Eliminating Complement Factor D Reduces Photoreceptor Susceptibility to Light-Induced Damage

Abstract

purpose. Genetic risk factors such as variations in complement factors H (CFH) and B (CFB) have been implicated in the etiology of age-related macular degeneration. It has been hypothesized that inadequate control of complement-driven inflammation may be a major factor in disease pathogenesis. The authors tested the involvement of the complement system in an experimental model for oxidative stress-mediated photoreceptor degeneration, the light-damage mouse model.

methods. Changes in gene expression were assessed in BALB/c retinas in response to constant-light (CL) exposure using microarrays and real-time PCR. Susceptibility to CL exposure was tested in CFD ^−/− mice on a BALB/c background. Eyes were analyzed using electrophysiologic and histologic techniques.

results. Genes encoding for proteins involved in complement activation were significantly upregulated after CL. The altered gene profiles were similar to proteins accumulated in drusen and to genes identified in the retina and RPE/choroid of patients with age-related macular degeneration. Cyclic-light reared CFD ^−/− and CFD ^+/+ mice had indistinguishable rod function and number; however, after CL challenge, CFD ^−/− photoreceptors were significantly protected.

conclusions. These results suggest that rod degeneration in the CL-damaged retina involves the activity of the alternative complement pathway and that eliminating the alternative pathway is neuroprotective. Thus, the light damage albino mouse model may be a good model to study complement-mediated photoreceptor degeneration.

Chronic light exposure as a possible risk factor for photoreceptor injury has been under investigation for a number of years. Early reports include data on lightly pigmented military personnel and prisoners from World War II exposed to tropical sunlight who experienced declines in vision and in whom retinal lesions developed over time. Damage was worsened in those also subjected to nutritional stress (for a review, see Rattner and Nathans ¹ ). Recently, the Rotterdam study and a study on watermen of the Chesapeake Bay both concluded that long-term chronic exposure to sunlight or its blue component, especially later in life, may be related to the development of age-related macular degeneration (ARMD). ^{2 3} To identify the pathologic cellular and molecular events caused by light exposure, retina light damage is under investigation in experimental animal models (for a review, see Wenzel et al. ⁴ ). Light as an environmental factor is toxic to rodent rod photoreceptors if the retina is exposed to intermediate to high light levels over a long period (for a review, see Penn and Anderson ⁵ ). Oxidative stress has been implicated as the main trigger for photoreceptor cell death. ^{6 7 8 9} The resultant cascade that leads to photoreceptor degeneration is not fully understood. 

The complement system is an essential part of the evolutionarily ancient innate immune system. Its main role is to eliminate foreign antigens and pathogens as part of the normal host response (for reviews, see Fearon ¹⁰ and Holers ¹¹ ). However, inappropriate or excessive complement activation is also involved in the pathogenesis of autoimmune, inflammatory, and ischemic disease states (for a review, see Holers ¹² ). The complement system can be activated by three distinct pathways: classical, lectin, and alternative. ¹³ Activation by either of the three pathways results in enzymatic cleavage of the serum protein C3 to C3a and C3b, with the latter triggering the common activation pathway through C5 that results in the formation of the membrane attack complex. The membrane attack complex can directly lyse or activate the cell to which it is bound. In addition, as part of the process of complement activation, the anaphylatoxins C3a and C5a are generated and released. Finally, cleavage fragments of C3 (C3b, iC3b, and C3d) that become covalently bound to activating surfaces act as ligands for receptors on immune effector cells. Pertinent to this study, alternative pathway activation can occur spontaneously in the fluid phase (known as “tick over”) and, if an activating surface is available, results in the covalent binding of C3b to this surface. On surfaces not adequately protected by complement inhibitory mechanisms, the deposited C3b acts as a catalyst for further C3 activation through the alternative pathway amplification loop. In addition, C3b deposited as a result of classical or lectin pathway activation can also be amplified by this same alternative pathway amplification loop. Factor B (CFB), properdin, and factor D (CFD) are alternative pathway proteins involved in activation through the tick over mechanism and the amplification loop, whereas factor H (CFH) is a fluid phase inhibitor of the alternative pathway. The generation of C3b and complement effector mechanisms is normally under tight control by fluid phase and by intrinsic membrane regulatory proteins to prevent damage to host cells. 

