After the post-surgical recovery period of 3-7 days, rats were habituated to the experimenter, the sound-attenuated recording cage (electrically shielded: 2.5 × 1.5 × 1.5 ft), and free-moving polygraphic recording conditions for 7-10 days (Grass model 15 Neurodata Amplifier System with 15A94 EEG/PSG quad amplifier, Grass Instrument Division, Astor-Med Inc., West Warwick, RI, USA). The last day of these adaptation recording sessions was determined when for 3 consecutive days, the day-to-day variation on the percentage of REM sleep was less than 5% of their total amount of REM sleep. All adaptation-recording sessions were performed between 9:00 and 16:00 h, when rats are normally sleeping. During recovery, habituation, and free moving recording conditions (adaptation recording sessions), all rats were housed under the same 12/12 hr light/dark cycle with free access to food and water.

For the purpose of determining possible effects on sleep and wakefulness, polygraphic data were captured on-line in a computer using “Gamma” software (Grass product group, Astro-Med, Inc., West Warwick, RI, USA). From this captured data, three behavioral states (W, SWS, and REM sleep) were distinguished and scored visually using “Rodent Sleep Stager” software (Grass product group, Astro-Med, Inc., West Warwick, RI, USA). These three states were scored in successive 5-s epochs. The physiological criteria for the identification of these wake-sleep states were described in detail in our earlier publications (Datta et al., 2002, 2004). This epoch length allowed us to quantify the short periods of REM sleep (3-5 sec) in the rats where REM sleep was selectively terminated (Datta et al., 2004; Bandyopadhya et al., 2006).

On the day after the last adaptation recording session, when day-to-day variation on the total amount of REM sleep stabilized, rats underwent a session of baseline sleep-wakefulness recordings (between 9:00 and 14:00 h). The following day, the rats were subjected to an experimental recording session in which each animal was connected to the polygraphic recording system at 8:55 h. In the next room, polygraphic signs of the sleep-wake cycle were continuously observed on a computer monitor (between 9:00 and 14:00 h). At this point, the 21 rats were randomly divided into three groups. Group 1 (n = 7 rats): These rats were recorded for 5 hr (between 9:00 and 14:00 h) of undisturbed sleep-wakefulness (hereafter group 1 is labeled as “control/baseline REM sleep”; “CR”). Group 2 (n = 7 rats): The experimental protocol for these animals was almost identical to the protocol described above for “CR” group, except that for group 2 animals, from 12:00 to 14:00 h REM sleep episodes were selectively terminated at the beginning (within 3-5 sec) of each episode while the animals were connected to the polygraphic recording system (hereafter group 2 is labeled as “Low REM sleep”; “LR”). Group 3 (n = 7 rats): In this group of animals, REM sleep episodes were selectively terminated between 9:00 and 12:00 h and after which they were allowed to have undisturbed sleep-wake (hereafter group 3 is labeled as “High REM sleep”; “HR”). In summary, the main difference between these three groups of animals is the amount of REM sleep present during the last two hours of recording (between 12:00 and 14:00 h). The “CR” group had normal amounts of REM sleep. The “LR” group had about 82.8% less REM sleep than the “CR” group. The “HR” group had 130.2% more REM sleep then the “CR” group (Table 1). The experiments were conducted using three recording chambers and thus running on three animals at one time. At each time, three different rats were from three different groups (CR, LR, and HR).

At the end of experimental recording session (at 14:00 h) all rats were transferred to a CO2 chamber and exposed to CO2 for 2 min to make them semi-conscious, as described earlier (Saha and Datta, 2005). These semi-conscious rats were deeply anaesthetized with sodium pentobarbital (70 mg/kg, i.p.) and perfused transcardially with heparinized cold phosphate buffer (0.1 M, pH 7.4) followed by 4% paraformaldehyde in 0.1 M phosphate-buffered saline (PBS). The brains were removed and post-fixed in 4% buffered paraformaldehyde for 3-4 hours at 4°C. Post-fixed brains were then stored overnight in a solution of 20% sucrose in 0.1 M PBS (cryoprotectant) at 4 °C.

