Activation of the Ventrolateral Preoptic Neurons Projecting to the Perifornical-Hypothalamic Area Promotes Sleep: DREADD Activation in Wild-Type Rats

The ventrolateral preoptic area (VLPO) predominantly contains sleep-active neurons and is involved in sleep regulation. The perifornical-hypothalamic area (PF-HA) is a wake-regulatory region and predominantly contains wake-active neurons. VLPO GABAergic/galaninergic neurons project to the PF-HA. Previously, the specific contribution of VLPO neurons projecting to the PF-HA (VLPO > PF-HAPRJ) in sleep regulation in rats could not be investigated due to the lack of tools that could selectively target these neurons. We determined the contribution of VLPO > PF-HAPRJ neurons in sleep regulation by selectively activating them using designer receptors exclusively activated by designer drugs (DREADDs) in wild-type Fischer-344 rats. We used a combination of two viral vectors to retrogradely deliver the Cre-recombinase gene, specifically, in VLPO > PF-HA neurons, and further express hM3Dq in those neurons to selectively activate them for delineating their specific contributions to sleep–wake functions. Compared to the control, in DREADD rats, clozapine-N-oxide (CNO) significantly increased fos-expression, a marker of neuronal activation, in VLPO > PF-HAPRJ neurons (2% vs. 20%, p < 0.01) during the dark phase. CNO treatment also increased nonREM sleep (27% vs. 40%, p < 0.01) during the first 3 h of the dark phase, when rats are typically awake, and after exposure to the novel environment (55% vs. 65%; p < 0.01), which induces acute arousal during the light phase. These results support a hypothesis that VLPO > PF-HAPRJ neurons constitute a critical component of the hypothalamic sleep–wake regulatory circuitry and promote sleep by suppressing wake-active PF-HA neurons.


Introduction
Multiple lines of studies support that the ventrolateral preoptic area (VLPO) is critically involved in sleep regulation [1][2][3][4]. VLPO lesions result in significant sleep fragmentation and decreased total sleep amount [5]. The VLPO contains sleep-active neurons or neurons that exhibit increasing discharge from the waking to nonREM sleep transition to stable nonREM/REM sleep [6]. The evidence further suggests that these neurons are not only involved in the initiation of sleep, but rather are also involved in the building up and dissipation of homeostatic sleep drive during sleep deprivation and recovery sleep, respectively [7]. VLPO neurons also exhibit sleep-associated fos-expression (Fos-IR), a marker of neuronal activation [8][9][10]. While VLPO contains a higher concentration of sleep-active maintained on a 12 h: 12 h light: dark cycle (lights on at 9:00 h), an ambient temperature of 24 ± 2 • C, and with food and water available ad libitum. All experiments were conducted in accordance with the National Research Council's "Guide for the Care and Use of Laboratory Animals" and approved by the Institutional Animal Care and Use Committee of the Veterans Affairs Greater Los Angeles Healthcare System; Protocol # 1615739.

Surgical Procedures
The details of the surgical procedures have been described previously [31]. Briefly, under surgical anesthesia (100 mg/Kg ketamine + 15 mg/Kg xylazine; maintenance with isoflurane) and under aseptic conditions, two holes (~1 mm diameter), one above the VLPO and another above the PF-HA, were drilled ipsilaterally into the skull. The viral vectors were injected unilaterally into the VLPO (AP: −0.5 mm, LM: 1.0 mm, and DV: 9.2 mm, relative to bregma) and into the PF-HA (AP: −3.0 mm, LM: 1.4 mm, and DV: 9.0 mm) [32] using a micro-syringe with 28 gauge needle (Hamilton Co., Reno, NV, USA) and attached to a motor-driven micromanipulator. Electroencephalogram (EEG) and electromyogram (EMG) electrodes were implanted stereotaxically for polygraphic monitoring of sleep-waking states. Stainless steel screw EEG electrodes were implanted through small, hand-drilled holes in the skull, and flexible silver stainless steel EMG wire electrodes were implanted into the dorsal cervical musculature.

Experimental Groups and Viral Vectors
DREADDs group: We used AAV2retro with the gene of Cre-recombinase, already validated to propagate retrogradely, to create the genetic switch to turn on recombinant dependent effector and reporter genes in VLPO > PF-HA PRJ neurons and then further express the double-floxed hM3Dq gene in them for their activation by CNO [33][34][35]. This group of rats was injected with AAV2retro-EF1a-Cre (Addgene #55636; volume, 1.0 µL; titer of vector, 3.5 × 10 12 vg/mL) into the PF-HA and Syn-driven, Cre-dependent, hM3D(Gq) receptor genes with an mCherry reporter, AAV2-hSyn-DIO-hM3D(Gq)-mCherry (Addgene #44361, volume 0.4 µL; titer of vector, 3.5 × 10 12 vg/mL) into the VLPO. The vectors were injected at 0.1 µL/2 min. After completion of the injection, the needle was left in place for another 20 min and then slowly withdrawn to avoid any backflow. AAV2-hSyn-DIO-hM3D(Gq)-mCherry injections into the VLPO resulted in Cre-dependent hM3Dq expression (as evident from the mCherry labeling) exclusively in the VLPO > PF-HA PRJ neurons ( Figure 1) Control group: Using the same protocol as in the DREADDs group, this group of rats was also injected with the same AAV2retro-EF1a-Cre into the PF-HA for delivering Crerecombinase into VLPO > PF-HA PRJ neurons retrogradely. As a control, we injected AAV2-DIO-GCaMP6s-dTomato, available on hand, into the VLPO, resulting in Cre-dependent GCaMP expression as a reporter (as evident from the dTomato labeling) exclusively in the VLPO > PF-HA PRJ neurons. Unlike hM3Dq in the DREADD group, the GCaMP6 protein created by the synthetic fusion of GFP, calmodulin, and the additional peptide could not be activated by CNO [36].

