Area postrema, a brain circumventricular organ, is the site of antidipsogenic action of circulating atrial natriuretic peptide in eels

Atrial natriuretic peptide (ANP) acts on various tissues to regulate body fluid balance and cardiovascular function in fish(Farrell and Olson, 2000; Loretz and Pollina, 2000; Takei and Hirose, 2002; Toop and Donald, 2004). In eels, ANP promotes adaptation to seawater (SW) by reducing excess Na⁺ uptake from the environment(Tsukada et al., 2005). The reduced Na⁺ uptake is accomplished by decreases in both drinking of environmental SW and subsequent absorption of Na⁺ ions by the intestine, while renal and branchial excretion of Na⁺ are unchanged by ANP infusion. In fact, removal of endogenous ANP and ventricular natriuretic peptide, which also binds to the type A ANP receptor (NPR-A) with similar affinity, from plasma by immunoneutralization increased plasma Na⁺ concentration with a concomitant increase in drinking rate in SW eels (Tsukada and Takei,2006). These results strongly support the hypo-osmoregulatory roles of ANP in SW. Despite such an important effect of ANP on drinking, the proximate mechanisms and the site of action of ANP in the brain have not been examined in detail.

The mechanisms regulating drinking have been intensively and extensively investigated in mammals, reflecting its important role in body fluid regulation (Fitzsimons, 1998; Balment, 2002). Accumulating evidence has shown that the central nervous system (CNS) controls drinking by integrating neuronal and chemical signals originating in the brain and from the periphery. Circulating hormones are major controllers of drinking through their actions on the brain; these factors include angiotensin II (ANG II),relaxin, bradykinin and ANP (Fitzsimons and Simons, 1968; Katsuura et al.,1986; Fregly and Rowland,1991; Sinnayah et al.,1999). However, since circulating hormones do not cross the blood–brain barrier (BBB), they must act on the target sites outside the BBB such as circumventricular organs (CVOs)(Ferguson and Bains, 1996; McKinley et al., 1999). In mammals, the CVOs such as the subfornical organ (SFO), area postrema (AP) and organum vasculosum of the lamina terminalis (OVLT) serve as a blood–brain interface for the action of circulating hormones(Ferguson and Bains, 1996). In particular, forebrain CVOs, including the SFO and OVLT, are known to be primary sites for the central regulation of body fluid balance in mammals(Johnson and Gross, 1993; McKinley et al., 1999) and in birds (Simon et al., 1992). For instance, administration of ANP and ANG II into the SFO influences the neuronal activities within the structure(Gutman et al., 1988) and modifies drinking behaviour in mammals and birds(Simpson and Routenberg, 1973; Lin and Hubbard, 1992; Takei, 1977). Further,receptors for ANP (NPR-A type) and ANG II (AT₁ subtype) are densely localised in the SFO of mammals and birds(Saavedra et al., 1986; Quirion, 1989; Gehlert et al., 1990; Simon et al., 1992). Thus,CVOs are the most probable sites of action of circulating hormones in the brain.

In contrast to mammals and birds, fishes are unique in that the hindbrain rather than the forebrain may be involved in the elicitation of drinking,since eels without forebrain and midbrain (`decerebrated' eels) respond to SW transfer (Hirano et al., 1972)and to peripheral injection of ANG II(Takei et al., 1979) by copious drinking. Furthermore, direct injection of ANG II into the third ventricle (Takei, 1992) or fourth ventricle (Kozaka et al.,2003) of eels induced drinking at much lower doses than peripheral injections, suggesting that the ANG II acts on the brain to induce drinking in fishes. Kozaka et al. further showed that ANP injection into the ventricle at the level of medulla oblongata (MO) inhibits drinking in SW eels(Kozaka et al., 2003), but the effect was rather weak compared with the potent inhibition after peripheral injection (Tsuchida and Takei,1998). Reflecting on all the results obtained thus far, it is likely that the major route of ANP action is from the periphery, and that circulating ANP acts on the brain, probably at the level of MO, in eels. The regulatory motor signals generated from the hindbrain may be transmitted via the vagus nerves, since in the eel bilateral vagotomy abolished drinking induced by SW transfer (Hirano et al., 1972) and by peripheral ANG II injection(Takei et al., 1979).

