the effects of nitric oxide on magnocellular neurons could involve multiple indirect cyclic...

The effects of nitric oxide on magnocellular neurons couldinvolve multiple indirect cyclic GMP-dependent pathways

C. M. Vacher,1,2 H. Hardin-Pouzet,2 H. W. M. Steinbusch,1 A. Calas2 and J. De Vente1

1Department of Psychiatry and Neuropsychology, POB 616, European School of Neuroscience (EURON), Universiteit Maastricht,6200 MD Maastricht, the Netherlands2Laboratoire de Neurobiologie des Signaux Intercellulaires, CNRS UMR 7101, Universite Paris VI, 7, quai Saint-Bernard, 75252 ParisCedex 05, France

Keywords: brain slices, cGMP, immunohistochemistry, nitric oxide, paraventricular nucleus, supraoptic nucleus

Abstract

Nitric oxide (NO) is known to regulate the release of arginine-vasopressin (AVP) and oxytocin (OT) by the paraventricular nucleus (PVN)and the supraoptic nucleus (SON). The aim of the current study was to identify in these nuclei the NO-producing neurons and the NO-receptive cells in mice. The determination of NO-synthesizing neurons was performed by double immunohistochemistry for theneuronal form of NO synthase (NOS), and AVP or OT. Besides, we visualized the NO-receptive cells by detecting cyclic GMP (cGMP),the major second messenger for NO, by immunohistochemistry on hypothalamus slices. Neuronal NOS was exclusively colocalizedwith OT in the PVN and the SON, suggesting that NO is mainly synthesized by oxytocinergic neurons in mice. By contrast, cGMP wasnot observed in magnocellular neurons, but in GABA-, tyrosine hydroxylase- and glutamate-positive fibers, as well as in GFAP-stainedcells. The cGMP-immunostaining was abolished by incubating brain slices with a NOS inhibitor (L-NAME). Consequently, we providethe first evidence that NO could regulate the release of AVP and OT indirectly by modulating the activity of the main afferents tomagnocellular neurons rather than by acting directly on magnocellular neurons. Moreover, both the NADPH-diaphorase activity and themean intensity of cGMP-immunofluorescence were increased in monoamine oxidase A knock-out mice (Tg8) compared to control mice(C3H) in both nuclei. This suggests that monoamines could enhance the production of NO, contributing by this way to the fine regulationof AVP and OT release and synthesis.

Introduction

The inorganic gas nitric oxide (NO) has been shown to modulate

arginine-vasopressine (AVP) and oxytocin (OT) release by the mag-

nocellular neurons, mainly localized in the paraventricular nucleus

(PVN) and the supraoptic nucleus (SON) of the hypothalamus (Carter

& Murphy, 1989; Kadowaki et al., 1994; Villar et al., 1994; Yang &

Hatton, 1999; Srisawat et al., 2000). This action principally consists of

the inhibition of neuronal activity since both the NO donor sodium

nitroprusside (SNP) and the NO precursor L-arginine inhibit the

supraoptic neurons, whereas NOS inhibitors enhance the neuronal

activity in the SON (Liu et al., 1997; Srisawat et al., 2000). However,

Yang & Hatton (1999) have demonstrated a positive action of NO on

dye coupling and excitability of supraoptic neurons.

NO is an intercellular messenger synthesized from L-arginine and

molecular oxygen by NO synthases (NOS) that require b nicotinamide

adenine dinucleotide phosphate (NADPH) as an electron donor (Bredt

& Snyder, 1992; Lowenstein & Snyder, 1992). A family of three NOS

types has been identified: neuronal, endothelial and inducible forms

(nNOS, eNOS and iNOS). NO may act in a paracrine manner (De

Vente et al., 1998), inducing cyclic guanosine 30,50-monophosphate

(cGMP) production in the cytosol by binding to the soluble guanylyl

cyclase (sGC) (Knowles et al., 1989; Murad, 1994; Bhat et al., 1996).

In the rat PVN and SON, AVP and OT neurons strongly coexpress

not only the neuronal NOS (Bredt et al., 1991; Arevalo et al., 1992;

Hatakeyama et al., 1996), the major form of NOS in these structures

(Bhat et al., 1996), but also both the a1 and b1 subunits of sGC

(Furuyama et al., 1993). Consequently, it has been suggested that NO

plays a cGMP-dependent autoregulatory role in these neurons. Never-

theless, indirect pathways could also be involved. So far, the only

evidence of an indirect pathway concerns the potentiation by NO of the

GABAergic transmission into these neurons in rats (Ozaki et al., 2000;

Stern & Ludwig, 2001). However, since the magnocellular neurons are

modulated by various excitatory and inhibitory inputs, mainly repre-

sented by glutamatergic, GABAergic, catecholaminergic and seroto-

nergic afferents, we hypothesized that the effects of NO on

magnocellular neurons could involve multiple indirect pathways.

In this context, the current study was undertaken to identify the

cellular targets for NO in the PVN and SON, visualizing cGMP by

immunohistochemistry on mouse brain slices. Here, we present evi-

dences that NO, mainly synthesized by OT neurons, could modulate

the GABAergic, catecholaminergic, and glutamatergic transmissions

into magnocellular neurons, as well as astrocytic functioning, in the

mouse PVN and SON, via cGMP-dependent mechanisms. In addition,

elevated serotonin and noradrenaline in transgenic mice knocked-out

for monoamine oxidase A were found to increase both NO and cGMP

productions.

European Journal of Neuroscience, Vol. 17, pp. 455–466, 2003 � Federation of European Neuroscience Societies

doi:10.1046/j.1460-9568.2003.02467.x

Correspondence: Dr Claire-Marie Vacher, 2Laboratoire de Neurobiologie des Signaux

Intercellulaires, as above.

E-mail: [email protected]

Received 19 September 2002, revised 13 November 2002, accepted 15 November 2002

Fig. 1. Double immunostaining of nNOS and AVP or OT in the PVN (A–F) and in the SON (G–L) in mouse. These photographs are obtained using confocal laserscanning microscopy (projection of eight successive 1-mm-thick optical sections in normal 10 mm tissue sections). nNOS is shown in red fluorescence (cyanine-3), andthe other markers (AVP and OT) are represented by green fluorescence (fluoroscein isothiocyanate). Colocalization appears as (pale) yellow. (A–C) The doubleimmunostaining for nNOS and AVP in the PVN shows no colocalization. (D–F) The combined detection of nNOS and OT in the PVN reveals an extensivecolocalization between these two markers. (G–I) No colocalization is found between nNOS and AVP in the SON. (J–L) nNOS and OT are colocalized in the SON. III,third ventricle; OC, optic chiasma. Bars represent 50 mm.

