Upregulation of Extracellular ATP-Induced Müller Cell Responses in a Dispase Model of Proliferative Vitreoretinopathy

Abstract

purpose. To test whether in an animal model of proliferative vitreoretinopathy (PVR) the Müller glial cells displayed an upregulation of purinergic P2 receptor–mediated responses.

methods. PVR was induced by intravitreal injection of the proteolytic enzyme, dispase, in the eyes of adult rabbits. The developing PVR was examined ophthalmoscopically. After 3 weeks, small retinal pieces were wholemounted and used for calcium imaging, freshly dissociated Müller cells were subjected to calcium imaging, and patch-clamp recordings were made. The presence of P2 receptor–mediated Ca²⁺ responses was determined both directly—that is, fluorometrically, and indirectly, by electrophysiological recording of Ca²⁺-activated K⁺ currents.

results. According to earlier observations in another model of retinal detachment and PVR, the reactive Müller cells displayed hypertrophy, downregulation of inwardly rectifying K⁺ currents, and depolarization of the resting membrane potential, all dependent on the severity of the PVR. Further, significant PVR-induced increase was observed in the number of Müller cells responding to adenosine 5′-triphosphate (ATP), with a transient elevation of their[ Ca²⁺]_i. If isolated Müller cells were exposed to ATP, 13% of the control cells, but 29% (moderate PVR) or 53% (massive PVR) of the reactive cells, showed fluorometric Ca²⁺ increases. An increase of Ca²⁺-activated K⁺ currents was measured in 11% of the control cells, but in 83% (moderate PVR) and 90% (massive PVR) of the reactive cells. Confocal images of retinal wholemounts revealed similar results. Because similar responses were elicited by uridine triphosphate (UTP), the dominant involvement of metabotropic (P2Y type) purinergic receptors is suggested.

conclusions. An upregulation of purinergic receptors is part of the reactive changes of Müller cells during PVR. It is suggested that ATP-evoked Ca²⁺ responses may support the proliferation of Müller cells during PVR.

Müller (radial glial) cells are the dominant glial cells in the retina of all vertebrates, where they are involved in a wealth of crucial glioneuronal interactions, in both physiological and pathologic conditions. ^{1 2} During reactive gliosis, glial cells undergo functional changes that are accompanied by distinct changes in their membrane conductances. Normal Müller cells express predominantly inwardly rectifying K⁺ (Kir) channels in their plasma membranes. ^{3 4} The Kir channels are crucially involved in the maintenance of several retinal homeostasis mechanisms—for example, in the spatial buffering of the extracellular K⁺ concentration and, through the maintenance of a hyperpolarized membrane potential, in the electrogenic neurotransmitter uptake. ^{1 2} During reactive gliosis (e.g., after a retinal detachment) or during massive proliferative gliosis (e.g., during proliferative vitreoretinopathy [PVR]), the Kir currents of Müller cells are partly or fully extinguished. ^{5 6 7 8} The reduction in Kir currents is accompanied by a membrane depolarization that causes an increase of the open probability of depolarization-activated channels (e.g., of Ca²⁺-activated K⁺ channels of big conductance [BK]). ⁶  

Activation of purinergic P2 receptors has been suggested to be involved in the induction or maintenance of gliosis. ⁹ In the rat brain, an activation of P2 receptors by purinergic agonists has been shown to induce astrogliosis involving an increase of the immunoreactivity of intermediate filaments, cellular hypertrophy, and proliferation of astrocytes. ^{10 11} Adenosine 5′-triphosphate (ATP) is a transmitter within the retina, ¹² where it contributes to neuronal information processing. ¹³ The expression of purinergic P2 receptors by Müller cells of different species has been described using different techniques. The presence of mRNA for several distinct ionotropic P2X receptors has been shown in Müller cells of the rat retina. ¹⁴ In isolated salamander Müller cells, activation of P2 receptors by extracellular ATP stimulates the release of Ca²⁺ from intracellular stores. ¹⁵ Müller cells of the human retina have been shown to express ionotropic P2X₇ receptors. ¹⁶ The currents through the P2X₇ receptor channels are significantly elevated in cells from patients with PVR compared with cells from healthy donors. ¹⁷ Moreover, the increase of P2X₇ receptor currents was found to be correlated with other disease-induced alterations of membrane properties, such as the reduction of Kir currents ¹⁷ (i.e., with the severity of gliosis in human PVR). To determine whether an increased responsiveness to extracellular ATP is a general phenomenon in Müller cell gliosis, we studied Müller cell responses to extracellular ATP in an established animal model of PVR. ¹⁸  

Methods

Surgical Procedure

All experiments were performed in accordance with the applicable German laws and with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. To ensure the development of PVR and for a better comparison with a previous model of retinal detachment ⁸ we used higher doses of dispase, as recommended by Frenzel et al., ¹⁸ and created additionally a retinal detachment. Nine adult pigmented rabbits (2–3 kg, both sexes) were anesthetized by an intramuscular application of a mixture of ketamine hydrochloride (50 mg/mL, 1 mL/kg body weight; Ratiopharm, Ulm, Germany) and xylazine hydrochloride (20 mg/mL, 0.15 mL/kg body weight; BayerVital, Leverkusen, Germany), and the pupils of the right eyes were dilated with a topical application of 1% tropicamide (Ursapharm, Saarbrücken, Germany) and 5% phenylephrine hydrochloride (Ankerpharm, Rudolstadt, Germany). The eyes were protruded and immobilized. After pars plana sclerotomy, a circumscript vitrectomy was performed in the area of the future detachment (i.e., in the ventronasal quadrant, just below the medullary rays). A thin glass micropipette attached to a 250-μL glass syringe (Hamilton, Reno NV) was used to create a small local retinal detachment by injecting phosphate-buffered saline (PBS) into the subretinal space. Another thin glass micropipette was placed into the vitreous near the surface of the detached retina, to inject 100 μL of the proteolytic enzyme, dispase I (0.5 U in PBS, pH 7.4; Roche Molecular Biochemicals, Mannheim, Germany). The sclerotomies and the overlying conjunctiva were then closed. The left eyes served as the control. 

