INTRODUCTION

Nicotine, the major addictive component of tobacco, is one of the most widely used psycho-stimulating drugs (Corrigall, 1991). Acute exposure to nicotine enhances cognitive functions such as attention, learning, and memory in both animals and humans (Levin and Simon, 1998; Rezvani and Levin, 2001; Levin and Rezvani, 2002), while repeated exposure to nicotine via tobacco smoke can lead to long-lasting changes in the biochemistry of the brain and concomitant changes in responsiveness in a variety of physiological measures (Laviolette and Van der Kooy, 2004). Nicotine exerts its effects on the CNS by interacting with nicotinic acetylcholine receptors (nAChRs). nAChRs are ligand-gated ion channels existing in multiple pentameric subunit combinations (for a review, see McGehee and Role, 1995). Nine of the known alpha subunits (α2–α10) and three of the known beta subunits (β2–β4) have been found in the central and peripheral nervous systems of animals (Itier and Bertrand, 2001). Despite the potential for wide diversity, the most abundant nicotinic receptors in the mammalian brain are the homomeric α7 and the heteromeric α4β2* receptors (The designation ‘α4β2*’ means that these receptors contain α4 and β2 subunits and that other subunits (eg, α5 or α6) may be incorporated in this pentameric complex (Lukas et al, 1999)). (Flores et al, 1992; Holladay et al, 1997; Paterson and Nordberg, 2000). Nicotine acts in vitro as a prototypic agonist for nAChRs, with very high affinity for the α4β2* subtype, and lower but significant affinity for several other nAChR subclasses, including the α7, α3* and α6* (Gotti et al, 1997; Kuryatov et al, 2000; Wonnacott and Sharples, 2001; Romanelli and Gualtieri, 2003). The high affinity of nicotine for the α4β2* nAChR subtype and its relative abundance in the brain make this receptor one of the predominant targets mediating the psychoactive properties of nicotine. However, recent evidence suggests that other subtypes of nAChRs might also modulate aspects of the addictive, reinforcing, and cognitive properties of nicotine (Le Novere et al, 1999; Schilstrom et al, 2003; Laviolette and Van der Kooy, 2004; Wonnacott et al, 2005). Until recently, lack of selective ligands able to discriminate between the different subtypes of nAChRs has hampered the characterization of functional differences between these receptors in vivo. Moreover, nicotine's pharmacological effects are further complicated by the fact that it may either stimulate or desensitize nAChRs and consequently the stimulatory or inhibitory nicotinic responses converge to produce varying effects in different brain structures (Dani and Heinemann, 1996; Chiodini et al, 1999). As a consequence, little is known of the relative contribution of the different nAChR subtypes to the pharmacological and behavioral effects of nicotine at a system level.

Brain-mapping methods permit to locate and quantify the central action of nicotine and abused substances in laboratory animals. 2-Deoxyglucose utilization studies and cerebral blood flow (CBF) mapping with iodo-antipyrene following acute or chronic nicotine administration have evidenced significant signal changes in several striatal and limbic structures (Grunwald et al, 1987, 1991; London et al, 1988a, 1990). Analogously, Fos protein immunohistochemistry has shown elevated signal in restricted brain regions, including target areas of the limbic system, areas related to the optic system, as well as several dopamine-innervated structures (Ren and Sagar, 1992; Matta et al, 1993; Pang et al, 1993; Kiba and Jayaraman, 1994; Nisell et al, 1997; Mathieu-Kia et al, 1998; Schilstrom et al, 2000; Seppa et al, 2001). Although such autoradiography and immunohistochemical techniques have a high spatial resolution, the ex-vivo nature of the approach restricts their use to animal studies and obviously limits temporal resolution to a single point per animal or an average change in activity integrated over a long time period. Recent neuroimaging techniques such as pharmacological magnetic resonance imaging (phMRI) have gained increasing use in animal experimentation as high spatial and temporal resolution images of the brain are attainable noninvasively from rodent brains in vivo (Jenkins et al, 2003; Schwarz et al, 2004a, 2004b; Gozzi et al, 2005b). Moreover, phMRI may be employed to link mechanistic studies of drug action in preclinical animal models with the clinical investigation of nicotine's functional activity.