The present study was undertaken to determine the central receptor sites for the antidipsogenic effect of ANP in the eel. Initially, the cellular localization of NPR-A, a specific receptor for biological actions of ANP, was determined in the brain nuclei by immunohistochemistry using an antiserum specific for eel NPR-A. The BBB-deficient structures were identified in the eel brain by peripheral injection of Trypan Blue, which is impermeable to the BBB after binding to albumin (Clasen et al., 1970). Since the AP was the only site possessing ANP receptors and lacks a functional BBB, heat-coagulative lesions or chemical lesions with kainic acid were made topically to the AP, and effects of the lesioning on the antidipsogenic effect of peripherally injected ANP were examined.

Cultured, immature eels Anguilla japonica Temminck and Schlegel 1847 of either sex were purchased from a local dealer. They were maintained in SW without feeding for at least 2 weeks before use. Water in the tank was continuously circulated, aerated, and regulated at 18°C. All conditions for fish maintenance conformed to the Guideline for Maintenance and Care of Experimental Animals, and the experiments were approved by the Committee for Animal Experiments in the University of Tokyo. Eels weighed 182±2 g(N=68) at the time of surgery.

The eel brain sections were stained to determine the cells possessing NPR-A using the Elite avidin-biotin-peroxidase complex kit (Vector Laboratories,Burlingame, CA, USA). After anesthetizing eels (N=7) in 0.1% (w/v)tricaine methanesulfonate (Sigma-Aldrich, St Louis, MO, USA) for 15 min, the skull was opened and the exposed brain was fixed by 4% paraformaldehyde in 0.01 mol l^–1 phosphate-buffered saline (PBS, pH 7.3) for 1 h in situ. The brain was then dissected out of the skull and immersed in the same fixative solution for 48 h at 4°C. The brain was immersed in 30% sucrose in PBS overnight at 4°C for cryoprotection and embedded in Tissue-Tek OCT compound (Sakura Finetechnical, Tokyo, Japan) at–20°C. Serial cross sections (10 μm thick) were cut and mounted onto gelatin-coated slides and dried overnight at 4°C. For immunohistochemistry, the sections were incubated sequentially with (1) 0.6%H₂O₂ in methanol for 30 min at room temperature, (2) 2%normal goat serum in PBS (NGS-PBS) containing 0.01% sodium azide for 1 h at room temperature, (3) the affinity-purified antiserum for NPR-A diluted at 1:9000 with NGS-PBS for 48 h at 4°C, (4) biotinylated, goat anti-rabbit IgG for 30 min at room temperature, (5) avidin-biotin-peroxidase complex for 45 min at room temperature, and (6) 0.05% diaminobenzidine tetrahydrochloride and 0.01% hydrogen peroxidase in 50 mmol l^–1 Tris buffer (pH 7.2) for 10 min at room temperature. Specificity of immunoreactive signals was confirmed by preabsorption of the antibody with antigen-conjugated CNBr-activated Sepharose 4B (GE Healthcare, Bucks, UK). The specificity of this antiserum in eel tissue was confirmed by western blotting of membrane fractions prepared from the gills as previously reported(Healy et al., 2005). Immunostained sections were counterstained with Haematoxylin and then dehydrated and cleared, and coverslips placed on top for observation with a light microscope (E800, Nikon, Tokyo, Japan). The brain loci were determined in the present study based on the brain atlas of Japanese eel(Mukuda and Ando, 2003a)(Fig. 1).

To identify BBB-deficient structures in the eel brain, we injected 0.5%Trypan Blue diluted with 0.9% NaCl (saline) intraperitoneally (1 ml/day for 4 days) into anesthetized eels (N=7). 24 h after the last injection,eels were anesthetized and perfused with 30 ml of saline containing heparin(200 units ml^–1) for 30 min to remove blood cells. The brains were fixed with Bouin's solution for 24 h after a brief fixation for 1 h in situ. The fixed brain was dehydrated through graded ethanol concentrations, cleared with benzene and then embedded in Paraplast™(Oxford Labware, St Louis, MO, USA). Serial sections were cut at 10 μm,mounted onto gelatin-coated slides, and observed under a fluorescence microscope (BX 51, Olympus, Tokyo, Japan).

