Genome-Wide Expression Profiling of the Retinoschisin-Deficient Retina in Early Postnatal Mouse Development

Abstract

purpose. The Rs1h knockout mouse displays retinal features typical for X-linked juvenile retinoschisis (RS). Consequently, this mouse line represents an excellent model to study early molecular events in RS.

methods. Whole genome expression profiling using DNA-microarrays was performed on total RNA extracts from retinoschisin-deficient and wild-type murine retinas from postnatal days 7, 9, 11, and 14. Quantitative real-time RT-PCR (qRT-PCR) analysis of additional time points facilitated the refinement of the temporal expression profile of differentially regulated transcripts. Differential protein expression was confirmed by Western blot analysis.

results. Based on biostatistic and knowledge-based DNA-microarray analyses we have identified differentially regulated retinal genes in early postnatal stages of the Rs1h-deficient mouse defining key molecular pathways including adhesion, cytoskeleton, vesicular trafficking, and immune response. A significant upregulation of Egr1 at P11 and several microglia/glia-related transcripts starting at P11 with a peak at P14 were identified in the diseased retina. The results provided evidence that macrophage/microglia activation precedes apoptotic photoreceptor cell death. Finally, the role of Egr1 in the pathogenesis of Rs1h-deficiency was investigated, and the results indicated that activation of the MAPK Erk1/2 pathway occurs as early as P7. Analyses of Rs1h ^−/Y /Egr1 ^−/− double-knockout mice suggest that Egr1 upregulation is not a prerequisite for macrophage/microglia activation or apoptosis.

conclusions. The findings imply that microglia/glia activation may be triggering events in the photoreceptor degeneration of retinoschisin-deficient mice. Furthermore, the data point to a role of Erk1/2-Egr1 pathway activation in RS pathogenesis.

Retinoschisis (RS) is a recessively inherited retinal dystrophy with macular disease often resulting in early-onset vision loss. A hallmark of the disease is a limited splitting of the central retina typically presenting as a spoke-wheel pattern. ¹ Peripheral schisis is seen in approximately 50% of affected males and is usually located in the inferior temporal quadrant. RS is caused by loss-of-function mutations in the RS1 gene on Xp22.13. ² It encodes a 24-kDa protein, termed retinoschisin, which is mainly secreted from photoreceptors as a homo-oligomeric complex. ^{3 4 5 6 7} Retinoschisin is almost entirely composed of a highly conserved discoidin domain frequently found in a wide range of membrane and extracellular proteins and most likely mediating cell adhesion and cellular signaling processes. ^{8 9}  

To study retinoschisin function, we have generated a gene-targeted mouse line with a disruption of Rs1h, the murine orthologue of the human RS1 gene. ¹⁰ The knockout mouse appears to be an excellent disease model that closely resembles human disease. For example, hemizygous male mice reveal a characteristic “negative electroretinogram” and a highly disorganized retinal architecture with schisis cavities within the inner nuclear layer. This is accompanied by progressive loss of cone and rod photoreceptor cells. ¹⁰ Recently, we have shown that the photoreceptor degeneration is due to apoptosis, more specifically to the receptor mediated extrinsic pathway, with a major burst of dying cells around postnatal day 18. ¹¹  

To elucidate the molecular events that precede photoreceptor disease in the Rs1h knockout mouse, we applied DNA-microarray-based mRNA profiling in retinal tissue of postnatal stages (P)7, 9, 11, and 14. Besides Rs1h, we identified several differentially expressed transcripts with a functional annotation to adhesion, cytoskeleton, vesicular trafficking, and immune response. Most of the genes in these groups are regulated before the onset of an appreciable expression of genes involved in apoptosis. Notably, many genes are implicated in microglia/glia activity and inflammatory processes. An upregulated expression of the early growth response gene 1 (Egr1) was found relatively early in the disease process at P9. This has further prompted us to investigate the role of Egr1 in RS disease by generating Rs1h^−/Y/Egr1^−/− double-knockout mice. 

Materials and Methods

Animals

Animals were maintained in an air-conditioned environment on a 12-hour light–dark schedule at 20°C to 22°C, and had free access to food and water. The health of the animals was regularly monitored, and all procedures strictly adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. The Egr1 ^{+/ −} mouse was purchased from Taconic (Ry, Denmark) and the Rs1h ^{− /Y} mouse has been described previously. ¹⁰ These mice were on a C57BL/6 background and backcrossed for 12 generations. 

RNA Isolation

Retinal tissue was isolated from eye bulbs and purified under the microscope from contaminating vitreous body and retinal pigment epithelium/choriocapillaris. Total RNA was extracted (RNeasy Mini Kit; Qiagen, Hilden, Germany). 

