Thrombin signalling through PAR (protease-activated receptor)-1 is involved in cellular processes, such as proliferation, differentiation and cell survival. Following traumatic injury to the eye, thrombin signalling may participate in disorders, such as PVR (proliferative vitreoretinopathy), a human eye disease characterized by the uncontrolled proliferation, transdifferentiation and migration of otherwise quiescent RPE (retinal pigment epithelium) cells. PARs activate the Ras/Raf/MEK/ERK MAPK pathway (where ERK is extracellular-signal-regulated kinase, MAPK is mitogen-activated protein kinase and MEK is MAPK/ERK kinase) through the activation of Gα and Gβγ heterotrimeric G-proteins, and the downstream stimulation of the PLC (phospholipase C)-β/PKC (protein kinase C) and PI3K (phosphoinositide 3-kinase) signalling axis. In the present study, we examined the molecular signalling involved in thrombin-induced RPE cell proliferation, using rat RPE cells in culture as a model system for PVR pathogenesis.
Our results showed that thrombin activation of PAR-1 induces RPE cell proliferation through Ras-independent activation of the Raf/MEK/ERK1/2 MAPK signalling cascade. Pharmacological analysis revealed that the activation of ‘conventional’ PKC isoforms is essential for proliferation, although thrombin-induced phosphorylation of ERK1/2 requires the activation of atypical PKCζ by PI3K. Consistently, thrombin-induced ERK1/2 activation and RPE cell proliferation were prevented completely by PI3K or PKCζ inhibition. These results suggest that thrombin induces RPE cell proliferation by joint activation of PLC-dependent and atypical PKC isoforms and the Ras-independent downstream stimulation of the Raf/MEK/ERK1/2 MAPK cascade. The present study is the first report demonstrating directly thrombin-induced ERK phosphorylation in the RPE, and the involvement of atypical PKCζ in this process.
The serine protease thrombin has been shown to affect a wide variety of processes within diverse cell types, including RPE (retinal pigment epithelium) cells [1,2]. Intracellular thrombin signalling is triggered by activation of the PARs (protease-activated receptors), a family of GPCRs (G-protein-coupled receptors) activated by proteolytic cleavage of the extracellular domain, which unmasks a new sequence that functions as an intramolecular ligand . Four members of this family have been identified: PAR-1, PAR-3 and PAR-4, which are activated by thrombin, and the closely related PAR-2, which is sensitive to cleavage by trypsin . PAR-1 is the prototype of this receptor family, and its cleavage at the Arg41–Ser42 bond by thrombin exposes a new N-terminus (S42FLLRN47) that acts as a tethered ligand . Synthetic ligands corresponding to the cleaved N-terminus can displace the tethered ligand from the binding site and fully activate PAR-1 in an intermolecular mode .
PAR-1 expression at the cell surface is highly regulated by ligand cleavage. In resting cells, an intracellular pool of functional receptors maintains surface expression through a largely undefined recycling mechanism. Upon thrombin stimulation, the cleaved receptors are targeted to the lysosome for degradation, and new functional receptors from the intracellular pool restore PAR membrane expression  and sensitivity to thrombin. Although in endothelial cells and fibroblasts receptor replenishment does not require new protein synthesis [7,8], in megakaryoblasts this process depends on de novo protein synthesis .
PARs have been linked to the activation of an extraordinarily diverse array of physiologic responses by interacting with several GPCR Gα subunits, in particular, Gq11α, G12/13α and Gαi, which accounts for the pleiotropic action of its ligands [3,10]. Activation of PARs by thrombin in platelets and CCL-39 cells showed receptor coupling to PI (phosphoinositide) hydrolysis and to the inhibition of adenylate cyclase via at least two distinct effectors, most likely Gq-like and Gi-like G-proteins . Consistent with previous studies, thrombin-induced PI hydrolysis and cell growth via PTX (pertussis toxin)-insensitive and -sensitive G-proteins has been observed in human airway smooth muscle . PAR-1 coupling to G12/13 activation has also been demonstrated , although the downstream modulation of G12/13 effectors, such as Rho, GEFs (guanine-nucleotide-exchange factors) and PLC (phospholipase C)-β, by PAR-1 stimulation remains to be established [14,15].
