Expression of Osteopontin in the Rat Retina: Effects of Excitotoxic and Ischemic Injuries

Abstract

purpose. The cytokine osteopontin (OPN) has been localized to the retinal ganglion cell layer in the normal rodent retina, prompting the suggestion that it could serve as a useful marker for identifying and quantifying such neurons in models of retinal and optic nerve neurodegeneration. In the present study, we characterized the time course and cellular localization of OPN expression in the rat retina after excitotoxic and ischemic injuries.

methods. Excitotoxicity and ischemia–reperfusion experiments were performed by using standard techniques. Rats were killed at various time points, and the retinas were removed either for mRNA analysis or to be processed for immunohistochemistry.

results. In the normal retina, double-labeling immunofluorescence indicated that OPN is expressed by the majority of, if not all, RGCs, since OPN was associated with more cells than Brn-3, but was colocalized with Thy1.1. NMDA, kainic acid, and ischemia–reperfusion all caused decreases in the total retinal levels of Thy1 and Brn-3 mRNAs, reflecting injury to RGCs, but a dramatic, short-lived upregulation in OPN mRNA. The source of the increased OPN signal after excitotoxic–ischemic insults is unlikely to be injured RGCs, as no alteration in the intensity of OPN immunostaining in RGCs was apparent. Instead, additional cells, mostly contained within the IPL, were identified as positive for OPN. Double-labeling immunofluorescence showed that ED1 always colocalized with OPN in these cells, indicating their status as activated microglia.

conclusions. OPN is exclusively expressed by RGCs in the physiological retina, but in response to retinal neurodegeneration is synthesized de novo by endogenous, activated microglia.

Ischemia-like injury to the retina and optic nerve head is a major cause of blindness worldwide and manifests in a variety of ophthalmic disorders, including arterial and venous occlusions, diabetic retinopathy, anterior ischemic optic neuropathy, and glaucomatous optic neuropathy. All these diseases display retinal ganglion cell (RGC) death and consequent thinning of the nerve fiber layer. ¹ A central feature of ischemic or traumatic injury to the retina is an uncontrolled release of the excitatory neurotransmitter glutamate, ^{2 3} which can overstimulate neurons expressing ionotropic glutamate receptors, resulting in their excessive depolarization and subsequent excitotoxic death. Of the neuronal classes in the retina, RGCs are particularly sensitive to glutamate toxicity, owing to the presence of such receptors ⁴ ; thus, glutamate and glutamate receptor agonists are toxic to RGCs, ^{5 6 7} and ionotropic glutamate receptor antagonists protect RGCs from ischemic, traumatic, and glutamatergic insults. ^{5 8 9 10} In recent years, much effort has gone into attempting to elucidate the signaling pathways that contribute to RGC death and survival so that effective neuroprotective strategies can be developed. ^{11 12} One putative modulator of RGC survival in conditions of ischemia/excitotoxicity that has received scant attention to date, yet is implicated in neurodegenerative processes in the brain, is osteopontin (OPN). 

OPN is a phosphorylated glycoprotein, originally described as an immobilized extracellular matrix protein in mineralized tissues. It is now known that a variety of cell types throughout the body constitutively synthesize and secrete OPN, whereupon it acts as a cytokine. Detailed knowledge both of the factors that regulate OPN expression and the biochemical pathways that are influenced by the molecule is lacking; nevertheless, it has been implicated in a diverse array of physiological and pathologic processes, including biomineralization, inflammation, cell-mediated immunity, cell survival, wound repair, and cancer biology. ^{13 14 15 16} OPN is widely expressed by neurons of the developing and adult central nervous system (CNS), ^{17 18 19} but its significance in the normal brain is as yet unknown. In the pathologic brain, however, a mounting body of evidence points to OPN as a key cytokine in the cellular response to injury. OPN has been consistently identified at sites of inflammation caused by tissue damage or disease, even in areas in which it is not constitutively synthesized. Of interest, the cell types responsible for de novo OPN expression in these situations are predominantly of the monocyte/macrophage lineage rather than neurons; for example, induction of OPN expression has been reported in cells identified as activated microglia after focal ^{20 21 22} and global ²³ ischemia and kainic acid. ²⁴ Although the precise actions of OPN in ischemia-like injuries are still to be identified, recent data suggest that it acts as a potent neuroprotectant. ²⁵  

Few data exist with regard to the role of OPN in the retina. In the normal embryonic ²⁶ and adult ^{27 28} rodent retina, it has been localized exclusively to the ganglion cell layer (GCL), prompting the suggestion that it could be a useful marker for identifying RGCs. The function of OPN in the normal retina, however, is unknown. It is of particular significance that it has been identified in RGCs, since these are the most vulnerable cells in the retina to ischemic and excitotoxic insults. ¹² As such, the major objective of the present study was to determine the effect of excitotoxic and ischemic injuries on the expression of OPN in the retina. To achieve this goal we compared the expression of OPN mRNA and protein with other established RGC markers after ischemia–reperfusion or injection of excitatory glutamate receptor agonists, and we investigated whether OPN is synthesized by retinal glial cells or infiltrating macrophages after these insults. 

