Ventromedial Arcuate Nucleus Communicates Peripheral Metabolic Information to the Suprachiasmatic Nucleus
http://www.100md.com
《内分泌学杂志》
Netherlands Institute for Brain Research (C.-X.Y., J.v.d.V., J.D., R.M.B.) 1105AZ, Amsterdam, The Netherlands
Tongji Medical College of Huazhong University of Science and Technology (J.D., G.Y., L.R.), 430030 Wuhan, Peoples Republic of China
Abstract
The arcuate nucleus (ARC) is crucial for the maintenance of energy homeostasis as an integrator of long- and short-term hunger and satiety signals. The expression of receptors for metabolic hormones, such as insulin, leptin, and ghrelin, allows ARC to sense information from the periphery and signal it to the central nervous system. The ventromedial ARC (vmARC) mainly comprises orexigenic neuropeptide agouti-related peptide and neuropeptide Y neurons, which are sensitive to circulating signals. To investigate neural connections of vmARC within the central nervous system, we injected the neuronal tracer cholera toxin B into vmARC. Due to variation of injection sites, tracer was also injected into the subependymal layer of the median eminence (seME), which showed similar projection patterns as the vmARC. We propose that the vmARC forms a complex with the seME, their reciprocal connections with viscerosensory areas in brain stem, and other circumventricular organs, suggesting the exchange of metabolic and circulating information. For the first time, the vmARC-seME was shown to have reciprocal interaction with the suprachiasmatic nucleus (SCN). Activation of vmARC neurons by systemic administration of the ghrelin mimetic GH-releasing peptide-6 combined with SCN tracing showed vmARC neurons to transmit feeding related signals to the SCN. The functionality of this pathway was demonstrated by systemic injection of GH-releasing peptide-6, which induced Fos in the vmARC and resulted in a reduction of about 40% of early daytime Fos immunoreactivity in the SCN. This observation suggests an anatomical and functional pathway for peripheral hormonal feedback to the hypothalamus, which may serve to modulate the activity of the SCN.
Introduction
ALTHOUGH MOST RESEARCH on nonphotic input to the suprachiasmatic nucleus (SCN) has focused on the intergeniculate leaflet and raphe nuclei, signals from other parts of the brain are also possible. The secretion of many hormones is synchronized by the SCN, and feedback from peripheral hormones to the hypothalamus could thus also influence this central clock and adjust its output. Although many metabolic active peripheral hormones/substances have transport systems to pass the blood-brain barrier, there are selected areas in the central nervous system (CNS) with special sensory functions for circulating substances, such as the arcuate nucleus (ARC). Several studies analyzing the possible ability of the ARC to sense blood-borne substances indicate that the ventromedial ARC (vmARC) may sense signals directly from the general circulation via the median eminence (ME) (1, 2, 3) and thus functions as an important sensory "window" for blood-born substances to the brain (4, 5, 6). This is firmly supported by the presence of receptors with high expression levels for nearly all known circulating metabolically active hormones, such as leptin, insulin, glucocorticoid, and ghrelin in the vmARC (7, 8, 9, 10). These hormones and also glucose are able to modulate the electrical activity of vmARC neurons (11, 12, 13) and thus affect the output of the vmARC in energy homeostasis. This special role of the vmARC in integrating hormonal signaling was confirmed by the fact that leptin signaling to the CNS crucially depends on normal ARC functioning (14, 15).
The anatomical mapping of ARC in the CNS is still updated with a variety of methods, such as transsynaptic (16) or monosynaptic tracing studies (14, 17). Because a specific vmARC tracing has not yet been performed, a systematic retrograde and anterograde tracing study is necessary to understand the anatomical relationship of the vmARC within the CNS.
We therefore studied the connections of vmARC with the CNS and used injection of fluorophore-conjugated cholera toxin B (CTB; CTB-Alexa Fluor 555) as retrograde and anterograde tracers into this area. The specificity of the projections of the vmARC to the SCN and nucleus of the solitary tract (NTS) were confirmed by control tracing from SCN and NTS, respectively.
The functional interaction between vmARC and SCN was demonstrated by the use of the natural leptin antagonist, the gut-brain peptide ghrelin (18), which is the only metabolic molecule found to date that is able to activate vmARC neurons by specifically targeting agouti-related peptide (AGRP)/neuropeptide Y (NPY) neurons and that stimulates food ingestion (19). With systemic administration of the ghrelin mimetic, GH-releasing peptide-6 (GHRP-6), combined with SCN tracing and AGRP/Fos staining, the present data showed that GHRP-6 activated vmARC neurons projecting to the SCN. Finally, the influence of systemic GHRP-6 stimulation on diurnal Fos immunoreactivity in the SCN was investigated.
Materials and Methods
CTB tracing
All the following experiments were conducted under the approval of the animal care committee of the Royal Netherlands Academy of Sciences. Male Wistar rats, weighing 300 ± 10 g (Harlan Nederland, Horst, The Netherlands), were housed at room temperature with a 12-h light, 12-h dark schedule (lights on at 0700 h). For tract tracing, rats were anesthetized [0.8 ml/kg Hypnorm (Janssen, High Wycombe, Buckinghamshire, UK), im, and 0.4 ml/kg Dormicum (Roche, Almere, The Netherlands), sc], mounted with their heads in a standard stereotaxic apparatus, tooth bar set at –3.4 mm, and received tracer injection. Because, in contrast to CTB, fluorophore-labeled CTB cannot be applied by iontophoresis, we used pressure injection. Thus, fluorophore-conjugated CTB (with Alexa Fluor 555, Molecular Probes, Eugene, OR; CTB will be used in the text as the abbreviation of CTB-Alexa Fluor 555 for convenience) was used for purpose of combination of neurotransmitter and tracer detection in target area. Finally, in 30 cases CTB injections were aimed at the vmARC. To confirm the reciprocity of connection, injections with CTB were also aimed at the SCN (n = 30) or NTS (n = 10); all coordinates for these regions were adapted from the atlas of Paxinos (20).
CTB was injected with pressure (10 mbar, 5 sec) using a glass pipette with a tip of maximally 50 μm in diameter to minimize damage of passing fibers; the glass pipette was filled with an injection volume of 50–100 nl. The coordinates for the vmARC injection were 3.3 mm caudal to bregma, 0.4 mm lateral to the midline, and 10.4 mm below the dura. The coordinates for SCN injection followed those reported by Buijs et al. (21); for NTS injection, the coordinate was 0.4 lateral to the midline, 13.8 mm caudal to bregma, and 5 mm ventral from dura. After the tracer was injected by pressure injection, the pipettes were fixed with dental cement to the skull and left in the brain to minimize leakage from the tract until the animals were killed. This procedure did not lead to visible extra discomfort of the animals, whereas the reduction in leakage along the pipette track is substantial. Fifteen rats of the SCN tracing were also combined with the placement of a jugular venous catheter for the iv injection of GHRP-6. See below for details on this operation protocol.
GHRP-6 iv infusion
For the iv infusion of GHRP-6, a silicone catheter was implanted in the right jugular vein according to the method of Steffens (22). The operation of the first group of rats (n = 15) was combined with SCN tracing, such that both operations were carried out at same time. The second group of rats (n = 8) received only catheter implantation for iv infusion of GHRP-6 (n = 4) or vehicle (Ringer solution; n = 4) to examine Fos immunoreactivity in the vmARC and SCN. After surgery, all rats were given 10 d to recover. After the operation, the animals were handled every day and connected to the jugular catheter to familiarize the rats with the experimental procedures. On d 9, the rats were connected to a drug administration catheter for 24 h, which was kept out of reach of the rats by means of counterbalanced beam. This allowed all manipulations to be carried out outside the cages without handling the rat. All experiments were performed in the rat home cage.
The drug administration catheter was connected to a syringe containing vehicle with or without GHRP-6. Infusion into the catheter was initiated at Zeitgeber time 1 (ZT1), with ZT0 being defined as the onset of the light period. GHRP-6 was used at a dose of 5 μg/rat dissolved in 0.4 ml Ringer solution, and 0.4 ml Ringer solution was used as the vehicle control infusion. The infusion speed was kept at 0.1 ml/min. All rats were killed by perfusion 100 min after the infusion.
Food deprivation
We found that neurons in the subependymal layer of ME (seME) have a similar projection pattern as those in the vmARC. Previous studies supposed them to be an extension of vmARC (23); therefore, we examined whether seME neurons show a comparable increase in the expression of NPY/AGRP after food deprivation as observed in vmARC neurons (24). Food, but not water, was removed at ZT1 from four intact rats for 48 h; subsequently, animals were killed by perfusion.
Immunocytochemical staining
All rats were deeply anesthetized with a lethal dose of sodium pentobarbital and perfused with saline, followed by a solution of 4% paraformaldehyde in 0.1 M PBS (pH 7.4) at 4 C. The brains were removed and kept in fixative at 4 C for overnight postfixation, equilibrated 48 h with 30% sucrose in 0.1 M Tris-buffered saline (TBS; pH 7.2). Brains were coronally cut in a cryostat into 30-μm sections; sections used for immunolabeling were collected and rinsed in 0.1 M TBS.
With ARC, ME, SCN, and NTS tracing, whole brain sections were incubated with rabbit anti-CTB primary antibody (Sigma-Aldrich Corp., St. Louis, MO) at a 1:100,000 dilution. Sections of fasted rats were incubated with rabbit anti-AGRP (1:1500; Phoenix Pharmaceuticals, Belmont, CA) or NPY (1:2000; Niepke 091188, Netherlands Institute for Brain Research) primary antibody. Sections of second group of GHRP-6 infusion rats were incubated with goat anti-Fos (1:1500; Santa Cruz Biotechnology, Inc., Santa Cruz, CA). After all these primary antibody incubations overnight at 4 C, sections were rinsed and incubated in biotinylated secondary antibody (goat antirabbit or horse antigoat IgG) for 1 h, then rinsed and incubated in avidin-biotin complex (Vector Laboratories, Inc., Burlingame, CA) for 1 h, and the reaction product was visualized by incubation in 1% diaminobenzidine (0.05% nickel ammonium sulfate was added to the diaminobenzidine solution to darken the reaction product) with 0.01% hydrogen peroxide for 5–7 min. Sections were mounted on gelatin-coated glass slides, dried, run through ethanol and xylene, and covered for observation by light microscope.
Confocal microscopy
To characterize the functionality of neurons in the vmARC that project to the SCN, we used triple-labeling immunofluorescence to show the colocalization of GHRP-6-induced Fos with AGRP in retrogradely CTB-labeled neurons in vmARC from SCN tracing. Sections were incubated overnight at 4 C with goat anti-Fos and rabbit anti-AGRP primary antibody, rinsed in 0.1 M TBS, incubated 1 h in biotinylated horse antigoat IgG, rinsed, incubated with streptoavidin-Cy5 and donkey antirabbit-Cy2 for 1 h, rinsed, mounted on gelatin-coated glass slides, dried, covered with glycerol in 0.1 M PBS (pH 9.0), and observed by confocal laser scanning microscopy.
Analysis of Fos immunoreactivity
The effect of GHRP-6 treatment on Fos immunoreactivity in the SCN was determined in four GHRP-6-infused or four Ringer solution-infused rats. From SCN sections, tiled images were captured by a computerized image analysis system consisting of an Axioskop 9811+ Sony XC77 black and white video camera (Sony Corp., Tokyo, Japan). In these images, the right side of middle portion of the SCN was manually outlined. The Fos-positive nuclear profiles were automatically (no user interaction) segmented by a dedicated macro written within the ImagePro programming environment. For each rat, three sections were measured 90 μm apart (from bregma –1.20 to –1.40 mm); the mean number of Fos-positive nuclear profiles from these three sections was calculated. All values are expressed as the mean ± SEM, and data were analyzed using one-way ANOVA. Statistical significance was set at P < 0.01.
