TNFα drives the proliferation and inflammatory response, but not regenerative potential of Müller cells in the mouse retina

Background Mouse Müller cells, considered dormant retinal progenitors, respond to retinal injury by undergoing reactive gliosis rather than displaying regenerative responses. Tumor necrosis factor alpha (TNFα) is a key cytokine induced after injury, and implicated in mediating inflammatory and regenerative responses. However, the molecular events driving reactive gliosis and regenerative responses in Müller cells, and the role of TNFα in these processes, remain unclear. In this study, we investigated the effects of TNFα on Müller cell responses following injury. Methods To investigate the involvement of TNFα in retinal injury, adult C57BL/6J mice were subjected to treatment with light (5,000 lux) for 14 consecutive days; induction of TNFα was confirmed by quantitative polymerase chain reaction (qPCR). TNFα effects on Müller-cell proliferation were evaluated via 5-ethynyl-2’-deoxyuridine (EdU) incorporation in culture. TNFα-mediated gene profile changes were examined using Affymetrix microarray, and gene ontology analysis was carried out to define the molecular pathways involved. Gene- and protein-expression changes were further verified by qPCR, western blot, and enzyme linked immunosorbent assay (ELISA). Results We showed that TNFα induced Müller cell proliferation and the expression of inflammatory and proliferation-related genes, including NFKBIA, Leukemia inhibitory factor, Interleukin-6, Janus kinase (Jak) 1, Jak2, Signal transducer and activator of transcription (Stat) 1, Stat2, Mitogen-Activated Protein Kinase (MAPK) 7, and MAP4K4. Blockade of Jak/Stat and MAPK pathways attenuated TNFα-induced Müller cell proliferation. Moreover, we detected TNFα drove A1 phenotype-reactive gliosis, while Wnt attenuated TNFα-mediated induction of A1 phenotype and promoted an Conclusion In Müller cells, TNFα triggered primarily inflammatory and reactive gliosis by activating Jak/Stat and MAPK-pathways without inducing progenitor cell/regeneration-related genes. Wnt signaling suppressed inflammation, and induced proliferation and expression of progenitor-cell genes in Müller cells. These results suggest that reactive gliosis and regenerative responses in Müller cells are regulated by independent mechanisms. Our study provides new insights into regulation of inflammatory and regenerative responses of Müller cells in the injured retina gene-chip TNFα stimulated the proliferation of Müller cells via Jak/Stat and MAPK We then used Jak3 inhibitor tofacitinib, and ERK inhibitor FR180204, to examine whether these pathways occur downstream of the TNFα signaling pathway, and whether they participate in regulating the proliferation of Müller cells. pathways. Our results indicate that TNFα signaling can transiently stimulate the proliferation of Müller cells by activating MAPK and Jak/Stat signaling pathways during early stages of light-induced retinal injury. Our present study had several limitations. TNFα-mediated promotion of mammalian Müller-cell proliferation and differentiation needs to be verified in vivo, which we will address in our future studies. However, this study provides new insights into how light-induced retinal injury regulates the proliferation of mammalian Müller cells, and helps us understand the mechanisms of Müller-cell proliferation in mammalian injured retina.

However, molecular pathways that drive reactive gliosis and regenerative responses in Müller cells are unknown.
In Müller cell, proliferation is associated with their dedifferentiation into progenitor cells and activation of their regenerative potential. Under certain conditions, such as following stimulation with α-aminoadipic acid [10], epidermal growth factor (EGF) [11,12], fibroblast growth factors (FGFs), or insulin [13], 13 Müller cells of adult mice reenter the cell cycle and generate new retinal neurons. Wnt and Notch signaling are well-known cellular events that participate in Müller-cell proliferation and regenerative processes of the mammalian retina [14][15][16]. The sonic hedgehog (Shh), and MAPK and Jak/Stat signaling pathways also play important roles in stimulating the proliferation of Müller cells [11,[17][18][19]. However, induction of Müller-cell proliferation by these factors occurs only in the presence of retinal injury such as that caused by light exposure [20,21] or treatment with N-methyl-D-aspartate (NMDA) [11,14] or N-methyl-N-nitrosourea (MNU) [15,17]. The mechanisms by which these stimulations interact with injury signals to induce Müller cell proliferation remain elusive.
TNFα, a potent cytokine induced after retinal injury, is involved in numerous biological processes such as cellular apoptosis, survival, and proliferation [22].
TNFα activates Stat3-mediated proliferation of Müller cells in the zebrafish retina [5] and in fetal human cortical neural progenitors [23]. In light-damaged zebrafish retina, enhanced proliferation of Müller cells is accompanied by upregulated expression of TNFα. In undamaged retina, administration of TNFα induces proliferation of Müller glia via ASCL1a and Stat3 signaling pathways. 5 Although TNFα is considered a key inflammatory cytokine in mammals [24,25], few studies have examined the effect of TNFα on Müller-cell proliferation. It is still unclear whether TNFα serves as key injury signal that promotes Müller-cell proliferation and, thereby, drives retinal neuroregeneration or repair following injury in the mouse retina. In the present study, we used EdU incorporation assay, gene chip analysis, real-time PCR, and western blotting to examine changes in the proliferation and gene expression of Müller cells obtained from the mouse retina and subjected to treatment with TNFα.
Primary Müller cells were cultured, and adherent cells were passaged twice before being used in further experiments.The purity of Müller cells was evaluated by immunocytochemistry using immunolabeling with primary antibodies specific for glutamine synthetase (GS) [28] and vimentin [29], which are biomarkers of Müller cells.

