Am J Physiol Cell Physiol 287: C1003–C1011, 2004.
First published June 9, 2004; 10.1152/ajpcell.00175.2004.
cAMP-activated maxi-Cl⫺ channels in native
bovine pigmented ciliary epithelial cells
Chi-Wai Do,1 Kim Peterson-Yantorno,1 Claire H. Mitchell,1 and Mortimer M. Civan1,2
1
Departments of Physiology and 2Medicine, University of Pennsylvania, Philadelphia, Pennsylvania 19104
Submitted 1 April 2004; accepted in final form 3 June 2004
Do, Chi-Wai, Kim Peterson-Yantorno, Claire H. Mitchell, and
Mortimer M. Civan. cAMP-activated maxi-Cl⫺ channels in native
bovine pigmented ciliary epithelial cells. Am J Physiol Cell Physiol
287: C1003–C1011, 2004. First published June 9, 2004; 10.1152/
ajpcell.00175.2004.—The eye’s aqueous humor is secreted by a
bilayered ciliary epithelium comprising pigmented (PE) and nonpig-
mented (NPE) epithelial cell layers. Stromal Cl⫺ enters the PE cells
and crosses gap junctions to the NPE cells for release into the aqueous
humor. Maxi-Cl⫺ channels are expressed in PE cells, but their
physiological significance is unclear. To address this question, excised
patches and whole native bovine PE cells were patch clamped, and
volume was monitored by calcein fluorescence. In symmetrical 130
mM NaCl, cAMP at the cytoplasmic surface of inside-out patches
produced concentration-dependent activation of maxi-Cl⫺ channels
with a unitary conductance of 272 ⫾ 2 pS (n ⫽ 80). Voltage steps
from 0 to ⫾80 mV, but not to ⫾40 mV, produced rapid channel
inactivation consistent with the typical characteristics of maxi-Cl⫺
channels. cAMP also activated the maxi-Cl⫺ channels in outside-out
patches. In both cases, maxi-Cl⫺ channels were reversibly inhibited
by SITS and 5-nitro-2-(phenylpropylamino)benzoate (NPPB). De-
creasing cytoplasmic Cl⫺ concentration reduced both open-channel
probability and unitary conductance. Similarly, the membrane-per-
meant 8-bromo-cAMP stimulated outward and inward whole cell
currents; the stimulation was larger at higher intracellular Cl⫺ con-
centration. As with unitary currents, cAMP-triggered whole cell
currents displayed inactivation at ⫾80 but not at ⫾40 mV. Moreover,
cAMP triggered NPPB-sensitive shrinkage of PE cells. The results
suggest that cAMP directly activates maxi-Cl⫺ channels of native PE
cells that contribute to Cl⫺ release particularly from Cl⫺-loaded cells.
These cAMP-activated channels provide a potential mechanism for
reducing and modulating net aqueous humor secretion by facilitating
Cl⫺ reabsorption into the ciliary stroma.
cell volume; chloride secretion; aqueous humor formation
GLAUCOMA IS A GROUP OF CONDITIONS in which retinal ganglion
cells are lost, at least in part, because of events triggered by an
increase in intraocular pressure (IOP). The level of IOP is
governed by the dynamic balance between aqueous humor
formation and drainage. Although glaucoma is believed to
stem from a primary increase in outflow resistance, a major
strategy for reducing IOP is to decrease the rate of aqueous
humor formation. Aqueous humor is secreted by the ciliary
epithelium, which comprises two epithelial layers: an outer
pigmented ciliary epithelial (PE) layer adjacent to the ciliary
stroma, and an inner nonpigmented ciliary epithelial (NPE)
layer abutting the aqueous humor. Cells within and between
these two layers are functionally coupled through intercellular
gap junctions (41). Aqueous humor formation is driven by
Address for reprint requests and other correspondence: M. M. Civan, Dept.
of Physiology, A303 Richards Bldg., Univ. of Pennsylvania, Philadelphia, PA
19104-6085 (E-mail: civan@mail.med.upenn.edu).
http://www.ajpcell.org 0363-6143/04 $5.00 Copyright © 2004 the American Physiological Society
active ion transport across the ciliary epithelium, followed by
osmotic water movement (30). Cl⫺ is the principal anion of the
aqueous humor, and transepithelial Cl⫺ secretion likely plays a
central role in aqueous secretion (12, 14, 18, 51). Figure 1
illustrates a consensus model of aqueous humor formation. Cl⫺
can enter from the ciliary stroma into the PE cells through two
main electroneutral pathways: the Na⫹-K⫹-2Cl⫺ cotransporter
and parallel Na⫹/H⫹ and Cl⫺/HCO⫺ 3 exchangers (14, 18, 20,
35, 52). Cl⫺ can then diffuse through the intercellular gap
junctions to the NPE cells and finally exit through Cl⫺ chan-
nels to the posterior chamber of the eye (11, 18, 54, 55). Many
transport mechanisms underlying aqueous humor secretion
have been identified, but the regulatory pathways remain elu-
sive.
cAMP has long been proposed to play an important role in
modulating the rate of aqueous humor formation and IOP but
by mechanisms and regulatory pathways that are yet unclear
(reviewed recently by Do and Civan, Ref. 17). Administration
of isoproterenol and forskolin, which trigger cAMP produc-
tion, reduces aqueous humor formation and IOP among differ-
ent species including rabbit, monkey, and human (7, 8, 25, 32).
The addition of terbutaline, another ␤-adrenergic agonist, to
the arterially perfused bovine eye reduces the rate of aqueous
humor formation (46). Recently, cAMP has been found to
activate whole cell Cl⫺ currents of immortalized bovine PE
cells (24). Indirect evidence also suggested that cAMP might
trigger activation by acting directly on Cl⫺ channels, indepen-
dent of cAMP-activated kinase (PKA). Large-conductance
maxi-Cl⫺ channels have been noted in PE cells (38), but their
role in ciliary epithelial secretion is unknown. As in other cells,
the physiological significance of these channels is unclear,
because their open-channel probability (Po) is maximum
within ⫾20 mV (6), far from the ciliary epithelial membrane
potential (Vm) of approximately ⫺60 mV (9, 26).
In the present study, we examine the effects of cAMP on
excised single-channel and whole cell currents and on cell
volume in native bovine PE cells. Our results suggest that
cAMP activates maxi-Cl⫺ channels in PE cells, facilitating Cl⫺
release from the PE cells to the ciliary stroma. This Cl⫺
recycling process may potentially serve as a pathway to regu-
late net aqueous humor formation.