Materials and Methods

Animals and Procedures

All experiments conformed to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. This project was approved by the Animal Ethics Committee of the Institute of Medical and Veterinary Science, Adelaide, and conforms with the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes, 2004. Adult Sprague-Dawley rats (200–250 g) were housed in a temperature- and humidity-controlled room with a 12-hour light, 12-hour dark cycle and were provided with food and water ad libitum. For excitotoxicity or lipopolysaccharide experiments, rats were anesthetized with isoflurane and intravitreous injections of 5 nanomoles of kainic acid or 30 nanomoles of NMDA (5 μL in sterile saline) or 0.2% lipopolysaccharide (5 μL in sterile saline) were performed in one eye. The control eye was injected with 5 μL of vehicle. For the induction of ischemia, rats were given an intraperitoneal injection of 100 mg/kg ketamine and 10 mg/kg xylazine. The anterior chamber of one eye was cannulated with a 30-gauge needle connected to a reservoir of sterile balanced salt solution above 120 mm Hg and maintained for 60 minutes before allowing reperfusion to commence. Rats were killed at various time points after treatment, and the retinas were removed either for mRNA analysis or to be processed for immunohistochemistry. 

Conventional and Quantitative RT-PCR

For reverse transcription–polymerization chain reaction (RT-PCR) studies, total RNA from retinas was isolated and quantified. First-strand cDNA was synthesized from 2 μg DNase-treated RNA and samples were diluted to a final concentration equivalent to 10 ng/μL total RNA. To assess genomic DNA contamination, we performed additional reactions in which the reverse transcriptase enzyme was omitted. PCR products were not observed for any of the primer pairs tested using these samples. 

Primer pairs were designed from sequences contained in the GenBank database using the primer design software Primer 3 (http://frodo.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi/ provided in the public domain by the Whitehead Institute, Massachusetts Institute of Technology, Cambridge, MA) and were selected to amplify sequences that spanned at least one intron. Primer sequences were analyzed for T _m (melting temperature), secondary structure, and primer–dimer formation with primer analysis software (Premier Biosoft, Palo Alto, CA) and verified for their specificity to the target sequence by using the BLAST database search program (www.ncbi.nlm.nih.gov/BLAST/ provided in the public domain by the National Center for Biotechnology Information, Bethesda, MD). The primer pair used to amplify Brn-3b was from Weishaupt et al., ²⁹ and the cyclophilin primers used for real-time PCR reactions were from Marini et al. ³⁰ All primers were optimized for annealing temperature and MgCl₂ concentration before being used in assays. The oligonucleotide primer sequences, their annealing temperatures, and MgCl₂ concentrations are shown in Table 1 . 

For conventional PCR, the individual cDNA species were amplified in a reaction containing the cDNA equivalent of 20 ng total RNA, 1× PCR buffer, MgCl₂, 0.8 mM each dNTP, the relevant sense and antisense primers at a final concentration of 4 ng/μL and 0.5 U Taq polymerase (AmpliTaq Gold; Applied Biosystems, Melbourne, Australia). PCRs (94°C, 12 seconds; annealing temperature, 30 seconds; 72°C, 30 seconds) were performed for a suitable number of cycles followed by a final extension at 72°C for 3 minutes. Prior experiments had determined the linear phase of amplification for each set of primers. All the PCR products yielded single bands corresponding to the expected molecular weights (Table 1) , and PCR products were sequenced to ensure their validity. 

Quantitative real-time PCR reactions were performed in 96-well optical reaction plates using the cDNA equivalent of 20 ng total RNA for each sample in a total volume of 25 μL containing 1× SYBR Green PCR master mix and forward and reverse primers at a final concentration of 400 nM. A gene-specific standard curve was included in each plate. This was generated by serial dilution of a pooled rat retinal cDNA obtained from the RNA of several eyes. The thermal cycling conditions of PCRs were as follows: 95°C for 3 minutes and 40 cycles of amplification comprising 95°C for 12 seconds, 63°C for 30 seconds, and 72°C for 30 seconds. After the final cycle of the PCR, primer specificity was checked by the dissociation (melting) curve method. PCR assays were performed with a thermal cycler (IQ5 iCycler; Bio-Rad, Hercules, CA), and all samples were run in duplicate. The results showed that all mRNAs amplified with high efficiency and linearity during real-time PCR. Mean amplification efficiencies, as determined by plotting cycle threshold as a function of initial cDNA quantity, were between 1.90 and 2.00. Heat dissociation of the amplified DNA detected a single peak in all cases, indicating that a single, specific PCR product had been synthesized. This was confirmed by electrophoresis of PCR products, in which a single band of the expected molecular weight was observed. 

