Oncostatin M Protects Rod and Cone Photoreceptors and Promotes Regeneration of Cone Outer Segment in a Rat Model of Retinal Degeneration

Retinitis pigmentosa (RP) is a group of photoreceptor degenerative disorders that lead to loss of vision. Typically, rod photoreceptors degenerate first, resulting in loss of night and peripheral vision. Secondary cone degeneration eventually affects central vision, leading to total blindness. Previous studies have shown that photoreceptors could be protected from degeneration by exogenous neurotrophic factors, including ciliary neurotrophic factor (CNTF), a member of the IL-6 family of cytokines. Using a transgenic rat model of retinal degeneration (the S334-ter rat), we investigated the effects of Oncostatin M (OSM), another member of the IL-6 family of cytokines, on photoreceptor protection. We found that exogenous OSM protects both rod and cone photoreceptors. In addition, OSM promotes regeneration of cone outer segments in early stages of cone degeneration. Further investigation showed that OSM treatment induces STAT3 phosphorylation in Müller cells but not in photoreceptors, suggesting that OSM not directly acts on photoreceptors and that the protective effects of OSM on photoreceptors are mediated by Müller cells. These findings support the therapeutic strategy using members of IL-6 family of cytokines for retinal degenerative disorders. They also provide evidence that activation of the STAT3 pathway in Müller cells promotes photoreceptor survival. Our work highlights the importance of Müller cell-photoreceptor interaction in the retina, which may serve as a model of glia-neuron interaction in general.


Introduction
Retinitis pigmentosa (RP) is a heterogeneous group of inherited retinal degenerative disorders that affect one in 3,500 to 4,000 people [1]. Mutations in 40 genes identified so far are implicated in RP, most of which selectively affect rods (http://www.sph.uth. tmc.edu/retnet/sum-dis.htm). In early stages of RP, patients typically experience night blindness and decline in peripheral vision, due to loss of rod photoreceptors in the peripheral retina. In many cases even though the mutations do not directly affect cone photoreceptors, cones undergo secondary degeneration following rods [1,2]. The secondary cone degeneration eventually affects the central vision, leading to total blindness. No effective treatments are currently available for RP. How to effectively protect photoreceptors, especially cone photoreceptors, is a major challenge in retinal degeneration research.
Pre-clinical studies have shown that photoreceptors can be protected by exogenous neurotrophic factors. LaVail and colleagues [3] first demonstrated in the RCS (the Royal College of Surgeon) rats that a single injection of FGF-2 successfully rescues photoreceptors from degeneration [3]. A subsequent screen of a panel of factors identified several neurotrophic factors that protect photoreceptors in a light-damaged model, including CNTF (ciliary neurotrophic factor), a member of the interleukin-6 (IL-6) family of cytokines [4]. Many studies since confirmed the protective effects of CNTF in animal models cross several species. These pre-clinical results led to a phase 1 clinical trial with promising results [5]. Two other members of the IL-6 family of cytokines, including CT-1 (cardiotrophin 1) and LIF (leukemia inhibitory factor), have been reported to protect photoreceptors in different animal models [6,7].
Oncostatin M (OSM), originally isolated from a tumor cell line and identified as a factor to suppress the growth of tumor cells, is also a member of the IL-6 family of cytokines [8,9,10]. It expresses mainly in activated T lymphocytes, monocytes [8,11], but also in other tissues [12], including in neurons, astrocytes, and microglia [13]. OSM has a variety of biological effects, including inhibition of solid tumor cells, inhibition of totipotent embryonic stem cells, and induction of acute phase protein synthesis in hepatic cells [10]. In the nervous system, OSM induces the expression of IL-6 and other cytokines, regulates inflammatory reactions as well as matrix remodeling [13]. It also suppresses brain tumor cells [13]. Elevated levels of OSM have been detected in several human CNS pathologies, such as multiple sclerosis, HIV-associated dementia, and epileptic seizure [14,15,16].
In the present work, we studied the neuroprotective feature of OSM in a transgenic rat model of retinal degeneration. Our data show that OSM not only protects both rod and cone photoreceptors, but also promotes regeneration of cone outer segments (COS) in degenerating cones. Furthermore, our results suggest that the effects of OSM are mediated through the Jak/STAT3 pathway in Müller cells.

