Adenosine inhibits basal forebrain cholinergic and noncholinergic neurons in vitro

Neuroscience 140 (2006) 403– 413

ADENOSINE INHIBITS BASAL FOREBRAIN CHOLINERGIC AND NONCHOLINERGIC NEURONS IN VITRO E. ARRIGONI,a* N. L. CHAMBERLIN,a C. B. SAPERa,b AND R. W. MCCARLEYc

Key words: adenosine, sleep, basal forebrain, inwardly rectifying potassium current, hyperpolarization-activated cation current, in vitro slices.

a Department of Neurology, Beth Israel Deaconess Medical Center, Harvard Institutes of Medicine, Room 814, 77 Louis Pasteur Avenue, Boston, MA 02115, USA

Adenosine has been proposed as an endogenous homeostatic sleep-promoting factor that accumulates during waking (Radulovacki et al., 1984; Rainnie et al., 1994; Benington and Heller, 1995). I.c.v. administration of adenosine increases sleep time and electroencephalographic (EEG) slow wave activity similar to that observed at the onset of sleep (Radulovacki et al., 1984; Radulovacki, 1995; Strecker et al., 2000; Steriade and McCarley, 2005) and systemic administration of adenosine antagonists, such as caffeine and theophylline, suppresses sleep (Fredholm et al., 1999). With respect to the mediating brain region, previous studies have focused on several ventral forebrain nuclei including the magnocellular preoptic nucleus and substantia innominata (MCPO/SI) of the basal forebrain, and the adjacent ventrolateral preoptic nucleus (VLPO), where the majority of the neurons show state dependent activities (Szymusiak and McGinty, 1986; Sherin et al., 1996; Szymusiak et al., 1998; Lee et al., 2005). In the SI region adenosine levels rise during forced waking and decline during subsequent recovery sleep in both cats and rats (Porkka-Heiskanen et al., 1997; Basheer et al., 1999; Kalinchuk et al., 2003; Murillo-Rodriguez et al., 2004). In the same region local administration of adenosine or blockade of adenosine reuptake increases EEG slow wave activity and sleep (Porkka-Heiskanen et al., 1997; Portas et al., 1997; Basheer et al., 1999). Conversely, application of A1 receptor antagonists, or A1 receptor antisense RNA reduces the amount of slow wave sleep and increases wakefulness (Strecker et al., 2000; Thakkar et al., 2003b; Basheer et al., 2004). The MCPO/SI contains both cholinergic and noncholinergic (GABAergic and putative glutamatergic) neurons that project to the cortical mantle in a topographical organization in which the rostromedial region projects to the hippocampus and cingular cortex, whereas the caudolateral region projects to the lateral cortex and amygdala (Saper, 1984; Gritti et al., 1997; Manns et al., 2001). The cholinergic population which provides the major cholinergic input to the cerebral cortex (Woolf, 1991), projects to both pyramidal neurons and interneurons (Beaulieu and Somogyi, 1991). Acetylcholine in vitro produces a short inhibition followed by a long lasting increase in excitability of pyramidal neurons (McCormick and Prince, 1985). The initial hyperpolarization is due to a rapid excitation of interneurons that in turn inhibit pyramidal neurons, whereas the slow excitatory response is mediated by direct depolariza-

b

Program in Neuroscience, Harvard Medical School, Beth Israel Deaconess Medical Center, 330 Brookline Avenue, Boston, MA 02215, USA c

Department of Psychiatry, Harvard Medical School, Veterans Affairs Medical Center, 940 Belmont Street, Brockton, MA 02401, USA

Abstract—Adenosine has been proposed as a homeostatic “sleep factor” that promotes the transition from waking to sleep by affecting several sleep–wake regulatory systems. In the basal forebrain, adenosine accumulates during wakefulness and, when locally applied, suppresses neuronal activity and promotes sleep. However, the neuronal phenotype mediating these effects is unknown. We used whole-cell patchclamp recordings in in vitro rat brain slices to investigate the effect of adenosine on identified cholinergic and noncholinergic neurons of the magnocellular preoptic nucleus and substantia innominata. Adenosine (0.5–100 ␮M) reduced the magnocellular preoptic nucleus and substantia innominata cholinergic neuronal firing rate by activating an inwardly rectifying potassium current that reversed at ⴚ82 mV and was blocked by barium (100 ␮M). Application of the A1 receptor antagonist 8-cyclo-pentyl-theophylline (200 nM) blocked the effects of adenosine. Adenosine was also tested on two groups of electrophysiologically distinct noncholinergic magnocellular preoptic nucleus and substantia innominata neurons. In the first group adenosine, via activation of postsynaptic A1 receptors, reduced spontaneous firing via inhibition of the hyperpolarization-activated cation current. Blocking the H-current with ZD7288 (20 ␮M) abolished adenosine effects on these neurons. The second group was not affected by adenosine. These results demonstrate that, in the magnocellular preoptic nucleus and substantia innominata region of the basal forebrain, adenosine inhibits both cholinergic neurons and a subset of noncholinergic neurons. Both of these effects occur via postsynaptic A1 receptors, but are mediated downstream by two separate mechanisms. © 2006 IBRO. Published by Elsevier Ltd. All rights reserved. *Corresponding author. Tel: ⫹1-617-667-0834; fax: ⫹1-617-667-0810. E-mail address: [email protected] (E. Arrigoni). Abbreviations: ACSF, artificial cerebrospinal fluid; AHP, afterhyperpolarization; cAMP, cyclic AMP; ChAT, choline acetyltransferase; CPT, 8-cyclo-pentyl-theophylline; Cy3, indocarbocyanine; DMSO, dimethyl sulfoxide; EEG, electroencephalographic; IAD, adenosine-induced current; IC50, concentration to produce half-inhibition; G, chord conductance; Ih, hyperpolarization-activated cation current; IK(A), “A” type K⫹ current; IKir, inwardly rectifying potassium current; IR-DIC, infrared differential interference contrast; k, slope constant; LDT, laterodorsal tegmental; MCPO/SI, magnocellular preoptic nucleus and substantia innominata; PBS, phosphate-buffered saline; REM, rapid eye movement sleep; TTX, tetrodotoxin; VLPO, ventrolateral preoptic nucleus; Vm, membrane potential; Vsag, depolarizing sag potential; V1/2, halfactivated potential; 192IgG, monoclonal antibody against p75NTR neutrophin receptors; 4-AP, 4-aminopyridine.

0306-4522/06$30.00⫹0.00 © 2006 IBRO. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.neuroscience.2006.02.010

403

404

E. Arrigoni et al. / Neuroscience 140 (2006) 403– 413

tion (McCormick and Prince, 1986; Haj-Dahmane and Andrade, 1996). Cholinergic MCPO/SI neurons fire in association with cortical activation (Duque et al., 2000; Lee et al., 2005) during waking and rapid eye movement (REM) sleep and are virtually silent during slow wave sleep (Jones, 2004; Lee et al., 2005). Accordingly, cortical acetylcholine release is maximal during cortical activation of waking and REM sleep (Jasper and Tessier, 1971; Marrosu et al., 1995). GABAergic neurons in the basal forebrain may also be important governors of cortical activity. GABAergic neurons account for about one-third of the MCPO/SI cortically projecting neurons, and are co-distributed with the cholinergic population. They are similar in size and morphology to the cholinergic neurons (Gritti et al., 1997). Unlike the cholinergic neurons, their cortical target is predominantly GABAergic interneurons (Freund and Gulyas, 1991; Semba, 2000). In the MCPO/SI there are two physiologically distinct groups of GABAergic cortically projecting neurons: one is active during cortical arousal and may drive cortical activation during waking, and a second group that specifically expresses ␣2A-adrenergic receptors, discharges in association with cortical slow wave activity and may attenuate cortical activation during slow wave sleep (Manns et al., 2000; Modirrousta et al., 2004). In the basal forebrain the effects of adenosine at the cellular level have not been fully investigated. Extracellular recordings from unidentified neurons of the horizontal limb of the diagonal band in vitro show that adenosine suppresses neuronal activity (Rainnie et al., 1994) and in vivo studies show that adenosine inhibits the discharge of MCPO/SI wake-active neurons (Alam et al., 1999; Thakkar et al., 2003a). In the current study we examined the effects of adenosine on cholinergic and noncholinergic MCPO/SI neurons in in vitro slices. We identified cholinergic neurons by prelabeling in vivo with a non-toxic fluorescent dye and by post hoc immunohistochemistry.