Results
CTB injections into the vmARC and seME
Of all the animals injected, two rats had a successful injection within the middle part of the vmARC along its rostrocaudal extension, and two rats had a successful injection within the seME. Due to the difficulties of accurate vmARC injection, we only focused on the middle part (in the anterior posterior direction) of the vmARC, which guaranteed maximal vmARC space for tracing. Misplaced injections were mainly located in a different coronal plane, but usually at the same distance from the bregma. The misplaced injections in the parenchymal area, including the external zone of the ME and the dorsal or ventrolateral ARC, served as controls for the specificity of the projections of the vmARC and seME. Misplacement and injection in the third ventricle served as control for leakage of tracer into cerebrospinal fluid, which is almost inevitable. The injections in which the small tip of glass pipette was blocked by brain tissue and the tracer could not be pushed out by pressure were used as specificity controls of the used antibodies. In none of these latter cases was any staining observed, which shows that all the staining observed was due to CTB.
Tracer injections into the vmARC resulted in the spread of CTB not only limited to the ventromedial part of the nucleus, but expanding also into the ipsilateral and contralateral subependymal layers of the ME (Fig. 1A). After seME injection, the CTB distribution covered the dorsal aspect of the ME, between the ependymal layer and the fibrous layer (Fig. 1D). AGRP- and NPY-containing cell bodies can more easily be detected in the seME as well as in the vmARC after 48 h of fasting (Fig. 1, C and F), whereas with a normal feeding schedule, we only observed AGRP or NPY fibers and terminals in these two areas. This observation together with the distribution pattern of the tracer could support the previous suggestion based on development that seME neurons are an extension of the ARC (23). Consequently, Table 1 shows retrograde and anterograde labeling in different areas of the CNS with CTB injections into the vmARC and seME.
Control injections
Injections in the ventrolateral ARC (Fig. 1B), dorsal ARC, external zone of the ME (Fig. 1E), and third ventricle were used as controls for specificity of the projections of the vmARC and seME and leakage of the tracer. Three rats had tracing into the external zone of the ME, three rats had ventricle injection only, and two rats received ventrolateral ARC and dorsal ARC tracing, respectively.
The labeling pattern of external zone tracing was consistent with those from Lechan et al. (25) and Wiegand and Price (26). In contrast with the seME injection, it revealed mainly retrogradely labeled neurons in the parvocellular paraventricular nucleus (PVN; Fig. 3F) and several magnocellular neurons in PVN and supraoptic nucleus (SON; Fig. 3I) as well as the periventricular nucleus and magnocellular dorsal ARC. The absence of detectable anterograde tracing confirms the external zone of ME as a pure neurosecretory pathway of the hypothalamus. Three control injections into the third ventricle resulted in labeling of the ependymal plexus and tanycytes or cerebrospinal fluid contacting neurons along the third and lateral ventricles, some neurons in the dorsal raphe nucleus were also labeled. This pattern was consistent with ventricle tracing experiments aiming at detecting cerebrospinal fluid contacting neurons in the CNS by CTB-horseradish peroxidase injection (27). Injections into the ventrolateral or dorsal part of the ARC had projection patterns distinct from that of injection into the vmARC, e.g. with ventrolateral ARC injection even when very close to the vmARC, we only observed some neurons present in the dorsal and medial parts of the SCN (Fig. 3G) and no projections to the SON or NTS.
Retrograde and anterograde labeling in the CNS with CTB injection into the vmARC and seME
The injection cases R122 and R173 had their injection tips localized within the vmARC, whereas R121 and R144 had their injection tips in the seME. Cases R122–R173 and R121–R144 showed some similar pattern of retrograde and anterograde labeling in the CNS, although with several differences exemplified by the presence of retrogradely filled neurons and anterogradely filled fibers in a number of additional areas after injection in the vmARC as compared with injection in the seME (see Table 1). The projections from the seME area as compared to the projections from the vmARC are less dense than vmARC.
Areas with reciprocal interaction with vmARC and seME
CTB tracer injection into the vmARC and seME revealed several sites that have extensive reciprocal interaction with this area. In the anterior part of the CNS, reciprocal connections with this area arise from the anteroventral PVN, ventromedial preoptic nucleus, and medial preoptic nucleus (lateral part), with labeled cells and fibers scattered throughout these areas, and the labeled neurons were mainly small in size. Both lateral septum and accumbens nucleus connect with vmARC reciprocally, but labeled fibers were not visible after seME tracing.
We observed that all parts of the SCN show labeled neurons, with the highest density in the medial and dorsal parts. Moreover, abundant labeling of fibers and terminals were found in the dorsal and ventral parts. A few labeled neurons and fibers were also observed in the area around the SCN (Fig. 2). From rostral to caudal, the whole extension of the periventricular hypothalamic nucleus is labeled by scattered neurons and fibers.
Consistent with previously published data about the projection from the ARC to the PVN (28, 29), with vmARC and seME tracing, the triangular shape of the PVN was full of labeled fibers and terminals (Fig. 3, A and B); in contrast with the dense labeling of fibers and terminals, only few PVN neurons were labeled. Although when the injection was in the external zone of the ME, PVN labeling was characterized by the presence of a large number of neurons that by location and appearance can easily be recognized as neuroendocrine (Fig. 3F). Also in the SON the contrast is obvious; vmARC and seME injections result in dense innervations in SON (Fig. 3H), whereas injection in the external zone only shows retrogradely filled neurons (Fig. 3I). After vmARC injection, a considerable number of fibers and small neurons in the sub-PVN (30) were also labeled (Fig. 3A). More laterally, the reciprocal connection also only exists between lateral hypothalamus and vmARC, but not with the seME.
Within the whole ARC, many anterogradely labeled fibers as well as a few retrogradely labeled neurons were present in the dorsal and ventrolateral parts of the ARC near the border that separates the ARC and ventromedial hypothalamic nucleus (VMH; Fig. 1). This labeling pattern of ARC was consistent from rostral to caudal.
Another important target of the vmARC and seME is the dorsomedial hypothalamic nucleus (DMH). A large number of labeled neurons, fibers, and terminals visualized throughout the DMH demonstrate the reciprocal projection between the vmARC and DMH (Fig. 3D). Interestingly, the vmARC and seME tracing revealed a very weak anatomical relationship with the VMH, with just a few neurons and fibers being labeled. In the posterior hypothalamus, numerous labeled neurons were present in the ventral premamillary nucleus, whereas fibers were relatively sparse. Of the tuberomamillary nucleus, the dorsal part received a denser innervation from the vmARC and seME than the ventral area.
In addition, neurons and some axonal terminals were observed in the other three circumventricular organs (CVOs) after vmARC and seME injection. In the subfornical organ (SFO), neurons and fibers were spread over the vascular area (Fig. 4A). In the organum vasculosum of the lamina terminals (OVLT), abundant labeling was present in neurons and fibers bilaterally located in the vascular zone and perinucleus area. A strong labeling of fibers, as in the ependymal plexus, was seen along the wall of the ventricle (Fig. 4B). Similarly, in the area postrema (AP), several fibers and neurons were observed in the vascular area (Fig. 4C). At the border area between the AP and NTS, we visualized strongly labeled fibers, probably derived from the ependymal plexus that is often labeled after injection in the seME (Fig. 4C), this labeling pattern in AP as well as in OVLT, however, was also seen with the third ventricle control injection, whereas with control incubation of anti-CTB antibody with sections from nontracing, we did not see any such staining, and we also have not found such a labeling pattern with lateral ARC and SCN CTB tracing animals.
In the brain stem, almost all subdivisions of the parabrachial nucleus (PB) contain labeled neurons and terminals from the most rostral to the caudal part of the nucleus (Fig. 4, E and F). In addition, scattered neurons and terminals were seen in the periaqueductal gray, locus coeruleus, Barrington’s nucleus, and prepositus nucleus.
Similar to the PB, another visceral integration nucleus, the NTS, contained several labeled neurons and fibers from the caudal to the middle part. Consistent with a previous study, after vmARC injection, we observed a dense labeling of fibers in a limited area at ipsilateral part in the caudal NTS (Fig. 4, H and I); in view of the paucity of retrogradely labeled neurons, we consider these fibers be largely due to anterograde transport and not derived from retrogradely labeled neurons as has been suggested to occur with CTB tracing (31). Interestingly, some small labeled neurons were also found in the dorsal motor nucleus of the vagus (DMV) (Fig. 4H). These finding corroborate our previous observation that the area of the DMV incorporates neurons without an autonomic motor function (32).
Areas with only input to or output from the vmARC and seME
A few regions seem to only provide input to the vmARC and seME, as demonstrated by the presence of merely retrogradely labeled cell bodies and no clear staining of fibers. Because other brain regions at the same level do show labeled fibers, this does not seem to be a technical limitation of the anterograde tracing. Only retrogradely filled neurons were detected in the dorsal endopiriform nucleus, claustrum, bed nucleus of the stria terminalis, and posterior part of the PVN. The whole extension of most of the medial amygdaloid complex was also strongly retrogradely labeled (Fig. 4D).
With vmARC and seME injections, a dense innervation of the SON, retrochiasmatic area, and tuber cinereum area, but without retrogradely labeled cell bodies, indicates that these areas only get input from the vmARC and seME. We also found that the vmARC and seME tracing resulted in projections to the intergeniculate leaflet (Fig. 4G).
Retrograde tracing from SCN and NTS
To provide additional evidence for the projections of the vmARC and seME to the SCN and NTS and to ensure that the projections are not due to uptake from passing fibers or to anterograde labeling by retrogradely labeled neurons (33, 34), CTB was injected into the SCN and NTS.
Injections of CTB into both the SCN and NTS revealed not only neurons in the vmARC (Fig. 5, A, B, and D), but also retrogradely labeled neurons in the subependymal layer of the ME with a location just under the ependymal lining of the third ventricle (Fig. 5C). We also considered the abundant presence of fibers as evidence for the reciprocity of the connection of the vmARC and seME with the SCN and NTS. The labeled neurons in the seME area were different from glial cell. Firstly, the location of these subependymal neurons is just under and very close to the ependymal layer lining the bottom of the third ventricle. In Fig. 5C we can see the difference with ependymal cells that were not labeled by CTB. Secondly, the orientation of the oval or fusiform-shaped labeled neurons was parallel with the border of the third ventricle, whereas glia cells have many more processes and may extend to every orientation, including the external zone. Thirdly, the size of subependymal neurons was far larger than glial cell, either tanycytes or astrocytes.
After CTB injection into the SCN, fibers and terminals were abundant in the lateral and dorsal parts of the ARC where few retrogradely labeled neurons were present (Fig. 5, A and B). This projection pattern confirmed the vmARC and seME tracing, which resulted in labeling of neurons, fibers, and terminals in the SCN, and the ventrolateral and dorsal ARC tracing, which resulted in only labeled neurons in the SCN without fibers or terminals (Fig. 3G).
Functionality of the connections between vmARC and SCN
We addressed the possible role of vmARC neurons in the regulation of metabolism in relation to their projection to the SCN by examining the induction of Fos in AGRP neurons after systemic administration of GHRP-6 combined with SCN tracing. In addition, we observed that GHRP-6 stimulation allowed an easier detection of AGRP in vmARC neurons, which is normally difficult. Most AGRP-positive neurons also expressed Fos, and some of these Fos-positive AGRP neurons were shown to project to the SCN, as revealed by colocalization with CTB from SCN tracing (Fig. 6). Interestingly, in both ad libitum-fed and feeding-restricted rats, it was difficult to detect AGRP fibers in the SCN. We observed only very thin AGRP-positive fibers in the ventral SCN after 2 h of refeeding following 48-h fasting (data not shown), which suggests that although AGRP neurons are the target of peripheral GHRP-6 and project to the SCN, the detection of this peptide is limited, probably because of the low quantity of peptide present in these nerve terminals.
After SCN tracing, there was no obvious Fos immunoreactivity in ARC neurons or in intact rat ARC that were housed and perfused under the same conditions as SCN tracing rats, which means that SCN tract tracing has no effect on Fos immunoreactivity in the ARC. Fos immunoreactivity in vmARC induced by GHRP-6 is apparent after a 5-μg iv infusion, whereas the Ringer solution infusion had no detectable Fos in the vmARC (Fig. 7, A and B). In addition, Fos immunoreactivity was checked in the geniculate area and raphe nucleus, which are responsible for nonphotic input to the SCN, and there was no significant difference between Ringer control and 5 μg GHRP-6 administration in both nuclei.
