- Email: [email protected]
Ectopic expression of non-catecholaminergic tyrosine hydroxylase in rat hypothalamic magnocellular neurons
Tyrosine hydroxylase in vasopressin neurons
Pergamon PII: S0306-4522(99)00252-3
Neuroscience Vol. 94, No. 1, pp. 151–161, 1999 151 Copyright q 1999 IBRO. Published by Elsevier Science Ltd Printed in Great Britain. All rights reserved 0306-4522/99 $20.00+0.00
ECTOPIC EXPRESSION OF NON-CATECHOLAMINERGIC TYROSINE HYDROXYLASE IN RAT HYPOTHALAMIC MAGNOCELLULAR NEURONS F. MARSAIS* and A. CALAS Laboratoire de Cytologie, Institut des Neurosciences, Universite´ Pierre et Marie Curie, CNRS UMR 7624, 7 quai St Bernard, 75240 Paris cedex 05, France
Abstract—Hypothalamic magnocellular neurons constitute a good model of neurochemical plasticity, because a single neuron can express various combinations of neuropeptides and enzymes under different physiological conditions. Tyrosine hydroxylase has been shown to occur ectopically in various non-catecholaminergic neurons. We investigated the expression of tyrosine hydroxylase and its possible role in the magnocellular neurons of the supraoptic and paraventricular nuclei in salt-loaded and lactating rats, using in situ hybridization and immunohistochemistry, alone or combined, in light and electron microscopy. Our results demonstrated that almost 25% of the magnocellular neurons in the supraoptic nucleus and 15% in the paraventricular nucleus expressed tyrosine hydroxylase in salt-loaded rats, and 10% in the supraoptic nucleus of two-day lactating rats. Double labelling showed that this tyrosine hydroxylase was essentially synthesized in magnocellular neurons expressing vasopressin. The ultrastructural localization of tyrosine hydroxylase was less homogeneous in the cytoplasm of magnocellular neurons than in periventricular neurons. In lactating and salt-loaded rats, magnocellular neurons were devoid of the catecholamine biosynthesis markers aromatic l-amino acid decarboxylase, l-3,4 dihydroxyphenylalanine, dopamine and GTP-cyclohydrolase I. Tyrosine hydroxylase expression did not increase after rats were injected with reserpine. Our results indicate that the phenotype of the magnocellular neurons expressing tyrosine hydroxylase in lactating and salt-loaded rats is non-catecholaminergic, and suggest that this tyrosine hydroxylase might be involved in osmoregulation. q 1999 IBRO. Published by Elsevier Science Ltd. Key words: tyrosine hydroxylase, salt loading, lactation, oxytocin, vasopressin, hypothalamus.
The magnocellular neurons (MCNs) of the hypothalamus constitute a well-known model of neuronal plasticity. 5 Although two distinct populations of MCNs which synthesize either vasopressin or oxytocin have been described, 30 many peptides and enzymes have been shown to co-exist in and be synthesized by the same neurons. 3 In response to physiological and experimental stimulations, tyrosine hydroxylase (TH), the rate-limiting enzyme in catecholamine synthesis, 2,28 was detected in MCNs by different authors. 17,23 This ectopic expression of TH in MCNs has already been described in different species. In human brain, abundant TH-immunoreactive cells were detected in the hypothalamic magnocellular part of the paraventricular nucleus (PVN) and supraoptic nucleus (SON). 21 In rabbit brain, MCNs of the PVN also exhibited immunoreactivity to TH. 32 In genetically modified rats, an increased TH expression has been described in homozygous Brattleboro (diabetes insipidus) rats 17 and in obese Zucker rats (fa/fa). 9 In normal rat brain, however, very few TH-immunoreactive MCNs were detected under normal conditions. 6 After salt loading, large amounts of immunoreactive TH and TH mRNA were found. 17,23,47 In the context of the chemical plasticity of neurons, we were interested in the expression of TH in MCNs of the
SON and PVN under two different physiological states during which MCNs are activated: salt loading and lactation. We first explored the cellular and subcellular localization of the ectopic TH expressed in MCNs. We therefore began by defining the peptidergic type of MCN that expressed TH preferentially using double in situ hybridization and immunohistochemistry. We further attempted to establish whether or not this ectopic TH is involved in catecholamine synthesis. We thus looked for a possible catecholaminergic phenotype, using immunohistochemistry to identify specific markers of catecholamine synthesis: TH, which catalyses the hydroxylation of tyrosine to l-3,4 dihydroxyphenylalanine (l-DOPA) and requires tetrahydrobiopterin (BH4) as a co-factor; GTP-cyclohydroxylase I (GCH), which synthesizes BH4 from GTP; and aromatic l-amino acid decarboxylase (AADC), the second enzyme involved in catecholamine synthesis, which converts l-DOPA into dopamine. EXPERIMENTAL PROCEDURES
Animals All animal experiments were performed in accordance with French Legal Requirements (Decree 87-848) and with the European Community Council Directive of 24 November 1986 (86/609/EEC). For salt-loading experiments, adult male Wistar rats (from the authors’ breeding colony; 300–400 g) were given composite food (Charles River) and 2% NaCl to drink (for salt loading) for one week before they were killed. All other rats were given food and water ad libitum. Adult lactating female rats (300–400 g) were killed on different days after parturition, the latter being day 1 of lactation. Animals were housed under a constant light–darkness cycle (lights on between 06.00 and 18.00) in a temperature- and humidity-controlled environment. Four male Wistar rats (300–400 g) were housed and injected either with 10 mg/kg reserpine (Sigma) or with vehicle, as described previously. 15,38 After two days of this treatment, TH
*To whom correspondence should be addressed. Tel.: 1 33 1.44.27.33.95; fax: 1 33 1.44.27.25.08. E-mail address: [email protected] (F. Marsais) Abbreviations: AADC, aromatic l-amino acid decarboxylase; ABC, avidin–biotin–peroxidase complex; BH4, tetrahydrobiopterin; DAB, 3,3 0 -diaminobenzidine; DTT, dithiothreitol; EDTA, ethylenediaminetetra-acetate; GCH, GTP-cyclohydrolase I; l-DOPA, l-3,4 dihydroxyphenylalanine; MCN, magnocellular neuron; NGS, normal goat serum; PBS, phosphate-buffered saline; PBSM, phosphate-buffered saline containing metabisulphite; PVN, paraventricular nucleus; SON, supraoptic nucleus; SSC, standard saline citrate; TH, tyrosine hydroxylase. 151
152
F. Marsais and A. Calas
immunohistochemistry was investigated as described below in order to check whether the expression of TH is inducible in MCNs after reserpine injection, as has been described in the locus coeruleus. 8 Probes TH-64 is a 25-mer oligonucleotide probe complementary to exon 2 of the TH mRNA sequence. 20 Vasopressin-41 40 is a 41-mer oligonucleotide probe complementary to the pro-vasopressin mRNA sequence coding for the glycopeptide region of the rat vasopressin gene, i.e. amino acid residues 115–128, as this sequence has no counterpart in oxytocin mRNA. Oxytocin-25 40,41 is a 25-mer oligonucleotide probe complementary to the oxytocin mRNA sequence coding for amino acids 97–105, as this sequence is very different from its counterpart in vasopressin mRNA. Our oligomeric probes, which were designed for oxytocin, vasopressin and TH mRNAs, had already been used and tested for their specificity. 20,40,41 Two picomoles of oligonucleotide were incubated for 45 min at 378C with 20 mCi of [ 35S]dATP (Amersham; 10 mCi/ml) and 25 units of terminal transferase (Boehringer Mannheim, Germany) in 10 ml of 100 mM potassium cacodylate, 2 mM cobalt chloride and 0.2 mM dithiothreitol (DTT). The reaction was stopped in 0.2 M EDTA–5 mg/ml yeast tRNA (Sigma) and the radiolabelled probe was precipitated with 4 M lithium chloride and cold ethanol. The specific activity of the probe was about 1–4 × 10 8 c.p.m./g. In situ hybridization with tyrosine hydroxylase-64 Rats were deeply anaesthetized with 0.1 ml/100 g body weight sodium pentobarbital, and perfused through the ascending aorta with 100 ml saline and 300 ml ice-cold 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). Brains were removed, postfixed for 4 h in the same fixative and immersed overnight at 48C in sterile phosphatebuffered saline (PBS; pH 7.4) containing 15% sucrose. Brains were then frozen in isopentane, cooled to 2608C and stored at 2808C. Frontal sections 20 mm thick were cut on a cryostat (Reichert–Jung) and collected on CML Superfrost Plus slides. Sections were rinsed several times in sterile PBS. The hybridization step was performed by overnight incubation of the slides at 428C in 4 × standard saline citrate (SSC; 1 × SSC 0.15 M NaCl and 0.0015 M sodium citrate), 1 × Denhardt (0.02% Ficoll, 0.02% polyvinyl pyrrolidone and 0.02% bovine serum albumin), 10% dextran sulphate, 100 mM DTT, 250 mg/ml tRNA, 250 mg/ml polyA, 250 mg/ml sonicated fish sperm DNA and 50% deionized formamide. This hybridization buffer contained 1 nM of the radiolabelled probe. Finally, slides were washed twice for 15 min at 558C in 1 × SSC/10 mM DTT and at room temperature in 0.5 × SSC/10 mM DTT for 15 min. Sections were air dried and covered with Amersham-bmax autoradiographic film for 14 days at 48C in the dark. Films were developed with Microdol X (Kodak, Rochester, NY, U.S.A.) for 10 min at 188C, rinsed in water and fixed with Kodak Max-Fix fixative for 10 min. Slides were dipped in Ilford K5 nuclear emulsion (Mobberly, Cheshire, U.K.) diluted 1:1 in water, exposed for one month at 48C in darkness, developed with Kodak D19 revelator for 4 min at 178C and fixed with 30% sodium thiosulphate solution for 10 min. Sections were then stained with 0.1% Toluidine Blue solution and mounted in Permount. Immunohistochemical processing for tyrosine hydroxylase, aromatic l-amino acid decarboxylase and GTP-cyclohydrolase I Rat brains were perfused as described above and dissected. Sections were cut on a Vibratome (Leica; 15–50 mm thick) and on a cryostat (Leica; 20 mm thick). Free-floating sections were immersed for 30 min in 0.01 M PBS and preincubated in PBS/50 mM NH4Cl and PBS/0.2% gelatin for 30 min. They were then incubated overnight at 48C with the primary antibodies for the detection of TH (monoclonal mouse antiTH, diluted 1:2000; Sigma), AADC (polyclonal antibody, diluted 1:1000; Institut J. Boy) or GCH (polyclonal antibody, diluted 1:8000; kindly donated by Dr Nagatsu) in 2% normal goat serum (NGS; Gibco BRL) containing 0.1 M PBS. Sections were rinsed for 30 min in PBS and the primary antibodies were then detected using the avidin–biotin–peroxidase complex (ABC) system (Biosys). In brief, sections were incubated in PBS, first with a biotinylated anti-mouse or anti-rabbit antibody diluted 1:250 for 1 h at room temperature and then with ABC system (ABC Vectastain kit, Biosys; 1:100). 14 After several washes in PBS, the peroxidase activity was developed by incubating the sections in 50 mM Tris–HCl buffer containing 0.05%