Recovery and Adaptation
Rats were allowed to recover from the surgical procedure for 10-14 days in Plexiglas recording cages placed in a sound-attenuated and temperature-controlled recording chamber. The rats were then connected with the recording cable and allowed an additional seven days for acclimatization to the recording setup and recording cables before the experiments were carried out. During this time, the rats were also adapted to the handling, similar to that they were subjected to during the injection procedure. In the recording chamber, the rats were maintained at the same 12 h: 12 h light: dark cycle, with ad libitum access to food and water.
Cre-recombinase was retrogradely transported to VLPO > PF-HA PRJ neurons (Figures 1  and 6). In the VLPO, we observed a relatively higher density of hM3Dq or GCaMP6 labeled neurons in the VLPO core and, to some extent, in the medial POA and dorsolateral POA, indicating that the sleep-wake changes observed in this study were predominantly due to the activation of VLPO > PF-HA PRJ neurons localized in the VLPO core and, to a lesser extent, in the extended VLPO. . This approach was used to specially target and examine the role of VLPO > PF-HA PRJ neurons and their interactions with the PF-HA neuronal groups in sleep-wake regulation. (C1,C2): AAV2retro-EF1a-Cre injection sites in the PF-HA. Although the AP locations of the injection sites varied by 100-200 µm, these are shown in one plane (Bregma −3.14 mm; adapted with permission from Paxinos and Watson, 1998 [32]). (C1): photomicrograph of a representative histological section showing the needle tract, as indicated by the increased autofluorescence, which was used to identify injection sites in the PF-HA. (C2): The locations of all injections in the DREADD and non-DREADD control groups. The locations of injection sites were broadly comparable in both the DREADD and control groups. Thus, it is likely that similar groups of VLPO > PF-HA PRJ neurons were retrogradely labeled in both DREADD and control groups. 3V, third ventricle; Arc, arcuate nucleus; DMH, dorsomedial hypothalamus; f, fornix; LH, lateral hypothalamus; mt, mammillothalamic tract; PeF, perifornical nucleus; TMN, tuberomammillary nucleus; VLPO, ventrolateral preoptic area; VMH, ventromedial hypothalamus.

Chemoactivation of VLPO > PF-HA PRJ Neurons Causes Sleep during the Dark Phase
The amounts of waking, nonREM sleep, and REM sleep during 0-3 and 3-6 h after the saline and CNO injections at dark-onset in the rats expressing hM3Dq (DREADD rats) or GFP (non-DREADD rats) in the VLPO > PF-HA PRJ neurons are shown in Figures 2 and 3.

Data Acquisition
The actual experiment was started after~4 weeks of surgical procedure/viral vector injections, allowing sufficient time for the retrograde labeling and the expression of hM3Dq in VLPO > PF-HA PRJ neurons. The amplified and filtered EEG and EMG signals were continuously digitized and stored on the computer's hard disk using an integrated computer interface device (Cambridge Electronic Design 1401, UK; supporting software, Spike 2) for subsequent sleep-wake scoring and analyses. To determine the specific contributions of the VLPO > PF-HA PRJ neurons in sleep regulation, we chemo-activated VLPO > PF-HA PRJ neurons by CNO. We conducted the following two sets of studies: Experiment 1: This study was conducted during the first half of the dark phase, when rats are largely awake and sleep propensity is the lowest. The DREADDs and the control group of rats were injected SQ with saline or CNO at dark-onset, and their EEG and EMG were continuously recorded in their home cages without any disturbance for 6 h. For SQ injection, the rats were gently taken out of the cage without disconnecting the recording cable and placed on a plastic platform (25 × 25 cm) near the cage. The skin flap around the lumbar part of the body and from the backside was gently pulled up, and the CNO was injected. After injection, the rats were gently returned to their cages. Saline or CNO was injected (200 nL) using an insulin syringe (needle 27 G), and between injections, each rat had a break of 2-3 days in the case of CNO and 1-2 days in the case of the saline treatments. The injection of saline and various doses of CNO were randomized. The injection of each rat took about 1 min. We chose SQ injection instead of IP, which has typically been used in other studies because it is faster, because rats exhibit much less behavioral stress and vocalization during SQ injection. Although we did not assess the difference between the latency of the effects after IP and SQ injections, the SQ route likely has longer latency. We started with a dose of 0.3 mg/kg of CNO and further determined the effects of increasing its amounts to 1.0 mg/kg and 3.0 mg/kg. Experiment 2: In rodents, acute exposure to the novel environment (e.g., bedding changes/new cage) suppresses sleep and causes arousal for 1-3 h [20,37,38]. We determined whether chemoactivation of the VLPO > PF-HA PRJ neurons helps improve sleep in this model of insomnia. Both the DREADDs and the control group of rats were injected with either saline or CNO (1 mg/kg) 6 h after lights-on. After 20 min of injection, the rats were transferred to a new cage, similar to the home cage, with fresh/new bedding and nesting materials, and their EEG and EMG were continuously recorded for 5-6 h or until the onset of the dark phase. We chose the middle of the light phase so that the animals would have less sleep pressure and were more prone to novel-environment-induced arousal. In addition, animals were exposed to the novel environment 20 min after saline or CNO injections so that the CNO-induced activation of the VLPO > PF-HA PRJ neurons coincided with the exposure to the novel environment.
To quantify the population of neurons that were activated by CNO, at the end of the experiment, 1.0 mg/kg of CNO was injected using the same experimental paradigm, and after 90 min, the rats were perfused and their brains were harvested for immunohistochemical processing.