Enzyme-linked immunosorbent assay (ELISA)
The levels of TNFα in mouse Müller-cell cultures were quantified using a TNFα mouse ELISA kit (Invitrogen) in accordance with manufacturer's instructions.
Absorbance was read at 450 nm using a microplate reader (please provide manufacturer's information for this instrument.

Cellular proliferation assay
To assess cellular proliferation, Müller cells were dissociated into single-cell suspension using 0.25% EDTA-trypsin (Invitrogen), seeded at a density of 4,000 cells per well in 96-well culture plates containing DMEM/F12 medium supplemented with 10% FBS. TNFα (50 ng/ml; R&D systems, Minneapolis, MN) was added into the culture media of the treated groups of Müller cells, and the cells were allowed to incubate for 24 hours at 37℃.Cell proliferation was evaluated via Click-iT EdU Alexa Fluor 555 Imaging Kit (Invitrogen) [30,31] in accordance with manufacturer's instructions. Briefly, EdU (10μΜ) was added to cell-culture media, and cells were incubated for 4 hours at 37℃; cells were then fixed and incubated with Hoechst Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as loading control.

Simple western analysis
Simple western analysis was performed using 12-230 kDa Wes Separation Module and 8x25 capillary cartridges in accordance with the manufacturer's protocol (ProteinSimple, California, USA). Briefly, after preparing standard pack reagents, samples, and primary antibodies (diluted 1:50) in accordance with manufacturer's instructions, the samples and biotinylated ladder were denatured in a heating block for 5 min and stored on ice until use. Luminol-S and peroxide, supplied in the detection module, were combined in a microcentrifuge tube. The biotinylated ladder, samples, primary antibodies, secondary antibodies, and reagents were dispensed into the assay plate using volumes shown in the plate diagram, and the plate was centrifuged for 5 minutes at 2500 rpm and RT. After completion of the analysis, results were evaluated using Compass software v3.1The following monoclonal primary antibodies were used in this procedure: rabbit anti-GAPDH (CST), rabbit anti-p44/42 MAPK (Erk1/2, CST), and rabbit anti-Phospho-p44/42 MAPK (CST). ERK inhibitor FR180204 (10 μM, Selleck) was added 2 hours, and exogenous TNFα (R&D systems) was added 30 min, before the cells were harvested.

Statistical analysis
All data were analyzed using SPSS 19.0 (IBM Corporation, Armonk, NY), and Student's t-test was used to compare differences between two groups All results are shown as mean ±SD; p <0.05 was considered statistically significant.