Coronal sections were cut through the selected portion of the brainstem (from interaural 1.5 mm to interaural -1.0 mm; based on the stereotaxic atlas of Paxinos and Watson, 1998) at 25-μm thickness on a cryostat (Bright Inst., Co., Huntington, UK). Four series of adjacent sections were collected and stored in 0.1 M PBS containing 0.005% sodium azide solution at 4 °C. In each series, adjacent sections were collected at interval of 100 μm. Next; three series of the tissue sections (first, third, and fourth) were processed on a rotary shaker inside a temperature-controlled see through refrigerator (GDM-3, Eastern Baker Supply, Boston, MA). The first series of sections was processed for pCREB expression as described elsewhere (Saha and Datta, 2005). In brief, free floating sections were first rinsed in three changes of PBS (10 min each) then immersed in 0.5% hydrogen peroxide for 30 min at 4 °C to remove endogenous peroxidase and to lyse red blood cells. Sections were then carried through rinses of PBS either three times until small gas bubbles were no longer present. Next, sections were treated with 0.5% Triton X-100 in PBS for 1 h to facilitate antibody penetration. Sections were then incubated for 30 min at 4 °C in a solution of PBS containing 10% normal goat serum and 0.2% Triton X-100 to block non-specific binding and to reduce background staining. The sections were then incubated in the presence of the monoclonal primary antiserum, phosphorylated CREB125-135 (1:5,000 dilution; Upstate Biotechnology, Lake Placid, NY) for 48 hours at 4°C. This primary antiserum solution was made in PBS containing 1.0% normal goat serum and 0.1% Triton X-100. Following incubation in the primary antiserum, sections were rinsed in three changes of PBS (10 min each) and then incubated in the secondary antibody (1:200, biotinylated goat anti-mouse-IgG; Vector Labs, Burlingame, CA) for 3 hours at 4 °C. The secondary antibody solution was made in PBS containing 1% normal goat serum and 0.1% Triton X-100. Following incubation with the secondary antibody, sections were washed in PBS (3 times, 10 min each) and then incubated in PBS containing avidin-biotin-peroxidase complex (1:100, Peroxidase Standard PK-4000, Vectastain ABC kit, Vector Labs, Burlingame, CA) for 1h at 4 °C. The tissue sections were then washed in PBS 3 times for 10 min each time. To visualize the immunoreaction product with light microscopy, free-floating sections were placed in 0.1 M PBS solution, containing 0.025% diaminobenzidine (DAB), 0.06% nickel ammonium sulfate, and 0.0027% hydrogen peroxide, for 4 min. This immunoreaction product is dark purple in color. In these sections, ongoing reaction was then stopped by three rinses in PBS containing 0.1% Triton X-100 and 0.1% sodium azide (PBST-Az). For double immunostaining experiments, some of these free-floating pCREB immunostained sections (between interaural 1.2 mm and 0.0 mm) were then incubated in rabbit anti-choline acetyltransferase (ChAT) antiserum (1:2000; Chemicon International, Temecula, CA) for 3 days at 4 °C. Following incubation in the primary antiserum, sections were rinsed in three changes of PBS (10 min each) and then incubated in the secondary antibody (1:200, biotinylated goat anti-rabbit-IgG; Vector Labs, Burlingame, CA) for 3 hours at 4 °C. Following incubation with the secondary antibody, sections were washed in PBS (3 times, 10 min each) and then incubated in PBS containing avidin-biotin-peroxidase complex (1:100, Peroxidase Standard PK-4000, Vectastain ABC kit, Vector Labs, Burlingame, CA) for 1h at 4 °C. The tissue sections were then washed in PBS 3 times for 10 min each time. To visualize the immunoreaction product with light microscopy, free-floating sections were placed in 0.1 M PBS solution, containing 0.025% diaminobenzidine (DAB), and 0.0027% hydrogen peroxide, for 15 min. This immunoreaction gives a brown color on the cholinergic cells. The sections were then mounted onto gelatin subbed glass slides and allowed to air dry. Slides were then sequenced through a series of graded ethanol followed by three washes in Histoclear (National Diagnostics, Atlanta, GA). From the final Histoclear wash, the sections were coverslipped using Permount (Fisher Scientific, Fair Lawn, NJ). Similarly, free-floating sections between interaural 1.2 mm and interaural 0.0 mm of the third series were double immunostained for c-Fos and ChAT using a protocol as described in another publication (Verret et al., 2005). In brief, free-floating sections were successively incubated in: (i) a rabbit antiserum to c-Fos (1:3000; Ab-5; Oncogene, CA) in PBST-Az for 72 h at 4 °C; (ii) a biotinylated goat antirabbit IgG (1:2000; Vector Laboratories, Burlingame, CA) for 90 min at room temperature; and (iii) an ABC-HRP solution (1:1000; Elite kit, Vector Laboratories) for 90 min at room temperature. The sections were immersed in a 0.05 M Tris-HCL buffer (pH 7.6) containing 0.025% DAB, 0.0027% hydrogen peroxide and 0.06% nickel ammonium sulfate for 4 min at room temperature. The reaction was stopped by three rinses in PBS containing 0.1% Triton X-100 and 0.1% sodium azide (PBST-Az). Free-floating sections were then immunoreacted for the ChAT, as described above. Immunoreaction product for c-Fos was dark purple and ChAT was brown, reflecting nuclear (purple) and cytoplasmic (brown) staining. To confirm the specificity of pCREB immunostaining, some sections of the fourth series were incubated with pCREB antibody, which was pre-absorbed overnight at 4°C with an excess (10 μM) of synthetic phosphopeptide corresponding to residues 125-135 (H-Arg-Glu-IIe-Leu-Ser-Arg-Arg-Pro-pSer-Tyr-Arg-OH) of CREB (Anaspec, San Jose, CA). Immunostaining of these sections was completely abolished when the pCREB antiserum was pre-absorbed with corresponding phosphopeptide. To ensure the absence of nonspecific single or dual immunostaining, controls in the absence of c-Fos and ChAT primary antibodies and in the presence of normal sera were routinely run on the fourth series of sections.