Since the AP was identified as a probable target site for circulating ANP(see Results), the involvement of the AP in ANP-induced antidipsogenesis was evaluated in vivo. For this purpose, eels (N=49) were anesthetized and cannulated with a polyethylene tube (o.d. 0.8 mm) in the ventral aorta for injection of ANP. Then, vinyl tubes (o.d. 1.5 mm) were inserted into the esophagus and stomach of eels for continuous measurement of drinking rate as described previously(Takei et al., 1998).

The skin was incised along the occipital region and skull exposed by removing attached muscles. The skull around the supraoccipital bone was drilled to expose the AP, and the arachnoid membrane was removed carefully so as not to break the blood vessels running mid-sagittally above the AP. The exposed AP was then heat-coagulated by electric cautery (Surgitron EMC,Ellman, Oceanside, NY, USA). After lesioning, the drilled hole was filled with bone wax (Lukens, Albuqueque, NM, USA) and the skin incision was sutured. Ten eels were subjected to the same surgical procedures without heat coagulation and served as sham-operated controls. Eight additional eels were subjected to the lesioning in the caudal part of MO adjacent to the AP and served as lesion controls. All surgical procedures were completed within 40 min/eel. Two to three animals were operated upon in the same day irrespective of the operation(sham control, lesion control or AP-lesion).

After surgery, eels were immediately placed, according to their individual size, in a Plexiglas™ trough through which aerated SW continuously circulated at 18°C to avoid deterioration of physiological conditions. The trough was covered with a black vinyl sheet to minimize visual stress to the animal during the experiments. Beginning at least 18 h after lesioning, each eel received an injection of saline as control via aortic cannula in a volume of 50 μl, followed by a flush of cannula with 50 μl saline. Subsequently, 50 μl of ANP containing 100 pmol kg^–1 eel mass was injected in the same fashion. This dose is sufficient for the antidipsogenic action of ANP (Takei and Tsukada, 2001). The saline and ANP injections were separated by an interval of at least 45 min. Drinking rate was measured throughout the experimental period. After the experiment, eels were anesthetized and decapitated for histological verification of the lesion sites(Fig. 2).

In addition to heat-coagulative lesioning by electric cautery, chemical lesioning with kainic acid was undertaken to confirm that effect was due to the lesion of AP neurons (Fig. 3). Kainic acid is an excitotoxic glutamate agonist that specifically induces degeneration of neurons(Tomioka et al., 2002). To perform a topical microinjection of kainic acid to the AP, a glass capillary injector was prepared from the glass capillary stock (o.d. 0.9 mm) with a puller and a microgrinder (Narishige, Tokyo, Japan). The capillary injector was connected to a 0.5 μl Hamilton microsyringe via polyethylene tubing (o.d. 0.61 mm) (Fig. 3B). After exposure of the AP as described above, anesthetized eels were placed on a stereotaxic apparatus (Narishige), and then the glass capillary filled with kainic acid (4 mmol l^–1, Sigma-Aldrich)containing 1% Chicago Sky Blue 6B (Sigma-Aldrich) in PBS was inserted into the AP to a depth of 0.1 mm by the manipulator of the stereotaxic apparatus. Kainic acid solution (50 nl) was injected into the AP (N=7; AP-lesion eels) or into an adjacent area (N=7; lesion control). Control injection was also performed into the AP with 1% Chicago Sky Blue 6B in PBS(N=7; sham control). The injections were accomplished within 1 min,but the capillary remained in place for 3 min to insure complete injection of volume. Beginning at least 18 h post-operatively, the eels were anesthetized and cannulae were inserted into the dorsal aorta for injection of ANP and into the esophagus and stomach for continuous measurement of drinking rate as mentioned above. On the next day of cannulation, ANP was injected into the dorsal aorta and the effect of AP lesioning on the ANP-induced antidipsogenesis was examined. After the experiment, the distribution of kainic acid was confirmed by staining with Chicago Sky Blue 6B(Fig. 3C).