The number of detectable Fos-positive nuclear profiles in the middle portion SCN was counted in controls as well as GHRP-6-infused animals. Negatively correlated with the same dose of 5 μg GHRP-6 action on the vmARC, the number of Fos-positive nuclei in the middle portion SCN was decreased by about 40% compared with that after Ringer solution infusion (195.33 ± 14.45 vs. 116.89 ± 4.7 in Ringer solution and 5 μg GHRP-6, respectively; P < 0.01; Fig. 7, C and D).
Discussion
In the present study the properties of CTB as anterograde and retrograde tracer were confirmed by the presence of tracer-labeled terminals and cell bodies, for example, in the NTS after injection into the vmARC, which is consistent with previous tracing studies on the ARC and NTS (17) and consequently allowed us to study the input as well as the output of one area simultaneously.
To avoid uptake of tracer by damage of fibers of passage as much as possible, glass pipettes with a tip approximately 50 μm in diameter were used for pressure injection instead of a Hamilton syringe (400 μm), which minimized the possibility of tracer uptake by passing fibers.
This is demonstrated by the fact that only injection in the external zone of the median eminence resulted in retrograde labeling of neuroendocrine neurons in the PVN and SON, which confirms previous studies (25, 26). Injections into other areas where neuroendocrine fibers pass, such as the subependymal layer-internal zone and the lateral ARC, did not result in cell body labeling. In contrast, injection into the subependymal layer-internal zone resulted only in labeling of fiber terminals in the SON and PVN. These findings suggest not only a high specificity of projections between the different layers of the median eminence, but also a minimal uptake of fibers of passage by the method used.
The present results confirm and extend previous anatomical studies of ARC and ME and provide a more comprehensive view of the organization of the ARC, including its interaction with the biological clock (Fig. 8).
Moreover, we demonstrate for the first time that both the vmARC and seME areas have a similar extensive interaction with the SCN, the sensory CVOs, and other parts of the brain. In both the vmARC and seME, the blood-brain barrier operates in a modified manner comparable with that of other CVOs that also have permeable capillary networks for penetration of circulating substances (35). Moreover, both areas not only share the same projection patterns, but also contain NPY and AGRP neurons, which respond similarly to, for instance, fasting. Considering these aspects, we propose that the vmARC-seME should be viewed as an arcuate-median eminence complex (AMC).
It is proposed that three typical sensory CVOs, SFO, OVLT, and AP, have active transport systems and receptors for blood-borne molecules and play a key role in sensing and relaying humoral information to other CNS areas (36). Our present data demonstrate that the AMC has similar specific neuronal connections to hypothalamic neuroendocrine and autonomic centers, such as the SFO, OVLT, and AP. Early studies have already suggested that the blood-brain barrier border could reside at the area between the ventromedial (or proximal) part and dorsolateral (or distal) part of the ARC, and that the perivascular spaces of vmARC could thus communicate with the neurohemal area of the ME, where fenestrated capillaries allow the passage of larger molecules (1, 2, 3). Therefore, the present finding that the SCN receives direct input from the AMC suggests a route of communication between the general circulation and the SCN. Intravenous injection of the ghrelin mimetic GHRP-6 results in Fos-positive neurons in a narrow band in the AMC. Of all tracer injections in the ARC (n = 30), many were very close to the vmARC; only four injections that were inside and covered the same area that is known to express Fos after GHRP-6 stimulation had reciprocal interaction with the SCN. This suggests that a relatively small area is involved in this exchange of information. Furthermore, injections into the SCN resulted in the combination of retrogradely labeled neurons and fibers only in this ventromedial area of the ARC, which confirms the reciprocity of this connection. In view of the (daily) time-dependent organization of metabolism, which depends so crucially on the presence of the SCN (37), the interaction of the AMC with the SCN could serve this timing purpose and provide the SCN with information about glucose concentration and the associated level of hormones.
That there are innervations from vmARC to SCN is a novel finding. However, previous electrophysiological studies on ARC (38) indicated this connection. Other proof for the functional interaction between vmARC and SCN comes from a number of studies indicating that physiological and behavioral circadian rhythms are also affected by ARC malfunction. The fact that ARC lesions by neonatal monosodium glutamate treatment, which mainly destroys NPY cells (39), is able to compromise the circadian activity while the SCN remains intact (40, 41), could be interpreted, together with the present data, to mean that the innervations from vmARC to the SCN may relay peripheral hormonal information to the SCN and may thus affect the circadian activity of the SCN. At the same time, input from the SCN to the ARC (38, 42, 43) may time the activity of the ARC, as shown by rhythmic Fos expression in the ARC (44) and the rhythmic activity of dopaminergic ARC neurons negatively driven by vasoactive intestinal peptide from the SCN (45). Actually, such an SCN-CVO reciprocal connection has already been demonstrated for the SCN-SFO and SCN-OVLT interaction (46, 47). Additional evidence for a more general interaction between SCN and CVOs, including the ARC, can be found in studies describing the action of prokineticin 2, which is synthesized in the SCN and for which receptors have been shown in the SFO and the vmARC (48, 49). Together, these studies suggest the capacity of the SCN to modulate the properties of the CVOs, probably to control or regulate the sensitivity of CVOs to peripheral signals, and the present study indicates that these signals may be relayed to the SCN as well not only via the SFO and OVLT, but also via the vmARC and seME. In view of the high sensitivity to peripheral metabolic hormones in the vmARC, the possible feedback pathway for these hormones to the SCN is proposed to be via this area.
Systemic administration of the ghrelin mimetic, GHRP-6, combined with SCN tracing showed that AGRP neurons activated by GHRP-6 project to the SCN; at the same time, the diurnal Fos immunoreactivity in the SCN was reduced by 40%, suggesting that the ghrelin mimetic, GHRP-6, could interfere with the activity of the SCN via the vmARC. The obvious decrease in Fos in the ventrolateral SCN suggests a possible regulation of the light-receiving part of the SCN. The fact that the staining of AGRP fibers in the SCN compared with the PVN is much less intense may indicate that a certain selection of neurons in vmARC projects to the SCN, or those collaterals of vmARC neurons to SCN and PVN may contain transmitters in different proportions.
It is well known that most nonphotic inputs to the SCN have the ability to reset the clock when they are provided during the day (50, 51). Nonphotic inputs to the SCN were shown to arise from the intergeniculate leaflet (52) and raphe nucleus (53). In addition, nonphotic inputs are known to inhibit photic-responsive Fos expression in the SCN at night (54). In the present study with peripheral GHRP-6 administration, there is no obvious variation in Fos immunoreactivity in the geniculate area and raphe nucleus, which indicates that these areas are not activated. Consequently, the GHRP-6 information relayed by the vmARC to the SCN, as indicated by Fos diminishment in the SCN, is another functional support for the anatomical connection between the AMC and SCN, and suggests that pharmacological effects of ghrelin may inhibit the activity of the SCN during the light period. This observation agrees with the fact that with a normal feeding schedule, ghrelin peaks after dark onset (39), suggesting another central function of ghrelin, such as interaction between locomotion and metabolism (55, 56).
Because the SCN is essential for the organization of daily metabolic activity (the daily rhythm of plasma glucose and metabolic hormones is driven by the SCN) (37), the present results provide an anatomical basis for our previous findings that metabolic cues could indeed influence circadian activity of, for example, vasopressin neurons in the SCN (57). Additional research is needed to elucidate the effect of the metabolism-related nonphotic input on SCN functionality and the possible neurotransmitters that are involved in the signaling.
Naturally the neuronal connections of AMC with other areas in the CNS may also implicate those areas as potential participants in metabolism. In this study we stress the role of interaction between AMC and CVOs; not only are the classic CVOs engaged in monitoring changes in metabolic, osmotic, ionic, and hormonal compositions, they also show neuronal activation or inhibition by cytokines (58, 59, 60). In this study we show that the SFO, OVLT, and AP have reciprocal connections with the vmARC and seME, centers for the control of food intake, which suggests a functional interaction of CVOs, for example, in cytokine activity and sickness behavior (61). This could be the anatomical basis for anorexia and weight loss during inflammations. In addition, the present finding that the amygdala complex and nucleus accumbens provide input to the AMC supports the significance of limbic structures in energy homeostasis. This possibility is supported by the fact that chemical manipulation of the nucleus accumbens induced high Fos expression in ARC NPY neurons (62).
Moreover, the AMC area has reciprocal connections with visceral sensory integration sites such as the PB and NTS, as was shown in part previously (17, 63). This indicates that the vmARC and seME may integrate hormonal information with signals from first order visceral sensory centers (64, 65). Logically, the outcome of this interaction should be signaled to many sites, including the SCN, and is necessary for the multilevel central control of visceral activities.
Our present data provide the anatomical and functional basis for the hypothesis that both humoral and neuronal metabolic information may be relayed back to the SCN via the AMC. Herein the AMC area is responsible for the first order integration of circulating signals. When the normal metabolic balance is seriously disturbed, e.g. by night-eating syndrome (66) or stress, this may influence the integrative activity of the AMC as well as the organizational activity of the SCN. Thus, a slight disturbance at the level of these essential integration sites raises a serious risk of an unbalanced energy state and possibly obesity, diabetes, or other metabolic diseases (67).
Acknowledgments
We thank Drs. Carolina Escobar and Marco van der Top for their valuable comments on the manuscript, Wilma Verweij for correction of the English, and Henk Stoffels for the art work.
Footnotes
This work was supported by a grant from the Royal Netherlands Academy of Arts and Sciences for a joint project between The Netherlands and China (03CDP014).
First Published Online September 29, 2005
Abbreviations: AGRP, Agouti-related peptide; AMC, arcuate-median eminence complex; AP, area postrema; ARC, arcuate nucleus; CNS, central nervous system; CTB, cholera toxin B; CVO, circumventricular organ; DMH, dorsomedial hypothalamic nucleus; DMV, dorsal motor nucleus of the vagus; GHRP-6, GH-releasing peptide-6; ME, median eminence; NPY, neuropeptide Y; NTS, nucleus of the solitary tract; OVLT, organum vasculosum of the lamina terminals; PB, parabrachial nucleus; PVN, paraventricular nucleus; SCN, suprachiasmatic nucleus; seME, subependymal layer of the median eminence; SFO, subfornical organ; SON, supraoptic nucleus; TBS, Tris-buffered saline; vmARC, ventromedial arcuate nucleus; VMH, ventromedial hypothalamic nucleus; ZT, Zeitgeber time.
Accepted for publication September 22, 2005.