PAR-1 signalling activates the MAPK (mitogen-activated protein kinase) cascade in several cell types [16,17]. MAPKs play a pivotal role in a variety of cellular functions [18,19], including the induction of cell proliferation, chemokine expression and epithelial–mesenchymal transformation [20–22].
Three major mammalian MAPK subfamilies have been described: ERK (extracellular-signal-regulated kinase) 1 and ERK2, JNKs (c-Jun N-terminal kinases) and the p38 kinases. Among these, p42/p44 MAPK (ERK1/2) activation is typically associated with cell survival, proliferation and differentiation, given their activation by mitogens and cell-survival factors [18,23]. The activation of ERK1/2 triggers their translocation from the cytoplasm to the nucleus, which appears to be an important regulatory step for mitogen-induced gene expression and cell cycle re-entry to promote cell proliferation .
The best-characterized MAPK linear pathway includes the small GTPase Ras and the kinases Raf, MEK (MAPK/ERK kinase) and ERK. Receptor and non-receptor tyrosine kinases, as well as GPCRs, lacking intrinsic tyrosine kinase activity, have been shown to activate Ras, the first step in this signalling cascade. Upon Ras-mediated activation of Raf, which, in turn, activates the dual-function kinase MEK, ERK1/2 phosphorylation on Tyr204 and Thr202 by MEK leads to the activation of nuclear and cytoplasmic ERK substrates . This signalling cascade has been shown to play an important role in the control of cell proliferation by ERK-induced activation of the transcription factors NF-κB (nuclear factor κB), c-Myc, CREB (cAMP-response-element-binding protein) and AP-1 (activator protein-1) , as well as by promoting the expression of regulatory proteins involved in the cell cycle [27,28]. Thus MAPKs are important integrators of GPCR- and tyrosine-kinase-receptor-mediated signals for cell proliferation .
In the RPE, thrombin stimulates PLC activity  and IL-8 (interleukin-8) gene expression , and also induces cell proliferation in vitro , suggesting a role for thrombin in the pathogenesis of proliferative disorders, such as PVR (proliferative vitreoretinopathy), a human retinal disease that involves the proliferation, de-differentiation and migration of RPE cells into the vitreous . The direct activation of the MAPK signalling pathway by thrombin in the RPE has not been demonstrated, although the induction of VEGF (vascular endothelial growth factor) expression in these cells by thrombin has been shown to depend on ERK1/2 activation , indirectly suggesting a role for MAPK in PAR-1 downstream signalling. In vitro experiments have shown that serum stimulates RPE cell proliferation by activating the Ras/Raf/MEK/ERK signalling pathway . Furthermore, chick , rat  and human  RPE cells in culture are also induced to proliferate by glutamate, through the activation of the MEK/ERK1/2 MAPK pathway.
GPCRs, such as PARs, activate the MAPK pathway through multiple signalling pathways . PKC (protein kinase C) and Ca2+/calmodulin-dependent pathways have been shown to regulate the MAPK cascade by the activation of the small GTPase Ras, the first step in this signalling mechanism . Additionally, it has been suggested that the sustained activation of ERK required for the induction of gene expression by MAPK pathway activity is MEK-independent [38,39] and might depend on PKC activation .
In the present study, we analysed the kinetics of thrombin-induced activation of the Raf/MEK/ERK1/2 MAPK cascade in rat RPE cells in culture, its contribution to RPE cell proliferation and its transactivation by PKC isoenzymes. Our results show, for the first time, the requirement for the joint activation of two distinct intracellular signalling pathways for thrombin stimulation of RPE cell proliferation.
MATERIALS AND METHODS
All reagents used were of cell-culture grade. The PAR-1 peptide (Ser-Phe-Leu-Leu-Arg-Asn-Pro-Asn-Asp-Lys-Tyr-Glu-Pro-Phe), thrombin, hirudine, Ro-32-0432 and the PKCζ pseudosubstrate (myristic acid–Ser-Ile-Tyr-Arg-Arg-Gly-Ala-Arg-Arg-Trp-Arg-Lys-Leu) were obtained from Calbiochem. Staurosporine, wortmannin and PD98058 were from Sigma, and AG-1478, U-73122 and U0126 were purchased from Tocris Bioscience. PAR-3 (Ser-Phe-Asn-Gly-Gly-Pro) and PAR-4 (Gly-Tyr-Pro-Gly-Lys-Phe) peptide agonists were obtained from Bachem. Serum-free Opti-MEM® (Invitrogen) was used as the standard medium for all assays.