The time-dependent changes in drinking rate after ANP injection were expressed in terms of percentage of the value over a 30 min period before ANP injection. Comparisons of the effects of AP lesioning on ANP-inhibited drinking among different groups were made by ANOVA, followed by Steel–Dwass test at each time point. The changes in drinking rate were measured for 20 min before and after ANP injection and compared by paired-sample Student's t-test. Significance was determined at P<0.05. All results were expressed as means ± s.e.m.

Immunoreactive NPR-A was detected in the capillary endothelial cells in the whole brain areas (Fig. 4A)including the vascular-rich saccus vasculosus (SV). Consistent with the immunohistochemical results, preliminary RT-PCR examination showed that the NPR-A gene was expressed minutely and ubiquitously in all regions of the eel brain including olfactory bulb (OB), telencephalon (Tel), optic tectum (TeO),cerebellum (Cb) and MO (data not shown). In addition to the vasculature, NPR-A positive neurons were observed at discrete loci in the MO, such as the reticular formation (RF), glossopharyngeal-vagal motor complex (GVC),commissural nucleus of Cajal (NCC) and AP(Fig. 4B–D). The perikarya of RF neurons were relatively large (ca. 20–30 μm in diameter), while the GVC neurons were ca. 10 μm and ovoid/ellipsoid in shape. The NCC and AP are neighboring nuclei in the caudal MO;NPR-A-immunopositive neurons (ca. 5–10 μm) were located predominantly near the boundary of the two loci. Neuronal staining was not detectable in this study in the more rostral parts, such as the OB, Tel, TeO and Cb(Fig. 1).

Intraperitoneally injected Trypan Blue was taken up into the blood stream and stained almost all peripheral tissues. In contrast, no staining was observed in the brain parenchyma except for the pineal gland and surrounding saccus dorsalis (SD), parvocellular and magnocellular part of preoptic nucleus, SV and AP (Fig. 5A–C). In addition to these loci, some RF and GVC neurons were stained, but the signals were much weaker than other Trypan Blue-positive loci (Table 1).

As the AP is the sole brain structure that is positive to both NPR-A immunohistochemical staining and vital staining with Trypan Blue(Table 1), it is the most probable target site of circulating ANP. Therefore, the effect of AP lesioning on antidipsogenesis caused by peripheral ANP injection was examined. Effects of the heat-coagulative and chemical lesions to the AP on drinking are summarized as changes in 20 min water intake before and after ANP injection(Table 2), since the effect of 100 pmol kg^–1 of ANP was usually continued for 20 min. After heat-coagulative lesioning of the AP, ANP injection into the circulation no longer inhibited drinking, although the inhibition was apparent in sham controls and lesion controls. In addition, the AP lesioning also reduced basal drinking in eels (Table 2). Since heat coagulation not only destroys neurons in the AP but also neural circuitry that traverses the AP, chemical lesioning that damages only AP neurons was attempted with kainic acid. Average basal drinking rate of the AP-lesioned eels with kainic acid was not different from those of sham controls and lesion controls (Table 2). Furthermore, the antidipsogenic effect of ANP was significantly suppressed by kainic acid lesioning of the AP, although slight inhibition of drinking still remained after ANP injection(Table 2). In order to highlight the effect of AP lesions on the ANP effects, the time course of changes in drinking rate was normalized to percentage change in drinking rate after intravascular injection of ANP (Fig. 6). In sham-operated eels and lesion controls with intact AP,drinking was strongly inhibited by ANP within 5 min after injection. However,total destruction of the AP by electric cauterization abolished the antidipsogenic effect of ANP injected into the circulation in SW eels(Fig. 6A). After chemical lesioning with kainic acid, the normalized time-course changes clearly shows that AP-lesioned group significantly attenuated the antidipsogenic effect of ANP compared with sham controls and lesion controls(Fig. 6B).

