Efferent Projections From the Median Preoptic Nucleus to Sleep- and Arousal Regulatory Nuclei in the Rat Brain

Abstract

The median preoptic nucleus (MnPO) has been implicated in the regulation of hydromineral balance and cardiovascular regulation. The MnPO also contains neurons that are active during sleep and in response to increasing homeostatic pressure for sleep. The potential role of these neurons in the regulation of arousal prompted an analysis of the efferent projections from the MnPO. Anterograde and retrograde neuroanatomical tracers were utilized to characterize the neural connectivity from the MnPO to several functionally important sleep- and arousal-regulatory neuronal systems in the rat brain. Anterograde terminal labeling from the MnPO was confirmed within the core and extended ventrolateral preoptic nucleus. Within the lateral hypothalamus, labeled axons were observed in close apposition to proximal and distal dendrites of hypocretin/orexin immunoreactive (IR) cells. Projections from the MnPO to the locus coeruleus were observed within and surrounding the tyrosine hydroxylase-IR cell cluster. Labeled axons from the MnPO were mostly observed within the lateral division of the dorsal raphé nucleus and heavily within the ventrolateral periaqueductal gray. Few anterogradely labeled appositions were present juxtaposed to choline acetylransferase-IR somata within the magnocellular preoptic area. The use of retrogradely transported neuroanatomical tracers placed within the prospective efferent terminal fields supported and confirmed findings from the anterograde tracer experiments. These anatomical findings support the hypothesis that MnPO neurons function to promote sleep by inhibition of orexinergic and monoaminergic arousal systems and disinhibition of sleep regulatory neurons in the ventrolateral preoptic area.

INTRODUCTION

The preoptic area/anterior hypothalamus participates in the regulation of homeostatic processes, including blood pressure, thirst, salt appetite, thermoregulation, sexual behavior and arousal (Boulant and Silva, 1988, Everitt, 1990, McGinty and Szymusiak, 1990, Johnson et al., 1996). The median preoptic nucleus (MnPO) and the ventrolateral preoptic area (vlPOA) contain neurons that are activated during both slow-wave (nonREM) and REM sleep, compared to waking (Sherin et al., 1998, Szymusiak et al., 1998, Gong et al., 2000, Szymusiak et al., 2001, Suntsova et al., 2002, McGinty and Szymusiak, 2003).

MnPO neurons innervate the paraventricular nucleus of the hypothalamus (PVH) (Silverman et al., 1981, Tanaka et al., 1987) the thalamus, the perifornical lateral hypothalamus (pLHA), the bed nucleus of the stria terminalis (BNST) (Gu and Simerly, 1997, Uschakov et al., 2001, Uschakov et al., 2006, Yoshida et al., 2006b), the suprachiasmatic nucleus (Oldfield et al., 1991b, Moga and Moore, 1997), the supraoptic nucleus (SON) (Miselis et al., 1979, Oldfield et al., 1991b), the dorsal raphé nuclei (DRN), the locus coeruleus (LC), the pericoerulear area (Zardetto-Smith and Johnson, 1995) and the vlPOA (Chou et al., 2002, Thompson and Swanson, 2003, Uschakov et al., 2006). Efferent connectivity with these areas of the brain may, in part, mediate the homeostatic processes attributed to the MnPO (Jones, 1988, Fitzsimons, 1998).

Sleep-related suppression of neural activity in some of these regions of the brain may be mediated by projections from MnPO GABAergic neurons that are activated during sleep (Gong et al., 2004, Uschakov et al., 2006). Several identified and potential targets of the MnPO are prominently implicated in the regulation of sleep and arousal, including the LC, DRN, posterior hypothalamus (PH), pLHA and magnocellular basal forebrain (see Jones, 2003, Jones, 2005 for review). However, projections from the MnPO to neurochemically characterized cell types in these areas have not been previously quantified. Therefore, we utilized anterograde and retrograde neuroanatomical tracers and the immunohistochemical identification of noradrenalin-, serotonin-, acetylcholine- and orexin/hypocretin-containing cells, to characterize the efferent neural projections from the MnPO to potential arousal regulatory cell groups.

METHODS

All experiments were conducted in accordance with the National Research Council Guide for the Care and Use of Laboratory Animals. All protocols described here were reviewed and approved by the Animal Care and Use Committee at the Veterans Administration of the Greater Los Angeles Health Care System.

Subjects were male, Sprague-Dawley rats, maintained on a 12:12 light:dark schedule at an ambient temperature of 21±3°C. Animals weighed 250 – 300g at the time of anatomical tracer injection, consistent with the neuroanatomical maps used to determine stereotaxic coordinates (Swanson, 1998). Injections were performed during a single, aseptic surgery under ketamine/xylazine (80/8 mg⁻¹kg⁻¹) anesthesia.

Anterograde studies utilizing biotin-dextran amine

Borosilicate glass pipettes (O.D. 1.2mm, I.D. 0.69mm; tip diameter 10–15μm) were filled with a 10% solution of biotin-dextran lysine (BDA-10,000, Sigma-Aldrich) in sterile isotonic saline. BDA was iontophoretically applied (10%, 7s on 7s off for 45 mins, +5μA) into the MnPO (A/P from bregma-0.0mm, lateral-0.0mm, depth from top of cortex–6.5mm). Animals were allowed to recover for 10 days after which they were sacrificed with an overdose of sodium pentobarbital (100 mg/kg, IP) and perfused through the heart with 100ml isotonic saline followed by 300ml of 4% paraformaldehyde in phosphate buffer (PB; 0.1M, pH 7.2), then by 100ml of 10% sucrose and 100ml 30% sucrose in PB. Brains were removed and stored in 30% sucrose for 24–48hrs. Serial coronal sections, 40 μm in thickness, were cut throughout the brain. A Vectastain ‘Elite’ biotin-avidin horseradish peroxidase conjugation followed by a nickel diaminobenzidine (Ni-DAB) visualization procedure was performed on all sections (Hsu and Soban, 1982).

Immunohistochemistry

Sections that had been processed for BDA staining were then separated into regions which are known to contain cholinergic neurons in the magnocellular preoptic area (MCP), orexin/hypocretin neurons in the pLHA, serotonergic neurons in the DRN, or noradrenergic neurons in the LC.