456 C. M. Vacher et al.

� 2003 Federation of European Neuroscience Societies, European Journal of Neuroscience, 17, 455–466

Materials and methods

Animals

All experiments were performed in agreement with European legal

requirements (Decree 86/609/EEC and approved by the local com-

mittee on Animal Welfare). Two different mouse strains were used in

this work: C3H/HeJ (C3H) and Tg8 mice. Tg8 mice are characterized

by the inactivation of the monoamine oxidase A (MAO-A) gene and

descend from C3H mice. The mutation results in increased amounts

of 5-HT and NA in the brain (Cases et al., 1995), and notably in the

PVN and SON (Vacher et al., 2002). The 12 C3H and 8 Tg8 mice

(3-month-old-males) used in this study were housed on a 12 : 12

light-dark cycle (lights on at 07:00 h) with free access to food and

water. They were always killed at the same time of the day (3–4 h after

lights on).

Tissue preparation for brain slice experiments

Four C3H and 4 Tg8 mice were rapidly decapitated without anaes-

thesia, and their brains immediately removed and transferred into ice-

cold oxygenated Krebs-Ringer bicarbonate buffer (121.1 mM NaCl,

1.87 mM KCl, 1.17 mM KH2PO4, 1.15 mM MgSO4�7H2O, 24.9 mM

NaHCO3, 2.0 mM CaCl2.2H2O and 11.0 mM glucose). Three hundred

mm-thick hypothalamus slices were prepared as described previously

(De Vente et al., 1998). Slices, slowly warmed to 35.5 8C, were

incubated in Krebs bicarbonate buffer. After 30-min incubation,

0.1 mM sodium nitroprusside (SNP; Sigma, Lyon, France) as a NO

donor was added in the incubation medium for 10 min. When added,

NG-nitro-L-arginine methyl ester (L-NAME; 3mM; Sigma) was present

for the duration of the incubation. In that case, SNP was not added in

the incubation medium.

Tissue preparation for immunohistochemistry and NADPH-diaphorase histochemistry

Double-immunostainings including the detection of cGMP were per-

formed on incubated brain slices. After the 40-min incubation, slices

were fixed for 30 min in 4% freshly depolymerized paraformaldehyde

in 0.1 M phosphate buffer (pH 7.4), and then in the same fixative

containing 10% sucrose for 2 h at 4 8C. Thereafter, they were frozen in

CO2, and cut with a cryostat at �21 8C into 10-mm thick sections.

These sections were thawed onto chrome-alumn/gelatin slides, and

stored at �20 8C until used.

Double-immunodetections including the staining of nNOS, as well

as the NADPH-diaphorase staining, were performed on perfused fixed

brains. Briefly, eight C3H mice were anaesthetized with sodium

pentobarbital (25 mg/kg) and perfused through the left ventricle with

50 mL of saline, followed by 50 mL of 4% paraformaldehyde in 0.1 M

phosphate buffer (pH 7.4). Brains were removed, post-fixed in the

same fixative for 4 h, and cryoprotected in 20% sucrose. Serial coronal

18 mm-thick sections were cut on a cryostat at �21 8C, thawed onto

chrome-alumn/gelatin slides and stored at �20 8C.

Immunohistochemistry

Sections were washed in three 5-min baths with Tris-buffered saline

(TBS) before being incubated overnight at 4 8C with sheep anti-

formaldehyde fixed cGMP (1 : 4000; De Vente et al., 1998), mouse

anti-nNOS (1 : 1000; Sigma), rabbit anti-AVP (1 : 4000; gift from

Dr G. Alonso; Alonso et al., 1988), rabbit anti-OT (1 : 4000; gift from

Dr G. Alonso; Alonso et al., 1988), mouse anti-g-aminobutyric acid

(GABA; 1 : 500; Sigma), rabbit anti-presynaptic glutamate transporter

(EAAC1; 1 : 1000; gift from Dr J.D. Rothstein; Furuta et al., 1997),

Fig. 2. Immunodetection of cGMP in the PVN (A and B) and in the SON (C and D) from brain slices incubated without SNP for 10 min. (A–C) Control experiment.(B–D) Incubation with L-NAME (3mM). In the presence of L-NAME, the cGMP-immunostaining is abolished both in the PVN and the SON. Bars represent 50 mm (Aand B), and 40mm (C and D).


NO targets in magnocellular nuclei 457

Fig. 4. Confocal laser scanning microscopy of double-immunostained sections from PVN (A–C and G–I) and SON (D–F and J–L) slices incubated in the presence of0.1 mM SNP. The green colour stands for cGMP-immunoreactivity, whereas the red colour represents either TH (B, C, E and F) or 5-HT (H, I, K and L). Colocalizationappears as yellow on the projection of eight successive 1-mm-thick optical sections (large photographs), or on detailed single 1-mm-thick optical sections (smallinserts). (A–F) There is extensive colocalization of cGMP-immunoreactivity in TH-positive fibers both in the PVN and the SON. (G–L) Colocalization betweencGMP and 5-HT is exceptional in the PVN and the SON. Bars represent 20 mm (A–C and G–L), and 15mm (D–F).

