Main

Throughout our life span, cells have to be repaired or regenerated in order to maintain tissue function. Whereas this happens frequently in most parts of the body, the regeneration capacity of the nervous system is generally very limited in mammals, and the consequences of neuronal disease or injury are mostly irreversible. Therefore, it is crucial to rescue neurons from devastating insults, a task that requires activation of endogenous neuroprotective systems. In the retina, several pathways exist to support the survival of neurons.1, 2, 3, 4, 5, 6, 7 Leukemia inhibitory factor (LIF) controls one of these pathways and is one of the most important endogenous factors for neuroprotection in the retina.1, 2 Photoreceptor injury or degeneration activates a subset of Muller glia cells to express Lif, which controls a cascade of neuroprotective signaling between photoreceptors and Muller cells.1, 8 These signaling events include activation of the Janus kinase/signal transducer and activator of transcription (JAK/STAT) pathway,9, 10, 11 and result in the upregulation of several genes important for neuroprotection, including signal transducer and activator of transcription-3 (Stat3), endothelin-2 (End2) and fibroblast growth factor-2 (Fgf2), and for gliosis (glial fibrillary acidic protein, Gfap).1, 12 In the absence of LIF, none of these factors is induced, and photoreceptor cell death is accelerated.1, 12

Although downstream events of LIF signaling have been studied in the retina,1, 12, 13, 14 the signal that initiates Lif expression has not been elucidated. One of the hypotheses for the initiation of survival pathways is based on the generation of reactive oxygen species (ROS) in stressed photoreceptors.15, 16, 17 Subtoxic levels of ROS have been shown to be neuroprotective for photoreceptors and ROS may act as signaling molecules for survival pathways in the retina.15, 16, 17, 18 Another hypothesis especially with respect to the involvement of Muller cells includes tumor necrosis factor-alpha (TNF) signaling, as TNF has been recently shown to be the key signaling molecule for Muller cell proliferation and differentiation into a photoreceptor fate in the degenerating zebrafish retina.19 However, its role during photoreceptor degeneration in the mammalian retina has not been identified in detail.

TNF was shown to regulate expression of several important factors that mediate a proinflammatory response. Likewise, TNF treatment upregulates several cytokines including Lif in various cell types.20, 21 The reported neuroprotective effect of TNF is mostly attributed to nuclear factor kappa-light-chain-enhancer of activated B cells (NFκB) activation and the resulting differential regulation of gene expression.22, 23, 24, 25, 26, 27 However, divergent data exist and various reports attribute TNF also a role in the promotion of neurodegeneration (reviewed in detail22, 23). Blocking TNF in a glaucoma model of retinal ganglion cell (RGC) death, for example, had beneficial effects on RGC survival.28, 29, 30 In contrast, increased survival of RGCs after axotomy of the optic nerve required elevated and not decreased TNF levels.31

In addition to NFκB, TNF has been shown to activate the p38 mitogen-activated protein kinase (MAPK) pathway in a variety of models, and a small number of genes have been identified that are regulated through p38 MAPK after TNF induction.27 However, similar to TNF, no consensus exists on the anti- or pro-apoptotic effects of p38 MAPK activity in neuronal tissues. Active p38 MAPK signaling has been shown to contribute to RGC death after ischemia or optic nerve axotomy.32, 33 In contrast, recent findings suggest that p38 MAPK activity is important for RGC survival after ischemia/reperfusion injury,34, 35 and that active crosstalk between NFκB and p38 MAPK pathways may be an important aspect of this neuroprotection.35

To date, the effects of TNF and p38 MAPK on photoreceptor protection and their possible role in the regulation of neuroprotective factors in the retina have not been studied. To gain insight into the mechanisms that regulate LIF-mediated neuroprotection, we investigated the contributions of TNF and p38 MAPK to Lif expression both in vitro and in vivo, and tested the relevance of the findings in the model of light-induced retinal degeneration (LIRD) in vivo. The results show that p38 MAPK activity is neuroprotective and required to upregulate expression of Lif in the injured retina.