We sought to study the interaction of light as a risk factor for photoreceptor cell death and the complement system. Gene expression changes in albino mouse retinas in response to constant light exposure were assessed in samples collected at different time points after the onset of light. Because we could not detect any differential effect on light-damage in animals in which the classical pathway had been eliminated (C1qα^−/−) when compared with their wild-type littermates (Grimm C, Rohrer B, unpublished results, 2005), we tested the hypothesis that the alternative pathway may be involved in the neurodegenerative mechanisms activated by CL, testing CFD knockout mice in the CL model. The findings are discussed herein in the context of the light-damage albino mouse model as a model to study complement-mediated photoreceptor degeneration and oxidative stress-mediated pathways in ARMD. 

Materials and Methods

Animals

Complement factor D–deficient mice ¹⁴ were backcrossed five times onto the BALB/c background and then intercrossed, rendering them more than 95% BALB/c. They were genotyped for the presence of the sensitizing Leu450 Rpe65 isoform using published protocols. ¹⁵ BALB/c mice used for microarray were generated from breeding pairs obtained from Harlan Laboratories (Indianapolis, IN). Animals were housed in the MUSC Animal Care Facility under a 12-hour light/12-hour dark cycle with access to food and water ad libitum. The ambient light intensity at the eye level of the animals was 85 ± 18 lux. All experiments were conducted in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by the University Animal Care and Use Committee. 

Light Damage

Adult mice (approximately 1 year of age) were moved into the light-treatment room and were dark adapted overnight. At 12:00 PM., animals were housed two animals per cage in clear acrylic glass (Plexiglas; Rohm and Haas, Philadelphia, PA) cages with free access to food and water. Light damage was induced by exposing the animals to 1000 lux of white light provided by two 30-W fluorescent bulbs (T30T12-CW-RS; General Electric, Piscataway, NJ) suspended approximately 40 cm above the cages. Light intensity was measured using a light meter (Extech Instruments, Waltham, MA) to ensure that equal luminance was provided to all animals. This amount of light reduces the numbers of rods to one row within 2 to 3 weeks in albino mice. ¹⁶  

Gene Expression Profiling

Three-month-old BALB/c mice exposed to cyclic light (control) or different amounts of constant light (24 and 48 hours; 4, 6, and 10 days) were killed by decapitation, and retinas were quickly isolated and stored in storage reagent (RNAlater; Ambion, Austin, TX) at −20°C. Eyes from four animals per time point were pooled to obtain good-quality RNA, to reduce sample variability, and to reduce the number of arrays needed to generate reliable data, ^{17 18} and each data point was obtained independently in duplicate. 

Total RNA was isolated using reagent (Trizol; Ambion), followed by a clean-up using RNeasy minicolumns (Qiagen, Valencia, CA). Sample preparation for microarray hybridization was carried out as described in the Expression Analysis Technical Manual (Affymetrix, Santa Clara, CA) and published previously. ¹⁸ In short, 4 μg total RNA was used to generate double-stranded cDNA (Invitrogen, Carlsbad, CA), which served as a template for the generation of biotinylated cRNA with the use of an RNA transcript labeling kit (BioArray HighYield; Enzo Diagnostics, New York, NY). The labeled probes were fragmented (8 M Na⁺-citrate buffer) and hybridized to U74A oligonucleotide arrays (Affymetrix) by the DNA Microarray Core Facility at the Medical University of South Carolina, using a fluidics station (Fluidics Station 400; Affymetrix). 

Arrays (U74Av2) were scanned using scanner software (Microarray Suite 5.0; Affymetrix) to obtain probe-level data, and outputs were scaled. Data analysis was performed using dChip, a model-based program that allows for the comparison of multiple arrays. ¹⁹ Raw expression data of all the arrays (n = 12) were normalized together using the BALB/c control sample as the baseline array and omitting Affymetrix control genes. Subsequent gene filtering was performed based on the following criteria: 1 < coefficient of variance < 1000; percentage present call >20% of the arrays in the array list file. To identify genes that were reliably differentially expressed, the data were further filtered based on group mean ± SE according to recommended criteria (the threshold for the required fold change between the group means was set at 1.2 [using the lower 90% confidence bound]; the absolute difference between the two group means was set at 100). 

Genes identified as differentially regulated at one time point or more were obtained and clustered based on K-means clustering using the Cluster and Treeview software by Eisen et al. ²⁰ Overrepresentation of gene ontology (GO) terms in the subset of identified genes when compared with their frequency in the entire mouse genome was confirmed by t-test. 

GO terms (biological process) and the relation scheme between the terms were obtained at the Gene Ontology Consortium (http://www.geneontology.org) for all genes on the array. 

