Download co-culture of hypothalamic neurons and melanotrope cells

Pergamon
PII:
Neuroscience Vol. 76, No. 1, pp. 203–214, 1997
Copyright ? 1996 IBRO. Published by Elsevier Science Ltd
Printed in Great Britain
0306–4522/97 $17.00+0.00
S0306-4522(96)00279-5
CO-CULTURE OF HYPOTHALAMIC NEURONS AND
MELANOTROPE CELLS: A MODEL TO STUDY
SYNAPTOGENESIS BETWEEN CENTRAL NEURONS AND
ENDOCRINE CELLS
F. RENE
u ,* P. POISBEAU, C. EGLES, R. SCHLICHTER and J. M. FE
u LIX
Laboratoire de Neurophysiologie et de Neurobiologie des Systèmes Endocrines,
Université Louis Pasteur, URA CNRS 1446, 21 rue René Descartes, 67084, Strasbourg, Cedex, France
Abstract––As a first step towards elucidating mechanisms involved in neuroendocrine synaptogenesis, we
developed a model of co-culture based on hypothalamic–intermediate pituitary interactions. Dissociated
hypothalamic neurons from fetal rats at embryonic day 15 were cultured in a defined medium together
with melanotrope cells of the pituitary intermediate lobe from neonatal rats. In these co-cultures,
establishment of synaptic contacts between GABAergic or dopaminergic neurons and an endocrine target
cell, the melanotrope cell, was studied by morphofunctional approaches.
Using double immunostaining with antibodies directed against glutamate decarboxylase or tyrosine
hydroxylase and á-melanocyte-stimulating hormone, we demonstrated morphological contacts between
GABAergic or dopaminergic neurons and melanotrope cells as early as three days in vitro. Furthermore,
using an antibody directed against synapsin I, we showed a modification of synapsin I immunoreactivity
from diffuse to punctate distribution correlated with the establishment of contacts and the observation of
characteristic neuroendocrine synapses by electron microscopy. These results were further confirmed by
electrophysiological studies. Patch-clamp recordings demonstrated that, at six days in vitro, some
melanotrope cells displayed GABAergic synaptic currents, which occurred either spontaneously and/or
could be evoked chemically by 50 mM KCl or 100 µM kainate. The proportion of the melanotrope cells
receiving functional synaptic inputs increased until 10 days in culture, a stage at which virtually all
melanotrope cells in contact with neurons possessed functional synapses.
The results presented here describe the establishment of neuroendocrine synapses in vitro, studied by
combining morphofunctional and electrophysiological approaches. Copyright ? 1996 IBRO. Published
by Elsevier Science Ltd.
Key words: neuroendocrine synapse, hypothalamus, intermediate lobe of the pituitary, dopamine, GABA,
patch-clamp recording.
Due to the complexity of the developing nervous
system and the technical difficulties involved in studying individual neurons or one neuron coupled to its
target cell, mechanisms underlying synaptogenesis
in vivo are poorly understood. The technical difficulties in studying individual neurons in situ are
being overcome with the development of in vitro
approaches. Until now, all in vitro studies concerned
synaptogenesis21 at the neuromuscular junction and
neuroneuronal synapse formation of a limited
number of neurons of the CNS.20,25 Furthermore, in
the neuroneuronal systems, precise identification of
*To whom correspondence should be addressed.
Abbreviations: DIV, days in vitro; E, embryonic
day; EDTA, ethylenediaminetetra-acetate; EGTA,
ethyleneglycolbis(â-aminoethyl ether)-N,N,N*,N*-tetraacetate; FCS, fetal calf serum; GAD, glutamate decarboxylase; GFAP, glial fibrillary acidic protein; HEPES,
N-2-hydroxyethylpiperazine-N*-2-ethanesulfonic acid; IL,
intermediate lobe; á-MSH, á-melanocyte-stimulating
hormone; NSE, neuron-specific enolase; PB, phosphate
buffer; PBS, phosphate-buffered saline; TH, tyrosine hydroxylase.
203
the nature of the target cell has been made difficult
because of the heterogeneity of the cell populations. In this work, we report a system of co-culture
offering the possibility of studying the genesis of a
neuroendocrine synapse. This system allows studies
of the reciprocal inductive interactions between
central neurons of the hypothalamus and endocrine
melanotrope cells from the intermediate lobe (IL) of
the pituitary.
In situ, melanotrope cells constituting the IL of the
pituitary are negatively regulated by a direct innervation originating from the rostral periventricular area
of the hypothalamus.5,11,39 This innervation is mainly
dopaminergic2 and GABAergic.31 Dopamine and
GABA are co-localized within the same axon
terminals,45 synaptically released, and exert a tonic
inhibition on melanotrope cells. In adult melanotrope cells, pharmacological activation of dopamine
D2 receptors9,24,28 and GABAA receptors27,40,42 lead
to a decrease of pro-opiomelanocortin mRNA expression and to a reduction of spontaneous electrical
and secretory activities. Dopamine D2 receptors
204
F. René et al.
activate a K+ conductance and inhibit a Ca2+ current
via a pertussis toxin-sensitive pathway,38,49,50
whereas GABA inhibits electrical and secretory
activities of melanotrope cells by activating GABAA
receptors40,41,42 and decreases the release of
á-melanocyte-stimulating hormone (á-MSH) by
activating GABAB receptors,15 probably by reducing
the intracellular free Ca2+ concentration.35
Maturation of synapses is characterized by the
increased expression of synapsin I. This protein belongs to a family that is highly specific for nerve
terminals.4 Synapsin I is selectively associated
with small synaptic vesicles and is not present on
other membrane structures.30 Moreover, in vivo,
the appearance of synapsin I in neurons during
ontogenesis of the nervous system correlates with the
onset of synaptogenesis and with the development of
synaptic contacts.29 For these reasons, synapsin I
immunoreactivity was used as a marker of synaptic
maturation in our model. The maturation was further confirmed by using electron microscopy and
patch-clamp recording.
EXPERIMENTAL PROCEDURES
Tissue culture
Hypothalamic neurons. Pregnant Wistar rats (Deprés,
France) at day 15 of gestation (day 0 of gestation corresponding to the day of fecondation) were anesthetized with
pentobarbital and the embryos removed. The diencephalic
area corresponding to the hypothalamus was isolated and
incubated with trypsin–EDTA (0.05–0.02%; Gibco) for
7 min at 37)C under constant shaking. The enzymatic
dissociation was stopped by adding 20% fetal calf serum
(FCS) and followed by centrifugation for 3 min at 180 g.