3,3 0 -diaminobenzidine tetrahydrochloride (DAB; Sigma) and 0.006% H2O2 (Sigma). Sections were mounted in Permount on gelatin-coated slides for examination by light microscopy. The TH and AADC antibodies used were characterized by the suppliers (Sigma, Institut J. Boy), and the GCH antibody by Nagatsu et al. 26 Immunohistochemistry dopamine
for
l-3,4
dihydroxyphenylalanine
and
Brains were perfused with 5% glutaraldehyde, and 15-mm-thick sections were cut on a Vibratome, washed in 0.1 M PBS containing 1% sodium metabisulphite (PBSM), incubated for three days at 48C with antisera against either l-DOPA (Chemicon, U.S.A.; 1:500) or dopamine (kindly donated by Dr Tillet; diluted 1:4000) diluted in PBSM containing 2% NGS and 1% Triton X-100. Sections were rinsed in PBSM and revealed with ABC peroxidase–antiperoxidase, according to Davidoff and Schulze. 7 Sections were incubated for 30 min at room temperature, first in biotinylated anti-rabbit antibody, as described above, then in peroxidase–antiperoxidase complex diluted 1:100 (Dako), and finally in ABC. DAB was used as the chromogen for l-DOPA and dopamine detection, as described above. The dopamine antibody was characterized by Tillet et al. 39 The l-DOPA antibody was characterized by the supplier (Chemicon). Relative quantification of magnocellular neurons expressing tyrosine hydroxylase in the supraoptic and paraventricular nuclei The purpose of our study was to evaluate the percentage of MCNs stained in relation to the total number of cells in the SON and PVN of salt-loaded, lactating and control rats. We were able to assess the number of cells containing TH from microscope examination (Zeiss) of every other section of the SON and PVN of three control rats versus three salt-loaded or lactating rats. Twenty sections in the medial and caudal areas of the nuclei were analysed per animal, and all stained profiles that had a distinguishable nucleus were counted. As a control, the total number of MCN profiles in the SON was counted in every other section of a sample of Toluidine Blue-counterstained sections from control and salt-loaded or lactating rats. Double labelling: in situ hybridization for oxytocin, vasopressin and tyrosine hydroxylase messenger RNAs, and immunocytochemistry for tyrosine hydroxylase For direct comparison of labelling in the same section, double labelling with in situ hybridization/immunohistochemistry was used. Floating sections of brains fixed with 4% paraformaldehyde were washed in 0.1 M PBS and incubated in sterile dishes for 1 h in 4 × SSC/ 1 × Denhardt/0.04% diethyl pyrocarbonate at 428C. Sections were then hybridized in 1 nM of the [ 35S]oxytocin-25, [ 35S]vasopressin-41 or [ 35S]TH-64 probes diluted and incubated overnight at 428C in a solution of 50% deionized formamide, 10 mM DTT, 0.04% diethyl pyrocarbonate, 600 mM NaCl, 80 mM Tris–HCl (pH 7.5), 4 mM EDTA, 0.1% sodium pyrophosphate and 0.2% N-lauryl sarcosyl. Hybridization was followed by several washes at 428C in SSC (2 × SSC for 2 × 30 min, 1 × SSC for 1 × 30 min and 0.5 × SSC for 1 × 30 min), and finally by two 30-min washes at room temperature in 0.1 × SSC with the oxytocin-25 and vasopressin-41 probes. With the TH-64 probe, sections were washed several times with SSC at 558C (1 × SSC and 10 mM DTT for 2 × 15 min, 0.5 × SSC and 10 mM DTT for 2 × 15 min) and at room temperature (0.5 × SSC and 10 mM DTT for 1 × 15 min). Sections were then processed for immunocytochemistry with TH antibodies as described above. After visualization of the DAB reaction, sections were mounted on slides, dipped into K5 Ilford nuclear emulsion diluted in water (v/v), exposed for five days at 48C and developed with Kodak D19. All sections were examined under a Leitz photomicroscope. Electron microscopy Brains were perfused with PBSM containing 5% glutaraldehyde, dissected and postfixed for 30 min in the same fixative. Frontal sections 15 mm thick were cut with a Vibratome (Leica) and collected in culture dishes. The floating sections were incubated overnight at 48C with TH antibodies diluted 1:1000 in PBSM containing 2% NGS and 0.1% Triton X-100. The primary antibody was detected by the ABC system and peroxidase activity was developed with DAB, as described above. After light microscope examination, sections of interest for
Tyrosine hydroxylase in vasopressin neurons
153
Fig. 1. Distribution of TH immunoreactivity on glutaraldehyde-fixed sections in the SON (A–C) and PVN (D, E) of salt-loaded rats (B, E), lactating rats (C) and controls (A, D). A larger number of MCNs in the PVN (E) and SON (B, C) expressed TH in salt-loaded and lactating rats than in the controls. Arrows show that some MCNs are more intensely labelled than others. Scale bars 50 mm.
electron microscopy were selected. They were osmicated for 10 min in 4% OsO4 solution and dehydrated in graded alcohols (70%, 95% and 100%). Sections were incubated for 1 h at 378C in Araldite–ethanol solution, overnight at 378C in Araldite solution and for 4 h in Araldite solution containing Araldite M accelerator 960 (Fluka) at the same temperature. Sections were then mounted on silicon-coated slides and incubated overnight at 608C. Ultrathin sections were obtained with an Ultracut (Leica) and observed with or without lead citrate counterstaining in a Jeol electron microscope.
RESULTS
Tyrosine hydroxylase expression in the magnocellular neurons of salt-loaded and lactating rats In female and male control rats, one or two perikarya were weakly stained in the dorsal part of the SON along the varicosities concentrated in the ventral part of the nucleus (Fig. 1A). No TH-labelled magnocellular neurons were
observed in the PVN, although areas densely packed with fibres occurred in the whole nucleus (Fig. 1D). Periventricular neurons were smaller and more deeply stained by TH antibodies than MCNs (Fig. 1D). In salt-loaded rats, the ventromedial part of the SON and the medial part of the PVN were characterized by large numbers of intensely labelled MCNs. The amount of immunoreactive TH varied from one neuron to another, and highly immunoreactive cells were mixed with cells displaying weaker immunoreactivity (Fig. 1B, E). Almost 25% of the MCNs in the SON and 15% in the PVN expressed TH. Their perikarya were stained more intensely than control perikarya, but were less intensely labelled than the small periventricular neurons. The number of immunoreactive fibres containing TH also increased in the medial and ventral parts of the SON after salt loading. In lactating rats, the increase in TH expression in the MCNs
154
F. Marsais and A. Calas
Fig. 2. Distribution of TH mRNA on paraformaldehyde-fixed sections in the SON (A, C) and PVN (C) of salt-loaded rats (B, C) and lactating rats (A). The density of silver grains is higher in the SON and PVN of salt-loaded than lactating rats. Combined in situ hybridization and TH immunohistochemistry (A) shows co-localization of TH and TH mRNA in lactating rats. Scale bars 50 mm.
of the SON and PVN was smaller than in salt-loaded rats, as reflected by both the number of cells stained and the intensity of staining (Fig. 1C). At day 2 of lactation, the proportion of immunoreactive MCNs was about 10%. Whatever the conditions, the brown precipitate of immunoreactive neurons containing TH was mainly located in the perikarya in saltloaded and lactating rats, but nuclear labelling was also observed. In control rats, autoradiographic staining of cells expressing TH mRNA was observed in the periventricular area, but the MCNs of the PVN, SON and the circularis nucleus were completely devoid of silver grains, except for a very few neurons in the SON and PVN. In salt-loaded and lactating rats, the MCNs of the PVN, SON and circularis nucleus, as well as the neurons of the periventricular nucleus, were labelled with the radioactive TH-64 probe. In the emulsiondipped sections, the silver grains overlaid the MCNs, but their density on each cell varied, as observed for the immunostaining (Fig. 2A–C). Two types of DAB labelling were distinguished in the areas examined by electron microscopy after treatment with antibodies against TH. In the periventricular area, intense, diffuse and uniform labelling was observed in small neurons (7 mm × 15 mm) and dendrites. The nuclei were not labelled (except for discrete staining of small and scattered precipitates), and neither were the mitochondria, Golgi apparatus or
inner cisternae of the endoplasmic reticulum (Fig. 3E). Intense and diffuse labelling was also observed occasionally in nerve terminals, in the SON and more frequently in the PVN, which displayed unlabelled clear vesicles and synapses. In the SON and PVN, a different type of labelling was observed, involving large (20 mm × 30 mm) neurosecretory perikarya (Fig. 3A) and dendrites (Fig. 3B, F). However, this labelling was more discrete and consisted mainly of clumps of dense precipitate which occurred in the cytoplasm, sometimes in apposition to membranes (plasma membrane, membrane of the endoplasmic reticulum, mitochondria and granules). Some large neurosecretory swellings, resembling Herring bodies, displayed this “clumped” labelling, which never affected the accumulated granules (Fig. 3D). MCNs with a hypertrophied reticulum were also observed to display the same clumped reaction (Fig. 3C). In addition, a few labelled and unlabelled perikarya displayed discrete precipitates in the nucleus (Fig. 3F). Phenotype of tyrosine hydroxylase-positive magnocellular neurons To establish the peptidergic phenotype of the TH-positive MCNs, we performed double labelling. In salt-loaded and lactating rats, the oxytocin-25 and vasopressin-41 probes showed a deeper staining of MCNs than of control rat MCNs.
Tyrosine hydroxylase in vasopressin neurons
155
Fig. 3. TH immunoreactivity on glutaraldehyde-fixed sections of MCNs (A) of the SON, in salt-loaded rats, at the ultrastructural level. Together with an MCN perikaryon, immunoreactive terminals (A, arrow) and dendrites (B, C, F) are visible. The arrow in B shows a non-immunoreactive synaptic terminal upon a reactive dendrite (B). Some MCNs displaying a hypertrophied endoplasmic reticulum exhibit clumped labelling (C, one star) while others are negative (two stars). Swellings are immunoreactive in the SON, with labelling in clumps, similar to that found in MCNs (D). Immunoreactive staining is denser and more homogeneous in periventricular neurons (E) than in MCNs. A few nuclear labellings of immunonegative MCNs are observed (F). Scale bars 1 mm.
In salt-loaded rats, all of the hypothalamic MCNs labelled with TH antibodies also displayed silver grains after labelling with the vasopressin-41 probe (Fig. 4A, C). Single-labelled MCNs displayed only silver grains. However, the oxytocin-25 probe revealed almost no co-localization of silver grain labelling and immunoreactive TH in MCNs which exhibited either a brown precipitate or silver grains (Fig. 4B, D).