EXPERIMENTAL PROCEDURES Animals Forty-seven Sprague–Dawley rats (Harlan, Indianapolis, IN, USA) of either sex (10 –19 days old) were used. Litters with their mothers were housed in a pathogen-free barrier facility maintained at 21.5– 22.5 °C with food and water available ad libitum. The Institutional Animal Care and Use Committee and the Committee on Microbiologic Safety of Harvard Medical School approved all the procedures. All efforts were made to minimize the number of animals and their suffering.

Prelabeling of cholinergic neurons Twenty-three rats (10 –14 days old) were anesthetized with choral hydrate (350 mg/kg) and injected i.c.v. with indocarbocyanine (Cy3) dye-coupled to 192IgG, a monoclonal antibody raised against p75NTR neutrophin receptors (Advanced Targeting Systems, San Diego, CA, USA). The p75 neurotrophin receptors are found only on cholinergic cells in the MCPO/SI and thus they are selectively labeled by Cy3-192IgG (Hartig et al., 1998). In vitro slices obtained from these animals can be used for electrophysiological recording (Wu et al., 2000). Cy3-192IgG was pressureinjected (2 ␮l, 0.4 mg/ml; delivery rate of 0.1 ␮l/min; Picospritzer II, General Valve, Fairfield, NJ, USA) into the left lateral ventricle

(AP: ⫺0.1 mm from bregma; RL: 1 mm; DV: 2 mm), from a silane-coated glass pipette (25–30 ␮m tip diameter). Two to five days later, the injected rats were used to prepare brain slices for in vitro recording.

In vitro brain slice preparation Twenty-three Cy3-192IgG-injected and 24 non-injected rats were anesthetized (isoflurane inhalation) to the point of respiratory arrest and then decapitated. Coronal brain sections (400 ␮m thickness) were cut with a vibrating microtome (VT1000 Leica, Bannockburn, IL, USA) in ice cold artificial cerebrospinal fluid (ACSF) oxygenated with 95% O2–5% CO2 and containing: (in mM) NaCl 124, KCl 2, MgSO4 1, CaCl2 2.5, NaHCO3 26, NaH2PO4 3, glucose 10 (pH 7.4). Slices were kept in a holding chamber at room temperature in oxygenated ACSF, and then transferred to the recording chamber where they were maintained fully submerged and perfused (2 ml/min) with oxygenated ACSF maintained at 32 °C using a temperature controller (TC-344B Warner Instruments, Hamden, CT, USA).

Fluorescence and infrared imaging The MCPO/SI region was identified at low magnification using tissue landmarks (anterior commissure, olfactory tubercle and optic chiasm). Whole-cell recordings were guided by combined fluorescence and infrared differential interference contrast (IRDIC) video microscopy using a fixed stage upright compound microscope (Axioscope 2-FS Carl Zeiss, Inc., Germany) equipped with a Nomarski water immersion lens (40⫻0.8 NA, working distance 3.6 mm). Fluorescence was visualized using the appropriate filters for Cy3 and Lucifer Yellow. IR-DIC and fluorescence images were detected with an IR-sensitive CCD camera (300-T-RC DAGE-MTI, Michigan City, IN, USA) and displayed on a computer screen in real time using frame grabber software (Beta 4.02 Scion Image, Frederick, MD, USA). Images (Fig. 1) were stored on hard disk and processed for brightness and contrast with Adobe Photoshop. When Cy3-192IgG-labeled neurons were targeted for recording, the glass electrode was visually advanced through the slice and the approach to the labeled neurons was monitored by switching between IR-DIC and fluorescence optical systems. Position and orientation were used as guidelines to recognize the designated neurons when switching between the fluorescence and the IR-DIC optics. Labeled neurons were found both ipsi- and contralateral to the injection side (Hartig et al., 1998) and recordings were made from both sides.

Whole-cell recordings Patch electrodes were prepared from borosilicate glass tubing (O.D. 1.5 mm; I.D. 0.86 mm, with filament) with a P97 pipette puller (Sutter Instrument, Novato, CA, USA). Electrodes were filled with (in mM): 120 K-gluconate, 10 KCl, 3 MgCl2, 10 HEPES, 2 K-ATP, 0.2 Na-GTP (pH 7.2 adjusted with KOH; 280 mOsm). Lucifer Yellow CH ammonium salt (0.1%) was added to the pipette solution to label the recorded cells. Recordings in current- and voltage-clamp mode were made using a Multiclamp 700A amplifier (Axon Instruments, Foster City, CA, USA). Signals were lowpass filtered at 1 kHz and digitized at 5–10 kHz with a Digidata 1322A interface and Clampex 8.0 software (Axon Instruments). Electrode resistance was 7–10 M ⍀. Series resistance was monitored at regular intervals with current or voltage pulses (⫺5 to ⫺10 nA, 50 –100 ms, or ⫺5 to ⫺10 mV, 20 – 40 ms). Data were discarded if neurons showed an unstable resting membrane potential (Vm) or if the series resistance changed by more than 25%. 8-Cyclo-pentyl-theophylline (CPT) stock solution was prepared in dimethyl sulfoxide (DMSO; ACSF final concentration of DMSO ⬍0.1%). Tetrodotoxin (TTX) was purchased from Alomone Laboratories (Jerusalem, Israel) and ZD7288 from Tocris (Ellisville,

E. Arrigoni et al. / Neuroscience 140 (2006) 403– 413

405

Fig. 1. Identification and electrophysiological properties of MCPO/SI cholinergic neurons. (A) Fluorescence photomicrograph of a Cy3-192IgG-labeled neuron visualized in a living brain slice before the recording starts. (B) IR-DIC image of the same neuron and the recording electrode (out of focus except for its tip). (C) By the end of the recording, the neuron is filled with Lucifer Yellow that has diffused in from the recording electrode. (D) Specific internalization of Cy3-192IgG by cholinergic neurons: Cy3-192IgG in vivo labeling (red), ChAT immunoreactivity using Alexa 488 (green). Scale bars⫽25 ␮m. (E, F) Whole-cell recordings from the same neuron (shown in panels A–C) in response to depolarizing (E) and hyperpolarizing (F) pulses. Note the delayed rebound discharge after hyperpolarizing steps due to IK(A) activation (F, arrow) and the pronounced AHP (F, double arrow). The responses to injection of negative current pulses exhibit a voltage-dependent rectification due to the activation of IKir, and absence of Vsag, indicating no Ih activation (asterisk). (G) Effect of 4-AP (5 mM) on the firing pattern. Application of 4-AP (thin trace) eliminates the delay of the rebound discharge (arrows) after a hyperpolarizing step (2.5 s).

MO, USA); all other reagents were purchased from Sigma-Aldrich (St. Louis, MO, USA).