Long-Evans rat RPE cell culture
RPE cells were isolated as described previously . Briefly, 8–10-day old Long-Evans rats were anaesthetized by inhaled chloroform and killed following the animal care and use guidelines established by our institution. The eyes were enucleated, rinsed in Dulbecco's modified Eagle's medium (Gibco BRL) containing 100 i.u./ml penicillin and 100 μg/ml streptomycin, and incubated for 30 min at 37°C in the presence of 2% (v/v) dispase. After removal of the sclera and the choroid, the RPE was detached from the neural retina in calcium- and magnesium-free HBSS (Hanks balanced salt solution), and incubated in the presence of 0.1% trypsin for 5 min at 37°C. Trypsin digestion was stopped by a 1:1 dilution with Opti-MEM®. The dissociated cells were suspended in Opti-MEM® containing 4% (v/v) FBS (fetal bovine serum), and seeded at a density of 50000 cells/cm2 in 96-well or 6-well culture plates for cell proliferation assays and Western blot analyses respectively. The purity of the culture was established by expression of the RPE marker RPE-65, and cell viability (>90%) was assessed by the Trypan Blue exclusion method.
RPE cell proliferation assay
RPE cells cultured for 24 h at 37°C in 4% (v/v) FBS-supplemented Opti-MEM® were serum-deprived for 24 h prior to stimulation with thrombin (0–4 units/ml) in serum-free Opti-MEM® for a further 24 h. Cell proliferation was quantified in nonconfluent cultures using the colorimetric MTS [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium] reduction method (Cell Titer 96® Aqueous One Solution Reagent; Promega) according to the manufacturer's instructions. Cultures maintained in serum-free Opti-MEM® were used as negative controls and the specificity of thrombin effects was assessed, in all cases, using the thrombin inhibitor hirudine (4 units/ml). When tested, enzyme inhibitors [AG-1478, U0126, U-73122, staurosporine, Ro-32-0432 (various concentrations, see the Figure legends for details), 20 μM manumycine (Calbiochem) and 50 nM Raf-1 inhibitor (Calbiochem)] were added 1 h prior to thrombin stimulation. The absorbance measured at 490 and 630 nm was corrected for background, and the absorbance from unstimulated cultures was arbitrarily set as 100% (basal) proliferation. When the effect of inhibitors was tested, proliferation in thrombin-stimulated cultures was set as 100% (control).
Western blot analysis
RPE cells from confluent 6-well plates were serum-deprived for 24 h and washed three times with RPE KRB (Krebs–Ringer bicarbonate) buffer (118 mM NaCl, 5 mM KCl, 1.2 mM MgSO4, 2.5 mM CaCl2, 1 mM NaHPO4, 5.6 mM glucose and 25 mM NaHCO3). Cultures were then incubated at 37°C in thrombin-supplemented (2 units/ml) serum-free Opti-MEM® for 0–72 h to determine the time course of ERK1/2 phosphorylation. When tested, inhibitors were added to the culture 1 h prior to thrombin stimulation (2 units/ml). PAR-1, PAR-3 and PAR-4 agonists were used at a final concentration of 25 μM in serum-free Opti-MEM®. At the indicated time points, cells were washed twice with KRB buffer, and disrupted in lysis buffer containing 50 mM Tris/HCl (pH 7.4), 150 mM NaCl, 10 mM EDTA, 0.1% SDS, 1% Triton X-100, 1% CHAPS, 0.5% NP-40 (Nonidet P40), 0.1% BSA, 40 mM β-glycerophosphate, 10 mM sodium pyrophosphate and a protease inhibitor cocktail (10%; Sigma). Proteins in total-cell lysates (25 μg) were resolved by SDS/PAGE (12% gels) and transferred on to PVDF membranes.