Quantitative RT-PCR

Quantitative RT-PCR was performed as described previously. ¹⁶ In short, 1 μg total RNA was used in reverse-transcription reactions (Invitrogen) to generate the template for the PCR amplifications (QuantiTect Syber Green; Qiagen) carried out with 0.2 μM forward and reverse primers. Primers used were GAPDH (forward, 5′-TGCACCACCAACTGCTTAGC-3′; reverse, 5′-GGCATGGACTGTGGTCATGAG-3′) and C3 (forward, 5′-GGAAACGGTGGAGAAAGC-3′; reverse, 5′-CTCTTGACAGGAATGCCATCGG-3′). Reactions were treated with 0.01 U/μL UNG enzyme (AmpErase; Applied Biosystems) to prevent carryover contamination. Real-time PCR was performed in triplicate in a sequence detection system (GeneAmp 5700; Applied Biosystems) using standard cycling conditions. Quantitative values were obtained by the cycle number (C _t value), and relative gene expression levels were calculated using y = (1 + AE)^ΔΔCt, where AE represented amplification efficiency of the target gene, set at 1.0 for all calculations, and ΔΔC _t represented the difference between mean experimental and control ΔC _t values. 

Electroretinographic Recordings

ERG recordings were performed as described previously. ^{21 22} In short, animals were anesthetized using xylazine and ketamine (20 and 80 mg/kg, respectively), and pupils were dilated with 1 drop of phenylephrine HCl (2.5%) and atropine sulfate (1%) and placed on a heated block held at 37°C within a light-tight Faraday cage. Light stimulation was provided using a device of our construction based on the design of Lyubarsky and Pugh, ²³ which integrates the tip of a liquid light guide, a cornea-shaped miniature ganzfeld device, and a recording electrode. The optical signal, driven by a 250-W halogen lamp, was controlled with mechanical shutters, manually operated neutral density, and a 500-nm bandpass filter. Light intensity per 10-ms flash provided in the stimulus path could be varied in steps of 0.3 log units from 3.0 × 10⁵ to 3.0 × 10¹¹ photons/mm². Scotopic electroretinograms were recorded in response to single-flash stimulation of increasing light intensities, averaging three to five responses. Peak a-wave amplitude was measured from baseline to the initial negative-going voltage, whereas peak b-wave amplitude was measured from the trough of the a-wave to the peak of the positive b-wave. ERG recordings were stored, displayed, and analyzed with a PC interface and software suite (pClamp; Axon Laboratories, Burlingame, CA) and Origin software. 

Anatomy

After the ERG recordings, the eyes were removed, fixed in 4% paraformaldehyde and 2% glutaraldehyde, and bisected dorsal to ventral through the optic nerve. The halves were embedded in mixture (Epon-Araldite; Electron Microscopy Sciences, Hatfield, PA), and sections were cut 1 μm through the horizontal meridian and stained with toluidine blue. Photoreceptor layers were counted as described previously, ²⁴ counting in two locations in the central retina (superior and inferior, within 350 μm of the optic nerve head) and two in the peripheral retina (superior and inferior, within 350 μm of the ciliary body). At each location we obtained three measurements, which were averaged to provide a single value for each area. 

Data Analysis

For all experiments, data were expressed as mean ± SEM. ERG data were analyzed using repeated-measures analysis of variance (ANOVA; StatView). After the ANOVA analyses, individual data pairs were compared using Student’s t-test analysis, accepting a significance level of P < 0.05. Morphologic and biochemical data were analyzed using the standard Student’s t-test, also accepting a significance level of P < 0.05. 

Results

Differentially Regulated Genes in Light-Damaged BALB/c Retinas

We have previously analyzed marker gene expression for different processes involved in photoreceptor degeneration, which allows for a limited analysis of the genomic response to constant light damage. ¹⁶ Here we sought to expand this analysis by using DNA microarrays (U74Av2; Affymetrix) for a more comprehensive analysis. Differentially expressed genes were identified with dChip software, a model-based analysis tool tailored to eliminate samples with high SD and low expression levels (see Materials and Methods for filtering parameters). This conservative filtering resulted in a list of 257 genes (Fig. 1 ; see also Supplementary Table S1). Cluster analysis (K-means clustering) on the average expression levels per time point revealed two large clusters that could be subdivided into multiple smaller clusters based on correlation distance. Overall, based on median gene expression across the five time points, 136 genes were identified as downregulated, whereas 121 genes were upregulated. Not surprisingly, the downregulated genes were described by gene ontology terms such as sensory perception of light stimulus (P < 0.001), phototransduction (P < 0.001), G-protein signaling, coupled to cGMP nucleotide second messenger (P < 0.001), and oxygen transport (P = 0.001). The upregulated genes were associated with gene ontology terms such as complement activation (P < 0.001), response to oxidative stress (P < 0.001), and cell ion homeostasis (P = 0.005). Genes whose products are part of the complement system are C1q components (a, b, and c polypeptides) C3, C4, and Serping1 (Fig. 1 , inset). 