Results

TNF upregulates Lif expression through p38 MAPK in cultured Muller cells in vitro

Previously, it has been shown that treatment with recombinant TNF induces upregulation of Lif expression in fibroblasts and other cells.20, 21 Since LIF is crucial for endogenous neuroprotection in the retina and is expressed by a subset of Muller cells upon photoreceptor injury, we tested whether Muller cells upregulate Lif in response to TNF administration in vitro. When cultured rat Muller cells (rMC-1)36 were treated with TNF, Lif and Tnf were simultaneously upregulated 10.7- and 21-fold, respectively (Figure 1). This transcriptional response was fast and reached its peak at 1 h before it gradually decreased towards basal levels even though TNF was still present in the culture medium. This suggests a transcriptional induction by TNF followed by suppression of expression. The transient upregulation of Lif and Tnf in Muller cells is consistent with results from previously studied models.20, 21 We also tested expression of genes that are known to be upregulated in activated Muller cells including Gfap, ciliary neurotrophic factor (Cntf) and Fgf2.1 However, none of these genes was activated by TNF demonstrating a specificity of TNF for the regulation of Lif expression in rMC-1 (Figure 1).

Figure 1
figure 1

TNF treatment transiently upregulates Lif expression in Muller cells in vitro. Real-time PCR analysis of gene expression in rMC-1 cells before (0 h) or at various timepoints during TNF treatment as indicated. TNF (10 ng/ml) treatment resulted in a strong upregulation of Lif expression as early as 30 min. Peak of expression was at 1 h. Tnf expression was upregulated similarly to Lif. In contrast, expression of Cntf, Gfap and Fgf2 was not affected by TNF. Shown are means±S.E.M. of N=3. ANOVA with Dunnett’s post-tests was used to compare control levels (at ‘0’) with expression levels of each gene at all timepoints after TNF treatment. Note that the color of the stars indicating significance match the color of the respective gene. *P<0.05

Full size image

Stimulation of cells with TNF has been previously shown to induce p38 MAPK signaling. Thus, we determined the role of p38 MAPK in the regulation of Lif expression. Since the expression of Lif in cultured Muller cells was robust, we first tested the effect of p38 MAPK activity on basal Lif expression in the absence of TNF by using two specific chemical inhibitors for p38 MAPK activity, SB239063 and SB202190.37, 38 Treatment with either SB compound downregulated Lif expression in a dose-dependent manner within 1 h of treatment and at a similar concentration range (Figures 2a and b). As expected, inhibitor treatment did not block phosphorylation of p38 MAPK (see also Figures 5b and c) but prevented its activity reducing activation of downstream targets like heat shock protein-27 (data not shown). The effect of p38 MAPK inhibition was specific for Lif as the expression of Gfap, Cntf and Fgf2 was not reduced (Figures 2a and b).

Figure 2
figure 2

Inhibition of p38 MAPK activity downregulates Lif expression in Muller cells in vitro. Real-time PCR analysis of gene expression in rMC-1 cells before (control) or at 1 h of treatment with various concentrations (as indicated) of p38 MAPK inhibitors SB239063 (a) or SB202190 (b). Expression levels of Lif inversely correlated with the concentration of p38 MAPK inhibitors. Expression levels of Cntf, Gfap and Fgf2 were not affected. Shown are means±S.E.M. of N=3. ANOVA with Dunnett’s post-tests was used to compare control levels with expression levels of each gene after inhibitor treatment. Note that the color of the stars indicating significance match the color of the respective gene. *P<0.05; **P<0.01; ***P<0.005

Full size image

To analyze whether TNF-mediated upregulation of Lif expression also involves p38 MAPK signaling, we co-treated Muller cells with TNF and SB239063. Consistent with our results above (Figure 1), TNF treatment induced Lif expression (Figure 3a). However, inhibition of p38 MAPK activity by SB239063 completely blocked Lif upregulation in the presence of TNF (Figure 3a), suggesting that p38 MAPK activity is crucial not only for basal Lif expression but also for TNF-induced Lif upregulation.

Figure 3
figure 3

Inhibition of p38 MAPK activity prevents TNF-induced upregulation of Lif expression in Muller cells in vitro. (a) RT PCR analysis of Lif expression in rMC-1 cells before (control, black bar) or at 1 h of treatment with TNF (10 ng/ml, blue bar), p38 MAPK inhibitor SB239063 (SB, 100 μM, pink bar) or a combination of both (green bar). Inhibition of p38 MAPK activity downregulated basal Lif expression and blocked TNF-induced Lif upregulation. Shown are means±S.E.M. of N=3. ANOVA with Dunnett’s post-tests was used to compare Lif levels after TNF injection with control levels or levels after SB or SB+TNF treatment. (b) Luciferase levels in rMC-1 cells transfected with a control vector or an NFκB reporter vector. Cells were treated for 6 h with TNF (10 ng/ml, blue bars), p38 MAPK inhibitor SB202190 (SB, 100 μM, pink bars), or a combination of both (green bars). Controls (black bars) were not treated. NFκB-mediated luciferase expression was upregulated by TNF treatment. The TNF-mediated upregulation was partially blocked by the inhibitor of p38 MAPK activity. ANOVA with Dunnett’s post-tests was used to compare luciferase levels after TNF injection with levels of other treatments. Statistics were calculated seperately for the ‘control vector’ group and the ‘NFκB vector’ group. Shown are means±S.E.M. of N=4