Mechanical dissociation was performed in Dulbecco modified Eagle’s medium (Gibco, France) containing 20% FCS
by aspiration into fire-polished Pasteur pipettes with decreasing tip diameters. After centrifugation for 7 min at
180 g, the pellet was resuspended in serum-free medium
according to Bottenstein and Sato.8 Hypothalamic cells
were seeded at a density of 400 cells/mm2 in 35-mm Petri
dishes for electrophysiological study, or at a density of
800 cells/mm2 on 8-mm diameter coverslips in four-well
multidishes (15 mm diameter) for immunocytochemistry
and electron microscopy studies. In each condition, both
35-mm dishes and glass coverslips were precoated with
poly--lysine (type VIIB, mol. wt 60,000, Sigma; 10 µg/ml in
0.1 M borate buffer, pH 8.4) and preincubated for 1 h in
phosphate-buffered saline (PBS) containing 10% FCS. Cultures were maintained in water saturated atmosphere (95%
air, 5% CO2) at 37)C. Medium was first renewed after five
days and every three days thereafter.
Intermediate lobe cells. Neonatal (one- to seven-day-old)
Wistar rats were killed by decapitation under deep diethyl
ether anesthesia. Pituitaries were taken out, extensively
washed in PBS and neurointermediate lobes were separated
from anterior lobes. Melanotrope cells of neurointermediate
lobes were dissociated as described above. Cells were seeded
with hypothalamic neurons at a density of 80 cells/mm2 for
electrophysiological studies and 160 cells/mm2 for immunocytochemical and electron microscopic analyses.
Immunostaining
After one, three, five, seven or 12 days in vitro (DIV),
cultures were fixed for 10 min in PBS containing 4%
paraformaldehyde (Kodak) at room temperature and
subsequently rinsed with PBS.. For the detection of tyrosine
hydroxylase (TH) immunoreactivity, cultures were permeabilized for 10 min in PBS containing 1% Triton X-100
followed by 30 min incubation in PBS containing 5% horse
serum and 0.1% Triton X-100.
For the detection of glial fibrillary acidic protein (GFAP)
immunoreactivity, cultures were permeabilized for 30 s in
ice-cold methanol and rinsed with PBS.
For the detection of neuron-specific enolase (NSE), vimentin, glutamate decarboxylase (GAD) and á-MSH
immunoreactivities, cultures were not pre-permeabilized.
Cultures were incubated overnight at 4)C with the primary
antibody diluted in PBS containing 0.1% Triton X-100. The
rabbit polyclonal antibodies used were directed against NSE
(1:2000; Immunotech, France), GFAP (1:1000; Dako,
Denmark), synthetic á-MSH (1:3000; the late Dr G.
Schmidt, Strasbourg, France) and GAD (1:500; Chemicon:
AB 108). The mouse monoclonal antibodies used were
directed against vimentin (1:20; Boerhinger) and TH
(1:10,000; Chemicon: MAB 318). Following antibody incubation, cells were rinsed three times with PBS and incubated
for 1 h at room temperature in biotinylated goat anti-rabbit
immunoglobulin G (1:500; Vector) for NSE, GFAP,
á-MSH and GAD detection, or in sheep anti-mouse immunoglobulin G (1:500; Vector) for vimentin and TH. Cultures
were rinsed three times for 10 min with PBS and incubated
for 1 h with an avidin–biotin–peroxidase complex (Elite kit,
Vector). After three washes, peroxidase was detected with
3,3*-diaminobenzidine tetrahydrochloride (Vectastain kit,
Vector). Reaction was stopped by washing with PBS. Cultures were then dehydrated in graded alcohols, cleared in
xylene and mounted.
To perform the double immunostaining for TH or GAD
and á-MSH, detection of TH or GAD was performed as
described above. Cultures were then incubated for 1 h at
37)C with rabbit antiserum against á-MSH (1:3000) and
rinsed with PBS. They were incubated for 1 h with goat
anti-rabbit alkaline phosphatase-coupled Fab fragments
(1:650; Sigma). After rinsing, á-MSH immunoreactivity was
detected with a chromogen solution containing 340 µg/ml
nitro blue tetrazolium (Sigma), 175 µg/ml bromo-chloroindol-phosphate (Sigma), 0.1 M NaCl, 0.05 M MgCl2 in
0.1 M Tris–HCl (pH 9.5). Cultures were rinsed and
mounted in aqueous mounting medium consisting of 7.5%
gelatin and 50% glycerol in PBS.
For the detection of synapsin I, cultures were incubated
overnight in rabbit polyclonal antibody directed against
bovine synapsin I (1:500; Dr J. Baudier, Strasbourg,
France) diluted in PBS containing 0.1% Triton X-100. In
western blot analysis of rat cortical neurons, the synapsin
antibody used revealed only two bands (mol. wts 80,000 and
86,000), corresponding to the isoforms described by De
Camilli et al.13 (not shown). After three washes in PBS
containing 0.1% Triton X-100, they were incubated for 1 h
at room temperature with fluorescein isothiocyanateconjugated anti-rabbit immunoglobulin G (1:200; Biosys,
France), rinsed with PBS and mounted.
Cultures used for semi-thin sections (detection of
synapsin I and á-MSH on adjacent sections) were fixed
for 10 min in 0.1 M phosphate buffer (PB; pH 7.4) containing 4% paraformaldehyde (Kodak). After dehydration
in graded alcohols, they were conventionally embedded
in Araldite–Epon mixture and semi-thin sections
(1.5 µm) were made. Embedding medium was removed
with sodium methoxide and sections were incubated overnight with the polyclonal antibodies directed against synapsin I or á-MSH (see above). The following steps for
immunoperoxidase detection were performed as decribed
above.
All the antisera used in culture have also been characterized in histological sections; no non-specific labeling
was observed. Moreover, cultures incubated without
Morphofunctional study of neuroendocrine synaptogenesis in vitro
205
primary antibody or with an irrelevant antidoby showed no
staining.
