Full text of DOI:10.1007/s004249900223

Pflügers Arch – Eur J Physiol (2000) 439:559–566
Digital Object Identifier (DOI) 10.1007/s004249900223
O R I G I N A L A RT I C L E
A. Rabe · E. Frömter
Micromolar concentrations of steroids and of aldosterone antagonists
inhibit the outwardly rectifying chloride channel with different kinetics
Received: 22 October 1999 / Received after revision and accepted: 15 November 1999 / Published online: 22 December 1999
© Springer-Verlag 2000
Key words ORIC Cl– channel · Open/close kinetics ·
Three closed time constants · Steroids · Aldosterone
antagonists · Cytosolic Cl– channel inhibitor
Abstract We used the patch-clamp technique to analyse
the open/close kinetics of single, outwardly rectifying,
intermediate-conductance (ORIC) Cl– channels from cul-
tured epithelial cells under control conditions and in
presence of different inhibitors. As observed previously
in excised inside/out patches under control conditions,
the switching kinetics were characterized by one open-
state time constant (τo≈30 ms) and three closed-state
time constants (τc1≈0.2 ms, τc2≈2 ms and τc3≈60 ms). Al-
dosterone, six further steroids and two aldosterone an-
tagonists inhibited channel open probability (NPo) con-
centration dependently with the potency at 10 µmol/l in-
creasing in the sequence: hydrocortisone, corticosterone,
β-oestradiol, cortisone, aldosterone, testosterone, proges-
terone, canrenone, spironolactone. Although all sub-
stances decreased τo, neither the steroids nor the aldo-
sterone antagonists affected τc1, τc2 or τc3 or induced ad-
ditional transitions with additional time constants. In-
stead, the steroids increased the prevalence of τc2 in the
dwell-time histograms and the aldosterone antagonists
increased the prevalence of τc3, both in a concentration-
dependent manner. These observations may be explained
by a model in which one open state leads to one of three
closed states with rate constants α, β and γ, and in which
β or γ increase under the influence of steroids or aldo-
sterone antagonists, respectively. Cytosol, which con-
tains a Cl– channel inhibitor of unknown molecular
structure, (Krick et al., Pflügers Arch 418:491, 1991)
was also tested, but the results did not conform to the
blocker mechanisms described above. This shows that
there are even further modes of channel inhibition and
argues against the cytosolic Cl– channel inhibitor being a
steroid.
Introduction
Although its molecular identity and its physiological role
are not yet known, the outwardly rectifying, intermedi-
ate-conductance (ORIC) Cl– channel is a well-character-
ized ion channel widely distributed in epithelial and non-
epithelial cells [2, 6]. Besides the typical current/voltage
relation in symmetrical solutions with plasma-like Cl–
concentration, it has further characteristic properties.
First, it is virtually only ever observed in excised patch-
es, and it opens either immediately upon excision (if the
experiment is performed at 37 °C [13]) or with some de-
lay after application of depolarizing voltages (if the ex-
periment is performed at room temperature [7, 26]). Sec-
ond, it displays typical switching kinetics with at least
one open state and three closed states [20] and third, it is
inhibited by a wide variety of substances including vari-
ous chloride channel blockers [9, 23, 24] and unsaturated
fatty acids such as arachidonic acid [1, 10]. Based on its
spontaneous activation after excision we have speculated
previously that the channel in intact cells probably is
blocked by a cytosolic inhibitor that diffuses away after
excision. Subsequently, and in collaboration with others
[11, 12, 14], we have succeeded in demonstrating the ex-
istence of such an inhibitory principle in cytoplasm. It
is presumably a low-molecular weight (approximately.
1000 Da), amphiphilic substance but its molecular iden-
tity is not yet known.
In the search for possible physiological inhibitor mol-
ecules we tested a number of substances, among them
steroids that block Cl– channels in brain [15]. Steroids do
indeed block the ORIC Cl– channel, albeit at higher con-
centrations than are expected inside cells. Irrespective of
their possible physiological role, however, we observed
an interesting blocking pattern, insofar as all seven ste-
roids tested shifted the channel into its second closed
A. Rabe · E. Frömter (
)
✉
Institut für Nieren- und Membranphysiologie
der J.W. Goethe-Universität, Theodor-Stern-Kai 7,
D-60590 Frankfurt/Main, Germany
e-mail: froemter@em.uni-frankfurt.de
Tel.: +49-69-63016093, Fax: +49-69-63017142

560
state, while the two aldosterone antagonists tested shift-
ed the channel into the third closed state. In addition we
observed that the potency of the steroids correlated well
with their lipophilicity.
