Sprague–Dawley male rats, weighing 280–380 g, were obtained from Janvier (Le Genest-Saint-Isle, France). They were housed under 12 h light: 12 h dark cycle at a constant room temperature and had free access to water and standard laboratory diet. All animal use and care protocols were in accordance with institutional guidelines (all protocols were approved and investigators authorized).

Rats were anesthetized with an intramuscular (IM) injection of a mixture of xylazine and ketamine (Vetokinol, 1 mg/100 g of body weight and 10 mg/100 g, respectively) and placed in a stereotaxic instrument.

Rats received a unilateral iontophoretic injection of 10% Fluorogold (FG, Interchim) solution diluted in 0.9% NaCl. Coordinates were taken according to the Paxinos and Watson stereotaxic atlas (Paxinos and Watson, 2005). Injections were made in different sites of the basal telencephalon (medial septal complex, nucleus of the diagonal band, substantia innominata, globus pallidus, and dorsal striatum). Glass micropipettes (tip diameter: 30–50 μm) were used to inject the FG iontophoretically into these regions using an intermittent current of 5 μA and 7 s on/off time for 5 min. The micropipette was left in place for another 5 min before being removed to avoid FG diffusion along the micropipette track.

After 10 days survival time, rats were deeply anesthetized with intraperitoneal injection (IP) of Pentobarbital (CEVA, 50 mg/kg). Animals were perfused transcardially with 0.9% NaCl followed by ice-cold 4% paraformaldehyde (PFA, Roth) fixative in 0.1 M phosphate buffer saline (PBS) at pH 7.4. Brains were removed, post-fixed in the same fixative for several hours at 4°C, immersed overnight at 4°C in a solution of 15% sucrose in 0.1 M PBS, and then quickly frozen. Brains were cut in four series of 30 μm coronal thick sections, collected in a cryoprotector solution (1:1:2 glycerol/ethylene glycol/PBS), and stored at −40°C.

After rinsing in PBS + 0.3% Triton X100, free-floating sections were incubated with the primary anti-FG antiserum raised in rabbit (polyclonal, Oncogene Research Products) at a dilution of 1:3000 in PBS containing 0.3% Triton X100, 1% bovine serum albumin, 10% lactoproteins, and 0.01% sodium azide, during 65 h at 4°C. Then, free-floating sections were incubated for 24 h at 4°C in a solution of biotinylated goat anti-rabbit IgG antibody (Vector Laboratories) at a dilution of 1:1000 in PBS Triton. Finally, sections were placed in the mixed avidin-biotin horseradish peroxidase (HRP) complex solution (ABC Elite Kit, Vector Laboratories) for 1 h at room temperature. The peroxidase complex was visualized by an exposure to a chromogen solution containing 0.04% 3,3′diaminobenzidine tetrahydrochloride (DAB, Sigma) and 0.006% hydrogen peroxide (Sigma) in PBS at pH 7.4. The reaction was stopped by extensive washing in PBS at pH 7.4. Free-floating sections were mounted on gelatin-coated slides, and then dehydrated and coverslipped with Canada balsam (Roth). An adjacent series was always stained in a solution of 1% toluidine blue (Roth) in water to serve as a reference series for cytoarchitectonic purposes.

Sections were incubated with the anti-MCH antibody (rabbit polyclonal, our laboratory, Risold et al., 1992) dissolved in PBS containing 0.3% Triton X100, 1% bovine serum albumin, 10% lactoproteins, and 0.01% sodium azide at 1:1000 for 65 h at 4°C. Tissues were then incubated with Alexa Fluor 488 goat anti-rabbit IgG antibody (Invitrogen) diluted in PBS-T at 1:1000 for 2 h at room temperature.

After the first staining, sections were incubated with the anti-CART antibody (Cocaine and Amphetamine Regulated Transcript, mouse monoclonal, generously provided by Dr J. T. Clausen, Novo Nordisk, Denmark, 1:1000) dissolved in PBS-T for 65 h at 4°C. Tissues were then incubated with Alexa Fluor 555 donkey anti-mouse IgG antibody (Invitrogen) diluted in PBS-T at 1:1000 for 2 h at room temperature. Finally, free-floating sections were mounted on gelatin-coated slides and coverslipped with 60:40 glycerol:PBS-T.

