Department of Physiology, University of Wisconsin Medical School,
Madison, Wisconsin (M.B.J.), and
Royal Danish School of Pharmacy,
Copenhagen, Denmark (S.L.H., B.F.)
Dehydroepiandrosterone sulfate (DHEAS) is a neuroactive steroid
with antagonist action at 
-aminobutyric acid type A
(GABAA) receptors. Patch-clamp techniques were used to
investigate DHEAS actions at GABAA receptors of the rat
pituitary gland at two distinct loci: posterior pituitary nerve
terminals and intermediate pituitary endocrine cells. The GABA
responses in these two regions were quite different, with posterior
pituitary responses having smaller amplitudes and desensitizing more
rapidly and more completely. DHEAS blockade of GABAA
receptors in the two regions also was different. In posterior
pituitary, a site with an apparent dissociation constant of 15 µM
accounted for most of the blockade, but a small fraction of blockade
may be related to a site with a dissociation constant in the nanomolar
range. In the intermediate lobe, DHEAS sensitivities in the nanomolar
and micromolar ranges were clearly evident, in proportions that varied
widely from cell to cell. Regardless of whether the GABA response of a
cell was highly sensitive or weakly sensitive to DHEAS, GABA alone
evoked currents that were indistinguishable in terms of amplitude,
desensitization kinetics, and GABA sensitivity. Thus, the structural
elements responsible for DHEAS blockade have a highly selective impact on receptor function. GABAA receptors with nanomolar
sensitivity to DHEAS have not been described previously. This suggests
that DHEAS may have an important role in the modulation of neuropeptide secretion, and the diverse properties of GABAA receptors in
the rat pituitary provide mechanisms for selective regulation of the different peptidergic systems of this gland.
 |
Introduction |
In
endocrine cells of the pituitary intermediate lobe (IL),
-aminobutyric acid type A (GABAA)
receptor-specific agonists modulate the release of
-melanocyte-stimulating hormone (Tomiko et al., 1983
; Taraskevich
and Douglas, 1985). In the nerve terminals of the posterior pituitary
(PP), GABAA receptor activation alters the
release of oxytocin and vasopressin (Dyball and Shaw, 1978; Fjalland et
al., 1987; Saridaki et al., 1989). In both IL (Demeneix et al., 1986;
Taleb et al., 1987; Schneggenburger and Konnerth, 1992) and PP (Zhang
and Jackson, 1993), the receptors have many of the properties of
classic neuronal GABAA receptors. The channels are selectively permeable to Cl
, muscimol acts
as an agonist, and the responses are blocked by bicuculline and
picrotoxin. IL GABAA receptors possess a pure type 2 benzodiazepine-binding site, and ribonuclease protection assays
have shown that the IL contains mRNA encoding for the 2, 3, 
1,
3, 2s, and 1 GABAA receptor subunits
(Berman et al., 1994). Sensitivity to benzodiazepines and insensitivity
to zinc indicate that PP GABAA receptors contain
subunits (Zhang and Jackson, 1994b). These different molecular and
pharmacological properties determine how GABA and
GABAA receptor-specific drugs regulate different
peptide systems in the pituitary gland.
Neuroactive steroids modulate the responsiveness of
GABAA receptors in many preparations, and this
represents an important nongenomic action of steroids (Harrison and
Simmonds, 1984; Paul and Purdy, 1992; Gee et al., 1995).
Allopregnanolone (5-pregnan-3-hydroxy-20-one) and alphaxolone
have been shown to act as positive GABAA receptor modulators in the PP (Zhang and Jackson, 1994a). In the IL,
allopregnanolone also enhances GABAA
receptor-mediated responses, and the neuroactive steroid pregnenolone
sulfate (5-pregnen-3-ol-20-one sulfate) antagonizes them (Poisbeau
et al., 1997). These findings raise the possibility that neuroactive
steroid modulation of neuropeptide release plays a role in some of the
endocrinological transitions mediated by these peptidergic systems.
Furthermore, neuroactive steroid potencies vary between brain regions
and species (Gee et al., 1995; Nguyen et al., 1995). This adds another
dimension to neuroactive steroid signaling by allowing for differential control over peptidergic versus synaptic processes and, possibly, differential control between various peptidergic systems.
