Forebrain projections of tuberoinfundibular peptide of 39 residues (TIP39)-containing subparafascicular neurons

Neuroscience 138 (2006) 1245–1263

FOREBRAIN PROJECTIONS OF TUBEROINFUNDIBULAR PEPTIDE OF 39 RESIDUES (TIP39)-CONTAINING SUBPARAFASCICULAR NEURONS J. WANG,a M. PALKOVITS,a,b T. B. USDINa* AND A. DOBOLYIa,b

Key words: immunohistochemistry, neuropeptide, tuberoinfundibular peptide of 39 residues, subparafascicular, neuroanatomical tract tracing, biotinylated dextran amine.

a Laboratory of Genetics, National Institute of Mental Health, Building 35, Room 1B215, 35 Convent Drive, Bethesda, MD 20892-3728, USA b Laboratory of Neuromorphology, Semmelweis University and Hungarian Academy of Sciences, Budapest, 1094, Tuzalto U. 58, Hungary

Tuberoinfundibular peptide of 39 residues (TIP39) was purified from bovine hypothalamus based on its activation of the parathyroid hormone 2 receptor (Usdin et al., 1999). Pharmacological data support assignment of TIP39 as the endogenous ligand of the parathyroid hormone 2 receptor (Usdin, 2000), which has a relatively widespread distribution in the CNS (Wang et al., 2000). Functional studies suggest that TIP39 has anxiolytic- and antidepressant-like actions (LaBuda et al., 2004) and also that it affects some components of the hypothalamo-pituitary axis (Ward et al., 2001; Sugimura et al., 2003). The distribution of TIP39 fibers has been described in detail (Dobolyi et al., 2003b). They have a relatively widespread distribution in the rat CNS including a high density of fibers in the limbic cortex, septal, and amygdaloid regions, some midline and intralaminar thalamic nuclei, hypothalamus, periaqueductal gray, parabrachial nuclei, brainstem auditory nuclei, and spinal cord (Dobolyi et al., 2003b). In contrast, TIP39 synthesizing cell bodies are almost entirely restricted to three regions of the rat brain (Dobolyi et al., 2003b): 1) an area in the medial part of the posterior thalamus that includes the subparafascicular nucleus (SPF) and the subparafascicular area (SPA); 2) an area medial to the medial geniculate body in the lateral part of the posterior thalamus that includes the posterior intralaminar nucleus of the thalamus and the parvicellular SPF; and 3) the medial paralemniscal nucleus in the lateral pons. In the posterior thalamus TIP39 neurons form a rostrocaudally elongated cell group. Anteriorly, these cells occupy the rostral half of the subparafascicular nucleus (rSPF) and posteriorly, a portion of the SPA. The rSPF is located within the thalamus, dorsal to the posterior hypothalamic nucleus and medial to the parvicellular part of the ventral posterior thalamic nucleus. The SPA includes the caudal part of the SPF and an area between the most caudal part of the third ventricle and the fasciculus retroflexus. It is continuous caudally with the periaqueductal gray of the midbrain (Moriizumi and Hattori, 1991). In contrast to the rSPF, the caudal part of the SPF contains a relatively low number of TIP39 cells. At this level, TIP39 cells are located more dorsomedially in the SPA. The SPA received major attention when A11 dopaminergic cells projecting to the spinal cord were identified in the area (Bjorklund and Skagerberg, 1979; Hökfelt et al., 1979; Skagerberg and Lindvall, 1985) and analgesia could be elicited by stimulating the region (Yeung et al., 1977;

Abstract—Neurons containing tuberoinfundibular peptide of 39 residues (TIP39) constitute a rostro-caudally elongated group of cells in the posterior thalamus. These neurons are located in the rostral part of the subparafascicular nucleus and in the subparafascicular area, caudally. Projections of the caudally located TIP39 neurons have been previously identified by their disappearance following lesions. We have now mapped the projections of the rat rostral subparafascicular neurons using injections of the anterograde tracer biotinylated dextran amine and the retrograde tracer cholera toxin B subunit, and confirmed the projections from more caudal areas previously inferred from lesion studies. Neurons from both the rostral subparafascicular nucleus and the subparafascicular area project to the medial prefrontal, insular, ecto- and perirhinal cortex, nucleus of the diagonal band, septum, central and basomedial amygdaloid nuclei, fundus striati, basal forebrain, midline and intralaminar thalamic nuclei, hypothalamus, subthalamus and the periaqueductal gray. The subparafascicular area projects more densely to the amygdala and the hypothalamus. In contrast, only the rostral part of the subparafascicular nucleus projects significantly to the superficial layers of prefrontal, insular, ectorhinal and somatosensory cortical areas. Double labeling showed that anterogradely labeled fibers from the rostral part of the subparafascicular nucleus contain TIP39 in many forebrain areas, but do not in hypothalamic areas. Injections of the retrograde tracer cholera toxin B subunit into the lateral septum and the fundus striati confirmed that they were indeed target regions of both the rostral subparafascicular nucleus and the subparafascicular area. In contrast, TIP39 neurons did not project to the anterior hypothalamic nucleus. Our data provide an anatomical basis for the potential involvement of rostral subparafascicular neurons in limbic and autonomic regulation, with TIP39 cells being major subparafascicular output neurons projecting to forebrain regions. © 2005 IBRO. Published by Elsevier Ltd. All rights reserved. *Corresponding author. Tel: ⫹1-301-402-6976; fax: ⫹1-301-402-0245. E-mail address: [email protected] (T. B. Usdin). Abbreviations: ABC, avidin– biotin– horseradish peroxidase complex; BDA, biotinylated dextran amine; CTB, cholera toxin B subunit; DAB, 3,3-diaminobenzidine; ir, immunoreactive; PBS, phosphate-buffered saline at pH⫽7.4; PBS/T, phosphate-buffered saline containing 0.5% Triton X-100; rSPF, rostral subparafascicular nucleus; SPA, subparafascicular area; SPF, subparafascicular nucleus; TB, 0.1 M Tris buffer at pH⫽8.0; TH, tyrosine hydroxylase; TIP39, tuberoinfundibular peptide of 39 residues.

0306-4522/06$30.00⫹0.00 © 2005 IBRO. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.neuroscience.2005.12.022

1245

1246

J. Wang et al. / Neuroscience 138 (2006) 1245–1263

Rhodes and Liebeskind, 1978; Boivie and Meyerson, 1982; Hosobuchi, 1983; Fleetwood-Walker et al., 1988). Labeled cells in the SPA were found following retrograde tracer injections into several regions of the forebrain, including the frontal cortex (Musil and Olson, 1988; Takada et al., 1988; Takada, 1993), the auditory cortex (Moriizumi and Hattori, 1991), the amygdala (Ottersen and Ben-Ari, 1979; Peschanski and Mantyh, 1983; Takada, 1990; Moriizumi and Leduc-Cross, 1992; Takada, 1993), striatal regions (Peschanski and Mantyh, 1983; Beckstead, 1984; Bentivoglio and Molinari, 1984; Takada, 1993), and the ventromedial nucleus of the hypothalamus (LeDoux et al., 1985). The few studies using anterograde tract tracing to examine the projections of this area only described projections to the brainstem (Yasui et al., 1992) and amygdala (Turner and Herkenham, 1991). A lesioning study aimed at identifying the projections of the recently discovered TIP39 neurons in this area revealed much more extensive projections (Dobolyi et al., 2003a). Large lesions encompassing both the medial and lateral areas at the diencephalon-midbrain junction that contain TIP39 cell bodies suggest that TIP39 fibers in the forebrain mostly originate in neurons of the SPA and the parvicellular SPF. These lesion studies also show that TIP39 fibers in the hindbrain mostly originate in the medial paralemniscal nucleus (Dobolyi et al., 2003a). The projections of the rSPF, in which there are a fairly high number of TIP39-containing neurons, have not yet been mapped. Since a number of axons from TIP39 neurons in the caudally located SPA as well as ascending sensory fibers pass through the rSPF, electrolytic lesions are not optimal for identifying the projections from this area. Because of the proximity to the third ventricle excitotoxic lesions are also problematic. Tract tracing methods may be better suited for defining the projections from this area. Investigation of the physiological functions of the SPF and SPA is limited by the difficult identification of the anatomical borders of these regions, and by the absence of detailed anterograde studies of the projections of neurons located there. To better define the region of interest in this study, we first examined the distribution of TIP39- and tyrosine hydroxylase (TH) -containing neurons in the SPF and SPA, because the A13 dopaminergic cell group is in the proximity of the rSPF and the A11 dopaminergic cell group is within the SPA. Next, we examined the projections of neurons in the rSPF using injections of the anterograde tracer biotinylated dextran amine (BDA) and compared them to projections of the SPA, with special attention to the distribution of TIP39-immunoreactive (ir) fibers. We performed double-labeling for BDA and TIP39 following BDA injections into the SPF to identify which of the projections were from TIP39 neurons. We then injected retrograde tracer into the lateral septum, the fundus striati, and the anterior hypothalamic nucleus to confirm that they are target areas of the SPF and area, and that projection neurons contain TIP39. The final goal of this study was to complete the map of the TIP39 innervation pattern throughout the forebrain in rat.