In this study, we employed a drug-naïve rat model to investigate the anatomical and temporal patterns of brain activation induced by nicotine using phMRI methods. Potential confounds arising from peripheral blood pressure changes were assessed performing a control experiment with norepinephrine, a potent non-brain-penetrant vasopressor (Oldendorf, 1971), at a dose that mimics the cardiovascular response induced by nicotine. The nAChR ion-channel-blocking agent mecamylamine (Papke et al, 2001) was used to prove the pharmacological specificity of the functional changes. In an attempt to highlight the relative contribution of different nAChR subtypes to the observed activation pattern of nicotine, we also investigated the central phMRI response to an acute challenge with two selective nAChR agonists: (R)-N-(1-azabicyclo[2.2.2]oct-3-yl)(5-(2-pyridyl)thiophene-2-carboxamide) (cpdA) and 5-iodo-A-85380 (3-[(2S)-2-azetidinylmethoxy]-5-iodopyridine) (5IA). CpdA is a potent and selective α7 nAChR agonist that crosses the blood–brain barrier (O'Neill et al, 2002; Bunnelle and Decker, 2003; Cilia et al, 2005). 5IA is a compound that features high affinity towards β2* vs α7 and other nAChR subtypes in vitro (ratio of Ki(α7)/Ki(α4β2)25 000 (Mukhin et al, 2000; Mogg et al, 2004)), and has been shown to interact with agonist binding sites on nAChRs in the brain (Vaupel et al, 1998; 2005). A preliminary account of this work has been presented (Gozzi et al, 2005a, 2005c).

MATERIALS AND METHODS

Animal Preparation

All experiments were carried out in accordance with Italian regulations governing animal welfare and protection. Protocols were also reviewed and consented to by a local animal care committee, in accordance with the guidelines of the Principles of Laboratory Animal Care (NIH publication 86-23, revised 1985). These studies were performed on male Sprague–Dawley rats (250–350 g) (Charles River, Como, Italy). Animals had free access to standard rat chow and tap water and were housed in groups of five in solid bottom cages with sawdust litter. Room temperature (20–22°C), relative humidity (45–65%), and dark–light cycles (12 h each, lights on at 06:00 h) were automatically controlled. After arrival, rats were allowed to acclimatize for at least 5 days.

On the day of MR experiment, the rats were anesthetized with 3% halothane (Wyeth, Italy) in a 30 : 70% O2 : N2 gas mixture, tracheotomized, and artificially ventilated with a mechanical respirator (Inspira, Oxford, UK). Upon tracheotomy and throughout surgery, the anesthetic level was maintained at 1.5%. The left femoral artery and vein were cannulated and animal paralysis was achieved by a 0.25 mg/kg intravenous (i.v.) bolus of D-tubocurarine (Sigma-Aldrich, Milano, Italy), followed by a continuous infusion of 0.25 mg/kg/h (in heparinized saline 25 IU/ml) through the artery. The rats employed in studies 2 and 3 (see below) were also provided with an intraperitoneal PE50 cannula. All wounds were infiltrated with 1% lidocaine before incision. At the end of surgery rats were secured into a customized stereotactic holder (Bruker, Ettlingen, Germany). In order to monitor arterial blood gases, 200 μl blood samples were withdrawn from the arterial catheter both before positioning the animal in the magnet and at the conclusion of the experiment, and analyzed for partial pressure of oxygen, carbon dioxide, and pH (AVL, Roswell, GA, USA). Following the former measurement, ventilation parameters were adjusted in order to keep the arterial blood gas values within physiological range (29<pCO2 <50 mmHg; pO2>100 mmHg). Student's t-test analysis of blood gas data (corrected for multiple comparisons using Hochberg's procedure) did not evidence systematic differences between pCO2 or pO2 values pre- and post-MRI, or between groups. Arterial blood pressure was monitored continually throughout the MRI experiment. The heart rate was calculated from arterial pressure pulses by the physiological data acquisition software (Biopac systems Inc., Goweta, CA, USA). An MR-compatible thermocouple probe was used to check rectal temperature at the beginning and at the end of the experiment. During the image acquisition, the maintenance level of anesthesia was set to 0.8%.