Dehydroepiandrosterone sulfate (DHEAS) is a neuroactive steroid that
has been the subject of wide-ranging discussions regarding potential
roles in cognitive function, aging, stress, and development (Baulieu
and Robel, 1996; Bastianetto and Quirion, 1997). In contrast to many
neuroactive steroids that enhance the responses of
GABAA receptors, DHEAS blocks
GABAA receptors in a number of preparations (Majewska et al., 1990; Spivak, 1994; Souza and Ticku, 1997). In the
present study, we investigated the actions of DHEAS at GABAA receptors in slices prepared from the rat
pituitary gland (Jackson et al., 1991; Schneggenburger and Konnerth,
1992), focusing on both the peptidergic nerve terminals of the PP and
the endocrine cells of the IL. DHEAS inhibits GABA responses in both PP
and IL but with a concentration dependence indicative of multiple binding sites. Variations in the DHEAS sensitivity of endocrine cells
within the IL suggest the presence of multiple molecular forms of
GABAA receptors. This would allow DHEAS to
influence neuropeptide release over a broad range of concentrations and possibly modulate different peptide systems selectively.
| |
Materials and Methods |
Pituitary Slices.
Slices of rat pituitary gland were
prepared as described previously (Jackson et al., 1991). Male
Sprague-Dawley rats 4 to 6 weeks old were sacrificed by decapitation
after CO2-induced narcosis. The pituitary gland
was quickly removed and placed into ice-cold artificial cerebrospinal
fluid (aCSF) consisting of 125 mM NaCl, 4 mM KCl, 26 mM
NaHCO3, 1.25 mM
NaH2PO4, 2 mM
CaCl2, 1 mM MgCl2, and 10 mM glucose, pH 7.3, bubbled with a mixture of 95%
O2/5% CO2 (carbogen). The
pituitary gland was glued to a cutting block, with the posterior lobe
facing upward, and immersed in chilled aCSF. Slices 70 to 80 µm thick
were then cut with a Vibratome. Slices were either kept in
carbogen-bubbled aCSF or transferred to a recording chamber. Recordings
were made with continuous perfusion of carbogen-bubbled aCSF, while
being viewed with a DIC microscope at 600×. Nerve terminals in the PP
ranging in size from 5 to 15 µm in diameter were selected for
patch-clamp recording. IL endocrine cells at the perimeter of the slice
were readily identified by appearance and location (Schneggenburger and
Konnerth, 1992).
Drug Application.
Drugs were dissolved in a solution
consisting of 140 mM NaCl, 3.5 mM KCl, 10 mM glucose, 1.25 mM
Na2HPO4, 2 mM
MgCl2, 2 mM CaCl2, and 20 mM HEPES, pH 7.3, and applied focally by one of two methods. In one
method, a drug-containing patch pipette was positioned ~5 µm from
the cell or terminal under recording, and the drug was ejected with
pressure (12 p.s.i.) from a Picospritzer (General Valve Corp.,
Fairfield, NJ). In the second method, drugs were applied by
gravity-feed through a ~100-µm-diameter tube positioned close to
the cell or terminal under recording, with different solutions selected
from one of seven channels by electrically operated valves. Between
drug applications, the system fed a drug-free solution to maintain
constant flow and reduce hydrodynamic switching artifacts. This also
hastened recovery after drug application. Because the Picospritzer
delivers drugs more rapidly than the multibarrel gravity-feed system,
we compared responses to GABA applied by the two methods. Each method
resulted in responses with the same amplitude in either PP nerve
terminals or IL endocrine cells (Fig. 1).
The multibarrel application system was chosen for most experiments
because this method makes it easy to test several drugs concentrations
in the same recording.
View larger version (20K):
[in this window]
[in a new window]
|
Fig. 1.
Responses to GABA in voltage-clamped PP nerve
terminals and IL endocrine cells, with pressure and gravity-feed
application. GABA (50 µM) was applied for the times indicated (bar).
In all experiments, the duration of GABA application was sufficient to
desensitize the receptors. The two drug-application methods gave
similar results when applied to the same cell, as judged by the
Mann-Whitney rank sum test (in PP: Picospritzer, n = 4; multibarrel, n = 10; in IL: Picospritzer,
n = 3; multibarrel, n = 8).
Response properties are summarized in Table 1.
|
|
All drugs and chemicals were purchased from Sigma Chemical Co. (St.