Electron microscopy
Cultures grown for 8 and 12 DIV on glass coverslips were
fixed with 5% glutaraldehyde (Fluka) in PB for 1 h at 4)C
and postfixed for 1 h with 1% osmium tetroxide in PB
containing 1.6% potassium ferrocyanide. After dehydration
in graded alcohols, they were flat-embedded in an Araldite–
Epon mixture. The coverslips were detached by immersion
in liquid nitrogen. Thin sections (50 nm), contrasted with
uranyl acetate and lead citrate, were examined under a Jeol
100 CX electron microscope.
á-Melanocyte-stimulating hormone measurement
á-MSH radioimmunoassay was performed as described
previously by Vuillez et al.46 Data are the mean & S.D.
from five cultures.
Neuron counting
Cells were counted using an inverted phase-contrast
microscope. The number of NSE-, vimentin-, GFAP-,
á-MSH-, GAD- or TH-positive neurons were counted in 10
randomly chosen observation fields corresponding to
0.39 mm2. In the same fields, total number of neurons was
counted on the basis of morphological criteria (ovoid cell
body, neurites). Data are expressed either as the mean of cell
type & S.D. or as the mean percentage of GAD-positive
(or TH-positive) neurons in the total neuronal
population & S.D. Data are the average from three cultures
in triplicate.
Electrophysiological recordings
Electrophysiological experiments were performed at
room temperature (20–22)C) three to 15 days after plating
of the cells. Patch-clamp recordings were made under
voltage-clamp in the whole-cell recording configuration22
using an Axopatch 200 A amplifier (Axon instruments,
U.S.A.). Low-resistance (3–4 MÙ) electrodes were used.
The external medium contained (in mM): NaCl 135; KCl 5;
CaCl2 5; MgCl2 1; HEPES 5; glucose 10; pH 7.3 with
NaOH. Pipettes were filled with an intracellular solution
containing (in mM): KCl or CsCl 125; CaCl2 5; MgCl2 2;
HEPES 10; EGTA 10; pH 7.3 (adjusted with KOH or
CsOH). The estimated intracellular free calcium concentration was 10"7 M and the equilibrium potential for
chloride ions (ECl) was "2 mV.
The competitive GABAA receptor antagonists SR95531
and bicuculline were prepared as 10 mM stock solutions in
distilled water and stored at "20)C. Just before the recording session, these substances were diluted to their final
concentrations in external medium. Local application of
drugs was performed with a ‘‘U-tube’’.18 SR95531 was a
kind gift of Sanofi-Recherche (Montpellier, France) and
bicuculline methiodide was purchased from Sigma (France).
Data storage and analysis
Synaptic currents were stored on videotape in digital form
(20 kHz) after being filtered at 5 kHz by the internal filter of
the Axopatch 200 A amplifier. Analysis was performed
using the Axograph software (Axon Instruments, U.S.A.)
on a Macintosh IIvx computer. For this purpose, stored
data were acquired off-line with the pclamp software
(Fetchex, Axon Instruments, U.S.A.) at a 10-kHz resolution
with no additional filtering. All statistical results are
expressed as mean & S.D.
Fig. 1. Phase-contrast micrographs showing the development of hypothalamic neurons from E15 rat embryos
co-cultured together with melanotrope cells. (A) Immediately after plating, all the cells are spherical and devoid of
any processes, making it impossible to indentify any type of
cell. (B) After 3 DIV, neurons have started to extend
neurites, whereas melanotrope cells remain spherical. Some
neurons (white arrows) have already entered into contacts
with melanotrope cells (black arrowheads). Scale
bar = 50 µm.
RESULTS
Co-culture characterization
Immediately after dissociation, when observed
under a phase-contrast microscope, hypothalamic
cells as well as IL cells appeared spherical and devoid
of any processes (Fig. 1A). Within hours after plating, they became adherent to the bottom of the dish.
After 24 h in vitro, different cell populations were
distinguishable on the basis of their morphological
appearance and their immunoreactivity. Two types
of non-refractive and strongly adhering flat cells
were present. One type displayed GFAP immunoreactivity and corresponded to glial cells, including
IL folliculostellate cells (Fig. 2A). The second type
was vimentin-immunoreactive and corresponded
both to fibroblasts and folliculostellate cells (Fig. 2B).
These cells proliferated (Table 1) and progressively
formed a continuous layer on the bottom of the dish.
The melanotrope cells kept their spherical appearance, without processes, and were strongly phase
bright. Throughout the culture period, the number
of á-MSH-positive increased slightly (Table 1).
These cells displayed a strong immunoreactivity to
á-MSH, as shown after 3 DIV (Fig. 2C), and secreted
206
F. René et al.
Table 1. Evolution of several cell types over time in culture
NSE-positive
GFAP-positive
Vimentin-positive
á-MSH-positive
TH-positive
GAD-positive
1 DIV
3 DIV
5 DIV
7 DIV
12 DIV
14.5 & 1.3*
4.3 & 0.3
44.8 & 4.2
8.8 & 3.7
2.3 & 0.4
32.5 & 1.9
116.8 & 5.1
18.4 & 1.2
61.1 & 4.2
15.6 & 2.2
7.5 & 1.9
97.8 & 7.2
130.3 & 7.9
20.3 & 1.4
75.1 & 1.5
19.2 & 2.17
9.2 & 0.7
110.7 & 6.3
98.9 & 5.7
55.4 & 6.8
33.3 & 1.4
23.2 & 1.6
5.7 & 0.6
81 & 4.2
58.2 & 7.5
62.5 & 2.3
35.7 & 4.8
21.4 & 5.4
3.2 & 0.9
46.7 & 2.8
After 1, 3, 5, 7 and 12 DIV, cultures were fixed and processed for immunocytochemical detection of NSE for neurons,
GFAP and/or vimentin for glial cells and fibroblasts, á-MSH for melanotrope cells, TH for dopaminergic neurons and
GAD for GABAergic neurons.
Immunoreactive cells were counted in 10 randomly chosen observation fields corresponding to 0.39 mm2 each. Data are
expressed as the mean of cell type & S.D. per field and are the average from three cultures in triplicate.
*
In the same fields, the number of neurons identified on the basis of morphological criteria was 111.6 & 13.2.
á-MSH in culture medium at a constant rate
(1.04 & 0.38 nM per day and per well; n = 20).
Usually, melanotrope cells were isolated, but
occasionally they tended to reaggregate and to form
small clusters.
The neuronal cells had a bright and ovoid cell
body, and extended one or two thin short processes.