Choline acetyltransferase (ChAT)

Sections collected through the MCP area were incubated for 48 hours in monoclonal mouse anti-ChAT (1:500, Chemicon) made up in tris buffered saline (0.05M, pH 7.2, 15% NaCl; TBS) with 2% triton-x 100 and 4% normal goat serum. Tissue was then washed 3 times in TBS for 5 minutes. Following this, sections were incubated in the secondary antibody, biotinylated goat-anti-rat (mouse) IgG (1:200, Vector Laboratories, Burlingame, CA) in TBS with 4% normal goat serum, for 2 hours. After washing in TBS, sections were incubated with avidin peroxidase for 2 hours (1:200 Vector ‘Elite’) and developed with DAB.

Orexin/Hypocretin

Sections spanning the pLHA area were incubated for 48 hours in polyclonal rabbit anti-Orexin A (1:5000 AB-1, Oncogene) in TBS with 2% triton-x 100 and 4% normal goat serum. Sections were washed in TBS followed by two hours incubation in biotinylated secondary antibody, goat anti-rabbit IgG (1:200 Vector ‘Elite’) in TBS with 4% normal goat serum. Sections were then incubated with avidin peroxidase for two hours (1:200 Vector ‘Elite’) and developed with DAB.

Serotonin

Sections spanning the DRN were incubated for 48 hours in primary antibody, polyclonal rabbit anti-5HT (1:20000 Immunostar), in TBS with 2% triton-x 100 and 4% normal donkey serum. Tissue was washed in TBS, and incubated in secondary antibody, donkey anti-rabbit-HRP (1:100, Jackson) in TBS with 4% normal donkey serum, for 2 hours. Sections were then incubated with avidin peroxidase for two hours (1:200 Vector ‘Elite’) and developed with DAB.

Tyrosine hydroxylase

Sections spanning the LC (~8.5mm caudal to 10.5mm caudal to the optic chiasm) were incubated for 48 hours in polyclonal rabbit anti-tyrosine hydroxylase (1:500 Chemicon) in TBS with 2% triton-x 100 and 4% normal goat serum. Following three washes in TBS, sections were incubated for 2 hours in biotinylated secondary antibody, goat anti-rabbit IgG (1:200 Vector ‘Elite’) in TBS with 4% normal goat serum. Sections were then incubated with avidin peroxidase for two hours (1:200 Vector ‘Elite’) and developed with DAB.

Retrograde tracer studies utilizing fluoro-gold

The retrograde tracer, fluoro-gold (Fluorochrome, Denver, Co.), was unilaterally pressure injected (0.2 μl of 4% Fluoro-Au, in isotonic saline) at a rate of 10nl/min into the pLHA, (A/P –3mm, lateral–1.4mm, depth-7.8mm), the LC (A/P–10mm, lateral–1.4mm, depth –7mm), the DRN (A/P–8mm, lateral–0.0mm, depth–6.4mm) or the MCP (A/P–0.1mm, lateral–1.2mm, depth -7.0mm). Following a 10-day survival period, animals were perfused with 100ml PBS, 300ml 4% paraformaldehyde made up in PB, 100ml of 10% sucrose and 100 ml of 30% sucrose in PB and stored in 30% sucrose overnight. Coronal sections throughout the injection site and the preoptic area were cut at 40μm. Collected sections were divided into two series; the first series were mounted onto gelatin coated glass slides, air-dried and then defatted in xylene for 10 minutes. Following this the slides were coverslipped with a xylene based media (DPX). These sections were analyzed using a Nikon Eclipse E-600 fluorescent microscope. The second series were incubated for 24 hours in primary antibody, polyclonal rabbit anti-Fluoro-Au (1:5000 Chemicon) in TBS with 2% Triton x-100 and 4% normal goat serum. Sections were washed three times in TBS and incubated for 2 hours in secondary antibody, goat anti-rabbit IgG (1:200 Jackson) conjugated to either fluorescent rhodamine red or fluorescein isothiocyanate (FITC). Sections were then mounted on gelatin coated glass slides and coverslipped with a water-based mounting medium (Fluromount).

BDA-10,000 analyses

The computer-aided plotting system Neurolucida (V. 6.0) with a Lucivid attachment (MicroBrightfield) and a Ludl motorized stage controller was used to analyze and represent the locations of injection sites and boutons/varicosities within terminal fields. Injection loci were defined as sites containing major deposits of Ni-DAB reaction product. Labeled neurons surrounding the main deposit of BDA were considered as part of the injection site. Sections spanning the rostral-caudal levels of the vlPOA, MCP, pLHA, DRN and LC were selected for analysis. BDA-labeled processes and the distribution of those processes with respect to immuno-labeled neurons were examined in each selected region.

BDA-labeled structures included 1) terminal boutons; those structures that displayed noticeable swellings with little or no apparent continuation of the axon fiber, 2) en passant fibers; axons with multiple swellings along the longitudinal axis and 3) fibers of passage; axons located within the plane of section with no apparent intumescence or nodular structures.

An initial graphical format is presented whereby rostral-caudal sections throughout the target areas display the locations of varicosities and/or boutons (diameter ≥ 1.2μm) as determined by cursor size. Sections mapped were referenced to the standardized maps (adapted from Swanson, 1998) on which they were to be superimposed. Scaling, alignment and placement of anatomical profiles were guided by readily recognizable landmarks including magnocellular divisions (e.g. the SON) chiasms, fiber bundles, distinct nuclei and the ventricular system. Areas of the brain with a larger lateral span (i.e. the lateral hypothalamus) were represented unilaterally. Anatomical areas with prescribed borders that were not immunohistochemically defined (e.g. the vlPOA) were reconfirmed as to the correct plane of section and the superposition of labeled axons upon the map. Sections were mapped at a magnification of 400x.

Secondly, to quantify the density of boutons/varicosities in each area, standardized grids were placed in areas of interest and all boutons/varicosities throughout the 40μm section were counted, ipsilateral to the bulk of each injection site. This analysis was conducted for each of the following areas:

vlPOA

50μm × 50μm counting frames were placed in the center of the vlPOA core and the extended vlPOA at 0.40mm caudal to bregma. The center of these divisions was determined by the placement of larger grids as described previously (Chou et al., 2002, Uschakov et al., 2006).