Louis, MO), except DHEAS, which was purchased from Research Biochemicals International (Natick, MA). In general, responses to 50 µM GABA were tested at regular intervals to check the stability of
control responses. Previous work from this laboratory has shown that
GABA responses in PP showed no run-down (Zhang and Jackson, 1994b), and
responses in IL examined here showed a similar stability. Application
times of 1 to 2 s were used for PP terminals, and a time of 8 s was used for for IL cells. These application times were based on
differences in desensitization kinetics in the two preparations (see
Results). Control GABA responses recovered completely by 1 min after DHEAS application ended, so 1 min was allowed between tests
with different drugs and concentrations. In experiments in which longer
times were allowed, no difference was observed. In studies of
concentration dependence, different concentrations were selected in
random sequences.
Except where noted, various concentrations of antagonists were applied
simultaneously with 50 µM GABA as premixed solutions. When 100 µM
DHEAS alone was applied 2 to 3 min before the application of a mixture
of 50 µM GABA and 100 µM DHEAS, the degree of blockade (~75%)
was indistinguishable from the degree of blockade without pretreatment
in both PP (n = 4) and IL (n = 4).
Thus, the action of DHEAS was sufficiently rapid to obtain full effects
with simultaneous presentations of DHEAS and GABA.
Electrophysiological Recording.
Patch-clamp recordings from
nerve terminals and endocrine cells in pituitary slices were made at
room temperature according to standard methods (Hamill et al., 1981).
All recordings were made with a holding potential of 
70 mV. Patch
electrodes with resistances between 2 and 5 M
were made from
borosilicate glass (1.1 mm I.D., 1.7 mm O.D.; Garner Glass, Claremont,
CA) and filled with a solution consisting of 130 mM KCl, 10 mM HEPES,
10 mM EGTA, 2 mM MgCl2, and 2 mM MgATP, pH 7.2. Tight-seal intracellular recordings were achieved with series
resistances ranging from 3 to 15 M. Whole-terminal and whole-cell
current was recorded under voltage clamp with an Axopatch 200 patch-clamp amplifier (Axon Instruments, Foster City, CA).
Data Acquisition and Analysis.
Data acquisition and analysis
were carried out with pCLAMP6 software (Axon Instruments) on a personal
computer. Curve fitting was performed with the computer program ORIGIN
(Microcal Software, Northampton, MA), and statistical analysis was
carried out with Sigma Stat 2.0 (Jandel Statistical Software, Erkrath, Germany).
Concentration-inhibition plots (for the blockade of GABA responses by
DHEAS) were fitted to models with one or two binding sites. The
one-binding-site model had the form
|
(1)
|
where I is the response as percent of control,
K is the apparent dissociation constant, and C is
blocker concentration. These fits were compared with fits to a
two-binding-site model of the form
|
(2)
|
where K1 and
K2 are the apparent dissociation
constants of two different sites, and
X1 is the percentage of receptor with an apparent dissociation constant of
K1. The statistical significance of
fits to these equations was evaluated by the 
2
goodness-of-fit test (Bevington and Robinson, 1992), using a cutoff
probability of .05.
Concentration-response plots (for activation of receptors by GABA) were
fitted to the Hill equation
|
(3)
|
where K is the apparent dissociation constant and
n is the Hill coefficient.
Open probability at the peak of a response was measured by fluctuation
analysis. According to the binomial distribution, the mean current is
Nip, where N is the number of channels,
i is the single channel current, and p is the
open probability. The variance is
Ni2p(1 p), and the
variance to mean ratio is i(1 p). Because i is known in both PP
(Zhang and Jackson, 1993) and IL (Taleb et al., 1987), p can
be calculated. We determined the variance from a flat segment of data
at the peak of the response typically lasting 20 to 100 ms. Baseline
variance was determined from current recorded before drug application
in the same experiment and subtracted from the variance at the peak.
Unless otherwise stated, arithmetic means from different groupings were
compared using the Student's t test, with a cutoff probability of .05.
| |
Results |
GABA Responses in PP and IL.
GABA responses in PP terminals
are mediated by GABAA receptors (Saridaki et al.,
1989; Zhang and Jackson, 1993). The pharmacology of IL
GABAA receptors has been described previously
(Demeneix et al., 1986; Schneggenburger and Konnerth, 1992), and we
verified this by showing that the current evoked by 50 µM GABA was
blocked completely by 100 µM bicuculline methbromide
(n = 8) and 89 ± 8% blocked (n = 8) by 100 µM picrotoxinin.