After blocking for 30 min at room temperature (25°C) with 5% (w/v) non-fat dried skimmed milk powder in 20 mM Tris/HCl (pH 7.5) containing 500 mM NaCl and 0.1% Tween 20, the PVDF membranes were probed at 4°C with the following primary antibodies: mouse anti-[MAPK pERK1/2 (pThr202/Tyr204)] antibody (where pERK is phospho-ERK and pThr202/Tyr204 are phospho-Thr202 and phospho-Tyr204) (1:1000 dilution; Calbiochem), rabbit anti-ERK1/2 antibody (1:10000 dilution; Calbiochem) and mouse anti-β-actin antibody (1:5000 dilution; Chemicon). HRP (horseradish peroxidase)-conjugated secondary antibodies (Invitrogen), raised in the corresponding species, were used at the same dilution as the corresponding primary antibodies and developed using the Immobilon Western AP Chemiluminescent Substrate (Millipore). Kodak® film images were digitized using an Alpha Digi-Doc system (Alpha Innotech), and densitometric analysis was performed using the Quantity One Software V4.6 from Bio-Rad. Basal ERK1/2 activation from unstimulated cells was arbitrarily set as 100% for the time course assays. When inhibitors were tested, ERK1/2 activation in thrombin-stimulated cultures was set as 100% (control).
Measurement of [
3H]inositol phosphate accumulation
PLC-β activation by thrombin was quantified by [3H]inositol phosphate production as described previously . RPE cell cultures were incubated at 37°C in the presence of 2 μCi of myo-[2-3H(N)]inositol (PerkinElmer) per well for 16–24 h. Cells were then washed three times with 1 ml of pre-warmed (37°C) KRB buffer containing 10 mM LiCl. RPE cells were stimulated with 2 units/ml thrombin for 10 min, and total-cell lysates were obtained by adding 1 ml of chloroform/methanol [1:2 (v/v)] to each well. The aqueous phase was extracted and 3H-labelled inositol phosphates were eluted individually, or in batches, using a Dowex AG1 (X8, 100–200 mesh) column. Radioactivity in the samples was determined usng a liquid scintillation counter (LS 6000 SC; Beckman Coulter). Total PI formation [inositol monophosphate+IP2 (inositol 1,4-bisphosphate)+IP3 (inositol 1,4,5-trisphosphate)] from non-stimulated cells was used as control wells for basal activation of PLC-β. The inhibition of inositol phosphate formation by hirudine (4 units/ml) was used as a control for specificity.
Raw data for analysis were obtained from pooled RPE cells from 10–15 Long-Evans rats in three independent experiments. An unpaired Student's t test was applied for statistical analysis, using the GraphPad Prism V4.0 program *P<0.05, **P<0.01 and ***P<0.001.
Thrombin induces RPE cell proliferation by activating the MAPK cascade
Thrombin has been shown to induce the proliferation of human RPE cells in culture . As a first approach towards understanding the role of thrombin in PVR pathogenesis, we studied the effect of thrombin on rat RPE cell proliferation. RPE cells were serum-deprived for 24 h prior to stimulation with thrombin at 1, 2 and 4 units/ml in serum-free Opti-MEM®, and cell proliferation was quantified using the colorimetric MTS assay after 24 h incubation in the presence of thrombin. As shown in Figure 1(a), the proliferative response of RPE cells was dose-dependent and was totally inhibited by the thrombin-specific inhibitor hirudine (4 units/ml).
Thrombin induces RPE cell proliferation through the MAPK pathway
In order to define the molecular mechanisms underlying thrombin stimulation of RPE proliferation, we assessed the involvement of the MAPK signalling pathway by testing the effect of MEK inhibition on thrombin-induced proliferation. As shown in Figure 1(b), U0126, known to interfere with ERK phosphorylation by MEK, induced a dose-dependent inhibition of thrombin-induced proliferation. Correspondingly, the inhibition of MEK activation by 30 mM PD98059 produced a similar effect (results not shown). U0126 was used in subsequent experiments, since it has been shown to inhibit ERK phosphorylation by active or inactive MEK, which eliminates any effect caused by resident-activated MEK.
Thrombin induces biphasic activation of ERK1/2 through PAR-1
The ERK1/2 MAPK signalling pathway has been linked to cell proliferation in a variety of cells, as well as to thrombin-mediated effects on platelets, in which distinct effects have been related to variations in the duration of ERK activation [16,42]. In order to determine the kinetics of ERK1/2 activation by thrombin, serum-deprived RPE cells were stimulated with 2 units/ml thrombin, and ERK phosphorylation was measured at early (5, 10, 15 and 30 min), intermediate (1, 2, 4, 8 and 12 h) and late time points (1, 2 and 3 days). Western blot analysis of phosphorylated ERK1/2 showed that thrombin induced transient and biphasic activation of ERK, which peaked at 10 and at 120 min post-stimulation (Figure 2a). The ERK phosphorylation level returned to baseline after 4 h, and remained unchanged up to 8–12 h. Replenishment of the medium with thrombin every 24 h did not modify the ERK activation profile at later time periods (results not shown).