MCP

150μm × 150μm frames were placed within the ChAT immunopositive region at approximately 0.25mm caudal to bregma. Three frames per animal were positioned at the level of the nucleus of the diagonal band of Broca (NDB), the MCP and the substantia innominata (SI). Frames incorporated a minimum of 10 ChAT-IR cells each.

pLHA

Grids were placed according to Yoshida et al. (2006), in the center of the orexin/hypocretin-IR field. Briefly, a medial (400μm × 500μm), central (200μm × 500μm) and lateral (400μm × 500μm) area was outlined with the base of the central partition located in the fornix. One 150μm × 150μm frame was then centered in each area (medial, central and lateral) and placed 100μm ventral from the top of each larger grid. Each counting frame incorporated a minimum of 6 immunopositive orexin/hypocretin soma.

DRN

150μm × 150μm counting frames were placed within the ventral part (between and dorsal to the medial longitudinal fasciculus), the dorsal core (directly beneath the cerebral aqueduct) and the lateral wing of the DRN, as determined by 5-HT immunoreactivity (approximately 7.9–8.1mm caudal to bregma). Each frame incorporated at least 15, 5-HT-IR cells. Additionally, 150μm × 150μm frames were placed upon the laterodorsal tegmental nucleus and within the ventral part of vlPAG (50μm lateral to the cerebral aqueduct). These frames contained less than 3, 5-HT-IR cells.

LC

50μm × 50μm counting frames were positioned within the TH-IR cell group approximately 9.8mm caudal to bregma. Each frame encompassed a minimum of 10, TH-IR cells.

Counts are expressed as the number of boutons and/or varicosities per 10μm². To correct counts for variation in the size of injection sites, the total area of each injection site was calculated with the Neurolucida system, and the area of the injection site for animal 1a (see Figure 1a and Figure 2a) was equated as one.

Figure 1.

Sketches of BDA-10,000 injection sites compiled using camera lucida imaging. a – f, representations of injection sites that were located within the MnPO, g – i, control injection sites that involved surrounding preoptic nuclei excluding the MnPO. aco-anterior commisure, ADP-anterodorsal preoptic nucleus, BNST-bed nucleus of the stria terminalis, fx- fornix, LPO-lateral preoptic area, LSv-ventral part lateral septal nucleus, LV-lateral ventricle, MPO-medial preoptic area, MS-medial septal nucleus, NDB-nucleus of the diagonal band of Broca, och-optic chiasm, OVLT-vascular organ of the lamina terminalis, PS-parastrial nucleus, 3v third ventricle.

Figure 2.

Photomicrographs of injection sites sketched in Figure 1a and 1b. a). a conservative injection of BDA into the MnPO, note the lack of tracer in surrounding structures, b). a larger injection into the MnPO; note the greater amount of tracer within the MnPO and the numerous labeled fibers surrounding the injection site. aco-anterior commisure, ADP-anterodorsal preoptic nucleus, BNST-bed nucleus of the stria terminalis, fx-fornical column, LSv-ventral part lateral septal nucleus, MPO-medial preoptic area, OVLT-vascular organ of the lamina terminalis LPO-lateral preoptic area. Scale bar = 400μm.

Retrograde tracer analyses

From a series of 18 rats that received Fluoro-Au injection into either the MCP, pLHA, DRN or LC, one optimal injection site from each targeted region was selected for analysis (see Figure 8). Retrogradely labeled neurons were plotted in coronal sections of the preoptic area at the level of the MnPO. Plots were conducted on a Nikon Eclipse E600 microscope with a fluorescent attachment and an EF-4 B-2E filter tube for FITC labeled tissue, a EF-4 G-2B filter tube for rhodamine labeled tissue or a UV-2A filter tube for Fluoro-Au labeled tissue. Photomicrographs were acquired on a Leica TCS-SP MP confocal microscope. For quantification of retrograde labeling within the MnPO, modified counting boxes that had been previously used to quantify sleep-related c-Fos-IR cells were positioned over the rostral MnPO (equilateral triangle with 600 μm sides, the base centered on the apex of the third ventricle) and the caudal MnPO (300μm × 600μm grid) (see Gong et al., 2000; 2004 for details). All retrogradely labeled neurons located within the grids were counted at a magnification of 400x.

Figure 8.

Neurolucida plots and confocal images of retrograde tracer injection sites and the location of retrogradely labeled neurons within the preoptic area. Schematic representation of injection sites into a) the magnocellular preoptic area, b) the perifornical lateral hypothalamus d) the dorsal raphe nucleus, and f) the locus coeruleus. a′,b′,d′,f′ location of retrogradely labeled neurons within the anterior or a″,b″,d″,f″ the posterior preoptic area. Single points represent retrogradely labeled neurons. c,e and g are confocal images of the injection sites in b,d and f respectively. c′e′ and g′ are confocal images of the MnPO at a rostral or c″e″ g″ caudal level respectively. Note that photomicrographs do not correspond to the level plotted above in every case. c-MnPO-central/medial division or l-MnPO lateral division of the MnPO. Scale bars = 100μm.

RESULTS

Anterograde Injection Sites

Nine anterograde injections were considered of most value in distinguishing the efferent neural projections from the MnPO or surrounding neural tissue. Figure 1 summarizes these experiments and indicates where the bulk of the injectant was localized. Figure 1, sections a through f, indicate injections that were well contained within the MnPO. Figure 1, sections g through i, presents injection sites that were considered as suitable control injections for areas partially labeled in cases a–f or as areas of interest within the preoptic area. Photomicrographs and/or Neurolucida plots throughout this manuscript refer to these injection sites/experiments and are labeled accordingly (i.e. animal 1a, 1b etc.).

Anterograde injections into the MnPO

Figures 2a and b, are photomicrographs of the injection sites traced in Figure 1a and b, respectively. Figure 2a depicts a medium well-localized injection into the ventral/rostral part of the MnPO, with a small deposit of tracer located in the dorsal division of the nucleus. Little if any spread of tracer was observed outside the MnPO. Figure 2b depicts a less conservative injection that displayed spread of tracer laterally on the right hand side of the brain and a portion dorsally within the septal nucleus. These two experiments were considered ideal to include for a comprehensive analysis of the efferent projections from the MnPO. Animal 1a provides an injection site that is confined to the MnPO, yet labels the nucleus throughout much of its rostral-caudal extent. The larger injection (animal 2a) provides a better overview of the absolute projection density from the MnPO and ensures that minor projections are accounted for. Injections shown in Figures 1c, d and e were well contained within the MnPO, but labeled only a portion of the nucleus. These cases were referenced to confirm projections revealed by the injections in animals 1a and 1b.