Microarray Analysis and Statistical Data Evaluation

Generation of double-stranded cDNA, preparation and labeling of cRNA, hybridization to 430A 2.0 or 430 2.0 mouse genome arrays (Affymetrix, Santa Clara, CA), washing, and scanning were performed according to the protocol. Scanned images were analyzed (Microarray Suite version 5.0; Affymetrix), to generate report files for quality control. For data analysis, expression levels of single probes derived from Affymetrix CEL-files were evaluated (ChipInspector software; Genomatix GmbH, Munich, Germany). Previous annotations of the single oligonucleotide probes by Affymetrix were disregarded, together with the grouping of the probes in probe sets. The sequence of each single probe was mapped against the current genome annotation (Build 36; National Center for Biotechnology Information [NCBI], Bethesda, MD) and only probes with uniqueness in the genome were used for analysis. Furthermore, single nucleotide polymorphisms were accounted for. Thereafter, ratios of single probe signals were calculated and logarithmic transformation was performed. For normalization, a linear total-intensity normalization algorithm was used. Significantly regulated transcripts were discovered by a single sided permutated t-test with a false discovery rate (FDR) calculation using the significance analysis of microarray (SAM) algorithm.¹²Scores were calculated for each gene based on the multiple of expression level changes relative to the standard deviation of repeated measurements. To estimate FDR, nonsense genes were identified by analyzing permutations of the measurements. An FDR of 5% was chosen in the first step of our statistical analysis and lists of genes that fulfill these criteria for data sets at P7, P9, P11, and P14 are given in Supplementary Tables S1, S2, S3, and S4, respectively. To increase stringency, a 1.8 fold-change and a minimum signal intensity of 50 were selected as inclusion criteria (Table 1) .¹³Functional annotation of transcripts was performed on computer (BiblioSphere, Pathway Edition; Genomatix, Munich, Germany).¹⁴The microarray data set of this study is publicly available at the NCBI Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/) through accession number GSE5581. 

Quantitative Real-Time RT-PCR Analysis

Total RNAs from time points P5, P7, P9, P11, P13, P14, P15, P18, and P22 from Rs1h ^{− /Y} mice and RNAs from time points P18 and P24 from Rs1h ^{− /Y}/Egr1 ^{+/ −} and Rs1h ^{− /Y}/Egr1 ^{− / −} mice were used to generate cDNA. For each condition two animals (four eyes) were analyzed in triplicate measurements. First strand cDNA synthesis was performed (Superscript II; Invitrogen, Karlsruhe, Germany). Quantitative real-time RT-PCR (qRT-PCR) was performed with a thermocycler (iCycler; Bio-Rad, Munich, Germany) as described previously, ¹¹ using primers listed in Supplementary Table S5. Input cDNA quantities were estimated by comparison to expression of five reference genes (Atp5b, Gusb, Hprt1, Rpl4, and Sdha). ¹⁵  

SDS Gel Electrophoresis and Western Blot Analysis

Retinal tissue was homogenized in buffer (PhosphoSafe; Novagen, Madison, WI). Quantitation of protein extracts was performed by a protein assay based on the Bradford method (Bio-Rad). Western blots were prepared after SDS-PAGE of 20 μg protein and labeled with phospho-ERK 1/2 pAb (dilution 1:1.000; Sigma-Aldrich, Munich, Germany), ERK 1/2 pAb (dilution 1:2.000; Sigma-Aldrich), or β-actin mAb (dilution 1:60.000; Sigma-Aldrich). After the blots were washed in PBS/Tween-20, they were labeled with goat anti-rabbit and goat anti-mouse Ig-peroxidase (dilution 1:10.000; Calbiochem, La Jolla, CA) for detection by enhanced chemiluminescence (ECL). 

Immunofluorescence Labeling and TUNEL-Assay

Immunofluorescence studies of retinal cryosections and TUNEL-staining of eye bulbs from Rs1h-deficient, Egr1-deficient (Taconic) or double-knockout mice (Rs1h^−/Y/Egr1^+/− and Rs1h^−/Y/Egr1^−/− ) was performed exactly as described recently. ¹¹  

Results

Microarray Analysis

To study molecular events accompanying retinoschisin deficiency, we first analyzed gene expression in retinal tissues from Rs1h-knockout and wild-type mice at P7, P9, and P11. In wild-type animals, expression of Rs1h mRNA is detectable at P5 and reaches adult levels at P11 (data not shown). Retinal cRNAs from three Rs1h ^{− /Y} mice and three wild-type littermates at P7, P9, and P11 were hybridized to microarrays (Mouse Genome 430A 2.0 GeneChips; Affymetrix) detecting 14,000 annotated mouse genes. For a second round analysis of stage P14, cRNAs from six Rs1h-knockout mice and seven wild-type littermates were hybridized to microarrays (Mouse Genome 430 2.0 GeneChips; Affymetrix) covering more than 34,000 mouse genes. 

Initial evaluation of single probe expression ratios with a false discovery rate (FDR) of 5% based on SAM analysis indicated differential expression of 55 genes on P7, 60 genes on P9, 87 genes on P11, and 262 genes on P14 (for gene lists see Supplementary Tables S1–S4). Based on own experience with previous microarray studies,¹³we next applied higher-stringency conditions selecting for genes with at least 1.8-fold expression changes and minimum signal intensities of 50 signal units. Thereafter, the number of significantly regulated genes was reduced to 17 for P7, 24 for P9, 27 for P11, and 41 for P14 (Table 1) . The top regulated genes apart from Rs1h were Tlr4 (4.0-fold) and Wasf3 (−2.5-fold) at P7, Ubqln2 (2.0-fold) and Ndrg3 (−4.2-fold) at P9, Crip3 (3.7-fold) and the unknown gene 4632413K17 (−3.0-fold) at P11, and Spp1 (4.3-fold) and Cryaa (−2.2-fold) at P14. 