Data analysis Data were analyzed using Clampfit 8.0 (Axon Instruments), IGOR Pro 4.0 (WaveMetrics, Lake Oswego, OR, USA) and Mini Analysis 6 (Synaptosoft, Leonia, NJ, USA) software. Depolarizing sags (Vsag) during 2-s hyperpolarizing current steps were calculated as: Vsag⫽Vm(start)⫺Vm(end). Where Vm(start) is the peak value reached by the Vm before the Vsag begins and Vm(end) is the Vm at the end of the 2-s hyperpolarizing current steps. Vsag were calculated for hyperpolarizing pulses that produced a Vm(start) ranging between ⫺95 and ⫺100 mV. The adenosine dose-response relationship was fitted using a sigmoid equation Y⫽min(inh)⫹{max(inh)/ [1⫹exp((E50⫺AD)/k)]}, where Y is the % of the firing rate inhibition, AD is the adenosine concentration (␮M), IC50 is the adenosine concentration that produced half-maximal inhibition and k is the slope constant. Fitting constraints: min(inh)⫽0% and max(inh)ⱕ100%. The standard definition of chord conductance (G) was used: G⫽I/(Vm⫺Vrev), where I is the current amplitude at the Vm and Vrev is the zero-current (reversal) potential. G was calculated on averaged successive groups of 400 samples. The plot of the adenosine-mediated G GAD vs. Vm was fit using the Bolzmann equation (Rainnie et al., 1994), Y⫽G(min)⫹{(G(max)⫺G(min))/{1⫹exp [(Vm⫺V1/2)/k]}}, where Y is the fractional voltage sensitive change in conductance; V1/2 is the half-activated potential. Data are presented as means⫾S.E.M. and statistical significance was established by one-way ANOVA or t-tests (two-tailed) with P values as reported.

Post hoc immunolabeling Following recording, slices were processed to determine the cholinergic phenotype of the neurons that had been filled with Lucifer

Yellow. The 400 ␮m slices were not resectioned as recorded cells lay near the slice surface within the range of antibody penetration. Slices were fixed in formalin and then incubated overnight in phosphate-buffered saline (PBS) containing 0.3% Triton X-100 and goat anti-choline acetyltransferase (ChAT) primary antibody (1:100; Chemicon International, Temecula, CA, USA). The next day the slices were incubated (1 h) in donkey Cy3-anti-goat IgG secondary antibody (1:500 in PBS; Jackson, West Grove, PA, USA), rinsed with PBS, and wet-mounted on glass microscope slides. Each recorded neuron labeled with Lucifer Yellow was examined with an epifluorescence microscope (Axioplan2, Carl Zeiss, Oberkochen, Germany), for localization with respect to the MCPO/SI area and for determination of double-labeling. Recorded ChAT-immunoreactive neurons contained both Lucifer Yellow and Cy3 immunofluorescence. Although both the 192IgG and ChAT secondary antibodies were tagged with Cy3, the two markers had distinct patterns of labeling: ChAT immunofluorescence was homogeneously distributed whereas internalized Cy3-192IgG had a granular or punctate appearance due to clustering in endosomes throughout the perikarya (Fig. 1A). Only neurons with homogeneous ChAT staining in the cell body were considered ChATimmunoreactive. Four unrecorded slices from two Cy3-192IgGinjected animals were processed using donkey Alexa Fluor 488anti-goat IgG secondary antibody (1:1000 in PBS; Molecular Probes, Eugene, OR, USA; Fig. 1D).

RESULTS Prelabeling of MCPO/SI cholinergic neurons We prelabeled in vivo the MCPO/SI cholinergic population using the fluorescent marker Cy3-192IgG (Hartig et al., 1998). Neurons that had internalized the Cy3-192IgG had fluorescent antibodies clustered in endosomes throughout

406

E. Arrigoni et al. / Neuroscience 140 (2006) 403– 413

Table 1. Electrophysiological properties of 192IgG-positive/ChAT-positive neurons and ChAT-positive neurons from uninjected animals are identical Property

192IgG/ChAT (⫹) neurons (n⫽45)

ChAT (⫹) neurons (n⫽14)

t-Test

Resting potential (mV) Input resistance (M⍀) Presence of spontaneous firing; rate (Hz) Spike half-width (ms) AHP amplitude (mV) AHP duration (ms) Presence of accommodation Depolarizing sag (mV) (Ih activation) Percentage of neurons inhibited by AD

⫺55.3⫾1.2 (n⫽45) 343⫾22 (n⫽45) 15% (n⫽40); 1.4⫾0.3 (n⫽6) 2.95⫾0.18 (n⫽45) ⫺12.9⫾0.90 (n⫽15) 360⫾32 (n⫽15) 100% (n⫽14) ⫹1.12⫾0.34 mV (n⫽30) 83% (n⫽35)

⫺55.2⫾1.6 (n⫽14) 381⫾26 (n⫽14) 15% (n⫽13); 1.6⫾0.1 (n⫽2) 2.91⫾0.33 (n⫽14) ⫺11.2⫾2.02 (n⫽8) 315⫾64 (n⫽8) 100% (n⫽6) ⫹0.46⫾0.46 mV (n⫽10) 83% (n⫽6)

P⫽0.959 P⫽0.265 P⫽0.517 P⫽0.907 P⫽0.443 P⫽0.546 P⫽0.260

Values are presented as means⫾SEM (n, number of cells). Significant differences were determined by unpaired t-test. Physiological parameters were calculated as described in the Experimental Procedures section.

the perikarya and we were able to selectively target them for electrophysiological recordings (Fig. 1A and B). Consistent with previous studies (Hartig et al., 1998; Wu et al., 2000), most but not all of the MCPO/SI cholinergic neurons were labeled with Cy3-192-IgG following i.c.v. injection. Post hoc double-labeling showed that 45 out of 47 Cy3192IgG-labeled-recorded neurons in the MCPO/SI were ChAT-immunoreactive. Data from the two 192IgG-labeled, but ChAT negative cells were excluded from the study. We did not observe any sign of phototoxicity in Cy3192IgG-labeled neurons (Wu et al., 2000) and the action potentials had normal overshoot (⫹25.3⫾1.3 mV; n⫽45) (Fig. 1). Accordingly, there were no significant differences in the electrophysiological properties between Cy3192IgG-labeled neurons (n⫽45) and cholinergic neurons from uninjected animals identified by post hoc ChAT immunohistochemistry (n⫽14; Table 1). Thus results from 192IgG-labeled and non-prelabeled cholinergic neurons recorded from uninjected animals were grouped together for analysis (n⫽59; Table 2). Firing properties of MCPO/SI cholinergic neurons Cholinergic neurons in the MCPO/SI displayed electrophysiological characteristics similar to those previously reported in in vitro studies (Table 2; Fig. 1). 1) The MCPO/SI cholinergic population exhibited a robust outward transient

rectification by an “A” type K⫹ current (IK(A)) (Khateb et al., 1995) that delayed the rebound firing following hyperpolarization. Application of 4-aminopyridine (4-AP; 5 mM) which blocks IK(A) (Rudy, 1988) eliminated the delay in the rebound firing after hyperpolarizing pulses (n⫽4; Fig. 1G). In addition, spikes were broadened (5.63⫾0.42 ms spike half-width, compared with 3.33⫾0.38 ms in control; n⫽4; P⫽0.016 paired t-test), indicating that IK(A) contributes to the spike repolarization of these neurons. 2) The MCPO/SI cholinergic neurons showed a relatively broad spike under control conditions (spike half-width: 2.94⫾0.16 ms; n⫽59) (Morris et al., 1999) probably due to activation of Ca2⫹ conductances. 3) They showed a pronounced and relatively long afterhyperpolarization (AHP; amplitude: ⫺12.3⫾0.9 mV; duration: 345⫾30 ms; n⫽23) due to the activation of a Ca2⫹-activated K⫹ conductance that limited their firing frequency (Khateb et al., 1995). 4) They displayed a rapid time- and voltage-dependent rectification in response to injection of negative current pulses compatible with the presence of an inwardly rectifying potassium (IKir) conductance (Wu et al., 2000). 5) They were mostly silent at the resting Vm (85%; n⫽53) and showed spike frequency accommodation in all cases (n⫽24). 6) They did not display an hyperpolarization-activated cation current (Ih)-mediated Vsag upon hyperpolarization (Table 2; Fig. 1F) (Alonso et al., 1996; Wu et al.,