With both of the application methods used in this study, responses to
GABA (50 µM) in the PP were strikingly different from those in IL
(Fig. 1). The response amplitudes were ~10- to 15-fold greater in IL
endocrine cells than in PP nerve terminals; responses in both
structures decayed in the presence of sustained GABA application, but
in the IL, this decay was ~18-fold slower. The basic parameters of
these responses are summarized in Table
1. (The high-sensitivity and
low-sensitivity groupings of IL cells in this table are based on
variations in DHEAS sensitivity explained below.) In addition, three
cells in the anterior pituitary were tested with 50 µM GABA. Responses were very small (3-6 pA), so the anterior pituitary was not
studied further.
View this table:
[in this window]
[in a new window]
|
TABLE 1
Response characteristics and kinetic parameters
Responses to 50 µM GABA were compared between PP and IL. Responses in
IL cells were divided into two groups based on amount of blockade by
0.01 µM DHEAS. Cells in which DHEAS reduced GABA responses to <30%
of control were classified as high sensitivity; cells in which GABA
responses remained within 60% of control were classified as low
sensitivity. D/P was ratio of the end point amplitude of current after
desensitization was complete (D) to peak response (P).
|
|
As noted above in Materials and Methods, drug application by
pressure and gravity-feed gave similar results within each pituitary region (Fig. 1). Furthermore, response rise times were clearly more
rapid than the decays seen with continuous drug application. This makes
it unlikely that drug presentation distorts response amplitudes or
desensitization kinetics and thus indicates that the differences
between PP and IL responses can be attributed to receptor properties.
The low-amplitude responses in PP are consistent with earlier reports
from this laboratory (Zhang and Jackson, 1993, 1995), and the larger
amplitudes of responses in IL are comparable to those seen in
dissociated IL cells (Poisbeau et al., 1997). PP and IL
GABAA receptor channels have similar single-channel conductances (Taleb et al., 1987; Zhang and Jackson, 1993). Capacitance measurements taken during these experiments gave
mean values of 7.2 pF in PP and 8.3 pF in IL, indicating that the
membrane areas are comparable. Fluctuation analysis (see Materials and Methods) gave open probabilities at the peak
of the response as .82 ± .05 (n = 6) in PP and
.76 ± .07 (n = 10) in IL. Thus, the different
response amplitudes most likely reflect differences in receptor density.
Table 1 summarizes the comparison of the kinetic parameters associated
with GABA responses in the two regions. As noted above, GABA responses
in IL were larger and desensitized more slowly than those in PP. In all
experiments, GABA application was maintained for a sufficient duration
to desensitize receptors, and the half-time for responses to decay was
taken as a simple quantitative index of this process. Curve fits of
exponential functions to these decay processes were used to estimate
the final end point level of desensitization, and the ratio of this
value to the peak amplitude (D/P in Table 1) provides a quantitative
indication of the extent of desensitization relative to peak. This
ratio was significantly lower in PP, indicating that desensitization of
PP GABAA receptors was more complete than
desensitization of IL GABAA receptors.
DHEAS Blockade of GABA Responses in PP.
DHEAS was an effective
antagonist at GABAA receptor-mediated responses
in PP nerve terminals (Fig. 2). The
response to 50 µM GABA was nearly completely blocked by 100 µM
DHEAS, and as noted, the GABA response recovered fully within 1 min
after DHEAS removal. DHEAS alone at this concentration had no effect.
Figure 2 shows that a lower concentration of DHEAS (0.1 µM) blocked
the GABA response by ~50%.
View larger version (23K):
[in this window]
[in a new window]
|
Fig. 2.
Blockade of PP GABA responses by DHEAS. GABA (50 µM) was applied alone, with 0.1 µM DHEAS and with 100 µM DHEAS.
|
|
The concentration dependence of DHEAS action at
GABAA receptors is shown more fully in Fig.
3. This concentration-inhibition plot
shows increasing blockade as DHEAS concentration increased from 30 pM
to 100 µM. The best-fitting two-binding-site model was drawn through
the data. The one-binding-site model (fit not shown) gave a
2 value that was ~7-fold higher than that
obtained from the two-binding-site model. Furthermore, the
two-binding-site model satisfied the 2 test
for goodness-of-fit to these data with P > .8. These
data clearly show a low-affinity site and provide some indication for the existence of a second high-affinity site. If there are two sites,
the low-affinity site with an apparent dissociation constant of 15 ± 4 µM predominates in determining the concentration dependence of
blockade (Table 2). A second site,
implied by the weak blockade at 1000-fold lower concentrations, has an
apparent dissociation constant of 0.21 ± 0.15 nM and accounts for
only 28 ± 3% of the response to GABA.