The rabbits were regularly examined ophthalmically to document the development and severity of PVR. After 3 weeks, animals anesthetized as described earlier were killed by an intravenous application of 3 mL T61 (0.2 g/mL embutramide, 0.05 g/mL mebezonium iodide, 5 mg/mL tetracaine hydrochloride; Hoechst, Unterschleissheim, Germany), and both the treated and the control eyes were excised. 

Müller Cell Isolation

All experiments (with the exception of the Ca²⁺-imaging experiments on wholemounted retinal pieces) were performed on acutely dissociated, noncultured Müller cells. ¹⁹ The retinas were isolated from the excised eye balls (corresponding areas were used in control and PVR-affected eyes), and incubated in papain (0.2 mg/mL; Roche Molecular Biochemicals), containing Ca²⁺- and Mg²⁺-free PBS for 30 minutes at 37°C, followed by several washing steps with PBS. After a short incubation in PBS supplemented with DNase I (200 U/mL; Sigma, Deisenhofen, Germany) the tissue pieces were titrated by a wide-pore pipette to obtain suspensions of isolated cells. The cells were stored at 4°C to 8°C in serum-free modified Eagle’s medium until use within 10 hours after isolation. 

Ca²⁺ Imaging: Retinal Wholemounts

For the experiments on retinal wholemounts (placed with their vitread surface up), the Ca²⁺ indicator fluo-4/acetomethoxyester (fluo-4/AM;Molecular Probes, Eugene, OR) was used. The dye was dissolved in dimethyl sulfoxide (DMSO) plus 50 μg fluo-4/AM in 50 μL DMSO plus 5 μL 20% nonionic detergent (Pluronic F-127; Biotium, Hayward, CA). The retinas were incubated for 45 minutes at room temperature in 400 μL of the bath solution (described later) to which 5 μL dissolved fluo-4/AM had been added. All experiments were performed at room temperature. Retinal pieces (5 × 5 mm) were placed and mechanically fixed in a perfusion chamber and perfused with a bath solution containing (in millimolar): 110 NaCl, 3 KCl, 2 CaCl₂, 1 MgCl₂, 1 Na₂HPO₄, 10 HEPES, 11 glucose, and 25 NaHCO₃, adjusted to pH 7.4 with tris (hydroxy-methyl) aminomethane (Tris-base). The bath solution was bubbled with carbogen (95% O₂-5% CO₂). Fluo-4/AM was excited with an argon laser at 488 nm, and confocal fluorescence images were collected with a 505-nm long-pass emission filter, using a laser scanning microscope (model 510; Carl Zeiss, Oberkochen, Germany). 

Ca²⁺ Imaging: Isolated Cells

For fluorescence measurements on acutely isolated cells, cells were loaded with fura-2/AM (10 μM AM; Molecular Probes) for 30 minutes at 37°C. Measurements were made at room temperature by using a bath solution containing (in millimolar): 129 NaCl, 3 KCl, 1 CaCl₂, 0.2 MgCl₂, 20 glucose, and 10 HEPES (pH 7.4 adjusted with NaOH). A fluorescence measurement system (Fucal 5.12B; Till-Photonics, Munich, Germany) was used. Fluorescence was excited at 340 nm (F₃₄₀) and 380 nm (F₃₈₀), and the ratio was calculated. Images were recorded every 6 seconds during the application of the test substances. 

Electrophysiological Recordings

To document the electrophysiological membrane properties of the cells, records were made in the whole-cell configuration of the patch-clamp technique. ²⁰ Isolated cells were pipetted into the recording chamber. The chamber was continuously perfused with bath solution, and test substances were added by fast changes of the perfusate. Patch pipettes were pulled from borosilicate glass and had resistances between 3 and 5 MΩ. Pipettes were used uncoated and without fire polishing. Seal resistances of 5 to 10 GΩ were obtained after slight suction was applied to the interior of the pipette. Voltage-clamp recordings were performed at room temperature (22–25°C) using patch-clamp amplifiers (EPC 7; List, Darmstadt, Germany; RK-400; Biological, Claix, France; Axopatch 200A; Axon Instruments, Foster City, CA) and software (TIDA 5.72; Heka Elektronik, Lambrecht, Germany; or ISO-2; MFK-Computer, Niedernhausen, Germany). The signals were low-pass filtered at 3 to 4 kHz (eight-pole Bessel filter); the sampling rate was 5 to 40 kHz. The series resistance was compensated as much as possible (30%–50%). Only recordings with a series resistance below 25 MΩ were accepted. The traces were not leak subtracted. Data were not corrected for liquid-junction potentials, because these did not exceed 3 mV. The membrane capacitance of the cells was measured by the integral of the uncompensated capacitive artifact evoked by a hyperpolarizing voltage step from −80 to −90 mV when Ba²⁺ ions (1 mM) were present in the bath solution to block the K⁺ conductance. For recordings of the capacitive artifact, the sampling rate was 30 kHz, and frequencies above 10 kHz were cut off. After establishing the whole-cell configuration, control currents were recorded for at least 3 minutes to be sure that these currents were stable (six cells from retinas with PVR were rejected from further investigation, because they already showed spontaneous alterations of the BK current amplitude under control conditions). 