Materials and methods
The experiments were performed on subconfluent monolayer cul-
tures of HT29 and T84 cells grown on collagen-coated cover-slips
and bathed in NaCl Ringer solution of plasma-like composition as
described previously [11, 20]. The bath solution consisted of
(in mmol/l): Na+ 145, K+ 5, Ca2+ 1, Mg2+ 1 Cl– 154 and N-(2-hy-
droxyethyl)piperazine-N’-(2-ethanesulphonic acid) (HEPES) 5
and was titrated with NaOH to pH 7.4. Conventional patch-clamp
techniques were used to measure single-channel currents in ex-
cised, inside-out or outside-out patches at room temperature, usu-
ally on the 1st or 2nd day after cell seeding. Patch pipettes consist-
ed of borosilicate glass tubing, were filled with 150 mmol/l NaCl
buffered with 5 mmol/l HEPES to pH 7.4, and had resistances of
5–15 MΩ. All measurements were performed with a patch-clamp
amplifier (EPC9, HEKA Lambrecht, Germany) in conjunction
with an Atari Mega ST4 computer. The data were digitized and
stored on video tape using a pulse-code modulation device (PCM
501, Sony) operating at 44.1 kHz with 16-bit resolution. They
were then replayed as analogue signals, passed through an 8-pole
Bessel filter (902 LPF, Frequency Devices, Haverhill, Mass.,
USA) and stored for further processing. For determination of open
probability (Po) of single channels or NPo (where N is the number
of simultaneously active channels) sampling rate and filter fre-
quency were set at 0.5 and 0.1 kHz respectively. Po and NPo were
calculated from amplitude histograms collected over at least 30 s
(typically ≈1 min) and changes were calculated by referring exper-
imental data to averages from control and recovery periods ob-
served on the same patch. For determination of open/close kinet-
ics, sampling rate and filter frequency were set at 5 and 1 kHz re-
spectively. Open and closed times were analysed with the TAC
program (Instrutec, Almond, N.Y., USA) which counts all transi-
tions of the mid level (50% threshold method) between open and
closed states and bridges the intervals between data points by
spline interpolation. Proper setting of open and closed levels was
monitored and, if necessary, adjusted over the entire length of all
records (typically 1 min or up to 32768 events, the maximum the
program can handle). The steroids were added to the bath perfu-
sate from stock solutions in dimethylsulphoxide (DMSO) to final
concentrations of 0.1–100 µmol/l. The maximal DMSO concentra-
tion was 0.1%. DMSO concentrations up to 1% did not affect NPo.
Cytosol was collected as described [12]. Data are given as mean
values±SD for n measurements. Where n=2, the mean±half the
difference is given. The significance of differences was tested with
Student’s t-test for paired or unpaired samples as appropriate.
Fig. 1 Effect of aldosterone (50 µmol/l) on an outwardly rectify-
ing, intermediate conductance (ORIC) Cl– channel in an inside-out
membrane patch from an HT29 cell. The holding potential was
–60 mV. Upper row: single-channel currents recorded at a sam-
pling rate of 500 Hz and a filter frequency of 100 Hz. Upward de-
flections of the current trace indicate channel opening. Middle
row: sections of upper trace at a higher time resolution (sampling
rate 5 kHz; filter setting 1 kHz). Lower row: amplitude histograms
used to calculate the open probability (Po) of the channel
Fig. 2 Concentration dependence of ORIC Cl– channel inhibition
by aldosterone (triangles) and spironolactone (circles). Reduction
of NPo as a percentage of control (mean of NPo of pre- and post-
experimental control periods) as function of inhibitor concentra-
tion on a logarithmic scale. Means±SD, n=2–7 or 5–7 respective-
ly. The data were fitted with single-site inhibition kinetics (solid
lines)
with virtually identical Cl– concentrations in pipette and
bath at a holding potential of –60 mV. Low- and high-
speed recordings are presented as well as current ampli-
tude histograms under control and experimental condi-
tions. As can be seen, aldosterone reversibly induced a
typical flicker block with reduction of Po from almost
0.9 to 0.4. The apparent slight contraction of the current
amplitude and the less clear-cut separation of open and
closed states in the presence of aldosterone reflect the
limited frequency resolution of these measurements.