MATERIALS AND METHODS
Isolation of fresh native bovine PE cells. Freshly nucleated bovine
eyes were collected from a local abattoir. The cornea and iris were
removed along the limbus. After that, ciliary processes were excised
into small pieces and rinsed with Dulbecco’s phosphate-buffered
The costs of publication of this article were defrayed in part by the payment
of page charges. The article must therefore be hereby marked “advertisement”
in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
C1003
C1004 ACTIVATION OF MAXI-CL⫺ CHANNELS IN BOVINE PE CELLS
Fig. 1. Formation of aqueous humor: abridged hypothesis,
emphasizing possible role of Cl⫺ recycling. Na⫹-K⫹-2Cl⫺
cotransporter and coupled Cl⫺/HCO⫺ 3 and Na /H
⫹ ⫹
parallel
exchangers are principal pathways for the uptake of Na⫹ and
Cl⫺ into the pigmented epithelial (PE) cells. Tight junctions are
present only between nonpigmented epithelial (NPE) cells and
not between PE cells.
saline (GIBCO-BRL, Grand Island, NY). The preparation was then
incubated with 0.25% trypsin at 37°C for 30 min. Singly dissociated
PE and NPE cells were obtained by trituration. Cells were washed
twice with Dulbecco’s phosphate-buffered saline before they were
seeded to the coverglass (Fisher Scientific, Pittsburgh, PA). Cells were
incubated in medium 199 containing 10% fetal bovine serum and
0.1% gentamicin (GIBCO-BRL) at 37°C in 5% CO2 before use.
Rounded PE cells, as noted by the presence of abundant pigment
granules, were chosen for patch clamp and cell volume measurements
after being incubated for at least 3 h but not ⬎36 h. Because of the
absence of the tight junctions in PE cells (Fig. 1), channels patched
either on the basolateral or apical surfaces were expected to have
identical effects on transepithelial transport across the ciliary epithe-
lium. In addition, pigment granules were generally localized along the
apical border of the PE cells; the area less heavily pigmented was
always chosen for patching.
Patch clamp measurements. Micropipettes for excised patches and
whole cell configurations with resistances of 5–10 and 2– 4 M⍀,
respectively, were prepared from Corning glass (cat. no. 7052; World
Precision Instruments, Sarasota, FL) using a Flaming/Brown micropi-
pette puller (model P-97; Sutter Instruments). Micropipettes were
coated with Sylgard (World Precision Instruments) and fire-polished
with a microforge (model MF-830; Narishige). Recordings were
conducted in a perfusion chamber connected to a Ag/AgCl pellet via
a 3 M KCl agar bridge at room temperature (22–24°C). Single-
channel and whole cell currents were measured by using either
Axopatch 1B or 1D patch clamps (Axon Instruments, Foster City,
CA) coupled to an external Bessel filter (model 990C; Frequency
Devices, Haverhill, MA). Single-channel recordings were acquired at
10 kHz and filtered at 2 kHz; whole cell measurements were obtained
at 2 kHz and filtered at 500 Hz. For excised patches, the holding
potential (Vh) was 0 mV, the command potentials were from ⫺80 to
⫹80 mV in 20-mV steps, and the upward current deflections depicted
inward currents, and vice versa. The duration of each command step
was 4 s, with intervening periods of 1.2 s at Vh. For whole cell
measurements (Vh was also 0 mV) the step pulses ranged either from
⫺100 to ⫹80 mV in 20-mV increments (the duration of each com-
mand potential was 300 ms) or from ⫺80 to ⫹80 mV in 40-mV steps
(each command pulse lasted for 4 s). In both cases, the upward current
deflections indicated outward currents and vice versa. Data were
acquired with the Axon Digidata using pClamp 8.2 software (Axon
AJP-Cell Physiol • VOL
Instruments, Foster City, CA). Analysis was conducted by using Axon
Instruments Clampfit 8.2 software.
Cell volume measurements. Cell volume was monitored by mea-
suring cell area using calcein fluorescence (37). Coverglasses were
mounted in a chamber connected to a Nikon Diaphot microscope.
Fresh PE cells were loaded with 4 ␮M calcein-AM and 0.02%
Pluronic for 30 – 40 min and then perfused with isotonic Tyrode’s
solution for 30 min before initiating the experiment. Calcein was
excited every 20 s at 488 nm, and light emitted at 520 nm was detected
with an IC-200 charge-coupled device camera (Photon Technology
International, Princeton, NJ). Cell area was determined from the
number of pixels detected above threshold within the cell region.
Results were analyzed by using Imagemaster software (Photon Tech-
nology International). Experiments were conducted at room temper-
ature.
Solutions and pharmacological agents. For single-channel record-
ings, the baseline solutions for the micropipette and for the bath were
identical and contained (in mM): 130 NaCl, 20 sucrose, 10 HEPES,
0.674 CaCl2, 1.1 EGTA, and 0.1 GTP (290 mosmol/kgH2O, pH 7.4).
Cl⫺ concentrations were reduced by replacing NaCl with isosmolal
amounts of sucrose. For whole cell measurements, the standard
bathing solution contained (in mM): 140 NMDGCl, 0.5 MgCl2, 10
HEPES and 1.2 CaCl2 (290 mosmol/kgH2O, pH 7.4). The micropi-
pette solution contained (in mM): 140 NMDGCl, 1.2 MgCl2, 10
HEPES, 1 EGTA, 2 ATP and 0.5 GTP (290 –295 mosmol/kgH2O, pH
7.3). Where appropriate, the Cl⫺ concentration in the micropipette
was reduced to 30 mM by equimolar substitution of aspartate.
For cell volume measurements, the isotonic Tyrode’s bathing
solution contained (in mM): 110 NaCl, 15 HEPES, 2.5 CaCl2, 1.2
MgCl2, 4.7 KCl, 1.2 KH2PO4, 30 NaHCO3, and 10 glucose (295–305
mosmol/kgH2O, pH 7.35). Solutions were made hypotonic by reduc-
ing the concentration of NaCl from 110 to 55 mM (240 –250 mosmol/
kgH2O).