PAR-1 activation by thrombin stimulates ERK1/2 phosphorylation
In order to identify the PAR subtype involved in the thrombin-mediated effects on ERK1/2 activation, we stimulated RPE cells with specific TRAPs (thrombin receptor agonist peptides). The PAR-1 agonist (25 μM) was able to induce the first (10 min) and second (120 min) peak of ERK1/2 phosphorylation at the same level as thrombin. Neither PAR-3 nor PAR-4 agonist peptides, alone or in combination, modified ERK activation, as shown in Figure 2(b) (only the first 10 min peak is shown).
Thrombin-induced ERK1/2 activation is specific and unrelated to EGFR [EGF (epidermal growth factor) receptor] transactivation
The specificity of the effect of thrombin on ERK1/2 activation at 10 and 120 min was demonstrated by the abolition of both phosphorylation peaks by the inclusion of the thrombin-specific inhibitor hirudine 30 min prior to stimulation (Figure 3a). Moreover, replacement of the medium with thrombin-free Opti-MEM® at 20 min post-stimulation, or the addition of hirudine following the first ERK1/2 activation peak, prevented ERK1/2 activation at 120 min (results not shown). These results rule out the possibility that the second (120 min) peak of ERK activation resulted from early thrombin-induced release of growth factors into the culture medium.
ERK1/2 activation by thrombin is unrelated to EGFR
Specifically, ERK1/2 activation by thrombin has been ascribed to the transactivation of EGFR in some cells [43–45]. To address this possibility in the RPE, cells were treated with the specific EGFR inhibitor AG-1478 prior to thrombin stimulation. Our results showed that ERK1/2 activation by thrombin does not result from EGFR transactivation (Figure 3b) and, moreover, that RPE cell proliferation was not prevented by AG-1478 (Figure 3c).
Thrombin-induced ERK1/2 activation bypasses Ras activation
Although the linear activation of Ras/Raf/MEK/ERK is the common sequence followed by receptor-induced MAPK pathway activation in most cell types, it has not been defined for thrombin-activated PAR-1 in the RPE. In order to analyse the sequential activation of the MAPK pathway components by thrombin in RPE cells, we used the Ras farnesylation inhibitor manumycine, the Raf-1 inhibitor and the MEK1/2 inhibitor U0126. The activation status of the terminal kinase ERK1/2 was analysed by Western blotting following 10 min stimulation with thrombin in cultures pre-treated with the above-mentioned inhibitors. As shown in Figure 4, although the Raf-1 inhibitor (50 nM) and the MEK inhibitor U0126 (10 μM) prevented thrombin-induced ERK1/2 activation, the Ras-activation inhibitor manumycine (20 and 40 μM) did not. The same result was observed for the late peak (120 min) of ERK1/2 activation (results not shown). These results suggest that thrombin activates ERK in a Ras-independent manner.
ERK1/2 activation by thrombin is Ras independent
Thrombin-induced RPE cell proliferation depends on PLC-β activation
PLC-β has been involved in thrombin-induced signalling in human RPE cells . We analysed the effect of thrombin stimulation on PLC-β activity by measuring [3H]inositol phosphate formation in cells incubated overnight in the presence of myo-[2-3H (N)]inositol. Figure 5(a) shows that thrombin (2 units/ml) increases 3H-labelled PI formation by approx. 200%. This effect was prevented by hirudine, as well as by the PLC-β inhibitor U-73122. Moreover, U-73122 also abolished thrombin-induced RPE cell proliferation (Figure 5b). Since the MEK inhibitors PD98058 and U0126 also prevented the effect of thrombin on proliferation, these results suggest that the upstream activation of PLC-β is required for thrombin induction of MEK activation and proliferation.
