Identification of discrete, intermingled hypocretin neuronal populations

Abstract

Neurons in the lateral hypothalamic area that express hypocretin (Hcrt) neuropeptides help regulate many behaviors including wakefulness and reward seeking. These neurons project throughout the brain, including to neural populations that regulate wakefulness, such as the locus coeruleus (LC) and tuberomammilary nucleus (TMN), as well as to populations that regulate reward, such as the nucleus accumbens (NAc) and ventral tegmental area (VTA). To address the roles of Hcrt neurons in seemingly disparate behaviors, it has been proposed that Hcrt neurons can be anatomically subdivided into at least two distinct subpopulations: a “medial group” that projects to the LC and TMN, and a “lateral group” that projects to the NAc and VTA. Here, we use a dual retrograde tracer strategy to test the hypotheses that Hcrt neurons can be classified based on their downstream projections and medial/lateral location within the hypothalamus. We found that individual Hcrt neurons were significantly more likely to project to both the LC and TMN or to both the VTA and NAc than would be predicted by chance. In contrast, we found that Hcrt neurons that projected to the LC or TMN were mostly distinct from Hcrt neurons that projected to the VTA or NAc. Interestingly, these two populations of Hcrt neurons are intermingled within the hypothalamus and cannot be classified into medial or lateral groups. These results suggest that Hcrt neurons can be distinguished based on their downstream projections but are intermingled within the hypothalamus.

Graphical Abstract

A substantial number of hypocretin neurons send bifurcating projections to the nucleus accumbens and ventral tegmental area. A separate group of hypocretin neurons send bifurcating projections to the locus coeruleus and tuberomammilary nucleus. Both groups of neurons are intermingled within the hypocretin field. Other bifurcating projections from hypocretin neurons to these four downstream populations occur much more rarely.

1. Introduction

The hypocretins (Hcrts) consist of a pair of neuropeptides, hypocretin-1 and hypocretin-2 (also known as orexin A and orexin B, respectively), that are processed from the same genetic precursor, preprohypocretin (de Lecea et al., 1998; Sakurai et al., 1998). In mammals, neurons that express Hcrts are located in the hypothalamus with species-specific expression patterns (Nixon & Smale, 2007; Swanson, Sanchez-Watts, & Watts, 2005). Most Hcrt neurons are glutamatergic and cause excitatory effects on their postsynaptic targets (Date et al., 1999; de Lecea et al., 1998; Rosin, Weston, Sevigny, Stornetta, & Guyenet, 2003; van den Pol, Gao, Obrietan, Kilduff, & Belousov, 1998). Through their downstream connections, Hcrt neurons regulate many behaviors and physiological processes (Graebner, Iyer, & Carter, 2015) and play an especially prominent role in regulating wakefulness and reward seeking behaviors (A. Adamantidis & de Lecea, 2008; Alexandre, Andermann, & Scammell, 2013; Carter, Borg, & de Lecea, 2009; Harris & Aston-Jones, 2006).

The role of Hcrt neurons in promoting wakefulness was first demonstrated by the finding that dysregulation of Hcrt neurons or their downstream receptors causes the sleep disorder narcolepsy (Chemelli et al., 1999; Lin et al., 1999; Nishino, Ripley, Overeem, Lammers, & Mignot, 2000; Thannickal et al., 2000). Indeed, loss of function of Hcrt neurons causes a decrease in wakefulness (Sasaki et al., 2011; Tabuchi et al., 2014; Tsunematsu et al., 2011), while intracerebroventricular injection of hypocretin-1 or stimulation of Hcrt neurons causes an increase in total wake time and frequency of sleep-to-wake transitions (A. R. Adamantidis, Zhang, Aravanis, Deisseroth, & de Lecea, 2007; Carter, Adamantidis, Ohtsu, Deisseroth, & de Lecea, 2009; Piper, Upton, Smith, & Hunter, 2000; Sasaki et al., 2011). To regulate wakefulness, Hcrt neurons project to a variety of downstream populations known to promote wakefulness and arousal, including the locus coeruleus (LC) and tuberomammilary nucleus (TMN) (Graebner et al., 2015; Peyron et al., 1998). Hcrt neuropeptides cause an increase in wakefulness when selectively microinjected into the LC (Hagan et al., 1999; Horvath et al., 1999; Walling, Nutt, Lalies, & Harley, 2004) or TMN (Huang et al., 2001; Ishizuka, Yamamoto, & Yamatodani, 2002; Schone et al., 2012). These two downstream structures also play a necessary role in Hcrt-mediated effects on wakefulness, as effects on wakefulness are diminished when there is loss of function in the LC (Carter et al., 2012; Chen et al., 2010; Choudhary, Khanday, Mitra, & Mallick, 2014) or TMN (Huang et al., 2001).

The role of Hcrt neurons in promoting reward seeking behaviors and cue-induced reinstatement has been established for a variety of drugs of abuse including cocaine (Borgland, Taha, Sarti, Fields, & Bonci, 2006; Boutrel et al., 2005; Harris, Wimmer, & Aston-Jones, 2005; Smith, See, & Aston-Jones, 2009; Smith, Tahsili-Fahadan, & Aston-Jones, 2010), ethanol (Mayannavar, Rashmi, Rao, Yadav, & Ganaraja, 2014, 2016; Shoblock et al., 2011; Srinivasan et al., 2012), nicotine (Dehkordi et al., 2017; Hollander, Lu, Cameron, Kamenecka, & Kenny, 2008; LeSage, Perry, Kotz, Shelley, & Corrigall, 2010; Plaza-Zabala, Martin-Garcia, de Lecea, Maldonado, & Berrendero, 2010), morphine (Georgescu et al., 2003; Harris et al., 2005; Harris, Wimmer, Randall-Thompson, & Aston-Jones, 2007; Narita et al., 2006; Sharf, Guarnieri, Taylor, & DiLeone, 2010; Sharf, Sarhan, & Dileone, 2008; Zarepour, Fatahi, Sarihi, & Haghparast, 2014), and heroin (Smith & Aston-Jones, 2012). These effects are thought to be mediated by direct projections to both components of the mesolimbic reward circuit, the ventral tegmental area (VTA) (Baimel & Borgland, 2015; Borgland, Storm, & Bonci, 2008; Borgland et al., 2006; España, Melchior, Roberts, & Jones, 2011; España et al., 2010; Hrabovszky et al., 2013; Muschamp et al., 2014; Srinivasan et al., 2012; Taslimi et al., 2012; Zarepour et al., 2014) and the nucleus accumbens (NAc) (Mayannavar et al., 2014, 2016; Mori, Kim, & Sasaki, 2011; Mukai et al., 2009; Sharf et al., 2008; Thorpe & Kotz, 2005).