The numbers reported below, of boutons/varicosities per 10um², refer to the rectified values that took into account the size of each particular injection site, with the area of the injection site from animal 1a (Figures 1a and 2a) equated as one.

Anterograde tracer injections surrounding the MnPO

Three experiments were considered suitable for comparisons that were located entirely outside the bounds of the MnPO. These injection sites included tissue located laterally (Figure 1g and h) and dorsolaterally to the MnPO (Figure 1i).

The injection site presented in Figure 1g was located lateral to the MnPO, within the medial preoptic area (MPO), the ventral part of the lateral septal nucleus (LS) and the anterodorsal preoptic nucleus. The projections from this injection are likely to be similar to those arising from any lateral spread of tracer from the injection in animal 1b. The injection in animal 1h was located laterally to that in animal 1g, the bulk of tracer located in the anterodorsal and oval nucleus of the BNST. The injection site that is depicted in Figure 1i was situated within the caudal part of the LS, and should yield labeling similar to that caused by dorsal spread of tracer in animal 1b.

Projections to the ventrolateral preoptic area (vlPOA)

The presence of terminal boutons and varicosities were readily observed within vlPOA and the SON following BDA injections into the MnPO (animals 1a and 1b). MnPO projections to the vlPOA were dense within both the vlPOA core (c-vlPOA; 2.51 boutons/varicosities per 10um²) and the extended vlPOA (ex-vlPOA; 1.44 boutons/varicosities per 10um²; Figure 3). Injections into either the dorsal, rostral or caudal divisions of the MnPO all resulted in significant densities of anterograde fibers and varicosities in the c-vlPOA and the ex-vlPOA. A higher projection density (approx. 74% greater) was recognized within the c-vlPOA compared to the ex-vlPOA (Figure 3, Table 1).

Figure 3.

Projections from the MnPO to the ventrolateral preoptic area. a). Anterograde labeling within the MPO and the vlPOA following a conservative injection of anterograde tracer into the MnPO, note the retrogradely labeled neurons in both the extended-vlPOA (ex-vlPOA) and the greater number in the core-vlPOA (c-vlPOA; animal 1a) b). darkfield photomicrograph of that in a). note the high density of projections with boutons/varicosities in the vlPOA. c, d and e, rostrocaudal representations of medial and ventrolateral preoptic tissue mapped with Neurolucida software, each point represents the location of a single bouton and/or varicosity from animal 1a. f, g and h, plots of single boutons and/or varicosities following a less conservative injection site (animal 1b). Scale bar = 300μm.

Table 1.

Table 1a Relative density of efferent projections following injections of anterograde tracer BDA-10,000 into the MnPO or surrounding preoptic areas.

Table 1b

Area	vlPOA core	extended vlPOA	ChAT NDB	ChAT MCP	ChAT SI	Orexin Medial	Orexin Central	Orexin Lateral	DR ventral	DR ventro- lateral	DR dorsal core	PAG ventro- lateral	LDT	LC
Table 1a
MnPO 1 (animal 1a)	2.61	1.54	0.08	0	0	0.27	0.37	0.27	0.03	0.04	0.02	0.21	0.12	0.24
MnPO 2 (animal 1b)	20.49	11.35	2.4	0.7	0.35	4.72	4.5	4.51	4.09	5.7	1.87	5.82	4.21	8.74
Medial Preoptic Area (animal 1g)	10.24	10.24	2.58	0.75	0.52	6.79	6.76	7.21	0.5	0.72	0.82	1.59	0.6	6.91
Bed Nucleus of the Stria Terminalis (animal 1h)	27.56	19.21	4.45	1.6	2.77	4.21	4.57	4.44	5.27	5.72	4.1	6.91	7.35	6.5
Lateral Septal Area (animal 1i)	30.96	17.97	7.48	4.65	2.01	8.26	7.73	6.51	2.19	1.45	0.32	1.06	0.64	2.17
Table 1b
MnPO 1 (animal 1a)	2.61	1.54	0.08	0	0	0.27	0.37	0.27	0.03	0.04	0.02	0.21	0.12	0.24
MnPO 2 (animal 1b)	2.41	1.34	0.28	0.08	0.04	0.56	0.53	0.53	0.48	0.67	0.22	0.69	0.5	1.03
Medial Preoptic Area (animal 1g)	0.82	0.82	0.21	0.06	0.04	0.55	0.55	0.58	0.04	0.06	0.07	0.13	0.05	0.56
Bed Nucleus of the Stria Terminalis (animal 1h)	1.25	0.87	0.2	0.07	0.13	0.19	0.21	0.2	0.24	0.26	0.19	0.31	0.33	0.29
Lateral Septal Area (animal 1i)	3.7	2.14	0.89	0.55	0.24	0.99	0.93	0.78	0.26	0.17	0.04	0.13	0.08	0.26

Labeled boutons and/or varicosities in 10um2, within structures of the neuraxis by injection sites.

Labeled boutons and/or varicosities in 10um2, within structures of the neuraxis by injection sites. Analysis corrected for apparent size of injection site by normalizing injection site volumes to that of animal 1a (Figure 1a and 2a).

Labeled axons projecting to the vlPOA from the MnPO coursed bilaterally through the MPO in a ventrally orientated arc. This arc of axonal projections increased in width with greater lateral inclusion of tissue at the injection site. The axons in the arced band displayed numerous boutons within the MPO as they coursed toward the vlPOA and SON. Efferent projections originating in the MPO area (animal 1g) to the vlPOA were consistent between the core and the extended vlPOA (0.82 boutons/varicosities per 10um² in both sites; Table 1b). Heavy labeling within the vlPOA was also observed following tracer application into the BNST (animal 1h) although factoring in the size of the injection site, the projection density was approximately 47% less than that from the MnPO (Table 1b). The greatest density of labeled axons within the vlPOA were apparent following the injection located within the lateral septal nucleus (animal 1k).

