The regulation of connective tissue growth factor expression influences the viability of human trabecular meshwork cells

Connective tissue growth factor (CTGF) induces extracellular matrix (ECM) synthesis and contractility in human trabecular meshwork (HTM) cells. Both processes are involved in the pathogenesis of primary open-angle glaucoma. To date, little is known about regulation and function of CTGF expression in the trabecular meshwork (TM). Therefore, we analysed the effects of different aqueous humour proteins and stressors on CTGF expression in HTM cells. HTM cells from three different donors were treated with endothelin-1, insulin-like growth factor (IGF)-1, angiotensin-II, H2O2 and heat shock and were analysed by immunohistochemistry, real-time RT-PCR and Western blotting. Viability after H2O2 treatment was measured in CTGF silenced HTM-N cells and their controls. Latrunculin A reduced expression of CTGF by about 50% compared to untreated HTM cells, whereas endothelin-1, IGF-1, angiotensin-II, heat shock and oxidative stress led to a significant increase. Silencing of CTGF resulted in a delayed expression of αB-crystallin and in reduced cell viability in comparison to the controls after oxidative stress. Conversely, CTGF treatment led to a higher cell viability rate after H2O2 treatment. CTGF expression is induced by factors that have been linked to glaucoma. An increased level of CTGF appears to protect TM cells against damage induced by stress. The beneficial effect of CTGF for viability of TM cells is likely associated with the effects on increased ECM synthesis and higher contractility of the TM, thereby contributing to reduced aqueous humour outflow facility causing increased intraocular pressure.


Introduction
Primary open-angle glaucoma (POAG), one of the major causes of blindness worldwide [1], is a neuropathy of the optic nerve leading to a loss of axons at the optic nerve head. The critical risk factor for POAG is intraocular pressure (IOP) which is frequently elevated [2][3][4]. Elevated IOP is caused by an abnormally high aqueous humour (AH) outflow resistance that is generated in the juxtacanalicular region of the human trabecular meshwork (HTM) [5,6]. To date, the mechanisms that are responsible for the increase in TM outflow resistance in POAG are not fully understood [7,8]. There is some evidence though that changes in the amounts and the composition of the HTM extracellular matrix (ECM) [9,10] as well as in the actomyosin system of HTM cells [11][12][13][14] are involved.
Currently, transforming growth factor (TGF)-b2 is one of the leading candidates among the multiple signalling molecules in the AH that may cause molecular changes leading to an increase in outflow resistance in POAG. Accordingly, patients suffering from POAG exhibit higher levels of TGF-b2 in the AH when compared to healthy controls [15][16][17][18]. TGF-b2 is a strong inducer of the HTM ECM and a modifier of the actin cytoskeleton in TM cells [19]. Perfusion of anterior segments with TGF-b2 results in an increase in outflow resistance [20].
In recent studies, we showed that connective tissue growth factor (CTGF) is mediating most of the ECM effects of TGF-b2 on TM cells [21]. In addition, CTGF is able to modulate the biological properties of the TM actin cytoskeleton and to increase its contractility [11]. CTGF belongs to a family of regulatory proteins that are upregulated in a substantial number of disorders associated with a pathological increase in ECM [22][23][24]. In the HTM, CTGF is among the most highly expressed genes [25]. The comparison of the CTGF expression in Schlemm's canal endothelial (SC) cells derived from glaucoma patients with SC cells from healthy donors reported a significant higher CTGF expression level in the glaucomatous SC cells [26]. CTGF has also been detected in the AH [27] and its amounts are increased in patients with pseudoexfoliation syndrome and glaucoma associated with it [28,29]. In patients with POAG, only a slight increase in CTGF was found in the AH in one study [28], whereas other authors reported a significant increased level of CTGF [30].
To investigate the influence of CTGF signalling in the living eye, we recently developed a mouse model with an eye-specific overexpression of CTGF by the use of a lens-specific promotor. Since the CTGF-mediated effects on ECM and the actin cytoskeleton that correlated with an increase in IOP and a successive loss of axons in the optic nerve, we concluded that high amounts of CTGF cause POAG in the mouse eye [11].
The dynamics of factors that may increase the trabecular expression of CTGF in patients with POAG remain unclear. TGF-b and cyclic mechanical stress are the only known stimuli that have been shown to increase the expression of CTGF in HTM cells [31,32]. In the present study, we show that the regulation of CTGF expression involves various factors including different kinds of stress, suggesting a protective role for CTGF in the HTM.