EXPERIMENTAL PROCEDURES Animals Animal procedures were performed according to U.S. National Institutes of Health and National Institute of Mental Health guidelines and were approved by the National Institute of Mental Health Institutional Animal Care and Use Committee. Adult male Sprague–Dawley rats (250 –290 g in body weight) were purchased from Taconic Breeding Laboratories and housed under standard laboratory conditions. All possible efforts were made to minimize the number of animals used and their suffering.

Tracer injections The anterograde tracer BDA (10,000 MW, Molecular Probes, Eugene, OR, USA) was targeted to the rSPF (n⫽25) and the SPA (n⫽8). Injections immediately dorsal, ventral and caudal to the rSPF (n⫽10) served as controls. Rats anesthetized with pentobarbital (50 mg/kg, i.p.) were positioned in a stereotaxic apparatus with the incisor bar set at ⫺3.3 mm. Glass micropipettes (10 ␮m i.d. for the rSPF and 15 ␮m i.d. for the SPA) filled with 10% BDA dissolved in phosphate-buffered saline at pH⫽7.4 (PBS) were lowered to the targets of interest using the following stereotaxic coordinates (Paxinos and Watson, 1998): AP⫽⫺3.8 mm, L⫽0.5 mm from the midline, V⫽6.9 mm from the surface of the skull for the rSPF and AP⫽⫺4.4 mm, L⫽0.4 mm, V⫽6.4 mm for the SPA. BDA was injected by iontophoresis using a constant current source (BAB-350, Kation Scientific, Minneapolis, MN, USA) that delivered a current (⫹10 ␮A for the rSPF and ⫹15 ␮A for the SPA) which pulsed for 5 s on and 5 s off for 15 min. The pipette was left in place for 5 min with no current and then withdrawn under negative current. The retrograde tracer cholera toxin B subunit (CTB; List Biological Laboratories, Campbell, CA, USA) was injected into the ventral part of the lateral septal nucleus (two animals), the fundus striati (two animals), and the anterior hypothalamic area (two animals). The coordinates were: AP⫽0.0, L⫽1.0, V⫽7.4 mm for the ventral lateral septum, AP⫽⫺2.1, L⫽4.7, V⫽8.6 mm for the fundus striati, and AP⫽⫺1.6, L⫽0.9, V⫽9.8 mm for the anterior hypothalamic nucleus. CTB (0.25% in PBS) was delivered from a glass micropipette (20 ␮m I.D.) by iontophoresis (⫹15 ␮A, 5 s on and 5 s off for 20 min).

Histology Seven days after tracer injection, rats deeply anesthetized with pentobarbital (80 mg/kg) were perfused transcardially with 200 ml PBS followed by 300 ml of ice-cold 4% paraformaldehyde in PBS. Brains were post-fixed in 4% paraformaldehyde for 24 h and then transferred to PBS. Coronal sections cut at 50 ␮m on a vibrating microtome (Leica S700) were collected in PBS/0.1% sodium azide and stored at 4 °C. Seven rats, which did not receive tracer injections, were perfused, and their brains were collected as described above. Coronal sections from one animal were Nissl-stained with Thionin and used as a reference for the identification and demonstration of the nuclear groups. Coronal (three animals) and sagittal (three animals) brain sections were collected for TIP39 and TH immunolabeling and TIP39/TH dual labeling.

Visualization of BDA Every fourth section was stained for BDA using either immunoperoxidase or immunofluorescence procedures with or without tyramide-mediated amplification. Sections were pretreated in PBS containing 0.5% Triton X-100 (PBS/T) for 1 h at room temperature. Sections were then incubated in avidin– biotin– horseradish peroxidase complex (ABC) at 1:500 (Vectastain ABC Elite kit, Vector Laboratories, Burlingame, CA, USA) for 2 h. Then the BDA-fibers

J. Wang et al. / Neuroscience 138 (2006) 1245–1263 were visualized using 0.02% 3,3-diaminobenzidine (DAB; Sigma, St. Louis, MO, USA), 0.08% nickel (II) sulfate and 0.0012% hydrogen peroxide in PBS. Alternatively, for biotin tyramide amplification, sections were rinsed extensively first in PBS and then in 0.1 M Tris at pH⫽8.0 (TB) after the first ABC incubation. Next, sections were placed in biotin–tyramide solution (1:1000) prepared in TB containing 0.0012% hydrogen peroxide for 20 min. Then a second ABC incubation was conducted at 1:1000 for 1 h. Finally, the reaction product was visualized by incubation for 10 min in DAB solution as described above. For FITC-tyramide amplification, after the ABC reaction sections were incubated in FITC-tyramide (1:10,000) and 0.0012% hydrogen peroxide in TB for 10 min. All sections were mounted on positively charged slides (Superfrost Plus; Fisher Scientific, Pittsburgh, PA, USA). Most DAB sections were counterstained with Nuclear Red (Vector Laboratories), and all DAB sections were dehydrated and coverslipped with Cytoseal 60 (Stephens Scientific, Riverdale, NJ, USA). Fluorescent sections were mounted using Prolong Antifade reagent (Molecular Probes).

Visualization of CTB Every second brain section from CTB-injected animals was incubated overnight in goat anti-CTB antibody (1:30,000; List Biological Laboratories) at room temperature and then in Alexa 594 donkey anti-goat secondary antibody (1:400; Molecular Probes) for 1 h.

TIP39 and TH immunocytochemistry A previously characterized (Dobolyi et al., 2003a,b) affinity-purified rabbit polyclonal antibody to rat TIP39, which labels cell bodies with exactly the same distribution as observed by in situ hybridization histochemistry and which can be absorbed with synthetic TIP39 (Dobolyi et al., 2003b) was used in this study. The antibody against TH (mouse monoclonal anti-TH, Novus Biologicals, Littleton, CO, USA) provided the distribution well established in the literature (Hökfelt et al., 1976) in this study and also in previous studies (Guo et al., 2002; Nocjar et al., 2002). Immunocytochemical controls were conducted with the omission of the primary antibody throughout the study. Brain sections were pretreated in PBS/T/1% bovine serum albumin for 1 h followed by incubation with TIP39 primary antibody (1:2000 for tyramide amplification and 1:400 for DAB reaction) or anti-TH (1:1000) in PBS/T/1% bovine serum albumin for 48 h at room temperature. Sections were then incubated in biotin-conjugated donkey anti-rabbit secondary IgG at 1:600 (Jackson Immunoresearch, West Grove, PA, USA) or biotin-conjugated donkey anti-mouse IgG at 1:600 (Jackson Immunoresearch) for 1 h, followed by incubation in ABC and DAB or FITC-tyramide as described above for the visualization of BDA.

Double labeling TIP39/TH double labeling on sagittal sections was done as for single labeling except that TIP39 was visualized by incubation in Alexa Fluor 594 anti-rabbit secondary antibody (1:400; Molecular Probes) for 2 h after the labeling of TH with FITC-tyramide. For BDA/TIP39 and BDA/TH colocalization every fourth section was labeled. TIP39 or TH immunocytochemistry was performed first and then followed by visualization of BDA. Briefly, sections were pretreated in PBS/T/3% normal donkey serum (Vector Laboratories) for 1 h. The sections were then incubated with primary antibody against TIP39 (1:1000) or with primary antibody against TH (1:800) for 48 h in PBS/T/3% normal donkey serum. Following thorough rinses in PBS, sections were incubated in polymerized peroxidase anti-rabbit or anti-mouse antibody (ImmPress Reagent, Vector Laboratories) for 30 min, then rinsed three times in PBS and three times in TB. TIP39-, or TH-immu-

1247

noreactivity was visualized using FITC-tyramide at 1:10,000 for 6 min. Subsequently, BDA was visualized by incubating the sections in Alexa Fluor 594 streptavidin at 1:500 (Molecular Probes) for 2 h at 37 °C. For CTB/TIP39 and CTB/TH double labeling, TIP39 or TH immunocytochemistry was performed first. TIP39 immunoreactivity was visualized with FITC-tyramide as described for single labeling. TH immunoreactivity was visualized by a 2-hour incubation of sections in Alexa Fluro 488 (1:400; Molecular Probes). Subsequently, sections were incubated in goat anti-CTB antibody (1:30,000; List Biological Laboratories) overnight and then in Alexa Fluor 594 donkey anti-goat antibody (1:400; Molecular Probes) for 1 h.