Histology
At the end of the experiment, under deep anesthesia (100 mg/kg, IP, pentobarbital, Vortech, Dearborn, MI, USA), the rats were perfused transcardially with 40 mL of 0.1 M phosphate buffer saline (PBS, pH 7.2), followed by 500 mL of 4% paraformaldehyde in PBS. Their brains were removed and equilibrated in 10%, 20%, and, finally, 30% sucrose until they sank. The coronal sections from the POA to the posterior hypothalamus, including the sites of viral vector injections in the VLPO and PF-HA, were freeze-cut at a 30 µm thickness. Freefloating sections were immunostained to visualize VLPO > PF-HA PRJ neurons expressing hM3Dq or GCaMP and c-fos immunoreactivity using standard protocols [8,12]. Briefly, the sections were incubated with anti-c-fos rabbit polyclonal antibody (1:2000; cat. #ABE475, EMD Millipore Corporation, Temecula, CA, USA) for 72 h at 4 • C, and then developed with secondary antibody (1:400, horse anti-rabbit DaiLight 647; Vector Laboratories Inc., Newark, CA, USA). The needle tracks and injection sites in the PF-HA were identified and mapped.

Data Analyses
Sleep-wake parameters: The sleep-wake quantity of the animals was scored manually in 10 s epochs in terms of waking (consisting of active and quiet-waking), nonREM sleep, and REM sleep using SleepSign software (version 2; Kissei Comtec Co., Tokyo, Japan) and standard criteria [39]. Six hours of night recording and five hours of day recordings after exposure to the novel environment were analyzed after saline and CNO injections. We found 1 mg CNO to be optimally effective in inducing sleep-wake changes. We examined the continuity and stability of the sleep-waking state after CNO treatments in the DREADDs and control group by further classifying episode duration of waking (active), nonREM sleep, and REM sleep stages into four subcategories: episodes with a duration of < 30 s (very short), > 30 s to 2 min (short), > 2 min to 5 min (medium), and > 5 min (long). The changes in sleep-wake profiles after CNO treatments were analyzed during 0-3 and 3-6 h during the dark phase and 0-2 and 2-5 h after exposure to the novel environment to determine if the effects were time-dependent.
Statistical analyses: The SigmaPlot (Systat Software, San Jose, CA, USA) software package was used for statistical analyses of the data. The data are presented as mean ± standard error of the mean (SEM). One-way repeated-measures analysis of variance (ANOVA) for within-group analysis and one-way ANOVA for between-group analysis, followed by multiple pairwise comparisons using the Holm-Sidak test, were used for comparing sleep-wake changes. In a few cases, the data failed normality or equal variance tests. Those data points were analyzed by both parametric and equivalent non-parametric methods (Friedman RM ANOVA on ranks or Kruskal-Wallis one-way ANOVA on ranks) and the more conservative of the two levels of significance was used. The sleep-wake changes or changes in Fos-IR in GFP+ or mCherry+ cells after CNO treatments between the DREADD and non-DREADD groups were compared using the Student's t-test or Mann-Whitney rank sum test in the case where the normality test failed.