Projections to the Magnocellular Preoptic Nucleus (MCP)

In contrast to the moderate to dense anterograde labeling seen in the vlPOA and SON following MnPO tracer injections, comparatively few axons were observed to continue dorsolaterally and innervate basal forebrain regions containing cholinergic neurons (Figure 4). This was particularly the case for animal 1a, for which labeled axons in cholinergic cell regions were scarce and close appositions among labeled axons and ChAT-IR cells were rare (Figure 4a). Following less conservative MnPO injections (animal 1b), labeled fibers were found among the most medially situated cholinergic cell groups in the lateral preoptic area (LPO) and the horizontal nucleus of the diagonal bands of Broca (NDB) (Figures 4b, 4c, 4d). Close appositions of labeled axons with terminal boutons were observed on some medially situated ChAT-IR neurons, but few labeled processes reached cholinergic regions in the MCP, SI or lateral portions of the NDB. Conversely, injections of anterograde tracer into the LS (0.56 boutons/varicosities per 10um²), the BNST (0.13 boutons/varicosities per 10um²) or the MPO (0.1 boutons/varicosities per 10um²) resulted in greater numbers of anterogradely labeled fibers with distinct varicosities in the ChAT positive areas of the basal forebrain investigated (Table 1a and b).

Figure 4.

Photomicrographs and Neurolucida plots of projections from the MnPO to ChAT-IR cells located within the magnocellular basal forebrain. a) few afferents were observed juxtaposed to ChAT-IR somata located within the lateral preoptic area following a conservative injection into the MnPO (animal 1a). b) ChAT-IR somata located closely to the supraoptic nucleus did receive close appositions following less conservative injections into the MnPO (animal 1b). c and d, neurolucida plots rostrocaudally through the preoptic-anterior hypothalamic area, ChAT-IR cells (green points) or labeled axons with boutons or varicosities (red points; animal 1b). Scale bar = 50μm.

Projections to the Perifornical Lateral Hypothalamus (pLHA)

Anterograde labeling within the pLHA was consistently observed following injections of BDA into the MnPO. Numerous labeled axons were seen coursing laterally to the pLHA through the periventricular nuclei, including the paraventricular and the dorsomedial nuclei of the hypothalamus (Figure 5). Labeled axons coursed dorsally around the fornix toward the lateral extent of the optic tract and the retrochiasmatic part of the SON. Fibers with boutons/varicosities were observed within the retrochiasmatic part of the SON, the tuberal nucleus and the more caudally located tuberomamillary nucleus. In the pLHA, labeled axons were frequently juxtaposed to orexin/hypocretin-IR somata (Figure 5 a and b). Putative synaptic contacts were observed proximal to cell somata and appositions were often located upon first order dendrites (Figure 5a). A few retrogradely labeled neurons that were immunopositive for orexin/hypocretin were noted following MnPO injections (Figure 5b arrowhead), indicating reciprocal connectivity with this neuronal class. Numerous labeled axons from the MnPO were also recognized in the surrounding parenchyma. Here, within the dendritic field of the orexin/hypocretin-IR cells, numerous terminal and en passant appositions were apparent (Figure 5).

Figure 5.

Projections from the MnPO to the perifornical lateral hypothalamic area and the orexin/hypocretin cluster. Photomicrographs of a). labeled axons with well defined boutons/varicosities located in close apposition to orexin/hypocretin-IR cells near the fornix, arrows indicate appositions upon proximal dendrites or cell bodies. Inset, large appositions in close proximity to an orexin/hypocretin-IR cell (animal 1a). b). a retrogradely labeled neuron from the MnPO that is double labeled for orexin/hypocretin-IR (arrowhead), note also the close appositions to orexin/hypocretin IR-cells (arrow, animal 1c). Lucivid plots of orexin/hypocretin-IR cells (green points) and individual boutons/varicosities (red points) rostrocaudally through the lateral hypothalamus following injections into the MnPO, c and d (animal 1a) or e and f, (animal 1b). Scale bar = 50μm.

The number of orexin/hypocretin-IR cells located in close apposition to labeled fibers was quantified and no differences were observed between the medial, central and lateral divisions of the field (Table 1). Projections to the pLHA were also apparent following injections of tracer into the LS and the BNST. Labeled projections from these sites were commonly located in close apposition to orexin/hypocretin neurons (Table 1). The greatest innervation to this area was from the LS where a high density of labeling was apparent medially/centrally within the orexin/hypocretin field (Table 1).

Projections to the Dorsal Raphé Nucleus (DRN)

Labeled processes were identified within the ventrolateral periaqueductal gray (vlPAG) (0.45 boutons/varicosities per 10um²) the ventrolateral division of the middle DRN (0.36 boutons/varicosities per 10um²) and the laterodorsal tegmental nucleus (0.31 boutons/varicosities per 10um²; Figure 6, Table 1). Efferent projections were largely located ipsilateral to the injection site. Fewer labeled projections from the MnPO were observed within the ventral division of the middle DRN (0.26 boutons/varicosities per 10um²) or the dorsal core (0.12 boutons/varicosities per 10um²), of which en passant boutons/varicosities predominated (Figure 6b,d). Anterograde labeling was also noted within the rostral and caudal portions of the DRN, albeit at a lesser density compared to the middle portion. Labeled varicose fibers were commonly seen in close apposition to 5HT-IR soma and proximal appendages within the ventrolateral DRN (Figure 6e,i,j). The greatest aggregation of labeled fibers was consistently located adjacent to the cerebral aqueduct along the border between the vlPAG and the ventrolateral division of the middle DRN (Figure 6d). The labeled projections to this area of the midbrain likely coursed from the MnPO via a periventricular route, either ventrally through the hypothalamus or dorsally through the thalamus alongside the third ventricle.

Figure 6.

Projections from the MnPO to the midbrain dorsal raphé/periaqueductal gray region. Anterogradely labeled axons from the MnPO to a). the middle ventral part of the DRN and b). the vlPAG and DRN from a conservative injection into the MnPO (animal 1a). (inset in a). represents higher magnification of the area depicted by the two arrows, note the well defined terminal swellings) c). labeled projections from the MnPO in close apposition to 5-HT positive neurons, though note that not all varicosities are adjacent to 5HT-IR neurons d). BDA positive axons within the vlPAG and the middle part of the DRN following a less conservative injection into the MnPO (animal 1b). e, f and g (animal 1a) or h, i and j, (animal 1b) Neurolucida plots of the locations of 5-HT cells (green points) and single boutons or varicosities (red points) rostrocaudally through the middle DRN following anterograde tracer injections into the MnPO.

