All the procedures were approved by the Institutional Animal Care and Use Committee of the University of California at Los Angeles and of the Veterans Administration Greater Los Angeles Health Care System. Prepro-hypocretin mutated C57BL/6J-129/SvEv mice were generated as reported previously(Chemelli et al., 1999
). They were maintained as heterozygotes and crossed to obtain null mutants (KO) and WT littermates. The genotype of Hcrt KO mice was identified by PCR using the previously described method(Kayaba et al., 2003
). Briefly, DNA extracted from a tail biopsy was amplified using the 5'primer, GAC CTA TCA GGA CAT AGC GTT GGC and the 3' primer, TCA CCC CCT TGG GAT AGC CCT TTC for the mutant allele and the 5' primer, GAC GAC GGC CTC AGA CTT CTT GGG with the same 3' primer to identify the WT allele. Mice used in this study were 2–4 months old at the beginning of the procedures. Male Hcrt KO mice (n=26) and male WT littermates (n=70) were used. Genotyping was further confirmed by brain immunohistochemistry using rabbit anti-Hcrt B (H-003-32, Phoenix Pharmaceutical, Belmont, CA, USA). Animals were kept inside sound attenuating light-tight chambers with white noise produced by ventilating fans in a laboratory room in a one story building with unoccupied adjacent rooms, and entry only by experimenters running the study. The room was maintained at 22±1°C on a 12 h light (45 lux) dark (0.03 lux) cycle (lights on 7 AM and off 7 PM). All animals were obtained from the UCLA animal facility.
Experimental sessions were conducted in mouse operant conditioning cages (17.8 cm × 17.8 cm × 30.5 cm, model H10-11M-TC, Coulbourn Instruments, Whitehall, PA, USA. All equipment for behavioral tests was obtained from this company). These were located within ventilated sound-attenuating isolation cubicles (65.4 cm × 52.7 cm × 61.6 cm, model H10-24 TA, with Modular Floor, model H10-11M-TC-SF). Each cubicle was equipped with an infrared CCD camera (KPC-S50NV, Sony) connected to a computer. Every experimental session was recorded for behavioral analysis. Each cage was equipped with one lever (1.9 cm × 0.9 cm) located 0.6 cm above the grid floor, triple LCD cue lights (H11-02M), a speaker on the opposite side of the recording chamber and a house light on top of the cage. For the food reinforcement paradigm, pellets were delivered from a motor driven feeder (model H14-23M) to either a pellet trough or, in animals implanted for the EEG studies, to a plastic tray (9.5 cm long × 4 cm wide × 2 cm high) located on the left side of the lever, so as to prevent the implant from impeding access to the food. For the water reinforcement paradigm, water was delivered by a motor-driven dipper (model H14-05M), which was accessible through a receptacle opening (2.3 cm × 2.5cm) located on the left side of the lever. The dipper contained a 0.01 cc volume cup machined into the end of the arm. For the negative reinforcement paradigm, electrical stimulation was delivered to the grid cage floor (1 s, 150–500 μA) (Model H13-15). Light in the chamber during light phase testing was 11.4 lux (white house light) and 0.07 lux (red cue light). During dark phase testing, only the red cue light was on. The red cue light flickered, 0.1 s on then 0.1 s off, for 3 s when a reward was delivered. For the comparison of Fos activation with and without light, only the red cue light was activated. When the sessions ended, all interior lights were off (0.03 lux) and the subjects were returned to their home cages.
Quantification of food intake
For the food bar press condition, animals were maintained at 85–90% of their initial body weight by restricting food intake. Food (LabDiet, PMI Nutrition, Brentwood, MO, USA) was rationed based on each animal's daily weight. Once the animal's weight became stable, food intake was restricted and measured for 120 min between 11:00 AM and 1:00 PM each day for 2 weeks in the home cage. Water was available ad libitum.