Thrombin activation of PLC-β stimulates RPE cell proliferation
Thrombin-induced RPE cell proliferation requires the activation of cPKC (classic PKC) and nPKC (novel PKC) isoenzymes
The activation of both cPKC (α, β and γ) and nPKC (δ, ε, η and θ) isoenzymes has been shown to require membrane-associated phosphatidylinositol bisphosphate breakdown of DAG (diacylglycerol) and IP3. The IP3-mediated release of Ca2+ from intracellular pools recruits classic isoenzymes to lipid rafts, where membrane-associated DAG fully activates the cPKC isoenzymes . nPKCs, however, do not require DAG for full activation.
In order to establish the downstream effectors of PLC-β activation leading to proliferation, we examined the participation of PKC isoenzymes in thrombin-stimulated RPE cells by measuring the effect of the broad-range protein kinase inhibitor staurosporine and that of Ro-32-0432, a specific inhibitor of the cPKCs (α, β and γ) and the nPKC PKCε on cell proliferation. As shown in Figure 6, both staurosporine (Figure 6a) and Ro-32-0432 (Figure 6b) inhibited thrombin-induced cell proliferation in a dose-dependent manner. However, although staurosporine partially inhibited proliferation, the effect of thrombin was completely prevented by Ro-32-0432. This result suggests that both cPKC (α, β and γ) and nPKC (ε) isoenzymes are involved in this process.
cPKC and nPKC isoenzymes are involved in thrombin-induced RPE cell proliferation
Thrombin-induced ERK1/2 phosphorylation is unrelated to PLC-β, cPKC and nPKC activity
The cPKC isoenzyme PKCα has been shown to activate Raf-1 by direct phosphorylation , and PKCε has also been shown to activate Raf-1 in a Ras-independent manner . Hence the Ras-independent activation of MEK/ERK by thrombin could be elicited by the activation of cPKC and nPKC isoenzymes derived from PLC-β activity shown in the present study (Figures 5 and 6). In order to explore this possibility, we measured the effect of thrombin on ERK1/2 phosphorylation in the presence of the PLC-β inhibitor U-73122 and the PKC inhibitors staurosporine and Ro-32-0432. The results showed that none of the inhibitors affected thrombin-induced activation of ERK at concentrations shown to inhibit cell proliferation (Figure 7).
ERK1/2 activation by thrombin is unrelated to PLC-β, cPKC and nPKC
These results reveal that thrombin stimulates RPE cell proliferation through the activation of a PLC-β/PKC signalling pathway, independent from ERK activation. Moreover, since cell proliferation was abolished by the independent inhibition of MEK or PKC, our results suggest a co-operative interaction of the MAPK pathway and PLC-β/PKC signalling in the induction of proliferation by thrombin.
Atypical PKCζ mediates Ras-independent ERK1/2 activation by thrombin
Within the PKC family of serine/threonine kinases, PKC isoenzymes ζ and λ/ι are considered atypical, since their activation does not require Ca2+ or DAG, but depends on the activity of PI3K (phosphoinositide 3-kinase) .
In order to analyse further the mechanism involved in Ras-independent activation of ERK by thrombin, we examined the possible involvement of atypical PKC isoforms in this process, using a myristoylated peptide pseudosubstrate for PKCζ (myristic acid–Ser-Ile-Tyr-Arg-Arg-Gly-Ala-Arg-Arg-Trp-Arg-Lys-Leu). As shown in Figure 8(a), this pseudosubstrate prevented Ras-independent ERK1/2 activation by thrombin. Moreover, inhibition of PI3K, the upstream activator of PKCζ, by wortmannin also prevented thrombin stimulation of ERK phosphorylation (Figure 8a). As expected, PKCζ inhibition also abolished RPE cell proliferation (Figure 8b).
PI3K and atypical PKCζ activation is required for thrombin-induced ERK1/2 phosphorylation and RPE cell proliferation
Proliferative eye diseases, which eventually lead to blindness, represent an important cause of failure in surgery aimed at correcting retinal detachment . In response to ocular stress, such as trauma, photocoagulation, retinal detachment or ischaemia, the blood–retina barrier is compromised, allowing ocular tissues, such as the RPE, to come in contact with blood constituents . Thrombin has been shown to stimulate cell proliferation via the activation of GPCRs in several cell types, including the RPE , although the molecular mechanisms leading to this outcome remain largely undefined.