To characterize further the observed mRNA regulations, molecular function annotation of pathways was performed. As displayed in Figure 1 , grouping of genes into pathways with at least two entries recurring in the time course (colored segments) yielded the main categories cytoskeleton, adhesion, vesicular trafficking, immune response, and apoptosis. At P7, genes involved in adhesion and vesicular trafficking were prominent, whereas at later stages immune response and microglia/glia-specific genes were mainly detected. Genes regulated at several time points include the induced immune regulatory gene Tlr4 and the downregulated vesicular trafficking gene Sh3gl2 (Table 2) . All commonly regulated transcripts were found at P11 and we therefore conclude that these molecular events represent initial as well as secondary changes related to Rs1h deficiency. The large number of upregulated genes at P14 related to immune response also indicates a significant involvement of microglia and glia cells in RS. Because transcription factors are of particular importance for regulatory responses, the upregulation of Egr1 was verified by qRT-PCR along with the immune response genes induced at P14 (Table 1and Fig. 1 ). Six transcripts which are increased at P14, namely Spp1, Emp1, S100a6, Lgals3, Cxcl12, and Vim contain verified Egr1 binding sites in their regulatory regions, suggesting that early Egr1 expression could induce transcriptional waves at later time points. 

Refined Expression Profiling by qRT-PCR and Immunohistochemistry

We next examined the upregulated transcripts at postnatal stages P7, P9, P11, P13, P15, P18, and P22 by qRT-PCR (Fig. 2) . Egr1 was significantly induced in the retinoschisin-deficient retina starting at P9 (P < 0.05), further upregulated at P11 and P13 (P < 0.001 and P < 0.05, respectively), unchanged at P15, and again strongly induced at P18 and P22 (Fig. 2a) . We then quantified mRNA levels of 1700063D05, a gene with an as yet unknown function (Fig. 2b) . Upregulation of 1700063D05 transcripts started in Rs1h knockout mice at P13 but showed a marked decline from P15 to P22. The expression profile 1 in Figure 2csubsumes the mRNA pattern of 10 upregulated genes identified in the DNA-microarray experiments at P14 including Spp1, Tyrobp, Clec7a, Cd68, Ccl6, Fcer1g, Lyzs, Mpeg1, Msr2, and Cd53 (Table 1 and Supplementary Table S4) comprising microglia-associated genes. There is a steady increase in gene expression starting between P11 and P13 with a maximum at P18 and a decline beyond this time point. Expression profile 2 (Fig. 2d)represents the expression pattern of five genes including S100a6, Cyp26a1, Tgm2, Tagln2, and Gfap revealing a sharp upregulation between P15 and P18 and elevated levels until P22. 

We then sought to define precisely the chronological expression of microglia-associated genes exemplified by Cd68 and Clec7a in comparison with transcripts involved in the extrinsic pathway of apoptosis (Casp8 and Tnfrsf6). There was a steady increase of expression profile 1 transcripts beginning at P13 while apoptosis-related upregulation started not until P16 (Fig. 3) . These data clearly indicate that microglia-related gene expression precedes expression of apoptosis-related transcripts in the retinoschisin-deficient retina by several days. 

To correlate mRNA profiles with protein expression, we performed immunohistochemistry on Rs1h knockout and wild-type retinal sections from three developmental stages (P14, P18, and P24). Anti-F4/80 was used as a marker of microglia activation, whereas the glial fibrillary acidic protein (Gfap) served as a stress indicator of retinal Müller glial cells. At P14, expression of Gfap (green) was similar in the retinoschisin-deficient and the wild-type retina, whereas the activation marker F4/80 (red) indicated isolated microglia in the inner nuclear layer of the Rs1h ^{− /Y} retina (Fig. 4 , left). At P18 and P24 there was prominent labeling of the two marker proteins in the Rs1h-knockout tissue (Fig. 4 , middle and right) fully corroborating our qRT-PCR data (Figs. 2c 2d 3) . 

Analysis of the Erk1/2 Pathway in the Retinoschisin-Deficient Retina

Our data demonstrate that upregulation of Egr1 at P9 may contribute to the pathogenesis of the retinoschisin-deficient retina. Egr1 expression is induced by external signals including growth factors, hormones, neurotransmitters, or toxins via the serine-threonine kinases Erk1 and Erk2 in the mitogen-activated protein kinase (MAPK) pathway. ^{16 17} We therefore analyzed the phosphorylation status of Erk1/2 in the Rs1h-deficient retina at stages P5, P7, P11, and P13 (Fig. 5) . Western blot analysis revealed a significant increase in Erk1/2 phosphorylation (p-Erk1/2) at P7, P11, and P13 but not at P5 in retinal extracts from retinoschisin-deficient versus wild-type mice (Fig. 5) . 