Fig. 3. cGMP-immunoreactivity in the PVN (B, D and F) and in the SON (I and J) in combination with AVP (A–G), OT (C and H), and GFAP (E) from brain slicesincubated with 0.1 mM SNP. No colocalization is found between cGMP and AVP, and between cGMP and OT. In the other hand, an extensive colocalization isobserved between cGMP and GFAP in the PVN. Bars represent 40 mm (A–F), and 25mm (G–J).



rabbit anti-tyrosine hydroxylase (TH; 1: 2000; J. Boy, Reims, France),

rabbit anti-serotonin (5-HT; 1: 2000; gift from Dr Y. Tillet; Tillet et al.,

1986), or mouse anti-glial fibrillary acid protein (GFAP; 1 : 1600;

Innogenetics), in TBS containing 0.3% Triton X-100 (TBS-T). Double

staining experiments were also performed, incubating simultaneously

the sections with a combination of these antibodies. The sheep anti-

body was visualized using Alexa-488 conjugated donkey anti-sheep

immunoglobulins (1 : 100; Molecular Probes, Leiden, The Netherlands),

while the rabbit and mouse antibodies were visualized using cyanine-3

conjugated donkey anti-rabbit or mouse antiserum (1 : 800; Jackson,

West Grove, USA), respectively. Control experiments were performed

by omitting primary or secondary antibodies.

Fluorescence microscopy

Sections were examined with an Olympus AX-70 TRF microscope,

and pictures were made with a Sony Power HAD 3CCD Colour Video

Camera. Semi-quantification analysis of cGMP immunostaining inten-

sity was performed using the computer program analySIS Version 3.0.

Briefly, a colour separation of the green image was carried out and the

mean grey value of each manually surrounded region was considered

as a measure for the cGMP content in each of these regions, namely the

PVN and the SON. Results for different groups, namely C3H and Tg8

mice, were represented in arbitrary unit (AU)� SEM, and compared

by one-way ANOVA, followed by a Scheffe’s test, and were considered

as significant if P <0.05. Moreover, double staining experiments were

visualized by confocal laser scanning microscopy with a Leica TCS NT

system (Leica Microsystems, Heidelberg, Germany). The argon-krypton

laser was used to excite the fluorochromes at 488 nm (fluoroscein

isothiocyanate) and 568 nm (cyanine-3). The fluoroscein isothiocya-

nate and the cyanine-3 fluorescences were selected by a 550–555 nm

and a 580–700 nm long-pass filter, respectively. Cross-over fluores-

cence was negligible. Each optical section (1 mm) was averaged four

times. Pictures were either optical sections or the projection of eight

successive optical sections into one image. Colocalization of the two

fluorescent markers is visualized in yellow in the combined pictures.

NADPH-diaphorase staining

NADPH-diaphorase staining was performed as described by Scherer-

Singler et al. (1983) by incubating slide-mounted sections with 1 mM

NADPH, 0.2 mM nitroblue tetrazolium, 0.1 M Tris-HCl (pH 7.2), 0.2%

Triton X-100 for 1 h at 37 8C. After being rinsed in a bath of Tris-HCl,

the sections were mounted in Permount for observation under a light

microscope (Leitz, Wetzlar, Germany).

Results

Identification of NO-producing neurons

The following immunohistochemical data for nNOS were obtained

from paraformaldehyde-perfused animals.

In the anterior PVN of C3H mice (bregma, �0.7 mm), nNOS-IR cell

bodies mainly surrounded the PVN, but they were quite rarely observed

in the core of the PVN (Fig. 1A). On the other hand, nNOS-IR

perikarya represented a dense neuronal population in the posterior

PVN (bregma, –0.94 mm), especially in its dorsal and ventral portions

(Fig. 1D). Moreover, nNOS-IR neurons were located in the dorsal part

of the SON throughout its whole antero-posterior axis (Fig. 1G and J).

In both regions, only few limited nNOS-IR processes were observed.

AVP-IR neurons were mainly regrouped in an extended portion of

the anterior PVN (Fig. 1B) and occupied the major ventral part of the

SON (Fig. 1H). In addition, varicose AVP-immunostained fibers were

observed inside the PVN and the SON as well as emerging laterally

from the PVN in direction to the SON. No colocalization was observed

between nNOS and AVP in the PVN (Fig. 1C) and in the SON (Fig. 1I).

OT-IR cell bodies were primarily located in the posterior portion of

the PVN (Fig. 1E) as well as in the dorsal SON (Fig. 1K). They were

less numerous and more scattered than AVP-positive neurons. Two

kinds of OT-stained processes were distinguished: short fibers limited

to the PVN and SON, and long processes leaving the PVN and heading

laterally and ventrally for the SON. Extensive colocalization between

nNOS immunoreactivity and OT was found both in the PVN (Fig. 1F)

and the SON (Fig. 1L). Rare perikarya (less than three per section)

exhibited a simple staining for nNOS or OT in the PVN (Fig. 1F).

Identification of NO receptive, cGMP-producing cells

The immunohistochemical data presented for cGMP are from slices

that were incubated in the presence of 0.1 mM SNP as a NO donor.

Without SNP, cGMP-immunoreactivity in the hypothalamus was

similarly distributed but less intense. Compared to control experiments

(Fig. 2A and C), the presence of 3 mM L-NAME during the incubation

of slices reduced dramatically the intensity of the fluorescent cGMP-

immunostaining both in the PVN (compare Fig. 2A and B) and the

SON (Fig. 2C vs. 2D).

Hypothalamus slices exhibited cGMP immunoreactivity in small

and sparse astrocyte-like cell bodies showing short immunopositive

processes. These immunostained cells, distributed all over the

hypothalamus, were particularly abundant in the PVN (Fig. 3B, D

and F). By contrast, they were rarely observed in the SON (Fig. 3I and

J). In addition, a diffuse or fibre-like cGMP labelling wrapped immu-

nonegative perikarya and blood vessels. This staining presented an

aspect of small particles homogeneously distributed in the hypotha-

lamus, notably in the PVN and SON. On the other hand, no neuronal

perikarya-like staining was observed in the PVN (Fig. 3B, D and F) and

in the SON (Fig. 3I and J).

No colocalization was found between cGMP and AVP in the PVN

(Fig. 3A and B) and the SON (Fig. 3G and I). As well, cGMP staining

was never colocalized with OT in the PVN (Fig. 3C and D) and in the

SON (Fig. 3H and J). These data confirm our first observations indicating

the absence of neuronal perikarya-like cGMP staining in these regions.

The immunofluorescent detection of GFAP revealed the presence of

sparse astrocyte-like cells and processes in the hypothalamus. In the

PVN, these cells were quite scattered (Fig. 3E), while in the SON,

GFAP-immunolabelling was concentrated in radial glia emerging from

the ventral borderline of the nucleus. In the PVN each GFAP-IR cell

was also cGMP-immunopositive (arrows Fig. 3E and F), although some

rare cGMP-IR astrocytic-like cell bodies remained GFAP-negative. No

colocalization was observed between cGMP and GFAP in the SON.