The vasopressin-41 probe showed that, in lactating rats, immunopositive TH was co-located with silver grains in the MCNs of the SON (Fig. 5B), but with the oxytocin-25 probe no such colocalization was seen (Fig. 5A). Double labelling with vasopressin-41 probe occurred in 100% of the MCNs displaying immunoreactive TH, in both lactating and salt-loaded rats. In both cases, the periventricular neurons only stained for TH.
156
F. Marsais and A. Calas
Fig. 4. Detection of TH and oxytocin or vasopressin mRNA in the PVN (A, B) and SON (C, D) of salt-loaded rats, by in situ hybridization combined with immunohistochemistry. In the PVN, vasopressin mRNA and TH co-exist in the same MCNs (A) and in the SON (C), although some neurons contain only vasopressin mRNA (arrow). Note that oxytocin mRNA and TH occur in different MCNs. Scale bars 100 mm.
Tyrosine hydroxylase in vasopressin neurons
157
Fig. 5. Detection of TH and oxytocin or vasopressin mRNA in the SON of lactating rats, by in situ hybridization combined with immunohistochemistry. Vasopressin mRNA and TH co-exist in the same MCNs (B), but oxytocin mRNA and TH do not (A). Scale bars 50 mm.
Catecholamine markers in the investigation of magnocellular neurons In reserpine-treated rats, no MCNs containing TH were observed, either in the PVN or in the SON. Staining was similar in reserpine- and vehicle-treated animals, although at day 2 the locus coeruleus exhibited an increased TH immunoreactivity. In salt-loaded and lactating rats, immunohistochemistry showed clear staining for GCH in both periventricular and parvocellular neurons in the PVN (Fig. 6B), but not in the SON (Fig. 6A). Immunoreactive perikarya stained strongly for AADC were observed in the suprachiasmatic nucleus and periventricular area (Fig. 6C). Neurons stained strongly for l-DOPA and dopamine occurred in the arcuate nucleus and periventricular area (Fig. 6D).
DISCUSSION
Our findings are mainly of interest in the context of chemical plasticity of neurons for the insight that they provide into (i) the cellular and subcellular phenotype of MCNs expressing TH in response to salt loading and lactation, and (ii) the nature and (iii) possible role of this TH. Neuronal plasticity in salt loading and lactation Administration of hyperosmotic stimuli is known to induce changes in MCNs, including increased cell size, 16 reduced glial cell contacts, 12 and increased mRNA production for a large pool of neuropeptides and enzymes, including vasopressin, 4,35 oxytocin, 43 TH, 47 galanin, 22 dynorphin, cholecystokinin 34 and corticotropin-releasing factor. 46 After salt loading, vasopressin and oxytocin concentrations were found to
158
F. Marsais and A. Calas
Fig. 6. Detection of immunoreactive GCH (A, B), AADC (C) and l-DOPA (D) in the SON or PVN of salt-loaded and lactating rats. MCNs are devoid of immunostaining with any of the antibodies tested, but periventricular and parvocellular neurons are stained for GCH and l-DOPA (B, D), and suprachiasmatic nuclei for AADC (C). Scale bars 300 mm.
decrease in the pituitary neural lobe and to increase in plasma. In addition, corticotropin-releasing factor, peptide histidine isoleucine and TH increased in MCNs, whereas vasopressin, oxytocin, galanin and dynorphin immunoreactivity 23 decreased, probably owing to an increase in their axonal transport. Microtubules have been shown to increase in neurohypophysis after osmotic stimulation. 11 Consequently, after such a stimulation there is a general increase in synthetic activity in MCNs. Furthermore, overexpression of c-fos and c-jun has been demonstrated in MCNs after osmotic stimuli or lactation, and is indicative of intense neuronal activity. 36,44 We chose to compare TH expression in lactating rats after two days of lactation
because, at this time, the most apparent changes in neuropeptide expression in MCNs were observed, with 17% of the SON neurons producing both oxytocin and vasopressin. 25 Mezey and Kiss suggested that, under these conditions, MCNs that synthesize oxytocin also produce the antidiuretic hormone vasopressin in order to help the animal to compensate for the loss of water associated with lactation. Cellular and subcellular phenotype of tyrosine hydroxylaseexpressing magnocellular neurons The observed TH immunostaining is similar to the pattern reported previously in rat brain. 6,13,37,42 Previous investigators
Tyrosine hydroxylase in vasopressin neurons
observed a few TH-immunoreactive MCNs in the SON of normal rats, but none in the PVN. 6 The present results confirm that, in salt-loaded rats, the number of these TH-immunoreactive MCNs increases. 17,23,24 There is good agreement between the data we obtained by immunohistochemistry and in situ hybridization. Furthermore, in situ hybridization using five different oligonucleotides dispersed throughout TH mRNA revealed the same staining (data not shown), strongly suggesting that the TH detected and synthesized in MCNs is tightly TH. To define the MCNs that synthesize TH, we further characterized the peptidergic phenotype of TH-expressing neurons in the SON and PVN. In both salt-loaded and lactating rats, TH was localized exclusively in vasopressinergic neurons. We found no TH in MCNs synthesizing oxytocin mRNA, either in salt-loaded rats, in agreement with Meister et al., 23 or in lactating rats. TH was also expressed in the PVN of control rabbits, 32 and in the SON and PVN of humans 21 and Zucker rats. 9 In contrast to our results, TH is expressed exclusively in oxytocin-containing MCNs of the PVN of control rabbits. 32 In adult human post mortem brains, Panayotacopoulou et al. 29 reported preferential expression of TH in the perikarya of oxytocin-containing MCNs, whereas only a few neurons containing immunoreactive TH appeared to contain vasopressin. These authors also demonstrated that the opposite occurred in human neonates, whose TH was preferentially localized in the vasopressin-containing cells of the SON and PVN. The same authors suggested that the enhancement of TH immunoreactivity in vasopressinexpressing neonate neurons implied the presence of perinatal hypoxia. In obese Zucker rats, TH was also expressed mainly in MCNs containing vasopressin, and sometimes in oxytocincontaining cells, as described recently by Fetissov et al. 9 Therefore, the peptidergic phenotype of the MCNs expressing TH varies, depending on the species and experimental conditions, although several results suggest that preferential expression of TH depends on osmoregulation, including data derived from studies of Brattleboro rats 17,25 and from our study of obese Zucker rats. 9 In the latter study, we suggested that the hyperosmolarity of these Zucker rats might be a cause of such expression, because it was stronger in MCNs containing vasopressin than in those containing oxytocin. In these models, the restoration of a normal hydric balance has not been performed to demonstrate that TH ectopic expression disappeared with normal hydric parameters. However, Yagita et al. 45 demonstrated that the expression of TH in MCNs decreased after rehydration of salt-loaded rats. Furthermore, vasopressin neurons which do not belong to the hypothalamo-neurohypophyseal system and whose role is independent of the osmoregulation do not express TH. For example, the parvocellular neurons of the PVN, which synthesize vasopressin and belong to other central projection systems, and vasopressin neurons of the suprachiasmatic nucleus do not express TH. The enhanced production of TH and its mRNA in salt-loaded and lactating rat MCNs probably arose from the same stimuli, and both could be related to osmoregulation. In electron microscopy, the labelled MCNs with a hypertrophied endoplasmic reticulum are probably hyperactivated vasopressinergic neurons, but all activated neurons do not appear to display TH. To ascertain the subcellular localization of TH, we compared, using electron microscopy, TH immunostaining in MCNs and in catecholaminergic neurons of the periventricular area, known to
159
be dopaminergic. 1 The difference between TH expression observed in these regions by light microscopy was confirmed at the ultrastructural level. The ultrastructural localization of TH was found to be cytoplasmic in both MCNs and catecholaminergic neurons, but its distribution was clearly different. The relatively small amount of TH in MCNs, in comparison with catecholaminergic neurons, probably leads to this peculiar distribution of immunoreactivity following fixation. Using light and electron microscopy, we showed staining in the nuclei of some immunopositive neurons expressing TH. However, as the same nuclear labelling is also observed in some of the immunonegative neurons, we may assume that this labelling was probably due to the high concentration of glutaraldehyde used to perfuse brains for immunohistochemistry, since the fixative is generally believed to cause strong non-specific staining. 19 Role of tyrosine hydroxylase in magnocellular neurons Our data indicate that TH expression in vasopressincontaining MCNs was quantitatively and perhaps qualitatively different from that observed in catecholaminergic neurons. MCNs did not, in fact, synthesize catecholamines, as indicated by the absence of dopamine. This could be due to a TH devoid of enzyme activity. Such TH was described previously by Schussler et al. 33 in the anterior pituitary, where its presence was probably due to an alternative splicing and/or putative additional post-translational modifications. In our lactating and salt-loaded rats, the apparent absence of TH activity may rather be due to the lack of constituents of the enzymatic oxidative reaction than to an inactive form of the enzyme itself. The absence of GCH in MCNs under both conditions suggests that the lack of enzyme activity could be related to the absence of the cofactor BH4. Such TH expression in the absence of GCH was demonstrated by Nagatsu et al. during postnatal murine development in the anterior olfactory nucleus, 27 striatum, 18 cerebellum 10 and spinal trigeminal nucleus. 31 The expression of TH in the MCNs of the SON and PVN of salt-loaded and lactating rats could be a resurgence of the transient expression of TH that occurs in postnatal life. However, this hypothesis does not exclude the possibility that this ectopic TH has specific functions and is subject to regulation, as suggested by its ultrastructural localization and its lack of response to reserpine treatment. CONCLUSIONS
Our data demonstrate the neuronal plasticity of MCNs with respect to the ectopic expression of TH in the SON and PVN. This TH corresponds to a non-catecholaminergic phenotype and is essentially expressed in vasopressin-containing neurons, suggesting a possible link to osmoregulation. The MCNs display an interesting novel form of TH expression which might be specifically regulated and be implicated in functions other than catecholamine synthesis.
Acknowledgements—We are grateful to R. Picart for valuable advice and to J. Taxi for help with electron microscopy. We thank I. Nagatsu for kindly donating anti-GCH antibody and Y. Tillet for anti-dopamine antibody. This work was supported by a grant from the Fondation Singer-Polignac (F.M.).