Quantification of water intake
For the water bar press condition, animals were water restricted, allowing access twice daily for a total of 90 min. Drinking sessions were divided into two periods: 60 min (11:00 AM-12:00PM) and then 30 min (3:00 PM-3:30 PM). Once the animals had adapted to the drinking schedule, water intake was measured for 60 min between 11:00 AM and 12:00 PM using a graduated cylinder each day for 2 weeks in the home cage. Food was available ad libitum.
Mice were food restricted to maintain 85–90% of their original body weight following the same procedure as for the food operant task. Once the animals' weight had stabilized, mice were placed in the operant conditioning chambers daily for 120 min, starting 2 h after the onset of the light phase (between 9:00 AM and 1:00 PM) or 2 h into the dark phase (between 9:00 PM and 1:00 AM). Both the house light and the cue light were on during the light sessions. All experimental sessions had a duration of 120 min and lasted for 5 weeks. No food, water or foot shock was given during the sessions.
Surgical procedures: EEG and EMG electrode implantation
Eight animals, 5 KO and 3 WT were implanted with cortical electroencephalogram (EEG) and neck muscle electromyogram (EMG) electrodes under aseptic conditions. Anesthesia was induced with a mixture of ketamine/xylazine (100 mg/kg/15 mg/kg, i.p.) and then maintained with a gas mixture of isoflurane in oxygen (0.6–1.2%) after the animals were placed in the stereotaxic device. Body temperature was maintained with a water-circulated heating pad (Gaymar Industries, Orchard Park, NY, USA). The head was positioned in the stereotaxic frame and the skull was exposed. Four stainless steel screw electrodes, two in the frontal (AP: +2.6mm, L: +/− 1.5mm, relative to bregma) and two in the parietal (AP: −1.5mm, L: +/− 1.5mm, relative to lambda) bones, were implanted to record the EEG. Two other stranded stainless steel wire electrodes were placed in the neck muscles to record EMG activity. All six electrode leads were inserted into a plastic head plug (SL6C/SB, Plastics One, Roanoke, VA, USA) that was then fixed to the skull with dental cement. A post surgical recovery period of two weeks was allowed before any training was conducted.
EEG and EMG recordings
Recordings were done of the mice both during performance in the operant chamber and during the spontaneous sleep/wake cycle in their home cages. Cortical EEG was filtered between 0.3 and 100 Hz and sampled at a rate of 128 Hz. The EMG was filtered between 30 Hz and 3 kHz and sampled at 1 kHz. EEG and EMG signals were digitized with the CED1401 Plus (Cambridge Electronic Design Ltd., Cambridge, UK) interface and recorded on a computer using Spike2 software (Cambridge Electronic Design Ltd.).
Operant tasks Positive reinforcement, food reward
The training schedules for all behavioral and histological experiments are depicted in the diagrams of (operant tasks) and (non-operant tasks). Mice were food restricted to maintain 85–90% of their original body weight (food amount: about 3–3.5 g/day vs. regular portion 4–4.5g/day) and were exposed daily to 20 mg food pellets (Dustless Precision Pellets F0071, BioServe, Frenchtown, NJ, USA). All training sessions and most experiments were carried out within this weight range. In selected experiments described above animals were kept at 70–75% or 100% of their initial weight.