Protection of Rod Photoreceptors by OSM
Mature human recombinant OSM, fused to a 6xHis tag at the amino terminus, was expressed in E. coli, purified by nickel columns, and buffer exchanged with phosphate buffered saline (PBS) (see Materials and Methods for details). The purified recombinant protein has an apparent size of ,25 kD after electrophoresis in acrylamide gel (Fig. 1). Although OSM is a glycoprotein, recombinant OSM expressed in E. coli is folded in the native conformation and fully functional as the native protein produced in human cells [17].
To examine the effects of OSM on photoreceptor degeneration, we used the heterozygous S334ter-3 rats as a model of retinal degeneration with rapid degeneration of rod photoreceptors [18]. Intravitreal injection of OSM (10 mg in 3 ml PBS) at PD9 resulted in significant protection of rod photoreceptors when examined at PD20. As shown in Fig. 2, the outer nuclear layer (ONL) of the retina in normal wild type animals has 10-12 rows of photoreceptor nuclei ( Fig. 2A), whereas the ONL of S334ter animals treated with PBS had only one row (Fig. 2B). In contrast, OSM-treated retinas typically had 6-8 rows of photoreceptor nuclei in the ONL (Fig. 2C). Measurement of the ONL thickness showed that the thickness in the OSM-treated retinas (23.1161.70 mm, n = 3) is significantly greater (P = 0.002, Student t test) than in PBS-treated control retinas (5.2260.19 mm).

Protection of Cone Photoreceptors by OSM
To examine whether OSM protects cone photoreceptors as well, we used the same rat model in which the secondary cone degeneration develops following rod degeneration, as has been characterized recently [19]. In these experiments, eyes of homozygous S334ter rats were injected with OSM (10 mg in 3 ml PBS) at PD 20 and retinas were collected at PD 30. COS were identified by PNA (peanut agglutinin) staining. In the PBS-treated control retinas, loss of COS was found in numerous small PNAnegative areas throughout the retina (Fig. 3A). This finding is consistent with previous observations [19]. In OSM-treated retinas, the PNA-negative areas became very small and in many cases completely disappeared (Fig. 3B). Quantitative analysis showed that the density of PNA-positive cells in the OSM-treated retinas is 905635.44/0.15 mm 2 (n = 6), significantly higher than that in the PBS-treated retinas (742.83636.14, n = 6) (P,0.01, Student t-test) (Fig. 3C). These results indicate that OSM

COS Regeneration Induced by OSM Treatment
The disappearance of PNA negative areas in the OSM-treated retinas ( Fig. 3B) also suggested to us that OSM promoted regeneration of COS in cones that had lost COS. To test this hypothesis, we performed experiments in which retinas were treated with OSM at PD35 and examined at PD45. As shown in Fig. 4, it appears that there are more PNA-positive cells in OSMtreated retinas at PD45 (Fig. 4C) than in untreated retinas at PD 35 ( Fig. 4A) or PBS-treated retinas at PD45 (Fig. 4B). Quantitative analysis showed that indeed, the density of PNA-positive cells in the OSM-treated retinas at PD45 (665616.56/0.15 mm 2 , n = 3) is significantly higher than that of the untreated retinas collected at PD35 (565.3366.36, n = 3) (P,0.01, ANOVA and Tukey test) or PBS-treated retinas at PD45 (464618, n = 3) (P,0.001, ANOVA and Tukey test) (Fig. 4D). The fact that the density of PNA-positive cells in the OSM-treated retinas at PD45 is higher than that in the retinas at PD35 provides evidence of COS regeneration in response to OSM treatment. This finding is consistent with our previous results with CNTF treatment [19].

Phosphorylation of STAT3 Induced by OSM
We next examined the phosphorylation of STAT3 in the retina after OSM treatment. A significant increase in STAT3 phosphorylation was detected as early as 30 min after OSM injection (Fig. 5A). The treatment also induced a small increase in STAT3 protein level in the retina (Fig. 5A). In contrast, PBS injection induced a much smaller increase with shorter duration in phosphorylated STAT3 (Fig. 5B), which is consistent with our previous observations [6].
To localize OSM-induced STAT3 phosphorylation, wild-type rats were treated with OSM in the left eyes and PBS in the right eyes. Eyes were collected 1 hour later. Phosphorylated STAT3 was detected in the OSM-treated retina in a group of cells in the inner nuclear layer (INL). Those cells were also positive for glutamine synthetase (GS), a Müller cell specific marker, (Fig. 6D-6F, high magnification 6G-6I). The colocalization of pSTAT3 and GS confirms that OSM induces STAT phosphorylation in Müller cells. In contrast, no pSTAT3 was detected in the outer nuclear layer (ONL) where photoreceptor cell bodies reside, indicating that photoreceptors are not directly responsive to OSM (Fig. 6D, 6F). Phosphorylated STAT3 was not detected in the control retina (Fig. 6A). These results provide evidence that OSM induces STAT3 activation specifically in Müller cells, not in photoreceptors.

