Gap junctional intercellular communication in bovine corneal endothelial cells

Experimental Eye Research 83 (2006) 1225e1237 www.elsevier.com/locate/yexer

Gap junctional intercellular communication in bovine corneal endothelial cells Priya Gomes a, Sangly P. Srinivas b, Johan Vereecke a,*, Bernard Himpens a a

Laboratory of Physiology, KU Leuven, Campus Gasthuisberg O/N, box 802, Herestraat 49, B-3000 Leuven, Belgium b School of Optometry, Indiana University, Bloomington, IN 47405, USA Received 28 February 2006; accepted in revised form 26 June 2006 Available online 28 August 2006

Abstract Gap junctions and/or paracrine mediators, such as ATP, mediate intercellular communication (IC) in non-excitable cells. This study investigates the contribution of gap junctions toward IC during propagation of Ca2þ waves in cultured bovine corneal endothelial cells (BCEC) elicited by applying a point mechanical stimulus to a single cell in a confluent monolayer. Changes in [Ca2þ]i were visualized using the fluorescent dye Fluo-4. The area reached by the Ca2þ wave, called the active area (AA), was determined as a measure of efficacy of IC. RTePCR and Western blotting showed expression of Cx43, a major form of connexin, in BCEC. In scrape-loading (using lucifer yellow) and fluorescence recovery after photobleaching (FRAP; using carboxyfluorescein) protocols, significant dye transfer of the hydrophilic dyes was evident indicating functional gap junctional IC (GJIC) in BCEC. Gap27 (300 mM), a connexin mimetic peptide that blocks gap junctions formed by Cx43, reduced the fluorescence recovery in FRAP experiments by 19%. Gap27 also reduced the active area of the Ca2þ wave induced by point mechanical stimulation from 73,689 mm2 to 26,936 mm2, implying that GJIC contribution to the spread of the wave is at least w63%. Inhibitors of ATP-mediated paracrine IC (PIC), such as a combination of apyrase VI and apyrase VII (5 U/ml each; exogenous ATPases), suramin (200 mM; P2Y antagonist), or Gap26 (300 mM; blocker of Cx43 hemichannels) reduced the active area by 91%, 67%, and 55%, respectively. Therefore, estimating the contribution of GJIC from the residual active area after PIC inhibition appears to suggest that GJIC contributes no more than w9% towards the active area of the Ca2þ wave. Gap27 did not affect the enhancement in active area induced by ARL-67156 (200 mM, ectonucleotidase inhibitor), ATP release induced by point mechanical stimulation, and zero [Ca2þ]o-induced lucifer yellow uptake, indicating that the peptide has no influence on PIC. Exposure to Gap27 in the presence of PIC inhibitors led to a significant further inhibition of the Ca2þ wave. The finding that the residual active area after inhibition of PIC by apyrases was much smaller than the reduction of the active area by Gap27, provides evidence for interaction between GJIC and PIC. These findings together suggest that functional gap junctions are present in BCEC, that both GJIC and PIC contribute significantly to IC, and that the two pathways interact. Ó 2006 Elsevier Ltd. All rights reserved. Keywords: corneal endothelium; intercellular communication; gap junction; connexin mimetic peptide; calcium wave

1. Introduction A coordinated response of a tissue to external stimuli requires robust intercellular communication (IC). In many non-excitable cells, two distinct modes of IC have been

* Corresponding author. Tel.: þ32 16 345732; fax: þ32 16 345991. E-mail address: [email protected] (J. Vereecke). 0014-4835/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.exer.2006.06.012

identified: gap junctional IC (GJIC) and paracrine IC (PIC) (Sanderson et al., 1994; Saez et al., 2003a,b) (Fig. 1). GJIC is a mode of IC involving direct connections between neighboring cells via gap junctions. Gap junctions form when two hemichannels of apposing cells dock together and form a pore that links the cytoplasm of the two cells. This enables a diffusional exchange of molecules between the two cells, as well as electrical coupling (Saez et al., 2003a) (Fig. 1). The pores formed by gap junctions are relatively non-selective and allow permeation of molecules of up to about 1 kDa,

1226

P. Gomes et al. / Experimental Eye Research 83 (2006) 1225e1237

Fig. 1. Paracrine intercellular communication (PIC) and gap junctional intercellular communication (GJIC). Connexins play a role in both forms of intercellular communication, since they form gap junctions, while connexin hemichannels are involved in release of extracellular messengers.

including ions, signaling molecules such as cAMP, IP3 and Ca2þ, and a number of hydrophilic dyes (Saez et al., 2003a). The hemichannels consist of homomeric/heteromeric complexes of the connexin (Cx) family gene products. In many cell types, Cx43 is a major type of Cx (Goodenough and Paul, 2003; Saez et al., 2003a). In contrast to GJIC, PIC involves the release of one or more chemical mediators into the extracellular milieu, and their subsequent interaction with receptors, such as G protein-coupled receptors (GPCRs), on the neighboring cells (Fig. 1). The recruitment of GPCRs for IC can act as a potent and possibly regenerative amplification mechanism via their downstream signaling pathways, and enables communication between distant cells. Although not much is known as to how neighboring cells act as relay centers for IC, evidence has accumulated in favor of ATP functioning as a paracrine mediator in many cell types (Cotrina et al., 2000; Moerenhout et al., 2001; Gomes et al., 2005b). The recent discovery that Cx hemichannels are involved in ATP release has further emphasized a functional role for the nucleotide as a paracrine mediator that is released involving proteins associated with GJIC (Cotrina et al., 1998; Braet et al., 2003b; Ebihara, 2003; Gomes et al., 2005a) (Fig. 1). Given the involvement of a chemical messenger in PIC, the intensity of the paracrine signal is likely to lose its strength during transmission by dilution and degradation of the messenger. This would be the case, except when the messenger has an autocrine effect on cells along the wave propagation path that causes a regenerative release of the messenger. However, since PIC, in contrast to GJIC, does not require cellcell apposition, it can bring about IC even when cells are separated, which may occur when intermediary cells are dead or absent. The primary objective of this study is to examine the relative contribution and role of GJIC and PIC in IC in corneal endothelial cells. These cells form a monolayer lining of the posterior surface of the cornea and serve as an interface between the corneal stroma and the aqueous humor. As a fluid

transporting epithelium, the endothelial monolayer is responsible for maintaining hydration of the stroma, which is essential for corneal transparency (Dikstein and Maurice, 1972; Maurice, 1972; Riley et al., 1998; Bonanno, 2003). Given the need of transparency for acute vision and the non-regenerative nature of the human corneal endothelium, it is of interest to examine how cells of this monolayer function as a syncytium and coordinate their responses to extracellular stresses, such as mechanical stress during intraocular surgery, or exposure to inflammatory mediators during immune rejection or uveitis. In our two recent studies investigating mechanisms underlying IC, we examined intercellular Ca2þ wave propagation induced by a point mechanical stimulus of single cells in a confluent monolayer of cultured bovine corneal endothelial cells (BCEC) (Gomes et al., 2005a,b). The first report characterized the Ca2þ wave and demonstrated its dependence on ATP-mediated PIC (Gomes et al., 2005b). The second study showed that ATP is released in response to point mechanical stimulation by a mechanism involving Cx hemichannels (Gomes et al., 2005a). While expression of Cx43 and dye transfer via gap junctions have been demonstrated in rabbit and rat corneal endothelial cells (Joyce et al., 1998; Laux-Fenton et al., 2003; Williams and Watsky, 2004), involvement of GJIC in Ca2þ wave propagation in the corneal endothelium has not been demonstrated. The objective of this study is, therefore, to investigate the contribution of GJIC and its possible interaction with PIC during point mechanical stimulation-induced Ca2þ wave propagation. Our results show that IC in BCEC is dependent on both GJIC and PIC, and provide evidence that both pathways interact. 2. Materials and methods 2.1. Cell culture Primary cultures of BCEC were established as previously described (Srinivas et al., 1998; Srinivas et al., 2002; Gomes