Müller cells derived from light-damaged mouse retina show increased proliferation
To investigate the effect of retinal damage induced by over exposure to light, C57BL/6J mice over the age of 6-8 weeks were subjected to 5,000 lux of cool white LED light, for 14 consecutive days. As shown by HE staining (Fig.1A-C), after 14 days of exposure to light, numerous photoreceptor cells were lost, and the thickness of the outer nuclear layer (ONL) was significantly (** p≤0.01) decreased in lightdamaged retina compared with that of the control retina. Retinas, obtained from mice subjected to light exposure, and from unexposed control mice, were used for primary cell culture. Some of the cells obtained from these retinas were preserved for passaging. The purity of these cells in passage 2 was evaluated by immunocytochemistry using antibodies specific for Müller-cell markers GS and vimentin (Fig 1D-F). More than 90% of cells showed positive labeling for GS and vimentin, indicating that these Müller cells here highly pure. We then used an EdU assay to examine the proliferation capacity of Müller cells from light-treated and control mice. Müller cells obtained from light-treated mice showed a higher EdU+/Hoechst+ ratio (23.0%) than that of Müller cells obtained from controls (17.2%, p≤0.01, Fig. 1G-I); this finding indicates that the proliferative capacity of Müller cells was increased in a light-damaged mouse retina.  Fig 2B). In addition, densitometric analyses via western blotting showed that TNFα protein levels were elevated in light-damaged retinas compared with the levels of control retinas (Fig 2C,D). Taken together, these results indicate that TNFα expression increased in Müller cells derived from the retinas of light-damaged mice.

TNFα promotes the proliferation of Müller cells in vitro
Previous studies indicate that Müller cell proliferation is enhanced by TNFα upregulation in the light-damaged retina. 5 To investigate whether increased TNFα concentration is required for accelerated proliferation of Müller cells in lightdamaged mice, we blocked the effect of TNFα by adding a TNFαneutralizing antibody to the culture medium of Müller cells isolated from adult lightdamaged mice; we then evaluated the proliferation capacity of these cells via EdU assays. Our results indicate that the ratio of EdU+ /Hoechst+ cells increased by 33.7 % in the light-damaged group (p≤0.01; vs. controls groups), and decreased by 16.2% in the cell-culture medium of light-induced mice treated with a TNFαneutralizing antibody (EdU+/Hoechst+ ratio =17.2% vs. light-damaged group (EdU+ /Hoechst+ ratio =23.0%) (Fig. 3A). This indicates that a neutralizing antibody inhibited Müller cell proliferation in the light-damaged retina.
We then added exogenous TNFα to Müller cell cultures to investigate its effect on cell proliferation in vitro. The results of proliferation assay showed that after treatment with 50 ng/ml TNFα, the relative proliferation rate of Müller cells isolated from PN 3-day mice increased by 42.6% (p< 0.001) (Fig 3B)

Microarray analysis of TNFα-regulated genes in Müller cells
Next, we aimed to determine the transcription factors and signaling pathways  Table 2).
We then examined the expression of MAPK and Jak/Stat signaling pathways involved in cellular proliferation. Microarray analysis revealed that six genes in these pathways (Jak1, Jak2, Stat1, Stat2, Map4k4, Mapk7) were upregulated after treatment with TNFα , as shown in Table 3. The results of real-time PCR and RT-PCR were consistent with those of microarray analysis (Fig. 4E, F, Table 3). Expression of inflammatory signaling pathways, including TNF and NF-kappa B signaling pathways, and related genes (Nfkbia, Lif, and Il-6) showed significantly enhanced expression after treatment with TNFα (Fig. 4C, F and Table 3).
The mRNA expression of genes related to neuronal stem cells is shown in Fig 6D,F and Table 3. Both microarray and real-time PCR analysis showed that mRNA expression of neuronal stem-cell markers, such as Hes1, Wnt2, Sox9, was significantly downregulated in Müller cells treated with TNF-α. This indicates that neuronal stem-cell signaling pathways may be downregulated by increased expression of TNF-α.

TNFα promotes Müller-cell proliferation via Jak/Stat and MAPK pathways
The results of gene-chip analysis indicate that TNFα stimulated the proliferation of Müller cells via Jak/Stat and MAPK signal pathways. We then used specific signalingpathway inhibitors, including Jak1/2 inhibitor ruxolitinib, Jak3 inhibitor tofacitinib, and ERK inhibitor FR180204, to examine whether these pathways occur downstream of the TNFα signaling pathway, and whether they participate in regulating the proliferation of Müller cells.
As shown in Fig 5A and  We additionally found that the phosphorylation level of ERK, an important factor in the MAPK pathway, was elevated after TNFα was added to Müller-cell culture. When TNFα was added to Müller-cell cultures containing an ERK inhibitor, the phosphorylation level of ERK was inhibited, and proliferation of Müller cells decreased compared with that of Müller cells treated only with TNFα (EdU+/ Hoechst+ cell ratio from 21.1 to 15.7%, p<0.001, Fig 5B,D). These results indicate that the MAPK signaling pathway acts downstream of the TNFα-pathway to regulate the proliferation of Müller cells.