The auto-fluorescence of FG was observed under UV illumination.

The MCH is revealed by the technique described above. Then, sections were incubated with the anti-parvalbumin (mouse monoclonal, Swant) or anti-ChAT (goat polyclonal, AB144P Chemicon) antibodies dissolved in PBS containing 0.3% Triton X100, 1% bovine serum albumin, 10% lactoproteins and 0.01% sodium azide at 1:1000 for 65 h at 4°C. Tissues were then incubated with Alexa Fluor 555 donkey anti-mouse IgG antibody (Invitrogen) or Cyanine-3 donkey anti-goat IgG antibody (Jackson Immunoresearch) diluted in PBS-T at 1:1000 for 2 h at room temperature. Finally, free-floating sections were mounted on gelatin-coated slides and coverslipped with 60:40 glycerol:PBS-T.

In this work, we mostly employed the nomenclature of Swanson (2004). The magnocellular cholinergic neurons in the basal forebrain are abundant across the borders of the medial septal complex (medial septal nucleus and nucleus of the diagonal band), the magnocellular preoptic nucleus, the substantia innominata and most the internal segment of the globus pallidus. However, in the globus pallidus and the medial septal complex, these cholinergic neurons are segregated within a medial sector (peripheral in the diagonal band nucleus), while an external part (central in the diagonal band nucleus) is rich in parvalbumin-containing neurons (Kiss et al., 1990; Hontanilla et al., 1998; Henderson et al., 2004; Risold, 2004; Croizier et al., 2010; Mallet et al., 2012). Such differentiation of parvalbumin positive vs. cholinergic-rich region is not as clear in the substantia innominata and in the magnocellular preoptic nucleus.

MCH projections are sparse within the dorsal striatum. Fibers labeled for MCH traveled essentially through ventral components of the internal capsule. Only the caudal tail of the caudoputamen nucleus contained some axons providing a very moderate to light innervations of the neuropil.

Four FG injections were aimed at the dorsal striatum, and three were restricted to respectively, dorsal, ventral, or ventrolateral portions of the caudoputamen nucleus (Figure 5A). Retrogradely labeled cells were observed in the cerebral cortex, the globus pallidus, intralaminar nuclei of the thalamus, the subthalamic nucleus, the dorsal raphe nucleus, and the substantia nigra. Only one or two per experiment were found within the LHA/zona incerta. These very few cells were labeled by the MCH-antiserum (not shown).

MCH labeled axons were examined in proximity of retrogradely labeled cells in the following regions.

Both the globus pallidus and the medial septal complex are divided into parvalbumin rich and parvalbumin poor compartments (Kiss et al., 1990; Hontanilla et al., 1998; Henderson et al., 2004; Risold, 2004; Croizier et al., 2010; Mallet et al., 2012). By contrast, the whole substantia innominata contains intermixed neuron populations (Lagos et al., 2012; Lu et al., 2013). MCH axons clearly targeted the cholinergic-rich/parvalbumin-poor regions of the medial septal nucleus, nucleus of the diagonal band, and of the globus pallidus, but they are diffuse within the substantia innominata. MCH and cholinergic neurons have been reported to interact in the medial septal complex and to be involved in sleep control and hippocampal functions such as memory formation (Chung et al., 2011; Lagos et al., 2012; Jego et al., 2013; Lu et al., 2013). However, MCH axons may as well-interact with cholinergic neurons in the substantia innominata and globus pallidus. The cholinergic rich basal telencephalon contains other neuronal types such as GABAergic and glutamatergic neurons that are far more abundant than cholinergic cells (Risold, 2004; Gritti et al., 2006), but little attention has been devoted to date to the innervation that MCH axons may provide to these cells. We observed buttons, and putative synaptic contacts targeting ChAT negative and therefore non-cholinergic cells. This is a topic that will clearly need to be investigated in the future to understand the role of the MCH projections into these regions.

Depending on the location of the injection sites, the numbers, and distributions of FG-labeled neurons were markedly different within the hypothalamus. Injection sites within the medial septal complex retrogradely labeled far more neurons in the hypothalamus than injections centered into the globus pallidus. This was expected because the medial forebrain bundle originates in and innervates the septal region and the substantia innominata, while the globus pallidus is associated to the internal capsule and the cerebral peduncle. We also observed a topographical arrangement in the origin of these projections, with neurons projecting to the medial septal complex being more medial than those projecting to the anterior substantia innominata or globus pallidus. Because some of these projections continue farther to reach the pallium, this topographical organization is clearly reminiscent of that described by Saper (1985) after cortical retrograde tracer injections.