P. Gomes et al. / Experimental Eye Research 83 (2006) 1225e1237

et al., 2005b). The growth medium contained Dulbecco’s modified Eagle’s medium (DMEM, 11960-044; InvitrogeneGibco, Karlsruhe, Germany) and 10% fetal bovine serum (F 7524, SigmaeAldrich, Deisenhofen, Germany), 6.6% L-glutamine (Glutamax; 35050-038; InvitrogeneGibco) and 1% antibiotic-antimycotic mixture (15240-096; InvitrogeneGibco). Cells were grown at 37
C in a humidified atmosphere containing 5% CO2. Cells of the second and third passages were harvested and seeded into two chambered glass slides (155380, Laboratory-Tek; Nunc, Roskilde, Denmark) at a density of 165,000 cells per chamber (4.2 cm2) unless otherwise stated. Cells were allowed to grow to confluence for 3 or 4 days before use. 2.2. Immunocytochemistry for Cx43 expression and localization Confluent monolayers were fixed with 2% paraformaldehyde and then permeabilized using 1% Triton X-100 for 5 min. This was followed by incubation with freshly made 3% BSA for 20 min, 1% glycine-containing solution for 10 min and 10% rabbit serum for 45 min. The cells were rinsed using Dulbecco’s phosphate-buffered saline (PBS; InvitrogeneGibco) between each treatment. Cells were incubated overnight with a mouse monoclonal antibody directed against goat a-Cx43. The secondary antibody was goat anti-mouse IgG labeled with Alexa FluorÒ 488. Negative controls were obtained using monoclonal mouse-derived IgG1 antibody to Aspergillus niger glucose oxidase, an enzyme which is neither present nor inducible in mammalian tissue. Cells were visualized using a confocal microscope (LSM510, Carl Zeiss Meditec, Jena, Germany). The 488 nm line of the Argon laser was used for excitation. Emission was recorded via a long-pass filter at 505 nm. 2.3. RTePCR assay for expression of Cx43 mRNA Total RNA was harvested using TRIzol reagent (18080093; InvitrogeneGibco, Carlsbad, CA, USA) from BCEC and quantified by absorption at 260 nm. First-strand cDNA synthesis was performed with the SuperScript III Reverse Transcriptase for RTePCR (18080-093; InvitrogeneGibco). To amplify Cx43 mRNA (J05535; homology: human/bovine; annealing temperature: 55.4
C), PCR was performed for 35 cycles [94
C for 30 s; 55
C for 30 s; and 72
C for 45 s (but 10 min in the final cycle)] using 5 0 -GGTGACTGGA GTGCCTTAGGCA-3 0 and 5 0 -TTCAGCTTCTCTTCCTTTC GCATC-3 0 with Taq DNA Polymerase (1732641; Roche, Indianapolis, IN). A 317 bp fragment, representing Cx43, was demonstrated after staining 1% agarose gels with ethidium bromide. The fragment representing Cx43 was prepared for sequencing using ABI Prism BigDye Terminator DNA Sequencing Kit (4303152; Applied Biosystems, Foster City, CA, USA). Sequencing reactions contained 8 ml of Termination Ready Reaction Mix, 1.6 pmol of sequencing primer and 300 ng of the DNA product. Sequenced data was submitted to BLAST nucleotide analysis.

1227

2.4. Dye transfer assays for GJIC 2.4.1. Scrape-loading Confluent cells grown on glass chambers cells were rinsed three times with PBS. A 27-gauge needle was then used to create a scrape through the monolayer in the presence of 0.5% hydrophilic dye lucifer yellow in the extracellular solution. After 3 min of incubation at room temperature, the culture was rinsed three times with PBS and then incubated for an additional 5 min. The cells were then fixed with 4% paraformaldehyde and viewed on the confocal microscope (LSM510; Carl Zeiss Meditec). 2.4.2. Fluorescence recovery after photobleaching (FRAP) Cells were loaded with the Ca2þ-insensitive dye 6-carboxyfluorescein diacetate (10 mM) for 5 min at room temperature. Fluorescence recovery after photobleaching (FRAP) was measured using the confocal microscope (LSM510; Carl Zeiss Meditec). The dye was excited at 488 nm and its emission was recorded at 570 nm. A neutral density filter was used to minimize photobleaching. Before bleaching, polygons were drawn around the cells chosen for bleaching, and two prebleach images were scanned. The cells chosen for bleaching were then exposed to 50 scans with the laser at 95% intensity, and the recovery of fluorescence in the bleached cells was measured every 10 s over a period of 5 min. The decrease of fluorescence in a square region of interest far from the bleached cells was measured as a reference in order to correct for any bleaching caused by the excitation light. After correction for background bleaching, recovery of fluorescence in the bleached cell at 3 min was compared with that of the prebleach scan, and the percentage recovery was calculated. 2.5. Mechanical stimulation for inducing Ca2þ wave Point mechanical stimulation of a single cell consisted of an acute deformation of the cell by briefly touching less than 1% of the cell membrane with a glass micropipette (tip diameter <1 mm) coupled to a piezoelectric crystal (Piezo Actuator P-280, Amplifier-E463; Polytech PI, Karlsruhe, Germany) mounted on a micromanipulator. 2.6. Measurement of [Ca2þ]i The Ca2þ wave propagation was assayed by imaging [Ca2þ]i, as described previously (Gomes et al., 2005b). Cells were loaded with the Ca2þ-sensitive dye Fluo-4 AM (10 mM) for 30 min at 37
C. The dye was excited at 488 nm and its fluorescence emission was collected at 530 nm. Spatial changes in [Ca2þ]i following point mechanical stimulation were measured with the confocal microscope (LSM510; Carl Zeiss Meditec) using a 40 objective unless otherwise mentioned. Images were collected and stored on a personal computer. Polygonal regions of interest were drawn to define the borders of each cell. The mechanically stimulated cell is called the MS cell. The neighboring cells (NB cells) immediately surrounding the mechanically stimulated cell are defined as neighboring

1228

P. Gomes et al. / Experimental Eye Research 83 (2006) 1225e1237

cell layer 1 (NB1), and the cells immediately surrounding the NB1 cells are defined as neighboring cell layer 2 (NB2), and so on. Fluorescence was averaged over the area of each regions of interest. Normalized fluorescence (NF) was defined as the ratio of fluorescence to average fluorescence before point mechanical stimulation. Intercellular Ca2þ waves were characterized by maximum normalized fluorescence, delay, and percentage of responsive cells (%RC), as well as the total area of responsive cells (active area, AA) defined as the total area occupied by cells with normalized fluorescence 1.1. 2.7. Measurement of ATP release The accumulation of released ATP in the solution bathing a monolayer was probed with the luciferin-luciferase bioluminescence assay. Photons emitted as a result of the oxidation of luciferin in the presence of ATP and O2 and catalyzed by luciferase were detected by a photon counting photomultiplier tube (H7360-01; Hamamatsu Photonics, Hamamatsu, Japan). Voltage pulses from the photomultiplier module were counted with a high-speed counter (PCI-6602; National Instruments, Austin, TX). Dark count of the photomultiplier tube was <80 counts/s. Experiments with ATP measurement after mechanical stimulation were performed on the confocal microscope. Samples of 25 ml of the 500 ml bathing solution were taken at different times and quickly transferred to the separate photon counting set-up. In order to measure the accumulated ATP released by the cells, a solution containing luciferin-luciferase was applied. Results of consecutive ATP measurements were corrected for the volume of extracellular fluid removed in previous samples. Calibration curves of luminescence count versus ATP were obtained with known concentrations of ATP in the presence of luciferineluciferase reagent (50 ml/ml) dissolved in the same solutions used in the experiments. Calibration experiments showed that Gap27 (300 mM) was without any effect on the ATP-induced luminescence. 2.8. Hemichannel assay by dye loading Monolayers of cells were incubated for 5 min in extracellular solution containing 5% lucifer yellow, in the presence of Ca2þ or in Ca2þ-free medium containing 2 mM EGTA. After a brief washout of the dye with a Ca2þ-containing solution, cellular fluorescence was recorded using the laser scanning confocal microscope (LSM510; Carl Zeiss Meditec) by excitation at 488 nm with emission recorded at 530 nm. 2.9. Chemicals Fluo-4 AM (F14217) and 6 carboxyfluorescein diacetate (cat. no. C1362) were obtained from Molecular Probes (Eugene, OR). Suramin (S2671), apyrase VI (A6410), apyrase VII (A6535), ARL-67156 (6-N,N-diethyl-b,g-dibromomethylene-D-ATP; A265), ATP assay mixture (FLAAM), ATP, Triton X-100 (T-9284), rabbit serum (R-4505), 18-a