The immunodetection of TH revealed the presence of a dense

network of varicose fibers encompassing large immunonegative peri-

karya both in the PVN (Fig. 4B) and the SON (Fig. 4E). The codetec-

tion for TH and cGMP visualized by confocal laser scanner

microscopy on the projection of eight successive 1-mm optical sec-

tions, and emphasized on unique focal planes, showed that each TH-

positive fibre was also cGMP-immunolabelled in the PVN (Fig. 4A–C)

and the SON (Fig. 4D–F).

Fig. 5. Confocal laser scanning microscopy of double-immunostained sections from PVN (A–C and G–I) and SON (D–F and J–L) slices incubated with 0.1 mM SNP.The green colour stands for cGMP-immunoreactivity, whereas the red colour represents either EAAC1 (B, C, E and F) or GABA (H, I, K and L). Colocalizationappears as yellow on the projection of eight successive 1-mm-thick optical sections (large photographs), or on detailed single 1-mm-thick optical sections (smallinserts). (A–F) There is extensive colocalization of cGMP-immunoreactivity with EAAC1-immunoreactivity both in the PVN and the SON. (G–L) Colocalizationbetween cGMP and GABA is also extensive in the PVN and the SON. Bars represent 30 mm (A–F), 20 mm (G–I), and 10 mm (J–L).



The immunostaining for 5-HT displayed varicose processes in the

PVN (Fig. 4H) and the dorsal SON (Fig. 4K). Colocalization between

5-HT and cGMP was an exception both in the PVN (Fig. 4G–I) and the

SON (Fig. 4J–L), as emphasized on the detailed focal planes.

We detected the presence of glutamatergic terminals by using an

antibody raised against the presynaptic glutamate transporter EAAC1.

The staining for EAAC1 took the particulary appearence of synaptic

boutons either wrapping cell bodies or forming fibers both in the

PVN (Fig. 5B) and the SON (Fig. 5E). As showed on the detailed

focal planes, some of these synaptic boutons were stained doubly

for EAAC1 and cGMP in the PVN (Fig. 5A–C) and the SON

(Fig. 5D–F).

The immunodetection of GABA revealed the presence of numerous

varicosities homogeneously scattered in the PVN (Fig. 5H) and the

Fig. 6. Effect of the lack of MAO-A on NADPH-diaphorase staining in the PVN (A and B) and in the SON (C and D), and on the mean cGMP-immunostainingintensity in the PVN (E) and in the SON (F). Compared to wild-type mice (C3H), Tg8 mice (KO for the MAO-A) exhibit an increase in the staining intensity ofNADPH-diaphorase both in the PVN (A and B) and the SON (C and D). Besides, the mutation is associated with an augmentation of the staining intensity for cGMP inthe PVN (E) and the SON (F). III, third ventricle; OC, optic chiasma. Bars represent 50 mm.



SON (Fig. 5K), and enclosing immunonegative perikarya. Each of

them exhibited a double staining for GABA and cGMP, as emphasized

on optical sections in the PVN (Fig. 5G–I) and the SON (Fig. 5J–L).

Effect of the lack of MAO-A on NO production

In C3H mice, the NADPH-diaphorase active cells were identified

primarily as neuron-like cells in the PVN (Fig. 6A) and the SON

(Fig. 6C). In the PVN, these stained-neurons were observed mainly in a

posterior position. In the SON, they were located in the dorsal portion

of the nucleus. Moreover, NADPH-diaphorase-active neurons exhib-

ited more or less long stained-processes in the PVN and the SON. In

both nuclei, NADPH-diaphorase-stained neurons were distributed

similarly to OT neurons (compare Fig. 6A and C, and Fig. 1E and

K). The colocalization between NADPH-diaphorase staining and OT

was confirmed by the codetection of NADPH-diaphorase activity and

OT immunolabelling (not shown). By contrast, no colocalization was

found between NADPH-diaphorase staining and AVP-immunoreac-

tivity both in the PVN and the SON (not shown). In mice lacking

MAO-A (Tg8 mice), the distribution of NAPDH-diaphorase-active

neurons was similar to that observed in C3H mice both in the PVN

(compare Fig. 6B and A) and the SON, although these neurons were

more numerous in the SON of Tg8 compared to C3H mice (Fig. 6D vs.

C). Besides, the mutation was associated with an increased staining

intensity in perikarya and processes in both nuclei. As well, the mean

staining intensity measured on cGMP-immunolabelled sections was

significantly increased by 102% (P< 0.01) in the PVN (Fig. 6E) and

by 66% (P< 0.05) in the SON (Fig. 6F) in Tg8 compared to C3H mice.

Discussion

The aim of our study was to identify the neurons that constitutively

synthesize NO and the NO-receptive/cGMP-producing cells in the

PVN and the SON in mouse. To address these questions, we detected

the neuronal form of NOS by classical immunohistochemistry, and

cGMP by immunohistochemistry on brain slices ex vivo. Evidence

presented here indicates that NO (i) is produced by OT but not AVP

neurons, and (ii) could induce cGMP synthesis in astrocytes as well as

in GABAergic, catecholaminergic, in a lesser extent, glutamatergic,

and, exceptionally, 5-HTergic fibers. Moreover, the use of the trans-

genic strain Tg8, knock-out for the MAO-A, allowed us to suggest a

positive action of monoamines on NO and cGMP productions both in

the PVN and the SON.

By performing multiple immunohistochemical stainings for nNOS,

AVP and OT, we demonstrated that nNOS is expressed in OT, but not in

AVP neurons, both in the PVN and the SON in mice. This observation

was confirmed by NADPH-diaphorase histochemistry, a method that

allows to visualize the NOSs, and especially the nNOS on fixed brain

sections. Previous studies carried out in rats have related that the nNOS

immunostaining and the NADPH-diaphorase staining are present both

in AVP and OT neurons (Bredt et al., 1991; Arevalo et al., 1992;

Hatakeyama et al., 1996). Consequently, our work suggests that the

nNOS could be differently expressed in the magnocellular nuclei,

depending on the species. Nevertheless, our data also reinforce a

number of studies reporting a preferential colocalization of nNOS

and NADPH-diaphorase staining with OT rather than with AVP in the

magnocellular nuclei (Torres et al., 1993; Miyagawa et al., 1994;

Sanchez et al., 1994; Hatakeyama et al., 1996).