To address the potentially dichotomous functions of Hcrt neurons in regulating both wakefulness and reward seeking behaviors, it has been proposed that Hcrt neurons can be anatomically subdivided into at least two separate subpopulations: one subpopulation is thought to reside in the medial Hcrt field and send projections to the LC and TMN; the other subpopulation is thought to reside in the lateral Hcrt field and send projections to the VTA and NAc (Harris & Aston-Jones, 2006). The hypothesis that medial Hcrt neurons mediate wakefulness is based on the observation that Fos, an indirect marker of neuronal activation, is preferentially expressed in medial Hcrt neurons during the active versus the inactive period (Estabrooke et al., 2001).

In contrast, the hypothesis that lateral Hcrt neurons mediate reward seeking is based on observations that Fos is expressed specifically in lateral Hcrt neurons following environmental cues conditioned with food or drug reward (Harris et al., 2005). Additionally, Fos is preferentially expressed in lateral Hcrt neurons following morphine administration, and these neurons project to the VTA (Richardson & Aston-Jones, 2012). In contrast, recent studies have shown that LC and VTA-projecting Hcrt neurons located in equal frequencies in the lateral and medial subdivisions of the Hcrt field (España, Reis, Valentino, & Berridge, 2005; Gonzalez, Jensen, Fugger, & Burdakov, 2012). Fos expression was observed in both lateral and medial Hcrt neurons in response to administration of antipsychotic drugs and to environmental stimuli causing positive reinforcement (Fadel, Bubser, & Deutch, 2002; McGregor, Wu, Barber, Ramanathan, & Siegel, 2011). Interestingly, Hcrt neurons can be functionally subdivided based on their electrophysiological properties (Schone, Venner, Knowles, Karnani, & Burdakov, 2011); however, retrograde tracers injected into either the LC or VTA do not seem to preferentially label either electrically active class of Hcrt neurons (Gonzalez et al., 2012). Hcrt neurons send bifurcating projections to the LC and basal forebrain (España et al., 2005), as well as to the paraventricular thalamic nucleus and NAc (E. Y. Lee & Lee, 2016), suggesting that specific Hcrt neurons could be classified based on collateral downstream projection patterns. However, it is unknown whether Hcrt neurons can be identified based on collateral projections to downstream populations, and the identification of lateral versus medial populations of Hcrt neurons remains unresolved.

The goal of this study was to test the hypothesis that Hcrt neurons can be classified by bifurcating downstream projections to the LC and VTA or to the NAc and VTA, and to further analyze whether Hcrt neurons projecting to these areas could be categorized by lateral/medial location within the Hcrt field. Specifically, we asked four overall questions: (1) Are individual Hcrt neurons likely to project to both the LC and TMN? (2) Are individual Hcrt neurons likely to project to both the VTA and NAc? (3) Are individual Hcrt neurons that project to the LC or TMN less likely to also project to the NAc and VTA (and vice versa)? (4) Finally, do Hcrt neurons that project to the LC, TMN, NAC, or VTA, or different combinations of these downstream nuclei, exhibit lateral/medial topography within the hypothalamus?

To answer these questions, we used a dual retrograde tracer approach by unilaterally injecting a green retrograde tracer into one downstream projection of Hcrt neurons and a red retrograde tracer into a separate downstream projection (Figure 1a–c). For each experiment, we hypothesized multiple potential outcomes: green and red retrograde signal could be located in distinct Hcrt neurons, each exhibiting a lateral or medial bias within the hypothalamus (Figure 1d); green and red retrograde signal could be located in distinct Hcrt neurons, with no lateral or medial bias (Figure 1e); green and red retrograde signal could be located in the same Hcrt neurons, with no lateral or medial bias (Figure 1f); green and red retrograde signal could be located in the same Hcrt neurons, with either a lateral or medial bias (Figure 1g). This dual retrograde tracer approach resolves the question of how Hcrt neurons can be classified based on downstream projections and lateral/medial topography.

Figure 1.

Potential outcomes of experiments to test for subpopulations of Hcrt neurons. (a) Representative coronal brain section immunolabeled for Hcrt neurons, showing precise location of Hcrt-positive neurons in the lateral hypothalamic area. (b) Higher power magnification of the full Hcrt field. (c) Dual retrograde tracer injection strategy. For each animal, a green retrograde tracer and a red retrograde tracer was injected into two distinct downstream brain regions to retrogradely label individual populations of Hcrt neurons. (d–g) Four potential outcomes for dual retrograde tracer experiments: (d) Hcrt neurons can be classified as projecting to two distinct downstream regions, with each population exhibiting a lateral or medial location bias; (e) Hcrt neurons can be classified as projecting to two distinct downstream regions, with no population exihibing a lateral or medial location bias; (f) Hcrt neurons send bifurcating projections to two distinct downstream regions, with no demonstration of lateral or medial location bias; (g) Hcrt neurons send bifurcating projections to two distinct downstream regions, with demonstration of a lateral or medial location bias. Grey ovals represent Hcrt neurons; green or red ovals represent green or red retrogradely labeled Hcrt neurons, respectively; white ovals represent dual labeled Hcrt neurons. fx, fornix. Scale bar = 250 μm.

2. Materials and Methods

2.1 Animals

All experiments were approved by the Institutional Animal Care and Use Committee at Williams College and were performed in accordance with the guidelines described in the U.S. National Institutes of Health Guide for the Care and Use of Laboratory Animals as well as the U.S. Public Health Service’s Policy on Humane Care and use of Laboratory Animals. We used male mice bred on a C57Bl/6 background. All mice were 7–9 weeks old at the time of surgery and perfused exactly 7 days following surgery. Mice were housed in individual cages with a 12h/12h light/dark cycle at 22 °C.

2.2 Stereotaxic surgery

Mice were anaesthetized with 4% isoflurane and placed on a stereotaxic frame (David Kopf Instruments). Once on the frame and throughout the remainder of surgical procedures, mice received 1–2% isoflurane trans-nasally. After the skull was exposed and leveled in the horizontal plane, green or red retrobeads (Lumofluor) were injected into the right LC, NAc, TMN, or VTA (Table 1). Per the recommendation of the manufacturer, red retrobeads were diluted 1:2 in 0.9% sterile saline solution and green retrobeads were undiluted. A total of 0.20 μl of retrobeads was injected at a rate of 0.10 μl/min using a microsyringe (Hamilton, #7105) controlled by an automated micropump (World Precision Instruments). For injections into the LC or NAc, the microsyringe was raised 100 μm immediately following the injection to allow for more dorsoventral spread of the tracer. For all injection sites, the microsyringe was not removed for an additional 8–10 min post-injection to allow the tracer time to diffuse. After withdrawal of the microsyringe delivering the first retrograde tracer, the second retrograde tracer was injected using a separate syringe into the same animal. Following surgical procedures, mice recovered for seven days to allow for retrograde transport prior to perfusion.