Analysis of Data

To allow a comparison to be made between the levels of expression of OPN, Thy1, and NF-L in control versus treated retinas, results obtained from the real-time PCR experiments were quantified by the standard curve method. All values were then normalized to that of the endogenous housekeeping gene cyclophilin. Cyclophilin thus acted as an internal standard to correct for any variations in RNA isolation and/or cDNA synthesis. The process essentially comprised three steps: first, the threshold cycles (C_T) of each sample for OPN, Thy1, NF-L, and cyclophilin were converted to known quantities by interpolation from the appropriate standard curve; second, the amount of OPN, Thy1, or NF-L in each sample was normalized to the level of cyclophilin; third, the normalized amount of OPN, Thy1, or NF-L in the retina of the treated eye was expressed relative to the normalized amount in the retina of the control eye. All results are expressed as the mean ± SEM. Differences between control and treated retinas were assessed with Student’s paired t-test, whereas comparisons between treatment groups were made with one-way ANOVA. In both cases, P < 0.05 was considered statistically significant. 

Immunohistochemistry

For immunohistochemistry, rats were transcardially perfused with PBS and subsequently with 4% paraformaldehyde. After enucleation, whole eyes were fixed in 10% neutral-buffered formalin and processed for routine paraffin-embedded sections. Eyes were embedded sagittally and 5-μm serial sections were cut. Tissue sections were deparaffinized, rinsed in 100% ethanol and treated for 30 minutes with 0.5% H₂O₂ to block endogenous peroxidase activity. Antigen retrieval was achieved by microwaving the sections in 10 mM citrate buffer (pH 6.0). Tissue sections were then blocked in PBS containing 3% normal horse serum, incubated overnight at room temperature in primary antibody (containing 3% normal horse serum; see Table 2 ), followed by consecutive incubations with biotinylated secondary antibody (Vector, Burlingame, CA) and streptavidin-peroxidase conjugate (Pierce, Rockford, IL). Color development was achieved with 3′-,3′-diaminobenzidine. Sections were counterstained with hematoxylin, dehydrated and mounted. The specificity of antibody staining was confirmed by incubating adjacent sections in the absence of the primary antibody. 

For immunohistochemical double labeling, sections were incubated overnight at room temperature with anti-OPN and one of anti-PGP 9.5 or anti-Brn-3, as described previously. On the following day, sections were incubated with Alexa Fluor 488 donkey anti-mouse IgG (1:250) together with Alexa Fluor 660 donkey anti-rabbit or goat IgG (1:250), as appropriate, for 1 hour. For double immunolabeling using primary antibodies raised in the same species, anti-OPN antibody was first biotinylated (courtesy of Dako, Botany, Australia), then incubated with anti-Thy1 or anti-ED1 antibodies overnight, as described previously. On the following day, sections were incubated with streptavidin-conjugated FITC (1:250) and donkey anti-mouse IgG-Cy5 (1:400). Sections were then mounted using anti-fade mounting medium (ProLong Gold; Invitrogen, Carlsbad, CA) and examined under a confocal fluorescence microscope. 

Results

OPN Immunoreactivity in Normal and Axotomized Rat Retinas

OPN immunoreactivity in the normal rat retina was primarily localized to cell bodies in the GCL (Figs. 1A 2A 2B) . Cells varied with regard to the intensity of staining, with larger cells typically containing the most signal. Similar patterns of labeling were obtained with two different OPN antibodies. In order both to confirm that OPN is expressed exclusively by RGCs rather than by displaced amacrine cells, which are abundant in the GCL of the rat retina, and to determine the extent that OPN colocalizes with other RGC markers, double-labeling experiments were performed in normal and axotomized retinas. In the normal retina, cells that expressed OPN also labeled for the RGC markers Brn-3 (Figs. 1A 1B 1C)and Thy1 (Figs. 1D 1E 1F) . A greater number of cells in the GCL were positive for OPN than for Brn-3, but there are two reasons for concluding that these cells are RGCs rather than displaced amacrine cells: First, Brn-3 is known to be expressed by only a subset of the total RGC population, and second, OPN did not colocalize with PGP 9.5-labeled amacrine cells or horizontal cells in the inner nuclear layer. More substantive evidence that OPN is not expressed by displaced amacrine cells was obtained by using rats in which the RGC population had been eliminated by prior optic nerve transection (Figs. 2C 2D 2E 2F 2G 2H 2I) . In these retinas, OPN immunolabeling in the GCL, like those of the specific RGC markers Brn-3 and Thy1, was essentially absent. The continuing presence of displaced amacrine cells in the GCL was indicated by labeling with NeuN and PGP 9.5. 