glycyrrhetinic acid and lucifer yellow were obtained from SigmaeAldrich. Paraformaldehyde (1.04005.1000) was obtained from Merck (Darmstadt, Germany). BSA (735 078) was obtained from Roche (Vilvoorde, Belgium). Monoclonal mouse anti-goat Cx43 antibody (C8093) was obtained from SigmaeAldrich (Deisenhoven, Germany), Alexa FluorÒ 488 labeled isotype specific secondary goat anti-mouse IgG antibody (A-21121) from Molecular Probes (Eugene, OR, USA) and monoclonal mouse-derived IgG1 antibody to Aspergillus niger glucose oxidase (X0931) from Dako (Glostrup, Denmark). Gap27 (SRPTEKTIFII), Gap26 (VCYDKSFPISHVR) and control (SRGGEKNVFIV) peptides were synthesized at the Laboratory of Biochemistry, KU Leuven. The peptides were analyzed by reverse-phase HPLC (Waters Corp., Milford, MA), on a C18-column (Phenomenex Luna 5u, 250  4.60 mm), using a linear gradient of acetonitrile-water, containing 0.06% TFA. The exact sequence of the peptide was confirmed by ESI-triple quadrupole mass spectrometry on an API-3000 mass spectrometer (PE-SCIEX, Applied Biosystems). The purity of the peptide was greater than 95%. All chemicals were applied in the extracellular solution. 2.10. Data analysis All data are given as mean  standard error of the mean (S.E.M.). Comparisons of means between groups were performed by unpaired t-tests. P < 0.05 was considered as a statistically significant difference. N indicates the number of independent experiments (the number of cells subjected to point mechanical stimulation), while n represents the total number of responsive cells. 3. Results 3.1. Expression of Cx43 Cx43 is a major connexin subtype expressed in many cell types and is well known to form gap junctions and hemichannels (Guo et al., 2003; Saez et al., 2003a). Furthermore, Cx43 hemichannels are also implicated in ATP release (Romanello and D’Andrea, 2001; Ebihara, 2003; Leybaert et al., 2003). Therefore, we investigated Cx43 expression in BCEC at protein and mRNA levels. Our findings from RTePCR and immunocytochemistry are shown in Fig. 2. The distinct band corresponding to 317 bp in Fig. 2A is the expected band size for Cx43. A BLAST analysis of the PCR product’s sequence indicated a primary nucleotide match to bovine Cx43 with sequence identity of 99% over 317 bp. Consistent with the expression of Cx43 mRNA, immunofluorescence staining of BCEC using a monoclonal antibody shows plaques of Cx43 localized at the cell boundaries (Fig. 2B). 3.2. Functional gap junctions Intercellular transfer of hydrophilic dyes (e.g., lucifer yellow and carboxyfluorescein) after scrape-loading or

P. Gomes et al. / Experimental Eye Research 83 (2006) 1225e1237

Fig. 2. Expression of Cx43 in BCEC. (A) RTePCR identification of Cx43 transcript in total RNA extracted from cultured cells. (B) Immunocytochemical localization of Cx43 in cultured cells. Note the predominant association of Cx43 plaques along the cell border. (C) negative control. Scale bar ¼ 100 mm.

photobleaching (as in the FRAP protocol) indicates functional GJIC (el-Fouly et al., 1987; Carruba et al., 2004). In scrapeloading experiments (N ¼ 10), extracellular lucifer yellow entered the cells at the border of the scratch and reached cells as far as 400 mm away from the scratch (Fig. 3A). In FRAP experiments, fluorescence recovery in the bleached cells reached 60e70% within 3 min after photobleaching. FRAP was also performed in the presence of the gap junction blocker Gap27, a connexin mimetic peptide that has a similar sequence to a highly conserved region of the second extracellular loop of Cx43 (Evans and Boitano, 2001). Pretreatment with Gap27 (300 mM) in the extracellular solution for 30 min reduced the fluorescence recovery by 19%. An inactive analog of the peptide (referred to as control peptide) which differs from Gap27 at three residues, was found to have no effect on the fluorescence recovery (Control: 64  2%, N ¼ 64; Control peptide: 63  4%, N ¼ 28;

1229

Fig. 3. Functional gap junction dye coupling in BCEC. (A) Dye transfer after scrape-loading. This is a typical experiment showing that extracellular lucifer yellow is taken up by the damaged cells at the border of the scratch (made by a 27-gauge needle, and seen running downwards from near the upper right corner of the picture), and is transferred to cells away from the scratch. The maximum distance reached by the dye exceeds 400 mm. (B) Summary of percent fluorescence recovery at 3 min in FRAP experiments with and without inhibition of Cx43 gap junctions by the selective connexin mimetic peptide, Gap27. Cells were exposed for 30 min to Gap27 (300 mM) or to an inactive Control peptide. The percent of the dye recovered (% Recovery) in the bleached cell was measured at 3 min after bleaching. Control conditions (black bar, N ¼ 64), Control peptide (white bar, N ¼ 28), or Gap27 (gray bar, N ¼ 37).

Gap27: 53  3%, N ¼ 37) (Fig. 3B). These results show the presence of dye-coupling via gap junctions and hence indicate functional GJIC in BCEC. 3.3. Inhibition of point mechanical stimulation-induced Ca2þ wave propagation by Gap27 Our previous studies demonstrated the important contribution of PIC in point mechanical stimulation-induced Ca2þ wave propagation in BCEC and suggested that GJIC could also play a role (Gomes et al., 2005b). In order to further investigate GJIC in BCEC, we examined Ca2þ wave