In a second place, the determination of NO-receptive cells was

realized by visualizing cGMP, the major second messenger of NO, on

brain slices incubated with SNP, a NO donor. Indeed, although NO

induces changes in signalling-related functions by several ways (like

the nitrosylation, the nitration or the oxidation of proteins), most of the

NO effects on neuronal activity is subordinated to cGMP synthesis

(reviewed by Prast & Philippu, 2001). By activating the sGC, NO

induces the augmentation of cGMP production and the subsequent

activation of cGMP-dependent kinases (PKG) in its cellular targets

(Wang & Robinson, 1997; Smolenski et al., 1998). In our study, we

demonstrate that AVP and OT neurons, known to be mainly inhibited

by NO (Carter & Murphy, 1989; Kadowaki et al., 1994; Villar et al.,

1994; Liu et al., 1997; Srisawat et al., 2000), do not represent a main

direct cellular target for NO since they do not synthesize cGMP in

response to NO. Nevertheless, magnocellular neurons are known to

express sGC in rats (Furuyama et al., 1993). This can suggest either the

existence of a dimorphism between rats and mice, or a particularly

important degradation of cGMP by phosphodiesterases in the mouse

magnocellular neurons. However, our work indicates that cGMP is

mainly produced by astrocytes, as well as by GABAergic, catecho-

laminergic and glutamatergic fibers. Since the cGMP-staining observed

without SNP was abolished by incubating the slices with L-NAME, an

NOS inhibitor, our work shows that NO could modulate the activity of

magnocellular neurons indirectly via their main cellular afferents.

Our study indicates that NO could modulate the GABAergic

transmission to the magnocellular nuclei. Such a regulation is sup-

ported by electrophysiological studies indicating that NO inhibits AVP

and OT neurons indirectly by activating the GABAergic synaptic

transmission in the SON (Ozaki et al., 2000; Stern & Ludwig,

2001). The NO/cGMP/cGMP-dependent protein kinase (PKG) path-

way has been shown to regulate the efficiency of the GABAergic

transmission by modulating (i) the release of GABA in the nucleus

accumbens (Kraus & Prast, 2002), possibly by the modulation of

cation channels as described in the hippocampus (Erdemli & Krnjevic,

1995) and in the locus coeruleus (Pineda et al., 1996), as well as (ii) the

postsynaptic GABA-A receptor function in the cerebellum (Zarri et al.,

1994; Robello et al., 1996). Since no cGMP-immunoreactivity was

found in the magnocellular neurons, that bear the GABA-A receptors,

but was colocalized with GABA, our work indicates that the regulation

of the GABA transmission by NO could involve, at least partly, the

cGMP-dependent modulation of the GABA release in the PVN and the

SON. Such a mechanism has never been described in the magnocel-

lular nuclei, but the NO/cGMP/PKG pathway has been demonstrated

to activate the release of GABA in the striatum (Trabace & Kendrick,

2000) and the nucleus accumbens (Kraus & Prast, 2002). Thus, we

hypothesize that NO could inhibit the activity of magnocellular

neurons as well as the release of AVP and OT by these neurons by

activating the synaptic GABA release in contact to the magnocellular

neurons. This may be a major effect of NO in the PVN and the SON

since 45% of synaptic boutons are GABA-IR in the SON in control

condition (Decavel & van den Pol, 1990).

Likewise, our study suggests that NO could modulate, via a cGMP-

dependent mechanism, the catecholaminergic transmission in the PVN

and the SON. The catecholaminergic input represents the most impor-

tant monoaminergic afferent in the PVN and the SON (Sawchenko &

Swanson, 1982; Cunningham & Sawchenko, 1988). This innervation

mainly activates magnocellular neurons as well as the synthesis and the

release of AVP and OT (Bridges et al., 1976; Mason, 1983; Willoughby

et al., 1987; Bealer & Crowley, 1998; Vacher et al., 2002). Because the

major effects of NO on magnocellular neurons are opposite to those of

catecholamines, our work indicates that NO could inhibit the cate-

cholaminergic transmission to AVP and OT neurons in the PVN and

the SON. By contrast, colocalizations between cGMP and 5-HT were

exceptional. This indicates that the scarce serotonergic input does not

represent a significant target for NO in the PVN and the SON.

Moreover, since cGMP was also colocalized with the presynaptic

transporter EAAC1, we hypothesize that NO could modulate the



glutamatergic transmission via a cGMP-dependent mechanism in the

PVN and the SON. The glutamatergic input, represented in a quarter of

synapses in the magnocellular nuclei (El Majdoubi et al., 1996),

activates the release of AVP and OT by the magnocellular neurons

(Parker & Crowley, 1993; Herbison et al., 1997). In addition, the effect

of NO being opposite to that of glutamate in regard to the release of

AVP and OT, we suggest that NO could inhibit the glutamatergic

transmission to magnocellular neurons.

Additionally, cGMP was detected in astrocytes, identified by their

GFAP-immunoreactivity, in the PVN but not the SON. Consequently,

astrocytes could also represent an important target for NO in the PVN.

We hypothesize that NO could modulate astrocytic functions in a

cGMP-dependent manner in the PVN. Astrocytes play a fundamental

role in the control of AVP and OT synthesis and release in the

magnocellular nuclei. By wrapping magnocellular neurons, astrocytes

contribute to the regulation of synaptic efficacity committing notably

glutamate and GABA (Hatton, 1999; Theodosis & Poulain, 1999; Oliet

et al., 2001). Moreover, the stimulation of the hypothalamo-neurohy-

pophysial system during lactation induces a decrease in the glial

coverage, resulting in increased numbers of glutamatergic, GABAer-

gic and noradrenergic synapses in the magnocellular nuclei (Hatton

et al., 1984; Theodosis & Poulain, 1989; Theodosis et al., 1998;

Theodosis & Poulain, 1999; Salm, 2000). NO, known to regulate the

intercellular Ca2þ waves in primary glial cell cultures via the produc-

tion of cGMP (Willmott et al., 2000), could also be involved in the

synchronization of the morphological and metabolic glial changes in

the PVN necessary to modulate the activity of magnocellular neurons.