Full size image

Another signaling pathway known to be activated by TNF centers around NFκB. Since multiple potential binding sites for NFκB are located in the Lif promoter region (not shown) and NFκB was identified to bind the Lif promoter upon stimulation of cells with Toll-like receptor-2 (TLR2) agonists,39 we determined the activation of NFκB in Muller cells upon TNF treatment using a luciferase reporter vector that contains NFκB binding elements. Treatment with TNF strongly increased luciferase levels suggesting that NFκB was activated in Muller cells under these conditions (Figure 3b). Importantly, addition of the p38 MAPK inhibitor SB202190 completely inhibited TNF-induced Lif upregulation (Figure 3a), but reduced TNF-induced NFκB activity only by 49% leaving it still fivefold above control levels (Figure 3b). This indicates that, although p38 MAPK and NFκB pathways may interact, increased NFκB activity may not be sufficient to induce Lif expression in Muller cells in response to TNF treatment (Figure 3b) in vitro.

Regulation of Lif expression in the neuronal retina in vivo

Our results show that TNF induces Lif expression via the p38 MAPK pathway in Muller cells in vitro. To determine whether Lif expression is similarly regulated in the healthy or injured neuronal retina in vivo, we injected TNF, a water-soluble p38 MAPK inhibitor (SB220025 trihydrochloride),40 or a combination of both into the vitreous of wild-type mice that were or were not exposed to damaging levels of white light (Figures 4, 5, 6, 7).

Figure 4
figure 4

Intravitreal injection of TNF induces Lif gene expression in the neuronal retina in vivo. Real-time PCR analysis of gene expression in retinas of C67BL/6 mice before (control) or at indicated timepoints after intravitreal injection of TNF (blue lines) or vehicle (maroon lines). Lif and Tnf were upregulated as early as 1 h (Tnf) or 2 h (Lif) after injection, before levels of Stat3, Edn2, Fgf2 and Gfap started to increase at 6 h. Expression of Cntf, Lifr and Pax6 was not remarkably affected by the injections. Shown are means±S.E.M. of N=3−4 per timepoint and treatment. Two-way ANOVA with Bonferroni post-test was used to test for statistical significance between treatments at each timepoint. *P<0.05; ***P<0.005; ****P<0.001

Full size image
Figure 5
figure 5

Activation of p38 MAPK is essential for TNF-induced upregulation of Lif expression in the neuronal retina in vivo. (a) RT PCR analysis of gene expression in retinas of C67BL/6 mice before (no injection, black bars) or at 2 and 6 h after intravitreal injection of TNF (blue bars), p38 MAPK inhibitor SB220025 (SB, pink bar), TNF+p38 MAPK inhibitor SB220025 (TNF+SB, green bar) or vehicle (maroon bar). TNF-mediated upregulation of Lif expression was significantly attenuated by the p38 MAPK inhibitor. Shown are means±S.E.M. of N=3−4 per timepoint and treatment. ANOVA with Bonferroni post-tests was used to compare expression levels after TNF injection with levels of remaining groups at the 2-h timepoint. Student’s t-test was used to compare expression at 6 h. *P<0.05; **P<0.01; ***P<0.005. (b) Detection of phospho-p38 MAPK (upper panel) and p38 MAPK (lower panel) by western blot of protein extracts from retinas before (control) or at 2 h after injection of vehicle, TNF, p38 MAPK inhibitor SB220025 (SB), or TNF+p38 MAPK inhibitor SB220025 (TNF+SB) as indicated. Injection of TNF, SB and TNF+SB increased levels of phospho-p38 MAPK compared with vehicle and uninjected retinas. Shown are representative blots. (c) Quantification of signals detected by western blot in (b) using Bio1D software. Phospho-p38 MAPK levels were highest after injection of TNF, of p38 MAPK inhibitor SB220025 (SB) and of TNF+p38 MAPK inhibitor SB220025 (TNF+SB). Shown are means±S.E.M. of N=3–9 per treatment. ANOVA with Bonferroni post-tests was used to compare levels to control. *P<0.05

