Alpha 2-Adrenergic Receptor Agonist Brimonidine Stimulates ERK1/2 and AKT Signaling via Transactivation of EGF Receptors in the Human MIO-M1 Müller Cell Line.

ABSTRACT Purpose: Alpha 2-adrenergic receptor (α2-ADR) agonists are used clinically for a range of indications including reducing elevated intraocular pressure in patients with open-angle glaucoma or ocular hypertension. Animal experiments show that α2-ADR agonists attenuate the injury-induced Müller cell dedifferentiation by a mechanism that involves activation and regulation of extracellular signal-regulated kinase (ERK) 1/2 leading to transactivation of epidermal growth factor receptors (EGFRs). The purpose of this study was to study and corroborate the activation of this system in human cells. Material and Methods: The human Müller cell line MIO-M1 was treated with the α2A-ADR agonist brimonidine in combination with inhibitors for Src-kinase, EGFR-kinase, matrix metalloproteinase (MMP) as well as small interfering RNAs (siRNAs) for the EGFR. The cells were analyzed using immunocytochemistry, quantitative PCR and western blot techniques. Results: Our results show that human MIO-M1 cells express α2A-ADRs and that stimulation of these receptors caused a robust increase of ERK1/2 and protein kinase B (PKB/AKT) (Thr-308) phosphorylation in MIO-M1 cells. P-ERK1/2 and P-AKT (Thr-308) signaling was mediated by Src-kinase and associated with phosphorylation of tyrosine residue of epidermal growth factor receptor (P-EGFR Y1173). In addition, the agonist caused activation of MMPs. These effects could be blocked by Src-kinase inhibitors (PP1, PP2), EGFR-kinase inhibitor (AG1478), EGFR-siRNA and a MMP inhibitor (GM6001). Conclusion: The results confirm that this human Müller cell line responds to ADR stimulation with phosphorylation of ERK and AKT, which suggests that it is possible to pharmacologically target ADR to modulate the early events in human Müller cell dedifferentiation in a similar fashion as been shown for chicken Müller cells. Abbreviations: CRALBP: cellular retinaldehyde binding protein; EGFR: epidermal growth factor receptor; ERK1/2: extracellular signal-regulated kinase 1/2; GS: glutamine synthetase; GPCR: G protein-coupled receptor; IR: immunoreactivity; MAPK: mitogen-activated protein kinase; MMP: matrix metalloproteinase; P-ERK1/2: phospho-ERK1/2; qRT-PCR: quantitative reverse transcriptase PCR


Introduction
Müller cells are the principal supporting cell in retina 1,2 and in response to retinal disease or injury, they also provide a source for regenerated neurons. 3,4 The adult mammalian retina has little or no capacity for regeneration while several cold-blooded vertebrates such as teleost fish functionally regenerate their retina following injury, with Müller cells as a source for new neurons. 5 The initial events after injury include Müller cell dedifferentiation to a retinal progenitor-like cell and proliferation. This process is regulated by several growth factors including fibroblast growth factors, Wnts and heparin-binding epidermal growth factor (HB-EGF) that act by stimulating a core of signaling cascades including mitogen-activated protein kinase (MAPK)/extracellular signalactivated kinases 1/2 (ERK1/2) and phosphatidyl inositide-3 kinase (PI3K) signaling. [6][7][8] Even though regeneration is absent in mammalian retina, the mammalian Müller cells respond to injury with dedifferentiation and proliferation 9 and in fact functional regeneration in the mammalian retina can be stimulated by transgenic expression of the proneural transcription factor Ascl1. 10,11 However, equipped with the capacity for dedifferentiation but not with the regenerative capacity, the injured mammalian retina will neither get support for the injured neurons from differentiated Müller cells nor it will get newly regenerated neurons. It has therefore been hypothesized that neurons in an injured retina will withstand injury better if the Müller cell dedifferentiation is blocked. 2 The stimulation of alpha 2-adrenergic receptors (α2-ADR) activates ERK1/2 MAPK signaling in Müller cells 12,13 and it was shown in chicken Müller cells that the ERK1/2 MAPK signaling is triggered by transactivation of epidermal growth factor receptors (EGFRs). 13 Interestingly, the transactivation triggers a negative feedback regulation that is able to attenuate both injury-induced ERK1/2 MAPK signaling and the dedifferentiation of the chicken Müller cells. 14 The initial activation of ERK1/2 MAPK signaling is important for the attenuation of the injury-induced events. The transactivation of EGFR occurs via the activation of cytosolic Src-kinase that causes shedding of the membrane-bound EGFR ligands such as heparin-binding epidermal growth factor (HB-EGF) by stimulation of matrix metalloproteinases (MMPs), which elicits an autocrine stimulation of the receptors. 13,15 In this study, we corroborated the results from chicken and tested if α2-ADRs transactivate EGFRs in a human Müller cell line. 'The MIO-M1 cells were established by spontaneous immortalization from adult human retina. 16,17 We analyzed the expression of α2-ADRs and stimulated the α2-ADR with the agonist brimonidine (BMD). The intracellular signaling pathways were then studied with focus on MAPK/ERK and PI3K/AKT-kinase signaling. Our results show that the human MIO-M1 cells express α2A-ADRs and stimulation by BMD triggers Src-kinase mediated ligand-dependent and ligandindependent transactivation of EGFRs with activation of ERK1/2 and AKT signaling. The results confirm that this human Müller cell line responds to ADR stimulation, in a similar way as chicken Müller cells, which when stimulated with BMD attenuate dedifferentiation in injured retina. 13,14 The results imply that the early events in human Müller cell dedifferentiation are similar in the chicken and human systems.

