 cM	56284	Mitochondrial ribosomal protein L19
Nr2c2	6D1	22026	Nuclear receptor subfamily 2, group C, member 2
Pon2	6A1	330260	Paraoxonase 2
Ptn	6B1	19242	Pleiotrophin
Repin1	6B2.3	58887	Replication initiator; IL-1R binding
Tnfrsf1a	6F3	21937	TNFR superfamily, member 1a
Tspan9	6F3	109246	Tetraspanin9
2700089E24Rik	6G1	381820	RIKEN cDNA 2700089E24
1810060J02Rik	6G3	67015	Protein transport in Golgi; GGA binding partner

 Shown are genes with SNPs in 5′ or 3′ mRNA UTRs between the C57BL/6J and A/J strains (NIH Build 125) with changes of expression in retina or RPE/choroid between C57BL/6J and B.A-Chr6 strains in either normoxia or hyperoxia.

Supplementary Materials

Appendix I - (PDF) 

Appendix II - (PDF) 

The authors thank Douglas Yasumura, Haidong Yang, and Michael Matthes for superb technical assistance with the quantification of ONL thickness.