The use of the transgenic mouse Tg8 allowed us to demonstrate that

an increase in 5-HT and NA concentration in the PVN and the SON

(Vacher et al., 2002) is associated with an augmentation in the staining

intensity of NADPH-diaphorase in OT neurons, as well as an increase

in the intensity of cGMP-immunoreactivity in the PVN and the SON.

This suggests that monoamines positively regulate the expression of

nNOS and/or stimulate the activity of nNOS in these neurons. Several

reports indicate that the expression of nNOS in the magnocellular

system is functionally regulated, notably in case of chronic salt loading

(Kadowaki et al., 1994; Villar et al., 1994) or during pregnancy,

parturition and lactation (Okere & Higuchi, 1996; Okere et al.,

1996; Xu et al., 1996; Popeski et al., 1999; Srisawat et al., 2000),

suggesting a possible activation of the nNOS expression by mono-

amines.

In conclusion, our data provide a precedent for the control by NO of

magnocellular neurons. The current study demonstrates that NO is

mainly synthesized by OT neurons in the mouse PVN and SON. In

addition, the detection of cGMP, the main second messenger for NO,

allowed us to suggest that NO may modulate the release of AVP and

OT by regulating the synaptic transmission of GABA, catecholamines,

glutamate, and, exceptionally, 5-HT to the magnocellular neurons,

rather than by acting directly on magnocellular neurons. Likewise,

astrocytes, that represent important NO-receptive/cGMP-producing

cells in the magnocellular hypothalamus, could also be involved in

the transmission of the NO signal to the magnocellular neurons. In

addition, our work indicates that the interactions between NO and the

main afferents to magnocellular neurons are bi-directional since

monoamines could activate the production of NO and cGMP. Taken

together, these data prove that the regulation by NO of the neuroendo-

crine hypothalamus involves a multiplicity of complex loops allowing

to adapt finely the release of AVP and OT to the organism needs

(Fig. 7).

Acknowledgements

This study was performed as Marie Curie Fellowship (‘Quality of Life andManagement of Living Resources’; contract no. QLK6-CT-2000–60042; grant# QLK6-GH-00-60042-04). We wish to thank M. Markerink for her skilfultechnical assistance in brain slice experiments, and R. Schwartzmann for hishelp in confocal microscopy. We are grateful to Dr I. Seif for providing the firstcouples of C3H and Tg8 mice, Dr Y. Tillet for his donation of 5-HT antiserum,Dr J.D. Rothstein for his gift of EAAC1 antiserum, and Dr G. Alonso for his giftof AVP and OT antisera.

Abbreviations

5-HT, serotonin; AVP, arginine-vasopressin; cGMP, cyclic guanosine mono-phosphate; EAAC1, presynaptic glutamate transporter molecule; GABA, gammaamino-butyric acid; GFAP, glial fibrillary acid protein; -IR, immunoreactive;L-NAME, NG-nitro-L-arginine methyl ester; MAO-A, monoamine oxidase A;NA, noradrenaline; NADPH, b nicotinamide adenine dinucleotide phosphate;NO, nitric oxide; NOS, NO synthase; eNOS, endothelial NOS; iNOS,inducible NOS; nNOS, neuronal NOS; OT, oxytocin; PKG, cGMP-dependentprotein kinase; PVN, paraventricular nucleus; sGC, soluble guanylyl cyclase;SNP, sodium nitroprusside; SON, supraoptic nucleus; TBS, Tris-bufferedsaline; TBS-T, TBS containing 0.3% Triton X-100; TH, tyrosine hydroxylase.

References

Alonso, G. (1988) Effects of colchicine on the intraneuronal transport ofsecretory material prior to the axon: a morphofunctional study in hypotha-lamic neurosecretory neurons of the rat. Brain Res., 453, 191–203.

Arevalo, R., Sanchez, F., Alonso, J.R., Carretero, J., Vazquez, R. & Aijon, J.(1992) NADPH-diaphorase activity in the hypothalamic magnocellularneurosecretory nuclei of the rat. Brain Res. Bull., 28, 599–603.

Bealer, S.L. & Crowley, W.R. (1998) Noradrenergic controls of central oxytocinrelease during lactation in rats. Am. J. Physiol., 274, E453–E458.

Bhat, G., Mahesh, V.B., Aguan, K. & Brann, D.W. (1996) Evidence that brainnitric oxide synthase is the major nitric oxide synthase isoform in thehypothalamus of the adult female rat and that nitric oxide potently regulateshypothalamic cGMP levels. Neuroendocrinology, 64, 93–102.

Fig. 7. Summary diagram. In the mouse magnocellular neurons, NO, mainlyproduced by OT neurons, is proposed to modulate the synaptic transmissionof GABA, catecholamines, glutamate, and exceptionally 5-HT, as well as theastrocytic activity via cGMP-dependent mechanisms. Additionally, mono-amines could regulate the production of NO from OT neurons. These regulatingloops could be involved in the fine regulation of the magnocellular systemactivity.



Bredt, D.S., Glatt, C.E., Hwang, P.M., Fotuhi, M., Dawson, T.M. & Snyder, S.H.(1991) Nitric oxide synthase protein and mRNA are discretely localized inneuronal populations of the mammalian CNS together with NADPH dia-phorase. Neuron, 7, 615–624.

Bredt, D.S. & Snyder, S.H. (1992) Nitric oxide, a novel neuronal messenger.Neuron, 8, 3–11.

Bridges, T.E., Hillhouse, E.W. & Jones, M.T. (1976) The effect of dopamine onneurohypophysial hormone release in vivo and from the rat neural lobe andhypothalamus in vitro. J. Physiol., 260, 647–666.

Carter, D.A. & Murphy, D. (1989) Cyclic nucleotide dynamics in the rathypothalamus during osmotic stimulation: in vivo and in vitro studies. BrainRes., 487, 350–356.