Characterization of the Rs1h ^{− /Y}/Egr1 ^{− / −} Double-Knockout Mouse

To investigate the role of Egr1 in RS, we generated double-knockout mice deficient for both retinoschisin and Egr1. Histologic examination of retinal sections from double knockout mice at P18 and P24 showed pathologic manifestations typical for the single Rs1h knockout. ¹⁰ Furthermore, three knockout genotypes (Rs1h ^+/Y/Egr ^{− / −}, Rs1h ^{− /Y}/Egr ^{+/ −}, and Rs1h ^{− /Y}/Egr ^{− / −}) were analyzed at P18 and P24 for mRNA expression of the upregulated marker genes 1700063D05, Cd68, and Gfap (Figs. 6a 6b 6c) . qRT-PCR demonstrates a significant increase of all three transcripts in the Rs1h-deficient retina independent of the Egr1 genotype (Egr1 ^{+/ −} or Egr1 ^{− / −}) similar to the data obtained for the Rs1h single knockout. This suggests that partial or full Egr1 deficiency does not prevent induction of microglia/glia activation markers in the Rs1h ^{− /Y} retina. 

We finally explored possible effects of Egr1-deficiency on the extrinsic apoptotic pathway and identified that the apoptosis markers Tnfrsf6 and Casp8 are upregulated in the retinoschisin-deficient retina, independent of Egr1 expression status (Egr1 ^{+/ −} or Egr1 ^{− / −}; Figs. 6d 6e ). To further corroborate these findings, TUNEL assays were performed in retinal sections of the three genotypes Rs1h ^+/Y/Egr^−/−, Rs1h ^{− /Y}/Egr ^{+/ −}, and Rs1h ^{− /Y}/Egr ^{− / −} at developmental stages P18 and P24. The number of apoptotic nuclei was significantly increased in Rs1h ^{− /Y} versus wild-type retinas, either on the Egr1 ^{+/ −} or the Egr1 ^{− / −} background (Fig. 7) . This suggests no direct link between transcriptional Egr1 upregulation and apoptotic events in the retinoschisin-deficient mouse. 

Discussion

In the present study, we performed microarray analysis and qRT-PCR in the Rs1h^−/Y retina compared with wild-type. At P7, where Rs1h is normally present and where early events in pathogenesis are expected, transcripts for adhesion molecules, photoreceptor-specific proteins, cytoskeleton, and neuroprotective proteins were detected. Among these, Nell2 is a neuron-specific secreted glycoprotein protecting against oxygen-glucose-deprivation- and β-amyloid-induced cell death, ¹⁸ whereas overexpression of the photoreceptor protein recoverin leads to photoreceptor degeneration. ¹⁹ The strongest upregulated gene at P7, Tlr4, is expressed in retinal pigment epithelium ²⁰ and in activated microglia. ²¹ Furthermore, human TLR4 genetic variants are associated with age-related macular degeneration, ²² suggesting functional consequences of these aberrant transcript patterns. 

We have identified 10 genes gradually increasing in expression starting at P11 (profile 1), a steep upregulation of 1700063D05 at P13, and induction of five genes in profile 2 (1700063D05, Cyp26a, S100a6, Tagln2, and Gfap) between P15 and P18. Notably, P14 is a time point where first histopathological features such as cystlike structures in the inner retina, disruption of the outer plexiform layer, and disorganization and displacement of retinal cells become apparent in Rs1h^−/Y animals. ¹⁰ Although 1700063D05 is of unknown function, Cyp26a1 plays a role in retinoic acid metabolism, ²³ S100a6 is functionally involved in cell cycle progression, ^{24 25} and Tagln2 represents a marker of smooth muscle differentiation. ²⁶ Whereas a correlation between RS and upregulation of transcripts 1700063D05, Cyp26a, S100a6, and Tagln2 may be challenging to reconcile, increased expression of the intermediate filament protein Gfap most likely represents a reactive response of Müller glia cells and astrocytes to the degenerative processes caused by the lack of retinoschisin. Accordingly, consistent upregulation of Gfap in other models of retinal injury has been reported, including glaucoma, ^{27 28} retinal tear injury, ²⁹ and diabetic retinopathy. ³⁰  

The expression data at time points P11 and P14 clearly implicate microglia-mediated processes in pathologic events of retinoschisin deficiency. Several genes are related to immune response and are expressed in microglia. Clec7a is a pattern-recognition receptor detecting β-glucans from fungi and plants, resulting in inflammation ^{31 32} and Fcer1g encodes the gamma chain of the IgE receptor CD23, which is important for allergic reactions. ³³ Lysz is an inducible marker of phagocytosis expressed during microglia activation after neuronal injury. ³⁴ The chemokine Ccl6 acts as mediator of astrocyte and microglia migration in vitro, ³⁵ whereas the secreted phosphoprotein Spp1, participates in cellular attachment. ³⁶ Egr1 induces the expression of inflammatory mediators such as TNF-α or macrophage inflammatory protein-2 (MIP-2). ^{16 37 38} Egr1, together with Tyrobp, also controls terminal differentiation and activation of macrophages. ^{39 40 41 42}  