Table 2. Electrophysiological properties of cholinergic and noncholinergic MCPO/SI neurons Property

Resting potential (mV) Input resistance (M⍀) Presence of spontaneous firing; rate (Hz) Spike half-width (ms) AHP amplitude (mV) AHP duration (ms) Presence of accommodation Depolarizing sag (mV) (Ih activation) Percentage of neurons inhibited by AD

ChAT (⫹) neurons (n⫽59)

⫺55.3⫾1.0 (n⫽59) 352⫾18 (n⫽59) 15% (n⫽53); 1.5⫾0.2 (n⫽8) 2.94⫾0.16 (n⫽59) ⫺12.34⫾0.90 (n⫽23)† 345⫾30 (n⫽23) 100% (n⫽24) ⫹0.96⫾0.28 mV† (n⫽40) 83% (n⫽41)

ChAT (⫺) neurons Group I (n⫽25)

Group II (n⫽14)

⫺47.4⫾2.1 (n⫽14) 369⫾29 (n⫽25) 79% (n⫽19); 5.00⫾0.70 (n⫽16)† 1.99⫾0.31 (n⫽22)† ⫺6.30⫾1.10 (n⫽8) 190⫾46 (n⫽8)† 2 of 11 neurons ⫹11.1⫾0.7 mV†** (n⫽25) 83% (n⫽12)

⫺50.9⫾1.7 (n⫽13) 735⫾123 (n⫽14)†† 50% (n⫽14); 1.7⫾0.4 (n⫽7) 3.17⫾0.40 (n⫽13) ⫺7.30⫾0.83 (n⫽13) 313⫾32 (n⫽13) 4 of 6 neurons ⫹3.6⫾0.4 mV† (n⫽14) 1 of 9 neurons

ChAT (⫹) neurons include both 192IgG-positive/ChAT-positive neurons and ChAT-positive neurons from uninjected animals. Values are presented as means⫾SEM (n, number of cells). Significant differences were determined by one-way ANOVA tests. Cross symbols denote statistically significant differences across all three groups: † for P⬍0.05; †† for P⬍0.01. Asterisks denote statistically significant differences between I group and II group of ChAT (⫺) neurons: * for P⬍0.05; ** for P⬍0.01.

E. Arrigoni et al. / Neuroscience 140 (2006) 403– 413

2000). 7) Finally, we found that low threshold Ca2⫹ spikes were only present in slices from rats over 16-days-old. This finding is in accord with previous studies showing the presence of low threshold Ca2⫹ spikes in MCPO/SI cholinergic neurons in mature (Khateb et al., 1995, 1997; Alonso et al., 1996; Lee et al., 2005), but not immature rodents (Wu et al., 2000; Eggermann et al., 2001). Adenosine inhibited MCPO/SI cholinergic neurons via postsynaptic A1 receptors Adenosine (0.5–100 ␮M) was tested on 41 cholinergic neurons. Adenosine inhibited 83% of these neurons and had no effect on the remaining 17%. Excitatory responses were never observed. Adenosine (50 –100 ␮M) hyperpolarized MCPO/SI cholinergic neurons (n⫽10; Fig. 2C). This effect was blocked by CPT (n⫽3), but not by TTX (1 ␮M; n⫽3), consistent with mediation by postsynaptic adenosine A1 receptors. When current was injected to initially hold the cells at ⫺60 mV, application of adenosine hyperpolarized the membrane to ⫺67.4⫾1.1 mV (n⫽11) and reduced the input resistance (372.3⫾30.0 M ⍀ in control; 277.4⫾16.0 M ⍀ with adenosine; n⫽7; P⫽0.007 paired t-test). Adenosine reduced the discharge rate of cholinergic MCPO/SI neurons (n⫽22) an effect that was blocked by bath application of the A1 receptor antagonist CPT (200 nM; n⫽3; Fig. 2A). The dose-response relationship for adenosine reduction of firing rate was determined by setting firing rate initially at 2 Hz by constant current injection (Fig. 2B; IC50⫽8.8⫾1.3 ␮M; n⫽15). The minimum concentration of adenosine required to produce a significant inhibition of neuronal firing rate was 0.5 ␮M, which reduced the firing rate from 1.96⫾0.1 Hz, in control to 1.7⫾0.15 Hz (n⫽5; P⫽0.022, paired t-test).

407

Adenosine-induced currents (IAD) Voltage ramps from ⫺100 to ⫺30 mV (5–10 mV/s) were recorded before and during adenosine application in the presence of TTX (1 ␮M; Fig. 3). Adenosine increased the current at all the Vm through the ramp protocol (n⫽7) and the effect was blocked by bath application of CPT (200 nM; n⫽3). IAD was determined by digital subtraction of the currents recorded under control conditions from those recorded during adenosine applications. Current-voltage relationships for IAD showed that IAD had a reversal potential of ⫺81.8⫾2.1 mV (n⫽7) and rectification properties compatible with an IKir conductance (Fig. 3C, D). Barium (100 ␮M), which blocks IKir conductance (Gerber et al., 1989), reduced the current at all tested Vm (n⫽6) and abolished the effects of adenosine (n⫽6; Fig. 3B, C). Adenosine effects on noncholinergic neurons We made recordings from 39 ChAT-negative neurons in the MCPO/SI. Most of these (n⫽33) were recorded in slices from animals not injected with Cy3-192IgG. The remaining six were unlabeled neurons recorded in slices from Cy3-192IgG-injected rats. All 39 were ChAT-negative by post hoc immunohistochemistry and none had electrophysiological characteristics resembling those of cholinergic neurons. On the basis of their electrophysiological properties we identified two groups of the ChAT-negative MCPO/SI neurons (Table 2). Group I (n⫽25) was characterized by a profound Vsag upon hyperpolarization that was abolished by ZD7288 (20 ␮M; n⫽3; Fig. 4), a specific blocker of the H-current (Harris and Constanti, 1995). The Ih-mediated Vsag ranged between ⫹7.0 and ⫹17.2 mV at the end of 2-s hyperpolarizing pulses (see Experimental Procedures) with a mean value of ⫹11.1⫾0.7 mV (n⫽25). This group of neurons

Fig. 2. Adenosine inhibits MCPO/SI cholinergic neurons. (A) Firing is produced by injection of a depolarizing holding current (25 pA), bath-applied adenosine (100 ␮M) reduces firing rate and the effect is reversed by CPT (200 nM). (B) Dose-response curve for depression of neuronal firing rate by adenosine (n⫽4 – 6 trials per dose). The firing rate was initially set at 2 Hz by injection of a depolarizing holding current (⫹50.3⫾13.3 pA; n⫽15 neurons). Data are fit (line) using a sigmoid equation, max inhibition⫽100⫾13.7%; IC50⫽8.8⫾1.3 ␮M; and k⫽2.5⫾0.3 nM. (C) Adenosine induces hyperpolarization associated with decreases in membrane input resistance. Hyperpolarizing current pulses (500 ms duration; 80 pA amplitude) are used to monitor membrane input resistance changes. Asterisks indicate when the potential is manually clamped back to ⫺60 mV.