Expression of OPN mRNA after Excitotoxic and Ischemic Injuries: RT-PCR Studies

Both NMDA and kainic acid cause injury to a population of RGCs, which is demonstrated by rapid downregulation of the retinal levels of mRNAs encoding the RGC-specific markers Thy-1, NF-L, and Brn-3 (Fig. 3A) . Of note, the OPN mRNA level was not reduced after these injuries. In fact, OPN mRNA expression increased dramatically after NMDA and kainic acid (Fig. 3A) . A similar effect was observed after ischemia–reperfusion (data not shown). Conventional RT-PCR is an extremely useful guide to gross changes in mRNA levels; however, it is only semiquantitative, mostly due to the small dynamic range of ethidium bromide. To facilitate accurate quantification of both the magnitude and duration of the upregulation of OPN mRNA and the difference between OPN and other RGC-specific markers, quantitative, real-time PCR experiments were performed. 

Initially, the suitability of cyclophilin as a housekeeping gene was evaluated. This was achieved by determining whether the absolute, non-normalized level of cyclophilin mRNA (in the cDNA equivalent of 20 ng total RNA) differed between control and treated retinas. The results showed no significant difference between the total retinal level of cyclophilin mRNA in control and treated eyes at any of the time points after NMDA or kainic acid injection, or after ischemia–reperfusion (data not shown). The data thus indicate the suitability of cyclophilin for use in this study. 

Using real-time, quantitative PCR we confirmed the results obtained using conventional PCR—namely, that NMDA and kainic acid cause a rapid decrease in the retinal levels of Thy1 and NF-L mRNAs, but a striking increase in OPN mRNA (Fig. 3B) . Detailed analysis of the effect of NMDA on OPN mRNA expression revealed that the upregulation of OPN mRNA was of considerable magnitude, but was acute in nature. There was a statistical increase in OPN mRNA expression within 2 hours of NMDA injection, the response had increased sharply by 6 hours and peaked by 12 hours, but then began to decline, so that by 3 days it was only moderately higher than in control eyes, with expression having returned to basal levels in the majority of eyes by 7 days (Fig. 3C) . Comparison of the relative effects of NMDA, kainic acid, and high IOP-induced ischemia–reperfusion after 24 hours indicated that treatment with kainic acid induced the largest increase in OPN mRNA production, with NMDA and ischemia–reperfusion eliciting responses of a similar magnitude (Fig. 3D) . 

The increase in total retinal OPN expression measured after excitotoxic and ischemic insults may occur as a result of upregulated expression by injured RGCs, increased expression by neighboring uninjured RGCs, or synthesis by retinal cell types that normally do not express OPN. To discover which of these scenarios occurred, OPN immunohistochemistry was evaluated after excitotoxic and ischemic injuries. 

OPN Immunoreactivity after Excitotoxic and Ischemic Injuries

It is known, and was evident in our study, that approximately 50% of RGCs are lost from the rat retina after excitotoxic challenge. Histologic examination 7 days after treatment with NMDA revealed fewer cells in the GCL and a marked thinning of the IPL. Immunohistochemical labeling with antibodies specific for the RGC markers Brn-3 (Figs. 4D 4E 4F)and Thy1 (Figs. 4G 4H 4I)revealed no change in the patterns of staining 12 hours after NMDA, but characteristic reductions in the number of cells labeling for these antigens after 7 days—a result consistent with a loss of approximately half of the RGC population. RGCs unaffected by the insult continued to express Brn-3 and Thy1. 

OPN (Figs. 4A 4B 4C)expression by RGCs after excitotoxic injury mirrored that of Brn-3 and Thy1—namely, that 12 hours after NMDA injection, no alterations in either the number of RGCs expressing OPN or the intensity by which they were labeled were apparent (Fig. 4B) . Likewise, 7 days after NMDA treatment, there was a pronounced reduction in staining for OPN in the GCL, due to the decreased number of RGCs present, but, as was the case for Brn-3 and Thy1, the remaining RGCs were still positive for OPN (Fig. 4C) . Significantly, however, and unlike Brn-3 or Thy1, additional cells, mostly contained within the IPL, were identified as OPN positive 12 hours after NMDA (Fig. 4B) . Thus, it appears probable that the dramatic increase in synthesis of OPN mRNA which occurs after excitotoxic injury is by retinal cell types that normally do not express OPN. As such, we determined the time course of OPN protein expression by these non-RGCs after NMDA injection and made a comparison with markers of activated glial cells. 

An increase in retinal OPN mRNA was measured as soon as 2 hours after NMDA injection, but OPN protein was not detected within cells of the IPL at this early time point (data not shown). OPN-positive cells were evident by 6 hours after injury, and a greater number were observed after 12 hours, but no immunoreactive cells were visible after 3 days (Figs. 5A 5B 5C) . The pattern of OPN labeling displays obvious similarity to that of the activated microglial marker ED1 (Figs. 5D 5E 5F) , suggesting that microglial cells synthesize OPN; yet, unlike staining for OPN, ED1 immunoreactivity increased throughout the 7-day period after NMDA, and even within the first 24 hours, a greater number of cells were positive for ED1 than for OPN. OPN labeling bore no resemblance to expression of GFAP (Figs. 5G 5H 5I) , which labels astrocytes and activated Müller cells. To establish whether microglia were responsible for synthesizing OPN, double-labeling experiments were performed. The results showed that in OPN-positive cells, ED1 was always colocalized, indicating their status as activated microglia (Fig. 6) . Nevertheless, only a subset of ED1-positive cells expressed OPN, and at later time points ED1-positive microglia did not express OPN. Of particular note was the fact that any putative infiltrating macrophages, which express ED1, were OPN negative. As expected, no colocalization was observed between OPN and GFAP (data not shown). 