References
Broadwell RD, Balin BJ, Salcman M, Kaplan RS 1983 Brain-blood barrier Yes and no. Proc Natl Acad Sci USA 80:7352–7356
Krisch B, Leonhardt H 1978 The functional and structural border of the neurohemal region of the median eminence. Cell Tissue Res 192:327–339
Shaver SW, Pang JJ, Wainman DS, Wall KM, Gross PM 1992 Morphology and function of capillary networks in subregions of the rat tuber cinereum. Cell Tissue Res 267:437–448
Berthoud HR 2002 Multiple neural systems controlling food intake and body weight. Neurosci Biobehav Rev 26:393–428
Horvath TL, Diano S 2004 The floating blueprint of hypothalamic feeding circuits. Nat Rev Neurosci 5:662–667
Elmquist JK, Elias CF, Saper CB 1999 From lesions to leptin: hypothalamic control of food intake and body weight. Neuron 22:221–232
Aronsson M, Fuxe K, Dong Y, Agnati LF, Okret S, Gustafsson JA 1988 Localization of glucocorticoid receptor mRNA in the male rat brain by in situ hybridization. Proc Natl Acad Sci USA 85:9331–9335
Unger J, McNeill TH, Moxley III RT, White M, Moss A, Livingston JN 1989 Distribution of insulin receptor-like immunoreactivity in the rat forebrain. Neuroscience 31:143–157
Schwartz MW, Seeley RJ, Campfield LA, Burn P, Baskin DG 1996 Identification of targets of leptin action in rat hypothalamus. J Clin Invest 98:1101–1106
Guan XM, Yu H, Palyha OC, McKee KK, Feighner SD, Sirinathsinghji DJ, Smith RG, Van der Ploeg LH, Howard AD 1997 Distribution of mRNA encoding the growth hormone secretagogue receptor in brain and peripheral tissues. Brain Res Mol Brain Res 48:23–29
Pinto S, Roseberry AG, Liu H, Diano S, Shanabrough M, Cai X, Friedman JM, Horvath TL 2004 Rapid rewiring of arcuate nucleus feeding circuits by leptin. Science 304:110–115
van den Top M, Lee K, Whyment AD, Blanks AM, Spanswick D 2004 Orexigen-sensitive NPY/AgRP pacemaker neurons in the hypothalamic arcuate nucleus. Nat Neurosci 7:493–494
Spanswick D, Smith MA, Mirshamsi S, Routh VH, Ashford ML 2000 Insulin activates ATP-sensitive K+ channels in hypothalamic neurons of lean, but not obese rats. Nat Neurosci 3:757–758
Bouret SG, Draper SJ, Simerly RB 2004 Formation of projection pathways from the arcuate nucleus of the hypothalamus to hypothalamic regions implicated in the neural control of feeding behavior in mice. J Neurosci 24:2797–2805
Dawson R, Pelleymounter MA, Millard WJ, Liu S, Eppler B 1997 Attenuation of leptin-mediated effects by monosodium glutamate-induced arcuate nucleus damage. Am J Physiol 273:E202–E206
DeFalco J, Tomishima M, Liu H, Zhao C, Cai X, Marth JD, Enquist L, Friedman JM 2001 Virus-assisted mapping of neural inputs to a feeding center in the hypothalamus. Science 291:2608–2613
Sim LJ, Joseph SA 1991 Arcuate nucleus projections to brainstem regions which modulate nociception. J Chem Neuroanat 4:97–109
Kojima M, Hosoda H, Date Y, Nakazato M, Matsuo H, Kangawa K 1999 Ghrelin is a growth-hormone-releasing acylated peptide from stomach. Nature 402:656–660
Nakazato M, Murakami N, Date Y, Kojima M, Matsuo H, Kangawa K, Matsukura S 2001 A role for ghrelin in the central regulation of feeding. Nature 409:194–198
Paxinos G, Watson C 1998 The rat brain in stereotaxic coordinates. San Diego: Academic Press
Buijs RM, Markman M, Nunes-Cardoso B, Hou YX, Shinn S 1993 Projections of the suprachiasmatic nucleus to stress-related areas in the rat hypothalamus: a light and electron microscopic study. J Comp Neurol 335:42–54
Steffens AB 1969 A method for frequent sampling blood and continuous infusion of fluids in the rat without disturbing the animals. Physiol Behav 4:833–836
Rethelyi M 1975 Neurons in the subependymal layer of the rat median eminence. Neuroendocrinology 17:330–339
Hahn TM, Breininger JF, Baskin DG, Schwartz MW 1998 Coexpression of Agrp and NPY in fasting-activated hypothalamic neurons. Nat Neurosci 1:271–272
Lechan RM, Nestler JL, Jacobson S 1982 The tuberoinfundibular system of the rat as demonstrated by immunohistochemical localization of retrogradely transported wheat germ agglutinin (WGA) from the median eminence. Brain Res 245:1–15
Wiegand SJ, Price JL 1980 Cells of origin of the afferent fibers to the median eminence in the rat. J Comp Neurol 192:1–19
Zhang LC, Zeng YM, Ting J, Cao JP, Wang MS 2003 The distributions and signaling directions of the cerebrospinal fluid contacting neurons in the parenchyma of a rat brain. Brain Res 989:1–8
Baker RA, Herkenham M 1995 Arcuate nucleus neurons that project to the hypothalamic paraventricular nucleus: neuropeptidergic identity and consequences of adrenalectomy on mRNA levels in the rat. J Comp Neurol 358:518–530
Bell ME, Bhatnagar S, Akana SF, Choi S, Dallman MF 2000 Disruption of arcuate/paraventricular nucleus connections changes body energy balance and response to acute stress. J Neurosci 20:6707–6713
Watts AG, Swanson LW 1987 Efferent projections of the suprachiasmatic nucleus. II. Studies using retrograde transport of fluorescent dyes and simultaneous peptide immunohistochemistry in the rat. J Comp Neurol 258:230–252
Chen S, Aston-Jones G 1998 Axonal collateral-collateral transport of tract tracers in brain neurons: false anterograde labelling and useful tool. Neuroscience 82:1151–1163
Buijs RM, Chun SJ, Niijima A, Romijn HJ, Nagai K 2001 Parasympathetic and sympathetic control of the pancreas: a role for the suprachiasmatic nucleus and other hypothalamic centers that are involved in the regulation of food intake. J Comp Neurol 431:405–423
Luppi PH, Fort P, Jouvet M 1990 Iontophoretic application of unconjugated cholera toxin B subunit (CTb) combined with immunohistochemistry of neurochemical substances: a method for transmitter identification of retrogradely labeled neurons. Brain Res 534:209–224
Aston-Jones G, Chen S, Zhu Y, Oshinsky ML 2001 A neural circuit for circadian regulation of arousal. Nat Neurosci 4:732–738
Gross PM 1992 Circumventricular organ capillaries. Prog Brain Res 91:219–233
Cottrell GT, Ferguson AV 2004 Sensory circumventricular organs: central roles in integrated autonomic regulation. Regul Pept 117:11–23
la Fleur SE, Kalsbeek A, Wortel J, Fekkes ML, Buijs RM 2001 A daily rhythm in glucose tolerance: a role for the suprachiasmatic nucleus. Diabetes 50:1237–1243
Saeb-Parsy K, Lombardelli S, Khan FZ, McDowall K, Au-Yong IT, Dyball RE 2000 Neural connections of hypothalamic neuroendocrine nuclei in the rat. J Neuroendocrinol 12:635–648
Meister B, Ceccatelli S, Hokfelt T, Anden NE, Anden M, Theodorsson E 1989 Neurotransmitters, neuropeptides and binding sites in the rat mediobasal hypothalamus: effects of monosodium glutamate (MSG) lesions. Exp Brain Res 76:343–368
Mistlberger RE, Antle MC 1999 Neonatal monosodium glutamate alters circadian organization of feeding, food anticipatory activity and photic masking in the rat. Brain Res 842:73–83
Schoelch C, Hubschle T, Schmidt I, Nuesslein-Hildesheim B 2002 MSG lesions decrease body mass of suckling-age rats by attenuating circadian decreases of energy expenditure. Am J Physiol 283:E604–E611
Horvath TL 1997 Suprachiasmatic efferents avoid phenestrated capillaries but innervate neuroendocrine cells, including those producing dopamine. Endocrinology 138:1312–1320
Saeb-Parsy K, Dyball RE 2003 Responses of cells in the rat suprachiasmatic nucleus in vivo to stimulation of afferent pathways are different at different times of the light/dark cycle. J Neuroendocrinol 15:895–903
Jamali KA, Tramu G 1999 Control of rat hypothalamic pro-opiomelanocortin neurons by a circadian clock that is entrained by the daily light-off signal. Neuroscience 93:1051–1061
Gerhold LM, Sellix MT, Freeman ME 2002 Antagonism of vasoactive intestinal peptide mRNA in the suprachiasmatic nucleus disrupts the rhythm of FRAs expression in neuroendocrine dopaminergic neurons. J Comp Neurol 450:135–143
Lind RW, Van Hoesen GW, Johnson AK 1982 An HRP study of the connections of the subfornical organ of the rat. J Comp Neurol 210:265–277
Trudel E, Bourque CW 2003 A rat brain slice preserving synaptic connections between neurons of the suprachiasmatic nucleus, organum vasculosum lamina terminalis and supraoptic nucleus. J Neurosci Methods 128:67–77
Cheng MY, Bullock CM, Li C, Lee AG, Bermak JC, Belluzzi J, Weaver DR, Leslie FM, Zhou QY 2002 Prokineticin 2 transmits the behavioural circadian rhythm of the suprachiasmatic nucleus. Nature 417:405–410
Cottrell GT, Zhou QY, Ferguson AV 2004 Prokineticin 2 modulates the excitability of subfornical organ neurons. J Neurosci 24:2375–2379
Horikawa K, Yokota S, Fuji K, Akiyama M, Moriya T, Okamura H, Shibata S 2000 Nonphotic entrainment by 5-HT1A/7 receptor agonists accompanied by reduced Per1 and Per2 mRNA levels in the suprachiasmatic nuclei. J Neurosci 20:5867–5873
Maywood ES, Mrosovsky N, Field MD, Hastings MH 1999 Rapid down-regulation of mammalian period genes during behavioral resetting of the circadian clock. Proc Natl Acad Sci USA 96:15211–15216
Card JP, Moore RY 1989 Organization of lateral geniculate-hypothalamic connections in the rat. J Comp Neurol 284:135–147
Meyer-Bernstein EL, Morin LP 1996 Differential serotonergic innervation of the suprachiasmatic nucleus and the intergeniculate leaflet and its role in circadian rhythm modulation. J Neurosci 16:2097–2111
Bradbury MJ, Dement WC, Edgar DM 1997 Serotonin-containing fibers in the suprachiasmatic hypothalamus attenuate light-induced phase delays in mice. Brain Res 768:125–134
Carlini VP, Monzon ME, Varas MM, Cragnolini AB, Schioth HB, Scimonelli TN, de Barioglio SR 2002 Ghrelin increases anxiety-like behavior and memory retention in rats. Biochem Biophys Res Commun 299:739–743
Wellman PJ, Davis KW, Nation JR 2005 Augmentation of cocaine hyperactivity in rats by systemic ghrelin. Regul Pept 125:151–154
Kalsbeek A, van Heerikhuize JJ, Wortel J, Buijs RM 1998 Restricted daytime feeding modifies suprachiasmatic nucleus vasopressin release in rats. J Biol Rhythms 13:18–29
Elmquist JK, Scammell TE, Jacobson CD, Saper CB 1996 Distribution of Fos-like immunoreactivity in the rat brain following intravenous lipopolysaccharide administration. J Comp Neurol 371:85–103
Ericsson A, Kovacs KJ, Sawchenko PE 1994 A functional anatomical analysis of central pathways subserving the effects of interleukin-1 on stress-related neuroendocrine neurons. J Neurosci 14:897–913
Kelly JF, Elias CF, Lee CE, Ahima RS, Seeley RJ, Bjorbaek C, Oka T, Saper CB, Flier JS, Elmquist JK 2004 Ciliary neurotrophic factor and leptin induce distinct patterns of immediate early gene expression in the brain. Diabetes 53:911–920
Hart BL 1988 Biological basis of the behavior of sick animals. Neurosci Biobehav Rev 12:123–137
Zheng H, Corkern M, Stoyanova I, Patterson LM, Tian R, Berthoud HR 2003 Peptides that regulate food intake: appetite-inducing accumbens manipulation activates hypothalamic orexin neurons and inhibits POMC neurons. Am J Physiol 284:R1436–R1444
Ricardo JA, Koh ET 1978 Anatomical evidence of direct projections from the nucleus of the solitary tract to the hypothalamus, amygdala, and other forebrain structures in the rat. Brain Res 153:1–26
Berthoud HR, Patterson LM, Neumann F, Neuhuber WL 1997 Distribution and structure of vagal afferent intraganglionic laminar endings (IGLEs) in the rat gastrointestinal tract. Anat Embryol 195:183–191
Schemann M, Grundy D 1992 Electrophysiological identification of vagally innervated enteric neurons in guinea pig stomach. Am J Physiol 263:G709–G718
Qin LQ, Li J, Wang Y, Wang J, Xu JY, Kaneko T 2003 The effects of nocturnal life on endocrine circadian patterns in healthy adults. Life Sci 73:2467–2475
Kreier F, Yilmaz A, Kalsbeek A, Romijn JA, Sauerwein HP, Fliers E, Buijs RM 2003 Hypothesis: shifting the equilibrium from activity to food leads to autonomic unbalance and the metabolic syndrome. Diabetes 52:2652–2656
Dhillo WS, Small CJ, Stanley SA, Jethwa PH, Seal LJ, Murphy KG, Ghatei MA, Bloom SR 2002 Hypothalamic interactions between neuropeptide Y, agouti-related protein, cocaine- and amphetamine-regulated transcript and -melanocyte-stimulating hormone in vitro in male rats. J Neuroendocrinol 14:725–730(Chun-Xia Yi, Jan van der Vliet, Jiapei D)
Tongji Medical College of Huazhong University of Science and Technology (J.D., G.Y., L.R.), 430030 Wuhan, Peoples Republic of China
Abstract
The arcuate nucleus (ARC) is crucial for the maintenance of energy homeostasis as an integrator of long- and short-term hunger and satiety signals. The expression of receptors for metabolic hormones, such as insulin, leptin, and ghrelin, allows ARC to sense information from the periphery and signal it to the central nervous system. The ventromedial ARC (vmARC) mainly comprises orexigenic neuropeptide agouti-related peptide and neuropeptide Y neurons, which are sensitive to circulating signals. To investigate neural connections of vmARC within the central nervous system, we injected the neuronal tracer cholera toxin B into vmARC. Due to variation of injection sites, tracer was also injected into the subependymal layer of the median eminence (seME), which showed similar projection patterns as the vmARC. We propose that the vmARC forms a complex with the seME, their reciprocal connections with viscerosensory areas in brain stem, and other circumventricular organs, suggesting the exchange of metabolic and circulating information. For the first time, the vmARC-seME was shown to have reciprocal interaction with the suprachiasmatic nucleus (SCN). Activation of vmARC neurons by systemic administration of the ghrelin mimetic GH-releasing peptide-6 combined with SCN tracing showed vmARC neurons to transmit feeding related signals to the SCN. The functionality of this pathway was demonstrated by systemic injection of GH-releasing peptide-6, which induced Fos in the vmARC and resulted in a reduction of about 40% of early daytime Fos immunoreactivity in the SCN. This observation suggests an anatomical and functional pathway for peripheral hormonal feedback to the hypothalamus, which may serve to modulate the activity of the SCN.