408

E. Arrigoni et al. / Neuroscience 140 (2006) 403– 413

Fig. 3. Adenosine (AD) inhibits the MCPO/SI cholinergic neurons by activating an IKir. (A) Voltage ramps from ⫺100 to ⫺30 mV (10 mV/s) recorded in the presence of TTX (1 ␮M) before (CON) and during application of AD (50 ␮M). (B) BaCl2 (Ba, 100 ␮M) completely blocked the effects of AD. Note that Ba alone and Ba plus AD traces overlap completely. (C) Current–voltage relationships of AD-evoked currents (IAD), barium-blocked current (IBa-blocked) and AD-evoked currents recorded in barium (IAD(in Ba)) calculated by digital subtractions: IAD⫽AD⫺CON of the current showed in A; IBa-blocked⫽CON⫺Ba of the currents displayed in B and IAD(in Ba)⫽AD(in BA)⫺Ba of the currents displayed in B. (D) Plot of AD G (GAD) as a function of Vm is fit to the Bolzmann equation (line) GAD(max)⫽1.82 nS; GAD(min)⫽0.05 nS; V1/2⫽⫺62.4 mV; and k⫽25.6 mV. Current traces obtained from voltage ramp protocols are the averages of three consecutive traces.

had more depolarized resting Vm than cholinergic neurons (⫺47 mV; Table 2). In addition, a larger percentage of them showed spontaneous firing (79%; 5.00⫾0.70 Hz, n⫽16) which was maintained even when they were held at the same resting Vm as the cholinergic population (⫺55 mV; n⫽4, ranging between 0.44 –1.44 Hz). Only two of 11 neurons showed spike frequency accommodation. They had narrow action potentials (spike half-width: 1.99⫾0.31 ms;

n⫽22), and relatively short AHPs (190⫾46 ms; n⫽8). In addition, these neurons rapidly discharged following hyperpolarizing pulses when the membrane repolarized (Fig. 4). We tested the effects of adenosine on twelve group I neurons. Adenosine (50 –100 ␮M) inhibited 10 of them, two neurons did not respond to adenosine, and excitatory responses were never observed. At the resting Vm adenosine inhibited their spontaneous firing (n⫽6), and the

Fig. 4. Electrophysiological properties of MCPO/SI noncholinergic neurons group I. (A) Typical firing pattern of group I ChAT negative MCPO/SI neurons. At resting Vm the neuron spontaneously fires. Responses to negative current pulses exhibit pronounced Vsag, produced by Ih activation (asterisk) and rapid discharges at the end of the hyperpolarizing steps (arrow). (B) ZD7288 (20 ␮M), a specific blocker of the H-current abolished the Vsag.

E. Arrigoni et al. / Neuroscience 140 (2006) 403– 413

409

Fig. 5. Adenosine (AD) inhibits group I MCPO/SI noncholinergic neurons by reducing Ih. (A) AD (100 ␮M) inhibits the spontaneous firing of a group I MCPO/SI neuron, and the effect is blocked by bath application of CPT (200 nM). (B) Voltage-clamp recordings (Vh⫽⫺50 mV) of a different neuron shows that AD (100 ␮M) reduces the Ih evoked by hyperpolarizing voltage steps (5 s; to ⫺70, ⫺85, ⫺100 mV). Iis, point where the instantaneous current is measured; Iss, point where the steady-state current was measured. (C) Current–voltage relationship for the experiment in B. Difference Iss⫺Iis used as measurement of the Ih is plotted against Vm during hyperpolarizing steps (⫺5 mV; from ⫺50 to ⫺105 mV). (D) Voltage-clamp recordings (Vh⫽⫺60 mV; pulses to ⫺70, ⫺80, ⫺90 mV) in control (left) and in the presence of ZD7288 (right). ZD7288 (20 ␮M) blocks the H-current and abolishes the effects of AD (100 ␮M). (E) ZD7288 blocks the effects of AD throughout the voltage range of the test ramp (⫺100 to ⫺40 mV; 10 mV/s; recorded in the presence of TTX 1 ␮M). Current traces are averages of three consecutive trials.

effect was blocked by application of the A1 receptor antagonist CPT (200 nM; n⫽5; Fig. 5A). Adenosine induced membrane hyperpolarization (n⫽5). In the presence of TTX (1 ␮M) and at ⫺60 mV using DC current injections (10.3⫾2.6 pA; n⫽3), application of adenosine hyperpolarized the membrane to ⫺64.5⫾1.5 mV (n⫽3). Thus, adenosine inhibits this subset of MCPO/SI noncholinergic neurons by activating postsynaptic A1 receptors. Recording in voltage-clamp mode we tested the effects of adenosine on the H-current. Hyperpolarizing voltage steps from ⫺50 mV holding potentials activated the slow onset, non-inactivating inward current Ih. Adenosine reduced Ih in three of four neurons tested (Fig. 5B and C). Application of ZD7288 (20 ␮M) blocked the H-current (n⫽3) and abolished the effects of adenosine (100 ␮M; n⫽6; Fig. 5D and E). Using voltage ramp protocols (5 mV/s) we tested the effects of adenosine on the residual current after blocking the Ih. No effects of adenosine were observed in the presence

of ZD7288 at potentials ranging from ⫺100 to ⫺40 mV; n⫽4). Group II MCPO/SI ChAT-negative neurons (n⫽14) showed small Vsag (average: ⫹3.6⫾0.4 mV; n⫽14; range: ⫹0.7 to ⫹5.7 mV) during 2-second hyperpolarizing current steps (Fig. 6A). This group of neurons had relatively high input resistance (735⫾123 M ⍀, n⫽14, measured at resting Vm). They showed spontaneous firing only in 50% of the cases (n⫽14), they had relatively broad spikes (spike half-width: 3.17⫾0.40 ms; n⫽13) and long AHPs (313⫾32 ms; n⫽13). In contrast to the cholinergic population they did not show outward transient rectification by IK(A) activation and they rapidly discharged following hyperpolarizing current pulses when the membrane returned to the resting potential (Fig. 6A). These neurons did not respond to adenosine. Adenosine (50 –200 ␮M) inhibited firing in only one of nine neurons, and the others did not respond to adenosine (Fig. 6B).

410

E. Arrigoni et al. / Neuroscience 140 (2006) 403– 413

Fig. 6. Adenosine does not affect group II ChAT negative MCPO/SI neurons. (A) Typical firing pattern of a group II ChAT negative MCPO/SI neuron. Current-clamp recordings at resting Vm in response to hyperpolarizing pulses. Note no Vsag, indicating the absence of Ih (asterisk), and no delayed in the rebound discharge at the end of the hyperpolarizing steps (arrow). (B) In the same neuron application of adenosine (100 ␮M) does not affect the neuronal firing produced by injection of a depolarizing holding current (60 pA).

DISCUSSION In the magnocellular MCPO/SI area of the basal forebrain, bath-applied adenosine consistently inhibited cholinergic neurons but only a subset of noncholinergic neurons. In both cases inhibitory effects were postsynaptic and were mediated by the activation of adenosine A1 receptors. Adenosine inhibited the cholinergic population by activating an IKir current whereas the affected noncholinergic neurons were inhibited by a reduction of the H-current. Adenosine inhibition of the cholinergic neurons Previous studies in the MCPO/SI region have shown that adenosine via A1 receptors inhibits wake-active neuron discharge (Alam et al., 1999; Thakkar et al., 2003a). At least a portion of wake-active neurons are cholinergic (Lee et al., 2005). Consistent with these studies we found that adenosine directly inhibits immunohistochemically identified cholinergic neurons in the MCPO/SI regions. This effect is mediated by adenosine A1 receptors, is dose dependent with an IC50 of 8.8 ␮M, and the threshold concentration of adenosine is 500 nM. Similarly, dose response studies in vitro in other brain regions found adenosine effects mediated by A1 receptors to have an IC50 ⬃1–10 ␮M (Dunwiddie and Hoffer, 1980; Oliet and Poulain, 1999). Furthermore, the adenosine concentrations that we found to inhibit basal forebrain cholinergic neurons are in the range of the physiological levels measured in vivo. Indeed, baseline extracellular adenosine levels estimated from the microdialysate samples (probe recovery of 10 –20%) in rats in the basal forebrain ranged from 160 to 750 nM (Porkka-Heiskanen et al., 2000; Strecker et al., 2000; Murillo-Rodriguez et al., 2004) which increased by 50 –200% after sleep deprivation (Porkka-Heiskanen et al., 2000; Kalinchuk et al., 2003). We also found that similar to the inhibitory actions of adenosine observed in other areas of the CNS (Pape, 1992; Rainnie et al., 1994) adenosine inhibits the MCPO/SI cholinergic neurons by activating the G-protein regulated IKir. In addition to adenosine A1 receptors the G-proteinregulated IKir channels are activated by a variety of Gi/Gocoupled inhibitory neurotransmitter receptors such as muscarinic (M2), serotoninergic (5-HT1A), GABAB, somatostatin, and opioid (mu, delta, kappa) (Isomoto et al., 1997) and within the same neuron several of these neurotransmitters