Discussion
The present work has shown that OSM protects both rod and cone photoreceptors from degeneration. OSM treatment also promotes regeneration of COS in degenerating cones. We have further demonstrated that OSM induced STAT3 activation in Müller cells, not in photoreceptors, suggesting that OSM-induced photoreceptor survival is mediated by Müller cells.
OSM is a member of the IL-6 family of cytokines with significant homology of amino acid sequence and structure with other members of the family, especially with LIF [9,10,20]. As a pleiotrophic cytokine, OSM exerts many functions that overlap those of related cytokines in a variety of systems, including  inflammation, cell proliferation, and hematopoiesis, [10,21,22]. The expression of OSM is found in various tissues, including the brain [12]. In the brain, OSM expression was reported in hippocampus [23] and pituitary gland [24]. Upregulation of OSM is found in multiple sclerosis lesions in the brain in microglia, reactive astrocytes, infiltrating leukocytes [25], as well as in peripheral blood mononuclear cells [26]. It is also found in interneurons and glial cells in the hippocampus after prolonged seizure [16]. Significant increase in OSM expression was also seen after sciatic nerve crush or transaction [27]. In the OSM-deficient mice, there is a significant decrease in neurons in the dorsal root ganglia [28]. The loss of neurons in the dorsal root ganglia is accompanied with reduced noxious responses to acute thermal, mechanical, chemical, and visceral pain, indicating the essential function of OSM in nociceptor neurons [28]. Furthermore, OSM have been shown to be neuroprotective in models of excitotoxic injury [29,30]. Our present finding that OSM protects photoreceptors is consistent with OSM being a neurotrophic cytokine, similar to CNTF, LIF, and CT-1.
Two OSM receptor complexes have been identified in human for OSM signal transduction. The type I receptor complex is a heterodimer of LIF receptor b (LIFRb) and gp130, the same receptor complex shared by several IL-6 family of cytokines, including CNTF, LIF, and CT-1 [31,32]. The type II OSM receptor complex is composed of the OSM specific receptor OSMRb and gp130. Human OSM binds to both the type I and type II human receptors [22,33]. This is not the case in mouse, however. Mouse OSM is unable to transduce signals through mouse LIFRb. It uses only the receptor complex of mouse OSMRb and gp130 for its biological activities. On the other hand, human OSM can activate mouse LIFRb, but not the mouse OSMRb [33]. In our experiments, we used recombinant human OSM on rat model of retinal degeneration. Although it is not clear whether human OSM interacts with rat OSMRb, it does induce biological effects on rat cells [34,35]. It is likely that in rat the human OSM effects are also mediated by the receptor complex of LIFRb and gp130.
The finding that OSM induces STAT3 phosphorylation in Müller cells but not in photoreceptors is in agreement with previous findings that CNTF and CT-1, two members of the IL-6 family of cytokines, activate the STAT3 pathway in Müller cells but not in photoreceptors [6,36]. These consistent results strongly suggest that the protective effects of OSM are mediated by Müller cells, which express receptors for cytokines of the IL-6 family [37]. Studies have shown that activation of the STAT3 pathway in Müller cells by CNTF regulates the phototransduction machinery in rod photoreceptors, similar to light-induced photoreceptor plasticity [36,38]. Together, these results emphasize the importance of Müller cell-photoreceptor interaction in regulating the behavior of photoreceptors. This glia-neuron interaction in the retina may serve as a model of such interaction in other parts of the nervous system.
In summary, our present study demonstrates the neuroprotective effects of OSM on photoreceptor degeneration and suggests the role of Müller cells in mediating the protection. Our work provides additional evidence to support the therapeutic approach of using members of the IL-6 family of cytokines, especially  CNTF, for retinal degenerative disorders. Recent results from clinical trials showed that CNTF treatment (CNTF-secreting implants) improves vision and stops the loss of cones in patients with retinal degeneration and geographic atrophy [5,39,40]. The clinical results demonstrate that neuroprotection is a viable approach for treating retinal degenerative disorders.