Complement System Misregulation in Light Damage

One of the main GO terms describing the genes misregulated in light damage is immune response (genes represented by this GO term are Sh2d1a, C3, C4, LOC56628, H2-D1, B2m, Spp1, C1qα, C1qβ, C1qc, Serping1), identifying members of the classical pathway of complement activation (C1qα, C1qβ, C1qc, C4, Serping1) and a gene common to all complement pathways (C3). We have previously confirmed the upregulation of C1qβ in the light-damaged albino mouse by quantitative real-time PCR (QRT-PCR) ¹⁶ and extended this analysis here to include the expression of the complement component C3 (Fig. 2) . Although elevated levels of C1qβ could not be detected before 12 hours after the onset of the bright light, ¹⁶ C3 mRNA was upregulated within 3 hours after light onset and plateaued at an elevated level between 4 and 10 days. 

Inactivation of Alternative Complement Pathway Signaling

In previous experiments, it was found that mice in which the classical pathway had been eliminated (C1qα^−/− on a C57BL/6 background) were equally as susceptible to bright or blue light damage as their wild-type littermates (Grimm C, Rohrer B, unpublished results, 2005). Thus, it appears that the classical complement system of innate immunity may not be essential for the degenerative process observed in the light-damaged retina. To test the potential role of the alternative complement pathway in the degenerative process, we analyzed the susceptibility of albino mice to light damage on a complement factor D null background (CFD ^−/−). A potential influence of the lack of activity in the alternative pathway on the light-induced degeneration of photoreceptor cells was assessed morphologically (Figs. 3 4)and electrophysiologically (Figs. 5 6)10 days after the onset of CL exposure. 

Lack of CFD did not have any apparent influence on normal retinal morphology at the light microscopic level (Figs. 3A 3C) , thus resulting in a number of rows of photoreceptors that were slightly higher than those of the age-matched BALB/c controls (Fig. 4 , ▴). Similarly, baseline ERG amplitudes recorded before light damage appeared to be similar in CFD ^−/− and wild-type mice (Figs. 5[left], 6A ), with the exception of the width of the b-wave, which seemed to be wider in the animals lacking CFD. 

Exposing the albino mice to 10 days of constant light, however, resulted in a significant phenotypic difference (Figs. 3B 3D) . In control BALB/c mice, constant light resulted in the elimination of most of the photoreceptors (average retina score, 2.0 ± 0.37 rows of photoreceptors), whereas the CFD ^−/− mice retained their photoreceptors (6.7 ± 0.40 rows of photoreceptors; P < 0.001; Fig. 4 ). ERG analysis confirmed that though the ERG consisted of only a measurable b-wave (26.8 ± 4.99 μV) and no a-wave in the BALB/c animals after light damage, the ERG of the CFD ^−/− mice consisted of a measurable a-wave (27.0 ± 4.4 μV) and a partially preserved b-wave (159 ± 22.0 μV; Figs. 5 6 ). 

Discussion

It has been hypothesized that in a number of neurodegenerative diseases, including ARMD, a major contributor to disease pathogenesis may be inadequate control of complement-driven inflammation. Here we examined the involvement of the alternative pathway of the complement system in an experimental model for oxidative stress-mediated photoreceptor degeneration, the light-damage mouse model. Our main findings can be summarized as follows: (1) light damage caused by CL leads to increased mRNA expression of components of the complement pathway and increased expression of C1qβ, ¹⁶ and C3 was confirmed by QRT-PCR; and (2) involvement of the alternative pathway in CL-mediated photoreceptor cell death was confirmed in CFD ^−/− mice, which were shown electrophysiologically and anatomically to be significantly less susceptible to CL damage. Taken together, these observations suggest that photoreceptor degeneration in the CL-damaged eye activates the alternative pathway of the complement system and that eliminating this pathway is neuroprotective. In addition, the results suggest that the CL-damaged albino mouse model may be a suitable model to study complement-mediated photoreceptor degeneration. 