Full size image
Figure 6
figure 6

p38 MAPK is involved in light-induced Lif upregulation in vivo. Real-time PCR analysis of retinal gene expression before (dark controls) or at 6 h after exposure to 2 h of 13 000 lux of white light. Vehicle (maroon bars) or p38 MAPK inhibitor SB220025 (SB, pink bars) was injected into the vitreous immediately after the end of light exposure. Injections in dark-exposed control mice were done at a similar timepoint during the day to control for potential circadian alterations in gene expression. Light exposure significantly upregulated Lif expression in non-injected (black bars) and vehicle-injected eyes. Upregulation was prevented by SB injections. Note that Lif levels in light-exposed mice after SB injections were not different from dark control mice after vehicle injections. Expression of Clc was regulated similarly to Lif. Levels of Cntf, Vim and Gfap were not affected by light exposure or injections. Shown are means±S.E.M. of N=3–7 per treatment. ANOVA with Bonferroni post-tests was used to compare levels after vehicle injection to levels after SB injections and to levels of mice that received no injection. Statistics were calculated seperately for the light-exposed groups and the control groups. **P<0.01; ***P<0.005

Full size image
Figure 7
figure 7

Inhibition of p38 MAPK activity increases cell death in the retina after light exposure in vivo. Quantification of cell death in retinas of non-exposed SV129S6 mice (dark controls) or in retinas of mice at 40 h after exposure to 2 h of 13 000 lux of white light (light damage). Twenty hours before light exposure, mice received an intravitreal injection of TNF (blue bars), p38 MAPK inhibitor SB220025 (SB, pink bars), TNF+p38 MAPK inhibitor SB220025 (TNF+SB, green bars) or vehicle (maroon bars). No-injection controls (black bars) were treated equally but did not receive intravitreal injections. The highest level of retinal cell death after light exposure was observed in non injected mice. Vehicle injections induced a protection, which was further strengthened by injection of TNF. p38 MAPK inhibitor SB220025 increased cell death almost to the level of non-injected control mice. Shown are means±S.E.M. of N=3–7 per treatment. ANOVA with Bonferroni post-tests was used to compare levels of cell death of the various groups to vehicle control. Statistics were calculated seperately for the light damage and the dark control group. *P<0.05; ***P<0.005

Full size image

Similar to our observations in cultured Muller cells, TNF injections increased the expression of Lif and Tnf in the retina within 1−2 h before levels gradually decreased again (Figure 4). Since Fgf2, End2 and Stat3 are part of the LIF-controlled endogenous neuroprotective signaling system,1, 12 we also analyzed their expression pattern after TNF injections. Expression of all three genes was significantly upregulated and peaked at around 12 h after injection and thus with a slight delay compared with Lif (Figure 4). Although not directly tested, this suggests that LIF may also be important for Fgf2, End2 and Stat3 expression after TNF injections, as it is in the injured retina.1, 12 Similarly, expression of Gfap, which has also been shown to depend on LIF signaling,1 was upregulated with a similar delay (Figure 4). In contrast, expression of paired box protein-6 (Pax6), Lif receptor (Lifr) and Cntf, genes that may not depend on LIF signaling, was not comparably regulated. Although CNTF is strongly neuroprotective in the retina2, 3 and belongs to the same interleukin-6 (IL-6) family of cytokines as LIF, consequences of TNF injections were thus rather specific for the Lif gene.

To determine whether p38 MAPK is also involved in TNF-induced Lif expression in the retina in vivo, we analyzed gene expression at 2 and 6 h after intravitreal injection of TNF and/or the p38 MAPK inhibitor SB220025. As expected, injection of TNF caused increased Lif RNA levels at both timepoints (Figure 5a). However, when SB220025 was co-injected with TNF, upregulation of Lif expression was reduced by 78 and 73% at the 2 and 6 h timepoints, respectively. Injection of vehicle or SB220025 alone slightly upregulated Lif expression probably due to injection-inflicted retinal injury (Figure 5a). Western blots of retinal extracts showed an approximately 7.5-fold increase in phosphorylation levels of p38 MAPK at 2 h after TNF injection as compared with controls, suggesting that TNF increases p38 MAPK activity in the retina (Figures 5b and c). Injection of the p38 MAPK inhibitor (which does not block phosphorylation but the activity of p38 MAPK)40 either alone or in combination with TNF also resulted in increased p38 phosphorylation levels which may suggest an attempt of inhibitor-treated retinal cells to increase p38 MAPK activity by a positive feedback. Since vehicle injections upregulated phospho-p38 MAPK levels by 3.5-fold, intravitreal injections per se may generate a stress response and activate p38 MAPK to a certain degree (Figures 5b and c). In summary, these data imply that the p38 MAPK pathway is also important for TNF-mediated Lif regulation in the neuronal retina in vivo.