The whole-cell currents were elicited by a standard step protocol (holding potential −80 mV, de- and hyperpolarizing voltage steps of 250 msec, with an increment of 10 or 20 mV). To investigate ATP-evoked responses, whole-cell currents were evoked with a continuous stimulation protocol. The holding potential was 0 mV, to minimize the activation of voltage-gated K⁺ currents and to reduce the space clamp problems during the evocation of BK currents. Alternating voltage steps of 50 msec to +120 mV and −120 mV were applied at a frequency of 2.5 Hz. 

To investigate the membrane currents of unstimulated cells, the bath solution contained (in millimolar): 110 NaCl, 3 KCl, 2 CaCl₂, 1 MgCl₂, 10 HEPES, and 11 glucose. The pH was adjusted to 7.4 by Tris-base. The pipette solution (intracellular) contained (in millimolar): 10 NaCl, 130 KCl, 1 CaCl₂, 2 MgCl₂, 10 EGTA, and 10 HEPES, adjusted to pH 7.1 with Tris-base. This composition resulted in a stable intracellular Ca²⁺ concentration ([Ca²⁺]_i) of approximately 20 nM. For measurements of the 2′-3′-O-(4-benzoylbenzoyl)-ATP (BzATP)–evoked currents, the following solutions were used (in millimolar): pipette solution: 10 NaCl, 130 CsCl, 1 CaCl₂, 2 MgCl₂, 10 EGTA, and 10 HEPES, adjusted to pH 7.1 with Tris-base; extracellular: 116 NaCl, 10 HEPES, and 11 glucose, pH 7.4 adjusted with Tris-base. To record ATP-evoked effects on the whole-cell currents, another pipette solution was used that allowed changes of the intracellular Ca²⁺ concentration. This solution contained (mM): 10 NaCl, 130 KCl, 3 MgCl₂, 0.1 EGTA, and 10 HEPES, pH 7.1 adjusted with Tris-base. ATP (Serva Electrophoresis, Heidelberg, Germany) and uridine 5′-triphosphate (UTP; Sigma) were used as sodium salts. Iberiotoxin was obtained from Alomone Laboratories (Jerusalem, Israel). Adenosine hemisulfate and all other substances were purchased from Sigma. 

Data Analysis

The steady state whole-cell currents were measured at the end of the 250-msec voltage steps. In continuous recordings, the currents were measured at the end of the 50-msec voltage steps. The membrane potentials were determined by measuring the zero-current potentials of the steady state current–voltage curves. Current densities were calculated by dividing current amplitudes through membrane capacitances. For the experiments with fura-2/AM, the fluorescence ratio F₃₄₀-F₃₈₀ is presented to describe relative changes in the[ Ca²⁺]_i. An increase of the ratio indicates an increase in[ Ca²⁺]_i and the[ Ca²⁺]_i can be estimated by a calibration method. ²¹ Increases of the ratio values of more than six times the basal noise were accepted as a response of the cell to the agonist application. Müller cells with a basal ratio of more than 0.7 were excluded from the statistical analysis. For the experiments with fluo-4/AM, the[ Ca²⁺]_i is given as the ratio of fluorescence -F₀, where F₀ indicates the baseline fluorescence. Statistical analysis (Mann-Whitney test, two-tailed) was performed on computer (Prism program; GraphPad Software Inc., San Diego, CA). Data are expressed as means ± SD. 

Results

Ophthalmoscopic Observations

During and at the end of the postoperative period (3 weeks after the surgery), the development and severity of the dispase-induced PVR was examined by indirect ophthalmoscopy. Although (e.g., due to different modes of retinal vascularization) the ophthalmoscopic appearance and stages of PVR are not fully identical in humans and rabbits, the similarities were sufficient to discriminate between two distinct degrees of severity, according to the nomenclature used in human PVR. ^{22 23} In four animals, moderate PVR was induced, as indicated by a wrinkling of the inner retinal surface and by the formation of a tractional retinal detachment in only one or two quadrants (<4 hours) of the retinal circumference (Figs. 1C 1D ). In most cases, focal and very thin preretinal membranes (posterior) were identified. In one animal, a membrane formation developed in the anterior part of the eye. Vitreous condensation and blood vessel tortuosity were also observed. 

In five animals, massive PVR was induced, as indicated by the formation of large preretinal cellular membranes and by the tractional detachment of nearly the entire sensory retina from the pigment epithelium (Figs. 1E 1F 1G 1H) . Distortion of the medullary wings, large retinal tears, and breaks with rolled edges and fixed retinal folds were observed. 

Vitreous hemorrhages, caused by the sclerotomy, occurred during the surgical procedure in three animals (two showed moderate PVR; one, massive PVR); however, after 2 weeks the hemorrhages were no longer visible. In one animal, cataract formed (possibly caused by trauma to the lens), but after 3 weeks, a clear view of the fundus was obtained. The cornea and the anterior chamber remained clear in all rabbits, and endophthalmitis or conjunctivitis did not occur. 