Taking this effect into account there was no significant
change in the current/voltage relationship of the channel
in presence of the inhibitor. Figure 2 shows (among oth-
er data) the concentration dependence of the aldosterone
effect. It follows Michaelis-Menten kinetics with a concen-
tration for half-maximal inhibition (IC50) of 18.1 µmol/l.
Results
Channel inhibition by steroids
Single ORIC Cl– channels were observed after excision
in inside-out membrane patches either spontaneously or
after application of a brief hyperpolarizing pulse. As re-
ported previously, the Po for this channel is rather con-
stant with little indication of run-down, and is indepen-
dent of membrane voltage, ambient pH or [Ca2+] [20,
25]. Figure 1 shows the effect of 50 µmol/l aldosterone
on single-channel currents of an excised inside-out patch
of a HT29 cell. Essentially identical observations were
also made on T84 cells. This experiment was performed

561
Table 1 Inhibitory potency of
steroids and aldosterone antag-
onists (all 10 µmol/l).
Substance
Reduction
of NPo
n
OH groups
per molecule
logPoct/water
Means±SD for n observations.
Inhibition is expressed as the
percentage reduction in NPo (N
number of active channels, Po
single-channel open probabili-
ty, Poct/water octanol: water par-
tition coefficient)
Hydrocortisone
Corticosterone
β-Oestradiol
Cortisone
7±3
18±5
24±10
29±9
30±6
33±9
46±12
70±7
81±7
6
3
3
6
7
4
6
4
7
3
2
2
2
2
1
0
0
0
1.53–1.68a
1.82b
2.61a
1.42–1.64a
1.08b
Aldosterone
Testosterone
Progesterone
Canrenone
3.32a
2.48–3.87a
2.68c
aData from [8], bdata from [19],
cdata from [4], ddata from [16]
Spironolactone
2.19d
Fig. 4 Typical trace record of control single-channel activity at
high time resolution with indication of closed times as analysed
with the TAC program (see text for details)
Inhibition kinetics
The mechanism of channel inhibition was investigated
by analysing the switching kinetics with dwell-time his-
tograms. As reported previously for control conditions,
ORIC Cl– channels exhibit open/close kinetics which in-
dicate the presence of at least one open and three closed
states [20]. Under control conditions the channel opening
is characterized by one open time constant τo≈30 ms, and
channel closure by at least three closed time constants:
τc1≈0.25 ms, τc2≈2 ms and τc3≈60 ms, where τc1 – be-
cause of the limited frequency resolution of the measure-
ments – is only poorly defined although the underlying
transition phenomenon is a physical reality (see Fig. 4).
When closing, the channel prefers to enter the first
closed state (incidence approximately 70–90%) or the
second closed state (incidence approximately 10–25%),
but is only rarely observed in the third closed state (1%
or less).
Fig. 3 Effect of spironolactone (3 µmol/l) on ORIC Cl– channel.
Details as in Fig. 1
The observed Cl– channel inhibition is not restricted
to aldosterone, but can also be elicited by other steroids.
Six further steroids were tested. At concentrations of 1,
10 or 100 µmol/l they showed similar inhibition kinetics
as aldosterone, although the potencies differed. Rather
than determining the complete concentration dependence
of inhibition for each steroid we focussed on the effects
at 10 µmol/l. As shown in Table 1 the order of potency
was progesterone>testosterone>aldosterone>cortisone
oestradiol>corticosterone>hydrocortisone. Even hydro-
cortisone, however, which had only a small effect (7%
inhibition at 10 µmol/l), reduced NPo strongly (by 55%)
at a 10 times higher concentration.
Figure 5 shows the effect of increasing aldosterone
concentrations (10 and 100 µmol/l) on dwell-time histo-
grams analysed according to Sigworth and Sine [22]. Ta-
ble 2 summarizes the corresponding time constants ob-
tained on three single channels. It can be seen that in-
creasing concentrations of aldosterone decreased τo but
did not induce transitions into an additional closed state.