160
F. Marsais and A. Calas 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. 34. 35. 36. 37. 38. 39.
Bjo¨rklund A. and Lindvall O. (1984) Dopamine-containing systems in the CNS. In Handbook of Chemical Neuroanatomy (eds Bjo¨rklund A. and Ho¨kfelt T.), Vol. 2, pp. 55–122. Elsevier, Amsterdam. Blaschko H. (1957) Formation of catecholamines in the animal body. Br. med. Bull. 13, 162–165. Brownstein M. J. and Mezey E. (1986) Multiple chemical messengers in hypothalamic magnocellular neurons. Prog. Brain Res. 68, 161–168. Burbach J. P. H., De Hoop M. J., Schmale H., Richter D., De Kloet E. R., Ten Haaf J. A. and De Wied D. (1984) Differential responses to osmotic stress of vasopressin-neurophysin mRNA in hypothalamic nuclei. Neuroendocrinology 39, 582–584. Calas A. (1994) La versatilite´ neuronale. C. r. hebd Se´anc. Acad. Sci., Paris (Se´rie Ge´ne´rale) 11, 271–285. Chan Palay V., Zaborsky L., Kohler C., Golgstein M. and Palay S. L. (1984) Distribution of tyrosine hydroxylase immunoreactive neurons in the hypothalamus of rats. J. comp. Neurol. 227, 467–496. Davidoff M. and Schulze W. (1990) Combination of the peroxidase–antiperoxidase (PAP) and avidin–biotin–peroxidase complex (ABC) techniques: an amplification alternative in immunocytochemical staining. Histochemistry 93, 351–356. Faucon-Biguet N., Buda M., Lamouroux A., Samolyk D. and Mallet J. (1986) Time course of TH mRNA in rat brain and adrenal medulla after a single injection of reserpine. Eur. molec. Biol. Org. J. 5, 287–291. Fetissov S., Marsais F., Nicolaı¨dis S. and Calas A. (1997) Expression of tyrosine hydroxylase in magnocellular hypothalamic neurons of obese (fa/fa) and lean heterozygous (Fa/fa) Zucker rats. Molec. Brain Res. 50, 314–318. Fujii T., Sakai M. and Nagatsu I. (1994) Immunohistochemical demonstration of expression of tyrosine hydroxylase in cerebellar Purkinje cells of the human and mouse. Neurosci. Lett. 165, 161–163. Grainger F. and Sloper J. C. (1974) Overactivity of the hypothalamo-neurohypophysial neurosecretory system and the problem of the mechanism of transporting neurosecretory material. In Neurosecretion—The Final Neuroendocrine Pathway (eds Knowles F. and Vollrath K.), p. 307. Springer, Berlin. Hatton G. I., Perlmutter L. S., Salm A. K. and Tweedle C. D. (1984) Dynamic neuronal–glial interactions in hypothalamus and pituitary: implications for control of hormone synthesis and release. Peptides 5, 121–138. Ho¨kfelt T., Johansson O. and Goldstein M. (1984) Central catecholamine neurons as revealed by immunohistochemistry with special reference to adrenaline neurons. In Handbook of Chemical Neuroanatomy (eds Bjo¨rklund A. and Ho¨kfelt T.), Vol. 2, pp. 157–276. Elsevier, Amsterdam. Hsu S.-M., Raine L. and Fager H. (1981) Use of avidin–biotin–peroxidase complex (ABC) in immunoperoxidase techniques: a comparison between ABC and unlabeled antibody (PAP) procedure. J. Histochem. Cytochem. 29, 577–580. Joh T. H., Gehgman C. and Reis D. (1973) Immunohistochemical demonstration of increased accumulation of tyrosine hydroxylase protein in sympathetic ganglia and adrenal medulla elicited by reserpine. Proc. natn. Acad. Sci. U.S.A. 70, 2767–2771. Kalimo H. (1975) Ultrastructural studies in hypothalamic neurons of rat. III. Paraventricular and supraoptic neurons during lactation and dehydration. Cell Tiss. Res. 163, 151–168. Kiss J. Z. and Mezey E. (1986) Tyrosine hydroxylase in magnocellular neurosecretory neurons: response to physiological manipulation. Neuroendocrinology 43, 519–525. Komori K., Sakai M., Karasawa N., Yamada K. and Nagatsu I. (1991) Evidence for transient expression of tyrosine hydroxylase immunoreactivity in the mouse striatum and the effect of colchicine. Acta histochem. cytochem. 24, 223–231. Kosaka T., Nagatsu I., Wu J. Y. and Hama K. (1986) Use of high concentrations of glutaraldehyde for immunocytochemistry of transmitter-synthesizing enzymes in the central nervous system. Neuroscience 18, 975–990. Lanie`ce P., Le Hir H., Bodeau-Pe´an S., Charon Y., Valentin L., Thermes C., Mallet J. and Dumas S. (1996) A novel rat tyrosine hydroxylase mRNA species generated by alternative splicing. J. Neurochem. 66, 1819–1825. Li Y. W., Halliday G. M., Joh T. H., Geffen L. B. and Blessing W. W. (1988) Tyrosine hydroxylase-containing neurons in the supraoptic and paraventricular nuclei of the adult human. Brain Res. 461, 75–86. Meister B., Cortes R., Villar M. J. and Ho¨kfelt T. (1989) Increase of galanin mRNA and decrease of galanin immunoreactivity in magnocellular neurons of the hypothalamic supraoptic and paraventricular nuclei after salt loading. Eur. J. Neurosci., Suppl. 2, 126. Meister B., Corte´s R., Villar M. J., Schalling M. and Ho¨kfelt T. (1990) Peptides and transmitter enzymes in hypothalamic magnocellular neurons after administration of hyperosmotic stimuli: comparison between messenger RNA and peptide/protein levels. Cell Tiss. Res. 260, 279–297. Meister B., Villar M. J., Ceccatelli S. and Ho¨kfelt T. (1990) Localization of chemical messengers in magnocellular neurons of the hypothalamic supraoptic and paraventricular nuclei: an immunohistochemical study using experimental manipulation. Neuroscience 37, 603–633. Mezey E. and Kiss J. Z. (1991) Coexpression of vasopressin and oxytocin in hypothalamic supraoptic neurons of lactating rats. Endocrinology 129, 1814–1820. Nagatsu I., Ichinose H., Sakai M., Titani K., Suzuki M. and Nagatsu T. (1995) Immunocytochemical localization of GTP cyclohydroxylase I in the brain, adrenal, and liver of mice. J. neural Transm. 102, 175–188. Nagatsu I., Komori K., Takeuchi T., Sakai M., Yamada K. and Karasawa N. (1990) Transient tyrosine hydroxylase-immunoreactive neurons in the region of the anterior olfactory nucleus of pre- and postnatal mice do not contain dopamine. Brain Res. 511, 55–62. Nagatsu I., Levitt M. and Udenfriend S. (1964) Tyrosine hydroxylase: the initial step in norepinephrine biosynthesis. J. biol. Chem. 239, 2910–2917. Panayotacopoulou M. T., Raadsheer F. C. and Swaab D. (1994) Colocalization of tyrosine hydroxylase with oxytocin or vasopressin in the human paraventricular and supraoptic nucleus. Devl Brain Res. 83, 59–66. Rhodes C. H., Morrell J. I. and Pfaff D. W. (1981) Immunohistochemical analysis of magnocellular elements in rat hypothalamus: distribution and numbers of cells containing neurophysin, oxytocin and vasopressin. J. comp. Neurol. 198, 45–64. Sakai M., Fujii T., Yamawaki Y. and Nagatsu I. (1996) Transient tyrosine hydroxylase immunoreactive neurons in the developmental mouse brain. Acta anat. nippon. 71, 373. Schimchowitsch S., Stoeckel M. E., Vigny A. and Porte A. (1981) Oxytocinergic neurons with tyrosine hydroxylase-like immunoreactivity in the paraventricular nucleus of the rabbit hypothalamus. Neurosci. Lett. 43, 55–59. Schussler N., Boularand S., Peillon F., Mallet J. and Faucon-Biguet N. (1995) Multiple tyrosine hydroxylase transcripts and immunoreactive forms in the rat: differential expression in the anterior pituitary and adrenal gland. J. Neurosci. 42, 846–854. Sherman T. G., Day R., Civelli O., Douglass J., Herbert E., Akil H. and Watson S. J. (1988) Regulation of hypothalamic magnocellular neuropeptides and their mRNAs in the Brattleboro rat: coordinated responses to further osmotic challenge. J. Neurosci. 8, 3785–3796. Sherman T. G., McKelvy J. F. and Watson S. J. (1986) Vasopressin mRNA regulation in individual hypothalamic nuclei: a northern and in situ hybridization study. J. Neurosci. 6, 1685–1694. Summy-Long J. Y., Gestl S., Terrell M. L., Wolz G. and Kadekaro M. (1997) Osmoregulation of the magnocellular neuroendocrine system during lactation. Am. J. Physiol. 41, 275–288. Swanson L. W., Sawchenko P. E., Berod A., Hartman B. K., Helle K. B. and Vanorden D. E. (1981) An immunohistochemical organization of catecholaminergic cells and terminal fields in the paraventricular and supraoptic nuclei of the hypothalamus. J. comp. Neurol. 196, 271–285. Thoenen H., Mueller R. A. and Axelrod J. (1969) Increase in tyrosine hydroxylase activity after reserpine administration. J. Pharmac. exp. Ther. 193, 775–784. Tillet Y., Batailler M., Krieger-Poullet M. and Thibault J. (1990) Presence of dopamine-immunoreactive cell bodies in the catecholaminergic group A15 of the sheep brain. Histochemistry 93, 327–333.