In situ hybridization histochemistry Two rat brains were removed, and the fresh tissue was quickly frozen on dry ice. Coronal sections were cut at 12 ␮m, mounted on positively charged slides, dried, and stored at ⫺80 °C. In situ hybridization was performed as previously described with a 35Slabeled antisense riboprobe to rat TIP39 that labels the same cell bodies as the antibody to TIP39 (Dobolyi et al., 2003a,b).

Data analysis The locations of the injection sites were plotted onto corresponding regions of a rat atlas (Paxinos and Watson, 1998). Animals with injection sites centered in the rSPF or the SPA, or that had adjacent but non-overlapping injection sites were included in the analysis. Images were captured at 1300⫻1030 pixel resolution with a Photometrix CoolSnap Fx digital camera on an Olympus IX70 light microscope equipped with fluorescent epi-illumination using 4⫻ objective. The individual pictures were then montaged using Photoshop CS 8.0. Drawings were prepared by aligning the montaged pictures with corresponding schematics adapted from Paxinos and Watson (2005). Confocal images were acquired with a Zeiss LSM 510 confocal microscope using a 10⫻ objective for the CTB/TIP39 and CTB/TH studies and a 63⫻ objective for the BDA/TIP39 and BDA/TH studies. The presented confocal images are of single 10 and 0.8 ␮m optical layers, respectively. Contrast was adjusted using the “levels” and the “unsharp mask” commands in Adobe Photoshop. Full resolution was maintained until the micrographs were cropped and assembled for printing, at which point images were adjusted to a resolution of 300 dpi.

RESULTS Topography of the SPF and the SPA The SPF is a cytoarchitectonically well-defined rostro-caudally elongated nucleus, with a circular profile in coronal sections, in the ventromedial part of the posterior thalamus (Fig. 1). The rostral part, which comprises almost half of the nucleus, is embedded in the posterior thalamic and hypothalamic nuclei (Fig. 1A, B), while the caudal part is situated within the SPA at the diencephalon–midbrain junction (Fig. 1C, D). No obvious landmarks separate these two parts. The rostral part of the nucleus is bordered by the posterior hypothalamic nucleus ventrally, the parvicellular part of the ventral posterior thalamic nucleus, a specific gustatory relay thalamic nucleus, dorsolaterally and the central medial thalamic nucleus dorsally. Rostrally, it is clearly separated from the reuniens thalamic nucleus and the compact A13 dopaminergic cell group. The caudal part of the nucleus extends into the SPA and occupies a

1248

J. Wang et al. / Neuroscience 138 (2006) 1245–1263

Fig. 1. Coronal view of the rostral part of the subparafascicular thalamic nucleus (A and B) and the SPA (C and D). (A, C) Schematic drawings modified from Paxinos and Watson (1998). (B, D) Thionin-stained coronal sections illustrate the topography of the rSPF (B) and the SPA (D). Arrowheads indicate the SPF, with its caudal end in D. Scale bars⫽200 ␮m.

small region just ventral and ventromedial to the fasciculus retroflexus. Here, the SPF is connected to a laterally and caudally extending group of small neurons, called the parvicellular SPF. Caudally, the SPF is bordered by the interstitial nucleus of Cajal.

Both parts of the SPF contain heterogeneous cell populations. We found TIP39- as well as TH-ir cell bodies in the rSPF (Fig. 2A–C), as well as in the SPA along the third ventricle, an area between the midline and the fasciculus retroflexus (Fig. 2D–F), but relatively few TIP39 cells were

Abbreviations used in the figures AA ac AcbS AH AMY A11 A13 BST cc Ce CM CP DM f fr FS GP H IC IL Ins IPF LH LS LV

anterior amygdaloid area anterior commissure nucleus accumbens, shell portion anterior hypothalamic nucleus amygdala A11 dopaminergic cell group A13 dopaminergic cell group bed nucleus of the stria terminalis corpus callosum central nucleus of the amygdala central medial thalamic nucleus caudate-putamen dorsomedial nucleus of the hypothalamus fornix fasciculus retroflexus fundus striati globus pallidus hippocampus inferior colliculus infralimbic cortex insular cortex interpeduncular fossa lateral hypothalamic area lateral septal nucleus lateral ventricle

MA ml MPA mt ot PAG pc PF PH PIL Pn PrL PVN PVT py RI RN SC SI SPFPC VPPC ZI 3V 4V

medial amygdaloid nucleus medial lemniscus medial preoptic area mamillothalamic tract optic tract periaqueductal gray posterior commissure parafascicular nucleus posterior hypothalamic nucleus posterior intralaminar thalamic nucleus pontine nuclei prelimbic cortex paraventricular nucleus of the hypothalamus paraventricular nucleus of the thalamus pyramidal tract interstitial nucleus of Cajal red nucleus superior colliculus substantia innominata parvicellular subparafascicular nucleus parvicellular part of the ventral posterior thalamic nucleus zona incerta third ventricle fourth ventricle

J. Wang et al. / Neuroscience 138 (2006) 1245–1263

1249

Fig. 2. Coronal sections through the rostral part of the subparafascicular thalamic nucleus (A–C) and the SPA (D–F). (A, D) TIP39 immunolabeling. The TIP39-ir cell bodies are more difficult to see at low magnification in the rSPF; the boxed area in A is enlarged in the inset. (B, E) TIP39 mRNA detected by in situ hybridization (high silver grain density causes the black centers in this darkfield image). (C, F) TH immunolabeling; the arrows indicate the caudal part of the SPF. Scale bars⫽500 ␮m in A and D, 50 ␮m in inset, 400 ␮m in B and E, and 200 ␮m in C and F.

located in the caudal part of the SPF (arrows on Fig. 2D). TIP39-ir neurons in the rSPF appeared somewhat smaller than those in the SPA. On sagittal sections, the neurochemical markers revealed a characteristic rostro-caudally expanded sigmoid shape of a group of cells that was continuously distributed throughout the periventricular area of the caudal thalamus (Fig. 3). Rostrally, TIP39 cells were located within the SPF, but caudally occupied a territory dorsal to it, in the upper part of the SPA (Fig. 3C). Although, the distributions of TIP39-ir and TH-ir cell bodies largely overlapped, the TH-ir cell bodies were situated somewhat more laterally than the TIP39 cells in the rSPF and more laterally and ventrally in the SPA (Fig. 3B, C, D). We found no double-labeled cells (Fig. 3D).

Injection sites Small injections of the anterograde tracer BDA were placed in the rSPF and dorsocaudally to it, in the SPA (Fig. 4A–F). An injection site is shown in Fig. 4G and H to illustrate the tracer uptake by neurons in the rSPF. Similarly, an injection site located in the SPA is shown in Fig. 4I and J. Injections in brain areas that border the SPF and area were analyzed as controls. The control injection sites did not overlap with the SPF or area (Fig. 4A–F) and include parts, in various sizes, of the parvicellular ventral posterior thalamic nucleus, the posterior hypothalamic nucleus, the caudal part of the central medial thalamic nucleus, or the interstitial nucleus of Cajal and the nucleus of Darkschewitsch. Cerebral cortex

Projections from the rostral part of the SPF The distribution of BDA-labeled fibers was relatively widespread following BDA injections into either the rSPF or the SPA (Table 1). The great majority of labeled fibers were always ipsilateral to the injections, and only a low or very low density of labeled fibers was present contralaterally. The distribution of labeled fibers was similar for the two injection sites with a few notable exceptions (Table 1).

BDA fibers from the rSPF and the SPA were present in restricted parts of the cerebral cortex. Neurons from both regions projected to limbic cortical areas including cingulate, prelimbic, infralimbic, and dorsal peduncular cortices (Fig. 5A, B, H and J), and to deeper layers of frontal association, including somatosensory, orbital, insular, ectorhinal, perirhinal and entorhinal cortices. In contrast, a remarkable difference was seen between the cortical pro-

1250

J. Wang et al. / Neuroscience 138 (2006) 1245–1263

Fig. 3. Sagittal view of the rostro-caudally elongated group of TIP39- and TH-ir cells in the SPF and SPA. (A) Parasagittal drawing of the rat brain modified from Paxinos and Watson (1998). (B) Immunolabeling demonstrates the distribution of the TH-immunoreactivity in the framed area in Fig. A. (C) Immunolabeling demonstrates the distribution of the TIP39-immunoreactivity in the SPF and SPA. (D) Dual immunofluorescence labeling of TIP39 (red) and TH (green). Note that the distribution of TIP39-immunoreactivity overlaps with that of TH, but they are not co-localized. Scale bars⫽800 ␮m in B, and 200 ␮m in C and D.