Concomitantly, they started to establish apparent
contacts with each other (Fig. 1B). The diameters of
their perikarya were smaller (<10 µm) than those of
the melanotrope cells (10–12 µm). A slight NSE
immunoreactivity was detectable in some neurons
24 h after plating (14.7 & 0.8%; n = 9; Table 1), but
the labeling became consistent in all of them after
3 DIV (Fig. 2D). From the third DIV, all neuronal
cells had processes and a dense and complex fiber
network was formed. The number of NSE-positive
neurons was constant until 5 DIV (Table 1), and then
started to slowly decrease to about 50% of the initial
population identified on the basis of morphological criteria after 1 DIV. GAD-positive neurons
represented 80–85% of the total population of
neurons over the time in culture (Table 1). GADpositive immunoreactivity was found over the cell
body and the proximal part of the processes (Fig.
2E). The level of GAD-positive immunoreactivity
was not homogeneous between the GAD-positive
cells, but the light staining observed corresponded
to positive cells compared to the control reactions
(not shown). In contrast to GAD-positive neurons,
TH-positive neurons constituted a small subpopulation of neurons which increased until 5 DIV and
represented 5–7% of the neuronal population (Table
1). They displayed a uniform TH immunoreactivity
distributed over the whole cell (Fig. 2F).
Morphofunctional study of synaptogenesis
Phase-contrast observation of co-cultures shows
apparent neuroneuronal contacts after only 1 DIV
(not shown). Thin neurites of a few neurons made
apparent contacts with melantrope cells after 3 DIV
(Fig. 1B). Both GAD-positive (Fig. 3A) and THpositive (Fig. 3B) neurons established contacts with
melanotrope cells. At 3 DIV, most of the neurons
were weakly labeled with the synapsin I antibody
(Fig. 4A). Immunoreactivity was diffusely distributed
in the cell body and in the proximal part of neurites.
Progressively, the number of neurons contacting
melanotrope cells increased with time in culture. In
addition, most of the TH-positive neurons sent two
or more processes towards several melanotrope cells
and established contact with them, as shown after
12 DIV in Fig. 3C. Numerous GAD-immunoreactive
contacts were also identified (data not shown); however, due to the high density of GABAergic neurons,
a clear-cut identification of the cell bodies of
GABAergic neurons establishing contacts with
melanotrope cells was difficult. At this time, synapsin
I immunoreactivity was highly concentrated in tiny
dots scattered on less intensely labeled neuronal cell
bodies and extensions (Fig. 4B). Synapsin I immunoreactivity had almost disappeared from the neuronal
cell bodies. By comparing synapsin (Fig. 4C) and
á-MSH (Fig. 4D) immunolabeled adjacent semi-thin
sections, synapsin I punctae were found in close
apposition with melanotrope cells.
After eight days in culture, electron microscopy
showed the presence of contacts exhibiting synaptic
specializations between neuronal elements (Fig. 4E).
At this stage, synaptic differentiation, evaluated by
clustering of synaptic vesicles and the presence of
presynaptic density, was not obvious in the contacts
between neuritic extensions and melanotrope cells. In
contrast, presynaptic specializations were commonly
detected after 12 days (Fig. 4F).
Electrophysiological
transmission
characterization
of
synaptic
As shown in the above sections, hypothalamic
neurons formed direct contacts with melanotrope
cells as well as with other hypothalamic neurons.
Moreover, these contacts seemed to have the
morphological characteristics of synapses. In order to
test the functionality of these contacts, we have
recorded from melanotrope cells and neurons which
were in apparent contact with one or several neurites
Morphofunctional study of neuroendocrine synaptogenesis in vitro
207
Fig. 2. Immunocytochemical characterization of co-culture after 3 DIV (A–D) and 5 DIV (E, F). Sister
cultures were stained separately for GFAP (A), vimentin (B), á-MSH (C) and NSE (D), or GAD (E) and
TH (F). Note that GAD labeling is restricted to the cytoplasm and the proximal part of neurites, whereas
TH immunolabeling is distributed in the whole neuron. Scale bars = 25 µm.
of hypothalamic neurons. The postsynaptic cells
were identified on the basis of morphological and
electrophysiological criteria (see below). Their
electrical activities were recorded with the patchclamp technique under whole-cell voltage-clamp, a
configuration which allowed optimal resolution of
synaptic currents of small amplitude. Figure 5A
shows a recording obtained in a melanotrope cell.
The current trace displays spontaneously and randomly occurring inward currents corresponding to
208
F. René et al.
Fig. 3. Double labeling of GAD/á-MSH (A) and TH/á-MSH (B, C). á-MSH is revealed by immunophosphatase, GAD and TH by immunoperoxidase. (A, B) After 3 DIV, few GAD-positive neurons (A)
and TH-positive neurons (B) are connected to melanotrope cells by a thin, weakly labeled, neurite (arrow).
(C) After 12 DIV, a single TH-positive neuron establishes contacts (arrows) with several melanotrope cells
and displays a highly developed neuritic arborization. Scale bars = 20 µm (A, B); 40 µm (C).
synaptic events. These synaptic currents had variable
amplitudes (10–100 pA), and were characterized by
fast rise times (<1 ms) and slowly decaying phases.
The total duration of these events was typically between 100 and 200 ms. These synaptic currents were
completely and reversibly blocked in postsynaptic
melanotrope cells after local application of 5 µM
SR95531 (n = 5; Fig. 5A) or 5 µM bicuculline (n = 18;
data not shown), two selective and competitive antagonists of GABAA receptors. At the same concentration (5 µM), bicuculline (n = 8) and SR95531
(n = 6) also reversibly suppressed the synaptic activity
recorded in postsynaptic neurons (data not shown).
These results indicated that the contacts formed between hypothalamic neurons and melanotrope cells
corresponded to functional GABAergic synapses and
that synaptically released GABA activated GABAA
receptors.
Pattern of synaptic activity at neuroneuronal versus
neuroendocrine GABAergic synapses
As shown on Fig. 5B, neuroneuronal and neuroendocrine synapses displayed distinct patterns of
synaptic activity. Spontaneous synaptic events were
observed only in a subset of melanotrope cells, i.e.
22.9% (19 of 83 cells), although all of these cells
(n = 83) were synaptically connected to a presynaptic
neuron. Indeed, synaptic events could be induced in
these cells by local application of an extracellular
solution containing 50 mM KCl. There was no
significant change in the fraction of cells displaying
spontaneous activity with time in culture. For
example, this fraction was similar after eight days
(23%; n = 13) and 13 days (30.7%; n = 13) in culture.