Labeling of VLPO > PF-HA PRJ Neurons and hM3Dq Receptor Expression
The locations of the microinjection sites of the AAV2-EF1a-Cre in the PF-HA and the distribution of the retrogradely labeled VLPO > PF-HA PRJ neurons along with hM3Dq or GFP expression are shown in Figure 1 and Figure 6. The AAV2retro-EF1a-Cre vector did not produce fluorescent tags; therefore, the locations of its injection sites were confirmed by the needle tracts ( Figure 1C).
Although the locations of the AAV2retro-EF1a-Cre injection sites varied rostrocaudally and dorsoventrally by 100-200 µm, all of them were localized in the perifornical area, in portions of the dorsomedial and ventromedial hypothalamus and the lateral hypothalamus. The extent of the parenchymal propagation of AAV2retro is similar to AAV2, which has minimal propagating capacity compared to other serotypes and is estimated to be ≤ 1 mm from the injection site [40][41][42]. Thus, it is likely that in this study, AAV2retro-EF1a-Cre diffused into~1 mm area from the point of injections including the HCRT and glutamatergic neuronal fields ( Figure 1C). In the VLPO, the accuracy of both VLPO and PF-HA injections was confirmed by the Cre-dependent expression of mCherry or dTomato, which were visible only in neurons projecting to the PF-HA from where the Cre-recombinase was retrogradely transported to VLPO > PF-HA PRJ neurons ( Figure 1 and Figure 6). In the VLPO, we observed a relatively higher density of hM3Dq or GCaMP6 labeled neurons in the VLPO core and, to some extent, in the medial POA and dorsolateral POA, indicating that the sleep-wake changes observed in this study were predominantly due to the activation of VLPO > PF-HA PRJ neurons localized in the VLPO core and, to a lesser extent, in the extended VLPO.

Chemoactivation of VLPO > PF-HA PRJ Neurons Causes Sleep during the Dark Phase
The amounts of waking, nonREM sleep, and REM sleep during 0-3 and 3-6 h after the saline and CNO injections at dark-onset in the rats expressing hM3Dq (DREADD rats) or GFP (non-DREADD rats) in the VLPO > PF-HA PRJ neurons are shown in Figures 2 and 3.
Sleep-wake changes during 0-3 h post-treatment: In the non-DREADD rats, the effects of various doses of CNO on waking, nonREM sleep, and REM sleep during the 0-3 h postinjection recording during the dark phase were comparable and similar to that observed after the saline injection (one-way RM ANOVA; F 5,3 = 1.61-2.29; p > 0.05). However, in the DREADD rats, CNO treatment significantly decreased the amount of waking (one-way RM ANOVA, F 6,3 = 34.35, p < 0.001) and increased the amount of nonREM sleep (F 6,3 = 38.89, p < 0.001) during the first three h of the post-injection period. CNO also caused a marginal increase in REM sleep (overall, F 6,3 = 6.26, p < 0.01; significant difference only between saline and 0.3 mg/kg CNO; Holm-Sidak multiple comparisons).
to non-DREADD group, the effects of CNO treatment on waking (F7,44 = 15.40; p < 0.001) and nonREM sleep (F7,44 = 12.07; p < 0.001) in the DREADD rats were significant ( Figures  2 and 3). The DREADD group exhibited only a marginal increase in REM sleep (p = 0.03; Kruskal-Wallis one-way ANOVA on ranks, no intergroup differences). Among the CNO doses, 1 mg/kg was the minimal dose that produced consistent effects on sleep-waking. While CNO at a 3 mg/kg dose induced a stronger response on sleep during 0-3 h, it was not significantly different from the other doses used in this study ( Figure 3).  The spontaneous sleep-waking profiles of both the non-DREADD and DREADD rats during the 0-3 h recording after saline injections were comparable. However, compared to non-DREADD group, the effects of CNO treatment on waking (F 7,44 = 15.40; p < 0.001) and nonREM sleep (F 7,44 = 12.07; p < 0.001) in the DREADD rats were significant (Figures 2 and 3). The DREADD group exhibited only a marginal increase in REM sleep (p = 0.03; Kruskal-Wallis one-way ANOVA on ranks, no intergroup differences). Among the CNO doses, 1 mg/kg was the minimal dose that produced consistent effects on sleepwaking. While CNO at a 3 mg/kg dose induced a stronger response on sleep during 0-3 h, it was not significantly different from the other doses used in this study (Figure 3).
Since CNO-induced sleep-wake changes occurred predominantly during the first 3 h, we further analyzed the extent of sleep-waking stability during 0-3 h after CNO treatment in the DREADD and non-DREADD rats. We counted the frequency of waking (active), nonREM sleep, and REM sleep episodes of very short (≤30 s), short (>30 s-2 min), medium (>2 min-5 min), and longer (>5 min) durations after 1 mg/kg of CNO treatments in the non-DREAD and DREAD rats, as described earlier [43]. Compared to the non-DREADD rats, the CNO-treated DREADD rats exhibited: (a) significantly decreased numbers of longer bouts (t = 2.61; p < 0.05) and an increased number of very short bouts of activewaking (t = 5.11; p < 0.01); and (b) significant increases in the number of short (t = 2.97; p < 0.05) and medium bouts of nonREM sleep (t = 2.34; p < 0.05). Although the frequency of nonREM sleep episodes of < 30 s duration was also higher, it was not significant (t = 1.83; p = 0.09). There was no significant change in the frequency of REM sleep episodes of any bout duration, although the number of small and medium bouts of REM sleep was relatively higher (Figure 4). Since CNO-induced sleep-wake changes occurred predominantly during the first 3 h, we further analyzed the extent of sleep-waking stability during 0-3 h after CNO treatment in the DREADD and non-DREADD rats. We counted the frequency of waking (active), nonREM sleep, and REM sleep episodes of very short (≤30 s), short (>30 s-2 min),  bers of longer bouts (t = 2.61; p < 0.05) and an increased number of very short bouts of active-waking (t = 5.11; p < 0.01); and (b) significant increases in the number of short (t = 2.97; p < 0.05) and medium bouts of nonREM sleep (t = 2.34; p < 0.05). Although the frequency of nonREM sleep episodes of < 30 s duration was also higher, it was not significant (t = 1.83; p = 0.09). There was no significant change in the frequency of REM sleep episodes of any bout duration, although the number of small and medium bouts of REM sleep was relatively higher (Figure 4).