J. Wang et al. / Neuroscience 138 (2006) 1245–1263 Table 1. Distribution of BDA-labeled fibers following BDA injections in the rSPF and the SPA compared to that of TIP39 fibers

Table 1. Continued Area

Area

Forebrain Cerebral cortex Frontal association cortex Superficial layer Cingulate cortex Prelimbic cortex Infralimbic cortex Dorsal peduncular cortex Tenia tccta Orbital cortex Superficial layer Insular cortex Superficial layer Somatomotor cortex Somatosensory cortex Visual cortex Auditory cortex Superficial layer Ectorhinal cortex Superficial layer Perirhinal cortex Superficial layer Entorhinal cortex Olfactory bulb Ant. olfact. nucleus Olfactory tubercle Nucleus of the diagonal band Hippocampus CA1 CA3 Dentate gyrus Subiculum Parasubiculum Septum Medial septal nucleus Lateral septal nucleus Dorsal Intermediate Ventral Bed nucleus of the stria terminalis Amygdala Central nucleus Basolateral nucleus Basomedial nucleus Lateral nucleus Medial nucleus Cortical nucleus Anterior amygdaloid area “Fundus striati” Amygdalo-hippocamp. trans. area Amygdalo-piriform trans. area Basal nuclei Caudate-putamen Ventromedial part Globus pallidus Endopiriform nucleus Claustrum

BDA fibers (rSPF)

BDA fibers (SPA)

1251

TIP39 fibers

BDA fibers (rSPF)

BDA fibers (SPA)

TIP39 fibers

⫹ 0 ⫹ ⫹

⫹⫹ ⫹⫹ ⫹⫹ ⫹⫹

⫹⫹ ⫹ ⫹ ⫹⫹

0 0 ⫹ ⫹ ⫹⫹⫹ ⫹ 0 ⫹ ⫹ Site ⫹⫹ ⫹⫹ ⫹ ⫹ 0 ⫹ ⫹ ⫹⫹ ⫹⫹ ⫹ ⫹

⫹⫹ ⫹ ⫹ ⫹⫹ ⫹⫹⫹ ⫹⫹ ⫹⫹ ⫹⫹ ⫹⫹ ⫹⫹ Site ⫹⫹ ⫹ ⫹ 0 ⫹⫹ ⫹⫹ ⫹⫹ ⫹⫹⫹ ⫹ ⫹

0 0 ⫹ ⫹⫹ ⫹ 0 ⫹ 0 0 Perikarya Perikarya Perikarya 0 0 0 0 ⫹ ⫹ Perikarya ⫹ ⫹⫹

⫹ ⫹ ⫹

⫹ ⫹ ⫹

⫹ ⫹⫹ 0

0 ⫹ ⫹⫹ ⫹ ⫹⫹ 0 ⫹ ⫹⫹ ⫹ ⫹⫹ ⫹ ⫹ 0 ⫹⫹ ⫹⫹ ⫹ ⫹ ⫹⫹ ⫹⫹ ⫹ ⫹ ⫹

0 ⫹⫹⫹ ⫹⫹⫹ ⫹ ⫹⫹ ⫹ ⫹ ⫹⫹⫹ ⫹ ⫹⫹ ⫹ ⫹ 0 ⫹⫹⫹ ⫹⫹⫹ ⫹⫹ ⫹⫹ ⫹⫹⫹ ⫹⫹ ⫹ ⫹ ⫹⫹

0 ⫹ ⫹ ⫹ ⫹⫹⫹ 0 ⫹ ⫹ ⫹⫹ ⫹ ⫹ ⫹⫹ 0 ⫹ ⫹ 0 ⫹⫹⫹ ⫹⫹ 0 ⫹ 0 ⫹

0 ⫹ ⫹ ⫹⫹

0 ⫹ ⫹ ⫹⫹⫹

0 0 ⫹ ⫹⫹

Nucleus accumbens

⫹ ⫹⫹⫹ ⫹ ⫹ ⫹ ⫹⫹ ⫹⫹ ⫹ ⫹⫹ ⫹⫹ ⫹⫹⫹ 0 ⫹ 0 0 ⫹ ⫹⫹ ⫹⫹ ⫹⫹ ⫹⫹ ⫹ 0 0 ⫹ ⫹⫹

⫹ 0 ⫹ ⫹ ⫹ ⫹⫹ ⫹⫹ ⫹ 0 ⫹ 0 ⫹ ⫹ 0 0 0 ⫹⫹ 0 ⫹ 0 ⫹ 0 ⫹ ⫹ ⫹⫹

⫹ 0 0 ⫹ ⫹ ⫹⫹ ⫹⫹ 0 0 ⫹ 0 0 0 0 0 0 ⫹⫹ 0 ⫹ 0 0 0 ⫹ ⫹ 0

0 0 0 0 ⫹

0 0 0 ⫹ ⫹

0 0 0 ⫹ ⫹

⫹

⫹⫹

0

0 ⫹⫹ ⫹ ⫹

0 ⫹⫹⫹ ⫹⫹ ⫹⫹

0 ⫹⫹⫹ ⫹⫹ ⫹⫹

⫹ 0 ⫹ 0 0 ⫹ 0 ⫹ 0 0

⫹⫹ ⫹ ⫹⫹ ⫹ ⫹ ⫹ ⫹ ⫹⫹ ⫹ ⫹

⫹⫹ 0 ⫹ ⫹ ⫹⫹ ⫹ ⫹ ⫹⫹ ⫹ ⫹

0 0 ⫹ ⫹ ⫹

⫹ ⫹⫹ ⫹⫹ ⫹ ⫹

0 0 0 ⫹ 0

Shell Core Ventral pallidum Substantia innominala Diencephalon Thalamus Anterodorsal nucleus Anteroventral nucleus Paratenial nucleus Paraventricular nucleus Reuniens nucleus Laterodorsal thalamic nucleus Central medial thalamic nucleus Mediodorsal nucleus Parafascicular nucleus SPF SPA Parvicell. subparafascicular nu. Ventral nuclei Reticular nucleus Medial habenular nucleus Lateral habenular nucleus Posterior thalamic nucleus Lateral posterior nucleus Post. intralaminar thalamic nu. Peripeduncular area Suprageniculate thalamic nu. Medial geniculate body Ventral nucleus Medial nucleus Lateral geniculate body Hypothalamus OVLT Medial preoptic area Lateral preoptic area Supraoptic nucleus Supraoptic decussations Suprachiasmatic nucleus Retrochiasmatic area Anterior hypothalamic nucleus Paraventricular nucleus Periparaventricular area Periventricular nucleus Arcuate nucleus Median eminence Lateral hypothalamic area Perifornical nucleus Ventromedial nucleus Dorsomedial nucleus Posterior hypothalamic nucleus Tuberomamillary nucleus Ventral premamillary nucleus Dorsal premamillary nucleus Supramamillary nucleus Mamillary body Medial mamillary nucleus Lateral mamillary nucleus Subthalamic nucleus Forel’s fields

1252

J. Wang et al. / Neuroscience 138 (2006) 1245–1263

Fig. 4. Illustrations of the locations and sizes of BDA injections into the rSPF, the SPA and in several surrounding control sites. (A–F) Schematic drawings show BDA injections restricted to the rSPF (injection #63, #28, #32, #41, and #62), centered in the rSPF but including surrounding areas (#39), involving SPF and some surrounding areas (#40, #34), restricted to the SPA (#72, #129), involving the SPA and some surrounding areas (#84, #103, #127), as well as control injections into the PH (#57, #37, #105), the VPPC (#35), the central median thalamic nucleus (#128), the PAG (#104, #126), and an area that includes the RI and the nucleus of Darkschewitsch (#124). (G) The coronal section shows an injection site (#28) centered in the rSPF. (H) Higher magnification of the injection site in G. (I) The coronal section shows an injection site (#129) centered in the SPA. (J) Higher magnification of the injection site in I. Scale bars⫽800 ␮m in G and I, and 200 ␮m in H and J. The schematic drawings are adapted and modified from the atlas of Paxinos and Watson (1998).

J. Wang et al. / Neuroscience 138 (2006) 1245–1263

1253

Fig. 5. (A–N) Schematic illustrations of the terminal patterns of labeled fibers after BDA injections into the rSPF (left column: A–G) and the SPA (right column: H–N). Maps were modified from the atlas of Paxinos and Watson, 1998.

1254

J. Wang et al. / Neuroscience 138 (2006) 1245–1263

Fig. 5. (Continued).

J. Wang et al. / Neuroscience 138 (2006) 1245–1263

1255

Fig. 5. (Continued).

jections of the rSPF and the SPA in the superficial layers of the frontal association, orbital, insular, ectorhinal, and perirhinal cortices. These areas contained a very high to high density of fibers following BDA injection into the rSPF but not the SPA (Fig. 5A–N). The olfactory bulb and some cortical areas, including the anterior olfactory nucleus and the visual and auditory cortices did not contain labeled fibers (Table 1).

Septum

Hippocampus

Amygdala and the bed nucleus of the stria terminalis

Few BDA-fibers were found in the subiculum or parasubiculum. No BDA-fibers were visualized in the CA1–CA3 regions of the hippocampus or the dentate gyrus (Fig. 5E–G and L–N).