Several of these immune-regulatory genes (e.g., Spp1, Lysz, and C1qb) and the immediate early response gene Egr1 have also been associated with retinal disease in DNA-microarray approaches similar to those used in our study, such as analyses of ischemic injury and light-induced damage to the retina. ^{43 44 45} Egr1, which is upregulated at P9 is involved in a large number of cellular events ranging from mitogenesis, differentiation, cell protection and survival, pro- and antiapoptotic processes to macrophage differentiation. ^{40 46} Although not unique to retinoschisin-deficiency, this suggests that Egr1 may act as a convergence point for multiple signaling cascades, which is also supported by our finding that the promoters of six commonly induced genes at P14 related to adhesion, immune response, and cytoskeleton contain verified Egr1-binding sites. It is therefore of great interest to identify further the events upstream and downstream of the initiation of Egr1 transcription. Toward this end, we analyzed the phosphorylation of the MAP Erk1/2 kinases and have found that this pathway was activated in the retinoschisin-deficient retina as early as P7. In mammalian cells, the MEK1/2/ERK signaling pathway plays a crucial role in immediate early gene (IEG) induction by directly activating IEG promoter-bound transcription factors such as Egr1. ⁴⁷ Further studies are needed to determine Egr1-expressing cell types in the diseased retina and to analyze the gene activation cascade upstream of Egr1. So far, the mechanism coupling the extracellular membrane–associated retinoschisin to intracellular signaling remains elusive, but it may be relayed via an as yet unknown membrane receptor protein. 

We were also interested in exploring downstream consequences of Egr1 activation. However, double-knockout mice deficient of both Rs1h and Egr1 were indistinguishable from Rs1h single-knockout mice, specifically with regard to retinal histology and expression profiles of microglia/glia- and apoptosis-related transcripts. Our experiments are in line with findings demonstrating that Egr1 ^{− / −} mice have no obvious defects in macrophage differentiation. ^{48 49} However, data from an antisense strategy demonstrate that Egr1 is essential for cell growth and macrophage differentiation. ^{40 50 51} These controversial results may be explained by functional redundancy of remaining family members Egr2, Egr3, and Egr4 which are often coordinately regulated. ⁵² To further unravel the role of Egr1 in retinoschisin-deficiency, it may therefore be necessary to simultaneously target several Egr genes. 

Commonly, programmed cell death is thought to activate microglia cells resulting in the clearance of dying cells, ⁵³ although in some cases macrophage stimulation conversely has been demonstrated to trigger apoptosis. ^{54 55 56} We recently reported that postnatal photoreceptor cell death in the Rs1h ^{− /Y} knockout mouse is due to extrinsic apoptosis at P18. ¹¹ We have now defined the relative time course of microglia- and apoptosis-related events in this model and show that the two events can be unambiguously separated. Whereas immune-related genes are upregulated starting as early as P11, apoptosis-associated transcripts are increased in the Rs1h-deficient retina only after P16. We therefore conclude that microglia activation precedes the onset of apoptotic cell death and may even be involved in triggering apoptosis in the diseased retina. 

The goal of the study was to analyze molecular events caused by retinoschisin-deficiency in a mouse model for RS. Our findings indicate activation of the Erk1/2 pathway at P7 followed by an upregulation of transcription factor Egr1 at P9. Functional clustering highlights key pathways including adhesion, cytoskeleton, vesicular trafficking and glia-cell related immune responses. As apoptosis-related gene expression in the retinoschisin-deficient retina is not initiated before P16, we hypothesize that microglia/glia-mediated processes precede and may even cause events of programmed photoreceptor cell death. 

 Supported by Grant WE 1259/12-3 from the Deutsche Forschungsgemeinschaft (BHFW), the Interdisciplinary Center of Clinical Research Tübingen (IZKF), and Grant 01KS9602 from the Federal Ministry of Education and Research (MB).

 Submitted for publication June 13, 2006; revised August 23 and October 12, 2006; accepted December 18, 2006.

 Disclosure: A. Gehrig, None; T. Langmann, None; F. Horling, None; A. Janssen, None; M. Bonin, None; M. Walter, None; S. Poths, None; B.H.F. Weber, 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.

 

Table 1.

 

List of Genes Analyzed by DNA-Microarrays and qRT-PCR Analysis in Mutant (Rs1h^−/Y) Versus Wild-Type Retinal Extracts