Figure 4. Effects of the chemoactivation of VLPO > PF-HA PRJ neurons on sleep-waking stability.
The numbers of episodes of waking, nonREM sleep, and REM sleep of various durations exhibited by individual rats (filled circles) and as a group (mean ± SEM; bars) during the first 3 h after 1 mg/kg of CNO treatment in the non-DREADD and DREADD rats. Compared to the non-DREAD rats, the DREADD rats exhibited a decline in the number of longer bouts of waking, although the frequency of very short episodes of waking increased significantly. These rats also exhibited significant increases in short and medium bouts of nonREM sleep. REM sleep episodes were largely not affected. * p < 0.05; ** p < 0.01; independent t-test.
Sleep-wake changes during 3-6 h post-treatment: During 3-6 h of post-treatment recording, in the non-DREADD rats, the effects of the various doses of CNO on waking, nonREM sleep, and REM sleep were comparable to that observed after the saline injection. In the DREADD rats, compared to the saline control, the effects of CNO on sleep-waking states remained significant (waking; one-way RM ANOVA, F7,44 = 7.13, p < 0.001; nonREM sleep, F7,44 = 5.72, p < 0.001; REM sleep, F7,44 = 5.34, p < 0.001) during the 3-6 h post-treatment The numbers of episodes of waking, nonREM sleep, and REM sleep of various durations exhibited by individual rats (filled circles) and as a group (mean ± SEM; bars) during the first 3 h after 1 mg/kg of CNO treatment in the non-DREADD and DREADD rats. Compared to the non-DREAD rats, the DREADD rats exhibited a decline in the number of longer bouts of waking, although the frequency of very short episodes of waking increased significantly. These rats also exhibited significant increases in short and medium bouts of nonREM sleep. REM sleep episodes were largely not affected. * p < 0.05; ** p < 0.01; independent t-test.
Sleep-wake changes during 3-6 h post-treatment: During 3-6 h of post-treatment recording, in the non-DREADD rats, the effects of the various doses of CNO on waking, nonREM sleep, and REM sleep were comparable to that observed after the saline injection. In the DREADD rats, compared to the saline control, the effects of CNO on sleep-waking states remained significant (waking; one-way RM ANOVA, F 7,44 = 7.13, p < 0.001; nonREM sleep, F 7,44 = 5.72, p < 0.001; REM sleep, F 7,44 = 5.34, p < 0.001) during the 3-6 h post-treatment period. However, compared to the non-DREADD rats, these effects were insignificant after the saline and 0.3 and 1.0 mg/kg injections of CNO. Only the rats treated with 3.0 mg/kg of CNO exhibited significantly decreased waking and increases in nonREM and REM sleep during the 3-6 h post-treatment period, suggesting that the CNO effect was dose-dependent ( Figure 3).

Chemoactivation of VLPO > PF-HA PRJ Neurons Increases Sleep in an Animal Model of Insomnia
Evidence supports that after acute exposure to a novel environment, rodents are not able to sleep well for 1-3 h [20,37,38]. We determined if the chemoactivation of VLPO > PF-HA PRJ neurons helps improve sleep during rats' exposure to a novel environment. Both the non-DREADD and DREADD rats were injected with either saline or CNO at 1 mg/Kg, a dose that was optimally effective in causing sleep-wake changes in experiment 1. After the injections, these rats were transferred to new cages similar to their home cage, but with new bedding and nesting materials, and their sleep-wake profiles were recorded for 5 h.
The sleep-wake profiles of the DREADD and non-DREADD rats after their exposure to a novel environment, plus the saline and CNO treatments, are shown in Figures 2 and 5. The effects of saline on waking, nonREM sleep, and REM sleep in the DREADD and non-DREADD rats were comparable during the first two h of the post-injection period. However, compared to the non-DREADD rats or the saline control, the DREADD rats treated with CNO exhibited significantly reduced amounts of waking (37% versus 29%; t = 2.92; p < 0.05) and increased amounts of nonREM sleep (55% versus 65%; t = 3.37; p < 0.01) during the first 2 h of the post-injection period (Figures 2 and 5A). REM sleep amounts changed marginally. The sleep-wake profiles of both groups of rats during the 2-5 h post-injection period were largely comparable ( Figure 5B). dependent (Figure 3).