All chemicals were reagent grade. cAMP, 8-bromo-cAMP (8-Br-
cAMP), 5-nitro-2-(phenylpropylamino)benzoate (NPPB), and SITS
were purchased from Sigma (St. Louis, MO). All solutions were
filtered through a 0.22-␮m Millipore filter before use.
Statistics. Data are presented as means ⫾ SE, where n is the
number of experiments. The statistical significance of the differences
was evaluated by using either one-way ANOVA followed by the
Student-Newman-Keuls test or Student’s paired t-test. P ⬍ 0.05 was
considered statistically significant.
287 • OCTOBER 2004 • www.ajpcell.org
 ACTIVATION OF MAXI-CL⫺ CHANNELS IN BOVINE PE CELLS C1005

Fig. 2. Activation of maxi-Cl⫺ channels by cAMP (500 ␮M) in an excised inside-out patch from native bovine PE cells. The
holding potential (Vh) was 0 mV, and patches were clamped at membrane potentials (Vm) of ⫺80 to ⫹80 mV in 20-mV steps. The
channel was usually open when Vm was within ⫾40 mV; voltage-dependent channel inactivation was observed when Vm was
beyond this range. The dotted and solid lines indicated closed (c) and open (o) states of the channel, respectively. Upward current
deflections indicated inward currents, and vice versa. No channel activity was observed either before adding or after removing
cAMP. A: before adding cAMP. B: during cAMP exposure. C: after removing cAMP.

RESULTS

Activation of maxi-Cl⫺ channels by cAMP in excised inside-

out and outside-out patches. Under baseline conditions in
which both bathing and micropipette solutions contained 130
mM Cl⫺, maxi-Cl⫺ channel activity was rarely observed in
cell-attached patches, consistent with previous observations
(38). Its occurrence was also infrequently detected on excision,
accounting for ⬃20% of the preparations that subsequently
demonstrated maxi-Cl⫺ channel activity in the presence of
cAMP. However, when Cl⫺ channels were seen under baseline
conditions, the conductance was large (⬃280 pS), correspond-
ing to values reported for maxi-Cl⫺ or voltage-dependent
anion channels (VDAC) (6, 27, 33, 44). Although the activity
of maxi-Cl⫺ channels was not always detected under baseline
conditions, the addition of cAMP to the cytoplasmic solution
activated the maxi-Cl⫺ channels reversibly, displaying a linear
conductance of 272 ⫾ 2 pS (n ⫽ 80). Figure 2 shows the
activation of maxi-Cl⫺ channel by cAMP in excised inside-out
patches. No channel activity was detected either before adding
or after removing the cAMP from the bathing solution. Maxi-
Cl⫺ channel activity was usually observed 1–2 min after the Fig. 3. Vm-dependence of open probability (Po) for maxi-Cl⫺ channels in the
presence of 500 ␮M cAMP. Averages have been obtained from all patches that
addition of cAMP. In most cases, only one maxi-Cl⫺ channel displayed open events at all applied voltages. The channel displayed Vm-
per patch was detected. Multiple subconductance states were dependent inactivation, especially when Vm was either greater than ⫹40 mV or
observed in some of the patches as in Fig. 2B when Vm ⫽ ⫺60 smaller than ⫺40 mV. The uppermost curve represented the baseline condi-
mV. The channel was open within ⫾40 mV, with rapid tions in which Cl⫺ concentrations in the micropipette and bath were 130 mM.
inactivation at higher membrane polarizations, reducing Po at Reducing the cytoplasmic Cl⫺ concentration from 130 mm to either 65 or 30
mm reduced the Po at all potentials. In all cases, the extracellular NaCl
Vm beyond ⫾40 mV. This resulted in an inverted U-shaped concentration was maintained constant at 130 mM, whereas the cytoplasmic
plot of Po as a function of Vm (Fig. 3, uppermost curve), which Cl⫺ concentration was varied. The curves were fitted by two Boltzmann
is highly characteristic of maxi-Cl⫺ channels found in other equations.

AJP-Cell Physiol • VOL 287 • OCTOBER 2004 • www.ajpcell.org

C1006 ACTIVATION OF MAXI-CL⫺ CHANNELS IN BOVINE PE CELLS

studies (6, 28, 39). It was noted that the plot is not symmetrical;
Po was always higher at positive than at negative Vm, espe-
cially at extreme Vm, as described for other cell types (4, 16).
The addition of cAMP had no effect on the channel unitary
conductance, changing from 277 ⫾ 4 to 275 ⫾ 4 pS on the
addition of cAMP in paired measurements (n ⫽ 25). Instead,
cAMP activated the maxi-Cl⫺ channels primarily by increas-
ing Po. In symmetrical 130 mM NaCl solutions, cAMP caused
a concentration-dependent increase in Po over the concentra-
tion range from 30 to 500 ␮M (Fig. 4), leading to a corre-
sponding increase in mean patch current (Fig. 5). The cAMP-
activated maxi-Cl⫺ channel of the excised patches was signif-
icantly inhibited by Cl⫺ channel blockers (Fig. 6). SITS (1
mM) produced a complete inhibition of maxi-Cl⫺ channel in 4
of 5 experiments, and the remaining preparation showed a
flickery inhibition at all potentials. Like SITS, NPPB (100 and
500 ␮M) caused a flickery/complete inhibition of cAMP-
activated maxi-Cl⫺ channels. In most cases, the inhibition was Fig. 5. A typical experiment showing the reversible effects of cytoplasmic
totally reversible. cAMP on mean patch current in an excised inside-out patch (Vm ⫽ ⫺80 mV).