p38 MAPK is an important regulator for increased Lif expression after exposure to damaging light in vivo

We showed that TNF-induced Lif upregulation depends on p38 MAPK activity in the healthy wild-type retina. To test whether Lif regulation depends on p38 MAPK also in a disease model, we exposed dark-adapted mice to high-intensity white light that has been shown to cause photoreceptor degeneration and to upregulate Lif expression in Muller cells.1 Six hours after light exposure, Lif levels were increased 62-fold (Figure 6). Intravitreal injection of vehicle immediately after light exposure slightly further increased Lif levels. Importantly, however, injection of p38 MAPK inhibitor SB220025 reduced Lif levels by 57% compared with vehicle injections. Similarly to observations made before (Figure 5a), vehicle injection induced Lif expression also in non-exposed mice (dark controls). As in the light-exposed retinas, injection of SB220025 reduced this induction by about 63% (Figure 6).

SB-mediated inhibition of Lif expression in the light-exposed retina was not as strong as in TNF+SB-injected retinas (Figure 5a), most probably because the SB inhibitor was injected for experimental reasons 2 h after the start of light exposure and thus only after the Lif-inducing signaling cascade had been activated by the light stimulus. Cardiotrophin-like cytokine (Clc) is another member of the IL-6 family of cytokines that is upregulated in response to light damage.12 Similarly to Lif, induction of Clc was partially inhibited by SB treatment (Figure 6). Analysis of Cntf, Gfap and vimentin (Vim), which are expressed in Muller cells,41, 42, 43 did not show any significant changes as a result of SB treatment (Figure 6). These results imply that p38 MAPK is involved in controlling Lif upregulation after light-induced photoreceptor injury in vivo.

p38 MAPK activity is neuroprotective in the model of light-induced photoreceptor degeneration

If p38 MAPK is indeed important for regulation of neuroprotective Lif in the injured retina, inhibition of its activity should increase photoreceptor damage after light exposure. To directly test this hypothesis, we quantified cell death in the retina at 40 h after exposure of wild-type mice to high levels of white light. Although intravitreal injections had a protective effect in general (vehicle), inhibition of p38 MAPK by SB220025 significantly increased cell death as compared with vehicle injections (compare ‘SB’ to ‘vehicle’). In contrast, TNF injections reduced cell death almost to dark control levels (Figure 7). Importantly, cell death showed a distinct negative correlation with levels of Lif mRNA after no, SB, vehicle or TNF injections (Figures 5a and 7), corroborating earlier findings that showed that dosage and timing of Lif expression is an important factor for retinal physiology and neuroprotection.13, 44, 45 These results strongly argue that LIF-mediated endogenous neuroprotection after light damage depends on p38 MAPK signaling in the retina.

Discussion

Regulation of Lif expression in Muller cells

Our results demonstrate that p38 MAPK signaling is required for the regulation of Lif expression in Muller cells in vitro and in the retina in vivo. We also show that inhibition of p38 MAPK activity reduces Lif expression and increases cell death in a model of photoreceptor injury supporting a direct role for p38 MAPK in LIF-mediated neuroprotection in the retina. Furthermore, treatment with recombinant TNF was sufficient to activate p38 MAPK, to increase Lif expression in vitro and in vivo, and to protect against light damage in vivo. However, whether endogenous TNF is involved in regulating Lif expression in the injured retina in vivo needs still to be determined, even though Tnf expression is induced early after light exposure in the eyecup.46

p38 MAPK activity has been shown to be one of the most important stress response factors by regulating the expression of several genes through activation of transcription factors such as ATF2, CHOP, CREB and ELK1.22 Additionally, p38 MAPK affects gene expression at the post-transcriptional level by enhancing the stability of target mRNAs containing AU-rich elements in their 3′ UTRs.47 In our experimental systems, however, we focused on the effects of p38 MAPK signaling on the regulation of Lif expression at the transcriptional level in vitro and in vivo.