Introduction
ALTHOUGH MOST RESEARCH on nonphotic input to the suprachiasmatic nucleus (SCN) has focused on the intergeniculate leaflet and raphe nuclei, signals from other parts of the brain are also possible. The secretion of many hormones is synchronized by the SCN, and feedback from peripheral hormones to the hypothalamus could thus also influence this central clock and adjust its output. Although many metabolic active peripheral hormones/substances have transport systems to pass the blood-brain barrier, there are selected areas in the central nervous system (CNS) with special sensory functions for circulating substances, such as the arcuate nucleus (ARC). Several studies analyzing the possible ability of the ARC to sense blood-borne substances indicate that the ventromedial ARC (vmARC) may sense signals directly from the general circulation via the median eminence (ME) (1, 2, 3) and thus functions as an important sensory "window" for blood-born substances to the brain (4, 5, 6). This is firmly supported by the presence of receptors with high expression levels for nearly all known circulating metabolically active hormones, such as leptin, insulin, glucocorticoid, and ghrelin in the vmARC (7, 8, 9, 10). These hormones and also glucose are able to modulate the electrical activity of vmARC neurons (11, 12, 13) and thus affect the output of the vmARC in energy homeostasis. This special role of the vmARC in integrating hormonal signaling was confirmed by the fact that leptin signaling to the CNS crucially depends on normal ARC functioning (14, 15).
The anatomical mapping of ARC in the CNS is still updated with a variety of methods, such as transsynaptic (16) or monosynaptic tracing studies (14, 17). Because a specific vmARC tracing has not yet been performed, a systematic retrograde and anterograde tracing study is necessary to understand the anatomical relationship of the vmARC within the CNS.
We therefore studied the connections of vmARC with the CNS and used injection of fluorophore-conjugated cholera toxin B (CTB; CTB-Alexa Fluor 555) as retrograde and anterograde tracers into this area. The specificity of the projections of the vmARC to the SCN and nucleus of the solitary tract (NTS) were confirmed by control tracing from SCN and NTS, respectively.
The functional interaction between vmARC and SCN was demonstrated by the use of the natural leptin antagonist, the gut-brain peptide ghrelin (18), which is the only metabolic molecule found to date that is able to activate vmARC neurons by specifically targeting agouti-related peptide (AGRP)/neuropeptide Y (NPY) neurons and that stimulates food ingestion (19). With systemic administration of the ghrelin mimetic, GH-releasing peptide-6 (GHRP-6), combined with SCN tracing and AGRP/Fos staining, the present data showed that GHRP-6 activated vmARC neurons projecting to the SCN. Finally, the influence of systemic GHRP-6 stimulation on diurnal Fos immunoreactivity in the SCN was investigated.
Materials and Methods
CTB tracing
All the following experiments were conducted under the approval of the animal care committee of the Royal Netherlands Academy of Sciences. Male Wistar rats, weighing 300 ± 10 g (Harlan Nederland, Horst, The Netherlands), were housed at room temperature with a 12-h light, 12-h dark schedule (lights on at 0700 h). For tract tracing, rats were anesthetized [0.8 ml/kg Hypnorm (Janssen, High Wycombe, Buckinghamshire, UK), im, and 0.4 ml/kg Dormicum (Roche, Almere, The Netherlands), sc], mounted with their heads in a standard stereotaxic apparatus, tooth bar set at –3.4 mm, and received tracer injection. Because, in contrast to CTB, fluorophore-labeled CTB cannot be applied by iontophoresis, we used pressure injection. Thus, fluorophore-conjugated CTB (with Alexa Fluor 555, Molecular Probes, Eugene, OR; CTB will be used in the text as the abbreviation of CTB-Alexa Fluor 555 for convenience) was used for purpose of combination of neurotransmitter and tracer detection in target area. Finally, in 30 cases CTB injections were aimed at the vmARC. To confirm the reciprocity of connection, injections with CTB were also aimed at the SCN (n = 30) or NTS (n = 10); all coordinates for these regions were adapted from the atlas of Paxinos (20).
CTB was injected with pressure (10 mbar, 5 sec) using a glass pipette with a tip of maximally 50 μm in diameter to minimize damage of passing fibers; the glass pipette was filled with an injection volume of 50–100 nl. The coordinates for the vmARC injection were 3.3 mm caudal to bregma, 0.4 mm lateral to the midline, and 10.4 mm below the dura. The coordinates for SCN injection followed those reported by Buijs et al. (21); for NTS injection, the coordinate was 0.4 lateral to the midline, 13.8 mm caudal to bregma, and 5 mm ventral from dura. After the tracer was injected by pressure injection, the pipettes were fixed with dental cement to the skull and left in the brain to minimize leakage from the tract until the animals were killed. This procedure did not lead to visible extra discomfort of the animals, whereas the reduction in leakage along the pipette track is substantial. Fifteen rats of the SCN tracing were also combined with the placement of a jugular venous catheter for the iv injection of GHRP-6. See below for details on this operation protocol.
GHRP-6 iv infusion
For the iv infusion of GHRP-6, a silicone catheter was implanted in the right jugular vein according to the method of Steffens (22). The operation of the first group of rats (n = 15) was combined with SCN tracing, such that both operations were carried out at same time. The second group of rats (n = 8) received only catheter implantation for iv infusion of GHRP-6 (n = 4) or vehicle (Ringer solution; n = 4) to examine Fos immunoreactivity in the vmARC and SCN. After surgery, all rats were given 10 d to recover. After the operation, the animals were handled every day and connected to the jugular catheter to familiarize the rats with the experimental procedures. On d 9, the rats were connected to a drug administration catheter for 24 h, which was kept out of reach of the rats by means of counterbalanced beam. This allowed all manipulations to be carried out outside the cages without handling the rat. All experiments were performed in the rat home cage.
The drug administration catheter was connected to a syringe containing vehicle with or without GHRP-6. Infusion into the catheter was initiated at Zeitgeber time 1 (ZT1), with ZT0 being defined as the onset of the light period. GHRP-6 was used at a dose of 5 μg/rat dissolved in 0.4 ml Ringer solution, and 0.4 ml Ringer solution was used as the vehicle control infusion. The infusion speed was kept at 0.1 ml/min. All rats were killed by perfusion 100 min after the infusion.
Food deprivation
We found that neurons in the subependymal layer of ME (seME) have a similar projection pattern as those in the vmARC. Previous studies supposed them to be an extension of vmARC (23); therefore, we examined whether seME neurons show a comparable increase in the expression of NPY/AGRP after food deprivation as observed in vmARC neurons (24). Food, but not water, was removed at ZT1 from four intact rats for 48 h; subsequently, animals were killed by perfusion.
Immunocytochemical staining
All rats were deeply anesthetized with a lethal dose of sodium pentobarbital and perfused with saline, followed by a solution of 4% paraformaldehyde in 0.1 M PBS (pH 7.4) at 4 C. The brains were removed and kept in fixative at 4 C for overnight postfixation, equilibrated 48 h with 30% sucrose in 0.1 M Tris-buffered saline (TBS; pH 7.2). Brains were coronally cut in a cryostat into 30-μm sections; sections used for immunolabeling were collected and rinsed in 0.1 M TBS.
With ARC, ME, SCN, and NTS tracing, whole brain sections were incubated with rabbit anti-CTB primary antibody (Sigma-Aldrich Corp., St. Louis, MO) at a 1:100,000 dilution. Sections of fasted rats were incubated with rabbit anti-AGRP (1:1500; Phoenix Pharmaceuticals, Belmont, CA) or NPY (1:2000; Niepke 091188, Netherlands Institute for Brain Research) primary antibody. Sections of second group of GHRP-6 infusion rats were incubated with goat anti-Fos (1:1500; Santa Cruz Biotechnology, Inc., Santa Cruz, CA). After all these primary antibody incubations overnight at 4 C, sections were rinsed and incubated in biotinylated secondary antibody (goat antirabbit or horse antigoat IgG) for 1 h, then rinsed and incubated in avidin-biotin complex (Vector Laboratories, Inc., Burlingame, CA) for 1 h, and the reaction product was visualized by incubation in 1% diaminobenzidine (0.05% nickel ammonium sulfate was added to the diaminobenzidine solution to darken the reaction product) with 0.01% hydrogen peroxide for 5–7 min. Sections were mounted on gelatin-coated glass slides, dried, run through ethanol and xylene, and covered for observation by light microscope.
Confocal microscopy
To characterize the functionality of neurons in the vmARC that project to the SCN, we used triple-labeling immunofluorescence to show the colocalization of GHRP-6-induced Fos with AGRP in retrogradely CTB-labeled neurons in vmARC from SCN tracing. Sections were incubated overnight at 4 C with goat anti-Fos and rabbit anti-AGRP primary antibody, rinsed in 0.1 M TBS, incubated 1 h in biotinylated horse antigoat IgG, rinsed, incubated with streptoavidin-Cy5 and donkey antirabbit-Cy2 for 1 h, rinsed, mounted on gelatin-coated glass slides, dried, covered with glycerol in 0.1 M PBS (pH 9.0), and observed by confocal laser scanning microscopy.
Analysis of Fos immunoreactivity
The effect of GHRP-6 treatment on Fos immunoreactivity in the SCN was determined in four GHRP-6-infused or four Ringer solution-infused rats. From SCN sections, tiled images were captured by a computerized image analysis system consisting of an Axioskop 9811+ Sony XC77 black and white video camera (Sony Corp., Tokyo, Japan). In these images, the right side of middle portion of the SCN was manually outlined. The Fos-positive nuclear profiles were automatically (no user interaction) segmented by a dedicated macro written within the ImagePro programming environment. For each rat, three sections were measured 90 μm apart (from bregma –1.20 to –1.40 mm); the mean number of Fos-positive nuclear profiles from these three sections was calculated. All values are expressed as the mean ± SEM, and data were analyzed using one-way ANOVA. Statistical significance was set at P < 0.01.
Results
CTB injections into the vmARC and seME
Of all the animals injected, two rats had a successful injection within the middle part of the vmARC along its rostrocaudal extension, and two rats had a successful injection within the seME. Due to the difficulties of accurate vmARC injection, we only focused on the middle part (in the anterior posterior direction) of the vmARC, which guaranteed maximal vmARC space for tracing. Misplaced injections were mainly located in a different coronal plane, but usually at the same distance from the bregma. The misplaced injections in the parenchymal area, including the external zone of the ME and the dorsal or ventrolateral ARC, served as controls for the specificity of the projections of the vmARC and seME. Misplacement and injection in the third ventricle served as control for leakage of tracer into cerebrospinal fluid, which is almost inevitable. The injections in which the small tip of glass pipette was blocked by brain tissue and the tracer could not be pushed out by pressure were used as specificity controls of the used antibodies. In none of these latter cases was any staining observed, which shows that all the staining observed was due to CTB.