Cell culture
Cultures of HTM cells were established from the eyes of three human donors according to protocols published previously [32]. The age of the donors ranged from 34 to 76 years. HTM cells of the third to fifth passage were seeded in 35-mm culture wells (4.0 9 10 5 cells/well) and grown to a confluent monolayer in F10-HAM medium plus 10% (v/v) foetal bovine serum without antibiotics in 5% CO 2 at 37°C (PAA, Pasching, Austria). The confluent cells were incubated in serum-free medium for 24 hrs followed by incubation in fresh serum-free medium. Endothelin-1 (ET-1), angiotensin-II (Ang II), insulin-like growth factor (IGF)-1, hydrogen peroxide (H 2 O 2 ) or latrunculin A (LatA; Sigma-Aldrich, Taufkirchen, Germany) were added at various concentrations at different time-points. Control cells were treated equally with corresponding vehicles.
Heat-shock experiments were performed by seeding immortalized HTM cells (HTM-N) [33] and pSiCTGF-HTM-N cells (stable plasmidbased siRNA silencing of the CTGF gene) into Petri dishes to a confluent level (1.0 9 10 6 cells/well). The pSiCTGF-HTM-N cell line has an 80% reduced CTGF expression in comparison to the HTM-N cell line as described previously [11]. The cells were kept for 24 hrs under serum-free conditions. Then, cells were heat shocked for 15 min. at 42°C in a water bath and harvested after a further 37°C incubation period at different time-points [0 (control), 15, 30, 45, 60 min.].
Each of the described experiments except heat shock was done with each of the three primary cell lines. Methods for securing human tissues were humane, included proper consent and approval, and complied with the Declaration of Helsinki.

RNA analysis
Human trabecular meshwork cells were harvested and total RNA was extracted with TRIzol (Invitrogen, Karlsruhe, Germany) according to manufacturer's recommendations. First strand cDNA was prepared from total RNA using the iScript cDNA Synthesis Kit (Bio-Rad, Munich, Germany) according to the manufacturer's instructions. Real-time RT-PCR was performed on a BioRad iQ5 Real-time PCR Detection System (Bio-Rad) using the following temperature profile: 40 cycles of 10 sec. melting at 95°C, 40 sec. of annealing and extension at 60°C. Primer pairs ( Table 1) were purchased from Invitrogen and extended over exonintron boundaries. RNA that was not reversely transcribed served as negative control for real-time RT-PCR. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and guanine nucleotide binding protein 2-like-1 (GNB2L1) were both used as housekeeping genes for relative quantification of the real-time RT-PCR experiments. Quantification was performed with iQ5 Standard-Edition (Version 2.0.148.60623) software (Bio-Rad).

Western blot analysis
To obtain protein extracts, cells were directly lysed in RIPA lysis buffer (150 mM NaCl, 1% NP-40, 0.5% deoxycholic acid, 0.1% SDS and 50 mM Tris, pH 8) and protein content was measured with the bicinchoninic acid protein assay (Pierce, Rockford, IL, USA). Proteins were separated by SDS-PAGE and transferred to polyvinylidene fluoride membranes. Antibodies were used as follows: rabbit anti-human aB-crystallin (1:200; Stressgen, San Diego, CA, USA), goat anti-human CTGF (1:500), donkey anti-rabbit-horseradish peroxidase (HRP) and chicken anti-goat-HRP (1:2000; all Santa Cruz, CA, USA). Chemiluminescence was detected on a LAS 3000 imaging workstation (Raytest, Straubenhardt, Germany). For normalization of the signals, blotted membranes were stained with coomassie blue and digitized. The total amount of protein per lane was determined and calculated using AIDA Image analyser software (Raytest). The values of the total amount of protein were used to normalize the signal intensity of the bands detected in Western blot analysis.