Activation (phosphorylation) of p38 MAPK by TNF was rapid, as was upregulation of Lif expression. The fast upregulation of Lif levels was blocked by an inhibitor of p38 MAPK activity, suggesting that the initial role of p38 MAPK involves transcriptional activation of Lif expression. Transcription factors involved in this regulatory pathway have not been defined and their identification will need further investigations. However, NFκB is a candidate factor that was upregulated by TNF treatment in a p38 MAPK-dependent manner (Figure 3), and NFκB binding to the Lif promoter has recently been shown in response to TLR2 agonists.39 Furthermore, a connection between p38 MAPK, NFκB and cytokine regulation has been suggested48 and TNF-mediated activation of p38 MAPK is well documented.22 Nevertheless, the contribution of NFκB to TNF-mediated Lif upregulation may not be major, or else may depend on p38 MAPK. In the presence of TNF and the p38 MAPK inhibitor SB202190, Müller cells retain 50% of NFκB activity but completely lack Lif upregulation (Figures 3a and b), at least in vitro.

Regulation of Lif expression may additionally include a second level of complexity. The fast increase in expression after TNF treatment in vitro (Figure 1) or after light exposure in vivo,11 was followed by a rapid decline of Lif mRNA levels towards basal levels. This argues for the presence of a regulatory feedback loop and that transcriptional activation of Lif expression might be followed by transcriptional inhibition or by a reduction of Lif mRNA stability, or both.

Neuroprotective role of p38 MAPK activity and Lif expression

Depending on the concept and experimental setup, activation of p38 MAPK has been shown to promote cell survival or apoptosis. Whereas inhibition of p38 MAPK activity has been reported to protect 661W photoreceptor cells against light damage in vitro,49 our results indicate that inhibition of p38 MAPK activity accelerates light damage in vivo. This discrepancy may be due to the lack of intercellular communication between photoreceptors, RPE and Muller cells in the in vitro cell culture system. Photoreceptor degeneration induced in vivo by high levels of white light depends on RPE65 in the RPE and on the regeneration kinetics of the bleached chromophore in the visual cycle.50, 51 Hence mechanisms of light-induced death in isolated 661W cells in vitro and photoreceptors in vivo differ and the two experimental systems cannot be directly compared. Of additional importance, photoreceptors injured by light or other stimuli signal to Muller cells, which induce expression of Lif to ignite a neuroprotective response in the retina leading to increased photoreceptor survival.1, 8 This intercellular communication is not possible in the cell culture system, and effects of p38 MAPK inhibition in 661W cells may thus not accurately reflect the in vivo situation where p38 MAPK activity seems important in Muller cells to upregulate expression of neuroprotective Lif.

Despite the neuroprotective role of p38 MAPK, inhibition of its activity in the presence of TNF was still protective against LIRD, in contrast to inhibitor injections alone (Figure 7). Protection correlated with increased levels of Lif in retinas of eyes injected with a combination of TNF and inhibitor, and with decreased Lif levels after injection of the p38 MAPK inhibitor alone (Figure 5a). This suggests that injection of the p38 MAPK inhibitor was not sufficient to completely block the strong Lif-inducing activity of TNF and raises the possibility that a mechanism for TNF-induced Lif upregulation in addition to p38 MAPK may be present in vivo. Alternatively, TNF may potentially have protective effects through other mechanisms such as NFκB activation and its downstream targets, or the regulation of heat shock proteins.52 It is interesting to note that p38 MAPK inhibitors only partially blocked TNF-induced NFκB activation whereas they completely abolished Lif induction, at least in vitro (Figure 5). This may argue that TNF could regulate the fate of photoreceptor cells via several mechanisms.

Due to the inhibition of pro-inflammatory gene expression, inhibitors of p38 MAPK are discussed as potential therapeutic agents in inflammatory diseases like psoriasis and rheumatoid arthriis.53 Moreover, p38 MAPK inhibitors were shown to block tumor growth and metastases formation.54 However, our results demonstrating that inhibition of p38 MAPK activity may be detrimental to injured photoreceptor cells ask for precautions when developing p38 MAPK inhibitors for therapeutic use. In cases of open or hidden retinal disease conditions, application of p38 MAPK inhibitors to patients may have adverse effects on vision due to its negative effect on Lif expression and LIF-dependent survival factors in the retina. Similar adverse effects may possibly be observed in other neurodegenerative diseases where LIF has a reported protective activity.55, 56, 57 Thus, as a safety measure, p38 MAPK inhibitors may be designed not to cross retina−blood or brain−blood barriers, which may prevent potential neurodegenerative effects.