As D-tubocurarine is a non-brain-penetrant competitive nAChR antagonist (Alyautdin et al, 1998), preliminary experiments in curarized vs freely breathing animals were carried out to rule out unfavorable crossinteractions between nicotine and the neuromuscular blocker. These studies did not evidence detectable interaction between the two compounds, as the cardiovascular response to nicotine was similar in the two groups, and the paralyzing effect of D-tubocurarine was not affected by the administration of the nicotine bolus.

Experiments and Compounds

Six studies were performed in order to evaluate the central phMRI response to nicotine and its pharmacological mechanism:

  1. 1)

    Acute i.v. challenge of nicotine ((−)-nicotine hydrogen tartrate, Sigma-Aldrich, Italy) at two doses (0.35 mg/kg nicotine free-base (This dose is equivalent to 1 mg/kg of the salt form) (n=9) or 0.1 mg/kg (n=7)) vs vehicle (saline (n=9)).

  2. 2)

    Acute i.p. challenge with mecamylamine (mecamylamine hydrochloride, Sigma-Aldrich, Italy, 0.35 mg/kg (n=6)) vs vehicle (saline (n=6)).

  3. 3)

    Acute pretreatment with either mecamylamine (1 mg/kg i.p. (n=7)) or vehicle (saline i.p. (n=9)) 30 min prior to an i.v. nicotine challenge (0.35 mg/kg). A group of four rats pretreated and challenged with vehicle was used as negative control.

  4. 4)

    Acute i.v. challenge with cpdA (5 (n=2), 10 (n=4), 20 (n=5) and 30 (n=4) mg/kg) vs vehicle (saline (n=5)).

  5. 5)

    Acute i.v. challenge with 5IA (Tocris Cookson Ltd, UK, 5 mg/kg i.v. (n=6)) vs vehicle (saline (n=5)).

  6. 6)

    Acute i.v. challenge with norepinephrine (Sigma-Aldrich, Italy, 0.5 μg/kg, i.v. (n=4)). This experiment was performed in order to assess potential confounds arising from peripheral blood pressure changes elicited by nicotine. Norepinephrine is a potent non-brain-penetrant vasopressor (Tuor et al, 2002), and the dose of 0.5 μg/kg was selected so as to mimic the cardiovascular response to nicotine.

To enhance the solubility of cpdA in saline, the medium was slightly acidified with a few microliters of HCl 1 N, and subsequently corrected to neutral pH (7±1) with NaOH 1 M. Nicotine, mecamylamine, 5IA, and norepinephrine were dissolved in normal saline. All the compounds were administered as 1 ml bolus infusion, followed by a 400 μl flush of saline, at a rate of 1 ml/min. The doses of nicotine were selected based on the results of a functional c-fos study (Mathieu-Kia et al, 1998). Mecamylamine, at the dose used in this study, has been previously shown to effectively antagonize both the functional and behavioral changes induced by nicotine (Reavill and Stolerman, 1990; Seppa et al, 2001). The dose of 5IA was empirically selected following a preliminary study in which lower quantities (0.3, 1 and 3 mg/kg) were also evaluated. Detectable signal changes were observed at 1 and 3 mg/kg, and at 5 mg/kg the changes were consistent and reproducible enough to yield robust group activation maps. All the doses of cpdA used in this study are behaviorally effective in the rat (Wadsworth et al, 2005; Cilia et al, 2005). At the end of the experiment rats were euthanized with an overdose of anesthetic agent.

Magnetic Resonance Imaging

The MRI data were acquired using a Bruker Avance 4.7 T system. The animals were positioned prone in the manufacturer's holding support and a ‘Rat Brain’ curved quadrature receive coil (Bruker, Ettlingen, Germany) was mounted on top of the animal skull and fixed to the animal holder. The animal holder was then fitted into a 72 mm birdcage resonator (Bruker, Ettlingen, Germany), which was used for radiofrequency transmit only. Both coils are standard components provided by the manufacturer.