Common Pathways in Light Damage

Genetic pathways activated in light-induced photoreceptor degeneration have been studied extensively using DNA microarrays. Light damage can be induced by a number of different protocols, but two approaches are used extensively. In the CL model of intermediate intensity (≤1000 lux), which we used, animals are exposed to light for an extended period, including the prospective nights, which will lead to the loss of approximately 50% of photoreceptor cells within about 10 days. ¹⁶ In the BL model, dark-adapted animals with dilated pupils are exposed to 0.5 to 2 hours of bright light (5000–15,000 lux) followed by complete darkness to allow for the execution of photoreceptor cell death. Molecular mechanisms activated by the two different light protocols were previously reported to vary. ²⁵ However, current experimental evidence suggests that molecular mechanisms activated in different models of photoreceptor dystrophies overlap to a certain degree (see, for example, Lohr et al. ¹⁶ and Rattner and Nathans ²⁶ ). It is reasonable, given current data, to assume that the molecular signatures of BL and CL damage may be more similar than dissimilar. 

To test this hypothesis, we compared our data with results from other publications using DNA microarray analysis in BL-exposed animals (see, for example, Rattner and Nathans, ²⁶ Chen et al., ²⁷ Roca et al. ²⁸ ). These three BL experiments were performed using short-term bright-light exposure (≤7 hours) followed by no or 24-hour dark adaptation. The two sets of experiments that did not allow any dark adaptation ^{27 28} did report the presence of upregulated but not downregulated genes. Rattner and Nathans, ²⁶ who allowed for 24 hours of dark adaptation, reported that approximately 22% of the identified genes were downregulated. Here we report that in the case of CL damage, over the 10 day time course, on average, 136 genes were downregulated, whereas 121 genes were upregulated. However, at the 24-hour time point, our data were comparable to those reported by Rattner and Nathans, ²⁶ with approximately 24% of the identified genes downregulated. Because the GO terms associated with the downregulated genes identified here appear to be mainly photoreceptor cell specific (i.e., reflecting the loss of cells), it appears that light damage might have been triggered by the upregulation of damaging genes rather than the loss of neuroprotective genes. 

To determine the potential overlap between the cell death mechanisms triggered by the two light paradigms, we first compared our gene expression data as documented here with those published by Rattner and Nathans, ²⁶ who allowed cell death mechanisms to develop for 24 hours after BL damage. Our study and that of Rattner and Nathans ²⁶ used two different arrays, U74 (our study; approximately 12,000 elements) and MOE430 ²⁶ (more than 39,000 elements). Approximately 72% of the genes and ESTs of the smaller array are contained on the larger array based on more than 95% identity between the sequences. Rattner and Nathans ²⁶ reported the misregulation of approximately 300 genes, with 57% (176/311) of those misregulated genes present on the U74 array. Of these 176 genes, 41 were found to be differentially regulated in the CL damage model (Table 1) . These genes range from transcription factors (e.g., Cebpδ, JunB, Stat3) to oxidative stress-related genes (e.g., Mt1, Mt2, Cp, A2m, B2m) and immune and defense genes (e.g., complement system [C3, C4, C1qα, Serping1] and others [Serpina3n, H2-D1, Sh2d1a, B2m, LOC56628, Spp1]) and more. Of these 41 genes, 17 genes were upregulated after 7 hours of BL, ²⁷ including the three transcription factors (Stat3, JunB, Cebpδ), the five oxidative stress genes (Mt1, Mt2, Cp, A2m, B2m), the three immune and defense genes (Serpina3n, B2m, Spp1), and others (Msn, Socs3, Crym, Osmr, Lcn2, Myo10, Edn2). Finally, three genes—two transcription factors (Cebpδ, JunB) and one oxidative stress-related gene (Mt2)—were upregulated by 3 hours of BL exposure. ²⁸ Incidentally, these three genes were upregulated in all four studies. Thus, as expected, early cellular responses to BL are mediated by changes in transcription factor levels and the upregulation of free-radical scavengers. Common cellular responses that occur 24 hours or later are mediated by genes whose products are involved in oxidative stress and the immune and defense responses. Taken together, the two light damage mouse models share a significant number of genes, suggesting that similar cellular pathways are activated. 