P. Gomes et al. / Experimental Eye Research 83 (2006) 1225e1237

1230

propagation in the presence of Gap27. Pretreatment with Gap27 (300 mM) for 30 min strongly reduced the spread of the Ca2þ wave (active area ¼ 26,936  2299 mm2, N ¼ 39 versus 73,689  2844 mm2, N ¼ 36 in control). Doubling the concentration of the peptide did not result in any further reduction of the active area (19,215  3089 mm2, N ¼ 21 at 600 mM versus 25,984  3587 mm2, N ¼ 20 at 300 mM). The peptide did not affect normalized fluorescence in the mechanically stimulated cells but significantly inhibited normalized fluorescence and percentage of responsive cells of neighboring cells. Application of 300 mM of the Control peptide did not significantly alter active area or percentage of responsive cells when compared to control conditions (Table 1). In addition to Gap27, we also investigated the influence of the gap junction blocker a-glycyrrhetinic acid (20 mM). Application of the drug for 45 min significantly reduced percentage of responsive cells and active area. In the presence of the drug the percentage responsive neighboring cells was reduced with respect to control conditions by 6% in neighboring layer NB1, by 14% in NB2, 26% in NB3, 30% in NB4 and 40% in NB5 (N ¼ 10). Heptanol (3.5 mM applied for 45 min) reduced the percentage responsive neighboring cells by 10%, 34% and 38% in NB1, NB2 and NB3 respectively compared to control condition (N ¼ 4). 3.4. Lack of effect of Gap27 on PIC In order to verify whether the effect of Gap27 on Ca2þ wave propagation is indeed through diminished GJIC and not via an effect on PIC, we performed a series of experiments to study the effects of Gap27 on the PIC pathway. In previous studies, we showed that exposure of the cells to the ectonucleotidase inhibitor ARL-67156 resulted in a threefold increase in active area of the Ca2þ wave evoked by point mechanical stimulation, via a marked enhancement of PIC (Gomes et al., 2005b). Therefore, we then investigated whether Gap27 could affect this PIC-mediated increase in

active area induced by ARL-67156. As shown in Fig. 4A, the active area of cells treated with a combination of Gap27 and 100 mM ARL-67156 (648,401  56,777 mm2, N ¼ 17) was not significantly different when compared to cells treated with ARL-67156 alone (593,440  33,528 mm2, N ¼ 14), providing evidence that Gap27 does not influence PIC-mediated Ca2þ waves. Treatment with Gap27 also did not affect the extracellular ATP level after point mechanical stimulation, as shown by a typical result in Fig. 4B. This indicates that the inhibition of the Ca2þ wave by Gap27 (Fig. 4A) is not due to an effect on ATP release. Since our previous study (Gomes et al., 2005a) had provided evidence for the involvement of hemichannels in ATP release, the current findings also indicate that Gap27 does not block Cx hemichannels involved in ATP release in BCEC. Since our previous studies provided evidence that uptake of lucifer yellow (applied in the extracellular solution) in the absence of extracellular Ca2þ is mediated via Cx hemichannels (Gomes et al., 2005a), we also used the lucifer yellow uptake assay to investigate whether Gap27 has any effect on the functional hemichannels. As shown in Fig. 5, BCEC did not take up lucifer yellow in the presence of Ca2þ (N ¼ 12), but a marked dye uptake is evident in Ca2þ-free medium in the presence of 2 mM EGTA (N ¼ 9). While uptake of lucifer yellow in Ca2þ-free EGTA-containing medium is inhibited by Gap26 (Gomes et al., 2005a), pretreatment with the gap junction blocker Gap27 did not affect the dye uptake (N ¼ 10). These data provide further evidence that Gap27 does not block the hemichannels in BCEC, consistent with our data showing that Gap27 does not affect the point mechanical stimulationinduced ATP release. Pretreatment of cells with Gap27 also did not significantly affect the [Ca2þ]i rise in response to exogenous ATP (data not shown), indicating that Gap27 does not affect purinergic receptors, their downstream signaling pathway or the intracellular stores.

Table 1 Average maximal normalized fluorescence (NF), percentage of responsive cells (%RC) and average active area (AA) in the mechanically stimulated (MS) cell and in neighboring cells layers 1 to 4 (NB1 to NB4) during point mechanical stimulation in control conditions and in cells treated with Control peptide (300 mM) or Gap27 (300 mM) MS

NB1

NB2

NB3

NB4

Active area (AA, mm2)

Control

NF S.E.M. n %RC

2.55 0.1 36 100

2.68 0.05 245 100

2.36 0.04 491 99

2.03 0.03 689 84

1.83 0.03 644 64

73,689  2844

Control peptide (300 mM)

NF S.E.M. n %RC

2.75 0.11 36 100

2.76 0.05 235 99

2.35 0.03 480 99

2.05 0.03 674 87

1.99 0.04 650 60

75,642  3359

Gap27 (300 mM)

NF S.E.M. n %RC

2.52 0.11 39 100

2.12* 0.04 252 90

1.76* 0.03 528 60

1.61* 0.03 739 25

1.48* 0.05 692 12

26,936  2299*

*P < 0.05 versus control.

P. Gomes et al. / Experimental Eye Research 83 (2006) 1225e1237

Fig. 4. (A) Gap27 does not influence PIC. Cells were treated with a combination of Gap27 and the ecto-ATPase inhibitor ARL-67156 (gray bar) and the active area (AA) was measured and compared to that of cells treated with ARL-67156 alone (white bar). (B) Treatment with Gap27 for 30 min does not influence point mechanical stimulation-induced ATP release: Gap27 (300 mM, 30 min), ARL-67156 (100 mM, 30 min). *P < 0.05 versus control.

Taken together, the above series of experiments with Gap27 provide evidence that the effect of Gap27 on Ca2þ wave in BCEC is not mediated by an effect on PIC. 3.5. Contributions of PIC and GJIC to intercellular Ca2þ wave propagation In order to assess the relative contributions of PIC and GJIC to the observed Ca2þ wave, we compared the effects of

1231

combinations of GJIC and PIC inhibitors with those obtained with either GJIC inhibition or PIC inhibition alone. Table 2 shows comparative data of point mechanical stimulation-induced Ca2þ waves when cells were treated with a combination of Gap27 and apyrase VI þ VII (N ¼ 30) against cells treated with Gap27 alone (N ¼ 27). The combined exposure results in a marked reduction of active area, percentage of responsive cells and normalized fluorescence of the responsive neighboring cells without affecting normalized fluorescence of mechanically stimulated cells, when compared to cells treated with Gap27 alone. Similarly, when cells are exposed to Gap27 together with suramin (N ¼ 22), active area of the Ca2þ wave and percentage of responsive cells of neighboring cells were also significantly lower than in cells treated with Gap27 alone (N ¼ 27), while no difference was observed in normalized fluorescence of mechanically stimulated cells (Table 2). These data suggest that combined inhibition of GJIC and PIC leads to a stronger reduction of the Ca2þ wave than GJIC inhibition alone, thus that the effect of PIC inhibition is at least partially additive to the effect of GJIC inhibition. We also compared the effects of combination of gap junction blocker (Gap27) and PIC inhibitors (apyrase VI þ VII, N ¼ 30) with those of PIC inhibitors alone (apyrase VI þ VII, N ¼ 21). Apyrases (VI þ VII) alone caused a pronounced inhibition of the Ca2þ wave, resulting in an active area of only 9% of the value observed in control conditions, demonstrating the major contribution of PIC towards the point mechanical stimulation-induced Ca2þ wave. Combining the apyrases with Gap27 resulted in a further small but significant decrease of active area when compared to cells treated with apyrases only, corresponding to a 33% reduction (Table 2). We also found that active area and percentage of responsive cells of BCEC cells treated with the combination of Gap27 and suramin (N ¼ 22) was significantly lower than in cells treated with suramin alone (N ¼ 11) (Table 2). Since in previous studies we demonstrated that a 30 min application of the Cx mimetic peptide Gap26 inhibits PIC by inhibiting ATP release involving Cx hemichannels, without having an effect on GJIC (Gomes et al., 2005a), we also

Fig. 5. Uptake of Lucifer yellow in Ca2þ-free solutions is not affected by Gap27. Cells were exposed to the fluorescent dye lucifer yellow for 5 min in the presence (N ¼ 12) and absence (N ¼ 9) of extracellular Ca2þ. The uptake of the dye in Ca2þ-free conditions (2 mM EGTA) is not affected in cells pre-treated for 30 min with 300 mM Gap27 (N ¼ 10).