Labeled processes within and near the DRN were evident following injections into the LS. Heavy innervations were apparent with predominance on the ipsilateral side to the injection site (Table 1). These projections tended to be dispersed slightly differently than those from the MnPO, noticeably, a greater density of fibers was apparent within the caudal division of the DRN. Significantly dense projections to the midbrain raphé nuclei were also evident following experiments in which the BNST was included in the injection site (Table 1). Numerous labeled axons with were observed coursing throughout the rostral to caudal extent of the DRN and the vlPAG, and high densities of labeled fibers were also located within the pontine reticular nucleus, the cuneiform nucleus and the parabrachial nucleus.

Projections to the Locus Coeruleus (LC)

Inferior to the cerebellum, within the brainstem at the lateral recess of the fourth ventricle, tyrosine hydroxylase positive IR neurons were readily seen tightly clustered within the LC (Figure 7). Axonal projections with terminal boutons and varicosities from the MnPO were observed throughout the LC (0.64 boutons/varicosities per 10um²). Numerous innervations were also noted ventrally within the subcoerulear area where compact aggregations of varicose fibers and terminal boutons were located (Figure 7). The labeled axons within the subcoerulear region including the pontine central gray and pontine reticular nucleus were orientated obliquely and dorso-ventrally from the midline indicating that the course of fibers to this region arose via the periventricular stratum. Labeled axons from the MnPO were observed in an around the A6 and A4 groups, with minimal labeling observed in other brainstem noradrenergic clusters. Labeled terminal processes within the A6-A4 groups were also observed following injections of anterograde tracer into the MPO (0.56 boutons/varicosities per 10um²), the BNST (0.29 boutons/varicosities per 10um²) and the LS (0.26 boutons/varicosities per 10um²).

Figure 7.

Projections from the MnPO to the locus coeruleus. a) and b) labeled projections to TH-IR neurons at the rostral part of the LC following a conservative injection into the MnPO (animal 1a) arrows indicate anterogradely labeled axons with boutons or varicosities, inset in a). lower arrow (inset) indicates where boutons/varicosities are apparent, c) dark field image of labeled projections within the locus coeruleus and within the d) subcoerulear area (animal 1b). e, f and g, Lucivid maps rostrocaudally of TH-IR cells (green dots) and boutons or varicosities on afferent fibers from the MnPO (red dots; animal 1a). h, i and j, Neurolucida maps rostrocaudally of TH-IR cells (green points) and boutons or varicosities on afferent fibers from the MnPO (red points; animal 1b). 4v – fourth ventricle, LC – locus coeruleus, PCG – pontine central gray, PRN – pontine reticular nucleus, Scale bar = 200μm.

Retrograde confirmation of efferent projections from the MnPO

The Magnocellular Preoptic Nucleus

Application of Fluoro-Au into the ChAT-IR somatic field within the MCP/NDB resulted in small numbers of retrogradely labeled axons in the rostral MnPO, and no labeling in the caudal portions of the nucleus (Figure 8a and Table 2). Greater numbers of retrogradely labeled neurons were located laterally within the MPO, the vlPOA and dorsally within the LS and lateral part of the medial septal nucleus (Figure 8a).

Table 2.

Retrogradely labeled cell counts and cell density in the rostral and caudal MnPO following fluorogold injections into the dorsal raphe nucleus, locus coeruleus, perifornical lateral hypothalamus and the magnocellular preoptic area

Injection Site	Rostral MnPO Total cell number	Caudal MnPO Total cell number	Rostral MnPO Density cells/0.1 mm2	Caudal MnPO Density cells/0.1 mm2
DRN	37	52	20.6	28.9
LC	39	61	21.7	33.9
pLHA	69	87	38.3	48.3
MCP	11	0	6.1	0

The Lateral Hypothalamus

Injection of Fluoro-Au into the pLHA resulted in heavy retrograde labeling within the preoptic area. Nuclei/areas that contained labeled neurons included the MnPO, the vlPOA, the medial and lateral preoptic areas and the ventromedial part of the BNST (Figure 7b,c). Within the rostral part of the MnPO, retrogradely labeled neurons were commonly located within a ventral-dorsal lamina of soma in the midline/central MnPO (c-MnPO), Further caudally labeled neurons were noted within the central lamina and a lateral lamina (l-MnPO) ipsilateral to the injection site (Figure 8c″). Greater densities of retrogradely labeled neurons were located ipsilateral to the injection site both within the MnPO and the surrounding preoptic area. A somewhat higher density of retrogradely labeled neurons was observed in the caudal versus the rostral portions of the MnPO, following the pLHA tracer injection (Table 2).

The Dorsal Raphé Nucleus

Iontophoretic application of Fluoro-Au targeting DRN/vlPAG resulted in tracer deposited primarily within the ventral, ventrolateral and dorsolateral part of the DRN nucleus (Figure 8d & e). Numerous retrogradely labeled neurons were found within the preoptic area, including the vlPOA, the LPO and MPO, the BNST, and the substantia innominata (SI). Within the rostral part of the MnPO, retrogradely labeled neurons were mostly localized within the central lamina, with few labeled cells in the l-MnPO. More caudally, retrogradely labeled neurons tended to be located within lateral margins of the nucleus (Figure 8e′& e″), . Rostrally within the MnPO, the density of retrogradely labeled neurons was slightly less than that observed at more caudal levels of the nucleus (Table 2). The distribution of labeled soma tended to be heaviest within the MnPO with a more dispersed pattern in adjacent areas such as the MPO. Subjacent to the MnPO retrogradely labeled neurons were observed within the dorsal cap of the OVLT. The distribution of labeled neurons was heaviest ipsilateral to the injection site, particularly within the MPO, LPO and vlPOA (Figure 8).

The Locus Coeruleus

Injections of Fluoro-Au into the LC resulted in significant labeling within the preoptic area (Figure 8f & g). Retrogradely labeled neurons were readily identified within the MnPO and within the subjacent OVLT. Retrograde labeling was moderate to dense within both the c-MnPO and the l-MNPO, throughout the rostral-caudal extent of the nucleus (Figure 8f and g). The overall density of retrogradely labeled neurons was approximately one-third greater in the caudal MnPN, compared to rostral MnPO (Table 2). Numerous retrogradely labeled neurons were also located within the LPO, MPO, vlPOA, SI and BNST. Retrogradely labeled neurons from the LC injection presented displayed greater numbers of labeled neurons more rostrally within the medial and lateral preoptic areas. Retrogradely labeled neurons were predominantly located ipsilateral to the injection site (Figure 8f & g).