Tyrosine hydroxylase in vasopressin neurons
161
40. Trembleau A., Calas A. and Fe`vre-Montange M. (1988) Combination of immunocytochemistry and in situ hybridization with synthetic oligonucleotide probes to localize simultaneously vasopressin, oxytocin and their mRNAs in hypothalamic magnocellular neurons. Bull. Assoc. Anat. 72, 101–106. 41. Trembleau A., Calas A. and Fe`vre-Montange M. (1990) Ultrastructural localization of oxytocin mRNA in the rat hypothalamus by in situ hybridization using a synthetic oligonucleotide. Molec. Brain Res. 8, 37–45. 42. Van den Pol A. N., Herbst R. S. and Powell J. F. (1984) Tyrosine hydroxylase-immunoreactive neurons of the hypothalamus: a light and electron microscopic study. Neuroscience 19, 1117–1156. 43. Van Tol H. H. M., Voorhus D. Th. A. M. and Burbach J. P. H. (1987) Oxytocin gene expression in discrete hypothalamic cell groups is stimulated by prolonged salt loading. Endocrinology 120, 71–76. 44. Wang K., Guldenaar E. F. S. and McCabe J. T. (1997) Fos and Jun expression in rat supraoptic nucleus neurons after acute vs. repeated osmotic stimulation. Brain Res. 746, 117–125. 45. Yagita K., Okamura H. and Ibata Y. (1994) Rehydration process from salt loading: recovery of vasopressin and its coexisting galanin, dynorphin and tyrosine hydroxylase immunoreactivities in the supraoptic and paraventricular nuclei. Brain Res. 667, 13–23. 46. Young W. S. (1986) Corticotropin-releasing factor mRNA in the hypothalamus is affected differently by drinking saline and by dehydration. Fedn Eur. biochem. Socs Lett. 208, 158–162. 47. Young W. S., Warden M. and Mezey E. (1987) Tyrosine hydroxylase mRNA is increased by hyperosmotic stimuli in the paraventricular and supraoptic nuclei. Neuroendocrinology 46, 439–444. (Accepted 19 April 1999)
Pergamon PII: S0306-4522(99)00252-3
Neuroscience Vol. 94, No. 1, pp. 151–161, 1999 151 Copyright q 1999 IBRO. Published by Elsevier Science Ltd Printed in Great Britain. All rights reserved 0306-4522/99 $20.00+0.00
ECTOPIC EXPRESSION OF NON-CATECHOLAMINERGIC TYROSINE HYDROXYLASE IN RAT HYPOTHALAMIC MAGNOCELLULAR NEURONS F. MARSAIS* and A. CALAS Laboratoire de Cytologie, Institut des Neurosciences, Universite´ Pierre et Marie Curie, CNRS UMR 7624, 7 quai St Bernard, 75240 Paris cedex 05, France
Abstract—Hypothalamic magnocellular neurons constitute a good model of neurochemical plasticity, because a single neuron can express various combinations of neuropeptides and enzymes under different physiological conditions. Tyrosine hydroxylase has been shown to occur ectopically in various non-catecholaminergic neurons. We investigated the expression of tyrosine hydroxylase and its possible role in the magnocellular neurons of the supraoptic and paraventricular nuclei in salt-loaded and lactating rats, using in situ hybridization and immunohistochemistry, alone or combined, in light and electron microscopy. Our results demonstrated that almost 25% of the magnocellular neurons in the supraoptic nucleus and 15% in the paraventricular nucleus expressed tyrosine hydroxylase in salt-loaded rats, and 10% in the supraoptic nucleus of two-day lactating rats. Double labelling showed that this tyrosine hydroxylase was essentially synthesized in magnocellular neurons expressing vasopressin. The ultrastructural localization of tyrosine hydroxylase was less homogeneous in the cytoplasm of magnocellular neurons than in periventricular neurons. In lactating and salt-loaded rats, magnocellular neurons were devoid of the catecholamine biosynthesis markers aromatic l-amino acid decarboxylase, l-3,4 dihydroxyphenylalanine, dopamine and GTP-cyclohydrolase I. Tyrosine hydroxylase expression did not increase after rats were injected with reserpine. Our results indicate that the phenotype of the magnocellular neurons expressing tyrosine hydroxylase in lactating and salt-loaded rats is non-catecholaminergic, and suggest that this tyrosine hydroxylase might be involved in osmoregulation. q 1999 IBRO. Published by Elsevier Science Ltd. Key words: tyrosine hydroxylase, salt loading, lactation, oxytocin, vasopressin, hypothalamus.
The magnocellular neurons (MCNs) of the hypothalamus constitute a well-known model of neuronal plasticity. 5 Although two distinct populations of MCNs which synthesize either vasopressin or oxytocin have been described, 30 many peptides and enzymes have been shown to co-exist in and be synthesized by the same neurons. 3 In response to physiological and experimental stimulations, tyrosine hydroxylase (TH), the rate-limiting enzyme in catecholamine synthesis, 2,28 was detected in MCNs by different authors. 17,23 This ectopic expression of TH in MCNs has already been described in different species. In human brain, abundant TH-immunoreactive cells were detected in the hypothalamic magnocellular part of the paraventricular nucleus (PVN) and supraoptic nucleus (SON). 21 In rabbit brain, MCNs of the PVN also exhibited immunoreactivity to TH. 32 In genetically modified rats, an increased TH expression has been described in homozygous Brattleboro (diabetes insipidus) rats 17 and in obese Zucker rats (fa/fa). 9 In normal rat brain, however, very few TH-immunoreactive MCNs were detected under normal conditions. 6 After salt loading, large amounts of immunoreactive TH and TH mRNA were found. 17,23,47 In the context of the chemical plasticity of neurons, we were interested in the expression of TH in MCNs of the
SON and PVN under two different physiological states during which MCNs are activated: salt loading and lactation. We first explored the cellular and subcellular localization of the ectopic TH expressed in MCNs. We therefore began by defining the peptidergic type of MCN that expressed TH preferentially using double in situ hybridization and immunohistochemistry. We further attempted to establish whether or not this ectopic TH is involved in catecholamine synthesis. We thus looked for a possible catecholaminergic phenotype, using immunohistochemistry to identify specific markers of catecholamine synthesis: TH, which catalyses the hydroxylation of tyrosine to l-3,4 dihydroxyphenylalanine (l-DOPA) and requires tetrahydrobiopterin (BH4) as a co-factor; GTP-cyclohydroxylase I (GCH), which synthesizes BH4 from GTP; and aromatic l-amino acid decarboxylase (AADC), the second enzyme involved in catecholamine synthesis, which converts l-DOPA into dopamine. EXPERIMENTAL PROCEDURES
Animals All animal experiments were performed in accordance with French Legal Requirements (Decree 87-848) and with the European Community Council Directive of 24 November 1986 (86/609/EEC). For salt-loading experiments, adult male Wistar rats (from the authors’ breeding colony; 300–400 g) were given composite food (Charles River) and 2% NaCl to drink (for salt loading) for one week before they were killed. All other rats were given food and water ad libitum. Adult lactating female rats (300–400 g) were killed on different days after parturition, the latter being day 1 of lactation. Animals were housed under a constant light–darkness cycle (lights on between 06.00 and 18.00) in a temperature- and humidity-controlled environment. Four male Wistar rats (300–400 g) were housed and injected either with 10 mg/kg reserpine (Sigma) or with vehicle, as described previously. 15,38 After two days of this treatment, TH
*To whom correspondence should be addressed. Tel.: 1 33 1.44.27.33.95; fax: 1 33 1.44.27.25.08. E-mail address: [email protected] (F. Marsais) Abbreviations: AADC, aromatic l-amino acid decarboxylase; ABC, avidin–biotin–peroxidase complex; BH4, tetrahydrobiopterin; DAB, 3,3 0 -diaminobenzidine; DTT, dithiothreitol; EDTA, ethylenediaminetetra-acetate; GCH, GTP-cyclohydrolase I; l-DOPA, l-3,4 dihydroxyphenylalanine; MCN, magnocellular neuron; NGS, normal goat serum; PBS, phosphate-buffered saline; PBSM, phosphate-buffered saline containing metabisulphite; PVN, paraventricular nucleus; SON, supraoptic nucleus; SSC, standard saline citrate; TH, tyrosine hydroxylase. 151
152
F. Marsais and A. Calas
immunohistochemistry was investigated as described below in order to check whether the expression of TH is inducible in MCNs after reserpine injection, as has been described in the locus coeruleus. 8 Probes TH-64 is a 25-mer oligonucleotide probe complementary to exon 2 of the TH mRNA sequence. 20 Vasopressin-41 40 is a 41-mer oligonucleotide probe complementary to the pro-vasopressin mRNA sequence coding for the glycopeptide region of the rat vasopressin gene, i.e. amino acid residues 115–128, as this sequence has no counterpart in oxytocin mRNA. Oxytocin-25 40,41 is a 25-mer oligonucleotide probe complementary to the oxytocin mRNA sequence coding for amino acids 97–105, as this sequence is very different from its counterpart in vasopressin mRNA. Our oligomeric probes, which were designed for oxytocin, vasopressin and TH mRNAs, had already been used and tested for their specificity. 20,40,41 Two picomoles of oligonucleotide were incubated for 45 min at 378C with 20 mCi of [ 35S]dATP (Amersham; 10 mCi/ml) and 25 units of terminal transferase (Boehringer Mannheim, Germany) in 10 ml of 100 mM potassium cacodylate, 2 mM cobalt chloride and 0.2 mM dithiothreitol (DTT). The reaction was stopped in 0.2 M EDTA–5 mg/ml yeast tRNA (Sigma) and the radiolabelled probe was precipitated with 4 M lithium chloride and cold ethanol. The specific activity of the probe was about 1–4 × 10 8 c.p.m./g. In situ hybridization with tyrosine hydroxylase-64 Rats were deeply anaesthetized with 0.1 ml/100 g body weight sodium pentobarbital, and perfused through the ascending aorta with 100 ml saline and 300 ml ice-cold 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). Brains were removed, postfixed for 4 h in the same fixative and immersed overnight at 48C in sterile phosphatebuffered saline (PBS; pH 7.4) containing 15% sucrose. Brains were then frozen in isopentane, cooled to 2608C and stored at 2808C. Frontal sections 20 mm thick were cut on a cryostat (Reichert–Jung) and collected on CML Superfrost Plus slides. Sections were rinsed several times in sterile PBS. The hybridization step was performed by overnight incubation of the slides at 428C in 4 × standard saline citrate (SSC; 1 × SSC 0.15 M NaCl and 0.0015 M sodium citrate), 1 × Denhardt (0.02% Ficoll, 0.02% polyvinyl pyrrolidone and 0.02% bovine serum albumin), 10% dextran sulphate, 100 mM DTT, 250 mg/ml tRNA, 250 mg/ml polyA, 250 mg/ml sonicated fish sperm DNA and 50% deionized formamide. This hybridization buffer contained 1 nM of the radiolabelled probe. Finally, slides were washed twice for 15 min at 558C in 1 × SSC/10 mM DTT and at room temperature in 0.5 × SSC/10 mM DTT for 15 min. Sections were air dried and covered with Amersham-bmax autoradiographic film for 14 days at 48C in the dark. Films were developed with Microdol X (Kodak, Rochester, NY, U.S.A.) for 10 min at 188C, rinsed in water and fixed with Kodak Max-Fix fixative for 10 min. Slides were dipped in Ilford K5 nuclear emulsion (Mobberly, Cheshire, U.K.) diluted 1:1 in water, exposed for one month at 48C in darkness, developed with Kodak D19 revelator for 4 min at 178C and fixed with 30% sodium thiosulphate solution for 10 min. Sections were then stained with 0.1% Toluidine Blue solution and mounted in Permount. Immunohistochemical processing for tyrosine hydroxylase, aromatic l-amino acid decarboxylase and GTP-cyclohydrolase I Rat brains were perfused as described above and dissected. Sections were cut on a Vibratome (Leica; 15–50 mm thick) and on a cryostat (Leica; 20 mm thick). Free-floating sections were immersed for 30 min in 0.01 M PBS and preincubated in PBS/50 mM NH4Cl and PBS/0.2% gelatin for 30 min. They were then incubated overnight at 48C with the primary antibodies for the detection of TH (monoclonal mouse antiTH, diluted 1:2000; Sigma), AADC (polyclonal antibody, diluted 1:1000; Institut J. Boy) or GCH (polyclonal antibody, diluted 1:8000; kindly donated by Dr Nagatsu) in 2% normal goat serum (NGS; Gibco BRL) containing 0.1 M PBS. Sections were rinsed for 30 min in PBS and the primary antibodies were then detected using the avidin–biotin–peroxidase complex (ABC) system (Biosys). In brief, sections were incubated in PBS, first with a biotinylated anti-mouse or anti-rabbit antibody diluted 1:250 for 1 h at room temperature and then with ABC system (ABC Vectastain kit, Biosys; 1:100). 14 After several washes in PBS, the peroxidase activity was developed by incubating the sections in 50 mM Tris–HCl buffer containing 0.05%