Diagram of the training schedules for operant conditions for WT and KO animals
Diagram of the training schedules for non-operant conditions for WT and KO animals
Once the animals' weight had stabilized (14 days; average weight for the WT was 28.5±0.7 g and 29.3±0.9 g for the KO), training began with one week of acclimation (120 min per day) to the operant conditioning cages, followed by two 30-min sessions of magazine training, during which a food pellet was delivered on a variable interval-1 min with no contingency on lever pressing (VI-1 min) (food pellet was delivered randomly between 20 s and 100 s by the operating computer) and upon bar pressing. Shaping then commenced to train the animals to press the lever for food. Once shaping was completed, the animals were placed on a fixed ratio of 1 (FR1) schedule in which a food pellet was dispensed upon a single press of the lever. The session ended after 30 min or 20 pellets, whichever came first. Successful completion of the FR1 schedule was achieved when the animals received all 20 pellets within 10 minutes in three consecutive sessions. The ratio was then increased to FR3 and finally FR5 and the session duration increased to 60 min. Upon successful completion of FR5 (3 consecutive sessions in which the animal finished the 20 pellets in 15 min), mice were trained on a progressive ratio of 1 (PR1) schedule of reinforcement, where the number of lever presses required to receive a reinforcement is increased by 1 after each reward, i.e., 1, then 2, then 3, then 4 presses etc. were needed to get successive pellets. Each session had a total duration of 120 min. The step size was increased from PR1 to PR3 and finally to PR5 with 5 consecutive sessions under each protocol. There was no limit to the number of food pellets that could be received in PR3 and PR5 sessions. However, in PR5 the apparatus was programmed to terminate the session if the animal did not press the lever for 15 min. The average weight of the WT was 29.1±1.0 g and 28.7±1.0 g for the KO after the initial training
For animals that completed the 120-min session, the final ratio of bar presses per pellet reached at the end of the session was used for analysis. For the animals that did not complete the 120-min PR5 session (early termination), the break point (BP) was operationally defined as the last ratio completed (resulting in the delivery of a food reward) before the session terminated. The BP or final ratio of the last three PR5 trials were averaged and used for statistical comparisons. Food supplements were given after each session to keep all animals at their respective body weights (70–75%, 85–90% or 100% of their initial weight).
Operant sessions for food reward were run during both the light and the dark phases. For the light phase sessions, animals were run starting 2 h after the onset of the light phase (between 9:00 AM and 1:00 PM). A house light and a red LCD cue light above the lever were on throughout the light sessions. For the dark phase sessions, animals were run 2 h after the onset of the dark phase (between 9:00 PM and 1:00 AM).
Positive reinforcement, water reward
Mice were trained for water reward on a regime similar to that used for food reward. Mice were first water restricted for 14 days, having free access to water for 90 min per day. The animals were then acclimated to the operant conditioning cages for 60 min per day for a week. This was followed by two days of 1 h sessions of magazine training during which water was available to the animals for 4 s per delivery under a VI-1min schedule and upon bar pressing. During the second phase of training, animals were first trained to press the lever for water (shaping) and then put on a FR1 schedule under which water was delivered upon a single press of the lever. The session ended after 60 min. After each training session, animals were allowed access to water twice a day for a total of 90 min. Drinking sessions were divided into two periods: 60 min and then 30 min with a 4h interval between the sessions. Successful completion of the FR1 schedule was achieved when the animals pressed at least 100 times within 60 min in three consecutive sessions. The ratio was then increased to FR3, then FR5. Upon successful completion of the FR5 schedule (at least 500 presses within 60 min) animals were trained on a PR schedule following the same procedure as in the food operant condition but session time remained at 60 min. The water access time for the mice was reduced to 3 s in PR schedules. Five consecutive sessions of each schedule (PR1 and PR3) were conducted. The BP and final ratio were defined using the same criteria as for the food paradigm and the last three PR3 trials were averaged and used for statistical comparisons.
Negative reinforcement, shock avoidance
Mice were given a period of one week (120 min per day) to acclimate to the operant conditioning chambers. We then determined the shock threshold for each mouse by varying levels of shock from 0.06 to 0.4 mA (at 0.05 mA increments) using the modified titration method of Turner(Turner et al., 1967
). We performed first an ascending shock intensity series, then a descending series and finally a random sequence for each mouse. Each shock lasted for 1 s with an inter-stimulus interval of 15 s and a 30 s pause between the series. An observer blind to the magnitude of the stimulus quantified the reaction on a scale from 0 to 4, 0=no response, 1=flinch, 2=hop/move forward or backward, 3=run and 4=jump. An intensity was chosen for each animal that produced a scale response of 2.