View larger version (13K):
[in this window]
[in a new window]
|
Fig. 3.
DHEAS concentration-inhibition plot in PP. Control
responses to 50 µM GABA and responses to GABA plus DHEAS were
recorded in the same nerve terminal. All responses were then normalized
to the control GABA response and plotted. Data from 17 nerve terminals
were used, with each DHEAS concentration tested in 3 to 6 nerve
terminals. Mean ± S.E. values were plotted. The curve is the best
fitting two-binding-site model (see Materials and
Methods). The parameters obtained from this fit are presented
in Table 2. The two-binding-site model satisfied the 2
goodness-of-fit test, and the one-binding-site model did not.
|
|
View this table:
[in this window]
[in a new window]
|
TABLE 2
Apparent dissociation constants for DHEAS
Parameters were obtained from fits of the two-binding-site model (see
Materials and Methods) to DHEAS concentration-inhibition
plots. PP fit is shown in Fig. 3. IL fits are shown in Fig. 6.
K1 and K2 are apparent
dissociation constants for two sites, and X1 is
percentage of block contributed by site 1. In each fit, the
two-binding-site model satisfied the 2 goodness-of-fit test,
and the one-binding-site model did not. High- and low-sensitivity IL
cells were classified on basis of degree of blockade by 0.01 µM DHEAS
of GABA responses to <30% or above 60% of control (see text).
|
|
DHEAS Blockade of GABA Responses in IL.
DHEAS also was an
effective antagonist at GABAA receptors in IL
endocrine cells. Blockade was reversible, and responses to GABA
recovered within 1 min. In contrast to the PP, in the IL, the potency
of blockade varied enormously for a given DHEAS concentration. Figure
4 shows the effect of 0.01 µM DHEAS on
GABA responses in two different IL cells. In one cell, the response was
reduced to 15% of control (Fig. 4A), and in another cell, the response was reduced to 95% of control (Fig. 4B). To illustrate this
variability more completely, data from 24 cells were plotted in scatter
form without averaging, with DHEAS concentrations ranging from 30 pM to
100 µM (Fig. 5). This plot shows that
between 1 nM and 10 µM DHEAS, the variation in potency was especially
large. Repeated tests of a given DHEAS concentration in a single IL
cell gave reproducible amounts of blockade. Furthermore, there was much less variability in blockade of the GABA responses in PP nerve terminals. Thus, the scatter in Fig. 5 is greater than can be accounted
for by experimental error.
View larger version (15K):
[in this window]
[in a new window]
|
Fig. 4.
Variable blockade of IL GABA responses by DHEAS.
Responses to 50 µM GABA are shown together with responses to GABA
plus 0.01 µM DHEAS. A, in a cell with high DHEAS sensitivity, the
response was blocked by ~90%. B, in a cell with low DHEAS
sensitivity, the response was blocked by less than 10%.
|
|
View larger version (26K):
[in this window]
[in a new window]
|
Fig. 5.
Concentration-inhibition scatterplot for DHEAS
blockade of GABA responses in IL. Responses were normalized to the
control GABA response from the same cell. Points from each measurement
were plotted individually to illustrate the variability. Blockade was
most variable between 0.001 and 10 µM.
|
|
One possible explanation for the large variability in DHEAS sensitivity
is that cells are heterogeneous and contain variable proportions of two
GABAA receptors with high DHEAS sensitivity and
low DHEAS sensitivity. To explore this possibility, we focused on the
portion of the concentration-inhibition curve with the greatest
variability. Figure 4 indicates that the blockade of IL
GABAA receptors by 0.01 µM DHEAS varies widely.
Furthermore, Fig. 5 shows that 0.01 µM is in a highly variable part
of the concentration range. Therefore, the degree of blockade of GABA responses by this concentration was used as the basis for dividing IL
cells into the following three groups: 1) cells with GABA responses blocked to less than 60% of control were classified as having low
DHEAS sensitivity. Cells with GABA responses blocked to between 30%
and 60% of control were classified as having intermediate DHEAS
sensitivity. Cells with GABA responses blocked to less than 30% of
control were classified as having high DHEAS sensitivity. According to
these criteria, 57% of the cells were of low, 10% were of
intermediate, and 33% were of high DHEAS sensitivity.