The distribution patterns of BDA-containing fibers in the amygdala were different following injections into the rSPF and the SPA (Table 1). Moderate or low labeling densities were seen in all of the amygdaloid nuclei and areas after

The septum was a major target of both the rSPF (Figs. 5C, 5D and 6A) and the SPA (Figs. 5J, 5K and 6B). Injections in the two sites led to similar labeling patterns. Labeled fibers were present in the intermediate and ventral subdivisions of the lateral septal nucleus but not in its dorsal part. BDA-fibers were moderately dense in the medial septal nucleus, as well as in the nucleus of the diagonal band.

Fig. 6. Fluorescence photomicrographs illustrate the distribution of the anterogradely labeled fibers following BDA injections into rSPF (A), and the distribution of TIP39-ir fibers (B) in the septum. Note the high density of BDA-containing fibers in the septum ipsilateral to the injection site. In contrast, BDA fibers are absent in the adjacent shell part of the nucleus accumbens. Also note the similar distribution of TIP39 fibers. Scale bars⫽1 mm.

1256

J. Wang et al. / Neuroscience 138 (2006) 1245–1263

and cortical nuclei, as well as in the fundus striati, but not in other amygdaloid nuclei or areas (Table 1). All of the divisions of the bed nucleus of the stria terminalis had labeled fibers following injection into either site (Fig. 5B–E and 5I–L). Basal ganglia The rSPF did not project to the caudate-putamen. There were large differences in the density of labeled fibers in the globus pallidus, the endopiriform nucleus, the claustrum, the nucleus accumbens, the ventral pallidum and the substantia nigra after injections into the rSPF and the SPA. SPA injections resulted in a high density of BDA fibers in these areas whereas rSPF injections resulted in only a low density (Fig. 5D, E, K and L; Table 1). Thalamus

Fig. 7. Fluorescence photomicrographs illustrate the distribution of labeled fibers following BDA injections into the SPA (A), and the distribution of TIP39-ir fibers (B) in the AMY. Arrows point to the FS where the density of BDA containing fibers as well as TIP39-ir fibers was high. Also note the higher density of TIP39 fibers in the MA. Scale bars⫽500 ␮m.

injections into the SPA (Fig. 5L and 5M), but only a limited number of labeled fibers appeared after injections into the rSPF (Fig. 5E and 5F). After injections into the SPA, the fundus striati (or amygdalo-striatal transitional zone) showed the highest density of BDA-containing fibers (Fig. 7A), while the distribution of the fibers was relatively uniform in other amygdaloid nuclei. Only the central and basomedial nuclei contained a somewhat higher density of BDA-containing fibers than other amygdaloid nuclei. Following injections into the rSPF we found labeled fibers in the central, basomedial

BDA-labeled fibers were present in several thalamic nuclei with a similar distribution pattern after injections into the rSPF and the SPA (Fig. 5E–G and L–N). The highest density of labeled fibers was present in the reuniens nucleus and the posterior intralaminar thalamic nucleus. We also observed a very high density of fibers in the laterodorsal thalamic nucleus, the lateral posterior nucleus, and the parvicellular SPF. In addition, a high density of fibers was present in the midline and intralaminar thalamic nuclei following SPA, but only at very low density after rSPF, injections. A low density of fibers was also present in the anterodorsal, anteroventral, parataenial, ventral, reticular, peripeduncular, and suprageniculate thalamic nuclei as well as in the medial and lateral geniculate bodies. BDA fibers were absent from the medial habenula while present in the lateral habenula. Hypothalamus The density of labeled fibers was high in many regions of the hypothalamus (Fig. 5D–G, K–N and Fig. 8) following both rSPF and SPA injections. There was a particularly high density of labeled fibers in the lateral preoptic area, the anterior hypothalamic nucleus, around the paraventricular nucleus, the lateral hypothalamic area, the perifornical nucleus, the

Fig. 8. Fluorescence photomicrographs demonstrate the distribution of BDA-containing fibers in the anterior hypothalamus following injections into the rSPF (A) and the SPA (B). The distribution of TIP39-ir fibers is shown for comparison (C). Note that TIP39-ir fibers are present in the paraventricular hypothalamic nucleus, which is devoid of BDA-containing fibers. Scale bars⫽500 ␮m.

J. Wang et al. / Neuroscience 138 (2006) 1245–1263

posterior hypothalamic nucleus, and the fields of Forel. The medial preoptic area contained a very high density of labeled fibers following injection into the SPA but not following rSPF injection. Labeled fibers were present, at only low density following rSPF injections, in the supraoptic, suprachiasmatic, paraventricular, periventricular, arcuate, ventromedial, dorsomedial, premamillary, tuberomamillary (Fig. 5D–G), and the lateral mamillary nucleus. The low density of anterogradely labeled fibers in the paraventricular nucleus stands out, particularly when compared with the high density of fibers in the periparaventricular area or the high density of TIP39 fibers in the paraventricular nucleus (Fig. 8). Anterogradely labeled fibers were absent from the organum vasculosum laminae terminalis, the median eminence, and the medial mamillary nucleus. Projections following BDA injections into regions adjacent to the rSPF or the SPA Injections into the parvicellular part of the ventral posterior thalamic nucleus, the central medial thalamic nucleus, and the interstitial nucleus of Cajal and the nucleus of Darkschewitsch resulted in a dramatically different pattern of anterograde labeling than did the injections into the SPF or area. Injections into these three sites did not result in a significant density of anterogradely labeled fibers in most of the regions receiving projections from the SPF or area. Instead, the parvicellular part of the ventral posterior thalamic nucleus projected heavily to the insular cortex (mostly lamina V but also lamina I); the central medial thalamic nucleus projected heavily to the insular cortex (lamina I and also lamina V), the ventral part of the caudate-putamen, the substantia innominata (Fig. 9), and the central and lateral amygdaloid nuclei. Differences are also pronounced between the distribution of BDA fibers following BDA injection into the rostral periaqueductal gray and the rSPF and SPA. BDA injections into the rSPF and SPA labeled prelimbic, infralimbic, and cingulate cortices, the lateral septum, the bed nucleus of the stria terminalis, and the medial preoptic area, whereas there was not appreciable anterograde labeling present in these regions following large BDA injections into the rostral periaqueductal gray.

1257

Table 2. Presence of BDA and TIP39-ir double-labeled fibers following BDA injections into the rSPF and the SPA

Forebrain Limbic cortex Insular cortex Ectorhinal cortex Olfactory tubercle Nucleus accumbens. shell part Lateral septal nucleus Bed nucleus of the stria terminalis Substantia innominata Fundus striati Central amygdaloid nucleus Diencephalon Paraventricular thalamic nucleus Medial preoptic area Anterior hypothalamic nucleus Supraoptic decussations Paraventricular hypothalamic nucleus Arcuate nucleus Dorsomedial hypothalamic nucleus Posterior intralaminar thalamic nucleus

rSPF

SPA

⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫺

⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹

⫺ ⫹ ⫺ ⫺ ⫺ ⫺ ⫺ ⫹

⫹ ⫹ ⫺ ⫹ ⫺ ⫺ ⫺ ⫹

The projections of the posterior hypothalamic nucleus were relatively similar to those of the rSPF and the SPA, except only the rSPF projected to the superficial layer of some cortical areas. Another difference was that neurons in the posterior hypothalamus projected to the hippocampus. TIP39 content of the efferent projections of the rSPF and SPA Following injections into the rSPF, as well as the SPA, varying numbers of the BDA containing fibers were immunopositive for TIP39 in several forebrain areas including cortical, basal forebrain, septal, extended amygdalar, and some diencephalic regions (Table 2). Double-labeled fibers typically demonstrated several double-labeled varicosities connected by regions with BDA content only (Fig. 10), suggesting that TIP39-ir is concentrated in the varicosities. Hypothalamic nuclei, including the paraventricular and the anterior hypothalamic nuclei, as well as hindbrain regions contained both BDA- and TIP39-labeled fibers. However, these areas did not contain double-labeled fibers (Table 2). The distribution of double-labeled fibers following rSPF and SPA tracer injections was generally similar, however some differences were observed in the amygdala and hypothalamus (Table 2). Retrograde labeling in the SPF and area following injections into their terminal fields

Fig. 9. Distribution of anterogradely labeled BDA-containing fibers following a control injection of BDA into the caudal part of the CM (injection #128). Dense labeling is present in the Ins, the caudal ventral part of the CP, and the SI, but BDA-containing fibers are absent from the hypothalamus. Scale bar⫽1 mm.