The presence of maxi-Cl⫺ channels was also detected in The increase of patch current is mediated by an increase in Po but not channel’s
conductance.
outside-out patches although its occurrence was considerably
lower than that of inside-out patches. Like excised inside-out
patches, the addition of the plasma-membrane permeable
cAMP-analog 8-Br-cAMP (500 ␮M) to the bath activated the maxi-Cl⫺ channels. This was done by reducing cytoplasmic
maxi-Cl⫺ channel in symmetrical 130 mM NaCl solutions. The NaCl concentration of inside-out patches from 130 to either 65
channel displayed a conductance of 271 ⫾ 4 pS (n ⫽ 14), or 30 mM while maintaining extracellular NaCl concentration
which was not different from that of the inside-out patches. constant at 130 mM. Decreasing cytoplasmic Cl⫺ concentra-
Similarly, the addition of 100 ␮M NPPB significantly inhibited tion from 130 to either 65 or 30 mM caused a stepwise
the maxi-Cl⫺ channels by decreasing Po (data not shown). inhibition of the maxi-Cl⫺ currents in the presence of cyto-
Taken together, our results indicated the presence of cAMP- plasmic 500 ␮M cAMP. The inhibition could be explained
activated maxi-Cl⫺ channels in PE cells. partly by a reduction of Po at all potentials. The inhibition was,
Effects of cytoplasmic Cl⫺ concentrations on maxi-Cl⫺ however, more pronounced when Vm was negative. The plots
channel activity. As noted above, PE cells are primarily in- of Po as a function of Vm under different cytoplasmic Cl⫺
volved in Cl⫺ uptake, thereby increasing the intracellular Cl⫺ concentrations (130, 65, and 30 mM) are summarized in Fig. 3.
concentration. The increased Cl⫺ might either diffuse to the In addition to the reduction of Po, cytoplasmic Cl⫺ concentra-
NPE cells for subsequent Cl⫺ secretion into the posterior tion reduced channel conductance. The decrease in Cl⫺ chan-
chamber or return to the ciliary stroma through Cl⫺ efflux nel conductance was observed with both inward and outward
pathways. We tested whether the intracellular Cl⫺ concentra- currents, the reduction being greater for inward currents (Fig.
tion itself might modify the behavior of the cAMP-activated 7). Reducing cytoplasmic Cl⫺ concentration from 130 to 30
mM lowered Cl⫺ conductance by ⬃30% when Vm was posi-
tive, whereas the reduction was ⬃70% when Vm was negative.
Taken together, decreasing the cytoplasmic Cl⫺ concentration
caused a significant inhibition of maxi-Cl⫺ channels in excised

Fig. 4. Concentration-response relationship for cytoplasmic cAMP on maxi-

Cl⫺ channel Po when Vm ⫽ ⫺80 mV. The Cl⫺ concentrations in the
micropipette and bath were 130 mM. Averages were obtained from all patches
that were studied under both control and cAMP-perfused conditions and where
open events were detected at ⫺80 mV either under control or experimental or
both conditions, whether or not openings were observed at other potentials.
Number in parentheses is number of patches studied.
Fig. 6. Effects of the Cl⫺ channel blockers SITS (n ⫽ 5) and 5-nitro-2-
(phenylpropylamino)benzoate (NPPB) (n ⫽ 5) on the Po of maxi-Cl⫺ channels
activated by 500 ␮M cAMP in inside-out patches (Vm ⫽ ⫺80 mV). **P ⬍
0.01.

AJP-Cell Physiol • VOL 287 • OCTOBER 2004 • www.ajpcell.org

 ACTIVATION OF MAXI-CL⫺ CHANNELS IN BOVINE PE CELLS
Fig. 7. Effects of reduced cytoplasmic Cl⫺ concentration on cAMP-stimulated
unitary single-channel currents. In the presence of cytoplasmic cAMP (500
␮M), the concentration of NaCl at the cytoplasmic side of inside-out patches
was reduced from 130 mM to either 65 or 30 mM, whereas the extracellular
NaCl concentration was kept constant at 130 mM. Data points are usually
means of 6 –12 patches, because channels were not always open long enough
for precise unitary-current measurement at all potentials.
inside-out patches, which could be ascribed to reductions in
both Po and channel conductance.
Reducing cytoplasmic NaCl concentration shifted the rever-
sal potential, in addition to inhibiting maxi-Cl⫺-channel activ-
ity. As shown in Fig. 7, the reversal potential was close to zero
(⫺0.7 mV) in excised inside-out patches under symmetrical
130 mM NaCl but shifted to ⫺9.1 and ⫺15.0 mV when the
cytoplasmic NaCl concentration was reduced to 65 and 30
mM, respectively. The negative shift of the reversal potential
toward the Cl⫺ equilibrium potential indicated that the channel
was ⬃3.0 times more permeable to Cl⫺ than Na⫹. Taking into
consideration the changes in junction potential, the selectivity
was estimated to be ⬃6. The present estimate is comparable to
the values of 3.7 to 10 reported by many investigators (5, 28,
29, 39, 45), but higher selectivities have also been reported (3,
22, 44). The basis for this broad range of apparent selectivities
is unclear.
Stimulation of whole cell Cl⫺ currents by 8-Br-cAMP. To
learn whether cAMP-activated maxi-Cl⫺ channels contributed
significantly to PE-cell macroscopic currents, we studied the
effects of 8-Br-cAMP and of the Cl⫺ channel blocker NPPB on
whole cell current measurements. In symmetrical 140 mM
NMDGCl, the addition of 500 ␮M 8-Br-cAMP reversibly
stimulated the whole cell currents at all potentials (Fig. 8A). It
was noted that the cAMP-activated currents displayed less
outward rectification than did the baseline currents (Fig. 8B),
suggesting that the cAMP-activated conductances were differ-
ent from the baseline Cl⫺ conductances. These measurements
were obtained with our standard whole cell protocol involving
step changes in voltage of relatively brief duration (300 ms).
For comparison with the single-channel records, we also mea-
sured whole cell currents with voltage steps of longer duration.
With the prolonged voltage pulses of 4 s, the cAMP-activated
whole cell currents showed significant current inactivation at
both ⫹80 (P ⬍ 0.01) and ⫺80 mV (P ⬍ 0.05) but not at ⫺40
and ⫹40 mV (P ⬎ 0.05, n ⫽ 12). A typical experiment
showing the voltage-dependent whole cell current inactivations
AJP-Cell Physiol • VOL 287 • OCTOBER 2004 •
C1007
is shown in Fig. 9. This is consistent with the channel inacti-
vation due to Vm-dependent closing of maxi-Cl⫺ channels as
displayed in excised inside-out patches (Fig. 3). These results
strongly suggest that the cAMP-activated Cl⫺ channels ob-
served with excised patches contribute currents that can be
detected by macroscopic measurements, as well.