Expression and Purification of Human Recombinant OSM Protein
The open reading frame of mature human OSM cDNA was cloned by PCR and subcloned into the expression vector pQE30 (Qiagen, Valencia, CA), fused to a 6xHis tag at the N-terminus to generate plasmid pQE-OSM. Recombinant human OSM protein was expressed in E. coli (XL-blue; Stratagene, La Jolla, CA) and purified by immobilized-metal affinity chromatography on Ni-NTA agarose columns (Qiagen) under native conditions. Eluted protein was buffer-exchanged to PBS, and stored at 280uC in small aliquots. The purified recombinant protein has an apparent size of about 25 kD after electrophoresis (Fig. 1).

Animals and Intravitreal Injections
Transgenic S334ter-3 rats (Sprague-Dawley background) carrying a murine rhodopsin mutant S334ter [18] and wild-type Sprague-Dawley (Charles River Labs, Wilmington, MA) rats were used in all experiments. Heterozygous S334ter-3 rats were produced by mating homozygous male breeders with wild-type Sprague-Dawley females. Intravitreal injections were delivered through 33gauge needles connected to 10-ml microsyringes (Hamilton, Reno, NV), as described previously [36]. The left eye of an animal was injected with OSM protein and the right eye with PBS.

Histology, PNA Staining, and Immunocytochemistry
Animals at a given endpoint were killed by CO 2 overdose, immediately followed by vascular perfusion with mixed aldehydes [18]. Eyes were embedded in an Epon/Araldite mixture, sectioned at 1 mm thickness to display the entire retina along the vertical meridian. Retina sections were examined by light microscopy.
Cone outer segments were identified by PNA staining, as described previously [19]. Briefly, animals were killed by CO 2 inhalation and perfused with PBS. Retina-lens preparations were post-fixed in cold 4% paraformaldehyde solution for 4 hours at 4uC. After rinse in PBS, the retina-lens preparations were incubated with Alexa fluor 488-conjugate PNA (Invitrogen, Carlsbad, CA). Retinas were flat-mounted on slides after lenses were removed and examined by confocal microscopy.
For immunostaining, animals were killed by CO 2 inhalation and perfused with 4% paraformaldehyde. Eyecups were obtained, cryoprotected with 30% sucrose, embedded in the OCT compound (Tissue-Tek; Miles Inc., Elkhart, IN), and frozen in liquid nitrogen. Cryo-sections (12 mm) were incubated with antiphospho-STAT3 (Tyr705) antibodies (Cell Signaling Technology, Beverly, MA). Immunoreactivity was visualized using the ABC Elite kit (Vector Laboratories, Burlingame, CA) and the Tyramide Signal Amplification Detection System (Invitrogen) according to manufacturers' instructions. Müller cells were identified by antibodies against glutamine synthetase (GS; Millipore Corporation, Temecula, CA). GS immunoreactivity was visualized by Cy3conjugated secondary antibodies (Jackson Immunoresearch; West Grove, PA). Sections were examined by confocal microscopy.

Immunoblotting Analysis
Retinas were dissected, snap-frozen in powdered dry ice, and stored at 280uC. Retinas were homogenized and concentration of total protein in each sample was determined by the Bio-Rad protein assay (Bio-Rad Labs, Hercules, CA). Total protein of 20 mg from each sample was electrophoresed in 10% NuPAGE gels (Invitrogen) and transferred to nitrocellular membranes (Bio-Rad Labs). Blots were probed with anti-phosphorylated STAT3 (Tyr705) antibodies (Cell Signaling Technology), visualized using SuperSignal chemiluminescent substrates (Thermo Fisher Scientific) and recorded on Hyperfilm (GE Healthcare, Piscataway, NJ). Blots were then stripped and re-probed with anti-STAT3 antibodies (Cell Signaling Technology).

Statistical Analysis
Results were analyzed by either Student t test or ANOVA followed by Tukey test for comparisons between different groups, using InStat3 (GraphPad Software Inc., San Diego, CA). Data are expressed as mean 6 SD.