In the cases of moderate PVR, most of the sensory retina was attached to the pigment epithelium, and isolated Müller cells were thus obtained from attached areas of the retina (Figs. 1C 1D ; white asterisks). In the cases of massive PVR, Müller cells were dissociated from the detached retina (Figs. 1E 1F , black asterisks). 

Basic Müller Cell Membrane Properties

Electrophysiological whole-cell recordings were used to determine whether in the present experimental series, the Müller cells changed their membrane features in a similar manner to that observed in another rabbit model of retinal detachment and PVR ⁸ and in human cases of PVR. ⁶ Cell hypertrophy is generally accepted as one indicator of Müller cell gliosis during PVR. As a marker for the cell membrane area, the membrane capacitance of Müller cells was measured electrophysiologically. The cell membrane capacitances of acutely isolated Müller cells differed significantly between control and PVR-affected eyes, as well as between cells from eyes with moderate and massive PVR (Fig. 2) . These data confirm the earlier reports and suggest that the hypertrophy of Müller cells increases with the severity of the PVR. 

Reduced Kir currents are a typical feature of Müller cells from human PVR-affected retinas. ^{6 7} A similar downregulation of Kir currents in Müller cells was observed in the present dispase-induced PVR model in rabbit eyes. Examples of whole-cell records of typical control and PVR cells are shown in Figure 3A . The cell from the control eye displayed large inwardly directed currents (downwardly depicted). These currents were largely mediated by Kir channels, in that extracellular application of Ba²⁺ (1 mM), a specific blocker of Kir currents in Müller cells, ²⁴ strongly reduced these currents (Fig. 4A ). The current–voltage relation of the cells from the control eyes (Fig. 3B) shows only a very slight inward rectification of the Kir currents at approximately −80 mV, as previously described to be typical in rabbit Müller cells. ²⁵ The downregulation of Kir currents in cells from eyes with PVR was found to be dependent on the severity of the disease. Cells from eyes with moderate PVR displayed Kir currents with significantly reduced amplitudes and densities compared with cells from control eyes, whereas Müller cells from the eyes with massive PVR displayed virtually no Kir currents (Figs. 3A 3B 3C) . Figure 4A shows the mean current density–voltage relation in cells from control and PVR-affected eyes before and during extracellular application of Ba²⁺ (1 mM). As indicated in Figure 4B , the Ba²⁺-sensitive currents of cells from control eyes were very similar to the difference of the whole-cell currents of cells from the control and PVR-affected eyes. This further supports the assumption that, at negative membrane potentials, mainly Kir currents are reduced in the course of PVR. An additional disease-related downregulation of other K⁺ channel types cannot be ruled out. 

The mean membrane potentials of the cells from the three groups investigated are shown in Figure 3D . Cells from control eyes were characterized by a mean membrane potential of approximately −80 mV. Compared with the control, the Müller cells from eyes with moderate PVR displayed a slight depolarization of the membrane (mean− 73 mV), which was even stronger in cells from eyes with massive PVR (mean −62 mV). As indicated by the scatterplots in Figure 3E , the membrane potentials of individual cells from eyes with PVR were scattered over a wide range between −30 and −90 mV, whereas the potentials of most cells from control retinas were within a relatively small range close to −80 mV. This finding is in agreement with previous data on human Müller cells from patients with PVR ⁶ and may be caused by the high membrane resistances (i.e., low Kir current densities: Fig. 3E ) in PVR cells that allow large membrane hyperpolarizations or depolarizations in response to opening or closing of only a small number of ion channels. 

Identification of BK Currents

To determine whether Müller cells from adult rabbits express BK currents, the effect of the BK channel opener phloretin on their whole-cell currents was tested. The flavonoid phloretin has been shown to enhance the amplitude of BK currents in porcine Müller cells and to increase the open probability of single BK channels in excised membrane patches. ²⁶ Extracellular application of phloretin (200 μM) increased the outward currents at positive membrane potentials in cells from both control and PVR-affected eyes (Fig. 5A ). The Kir currents (elicited at negative membrane potentials, depicted as downward deflections) were not affected by the substance, but a phloretin-induced depression of voltage-gated, delayed rectifying K⁺ outward currents was regularly observed. A similar effect has been described for Xenopus nerve fibers. ²⁷ To verify that the phloretin-evoked currents were BK currents, the specific BK channel blocker iberiotoxin (100 nM) was tested in the presence of phloretin. Indeed, iberiotoxin inhibited the phloretin-evoked currents (Figs. 5A 5B 5C) but failed to abolish the depression of delayed rectifying K⁺ currents by phloretin. BK channel–mediated currents were recordable from most of the cells, even in the control eyes ( in 61 of 69 cells from control eyes, independent of whether they responded to ATP or not; discussed later). The densities of the phloretin-evoked currents were similar in cells from control and PVR-affected retinas (Fig. 5D) as previously described in human Müller cells. ⁶  

ATP-Evoked BK Current Responses

BK channels are both voltage sensitive and activated by elevated[ Ca²⁺]_i; thus, an activation of BK channels at a given voltage can be caused by an increase of [Ca²⁺]_i. For this reason, whole-cell current records of BK currents can be used as an indirect indicator to detect ATP-induced alterations in[ Ca²⁺]_i. ¹⁷ Whole-cell currents were recorded in Müller cells from eyes with PVR during exposure to extracellular ATP (500 μM). Examples of such records in three cells are shown in Figure 6A . The majority of cells from control retinas did not respond with any changes of the whole-cell currents (Figs. 6B 6C) . By contrast, most of the investigated cells from retinas with PVR responded to extracellular ATP with a transient increase of the BK currents (at +120 mV; Fig. 6A , middle and right side). The percentages of cells responding to extracellular ATP were significantly different in cells from control eyes (11.5%) and from eyes with PVR (moderate: 83.3%; massive: 90.5%; Fig. 6C ). The absence of ATP responses in some of the cells was not caused by the nonexistence of BK channels in these cells (Fig. 6B , left trace). Figure 6D illustrates mean values of the BK currents (measured at +120 mV) in cells from control and PVR-affected eyes. Provided that a given cell responded to ATP, the amplitudes of the ATP-evoked peak currents were similar in all three cell populations investigated. 