3,3 0 -diaminobenzidine tetrahydrochloride (DAB; Sigma) and 0.006% H2O2 (Sigma). Sections were mounted in Permount on gelatin-coated slides for examination by light microscopy. The TH and AADC antibodies used were characterized by the suppliers (Sigma, Institut J. Boy), and the GCH antibody by Nagatsu et al. 26 Immunohistochemistry dopamine
for
l-3,4
dihydroxyphenylalanine
and
Brains were perfused with 5% glutaraldehyde, and 15-mm-thick sections were cut on a Vibratome, washed in 0.1 M PBS containing 1% sodium metabisulphite (PBSM), incubated for three days at 48C with antisera against either l-DOPA (Chemicon, U.S.A.; 1:500) or dopamine (kindly donated by Dr Tillet; diluted 1:4000) diluted in PBSM containing 2% NGS and 1% Triton X-100. Sections were rinsed in PBSM and revealed with ABC peroxidase–antiperoxidase, according to Davidoff and Schulze. 7 Sections were incubated for 30 min at room temperature, first in biotinylated anti-rabbit antibody, as described above, then in peroxidase–antiperoxidase complex diluted 1:100 (Dako), and finally in ABC. DAB was used as the chromogen for l-DOPA and dopamine detection, as described above. The dopamine antibody was characterized by Tillet et al. 39 The l-DOPA antibody was characterized by the supplier (Chemicon). Relative quantification of magnocellular neurons expressing tyrosine hydroxylase in the supraoptic and paraventricular nuclei The purpose of our study was to evaluate the percentage of MCNs stained in relation to the total number of cells in the SON and PVN of salt-loaded, lactating and control rats. We were able to assess the number of cells containing TH from microscope examination (Zeiss) of every other section of the SON and PVN of three control rats versus three salt-loaded or lactating rats. Twenty sections in the medial and caudal areas of the nuclei were analysed per animal, and all stained profiles that had a distinguishable nucleus were counted. As a control, the total number of MCN profiles in the SON was counted in every other section of a sample of Toluidine Blue-counterstained sections from control and salt-loaded or lactating rats. Double labelling: in situ hybridization for oxytocin, vasopressin and tyrosine hydroxylase messenger RNAs, and immunocytochemistry for tyrosine hydroxylase For direct comparison of labelling in the same section, double labelling with in situ hybridization/immunohistochemistry was used. Floating sections of brains fixed with 4% paraformaldehyde were washed in 0.1 M PBS and incubated in sterile dishes for 1 h in 4 × SSC/ 1 × Denhardt/0.04% diethyl pyrocarbonate at 428C. Sections were then hybridized in 1 nM of the [ 35S]oxytocin-25, [ 35S]vasopressin-41 or [ 35S]TH-64 probes diluted and incubated overnight at 428C in a solution of 50% deionized formamide, 10 mM DTT, 0.04% diethyl pyrocarbonate, 600 mM NaCl, 80 mM Tris–HCl (pH 7.5), 4 mM EDTA, 0.1% sodium pyrophosphate and 0.2% N-lauryl sarcosyl. Hybridization was followed by several washes at 428C in SSC (2 × SSC for 2 × 30 min, 1 × SSC for 1 × 30 min and 0.5 × SSC for 1 × 30 min), and finally by two 30-min washes at room temperature in 0.1 × SSC with the oxytocin-25 and vasopressin-41 probes. With the TH-64 probe, sections were washed several times with SSC at 558C (1 × SSC and 10 mM DTT for 2 × 15 min, 0.5 × SSC and 10 mM DTT for 2 × 15 min) and at room temperature (0.5 × SSC and 10 mM DTT for 1 × 15 min). Sections were then processed for immunocytochemistry with TH antibodies as described above. After visualization of the DAB reaction, sections were mounted on slides, dipped into K5 Ilford nuclear emulsion diluted in water (v/v), exposed for five days at 48C and developed with Kodak D19. All sections were examined under a Leitz photomicroscope. Electron microscopy Brains were perfused with PBSM containing 5% glutaraldehyde, dissected and postfixed for 30 min in the same fixative. Frontal sections 15 mm thick were cut with a Vibratome (Leica) and collected in culture dishes. The floating sections were incubated overnight at 48C with TH antibodies diluted 1:1000 in PBSM containing 2% NGS and 0.1% Triton X-100. The primary antibody was detected by the ABC system and peroxidase activity was developed with DAB, as described above. After light microscope examination, sections of interest for
Tyrosine hydroxylase in vasopressin neurons
153
Fig. 1. Distribution of TH immunoreactivity on glutaraldehyde-fixed sections in the SON (A–C) and PVN (D, E) of salt-loaded rats (B, E), lactating rats (C) and controls (A, D). A larger number of MCNs in the PVN (E) and SON (B, C) expressed TH in salt-loaded and lactating rats than in the controls. Arrows show that some MCNs are more intensely labelled than others. Scale bars 50 mm.
electron microscopy were selected. They were osmicated for 10 min in 4% OsO4 solution and dehydrated in graded alcohols (70%, 95% and 100%). Sections were incubated for 1 h at 378C in Araldite–ethanol solution, overnight at 378C in Araldite solution and for 4 h in Araldite solution containing Araldite M accelerator 960 (Fluka) at the same temperature. Sections were then mounted on silicon-coated slides and incubated overnight at 608C. Ultrathin sections were obtained with an Ultracut (Leica) and observed with or without lead citrate counterstaining in a Jeol electron microscope.
RESULTS
Tyrosine hydroxylase expression in the magnocellular neurons of salt-loaded and lactating rats In female and male control rats, one or two perikarya were weakly stained in the dorsal part of the SON along the varicosities concentrated in the ventral part of the nucleus (Fig. 1A). No TH-labelled magnocellular neurons were
observed in the PVN, although areas densely packed with fibres occurred in the whole nucleus (Fig. 1D). Periventricular neurons were smaller and more deeply stained by TH antibodies than MCNs (Fig. 1D). In salt-loaded rats, the ventromedial part of the SON and the medial part of the PVN were characterized by large numbers of intensely labelled MCNs. The amount of immunoreactive TH varied from one neuron to another, and highly immunoreactive cells were mixed with cells displaying weaker immunoreactivity (Fig. 1B, E). Almost 25% of the MCNs in the SON and 15% in the PVN expressed TH. Their perikarya were stained more intensely than control perikarya, but were less intensely labelled than the small periventricular neurons. The number of immunoreactive fibres containing TH also increased in the medial and ventral parts of the SON after salt loading. In lactating rats, the increase in TH expression in the MCNs
154
F. Marsais and A. Calas
Fig. 2. Distribution of TH mRNA on paraformaldehyde-fixed sections in the SON (A, C) and PVN (C) of salt-loaded rats (B, C) and lactating rats (A). The density of silver grains is higher in the SON and PVN of salt-loaded than lactating rats. Combined in situ hybridization and TH immunohistochemistry (A) shows co-localization of TH and TH mRNA in lactating rats. Scale bars 50 mm.
of the SON and PVN was smaller than in salt-loaded rats, as reflected by both the number of cells stained and the intensity of staining (Fig. 1C). At day 2 of lactation, the proportion of immunoreactive MCNs was about 10%. Whatever the conditions, the brown precipitate of immunoreactive neurons containing TH was mainly located in the perikarya in saltloaded and lactating rats, but nuclear labelling was also observed. In control rats, autoradiographic staining of cells expressing TH mRNA was observed in the periventricular area, but the MCNs of the PVN, SON and the circularis nucleus were completely devoid of silver grains, except for a very few neurons in the SON and PVN. In salt-loaded and lactating rats, the MCNs of the PVN, SON and circularis nucleus, as well as the neurons of the periventricular nucleus, were labelled with the radioactive TH-64 probe. In the emulsiondipped sections, the silver grains overlaid the MCNs, but their density on each cell varied, as observed for the immunostaining (Fig. 2A–C). Two types of DAB labelling were distinguished in the areas examined by electron microscopy after treatment with antibodies against TH. In the periventricular area, intense, diffuse and uniform labelling was observed in small neurons (7 mm × 15 mm) and dendrites. The nuclei were not labelled (except for discrete staining of small and scattered precipitates), and neither were the mitochondria, Golgi apparatus or
inner cisternae of the endoplasmic reticulum (Fig. 3E). Intense and diffuse labelling was also observed occasionally in nerve terminals, in the SON and more frequently in the PVN, which displayed unlabelled clear vesicles and synapses. In the SON and PVN, a different type of labelling was observed, involving large (20 mm × 30 mm) neurosecretory perikarya (Fig. 3A) and dendrites (Fig. 3B, F). However, this labelling was more discrete and consisted mainly of clumps of dense precipitate which occurred in the cytoplasm, sometimes in apposition to membranes (plasma membrane, membrane of the endoplasmic reticulum, mitochondria and granules). Some large neurosecretory swellings, resembling Herring bodies, displayed this “clumped” labelling, which never affected the accumulated granules (Fig. 3D). MCNs with a hypertrophied reticulum were also observed to display the same clumped reaction (Fig. 3C). In addition, a few labelled and unlabelled perikarya displayed discrete precipitates in the nucleus (Fig. 3F). Phenotype of tyrosine hydroxylase-positive magnocellular neurons To establish the peptidergic phenotype of the TH-positive MCNs, we performed double labelling. In salt-loaded and lactating rats, the oxytocin-25 and vasopressin-41 probes showed a deeper staining of MCNs than of control rat MCNs.
Tyrosine hydroxylase in vasopressin neurons
155
Fig. 3. TH immunoreactivity on glutaraldehyde-fixed sections of MCNs (A) of the SON, in salt-loaded rats, at the ultrastructural level. Together with an MCN perikaryon, immunoreactive terminals (A, arrow) and dendrites (B, C, F) are visible. The arrow in B shows a non-immunoreactive synaptic terminal upon a reactive dendrite (B). Some MCNs displaying a hypertrophied endoplasmic reticulum exhibit clumped labelling (C, one star) while others are negative (two stars). Swellings are immunoreactive in the SON, with labelling in clumps, similar to that found in MCNs (D). Immunoreactive staining is denser and more homogeneous in periventricular neurons (E) than in MCNs. A few nuclear labellings of immunonegative MCNs are observed (F). Scale bars 1 mm.
In salt-loaded rats, all of the hypothalamic MCNs labelled with TH antibodies also displayed silver grains after labelling with the vasopressin-41 probe (Fig. 4A, C). Single-labelled MCNs displayed only silver grains. However, the oxytocin-25 probe revealed almost no co-localization of silver grain labelling and immunoreactive TH in MCNs which exhibited either a brown precipitate or silver grains (Fig. 4B, D).