Immunohistochemistry
Cultured HTM-N and pSiCTGF-HTM-N cells were grown on microscope slides and treated with heat shock as described above. After incubation,  cells were fixed with 4% paraformaldehyde for 15 min. and subsequently washed twice with PBS containing 0.1% Triton X-100. Rabbit anti-human aB-crystallin antibodies were added at a 1:50 dilution in PBS/bovine serum albumin (BSA; 5%) for 4 hrs at room temperature. After three wash steps with PBS, fluorescein-conjugated secondary pig anti-rabbit IgG (Dako, Glostrup, Denmark) was added at a 1:1000 dilution in PBS/BSA (5%) for 1 hr at room temperature. Actin stress fibres were stained using phalloidin-TRITC (Sigma-Aldrich) at a dilution of 1:1000 for 1 hr at room temperature. After immunohistochemical labelling, Vectashield mounting medium (Vector Laboratories, Burlingame, CA, USA) was used for mounting slides. Slides were analysed under an Axio Imager fluorescence microscope (Carl Zeiss AG, Oberkochen, Germany). Corresponding negative controls to estimate unspecific binding of secondary antibodies were handled similarly, but incubated in PBS/ BSA without primary antibody.

Cell viability assay
To analyse the CTGF effect on the viability of HTM-N and pSiCTGF-HTM-N cells after treatment with H 2 O 2 , cells were seeded into 96-well plates at 3.0 9 10 4 cells/well. The cells were treated with 50 lM of H 2 O 2 alone or in combination with 50 ng/ml of recombinant CTGF (rCTGF) [21] for 24 hrs under serum-free conditions. Viability of cells was analysed by a MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) assay. The yellow MTT is reduced by a mitochondrial reductase to purple formazan in the mitochondria of living cells. The absorbance of the formed fomazan is measured by 570 nm wave length.The cells were washed with PBS, and 80 ll/well MTT solution (tetrazolium dye, PBS and serum-free medium) was added for 2 hrs. The formazan crystals that formed were dissolved by the addition of dimethyl sulfoxide (DMSO). Absorption was measured with a spectrophotometer (Sunrise, Tecan, Maennedorf, Switzerland) at 570 nm and analysed using Magellan software (Tecan). Results are expressed as the mean percentage of the controls. Untreated HTM-N and pSiCTGF-HTM-N cells served as controls. Values of each sample were normalized to a 'blank' containing DMSO only.

Number of experiments and statistical analysis
To assess the effects of the different treatments, each experiment was repeated at least three times from primary HTM cell lines of three different donors and with the HTM-N and pSiCTGF-HTM-N cells. Student's t-test was used for statistical analysis and differences with a P-value smaller 0.05 were regarded as significant, and with a P < 0.01 as highly significant.

Heat shock induces CTGF expression
As mechanical stress induces CTGF in HTM cells [31], we investigated whether CTGF is also induced by other stressors. To this end, HTM-N cells were treated with heat shock (42°C) and were then allowed to recover at 37°C. After 30 min., the expression of CTGF was significantly increased (3.3 AE 0.4-fold, P < 0.05) and remained at this level during the entire 60 min. of the recovery phase (Fig. 1A).

CTGF induces aB-crystallin expression
Amounts of the small heat-shock protein (sHSP) aB-crystallin are higher in the TM of POAG eyes [34] and are induced by TGF-b in cultured HTM cells [35]. Here, we were interested to learn whether there is a link between aB-crystallin and CTGF in HTM cells. HTM cells treated with recombinant CTGF showed a dose-dependent increase in the amounts of aB-crystallin, which were 3.2 AE 0.3-fold higher after treatment with 10 and 50 ng/ml of CTGF for 24 hrs (P < 0.05, Fig. 1B).
We now wanted to know whether the expression of aB-crystallin is changed in the presence of lower amounts of CTGF. For that purpose, we used the pSiCTGF-HTM-N cell line and compared it to HTM-N cells [11]. HTM-N cells treated with 42°C for 15 min. showed an increased immunohistochemical staining for aB-crystallin and an increase in actin stress fibre formation after 30 min. of recovery phase (Fig. 1E). Following Western blot experiments, densitometric analysis of the immunoblots identified a significant up-regulation of aB-crystallin in HTM-N cells (4.1 AE 0.3-fold, P < 0.05, 45 min., Fig. 1D). In contrast in pSiCTGF-HTM-N cells, aB-crystallin staining and stress fibre formation was unchanged after heat shock exposure (Fig. 1E). Western blot analysis of aB-crystallin synthesis in pSiCTGF-HTM-N cells showed a delayed and reduced induction in comparison to HTM-N cells. An induction (2.1 AE 0.5-fold) of aBcrystallin was only observed after 60 min. of recovery phase (Fig. 1D).