Tracer injections into the vmARC resulted in the spread of CTB not only limited to the ventromedial part of the nucleus, but expanding also into the ipsilateral and contralateral subependymal layers of the ME (Fig. 1A). After seME injection, the CTB distribution covered the dorsal aspect of the ME, between the ependymal layer and the fibrous layer (Fig. 1D). AGRP- and NPY-containing cell bodies can more easily be detected in the seME as well as in the vmARC after 48 h of fasting (Fig. 1, C and F), whereas with a normal feeding schedule, we only observed AGRP or NPY fibers and terminals in these two areas. This observation together with the distribution pattern of the tracer could support the previous suggestion based on development that seME neurons are an extension of the ARC (23). Consequently, Table 1 shows retrograde and anterograde labeling in different areas of the CNS with CTB injections into the vmARC and seME.
Control injections
Injections in the ventrolateral ARC (Fig. 1B), dorsal ARC, external zone of the ME (Fig. 1E), and third ventricle were used as controls for specificity of the projections of the vmARC and seME and leakage of the tracer. Three rats had tracing into the external zone of the ME, three rats had ventricle injection only, and two rats received ventrolateral ARC and dorsal ARC tracing, respectively.
The labeling pattern of external zone tracing was consistent with those from Lechan et al. (25) and Wiegand and Price (26). In contrast with the seME injection, it revealed mainly retrogradely labeled neurons in the parvocellular paraventricular nucleus (PVN; Fig. 3F) and several magnocellular neurons in PVN and supraoptic nucleus (SON; Fig. 3I) as well as the periventricular nucleus and magnocellular dorsal ARC. The absence of detectable anterograde tracing confirms the external zone of ME as a pure neurosecretory pathway of the hypothalamus. Three control injections into the third ventricle resulted in labeling of the ependymal plexus and tanycytes or cerebrospinal fluid contacting neurons along the third and lateral ventricles, some neurons in the dorsal raphe nucleus were also labeled. This pattern was consistent with ventricle tracing experiments aiming at detecting cerebrospinal fluid contacting neurons in the CNS by CTB-horseradish peroxidase injection (27). Injections into the ventrolateral or dorsal part of the ARC had projection patterns distinct from that of injection into the vmARC, e.g. with ventrolateral ARC injection even when very close to the vmARC, we only observed some neurons present in the dorsal and medial parts of the SCN (Fig. 3G) and no projections to the SON or NTS.
Retrograde and anterograde labeling in the CNS with CTB injection into the vmARC and seME
The injection cases R122 and R173 had their injection tips localized within the vmARC, whereas R121 and R144 had their injection tips in the seME. Cases R122–R173 and R121–R144 showed some similar pattern of retrograde and anterograde labeling in the CNS, although with several differences exemplified by the presence of retrogradely filled neurons and anterogradely filled fibers in a number of additional areas after injection in the vmARC as compared with injection in the seME (see Table 1). The projections from the seME area as compared to the projections from the vmARC are less dense than vmARC.
Areas with reciprocal interaction with vmARC and seME
CTB tracer injection into the vmARC and seME revealed several sites that have extensive reciprocal interaction with this area. In the anterior part of the CNS, reciprocal connections with this area arise from the anteroventral PVN, ventromedial preoptic nucleus, and medial preoptic nucleus (lateral part), with labeled cells and fibers scattered throughout these areas, and the labeled neurons were mainly small in size. Both lateral septum and accumbens nucleus connect with vmARC reciprocally, but labeled fibers were not visible after seME tracing.
We observed that all parts of the SCN show labeled neurons, with the highest density in the medial and dorsal parts. Moreover, abundant labeling of fibers and terminals were found in the dorsal and ventral parts. A few labeled neurons and fibers were also observed in the area around the SCN (Fig. 2). From rostral to caudal, the whole extension of the periventricular hypothalamic nucleus is labeled by scattered neurons and fibers.
Consistent with previously published data about the projection from the ARC to the PVN (28, 29), with vmARC and seME tracing, the triangular shape of the PVN was full of labeled fibers and terminals (Fig. 3, A and B); in contrast with the dense labeling of fibers and terminals, only few PVN neurons were labeled. Although when the injection was in the external zone of the ME, PVN labeling was characterized by the presence of a large number of neurons that by location and appearance can easily be recognized as neuroendocrine (Fig. 3F). Also in the SON the contrast is obvious; vmARC and seME injections result in dense innervations in SON (Fig. 3H), whereas injection in the external zone only shows retrogradely filled neurons (Fig. 3I). After vmARC injection, a considerable number of fibers and small neurons in the sub-PVN (30) were also labeled (Fig. 3A). More laterally, the reciprocal connection also only exists between lateral hypothalamus and vmARC, but not with the seME.
Within the whole ARC, many anterogradely labeled fibers as well as a few retrogradely labeled neurons were present in the dorsal and ventrolateral parts of the ARC near the border that separates the ARC and ventromedial hypothalamic nucleus (VMH; Fig. 1). This labeling pattern of ARC was consistent from rostral to caudal.
Another important target of the vmARC and seME is the dorsomedial hypothalamic nucleus (DMH). A large number of labeled neurons, fibers, and terminals visualized throughout the DMH demonstrate the reciprocal projection between the vmARC and DMH (Fig. 3D). Interestingly, the vmARC and seME tracing revealed a very weak anatomical relationship with the VMH, with just a few neurons and fibers being labeled. In the posterior hypothalamus, numerous labeled neurons were present in the ventral premamillary nucleus, whereas fibers were relatively sparse. Of the tuberomamillary nucleus, the dorsal part received a denser innervation from the vmARC and seME than the ventral area.
In addition, neurons and some axonal terminals were observed in the other three circumventricular organs (CVOs) after vmARC and seME injection. In the subfornical organ (SFO), neurons and fibers were spread over the vascular area (Fig. 4A). In the organum vasculosum of the lamina terminals (OVLT), abundant labeling was present in neurons and fibers bilaterally located in the vascular zone and perinucleus area. A strong labeling of fibers, as in the ependymal plexus, was seen along the wall of the ventricle (Fig. 4B). Similarly, in the area postrema (AP), several fibers and neurons were observed in the vascular area (Fig. 4C). At the border area between the AP and NTS, we visualized strongly labeled fibers, probably derived from the ependymal plexus that is often labeled after injection in the seME (Fig. 4C), this labeling pattern in AP as well as in OVLT, however, was also seen with the third ventricle control injection, whereas with control incubation of anti-CTB antibody with sections from nontracing, we did not see any such staining, and we also have not found such a labeling pattern with lateral ARC and SCN CTB tracing animals.
In the brain stem, almost all subdivisions of the parabrachial nucleus (PB) contain labeled neurons and terminals from the most rostral to the caudal part of the nucleus (Fig. 4, E and F). In addition, scattered neurons and terminals were seen in the periaqueductal gray, locus coeruleus, Barrington’s nucleus, and prepositus nucleus.
Similar to the PB, another visceral integration nucleus, the NTS, contained several labeled neurons and fibers from the caudal to the middle part. Consistent with a previous study, after vmARC injection, we observed a dense labeling of fibers in a limited area at ipsilateral part in the caudal NTS (Fig. 4, H and I); in view of the paucity of retrogradely labeled neurons, we consider these fibers be largely due to anterograde transport and not derived from retrogradely labeled neurons as has been suggested to occur with CTB tracing (31). Interestingly, some small labeled neurons were also found in the dorsal motor nucleus of the vagus (DMV) (Fig. 4H). These finding corroborate our previous observation that the area of the DMV incorporates neurons without an autonomic motor function (32).
Areas with only input to or output from the vmARC and seME
A few regions seem to only provide input to the vmARC and seME, as demonstrated by the presence of merely retrogradely labeled cell bodies and no clear staining of fibers. Because other brain regions at the same level do show labeled fibers, this does not seem to be a technical limitation of the anterograde tracing. Only retrogradely filled neurons were detected in the dorsal endopiriform nucleus, claustrum, bed nucleus of the stria terminalis, and posterior part of the PVN. The whole extension of most of the medial amygdaloid complex was also strongly retrogradely labeled (Fig. 4D).
With vmARC and seME injections, a dense innervation of the SON, retrochiasmatic area, and tuber cinereum area, but without retrogradely labeled cell bodies, indicates that these areas only get input from the vmARC and seME. We also found that the vmARC and seME tracing resulted in projections to the intergeniculate leaflet (Fig. 4G).
Retrograde tracing from SCN and NTS
To provide additional evidence for the projections of the vmARC and seME to the SCN and NTS and to ensure that the projections are not due to uptake from passing fibers or to anterograde labeling by retrogradely labeled neurons (33, 34), CTB was injected into the SCN and NTS.
Injections of CTB into both the SCN and NTS revealed not only neurons in the vmARC (Fig. 5, A, B, and D), but also retrogradely labeled neurons in the subependymal layer of the ME with a location just under the ependymal lining of the third ventricle (Fig. 5C). We also considered the abundant presence of fibers as evidence for the reciprocity of the connection of the vmARC and seME with the SCN and NTS. The labeled neurons in the seME area were different from glial cell. Firstly, the location of these subependymal neurons is just under and very close to the ependymal layer lining the bottom of the third ventricle. In Fig. 5C we can see the difference with ependymal cells that were not labeled by CTB. Secondly, the orientation of the oval or fusiform-shaped labeled neurons was parallel with the border of the third ventricle, whereas glia cells have many more processes and may extend to every orientation, including the external zone. Thirdly, the size of subependymal neurons was far larger than glial cell, either tanycytes or astrocytes.
After CTB injection into the SCN, fibers and terminals were abundant in the lateral and dorsal parts of the ARC where few retrogradely labeled neurons were present (Fig. 5, A and B). This projection pattern confirmed the vmARC and seME tracing, which resulted in labeling of neurons, fibers, and terminals in the SCN, and the ventrolateral and dorsal ARC tracing, which resulted in only labeled neurons in the SCN without fibers or terminals (Fig. 3G).
Functionality of the connections between vmARC and SCN
We addressed the possible role of vmARC neurons in the regulation of metabolism in relation to their projection to the SCN by examining the induction of Fos in AGRP neurons after systemic administration of GHRP-6 combined with SCN tracing. In addition, we observed that GHRP-6 stimulation allowed an easier detection of AGRP in vmARC neurons, which is normally difficult. Most AGRP-positive neurons also expressed Fos, and some of these Fos-positive AGRP neurons were shown to project to the SCN, as revealed by colocalization with CTB from SCN tracing (Fig. 6). Interestingly, in both ad libitum-fed and feeding-restricted rats, it was difficult to detect AGRP fibers in the SCN. We observed only very thin AGRP-positive fibers in the ventral SCN after 2 h of refeeding following 48-h fasting (data not shown), which suggests that although AGRP neurons are the target of peripheral GHRP-6 and project to the SCN, the detection of this peptide is limited, probably because of the low quantity of peptide present in these nerve terminals.
After SCN tracing, there was no obvious Fos immunoreactivity in ARC neurons or in intact rat ARC that were housed and perfused under the same conditions as SCN tracing rats, which means that SCN tract tracing has no effect on Fos immunoreactivity in the ARC. Fos immunoreactivity in vmARC induced by GHRP-6 is apparent after a 5-μg iv infusion, whereas the Ringer solution infusion had no detectable Fos in the vmARC (Fig. 7, A and B). In addition, Fos immunoreactivity was checked in the geniculate area and raphe nucleus, which are responsible for nonphotic input to the SCN, and there was no significant difference between Ringer control and 5 μg GHRP-6 administration in both nuclei.