can activate the same IKir (Luscher et al., 1997). Previous in vitro studies have shown that MCPO/SI cholinergic neurons are inhibited by acetylcholine via muscarinic receptors (Khateb et al., 1997) and by serotonin via 5-HT1A receptors (Khateb et al., 1993), and possibly by all converging to the same IKir activated by adenosine. Functionally, G-protein-regulated IKir channels maintain the resting Vm and mediate the receptor-dependent inhibition of cellular excitability (Isomoto et al., 1997). Activation of IKir inhibits neuronal firing primarily by membrane hyperpolarization, but the conductance increase alone can shunt synaptic inputs. In the hippocampus for example, the activation of dendritic G-protein-regulated IKir reduces the size of the synaptic potential reaching the soma (Seeger and Alzheimer, 2001). Thus increases in extracellular adenosine levels could produce both reduction of spontaneous neuronal firing and reduction in the response to incoming synaptic inputs in MCPO/SI cholinergic neurons. Adenosine effects on noncholinergic neurons We found that a portion of noncholinergic neurons in the MCPO/SI was inhibited by adenosine. Neurons in group I were inhibited by adenosine via the reduction of the Hcurrent. Adenosine-mediated-activation of IKir was not observed in these neurons. Adenosine-mediated inhibition of Ih has been previously described in thalamic neurons and laterodorsal tegmental (LDT) cholinergic neurons where adenosine also induces activation of IKir (Pape, 1992; Rainnie et al., 1994). Thus, in thalamic and LDT cholinergic neurons, activation of IKir and inhibition of Ih are combined, whereas in the basal forebrain the two effects segregate in two different neuronal populations. The H-current is involved in setting the resting Vm: increasing Ih causes membrane depolarization while decreasing Ih causes membrane hyperpolarization (Robinson and Siegelbaum, 2003). In GABAergic neurons of the medial septum, consequences of blocking Ih include firing inhibition and membrane hyperpolarization (Xu et al., 2004). Ih gating is directly regulated by intracellular cyclic AMP (cAMP), and neurotransmitters that reduce cAMP levels reduce Ih activation (Robinson and Siegelbaum, 2003). Adenosine, via A1 receptors and G-proteins (Gi), inhibits cAMP production (Brundege and Dunwiddie,

E. Arrigoni et al. / Neuroscience 140 (2006) 403– 413

1997), and, in the thalamus, adenosine inhibits Ih through the cAMP pathway (Pape, 1992). Thus, analogous to the effect on thalamic neurons, adenosine in the MCPO/SI noncholinergic neurons may mediate Ih inhibition through cAMP signaling. Group I MCPO/SI ChAT-negative neurons have electrophysiological properties similar to GABAergic neurons in the medial septum and diagonal band of Broca that project to the hippocampus and are characterized by the presence of an H-current, spontaneous firing, no potassium type-A current and no firing accommodation (Morris et al., 1999; Wu et al., 2000). If Group I MCPO/SI ChATnegative neurons are GABAergic cortically projecting neurons, they may correspond to the GABAergic cortically projecting neurons that fire during cortical activation (Manns et al., 2000; Modirrousta et al., 2004) and may represent a portion of the wake-active neurons that in unanesthetized animals were inhibited by adenosine (Alam et al., 1999; Thakkar et al., 2003a). The neurons that we termed group II based on electrophysiological properties and absence of ChAT-immunoreactivity did not respond to adenosine. These neurons may correspond to the adenosine-non-responding group found in the MCPO/SI regions in in vivo single unit recording studies (Alam et al., 1999; Thakkar et al., 2003a). These neurons represent a minority and showed a sleep active profile (Thakkar et al., 2003a). These neurons may be glutamatergic (Manns et al., 2001) and/or neuropeptide-containing neurons (Zaborszky et al., 1991; Duque et al., 2000). However, glutamatergic neurons and neuropeptide-containing neurons are active during cortical activation (Duque et al., 2000; Manns et al., 2000; Modirrousta et al., 2004). Thus, our hypothesis is that group II ChAT-negative neurons are more likely to correspond to the slow-wave– sleep-active neurons recorded in unanesthetized animals (Szymusiak et al., 2000) and/or the subset of GABAergic neurons that in anesthetized animals discharge in association with cortical slow wave activity and bear ␣2A-adrenergic receptors (Osaka and Matsumura, 1995; Manns et al., 2000; Modirrousta et al., 2004). Only a portion of these sleep active neurons are cortically projecting neurons (Szymusiak and McGinty, 1989; Manns et al., 2000); those that are not antidromically activated by cortical stimulation project caudally (Szymusiak and McGinty, 1989), possibly to the lateral hypothalamus or brainstem (Moga et al., 1990; Gritti et al., 1994). Basal forebrain cholinergic and noncholinergic neurons and sleep Basal forebrain cholinergic neurons appear to be closely involved in processes of cortical activation, but recent studies have shown that, in the ventral forebrain area, other neuronal populations may also be involved (Detari et al., 1999; Saper et al., 2001; Jones, 2004). Cholinergic neuron activity and cortical acetylcholine release correlate with tonic cortical activation of waking and REM sleep (Jasper and Tessier, 1971; Marrosu et al., 1995; Duque et al., 2000; Lee et al., 2005), but a portion of basal forebrain GABAergic neurons also fires in association with cortical

411

activation (Manns et al., 2000). Furthermore, specific immunotoxic lesions of basal forebrain cholinergic neurons do not alter the amount of sleep and waking (Kapas et al., 1996; Berntson et al., 2002), although recent data suggest the importance of cholinergic neurons for gamma and theta EEG activity during REM and wakefulness (Lee et al., 2005). Non-cholinergic neurons also appear important for EEG activation, since local injections of procaine (Cape and Jones, 2000) or, non-selective excitotoxic lesions (Buzsaki et al., 1988; Riekkinen et al., 1990) that affect both cholinergic and noncholinergic neurons may, depending upon the extent of the lesion, induce significant slowing of the cortical EEG. In addition, after lesioning or inactivating the basal forebrain, cortical activation can still be evoked by stimulation of the reticular formation through the brainstem–thalamic– cortical system (Steriade et al., 1993; Rasmusson et al., 1994), suggesting that multiple pathways and neuronal types are involved in cortical activation and behavioral state control.

CONCLUSION In summary, we found that in the MCPO/SI regions, in addition to the cholinergic neurons, at least one other neuronal population (Group I), presumably GABAergic, is inhibited by adenosine. Thus the increase in extracellular adenosine that occurs during prolonged waking may suppress cortical activation and promote sleep by inhibiting cholinergic and GABAergic cortically projecting neurons in the MCPO/SI. In addition, along the medial edge of the horizontal limb of the diagonal band and MCPO/SI regions of the basal forebrain lies the VLPO nucleus, a cluster of GABAergic, sleep-active neurons (Szymusiak et al., 1998; Saper et al., 2001) that are disinhibited by adenosine (Chamberlin et al., 2003; Morairty et al., 2004). Thus, at least two populations of cells in MCPO/SI and VLPO are substrates for adenosine action in the ventral forebrain and can potentially mediate the effects of adenosine when it is locally applied or accumulates during prolonged waking. In addition, adenosine increases in these regions may contribute to different complementary components of homeostatic sleep drive including the propensity to sleep, the increase in cortical synchronization and the neurocognitive deficits associated with sleep deprivation. Acknowledgments—This work was supported by P50-HL60292, a Sleep Medicine Education and Research Foundation Young Investigator award, and R37 MH39683. The authors wish to thank Dr. Jun Lu for helpful discussions and Quan H. Ha for providing excellent technical assistance.