Intravitreous injection of saline had no effect on OPN immunoreactivity compared with control, noninjected eyes (data not shown). This was consistent with the RT-PCR results, in which injection of saline elicited no change in retinal OPN mRNA expression. 

Comparison of Microglial-Derived mRNAs after Excitotoxic Injury and OPN Expression after Lipopolysaccharide Administration

Microglia are thought to be the principal source of proinflammatory mediators, such as interleukin (IL)-1β, tumor necrosis factor (TNF)-α, and nitric oxide (NO), which are elevated in the retina in a variety of pathologic conditions. Regulatory interactions between OPN and all three of these mediators have been described in other tissues; thus, it is of importance to ascertain in the retina whether production of OPN after excitotoxic injuries is linked to these inflammatory events. Initially, we determined whether the time-course of OPN upregulation in response to excitotoxic injury overlapped with increased expression of IL-1β, TNF-α, and iNOS. We found that NMDA injection resulted in strikingly similar patterns of expression of all four mRNAs: within 6 hours of excitotoxic injury, there was dramatic upregulation of the mRNAs, yet 3 days after the insult, the levels of all four transcripts had essentially returned to normal (Fig. 7A) . Subsequently, we established whether OPN is synthesized in the retina after administration of the prototypical proinflammatory toxin LPS and whether the same microglia that release proinflammatory mediators also express OPN. The results showed that OPN and TNF-α immunoreactivities were associated with activated microglia by 6 hours after injection (Figs. 7B 7C) , and double-labeling immunohistochemistry revealed a colocalization of the two molecules (Figs. 7D 7E 7F) . The combined data are consistent with the possibility that OPN regulates, or is regulated by, proinflammatory mediators. 

Discussion

OPN has been localized to the GCL of rat and mouse retinas, ^{27 28} but the cell type(s) responsible for expression were not confirmed by double immunolabeling. This result is important, as the rodent GCL contains numerous displaced amacrine cells, as well as microglia. Results presented herein demonstrate that in the normal retina OPN is only associated with RGCs, not with other cell types. Our findings, moreover, indicate that OPN is expressed by most, if not all, RGCs, since OPN was labeled in more cells than Brn-3, which is only expressed by a subset of the RGC population, but was colocalized with Thy1, which is found in all RGCs. The function of OPN in RGCs is unknown. 

In previous studies we, and others, have demonstrated that reliable and sensitive assessment of RGC injury can be achieved by measurement of the total retinal content of mRNAs encoding the RGC-specific markers Thy1, NF-L, and Brn-3. It was shown that downregulation of these mRNAs in response to damaging stimuli occurs in advance of histologic observation of cell loss and can be viewed as an early indicator of RGCs that are acutely stressed and dysfunctional. ^{7 29 31 32} As such, quantification of the levels of Thy1, NF-L, and Brn-3 mRNAs, together with histologic assessment of RGC loss comprise an ideal combination for evaluating the potential benefits of neuroprotective drugs on RGCs. In the present study we have confirmed these findings; however, we were unable to ascertain whether OPN mRNA is similarly downregulated by RGCs after ischemic–excitotoxic insults, since, unlike Thy1, NF-L, and Brn-3, OPN is additionally synthesized by activated microglia after injury. Thus, measurement of total retinal OPN mRNA is not solely a reflection of synthesis within RGCs. Transient upregulation of OPN expression by neurons has been documented after spinal root avulsion ³³ and experimental cryolesions, ³⁴ but not after kainic acid ²⁴ or ischemic insults. ^{20 21 22} In agreement with these latter findings, we did not observe increased OPN immunoreactivity in RGCs after excitotoxicity–ischemia. 

During the preparation of this article a study was published by Hikita et al., ²⁸ who investigated the involvement of OPN in murine experimental autoimmune uveitis (EAU), a well-established model of ocular inflammation. They showed that after induction of EAU, OPN is associated with putative resident microglia throughout the neural retina and choroid. In the present study, we have shown a similar association of OPN with resident microglia after excitotoxic and ischemic insults; thus, it appears likely that any retinal injury with an inflammatory component would result in synthesis of OPN by microglia. This conclusion is consistent with the results of other central nervous system studies (CNS) studies, in which an elevated tissue level of OPN was observed at sites of inflammation in the early period after various injuries. ^{20 21 22 23 24 33 35 36 37 38} Our study is the first to document increased expression of OPN after overactivation of NMDA receptors. Excitotoxicity resulting from prolonged NMDA receptor stimulation is implicated in the pathogenesis of neurodegeneration occurring in various acute and chronic disorders of the CNS. 