Table 1.

Stereotaxic injection coordinates relative to bregma.

Structure	Anteroposterior (mm)	Mediolateral (mm)	Dorsoventral (mm)
LC	−5.3	0.9	−4.0
NAc	+0.9	1.2	−4.5
TMN	−2.7	1.0	−5.5
VTA	−3.1	0.4	−4.8

2.3 Histology

Seven days following stereotaxic surgery, mice were anaesthetized with intraperitoneal injection of 2,2,2 tribromoethanol (Sigma, #48402) dissolved in Tert-amyl alcohol and sterile 0.9% saline. Mice were then transcardially perfused with cold 0.01M phosphate buffered saline (PBS), pH 7.4, followed by 4% paraformaldehyde in PBS. The brains were extracted, allowed to postfix overnight in 4% paraformaldehyde at 4 °C, and cryoprotected in 30% sucrose dissolved in PBS for an additional 24 h at 4 °C. Each brain was sectioned at 30 μm in the coronal plane on a microtome (Leica Microsystems). Sections were stored in PBS for no more than two days prior to immunohistochemical procedures.

To identify Hcrt neurons using immunohistochemistry, sections were washed three times in PBS with 0.2% Triton X-100 (PBST) for 10 min at room temperature. Sections were then incubated in a blocking solution composed of PBST with 3% normal donkey serum (Jackson ImmunoResearch, #017-000-121) for 15 min at room temperature. For primary antibody exposure, sections were incubated in rabbit anti-Orexin-A (1:1000; Phoenix Pharmaceuticals #H-003-30) in blocking solution overnight at 4 °C. After three 5-min washes in blocking solution, sections were incubated in Alexa Fluor 350 donkey anti-rabbit (1:250; Life Technologies, #A10039) in block solution for 1 h at room temperature. Finally, sections were washed three times in PBS. Brain sections were mounted in PBS onto SuperFrost Plus glass slides (VWR, #48311-703), coverslipped with Fluoromount-G (Southern Biotech, #0100-01), and stored in the dark at 4 °C before microscopy and imaging.

2.4 Antibody characterization

To visually identify Hcrt neurons using immunohistochemistry, we used a rabbit polyclonal antibody against Orexin A (Phoenix Pharmaceuticals #H-003-30; Lot # 01651-6; RRID:AB_2315019). The manufacturer reports this antibody shows 100% cross reactivity with Orexin A from human, rat, mouse, porcine, ovine, bovine, and monkey, and 0% cross reactivity with Orexin B and other neuropeptides including agouti-related peptide, α-MSH, and leptin, as determined by competitive radioimmunoassay and enzyme immunoassay (EIA). The manufacturer also reports that this antibody recognizes bands of 15 kDa and 3.5 kDa molecular weight, representing prepro-ORX-A and ORX-A respectively, on immunoblots of rat brain. A previous study used a dot blot assay to demonstrate that this antibody was specific to Orexin A but showed no cross reactivity with a panel of 21 other neuropeptides expressed in the hypothalamus (Fronczek, Lammers, Balesar, Unmehopa, & Swaab, 2005). Pre-adsorption with ORX-A peptide eliminates staining from hamster (Helwig et al., 2006) and human brain (Fronczek et al., 2007). This antibody was used in previous studies to visually identify Hcrt neurons using immunohistochemistry (Fronczek et al., 2005; Fronczek et al., 2007; Fronczek et al., 2012; Hamlin, Newby, & McNally, 2007; S. Lee, Lee, Elias, & Elmquist, 2009; Satoh et al., 2006; Solomon, De Fanti, & Martinez, 2007). In our experiments, immunolabeled cells were confined to the lateral hypothalamic area (Figure 1a), consistent with the well-documented expression pattern of Hcrt neurons in mice.

2.5 Microscopy and data analysis

Specimens were examined using an Eclipse 80i epifluorescent microscope (Nikon). To image in the blue spectrum, we used a Nikon 96310 UV-2E filter cube with 360nm emission (40nm bandpass) and 460nm barrier (50nm bandpass). To image in the green spectrum, we used a Nikon 96320 FITC filter cube with 480 nm emission (40 nm bandpass) and 535 nm barrier (50 nm bandpass). To image in the red spectrum, we used a Nikon 96365 mCH/TR filter cube with 560 nm emission (40nm bandpass) and a 630 nm barrier (75 nm bandpass). Throughout the data analysis, specimens were routinely surveyed for possible crosstalk between fluorescent channels, and no crossover was detected.

Quantification of colocalization of retrograde signal and Hcrt immunolabeling in the hypothalamus was performed on brain sections from approximately −1.23 to −1.91 mm from Bregma (exactly 21 serial sections per mouse) such that every section through the Hcrt population was included. Cell counts were performed by an investigator (MI or MEC) blinded to the identity of the injection sites. Quantification of single labeled, double labeled, or triple labeled neurons was performed using the microscope (i.e. not based off digital images) using 10x or 20x objectives. We considered cells to be double or triple labeled if the green or red retrograde signal was substantially brighter than background fluorescence and conformed to the shape of a blue immunolabelled Hcrt neuron. In initial experiments, we also confirmed colocalization between retrograde signal and Hcrt immunostaining in the same cell using z-stack analysis with an Observer Z1 epifluorescent microscope with Zen 2012 imaging software (Zeiss), however, due to the general distribution of Hcrt neurons within a single brain section, colocalization of retrograde signal with Hcrt neurons was determined to be accurate by adjusting the focus within the plane of the section. All quantification was performed on the Hcrt field ipsilateral to the injection sites.

To quantify the lateral versus medial distribution of retrogradely labeled Hcrt neurons, we used the fornix as a consistent anatomical reference point within each coronal section to divide the Hcrt field into lateral and medial subpopulations (as opposed to fixed mediolateral coordinates throughout the rostrocaudal axis), consistent with previous studies (Fadel et al., 2002; Gonzalez et al., 2012; Kessler, Stanley, Frederick-Duus, & Fadel, 2011; Williams, Alexopoulos, Jensen, Fugger, & Burdakov, 2008). Because we quantified immunolabelled cells versus a defined three-dimensional volume of brain tissue, and because the fornix serves as a reliable anatomical marker across the rostrocaudal axis, we did not employ further stereological analysis.

All cell counts were corrected for potential double counting using Abercrombie’s formula, N = n(T/T+D) in which N = the corrected cell count, n = the observed cell count, T = the section thickness (30 μm), and D = the diameter of the cell (Guillery, 2002). We estimated the diameter of a Hcrt neuron as 27.5 μm based on previous measurements (Peyron et al., 1998). Because red or green retrobeads tended to completely fill the cytoplasm, we also estimated the diameter of red or green retrogradely labeled neurons to be 27.5 μm.