Rather aldosterone increased the probability of finding
the channel in the second closed state while the closing
kinetics (τc1, τc2, and τc3) remained virtually unchanged.
The same pattern of channel inhibition, which also
agrees with our previous observations with glibencla-
mide [20, 25], was observed with all steroids that could
be tested on single channel activity, i.e. – besides aldo-
sterone – progesterone, testosterone, cortisone and hy-
drocortisone (see also Table 3).
Figure 3 shows essentially identical experiments as
Fig. 1, but here, instead of aldosterone, the aldosterone
antagonist spironolactone (3 µmol/l) was used; it was
10 times more potent (IC50 2.1 µmol/l, see Fig. 2). A
comparison of Figs. 1 and 3 reveals that spironolactone
did not cause a fast flicker block but rather induced lon-
ger closed periods without eliciting the typical bursting
behaviour often seen in channels with multiple closed
states. Very similar observations were also made with the
aldosterone antagonist canrenone which reduced NPo
with an IC50 of 3.6 µmol/l.

562
Table 3 Effect of different steroids (10 or 100 µmol/l) on τo, τc2
and A2, the relative preference of switching into the second closed
state. τc1 and τc3 were not affected, nor did the ratio A1/A3 change.
Mean of n observations
Blocker
Concentration
(µmol/l)
n
τo
τc2
A2
(ms) (ms) (%)
Control
–
10
100
–
10
–
100
–
10
–
10
3
1
2
2
2
1
1
3
3
1
1
32
1.2
16
57
94
12
44
Aldosterone
7.7 2.2
0.8 2.5
15.3 1.6
5.5 1.6
Control
Cortisone
Control
Hydrocortisone
Control
Progesterone
Control
Testosterone
45
1.6
8.9
1.5 1.2
90
12
47
6.5
29
13
2.3
4.9 6.7
13
3.8
1.3
3
constants unchanged. However, it increased progressive-
ly the probability of finding the channel in the third
closed state. Essentially identical observations were
made with canrenone (see also Table 2).
Channel inhibition by cytosol
We have shown previously that cytosol isolated from cell
cultures, pig kidney cortex or human placenta [11, 12]
inhibits the ORIC Cl– channel either totally or partially
by inducing some kind of flicker block, and that by dilut-
Fig. 5A–C Effect of aldosterone on dwell-time histogram of ing the cytosol we obtain Michaelis-Menten-type inhibi-
ORIC Cl– channel. Open and closed time histograms are presented
according to Sigworth and Sine [22] with logarithmic binning of
or spironolactone above. From these and other observa-
time intervals on the abscissa and square-root scaling of the num-
ber of events (n) on the ordinate. A Control state, B and C in the
presence of 10 and 100 µmol/l aldosterone respectively
tion kinetics similar to those observed with aldosterone
tions we have concluded that cytosol contains one (or
more) natural, low-molecular weight Cl– channel inhibi-
tors of amphiphilic structure. To characterize further the
inhibitor we tested the effect of cytosol on Cl– channel
The aldosterone antagonists spironolactone and canre- switching kinetics. As seen clearly in Fig. 7 and Table 2,
none, on the other hand, produced very different dwell- cytosol inhibited the Cl– channel with a blocking mecha-
time histograms. As shown in Fig. 6 and Table 2, in- nism different to that of the steroids or the aldosterone
creasing concentrations of spironolactone also progres- antagonists. It also reduced the open time constant, but
sively decreased τo and also left the three closed time did not affect the probability of finding the channel in
Table 2 Effect of different inhibitors on single-channel kinetics. the relative preferences of dwelling in the closed states C1–C3.