For each MRI experiment, high-resolution T2-weighted anatomical images were acquired using the RARE sequence with RARE factor 32, matrix 256 × 256, FOV 40 mm, slice thickness 1 mm, 32 contiguous slices. These were acquired in the coronal plane, centered 10 mm caudal from the posterior edge of the olfactory bulb. The phMRI time series images were acquired using the same T2-weighted RARE sequence with the same FOV and geometrical center, but a smaller matrix size of 128 × 128 and eight contiguous 2-mm-thick coronal slices. The rostro-caudal coverage of the time series acquisition was approximately from zbregma+4 mm (anterior edge of slice 1) to zbregma−12 mm (posterior edge of slice 8). The time parameters were TEeff=110ms and TR=2700 ms, with four or eight inline averages, yielding a temporal resolution δt of 40 or 80s, with 64 or 90 sequential time frames (depending on the study). The use of T2-weighted images for the time-series acquisition minimizes sensitivity both to large blood vessels and inhomogeneities of the static magnetic field (Ogawa et al, 1993; Weisskoff et al, 1994; Reese et al, 2000), and also ensured that the geometry of the time-series images matched the anatomical reference images, facilitating subsequent image analysis.

A 2.67 ml/kg dose of Endorem blood pool contrast agent (Guerbet, France) was administered i.v. following five reference image frames, to sensitize the acquisition to changes in CBV. Prior to the injection of any pharmacological agent, an equilibration period of 15 min was allowed. The MRI data were acquired over a period of 25–45 min (depending on the study) following the administration of the pharmacological challenge.

Data Analysis

The phMRI time-series data for each experiment were analyzed on both an image and region of interest (ROI) basis within the framework of the general linear model (GLM) using AFNI v.2.23 (Cox and Hyde, 1997).

For each study, the anatomical images from each subject were first spatially coregistered together by rigid body alignment to a template animal chosen from within the study. The transformation parameters for each subject were subsequently applied to the respective time-series data. The use of the paralyzing agent ensured that no motion-related effects were present in the time series. Data from all animals were checked for motion following acquisition by subtraction of image frames at the beginning and end of the time series, and at appropriate intermediate points (eg, before and after injection). No motion artifacts were detected.

Signal intensity changes in the time series were then converted into fractional relative cerebral blood volume (rCBV) changes on a pixel-wise basis using the transform (Mandeville et al, 1998)

where S(t) is the measured signal intensity, B(t) is the estimated background signal in the absence of transient functional stimuli, and SPRE is the signal intensity prior to administration of the contrast agent. B(t) was set equal to the mean signal intensity B0 during a 3-min period prior to compound injection. To quantify the response to the i.v. challenges, rCBV time series were calculated covering an 8 min preinjection baseline and 14 min post-injection (16 or 32 time points), normalized to a common injection time point within each study. This time window captured the relatively rapid phMRI signal changes observed following injection of the compounds investigated in this study.

For image-based analysis, the rCBV time series were smoothed by a 0.6 mm (2 pixels) FWHM Gaussian kernel in order to increase robustness to small differences in brain shape for subsequent group analysis. The functional response was then quantified by a pixel-wise multiple linear regression. For the nicotine and 5IA experiments, in which a clear response was observed, the regressor of interest was determined from a time course clustering method, wavelet cluster analysis (Whitcher et al, 2005) applied at the group level. This approach identifies a temporal component reflecting a common post-injection response across all subjects, suitable for use as a signal model in the GLM analysis (Schwarz et al, 2005). For the analysis of the mecamylamine and cpdA challenges, no clear signal changes were observed and so a post-injection boxcar model was used in order to identify regions systematically implicated in the response. Constant and linear polynomial terms were also included in the design matrix as regressors of no interest, allowing for background changes such as the slow negative drift in the rCBV signal due to contrast agent washout (Schwarz et al, 2003). The fitted coefficients of the regressor of interest were then analyzed at the group level by pixel-wise unpaired t-tests between appropriate groups (eg, compound vs vehicle) in a random-effects analysis. Statistical threshold of effect for group maps was determined using the Benjamini–Hochberg procedure (Genovese et al, 2002), with the threshold set at a false discovery rate of 1% (q=0.01).

RCBV time courses were extracted a posteriori for specific ROIs based on correspondence between the anatomical images and an anatomical atlas (Paxinos and Watson, 1998; see Figure 1 for ROI localization). From each ROI, the rCBV time course was extracted as the mean value within the ROI for each time point, for each animal. Quantification of the response and group comparisons within each ROI were performed using the same random effects GLM approach as outlined above.