Images were captured using a RETIGNA 2000R digital camera. Digital images were minimally processed using Photoshop CS5 (Adobe Systems) to enhance the brightness and contrast for optimal representation of the data. All digital images were processed in the same way between experimental conditions to avoid artificial manipulation between different datasets. Representative images for data presentation in Figures were taken from a consistent coronal plane, approximately bregma = −1.55 mm. For wide-field images of coronal brain sections, individual photomicrographs were taken and montaged together in Adobe Photoshop CS5.

To visually depict and compare injection sites across animals, injection areas (visible as a bright fluorescent signal) were overlaid and traced onto a single coronal brain diagram (LC, bregma = 5.30 mm; NAc, bregma = 0.90 mm; TMN, bregma = 2.70 mm; VTA, bregma = −3.10 mm) in Adobe Photoshop CS5. To standardize injection site analysis across different brains, we used the fourth ventricle as an anatomical reference point for the LC, the lateral ventricles and anterior commissure as reference points for the NAc, the third ventricle and dorsal third ventricle as reference points for the TMN, and the medial mammillary nucleus and interpeduncular nucleus as reference points for the VTA. We analyzed and presented injection sites relative to observable anatomical landmarks based on conventional mouse brain atlases (Dong, 2008; Paxinos & Franklin, 2004). Only animals for which both green and red tracer injection sites showed good accuracy and precision were included in the data analysis (about one out of every three animals).

2.6 Statistics

Statistical tests included a paired t-test, an unpaired t-test, Chi-squared tests, and two-way ANOVA without repeated measures, as described in the Results.

To assess the degree to which Hcrt neurons project to either one, both, or neither downstream population, we compared expected and observed numbers of Hcrt neurons using Chi-squared tests. We obtained the expected number by assuming that projections to the two downstream populations are independent events. Under independence, the probability of projecting to both populations is given by the product of the probability of projecting to one population times the probability of projecting to the other population. As a consequence, under independence, the expected number of Hcrt neurons that project to both populations is given by (# green retrolabeled Hcrt neurons) x (# red retrolabeled Hcrt neurons)/(# total Hcrt neurons). Table 2 shows the method used to determine expected counts from observed counts in each experiment. For each experiment we used the observed counts for all four possible outcomes: (red+/green+; red+/green–; red–/green+; red–/green–). We ran a Chi-squared test of independence for each pairwise comparison of two downstream populations. To adjust for the fact that we made a total of 6 such comparisons, we multiplied the p-value from the Chi-squared test by 6 (Bonferroni adjustment).

Table 2.

Observed and expected frequencies for Chi squared analysis^†.

Observed frequencies					Expected (under independence) frequencies
		Population 2					Population 2
		Projecting (red)	Not projecting				Projecting (red)	Not projecting
Population 1	Projecting (green)	a	b	e	Population 1	Projecting (green)	e*g/N	e*h/N	e
	Not projecting	c	d	f		Not projecting	f*g/N	f*h/N	f
		g	h	N			g	h	N

^†We defined a as the number of neurons that are labeled both green and red, b the number that are labeled solely green, c the number of neurons that are labeled solely red, d the number of unlabeled (non-projecting) neurons, with N the total number of Hcrt neurons for a given experiment. e, f and g, h are the row and column sums, respectively. The left table shows observed frequencies; the right table shows the frequencies we would expect to see if projections to population 1 and population 2 are independent.

3. Results

3.1 Red and green retrograde tracers label Hcrt neurons with equivalent efficiencies

Before performing retrolabeling experiments, we compared the efficiencies of red and green retrograde tracers. To compare the relative efficiencies, we injected either green (Figure 2a–d) or red (Figure 2e–h) retrograde tracers into the LC and quantified the number of retrogradely labeled Hcrt neurons (n=6 animals per condition). There was no statistical difference between the mean number of green retrogradely labeled Hcrt neurons (110.7 ± 6.9) versus green retrogradely labeled Hcrt neurons (118.2 ± 6.7), suggesting similar efficiencies between green and red retrograde tracers (Figure 2i; unpaired t-test, t₁₀ = 0.78, p = 0.45).

Figure 2.

Equivalent efficiency of green and red retrograde tracers. (a) Injection of green retrograde tracer (diluted 1:2 with 0.9% saline) into the LC and (b–d) photomicrographs of the Hcrt field showing overlap of green retrograde signal and immunolabeled Hcrt neurons. (e) Injection of red retrograde tracer into the LC and (f–h) photomicrographs of the Hcrt field showing overlap of red retrograde signal and immunolabeled Hcrt neurons. (i) Quantification of the number of green versus red retrolabeled Hcrt neurons. Data represent mean ± standard deviation. Black dots represent individual animals (n=6 animals per condition). n.s., not significant; p=0.453. (j) Injection of mixture of green and red retrograde tracer (1:2 dilution) into the LC and (k–m) photomicrographs of the Hcrt field showing overlap of green and red retrograde signal. (n) Quantification of the number of green versus red retrolabeled Hcrt neurons. Data represent mean ± standard deviation. Black dots represent individual animals (n=6 animals per condition). n.s., not significant; p=0.875. Scale bar = 250 μm. fx, fornix.

To further determine the relative efficiencies of green and red retrograde tracers, we injected a mixture of green and red tracers together into the LC of the same animal (n=6 animals; Figure 2j–m). We found greater than 98% of green-labeled neurons also labeled with red tracer and greater than 98% of red-labeled neurons also labeled with green tracer. There was no statistical difference between the mean number of green labeled Hcrt neurons (111.8 ± 5.7) versus red labeled Hcrt neurons (113.2 ± 5.9), indicating similar efficiencies between green and red retrograde tracers (Figure 2n; paired t-test, t₅ = 1.51 p = 0.0955). Taken together, the similar efficiency of green or red retrograde tracer allows for the relative quantification of green or red retrogradely labeled Hcrt neurons from different tracer injection sites.

3.2 Hcrt neurons project to the TMN and LC, or the NAc and VTA, with high fidelity

To determine if Hcrt neurons send bifurcating projections to the TMN and LC, we injected green or red retrograde tracers into each downstream population (n=6). Green retrograde tracer injections were accurately targeted to the TMN region (Figures 3a, b). Because the TMN is a small, ventrally located population, some green tracer was observed dorsally toward the lateral hypothalamus. Red retrograde tracer was also accurately targeted to the LC (Figures 3c, d). In 2–3 animals, there was minimal tracer spillover into Barrington’s nucleus (Ba) and the mesencephalic trigeminal nucleus (Me5). These patterns were consistent in the adjacent 4–6 rostrocaudal sections in which tracer spread along the anterior-posterior axis.

Figure 3.