Age	Unigene ID	Gene Symbol	Gene Name	Function	Array FC
P7*	Upregulated
	Mm.38049	Tlr4	Toll-like receptor 4	Immune response	4.0
	Mm.17484	Snca	Synuclein, alpha	Synaptic vesicle protein	4.0nc
	Mm.3959	Nell2	Nel-like 2 homolog (chicken)	Adhesion	2.3
	Mm.66264	Stx6	Syntaxin 6	Vesicular traffic	2.2
	Mm.333349	Wnk1	WNK lysine deficient protein kinase 1	Signaling	1.9
	Mm.235863	Rcvrn	Recoverin	Photoreceptor protein	1.8
	Mm.35650	Tspan31	Tetraspanin 31	Unknown	1.8
	Mm.201322	Adnp	Activity-dependent neuroprotective protein	Neuroprotection	1.8
	Mm.2020	Csrp2	Cysteine and glycine-rich protein 2	Proliferation/differentiation	1.8
	Mm.233799	Igfbp4	Insulin-like growth factor binding protein 4	Proliferation/differentiation	1.8
	Mm.207367	Zdhhc9	Zinc finger, DHHC domain containing 9	Signaling	1.8
	Mm.1574	Wasl	Wiskott-Aldrich syndrome-like (human)	Cytoskeleton	1.8
	Downregulated
	Mm.12145	Rbbp4	Retinoblastoma binding protein 4	Nuclear protein	−1.9
	Mm.3781	Piga	Phosphatidylinositol glycan, class A	Adhesion	−2.0
	Mm.4559	Ggt1	Gamma-glutamyltransferase 1	Glutatione metabolism	−2.4
	Mm.332849	Wasf3	WAS protein family, member 3	Cytoskeleton	−2.5
	Mm.41982	Rs1h	Retinoschisis 1 homolog (human)	Adhesion	−3.8c
P9*	Upregulated
	Mm.92529	Ubqln2	Ubiquilin 2	Ubiquitin metabolism	2.0
	Downregulated
	Mm.34405	Casp3	Caspase 3	Apoptosis	−1.8
	Mm.103669	Cabp5	Calcium binding protein 5	Calcium binding	−1.9
	Mm.5061	Arf2	ADP-ribosylation factor 2	Vesicular traffic	−2.1
	Mm.256034	Cct3	Chaperonin subunit 3 (gamma)	Chaperon	−2.1
	Mm.378902	Cttn	Cortactin	Cytoskeleton	−2.1
	Mm.278357	Kns2	Kinesin 2	Cytoskeleton	−2.2
	Mm.3527	Ccng2	Cyclin G2	Cell cycle	−2.2
	Mm.41642	Rgs4	Regulator of G-protein signaling 4	Signaling	−2.3
	Mm.379376	Syt11	Synaptotagmin XI	Synaptic vesicle protein	−2.3
	Mm.295284	Stom	Stomatin	Cytoskeleton	−2.3
	Mm.278865	Stxbp1	Syntaxin binding protein 1	vesicular traffic	−2.3
	Mm.2154	Mxi1	Max interacting protein 1	Transcription factor	−2.5
	Mm.3608	Pax6	Paired box gene 6	Transcription factor	−2.6
	Mm.21281	Rnf4	Ring finger protein 4	Transcription factor	−2.7
	Mm.143603	Sh3gl2	SH3-domain GRB2-like 2	Vesicular traffic	−2.7
	Mm.3193	Pkia	Protein kinase inhibitor, alpha	Signaling	−2.7
	Mm.308180	Ptplad1	Protein tyrosine phosphatase-like A domain containing 1	Signaling	−2.7
	Mm.43871	Trim35	Tripartite motif-containing 35	Apoptosis	−2.8
	Mm.4173	Mtap1b	Microtubule-associated protein 1 B	Cytoskeleton	−2.8
	Mm.218820	Dnm1l	Dynamin 1-like	Cytoskeleton	−3.0
	Mm.291928	Ctnnb1	Catenin (cadherin associated protein), beta 1, 88kDa	Cytoskeleton	−3.3
	Mm.41982	Rs1h	Retinoschisis 1 homolog (human)	Adhesion	−4.1c
	Mm.279256	Ndrg3	N-myc downstream regulated gene 3	Proliferation/differentiation	−4.2
P11*	Upregulated
	Mm.25168	Crip3	Cysteine-rich protein 3	Immune response	3.7
	Mm.92529	Ubqln2	Ubiquilin 2	Ubiquitin metabolism	2.2
	Mm.38049	Tlr4	Toll-like receptor 4	Immune response	2.2
	Mm.181959	Egr1	Early growth response 1	Transcription factor	2.0c
	Mm.234266	Sdc2	Syndecan 2	Vesicular traffic	1.9
	Mm.362054	Pigq	Phosphatidylinositol glycan, class Q	Adhesion	1.8
	Downregulated
	Mm.263396	Itgb1	Integrin beta 1 (fibronectin receptor beta)	Adhesion	−1.8
	Mm.5028	Inppl1	Inositol polyphosphate phosphatase-like 1	Immune response	−1.8
	Mm.256034	Cct3	Chaperonin subunit 3 (gamma)	Chaperon	−1.8
	Mm.4950	Glp2	Interferon, alpha-inducible protein	Immune response	−1.9
	Mm.253090	Ap2a2	Adaptor protein complex AP-2, alpha 2 subunit	Vesicular traffic	−1.9
	Mm.24724	Ppp1r3c	Protein phosphatase 1, regulatory (inhibitor) subunit 3C	Glycogen metabolism	−1.9
	Mm.294821	0610042115	RIKEN cDNA 0610042115 gene	Unknown	−1.9
	Mm.30837	Ndrg1	N-myc downstream regulated gene 1	Differentiation	−1.9