Statistical analysis
GraphPad Prism 6 (GraphPad Software Inc., La Jolla, CA, USA) software was used for statistical testing by one-way analysis of variance (ANOVA) followed by Tukey's post hoc test as indicated in the figure legends.

Expression of α2-ADRs in MIO-M1 cells
The expression of the three α2-ADR subtypes α2A, α2B, and α2C was studied by using immunocytochemistry and α2A-ADR immunoreactivity (IR) was diffuse in all cells over the entire cell as expected for a plasma membrane location (Figure 1a, b). Müller cell markers glutamine synthetase (GS) and cellular retinaldehyde-binding protein (CRALBP) showed granular cytoplasmic and perinuclear IR (Figure 1a, b). 17 Very low α2B-ADR IR was detected in the cells (Figure 1c, d). The α2C-ADR IR was mainly nuclear (Figure 1e). The qRT-PCR analysis was used to quantify the relative mRNA levels of the three α2-ADR subtypes. We found relatively high levels of α2A-ADR, very low levels of α2B-ADR and moderate levels of α2C-ADR mRNA ( Figure 1f).

Stimulation of α2-ADRs causes phosphorylation of ERK1/ 2 MAPK in MIO-M1 cells
We used BMD to stimulate α2-ADRs and studied the activation of ERK1/2 MAPK by P-ERK1/2 immunocytochemistry and western blot analysis. Cells were serum-starved for 16 h to attain the basal P-ERK1/2 levels, treated with BMD or vehicle, and analyzed at several times ( Figure 2a). A strong cytoplasmic and perinuclear P-ERK1/2 IR was observed in the MIO-M1 cells at 60 min after BMD treatment and background levels at 180 min (Figure 2b-g). The P-ERK1/2 IR co-labeled with GS and CRALBP IR (Figure 2d-g). Western blot analysis confirmed the immunocytochemistry results ( Figure 2h). P-ERK1/2 levels were normalized to total ERK1/2 levels and densitometric analysis showed that P-ERK1/2 levels increased within 10 min after BMD treatment, with peak levels at 60 min and gradually decreased to lower levels by 120 min (Figure 2i). These results showed that stimulation of α2-ADRs triggered ERK1/2 phosphorylation in the MIO-M1 cells.

α2-ADR stimulation accelerates phosphorylation of AKT kinase in MIO-M1 cells
We stimulated the MIO-M1 cells with BMD and studied the PI3K pathway by P-AKT (Thr-308)-western blot analysis. Phosphorylation of threonine residue 308 of AKT is indicative of an active PI3K-pathway after growth factor stimulation. 25 Serum-starved human cells were treated with BMD and analyzed ( Figure 3a). Western blot analysis was performed to quantify the P-AKT (Thr-308) levels (Figure 3b), which were normalized to non-phosphorylated (pan) AKT levels ( Figure 3b). Densitometric analysis showed that P-AKT (Thr-308) levels increased within 10 min after BMD treatment, with peak levels at 60 min and lower levels by 120 min (Figure 3c). This result showed that the AKT is phosphorylated after α2-ADR stimulation in MIO-M1 cells.