High frequency of bifurcating Hcrt neuronal projections to the TMN and LC. (a) Diagram showing superimposed areas of green retrograde tracer injections into the TMN (bregma = −2.70 mm) across six animals. (b) Representative histological image of green retrograde tracer injection in the TMN. (c) Diagram showing superimposed areas of red retrograde tracer injections into the LC (bregma = −5.30 mm) across six animals. (d) Representative histological image of red retrograde tracer injection in the LC. (e–f) Representative merged images showing (e) green and red retrogradely labeled neurons in the Hcrt field and (f) green and red retrogradely labeled neurons with blue immunolabeled Hcrt neurons. Yellow arrows indicate Hcrt-expressing neurons co-labeled with both green and red retrograde signal. (g–j) Higher power images of the respective boxes in (e–f) showing (g) a Hcrt-positive neuron with only green retrograde signal, (h) a Hcrt-positive neuron with only red retrograde signal, (i) a Hcrt-positive neuron with both green and red retrograde signal, and (j) a Hcrt-positive neuron with no retrograde signal. (k) Expected and observed percentages of projecting Hcrt neurons with only red, only green, or both red and green retrograde signal. Expected percentages were calculated under the assumption that projections to the two populations are independent. ***p<0.0001, Chi Squared test comparing expected and observed numbers of Hcrt neurons under the assumption of independent projections to the two populations. Scale bar in a, b, c, d = 500 μm. Scale bar in e, f = 100 μm. See Abbreviations for full anatomical definitions.

As expected, a number of Hcrt neurons were labeled with green or red signal (Table 3 and Figures 3e, f), and we identified Hcrt neurons that expressed only green, only red, both green and red, and neither green nor red signal (Figures 3g–j). To test the null hypothesis that projecting to the TMN is independent of projecting to the LC, we compared the expected versus observed frequencies of these four possible outcomes using a Chi-squared test (Table 2). Under independence, we expect 5.2% of retrogradely labeled Hcrt neurons to be labeled both green and red (Figure 3k). In fact, in our experiment, we observed 21.7% dual labeled Hcrt neurons (Figure 3k). Therefore, Hcrt neurons project to both the TMN and LC populations much more frequently than would be expected if they were independent populations (see Table 4 for Chi-squared statistics, p < 0.0001).

Table 3.

Quantification of retrogradely labeled Hcrt neurons.

Injection sites†	Hcrt neurons	Green retrogradely labeled Hcrt neurons		Red retrogradely labeled Hcrt neurons		Dual labeled Hcrt neurons
	Mean ± standard deviation	Mean ± standard deviation	Percentage of Hcrt neurons	Mean ± standard deviation	Percentage of Hcrt neurons	Mean ± standard deviation	Percentage of Hcrt neurons
	622.8 ± 60.0	55.9 ± 5.7	9.0 ± 0.85	67.3 ± 4.8	10.9 ± 0.76	22.0 ± 3.4	3.5 ± 0.31
	679.1 ± 45.2	73.7 ± 5.3	10.9 ± 1.1	87.4 ± 10.3	12.9 ± 1.9	28.3 ± 2.2	4.2 ± 0.23
	685.6 ± 87.2	95.7 ± 9.0	14.0 ± 1.2	78.2 ± 10.0	11.4 ± 0.6	2.2 ± 0.51	0.32 ± 0.05
	647.0 ± 74.1	88.1 ± 7.9	13.7 ± 1.7	55.9 ± 5.3	8.7 ± 0.31	2.3 ± 0.6	0.35 ± 0.11
	644.4 ± 125.5	54.8 ± 3.8	8.7 ± 1.3	69.4 ± 9.2	11.0 ± 1.4	2.1 ± 0.33	0.34 ± 0.12
	638.1 ± 91.8	64.8 ± 5.9	10.3 ± 1.7	54.5 ± 5.2	8.7 ± 1.1	2.7 ± 0.39	0.43 ± 0.09

^†n=6 animals for each pair of injection sites.

Table 4.

Chi squared analysis.

Injection sites†	Sample size (N)	X2	Degrees of freedom	Multiplicity- adjusted p value
	3737	311.4	1	< 0.0000
	4075	287.4	1	< 0.0000
	4114	55.1	1	< 0.0000
	3882	28.6	1	< 0.0000
	3866	18.1	1	0.0001
	3828	10.7	1	0.0066

^†n=6 animals for each pair of injection sites.

To determine if Hcrt neurons send bifurcating projections to the NAc and VTA, we injected green or red retrograde tracers into each downstream population (n=6). Green retrograde tracer injections were accurately targeted to the NAc (Figures 4a, b), with most tracer injected into the core region (NAcC) and lateral nucleus accumbens shell (NAcShl). In all six animals, a small amount of tracer was observed dorsally in the caudate putamen (CPu), however, Hcrt neurons do not project to this region (Peyron et al., 1998). Red retrograde tracer was also accurately targeted to the VTA (Figures 4c, d). In one animal there was potential spillover into the median reticular nucleus. These patterns were consistent in the adjacent 4–6 rostrocaudal sections in which tracer spread along the anterior-posterior axis.

Figure 4.

High frequency of bifurcating Hcrt neuronal projections to the NAc and VTA. (a) Diagram showing superimposed areas of green retrograde tracer injections into the NAc (bregma = 0.90 mm) across six animals. (b) Representative histological image of green retrograde tracer injection in the NAc. (c) Diagram showing superimposed areas of red retrograde tracer injections into the VTA (bregma = −3.10 mm) across six animals. (d) Representative histological image of red retrograde tracer injection in the VTA. (e–f) Representative merged images showing (e) green and red retrogradely labeled neurons in the Hcrt field and (f) green and red retrogradely labeled neurons with blue immunolabeled Hcrt neurons. Yellow arrows indicate Hcrt-expressing neurons co-labeled with both green and red retrograde signal. (g–j) Higher power images of the respective boxes in (e–f) showing (g) a Hcrt-positive neuron with only green retrograde signal, (h) a Hcrt-positive neuron with only red retrograde signal, (i) a Hcrt-positive neuron with both green and red retrograde signal, and (j) a Hcrt-positive neuron with no retrograde signal. (k) Expected and observed percentages of projecting Hcrt neurons with only red, only green, or both red and green retrograde signal. Expected percentages were calculated under the assumption that projections to the two populations are independent. ***p<0.0001, Chi Squared test comparing expected and observed numbers of Hcrt neurons under the assumption of independent projections to the two populations. Scale bar in a, b, c, d = 500 μm. Scale bar in e, f = 100 μm. See Abbreviations for full anatomical definitions.

We quantified the number and percentage of Hcrt neurons that were labeled by green or red signal (Table 3 and Figures 4g–j). Under independence, we expect 6.3% of retrogradely labeled Hcrt neurons to be labeled both green and red (Figure 4k). In fact, in our experiment, we observed 21.3% dual labeled Hcrt neurons (Figure 4k), indicating that Hcrt neurons project to both the NAc and VTA more frequently than would be expected if projections were independent (see Table 4 for Chi-squared statistics, p < 0.0001).