Cases, O., Seif, I., Grimbsy, J., Gaspar, P., Chen, K., Pournin, S., Muller, U.,Aguet, M., Babinet, C., Chen Shih, J. & de Maeyer, E. (1995) Aggressivebehavior and altered amounts of brain serotonin and norepinephrine in micelacking MAOA. Science, 268, 1763–1766.

Cunningham,E.T. Jr & Sawchenko, P.E. (1988) Anatomical specificity ofnoradrenergic inputs to the paraventricular and supraoptic nuclei of therat hypothalamus. J. Comp. Neurol., 274, 60–76.

De Vente, J., Hopkins, D.A., Markerink-Van Ittersum, M., Emson, P.C.,Schmidt, H.H.H.W. & Steinbusch, H.W.M. (1998) Distribution of nitricoxide synthase and nitric oxide-receptive, cyclic GMP-producing structuresin the rat brain. Neuroscience, 87, 207–241.

Decavel, C. & van den Pol, A.N. (1990) GABA: a dominant neurotransmitter inthe hypothalamus. J. Comp. Neurol., 302, 1019–1037.

El Majdoubi, M., Poulain, D.A. & Theodosis, D.T. (1996) The glutamatergicinnervation of oxytocin- and vasopressin-secreting neurons in the ratsupraoptic nucleus and its contribution to lactation-induced synaptic plas-ticity. Eur. J. Neurosci., 8, 1377–1389.

Erdemli, G. & Krnjevic, K. (1995) Nitric oxide tonically depresses a voltage-and Ca-dependent outward current in hippocampal slices. Neurosci. Lett.,201, 57–60.

Furuta, A., Rothstein, J.D. & Martin, L.J. (1997) Glutamate transporter proteinsubtypes are expressed differentially during rat CNS development. J. Neu-rosci., 17, 8363–8375.

Furuyama, T., Inagaki, S. & Tagaki, H. (1993) Localizations of alpha 1 and beta1 subunits of soluble guanylate cyclase in the rat brain. Brain Res. Mol. BrainRes., 20, 335–344.

Hatakeyama, S., Kawai, Y., Ueyama, T. & Senba, E. (1996) Nitric oxidesynthase-containing magnocellular neurons of the rat hypothalamus synthe-size oxytocin and vasopressin and express Fos following stress stimuli. J.Chem. Neuroanat., 11, 243–256.

Hatton, G.I. (1999) Astroglial modulation of neurotransmitter/peptide releasefrom the neurohypophysis: present status. J. Chem. Neuroanat., 16, 203–221.

Hatton, G.I., Perlmutter, L.S., Salm, A.K. & Tweedle, C.D. (1984) Dyna-mic neuronal–glial interactions in hypothalamus and pituitary: implica-tions for control of hormone synthesis and release. Peptides, 5 (Suppl. 1),121–138.

Herbison, A.E., Voisin, D.L., Douglas, A.J. & Chapman, C. (1997) Profile ofmonoamine and excitatory amino acid release in rat supraoptic nucleus overparturition. Endocrinology, 138, 33–40.

Kadowaki, K., Kishimoto, J., Leng, G. & Emson, P.C. (1994) Up-regulation ofnitric oxide synthase (NOS) gene expression together with NOS activity inthe rat hypothalamo-hypophysial system after chronic salt loading: evidenceof a neuromodulatory role of nitric oxide in arginine vasopressin andoxytocin secretion. Endocrinology, 134, 1011–1017.

Knowles, R.G., Palacios, M., Palmer, R.M. & Moncada, S. (1989) Formation ofnitric oxide from 1-arginine in the central nervous system: a transductionmechanism for stimulation of the soluble guanylate cyclase. Proc. Natl.Acad. Sci. USA, 86, 5159–5162.

Kraus, M.M. & Prast, H. (2002) Involvement of nitric oxide, cyclic GMP andphosphodiesterase 5 in excitatory amino acid and GABA release in thenucleus accumbens evoked by activation of the hippocampal fimbria. Neu-roscience, 12, 331–343.

Liu, Q.S., Jia, Y.S. & Ju, G. (1997) Nitric oxide inhibits neuronal activity inthe supraoptic nucleus of the rat hypothalamic slices. Brain Res. Bull., 43,121–125.

Lowenstein, C.J. & Snyder, S.H. (1992) Nitric oxide, a novel biologic mes-senger. Cell, 70, 705–707.

Mason, W.T. (1983) Excitation by dopamine of putative oxytocinergic neuronesin the rat supraoptic nucleus in vitro: evidence for two classes of continuouslyfiring neurones. Brain Res., 267, 113–121.

Miyagawa, A., Okamura, H. & Ibata, Y. (1994) Coexistence of oxytocin andNADPH-diaphorase in magnocellular neurons of the paraventricular andsupraoptic nuclei of the rat hypothalamus. Neurosci. Lett., 171, 13–16.

Murad, F. (1994) Cyclic GMP: synthesis, metabolism, and function. Introduc-tion and some historical comments. Adv. Pharmacol., 26, 1–5.

Okere, C.O. & Higuchi, T. (1996) Down-regulation of endogenous nitric oxidesynthase in late-pregnancy and parturition in the rat hypothalamic magno-cellular neurons and neurohypophysis. Neurosci. Lett., 220, 133–136.

Okere, C.O., Higuchi, T., Kaba, H., Russell, J.A., Okutani, F., Takahashi, S. &Murata, T. (1996) Nitric oxide prolongs parturition and inhibits maternalbehavior in rats. Neuroreport, 7, 1695–1699.

Oliet, S.H., Piet, R. & Poulain, D.A. (2001) Control of glutamate clearance andsynaptic efficacity by glial coverage of neurons. Science, 292, 923–926.

Ozaki, M., Shibuya, I., Kabashima, N., Isse, T., Noguchi, J., Ueta, Y., Shige-matsu, A. & Yamashita, H. (2000) Preferential potentiation by nitric oxide ofspontaneous inhibitory postsynaptic currents in rat supraoptic neurones. J.Neuroendocrinol., 12, 273–281.

Parker, S.L. & Crowley, W.R. (1993) Stimulation of oxytocin release in thelactating rat by central excitatory amino acid mechanisms: evidence forspecific involvement of R.,S-alpha-amino-3-hydroxy-5-methylisoxazole-4-propionic acid-sensitive glutamate receptors. Endocrinology, 133, 2847–2854.