Neurons double labeled with MCH and FG display similar distribution patterns in the hypothalamus. There are more MCH neurons projecting into the medial septal complex than into the globus pallidus. However, several points need to be emphasized:

First, although more abundant after the medial septal complex injections, overall these MCH/FG double-labeled neurons were intermixed with many retrogradely labeled non-MCH neurons in the hypothalamus. By contrast, after FG injections into the globus pallidus, most retrogradely labeled neurons in the hypothalamus were MCH positive. This indicates that MCH neurons may have an important and specific role in relaying the results of the hypothalamic (lateral hypothalamic) processing into the striatonigral system.

Neurons that were retrogradely labeled after FG injections centered into the caudoputamen nucleus received overall a very modest MCH innervation. Only the very ventromedial sector of the pars compacta of the substantia nigra received a clear MCH input, but this region might very well be also associated to the ventral striatum (nucleus accumbens) (Gerfen and Wilson, 1996).

Only significant putative contacts concerned nuclei for which the dorsal striatum is a secondary target or that have multiple projection sites including the striatum. For example, the dorsal raphe nucleus contained quite abundant MCH buttons, some of which adjacent to FG positive neurons after striatal injections. However, the dorsal raphe nucleus project in multiple telencephalic sites and its neurons often send collaterals in many nuclei or territories (Waselus et al., 2011). Therefore, the MCH input that we observed in this nucleus cannot be interpreted as a strong evidence of exclusive indirect dorsal striatal control.

We noted an intense MCH innervation in the subthalamic nucleus. Some neurons of this nucleus were retrogradely labeled after striatal injections, but its main projections are for the globus pallidus as well as the pars reticulata of the substantia nigra. MCH projections in the subthalamic nucleus have not been investigated so far. MCH-R1 seems not to be abundantly expressed in the rodent subthalamic nucleus (Saito et al., 2001). However, MCH neurons contains other neuropeptides as CART and Nesfatin, but most importantly, they are GABAergic (Sapin et al., 2010; Del Cid-Pellitero and Jones, 2012). Therefore, they may act as inhibitory GABAergic neurons. These projections into the subthalamic nucleus were noted in other species (Croizier et al., 2010; Chometton et al., 2014), and should deserve an increased attention in the future.

The hypothesis that the MCH system is anatomically associated with the extrapyramidal pathway was advanced by Karl Knigge which unfortunately published too few papers on this topic (Knigge et al., 1996). He advocated quite elegantly that these neurons are in good position to influence the mesostriatal pathway. Since, other authors have convincingly shown that MCH acts on parts of the nucleus accumbens in which a high expression of MCH-R1 is reported (Georgescu et al., 2005; Pissios et al., 2008; Guesdon et al., 2009; Sears et al., 2010; Hopf et al., 2013). Interactions between MCH and mesostriatal pathways have also been reported (Chung et al., 2011). In the present study, we observed that MCH influence on the basal nuclei/extrapyramidal circuitry involved mostly pallidal nuclei as well as the subthalamic nucleus in the rat.

Classically, descending pathways from the striatum take two distinct routes: a direct one that innervates the internal segment of the globus pallidus and the substantia nigra, and an indirect one that involves the external segment of the globus pallidus, subthalamic nucleus, and the substantia nigra (Figure 6). Therefore, MCH axons are susceptible to influence both descending routes from the striatum in the rat: into the internal segment of the globus pallidus for the direct pathway, or by innervating the subthalamic nucleus for the indirect pathway. The role of the subthalamic nucleus within this circuitry has been reevaluated in the last decade. From a motor function, it is now admitted that this nucleus has significant cognitive roles as well (Baunez et al., 2011). It is seen now as a central hub controlling important aspects of voluntary motor output. It is an important target for treatment of Parkinson disease by deep brain stimulation approaches. The role of MCH projections into this nucleus is unclear yet, but considering that MCH neurons are active during REM sleep, these projections into the globus pallidus, and subthalamic nucleus may be important to control motor responses as the cerebral cortex is the siege of an intense electrical activity.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.