Mean±SD for n single-channel experiments. Time is the total time Observations with 0.1 µmol/l spironolactone are not included be-
of channel activity analysed at each concentration. τo is the open cause the data were not different from control
time constant, τc1–τc3 are the closed time constants and A1–A3 are
Substance
Concentration
(µmol/l)
n
Time
(s)
τo
(ms)
τc1
(ms)
A1
(%)
τc2
(ms)
A2
(%)
τc3
(ms)
A3
(%)
Control
–
10
100
–
3
1
2
2
2
2
1
1
5
6
175
74
32±12
7
0.15±0.02
0.23
83±12
43
5.9±0.2
91±2.2
82±12
50±28
88
67
78±10
78±10
1.2±0.5
2.2
16±13
57
46±27
50
85±62
55±1
61±9.2 8.6±7.7
60±6.7 42±25
56
36
84±54
77±23
1±0.5
0.1
Aldosterone
Aldosterone
Control
93
0.84±0.07 0.26±0.03
2.2±1.3
94
0.2
1404
209
200
135
89
261
271
40±20
28±22
8.0±2.0
29
5.7
29±12
6.7±3
0.13±0.02
0.13±0.04
0.13±0.05
0.15
0.11
0.22±0.06
0.37±0.1
1.1±0.05 7.5±1.2
1.9±1.0
Spironolactone
Spironolactone
Control
Canrenone
Control
Cytosol
1
1.0±0.1
1.1±0.9
1.7
1.3
1.8±0.4
4.6±2.0
9.5±3.5
8.0±2.8
9.3
10
–
10
–
–
2.8
28
2.7±1.7
3.5±1.5
5
19±9
19±9

563
one of its natural closed states. Rather, cytosol appeared
to induce (at least) one new transition phenomenon with
an additional closed time constant located between τc1
and τc2. Because of the close proximity to τc1, and be-
cause of the limited high-frequency resolution of the
measurements, however, the significance of this interpre-
tation is not as clear as in the blocker experiments with
steroids and aldosterone antagonists.
Discussion
The present experiments have shown that the ORIC Cl–
channel is blocked by micromolar concentrations of ste-
roids or aldosterone antagonists. This channel has pecu-
liar switching kinetics that are characterized by one open
time constant and three closed time constants. Both the
steroids and the aldosterone antagonists reduced the
open time constant, but did not affect the closed time
constants or induce an additional exponential in the
closed-state dwell-time histograms, as expected from the
classical model of blocker action.
In the classical blocker model [3, 5] the channel is
represented by three states: shut (C) open (O) and
blocked (B)
where α, β‘, k–B and k+B·XB are
transition rates (second–1) and XB the molar concentra-
tion of the blocker. This model predicts firstly, that the
open time constant τo=1/(α+k+B·XB) decreases with in-
creasing blocker concentration; second, that the blocker
induces an extra exponential with time constant
τc2=1/k–B in the closed-channel dwell-time histogram
that would not normally be observed and third, that the
relative incidence of the blocking events described by
this time constant (a2) should increase with blocker con-
centration according to the probability density function
(pdfc):
Fig. 6A–C Effect of spironolactone on dwell-time histogram of
ORIC Cl– channel. Details as in Fig. 5 except that concentrations
were 1 and 10 µmol/l in panels B and C respectively. All data
were obtained on the same patch
pdfc=a1·β’·exp(–β’·t)+a2·k–B·exp(–k–B·t)
where t is the time and
(1)
a1=α/(α+k+B·XB) and a2=k+·XB/(α+k+B·XB)
(2)
Although the first prediction would agree with the pres-
ent observations and, although similar to the third pre-
diction, the relative incidence of staying in one closed
state increased with blocker concentration, the second
prediction was not fulfilled. It is true, of course, that a
fourth exponential would have been difficult to discern
against the background of the other three exponentials
observed already under control conditions; however the
fact that the dwell-time histograms in presence of high
concentrations of the steroids or aldosterone antagonists
peaked at the same values of τc2 (≈1.5–2.5 ms) or of τc3
(≈60 ms), respectively, that are observed in absence of
these blockers speaks strongly against an additional
blocked state.
Fig. 7A,B Effect of cytosol on dwell-time histograms of ORIC
Cl– channel. Details as in Figs. 5 and 6. A Control, B cytosol
We conclude therefore that, in the present experi-
ments, the blockers force the channel to dwell preferably
in the spontaneously occurring closed states 2 or 3 rather
than in a new closed state.

564
As will be justified further below, the most simple phenomenon (with an additional time constant). More-
model to explain our findings is:
over, as observed with aldosterone and the other steroids,
it concentration dependently increased the probability of
finding the channel in closed state C2.