The retrograde tracer CTB was injected into three areas: the ventral part of the lateral septal nucleus, the fundus striati, and the anterior hypothalamic nucleus. Although the injection sites were not very small, they were restricted to the target areas, as shown for the ventral part of the lateral septal nucleus (Fig. 11A). All three injections resulted in

1258

J. Wang et al. / Neuroscience 138 (2006) 1245–1263

Fig. 10. Confocal microscope image demonstrates the co-localization of TIP39-ir (green) with BDA-containing fibers (red) in the lateral septum after injection of BDA into the rSPF. Arrowheads indicate double-labeled fibers and terminals that contain both BDA and TIP39-ir. Note that there are fibers or terminals containing only either TIP39 immunoreactivity or BDA as well as double labeled ones. Scale bar⫽10 ␮m.

dominantly ipsilateral labeling of cells distributed uniformly in the rSPF (Fig. 11B) and the SPA (Fig. 11C). The number of retrogradely labeled cells was comparable to the number of TIP39-labeled cells following injection into the ventral part of the lateral septal nucleus (Fig. 11B, C), and somewhat less following injections into the anterior hypothalamic nucleus (Fig. 11E) and fundus striati (Fig. 11F). Double-labeling of CTB with TIP39 revealed that almost all retrogradely labeled cells were positive for TIP39 following injection into the fundus striati (Fig. 11F), about half of the retrogradely labeled cells were positive for TIP39 following injection into the ventral part of the lateral septal nucleus (Fig. 11B, C), and very few double-labeled cells were found following injection into the anterior hypothalamic nucleus (Fig. 11E). We found no double-labeling of CTB and TH immunoreactivity in the SPA following injection of CTB into any of the three projection areas, as demonstrated for the ventral part of the lateral septum (Fig. 11D). A uniform distribution of double-labeled cells in the SPF and SPA was observed following injections into the ventral part of the lateral septal nucleus and the fundus striati (Fig. 11B, C, E).

DISCUSSION The SPF and the SPA The SPF has been considered a thalamic nucleus with unknown function (Faul and Mehler, 1985; Turner and Herkenham, 1991), or a member of the posterior intralaminar group of nuclei (LeDoux et al., 1985; Rub et al., 2002). Most authors, however, do not consider it to be one of the midline or intralaminar nuclei (Berendse and Groenewegen, 1991; Van der Werf et al., 2002). Modern brain atlases are consistent in their definition of the SPF, a cytoarchitectonically distinct structure with circular appearance on coronal sections (Paxinos and Watson, 1998; Swan-

son, 2004; Paxinos and Watson, 2005). This terminology was followed in this manuscript. This is a confined use as compared with the original definition of the SPF as a large area below the fasciculus retroflexus (Rioch, 1929). Moreover, influential comprehensive descriptions of the thalamus do not mention the SPF or the area corresponding to the rostral central gray as part of the thalamus (Jones, 1985; Price, 1995). The complex shape and the difficulty in defining the area without neurochemical markers probably contributed to the introduction of different terminologies. The term “SPA” and its “compartments” corresponding to the SPF and the rostral central gray were introduced to describe connectional data (Moriizumi and Hattori, 1992). The area caudal and dorsal to the SPF has also been called “subfascicular area” (Peschanski and Mantyh, 1983). The term “periventricular gray of the caudal thalamus” was used to describe the location of A11 dopaminergic cells (Hökfelt et al., 1979; Skagerberg and Lindvall, 1985), and to describe sites for stimulus-induced anesthesia in the region (Rhodes and Liebeskind, 1978; Boivie and Meyerson, 1982). Another classification included the SPF together with the interstitial nucleus of Cajal and the nucleus of Darkschewitsch as the “perifascicular region” based on their locations and their projections to the inferior olive (Saint-Cyr, 1987). However, the inclusion of the SPF was later debated (Onodera and Hicks, 1995). In our definition, the SPF is a continuous group of cells in the posterior thalamus. The rostral part of the nucleus, which contains TIP39- and TH-immunopositive neurons is located rostral to the level of the fasciculus retroflexus. The caudal part can be followed caudally just ventral to the fasciculus retroflexus at the diencephalic–midbrain junction and it is topographically connected to the laterally and caudally located parvicellular SPF. This part of the SPF is considered a component of the SPA. The major portion of this area is located between the fasciculus retroflexus and

J. Wang et al. / Neuroscience 138 (2006) 1245–1263

1259

Fig. 11. Confocal images of retrogradely labeled cell bodies are shown in the SPA following injections into the lateral septum (A–D), the AH (E), and the FS (F). Cells double labeled with the retrograde tracer CTB (red) and with TIP39 (green) are indicated by white arrowheads. (A) An injection site is shown in the ventral part of the lateral septum. (B) CTB-labeled cells are present in the SPF following injection of the retrograde tracer into the ventral part of lateral septum. (C) CTB-labeled cells are demonstrated in the SPA following the same retrograde tracer injection. Numerous double-labeled cells are evenly distributed ipsilateral to the injection site but only a few retrogradely labeled cells are present contralaterally. (D) Cells labeled with either CTB (red) or by a TH antibody (green) but no double-labeled cells are observed following injection of the retrograde tracer into the ventral part of lateral septum. (E) Cells containing either CTB (red) or TIP39 (green) immunoreactivity but no double-labeled cells are observed following injection of the retrograde tracer into the AH. (F) CTB-labeled cells are demonstrated in the SPA following injection of the retrograde tracer into the FS. Note that most CTB-labeled cells are also TIP39 ir. Scale bars⫽100 ␮m.

the most caudal portion of the third ventricle (Fig. 1C, D) as far dorsal as the level of the parafascicular nucleus. We found two neurochemical markers, TIP39 and TH, which are both contained in perikarya distributed uniformly in the rSPF and the dorsomedial portion of the SPA, but

not in neighboring areas. It has been previously demonstrated that TIP39 immunoreactivity corresponds to TIP39 expressing cells bodies (Dobolyi et al., 2003b) whereas TH immunoreactivity in this region corresponds to the A11 dopaminergic cells (Hökfelt et al., 1976).

1260

J. Wang et al. / Neuroscience 138 (2006) 1245–1263

Efferent projections of the rSPF and the SPA Neurons in the rSPF send a large number of efferent fibers to forebrain areas with a somewhat different projection pattern than neurons located in the SPA. In general, the density of anterogradely labeled fibers observed in most of the forebrain regions is lower for rostral SPF injections than for the caudal ones. The superficial layers of the frontal association, orbital, insular, auditory, ectorhinal, and perirhinal cortices receive many fibers from the rSPF but not from the SPA, or areas around the SPF. In turn, the core portion of the nucleus accumbens, the globus pallidus, the anterodorsal, mediodorsal, and central medial thalamic nuclei, the medial preoptic area, as well as the midbrain raphe nuclei receive no, or only solitary, fibers from the rSPF. Furthermore, differences in projections to the amygdala from the rSPF and the SPA are significant. The innervation of the amygdala by the rSPF is minor. The specificity of the projections to the injection sites is likely, because high molecular weight BDA as anterograde tracer is not taken up by passing fibers (Reiner et al., 2000), and because we observed profoundly different projection patterns following injections into neighboring nuclei. Notably, injections into the parvicellular part of the ventral posterior thalamic nucleus, the central medial thalamic nucleus, the interstitial nucleus of Cajal, and the nucleus of Darkschewitsch resulted in restricted projections that were most significant in deep layers of the insular cortex, the insular cortex–ventral caudate-putamen– central amygdala regions, and in the oculomotor nuclei, respectively. These projections are in agreement with published data for these sites (Berendse and Groenewegen, 1991; Onodera and Hicks, 1995; Wang and Spencer, 1996; Nakashima et al., 2000; Van der Werf et al., 2002). We also observed that the ascending projections of the rostral periaqueductal gray are also much more restricted than those of the SPA and rSPF, which is in agreement with published data (Cameron et al., 1995). In addition, the projections from rSPF to the lateral septum and the fundus striati were confirmed by injecting retrograde tracer into these three areas and mapping labeled cell bodies in the SPF. There are many brain areas where more fibers contained the anterograde tracer than TIP39 immunoreactivity, including the superficial layers of the cortex, the medial septal nucleus, the caudate-putamen, several midline, intralaminar and ventral thalamic nuclei, the lateral preoptic and lateral hypothalamic areas, the area around the paraventricular nucleus of the hypothalamus, the anterior, ventromedial hypothalamic nuclei and some of the mamillary nuclei. The absence or very low density of TIP39 fibers in these regions suggests that the rSPF and the SPA contain neurons other than TIP39 synthesizing ones that project to these regions. This finding is further supported by the very low percentage of TIP39-positive cells among the retrogradely labeled cells in the rSPF and the SPA following injection of retrograde tracer into the anterior hypothalamic nucleus. We found no fibers in the brain double labeled with BDA and TH. This finding is consistent with previous stud-