Alternative Pathway Activation in Photoreceptor Degeneration

Immune and defense genes were highly overrepresented in CL, particularly genes whose products are involved in the complement pathway. The lack of the alternative complement pathway did not appear to affect retinal development and physiology significantly; however, eliminating this pathway significantly protected the photoreceptors against CL-induced cell death. After 10 days of CL, no photoreceptor nuclei were lost in the CFD ^−/− mice, whereas wild-type animals lost more than 50% of their photoreceptors in that same period. In ERG recordings, a- and b-wave amplitudes were significantly preserved. The reduced ERG amplitudes after light damage are in part a reflection of photostasis, a mechanism whereby the photoreceptors adjust their photon catch to the amount of light present in the environment. ²⁹ Given sufficient time for the retina to stabilize, the ERG waves will recover. ³⁰ However, the possibility that the small reduction in photoreceptor activity prior to light damage could blunt the sensitivity of the photoreceptors to light damage requires consideration. Hao et al. ²⁵ have shown that the kinetics by which bleached rhodopsin is regenerated, and thus how much pigment is available for bleaching at any given time, dictates the cell’s sensitivity to light-induced photoreceptor degeneration. To our knowledge, no studies have titrated the amount of rhodopsin present in the retina with sensitivity to light damage (a study that could be performed in the Rho ^−/− mouse). 

As presented in the Introduction, experimental evidence has suggested that the complement system, particularly the alternative pathway, is involved in inflammation and host tissue injury. Low-level activation of the alternative pathway is typically present in tissues through the tick-over process, generating complement component C3b. C3b bound to activating surfaces becomes a target for CFB, which then initiates, through the amplification loop, the generation of additional C3 convertases and C3b fixation. ¹³ Under normal conditions such as cyclic light, complement inhibitors such as complement factor H, as well as complement receptor-1–related protein y (Crry), decay-accelerating factor (DAF), and CD59, prevent runaway amplification and thus tissue injury. In light damage, the rapid increase in C3 without the corresponding increase in complement inhibitors (no increase in expression of CFH, Crry, CD59, DAF1, or DAF2 was observed in the CL-exposed animals according to the microarray study) could result in the activation of the alternative pathway directly or the engagement of the amplification loop by classical or lectin pathway activation. In addition to the molecular mechanism of amplification, the alternative pathway is also potentially enhanced by a cellular component. In response to initial complement activation, inflammatory cells are recruited to the sites of local injury, which can carry preformed C3 and properdin, adding to the activation specifically at the site of injury. BL damage has been shown to cause infiltration of activated-microglia cells, ³¹ and CD45⁺ cells, characteristic of the central nervous system microglia, were identified in the outer nuclear layer after 2 days of CL exposure (data not shown). In the rd1 retina, activated microglia have been shown to express different signaling molecules such as chemokines MCP-1, MCP-3, MIP-1α, MIP-1β, RANTES, and TNF-α. ³² Thus, in addition to amplifying the complement cascade, microglia may indirectly contribute to photoreceptor death by releasing cytotoxic substances. Activation of the complement system may lead to the formation of the membrane attack complex possibly on rod photoreceptors, leading to the lysis of cells. In sum, the precise mechanisms whereby the alternative pathway participates in mediating cell death in response to light damage and whereby eliminating the alternative pathway is neuroprotective still must be elucidated. 

Implications for Age-Related Macular Degeneration

ARMD occurs in two forms, wet and dry. Wet ARMD is associated with neovascularization, leakage of these new vessels, and rapid photoreceptor degeneration. Dry ARMD leads to the slow degeneration of the photoreceptors in the macula, for reasons that are not yet understood. Studies in the pathogenesis of ARMD have indicated that inflammation, particularly the alternative pathway of complement, is a fundamental component of the disease. Recent genetic evidence has strongly implicated a variation in the complement control protein CFH or CFB gene as a major risk factor for the disease. ^{33 34 35 36} Similarly, in an animal model of ARMD, the laser-damage model of choroidal neovascularization, the involvement of the alternative pathway has been confirmed in specific complement–component-deficient mice ^{37 38} using blocking antibodies ³⁷ or silencing RNA. ³⁹ Thus, it has been hypothesized that inadequate control of complement-driven inflammation may be a major factor in disease pathogenesis. However, patients with a variation in complement proteins can have dry or wet ARMD. It has been unclear how activation of the alternative pathway of the complement cascade could lead to the pathogenesis seen in dry ARMD. Our experiments, which revealed that blocking alternative pathway activation is neuroprotective against light-induced photoreceptor degeneration, may provide further clues. Here, we have compared the altered genes in the retina of the light-damaged mouse (see Supplementary Table S1) with those proteins accumulated in drusen. ⁴⁰ Both wet and dry ARMD are characterized by the presence of drusen, which consist of extracellular deposits of material. It was found that 24% of available matches on the array (proteins in drusen for which there were genes on the array) have altered expression levels, including vitronectin, complement components, clusterin, apolipoprotein E, and crystalline. The altered genes are similar to those identified in retina and RPE/choroid of patients with ARMD. ⁴¹ In summary, though no drusen development can be observed in the light-damaged mouse retina—indicating that the model can therefore not be used as a model for human ARMD—data derived from our study may nevertheless be informative because they provide new information about the complement system in oxidative stress-mediated photoreceptor degeneration. 