Regulation of whole cell cAMP-activated Cl⫺ currents by
cytoplasmic Cl⫺ concentration. As demonstrated earlier, the
activities of cAMP-activated maxi-Cl⫺ channels were modi-
fied by the Cl⫺ concentration at the cytoplasmic surface of
excised patches (Figs. 3 and 7). We tested whether the stimu-
lation of cAMP-activated whole cell currents is also sensitive
to changes of cytoplasmic Cl⫺ concentration. At constant bath
Cl⫺ concentration of 140 mM, reducing micropipette (cyto-
plasmic) Cl⫺ concentration from 140 to 30 mM decreased the
magnitude of cAMP-activated currents (Fig. 10). When the
cytoplasmic solution contained 140 mM Cl⫺, the inward cur-
rent was increased by 48 ⫾ 6% over the baseline values at
Vm ⫽ ⫺80 mV (n ⫽ 14). The stimulation was approximately
twofold larger than a separate set of experiments with 30 mM
Cl⫺ in the micropipette (25 ⫾ 4%, n ⫽ 6). In both cases, the
addition of NPPB to the bath completely blocked the cAMP-
activated current stimulation. Similarly, the stimulation of
outward current was approximately twofold larger when higher
cytoplasmic Cl⫺ concentration was used, increasing from 14 ⫾
2% (30 mM Cl⫺, n ⫽ 6) to 34 ⫾ 5% (140 mM Cl⫺, n ⫽ 14)
at ⫹80 mV. These results indicated that the cAMP-activated
whole cell Cl⫺ conductance could be modulated by cytoplas-
mic Cl⫺ concentration, consistent with the results from excised
inside-out patches.
cAMP-triggered shrinkage of PE cells. The electrophysio-
logical data indicated that cAMP activated Cl⫺ channels so
that we wondered whether cAMP-activated Cl⫺ efflux was
sufficiently great to produce volume changes in the PE cells.
Under isotonic conditions, the addition of 8-Br-cAMP (500
␮M) to the extracellular bath produced a gradual, time-depen-
dent shrinkage of PE cells (Fig. 11). 8-Br-cAMP shrank cells
by 6.3 ⫾ 1.3% over a period of 25 min (n ⫽ 18). We tested
whether the shrinkage was mediated by Cl⫺ release through
cAMP-activated Cl⫺ channels. In a separate set of experi-
ments, 100 ␮M NPPB significantly blocked cAMP-triggered
PE cell shrinkage; in the presence of the Cl⫺-channel blocker,
cells shrank by 0.9 ⫾ 1.5% by 25 min (n ⫽ 13). Subsequent
removal of NPPB from these preparations restored cAMP-
triggered cell shrinkage similar to that produced in cells with-
out preexposure to NPPB (data not shown).
DISCUSSION
The major new findings of the present study are as follows:
1) cAMP reversibly activates maxi-Cl⫺ channels of excised
patches from native bovine PE cells, 2) the Cl⫺ concentration
on the cytoplasmic surface triggers modification in these chan-
nels, and 3) the cAMP-activated maxi-Cl⫺ channels transfer
sufficient Cl⫺ to contribute significantly to whole cell currents
and participate in cell-volume regulation.
Maxi-Cl⫺ channels share many similarities with VDAC
including unitary channel conductance, multiple subconduc-
tance states, linear current-voltage (I-V) relationship, voltage
dependence, anion selectivity, and sensitivity to Cl⫺ channel
blockers (6, 33, 42, 44). The molecular identity of maxi-Cl⫺
www.ajpcell.org
C1008 ACTIVATION OF MAXI-CL⫺ CHANNELS IN BOVINE PE CELLS
Fig. 8. Effects of the cell membrane-permeable cAMP analog
8-bromo-cAMP (8-Br-cAMP; 500 ␮M) and of the Cl⫺-channel
blocker NPPB (100 ␮M) on whole cell currents. A: represen-
tative experiment showing the effects of 500 ␮M 8-Br-cAMP
on whole cell currents. Voltage pulses were from ⫺100 to ⫹80
mV in 20-mV steps, and each voltage pulse lasted 300 ms.
Upward current deflections indicated outward currents and vice
versa. The Cl⫺ concentrations in the micropipette and in the
bath were 140 mM. B: Current-voltage (I-V) relationship gen-
erated from the experiment in (A).
channels is likely porin-1 (VDAC-1, porin-31HL), which has
been localized in the plasma membrane of cultured T-lympho-
blastic leukemia CEM cells (2) and astrocytes (15). Although
plasma-membrane maxi-Cl⫺ channels have been recognized in
many cells (6), including PE cells (38) for two decades, their
physiological roles have been uncertain (1, 44, 47, 50), pri-
marily because of the voltage-dependence of their activity.
Their maximum Po is displayed within a narrow voltage range
(⫺20 to ⫹20 mV), far from the Vm of many cells.
It has been shown that Vm of both PE and NPE cells is
approximately ⫺60 mV in rabbit and shark ciliary epithelium
(9, 26, 53). A membrane-permeant form of cAMP, 8-Br-
cAMP, was reported to depolarize porcine ciliary epithelium
by ⬃10 mV (23). That action could reflect activation of Cl⫺
channels of either the NPE (10, 11, 20) or of PE cells (24).
Consistent with the observation in transformed cultured cells
(24), we have now documented that 8-Br-cAMP activates Cl⫺
channels of native bovine PE cells, facilitating Cl⫺ release and
thereby reducing the cell volume. At least one of the channel
targets is the maxi-Cl⫺ channel and the activation is likely to
be direct, because perfusion with cAMP activates the channels
AJP-Cell Physiol • VOL
in excised patches. Interestingly, cAMP has been recently
reported (16) to have no effect on baseline whole cell currents
of C1300 mouse neuroblastoma cells but to block the anties-
trogen-induced activation of maxi-Cl⫺ channels; the inhibitory
effect of cAMP was prevented by staurosporine. The basis for
the diverse effects of cAMP on the two cell types is not
apparent but might reflect differential expression of two dis-
tinct pathways for cAMP to modulate maxi-Cl⫺ channels,
either by direct activation or by indirect inhibition through
protein kinases (49).
Several lines of evidence suggest that the cAMP-activated
maxi-Cl⫺ channels observed with excised patches contribute
significantly to PE whole cell currents. First, both unitary
maxi-Cl⫺ and whole cell currents are stimulated by the addi-
tion of cAMP, and the effects are prevented by NPPB. Second,
measured as the mean value of initial 300 ms, cAMP-stimu-
lated whole cell currents displayed a more linear I-V relation-
ship than did baseline currents, consistent with the linear
relationship displayed by cAMP-activated maxi-Cl⫺ channels.