Purinergic Receptor Type Determination

To determine the types of purinergic receptors mediating the ATP-induced BK current responses, different agonists were tested. The application of BzATP (50 μM), a more specific agonist of P2X receptors, ²⁸ evoked only small cationic inward currents at a holding potential of −80 mV (0.14 ± 0.09 pA/pF, n = 6) in six of seven cells from the control retinas (Fig. 7A ). There were no significant differences compared with the currents recorded in Müller cells from retinas with PVR (0.20 ± 0.12 pA/pF, six of nine cells; Fig. 7A ). Adenosine did not evoke transient BK current responses in control Müller cells and in all investigated cells of retinas with PVR (14 cells from eyes with moderate or with massive PVR; Fig. 7B ). However, in only 20% (2/9) of control cells but in nearly 90% (18/19) of the PVR cells, an extracellular application of UTP (100 μM) evoked a transient BK current increase similar to that induced by extracellular ATP (Figs. 6A 6B 7B 7C) , in a dose-dependent manner (Figs. 7C 7D) . The EC₅₀ (i.e., the concentration needed to induce a peak current increase) at +120 mV, half the size of the maximal current increase, was estimated at 2.2 μM UTP. A maximal response was observed at 10 μM UTP. Furthermore, virtually identical current responses to both ATP and UTP were recorded in Ca²⁺-free extracellular solutions (data not shown). Together, the results indicate a dominant role of P2Y receptors in rabbit Müller cells, and an upregulation of these receptors in cells from retinas with PVR. 

ATP-Evoked Ca²⁺ Responses in Retinal Wholemounts

To study possible differences in ATP-evoked responses between control and PVR-affected retinas directly, confocal images were recorded by laser scanning microscopy of retinal wholemounts. The optical focus was approximately 10 μm below the vitread surface—that is, on the cobblestone pattern of Müller cell end feet, forming an almost uninterrupted sheet. ² The application of 200μ M adenosine did not evoke any changes in[ Ca²⁺]_i in Müller cell end feet in the control retinas. However, extracellular application of ATP (200 μM) and UTP (200 μM) induced Ca²⁺ transients within (some of) the Müller cell end feet (Fig. 8) . The numbers of end feet responding to ATP were different between control and PVR-affected retinas. In the control retina, only a few Müller cells responded to ATP and UTP by an elevation of[ Ca²⁺]_i, whereas in the retina with PVR, most of the identifiable end feet were responsive to ATP. To verify that the ATP-induced Ca²⁺ responses were located in Müller cells and to obtain better quantitative results (the total number of, possibly nonresponding, Müller cell end feet within the studied area could not be evaluated precisely), further experiments were performed on acutely isolated Müller cells. 

ATP-Evoked Ca²⁺ Responses in Isolated Cells

Extracellular application of ATP (500 μM) induced a transient elevation of the [Ca²⁺]_i in a subpopulation of rabbit Müller cells, with peaks concentrations occurring between 30 and 90 seconds after the beginning of drug exposure (Fig. 9A ). After several minutes, the[ Ca²⁺]_i decreased, regardless of continuous ATP application. Very similar responses were elicitable by UTP (500 μM), and in the absence of extracellular Ca²⁺ (Fig. 9A) . Whereas such responses were obtained only from a minority of control cells (13%), an increased incidence of responding cells from eyes with PVR was observed (Fig. 9C) ; 29% of the cells from eyes with moderate PVR and 53% from eyes with massive PVR showed a response. The peak amplitudes of the[ Ca²⁺]_i increases, if elicitable, were similar in all cell populations investigated (Fig. 9D) . The calculated basal[ Ca²⁺]_i of the nonresponding cells (control cells: 85 ± 38 nM, n = 62; PVR 72 ± 40 nM, n = 43) and of the responding cells (control cells: 87 ± 63 nM n = 9, PVR 66 ± 31 nM n = 37) do not differ significantly from each other. The mean increases of[ Ca²⁺]_i in the ATP-responding cells from control and PVR-affected retinas were also not significantly different (control cells: 223 ± 126 nM, n= 9; PVR 193 ± 166 nM, n = 37). Therefore, the[ Ca²⁺]_i increased 2.5- to 3-fold during an ATP application. We never observed a BzATP-evoked increase of [Ca²⁺]_i in extracellular control solution, either in Müller cells from control or retinas with PVR (Fig. 9A) . 

Discussion

ATP Receptors in Rabbit Müller Cells

Although the expression of purinergic receptors by Müller cells has been demonstrated in cells from several species, ^{14 15 16 17} this is the first report of their presence in rabbit Müller cells in situ. Our results allow for at least a partial identification of the type of receptors involved in the cellular responses. 