 Supported by the National Institutes of Health (NIH) Grant EY13520 and Vision Core Grant EY014793 and by an unrestricted grant to the Medical University of South Carolina from Research to Prevent Blindness. BR is a Research to Prevent Blindness Olga Keith Weiss Scholar. The authors used the DNA Microarray Facility, which is supported by NIH/National Center for Research Resources (NCRR) South Carolina COBRE for Cardiovascular Disease Grant RR-16434 and National Cancer Institute Shared Resource Grant R24 CA95841. Animal studies were conducted in a facility constructed with the support of NIH Grant C06 RR015455 from the Extramural Research Facilities Program of the NCRR.

 Submitted for publication March 7, 2007; revised May 16, 2007; accepted August 27, 2007.

 Disclosure: B. Rohrer, None; Y. Guo, None; K. Kunchithapautham, None; G.S. Gilkeson, 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.

 

Cluster analysis. Hierarchical clustering of gene expression profiles were filtered with the use of dChip. ¹⁹ Values of differentially regulated genes (at one time point or more) were averaged per time point and clustered, based on K-means clustering using the Cluster and TreeView software. ²⁰ Note that the clustering tree of genes shows a striking modular behavior. Data are expressed as the log₂ (expression ratio) from <−2 (red) to >2 (green). Inset: expression pattern and gene names of clustered genes involved in complement activation.

Figure 2.

 

Verification of C3 mRNA expression. Quantitative RT-PCR was performed on whole retina samples obtained from animals exposed to bright light ranging from 3 hours to 21 days. Quantitative values were obtained by the cycle number (C _t value), determining the difference between the mean experimental (C3) and control (GAPDH) ΔC _t values. C3 mRNA expression was significantly increased as early as 6 hours after light onset (light onset at 12 noon) and reached a plateau by 48 hours.

Figure 3.

 

Retinal histology. Comparison of CFD ^+/+ (left) and CFD ^−/− (right) mouse retinas. Photographs of sections were taken in the central retina for comparison. Sections stained with toluidine blue revealed that the 10-day light-damage regimen (B, D) resulted in significant loss of cells in the outer nuclear layer of the CFD ^+/+ retina but did not affect the CFD ^−/− retina compared with their respective controls (A, C).

Figure 4.

 

Photoreceptor cell counts. Rows of photoreceptor cells in CFD ^+/+ and CFD ^−/− mice before and after light damage. Four different measurements were taken across the retina, counting rows of photoreceptor nuclei in the superior periphery, superior central, inferior central, and inferior periphery. No change was observed after light damage in the CFD ^−/− mouse retina, but there was severe loss of photoreceptors in the CFD ^+/+ retina (P < 0.001 based on the comparison of overall averages). Data are plotted as mean ± SEM (n = 4–5 animals per group).

Figure 5.

 

Family of ERGs. Families of ERGs elicited from CFD ^+/+ (A, C) and CFD ^−/− (B, D) mice before (A, B) and after light damage (C, D) using increasing light intensities (maximum light intensity, 3.0 × 10¹¹ photons/mm², attenuated ⅓ log unit per recording). The four light intensities used in (C) and (D) correspond to the top four light intensities in (A) and (B). Rod ERGs of mice raised in cyclic light consisted of the hyperpolarizing response of the rods (a-wave) followed by the depolarizing response of the rod bipolar cells (b-wave). Rod ERG amplitudes were similar in the CFD ^−/− mice compared with their age-matched littermates, with the exception of the shape of the b-wave, which seems to be wider. The 10-day period of light damage resulted in the elimination of the a-wave and a reduction in the b-wave amplitudes in the CFD ^+/+ mouse retina but affected the ERG elicited from the CFD ^−/− retina significantly less (P < 0.004), resulting in significantly preserved a- and b-wave amplitudes. For complete data analysis, see Figure 6and text.

Figure 6.

 

ERG amplitudes. ERG amplitudes were determined for a-waves (A) and b-waves (B) at the maximum light intensity (3.0 × 10¹¹ photons/mm²) before (cyclic) and after (constant) light damage. Data are plotted as mean ± SEM.