Third, the cAMP-activated unitary maxi-Cl⫺ and whole cell
currents are both dependent on cytoplasmic Cl⫺ concentration,
287 • OCTOBER 2004 • www.ajpcell.org
 ACTIVATION OF MAXI-CL⫺ CHANNELS IN BOVINE PE CELLS
Fig. 10. Effect of cytoplasmic Cl⫺ concentration on cAMP-activated whole
cell currents. Pulses were of brief duration (300 ms) at Vm ⫽ ⫺80 mV. The
Cl⫺ concentration in the external bath was always 140 mM. Currents have
been normalized and averaged. Mean ⫾ SE values are shown by vertical bars.
The absolute baseline currents with 30 and 140 mM NaCl were ⫺72 ⫾ 10 and
⫺70 ⫾ 15 pA 䡠 pF⫺1 (P ⫽ 0.96), respectively. **P ⬍ 0.01.
AJP-Cell Physiol • VOL 287 • OCTOBER 2004 •
C1009
Fig. 9. Time courses of whole cell cAMP-stimulated currents
following step changes in Vm. Voltage pulses were from ⫺80 to
⫹80 mV in 40-mV steps and each voltage step was 4 s, as in
excised patches (Vh ⫽ 0 mV). The Cl⫺ concentrations in the
micropipette and in the bath were identical (140 mM). A:
representative experiment showing the time-dependent whole
cell cAMP-stimulated currents inactivating at ⫺80 and ⫹80
mV but not at ⫺40 and ⫹40 mV. B: I-V relationship from the
experiment in (A). Currents were calculated from the mean
values of the initial 500 ms (early endpoint) and also from the
last 500 ms (late endpoint) in each 4-s voltage pulse at differ-
ent Vm.
Fig. 11. Effects of 8-Br-cAMP on PE-cell volume. 8-Br-cAMP (500 ␮M)
triggered a gradual, time-dependent shrinkage of native bovine PE cells under
isosmotic conditions (n ⫽ 18). The SE values are shown by vertical bars. *P ⬍
0.05; **P ⬍ 0.01.
www.ajpcell.org
C1010 ACTIVATION OF MAXI-CL⫺ CHANNELS IN BOVINE PE CELLS
displaying an increase of currents at higher intracellular Cl⫺
concentration. Fourth, time- and voltage-dependent current
inactivations are observed at Vm of ⫾ 80 mV and not at ⫾ 40
mV in both excised single-channel and whole cell measure-
ments, consistent with the typical characteristics of maxi-Cl⫺
channels. In addition to maxi-Cl⫺ channels, it is possible that
cAMP enhances whole cell currents by activating other Cl⫺
channels in PE cells. However, if an additional channel is
targeted, it is unlikely to be CFTR, because Northern blot
analysis has not detected mRNA for CFTR in human ciliary
body (13). Moreover, patients with cystic fibrosis demonstrate
normal aqueous flow rate compared with healthy subjects (34).
The cAMP-activated maxi-Cl⫺ channels observed in the
present study were strongly influenced by cytoplasmic Cl⫺
concentration. Raising intracellular Cl⫺ concentration en-
hanced both inward and outward currents by a dual effect on Po
and channel conductance. Under physiological conditions, the
PE-cell cytoplasmic Cl⫺ concentration is determined by the
rate of uptake from the stroma by electroneutral transporters
and the rate of Cl⫺ diffusion through gap junctions to the NPE
cells and final release into the aqueous humor. Thus the Cl⫺
concentration is expected to rise when the rate of PE-cell Cl⫺
uptake from the stroma exceeds the NPE-cell transport capac-
ity to secrete Cl⫺ into the aqueous humor. The rise in PE-cell
Cl⫺ concentration, accompanied by a rise in the counter-ion
Na⫹, is predicted to present an osmotic gradient for water and
cell swelling. Hypotonic swelling has been found to trigger
NPPB-sensitive ATP release from cultured bovine PE cells
(36). The ATP can occupy P2Y2 receptors, initiating sequen-
tial increases in free Ca2⫹ concentration, phospholipase A2
activity, PGE2 formation and release, occupancy of PGE2
receptors, formation of cAMP in Madin-Darby canine kidney
epithelial cells (40), and transformed bovine PE cells (24). In
other words, Cl⫺-overloaded PE cells would be expected to
swell, triggering a cascade of events leading to cAMP forma-
tion, and thereby activating the maxi-Cl⫺ channels directly. In
agreement with the electrophysiological measurements, cAMP
triggers a reduction of cell volume, which can be prevented by
pretreatment with NPPB. Increased fluid transfer from the PE
cells to the ciliary stroma is expected to reduce net aqueous
humor secretion.
Although Cl⫺ overload could provide one source of delivery
of cAMP to the PE cells through an ATP-release-triggered
cascade, other mechanisms may prove of even greater physi-
ological importance. For example, PE cells express ␤2-adren-
ergic receptors (21) so that sympathetic nerve endings located
close to the ciliary epithelium in the ciliary processes could
stimulate cAMP production in these cells. Multiple biologi-
cally active peptides are also thought to be released from those
nerve endings (48). Included among the neuropeptides are VIP
and substance P, both of which trigger increased production of
cAMP (e.g., see Refs. 31 and 43).
The focus of the current work has been on the cAMP
activation of the PE cell maxi-Cl⫺ channels and their potential
physiological role in modifying net Cl⫺ secretion by the ciliary
epithelium. However, strong published evidence indicates that
maxi-Cl⫺ channels also mediate swelling-triggered ATP re-
lease by mouse mammary C127 cells (19, 44). Consistent with
their findings, and in contrast to the cAMP insensitivity of
release by transformed bovine PE cells (36), we have recently
found that 8-Br-cAMP (500 ␮M) enhances isotonic ATP
AJP-Cell Physiol • VOL
release by native bovine PE cells by ⬃50% (Do CW, Reigada
D, Mitchell CH, and Civan MM, unpublished data). The
potential role of maxi-Cl⫺-mediated ATP release in regulating
aqueous humor formation is currently under study.
GRANTS
The work was supported by National Eye Institute research Grants EY-
08343 (to M. M. Civan) and EY-13434 (to C. H. Mitchell), and Core Grant
EY-01583.