Chemoactivation of VLPO > PF-HA PRJ Neurons Increases Sleep in an Animal Model of Insomnia
Evidence supports that after acute exposure to a novel environment, rodents are not able to sleep well for 1-3 h [20,37,38]. We determined if the chemoactivation of VLPO > PF-HA PRJ neurons helps improve sleep during rats' exposure to a novel environment. Both the non-DREADD and DREADD rats were injected with either saline or CNO at 1 mg/Kg, a dose that was optimally effective in causing sleep-wake changes in experiment 1. After the injections, these rats were transferred to new cages similar to their home cage, but with new bedding and nesting materials, and their sleep-wake profiles were recorded for 5 h.
The sleep-wake profiles of the DREADD and non-DREADD rats after their exposure to a novel environment, plus the saline and CNO treatments, are shown in Figures 2 and  5. The effects of saline on waking, nonREM sleep, and REM sleep in the DREADD and non-DREADD rats were comparable during the first two h of the post-injection period. However, compared to the non-DREADD rats or the saline control, the DREADD rats treated with CNO exhibited significantly reduced amounts of waking (37% versus 29%; t = 2.92; p < 0.05) and increased amounts of nonREM sleep (55% versus 65%; t = 3.37; p < 0.01) during the first 2 h of the post-injection period (Figures 2 and 5A). REM sleep amounts changed marginally. The sleep-wake profiles of both groups of rats during the 2-5 h post-injection period were largely comparable ( Figure 5B).  , (B)) of recording after exposure to the novel environment plus saline and 1 mg/kg of CNO treatments. The CNO-treated DREADD rats spent significantly less time in waking and more time in nonREM sleep after exposure to the novel environment than the non-DREADD and saline control groups. * p < 0.05; ** p < 0.01 of significance (paired t-test for within-group and t-test for between-groups comparisons).

Chemoactivation Increases Fos-IR in VLPO > PF-HA PRJ Neurons
To determine if the sleep-wake changes observed after CNO treatment in the DREADD rats during the dark phase were due to the activation of VLPO > PF-HA PRJ neurons in a subset of rats, we examined the effects of 1 mg/kg of CNO on Fos-IR in VLPO > PF-HA PRJ neurons in the DREADD and non-DREADD rats. The effects of CNO on Fos-IR in VLPO > PF-HA PRJ neurons and VLPO neurons not projecting to the PF-HA (VLPO non-PRJ ) in these two groups of rats during the dark phase are shown in Figure 6.

Discussion
Although the role of VLPO in sleep regulation is well recognized, the relative contributions of its various neuronal phenotypes and their interactions with different wakepromoting systems across the neuraxis remain not fully understood. In this study, we determined specific contributions of VLPO neurons projecting to the wake-regulatory PF-HA in sleep regulation by selectively activating them using DREADDs in WT Fischer-344 The number of labeled VLPO > PF-HA PRJ neurons in the DREADD and non-DREADD rats in the POA region were comparable (p > 0.05; Figure 6(B1,B3 and C)). No labeled VLPO > PF-HA PRJ neurons were observed in the VLPO or extended VLPO regions on the contralateral side in both groups of rats ( Figure 6(B2,B4)). In the non-DREADD rats that were mostly awake after CNO treatment, fewer VLPO > PF-HA PRJ and VLPO non-PRJ neurons exhibited Fos-IR on the ipsilateral side ( Figure 6B,C). However, in the CNO-treated DREADD rats, the number of VLPO > PF-HA PRJ and VLPO > PF-HA non-PRJ neurons exhibiting Fos-IR was significantly higher than the ipsilateral side of the CNO-treated non-DREADD rats ( Figure 6B,C). Although the number of VLPO > PF-HA PRJ +/Fos+ neurons was highest in the VLPO core, VLPO > PF-HA PRJ and VLPO non-PRJ neurons expressing Fos-IR were also observed in its adjacent regions.
The effects of CNO on Fos-IR in the DREADD rats were mainly restricted to the ipsilateral side ( Figure 6B,C). In the DREADD rats, the number of total Fos-IR neurons was significantly higher on ipsilateral side than on the contralateral side (35 versus 15; p < 0.05, Mann-Whitney rank sum test) or those observed ipsilaterally in the CNO-treated non-DREADD rats (35 versus 18; p < 0.05; Mann-Whitney rank sum test). The number of Fos+ neurons in equivalent regions on the contralateral side was comparable in the DREADD and non-DREADD rats.