The vasopressin-41 probe showed that, in lactating rats, immunopositive TH was co-located with silver grains in the MCNs of the SON (Fig. 5B), but with the oxytocin-25 probe no such colocalization was seen (Fig. 5A). Double labelling with vasopressin-41 probe occurred in 100% of the MCNs displaying immunoreactive TH, in both lactating and salt-loaded rats. In both cases, the periventricular neurons only stained for TH.
156
F. Marsais and A. Calas
Fig. 4. Detection of TH and oxytocin or vasopressin mRNA in the PVN (A, B) and SON (C, D) of salt-loaded rats, by in situ hybridization combined with immunohistochemistry. In the PVN, vasopressin mRNA and TH co-exist in the same MCNs (A) and in the SON (C), although some neurons contain only vasopressin mRNA (arrow). Note that oxytocin mRNA and TH occur in different MCNs. Scale bars 100 mm.
Tyrosine hydroxylase in vasopressin neurons
157
Fig. 5. Detection of TH and oxytocin or vasopressin mRNA in the SON of lactating rats, by in situ hybridization combined with immunohistochemistry. Vasopressin mRNA and TH co-exist in the same MCNs (B), but oxytocin mRNA and TH do not (A). Scale bars 50 mm.
Catecholamine markers in the investigation of magnocellular neurons In reserpine-treated rats, no MCNs containing TH were observed, either in the PVN or in the SON. Staining was similar in reserpine- and vehicle-treated animals, although at day 2 the locus coeruleus exhibited an increased TH immunoreactivity. In salt-loaded and lactating rats, immunohistochemistry showed clear staining for GCH in both periventricular and parvocellular neurons in the PVN (Fig. 6B), but not in the SON (Fig. 6A). Immunoreactive perikarya stained strongly for AADC were observed in the suprachiasmatic nucleus and periventricular area (Fig. 6C). Neurons stained strongly for l-DOPA and dopamine occurred in the arcuate nucleus and periventricular area (Fig. 6D).
DISCUSSION
Our findings are mainly of interest in the context of chemical plasticity of neurons for the insight that they provide into (i) the cellular and subcellular phenotype of MCNs expressing TH in response to salt loading and lactation, and (ii) the nature and (iii) possible role of this TH. Neuronal plasticity in salt loading and lactation Administration of hyperosmotic stimuli is known to induce changes in MCNs, including increased cell size, 16 reduced glial cell contacts, 12 and increased mRNA production for a large pool of neuropeptides and enzymes, including vasopressin, 4,35 oxytocin, 43 TH, 47 galanin, 22 dynorphin, cholecystokinin 34 and corticotropin-releasing factor. 46 After salt loading, vasopressin and oxytocin concentrations were found to
158
F. Marsais and A. Calas
Fig. 6. Detection of immunoreactive GCH (A, B), AADC (C) and l-DOPA (D) in the SON or PVN of salt-loaded and lactating rats. MCNs are devoid of immunostaining with any of the antibodies tested, but periventricular and parvocellular neurons are stained for GCH and l-DOPA (B, D), and suprachiasmatic nuclei for AADC (C). Scale bars 300 mm.
decrease in the pituitary neural lobe and to increase in plasma. In addition, corticotropin-releasing factor, peptide histidine isoleucine and TH increased in MCNs, whereas vasopressin, oxytocin, galanin and dynorphin immunoreactivity 23 decreased, probably owing to an increase in their axonal transport. Microtubules have been shown to increase in neurohypophysis after osmotic stimulation. 11 Consequently, after such a stimulation there is a general increase in synthetic activity in MCNs. Furthermore, overexpression of c-fos and c-jun has been demonstrated in MCNs after osmotic stimuli or lactation, and is indicative of intense neuronal activity. 36,44 We chose to compare TH expression in lactating rats after two days of lactation
because, at this time, the most apparent changes in neuropeptide expression in MCNs were observed, with 17% of the SON neurons producing both oxytocin and vasopressin. 25 Mezey and Kiss suggested that, under these conditions, MCNs that synthesize oxytocin also produce the antidiuretic hormone vasopressin in order to help the animal to compensate for the loss of water associated with lactation. Cellular and subcellular phenotype of tyrosine hydroxylaseexpressing magnocellular neurons The observed TH immunostaining is similar to the pattern reported previously in rat brain. 6,13,37,42 Previous investigators
Tyrosine hydroxylase in vasopressin neurons
observed a few TH-immunoreactive MCNs in the SON of normal rats, but none in the PVN. 6 The present results confirm that, in salt-loaded rats, the number of these TH-immunoreactive MCNs increases. 17,23,24 There is good agreement between the data we obtained by immunohistochemistry and in situ hybridization. Furthermore, in situ hybridization using five different oligonucleotides dispersed throughout TH mRNA revealed the same staining (data not shown), strongly suggesting that the TH detected and synthesized in MCNs is tightly TH. To define the MCNs that synthesize TH, we further characterized the peptidergic phenotype of TH-expressing neurons in the SON and PVN. In both salt-loaded and lactating rats, TH was localized exclusively in vasopressinergic neurons. We found no TH in MCNs synthesizing oxytocin mRNA, either in salt-loaded rats, in agreement with Meister et al., 23 or in lactating rats. TH was also expressed in the PVN of control rabbits, 32 and in the SON and PVN of humans 21 and Zucker rats. 9 In contrast to our results, TH is expressed exclusively in oxytocin-containing MCNs of the PVN of control rabbits. 32 In adult human post mortem brains, Panayotacopoulou et al. 29 reported preferential expression of TH in the perikarya of oxytocin-containing MCNs, whereas only a few neurons containing immunoreactive TH appeared to contain vasopressin. These authors also demonstrated that the opposite occurred in human neonates, whose TH was preferentially localized in the vasopressin-containing cells of the SON and PVN. The same authors suggested that the enhancement of TH immunoreactivity in vasopressinexpressing neonate neurons implied the presence of perinatal hypoxia. In obese Zucker rats, TH was also expressed mainly in MCNs containing vasopressin, and sometimes in oxytocincontaining cells, as described recently by Fetissov et al. 9 Therefore, the peptidergic phenotype of the MCNs expressing TH varies, depending on the species and experimental conditions, although several results suggest that preferential expression of TH depends on osmoregulation, including data derived from studies of Brattleboro rats 17,25 and from our study of obese Zucker rats. 9 In the latter study, we suggested that the hyperosmolarity of these Zucker rats might be a cause of such expression, because it was stronger in MCNs containing vasopressin than in those containing oxytocin. In these models, the restoration of a normal hydric balance has not been performed to demonstrate that TH ectopic expression disappeared with normal hydric parameters. However, Yagita et al. 45 demonstrated that the expression of TH in MCNs decreased after rehydration of salt-loaded rats. Furthermore, vasopressin neurons which do not belong to the hypothalamo-neurohypophyseal system and whose role is independent of the osmoregulation do not express TH. For example, the parvocellular neurons of the PVN, which synthesize vasopressin and belong to other central projection systems, and vasopressin neurons of the suprachiasmatic nucleus do not express TH. The enhanced production of TH and its mRNA in salt-loaded and lactating rat MCNs probably arose from the same stimuli, and both could be related to osmoregulation. In electron microscopy, the labelled MCNs with a hypertrophied endoplasmic reticulum are probably hyperactivated vasopressinergic neurons, but all activated neurons do not appear to display TH. To ascertain the subcellular localization of TH, we compared, using electron microscopy, TH immunostaining in MCNs and in catecholaminergic neurons of the periventricular area, known to
159
be dopaminergic. 1 The difference between TH expression observed in these regions by light microscopy was confirmed at the ultrastructural level. The ultrastructural localization of TH was found to be cytoplasmic in both MCNs and catecholaminergic neurons, but its distribution was clearly different. The relatively small amount of TH in MCNs, in comparison with catecholaminergic neurons, probably leads to this peculiar distribution of immunoreactivity following fixation. Using light and electron microscopy, we showed staining in the nuclei of some immunopositive neurons expressing TH. However, as the same nuclear labelling is also observed in some of the immunonegative neurons, we may assume that this labelling was probably due to the high concentration of glutaraldehyde used to perfuse brains for immunohistochemistry, since the fixative is generally believed to cause strong non-specific staining. 19 Role of tyrosine hydroxylase in magnocellular neurons Our data indicate that TH expression in vasopressincontaining MCNs was quantitatively and perhaps qualitatively different from that observed in catecholaminergic neurons. MCNs did not, in fact, synthesize catecholamines, as indicated by the absence of dopamine. This could be due to a TH devoid of enzyme activity. Such TH was described previously by Schussler et al. 33 in the anterior pituitary, where its presence was probably due to an alternative splicing and/or putative additional post-translational modifications. In our lactating and salt-loaded rats, the apparent absence of TH activity may rather be due to the lack of constituents of the enzymatic oxidative reaction than to an inactive form of the enzyme itself. The absence of GCH in MCNs under both conditions suggests that the lack of enzyme activity could be related to the absence of the cofactor BH4. Such TH expression in the absence of GCH was demonstrated by Nagatsu et al. during postnatal murine development in the anterior olfactory nucleus, 27 striatum, 18 cerebellum 10 and spinal trigeminal nucleus. 31 The expression of TH in the MCNs of the SON and PVN of salt-loaded and lactating rats could be a resurgence of the transient expression of TH that occurs in postnatal life. However, this hypothesis does not exclude the possibility that this ectopic TH has specific functions and is subject to regulation, as suggested by its ultrastructural localization and its lack of response to reserpine treatment. CONCLUSIONS
Our data demonstrate the neuronal plasticity of MCNs with respect to the ectopic expression of TH in the SON and PVN. This TH corresponds to a non-catecholaminergic phenotype and is essentially expressed in vasopressin-containing neurons, suggesting a possible link to osmoregulation. The MCNs display an interesting novel form of TH expression which might be specifically regulated and be implicated in functions other than catecholamine synthesis.
Acknowledgements—We are grateful to R. Picart for valuable advice and to J. Taxi for help with electron microscopy. We thank I. Nagatsu for kindly donating anti-GCH antibody and Y. Tillet for anti-dopamine antibody. This work was supported by a grant from the Fondation Singer-Polignac (F.M.).