TNF signaling in the retina

Although most reports attribute TNF a devastating role in the retina and retinal diseases through the modulation of an inflammatory response,28, 29 our results demonstrate that intravitreal injection of TNF before light exposure did not accelerate degeneration. Rather, TNF reduced the relative average cell death by almost fourfold (0.27±0.049 S.E.M., N=6 for vehicle versus 0.074±0.07 S.E.M. for TNF, N=5). Even though these values did not reach statistical significance when tested in context of the additional experimental paradigms using analysis of variance (ANOVA), they showed a tendency for TNF-mediated protection of photoreceptors against light-induced degeneration (Figure 7). At least part of this effect may be explained by increased expression of Lif which leads to the upregulation of neuroprotective factors like Fgf2, End2 and Stat3 (Figure 4).1, 12

Emerging evidence indicates that TNF may be a signaling molecule of general importance that has differential effects on disease outcome depending on the interaction with its receptors. In a retinal ischemia-reperfusion model, TNF-R2 signaling was neuroprotective whereas TNF-R1 increased neuronal death.58 Moreover, a recent report suggests that TNF expression in dying photoreceptors and Muller cells is important for Muller cell proliferation and photoreceptor regeneration in a zebrafish model of LIRD.19, 59 Although the mammalian retina diverges from zebrafish in several ways including its limited proliferation and regeneration capacity of Muller cells,60, 61 these results together with our findings may nevertheless indicate a possible role of TNF for the protection of function in the stressed or injured mammalian retina. Clearly, further studies are warranted to elucidate the role of TNF signaling during photoreceptor degeneration in detail.

Materials and Methods

Animals

All experimental protocols were accepted by the Veterinary Authorities of Zurich and experiments were conducted in accordance with the statement of ‘The Association for Research in Vision and Ophthalmology’ for the use of animals in research. All mice had access to food and water ad libitum and were housed in a light–dark cycle of 12 : 12 h with 60 lux at cage level. All experimental conditions and time points were tested with a minimum of N=3 mice. All experimental mice were on a C57BL/6 background except for the light damage experiments in which mice of the SV129/S6 strain were used (Figure 7). For intravitreal injections after light exposure Rlbp::GFP mice62 on a C57BL/6 background were used (Figure 6).

Cell culture assays

TNF and p38 MAPK inhibitor treatment

rMC-1 cells36 were cultured in Dulbecco’s modified Eagle’s medium (Life Technologies, Grand Island, NY, USA) supplemented with 10% fetal bovine serum (Life Technologies), 100 U/ml penicillin and 100 μg/ml streptomycin (Life Technologies), and grown in a humidified 5% CO2 incubator as described in Sarthy et al.36 Rat recombinant TNF (R&D Systems, Minneapolis, MN, USA) was dissolved in PBS containing 0.1% bovine serum albumin (BSA) (Sigma Aldrich, St Louis, MO, USA), the p38 MAPK inhibitors SB202190 and SB239063 (Sigma Aldrich) were dissolved in water or in DMSO (Sigma Aldrich), respectively. Compounds were added alone or in combinations directly to growth media at concentrations and times indicated in the Results section.

Luciferase assay

rMC-1 cells were transfected using FuGENE6 reagent (Promega, Madison, WI, USA). Renilla luciferase expressing vector, pRL-CMV (Promega), was used as an internal control to normalize the transfection efficiency. Transfection solution was prepared in 91 μl Optimem (Life Technologies) media using 3 μg/ml of total plasmid DNA and 9 μl of FuGENE6 reagent according to manufacturer’s instructions. The ratio of renilla to firefly plasmids was 1 : 9. Ten thousand cells in 100 μl growth media were seeded on a 96-well plate and each well was transfected with 3.3 μl of transfection solution after attachment. An EGFP reporter vector, EGFP-C1 (Clontech, Mountain View, CA, USA), was used to assess the transfection efficiency of rMC-1 cells, which was between 15 and 20%. Cells were transfected either with a control vector, pTAL-luc (Clontech), or with an NFκB reporter vector, pNFκB-luc (Clontech). TNF, SB202190 and TNF+SB202190 treatments were performed for 6 h starting at 24 h after transfection. TNF and SB202190 had final concentrations of 10 ng/ml and 100 μM, respectively. Each treatment was performed in triplicates in four independent experiments. Luciferase levels were measured using the Dual Luciferase kit (Promega).