Oxidative stress induces CTGF expression
As increasing age comes along with higher amounts of reactive oxygen species (ROS) in the eye [36], we investigated whether oxidative stress influences the expression of CTGF in HTM cells. To this end, cells were treated with 50 lM H 2 O 2 for 1, 3, 6 or 24 hrs and CTGF and its mRNA were determined by real-time RT-PCR and Western blotting. A 3-hrs treatment resulted in a significant increase in CTGF mRNA expression up to 1.8 AE 0.2-fold (P < 0.05) which remained upregulated for at least 24 hrs ( Fig. 2A). Moreover, after 3 hrs of oxidative stress, the amounts of CTGF were 2.2 AE 0.3-fold higher than in control cells (P < 0.05). After 24 hrs, the amounts of CTGF were still 1.9 AE 0.2-fold higher than in controls (P < 0.05; Fig. 2B).

CTGF protects HTM cells against oxidative stress
As CTGF appeared to be a primary response gene to stress and as the amounts of CTGF correlate with sHSP expression, we investigated whether CTGF affects the viability of H 2 O 2 -treated HTM-N cells. For this purpose, we treated HTM-N and pSiCTGF-HTM-N cells with 50 lM H 2 O 2 alone or in combination with 50 ng/ml of CTGF for 24 hrs. Measuring cell viability was conducted by a MTT assay.

Treatment with LatA causes a reduction in the amounts of CTGF and its mRNA
Treatment of HTM cells with 30 nM LatA for 24 hrs led to a highly significant 0.55 AE 0.1-fold reduction in CTGF mRNA expression when compared to untreated control cells ( Fig. 3A; P < 0.01. A comparable effect was observed after treatment with 60 nM LatA, which did also cause a reduction in the amounts of CTGF (P < 0.01, Fig. 3B).
Treatment with ET-1, Ang II or IGF-1 induces CTGF and its mRNA To assess the effects of ET-1 on the expression of CTGF in the TM, HTM cells were treated with ET-1 at a concentration of 100 nM for 1-24 hrs and 1 day to 3 days. The amounts of CTGF and its mRNA were analysed by real-time RT-PCR and Western blot analysis. Treatment for 1, 3 and 12 hrs did not cause significant changes in the expression of CTGF (data not shown). After 3 days of treatment, the expression of mRNA for CTGF increased significantly by 1.6 AE 0.2fold ( Fig. 4A; P < 0.05). In addition, higher amounts of CTGF were detected in cell extracts of treated cells compared to untreated control cells (Fig. 4B).
To study the effects of Ang II, HTM cells were treated for 1, 3 or 6 hrs with concentrations of 10 nM and 1 lM of Ang II and analysed as described above. Following treatment with 10 nM Ang II, the amounts of CTGF and its mRNA remained unaffected (P > 0.05, Fig. 5B). In contrast, treatment with 1 lM of Ang II led to significant changes after 3 hrs of treatment. The expression of mRNA for CTGF increased up to 2.16 AE 0.18-fold ( Fig. 5A; P < 0.05), an effect that correlated with the presence of higher amounts of CTGF (Fig. 2B). When cells were treated for 6 hrs, no significant changes were observed when compared with control cells (Fig. 5B).
For studies on the CTGF-inducing effect of IGF-1, HTM cells were treated with IGF-1 at concentrations of 5 and 50 ng/ml for 6 and 24 hrs respectively. After 6 hrs of incubation, there was no significant increase between IGF-1-treated and control cells with respect to the amounts of CTGF and its mRNA. In contrast, treatment with 50 ng/ml of IGF-1 for 24 hrs led to a significant 2.2 AE 0.2-fold increase of CTGF mRNA when compared to untreated control cells ( Fig. 6A; P < 0.05). The increase in mRNA caused a significant increase in the amounts of CTGF (P < 0.05, 2.1 AE 0.3-fold, Fig. 6B).