Table 1.

 

Genes Differentially Regulated in the CL Damage Model

Accession	Gene Symbol	24 Hours	48 Hours	4 Days	6 Days	10 Days	Bright Light
AF097632	Sh2d1a	−1.7	−3.63	−3.38	−2.56	−3.4	−2.4
U19860	Gas7	−1.82	−2.61	−2.8	−1.98	−3.69	−2.8
AF100777	Wisp1	−1.45	−2.5	−2.42	−2.28	−3.25	−2.2
L10666	Gnat2	−1.19	−2.4	−1.62	−1.4	−1.16	−2.1
U92562	Opn1sw	−1.13	−1.87	−1.39	−1.21	−1.22	−3.1
AB031386	1810009M01Rik	−1.03	1.06	1.4	1.71	2.11	2.3
X04663	Tubb5	1.06	1.14	1.22	1.4	1.04	2.4
AW123952	Atp1a1	1.52	1.44	1.37	1.84	1.65	2.8
X68193	Nme2	1.38	1.49	1.72	1.9	1.69	4.2
X57337	Pcolce	1.13	1.28	1.85	2.27	1.97	3.4
U08378	Stat3	1.51	1.89	1.79	1.76	1.65	3.8
AW046965	Mak10	1.58	1.98	1.93	1.83	1.59	2.6
X06454	C4	1.22	1.68	1.79	2.24	2.47	2.4
AI839417	Msn	1.63	1.91	1.75	2.09	2.26	4.6
AA612450	Antxr2	2.06	1.78	1.62	2.48	1.76	3.1
AI843709	Il6st	2	1.61	1.65	2.14	2.44	2.2
U88328	Socs3	1.31	2.19	2.09	2.02	2.64	13.5
AF010254	Serping1	1.17	1.76	1.91	2.34	3.07	3.9
AC002397	Gnb3	1.73	2.04	2.34	2.42	2.12	2
C77227	Mrps25	1.56	2.59	1.95	2.57	2.2	2.3
X02801	Gfap	2.09	1.9	2.04	2.56	2.8	3.8
M58156	LOC56628	1.49	2.03	2.48	2.81	2.87	3.2
U20735	Junb	2.2	2.57	2.13	2.87	2.67	4.5
M69069	H2-D1	1.29	2.14	3.05	2.95	3.07	2.8
M64086	Serpina3n	2.17	1.84	2.34	2.82	4.07	5.7
K02782	C3	1.07	1.27	2.37	4.36	4.65	3.9
AF039391	Crym	1.84	2.26	3.32	3.14	3.28	7.1
AJ249706	Myo10	2.42	2.79	2.87	3.52	2.97	3
X13986	Spp1	1.31	4.55	2.96	3.18	3.69	13.9
X58861	C1qa	1.03	1.45	3.61	4.56	5.22	2.3
AB015978	Osmr	3.32	2.34	2.58	3.61	4.83	8.2
V00835	Mt1	2.44	2.63	3.36	3.62	5.53	5
U49430	Cp	2.73	3.4	3.74	4.3	5.13	8.2
K02236	Mt2	2.26	2.63	3.9	3.74	6.84	7.5
X01838	B2m	1.08	3.07	5.45	4.64	7.25	4.9
AW125390	Ifitm3	1.9	2.66	3.79	6.35	9.43	5.1
X00246	H2-D1	2.11	5.16	7.29	7.51	7.24	3.2
X81627	Lcn2	1.88	3.07	7.23	10.94	21.52	17.9
AI850558	A2m	4.23	7.26	8.1	10.32	19.96	5.2
X61800	Cebpd	9.32	11.48	9.91	10.36	9.79	19.3
X59556	Edn2	12.28	21.14	34.17	35.08	29.92	12.9

 Listed here are gene accession numbers (column 1), gene symbols (column 2), fold difference in gene expression for the different time points with respect to control (columns 3–7), and the fold difference in gene expression for BL exposure ²⁶ (column 8). The genes are ranked based on the median-fold difference for the CL samples.

Supplementary Materials

Supplementary Table S1 - (.xls) 

The authors thank Yuanyuan Xu for generously providing the CFD ^−/− mice, Michael Holers, Stephen Tomlinson, and Rosalie Crouch for helpful discussions, and Luanna Bartholomew for critical review. In addition, they thank Yang Juncong, Jing Yang, Heather Lohr, Christina Demos, and Coni Imsand for excellent technical assistance.