Blocking of α2-ADRs inhibits ERK1/2 and AKT phosphorylation in MIO-M1 cells
To corroborate that BMD-induced ERK1/2 and AKT phosphorylation in the MIO-M1 cells was due to stimulation of α2-ADR signaling, we studied the effects of the selective α2-ADR blocker yohimbine. 26 Serum-starved human cells were pretreated with yohimbine and then treated with BMD ( Figure 4a). Western blot analysis was used to quantify the P-ERK1/2 and P-AKT (Thr-308) levels and densitometric analysis showed that yohimbine treatment reduced the P-ERK1/2 (Figure 4b, c) and P-AKT (Thr-308) (Figure 4d, e) levels to control levels. Cells treated only with yohimbine did not alter the basal P-ERK1/2 (Figure 4b, c) or P-AKT (Thr-308) (Figure 4d, e) levels.

BMD-induced transactivation of EGFRs in MIO-M1 cells
α2-ADRs are G-protein-coupled receptors (GPCRs) 27 that have been showed to be involved in transactivation of EGFR in different cell types. 28 To assess whether α2-ADRs can transactivate EGFRs in MIO-M1 cells, we studied the effect of the EGFR kinase inhibitor AG1478 29 on BMD-induced ERK1/2 and AKT phosphorylation with EGF-induced ERK1/2 phosphorylation in the MIO-M1 cells as a positive control. Serum-starved MIO-M1 cells were treated with AG1478 for 30 min and then with BMD or EGF (Figure 5a). Western blot analysis was used to assay the effects of AG1478 on the cells (Figure 5b-g). We found that AG1478 pretreatment in MIO-M1 cells reduced the P-ERK1/2 (Figure 5b, c) levels to basal levels and inhibited the EGF-induced ERK1/2 phosphorylation (Figure 5d, e). AG1478 pretreatment also reduced P-AKT (Thr-308) levels to basal levels (Figure 5f, g), indicating that phosphorylation of both ERK and AKT was downstream of the EGFR. We did not observe any alternation of basal P-ERK1/2 and P-AKT (Thr-308) levels in cells treated only with AG1478 (Figure 5b, f).
To corroborate the involvement of the EGFR, we performed siRNA-mediated knock down of the EGFR expression in the MIO-M1 cells. Cells were transfected with EGFR-siRNA or control non-target siRNA for 48 h, followed by treatment with BMD for 60 min (Figure 6h). Western blot analysis was used to quantify the effects. We found that EGFR-siRNA specifically knocked down the EGFR levels ( Figure 6i) and consistent with reduced EGFR levels, EGFR-siRNA-transfected cells showed significantly reduced P-ERK1/2 (Figure 5i, J) and P-AKT (Thr-308) levels (Figure 5i, k) compared to non-transfected or non-target siRNA.
Studies have shown that phosphorylation of residue tyrosine 1173 in the EGFR is important for EGFR signaling. 30 Serum-starved human MIO-M1 cells had distinct P-EGFR (Y1173) IR at 60 min after BMD treatment with levels comparable with cells stimulated by EGF as a positive control (Figure 6a-j). Co-labeling with GS supported the Müller cell-phenotype of the cells (Figure 6e-j). There was no increase of P-EGFR (Y1173) IR in serumstarved control cells (Figure 6b, d). Western blot analysis corroborated the increase of P-EGFR (Y-1173) levels after BMD treatment (Figure 6k). P-EGFR (Y1173) levels were normalized to EGFR levels and we found a similar pattern of P-EGFR (Y-1173) as P-ERK1/2, with peak levels at 60 min that gradually decreased to background levels by 180 min (Figure 6l). The results are consistent with BMDinduced ERK1/2-, AKT phosphorylation and EGFR transactivation in MIO-M1 cells.