Discussion
We conclude that CTGF is a primary response gene after exposure to stress that protect HTM cells from injury. The protective effect of CTGF appears to be associated with a stabilization of the actin cytoskeleton. This conclusion is based on (i) the findings that CTGF is immediately upregulated in HTM cells after treatment with different stressors and (ii) that CTGF induces the expression of the sHSP aBcrystallin. (iii) The conclusion is supported by the observation that a knockdown of CTGF leads to a reduced cell viability in HTM cells after exposure to stress and to an attenuated up-regulation of aB-crystallin, (iv) whereas a pretreatment with CTGF could prevent the stressinduced cell loss.
The cellularity of the TM decreases throughout life [37,38] and the loss of TM cells is more pronounced in patients with POAG [39]. The causes for the loss of TM cells are not clear, but it is known that TM cells are exposed to many different stressors during the ageing process [40,41]. Among the stressors is mechanical stress mediated by ciliary muscle contraction and aqueous flow rate changes. Mechanical stress is thought to be elevated in POAG patients with increasing IOP [40], a major risk factor for glaucoma. Besides mechanical stress, TM is constantly exposed to oxidative stress [41].
In many diseases, it is accepted that the ageing process is associated with a higher exposure rate to free radicals, which would lead to a decline of physiological functions of various tissues [42] and several studies suggest that an increased formation of ROS might be associated with glaucoma [36,43]. In POAG patients, a significant reduction in mitochondrial respiratory activity and a significant decrease in anti-oxidative capacity were detected [41,[44][45][46]. Those findings resulted in the hypothesis that oxidative stress could be involved in onset and progression of POAG. Here, we show that CTGF is a primary response gene after oxidative stress in TM cells. The data are supported by in vivo analysis of early response genes after oxidative stress. In mice, the induction of oxidative stress in the cerebellum led to a substantial increase in CTGF within 6 hrs [47]. The immediate up-regulation of CTGF under stress conditions was further confirmed by our heat-shock experiments. Together with the known findings that also mechanical stress is able to induce CTGF expression [31], we conclude that CTGF might be a general primary response gene to various kinds of stressors in HTM cells.
The physiological function of the early up-regulation of CTGF seems to be a protective mechanism in HTM cells. The supplementation of CTGF prior to H 2 O 2 treatment had a beneficial effect on the viability of TM cells. A potential role for CTGF in cell survival was shown in gallbladder cancer cells, where silencing of CTGF led to a reduced cell viability [48]. We could observe a similar effect in TM cells, where reduced levels of CTGF led to a decline in cell viability rate after oxidative stress, whereas adding CTGF partially rescued the loss of TM cells. A protective function of CTGF was previously shown in the kidney, where supplementation of CTGF guarded puromycintreated podocytes from cell death [49].
The protective effect of CTGF might be directly linked to the expression of the sHSP aB-crystallin, as CTGF treatment led to a significant up-regulation of aB-crystallin in HTM cells. aB-crystallin belongs to the family of sHSPs, and it is known to be up-regulated in the TM of POAG patients [34]. The increased presence of sHSPs might have a protective effect, as TM cells respond to oxidative stress and heat shock by aB-crystallin induction [50], whereas silencing of CTGF in TM cells blocked the stress-induced up-regulation of the aB-crystallin. As both proteins are simultaneously regulated during the exposure to heat shock, we assume that CTGF acts as modulator of the aB-crystallin synthesis, because of the matricellular character of CTGF [51]. sHSPs are able to protect cells by different mechanisms depending on their subcellular localization. Under stress conditions, aB-crystallin can translocate to the mitochondria and thereby interacting with various components of the mitochondrial apoptotic machinery and preventing cell death [52,53], whereas the cytosolic aB-crystallin can inhibit actin depolymerization, thereby leading to an increased cell survival [54]. We assume that CTGF protects the cells against the oxidative stress-induced disruption of the cytoskeleton and disaggregation of actin fibres, a critical point for cell survival [54]. In an earlier study, we could already show the positive effect of CTGF on formation of actomyosin fibres and the contractility in HTM cells [11], whether the mitochondrial apoptotic events are also altered after CTGF treatment have to be investigated in the future.