ies demonstrating that the major target of the A11 dopaminergic cells is the spinal cord (Bjorklund and Skagerberg, 1979; Hökfelt et al., 1979; Skagerberg and Lindvall, 1985). Since there were brain regions with high BDA fiber density that did not contain either TIP39 or TH, it is likely that there are projection neurons in the SPA-rSPF that do not contain either of these markers. TIP39 innervation of the forebrain: TIP39 cells as major output neurons of the posterior thalamus The patterns of anterogradely labeled fibers following injections into the rSPF and the SPA were similar to the distribution of TIP39-ir fibers in many areas of the brain (Table 1). There was a high density and similar distribution of both anterogradely labeled fibers and TIP39-ir fibers in the limbic and ectorhinal cortices, the olfactory tubercle, the shell part of the nucleus accumbens, the lateral septal nucleus and the bed nucleus of the stria terminalis, the substantia innominata, the central amygdaloid nucleus and the fundus striati, the paraventricular and posterior intralaminar nuclei of the thalamus, the medial preoptic area, the posterior hypothalamic nucleus, the fields of Forel, the zona incerta and the periaqueductal gray. Fibers doublelabeled with BDA and TIP39 were present in all of these areas. Together these data suggest that TIP39-ir fibers in these regions originate in the medial part of the posterior thalamus. This finding was further confirmed by injections of retrograde tracers into the lateral septal nucleus and the fundus striati that resulted in labeled TIP39 perikarya both in the SPF and SPA. The percentage of double-labeled cells was high, supporting the idea that TIP39 neurons are a major type of output neuron in both the SPF and SPA. The fact that many brain areas contained fibers double-labeled with BDA and TIP39 following injections in both subparafascicular regions, and the finding that labeled TIP39 cell bodies were evenly distributed in the SPF and the SPA following injections of retrograde tracer into the lateral septum and the fundus striati suggest that TIP39 neurons in the rSPF and the SPA have similar projection patterns, and that TIP39 neurons constitute a single population. Some of the hypothalamic cell areas including the anterior, paraventricular, arcuate and dorsomedial nuclei did not contain fibers double-labeled with TIP39 and BDA. Most of these nuclei actually contain a higher density of TIP39-ir than BDA-labeled fibers. Together these data suggest that TIP39 fibers in these regions originate in TIP39 cells other than those in the rSPF or the SPA. TIP39 cell bodies are present in the parvicellular subparafascicular and the posterior intralaminar thalamic nuclei and the surrounding area medial to the medial geniculate body (Dobolyi et al., 2003b). Lesions of this area markedly reduced the density of TIP39-ir fibers in the hypothalamus (Dobolyi et al., 2003a). In addition, these lesions clearly showed that the majority of TIP39-ir fibers and terminals in the amygdala emanate from neurons in the parvicellular subparafascicular and the posterior intralaminar thalamic nuclei. Part of the trajectory of anterogradely labeled fibers can be followed from the rSPF to the forebrain. A group of fibers leaves the nucleus dorsolaterally, then after a hook-shaped

J. Wang et al. / Neuroscience 138 (2006) 1245–1263

1261

Fig. 12. Summary of major projections of neurons in the SPF and area. Projections to the infralimbic, prelimbic and insular cortices, the basal forebrain, the lateral septum, the BST, the MPA, the AH, the PH, the FS, the AMY, and the paraventricular thalamic nucleus are indicated.

turn they enter the epithalamus and run rostralwards in the paraventricular nucleus of the thalamus (Fig. 12). Some other fibers run rostrally from the nucleus, just dorsal to the medial lemniscus and terminate, most probably in thalamic nuclei. Another group, which comprises the most substantial number of fibers, joins the zona incerta and the supraoptic decussations and enters the basal forebrain/substantia innominata area. From here, the fibers divide into branches to limbic cortical, septal and amygdaloid areas (Fig. 12). The fine topography and the terminal pattern of these fibers will be the subjects of further studies. Possible functions of the SPF and area Little functional data are available regarding the SPF and SPA. Some evidence suggests a role of the SPA in nociceptive modulation. Cells in the area are activated in response to noxious stimuli (Dong et al., 1978; Sugiyama et al., 1992) and analgesia can be induced by stimulating the area. In addition to the activation of cells in the SPA in response to nociceptive stimuli (Dong et al., 1978; Sugiyama et al., 1992), cells in the area also exhibit c-fos activation following exposure to cold (Kiyohara et al., 1995; Miyata et al., 1995) and high intensity noise (Palkovits et al., 2004). The resulting corticosterone response might involve projections to the medial prefrontal cortex, the lateral septum, the bed nucleus of the stria terminalis, the medial preoptic area, the periparaventricular region of the hypothalamus, the amygdala, and the paraventricular thalamic nucleus, areas which have all been shown to be able to affect the corticosterone response driven by the paraventricular

nucleus (Herman et al., 2003). Other projections of the SPF and the SPA reach regions, such as the intermediate part of the lateral septum, the anteromedial division of the medial amygdaloid nucleus ventromedial and anterior hypothalamic nuclei, that are parts of a neural network controlling defensive behavior (Fuchs et al., 1985; Risold and Swanson, 1997; Canteras, 2002). Amygdaloid projections could affect emotional responses including the fear response (LeDoux, 2003). Projections to midline thalamic nuclei and to cholinergic areas of the basal forebrain could contribute to awareness (Smythies, 1997). Several areas receiving projections from the SPA are involved in cardiovascular and respiratory regulation. Such target regions include the amygdala, the A5 noradrenergic cell group, the subcoeruleus area, and the pontine raphe nucleus (Loewy and McKellar, 1980). Based on all of the projections together, we hypothesize that the SPA is involved in the central regulation of the generalized stress response (Chrousos and Gold, 1992). Although it is premature to speculate about functions of subgroups of cells within the SPF and the SPA, it is worth mentioning that TIP39 cells preferentially project to limbic areas that potentially affect the corticosterone response.

CONCLUSION In conclusion, our results allowed us to identify the SPF and SPA neurochemically using TIP39 and tyrosine-hydroxylase as markers distributed throughout the elongated rostrocaudal axis of the region. This study is the first to

1262

J. Wang et al. / Neuroscience 138 (2006) 1245–1263

systematically describe the efferent connections of both the SPF and the SPA, and to demonstrate dense limbic and autonomic projections. In addition we identified TIP39 neurons as major output neurons of the rSPF and the SPA. Acknowledgments—Support was provided by the National Institute of Mental Health Intramural Research Program. We are grateful to Dr. Éva Mezey for advice and assistance. We also appreciate help from Dr. Carolyn Smith and the NINDS Light Imaging Facility with confocal microscopy.

REFERENCES Beckstead RM (1984) The thalamostriatal projection in the cat. J Comp Neurol 223(3):313–346. Bentivoglio M, Molinari M (1984) The interrelations between cell groups in the caudal diencephalon of the rat projecting to the striatum and to the medulla oblongata. Exp Brain Res 54(1): 57– 65. Berendse HW, Groenewegen HJ (1991) Restricted cortical termination fields of the midline and intralaminar thalamic nuclei in the rat. Neuroscience 42:73–102. Bjorklund A, Skagerberg G (1979) Evidence for a major spinal cord projection from the diencephalic A11 dopamine cell group in the rat using transmitter-specific fluorescent retrograde tracing. Brain Res 177:170 –175. Boivie J, Meyerson BA (1982) A correlative anatomical and clinical study of pain suppression by deep brain stimulation. Pain 13: 113–126. Cameron AA, Khan IA, Westlund KN, Cliffer KD, Willis WD (1995) The efferent projections of the periaqueductal gray in the rat: a Phaseolus vulgaris-leucoagglutinin study. I. Ascending projections. J Comp Neurol 351:568 –584. Canteras NS (2002) The medial hypothalamic defensive system: hodological organization and functional implications. Pharmacol Biochem Behav 71:481– 491. Chrousos GP, Gold PW (1992) The concepts of stress and stress system disorders. Overview of physical and behavioral homeostasis. JAMA 267:1244 –1252. Dobolyi A, Palkovits M, Bodnar I, Usdin TB (2003a) Neurons containing tuberoinfundibular peptide of 39 residues project to limbic, endocrine, auditory and spinal areas in rat. Neuroscience 122: 1093–1105. Dobolyi A, Palkovits M, Usdin TB (2003b) Expression and distribution of tuberoinfundibular peptide of 39 residues in the rat central nervous system. J Comp Neurol 455:547–566. Dong WK, Ryu H, Wagman IH (1978) Nociceptive responses of neurons in medial thalamus and their relationship to spinothalamic pathways. J Neurophysiol 41:1592–1613. Faul RLM, Mehler WR (1985) Thalamus. In: The rat nervous system, Vol. I (Paxinos G, ed), pp 129 –168. Sydney: Academic Press. Fleetwood-Walker SM, Hope PJ, Mitchell R (1988) Antinociceptive actions of descending dopaminergic tracts on cat and rat dorsal horn somatosensory neurones. J Physiol 399:335–348. Fuchs SA, Edinger HM, Siegel A (1985) The organization of the hypothalamic pathways mediating affective defense behavior in the cat. Brain Res 330:77–92. Guo ZL, Li P, Longhurst JC (2002) Central pathways in the pons and midbrain involved in cardiac sympathoexcitatory reflexes in cats. Neuroscience 113:435– 447. Herman JP, Figueiredo H, Mueller NK, Ulrich-Lai Y, Ostrander MM, Choi DC, Cullinan WE (2003) Central mechanisms of stress integration: hierarchical circuitry controlling hypothalamo-pituitaryadrenocortical responsiveness. Front Neuroendocrinol 24:151– 180. Hökfelt T, Johansson O, Fuxe K, Goldstein M, Park D (1976) Immunohistochemical studies on the localization and distribution of