BMD-induced ERK1/2 and AKT phosphorylation requires SRC-kinase activity
Cytosolic Src is a non-receptor tyrosine kinase and mediates GPCR-triggered ERK1/2 activation by transactivation of EGFRs. 31,32 To investigate whether BMD-induced ERK1/2 and AKT phosphorylation is Src-kinase-dependent, we studied the effects of the Src-kinase inhibitors PP1 and PP2 on BMD-induced ERK1/2 and AKT phosphorylation. As a control, we studied the effect of PP1 or PP2 on EGF-induced ERK1/2 phosphorylation. Serum-starved cells were pretreated 20 min with PP1 or PP2 before treatment with BMD or EGF (Figure 7a). Western blot analysis showed that blocking of Src-kinase by PP1 or PP2 reduced the P-ERK1/2 and P-AKT levels to control levels in the MIO-M1 cells (Figure 7b, c, f, g). We did not observe any effect of PP1 or PP2 on EGF-induced ERK1/2 phosphorylation (Figure 7d, e). Cells treated only with PP1 or PP2 expressed basal P-ERK1/2 and AKT levels (Figure 7b, c, f, g). These results show that cytosolic Srckinase is involved in BMD-induced ERK1/2 and AKT phosphorylation in MIO-M1 cells. To assess whether Src-kinase inhibitors altered the transactivation of EGFR after BMD treatment, we studied the effects of PP1 or PP2 on BMD-induced P-EGFR (Y-1173) and western blot analysis showed that PP1 as well as PP2 treatments reduced BMD-induced P-EGFR (Y-1173) levels in the cells (Figure 7h, i). These results demonstrate that BMD-induced EGFR-transactivation is Src-kinase dependent in MIO-M1 cells.
HB-EGF is the one of several endogenous ligands that may activate the EGFR. 36 We confirmed the expression of HB-EGF in the MIO-M1 cells and analyzed whether BMD induced the expression of HB-EGF or EGFR. HB-EGF and EGFR mRNA expression remained at the same level at different time points after BMD treatment (Figure 8h, i). The results corroborated expression of HB-EGF and EGFR mRNAs in MIO-M1 cells and showed that their expression is not transcriptionally regulated by the BMD treatment.

Discussion
In this work, we have studied the expression of α2-ADRs and signal transduction responses after stimulation of the human MIO-M1 Müller cell line, with the α2-ADR agonist BMD. Our results confirm that the cells express α2A-ADR and that BMD caused robust MAPK-and PI3K signaling with phosphorylation of ERK1/2 and AKT. The signaling was mediated by Src-kinase with involvement of both ligand-dependent and ligand-independent transactivation of EGFRs.
The data showed that MIO-M1 cells mainly express α2A-ADR localized to cell somata with lower levels of α2B and α2C (Figure 1). α2A-ADR expression is found in ocular tissues in different species, 37,38 and the expression in MIO-M1 cells is similar to chicken primary Müller cells, studied both in vitro and in vivo. 13 The nuclear localization of the α2C-ADR IR was unexpected because GPCRs are normally in the plasma membrane. However, nuclear localization has been demonstrated for several GPCRs including ADRs. 39,40 The qRT-PCR results for the α2-ADRs were consistent with that of the cytochemical analyses.