Based on our observations, we wanted to address additionally the question whether CTGF regulation in HTM cells is also linked to other factors, which are present in the AH and/or are involved in the outflow facility regulation and are assumed to be involved in CTGF regulation in other tissues.
In the context of a CTGF-mediated induction of ECM synthesis, we also investigated the effect of IGF-1 on CTGF expression. IGF-1 stimulates CTGF to induce collagens via binding to the IGF-binding domain of CTGF [55]. IGF-1 is present in the AH [56] and is expressed in the TM together with its receptors [57]. In studies on the signalling pathways of IGF-1, the RhoA/ROCK signalling pathway is among the most commonly highlighted [58]. In our study, physiological concentrations of IGF-1 led to an increased expression of CTGF in TM cells [56]. Little is known though about the function of IGF-1 and its receptor within the TM, an avenue that should be analysed in the future.
Clearly, the common pathway between the molecules investigated here appears to be the RhoA/ROCK pathway that is involved in the regulation of outflow facility by altering the actin cytoskeleton of the TM. Previous studies showed that Ang II and ET-1 induce CTGF expression via the small GTPase RhoA [59,60]. In this study, we obeserved a late induction of CTGF after ET-1 treatment indicating that the increase in CTGF could be a secondary effect. On the other hand, Horstmeyer et al. observed a pronounced late induction of CTGF after ET-1 in comparison to short time treatments. This finding was explained with different expression patterns of the two endothelin receptors A and B at the different points [61]. In contrast, a rapid increase in CTGF after Ang II treatment clearly implicates a direct The CTGF-inducing effect of RhoA was shown by the treatment of HTM cells with a constitutive active form of RhoA, whereas the inhibition of the Rho kinase, a downstream mediator of RhoA, lead to a substantial down-regulation of CTGF [62]. The CTGF expression is coupled to the stability of the actin cytoskeleton [63], the formation of F-actin stress fibres led to an increased CTGF expression. This regulation was dependent on the availability of G-actin. Actin disrupting agents like LatA and B, which cause an enhanced cellular content of G-actin, lead to a reduction in CTGF expression in different cell lines [64]. Quite similarly, the LatA treatment of HTM cells caused a significant down-regulation of CTGF. As latrunculin treatment of perfused anterior chambers lead to an increase in outflow facility [65,66], we speculate that, given the loss of contractility in TM cells, the decrease in CTGF-mediated ECM deposition might contribute in the long run to the outflow increasing effect of latrunculins. On the other hand, under stress, like increased mechanical load or oxidative stress, the TM cells react with increased F-actin stress fibres formation, leading to decreased G-actin levels and thereby to an enhanced CTGF expression. Thereby, inducing a vicious circle, as CTGF by itself can activate the RhoA/ROCK signalling pathway [11]. If the stress last only for a short period, the expression of CTGF declines rapidly to a basal expression, which led to the assumption that compensatory mechanisms are activated to normalize the CTGF expression levels [67]. Under pathological conditions, which were shown in other tissues, those mechanisms seem to fail and the CTGF up-regulation maintains in those tissues leading to fibrotic changes [68,69]. We conclude that CTGF is an important regulatory molecule in the TM and that the RhoA/ROCK signalling pathway is involved in some CTGF-mediated TM effects [11,62,70]. The newly identified molecules and stressors, together with the findings that CTGF is induced by dexamethasone, TGF-b1 and 2 in HTM cells [31,32], brings up the idea that CTGF may play an important role in the pathogenesis of glaucoma. Along this line, the reported constitutive basal expression of CTGF in the TM [25] could be necessary for maintenance of the TM actin cytoskeleton and ECM. An immediate upregulation of CTGF following short-term stress might be an important mechanism to protect TM cells from damage and death. In contrast, a continuous upregulation of CTGF after chronic cellular stress might significantly alter TM homoeostasis and lead to an increase in TM ECM, actin-mediated contractility and finally stiffness. Clearly, such a scenario is likely to result in increased outflow resistance and IOP, and finally POAG.