The mean percentage of melanotrope cells exhibiting
spontaneous synaptic events was 29 & 11% as determined between seven and 14 days in culture in a total
of 83 synaptically connected cells from six different
co-cultures. Moreover, in this subset of melanotrope
cells, the spontaneous synaptic currents occurred at a
low frequency, i.e. 0.31 & 0.20 Hz (n = 19). In contrast, the mean frequency of spontaneous activity
recorded in neurons was about four-fold higher
(1.19 & 0.62 Hz; n = 16). In addition, the fraction
of neurons displaying spontaneous synaptic events
increased markedly with time in culture. For
example, 12.5% (n = 8) of the neurons possessed
spontaneous synaptic currents after eight days in
culture, whereas all of the neurons examined (n = 13)
displayed spontaneous activity after day 13.
Morphofunctional study of neuroendocrine synaptogenesis in vitro
209
Fig. 4. (A, B) Synapsin I labeling revealed by immunofluorescence in co-cultures. At 4 DIV (A), labeling
is diffusely distributed in the neuronal cell bodies and extensions; at 12 DIV (B), strongly labeled punctae
are apposed against weakly positive perikarya and neuritic extensions. (C, D) Adjacent semi-thin sections
of a 12-DIV co-culture, immunoperoxidase labeled for synapsin I (C) and á-MSH (D), revealing
appositions of synapsin-positive punctae to melanotrope cells (arrows). (E, F) Electron micrographs from
co-cultures showing synaptic differentiations between two neurites at 8 DIV (E) and a neuroglandular
contact at 12 DIV (F). The synaptic differentiations are indicated by the postsynaptic density, the synaptic
cleft (E), the accumulation of synaptic vesicles and presynaptic density (F). Scale bars = 20 µm (A–D);
500 nm (E, F).
Electrophysiological identification of cell types
As shown in the first part of this report, neurons
rapidly extended neurites after plating. The size of
these neurites increased with time in culture, whereas,
under the same conditions, melanotrope cells kept a
rounded phase-bright appearence and a constant
diameter of approximately 12 µm. This suggested
that the membrane surface of neurons should
increase with time in culture, whereas that of melanotrope cells should stay relatively constant. Since
the membrane capacitance of a cell is directly
proportional to the membrane area, this parameter is
a good index of the cell surface. Therefore, we
determined the value of the membrane capacitance of
210
F. René et al.
Fig. 5. Hypothalamic neurons form functional GABAergic synapses with melanotrope cells. (A) Example
of a recording obtained under whole-cell voltage-clamp in a melanotrope cell which had been contacted
by a neurite of a hypothalamic neuron. The transient downward deflections of the current trace reflect the
presence of spontaneous synaptic currents. These synaptic events were completely and reversibly blocked
by the competitive GABAA receptor antagonist SR95531 (5 mM), which was applied locally for the time
indicated by the horizontal bar. The inset (right) shows, on a faster time scale, some examples of synaptic
currents recorded in the same cell. Note the presence of fast rise times and slowly decaying phases which
are characteristic of synaptic currents. The recording was performed after eight days in culture and the
steady-state holding potential (HP) was set at "60 mV. (B) Comparison of the patterns of spontaneous
synaptic activity observed at synapses formed between two hypothalamic neurons (top trace) and between
a hypothalamic neuron and a melanotrope cell (bottom trace) after 12 DIV. Note the marked difference
in the frequency of synaptic events, which was higher in the case of the neuroneuronal synapse. The HP
was "60 mV in both cases. (C) Evolution of the value of membrane capacitance of hypothalamic neurons
(open circles) and of melanotrope cells (filled circles) as a function of time in culture (expressed in days).
The membrane capacitance was taken as an index of membrane area and was determined in separate
cultures of neurons and melanotrope cells. Note that after six days in culture the value of membrane
capacitance of neurons is markedly higher than that of melanotrope cells, and continues to increase at
later days in culture. In contrast, the value of membrane capacitance of melanotrope cells stayed at a
constant level over the same period. The symbols represent mean values and the vertical bars are the
standard deviations.
neurons and melanotrope cells with time in culture.
The measurements were performed in separate cultures of each cell type in order to avoid any possible
error. The membrane capacitance was deduced
directly from the capacitance compensation settings
of the patch-clamp amplifier. Figure 5C summarizes
the results obtained. The values of membrane capacitance of neurons and melanotrophs were relatively
close on the fifth day in culture, when capacitance
measured 13.5 & 5.3 pF (n = 7) for neurons and
7.8 & 0.7 pF (n = 15) for melanotrope cells. However, with time in culture, the membrane capacitance
of neurons continued to increase regularly (e.g.,
30 & 8 pF, n = 15, at day 14), whereas that of
melanotrope cells remained constant (7 & 1.7 pF,
n = 25, at day 14) over the same period. These results
suggest that, together with morphological characteristics (presence or absence of neurites), the value of
membrane capacitance and the pattern of synaptic
activity are useful criteria to distinguish melanotrope
cells from neurons in our co-culture system.
Development of functional synaptic connections
The pattern of development of functional neuroendocrine synapses was obtained by recording, at
different days in culture, melanotrope cells displaying
apparent contacts with neurites of hypothalamic
neurons. The presence of both spontaneously occurring or chemically evoked synaptic activities was
investigated. The latter was induced by applying
locally a solution containing 50 mM KCl, which
depolarized the presynaptic terminal and thereby
caused a calcium-dependent release of neurotransmitter. Figure 6 shows the general time-course of
functional synapse formation in our co-culture
Morphofunctional study of neuroendocrine synaptogenesis in vitro
211
Fig. 6. Development of functional synaptic contacts between hypothalamic neurons and melanotrope
cells. In these experiments, only melanotrope cells which had been contacted by one or several neurites
were taken into consideration. The existence of functional synapses was assessed by the presence of either
spontaneously occurring synaptic currents and/or the presence of synaptic events which could be evoked
by local application of an extracellular solution containing 50 mM KCl. The graphical representation
illustrates the percentage of cells displaying synaptic activity as a function of time in culture (in days).
Note the absence of functional synapses before day 6 and a steep subsequent increase between days 6 and
10, a stage at which virtually all melanotrope cells tested displayed synaptic events. The results illustrated
were obtained in a total of 137 cells. The symbols represent the mean values and the vertical bars are the
standard deviations.
system. The results illustrated represent recordings
from a total of 137 melanotrope cells in 26 different
co-cultures over a period ranging from the third to
14th days in culture. We never observed any spontaneous or chemically evoked synaptic activity before
day 6 (n = 16). At 6 DIV, 8.3 & 11.7% (n = 11) of the
melanotrope cells displayed synaptic currents. The
proportion of pituitary cells receiving synaptic inputs
continued to increase until day 10, after which
virtually all melanotrope cells (n = 48) in contact with
neurites were found to possess functional GABAergic
synapses.