Whether other blockers might have similar effects is
not absolutely clear at present. Hwang et al [10] have
analysed the switching kinetics of ORIC Cl– channels in
presence of arachidonic acid which, like other polyunsat-
urated fatty acids, blocks these channels. In their experi-
ments, however, only one exponential was resolved in
both the closed- and open-time histograms. Their data
show that τo decreased inversely with blocker concentra-
tion, while τc (approximately 2 ms and thus correspond-
ing to τc2 in the present experiments) was not affected.
These findings suggest that arachidonic acid increased
the probability of finding the channel in closed state 2
and this implies that arachidonic acid and the other poly-
unsaturated fatty acids tested (linoleic, oleic and ricin-
oleic acids) exert the same blocker effect as the steroids
and glibenclamide. In addition, Singh et al. have tested
three classes of Cl– channel blockers, anthranilic and in-
danyl alkanoic acids and disulphonic stilbenes, on colon-
ic epithelial Cl– channels reconstituted into planar lipid
bilayers [23]. These channels exhibit rectifying cur-
rent/voltage relations and thus may be ORIC Cl– chan-
nels. However, again only one closed-time constant
(about 3.6 ms) was identified under control conditions.
This may be related to the fact that simple, least-squares
fitting to exponential histograms was used rather than
the fitting procedure of Sigworth and Sine [22] which
should have provided better high- and low-frequency
resolution. Because of this limitation it is not possible at
present to decide whether the observed flicker block
might have corresponded to shifting the channel into the
presently observed blocked state C1.
With this model the open time constant in absence of
blockers is:
τo=1/(α+β+γ)
(3)
where α, β, and γ are rate constants of the spontaneous
transitions from the open state into each closed state and
the distribution of closed times is:
pdfc=A1·α’·exp(–α’·t)+A2·β’·exp(–β’·t)
+A3·γ’·exp(–γ’·t)
(4)
where α‘, β‘, and γ‘ are the rate constants of the reverse
(opening) reactions and
A1=α/(α+β+γ), A2=β/(α+β+γ) and A3=γ/(α+β+γ)
(5)
codetermine the relative weight of the three exponen-
tials.
In presence of a blocker B2 which shifts the channel
into state C2 or in presence of a blocker B3 which shifts it
into state C3 the rate constants β and γ will be β+k2·B2
and γ+k3·B3, respectively, where B2 and B3 are the block-
er concentrations and k2·B2 and k3·B3 the transition rates
(second–1) that add to the channels’ inherent blocking
rate constants. Under these conditions Eq. 3 may be writ-
ten:
τo=1/N with N=α+β+k2·B2+γ+k3·B3
and Eqs. 5 yield:
(6)
A1=α/N, A2=(β+k2·B2)/N and A3=(γ+k3·B3)/N
(7)
In contrast to linear reaction schemes, such as C1↔
O↔C2↔C3 or similar, our tridirectional model describes
the present observations well in as much as it predicts
that τo should decrease with increasing concentrations of
the blocker B2, as was observed, while the probability of
the channel dwelling in state C2 should increase with B2,
as was also observed, but the relative probability (proba-
bility ratio) of finding the channel in the other two states
(C1 and C3) should not be affected. In view of the rela-
tively large scatter in determining τc3 with the corre-
sponding area under the curve the latter prediction could
not be tested rigorously in case of the B2 blockers. How-
ever, the validity of the model is also supported by the
experiments with blockers of type B3, for which corre-
sponding observations hold and in which case the relative
probability of the channel staying in state C1 or C2 could
be tested and was not significantly affected.
Blocking action of cytosolic Cl– channel inhibitor
In contrast to steroids and aldosterone antagonists, the cy-
tosol preparations elicited a completely different blocking
pattern: rather than shifting the channel into closed state
C2 or C3 the cytosol appeared to induce at least one new
transition event with a time constant that was slightly
larger than τc1. In view of the limited frequency resolu-
tion, however, this time constant could not be separated
clearly from τc1. It is true that our fit to the closed time
histograms also indicated an increase in τc2 which, on sta-
tistical grounds, was even significant. In view of the rath-
er uncharacteristic shape of the corresponding region in
the histogram, however, we are reluctant to accept this as
a physically significant effect. We conclude therefore that
the blocking mechanism of the cytosolic inhibitor cannot
be resolved on the basis of the present experiments. How-
ever, the observation that the cytosolic inhibitor blocked
differently than all the steroids and aldosterone antago-
nists tested is of further interest since it allows us to con-
clude that the inhibitor is most likely neither a steroid nor
a polyunsaturated fatty acid.