The data from high- and low-sensitivity IL cells were used to construct
two separate concentration-inhibition plots (Fig. 6). In both cases, the one-binding-site
model fitted the data poorly. The 2 value was
reduced 6- or 7-fold by fitting to the two-binding-site model (solid
curves in each graph). The 2 goodness-of-fit
test for the two-binding-site model gave values of P > .8 and P > .20 for the high- and low-sensitivity data, respectively. For the high-affinity site, these fits yielded apparent dissociation constants of 0.23 ± 0.05 and 1.8 ± 1.5 nM in
the high- and low-sensitivity cells, respectively. The low-affinity binding sites had apparent dissociation constants of 0.7 ± 0.6 and 5 ± 3 µM for the high- and low-sensitivity cells,
respectively. In the high-sensitivity cells, the high-affinity binding
site accounted for 89 ± 3% of the inhibition, and in the
low-sensitivity cells, the high-affinity binding site accounted for
39 ± 9% of the inhibition. It is significant that in each plot,
two apparent dissociation constants were obtained: one in the nanomolar
range and the other in the micromolar range. The relative proportions of nanomolar and micromolar components in these two plots are consistent with the hypothesis that IL cells contain two distinct GABAA receptor variants, with DHEAS sensitivities
differing by ~3 orders of magnitude. The apparent dissociation
constants and fractions of high-affinity site are listed in Table 2,
together with the results for PP.
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 6.
Concentration-inhibition plots in IL cells with high
DHEAS sensitivity (A) and low DHEAS sensitivity (B). Based on blockade
by 0.01 µM DHEAS, IL cells were separated into the two groups as
described in the text. Fits to the two-binding-site model (see
Materials and Methods) satisfied the 2
goodness-of-fit test in both plots, but fits to the one-binding-site
model did not. See Table 2 for parameter values.
|
|
Comparisons of High- and Low-Sensitivity Receptors.
Because
the results above suggested that there are different types of
GABAA receptors in different IL cells, we
compared various features of GABA responses after classifying cells on
the basis of high sensitivity and low sensitivity to DHEAS. There were
no differences evident between the two groups of IL cells on
qualitative examination, and quantitative comparisons of response
amplitudes and desensitization behavior also show no differences (Table
1). The kinetics of GABAA receptor-mediated
responses in the PP have been analyzed previously and shown to
desensitize with simple exponential kinetics (Zhang and Jackson,
1994b). In IL cells with either high or low DHEAS sensitivity, the time
courses of desensitization of responses to GABA alone were also well
described by a single exponential in most cases (data not shown). The
results in Table 1 show that IL cells with high and low sensitivity to
DHEAS are indistinguishable in terms of speed and extent of
desensitization. (The half-time in the table is taken as a measure of
speed, and the ratio D/P of final to peak current is taken as a measure
of extent.)
Responses to different concentrations of GABA were measured and used to
construct concentration-response plots for IL cells with either high or
low DHEAS sensitivity (Fig. 7). Fits to
the Hill equation (see Materials and Methods) indicated that
the apparent dissociation constants for GABA were 55 ± 11 and
66 ± 11 µM and the Hill coefficients were 1.3 ± 0.3 and
1.5 ± 0.3 in the high-sensitivity and low-sensitivity groups,
respectively. These values are indistinguishable between the two
groups, indicating these two types of receptors differ very
specifically with regard to DHEAS sensitivity. Other domains of the
receptor with roles in determining GABA sensitivity and response
kinetics appear to be unaltered.
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 7.
GABA concentration-response plots from IL cells with
high (A) or low (B) DHEAS sensitivity. Responses were normalized to the
response to 1 mM GABA. Data for all five concentrations of GABA were
averaged from five cells for the high-sensitivity group and eight cells
for the low-sensitivity group. Mean ± S.E. values are shown
together with the best-fitting Hill equation. The apparent dissociation
constants were 55 ± 11 and 66 ± 11 µM for the
high-sensitivity and low-sensitivity groups, respectively. The Hill
coefficients were 1.3 ± 0.3 and 1.5 ± 0.3, respectively.