Discussion
Although the role of VLPO in sleep regulation is well recognized, the relative contributions of its various neuronal phenotypes and their interactions with different wakepromoting systems across the neuraxis remain not fully understood. In this study, we determined specific contributions of VLPO neurons projecting to the wake-regulatory PF-HA in sleep regulation by selectively activating them using DREADDs in WT Fischer-344 rats. We found that the selective activation of VLPO > PF-HA PRJ neurons in freely behaving rats significantly increased sleep, especially nonREM sleep, with a concomitant decrease in waking during the early dark phase, a time when rats are typically awake and sleep propensity is minimal. The activation of VLPO > PF-HA PRJ neurons also induced sleep during exposure to the novel environment during the light phase, which caused acute arousal and has been used as a rodent model of insomnia. These findings suggest that VLPO > PF-HA PRJ neurons and their interactions with wake-regulatory neurons in the PF-HA constitute a critical component of the preoptic-hypothalamic sleep-wake regulatory circuitry.
So far, transgenic mice models are being extensively used to dissect the role of specific cell types in regulating behavioral functions, including sleep-waking. Currently, transgenic rat models that may help address such questions are unavailable, although rats have been the animal model of choice for studying sleep and sleep disorders for decades. To our knowledge, this is the first study in WT rats where combinations of two viral vectors have been used to retrogradely deliver the Cre-recombinase gene specifically in VLPO > PF-HA PRJ neurons and further express hM3Dq in those neurons to selectively activate them for delineating their specific contributions to sleep-wake functions. It is pertinent to note that while the AAV2-DIO-hM3D(Gq) vector injected into the VLPO likely transfected most neurons in its diffusion field, hM3Dq expression was restricted to only VLPO > PF-HA PRJ neurons containing a Cre-dependent recombinant switch of this gene. In addition, hM3Dq expression in VLPO > PF-HA PRJ neurons may depend on multiple factors, including efficiencies of the AAVret-vector transfection in terminals, the retrograde transportation via axons, expression of Cre and hM3D(Gq) genes, "unlocking" of hM3D(Gq) gene via recombination, and, finally, expression of the hM3D(Gq) receptors. The strength of this approach is that we were able to precisely target VLPO > PF-HA PRJ neurons in WT rats and reliably study them across experimental paradigms. Such approaches, i.e., using combinations of viral vectors for targeting specific/local neuronal populations in WT animals, may also be a valuable tool in studies with limited availability of transgenic mice models or studies where maintaining transgenic models may be cost prohibitory, if available, e.g., aging research.
The sleep-wake changes observed in this study were specifically due to the CNOmediated activation of VLPO > PF-HA PRJ neurons and not a nonspecific effect, i.e., caused by VLPO > PF-HA nonPRJ neurons or CNO itself, as is evident from the following observations: (i) increased sleep with a concomitant decrease in waking after CNO injection in the DREADD group was consistent across two different physiological paradigms, reversible, and somewhat dose-dependent; (ii) various doses of CNO in the non-DREADD group did not produce any significant sleep-wake changes; (iii) saline injection in both the DREADD and non-DREADD rats did not produce any significant sleep-wake changes, and both groups exhibited comparable sleep-wake organization; (iv) CNO caused increased Fos-IR in VLPO > PF-HA PRJ neurons in the DREADD group and not in the non-DREADD group, indicating that the sleep-wake changes observed after CNO injection in the DREADD group were due to the hM3Dq-mediated selective activation of VLPO > PF-HA PRJ neurons; and (v) the efficacy of AAV2retro-EF1a-Cre in retrogradely delivering Cre-recombinase into projecting neurons, and that AAV-hSyn-DIO-hM3D(Gq) or AAV-DIO-GCAMP6s is selectively expressed in Cre expressing neurons, has already been validated [35,44]. However, in this study, instead of an AAV-hSyn-DIO-hM3D(Gq) control, we used AAV-DIO-GCaMP6s as a control, and we believe that it did not have any significant bearing on the outcome of this study because CNO cannot activate VLPO > PF-HA PRJ neurons expressing GCaMP6s, as well as the other control findings we have described above.
In this study, CNO treatment significantly increased sleep, especially nonREM sleep, with a concomitant decrease in waking in the DREADD rats during the early dark phase, when rats are primarily awake and sleep propensity is minimal. Further, CNO treatment significantly increased c-fos expression ipsilaterally in VLPO > PF-HA PRJ neurons in the DREADD animals, indicating that only the unilateral activation of VLPO > PF-HA PRJ neurons contributed to the observed sleep-wake changes, and hence they were moderate. Further analysis of sleep-wake organization revealed that while the frequency of very short waking episodes increased significantly, there was a significant decline in the number of longer bouts. On the contrary, the frequency of both >30-120 s and >120-300 s episodes of nonREM sleep increased significantly. These findings are consistent with the role of VLPO sleep-active neurons in the initiation and maintenance of nonREM sleep, as their discharge activity increases during the transition from waking to sleep and remains elevated during nonREM sleep [6,7]. VLPO lesions severely shorten nonREM sleep bout duration [5]. It is plausible that while under circadian influences, the animals kept attempting to enter waking but could not maintain sustained waking due to the activation of VLPO > PF-HA PRJ neurons with inhibitory influences on the wake-promoting systems in the PF-HA. This interpretation is further supported by another finding of this study that the activation of VLPO > PF-HA PRJ neurons also suppressed waking and increased sleep after exposure to the novel environment, which caused acute behavioral arousal and suppression of sleep. Although we could not study the effects of CNO on Fos-IR after exposure to the novel environment, it is likely that, similar to the dark phase, sleep induced after exposure to the novel environment was also caused by the activation of VLPO > PF-HA PRJ neurons by CNO.