The application of adenosine neither changed the[ Ca²⁺]_i nor evoked transient BK current responses in the cells studied. This suggests either that rabbit Müller cells do not express P1 receptors or that activation of P1 receptors in Müller cells does not result in a transient increase of[ Ca²⁺]_i. The activation of adenosine receptors is coupled to multiple second-messenger signal-transducing mechanisms and does not result necessarily in a transient change of[ Ca²⁺]_i. ⁹ Thus, we cannot exclude the expression of P1 receptors, but it is clear that they did not contribute to the PVR-induced Müller cell changes in the current study. 

The application of ATP did not evoke an increase of inwardly directed currents in Müller cells at a potential of −120 mV, which would reflect the activation of nonselective cation currents through P2X receptor channels (Figs. 6A 6B) . BzATP, a potent agonist for several P2X receptors, ²⁸ evoked only small cationic inward currents near the resting membrane potential, even in nominally Ca²⁺ free solutions reported to provide optimal conditions for currents through P2X₇ receptor channels. ^{16 29} BzATP-evoked Ca²⁺ responses were never observed under control conditions (Fig. 9A) . Thus, any physiological or pathophysiological role of ionotropic (P2X type) purinergic receptors in rabbit Müller cells remains to be elucidated. 

All results, particularly the large responses to UTP and the occurrence of large responses in the absence of extracellular Ca²⁺, indicate an activation of metabotropic P2Y receptors ³⁰ in rabbit Müller cells. There may be species differences in regard to the expression of purinergic receptors by mammalian Müller cells. All human Müller cells (control and pathologically altered) express both P2Y and P2X₇ receptors, ^{16 17} whereas only a small subpopulation of Müller cells from the rabbit retina regularly express P2Y receptors, and the P2X responses, if any, are negligible (present results). The P2X₇ receptor currents are upregulated in human Müller cells from retinas with PVR, ¹⁷ whereas the very small BzATP-evoked currents in rabbit Müller cells did not change significantly in retinas with PVR, compared with the control eyes. However, the different kinetics of the rabbit Müller cell responses (some cells displayed an oscillating, others a sustained response; Figs. 6 7 8 ) suggest an expression of more than one distinct purinergic receptor subtype in rabbit retinal Müller cells. Further studies are being performed to characterize the particular subtype of purinergic receptors in rabbit Müller cells of the normal and the pathologically changed rabbit retina. 

PVR-Induced Changes of Membrane Features

The present study reveals similar alterations of membrane features as observed in Müller cells from human retinas with PVR ^{6 7} and from an animal model of experimental retinal detachment. ⁸ We also confirmed that the hypertrophy of Müller cells, the downregulation of Kir currents, and the depolarization of the resting membrane potential all depend on the severity of the PVR. The mechanisms of the downregulation of the Kir currents during proliferative gliosis are not yet clear. In cultured Müller cells, the presence of both the extracellular matrix protein laminin and insulin has been described as necessary to induce the expression of Kir channels. ³¹ In cultured astrocytes, it has been shown that addition of tumor necrosis factor-α ³² or immunologically active substances ³³ to the culture medium may induce a reduction of Kir currents and a depolarization of the membrane potential. This suggests that a loss of cell–cell contacts and/or exposure to soluble factors could be involved in the changes of membrane properties in reactive Müller cells. It is interesting to note that significant alterations of membrane conductances (this study) and an increased immunoreactivity for glial fibrillary acidic protein (GFAP, data not shown) have already been observed during moderate PVR in Müller cells that were isolated from attached retinal pieces; a similar observation was made in an animal model of retinal detachment. ⁸  

Increased ATP-Evoked Responses in Müller Cells from Retinas with PVR

The key finding of the present study is that a significantly increased number of rabbit Müller cells responded to ATP when PVR developed in a retina. Because neither the amplitudes of the ATP-induced Ca²⁺ increases (Fig. 9D) nor those of the ATP-activated BK currents (Fig. 6D) were larger in responsive cells from retinas with PVR than in responding control cells, it is concluded that the capacities of the intracellular Ca²⁺ stores as well as the number of BK channels per cell do not change much in reactive gliosis. Rather, the expression of P2Y receptors is upregulated among the Müller cell population. 

In the human retina, Müller cells upregulate currents through P2X₇ receptor channels during PVR, whereas the P2Y receptor-mediated transient BK current responses were similar in cells from PVR-affected and control retinas. ¹⁷ Although the functional meaning of this species difference remains to be elucidated, our data support the hypothesis that an upregulation of P2-type purinergic receptors is a common feature of gliotic cells, including Müller cells. The involvement of purinergic receptors in the induction or maintenance of gliosis in vivo has been discussed, ⁹ and the activation of P2 receptors may induce astrogliosis and proliferation of astrocytes in the brain. ¹¹ Thus, the release of nucleotides from dying or degenerating neurons could trigger the induction of gliosis in Müller cells. Furthermore, released ATP may be a candidate for (one of) the soluble factors that induce the changes in membrane properties of Müller cells, in the degenerating areas as well as in the peripheral (attached) parts of the retina (see earlier discussion). The present results also indicate that purinergic receptors may be implicated in the induction or maintenance of proliferative Müller cell gliosis. In cultured Müller cells from guinea pig ³⁴ and human retinas, ¹⁷ extracellular ATP has been shown to enhance the DNA synthesis rate. Furthermore, the BK channel blocker iberiotoxin depresses the stimulating effect of ATP on DNA synthesis. It has also been shown in other cell types that the activation of BK channels may serve to enhance ATP-induced Ca²⁺ entry from extracellular space. ³⁵  

Conclusion

We demonstrate the presence of P2Y-type ATP receptors in rabbit Müller cells, and an upregulation of these receptors during proliferative gliosis in experimentally induced PVR. In most Müller cells from eyes with PVR, extracellular ATP application evoked a transient increase of the intracellular Ca²⁺ concentration and a stimulation of Ca²⁺-activated K⁺ currents. As in Müller cells of the human retina, both ATP-evoked responses may support the proliferation of Müller cells during PVR. 