The results in the present study demonstrate that thrombin, which is contained in blood serum, induces rat RPE cell proliferation through the GPCR PAR-1, via Ras-independent activation of the Raf/MEK/ERK cascade. We provide evidence for the involvement of PLC-β, PI3K and PKC isoenzymes in the regulation of this signalling pathway. We demonstrate, for the first time, the requirement for PKCζ in thrombin-induced effects on RPE cells.
MAPK signalling plays a central role in several cellular processes, including proliferation [16,53–55]. Our results demonstrate that thrombin induces RPE cell proliferation by PAR-1 activation of the MEK/ERK1/2 module of the MAPK pathway. The specificity of the effect of thrombin through PAR-1 was established by the inhibition of proliferation and MAPK activation by hirudine, and by the lack of effect of the agonist peptides for PAR-3 and PAR-4 on ERK1/2 phosphorylation (Figure 2).
ERK1/2 activation was found to be both transient and biphasic, peaking at 10 and 120 min following stimulation, which is in agreement with previous observations in astrocytes , fibroblasts  and endothelial cells . In astrocytes, the late response has been ascribed to receptor recycling, whereas thrombin promotes the cis-transactivation of PAR-1 gene expression in endothelial cells . Although the membrane recycling dynamics for PAR-1 in the RPE remain to be established, unresponsiveness to thrombin up to 72 h following the late ERK phosphorylation peak (Figure 2) rules out the appearance of newly synthesized uncleaved receptors at the membrane. Additionally, suppression of the increase in ERK phosphorylation at both time points by hirudine eliminates a possible non-specific effect mediated by thrombin-induced release of neurotrophins or neuroactive compounds known to activate the MAPK pathway. Among these compounds, EGF activation of the MAPK pathway in the RPE has been reported , and, in order to determine if an early release of EGF by thrombin could be responsible for the late ERK activation peak, we tested the effect of the specific EGFR inhibitor AG-1478 and showed that it did not affect the effect of thrombin on ERK or proliferation (Figure 3).
Canonical activation of the MEK/ERK module by ligand-gated receptors with intrinsic receptor tyrosine kinase activity has been studied extensively and shown to depend on the activation of the monomeric GTPase Ras by GEFs, such as SOS (Son of sevenless)-1/SOS-2 . A major downstream target of activated Ras-GTP is the serine/threonine kinase Raf-1, which subsequently activates MEK/ERK1/2. Thrombin has been shown to activate the classic Ras-dependent MAPK pathway in human  and canine  tracheal smooth muscle cells. Our results, however, showed that thrombin-induced ERK phosphorylation in RPE cells bypasses Ras, since it was not prevented by the Ras farnesylation inhibitor manumycine, which has been shown to suppress the membrane anchoring of Ras  and Ras activation by serum in RPE cells .
Ras-independent activation of MEK/ERK has been observed in human platelets and ascribed to PLC-β activity, leading to the downstream activation of PKC . Upon activation, PAR-1 signalling to Gq/11α, Gαi and G12/13α could be involved in ERK activation and/or cell proliferation. We showed that, in contrast with fibroblasts in which ERK activation has been shown to depend on the activation of Gi by PAR-1 , in RPE cells, neither PTX nor the inhibitor of Rho kinase (Y-27632), the downstream effector of G12/13α, inhibited the effect of thrombin on ERK or cell proliferation (results not shown), suggesting that Gq/11α could be responsible for thrombin actions through PAR-1.
We explored this possibility and demonstrated that the activation of PLC-β by thrombin increases IP3 formation in RPE cells by approx. 200%. This effect was prevented by hirudine as well as by the PLC-β inhibitor U-73122. Furthermore, the inhibition of PLC-β also prevented thrombin-induced proliferation, suggesting a causal relationship from PAR-1 to PLC-β and proliferation.
MEK activation , as well as direct phosphorylation of Raf-1 by cPKCα , has been demonstrated, and the activation of Raf-1 by PKCε in a Ras-independent manner has also been reported . Consistent with these results, we demonstrated that the inhibition of DAG-dependent PKC isoforms by staurosporine and Ro-32-0432 prevented thrombin-induced proliferation of RPE cells. However, in contrast with results showing that Ras-independent ERK1/2 phosphorylation depends on the activation of cPKC or nPKC isoenzymes [48,61,63], neither of these compounds prevented Ras-independent activation of ERK1/2 by thrombin. Since Ro-32-0432 has been shown to inhibit the cPKC (α, βI, βII and γ) and nPKC (ε) isoforms, but not PKCδ or PKCθ, the possible involvement of these isoforms in ERK1/2 activation by thrombin in our system cannot be disregarded.