Figure 1
figure 1

Coronal rat brain atlas diagrams from Paxinos and Watson (1998) corresponding to the middle of each 2 mm phMRI slice and indicating the ROIs evaluated in this study (1: nucleus accumbens, 2: orbital cortex, 3: mPFC, 4: primary motor cortex M1, 5: caudate putamen, 6: cingulate cortex, 7: amygdala, 8: lateral hypothalamus, 9: rhinal cortex).

Relative changes in mean arterial blood pressure (MABP) measures were compared between groups at each time point (20 s interval) using the Student's t-test. The level of significance was adjusted for multiple comparisons using Hochberg's method.

RESULTS

Acute nicotine injection at 0.35 mg/kg yielded increased rCBV in a distinct and reproducible pattern of activation. This response included the medial prefrontal cortex (mPFC) and cingulate cortex, with extension into the orbitofrontal, insular, piriform, and rhinal cortices (Figure 2). Activation of subcortical structures was also evident, with discrete regional increases in the amygdala, ventrocaudal hippocampal structures, and the anterior portion of the nucleus accumbens. At no time after nicotine infusion were any significant changes in CBV detected in any discernable portion of cerebellum, midbrain tegmentum, medulla, or pons. The lower dose of nicotine (0.1 mg/kg) produced a much weaker response that did not reach statistical significance in all the regions described above (Figure 4).

Figure 2
figure 2

Maps of rCBV response following acute nicotine challenge (0.35 mg/kg i.v.; n=9) relative to vehicle (n=7) and thresholded at a false discovery rate of 1% (q=0.01). The color scale indicates the magnitude of the fractional rCBV change. The temporal profile of the regressor used as a signal model in the GLM analysis (see Materials and methods) is shown at the top right of the activation map, with an arrow indicating the time of injection. (Two additional caudal slices covering structures of the hindbrain and the cerebellum were analyzed but did not show significant activation, and are therefore not reported here.)

Figure 4
figure 4

Effect of pretreatment with mecamylamine (1 mg/kg i.p., n=7) on the rCBV response induced by acute nicotine challenge (0.1 and 0.35 mg/kg i.v., n=7 and 9, respectively) in three representative anatomical regions (Amy, amygdala; mPFC, medial prefrontal cortex; Orb, orbitofrontal cortex). Baseline data of control animals who received a saline injection are reported for comparison (n=9). Data are expressed as mean±SEM within each group. veh: control rats challenged with vehicle; Mec + 0.35: nicotine challenge (0.35 mg/kg i.v.) following pretreatment with mecamylamine (1 mg/kg i.p.). *p<0.05 vs vehicle, **p<0.01 vs vehicle, ***p<0.001 vs vehicle, $p<0.05 vs vehicle pretreated nicotine challenge (0.35), $$p<0.01 vs vehicle pretreated nicotine challenge (0.35) $$$p<0.001 vehicle pretreated nicotine challenge (0.35).

The time course profile of nicotine-induced activation was analyzed in (nine ROIs) selected a posteriori (see Figure 1 for ROI localization). Since no systematic difference in the profile and the extent of rCBV response between the two brain hemispheres was observed, the ROI analysis was limited to the right hemisphere only. All the regions examined except the caudate putamen, the somatosensory cortex, and the hypothalamus showed a rapid increase in rCBV peaking approximately 2 min after injection, followed by a gradual return toward preinjection baseline values (Figure 3). The strongest signal changes were seen in the mPFC (+19.0±4.4%) and in the orbital cortex (+13.9±4.5%). Despite the difference in the magnitude of the effect, the overall profile and duration of the response did not appear to be region-dependent.

Figure 3
figure 3

ROI time courses showing rCBV changes following nicotine challenge with and without pretreatment with mecamylamine, in three representative regions (nic: rats pretreated with vehicle and challenged with nicotine 0.35 mg/kg i.v. (n=9); mec+nic: pretreatment with mecamylamine (1 mg/kg i.p., n=7) followed by a nicotine challenge 0.35 mg/kg i.v.; veh: control rats challenged and pretreated with vehicle (n=4)). Data are plotted as mean±SEM within each group.