1232

P. Gomes et al. / Experimental Eye Research 83 (2006) 1225e1237

Table 2 Average maximal normalized fluorescence (NF), percentage of responsive cells (%RC) and average active area (AA) in the mechanically stimulated (MS) cell and in neighboring cells layers 1 to 4 (NB1 to NB4) during mechanical stimulation in control conditions and in cells treated with Gap27, suramin, apyrase VI þ VII, a combination of Gap27 together with suramin, or a combination of Gap27 together with apyrase VI þ VII MS

NB1

NB2

NB3

NB4

Active area (AA, mm2)

Control

NF S.E.M. n %RC

2.82 0.13 24 100

2.62 0.06 165 99

2.15 0.04 294 93

1.83 0.04 338 68

1.56 0.05 271 38

66,855  4112

Gap27 (300 mM)

NF S.E.M. n %RC

2.5 0.08 27 100

2.02* 0.05 160 97

1.55* 0.03 312 66

1.32* 0.03 337 31

1.19* 0.01 240 15

27,514  3876*

Suramin (200 mM)

NF S.E.M. n %RC

2.67 0.17 11 100

2.34* 0.12 75 83

1.67* 0.08 122 41

1.28* 0.06 155 6

1.28 0.12 114 4

21,827  3345*

Apyrase VI þ VII (5 U/ml each)

NF S.E.M. n %RC

2.11* 0.1 21 100

1.88* 0.09 130 47

1.18* 0.06 262 2

1.14* 0.03 287 1

1.15* 0.02 196 2

5922  850*

Gap27 þ Suramin (300 mM þ 200 mM)

NF S.E.M. n %RC

2.67 0.19 22 100

2.06* 0.09 144 82

1.50* 0.06 247 31

1.21* 0.04 309 7

1.14* 0.01 216 5

10,686  1545*

Gap27 þ Apyrase VI þ VII (300 mM þ 5 U/ml each)

NF S.E.M. n %RC

2.45 0.11 30 100

1.60* 0.07 196 29

1.22* 0.06 356 2

1.13* 0.03 342 1

1.14* 0.00 177 1

3831  469*

Gap27 (300 mM), suramin (200 mM), apyrase VI þ VII (5 U/ml each). *P < 0.05 versus control.

studied the effects of combination of Gap26 and Gap27. Our results showed that the combination of Gap27 together with Gap26 (N ¼ 29) significantly decreased active area as well as percentage of responsive neighboring cells when compared to Gap27 (N ¼ 25) or Gap26 alone (N ¼ 24) (Fig. 6 and Table 3). These experiments with combinations of inhibitors of PIC and GJIC demonstrate that the effects of PIC and GJIC inhibition are at least partially additive. Furthermore, the effect of PIC inhibition in the presence of gap junction block is more pronounced than the effect of gap junction block in the presence of PIC inhibition.

4. Discussion Several studies on the mechanisms underlying Ca2þ wave propagation in cellular monolayers show that both PIC and GJIC may contribute towards IC (e.g., osteoblastic cells (Romanello and D’Andrea, 2001); thymic cells (Alves et al., 2000); glia (Zhang et al., 2003); alveolar epithelial cells (Isakson et al., 2001) and astrocytes (Hassinger et al., 1996; Braet et al., 2003b). In recent studies, we demonstrated the contribution of PIC in Ca2þ wave propagation induced by point mechanical stimulation in BCEC. Specifically, our studies established that ATP acts as a paracrine mediator for PIC

(Gomes et al., 2005b) and that the release of the nucleotide involves Cx hemichannels. In this study, we investigated the contribution and role of gap junctions in IC, and in particular the role GJIC plays in point mechanical stimulation-induced Ca2þ wave propagation. Our findings show that partially additive and interacting GJIC and PIC pathways promote the Ca2þ wave in BCEC. 4.1. Functional gap junctions in corneal endothelium The presence of functional gap junctions in BCEC is evident from scrape-loading (Fig. 3A) and FRAP experiments (Fig. 3B). The observed dye transfer in these experiments is indicative of GJIC, as is the dye transfer after a single cell microinjection of lucifer yellow in rabbit corneal endothelial cells (Williams and Watsky, 2004). We supplemented our findings with studies on expression of Cx43, which forms a major subtype of Cx in many cell types (Guo et al., 2003; Saez et al., 2003a) and is known to be expressed in corneal endothelium of rats (Joyce et al., 1998; Laux-Fenton et al., 2003), rabbits (Williams and Watsky, 2004), cows (Mohay and McLaughlin, 1995) and humans (Williams and Watsky, 2002). RTePCR showed Cx43 expression in BCEC at the mRNA level (Fig. 2A). Immunocytochemistry demonstrated abundant localization of Cx43 in regions of the plasma membrane

P. Gomes et al. / Experimental Eye Research 83 (2006) 1225e1237

1233

Fig. 6. The combined effect of the hemichannel blocker Gap26 and the GJIC blocker Gap27. Cells were treated with Gap26, Gap27 or a combination of Gap26 together with Gap27 for 30 min. The Ca2þ wave propagation in response to point mechanical stimulation is represented in control (black bars) (N ¼ 30), Gap27 (light gray bars) (N ¼ 25), and Gap26 (white bars) (N ¼ 24) and combination of Gap27 and Gap26 (dark gray bars) (N ¼ 29). Top: Bar graph showing maximal normalized fluorescence (NF) values for mechanically stimulated (MS) cells and neighboring cell layers 1 to 4 (NB1 to NB4); middle: Bar graph of the percentage of responsive cells (%RC) for the mechanically stimulated cell and NB1 to NB4; bottom: active area (AA) for the different experimental conditions. Gap27 (300 mM), Gap26 (300 mM). *P < 0.05 versus control. ^P < 0.05 versus Gap26 and versus Gap27.

bordering adjacent cells (Fig. 2B). Finally, Gap27, a connexin mimetic peptide with sequence homology to a part of the second extracellular loop of Cx43 that is known to block Cx43 gap junctions (Evans and Boitano, 2001), reduced the recovery of fluorescence in FRAP experiments (Fig. 3B). Taken together, these observations confirm that Cx43 gap junctions contribute to GJIC in BCEC. 4.2. GJIC-dependent Ca2þ wave in corneal endothelium Gap27 significantly inhibited the point mechanical stimulation-induced Ca2þ wave propagation, resulting in a marked reduction in active area (63%; Table 1), which points to an important role played by gap junctions. The reduction in active area by Gap27 is pronounced when compared to inhibition of

fluorescence recovery (19%) in the FRAP experiments (Fig. 3B). Although active area cannot be quantitatively related to gap-junctional conductance, it is a qualitative indicator of the efficacy of the Ca2þ wave. Thus, the marked decrease in active area by Gap27 in Table 1 suggests an important contribution of Cx43-mediated GJIC to the Ca2þ wave. Since other Gap27-insensitive Cx isoforms are likely to be present in BCEC, as is the case in many other cell types, the inhibition of the Ca2þ wave by Gap27 probably underestimates the overall contribution of GJIC in BCEC. Our previous study indicated that PIC partially mediates IC in BCEC and that Cx hemichannels are involved in point mechanical stimulation-induced ATP release. Braet et al. (2003a,b) provided evidence that Gap27 inhibits Cx43 hemichannels. With these observations in mind, we examined

1234

P. Gomes et al. / Experimental Eye Research 83 (2006) 1225e1237

Table 3 Average maximal normalized fluorescence (NF), percentage of responsive cells (%RC) and average active area (AA) in the mechanically stimulated (MS) cell and in neighboring cells layers 1 to 4 (NB1 to NB4) during mechanical stimulation in control conditions and in cells treated with Gap27, Gap26 and a combination of Gap27 and Gap26 MS

NB1

NB2

NB3

NB4

Active area (AA, mm2)