Taken together, these dual retrograde tracing experiments indicate that Hcrt neurons project to the LC and TMN, as well as to the NAc and VTA, at a much higher frequency than would be predicted under independence, suggesting a substantial collateral projection pattern for these subpopulations of Hcrt neurons.

3.3 Hcrt neurons that project to the LC or TMN are mostly distinct from Hcrt neurons that project to the NAc or VTA

Because we found that individual Hcrt neurons project to the LC or TMN or to the NAc or VTA at a higher frequency than would be predicted by chance (Figures 3,4), we examined the number of Hcrt neurons that send bifurcating projections to different permutations of these downstream populations. We tested four pairwise combinations of the neural populations previously described: VTA vs. LC (Figure 5), TMN vs. VTA (Figure 6), NAc vs. LC (Figure 7), and NAc vs. TMN (Figure 8). We only included animals in the data analysis that received accurate injection of retrograde tracers into the target regions. As before, there was occasional minor spillover of retrograde tracer into surrounding regions (Figures 5a–d, 6a–d, 7a–d, 8a–d). After confirming injection sites, we compared expected and observed frequencies for the four possible labeling outcomes in each experiment. In contrast to Hcrt neurons that project to the LC and TMN or to the NAc and VTA (Figures 3,4), we found that Hcrt neurons projected to different permutations of these populations less frequently than would be expected under independence (all four Chi-squared tests, p<0.01; for percentages and statistical values, respectively, see Figures 5k, 6k, 7k, 8k and Table 4). For example, comparing VTA and LC populations, we expect 6.7% of retrogradely labeled Hcrt neurons to be labeled both green and red. In fact, in our experiment, we observed only 1.3% dual labeled Hcrt neurons (Figure 5k), indicating that Hcrt neurons project to both the VTA and LC much less frequently than would be expected if projections were independent.

Figure 5.

Low frequency of bifurcating Hcrt neuronal projections to the VTA and LC. (a) Diagram showing superimposed areas of green retrograde tracer injections into the VTA (bregma = −3.10 mm) across six animals. (b) Representative histological image of green retrograde tracer injection in the VTA. (c) Diagram showing superimposed areas of red retrograde tracer injections into the LC (bregma = −5.30 mm) across six animals. (d) Representative histological image of red retrograde tracer injection in the LC. (e–f) Representative merged images showing (e) green and red retrogradely labeled neurons in the Hcrt field and (f) green and red retrogradely labeled neurons with blue immunolabeled Hcrt neurons. Yellow arrows indicate Hcrt-expressing neurons co-labeled with both green and red retrograde signal. (g–j) Higher power images of the respective boxes in (e–f) showing (g) a Hcrt-positive neuron with only green retrograde signal, (h) a Hcrt-positive neuron with only red retrograde signal, (i) a Hcrt-positive neuron with both green and red retrograde signal, and (j) a Hcrt-positive neuron with no retrograde signal. (k) Expected and observed percentages of projecting Hcrt neurons with only red, only green, or both red and green retrograde signal. Expected percentages were calculated under the assumption that projections to the two populations are independent. ***p<0.0001, Chi Squared test comparing expected and observed numbers of Hcrt neurons under the assumption of independent projections to the two populations. Scale bar in a, b, c, d = 500 μm. Scale bar in e, f = 100 μm. See Abbreviations for full anatomical definitions.

Figure 6.

Low frequency of bifurcating Hcrt neuronal projections to the VTA and TMN. (a) Diagram showing superimposed areas of green retrograde tracer injections into the VTA (bregma = −3.10 mm) across six animals. (b) Representative histological image of green retrograde tracer injection in the VTA. (c) Diagram showing superimposed areas of red retrograde tracer injections into the TMN (bregma = −2.70 mm) across six animals. (d) Representative histological image of red retrograde tracer injection in the TMN. (e–f) Representative merged images showing (e) green and red retrogradely labeled neurons in the Hcrt field and (f) green and red retrogradely labeled neurons with blue immunolabeled Hcrt neurons. Yellow arrows indicate Hcrt-expressing neurons co-labeled with both green and red retrograde signal. (g–j) Higher power images of the respective boxes in (e–f) showing (g) a Hcrt-positive neuron with only green retrograde signal, (h) a Hcrt-positive neuron with only red retrograde signal, (i) a Hcrt-positive neuron with both green and red retrograde signal, and (j) a Hcrt-positive neuron with no retrograde signal. (k) Expected and observed percentages of projecting Hcrt neurons with only red, only green, or both red and green retrograde signal. Expected percentages were calculated under the assumption that projections to the two populations are independent. ***p<0.0001, Chi Squared test comparing expected and observed numbers of Hcrt neurons under the assumption of independent projections to the two populations. Scale bar in a, b, c, d = 500 μm. Scale bar in e, f = 100 μm. See Abbreviations for full anatomical definitions.

Figure 7.

Low frequency of bifurcating Hcrt neuronal projections to the NAc and LC. (a) Diagram showing superimposed areas of green retrograde tracer injections into the NAc (bregma = 0.90 mm) across six animals. (b) Representative histological image of green retrograde tracer injection in the NAc. (c) Diagram showing superimposed areas of red retrograde tracer injections into the LC (bregma = −5.30 mm) across six animals. (d) Representative histological image of red retrograde tracer injection in the LC. (e–f) Representative merged images showing (e) green and red retrogradely labeled neurons in the Hcrt field and (f) green and red retrogradely labeled neurons with blue immunolabeled Hcrt neurons. Yellow arrows indicate Hcrt-expressing neurons co-labeled with both green and red retrograde signal. (g–j) Higher power images of the respective boxes in (e–f) showing (g) a Hcrt-positive neuron with only green retrograde signal, (h) a Hcrt-positive neuron with only red retrograde signal, (i) a Hcrt-positive neuron with both green and red retrograde signal, and (j) a Hcrt-positive neuron with no retrograde signal. (k) Expected and observed percentages of projecting Hcrt neurons with only red, only green, or both red and green retrograde signal. Expected percentages were calculated under the assumption that projections to the two populations are independent. ***p<0.0001, Chi Squared test comparing expected and observed numbers of Hcrt neurons under the assumption of independent projections to the two populations. Scale bar in a, b, c, d = 500 μm. Scale bar in e, f = 100 μm. See Abbreviations for full anatomical definitions.

Figure 8.