DISCUSSION

These studies describe the distribution of efferent projections from the MnPO to immunohistochemically identified neural populations involved in regulation of behavioral and electrographic arousal. Neural projections from the MnPO were observed juxtaposed to cells IR for hypocretin/orexin within the pLHA (~0.42 boutons/varicosities per 10um²). Anterogradely labeled axons from the MnPO were observed within the DRN where appositions to serotonin-IR cells within the ventral, dorsal and lateral divisions of the nucleus were characterized. A greater density of labeled axons with varicosities/boutons was observed dorsolateral to the DRN, within the vlPAG (~88% greater). Distinct anterogradely labeled axons with clear varicosities/boutons were observed within the tyrosine hydroxylase immuno-positive region of the LC (~0.64 boutons/varicosities per 10um²). Labeled axons were also observed within the pericoerulear area, medially, laterally and subjacent to the LC.

Labeled terminals from the MnPO to magnocellular preoptic ChAT-IR somata were found to be limited. Close appositions to ChAT-IR cells and labeled MnPO projections were only observed at the medial boundaries of the cholinergic neuronal field, lateral to the vlPOA and ventral BNST.

Local projections from the MnPO to the vlPOA, an area that contains functionally important sleep-active neurons, were also analyzed. Findings indicate a 74% greater projection density to the core-vlPOA (2.51 boutons/varicosities per 10um²) than that to the extended-vlPOA (1.44 boutons/varicosities per 10um²).

Technical Considerations

BDA-10,000 is commonly used as an anterograde neuroanatomical tracer, but may also be taken up by terminal processes and transported in a retrograde manner. The potential for confusion of anterogradely and retrogradely labeled axonal fibers was considered negligible in the present study due to the relatively low number of retrogradely labeled neurons observed.

Application of tracers for either anterograde or retrograde analyses should be considered semi-quantitative as minor variations in injection site locale and/or size are unavoidable. We attempted to minimize this variation by normalizing all injection site volumes to that of our most conservative anterograde MnPO injection (animal 1a; Figure 1a and 2a). However, we found that following rectification for the volume of different MnPO injection sites, there was greater variation with respect to the density of projections in the more caudal areas investigated. This difference was minimal at the level of the vlPOA, yet for the larger injection site (animal 1b) there was a 1.8x increase in density of labeling at the level of the LHA, 3.3x greater density at the level of the vlPAG and 4.3x greater density at the level of the LC, compared to the more conservative MnPO injection (Table 1b). We attribute the decline in label at more caudal regions to limited quantities of tracer available to particular neurons located within the MnPO. The large variation between DRN and LC projection density between animal 1a and 1b may indicate that a particular zone within the MnPO failed to acquire significant amounts of tracer following the smaller injection, most probably the more caudal region of the nucleus. This supposition is supported by the finding that the density of retrogradely labeled neurons was somewhat higher in the caudal versus the rostral MnPO following Fluoro-Au injections into the DRN and LC (Table 2).

The placement of the frames used for cell counts was carefully checked, and was, to the best of our ability, consistent across animals. The use of immunohistochemistry was helpful in equating brain areas between animals, although areas with widely dispersed IR-soma are unavoidably troubled by minor tissue inconstancies. Therefore, although the maps and counts are standardized and the locations of labeled projections true with respect to surrounding major landmarks, slight rostrocaudal variations in tissue section are inevitable. It is for this reason that the locations of IR neurons were clearly demarcated on the plots to guide review and that minimum numbers of IR labeled cells in each counting frame were required.

Finally, there can be a tendency to underestimate the extent of axonal labeling in the distal dendritic fields of immunohistochemically stained cell types, given that distal dendritic processes often fail to stain with high contrast.

Comparison with earlier works

We have quantified the distribution of anterogradely-labeled axons and varicosities with respect to cholinergic neurons in the MCP, orexin/hypocretin- neurons in the pLHA, serotonin neurons in the DRN and noradrengergic neurons in the LC. One previous study quantified the density of boutons/varicosities from the MnPO to the vlPOA (Chou et al., 2002), to which our data corresponds well.

The axonal projections from the MnPO to the monoaminergic neural clusters within the DRN and the LC have been described. Our work confirms these projections with both anterograde and retrograde neuroanatomical tracers (Zardetto-Smith et al., 1993, Zardetto-Smith and Johnson, 1995).

Projections from the MnPO to the pLHA have been described (Saper and Levisohn, 1983, Gu and Simerly, 1997, Thompson and Swanson, 2003, Sakurai et al., 2005, Uschakov et al., 2006, Yoshida et al., 2006b). We confirmed and quantified MnPO efferent projections to the lateral hypothalamic area, provided evidence of synaptic contact upon orexin/hypocretin neurons and provided preliminar evidence of orexinergic projections to the MnPO.

This study also recognized by way of anterograde tracer only, projections from the MnPO to the OVLT (Camacho and Phillips, 1981, Miselis et al., 1987, Tanaka et al., 1987, Oldfield et al., 1992, Zardetto-Smith et al., 1993, Gu and Simerly, 1997), the SON (Miselis et al., 1979, Oldfield et al., 1991b), the subfornical organ (Lind et al., 1982), the dorsomedial hypothalamic nucleus, the tuberomamillary nucleus (Thompson and Swanson, 1998) and the paraventricular nucleus of the hypothalamus (Silverman et al., 1981, Tanaka et al., 1987, Tanaka and Nomura, 1993). Efferent projections were also confirmed within the periaqueductal gray (Gu and Simerly, 1997) and the medial preoptic area (Oldfield et al., 1991b).

Functional Implications

The MnPO is a multifunctional area that is active during a number of homeostatic challenges (McKinley et al., 1996b, McKinley et al., 2001, Szymusiak et al., 2001, Gvilia et al., 2005, Gvilia et al., 2006). The way in which this nucleus exerts influence to correct homeostatic imbalance involves centers within the brain capable of orchestrating widespread and dynamic systems. Therefore the efferent neural connectivity of the MnPO to autonomic, endocrine, monoaminergic and associated preoptic and hypothalamic effectors provides suitable scaffolding for the modulation of both physiological and behavioral parameters to achieve homeostasis (Uschakov et al., 2001, Thompson and Swanson, 2003).