The model is further supported by our observations
on ORIC Cl– channels of HT29 [20] and M1 [25] cells, in
which glibenclamide induced exactly the same blocker
effects as the steroids: it decreased τo concentration de-
pendently, it did not affect the three closed time con-
stants (τc1–τc3) nor did it induce an additional transition

565
Acknowledgements The authors thank J.P. Koch for help with
the data presentation, I. Harward for secretarial assistance and Dr.
K-H. Bahringhaus, Hoechst-Marion-Roussel, Frankfurt for infor-
mation on partition coefficients.
References
1. Anderson MP, Welsh MJ (1990) Fatty acids inhibit apical
membrane chloride channels in airway epithelia. Proc Natl
Acad Sci USA 87:7334–7338
2. Anderson MP, Sheppard DN, Berger HA, Welsh MJ (1992)
Chloride channels in the apical membrane of normal and cys-
tic fibrosis airway and intestinal epithelia. Am J Physiol
263:L1–L14
3. Armstrong CM (1971) Interaction of tetraethylammonium ion
derivatives with the potassium channels of giant axons. J Gen
Physiol 58:413–437
4. Chien YW, Hofmann LM, Lambert HJ, Tao LC (1976) Bind-
ing of spirolactones to human plasma proteins. J Pharm Sci
65:1337–1340
5. Colquhoun D, Hawkes AG (1995) The principles of the sto-
chastic interpretation of ion channel mechanisms. In: Sakman
B, Neher E (eds) Single-channel recording, 2nd edn. Plenum
Press, New York, pp 397–482
Fig. 8 Inhibitory potency of different steroids on ORIC Cl– chan-
nels (reduction of NPo in percent) as function of the logarithm of
the hexadecane:water partition coefficient [21]. Steroid concentra-
tions were 10 µmol/l throughout, and the data represent (from left
to right) hydrocortisone, corticosterone, cortisone, β-oestradiol,
testosterone and progesterone
6. Frizzell, RA, Halm DR (1990) Chloride channels in epithelial
cells. Curr Top Membr 37: 247–282
Mechanism of steroid action
7. Frizzell RA, Rechkemmer G, Shoemaker RL (1986) Altered
regulation of airway epithelial cell chloride channels in cystic
fibrosis. Science 233:558–560
8. Hansch C, Leo A (1979) Substituent constants for correlation
analysis in chemistry and biology. Wiley, New York
9. Hayslett JP, Gögelein H, Kunzelmann K, Greger R (1987)
Characteristics of apical chloride channels in human colon
cells (HT29). Pflügers Arch 410:487–494
10. Hwang TC, Guggino SE, Guggino WB (1990) Direct modula-
tion of secretory chloride channels by arachidonic and other
cis-unsaturated fatty acids. Proc Natl Acad Sci USA 87:5706–
5709
With respect to the mechanism of action of steroids and
aldosterone antagonists, two observations are particular-
ly relevant. First, the inhibitory concentrations are high
(10–100 µmol/l) and second, the inhibitory potency ap-
pears to increase with lipophilicity. The latter follows
from a comparison of the percentage reduction of NPo
with the number of OH– groups per molecule, or with the
octanol:water partition coefficient (Poct/w, Table 1). This
is also supported by our observation with glibenclamide,
for which logPoct/w is 3.08 [18]; it reduced NPo by 11. Krick W, Disser J, Hazama A, Burckhardt G, Frömter E
(1991) Evidence for a cytosolic inhibitor of epithelial chloride
37±7% (n=4) at 10 µmol/l [20]. The characterization of
lipophilicity is not straight forward, however, because
logPoct/w values from the literature vary, and because no
channels. Pflügers Arch 418:491–499
12. Krick W, Disser J, Rabe A, Frömter E, Hansen CP, Roch B,
Kunzelmann K, Greger R, Fehlhaber HW, Burckhardt G
(1995) Characterization of cytosolic Cl– channel-inhibitors by
size exclusion chromatography. Cell Physiol Biochem 5:259–
268
systematic study of all compounds we used is available.