 Supported by Grant 01KS9504, Project C5, from the Bundesministerium für Bildung und Forschung (BMBF), Interdisciplinary Center for Clinical Research at the University of Leipzig, and by Grants Re. 849/8-1 and Graduiertenkolleg “Intercell,” GRK 250/1-96 from the Deutsche Forschungsgemeinschaft.

 Submitted for publication June 29, 2001; revised October 23, 2001; accepted November 2, 2001.

 Commercial relationships policy: N.

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked“ advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Fundus photographs and photographs of the open eyes (B, D, F) of healthy rabbits and of rabbits with moderate or massive PVR. Fundus photograph (A) and a photograph of the open eye (B) of the same healthy, untreated rabbit. White asterisks: Attached, normal retina of the control eyes. (C, D) Moderate PVR: The fundus photograph (C) shows the border between the attached part (white asterisk) and the detached part (black asterisk) of the retina. The attached part of the retina looks normal compared with the control retina. (D) In the open eye (same eye as in C), the detached parts of the folded retina are easily recognizable (arrows); white asterisks: attached retina. (E– H) Massive PVR: Fundus photographs (E, G, H) and a photograph of the open eye (F) of three different rabbits (E, F, same eye). The PVR is characterized by detachment of nearly the whole retina (E, F, black asterisks). Full-thickness retinal folds (G, arrows) , large retinal tears (E, arrowhead), areas of retinal degeneration (G, arrowheads), and large epiretinal membranes (H, arrows). mr, medullary rays.

Figure 2.

 

Mean ± SD membrane capacitances of rabbit Müller cells that were acutely isolated from control eyes, from eyes with moderate PVR, and from eyes with massive PVR. Cell numbers are in parentheses.• P < 0.05; •••P < 0.001.

Figure 3.

 

Müller cells from retinas with PVR show a reduction of whole-cell K⁺ currents when compared with cells from control eyes. (A) Examples of the whole-cell currents of three cells derived from a control eye (left) and from eyes with moderate (middle) and massive (right) PVR. The inwardly directed currents (downwardly depicted) were reduced in Müller cells from the PVR-affected eyes. Voltage steps were applied from a holding potential of −80 mV to increasing de- and hyperpolarizing potentials between −180 and +140 mV (250 msec, 20-mV increment). Small bars at left: zero-current levels. (B) Mean current density–voltage relationships of the whole-cell currents of Müller cells derived from control eyes and from eyes with moderate and massive PVR. The steady state currents were measured at the end of 250-msec voltage steps. Inset: part of the current-voltage curves, showing the disease-dependent shifts of the zero-current potential. (C) Mean density ± SD of the inwardly directed currents of the three cell populations. Cell numbers are in parentheses. The currents were measured between the voltage steps to −100 and −160 mV. (D) Mean ± SD membrane potentials of the cells from the three populations investigated. The potentials were determined by measuring the zero-current potentials in the steady state current-voltage curves. (E) Scatterplots of the density of the inwardly directed currents versus membrane potential in all cells investigated. One filled circle represents the relation of one cell.• P < 0.05; •••P < 0.001.

Figure 4.

 

The reduced K⁺ currents in cells from retinas with PVR are mainly Kir currents, as indicated by the blocking effect of Ba²⁺ ions. (A) Mean (± SD) current density–voltage relationships of the whole-cell currents of 23 Müller cells from control eyes (left) and of 15 cells from eyes with massive PVR (right) that were evoked before (control) and during extracellular application of Ba²⁺ ions (1 mM). Inset: part of the curves, demonstrating the Ba²⁺-induced shift of the zero-current potential. (B) Comparison of the mean Ba²⁺-sensitive currents with the difference of the current–voltage curves between cells from control eyes and cells from eyes with massive PVR. The Ba²⁺-sensitive currents were calculated by subtraction of the whole-cell currents that were evoked before and during application of Ba²⁺ ions. The difference currents from the healthy and massive PVR-affected eyes were calculated by subtracting the noninfluenced control whole-cell currents of control and massive PVR-affected eyes (A, ○). This clearly demonstrates that the Ba²⁺-sensitive currents of the control cells were very similar to the currents that disappeared in the PVR cells.

Figure 5.

 

Rabbit Müller cells from both control and PVR-affected eyes displayed BK currents that were increased by the BK channel activator phloretin and decreased by iberiotoxin. (A) Examples of current records in a cell from a control eye (top) and from an eye with moderate PVR (bottom). Extracellular exposure of phloretin (200 μM) increased outwardly directed currents in both cells. Simultaneous application of iberiotoxin (100 nM) blocked the phloretin-evoked currents. The cells were held at −80 mV, and de- and hyperpolarizing voltage steps were applied at an increment of 20 mV. Small bars at left: zero-current levels. (B, C) Mean ± SD current–voltage relationships in five cells from control eyes (B) and in eight cells from eyes with moderate PVR (C). The currents were recorded before (control) and during extracellular application of phloretin and during simultaneous application of phloretin and iberiotoxin. (D) Mean densities of the phloretin-influenced currents of cells from control eyes and from moderate PVR. The currents were calculated by subtracting the control currents from the currents that were recorded during phloretin exposure.