The number of detectable Fos-positive nuclear profiles in the middle portion SCN was counted in controls as well as GHRP-6-infused animals. Negatively correlated with the same dose of 5 μg GHRP-6 action on the vmARC, the number of Fos-positive nuclei in the middle portion SCN was decreased by about 40% compared with that after Ringer solution infusion (195.33 ± 14.45 vs. 116.89 ± 4.7 in Ringer solution and 5 μg GHRP-6, respectively; P < 0.01; Fig. 7, C and D).
Discussion
In the present study the properties of CTB as anterograde and retrograde tracer were confirmed by the presence of tracer-labeled terminals and cell bodies, for example, in the NTS after injection into the vmARC, which is consistent with previous tracing studies on the ARC and NTS (17) and consequently allowed us to study the input as well as the output of one area simultaneously.
To avoid uptake of tracer by damage of fibers of passage as much as possible, glass pipettes with a tip approximately 50 μm in diameter were used for pressure injection instead of a Hamilton syringe (400 μm), which minimized the possibility of tracer uptake by passing fibers.
This is demonstrated by the fact that only injection in the external zone of the median eminence resulted in retrograde labeling of neuroendocrine neurons in the PVN and SON, which confirms previous studies (25, 26). Injections into other areas where neuroendocrine fibers pass, such as the subependymal layer-internal zone and the lateral ARC, did not result in cell body labeling. In contrast, injection into the subependymal layer-internal zone resulted only in labeling of fiber terminals in the SON and PVN. These findings suggest not only a high specificity of projections between the different layers of the median eminence, but also a minimal uptake of fibers of passage by the method used.
The present results confirm and extend previous anatomical studies of ARC and ME and provide a more comprehensive view of the organization of the ARC, including its interaction with the biological clock (Fig. 8).
Moreover, we demonstrate for the first time that both the vmARC and seME areas have a similar extensive interaction with the SCN, the sensory CVOs, and other parts of the brain. In both the vmARC and seME, the blood-brain barrier operates in a modified manner comparable with that of other CVOs that also have permeable capillary networks for penetration of circulating substances (35). Moreover, both areas not only share the same projection patterns, but also contain NPY and AGRP neurons, which respond similarly to, for instance, fasting. Considering these aspects, we propose that the vmARC-seME should be viewed as an arcuate-median eminence complex (AMC).
It is proposed that three typical sensory CVOs, SFO, OVLT, and AP, have active transport systems and receptors for blood-borne molecules and play a key role in sensing and relaying humoral information to other CNS areas (36). Our present data demonstrate that the AMC has similar specific neuronal connections to hypothalamic neuroendocrine and autonomic centers, such as the SFO, OVLT, and AP. Early studies have already suggested that the blood-brain barrier border could reside at the area between the ventromedial (or proximal) part and dorsolateral (or distal) part of the ARC, and that the perivascular spaces of vmARC could thus communicate with the neurohemal area of the ME, where fenestrated capillaries allow the passage of larger molecules (1, 2, 3). Therefore, the present finding that the SCN receives direct input from the AMC suggests a route of communication between the general circulation and the SCN. Intravenous injection of the ghrelin mimetic GHRP-6 results in Fos-positive neurons in a narrow band in the AMC. Of all tracer injections in the ARC (n = 30), many were very close to the vmARC; only four injections that were inside and covered the same area that is known to express Fos after GHRP-6 stimulation had reciprocal interaction with the SCN. This suggests that a relatively small area is involved in this exchange of information. Furthermore, injections into the SCN resulted in the combination of retrogradely labeled neurons and fibers only in this ventromedial area of the ARC, which confirms the reciprocity of this connection. In view of the (daily) time-dependent organization of metabolism, which depends so crucially on the presence of the SCN (37), the interaction of the AMC with the SCN could serve this timing purpose and provide the SCN with information about glucose concentration and the associated level of hormones.
That there are innervations from vmARC to SCN is a novel finding. However, previous electrophysiological studies on ARC (38) indicated this connection. Other proof for the functional interaction between vmARC and SCN comes from a number of studies indicating that physiological and behavioral circadian rhythms are also affected by ARC malfunction. The fact that ARC lesions by neonatal monosodium glutamate treatment, which mainly destroys NPY cells (39), is able to compromise the circadian activity while the SCN remains intact (40, 41), could be interpreted, together with the present data, to mean that the innervations from vmARC to the SCN may relay peripheral hormonal information to the SCN and may thus affect the circadian activity of the SCN. At the same time, input from the SCN to the ARC (38, 42, 43) may time the activity of the ARC, as shown by rhythmic Fos expression in the ARC (44) and the rhythmic activity of dopaminergic ARC neurons negatively driven by vasoactive intestinal peptide from the SCN (45). Actually, such an SCN-CVO reciprocal connection has already been demonstrated for the SCN-SFO and SCN-OVLT interaction (46, 47). Additional evidence for a more general interaction between SCN and CVOs, including the ARC, can be found in studies describing the action of prokineticin 2, which is synthesized in the SCN and for which receptors have been shown in the SFO and the vmARC (48, 49). Together, these studies suggest the capacity of the SCN to modulate the properties of the CVOs, probably to control or regulate the sensitivity of CVOs to peripheral signals, and the present study indicates that these signals may be relayed to the SCN as well not only via the SFO and OVLT, but also via the vmARC and seME. In view of the high sensitivity to peripheral metabolic hormones in the vmARC, the possible feedback pathway for these hormones to the SCN is proposed to be via this area.
Systemic administration of the ghrelin mimetic, GHRP-6, combined with SCN tracing showed that AGRP neurons activated by GHRP-6 project to the SCN; at the same time, the diurnal Fos immunoreactivity in the SCN was reduced by 40%, suggesting that the ghrelin mimetic, GHRP-6, could interfere with the activity of the SCN via the vmARC. The obvious decrease in Fos in the ventrolateral SCN suggests a possible regulation of the light-receiving part of the SCN. The fact that the staining of AGRP fibers in the SCN compared with the PVN is much less intense may indicate that a certain selection of neurons in vmARC projects to the SCN, or those collaterals of vmARC neurons to SCN and PVN may contain transmitters in different proportions.
It is well known that most nonphotic inputs to the SCN have the ability to reset the clock when they are provided during the day (50, 51). Nonphotic inputs to the SCN were shown to arise from the intergeniculate leaflet (52) and raphe nucleus (53). In addition, nonphotic inputs are known to inhibit photic-responsive Fos expression in the SCN at night (54). In the present study with peripheral GHRP-6 administration, there is no obvious variation in Fos immunoreactivity in the geniculate area and raphe nucleus, which indicates that these areas are not activated. Consequently, the GHRP-6 information relayed by the vmARC to the SCN, as indicated by Fos diminishment in the SCN, is another functional support for the anatomical connection between the AMC and SCN, and suggests that pharmacological effects of ghrelin may inhibit the activity of the SCN during the light period. This observation agrees with the fact that with a normal feeding schedule, ghrelin peaks after dark onset (39), suggesting another central function of ghrelin, such as interaction between locomotion and metabolism (55, 56).
Because the SCN is essential for the organization of daily metabolic activity (the daily rhythm of plasma glucose and metabolic hormones is driven by the SCN) (37), the present results provide an anatomical basis for our previous findings that metabolic cues could indeed influence circadian activity of, for example, vasopressin neurons in the SCN (57). Additional research is needed to elucidate the effect of the metabolism-related nonphotic input on SCN functionality and the possible neurotransmitters that are involved in the signaling.
Naturally the neuronal connections of AMC with other areas in the CNS may also implicate those areas as potential participants in metabolism. In this study we stress the role of interaction between AMC and CVOs; not only are the classic CVOs engaged in monitoring changes in metabolic, osmotic, ionic, and hormonal compositions, they also show neuronal activation or inhibition by cytokines (58, 59, 60). In this study we show that the SFO, OVLT, and AP have reciprocal connections with the vmARC and seME, centers for the control of food intake, which suggests a functional interaction of CVOs, for example, in cytokine activity and sickness behavior (61). This could be the anatomical basis for anorexia and weight loss during inflammations. In addition, the present finding that the amygdala complex and nucleus accumbens provide input to the AMC supports the significance of limbic structures in energy homeostasis. This possibility is supported by the fact that chemical manipulation of the nucleus accumbens induced high Fos expression in ARC NPY neurons (62).
Moreover, the AMC area has reciprocal connections with visceral sensory integration sites such as the PB and NTS, as was shown in part previously (17, 63). This indicates that the vmARC and seME may integrate hormonal information with signals from first order visceral sensory centers (64, 65). Logically, the outcome of this interaction should be signaled to many sites, including the SCN, and is necessary for the multilevel central control of visceral activities.
Our present data provide the anatomical and functional basis for the hypothesis that both humoral and neuronal metabolic information may be relayed back to the SCN via the AMC. Herein the AMC area is responsible for the first order integration of circulating signals. When the normal metabolic balance is seriously disturbed, e.g. by night-eating syndrome (66) or stress, this may influence the integrative activity of the AMC as well as the organizational activity of the SCN. Thus, a slight disturbance at the level of these essential integration sites raises a serious risk of an unbalanced energy state and possibly obesity, diabetes, or other metabolic diseases (67).
Acknowledgments
We thank Drs. Carolina Escobar and Marco van der Top for their valuable comments on the manuscript, Wilma Verweij for correction of the English, and Henk Stoffels for the art work.
Footnotes
This work was supported by a grant from the Royal Netherlands Academy of Arts and Sciences for a joint project between The Netherlands and China (03CDP014).
First Published Online September 29, 2005
Abbreviations: AGRP, Agouti-related peptide; AMC, arcuate-median eminence complex; AP, area postrema; ARC, arcuate nucleus; CNS, central nervous system; CTB, cholera toxin B; CVO, circumventricular organ; DMH, dorsomedial hypothalamic nucleus; DMV, dorsal motor nucleus of the vagus; GHRP-6, GH-releasing peptide-6; ME, median eminence; NPY, neuropeptide Y; NTS, nucleus of the solitary tract; OVLT, organum vasculosum of the lamina terminals; PB, parabrachial nucleus; PVN, paraventricular nucleus; SCN, suprachiasmatic nucleus; seME, subependymal layer of the median eminence; SFO, subfornical organ; SON, supraoptic nucleus; TBS, Tris-buffered saline; vmARC, ventromedial arcuate nucleus; VMH, ventromedial hypothalamic nucleus; ZT, Zeitgeber time.
Accepted for publication September 22, 2005.