Low frequency of bifurcating Hcrt neuronal projections to the NAc and TMN. (a) Diagram showing superimposed areas of green retrograde tracer injections into the NAc (bregma = 0.90 mm) across six animals. (b) Representative histological image of green retrograde tracer injection in the NAc. (c) Diagram showing superimposed areas of red retrograde tracer injections into the TMN (bregma = −2.70 mm) across six animals. (d) Representative histological image of red retrograde tracer injection in the TMN. (e–f) Representative merged images showing (e) green and red retrogradely labeled neurons in the Hcrt field and (f) green and red retrogradely labeled neurons with blue immunolabeled Hcrt neurons. Yellow arrows indicate Hcrt-expressing neurons co-labeled with both green and red retrograde signal. (g–j) Higher power images of the respective boxes in (e–f) showing (g) a Hcrt-positive neuron with only green retrograde signal, (h) a Hcrt-positive neuron with only red retrograde signal, (i) a Hcrt-positive neuron with both green and red retrograde signal, and (j) a Hcrt-positive neuron with no retrograde signal. (k) Expected and observed percentages of projecting Hcrt neurons with only red, only green, or both red and green retrograde signal. Expected percentages were calculated under the assumption that projections to the two populations are independent. ***p<0.0001, Chi Squared test comparing expected and observed numbers of Hcrt neurons under the assumption of independent projections to the two populations. Scale bar in a, b, c, d = 500 μm. Scale bar in e, f = 100 μm. See Abbreviations for full anatomical definitions.

Taken together, these experiments demonstrate that Hcrt neurons that project to the LC or TMN are almost entirely anatomically separate from Hcrt neurons that project to the NAc or VTA.

3.4 Hcrt neuronal populations projecting to the LC, NAC, TMN, or VTA are not anatomically segregated into lateral and medial groups

To determine whether Hcrt neurons that project to the LC, NAc, TMN, or VTA are segregated into medial or lateral subpopulations, we quantified the number of retrogradely labeled neurons across all experiments (Figures 3–8) within the lateral or medial Hcrt field. To discriminate between the lateral and medial Hcrt field, we used the fornix as a consistent anatomical landmark in each coronal section (Figure 9a), consistent with previous studies (Fadel et al., 2002; Gonzalez et al., 2012; Kessler et al., 2011; Williams et al., 2008). Retrogradely labeled Hcrt neurons did not show a statistically significant lateral/medial bias within the Hcrt field (Figure 9b; no interaction between injection site and lateral/medial location: two way ANOVA, F_3,136 = 0.99, p = 0.4013; n=18 mice per condition). Dual retrogradely labeled Hcrt neurons also did not show a statistically significant lateral/medial bias (Figure 9c; no interaction between injection sits and lateral/medial location: two way ANOVA, F_1,20 = 0.7375; p = 0.4006; n=6 mice per condition). Taken together, these data do not support the hypothesis that Hcrt neuronal populations projecting to the LC, NAc, TMN, or VTA are anatomically segregated into lateral and medial groups.

Figure 9.

Hypocretin neurons projecting to the LC, NAc, TMN, or VTA do not segregate into lateral or medial populations. (a) We used the fornix (fx) as a consistent anatomical marker to divide the Hcrt field into lateral and medial populations within each coronal section. (b) The percentage of retrogradely labeled Hcrt neurons in the lateral or medial Hcrt field. Data represent mean ± standard deviation. Black dots represent individual animals (n=18 animals per injection site). (c) The percentage of dual retrogradely labeled Hcrt neurons in the lateral or medial Hcrt field. Data represent mean ± standard deviation. Black dots represent individual animals (n=6 animals per injection sites).

4. Discussion

4.1 Identification of distinct populations of Hcrt neurons based on downstream projections

Our data identify Hcrt neurons that project to both the LC and TMN or to both the VTA and NAc at a higher frequency than would be predicted by independent overlap of retrograde signal (modeled in Figure 1g). In contrast, we show that Hcrt neurons that project to the LC or TMN are unlikely to also project to the VTA or NAc (and vice versa; modeled in Figure 1e). Finally, our data do not support the hypothesis that Hcrt neurons that project to the LC, TMN, VTA, and NAc are anatomically segregated within the lateral or medial Hcrt field (Figure 9). These results are consistent with other studies that identified subpopulations of Hcrt neurons, but did not identify lateral/medial topography within the hypothalamus (España et al., 2005; Gonzalez et al., 2012). Taken together, these data suggest a model in which Hcrt neurons can be classified based on their downstream projections but that these classifications do not show a topographic location within the hypothalamus (Figure 10).

Figure 10.

Model depicting subpopulations of Hcrt neurons within the Hcrt field. A substantial subset projects to both the NAc and VTA (green), and another substantial subset to both the LC and TMN (red). A small number of Hcrt neurons (orange) project to different combinations of downstream structures. Not depicted are Hcrt neurons that project to only a single downstream area. Note that the majority of Hcrt neurons do not project to any of the four downstream populations.

It is important to note that although we found that Hcrt neurons project to both the LC and TMN or to the NAc and VTA much more frequently than would be expected by chance (and much more frequently than other permutations of the downstream areas studied), most Hcrt neurons retrogradely labeled from a particular downstream population did not show colocalization with retrograde tracer from the other downstream population. Although we believe there are more dual-projecting Hcrt neuron populations than those observed (see below for technical limitations), our study shows that non-dual projecting Hcrt neurons to the LC and TMN or to the NAC and VTA are more prominent than dual projecting Hcrt neurons (Table 3 and Figures 3k and 4k).

It is also important to note that our study categorized Hcrt neurons based on their downstream projections, not based on their functional activity. Previous studies showing that Hcrt neurons can be classified into lateral/medial groups based on Fos expression during the active/inactive period (Estabrooke et al., 2001), morphine administration (Richardson & Aston-Jones, 2012), or rewarding environmental cues (Harris et al., 2005) are not inconsistent with our present results, and it is possible that Hcrt neurons can be functionally subdivided based on activity but not anatomically subdivided based on projections to the LC, TMN, NAc, and VTA. Furthermore, our study also did not consider other potential classifications of Hcrt neurons within the lateral hypothalamus other than lateral/medial topography, and it is possible that other projection patterns might emerge from different methods of analysis. For example, our immunofluorescent identification of Hcrt neurons did not allow for a formal analysis of cytoarchitecture or the ability to identify different populations of Hcrt neurons based on gene or protein expression. Therefore, our analysis only broadly distinguished between lateral and medial populations across the Hcrt field. Our study also does not account for differential expression of the two downstream Hypocretin receptors, the possibility of differential levels of Hcrt or Hcrt receptor protein expression, or changes in connectivity during different physiological states. It is possible that, during certain physiological or environmental conditions, synaptic strength to some downstream regions are emphasized over others.