The extrapolated maximum responses were 1.04 times greater than the 1 mM responses in both groups.
|
|
We also examined the kinetics of GABA responses blocked ~50% by
DHEAS. The responses were generally more complex than the responses to
GABA alone and were difficult to interpret quantitatively. The only
noticeable feature of these responses was that in both high-sensitivity
and low-sensitivity cells, DHEAS made the GABA responses desensitize
more slowly (data not shown). Another neurosteroid antagonist,
pregnenolone sulfate, was found to accelerate desensitization in IL
cells (Poisbeau et al., 1997). In PP nerve terminals, there was no
difference in the kinetic parameters between GABA alone and GABA plus 1 µM DHEAS (data not shown). The decays were well fitted by one
exponential in both the presence and absence of DHEAS. In PP, the
enhancing neuroactive steroids allopregnanolone and alphaxolone also
left unchanged the time constant for desensitization of GABA
responses (Zhang and Jackson, 1994a). Thus, in this respect, the
antagonist and enhancing neuroactive steroids are similar.
| |
Discussion |
These experiments demonstrated that the neuroactive steroid DHEAS
has antagonist activity at GABAA receptors in two
distinct parts of the pituitary gland. In IL endocrine cells, multiple DHEAS binding sites were evident with sensitivities in the micromolar and nanomolar ranges. GABAA receptor properties
differed between the PP and IL, as well as within the IL. The
sensitivity of these receptors to neuroactive steroids provides a
potentially important signaling pathway in the regulation of
neuropeptide release.
Differences between PP and IL GABAA Receptors.
GABAA receptors in PP nerve terminals and IL
endocrine cells differed in their basic response characteristics. The
current evoked by 50 µM GABA was much larger in IL than in PP. The
smaller responses in PP probably reflect a low density of channels. The comparison of response properties revealed that desensitization is
faster and more complete in PP than in IL (Table 1). The slow desensitization of IL GABAA receptors was similar
to that observed by Poisbeau et al. (1997) in dissociated rat IL cells.
The contrast between slow and fast desensitization of IL and PP
GABAA receptors is reminiscent of the difference
between Drosophila GABA receptors assembled as homomeric
complexes from the Rdl gene product or as heteromeric complexes with
subunits (Zhang et al., 1995). Desensitization is virtually absent
in vertebrate receptors where the only subunit is 6, and removal
of the subunit results in a large nondesensitizing component of
current (Tia et al., 1996). The inclusion of a 
subunit also slows
desensitization (Saxena and Macdonald, 1994). Thus, a molecular basis
for different desensitization rates in GABA receptors is well
established and could account for the differences observed here.
The physiology of GABA-mediated inhibition could be very different in
the PP and IL, and the different receptor properties in the two regions
may have an adaptive significance. Release of vasopressin and oxytocin
from the PP is under the control of rhythmic activity in the
hypothalamus (Poulain and Wakerley, 1982). GABA in the PP cannot
influence this distant location, so to modulate release, GABA must
block action potentials or alter their shape. Under some conditions,
GABA may be able to block action potential propagation through the PP
nerve terminal arbor (Zhang and Jackson, 1993; Jackson and Zhang,
1995), but the timing would be important because an action potential is
a very brief event. The rapid kinetics of the PP
GABAA receptor could produce a brief episode of
inhibition, which would be effective only if well timed. On the other
hand, impulse propagation is irrelevant in IL endocrine cells because of their compact morphology. Receptor-mediated control of release from
endocrine cells results from a modulation of endogenous rhythmic electrical activity (Douglas, 1968; Davis and Hadley, 1978), and the
slower kinetics in IL GABAA receptors may have an
adaptive value for the regulation of oscillation frequency. These two
situations place different demands on the biophysical properties of
inhibitory receptors, and the IL and PP GABAA
receptor characteristics described here may be tailored for specific
physiological roles.
Multiple Affinities for DHEAS in PP and IL.
Concentration-inhibition plots for both PP and IL
GABAA receptors were better fitted by a
two-binding-site model than by a one-binding-site model. In PP,
blockade was dominated by a low-affinity component with an apparent
dissociation constant of 15 µM. However, the PP may also have a
high-affinity site because a small amount of blockade (~28%) could
be seen with DHEAS in the nanomolar range. This may reflect an
additional DHEAS binding site with weak negative control over receptor
function. A high-affinity site could also be an indication of two
receptor populations, as might be expected from the diversity of
neuroactive steroid efficacies in a number of tissues (Gee et al.,
1995; Nguyen et al., 1995).