Immunoblotting

Protein homogenates were prepared by sonication (Branson sonifier, 10 strokes of 0.3 s with 30% output) of isolated retinas in Tris-HCl (100 mM, pH 7.5 or pH 8.0). Protein concentrations were determined by Bradford using BSA as standard. Homogenates were mixed with an equal volume of 4X Laemmli sample buffer and a total of 40 μg/ml protein was loaded in each lane of 10% SDS-polyacrylamide gels. Samples were electrophoresed, blotted and probed as described previously in Bürgi et al.12 To detect phospho-p38 MAPK (cat no. 840771; R&D Systems) and p38 MAPK (cat no. 9212; Cell Signaling, Danvers, MA, USA) primary antibodies were used at a dilution of 1 : 500 and 1 : 1000, respectively. The secondary antibody, anti-rabbit IgG peroxidase-linked (cat no. NA934; GE Healthcare, Pittsburgh, PA, USA), was used at a dilution of 1 : 10 000. We have used WesternBright Sirius HRP substrate (Advansta, Menlo Park, CA, USA) for chemiluminescence reaction. Fusion FX7 Advance imaging system (Vilber Lourmat, Torcy, France) with a CCD camera was used for digital signal detection. Recordings were taken at the dynamic range of exposure without binning. Calculations for exposure levels were performed using BioD1 software (Vilber Lourmat) without background subtractions.

RT-PCR analysis

Retinas were collected through a slit in the cornea, and total RNA was prepared and analyzed by real-time PCR (RT-PCR) as described previously in Bürgi et al.12 Briefly, 10 ng cDNA was amplified in a LightCycler instrument (Roche Diagnostics, Basel, Switzerland) using appropriate primer pairs (Table 1) and SybrGreen Master mix (Roche). Actb was used as reference gene. For the analysis of gene expression in rMC-1 cells, total RNA was extracted using the Megamax RNA isolation kit (Life Technologies) according to manufacturer’s instructions. cDNA was prepared using the high-capacity cDNA reverse transcription kit (Life Technologies). Real-time PCR reactions were conducted using appropriate primer pairs (Table 1) and Actb as internal control. Additional internal controls, Gapdh and Rpl32, were used for each new treatment. Real-time PCR reactions were performed in a StepOne Real-Time PCR system with Fast SybrGreen master mix (Life Technologies). The comparative cycle threshold method was used to calculate relative transcript levels for both mouse and rat experiments.

Table 1 Real-time PCR primer sequences
Full size table

Intravitreal injections

Intravitreal injections were performed as previously described in Joly et al.1 Rat recombinant TNF (R&D Systems) was reconstituted in sterile PBS containing 0.1% BSA (Sigma Aldrich) and injected at a concentration of 10 μg/ml. SB220025 trihydrochloride (Sigma Aldrich) was dissolved in water and adjusted to 6 mM using sterile PBS containing 0.1% BSA for injections. Sterile PBS containing 0.1% BSA was used for vehicle control injections. For the injection of TNF+SB220025, stock solutions for SB220025 and TNF were mixed and diluted with PBS containing 0.1% BSA to reach final injection concentrations of 10 μg/ml/ml and 6 mM for TNF and SB220025, respectively. Injection volume was 1 μl.

Light damage and cell death assay

Light damage and quantification of cell death by an ELISA-based cell death assay (Roche Diagnostics, Basel, Switzerland) were performed essentially as previously described in Bürgi et al.12 and Samardzjia et al.10 with minor modifications. White light intensity for light damage was set to 13 000 lux, and 6–10-week-old animals with dilated pupils12 were exposed for 2 h. Animals were kept in darkness overnight before and after light exposure. Animals were sacrificed and retinas were isolated for cell death assay at 40 h after light exposure.

Data analysis

Statistical analysis was performed using ANOVA with Dunnett’s Multiple Comparison tests where multiple comparisons were made to a single value. ANOVA with Bonferroni post-test was performed for multiple comparisons against independent controls. Student’s t-tests were used for individual pairwise comparisons. P-values lower than 0.05 were considered to be significant. Error bars represent the standard error of the mean (S.E.M.). We also used the ROUT algorithm, with the coefficient Q value of 0.1, to detect possible outliers related to intravitreal injections. Graph Pad 6 software or Prism 5 (GraphPad Inc., San Diego, CA, USA) were used for all statistical analyses.