Pineda, J., Kogan, J.H. & Aghajanian, G.K. (1996) Nitric oxide and carbonmonoxide activate locus coeruleus neurons through a cGMP-dependentprotein kinase: involvement of a nonselective cationic channel. J. Neurosci.,16, 1389–1399.

Popeski, N., Amir, S. & Woodside, B. (1999) Changes in NADPH-d staining inthe paraventricular and supraoptic nuclei during pregnancy and lactation inrats: role of ovarian steroids and oxytocin. J. Neuroendocrinol., 11, 53–61.

Prast, H. & Philippu, A. (2001) Nitric oxide as modulator of neuronal function.Prog. Neurobiol., 64, 51–68.

Robello, M., Amico, C., Bucossi, G., Cupello, A., Rapallino, M.V. & Thellung,S. (1996) Nitric oxide and GABAA receptor function in the rat cerebralcortex and cerebellar granule cells. Neuroscience, 74, 99–105.

Salm, A.K. (2000) Mechanisms of glial retraction in the hypothalamo-neuro-hypophysial system of the rat. Exp. Physiol., 85, 197S–202S.

Sanchez, F., Alonso, J.R., Arevalo, R., Blanco, E., Aijon, J. & Vazquez, R.(1994) Coexistence of NADPH-diaphorase with vasopressin and oxytocin inthe hypothalamic magnocellular neurosecretory nuclei of the rat. Cell TissueRes., 276, 31–34.

Sawchenko, P.E. & Swanson, L.W. (1982) The organization of noradrenergicpathways from the brainstem to the paraventricular and supraoptic nuclei inthe rat. Brain Res. Rev., 4, 275–325.

Scherer-Singler, U., Vincent, S.R., Kimura, H. & McGeer, E.G. (1983)Demonstration of a unique population of neurons with NADPH-diaphorasehistochemistry. J. Neurosci. Methods, 9, 229–234.

Smolenski, A., Burckhardt, A.M., Eigenthaler, M., Butt, E., Gambaryan, S.,Lohmann, S. & Walter, U. (1998) Functional analysis of cGMP-dependentprotein kinases I and II as mediators of NO/cGMP effects. Naunyn-Schmie-deberg’s Arch. Pharmacol., 358, 134–139.

Srisawat, R., Ludwig, M., Bull., M.B., Douglas, A.J., Russell, J.A. & Leng, G.(2000) Nitric oxide and the oxytocin system in pregnancy. J. Neurosci., 20(17), 6721–6727.

Stern, J.E. & Ludwig, M. (2001) NO inhibits supraoptic oxytocin and vaso-pressin neurons via activation of GABAergic synaptic inputs. Am. J. Physiol.Regulatory Integrative Comp. Physiol., 280, R1815–R1822.

Theodosis, D.T., El Majdoubi, M., Pierre, K. & Poulain, D.A. (1998) Factorsgoverning activity-dependent structural plasticity of the hypothalamoneur-ohyphysial system. Cell. Mol. Neurobiol., 18, 285–298.

Theodosis, D.T. & Poulain, D.A. (1989) Neuronal-glial and synaptic plasticityin the adult rat paraventricular nucleus. Brain Res., 484, 361–366.

Theodosis, D.T. & Poulain, D.A. (1999) Contribution of astrocytes to activity-dependent structural plasticity in the adult brain. Adv. Exp. Med. Biol., 468,175–182.

Tillet, Y., Ravault, J.P., Selve, C., Evin, G., Castro, B. & Dubois, M.P. (1986)Conditions for the use of specific antibodies for immunohistochemicalvisualization of serotonin and melatonin in the pineal gland of sheep. C.R. Acad. Sci. III, 303, 77–82.

Torres, G., Lee, S. & Rivier, C. (1993) Ontogeny of the rat hypothalamic nitricoxide synthase and colocalization with neuropeptides. Mol. Cell. Neurosci.,4, 155–163.

Trabace, L. & Kendrick, K.M. (2000) Nitric oxide can differentially modulatestriatal neurotransmitter concentrations via soluble guanylate cyclase andperoxinitrite formation. J. Neurochem., 75, 1664–1674.

Vacher, C.M., Fretier, P., Creminon, C., Calas, A. & Hardin-Pouzet, H. (2002)Activation by serotonin and noradrenaline of vasopressin and oxytocinexpression in the mouse paraventricular and supraoptic nuclei. J. Neurosci.,22, 1513–1522.



Villar, M.J., Ceccatelli, S., Ronnqvist, M. & Hokfelt, T. (1994) Nitric oxidesynthase increases in hypothalamic magnocellular neurons after salt loadingin the rat. An immunhistochemical and in situ hybridization study. BrainRes., 644, 273–281.

Wang, X. & Robinson, P.J. (1997) Cyclic GMP-dependent protein kinase andcellular signalling in the nervous system. J. Neurochem., 68, 443–456.

Willmott, N.J., Wong, K. & Strong, A.J. (2000) A fundamental role for the nitricoxide-G-kinase signaling pathway mediating intercellular Ca2þ waves inglia. J. Neurosci., 20, 1767–1779.

Willoughby, J.O., Jervois, P.M., Menadue, M.F. & Blessing, W.W. (1987)Noradrenaline, by activation of a-1-adrenoreceptors in the region of the

supraoptic nucleus, causes secretion of vasopressin in the anaesthetized rat.Neuroendocrinology, 45, 219–226.

Xu, D.L., Martin, P.Y., St John, J., Tsai, P., Summer, S.N., Ohara, M., Kim, J.K.& Schrier, R.W. (1996) Upregulation of endothelial and neuronal constitutivenitric oxide synthase in pregnant rats. Am. J. Physiol., 271, R1739–R1745.

Yang, Q.Z. & Hatton, G.I. (1999) Nitric oxide via cGMP-dependent mechan-isms increases dye coupling and excitability of rat supraoptic nucleusneurons. J. Neurosci., 19, 4270–4279.

Zarri, I., Bucossi, G., Cupello, A., Rapallino, M.V. & Robello, M. (1994)Modulation by nitric oxide of rat GABAA receptors. Neurosci. Lett., 180,239–242.



the effects of nitric oxide on magnocellular neurons could involve multiple indirect cyclic...

Documents