	Mm.1635	Pias3	Protein inhibitor of activated STAT 3	Immune response	−1.9
	Mm.38993	Clstn1	Calsyntenin 1	Vesicular traffic	−2.0
	Mm.1682	Ptpns1	Protein tyrosine phosphatase, non-receptor type substrate 1	Adhesion	−2.0
	Mm.270484	Mkrn1	Makorin, ring finger protein, 1	Ubiquitin metabolism	−2.0
	Mm.295284	Stom	Stomatin	Cytoskeleton	−2.0
	Mm.347910	Ing4	Inhibitor of growth family, member 4	Cell cycle	−2.0
	Mm.5061	Arf2	ADP-ribosylation factor 2	Vesicular traffic	−2.1
	Mm.332967	Ube2h	Ubiquitin-conjugating enzyme E2H	Ubiquitin metabolism	−2.1
	Mm.3781	Piga	Phosphatidylinositol glycan, class A	Adhesion	−2.1
	Mm.291928	Ctnnb1	Catenin (cadherin associated protein), beta 1, 88kDa	Cytoskeleton	−2.2
	Mm.143603	Sh3gl2	SH3-domain GRB2-like 2	Vesicular traffic	−2.3
	Mm.295246	4632413K17	RIKEN cDNA 4632413K17 gene	Unknown	−3.0
	Mm.41982	Rs1h	Retinoschisis 1 homolog (human)	Adhesion	−4.0c
P14, †	Upregulated
	Mm.288474	Spp1	Secreted phosphoprotein 1	Immune response	4.3c
	Mm.182785	Emp1	Epithelial membrane protein 1	Adhesion	2.5
	Mm.100144	S100a6	S100 calcium binding protein A6 (calcyclin)	Calcium binding	2.4c
	Mm.46301	Tyrobp	TYRO protein tyrosine kinase binding protein	Immune response	2.3c
	Mm.38498	Krt1-15	Keratin complex 1, acidic, gene 15	Cytoskeleton	2.3
	Mm.383993	Krt2-5	Keratin complex 2, basic, gene 5	Cytoskeleton	2.2
	Mm.239516	Clec7a	C-type lectin domain family 7, member a	Macrophage protein	2.2c
	Mm.175661	Ifitm1	Interferon induced transmembrane protein 1	Immune response	2.2
	Mm.378931	Gsto1	Glutathione S-transferase omega 1	Detoxification	2.1
	Mm.15819	Cd68	CD68 antigen	Macrophage protein	2.1c
	Mm.3999	Mpeg1	Macrophage expressed gene 1	Macrophage protein	2.1c
	Mm.347407	Cebpd	CCAAT/enhancer binding protein (C/EBP), delta	Transcription factor	2.1
	Mm.2409	Adh1	Alcohol dehydrogenase 1 (class 1)	Detoxification	2.0
	Mm.1282	Ccl3	Chemokine (C-C motif) ligand 3	Immune response	2.0
	Mm.45436	Lyzs	Lysozyme	Immune response	2.0c
	Mm.355327	Dsp	Desmoplakin	Cytoskeleton	2.0
	Mm.4646	Krt1-13	Keratin complex 1, acidic, gene 13	Cytoskeleton	2.0
	Mm.22673	Fcer1g	Fc receptor, IgE, high affinity 1, gamma polypeptide	Macrophage protein	1.9c
	Mm.1239	Gfap	Glial fibrillary acidic protein	Cytoskeleton	1.9c
	Mm.248615	Lgals3	Lectin, galactose binding, soluble 3	Adhesion	1.9
	Mm.303231	Cxcl12	Chemokine (C-X-C motif) ligand 12	Immune response	1.9
	Mm.45173	Msr2	Macrophage scavenger receptor 2	Macrophage protein	1.9c
	Mm.275715	1700063D05	RIKEN cDNA 1700063D05 gene	Unknown	1.9c
	Mm.42230	Cyp26A1	Cytochrome P450, family 26, subfamily a, polypeptide 1	Retinoid metabolism	1.9c
	Mm.22723	Gpr126	G protein-coupled receptor 126	Signaling	1.9nc
	Mm.6974	Krt1-14	Keratin complex 1, acidic, gene 14	Cytoskeleton	1.9
	Mm.268000	Vim	Vimentin	Cytoskeleton	1.8
	Mm.244003	Scel	Sciellin	Epithelial barrier	1.8
	Mm.137	Ccl6	Chemokine (C-C motif) ligand 6	Immune response	1.8c
	Mm.271711	Tagln2	Transgelin 2	Cytoskeleton	1.8c
	Mm.2570	C1qb	Complement component 1q, beta polypeptide	Immune response	1.8
	Downregulated
	Mm.336111	Syt1	Synaptotagmin 1	Synaptic vesicle protein	−1.8
	Mm.259879	Birc4	Baculoviral IAP repeat-containing 4	Apoptosis/antiApoptosis	−1.8
	Mm.272264	Stx3	Syntaxin 3	Vesicular traffic	−1.8
	Mm.218473	Serinc3	Srine incorporator 3	Apoptosis/antiApoptosis	−1.8
	Mm.296520	Vps35	Vacuolar protein sorting 35	Vesicular traffic	−1.8
	Mm.306026	Epb4.1l2	Erythrocyte protein band 4.1-like 2	Cytoskeleton	−1.9
	Mm.30837	Ndrg1	N-myc downstream regulated gene 1	Differentiation	−1.9
	Mm.279923	Nedd4	Neural developmentally downregulated gene 4	Ubiquitin metabolism	−2.1
	Mm.1228	Cryaa	Crystallin, alpha A	Eye lens protein	−2.2
	Mm.41982	Rs1h	Retinoschisis 1 homologue (human)	Adhesion	−3.6c