DISCUSSION
In the present study, we show that our culture
conditions in serum-free medium allow the maturation of hypothalamic neurons and the establishment of functional neuroendocrine synapses with
melanotrope cells.
The hypothalamic neurons, which are poorly differentiated when dissociated, displayed signs of
maturation over time in culture. The number of
NSE-positive cells, very low just after plating (fewer
than 15% of cells with neuronal morphology),
increased thereafter. From this stage, the number of
neurons slowly decreased, but at 12 DIV, about 50%
of the initial neuronal population still displayed the
characteristics of differentiated neurons: TH or GAD
immunoreactivity, development of neuritic processes
and neuroneuronal or neuroglandular functional
contacts, attesting to the overall health of the
cells and the suitability of the co-culture for
synaptogenesis studies. However, the reduction of
postsynaptic activities (percentage of inhibitory postsynaptic currents; Fig. 6) from 14 DIV may indicate
that the co-culture system does not allow such studies
after further maintenance in culture under the prevailing conditions. As early as after 1 DIV, neurons
extended neurites, formed homotypic contacts and
established a dense neuritic network with a pattern
comparable to those described in embryonic day (E)
16 mouse hypothalamic cultures.3,17,26 We did not
investigate all possible types of neurons which could
be present in our cultures. In view of the specific
innervation of the IL, we first focused on dopaminergic (TH-positive) and GABAergic (GAD-positive)
neurons. The increase of TH and GAD expression
(until 7 DIV) is probably due to an increase in the
number of TH- and GAD-immunoreactive neurons
and an increase of immunoreactivities themselves,
which correlates with the maturation in vivo. Indeed,
during development, expression of TH increases
between E13 and postnatal day 3 within the rat
hypothalamus.7,33,36,37,43,44 GAD expression is first
detected around E156 in the hypothalamus. However,
in our system, we have a large proportion of GADpositive neurons (§80%) compared to that found
by Wahle et al.47 (30%) in cultures of E14–E15 rat
hypothalamic neurons alone. This discrepancy is not
due to a non-specific labeling with the anti-GAD
antibody used, because a specific labeling of
GABAergic fibers and cell bodies was found on
brain semi-thin and free-floating sections. However,
212
F. René et al.
this discrepancy could be related to differences in
tissue sampling, cell dissociation, culture conditions
or to the presence of the melanotrope cells in the
co-culture.
The melanotrope cells were taken from neonatal
rats (one to seven days old) at the time when
dopaminergic and GABAergic fibers invade the IL
and contact these cells.12 As demonstrated by in
vivo23,32,34 and in vitro experiments,10,12,46 at this
stage, almost all of the IL cells display the main
characteristics of differentiated melanotrope cells,
including pro-opiomelanocortin gene expression,
spontaneous release of á-MSH and â-endorphin,
functional dopamine D2 receptors, and ability to
form neuroendocrine synapses. Some of the phenotypic characteristics of melantrope cells (the expression and secretion of á-MSH) were conserved during
the time in culture, so it is likely that we reconstitute
in vitro the neuron–target interactive system, including differentiated melanotrope cells.
The apparent formation of contacts between
neurons and melanotrope cells was observed only
after 3 DIV and the proportion of melanotrope cells
receiving neuronal contacts increased through the
time in culture, despite the decrease of the neuronal
population after 7 DIV. The establishment of these
neuroendocrine contacts is subsequent to the formation of the first neuroneuronal appositions, which
occurred as early as 1 DIV. The delay observed in
the formation of neuroendocrine versus neuroneuronal contacts might be due to the lower density
of melanotrope cells compared to that of the
neurons over the time in vitro (from 1:10 after
1 DIV to 1:2.5 after 12 DIV). This delay may also
be related to the need of specific factors from the
endocrine cells, such as neurotrophic/neurotropic
factors or cell surface molecules which might
require some time in culture after dissociation prior
to their functioning.20
At the time of the first neuroendocrine contacts,
synapsin immunoreactivity is diffusely distributed in
the neurons. The diffuse synapsin I immunoreactivity in nerve cell bodies has been observed previously
both in vivo14 and at early times in vitro,1,16,19,48
and corresponds, as described in this report, to an
immature state where synapses are not functional.
The evolution of synapsin I immunoreactivity to a
punctate labeling is correlated with the appearance
of morphological and functional characteristics of
more mature neuroendocrine synapses. Indeed, we
observed ultrastructural features of mature synapses: clustering of electron-lucent vesicles in presynaptic axon terminals and symmetrical moderate
membrane thickening, as described in vivo by
Baumgarten et al.2 These different morphological
aspects of synapse maturation are similar to those
described in cultures of striatal48 and hippocampal19 neurons.
Moreover, patch-clamp recordings revealed that
the contacts formed between the neurons and the
endocrine cells corresponded to functional synapses.
The synaptic currents were specifically blocked by
competitive antagonists of the GABAA receptor,
which demonstrated that these synapses were
GABAergic. This electrophysiological approach
allowed us to describe the time-course of the formation of functional synapses in culture. Synaptic
currents were never observed in melanotrope cells
before 6 DIV. Between 6 and 10 DIV, the number
of melanotrope cells in contact with neurites, presenting spontaneous and chemically evoked synaptic
currents, increased markedly, reflecting the functional maturation of GABAergic synapses. This
direct evaluation of the development of functional
GABAergic synapses with time in culture matched
well with the morphofunctional synaptic maturation indicated by the redistribution of synapsin I
immunoreactivity (from the cell body to terminals)
and ultrastructural observations. Spontaneous synaptic currents were recorded only in a subset of
melanotrope cells (30%) and displayed specific
features (low frequency) as compared to those
recorded in neurons. We have no clear explanation
for this phenomenon, but it could be due to a
difference in electrical activity of neurons involved in
these connections, or related to an effect of the
postsynaptic element (melanotrope cell or neuron),
which could respond differently to, or regulate the
properties of, the presynaptic neuron. In the future,
this co-culture system might offer the possibility to
study the effects of different targets (neurons or
endocrine cells) on the last steps of presynaptic
differentiation.