REFERENCES
1. Bathori G, Parolini I, Szabo I, Tombola F, Messina A, Oliva M,
Sargiacomo M, De Pinto V, and Zoratti M. Extramitochondrial porin:
facts and hypotheses. J Bioenerg Biomembr 32: 79 – 89, 2000.
2. Bathori G, Parolini I, Tombola F, Szabo I, Messina A, Oliva M, De
Pinto V, Lisanti M, Sargiacomo M, and Zoratti M. Porin is present in
the plasma membrane where it is concentrated in caveolae and caveolae-
related domains. J Biol Chem 274: 29607–29612, 1999.
3. Becq F, Fanjul M, Mahieu I, Berger Z, Gola M, and Hollande E.
Anion channels in a human pancreatic cancer cell line (Capan-1) of ductal
origin. Pflügers Arch 420: 46 –53, 1992.
4. Berrier C, Besnard M, and Ghazi A. Electrophysiological characteris-
tics of the PhoE porin channel from Escherichia coli. Implications for the
possible existence of a superfamily of ion channels. J Membr Biol 156:
105–115, 1997.
5. Bertran GC and Kotsias BA. Chloride channels in toad skeletal muscle
fibers. J Exp Zool 287: 423– 431, 2000.
6. Blatz AL and Magleby KL. Single voltage-dependent chloride-selective
channels of large conductance in cultured rat muscle. Biophys J 43:
237–241, 1983.
7. Caprioli J and Sears M. Forskolin lowers intraocular pressure in rabbits,
monkeys, and man. Lancet 1: 958 –960, 1983.
8. Caprioli J, Sears M, Bausher L, Gregory D, and Mead A. Forskolin
lowers intraocular pressure by reducing aqueous inflow. Invest Ophthal-
mol Vis Sci 25: 268 –277, 1984.
9. Carre DA, Tang CS, Krupin T, and Civan MM. Effect of bicarbonate
on intracellular potential of rabbit ciliary epithelium. Curr Eye Res 11:
609 – 624, 1992.
10. Chen S, Inoue R, Inomata H, and Ito Y. Role of cyclic AMP-induced Cl
conductance in aqueous humour formation by the dog ciliary epithelium.
Br J Pharmacol 112: 1137–1145, 1994.
11. Chen S and Sears M. A low conductance chloride channel in the
basolateral membranes of the non-pigmented ciliary epithelium of the
rabbit eye. Curr Eye Res 16: 710 –718, 1997.
12. Civan MM. Transport by the ciliary epithelium of the eye. News Physiol
Sci 12: 158 –162, 1997.
13. Coca-Prados M, Anguita J, Chalfant ML, and Civan MM. PKC-
sensitive Cl⫺ channels associated with ciliary epithelial homologue of
pICln. Am J Physiol Cell Physiol 268: C572–C579, 1995.
14. Crook RB, Takahashi K, Mead A, Dunn JJ, and Sears ML. The role
of NaKCl cotransport in blood-to-aqueous chloride fluxes across rabbit
ciliary epithelium. Invest Ophthalmol Vis Sci 41: 2574 –2583, 2000.
15. Dermietzel R, Hwang TK, Buettner R, Hofer A, Dotzler E, Kremer M,
Deutzmann R, Thinnes FP, Fishman GI, Spray DC, and Siemen D.
Cloning and in situ localization of a brain-derived porin that constitutes a
large-conductance anion channel in astrocytic plasma membranes. Proc
Natl Acad Sci USA 91: 499 –503, 1994.
16. Diaz M, Bahamonde MI, Lock H, Munoz FJ, Hardy SP, Posas F, and
Valverde MA. Okadaic acid-sensitive activation of maxi Cl⫺ channels by
triphenylethylene antioestrogens in C1300 mouse neuroblastoma cells.
J Physiol 536: 79 – 88, 2001.
17. Do CW and Civan MM. Basis of chloride transport in ciliary epithelium.
J Membr Biol. In press.
18. Do CW and To CH. Chloride secretion by bovine ciliary epithelium: a
model of aqueous humor formation. Invest Ophthalmol Vis Sci 41: 1853–
1860, 2000.
19. Dutta AK, Okada Y, and Sabirov RZ. Regulation of an ATP-conductive
large-conductance anion channel and swelling-induced ATP release by
arachidonic acid. J Physiol 542: 803– 816, 2002.
20. Edelman JL, Loo DD, and Sachs G. Characterization of potassium and
chloride channels in the basolateral membrane of bovine nonpigmented
ciliary epithelial cells. Invest Ophthalmol Vis Sci 36: 2706 –2716, 1995.
287 • OCTOBER 2004 • www.ajpcell.org
 ACTIVATION OF MAXI-CL⫺ CHANNELS IN BOVINE PE CELLS
21. Elena PP, Fredj-Reygrobellet D, Moulin G, and Lapalus P. Pharma-
cological characteristics of ␤-adrenergic-sensitive adenylate cyclase in
non pigmented and in pigmented cells of bovine ciliary process. Curr Eye
Res 3: 1383–1389, 1984.
22. Falke LC and Misler S. Activity of ion channels during volume regula-
tion by clonal N1E115 neuroblastoma cells. Proc Natl Acad Sci USA 86:
3919 –3923, 1989.
23. Fleischhauer JC, Beny JL, Flammer J, and Haefliger IO. Cyclic AMP
and anionic currents in porcine ciliary epithelium. Klin Monatsbl Augen-
heilkd 218: 370 –372, 2001.
24. Fleischhauer JC, Mitchell CH, Peterson-Yantorno K, Coca-Prados M,
and Civan MM. PGE2, Ca2⫹, and cAMP mediate ATP activation of Cl⫺
channels in pigmented ciliary epithelial cells. Am J Physiol Cell Physiol
281: C1614 –C1623, 2001.
25. Gaasterland D, Kupfer C, Ross K, and Gabelnick HL. Studies of
aqueous humor dynamics in man. III. Measurements in young normal
subjects using norepinephrine and isoproterenol. Invest Ophthalmol 12:
267–279, 1973.
26. Green K, Bountra C, Georgiou P, and House CR. An electrophysi-
ologic study of rabbit ciliary epithelium. Invest Ophthalmol Vis Sci 26:
371–381, 1985.
27. Hurnak O and Zachar J. Maxi chloride channels in L6 myoblasts. Gen
Physiol Biophys 11: 389 – 400, 1992.
28. Hurnak O and Zachar J. Conductance-voltage relations in large-con-
ductance chloride channels in proliferating L6 myoblasts. Gen Physiol
Biophys 13: 171–192, 1994.