 c, confirmed; nc, not confirmed.

*  Three Rs1h ^−/Y and three wild-type retinal cRNA samples were hybridized to the Mouse Genome 430A 2.0 Array.

†  Six Rs1h ^−/Y and seven wild-type retinal cRNA samples were hybridized to the Mouse Genome 430 2.0 Array.

Figure 1.

 

Molecular function annotation of significantly regulated genes at time points P7 (a), P9 (b), P11 (c), and P14 (d) in retinoschisin-deficient versus wild-type retina. Colored segments: common pathways recurring at different stages; gray segments: unique at one time point. All genes included in these pathways are listed in Table 1 .

Table 2.

 

List of Commonly Regulated Genes at Different Stages in Mutant (Rs1h ^−/Y) versus Wild-Type Retinal Extracts

Unigene ID	Gene Symbol	Gene Name	Array FC	Day
Upregulated genes
Mm.38049	Tlr4	Toll-like receptor 4	2.2	P7, P11
Mm.92529	Ubqln2	Ubiquilin 2	2.2	P9, P11
Downregulated genes
Mm.256034	Cct3	Chaperonin subunit 3 (gamma)	−1.8	P9, P11
Mm.30837	Ndrg1	N-myc downstream regulated gene 1	−1.9	P11, P14
Mm.5061	Arf2	ADP-ribosylation factor 2	−2.1	P9, P11
Mm.3781	Piga	Phosphatidylinositol glycan. class A	−2.1	P7, P11
Mm.291928	Ctnnb1	Catenin (cadherin associated protein), beta 1, 88kDa	−2.2	P9, P11
Mm.295284	Stom	Stomatin	−2.3	P9, P11
Mm.143603	Sh3gl2	SH3-domain GRB2-like 2	−2.3	P9, P11
Mm.41982	Rs1h	Retinoschisis 1 homologue (human)	−3.6	P7–P14

Figure 2.

 

Quantitative RT-PCR analysis of differentially regulated genes in the retinoschisin-deficient (ko) versus wild-type (wt) retina. (a) Developmental time course of expression for Egr1 and (b) for the hypothetical transcript 1700063D05, respectively. (c) Profile 1 depicts the expression pattern of 10 genes, including Ccl6, Cd53, Cd68, Clecsf12, Fcer1g, Lyzs, Mpeg1, Msr2, Spp1, and Tyrobp whereas (d) profile 2 comprises five genes, including S100a6, Gfap, Cyp26a1, Tgm2, and Tagln2. The y-axis depicts the relative expression levels of the respective genes either separately for Rs1h^−/Y and wild-type (wt) or for both conditions calculated by subtracting relative expression values of wt from ko. Error bars, SD. Statistical significance is given as *P < 0.05 or **P < 0.001 (Student’s t-test). Experiments were performed with retinas from two different animals (four eyes) and each were performed in triplicate measurements.

Figure 3.

 

Quantitative real-time RT-PCR analysis of macrophage- (Cd68 and Clec7a) and apoptosis-related (Tnfrsf6 and Casp8) transcripts. The onset of differential expression in retinoschisin-deficient (Rs1h^−/Y) versus wild-type (wt) is distinctly earlier for Cd68 and Clec7a compared with Tnfrsf6 and Casp8.

Figure 4.

 

Immunofluorescence microscopy on cryosections of Rs1h^−/Y and wild-type mouse retinas at P14, P18, and P24. Macrophages-microglia were stained with the specific marker F4/80 (red), showing initial macrophage activation at P14 and more prominent labeling at postnatal day P18 and P24 in the Rs1h^−/Y retina. Müller glial cells are stained with the GFAP antibody (green). Retinal sections of Rs1h knockout retina reveal cellular disorganization and small cavities starting at P14. Retinal layers are shown in the wild-type P14 image: outer nuclear layer (ONL), outer plexiform layer (OPL), inner nuclear layer (INL), inner plexiform layer (IPL), and ganglion cell layer (GCL).

Figure 5.

 

Western blot analysis of phosphorylated Erk1/2 (pErk1/2) and nonphosphorylated Erk1/2 (Erk1/2) performed on retinal protein extracts from single Rs1h-deficient (Rs1h^−/Y ) and wild-type (wt) mice at P5, P7, P11, and P13. In comparison to the wild-type extracts, increased expression levels of pErk1/2 are apparent at P7 and P11 in the retinoschisin-deficient samples. The blot was reprobed with an anti-β-actin (Actb) antibody as a loading control.

Figure 6.

 

Quantitative real-time RT-PCR analysis of 1700063D05, Cd68, and Gfap as well as genes involved in the extrinsic pathway of apoptosis (Tnfrsf6 and Casp8) in retinal extracts from three mouse genotypes Rs1h^+/Y/Egr1^−/− , Rs1h^−/Y/Egr1^+/− , and Rs1h^−/Y/Egr1^−/− . The analysis was done for two stages at P18 and P24, respectively. Expression of succinate dehydrogenase subunit A (Sdha) serves as control for cDNA integrity. Experiments were performed with retinas from two different animals (four eyes) and were performed in triplicate measurements.

Figure 7.

 

TUNEL nuclei count for postnatal stages P18 and P24 indicates that apoptosis in the outer nuclear layer (ONL) depends on the Rs1h genotype but not on the Egr1 genotype (**P < 0.001, Student’s t-test). Each column represents the mean of apoptotic nuclei counted in three retinal sections from three animals at P18 and P24, respectively.

Supplementary Materials

Supplementary Table S5 - (PDF) 

Supplementary Table S1 - (.xls) 

Supplementary Table S2 - (.xls) 

Supplementary Table S3 - (.xls) 

Supplementary Table S4 - (.xls) 

The authors thank Heidi Schulz for help with the Exiqon Probe Library Kit.