Animals were first shaped to press the lever to avoid the shock. After shaping, animals were put under FR1 schedule with the following protocol: thirty seconds after the animal was placed in the cage, a green warning light and an 80 dB tone (2 KHz) came on for 10 s indicating the beginning of the session. Fifteen seconds later, a warning signal that consisted of an 80 dB, 4 KHz tone and the red cue light came on for 10 s. When the lever was pressed once before the end of the 10 s warning period, a 20 s break was given before another warning signal was delivered. Otherwise, a foot shock was given. When a shock occurred, 5 s thereafter another warning signal-shock cycle was presented. The session terminated if the animal failed to press the lever in 10 consecutive warning signal-shock cycles. The training sessions lasted 30 min. Training continued until there were 3 consecutive sessions in which the animal avoided at least 80% of the programmed shocks for the session. The number of presses required to postpone the shock was then increased to FR2, FR3 and finally FR5 and the session duration was increased to 60 min. Shock intensity was increased by 0.1 mA (to a maximum 0.4 mA) each time the animal failed to complete the avoidance ratio for 5 consecutive trials. Upon successful completion of the FR5 schedule, animals were transferred to the PR schedules and the session duration increased to 120 min. For PR schedules that required multiple lever presses to avoid being shocked, each press within the 10 s warning period postponed the shock by another 10 s, until the ratio was completed and then a 20 s break was provided before another warning signal-shock cycle began. Presses during the 20 s break would postpone the shock for another 10 s as in the warning period and count towards completing the ratio. A modified PR schedule was adopted(Grasing et al., 2003
) in which an exponential formula was used to generate a more gradual increase of bar press steps than the regular PR bar press schedule, in order to minimize shock administration while still requiring progressively higher rates of response. For example, in PR1–10 the first 10 shocks could be avoided by 1 lever press, the subsequent 9 required 2 presses, the following 8 required 3 presses and so on. The same principle was applied for PR2–10 and PR3–10. The session was terminated if the animal failed to avoid 10 consecutive shocks. Once the animal completed PR2–10 schedule, 5 trials were run under PR3–10 schedule and the BP/final ratio scores of the last three trials were averaged for statistical analysis.
Operant sessions for shock avoidance were run during both the light and dark phases. For the light phase sessions, animals were run starting 2 h after the onset of the light phase (between 9:00 AM and 1:00 PM). For the dark phase sessions, animals were run starting 2 h after the onset of the dark phase (between 9:00 PM and 1:00 AM).
Expected and unexpected food reward
Mice were given a week of acclimation (120 min per day) to the recording chambers with the house light on during the light phase, followed by daily exposure to 20 mg food pellets. Food rationing was then started to maintain 85–90% of their original body weight. Once the animals' weight had stabilized for 14 days, the experimental sessions began and animals were divided into two groups. In the expected reward group animals were placed in the conditioning chamber daily and food pellets were delivered at VI-4 min. Sessions were conducted daily for 4 weeks. In the unexpected reward group animals were placed in the conditioning chamber daily but pellets were delivered at VI-4 min only on the day they were sacrificed. For this group, the pellets were delivered to a food tray on the cage floor rather than the usual food trough so the animals could see the pellets when they were delivered. In both cases no bar press was required or performed to obtain the pellets. All experimental sessions had a duration of 120 minutes.
Unavoidable foot shock
Mice were given a period of one week (120 min per day) to acclimate to the operant chambers. We then determined the shock threshold for each mouse using the same procedure as described in the shock avoidance group. An intensity (range 0.25–0.4 mA) was chosen for each animal that produced a scale response of 2. Thereafter animals were placed in the chamber daily (120 min) daily for 4 weeks. Only on the day of the sacrifice, an unavoidable foot shock (1 s) was delivered at a fixed interval of 5 min (FI-5 min) during the 120 min session.
Double immunohistochemistry for the detection of Fos protein and Hcrt was performed on animals after 120 min of operant tasks, non-operant tasks or chamber stay. Eleven different experimental groups were used; eight during the light phase and three during the dark phase. For the food reinforcement task, a PR3 schedule was used.