29. Hurnak O and Zachar J. Selectivity of maxi chloride channels in the L6
rat muscle cell line. Gen Physiol Biophys 14: 91–105, 1995.
30. Jacob TJ and Civan MM. Role of ion channels in aqueous humor
formation. Am J Physiol Cell Physiol 271: C703–C720, 1996.
31. Karaki SI and Kuwahara A. Regulation of intestinal secretion involved
in the interaction between neurotransmitters and prostaglandin E2. Neu-
rogastroenterol Motil 16, Suppl 1: 96 –99, 2004.
32. Lee PY, Podos SM, Mittag T, and Severin C. Effect of topically applied
forskolin on aqueous humor dynamics in cynomolgus monkey. Invest
Ophthalmol Vis Sci 25: 1206 –1209, 1984.
33. McCann FV, McCarthy DC, Keller TM, and Noelle RJ. Characteriza-
tion of a large conductance non-selective anion channel in B lymphocytes.
Cell Signal 1: 31– 44, 1989.
34. McCannel CA, Scanlon PD, Thibodeau S, and Brubaker RF. A study
of aqueous humor formation in patients with cystic fibrosis. Invest Oph-
thalmol Vis Sci 33: 160 –164, 1992.
35. McLaughlin CW, Peart D, Purves RD, Carre DA, Macknight AD, and
Civan MM. Effects of HCO⫺ 3 on cell composition of rabbit ciliary
epithelium: a new model for aqueous humor secretion. Invest Ophthalmol
Vis Sci 39: 1631–1641, 1998.
36. Mitchell CH, Carre DA, McGlinn AM, Stone RA, and Civan MM. A
release mechanism for stored ATP in ocular ciliary epithelial cells. Proc
Natl Acad Sci USA 95: 7174 –7178, 1998.
37. Mitchell CH, Fleischhauer JC, Stamer WD, Peterson-Yantorno K,
and Civan MM. Human trabecular meshwork cell volume regulation.
Am J Physiol Cell Physiol 283: C315–C326, 2002.
38. Mitchell CH, Wang L, and Jacob TJC. A large-conductance chloride
channel in pigmented ciliary epithelial cells activated by GTP␥S. J Membr
Biol 158: 167–175, 1997.
AJP-Cell Physiol • VOL 287 • OCTOBER 2004 •
C1011
39. O’Donnell MJ, Kelly SP, Nurse CA, and Wood CM. A maxi Cl⫺
channel in cultured pavement cells from the gills of the freshwater rainbow
trout Oncorhynchus mykiss. J Exp Biol 204: 1783–1794, 2001.
40. Ostrom RS, Gregorian C, and Insel PA. Cellular release of and response
to ATP as key determinants of the set-point of signal transduction
pathways. J Biol Chem 275: 11735–11739, 2000.
41. Raviola G and Raviola E. Intercellular junctions in the ciliary epithelium.
Invest Ophthalmol Vis Sci 17: 958 –981, 1978.
42. Reymann S, Florke H, Heiden M, Jakob C, Stadtmuller U, Steinacker
P, Lalk VE, Pardowitz I, and Thinnes FP. Further evidence for multi-
topological localization of mammalian porin (VDAC) in the plasmalemma
forming part of a chloride channel complex affected in cystic fibrosis and
encephalomyopathy. Biochem Mol Med 54: 75– 87, 1995.
43. Roush ED and Kwatra MM. Human substance P receptor expressed in
Chinese hamster ovary cells directly activates G␣q/11, G␣s, and G␣o. FEBS
Lett 428: 291–294, 1998.
44. Sabirov RZ, Dutta AK, and Okada Y. Volume-dependent ATP-conduc-
tive large-conductance anion channel as a pathway for swelling-induced
ATP release. J Gen Physiol 118: 251–266, 2001.
45. Schwiebert EM, Mills JW, and Stanton BA. Actin-based cytoskeleton
regulates a chloride channel and cell volume in a renal cortical collecting
duct cell line. J Biol Chem 269: 7081–7089, 1994.
46. Shahidullah M, Wilson WS, and Millar C. Effects of timolol, terbutaline
and forskolin on IOP, aqueous humour formation and ciliary cyclic AMP
levels in the bovine eye. Curr Eye Res 14: 519 –528, 1995.
47. Shimizu S, Matsuoka Y, Shinohara Y, Yoneda Y, and Tsujimoto Y.
Essential role of voltage-dependent anion channel in various forms of
apoptosis in mammalian cells. J Cell Biol 152: 237–250, 2001.
48. Stone RA. Nervous system and intraocular pressure. In: The Glaucomas
(2nd ed.), edited by Ritch R, Shields MB, and Krupin T. St. Louis, MO:
Mosby, 1996, p. 357–383.
49. Strange K, Emma F, and Jackson PS. Cellular and molecular physiol-
ogy of volume-sensitive anion channels. Am J Physiol Cell Physiol 270:
C711–C730, 1996.
50. Thinnes FP and Reymann S. New findings concerning vertebrate porin.
Naturwissenschaften 84: 480 – 498, 1997.
51. To CH, Do CW, Zamudio AC, and Candia OA. Model of ionic
transport for bovine ciliary epithelium: effects of acetazolamide and
HCO⫺ 3 . Am J Physiol Cell Physiol 280: C1521–C1530, 2001.
52. Wiederholt M, Helbig H, and Korbmacher C. Ion transport across the
ciliary epithelium: lessons from cultured cells and proposed role of the
carbonic anhydrase. In: Carbonic Anhydrase, edited by Botre F, Gross G,
and Storey BT. New York: VCH, 1991, p. 232–244.
53. Wiederholt M and Zadunaisky JA. Membrane potentials and intracel-
lular chloride activity in the ciliary body of the shark. Pflügers Arch 407,
Suppl 2: S112–S115, 1986.
54. Yantorno RE, Carre DA, Coca-Prados M, Krupin T, and Civan MM.
Whole cell patch clamping of ciliary epithelial cells during anisosmotic
swelling. Am J Physiol Cell Physiol 262: C501–C509, 1992.
55. Zhang JJ and Jacob TJ. Three different Cl⫺ channels in the bovine
ciliary epithelium activated by hypotonic stress. J Physiol 499: 379 –389,
1997.
www.ajpcell.org