A functional role of the MnPO in detecting/mediating response to hydromineral imbalance and Angiotensin II has been well documented (Johnson et al., 1981, Culman et al., 1995, McKinley et al., 1996a, Gvilia et al., 2005). Of particular interest is the role of the MnPO in motivational drive whereby its chemical destruction results in significant reductions in water intake (Jones, 1988, Cunningham et al., 1992). The neural pathways that may mediate such drive, including the recollection of a water source, heightened vigilance state to avoid predators and the motivation to do such should be reflected anatomically (McKinley et al., 2004). MnPO projections to the DRN LC, and the pLHA are potential substrates of arousal and motivational state regulation. As the ascending projections from these areas modulate activity throughout the limbic system and neocortex, it is hypothesized that by way of these intermediary structures, the MnPO coordinates and modulates behavioral/motivational influences to drive response.

Central and lateral subdivisions of the MnPO are readily apparent following retrograde tracer administration into terminal areas from this nucleus. A distinct midline or central division of the MnPO has been reported previously following retrograde tracer administration into the SON (Oldfield et al., 1992). The central partition of the MnPO also stains positive for c-Fos following 24 hours of water deprivation (McKinley et al., 1994) and is innervated heavily from the periphery or outer shell of the subfornical organ (Miselis et al., 1979, Oldfield et al., 1991a), whereas the lateral divisions of the MnPO corresponds to the staining observed for the prostaglandin EP3 receptor (Nakamura et al., 2000). Neurons expressing c-Fos during sleep are located in both the central and lateral subdivisions of the MnPO, but c-Fos labeling is more prominent in the lateral subdivision in animals sleeping in a warm environment (Gong et al., 2000). The lack of a febrile response to lipopolysaccharide subsequent to lesions of the MnPO supports a possible role in fever and/or thermoregulation (Yoshida et al., 2006a). As EP3 receptors tend to be heavily distributed within the lateral margins of the MnPO (Nakamura et al, 2000) this subdivision may be a site of integration among thermoregulatory and sleep regulatory mechanisms.

The MnPO contains neurons that are active during nonREM and REM sleep, compared to waking (Suntsova et al., 2002). The sleep-waking discharge pattern of MnPO neurons is reciprocal to that described for DRN neurons (McGinty and Harper, 1976), LC neurons (Aston-Jones and Bloom, 1981), and hypocretin/orexin neurons (Lee et al., 2005, Mileykovskiy et al., 2005), all of which exhibit greatly diminished activity during nonREM and REM sleep, compared to waking. A proportion of MnPO neurons that express c-Fos protein immunoreactivity during sleep are GABAergic (Gong et al., 2004). To the extent that GABAergic, sleep active neurons are a source of MnPO projections to hypocretin/orexin and monoaminergic cell groups, they may contribute to sleep-related suppression of activity in these arousal systems, in parallel with projections from vlPOA sleep-regulatory neurons (Sherin et al., 1998; Steininger et al., 2001). It has been demonstrated that a subset of MnPO-to-pLHA projection neurons do exhibit sleep related c-Fos protein immunoreactivity (Uschakov et al., 2006) as do sleep-active neurons within the MnPO that project to the DRN/vlPAG and the BNST (unpublished observations).

The efferent projections from the MnPO to the vlPOA contains neurons that express c-Fos during sleep as well as significant numbers of neurons that express c-Fos during waking (Uschakov et al., 2006). Evidence from in vitro recordings indicates that vlPOA neurons are subject to tonic, local GABA-mediated inhibition (Chamberlin et al., 2003). The subset of GABAergic, sleep-active neurons in the MnPO that project to vlPOA may function to disinhibit GABAergic/galaninergic vlPOA neurons during nonREM and/or REM sleep.

List of anatomical abbreviations

3v

third ventricle

4v

fourth ventricle

a

anterior

aco

anterior commisure

ad

anterodorsal area

al

anterolateral area

ADP

anterodorsal preoptic nucleus

ARH

arcuate nucleus

av

anteroventral area

AVPV

anteroventral periventricular nucleus

AQ

cerebral aqueduct

B

Barrington’s nucleus

BST or BNST

bed nuclei stria terminalis

c

core or central, see text

cc

central core

cpd

cerebral peduncle

d

dorsal division

DC

dorsal core

dl

dorsolateral division

dm

dorsomedial area

DMH

dorsomedial nucleus hypothalamus

DRN and DR

dorsal raphe nucleus

DTN

dorsal tegmental nucleus

ex

extended

fu

fusiform nucleus

fx

fornix

l

lateral

LC

locus coeruleus

LDT

laterodorsal tegmental nucleus

LHA

lateral hypothalamic area

LPO

lateral preoptic area

LS

lateral septal nucleus

LV

lateral ventricle

m

medial division

MA

magnocellular preoptic nucleus

MCP

magnocellular preoptic area

ME

median eminence

MEA

medial nucleus amygdala

MEV

mesencephalic nucleus of the trigeminal nerve

mlf

medial longitudinal fasciculus

MnPO

median preoptic nucleus

MPN

medial preoptic nucleus

MPO

medial preoptic area

MRN

mesencephalic reticular nucleus

MS

medial septal nucleus

mtt

mammillothalamic tract

NDB

nucleus of the diagonal band of Broca

och

optic chiasm

opt

optic tract

OVLT

vascular organ of the lamina terminalis

ov

oval nucleus

PAG

periaqueductal gray

PB

parabrachial nucleus

pLHA

perifornical lateral hypothalamic area

lcPB

parabrachial nucleus: central lateral part

lvPB

parabrachial nucleus: ventral lateral part

mmPB

parabrachial nucleus: medial medial part

PCG

pontine central gray

PH

posterior hypothalamic nucleus

PRN

pontine reticular nucleus

PS

parastrial nucleus

PSCH

suprachiasmatic preoptic nucleus

rh

rhomboid nucleus

RMVE

rostral medullary velum

sc

subcommissural zone

scp

superior cerebellar peduncle

SG

supragenual nucleus

SI

substantia innominata

SLC

subcoeruleus nucleus

SO or SON

supraoptic nucleus

STN

subthalamic nucleus

sut

supratrigeminal nucleus

TU

tuberal nucleus

vl

ventrolateral division

POA

preoptic area

VMH

ventromedial nucleus hypothalamus

ZI

zona incerta