References
Broadwell RD, Balin BJ, Salcman M, Kaplan RS 1983 Brain-blood barrier Yes and no. Proc Natl Acad Sci USA 80:7352–7356
Krisch B, Leonhardt H 1978 The functional and structural border of the neurohemal region of the median eminence. Cell Tissue Res 192:327–339
Shaver SW, Pang JJ, Wainman DS, Wall KM, Gross PM 1992 Morphology and function of capillary networks in subregions of the rat tuber cinereum. Cell Tissue Res 267:437–448
Berthoud HR 2002 Multiple neural systems controlling food intake and body weight. Neurosci Biobehav Rev 26:393–428
Horvath TL, Diano S 2004 The floating blueprint of hypothalamic feeding circuits. Nat Rev Neurosci 5:662–667
Elmquist JK, Elias CF, Saper CB 1999 From lesions to leptin: hypothalamic control of food intake and body weight. Neuron 22:221–232
Aronsson M, Fuxe K, Dong Y, Agnati LF, Okret S, Gustafsson JA 1988 Localization of glucocorticoid receptor mRNA in the male rat brain by in situ hybridization. Proc Natl Acad Sci USA 85:9331–9335
Unger J, McNeill TH, Moxley III RT, White M, Moss A, Livingston JN 1989 Distribution of insulin receptor-like immunoreactivity in the rat forebrain. Neuroscience 31:143–157
Schwartz MW, Seeley RJ, Campfield LA, Burn P, Baskin DG 1996 Identification of targets of leptin action in rat hypothalamus. J Clin Invest 98:1101–1106
Guan XM, Yu H, Palyha OC, McKee KK, Feighner SD, Sirinathsinghji DJ, Smith RG, Van der Ploeg LH, Howard AD 1997 Distribution of mRNA encoding the growth hormone secretagogue receptor in brain and peripheral tissues. Brain Res Mol Brain Res 48:23–29
Pinto S, Roseberry AG, Liu H, Diano S, Shanabrough M, Cai X, Friedman JM, Horvath TL 2004 Rapid rewiring of arcuate nucleus feeding circuits by leptin. Science 304:110–115
van den Top M, Lee K, Whyment AD, Blanks AM, Spanswick D 2004 Orexigen-sensitive NPY/AgRP pacemaker neurons in the hypothalamic arcuate nucleus. Nat Neurosci 7:493–494
Spanswick D, Smith MA, Mirshamsi S, Routh VH, Ashford ML 2000 Insulin activates ATP-sensitive K+ channels in hypothalamic neurons of lean, but not obese rats. Nat Neurosci 3:757–758
Bouret SG, Draper SJ, Simerly RB 2004 Formation of projection pathways from the arcuate nucleus of the hypothalamus to hypothalamic regions implicated in the neural control of feeding behavior in mice. J Neurosci 24:2797–2805
Dawson R, Pelleymounter MA, Millard WJ, Liu S, Eppler B 1997 Attenuation of leptin-mediated effects by monosodium glutamate-induced arcuate nucleus damage. Am J Physiol 273:E202–E206
DeFalco J, Tomishima M, Liu H, Zhao C, Cai X, Marth JD, Enquist L, Friedman JM 2001 Virus-assisted mapping of neural inputs to a feeding center in the hypothalamus. Science 291:2608–2613
Sim LJ, Joseph SA 1991 Arcuate nucleus projections to brainstem regions which modulate nociception. J Chem Neuroanat 4:97–109
Kojima M, Hosoda H, Date Y, Nakazato M, Matsuo H, Kangawa K 1999 Ghrelin is a growth-hormone-releasing acylated peptide from stomach. Nature 402:656–660
Nakazato M, Murakami N, Date Y, Kojima M, Matsuo H, Kangawa K, Matsukura S 2001 A role for ghrelin in the central regulation of feeding. Nature 409:194–198
Paxinos G, Watson C 1998 The rat brain in stereotaxic coordinates. San Diego: Academic Press
Buijs RM, Markman M, Nunes-Cardoso B, Hou YX, Shinn S 1993 Projections of the suprachiasmatic nucleus to stress-related areas in the rat hypothalamus: a light and electron microscopic study. J Comp Neurol 335:42–54
Steffens AB 1969 A method for frequent sampling blood and continuous infusion of fluids in the rat without disturbing the animals. Physiol Behav 4:833–836
Rethelyi M 1975 Neurons in the subependymal layer of the rat median eminence. Neuroendocrinology 17:330–339
Hahn TM, Breininger JF, Baskin DG, Schwartz MW 1998 Coexpression of Agrp and NPY in fasting-activated hypothalamic neurons. Nat Neurosci 1:271–272
Lechan RM, Nestler JL, Jacobson S 1982 The tuberoinfundibular system of the rat as demonstrated by immunohistochemical localization of retrogradely transported wheat germ agglutinin (WGA) from the median eminence. Brain Res 245:1–15
Wiegand SJ, Price JL 1980 Cells of origin of the afferent fibers to the median eminence in the rat. J Comp Neurol 192:1–19
Zhang LC, Zeng YM, Ting J, Cao JP, Wang MS 2003 The distributions and signaling directions of the cerebrospinal fluid contacting neurons in the parenchyma of a rat brain. Brain Res 989:1–8
Baker RA, Herkenham M 1995 Arcuate nucleus neurons that project to the hypothalamic paraventricular nucleus: neuropeptidergic identity and consequences of adrenalectomy on mRNA levels in the rat. J Comp Neurol 358:518–530
Bell ME, Bhatnagar S, Akana SF, Choi S, Dallman MF 2000 Disruption of arcuate/paraventricular nucleus connections changes body energy balance and response to acute stress. J Neurosci 20:6707–6713
Watts AG, Swanson LW 1987 Efferent projections of the suprachiasmatic nucleus. II. Studies using retrograde transport of fluorescent dyes and simultaneous peptide immunohistochemistry in the rat. J Comp Neurol 258:230–252
Chen S, Aston-Jones G 1998 Axonal collateral-collateral transport of tract tracers in brain neurons: false anterograde labelling and useful tool. Neuroscience 82:1151–1163
Buijs RM, Chun SJ, Niijima A, Romijn HJ, Nagai K 2001 Parasympathetic and sympathetic control of the pancreas: a role for the suprachiasmatic nucleus and other hypothalamic centers that are involved in the regulation of food intake. J Comp Neurol 431:405–423
Luppi PH, Fort P, Jouvet M 1990 Iontophoretic application of unconjugated cholera toxin B subunit (CTb) combined with immunohistochemistry of neurochemical substances: a method for transmitter identification of retrogradely labeled neurons. Brain Res 534:209–224
Aston-Jones G, Chen S, Zhu Y, Oshinsky ML 2001 A neural circuit for circadian regulation of arousal. Nat Neurosci 4:732–738
Gross PM 1992 Circumventricular organ capillaries. Prog Brain Res 91:219–233
Cottrell GT, Ferguson AV 2004 Sensory circumventricular organs: central roles in integrated autonomic regulation. Regul Pept 117:11–23
la Fleur SE, Kalsbeek A, Wortel J, Fekkes ML, Buijs RM 2001 A daily rhythm in glucose tolerance: a role for the suprachiasmatic nucleus. Diabetes 50:1237–1243
Saeb-Parsy K, Lombardelli S, Khan FZ, McDowall K, Au-Yong IT, Dyball RE 2000 Neural connections of hypothalamic neuroendocrine nuclei in the rat. J Neuroendocrinol 12:635–648
Meister B, Ceccatelli S, Hokfelt T, Anden NE, Anden M, Theodorsson E 1989 Neurotransmitters, neuropeptides and binding sites in the rat mediobasal hypothalamus: effects of monosodium glutamate (MSG) lesions. Exp Brain Res 76:343–368
Mistlberger RE, Antle MC 1999 Neonatal monosodium glutamate alters circadian organization of feeding, food anticipatory activity and photic masking in the rat. Brain Res 842:73–83
Schoelch C, Hubschle T, Schmidt I, Nuesslein-Hildesheim B 2002 MSG lesions decrease body mass of suckling-age rats by attenuating circadian decreases of energy expenditure. Am J Physiol 283:E604–E611
Horvath TL 1997 Suprachiasmatic efferents avoid phenestrated capillaries but innervate neuroendocrine cells, including those producing dopamine. Endocrinology 138:1312–1320
Saeb-Parsy K, Dyball RE 2003 Responses of cells in the rat suprachiasmatic nucleus in vivo to stimulation of afferent pathways are different at different times of the light/dark cycle. J Neuroendocrinol 15:895–903
Jamali KA, Tramu G 1999 Control of rat hypothalamic pro-opiomelanocortin neurons by a circadian clock that is entrained by the daily light-off signal. Neuroscience 93:1051–1061
Gerhold LM, Sellix MT, Freeman ME 2002 Antagonism of vasoactive intestinal peptide mRNA in the suprachiasmatic nucleus disrupts the rhythm of FRAs expression in neuroendocrine dopaminergic neurons. J Comp Neurol 450:135–143
Lind RW, Van Hoesen GW, Johnson AK 1982 An HRP study of the connections of the subfornical organ of the rat. J Comp Neurol 210:265–277
Trudel E, Bourque CW 2003 A rat brain slice preserving synaptic connections between neurons of the suprachiasmatic nucleus, organum vasculosum lamina terminalis and supraoptic nucleus. J Neurosci Methods 128:67–77
Cheng MY, Bullock CM, Li C, Lee AG, Bermak JC, Belluzzi J, Weaver DR, Leslie FM, Zhou QY 2002 Prokineticin 2 transmits the behavioural circadian rhythm of the suprachiasmatic nucleus. Nature 417:405–410
Cottrell GT, Zhou QY, Ferguson AV 2004 Prokineticin 2 modulates the excitability of subfornical organ neurons. J Neurosci 24:2375–2379
Horikawa K, Yokota S, Fuji K, Akiyama M, Moriya T, Okamura H, Shibata S 2000 Nonphotic entrainment by 5-HT1A/7 receptor agonists accompanied by reduced Per1 and Per2 mRNA levels in the suprachiasmatic nuclei. J Neurosci 20:5867–5873
Maywood ES, Mrosovsky N, Field MD, Hastings MH 1999 Rapid down-regulation of mammalian period genes during behavioral resetting of the circadian clock. Proc Natl Acad Sci USA 96:15211–15216
Card JP, Moore RY 1989 Organization of lateral geniculate-hypothalamic connections in the rat. J Comp Neurol 284:135–147
Meyer-Bernstein EL, Morin LP 1996 Differential serotonergic innervation of the suprachiasmatic nucleus and the intergeniculate leaflet and its role in circadian rhythm modulation. J Neurosci 16:2097–2111
Bradbury MJ, Dement WC, Edgar DM 1997 Serotonin-containing fibers in the suprachiasmatic hypothalamus attenuate light-induced phase delays in mice. Brain Res 768:125–134
Carlini VP, Monzon ME, Varas MM, Cragnolini AB, Schioth HB, Scimonelli TN, de Barioglio SR 2002 Ghrelin increases anxiety-like behavior and memory retention in rats. Biochem Biophys Res Commun 299:739–743
Wellman PJ, Davis KW, Nation JR 2005 Augmentation of cocaine hyperactivity in rats by systemic ghrelin. Regul Pept 125:151–154
Kalsbeek A, van Heerikhuize JJ, Wortel J, Buijs RM 1998 Restricted daytime feeding modifies suprachiasmatic nucleus vasopressin release in rats. J Biol Rhythms 13:18–29
Elmquist JK, Scammell TE, Jacobson CD, Saper CB 1996 Distribution of Fos-like immunoreactivity in the rat brain following intravenous lipopolysaccharide administration. J Comp Neurol 371:85–103
Ericsson A, Kovacs KJ, Sawchenko PE 1994 A functional anatomical analysis of central pathways subserving the effects of interleukin-1 on stress-related neuroendocrine neurons. J Neurosci 14:897–913
Kelly JF, Elias CF, Lee CE, Ahima RS, Seeley RJ, Bjorbaek C, Oka T, Saper CB, Flier JS, Elmquist JK 2004 Ciliary neurotrophic factor and leptin induce distinct patterns of immediate early gene expression in the brain. Diabetes 53:911–920
Hart BL 1988 Biological basis of the behavior of sick animals. Neurosci Biobehav Rev 12:123–137
Zheng H, Corkern M, Stoyanova I, Patterson LM, Tian R, Berthoud HR 2003 Peptides that regulate food intake: appetite-inducing accumbens manipulation activates hypothalamic orexin neurons and inhibits POMC neurons. Am J Physiol 284:R1436–R1444
Ricardo JA, Koh ET 1978 Anatomical evidence of direct projections from the nucleus of the solitary tract to the hypothalamus, amygdala, and other forebrain structures in the rat. Brain Res 153:1–26
Berthoud HR, Patterson LM, Neumann F, Neuhuber WL 1997 Distribution and structure of vagal afferent intraganglionic laminar endings (IGLEs) in the rat gastrointestinal tract. Anat Embryol 195:183–191
Schemann M, Grundy D 1992 Electrophysiological identification of vagally innervated enteric neurons in guinea pig stomach. Am J Physiol 263:G709–G718
Qin LQ, Li J, Wang Y, Wang J, Xu JY, Kaneko T 2003 The effects of nocturnal life on endocrine circadian patterns in healthy adults. Life Sci 73:2467–2475
Kreier F, Yilmaz A, Kalsbeek A, Romijn JA, Sauerwein HP, Fliers E, Buijs RM 2003 Hypothesis: shifting the equilibrium from activity to food leads to autonomic unbalance and the metabolic syndrome. Diabetes 52:2652–2656
Dhillo WS, Small CJ, Stanley SA, Jethwa PH, Seal LJ, Murphy KG, Ghatei MA, Bloom SR 2002 Hypothalamic interactions between neuropeptide Y, agouti-related protein, cocaine- and amphetamine-regulated transcript and -melanocyte-stimulating hormone in vitro in male rats. J Neuroendocrinol 14:725–730(Chun-Xia Yi, Jan van der Vliet, Jiapei D)