Despite the obvious similarities between the results of our study and those of Hikita et al., ²⁸ three notable differences were apparent that related to the tissue location, the duration and the cell types responsible for expression of OPN after injury. With regard to location, we identified OPN-positive microglia solely within the inner retina rather than the entire retinochoroid. This finding would be expected, however, as NMDA is detrimental only to retinal neurons expressing ionotropic glutamate receptors, which are subsets of RGCs and amacrine cells found in the inner retina; the outer retina is completely spared by NMDA. ³⁹ In contrast, EAU primarily affects the outer retina and choroid. ⁴⁰ It can be concluded therefore that synthesis of OPN by retinal microglia is limited to the immediate microenvironment of the injury. In relation to the other differences, expression of OPN was documented in EAU retinas at an advanced disease stage, and by infiltrating CD4-positive T cells and macrophages as well as activated microglia. ²⁸ Conversely, we observed immunoreactive OPN to be associated solely with activated microglia and only for a day or two after excitotoxic or ischemic injury, a time period that is likely to be before neuronal death. We did not observe OPN expression by any putative infiltrating monocytes/macrophages; nevertheless, it should be acknowledged that immunohistochemical delineation of macrophages from activated microglia is highly problematic because both cell types express many of the same markers, including ED1, OX-42, and F4/80. ⁴¹ The conclusion was based largely on the density of cells and assessment of morphology, infiltrating cells being characterized by their large, round shape and possessing few cytoplasmic processes. An explanation for the differing results is afforded by understanding the distinct characteristics of the paradigms of injury used. EAU is a T-cell-mediated autoimmune disease characterized by chronic inflammation, which displays similarities to other clinical and experimental autoimmune diseases. Prolonged expression of OPN at the site of injury by infiltrating leukocytes and macrophages, as well as resident microglia, has been described for several of these conditions. ^{42 43 44 45} The excitotoxic and ischemic paradigms of neurodegeneration used in our study display few pathologic similarities to EAU; for example, infiltration of CD4-positive T cells is not a hallmark of these conditions. Instead, they are analogous to many brain models of neurodegeneration. The short-lived time course of OPN expression observed in our study is comparable to those found in CNS models of focal ^{20 21 22} and global ²³ ischemia and kainic acid administration, ²⁴ and suggests that OPN may play distinct roles in autoimmune diseases and neurodegenerative conditions. 

The biological roles of OPN in pathologic conditions are hitherto poorly understood, a situation caused, in part, by the pleiotropic nature of the molecule; nevertheless, it is becoming apparent that OPN can have different effects on neuronal survival, depending on the nature of the injury. In experimental autoimmune diseases, a persuasive case can be made that OPN functions as a proinflammatory Th1 cytokine, that contributes to the initiation and propagation of the inflammatory process and thereby the progression of the disease. ^{28 42 46 47} However, an accumulating body of evidence indicates that OPN delivers an anti-inflammatory, prosurvival action in pathologic situations similar to those used in our study. ^{20 25 48 49 50 51 52 53 54 55 56} The function of OPN in the retina after excitotoxic and ischemic injuries remains to be established, but the rapid induction and transient expression of the OPN gene suggests that it is involved in early inflammatory events rather than in later events, such as tissue remodeling. Of particular interest to the retina is the well-documented capacity of OPN in macrophages to inhibit inflammatory mediator-induced synthesis of NO, ^{57 58 59 60} as microglia-derived NO is strongly implicated in the death process of RGCs in ischemia-like injuries. ^{61 62} Also of interest is the ability of OPN to counteract IL-1β-mediated toxicity and inflammation, ^{63 64} since IL-1β plays an important role in mediating ischemic and excitotoxic damage in the retina. ⁶⁵ In the present study, we have shown that the time-courses of OPN, iNOS, IL-1β, and TNF-α expression in the retina after excitotoxic injury were remarkably similar; moreover, we have demonstrated that microglia simultaneously express both OPN and TNF-α proteins after injection of the proinflammatory toxin LPS. The combined results are consistent with the hypothesis that OPN regulates, or is regulated by, these proinflammatory mediators. 

 Submitted for publication June 15, 2007; revised August 15, 2007; accepted December 17, 2007.

 Disclosure: G. Chidlow, None; J.P.M. Wood, None; J. Manavis, None; N.N. Osborne, None; R.J. Casson, 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.