Figure 6.

 

Extracellular ATP caused transient increases of BK currents, mainly in cells from retinas with PVR. (A) Examples of whole-cell currents that were recorded in three cells from a control retina (left) and from eyes with moderate PVR (middle and right). Thick bars: extracellular application of ATP. Small bars at left of each trace: zero-current levels. The cells were held at 0 mV, and de- (to +120 mV) and hyperpolarizing (to −120 mV) 50-msec voltage steps were applied at a frequency of 2.5 Hz (B, inset). The currents were measured at the end of the 50-msec voltage steps. (B) Example of a control Müller cells that did not respond to ATP, but expressed BK channels. The BK channels could be activated by phloretin and were blocked by iberiotoxin. In the Müller cells from the eye with PVR, ATP and phloretin evoked outward currents. The currents that were activated by ATP represent mainly BK currents, as indicated by increased outward currents at very positive potentials. (C) Percentages of cells that responded to extracellular ATP with a transient BK current increase. Numbers of all cells investigated in parentheses. (D) The ATP-evoked peak current increases were similar in their amplitudes in cells from control eyes and from eyes with PVR. Mean ± SD amplitudes of the currents at +120 mV before (control) and during extracellular application of ATP (500μ M). The ATP values of the nonresponding cells were measured 15 to 20 seconds after beginning of ATP exposure. For responding cells, the peaks of the responses are shown. Number of cells investigated is in parentheses. •P < 0.05; •••P < 0.001.

Figure 7.

 

BzATP- and UTP-evoked currents. (A) BzATP-evoked currents were found in some rabbit Müller cells. Cells were held at −80 mV, and 50 μM BzATP was applied to a control cell (left) and a cell from an eye with massive PVR (right) in a divalent cation-free extracellular solution. In both cases, similar inward currents of small amplitudes were recorded. These currents were probably due to the activation of P2X receptors. (B) Examples of current recorded in one cell from a control eye (left) and an eye with moderate PVR (right). Extracellular application of adenosine did not alter the whole-cell currents, whereas UTP induced a transient increase of the currents at 0 and at +120 mV. (C) Example of current recorded in a cell from an eye with massive PVR. Extracellular UTP was applied in increasing concentrations. Between each application, there were washout periods of 5 minutes. (D) Mean dose–response relation of the UTP-induced increase in the currents at +120 mV, which was measured in four cells from eyes with massive PVR. The values show the relative peak amplitude increase over the basal level that was measured immediately before agonist application. The response at 100 μM UTP was set as 1.

Figure 8.

 

Confocal images of retinal wholemounts from a healthy eye (upper left), an eye with moderate PVR (middle), and an eye with massive PVR (upper right). The images were recorded within the ganglion cell layer and show apparent somata of ganglion cells (dark circles) and end feet of Müller cells. The fluorescence emission of fluo-4/AM (Ca²⁺ response) is shown after addition of ATP (200 μM) to the bath solution. In the control retina, only a few end feet of Müller cells responded to ATP with increased[ Ca²⁺]_i, whereas in the moderate retina with PVR the number of responding cells increased and in the massive retina with PVR, nearly all Müller cell end feet were responsive to ATP. The two diagrams (bottom left) show the time-dependent alterations of the fluorescence emission at selected regions indicated in the images and diagrams. Dotted vertical lines: time point at which confocal images were made. Numbers 1, 2 (left) and 4, 5 (right) are Müller end feet. Numbers 3 and 6 are ganglion cells that did not respond to the ATP application (thick lines in the diagrams). The Müller cell response oscillated during the application of ATP. The two smaller confocal images of retinal wholemounts (bottom right) show the response of a control (left) and massive PVR–affected (right) retina to an application of 200 μM UTP. The number of responding Müller cells increased in the retina with PVR. Nearly all Müller cells that responded to ATP responded to UTP. The images of the retina with massive PVR (ATP response, upper right; UTP response, bottom right) were made from the same retina in one series of experiments.

Figure 9.

 

Extracellular application of ATP and UTP (500 μM) may transiently increase the intracellular Ca²⁺ concentration in acutely isolated Müller cells, as indicated by fura-2/AM imaging. (A) Time dependence of Müller cell responses to ATP and UTP from a control (left) and PVR (middle) retina. Müller cells that responded to ATP also responded to UTP (solid line). The ATP response of Müller cells was evocable in extracellular Ca²⁺-free solution. We never observed a BzATP-evoked increase of[ Ca²⁺]_i in extracellular control solution (right). Dashed lines: nonresponding cells of the corresponding retinas. (B) Example of ratio images of one Müller cell from a healthy retina before (left) and during (right) ATP application. This Müller cell responded to ATP with a slight increase of[ Ca²⁺]_i in the end foot (at the top) and the somatic region (middle). These images are the ratio images calculated by dividing the images of F₃₄₀ and F₃₈₀. The gray scale (middle) represents the ratio calibration. (C) Percentages of cells that responded to extracellular ATP with a transient increase of the[ Ca²⁺]_i. Numbers of cells investigated in parentheses. (D) Mean ± SD of the ratio of F₃₄₀ to F₃₈₀ at basal conditions (immediately before ATP application) and during ATP exposure. The ATP levels in the nonresponding cells were measured 1 minute after the beginning of ATP exposure. In responding cells, the peaks of the responses in the somatic region are shown. Numbers of all cells investigated are in parentheses. •P < 0.05;•• P < 0.01; •••P < 0.001.