ANP is now established as an antidipsogenic hormone in mammals(Antunes-Rodriguez et al., 1986; Balment,2002). Although several ANP-binding sites are identified in the mammalian brains (Saavedra et al.,1986; Quirion,1989), little is known about the mechanisms of how circulating ANP acts on the brain to inhibit drinking compared with the many studies on the cerebral mechanisms for elicitation of drinking by ANG II(Fitzsimons, 1998). In contrast to the relatively weak antidipsogenic effect of ANP in mammals, ANP potently inhibits drinking in the eel, even more potently than the dipsogenic effect of ANG II (Tsuchida and Takei,1998; Takei,2002). Furthermore, ANP is highly potent for inhibition of drinking when injected systemically, which is in contrast to the much greater potency of ANG II after intracranial injection compared with peripheral injection. Therefore, the eel is an excellent model animal to investigate the cerebral mechanisms underlying the antidipsogenic effect of circulating ANP. In the present study, we pursued the site of action of peripherally generated ANP in the brain by the two histological methods. The immunohistochemical study using specific antiserum raised against eel NPR-A found immunoreactive neurons only in MO loci, such as the RF, GVC, NCC and AP. Furthermore, the AP was intensively stained by intraperitoneally injected Trypan Blue in eels, as previously reported using Evans Blue(Mukuda et al., 2005). These results show that circulating ANP can act on NPR-A in the BBB-deficient circumventricular structure in the MO. Therefore, the AP appears to be directly involved in the potent antidipsogenic action of ANP delivered from the periphery in the eel.

In addition to the histological approach, the ANP target site was also pursued by physiological studies in vivo. Heat-coagulation of the AP by electric cautery completely nullified the antidipsogenic effect of systemic injection of ANP, whereas ANP delivered by the same route strikingly inhibited drinking in sham-operated eels and in eels with lesions outside the AP. These results strongly suggest that the AP is involved in the antidipsogenic effect of circulating ANP in eel. However, heat-coagulative lesions to the AP considerably decreased basal drinking compared with sham controls and with controls experiencing lesions outside the AP. Since heat coagulation by electric cautery destroyed not only AP neurons, but also nerve fibres running through the AP, it is possible that the extensive destruction of the AP may have affected the neural circuitry that is responsible for basal drinking in eels. As an alternative method, therefore, the AP neurons were chemically destroyed by microinjection of kainic acid into the AP. Kainic acid has been successfully used to block the AP neurons in the cat(Bonigut et al., 1997). With this method, the AP lesion itself did not decrease basal drinking of SW eels. As observed after heat-coagulative lesioning, the chemical lesioning significantly attenuated ANP-induced inhibition of drinking in SW eels. Since injection of kainic acid to other sites in the MO did not affect the ANP effect, the AP neurons are most likely to be involved in ANP-induced antidipsogenesis. Compared with heat-coagulative lesioning, however, the inhibition of the ANP effect by kainic acid was partial in AP-lesioned eels. This partial inhibition is most probably due to the partial destruction of the AP, since Chicago Sky Blue 6B dye did not spread throughout the entire AP(Fig. 3C). Although there are technical limitations with the two methods employed in the present study,these data evaluated by two distinct methods clearly showed that the AP is the responsible locus for the antidipsogenic action of circulating ANP in eels.

The mechanisms of ANG II-induced drinking in terrestrial species have been well documented. In mammals and birds, the SFO and OVLT are forebrain CVOs important for controlling drinking by systemic hormones(Takei, 1992; Fitzsimons, 1998). The SFO and OVLT are the target sites for circulating ANG II(Kobayashi and Takei, 1996; Ferguson and Bains, 1996; Fitzsimons, 1998), since local injection of femtomolar doses of ANG II into these sites induces drinking, and since local lesioning of these structures eliminates the dipsogenic effects of systemic injection of ANG II (Simpson and Routenberg, 1973; Takei,1977; Ehrlich and Fitts,1990; Schmid and Simon,1992; Li and Ferguson,1993). Further, the ANG II subtype 1 receptor (AT1) is localized in the SFO (Song et al., 1992)and ANG II enhances neuronal activities of SFO(Gutman et al., 1988; Hattori et al., 1988) in the rat. Since autoradiographic study showed that ANP-binding neurons are located in the SFO in the rat (Saavedra et al.,1986), a similar cerebral mechanism may be applicable to the antidipsogenic action of ANP in terrestrial vertebrates (Antunes-Rodorigues et al., 1986), although the exact site of action of ANP has not been identified in mammals. In fishes, on the other hand, the site of action of ANG II is thought to be in the hindbrain, since `decerebration' of the entire forebrain does not influence ANG II-induced drinking in eels(Takei et al., 1979). ANP and ANG II injection into the ventricle near the AP inhibited and stimulated drinking, respectively, in a dose-related manner in eels(Kozaka et al., 2003). However, the exact target sites for circulating ANP and ANG II have not been determined yet for fishes. Since the site of action of ANP in the brain has also not been determined in mammals, the present study on eel is the first report to show the cerebral target site for the antidipsogenic action of circulating ANP in vertebrates, which is localized in the AP, a hindbrain CVO.