 

Primer Sequences for mRNAs Amplified by Conventional and Real-Time RT-PCR Assays

Assay/mRNA	Primer Sequences	Product Size (bp)	Mg2+ Conc. (mM)	Annealing Temperature (°C)	Accession Number
Conventional PCR
Cyclophilin	5′-GAGAGAAATTTGAGGATGAGAAC-3′	373	5	59.5	M19533
	5′-AAAGAACTTCAGTGAGAGCAGAG-3′
NF-L	5′-TGCAGCTTACAGGAAACTCTT-3′	378	4.5	58.5	AF031880
	5′-TCACCACCTTCTTCTTCTTTG-3′
Brn-3b	5′-GGCTCGGAGGCGATGCGGAG-3′	183	4.5	58	XM344756
	5′-GTGGTAAGTGGCGTCCGCTTG-3′
Thy1	5′-CGCTTTATCAAGGTCCTTACTC-3′	344	4	52	X03150
	5′-GCGTTTTGAGATATTTGAAGGT-3′
Osteopontin	5′-GGAGTTTCCCTGTTTCTGATG-3′	372	4.5	55	M14656
	5′-ACTCGTGGCTCTGATGTTCC-3′
iNOS	5′-CGCTACACTTCCAACGCAAC-3′	407	4	55	L12562
	5′-AGGAAGTAGGTGAGGGCTTG-3′
IL-1β	5′-GCTACCTATGTCTTGCCCGT-3′	542	4	60	M98820
	5′-GACCATTGCTGTTTCCTAGG-3′
TNF-α	5′-TACTGAACTTCGGGGTGATTGGTCC-3′	295	4	60	X66539
	5′-CAGCCTTGTCCCTTGAAGAGAACC-3′
Real-Time PCR
Cyclophilin	5′-GTGTTCTTCGACATCACGGCT-3′	82	3	63	M19533
	5′-CTGTCTTTGGAACTTTGTCTGCA-3′
Osteopontin	5′-CCGATGAGGCTATCAAGGTC-3′	135	4	63	M14656
	5′-ACTGCTCCAGGCTGTGTGTT-3′
NF-L	5′-ATGGCATTGGACATTGAGATT-3′	105	4	63	AF031880
	5′-CTGAGAGTAGCCGCTGGTTAT-3′
Thy1	5′-CAAGCTCCAATAAAACTATCAATGTG-3′	83	3.5	63	X03150
	5′-GGAAGTGTTTTGAACCAGCAG-3′

 Conc., concentration.

Table 2.

 

Antibodies Used for Immunohistochemistry

Target	Host	Clone or Catalog No.	Dilution	Source
Osteopontin	Mouse	MPIIIB10	1:500	DSHB, Iowa City, IA
Brn-3	Goat	C-13	1:1,000	Santa Cruz Biotechnology, Inc., Santa Cruz, CA
Thy1.1	Mouse	OX-7	1:1,000	Advanced Targeting Systems, San Diego, CA
ED1	Mouse	MCA341	1:500	Serotec, Oxford, UK
NeuN	Mouse	A60	1:2,000	Chemicon, Boronia, Australia
PGP 9.5	Rabbit	RA95101	1:10,000	Ultraclone, Wellow, Isle of Wight, UK
TNF-α	Rabbit	HP8001	1:200	Hycult, Uden, The Netherlands
GFAP	Rabbit	Z0334	1:40,000	Dako, Botany, Australia
Osteopontin	Mouse	01-20002	1:200	American Research Products, Palos Verdes, CA

 DSHB, Developmental Studies Hybridoma Bank.

Figure 1.

 

Double labeling immunofluorescence of OPN with Brn-3, Thy1, and PGP 9.5 in the normal rat retina. OPN (A, D, G; green) was observed in numerous cells within the GCL. Brn-3 immunoreactivity (B; red) localized to the nuclei of a subset of RGCs, all of which were positive for OPN as shown in (C), the merged image. There was also a clear colocalization of OPN and Thy1 (E; red) immunoreactivities in RGC bodies, but not axon bundles (arrow), as shown in (F), the merged image. PGP 9.5 (H; red) labeled both RGCs and putative displaced amacrine cells in the GCL, as well as amacrine cells in the inner part of the INL (long arrows) and horizontal cells in the outer part of the INL (short arrows). The merged image (I) shows colocalization of OPN and PGP 9.5 in RGCs, but not in amacrine or horizontal cells. GCL, ganglion cell layer; INL, inner nuclear layer. Scale bars, 50 μm.

Figure 2.

 

Localization of OPN and RGC markers in the normal rat retina (A, B) and after axotomy (C–I). OPN immunoreactivity in the normal rat retina was localized to the cell body and, to a lesser extent, the dendrites of cells in the GCL (A). Some labeling was also evident in synapses between the dendrites of cells in the GCL and cell bodies in the innermost part of the inner nuclear layer (B; arrows). OPN (C), Brn-3 (D), and Thy1 (E) immunoreactivities were drastically reduced 28 days after axotomy. The continuing presence of displaced amacrine cells in the GCL after axotomy is indicated by labeling with NeuN (F) and PGP 9.5 (H). No colocalization was apparent between OPN and PGP 9.5 in axotomized retinas (G–I). GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; outer plexiform layer; ONL, outer nuclear layer. Scale bars: (A, C–F) 30 μm; (B) 20 μm; (G–I) 50 μm.

Figure 3.