Control

NF S.E.M. n %RC

2.45 0.1 30 100

2.52 0.05 197 99

2.24 0.04 404 99

1.93 0.03 602 88

1.74 0.03 640 69

73,939  2759

Gap27 (300 mM)

NF S.E.M. n %RC

2.46 0.12 25 100

2.20* 0.06 163 90

1.80* 0.04 341 74

1.57* 0.03 502 40

1.46* 0.05 540 16

27,961  2624*

Gap26 (300 mM)

NF S.E.M. n %RC

2.46 0.14 24 100

2.38 0.06 164 94

2.02* 0.04 333 80

1.65* 0.03 488 48

1.50* 0.03 515 28

32,978  2567*

Gap27 þ Gap26 (300 mM each)

NF S.E.M. n %RC

2.24 0.12 29 100

1.97* 0.05 189 80

1.72* 0.04 403 48

1.56* 0.05 612 14

1.43* 0.05 649 7

18,134  2307*

Gap27 (300 mM), Gap26 (200 mM). *P < 0.05 versus control.

whether the peptide influences the PIC pathway, and, in particular, if it blocks the Cx hemichannels in BCEC. We found no effect of Gap27 on the magnitude of the Ca2þ transient evoked by the application of extracellular ATP (data not shown), which indicates Gap27 does not influence P2Y receptors, its downstream signaling pathways or the IP3-sensitive Ca2þ stores. Furthermore, we found that the ARL-induced enhancement in active area is unaffected by Gap27 (Fig. 4A). Since ARL exerts its effect on active area by inhibiting the degradation of extracellular ATP by ectoATPases (Gomes et al., 2005a), the lack of an effect of Gap27 on the ARL-induced enhancement in active area (Fig. 4A) indicates that ATP release is unperturbed by the peptide. Consistent with these findings, point mechanical stimulation-induced ATP release, measured using the luciferineluciferase protocol, was also found to be unaffected by Gap27 (Fig. 4B). Since the main pathway of the point mechanical stimulation-induced ATP release in BCEC is through hemichannels (Gomes et al., 2005a), the absence of an effect of Gap27 on the ARL-induced increase in active area also suggests that the peptide does not influence the hemichannels. Therefore, we conclude that the hemichannels involved in ATP release in BCEC are unaffected by the peptide, in contrast to the findings by Braet et al. (2003b) on rat brain endothelial cells and on Cx32 expressing cells (De Vuyst et al., 2006). To obtain additional evidence, we tested the effect of Gap27 on the lucifer yellow uptake in response to Ca2þ-free extracellular solution. Our previous studies showed that Gap26, a connexin mimetic peptide with a sequence identical to part of the highly conserved first extracellular loop of Cx43, completely blocked dye uptake (Gomes et al., 2005a). This demonstrates that, similar to point mechanical stimulation-induced ATP release, the lucifer yellow uptake is mediated via Cx hemichannels. Gap27 did not block

the dye uptake in Ca2þ-free extracellular solution, supporting the conclusion that the Cx hemichannels involved in point mechanical stimulation-induced ATP release are not inhibited by Gap27. Our findings thus exclude that the effect of Gap27 on the Ca2þ wave propagation in BCEC is mediated via the Cx hemichannels involved in ATP release. We therefore conclude that in BCEC, Cx-mediated GJIC contributes towards the point mechanical stimulation-induced Ca2þ wave propagation. 4.3. Additivity of GJIC and PIC to Ca2þ wave propagation Since our experiments demonstrated a contribution of GJIC to the Ca2þ wave and our previous studies (Gomes et al., 2005a) also showed that PIC contributes to the Ca2þ wave, we investigated whether the two effects are additive. The combination of Gap27 and apyrases induced a significant further reduction in active area as compared to Gap27 or apyrases alone. This finding indicates that the inhibitors of PIC and GJIC act at least partially via separate mechanisms. Similar conclusions can be drawn from experiments in which suramin was used to inhibit PIC (Table 2). We also performed experiments to investigate the effect of combining Gap27 and Gap26 (Fig. 6). The results also suggest that the effects of Gap27 and Gap26 are partially additive. Since we have previously shown that in the conditions of our experiments Gap26 inhibits ATP release by selectively blocking hemichannels without affecting gap junctions or P2Y signaling, and since Gap27 does not influence the PIC pathway, as discussed earlier, the significant reduction of the active area by the combination of Gap26 and Gap27 provides further evidence for partial additivity of the PIC and GJIC pathways.

P. Gomes et al. / Experimental Eye Research 83 (2006) 1225e1237

4.4. Interaction between GJIC and PIC to Ca2þ wave propagation The results shown in Table 2 demonstrate that the Ca2þ wave is strongly reduced by blocking the PIC pathway with apyrases, causing a reduction of the active area to less than 10% of the control value. This suggests that PIC is by far the major IC pathway and that GJIC hardly contributes to the Ca2þ wave. On the other hand, the results with Gap27, which inhibits gap junctions formed by Cx43 and does not affect the PIC pathway in BCEC, demonstrate a reduction of the active area as large as 63%, suggesting a major contribution of GJIC. Although comparison of the active area cannot be used as a quantitative estimate of contribution to IC, these apparently conflicting results indicate that PIC and GJIC are not independent in promoting Ca2þ wave propagation, but that they interact with each other by sharing parts of their downstream pathways. Such interactions between the PIC and GJIC pathways have also been recently described in other tissues (Moorby and Patel, 2001; Plotkin and Bellido, 2001; Oviedo-Orta and Evans, 2002; Plotkin et al., 2002; Krysko et al., 2004; Stuhlmann et al., 2004; Gittens et al., 2005). Although our data do not provide a clear-cut answer to the pathways of interactions, it is evident that the signal transduction pathways involved in the rise in [Ca2þ]i by PIC and GJIC share some second messengers. Ca2þ wave propagation promoted by GJIC is dependent on the availability of the second messenger IP3, well known to permeate gap junctions (Leybaert et al., 1998). IP3 also forms downstream of activated P2Y receptors, since these receptors are coupled to Gaq/11 proteins. Blocking gap junctions limits the flux of IP3 and/or Ca2þ from the mechanically stimulated cell into adjacent neighboring cells upon point mechanical stimulation, thus leading to a reduced [Ca2þ]i rise in neighboring cells. If [Ca2þ]i acts as the signal for ATP release, as has been demonstrated in astrocytes (Braet et al., 2003a,b), the reduced Ca2þ rise in neighboring cells would also reduce ATP release from neighboring cells. Thus, blocking gap junctions eventually reduces the extent of wave propagation via the PIC pathway. Blocking the PIC pathway would result in reducing P2Y receptor activation and IP3 release in neighboring cells, which would also result in a smaller Ca2þ transient in neighboring cells and thereby reduce gap junction-mediated Ca2þ wave propagation. An eventual autocrine effect of ATP, which could be due to ATP release because of the rise in [Ca2þ]i (Braet et al., 2003a), could also be expected to result in an increase of the [Ca2þ]i transient in neighboring cells, and this effect would be absent when PIC is blocked. Except in the presence of apyrases, we did not find a significant effect of the PIC and GJIC blockers on normalized fluorescence in the mechanically stimulated cell. The finding of a significant reduction in normalized fluorescence of the mechanically stimulated cell in the presence of apyrases suggests an autocrine effect. This is because the [Ca2þ]i transient in the mechanically stimulated cell is likely to be enhanced by ATP, released upon mechanical stimulation, via P2Y receptormediated IP3 release in the mechanically stimulated cell.