The DREADD animals also exhibited increased Fos-IR in VLPO > PF-HA non-PRJ neurons ipsilaterally after CNO treatment. It seems unlikely that these neurons exhibited Fos-IR due to increased sleep as no such increase occurred contralaterally. It is plausible that these may be local sleep-active neurons activated by VLPO > PF-HA PRJ neurons. However, it seems likely that these are VLPO > PF-HA PRJ neurons with very little or no visible expression of mCherry. hM3D(Gq) and mCherry are expressed using the bi-cistronics principle. When the same promoter regulates two gene expressions, the expression of (a second) mCherry gene could be considerably weaker compared to (the first) hM3D(Gq). Thus, some neurons, while visibly mCherry-negative (single Fos), may express enough hM3D(Gq) to be activated by CNO.
In this study, we did not ascertain the neuronal phenotypes in the PF-LHA that were affected by AAV-EF1a-Cre injections and to which the VLPO > PF-HA PRJ neurons projected to. However, based on the microinjection needle tips' localization and the viral vector's estimated diffusion field, it is likely that neurons in the PF-HA, including the perifornical area, LH, and parts of dorso-and ventro-medial hypothalamic areas, were affected. These regions predominantly contain wake-active neuronal groups, including HCRT and glutamatergic neurons, although some sleep-active neuronal groups, e.g., MCH and a subpopulation of GABAergic neurons, have also been reported [17,18,[45][46][47]. The glutamatergic activation of PF-HA induces arousal and suppresses sleep in various animal models, whereas the adenosinergic inhibition of neurons in this area promotes sleep [48][49][50]. Moreover, the activation of MnPO neurons projecting to the PF-LHA predominantly inhibited wake-active neurons while exerting minimal effects on sleep-active neurons [29]. Our findings that the activation of VLPO > PF-HA PRJ neurons induced sleep with a concomitant decrease in waking are consistent with these studies.
The VLPO > PF-HA PRJ neurons that were activated by CNO were localized mainly in the VLPO core. The ability of only a subset of VLPO > PF-HA PRJ neurons to promote sleep, especially nonREM sleep, suggests that these neurons constitute a critical component of hypothalamic sleep-regulatory circuitry. This is consistent with findings from earlier studies that found that fos expression in the VLPO cluster neurons correlates primarily with nonREM, but not REM, sleep, and a loss of neurons in the VLPO cluster is closely related to a loss of nonREM, but not REM, sleep behavior in rats [5]. Although a significant majority of VLPO PRJ neurons are GABAergic/galaninergic [2,9,12,[19][20][21], the neurotransmitter phenotype(s) of VLPO > PF-HA PRJ neurons examined in this study and their relative contributions or the role of co-transmitters in these neurons in influencing PF-HA neuronal groups and this circuitry remains to be fully characterized.
We had shown earlier that POA warm-sensitive neurons are predominantly sleepactive and exhibit increased sensitivity during non-REM sleep. Thus, their activation not only causes sleep, but also causes a decline in body temperature that accompanies sleep [51]. We also found that the activation of POA warm-sensitive neurons by local warming inhibits wake-active neuronal groups, including those in the PF-LHA and dorsal raphe nucleus [28,52]. Some of these warm-sensitive neurons are likely GABAergic/galaninergic VLPO PRJ neurons. Thus, although we did not record body temperature, it is likely that activating VLPO > PF-HA PRJ neurons would also induce hypothermia, as reported earlier after the activation of VLPO galaninergic neurons [20].
Further, the evidence supports the conclusion that VLPO neurons project to multiple wake-promoting regions in the hypothalamus and brainstem. The activation of VLPO neurons has been shown to inhibit PF-HA neurons and neurons in other arousal systems they project to [28,52]. It is plausible that in this study, the VLPO > PF-HA PRJ neurons that were activated by CNO also projected to other wake-promoting regions; thus, the possibility that their activation caused the simultaneous inhibition of other wake-promoting systems, contributing to the evoked sleep response, cannot be ruled out. The validation of such interpretations requires further studies. In addition, the evidence suggests that gonadal hormones may affect the sleep-wake organization, as well as Fos-IR in VLPO neurons, differently in male and female subjects [53][54][55]. To avoid these confounds, we used only male animals, which created a limitation of this study.

Conclusions
This study found that the selective activation of VLPO > PF-HA PRJ neurons induces sleep, with a concomitant decrease in waking during the dark phase, when rats are typically awake, and after exposure to a novel environment during the light phase, which causes acute arousal, and has been used as a rodent model of insomnia. These findings suggest that VLPO > PF-HA PRJ neurons and their interactions with wake-regulatory PF-HA neurons constitute a critical component of the preoptic-hypothalamic sleep-wake regulatory circuitry.