TNFα drives Müller cells toward A1 phenotype while Wnt counteracts this effect
Neuroinflammation and ischemia can induce two different types of reactive astrocytes called "A1" and "A2," respectively. The A1 type is induced by activated microglia. A1 astrocytes lack most of the normal astrocyte functions, but gain a new neurotoxic function that induces rapid death in neurons and oligodendrocytes. A2s, which are induced by ischemia, are considered neuroprotective and can regulate the activity of numerous neurotrophic factors [32]. Because gene chip analysis indicated that neuronal-stem-cell signaling pathways were downregulated in Müller cells treated with TNFα, we hypothesized that TNFα may induce Müller cells toward reactive gliosis rather than toward neurogenesis. Therefore, we performed RT-qPCR to examine the glial phenotype after Müller cells were induced toward reactive gliosis.
After Müller cells were treated with 50 ng/ml TNFα for 24 h, mRNA expression of most genes in cells with A1 phenotype was significantly upregulated compared with that of controls ( Fig 6A); mRNA expression of most genes in cells with an A2 phenotype showed no significant changes after treatment with TNFα ( Fig 6B).After Müller cells were treated with 100 ng/ml Wnt3a for 24 h, mRNA expression of several genes in cells with an A1 phenotype was significantly downregulated compared with that of controls; other genes in cells with an A1 phenotype showed upregulated expression, but these changes were not statistically significant ( Fig   6A). Cells treated with 100 ng/ml Wnt3a and induced toward A2 phenotype showed significant up-and downregulation of several genes, as shown in Fig 6B. When Müller cells were treated with a combination of 50 ng/ml TNFα and 100 ng/ml Wnt3a for 24 h, most genes in cells with an A1 phenotype showed significantly upregulated expression compared with that of controls, but the foldchange was less than that observed using 50 ng/ml TNFα alone (Fig 6A). The expression of genes in cells with A2 phenotype was significantly upregulated compared with that of controls (Fig 6B). Taken together, these results indicate that treatment with TNFα drove Müller cells toward A1 phenotype, while treatment with Wnt counteracted the effects of TNFα. Treatment with both TNFα and Wnt promoted an A2-like phenotype, but Wnt alone was not sufficient for driving cells toward either an A1 or A2 phenotype.

DISCUSSION
Our present study shows that TNFα expression was increased in mouse lightdamaged retina, and that TNFα regulated the proliferation of Müller cells via MAPK and Jak/Stat pathways. Treatment with TNFα alone promoted an inflammatory response and glial-cell proliferation, but did not promote neurogenesis, in the mouse retina.
Subjecting Müller glia of adult rats to light injury drives these cells toward reactive gliosis, which enables them to survive long-term in culture [33]. Studies using retinal explant cultures derived from light-injured rat retina show that Müller glia reenter the cell cycle in response to light injury [15]. In our present study, Müller cells isolated from light-damaged retina of adult mice showed a tendency toward survival and proliferation. Results obtained using an EdU assay demonstrated that Müller cells subjected to light damage showed a higher proliferation ratio than that of controls. The mRNA and protein expression of TNFα increased significantly after induction of light exposure, indicating that TNFα is involved in Müller-cell proliferation in injured mice retina.
TNFα is a pro-inflammatory cytokine involved in inflammation, immune response, synaptic function and cellular proliferation [34][35][36][37]. shown that TNFα expression is increased in injured or degenerated mammalian retinal Müller cells [24,25]. In this study, we found that TNFα expression was found that exogenous TNFα can induce in-vivo proliferation of Müller glia in undamaged zebrafish retina via Stat3 and ASCL1a signaling pathways [5]. In rodent and chick retinas, ciliary neurotrophic factor (CNTF) /Jak/Stat signaling appears to stimulate glial reactivity and enhance neuroprotection [38,39]. pStat3 levels are upregulated in HB-EGF-treated Müller glia obtained from NMDA-damaged mouse retinas, indicating that the Jak/Stat pathway is involved in proliferation of mammalian Müller cells [11]. In our present study, we found that numerous genes involved in activation of the Jak/Stat pathways, including Jak1, Jak2, Stat 1, and  FGF2: basic fibroblast growth factor 2; IGF1: insulin-like growth factor 1