In situ, the innervation of melanotrope cells arises
from neurons in which GABA and dopamine are
co-localized.45 This situation has to be confirmed in
our co-culture system. In this case, it will be interesting to study the formation of functional dopaminergic synapses as well as the possible co-release of
GABA and dopamine at the synaptic level. Double
immunostaining studies are currently underway to
investigate the presence of GABAergic/dopaminergic
neurons.
CONCLUSION
We report, for the first time, the development of a
functional neuroendocrine synapse in vitro. Furthermore, we describe a model of co-culture which
appears particularly well adapted to analyse the
mechanisms of synaptogenesis and synaptic transmission between pairs of connected cells. In the continuity of this work, a study concerning functional
GABAergic neuroendocrine synaptic transmission is
currently in progress.
Acknowledgements—The authors thank Dr M. E. Stoeckel
for helpful advice on electron microscopy and are indebted
to J. M. Gachon for excellent photographic work. This
study was supported by a grant from the DRET (no.
93086), ULP, CNRS and the European Community (association contract no. ERBCHRXCT 940569).
Morphofunctional study of neuroendocrine synaptogenesis in vitro
213
REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
Basarsky T. A., Parpura V. and Haydon G. (1994) Hippocampal synaptogenesis in cell culture: developmental time
course of synapse formation, calcium influx and synaptic protein distribution. J. Neurosci. 14, 6402–6411.
Baumgarten A. G., Björklund A., Holstein A. F. and Nobin A. (1972) Organization and ultrastructural identification
of catecholamine nerve terminals in the neural lobe and pars intermedia of the rat pituitary. Z. Zellforsch. mikrosk.
Anat. 126, 483–517.
Benda P., De Vitry F., Picart R. and Tixier-Vidal A. (1975) Dissociated cell cultures from fetal mouse hypothalamus
patterns of organization and ultrastructural features. Expl Brain Res. 23, 29–47.
Benfenati F. and Valtorta F. (1993) Synapsin and synaptic transmission. News physiol. Sci. 8, 18–23.
Björklund A., Moore R. Y., Nobin A. and Steveni U. (1973) The organization of tubero-hypophyseal and
reticulo-infundibular catecholamine neuron systems in the rat brain. Brain Res. 51, 171–191.
Bond R. W., Wyborski R. J. and Gottlieb D. I. (1990) Developmentally regulated expression of an exon containing
a stop codon in the gene for glutamic acid decarboxylase. Proc. natn. Acad. Sci. U.S.A. 87, 8771–8775.
Borisova N. A., Sapronova A. Y., Proshlyakova E. V. and Ugrumov M. V. (1990) Ontogenesis of hypothalamic
catecholaminergic system in rats: syntesis, uptake and release of catecholamines. Neuroscience 43, 223–229.
Bottenstein J. E. and Sato G. E. (1979) Growth of a rat neuroblastoma cell line in serum-free supplemented medium.
Proc. natn. Acad. Sci. U.S.A. 76, 514–519.
Chen C. L. C., Dionne F. T. and Roberts J. L. (1983) Regulation of the pro-opiomelanocortin mRNA levels in rat
pituitary by dopaminergic compounds. Proc. natn. Acad. Sci. U.S.A. 80, 2211–2215.
Chronwall B. M., Milligton W. R., Griffin W. S. T., Unnerstall J. R. and O’donohue T. L. (1987) Histological
evaluation of the dopaminergic regulation of proopiomelanocortin gene expression in the intermediate lobe of the rat
pituitary, involving in situ hybridization and (3H)thymidine uptake measurement. Endocrinology 120, 1201–1211.
Dahlström A. and Fuxe K. (1966) Monoamines and the pituitary gland. Acta endocr. 51, 301–314.
Davis M. D., Lichtensteiger W., Schlumpf M. and Bruinink A. (1984) Early postnatal development of pituitary
intermediate lobe control in the rat by dopaminergic neurons. Neuroendocrinology 39, 1–12.
De Camilli P., Cameron R. and Greengard P. (1983) Synapsin I (Protein I), a nerve terminal specific phosphoprotein.
I. Its general distribution in synapses of central nervous system in frozen and plastic sections. J. Cell Biol. 96,
1337–1354.
De Camilli P. and Greengard P. (1986) Synapsin I: a synaptic vesicle-associated neuronal phosphoprotein. Biochem.
Pharmac. 35, 4349–4357.
Demeneix B. A., Taleb O., Loeffler J. P. and Feltz P. (1986) GABA-A and GABA-B receptors on porcine pars
intermedia cells in primary culture: functional role in modulating peptide release. Neuroscience 17, 1275–1285.
Dubinsky J. M. (1989) Development of inhibitory synapses among striatal neurons in vitro. J. Neurosci. 9,
3955–3965.
Faivre-Bauman A., Rosenbaum E., Puymirat J., Grouselle D. and Tixier-Vidal A. (1981) Differentiation of fetal
mouse hypothalamic cells in serum-free medium. Devl Neurosci. 4, 118–129.
Fenwick E. M., Marty A. and Neher E. (1982) A patch-clamp study of bovine chromaffin cells and of their sensitivity
to acetylcholine. J. Physiol., Lond. 331, 577–597.
Fletcher T. L., Cameron P., De Camilli P. and Banker G. (1991) The distribution of synapsin I and synaptophysin in
hippocampal neurons developing in culture. J. Neurosci. 11, 1617–1626.
Fletcher T. L., De Camilli P. and Banker G. (1994) Synaptogenesis in hippocampal cultures: evidence indicating that
axons and dendrites become competent to form synapses at different stages of neuronal development. J. Neurosci. 14,
6695–6706.
Hall Z. W. and Sanes J. R. (1993) Synaptic structure and development: the neuromuscular junction. Cell/Neuron 72/10
Suppl., 99–121.
Hamil O. P., Marty A., Neher E., Sakmann B. and Sigworth F. J. (1981) Improved patch-clamp technique for high
resolution current recording from cell and cell-free membrane patches. Pflügers Arch. 391, 85–100.
Hindelang C., Felix J. M., Laurent F. M., Klein M. J. and Stoeckel M. E. (1990) Ontogenesis of proopiomelanocortine
gene expression and regulation in the rat pituitary intermediate lobe. Molec. cell. Endocr. 70, 225–235.