α2-ADR
agonists (e.g., Clonidine, Tizanidine, Dexmedetomodine) are used for various clinical indications including pain, sedation and to reduce anesthetic requirements, withdrawal symptoms, 41 and BMD specifically has been used for the treatment of open-angle glaucoma and ocular hypertension. 42 A large number of experimental preclinical studies suggest that BMD can protect retinal ganglion Bar graphs with densitometry of (C, E) P-ERK1/2 levels normalized to total ERK1/2 levels and (G) P-AKT (Thr-308) levels normalized to AKT (Pan) levels and (I) P-EGFR (Y1173) levels normalized to total EGFR levels. Bar graphs are mean ± SEM, n = 3 (***P < 0.0001) analyzed by one-way ANOVA and Tukey's post hoc test. Significance is only indicated for the comparisons: BMD 60 min-BMD + PP1 60 min, and BMD 60 min-BMD + PP2 60 min. cells against injury and the function of Müller cells has been implied in the protection. However, the cellular mechanisms are not fully resolved. Our results show that stimulation of the MIO-M1 human Müller cell line with BMD resulted in a robust ERK1/2 and AKT phosphorylation with transactivation of EGFRs. The specificity of the stimulation was confirmed by the ability of the α2-ADR-specific blocker yohimbine to inhibit the activation. The engagement of EGFR in the response was shown both by the ability of the EGFR-kinase inhibitor AG1478 and siRNA knock down of EGFR expression, to inhibit the phosphorylation of ERK1/2 and AKT after BMD stimulation. The Src-kinase blockers, PP1 and PP2, abolished the BMD-induced ERK1/2 phosphorylation but were not able to block EGF-induced ERK1/2 phosphorylation, showing that the BMD-induced transactivation require Src activity, and confirmed the well-known fact that EGFR signaling in itself does not. Studies have shown that Src-kinase can directly associate with the catalytic domain of EGFRs and activate EGFRs in a ligand-independent pathway in epithelial and fibroblast cells. 43,44 The BMD stimulation of the MIO-M1 cells resulted in phosphorylation of 1173 tyrosine residue of EGFR, which is one of the major autophosphorylation site that allows interaction of adapter proteins GRB2, SOS and SHC to intracellular domain of EGFR and that facilitates Ras-Raf activated MAPK/ERK and PI3K-activated AKT signaling. 45,46 Membrane-bound extracellular MMPs have key roles in α2-ADR-induced EGFR transactivation. 28,34 Activation of MMP leads to proteolytic release of membrane-bound HB-EGF or TGFα, which stimulate EGFRs in an auto-or paracrine mode of action. 15 Cytosolic Src-kinase activates MMPs 47 and the abrogated α2-ADR agonist-induced phosphorylation of ERK1/2 and AKT using the MMP blocker GM6001, indicated a ligand-dependent mechanism for EGFR transactivation in these cells. However, the phosphorylation of ERK1/2 and AKT was not completely blocked by MMP inhibitor treatment, indicating that both liganddependent and ligand-independent EGFR activation occurs in the MIO-M1 cells. HB-EGF is one of the endogenous EGFR ligands and we show that MIO-M1 cells express it. The expression was not induced or increased by the BMD treatment, indicating that the increased EGFR signaling was not due to increased HB-EGF synthesis. Instead, the increased EGFR signaling was likely due to increased shedding of plasma membrane-bound HB-EGF by MMPs as indicated by the blocking effect of the MMP-inhibitor (A) Experimental outline. Serum-starved MIO-M1 cells pretreated with 50 μM GM6001 or control (vehicle) for 30 min followed by treatment with 300 μM BMD or vehicle for 60 min or with 100 ng/mL EGF or vehicle for 10 min. Western blot analyses of P-ERK1/2 levels treated with (B) BMD + GM6001 and (D) EGF + GM6001 and P-AKT (Thr-308) levels in MIO-M1 cells. Bar graphs with densitometry of (C, E) P-ERK1/2 levels normalized to total ERK1/2 levels and (G) P-AKT (Thr-308) levels normalized to AKT (Pan) levels. (H, I) qRT-PCR analysis of HB-EGF and EGFR mRNA levels in normal and BMD-treated MIO-M1 cells. Serum-starved cells treated with 300 μM BMD or vehicle and analyzed at 0, 10, 30, 60 and 120 min time points. Bar graph shows normalized mRNA levels relative to β-actin for (H) HB-EGF and (I) EGFR mRNA. Bar graphs are mean ± SEM, n = 3 (*P < 0.01, **P < 0.001) analyzed by one-way ANOVA and Tukey's post hoc test. Significance is only indicated for the comparisons: Control 60 min-BMD + GM6001 60 min, and BMD 60 min-BMD + GM6001 60 min.
GM6001. Further studies are required to demonstrate that HB-EGF is the only endogenous ligand that mediates the transactivation of EGFRs in human MIO-M1 cells.
In conclusion, our data show that the human MIO-M1 Müller cells express α2A-ADR and their stimulation triggers Src-kinase mediated ERK1/2 and AKT signaling, which involve both ligand-dependent and ligand-independent EGFR transactivation. The results from human cells are in line with our previous data based on animal cells, showing that stimulation of ADRs on retinal Müller cells may be used to modulate the immediate events in Müller cell dedifferentiation after an acute injury. The result also adds to an evolutionary perspective. Müller cells are a source for retinal regeneration after injury in several vertebrates, e.g. fish, frog and chicken, which is not observed to the same extent in mammals. However, these results indicate that the response by human Müller cells after injury may be targeted by α2A-ADR agonists as was shown for chicken Müller cells. The results provide knowledge that contributes to the development of a potential pharmacological strategy to regulate the responses of human Müller cell in a situation with injury or disease.