To delineate architectonic borders, a second series of sections were processed for Nissl staining using thionin as described earlier (Saha and Datta, 2005). Sections mounted on gelatin-coated slides were airdried then defatted by dipping them in a solution containing 100% ethanol and chloroform (in a ratio of 1:1) for one hour. Sections were then re-hydrated through a series of graded alcohols and distilled deionized water, stained in 0.05% thionin for 1.5 min, differentiated through graded alcohols, cleared in Histoclear, and coverslipped using Permount.

The sections were microscopically examined (Eclipse E800, Nikon Corp., Tokyo, Japan) and images were captured using a color CCD camera (Optronics, Goleta, CA) using Image-Pro Plus 5.1 imaging software (Media Cybernatics, Inc. Silver Spring, MD). The experimenter, while performing immunohistochemical staining and cell counting, did not have knowledge of the experimental group to which the individual brains belonged. ChAT-immunoreactive cells and c-Fos-immunoreactive nuclei in the PPT and LDT and pCREB-immunoreactive (p-CREB-IR) nuclei in seven other areas of the brain were counted in a box (300 × 300 μm) centrally placed on those selected areas on both sides of the brain in three consecutive sections 100 μm apart (Fig. 1). The results of these six counts per animal were then averaged. For the PPT, ventrolateral periaqueductal gray (vlPAG), lateral pontine tegmentum (LPT), dorsal raphe nucleus (DRN), pontine reticular nucleus oral (PnO), cells were counted at levels corresponding to about ~1.00, ~1.10, and ~1.20 mm anterior to interaural zero in the atlas of Paxinos and Watson (1998). For the laterodorsal tegmental nucleus (LDT) and medial pontine reticular formation (mPRF), cells were counted at levels ~0.10, ~0.20, and ~0.30 mm anterior to interaural zero. For the locus coeruleus nucleus (LC) and dorsal subcoeruleus nucleus (SubCD), cells were counted at levels ~0.6, ~0.7, and ~0.8 mm posterior to interaural zero.

In the past, all single cell-recording studies have demonstrated that REM-off cells are located within the LC and DRN, but some recent immunohistochemical studies have offered contrasting results by suggesting REM-off cells are located mainly within the vlPAG and LPT. As a consequence of this discrepancy, in addition to the LC and DRN, we also examined the changes in pCREB expression in the vlPAG, and LPT as a function of three different REM sleep conditions. The numbers of pCREB+ cells, both in the LC and DRN varied significantly as a function of REM sleep and were significantly greater in the LR group compared to the CR and HR groups (Fig. 4 and Table 3). In the LC, the number of pCREB+ cells in the LR group was 1543.32% more than in the CR group and 2011.67% more than in the HR group (Table 3). Similarly, in the DRN, the numbers of pCREB+ cells in the LR group was 437.40% more than in the CR group and 2176.08% more than in the HR group (Table 3). These results demonstrated that in the LC and DRN, pCREB expression increases with low REM sleep. In contrast, in the vlPAG and LPT, the numbers of pCREB+ cells were not significantly different from the CR and HR groups (Fig. 5 and Table 3). These results indicated that changes in the total percentages of REM sleep does not significantly change the numbers of pCREB+ cells in the vlPAG and LPT.