Although the SFO and OVLT have not been identified in eels, the magnocellular preoptic nuclei (PON) in the forebrain appears to be outside the BBB as its neurons were stained by Trypan Blue in the present study and by Evans Blue in a previous study (Mukuda et al., 2005). However, the NPR-A antibody did not stain neurons in the PON and thus this nucleus may not be a target site for circulating ANP. We show in the present study that the GVC and RF, as well as AP, were stained by Trypan Blue vital staining in the eel brain, although the intensity of staining was weaker and the number of positive neurons was smaller in the RF and GVC than those of the AP. It is possible that the RF and GVC neurons have taken up the dye from the peripheral tissues. Indeed, when Evans Blue, a retrograde tracer of neurons and structural isomer of Trypan Blue, is injected into the esophageal muscles, some neurons in the GVC and spino-occipital motor nucleus (NSO) within the RF are stained(Mukuda and Ando, 2003b).

The functional role of the AP in drinking regulation in fishes remains largely unknown at present. However, previous histological studies showed that the sinus capillaries in the AP are fenestrated, and the dendrites of AP neurons extend dorsally to reach the basal lamina of the capillaries in goldfish (Morita and Finger,1987), suggesting that teleost AP serves as an avenue for endocrine signals such as ANP and ANG II to communicate with the CNS. Furthermore, it has been shown that the GVC neurons project directly to the striated muscles of the opercular membrane, pharynx, upper esophageal sphincter and esophageal body, all of which are implicated in swallowing(Mukuda and Ando, 2003b). It was also shown that the AP neurons project ventrally to the GVC(Ito et al., 2006). Although further experiments are required to identify precise neuronal networks that are involved in ANP-induced inhibition of drinking, it is possible that the AP acts as a center for integrating the information of dipsogenic and antidipsogenic hormones from the periphery and that it regulates the coordinated behaviour of swallowing through the vagus nerves in eels.

List of abbreviations

 
• ANG II

  angiotensin II

 
• ANP

  atrial natriuretic peptide

 
• AP

  area postrema

 
• BBB

  blood-brain barrier

 
• Cb

  cerebellum

 
• CNS

  central nervous system

 
• CVO

  circumventricular organ

 
• GVC

  glossopharyngeal-vagal motor complex

 
• MO

  medulla oblongata

 
• NCC

  commissural nucleus of Cajal

 
• NPR-A

  type A ANP receptor

 
• NSO

  spino-occipital motor nucleus

 
• OB

  olfactory bulb

 
• OVLT

  organum vasculosum of the lamina terminalis

 
• PBS

  phosphate-buffered saline

 
• PON

  magnocellular preoptic nuclei

 
• RF

  reticular formation

 
• SD

  saccus dorsalis

 
• SFO

  subfornical organ

 
• SV

  saccus vasculosus

 
• SW

  seawater

 
• Tel

  telencephalon

 
• TeO

  optic tectum

The authors thank Dr Christopher A. Loretz of State University of New York at Buffalo for critical reading of the manuscript. We are indebted to Ms Sanae Hasegawa of Ocean Research Institute for technical assistance. This work was supported by Grant-in-Aid for Scientific Research (A) (13304063) from Japan Society for the Promotion of Science to Y.T. and for Creative Basic Research(12NP0201) from the Ministry of Education, Science, Sports and Culture of Japan. T.T. was supported by a Research Fellowship for Young Scientists awarded by the Japan Society for the Promotion of Science.