For the light phase, the following conditions were employed: 1) PR food (), 2) Shock avoidance (), 3) Chamber control (), 4) Expected food. 5) Unexpected food, 6) Shock extinction animals trained to avoid the foot shock until they reached the PR3-10 schedule but with shock not delivered on the day of sacrifice (Shock extinction), 7) Unavoidable shock and 8) Subjective day; PR food in light phase without the house light on the day of sacrifice.
Fig 6 Distribution of Hcrt+ and cFos+/Hcrt+ neurons in the hypothalamus of WT mice under different behavioral conditions. Hcrt neurons express Fos during food motivated task in the light phase. Neither food nor shock avoidance tasks increase Fos expression (more ...)
For the dark phase: 1) PR food (), 2) Shock avoidance (), 3) Chamber control ().
Immediately after finishing the sessions, animals were deeply anesthetized with Nembutal (100 mg/kg, i.p.) and transcardially perfused with 0.03 l of heparinized (1000 units/l) phosphate buffered saline (PBS, 0.1M, pH 7.4) followed by 0.07 l of 4% paraformaldehyde in phosphate buffer (PB, 0.1M, pH 7.4). The brain was removed and immersed for 15 min in 4% paraformaldehyde in PB. After post-fixation, the tissue was transferred to 20% and then 30% sucrose solution for cryoprotection. Forty-eight hours later, the brain was frozen and cut into 35 μm sections using a cryostat (Reichert-Jung Cryocut 1800, West Germany). Each section was placed in one well of a 6-well tray containing PBS. Immunohistochemical procedures were performed immediately.
Detection of Fos was performed by sequential incubation of free-floating sections. The sections were first incubated in rabbit anti-Fos serum (PC-38, Calbiochem, La Jolla, CA, USA), 1:10000 in PBS, 0.3% triton X-100 (PBST). They were then exposed to biotinylated goat anti-rabbit IgG (Jackson Immunoresearch, West Grove, PA, USA) 1:600 in PBST, followed by incubation in standard ABC (Vector Laboratories, Burlingame, CA, USA) 1:400. The tissue was processed by the diaminobenzidine tetrahydrochloride (DAB) nickel-enhanced method, which consisted of tissue immersion in 0.6% nickel ammonium sulfate, 0.02% DAB and 0.03% hydrogen peroxide in 10 ml PBS, for 4–5 min. After Fos labeling, tissue was rinsed in PBST and incubated in rabbit anti-Hcrt B (H-003-32, Phoenix Pharmaceutical, Belmont, CA, USA), 1:10000, PBST. Subsequently, the tissue was immersed in biotinylated goat anti-rabbit IgG 1:600 in PBST, followed by incubation in standard ABC 1:400. The tissue was then processed with the DAB method for 4–5 min.
The number and distribution of Hcrt+ and Fos+/Hcrt+ cells were determined in every third section throughout the hypothalamus. A Nikon Eclipse 80i microscope with three axis motorized stage, video camera, Neurolucida interface and Stereoinvestigator software (MicroBrightField Corp.) was used. Cell counting was performed using the x60 objective by a trained histologist, always blind to the treatments. The criteria used in the present work to define the anatomical parcellation of the hypothalamus were based on the Franklin and Paxinos atlas of the mouse brain. The perifornical area (PFA) was defined as the region surrounding the fornix (140 μm from the perimeter of the structure). The rest of the hypothalamus was further divided into a medial subdivision (MH) that comprised the area from the medial limit of the fornix to the third ventricle (3V) and a lateral subdivision (LH) that extended from the medial boundary of the fornix to the lateral edge of the hypothalamus ().
Data were subjected to either Analysis of Variance (ANOVA) followed by Newman-Keuls post-hoc comparisons or t-test. All such tests were 2 tailed. Results were considered to be statistically significant if P<0.05. EEG power spectra were computed for consecutive 4-s epochs by Fast Fourier Transform (FFT, Hanning window, 0.25 Hz bands) with Spike2 software. For comparing power spectrum among behavior states the power within each state was divided by the total power of all states and presented as the percentage of total power. Frequencies between 0.5 and 20 Hz were analyzed. Video analysis to quantify behavior was performed by an observer blind to the genotype and the experimental condition.