monoamine neuron systems in the rat brain. I. Tyrosine hydroxylase in the mes- and diencephalon. Med Biol 54:427– 453. Hökfelt T, Phillipson O, Goldstein M (1979) Evidence for a dopaminergic pathway in the rat descending from the A11 cell group to the spinal cord. Acta Physiol Scand 107:393–395. Hosobuchi Y (1983) Combined electrical stimulation of the periaqueductal gray matter and sensory thalamus. Appl Neurophysiol 46:112–115. Jones EG (1985) The thalamus. New York: Plenum. Kiyohara T, Miyata S, Nakamura T, Shido O, Nakashima T, Shibata M (1995) Differences in Fos expression in the rat brains between cold and warm ambient exposures. Brain Res Bull 38:193–201. LaBuda CJ, Dobolyi A, Usdin TB (2004) Tuberoinfundibular peptide of 39 residues produces anxiolytic and antidepressant actions. Neuroreport 15:881– 885. LeDoux J (2003) The emotional brain, fear, and the amygdala. Cell Mol Neurobiol 23:727–738. LeDoux JE, Ruggiero DA, Reis DJ (1985) Projections to the subcortical forebrain from anatomically defined regions of the medial geniculate body in the rat. J Comp Neurol 242:182–213. Loewy AD, McKellar S (1980) The neuroanatomical basis of central cardiovascular control. Fed Proc 39:2495–2503. Miyata S, Ishiyama M, Shido O, Nakashima T, Shibata M, Kiyohara T (1995) Central mechanism of neural activation with cold acclimation of rats using Fos immunohistochemistry. Neurosci Res 22:209 –218. Moriizumi T, Hattori T (1991) Non-dopaminergic projection from the subparafascicular area to the temporal cortex in the rat. Neurosci Lett 129:127–130. Moriizumi T, Hattori T (1992) Anatomical and functional compartmentalization of the subparafascicular thalamic nucleus in the rat. Exp Brain Res 90:175–179. Moriizumi T, Leduc-Cross B (1992) Is there a dopaminergic projection from the A11 catecholamine cell group to the amygdala? Exp Brain Res 88(2):451– 454. Musil SY, Olson CR (1988) Organization of cortical and subcortical projections to anterior cingulate cortex in the cat. J Comp Neurol 272(2):203–218. Nakashima M, Uemura M, Yasui K, Ozaki HS, Tabata S, Taen A (2000) An anterograde and retrograde tract-tracing study on the projections from the thalamic gustatory area in the rat: distribution of neurons projecting to the insular cortex and amygdaloid complex. Neurosci Res 36:297–309. Nocjar C, Roth BL, Pehek EA (2002) Localization of 5-HT(2A) receptors on dopamine cells in subnuclei of the midbrain A10 cell group. Neuroscience 111:163–176. Onodera S, Hicks TP (1995) Patterns of transmitter labelling and connectivity of the cat’s nucleus of Darkschewitsch: a wheat germ agglutinin-horseradish peroxidase and immunocytochemical study at light and electron microscopical levels. J Comp Neurol 361: 553–573. Ottersen OP, Ben-Ari Y (1979) Afferent connections to the amygdaloid complex of the rat and cat. I. Projections from the thalamus. J Comp Neurol 187(2):401– 424. Palkovits M, Dobolyi A, Helfferich F, Usdin TB (2004) Localization and chemical characterization of the audiogenic stress pathway. Ann N Y Acad Sci 1018:16 –24. Paxinos G, Watson C (1998) The rat brain in stereotaxic coordinates. San Diego: Academic Press. Paxinos G, Watson C (2005) The rat brain in stereotaxic coordinates. San Diego: Academic Press. Peschanski M, Mantyh PW (1983) Efferent connections of the subfascicular area of the mesodiencephalic junction and its possible involvement in stimulation-produced analgesia. Brain Res 263: 181–190. Price JL (1995) Thalamus. In: The rat nervous system (Paxinos G, ed), pp 629 – 648. San Diego: Academic Press.

J. Wang et al. / Neuroscience 138 (2006) 1245–1263 Reiner A, Veenman CL, Medina L, Jiao Y, Del Mar N, Honig MG (2000) Pathway tracing using biotinylated dextran amines. J Neurosci Methods 103:23–37. Rhodes DL, Liebeskind JC (1978) Analgesia from rostral brain stem stimulation in the rat. Brain Res 143:521–532. Rioch DK (1929) Studies on the diencephalon of carnivora. Part I. The nuclear configuration of the thalamus, epithalamus, and hypothalamus of the dog and cat. J Comp Neurol 49:1–120. Risold PY, Swanson LW (1997) Connections of the rat lateral septal complex. Brain Res Brain Res Rev 24:115–195. Rub U, Del Tredici K, Del Turco D, Braak H (2002) The intralaminar nuclei assigned to the medial pain system and other components of this system are early and progressively affected by the Alzheimer’s disease-related cytoskeletal pathology. J Chem Neuroanat 23:279 –290. Saint-Cyr JA (1987) Anatomical organization of cortico-mesencephalo-olivary pathways in the cat as demonstrated by axonal transport techniques. J Comp Neurol 257:39 –59. Skagerberg G, Lindvall O (1985) Organization of diencephalic dopamine neurones projecting to the spinal cord in the rat. Brain Res 342:340 –351. Smythies J (1997) The functional neuroanatomy of awareness: with a focus on the role of various anatomical systems in the control of intermodal attention. Conscious Cogn 6:455– 481. Sugimura Y, Murase T, Ishizaki S, Tachikawa K, Arima H, Miura Y, Usdin TB, Oiso Y (2003) Centrally administered tuberoinfundibular peptide of 39 residues inhibits arginine vasopressin release in conscious rats. Endocrinology 144:2791–2796. Sugiyama K, Ryu H, Uemura K (1992) Identification of nociceptive neurons in the medial thalamus: morphological studies of nociceptive neurons with intracellular injection of horseradish peroxidase. Brain Res 586:36 – 43. Swanson LW (2004) Brain maps: structure of the rat brain. San Diego: Academic Press. Takada M (1990) The A11 catecholamine cell group: another origin of the dopaminergic innervation of the amygdala. Neurosci Lett 118(1):132–135.

1263

Takada M (1993) Widespread dopaminergic projections of the subparafascicular thalamic nucleus in the rat. Brain Res Bull 32(3):301–309. Takada M, Li ZK, Hattori T (1988) Single thalamic dopaminergic neurons project to both the neocortex and spinal cord. Brain Res 455(2):346 –352. Turner BH, Herkenham M (1991) Thalamoamygdaloid projections in the rat: a test of the amygdala’s role in sensory processing. J Comp Neurol 313:295–325. Usdin TB (2000) The PTH2 receptor and TIP39: a new peptidereceptor system. Trends Pharmacol Sci 21:128 –130. Usdin TB, Hoare SR, Wang T, Mezey E, Kowalak JA (1999) TIP39: a new neuropeptide and PTH2-receptor agonist from hypothalamus. Nat Neurosci 2:941–943. Van der Werf YD, Witter MP, Groenewegen HJ (2002) The intralaminar and midline nuclei of the thalamus. Anatomical and functional evidence for participation in processes of arousal and awareness. Brain Res Brain Res Rev 39:107–140. Wang SF, Spencer RF (1996) Spatial organization of premotor neurons related to vertical upward and downward saccadic eye movements in the rostral interstitial nucleus of the medial longitudinal fasciculus (riMLF) in the cat. J Comp Neurol 366:163–180. Wang T, Palkovits M, Rusnak M, Mezey E, Usdin TB (2000) Distribution of parathyroid hormone-2 receptor-like immunoreactivity and messenger RNA in the rat nervous system. Neuroscience 100: 629 – 649. Ward HL, Small CJ, Murphy KG, Kennedy AR, Ghatei MA, Bloom SR (2001) The actions of tuberoinfundibular peptide on the hypothalamo-pituitary axes. Endocrinology 142:3451–3456. Yasui Y, Nakano K, Mizuno N (1992) Descending projections from the subparafascicular thalamic nucleus to the lower brain stem in the rat. Exp Brain Res 90:508 –518. Yeung JC, Yaksh TL, Rudy TA (1977) Concurrent mapping of brain sites for sensitivity to the direct application of morphine and focal electrical stimulation in the production of antinociception in the rat. Pain 4:23– 40.

(Accepted 1 December 2005) (Available online 3 February 2006)