DHEAS blockade also revealed two binding sites in IL, differing by
~1000-fold in sensitivity. High variability in DHEAS sensitivity prompted us to divide IL cells into two groups. The segregation of this
property between different IL cell populations suggests some form of
structural distinction, such as subunit composition, receptor
phosphorylation, associated proteins, or lipid environment. The most
likely explanation is subunit composition, which has been shown to
underlie much of the pharmacological diversity of GABAA receptors (Whiting et al., 1995). mRNAs
encoding 2, 3, 1, 3, 2s, and 1
GABAA receptor subunits have been detected in the
IL (Berman et al., 1994), and these subunits could be combined in many
ways. The presence of subunits is significant because receptors
formed from a subunit alone can be modulated by neuroactive steroids (Puia et al., 1990). However, other subunits are likely to be
involved, and a more detailed understanding of the structural determinants of neuroactive steroid sensitivity will require further investigation (Whiting et al., 1995).
The two IL GABAA receptors generated responses
with similar kinetic properties and sensitivities to GABA. The
amplitudes and desensitization rates were indistinguishable (Table 1),
as were the apparent dissociation constants and Hill coefficients for activation by GABA (Fig. 7). This suggests that the structural factors
responsible for differences in DHEAS sensitivity are highly localized
and specialized. It may just be a single subunit that is different
between the two receptor forms. Thus, the structural elements that
determine GABA sensitivity and desensitization kinetics could be the
same in both high-sensitivity and low-sensitivity IL cells (but
different in PP nerve terminals). The structural elements that
determine DHEAS binding would then appear to have little influence over
these other parameters. Once the molecular structures of these two
closely related IL GABAA receptors are determined, sequence
comparisons are likely to provide valuable clues about DHEAS binding domains.
DHEAS has previously been shown to act as a noncompetitive antagonist
at GABAA receptors in binding assays and
electrophysiological studies. In binding assays in neuronal
synaptosomal membranes, a two-binding-site model fitted the data,
yielding apparent dissociation constants of 2.9 and 554 µM (Majewska
et al., 1990). Electrophysiological studies suggested blockade at a
single binding site, with an IC50 value of 13 µM (Majewska et al., 1990) or 4.5 µM (Spivak, 1994). In another
study, DHEAS was found to inhibit GABA-induced
36Cl influx in cortical
neurons with an IC50 value of 10 µM (Souza and
Ticku, 1997). The low-affinity binding site seen in both IL and PP is
consistent with these earlier studies, but to our knowledge a nanomolar
sensitivity of GABAA receptors to DHEAS, as
described here, has not been reported previously.
Physiological Roles for DHEAS Blockade of Pituitary
GABAA Receptors.
Because the neurointermediate lobe of
the pituitary gland lies outside the blood-brain barrier, it will be
exposed to circulating DHEAS. In humans, circulating DHEAS levels are
generally in the micromolar range (Sulcová et al., 1997), but in
rodents, they are only ~1 nM (Corpéchot et al., 1981). Thus,
the high-sensitivity site described here in rat could place IL
GABAA receptors under the regulatory control of
circulating DHEAS. The low-sensitivity site would then be free to
respond if DHEAS levels rise. The IL secretes several hormones derived
from the proopiomelanocortin precursor, including
-melanocyte-stimulating hormone and -endorphin. The variable
sensitivity to DHEAS raises the interesting possibility of differential
control of these different peptide hormones. Alternatively, the
variable sensitivity may provide for graded control of each peptide
hormone over a wide range of DHEAS levels.
Both the IL and PP receive GABAergic innervation from the hypothalamus
(Oertel et al., 1982; Vincent et al., 1982), and it has been shown that
GABA can either inhibit or enhance hormone release from the pituitary
gland depending on the stimulation protocol (Tomiko et al., 1983;
Fjalland et al., 1987; Saridaki et al., 1989). With the receptor
sensitivities reported here, physiological levels of DHEAS would be
capable of modulating GABAA receptors in both the
IL and PP. Actions on these and other endocrine systems could be
relevant to some of the physiological functions proposed for this
neuroactive steroid (Baulieu and Robel, 1996; Bastianetto and Quirion,
1997).
We thank Drs. Cynthia Czajkowski and Russ Wilke for comments on
this manuscript.
This work was supported by National Institutes of Health Grant
NS30016 and by the Royal Danish School of Pharmacy.
![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
-Aminobutyric Acid Type A
Receptors by the Neuroactive Steroid Dehydroepiandrosterone Sulfate in
Posterior and Intermediate Pituitary
Top
Summary
Introduction