160
F. Marsais and A. Calas 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. 34. 35. 36. 37. 38. 39.
Bjo¨rklund A. and Lindvall O. (1984) Dopamine-containing systems in the CNS. In Handbook of Chemical Neuroanatomy (eds Bjo¨rklund A. and Ho¨kfelt T.), Vol. 2, pp. 55–122. Elsevier, Amsterdam. Blaschko H. (1957) Formation of catecholamines in the animal body. Br. med. Bull. 13, 162–165. Brownstein M. J. and Mezey E. (1986) Multiple chemical messengers in hypothalamic magnocellular neurons. Prog. Brain Res. 68, 161–168. Burbach J. P. H., De Hoop M. J., Schmale H., Richter D., De Kloet E. R., Ten Haaf J. A. and De Wied D. (1984) Differential responses to osmotic stress of vasopressin-neurophysin mRNA in hypothalamic nuclei. Neuroendocrinology 39, 582–584. Calas A. (1994) La versatilite´ neuronale. C. r. hebd Se´anc. Acad. Sci., Paris (Se´rie Ge´ne´rale) 11, 271–285. Chan Palay V., Zaborsky L., Kohler C., Golgstein M. and Palay S. L. (1984) Distribution of tyrosine hydroxylase immunoreactive neurons in the hypothalamus of rats. J. comp. Neurol. 227, 467–496. Davidoff M. and Schulze W. (1990) Combination of the peroxidase–antiperoxidase (PAP) and avidin–biotin–peroxidase complex (ABC) techniques: an amplification alternative in immunocytochemical staining. Histochemistry 93, 351–356. Faucon-Biguet N., Buda M., Lamouroux A., Samolyk D. and Mallet J. (1986) Time course of TH mRNA in rat brain and adrenal medulla after a single injection of reserpine. Eur. molec. Biol. Org. J. 5, 287–291. Fetissov S., Marsais F., Nicolaı¨dis S. and Calas A. (1997) Expression of tyrosine hydroxylase in magnocellular hypothalamic neurons of obese (fa/fa) and lean heterozygous (Fa/fa) Zucker rats. Molec. Brain Res. 50, 314–318. Fujii T., Sakai M. and Nagatsu I. (1994) Immunohistochemical demonstration of expression of tyrosine hydroxylase in cerebellar Purkinje cells of the human and mouse. Neurosci. Lett. 165, 161–163. Grainger F. and Sloper J. C. (1974) Overactivity of the hypothalamo-neurohypophysial neurosecretory system and the problem of the mechanism of transporting neurosecretory material. In Neurosecretion—The Final Neuroendocrine Pathway (eds Knowles F. and Vollrath K.), p. 307. Springer, Berlin. Hatton G. I., Perlmutter L. S., Salm A. K. and Tweedle C. D. (1984) Dynamic neuronal–glial interactions in hypothalamus and pituitary: implications for control of hormone synthesis and release. Peptides 5, 121–138. Ho¨kfelt T., Johansson O. and Goldstein M. (1984) Central catecholamine neurons as revealed by immunohistochemistry with special reference to adrenaline neurons. In Handbook of Chemical Neuroanatomy (eds Bjo¨rklund A. and Ho¨kfelt T.), Vol. 2, pp. 157–276. Elsevier, Amsterdam. Hsu S.-M., Raine L. and Fager H. (1981) Use of avidin–biotin–peroxidase complex (ABC) in immunoperoxidase techniques: a comparison between ABC and unlabeled antibody (PAP) procedure. J. Histochem. Cytochem. 29, 577–580. Joh T. H., Gehgman C. and Reis D. (1973) Immunohistochemical demonstration of increased accumulation of tyrosine hydroxylase protein in sympathetic ganglia and adrenal medulla elicited by reserpine. Proc. natn. Acad. Sci. U.S.A. 70, 2767–2771. Kalimo H. (1975) Ultrastructural studies in hypothalamic neurons of rat. III. Paraventricular and supraoptic neurons during lactation and dehydration. Cell Tiss. Res. 163, 151–168. Kiss J. Z. and Mezey E. (1986) Tyrosine hydroxylase in magnocellular neurosecretory neurons: response to physiological manipulation. Neuroendocrinology 43, 519–525. Komori K., Sakai M., Karasawa N., Yamada K. and Nagatsu I. (1991) Evidence for transient expression of tyrosine hydroxylase immunoreactivity in the mouse striatum and the effect of colchicine. Acta histochem. cytochem. 24, 223–231. Kosaka T., Nagatsu I., Wu J. Y. and Hama K. (1986) Use of high concentrations of glutaraldehyde for immunocytochemistry of transmitter-synthesizing enzymes in the central nervous system. Neuroscience 18, 975–990. Lanie`ce P., Le Hir H., Bodeau-Pe´an S., Charon Y., Valentin L., Thermes C., Mallet J. and Dumas S. (1996) A novel rat tyrosine hydroxylase mRNA species generated by alternative splicing. J. Neurochem. 66, 1819–1825. Li Y. W., Halliday G. M., Joh T. H., Geffen L. B. and Blessing W. W. (1988) Tyrosine hydroxylase-containing neurons in the supraoptic and paraventricular nuclei of the adult human. Brain Res. 461, 75–86. Meister B., Cortes R., Villar M. J. and Ho¨kfelt T. (1989) Increase of galanin mRNA and decrease of galanin immunoreactivity in magnocellular neurons of the hypothalamic supraoptic and paraventricular nuclei after salt loading. Eur. J. Neurosci., Suppl. 2, 126. Meister B., Corte´s R., Villar M. J., Schalling M. and Ho¨kfelt T. (1990) Peptides and transmitter enzymes in hypothalamic magnocellular neurons after administration of hyperosmotic stimuli: comparison between messenger RNA and peptide/protein levels. Cell Tiss. Res. 260, 279–297. Meister B., Villar M. J., Ceccatelli S. and Ho¨kfelt T. (1990) Localization of chemical messengers in magnocellular neurons of the hypothalamic supraoptic and paraventricular nuclei: an immunohistochemical study using experimental manipulation. Neuroscience 37, 603–633. Mezey E. and Kiss J. Z. (1991) Coexpression of vasopressin and oxytocin in hypothalamic supraoptic neurons of lactating rats. Endocrinology 129, 1814–1820. Nagatsu I., Ichinose H., Sakai M., Titani K., Suzuki M. and Nagatsu T. (1995) Immunocytochemical localization of GTP cyclohydroxylase I in the brain, adrenal, and liver of mice. J. neural Transm. 102, 175–188. Nagatsu I., Komori K., Takeuchi T., Sakai M., Yamada K. and Karasawa N. (1990) Transient tyrosine hydroxylase-immunoreactive neurons in the region of the anterior olfactory nucleus of pre- and postnatal mice do not contain dopamine. Brain Res. 511, 55–62. Nagatsu I., Levitt M. and Udenfriend S. (1964) Tyrosine hydroxylase: the initial step in norepinephrine biosynthesis. J. biol. Chem. 239, 2910–2917. Panayotacopoulou M. T., Raadsheer F. C. and Swaab D. (1994) Colocalization of tyrosine hydroxylase with oxytocin or vasopressin in the human paraventricular and supraoptic nucleus. Devl Brain Res. 83, 59–66. Rhodes C. H., Morrell J. I. and Pfaff D. W. (1981) Immunohistochemical analysis of magnocellular elements in rat hypothalamus: distribution and numbers of cells containing neurophysin, oxytocin and vasopressin. J. comp. Neurol. 198, 45–64. Sakai M., Fujii T., Yamawaki Y. and Nagatsu I. (1996) Transient tyrosine hydroxylase immunoreactive neurons in the developmental mouse brain. Acta anat. nippon. 71, 373. Schimchowitsch S., Stoeckel M. E., Vigny A. and Porte A. (1981) Oxytocinergic neurons with tyrosine hydroxylase-like immunoreactivity in the paraventricular nucleus of the rabbit hypothalamus. Neurosci. Lett. 43, 55–59. Schussler N., Boularand S., Peillon F., Mallet J. and Faucon-Biguet N. (1995) Multiple tyrosine hydroxylase transcripts and immunoreactive forms in the rat: differential expression in the anterior pituitary and adrenal gland. J. Neurosci. 42, 846–854. Sherman T. G., Day R., Civelli O., Douglass J., Herbert E., Akil H. and Watson S. J. (1988) Regulation of hypothalamic magnocellular neuropeptides and their mRNAs in the Brattleboro rat: coordinated responses to further osmotic challenge. J. Neurosci. 8, 3785–3796. Sherman T. G., McKelvy J. F. and Watson S. J. (1986) Vasopressin mRNA regulation in individual hypothalamic nuclei: a northern and in situ hybridization study. J. Neurosci. 6, 1685–1694. Summy-Long J. Y., Gestl S., Terrell M. L., Wolz G. and Kadekaro M. (1997) Osmoregulation of the magnocellular neuroendocrine system during lactation. Am. J. Physiol. 41, 275–288. Swanson L. W., Sawchenko P. E., Berod A., Hartman B. K., Helle K. B. and Vanorden D. E. (1981) An immunohistochemical organization of catecholaminergic cells and terminal fields in the paraventricular and supraoptic nuclei of the hypothalamus. J. comp. Neurol. 196, 271–285. Thoenen H., Mueller R. A. and Axelrod J. (1969) Increase in tyrosine hydroxylase activity after reserpine administration. J. Pharmac. exp. Ther. 193, 775–784. Tillet Y., Batailler M., Krieger-Poullet M. and Thibault J. (1990) Presence of dopamine-immunoreactive cell bodies in the catecholaminergic group A15 of the sheep brain. Histochemistry 93, 327–333.
Tyrosine hydroxylase in vasopressin neurons
161
40. Trembleau A., Calas A. and Fe`vre-Montange M. (1988) Combination of immunocytochemistry and in situ hybridization with synthetic oligonucleotide probes to localize simultaneously vasopressin, oxytocin and their mRNAs in hypothalamic magnocellular neurons. Bull. Assoc. Anat. 72, 101–106. 41. Trembleau A., Calas A. and Fe`vre-Montange M. (1990) Ultrastructural localization of oxytocin mRNA in the rat hypothalamus by in situ hybridization using a synthetic oligonucleotide. Molec. Brain Res. 8, 37–45. 42. Van den Pol A. N., Herbst R. S. and Powell J. F. (1984) Tyrosine hydroxylase-immunoreactive neurons of the hypothalamus: a light and electron microscopic study. Neuroscience 19, 1117–1156. 43. Van Tol H. H. M., Voorhus D. Th. A. M. and Burbach J. P. H. (1987) Oxytocin gene expression in discrete hypothalamic cell groups is stimulated by prolonged salt loading. Endocrinology 120, 71–76. 44. Wang K., Guldenaar E. F. S. and McCabe J. T. (1997) Fos and Jun expression in rat supraoptic nucleus neurons after acute vs. repeated osmotic stimulation. Brain Res. 746, 117–125. 45. Yagita K., Okamura H. and Ibata Y. (1994) Rehydration process from salt loading: recovery of vasopressin and its coexisting galanin, dynorphin and tyrosine hydroxylase immunoreactivities in the supraoptic and paraventricular nuclei. Brain Res. 667, 13–23. 46. Young W. S. (1986) Corticotropin-releasing factor mRNA in the hypothalamus is affected differently by drinking saline and by dehydration. Fedn Eur. biochem. Socs Lett. 208, 158–162. 47. Young W. S., Warden M. and Mezey E. (1987) Tyrosine hydroxylase mRNA is increased by hyperosmotic stimuli in the paraventricular and supraoptic nuclei. Neuroendocrinology 46, 439–444. (Accepted 19 April 1999)