The best correlation is observed if we plot the inhibitory
potency against the logarithm of the hexadecane:water
partition coefficient (Fig. 8). This plot excludes aldoster-
one and glibenclamide, for which no data are available.
At the high concentrations required, and in view of the
potency increasing with lipophilicity, one might speculate
that the steroids flood the lipid phase of the cell mem-
brane and inhibit the channel by altering membrane fluid-
ity [17]. This is unlikely, however, because steroids and
aldosterone antagonists exhibited different inhibition pat-
terns. In addition it is unlikely that a change in membrane
fluidity following incorporation of lipophilic substances
into the membrane would occur instantaneously, and
would also reach a new distribution equilibrium as rapid-
ly as channel inhibition occurred (see Fig. 3). Alternative-
ly, we speculate that the channel protein has at least two
different lipophilic binding sites: one which binds ste-
roids, glibenclamide or polyunsaturated fatty acids and
probably other inhibitors, and another one which binds
the aldosterone antagonists and, probably, other inhibi-
tors. Binding then increases the rate of transition into the
respective closed state of the channel.
13. Kunzelmann K, Pavenstädt H, Greger R (1989) Properties and
regulation of chloride channels in cystic fibrosis and normal
airway cells. Pflügers Arch 415:172–182
14. Kunzelmann K, Tilmann M, Hansen CP, Greger R (1991) In-
hibition of epithelial chloride channels by cytosol. Pflügers
Arch 418:479–490
15. Lambert JJ, Peters JA, Cottrell GA (1987) Actions of synthetic
and endogenous steroids on the GABAA receptor. Trends
Pharmacol Sci 8:224–227
16. Ng AS, Kluza RB, Newton DW (1980) Distribution coeffi-
cients and in vitro human serum protein binding of spironolac-
tone and its 7α-carboxymethyl analog. J Pharm Sci 69:30
17. Ordway RW, Singer JJ, Walsh JV Jr (1991) Direct regulation
of ion channels by fatty acids. Trends Neurosci 14:96–100
18. Panten U, Burgfeld J, Goerke F, Rennicke M, Schwanstecher
M, Wallasch A, Zünkler BJ, Lenzen S (1989) Control of insu-
lin secretion by sulfonylureas, meglitinide and diazoxide in re-
lation to their binding to the sulfonylurea receptor in pancreat-
ic islets. Biochem Pharmacol 38:1217–1229
19. Pardridge WM, Mietus LJ (1979) Transport of steroid hor-
mones through the rat blood-brain barrier. Primary role of al-
bumin-bound hormone. J Clin Invest 64:145–154
20. Rabe A, Disser J, Frömter E (1995) Cl– channel inhibition by
glibenclamide is not specific for the CFTR-type Cl– channel.
Pflügers Arch 429:659–662

566
21. Scheuplein RJ, Blank IH, Brauner GJ, MacFarlane DJ (1960) 24. Tilmann M, Kunzelmann K, Frobe U, Cabantchik I, Lang HJ,
Percutaneous absorption of steroids. J Invest Dermatol 52:63–
Englert HC, Greger R (1991) Different types of blockers of
the intermediate conductance outwardly rectifying chloride
channel in epithelia. Pflügers Arch 418:556–563
70
22. Sigworth FJ, Sine SM (1987) Data transformations for im-
proved display and fitting of single-channel dwell time histo- 25. Volk T, Rabe A, Korbmacher C (1995) Glibenclamide inhibits
grams. Biophys J 52:1047–1054
an outwardly rectifying chloride channel in M-1 mouse corti-
cal collecting duct cells. Cell Physiol Biochem 5:222–231
23. Singh AK, Afink GB, Venglarik CJ, Wand R, Bridges RJ
(1991) Colonic Cl channel blockade by three classes of com- 26. Welsh MJ, Liedtke CM (1986) Chloride and potassium chan-
pounds. Am J Physiol 261:C51–C63
nels in cystic fibrosis airway epithelia. Nature 322:467–470