Together, these results suggested that thrombin action on PAR-1 promotes proliferation by activating two distinct intracellular signalling pathways: PLC-β/PKC and the Ras-independent Raf/MEK/ERK MAPK cascade. Since the pharmacologic inhibition of either individual pathway completely suppressed proliferation, we conclude that the joint activation of these pathways is required for the effect of thrombin on proliferation. In support of this assumption, although the inhibition of cPKC isoenzymes prevented proliferation, the direct activation of PKC by PMA alone, or in combination with a Ca2+ ionophore (ionomycin), had no effect (results not shown).
Previous studies on this subject have shown that, although insulin-induced ERK phosphorylation does not relate to DAG-dependent PKC isoforms, PI3K and PKCζ (or PKCλ, which is 72% homologous with PKCζ and shares an identical pseudosubstrate sequence), as well as MEK1, are required for insulin-induced activation of ERK in rat adipocytes . Because atypical PKCζ is known to serve as an effector of PI3K in different cell types , we tested the involvement of PKCζ in Ras-independent activation of ERK by thrombin.
Our results demonstrate that the pseudosubstrate peptide for the inhibitory region of PKCζ concomitantly inhibited ERK phosphorylation and cell proliferation induced by thrombin. Although direct phosphorylation of ERK by PKCζ was not observed, since MEK is considered to be the exclusive upstream kinase for ERK, the phosphorylation of 14-3-3 scaffold protein by PKCζ has been associated with Raf-1 activation , which could explain the requirement for PKCζ activity in Ras-independent ERK phosphorylation by thrombin in the RPE.
Findings regarding the participation of PKCζ in ERK activation are controversial, possibly as a result of the particular cell types or receptors analysed. Inhibition of PI3K, the upstream activator of PKCζ, has been shown to inhibit ERK activity induced by interleukins [66,67] and growth factors in different cell lines [68,69]. In contrast, PI3K inhibition by wortmannin and LY-294002 had no effect on ERK phosphorylation in response to the ligand-stimulated chemokine receptor CXCR3 (CXC chemokine receptor 3) in hepatic stellate cells  or on activation of ERK in response to EGF in glioblastoma cells .
Gβγ subunits couple PAR-1 to distinct signalling pathways, notably activation of PI3K. In astrocytes, the effect of PAR-1 agonists on activation of ERK1/2 and proliferation are strongly inhibited by the PI3K blocker wortmannin. PAR-1 activation of ERK and proliferation in these cells depends on a PTX-sensitive pathway mediated by Gβγ, PI3K and Ras, and a PTX-insensitive pathway involving PKC and Raf . These mixed observations led us to examine whether PI3K in RPE cells may be activated in response to stimulation of PAR-1 by thrombin.
In agreement with the requirement of PKCζ activity for the activation of ERK and proliferation by thrombin (Figure 8), inhibition of PI3K by wortmannin also inhibited ERK1/2 activation in RPE cells. These findings demonstrate that, although PI3K activity is essential for the activation of ERK, its activity may also be required to stimulate additional proliferative pathways that do not involve ERK and which remain to be defined.
In conclusion, in the present study we show that the activation of the GPCR PAR-1 by thrombin triggers RPE cell proliferation by the joint activation of the ERK1/2 MAPK signalling cascade, in a Ras-independent manner, and that of PI/PLC-β-PKC, upstream of MAPK activation. The present study shows, for the first time, the involvement of PKCζ-mediated phosphorylation of ERK1/2 in the proliferative response of RPE cells to thrombin, and further supports an important role for thrombin in the pathogenesis of PVR induced by injury or retinal surgery.
This work was supported by the Programa de Apoyo a Proyectos de Investigación e Innovación Tecnológica/Universidad Nacional Autónoma de México [grant number IN203507]; and the Consejo Nacional de Ciencia y Tecnología [grant number 80398] to A.M.L.-C.
epidermal growth factor
epidermal growth factor receptor
fetal bovine serum
mitogen-activated protein kinase
protein kinase C
retinal pigment epithelium
Son of sevenless