Although the LC and TMN are often described as wakefulness/arousal populations and the VTA and NAc are often described as reward populations, wakefulness and reward-seeking behaviors are linked, and these downstream nuclei can indeed regulate behaviors much more broadly. For example, activity in the LC-norepinephrine system is sufficient to cause cue-induced reinstatement of reward seeking (Al-Hasani, McCall, Foshage, & Bruchas, 2013; Mi et al., 2017; Shaham, Highfield, Delfs, Leung, & Stewart, 2000). Likewise, stimulants that enhance VTA activity and dopaminergic tone promote wakefulness (Boutrel & Koob, 2004), and activation of the VTA has recently been shown to increase arousal during the presentation of salient and rewarding stimuli (Eban-Rothschild, Rothschild, Giardino, Jones, & de Lecea, 2016). Therefore, distinguishing Hcrt neurons that project to the LC and TMN as “wakefulness projections” and Hcrt neurons that project to the VTA and NAc as “reward projections” is overly simplistic. Nevertheless, the fact that the majority of these projections are segregated suggests that many Hcrt neurons can potentially engage with their targets selectively and specifically.

Finally, it is important to note that our experiments were conducted using mice, whereas many other studies were conducted using rats. Because neurons that express Hcrts show species-specific expression patterns (Nixon & Smale, 2007; Swanson et al., 2005), there may also be differences in the hodological or functional classification of these neurons across species, and indeed across different strains of mice.

4.2 Technical limitations

A technical limitation of this study is the inherent variability in successfully targeting discrete structures with injections of retrograde tracers. Because the LC and TMN are compact structures, we observed occasional minimal spillover into neighboring brain regions (Figure 3a–d). Additionally, the three-dimensional topography of the four neural populations studied makes 100% penetrance of tracer injections challenging. Therefore, the percentage of Hcrt neurons that project to each downstream structure is likely higher than our observations. Because the number of Hcrt neurons labeled with green or red retrograde tracer is likely smaller than the true value of Hcrt neurons that project to each target region, the percentage of Hcrt neurons projecting to two downstream populations is also likely to be greater than our observed values. Therefore, even though we observed 21.7% of retrogradely labeled Hcrt neurons projecting to both the LC and TMN (Figure 3) and 21.3% of retrogradely labeled Hcrt neurons projecting to both the VTA and NAc (Figure 4), the actual percentages of dual labeled Hcrt neurons are likely to be higher than these values.

Injections into the NAc were particularly challenging because of the presence of the lateral ventricles and anterior commissure (ac): in pilot experiments we found that more medial injections spread to other brain regions, likely due to spread through the ventricular system or passage along the commissural fiber tract. Therefore, we biased our injections towards the NAc core (NAcC) and lateral shell (NAcShl). Although previous studies indicate that Hcrt neurons project to the NAcC and NAcShl (Brown, Khoo, & Lawrence, 2013; Peyron et al., 1998), Hcrt neurons send especially dense projections to the medial shell (NAcShm) (Baldo, Daniel, Berridge, & Kelley, 2003; Peyron et al., 1998). In fact, the NAc region can be anatomically divided into many subregions with functional differences. Therefore, this study may have missed functional aspects of NAc topology based on our stereotaxic injection coordinates.

4.3 Categorizing peptidergic hypothalamic systems based on downstream projections

As shown for Hcrt neurons in the present study, other peptidergic hypothalamic neurons have also been reported to be segregated into subpopulations based on their downstream projections. For example, neurons that express agouti-related protein (AgRP) are exclusively expressed in the arcuate nucleus of the hypothalamus and are well known to regulate feeding behavior (Andermann & Lowell, 2017; Sternson & Eiselt, 2017). Using dual retrograde tracers and viral strategies, Betley et al. demonstrated that AgRP neurons could be divided into functional subpopulations based on projections to downstream brain regions that differentially affect feeding (Betley, Cao, Ritola, & Sternson, 2013). Interestingly, unlike Hcrt neurons, AgRP neurons showed almost no collateralization to downstream populations. Instead, AgRP neurons that project to distinct brain regions are subdivided along the anteroposterior axis of the brain. Furthermore, AgRP neurons only project to about ten distinct target populations, while Hcrt neurons project much more diffusely throughout the brain (Peyron et al., 1998). Therefore, classifying Hcrt neurons based on downstream projections is likely to be much more complicated than other neural populations, with potential collateral projections to many downstream brain regions.

Indeed, a major question stemming from our results is the degree to which Hcrt neurons send collateral projections to many other brain regions. Our results suggest that Hcrt neurons that send collateral projections to the LC and TMN or to the VTA and NAc are distinct, but there are likely several other categories of Hcrt neurons, each with distinct collateral projections. España et al. found collateral projections from Hcrt neurons to both the LC and basal forebrain (España et al., 2005), another downstream structure known to promote wakefulness (Han et al., 2014; Hassani, Lee, Henny, & Jones, 2009). It would be interesting to determine the frequency of Hcrt neurons that send bifurcating projections to the LC, basal forebrain, and TMN. In addition to regulating wakefulness and reward-seeking, Hcrt neurons influence many other phenotypes such as stress (Winsky-Sommerer et al., 2004) and sexual behavior, and do so through projections to several additional brain regions (Di Sebastiano, Yong-Yow, Wagner, Lehman, & Coolen, 2010; Muschamp, Dominguez, Sato, Shen, & Hull, 2007). For example, Hcrt neurons are known to influence the behavioral and physiological stress response through projections to the paraventricular thalamic nucleus (Sakamoto, Yamada, & Ueta, 2004; Samson, Taylor, Follwell, & Ferguson, 2002), bed nucleus of the stria terminalis (Lungwitz et al., 2012), and central amygdala (Bisetti et al., 2006), and influence sexual behavior through efferent projections to the median preoptic area (Gulia, Mallick, & Kumar, 2003). Future studies should use these and other functional downstream projections to investigate the possibility of other discrete populations of Hcrt neurons.

Abbreviations

3V

3^rd ventricle

4V

4^th ventricle

ac

anterior commissure

Arc

arcuate nucleus

Ba

Barrington’s nucleus

CPu

caudate putamen (striatum)

fr

fasciculus retroflexus

fx

fornix

Hcrt

hypocretin

IF

interfascicular nucleus

IP

interpeduncular nucleus

LC

locus coeruleus

LDTg

laterodorsal tegmental nucleus

LH

lateral hypothalamus

Me5

mesencephalic trigeminal nucleus

MM

medial mammillary nucleus

MPB

medial parabrachial nucleus

MRN

midbrain reticular nucleus

MVe

medial vestibular nucleus

NAc

nucleus accumbens

NAcC

nucleus accumbens core region

NAcShl

nucleus accumbens shell, lateral

NAcShm

nucleus accumbens shell, medial

PH

posterior hypothalamic nucleus

PMD

premammillary nucleus, dorsal

PMV

premammillary nucleus, ventral

RLi

rostral linear nucleus

RM

retromammillary nucleus

SNR

substantia nigra pars reticulata

TMN

tuberomammilary nucleus

VTA

ventral tegmental area