REFERENCES Alam MN, Szymusiak R, Gong H, King J, McGinty D (1999) Adenosinergic modulation of rat basal forebrain neurons during sleep and waking: neuronal recording with microdialysis. J Physiol 521(Pt 3):679 – 690. Alonso A, Khateb A, Fort P, Jones BE, Muhlethaler M (1996) Differential oscillatory properties of cholinergic and noncholinergic nucleus basalis neurons in guinea pig brain slice. Eur J Neurosci 8:169 –182.

412

E. Arrigoni et al. / Neuroscience 140 (2006) 403– 413

Basheer R, Porkka-Heiskanen T, Stenberg D, McCarley RW (1999) Adenosine and behavioral state control: adenosine increases cFos protein and AP1 binding in basal forebrain of rats. Brain Res Mol Brain Res 73:1–10. Basheer R, Strecker RE, Thakkar MM, McCarley RW (2004) Adenosine and sleep-wake regulation. Prog Neurobiol 73:379 –396. Beaulieu C, Somogyi P (1991) Enrichment of cholinergic synaptic terminals on GABAergic neurons and coexistence of immunoreactive GABA and choline acetyltransferase in the same synaptic terminals in the striate cortex of the cat. J Comp Neurol 304: 666 – 680. Benington JH, Heller HC (1995) Restoration of brain energy metabolism as the function of sleep. Prog Neurobiol 45:347–360. Berntson GG, Shafi R, Sarter M (2002) Specific contributions of the basal forebrain corticopetal cholinergic system to electroencephalographic activity and sleep/waking behaviour. Eur J Neurosci 16: 2453–2461. Brundege JM, Dunwiddie TV (1997) Role of adenosine as a modulator of synaptic activity in the central nervous system. Adv Pharmacol 39:353–391. Buzsaki G, Bickford RG, Ponomareff G, Thal LJ, Mandel R, Gage FH (1988) Nucleus basalis and thalamic control of neocortical activity in the freely moving rat. J Neurosci 8:4007– 4026. Cape EG, Jones BE (2000) Effects of glutamate agonist versus procaine microinjections into the basal forebrain cholinergic cell area upon gamma and theta EEG activity and sleep-wake state. Eur J Neurosci 12:2166 –2184. Chamberlin NL, Arrigoni E, Chou TC, Scammell TE, Greene RW, Saper CB (2003) Effects of adenosine on gabaergic synaptic inputs to identified ventrolateral preoptic neurons. Neuroscience 119:913–918. Detari L, Rasmusson DD, Semba K (1999) The role of basal forebrain neurons in tonic and phasic activation of the cerebral cortex. Prog Neurobiol 58:249 –277. Dunwiddie TV, Hoffer BJ (1980) Adenine nucleotides and synaptic transmission in the in vitro rat hippocampus. Br J Pharmacol 69: 59 – 68. Duque A, Balatoni B, Detari L, Zaborszky L (2000) EEG correlation of the discharge properties of identified neurons in the basal forebrain. J Neurophysiol 84:1627–1635. Eggermann E, Serafin M, Bayer L, Machard D, Saint-Mleux B, Jones BE, Muhlethaler M (2001) Orexins/hypocretins excite basal forebrain cholinergic neurones. Neuroscience 108:177–181. Fredholm BB, Battig K, Holmen J, Nehlig A, Zvartau EE (1999) Actions of caffeine in the brain with special reference to factors that contribute to its widespread use. Pharmacol Rev 51:83–133. Freund TF, Gulyas AI (1991) GABAergic interneurons containing calbindin D28K or somatostatin are major targets of GABAergic basal forebrain afferents in the rat neocortex. J Comp Neurol 314: 187–199. Gerber U, Greene RW, Haas HL, Stevens DR (1989) Characterization of inhibition mediated by adenosine in the hippocampus of the rat in vitro. J Physiol 417:567–578. Gritti I, Mainville L, Jones BE (1994) Projections of GABAergic and cholinergic basal forebrain and GABAergic preoptic-anterior hypothalamic neurons to the posterior lateral hypothalamus of the rat. J Comp Neurol 339:251–268. Gritti I, Mainville L, Mancia M, Jones BE (1997) GABAergic and other noncholinergic basal forebrain neurons, together with cholinergic neurons, project to the mesocortex and isocortex in the rat. J Comp Neurol 383:163–177. Haj-Dahmane S, Andrade R (1996) Muscarinic activation of a voltagedependent cation nonselective current in rat association cortex. J Neurosci 16:3848 –3861. Harris NC, Constanti A (1995) Mechanism of block by ZD 7288 of the hyperpolarization-activated inward rectifying current in guinea pig substantia nigra neurons in vitro. J Neurophysiol 74:2366 –2378.

Hartig W, Seeger J, Naumann T, Brauer K, Bruckner G (1998) Selective in vivo fluorescence labelling of cholinergic neurons containing p75(NTR) in the rat basal forebrain. Brain Res 808:155–165. Isomoto S, Kondo C, Kurachi Y (1997) Inwardly rectifying potassium channels: their molecular heterogeneity and function. Jpn J Physiol 47:11–39. Jasper HH, Tessier J (1971) Acetylcholine liberation from cerebral cortex during paradoxical (REM) sleep. Science 172:601– 602. Jones BE (2004) Activity, modulation and role of basal forebrain cholinergic neurons innervating the cerebral cortex. Prog Brain Res 145:157–169. Kalinchuk AV, Urrila AS, Alanko L, Heiskanen S, Wigren HK, Suomela M, Stenberg D, Porkka-Heiskanen T (2003) Local energy depletion in the basal forebrain increases sleep. Eur J Neurosci 17:863– 869. Kapas L, Obal F Jr, Book AA, Schweitzer JB, Wiley RG, Krueger JM (1996) The effects of immunolesions of nerve growth factor-receptive neurons by 192 IgG-saporin on sleep. Brain Res 712:53–59. Khateb A, Fort P, Alonso A, Jones BE, Muhlethaler M (1993) Pharmacological and immunohistochemical evidence for serotonergic modulation of cholinergic nucleus basalis neurons. Eur J Neurosci 5:541–547. Khateb A, Fort P, Serafin M, Jones BE, Muhlethaler M (1995) Rhythmical bursts induced by NMDA in guinea-pig cholinergic nucleus basalis neurones in vitro. J Physiol 487(Pt 3):623– 638. Khateb A, Fort P, Williams S, Serafin M, Jones BE, Muhlethaler M (1997) Modulation of cholinergic nucleus basalis neurons by acetylcholine and N-methyl-D-aspartate. Neuroscience 81:47–55. Lee MG, Hassani OK, Alonso A, Jones BE (2005) Cholinergic basal forebrain neurons burst with theta during waking and paradoxical sleep. J Neurosci 25:4365– 4369. Luscher C, Jan LY, Stoffel M, Malenka RC, Nicoll RA (1997) G protein-coupled inwardly rectifying K⫹ channels (GIRKs) mediate postsynaptic but not presynaptic transmitter actions in hippocampal neurons. Neuron 19:687– 695. Manns ID, Alonso A, Jones BE (2000) Discharge profiles of juxtacellularly labeled and immunohistochemically identified GABAergic basal forebrain neurons recorded in association with the electroencephalogram in anesthetized rats. J Neurosci 20:9252–9263. Manns ID, Mainville L, Jones BE (2001) Evidence for glutamate, in addition to acetylcholine and GABA, neurotransmitter synthesis in basal forebrain neurons projecting to the entorhinal cortex. Neuroscience 107:249 –263. Marrosu F, Portas C, Mascia MS, Casu MA, Fa M, Giagheddu M, Imperato A, Gessa GL (1995) Microdialysis measurement of cortical and hippocampal acetylcholine release during sleep-wake cycle in freely moving cats. Brain Res 671:329 –332. McCormick DA, Prince DA (1985) Two types of muscarinic response to acetylcholine in mammalian cortical neurons. Proc Natl Acad Sci U S A 82:6344 – 6348. McCormick DA, Prince DA (1986) Mechanisms of action of acetylcholine in the guinea-pig cerebral cortex in vitro. J Physiol 375: 169 –194. Modirrousta M, Mainville L, Jones BE (2004) Gabaergic neurons with alpha2-adrenergic receptors in basal forebrain and preoptic area express c-Fos during sleep. Neuroscience 129:803– 810. Moga MM, Herbert H, Hurley KM, Yasui Y, Gray TS, Saper CB (1990) Organization of cortical, basal forebrain, and hypothalamic afferents to the parabrachial nucleus in the rat. J Comp Neurol 295: 624 – 661. Morairty S, Rainnie D, McCarley R, Greene R (2004) Disinhibition of ventrolateral preoptic area sleep-active neurons by adenosine: a new mechanism for sleep promotion. Neuroscience 123:451– 457. Morris NP, Harris SJ, Henderson Z (1999) Parvalbumin-immunoreactive, fast-spiking neurons in the medial septum/diagonal band complex of the rat: intracellular recordings in vitro. Neuroscience 92:589 – 600. Murillo-Rodriguez E, Blanco-Centurion C, Gerashchenko D, SalinPascual RJ, Shiromani PJ (2004) The diurnal rhythm of adenosine