 

(A) Gel analyses of conventional RT-PCR of retinas from eyes injected with NMDA (30 nanomoles), kainic acid (5 nanomoles) or vehicle. Both NMDA and kainic acid caused rapid decreases in the levels of the RGC-specific mRNAs Thy1, NF-L, and Brn-3 in treated (T) compared with control (C) eyes. In contrast, these excitotoxic agents induced a dramatic upregulation in the total retinal level of OPN mRNA. The results shown are from six representative animals, three of which were analyzed 12 hours after NMDA injection, and the remaining three 24 hours after kainic acid injection. (B–D) Quantitative real-time RT-PCR results: (B) Comparison of the levels of NF-L, Thy1, and OPN mRNAs after intravitreous injection of NMDA (30 nanomoles) and kainic acid (5 nanomoles). (C) Effect of intravitreous injection of NMDA (30 nanomoles) on the total retinal level of OPN mRNA after various time periods. (D) Comparison of NMDA (30 nanomoles), kainic acid (5 nanomoles), and ischemia–reperfusion on the total retinal level of OPN mRNA after 24 hours. In all cases, values (mean ± SEM) are normalized for the housekeeping gene cyclophilin and expressed as a fraction of the control eye, where n = 6 to 9. ***P < 0.001; **P < 0.01; *P < 0.05 by Student’s paired t-test (treated versus control eyes).

Figure 4.

 

OPN, Brn-3, and Thy1 immunoreactivities in control retinas (A, D, G, respectively), and in retinas 12 hours and 7 days after intravitreous injection of NMDA (30 nanomoles). Twelve hours after NMDA injection, there were no alterations in OPN, Brn-3, and Thy1 (B, E, H, respectively) labeling of RGCs. Unlike Brn-3 or Thy1, additional cells mostly contained within the IPL, were identified as OPN positive after 12 hours (B; arrows). Seven days after NMDA treatment, significantly fewer RGCs were present in the retina and there was a thinning of the IPL, which resulted in pronounced reductions in staining for OPN, Brn-3, and Thy1 (C, F, I, respectively). Remaining RGCs were still positive for all three markers, but no additional OPN-positive cells were present. Scale bar: 30 μm.

Figure 5.

 

(A–C) Time course of increase in OPN immunoreactivity by non-RGCs after intravitreous injection of NMDA (30 nanomoles). OPN immunoreactivity was associated with cells in the IPL by 6 hours after NMDA injection (A; arrows); this effect was more pronounced after 12 hours (B; arrows), but was no longer evident after 3 days (C; arrows). (D–I) Effect of intravitreous injection of NMDA (30 nanomoles) on ED1 and GFAP immunoreactivities. Ramified cells within the inner retina were positive for the activated microglia marker ED1 by 6 hours after NMDA injection (D). A slightly greater number of cells were positive for ED1 after 12 hours (E), and an even greater number were observed after 3 days (F). GFAP immunoreactivity was associated with astrocytes and Müller cell end feet, but only with occasional Müller cell process 6 hours after NMDA treatment (G). There was a small increase in GFAP in Müller cell processes by 12 hours (E). After 3 days, both astrocytes and Müller cell processes throughout the retina were strongly labeled for GFAP. Scale bars: (A–F) 20 μm; (G–I) 30 μm.

Figure 6.

 

Double-labeling immunofluorescence of OPN with the activated microglia marker ED1 in rat retina 12 hours after intravitreous injection of NMDA (30 nanomoles) (A–F) or 24 hours after ischemia (G–I). OPN immunoreactivity (A, D, G; green) was observed in ramified cells within the IPL. ED1 (B, E; red) always colocalized with OPN in these cells, as shown in (C, F) the merged images. Only a subset of ED1-positive cells were observed to express OPN (D–F; long arrows). Putative infiltrating macrophages were negative for OPN (D–F, G–I; short arrows). Scale bar: 50 μm.

Figure 7.

 

(A) Effect of intravitreous injection of NMDA (30 nanomoles) or vehicle on retinal expression of OPN, iNOS, IL-1β, and TNF-α mRNAs. NMDA induced considerable increases in all four mRNAs after 6 hours in treated (T) retinas relative to controls (C). By 3 days, however, there was little apparent difference between NMDA- and vehicle-injected eyes for any of the mRNAs with the exception of OPN which was still slightly elevated. The results shown are from six representative animals, three of which were analyzed 6 hours, and the remaining three 3 days, after NMDA injection. The level of the housekeeping gene cyclophilin was similar in treated and control eyes at both time points. (B–F) Localization of OPN and TNF-α 6 hours after intravitreous injection of lipopolysaccharide (LPS; 10 μg). OPN (B) and TNF-α (C) immunoreactivities were associated with cells in the GCL and IPL (arrows). There was a clear colocalization of OPN (D; green) and TNF-α (E; red) immunoreactivities in cells within the IPL, as shown in (F) the merged image. Scale bars: 50 μm.

The authors thank Ghafar Sarvestani of the Detmold Imaging Facility for expert technical assistance.