A number of studies using single cell recordings, localized lesion, and pharmacological manipulations have shown that the mPRF, PnO, and SubCD participate positively in the regulation of REM sleep. In this study, we also examined the changes in pCREB expression in the mPRF, PnO, and SubCD as a function of three different conditions of REM sleep (Fig. 6). The results show that the numbers of pCREB+ cells in these regions varied significantly (one-way ANOVAs) as a function of REM sleep conditions (Table 3). The results of individual comparisons with the CR group show that the number of pCREB+ cells in the mPRF was 186.28% greater in the HR group and 72.77% less in the LR group. Similarly, the numbers of pCREB+ cells in the PnO and SubCD, were significantly more in the HR group and less in the LR group compared to the CR group (Table 3). These results suggest that the expression of pCREB in the mPRF, PnO, and SubCD, increased with increased REM sleep and decreased with the reduced REM sleep.

In the DRN and LC, low REM sleep resulted in increased numbers of neurons expressing pCREB. The numbers of pCREB expressing neurons in the DRN in the high REM sleep condition was very low compared to control and low REM sleep conditions. These results are consistent with single cell recording studies in the cat and rat. For example, a number of studies have recorded single cell activity patterns of DRN neurons in the cat across sleep-wake cycle and all of those studies have shown DRN neurons remain active throughout the W and SWS period and stop firing as soon as animal enters into REM sleep (Hobson et al., 1975; McCarley and Hobson, 1975; McGinty and Harper, 1976; Trulson and Jacobs, 1979; Lydic et al., 1983). Similarly, a number of single cell recording studies in the rat and cat have shown that most of the neurons of the LC are active during W and SWS but remain silent during REM sleep (Chu and Bloom, 1973; Aston-Jones and Bloom, 1981; Foote et al., 1983). The results of the present study also demonstrated that the numbers of pCREB expressing neurons in the LC and DRN of individual animals are negatively correlated with their total percentages of REM sleep. Collectively, the results of present study and earlier single cell recoding studies could be interpreted as activation of DRN and LC neurons promoting the reduction of REM sleep. Indeed, the local application of a selective 5-HT1A receptor agonist, which suppresses neuronal activity of the DRN neurons (Sprouse and Aghajanian, 1987), increases REM sleep in the cat and rat (Portas et al., 1996). In another study, it has been shown that the localized cooling of the LC induces REM sleep in the behaving cats (Cespuglio et al., 1982). Collectively, these results suggest that the withdrawal of LC and DRN neuronal activity facilitate REM sleep.

One recent study has suggested that the vlPAG and LPT neurons in the rat are of the REM-on type by showing rebound REM sleep increases number of cells expressing c-Fos in the vlPAG and LPT (Verret et al., 2005). One single cell recording study in the PAG of cat had encountered few cells in the vlPAG that showed bursting type of activity during REM sleep, an indication of REM-on type of activity (Thakkar et al., 2002). Consistent with the results of this single cell recording study, during the low REM sleep condition, the numbers of pCREB labeled cells in the vlPAG and LPT were slightly less (see Table 3). On the contrary, another study had shown that lesions of the vlPAG and LPT increase REM sleep in the rat (Lu et al., 2006). Based on the lesion results, it has been proposed that vlPAG and LPT neurons in the rat are REM-off type (Lu et al., 2006). To our surprise, results of the present study did not reveal any significant changes in the number of pCREB expressing cells in the vlPAG or LPT. Given that the vlPAG is involved in the regulation of pain (Reynolds, 1969; Dostrovsky et al., 1983; Fardin et al., 1984; Auerbach et al., 1985), the increased c-Fos activation in rebound REM sleep after flowerpot exposure might have been the effect of a homeostatic phenomenon to attenuate stress and pain caused by the flowerpot method. Since the anticipation of stress and pain reduces REM sleep and since vlPAG and LPT neurons are involved in controlling stress and pain, the decreased REM sleep after vlPAG and LPT lesion might have been caused by the absence of such pain attenuating systems of our brain. Thus, at this time, it is reasonable to suggest that the vlPAG and LPT may not be involved in the regulation of REM sleep.