Höllt V., Haarmann I., Seizinger B. R. and Herz A. (1982) Chronic haloperidol treatment increases the level of in vitro
translatable messenger RNA coding for the â-endorphin-adrenocorticotropin precursor POMC in the pars intermedia
of the rat pituitary. Endocrinology 110, 1885–1891.
Jackson M. B., Lecar H., Brenneman D. E., Fitfzgerald S. and Nelson P. G. (1982) Electrical development in spinal
cord cell culture. J. Neurosci. 2, 1121–1141.
Legendre P., Tixier-Vidal A., Brigant J. L. and Vincent J. D. (1988) Electrophysiology and ultrastructure of mouse
hypothalamic neurons in culture: a correlative analysis during development. Devl Brain Res. 43, 273–285.
Loeffler J. P., Demeneix B. A., Pittius C. W., Kley N., Haegele K. D. and Höllt V. (1986) GABA differentially
regulates the gene expression of proopiomelanocortin in rat intermediate and anterior pituitary. Peptides 7,
253–258.
Loeffler J. P. P., Demeneix B. A., Kley N. A. and Höllt V. (1988) Dopamine inhibition of proopiomelanocortine gene
expression in the intermediate lobe of pituitary. Neuroendocrinology 47, 95–101.
Melloni R. H. and de Gennaro L. J. (1994) Temporal onset of synapsin I gene expression coincides with neuronal
differentiation during the development of the nervous system. J. comp. Neurol. 342, 449–462.
Navone F., Greengard P. and De Camilli P. (1984) Synapsin I in nerve terminals: selective association with small
synaptic vesicles. Science 226, 1209–1211.
Oertel W. H., Mugnaini E., Tappaz M. L., Weise W. K., Dahl A. L., Schmechel D. E. and Kopin I. J. (1982) Central
GABAergic innervation of neurointermediate pituitary lobe: biochemical and immunocytochemical study in the rat.
Proc. natn. Acad. Sci. U.S.A. 879, 675–679.
Pintar J. E. and Lugo D. I. (1987) Proopiomelanocortin gene expression, prohormone processing and secretion during
rat pituitary development. Ann. N. Y. Acad. Sci. 512, 318–327.
Reisert I., Schuster R., Zienecker R. E. and Pilgrim C. (1990) Prenatal development of mesencephalic and diencephalic
dopaminergic systems in the male and female rat. Devl Brain Res. 53, 222–229.
214
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
F. René et al.
René F., Hindelang C., Stoeckel M. E. and Felix J. M. (1994) Ontogeny of glucocorticoid and D2 receptors in the rat
pituitary: an in situ hybridization study. Molec. cell. Endocr. 105, 65–75.
Shibuya J., Kongsamut S. and Douglas W. W. (1992) Effectiveness of GABA B antagonists in inhibiting
baclofen-induced reductions in cytosolic free-calcium concentration in isolated melanotrophs of rat. Br. J. Pharmac.
105, 893–898.
Specht L. A., Pickel M. V., Joh T. H. and Reis D. J. (1981) Light-microscopic immunocytochemical localization of
tyrosine hydroxylase in prenatal rat brain. I. Early ontogeny. J. comp. Neurol. 199, 233–253.
Specht L. A., Pickel M. V., Joh T. H. and Reis D. J. (1981) Light-microscopic immunocytochemical localization of
tyrosine hydroxylase in prenatal rat brain. I. Late ontogeny. J. comp. Neurol. 199, 255–276.
Stack J. and Surprenant A. (1991) Dopamine actions on calcium currents, potassium currents and hormone release in
rat melanotrophs. J. Physiol. 439, 37–58.
Stoeckel M. E., Tappaz M., Hindelang C., Seweryn C. and Porte A. (1985) Opposite effects of monosodium glutamate
on the dopaminergic and GABAergic innervation of the median eminence and the intermediate lobe in the mouse.
Neurosci. Lett. 56, 249–255.
Taleb O., Loeffler J. P., Demeneix B. A., Kley N., Höllt V. and Feltz P. (1986) Ionic conductances related to GABA
action on secretory and biosynthetic activity of pars intermedia cells. Brain Res. Bull. 17, 725–730.
Taraskevitch P. S. and Douglas W. W. (1982) GABA directly affects electrophysiological properties on pituitary pars
intermedia cells. Nature 276, 733–734.
Tomiko S. A., Taraskevich P. S. and Douglas W. W. (1983) GABA acts directly on cells of pituitary pars intermedia
to alter hormone output. Nature 301, 706–707.
Ugrumov M. V., Tixier-Vidal A., Thibault J. and Mitskevich M. S. (1989) Ontogenesis of tyrosine hydroxylaseimmunopositive structures in the rat hypothalamus. An atlas of neuronal cell bodies. Neuroscience 29, 135–156.
Ugrumov M. V., Tixier-Vidal A., Thibault J. and Mitskevich M. S. (1989) Ontogenesis of tyrosine hydroxylaseimmunopositive structures in the rat hypothalamus. Fiber pathway and terminal fields. Neuroscience 29, 157–166.
Vuillez P., Carbejo-Perez S. and Stoeckel M. E. (1987) Colocalization of GABA and tyrosine hydroxylase
immunoreactivities in the axons innervating the neurointermediate lobe of the rat pituitary: an ultrastructural
immunogold study. Neurosci. Lett. 79, 53–58.
Vuillez P., Rene F., Plante M., Hindelang C., Klein M. J., Felix J. M. and Stoeckel M. E. (1992) Differentiation of
the melanotrophic cells of rat pituitary primordium in organotypic culture in defined medium. Cell Tiss. Res. 267,
169–183.
Wahle P., Müller T. H. and Swandulla D. (1993) Characterization of neurochemical phenotypes in cultured
hypothalamic neurons with immunocytochemistry and in situ hybridization. Brain Res. 611, 37–45.
Weiss S., Pin J. P., Sebben M., Kemp D. E., Sladeczek F., Gabrion J. and Bockaert J. (1986) Synaptogenesis of
cultured striatal neurons in serum-free medium: a morphological and biochemical study. Proc. natn. Acad. Sci.
U.S.A. 83, 2238–2242.
Williams P. J., MacVicar B. A. and Pittman Q. J. (1989) A dopaminergic inhibitory postsynaptic potential mediated
by an increased potassium conductance. Neuroscience 31, 673–681.
Williams P. J., MacVicar B. A. and Pittman Q. J. (1990) Synaptic modulation by dopamine of calcium currents in the
rat pars intermedia. J. Neurosci. 10, 757–763.
(Accepted 1 May 1996)