E. Arrigoni et al. / Neuroscience 140 (2006) 403– 413 levels in the basal forebrain of young and old rats. Neuroscience 123:361–370. Oliet SH, Poulain DA (1999) Adenosine-induced presynaptic inhibition of IPSCs and EPSCs in rat hypothalamic supraoptic nucleus neurones. J Physiol 520(Pt 3):815– 825. Osaka T, Matsumura H (1995) Noradrenaline inhibits preoptic sleepactive neurons through alpha 2-receptors in the rat. Neurosci Res 21:323–330. Pape HC (1992) Adenosine promotes burst activity in guinea-pig geniculocortical neurones through two different ionic mechanisms. J Physiol 447:729 –753. Porkka-Heiskanen T, Strecker RE, McCarley RW (2000) Brain sitespecificity of extracellular adenosine concentration changes during sleep deprivation and spontaneous sleep: an in vivo microdialysis study. Neuroscience 99:507–517. Porkka-Heiskanen T, Strecker RE, Thakkar M, Bjorkum AA, Greene RW, McCarley RW (1997) Adenosine: a mediator of the sleepinducing effects of prolonged wakefulness. Science 276:1265– 1268. Portas CM, Thakkar M, Rainnie DG, Greene RW, McCarley RW (1997) Role of adenosine in behavioral state modulation: a microdialysis study in the freely moving cat. Neuroscience 79:225–235. Radulovacki M (1995) The pharmacology of the adenosine system. In: The pharmacology of sleep (Kales A, ed), pp 307–319. Berlin: Springer-Verlag. Radulovacki M, Virus RM, Djuricic-Nedelson M, Green RD (1984) Adenosine analogs and sleep in rats. J Pharmacol Exp Ther 228: 268 –274. Rainnie DG, Grunze HC, McCarley RW, Greene RW (1994) Adenosine inhibition of mesopontine cholinergic neurons: implications for EEG arousal. Science 263:689 – 692. Rasmusson DD, Clow K, Szerb JC (1994) Modification of neocortical acetylcholine release and electroencephalogram desynchronization due to brainstem stimulation by drugs applied to the basal forebrain. Neuroscience 60:665– 677. Riekkinen P Jr, Sirvio J, Riekkinen P (1990) Relationship between the cortical choline acetyltransferase content and EEG delta-power. Neurosci Res 8:12–20. Robinson RB, Siegelbaum SA (2003) Hyperpolarization-activated cation currents: from molecules to physiological function. Annu Rev Physiol 65:453– 480. Rudy B (1988) Diversity and ubiquity of K channels. Neuroscience 25:729 –749. Saper CB (1984) Organization of cerebral cortical afferent systems in the rat. II. Magnocellular basal nucleus. J Comp Neurol 222: 313–342. Saper CB, Chou TC, Scammell TE (2001) The sleep switch: hypothalamic control of sleep and wakefulness. Trends Neurosci 24: 726 –731.

413

Seeger T, Alzheimer C (2001) Muscarinic activation of inwardly rectifying K(⫹) conductance reduces EPSPs in rat hippocampal CA1 pyramidal cells. J Physiol 535:383–396. Semba K (2000) Multiple output pathways of the basal forebrain: organization, chemical heterogeneity, and roles in vigilance. Behav Brain Res 115:117–141. Sherin JE, Shiromani PJ, McCarley RW, Saper CB (1996) Activation of ventrolateral preoptic neurons during sleep. Science 271: 216 –219. Steriade M, Amzica F, Nunez A (1993) Cholinergic and noradrenergic modulation of the slow (approximately 0.3 Hz) oscillation in neocortical cells. J Neurophysiol 70:1385–1400. Steriade M, McCarley RW (2005) Brain control of wakefulness and sleep. New York: Kluwer Academic/Plenum. Strecker RE, Morairty S, Thakkar MM, Porkka-Heiskanen T, Basheer R, Dauphin LJ, Rainnie DG, Portas CM, Greene RW, McCarley RW (2000) Adenosinergic modulation of basal forebrain and preoptic/anterior hypothalamic neuronal activity in the control of behavioral state. Behav Brain Res 115:183–204. Szymusiak R, Alam N, McGinty D (2000) Discharge patterns of neurons in cholinergic regions of the basal forebrain during waking and sleep. Behav Brain Res 115:171–182. Szymusiak R, Alam N, Steininger TL, McGinty D (1998) Sleep-waking discharge patterns of ventrolateral preoptic/anterior hypothalamic neurons in rats. Brain Res 803:178 –188. Szymusiak R, McGinty D (1986) Sleep-related neuronal discharge in the basal forebrain of cats. Brain Res 370:82–92. Szymusiak R, McGinty D (1989) Sleep-waking discharge of basal forebrain projection neurons in cats. Brain Res Bull 22:423– 430. Thakkar MM, Delgiacco RA, Strecker RE, McCarley RW (2003a) Adenosinergic inhibition of basal forebrain wakefulness-active neurons: a simultaneous unit recording and microdialysis study in freely behaving cats. Neuroscience 122:1107–1113. Thakkar MM, Winston S, McCarley RW (2003b) A1 receptor and adenosinergic homeostatic regulation of sleep-wakefulness: effects of antisense to the A1 receptor in the cholinergic basal forebrain. J Neurosci 23:4278 – 4287. Woolf NJ (1991) Cholinergic systems in mammalian brain and spinal cord. Prog Neurobiol 37:475–524. Wu M, Shanabrough M, Leranth C, Alreja M (2000) Cholinergic excitation of septohippocampal GABA but not cholinergic neurons: implications for learning and memory. J Neurosci 20:3900 –3908. Xu C, Datta S, Wu M, Alreja M (2004) Hippocampal theta rhythm is reduced by suppression of the H-current in septohippocampal GABAergic neurons. Eur J Neurosci 19:2299 –2309. Zaborszky L, Cullinan W, Braun A (1991) Afferents to basal forebrain cholinergic projection neurons: an update. In: The basal forebrain (Napier T et al., eds), pp 43–100. New York: Plenum Press.

(Accepted 4 February 2006) (Available online 20 March 2006)