1235

Furthermore, if ATP release is initiated by rise in [Ca2þ]i (Leybaert et al., 2003; De Vuyst et al., 2006), GJIC as well as PIC could contribute to an enhancement of normalized fluorescence in mechanically stimulated cells via ATP released by the neighboring cells. Both mechanisms could explain the reduction of normalized fluorescence in the mechanically stimulated cell by apyrases. However, the decremental nature of the intercellular Ca2þ wave in BCEC suggests either the absence of an ATP-induced release of ATP (in line with the findings of Simard et al., 2003 in glial cells) or at least that such an effect has limited efficacy. When apyrases and Gap27 are combined, we see no inhibition in normalized fluorescence of the mechanically stimulated cell versus control. The combination of Gap27 with apyrases tends to significantly enhance normalized fluorescence in the mechanically stimulated cell as compared to apyrases alone, which is likely to be due to limited loss of IP3 and/or Ca2þ to neighboring cells via gap junctions by Gap27.

4.5. Physiological significance in corneal endothelium The corneal endothelium is non-regenerative in the human cornea (Joyce, 2003). Deadhesion, shape changes, and loss of functional integrity of the corneal endothelium are documented in response to mechanical injury (e.g., during intraocular surgery such as phacoemulsification), aging, Fuch’s dystrophy, inflammation, and hypoxia (Edelhauser, 2000; Bonanno, 2003; Bourne, 2003; Bourne and McLaren, 2004). As demonstrated in many tissues, the response of a monolayer to extracellular stimuli is coordinated via IC, which can be mediated through a rapid intercellular transfer of intracellular messengers or by means of secretion of paracrine factors. In addition, IC is also known to be involved in regulation of cell growth and cell death (Abbracchio and Burnstock, 1998). Exchange of molecules via gap junctions can protect neighboring cells from cytotoxic substances. This effect has been called ‘‘metabolic cooperation’’ or ‘‘kiss of life’’. On the other hand, such exchange of molecules may also play a negative role by spreading factors that inhibit cell proliferation or by spreading focal toxic/apoptotic stimuli to adjacent cells, and the resulting spread of the stimuli has been termed the ‘‘bystander effect’’ or ‘‘kiss of death’’ (Andrade-Rozental et al., 2000; Chipman et al., 2003; Krysko et al., 2004; Mitchell et al., 2004). Cx-mediated GJIC has been demonstrated to be involved in the bystander effect (Fick et al., 1995; Vrionis et al., 1997; Krysko et al., 2004). Also, PIC plays an important role in such processes as cell growth, differentiation and migration. For example, propagation of injury-induced Ca2þ waves in corneal epithelial cells was demonstrated to occur via extracellular release of ATP acting on P2Y receptors, which lead to changes in cell migration and proliferation (Klepeis et al., 2001, 2004). These observations, taken together, suggest a specific role for GJIC and PIC in regulation of the cell cycle. Therefore, our present findings implicate roles for both PIC and GJIC in corneal endothelial pathophysiology.

1236

P. Gomes et al. / Experimental Eye Research 83 (2006) 1225e1237

Acknowledgments The authors wish to thank C. Beyers for his contribution in the immunocytochemistry experiments and for his help in providing the corneas, Raf Ponsaerts and Catheleyne D’hondt for helpful discussions and help in preparing the manuscript, and Professor W. Van Driessche en E. Larivie`re for help with the extracellular ATP assay. Supported by NIH grant EY14415 (S.P.S.), FWO-Vlaanderen G.0218.03, GOA/2004/07, IAP program 5/05 (B.H. and J.V.), IRO KU Leuven (P.G.). References Abbracchio, M.P., Burnstock, G., 1998. Purinergic signalling: pathophysiological roles. Jpn. J. Pharmacol. 78, 113e145. Alves, L.A., Nihei, O.K., Fonseca, P.C., Carvalho, A.C., Savino, W., 2000. Gap junction modulation by extracellular signaling molecules: the thymus model. Braz. J. Med. Biol. Res. 33, 457e465. Andrade-Rozental, A.F., Rozental, R., Hopperstad, M.G., Wu, J.K., Vrionis, F.D., Spray, D.C., 2000. Gap junctions: the ‘‘kiss of death’’ and the ‘‘kiss of life’’. Brain Res. Brain Res. Rev. 32, 308e315. Bonanno, J.A., 2003. Identity and regulation of ion transport mechanisms in the corneal endothelium. Prog. Retin. Eye Res. 22, 69e94. Bourne, W.M., 2003. Biology of the corneal endothelium in health and disease. Eye 17, 912e918. Bourne, W.M., McLaren, J.W., 2004. Clinical responses of the corneal endothelium. Exp. Eye Res. 78, 561e572. Braet, K., Aspeslagh, S., Vandamme, W., Willecke, K., Martin, P.E., Evans, W.H., Leybaert, L., 2003a. Pharmacological sensitivity of ATP release triggered by photoliberation of inositol-1,4,5-trisphosphate and zero extracellular calcium in brain endothelial cells. J. Cell Physiol. 197, 205e213. Braet, K., Vandamme, W., Martin, P.E., Evans, W.H., Leybaert, L., 2003b. Photoliberating inositol-1,4,5-trisphosphate triggers ATP release that is blocked by the connexin mimetic peptide gap 26. Cell Calcium 33, 37e48. Carruba, G., Cocciadiferro, L., Bellavia, V., Rizzo, S., Tsatsanis, C., Spandidos, D., Muti, P., Smith, C., Mehta, P., Castagnetta, L., 2004. Intercellular communication and human hepatocellular carcinoma. Ann. N.Y. Acad. Sci. 1028, 202e212. Chipman, J.K., Mally, A., Edwards, G.O., 2003. Disruption of gap junctions in toxicity and carcinogenicity. Toxicol. Sci. 71, 146e153. Cotrina, M.L., Lin, J.H., Alves-Rodrigues, A., Liu, S., Li, J., AzmiGhadimi, H., Kang, J., Naus, C.C., Nedergaard, M., 1998. Connexins regulate calcium signaling by controlling ATP release. Proc. Natl. Acad. Sci. USA 95, 15735e15740. Cotrina, M.L., Lin, J.H., Lopez-Garcia, J.C., Naus, C.C., Nedergaard, M., 2000. ATP-mediated glia signaling. J. Neurosci. 20, 2835e2844. De Vuyst, E., Decrock, E., Cabooter, L., Dubyak, G.R., Naus, C.C., Evans, W.H., Leybaert, L., 2006. Intracellular calcium changes trigger connexin 32 hemichannel opening. EMBO J. 25, 34e44. Dikstein, S., Maurice, D.M., 1972. The metabolic basis to the fluid pump in the cornea. J. Physiol. 221, 29e41. Ebihara, L., 2003. New roles for connexons. News Physiol. Sci. 18, 100e103. Edelhauser, H.F., 2000. The resiliency of the corneal endothelium to refractive and intraocular surgery. Cornea 19, 263e273. el-Fouly, M.H., Trosko, J.E., Chang, C.C., 1987. Scrape-loading and dye transfer. A rapid and simple technique to study gap junctional intercellular communication. Exp. Cell Res. 168, 422e430. Evans, W.H., Boitano, S., 2001. Connexin mimetic peptides: specific inhibitors of gap-junctional intercellular communication. Biochem. Soc. Trans. 29, 606e612. Fick, J., Barker 2nd, F.G., Dazin, P., Westphale, E.M., Beyer, E.C., Israel, M.A., 1995. The extent of heterocellular communication mediated by gap junctions is predictive of bystander tumor cytotoxicity in vitro. Proc. Natl. Acad. Sci. USA 92, 11071e11075.

Gittens, J.E., Barr, K.J., Vanderhyden, B.C., Kidder, G.M., 2005. Interplay between paracrine signaling and gap junctional communication in ovarian follicles. J. Cell Sci. 118, 113e122. Gomes, P., Srinivas, S.P., Van Driessche, W., Vereecke, J., Himpens, B., 2005a. ATP release through connexin hemichannels in corneal endothelial cells. Invest. Ophthalmol. Vis. Sci. 46, 1208e1218. Gomes, P., Srinivas, S.P., Vereecke, J., Himpens, B., 2005b. ATP-dependent paracrine intercellular communication in cultured bovine corneal endothelial cells. Invest. Ophthalmol. Vis. Sci. 46, 104e113. Goodenough, D.A., Paul, D.L., 2003. Beyond the gap: functions of unpaired connexon channels. Nat. Rev. Mol. Cell. Biol. 4, 285e294. Guo, Y., Martinez-Williams, C., Rannels, D.E., 2003. Gap junction-microtubule associations in rat alveolar epithelial cells. Am. J. Physiol. Lung Cell Mol. Physiol. 285, L1213eL1221. Hassinger, T.D., Guthrie, P.B., Atkinson, P.B., Bennett, M.V., Kater, S.B., 1996. An extracellular signaling component in propagation of astrocytic calcium waves. Proc. Natl. Acad. Sci. USA 93, 13268e13273. Isakson, B.E., Evans, W.H., Boitano, S., 2001. Intercellular Ca2þ signaling in alveolar epithelial cells through gap junctions and by extracellular ATP. Am. J. Physiol. Lung Cell Mol. Physiol. 280, L221eL228. Joyce, N.C., 2003. Proliferative capacity of the corneal endothelium. Prog. Retin. Eye Res. 22, 359e389. Joyce, N.C., Harris, D.L., Zieske, J.D., 1998. Mitotic inhibition of corneal endothelium in neonatal rats. Invest. Ophthalmol. Vis. Sci. 39, 2572e2583. Klepeis, V.E., Cornell-Bell, A., Trinkaus-Randall, V., 2001. Growth factors but not gap junctions play a role in injury-induced Ca2þ waves in epithelial cells. J. Cell Sci. 114, 4185e4195. Klepeis, V.E., Weinger, I., Kaczmarek, E., Trinkaus-Randall, V., 2004. P2Y receptors play a critical role in epithelial cell communication and migration. J. Cell Biochem. 93, 1115e1133. Krysko, D.V., Mussche, S., Leybaert, L., D’Herde, K., 2004. Gap junctional communication and connexin43 expression in relation to apoptotic cell death and survival of granulosa cells. J. Histochem. Cytochem. 52, 1199e1207. Laux-Fenton, W.T., Donaldson, P.J., Kistler, J., Green, C.R., 2003. Connexin expression patterns in the rat cornea: molecular evidence for communication compartments. Cornea 22, 457e464. Leybaert, L., Paemeleire, K., Strahonja, A., Sanderson, M.J., 1998. Inositol-trisphosphate-dependent intercellular calcium signaling in and between astrocytes and endothelial cells. Glia 24, 398e407. Leybaert, L., Braet, K., Vandamme, W., Cabooter, L., Martin, P.E., Evans, W.H., 2003. Connexin channels, connexin mimetic peptides and ATP release. Cell Commun. Adhes. 10, 251e257. Maurice, D.M., 1972. The location of the fluid pump in the cornea. J. Physiol. 221, 43e54. Mitchell, S.A., Randers-Pehrson, G., Brenner, D.J., Hall, E.J., 2004. The bystander response in C3H 10T1/2 cells: the influence of cell-to-cell contact. Radiat. Res. 161, 397e401. Moerenhout, M., Himpens, B., Vereecke, J., 2001. Intercellular communication upon mechanical stimulation of CPAE- endothelial cells is mediated by nucleotides. Cell Calcium 29, 125e136. Mohay, J., McLaughlin, B.J., 1995. Corneal endothelial wound repair in normal and mitotically inhibited cultures. Graefes Arch. Clin. Exp. Ophthalmol. 233, 727e736. Moorby, C., Patel, M., 2001. Dual functions for connexins: Cx43 regulates growth independently of gap junction formation. Exp. Cell Res. 271, 238e248. Oviedo-Orta, E., Evans, W.H., 2002. Gap junctions and connexins: potential contributors to the immunological synapse. J. Leukoc. Biol. 72, 636e642. Plotkin, L.I., Bellido, T., 2001. Bisphosphonate-induced, hemichannelmediated, anti-apoptosis through the Src/ERK pathway: a gap junctionindependent action of connexin43. Cell Commun. Adhes. 8, 377e382. Plotkin, L.I., Manolagas, S.C., Bellido, T., 2002. Transduction of cell survival signals by connexin-43 hemichannels. J. Biol. Chem. 277, 8648e8657. Riley, M.V., Winkler, B.S., Starnes, C.A., Peters, M.I., Dang, L., 1998. Regulation of corneal endothelial barrier function by adenosine, cyclic AMP, and protein kinases. Invest. Ophthalmol. Vis. Sci. 39, 2076e2084. Romanello, M., D’Andrea, P., 2001. Dual mechanism of intercellular communication in HOBIT osteoblastic cells: a role for gap-junctional hemichannels. J. Bone Miner. Res. 16, 1465e1476.

P. Gomes et al. / Experimental Eye Research 83 (2006) 1225e1237 Saez, J.C., Berthoud, V.M., Branes, M.C., Martinez, A.D., Beyer, E.C., 2003a. Plasma membrane channels formed by connexins: their regulation and functions. Physiol. Rev. 83, 1359e1400. Saez, J.C., Contreras, J.E., Bukauskas, F.F., Retamal, M.A., Bennett, M.V., 2003b. Gap junction hemichannels in astrocytes of the CNS. Acta Physiol. Scand. 179, 9e22. Sanderson, M.J., Charles, A.C., Boitano, S., Dirksen, E.R., 1994. Mechanisms and function of intercellular calcium signaling. Mol. Cell. Endocrinol. 98, 173e187. Simard, M., Arcuino, G., Takano, T., Liu, Q.S., Nedergaard, M., 2003. Signaling at the gliovascular interface. J. Neurosci. 23, 9254e9262. Srinivas, S.P., Yeh, J.C., Ong, A., Bonanno, J.A., 1998. Ca2þ mobilization in bovine corneal endothelial cells by P2 purinergic receptors. Curr. Eye Res. 17, 994e1004. Srinivas, S.P., Ong, A., Goon, L., Bonanno, J.A., 2002. Lysosomal Ca2þ stores in bovine corneal endothelium. Invest. Ophthalmol. Vis. Sci. 43, 2341e2350.

1237

Stuhlmann, D., Steinbrenner, H., Wendlandt, B., Mitic, D., Sies, H., Brenneisen, P., 2004. Paracrine effect of TGF-beta1 on downregulation of gap junctional intercellular communication between human dermal fibroblasts. Biochem. Biophys. Res. Commun. 319, 321e326. Vrionis, F.D., Wu, J.K., Qi, P., Waltzman, M., Cherington, V., Spray, D.C., 1997. The bystander effect exerted by tumor cells expressing the herpes simplex virus thymidine kinase (HSVtk) gene is dependent on connexin expression and cell communication via gap junctions. Gene Ther. 4, 577e585. Williams, K., Watsky, M., 2002. Gap junctional communication in the human corneal endothelium and epithelium. Curr. Eye Res. 25, 29e36. Williams, K.K., Watsky, M.A., 2004. Bicarbonate promotes dye coupling in the epithelium and endothelium of the rabbit cornea. Curr. Eye Res. 28, 109e120. Zhang, W., Segura, B.J., Lin, T.R., Hu, Y., Mulholland, M.W., 2003. Intercellular calcium waves in cultured enteric glia from neonatal guinea pig. Glia 42, 252e262.