Pretreatment with MEC (1 mg/kg i.p.) significantly attenuated the nicotine-induced rCBV increase in all activated regions (Figures 3 and 4). Pretreatment with vehicle (saline) did not affect the magnitude and spatial distribution of the nicotine response. MEC per se did not produce significant rCBV changes anywhere in the brain. Neither MEC nor vehicle affected the slope and magnitude of basal CBV values prior to nicotine injection.

Acute challenge with 5IA produced reproducible and region-specific positive rCBV changes in all the cortical and subcortical structures activated by nicotine (Figure 5). 5IA also produced positive activation in regions of the dorsomedial thalamus, and in the hypothalamic areas. The time profile of CBV in the activated structures was similar to the one observed with nicotine (Figure 6). At no time after 5IA infusion were any significant changes in CBV detected in any discernable portion of cerebellum, medulla, or pons.

Figure 5
figure 5

Maps of the rCBV response following acute 5IA challenge (5 mg/kg i.v., n=6), relative to vehicle (n=5) and thresholded at a false discovery rate of 1% (q=0.01). The color scale indicates the magnitude of the fractional rCBV change. The temporal profile of the regressor used as a signal model in the GLM analysis (see Materials and methods) is shown at the top right of the activation map, with an arrow indicating the time of injection. (Two additional caudal slices covering structures of the hindbrain and the cerebellum were analyzed but did not show significant activation, and are therefore not reported here.)

Figure 6
figure 6

ROI time courses showing rCBV changes following acute 5IA challenge (5 mg/kg iv, n=6), cpdA (30 mg/kg i.v., n=4), or vehicle (saline i.v., n=5) in three representative anatomical regions. Data are plotted as mean±SEM within each group.

CBV group maps of animals challenged with cpdA did not show significant rCBV changes at any of the doses tested (5, 10, 20, and 30 mg/kg i.v.) The temporal profiles of activation in ROIs also did not evidence significant activation in any of the nine regions analyzed (Figures 6 and 7).

Figure 7
figure 7

RCBV response to acute challenge with 5IA (5 mg/kg i.v., n=6) and cpdA (20 mg/kg i.v., n=5) in three representative anatomical regions (Amy, amygdala; mPFC, medial prefrontal cortex; Orb, orbitofrontal cortex). Baseline data of control animals who received a saline injection are reported for comparison (n=5). Data are expressed as mean±SEM within each group. ***p<0.001 vs vehicle.

I.v. administration of saline (vehicle) did not give rise to significant changes in peripheral blood pressure. Nicotine (0.35 mg/kg) induced a transient increase in arterial blood pressure, peaking approximately 90 s after the injection (+27.1±9.8%, p<0.05 at peak compared to vehicle (Figure 8)) and returning to preinjection baseline levels within 4 min form the injection. No correlation between response amplitude and absolute or relative MABP changes was found in any of the ROIs examined. The lower dose of nicotine (0.1 mg/kg) did not produce detectable cardiovascular changes. 5IA induced a fast-onset increase in arterial blood pressure (+25.8±2.1%, p<0.05 at peak compared to vehicle) very similar in shape and duration to the one previously observed with nicotine (Figure 9). Administration of MEC was accompanied by a slow and progressive decrease in blood pressure, which reached a stable negative plateau ca. 20 min after injection (−20.3±4.6%, p<0.01 compared to predrug baseline). The effect persisted for more than 30 min and was still present at the time of the nicotine challenge. CpdA produced a dose-dependent decrease in MBP, which reached a minimum approximately 1 min after the start of the injection (Figure 9). At the two highest doses the magnitudes of the effect were similar (−34.5±12.8 and −36.4±5.4% at 30 and 20 mg kg, respectively), but at 30 mg/kg the effect was slightly more sustained. At 10 and 3 mg/kg, the changes were not statistically different from vehicle (p>0.112 and 0.217, respectively, all time-points). The absolute magnitude of the blood pressure changes observed with nicotine, 5IA, and cpdA (at all doses) was within the range of CBF autoregulation (50–140 mmHg) under halothane anesthesia reported by Zaharchuk et al (1999).

Figure 8
figure 8

Effect of nicotine (0.35 mg/kg, n=9), norepinephrine (0.5 μg/kg, n=4), and vehicle (saline, n=9) infusion on the time course of MABP (relative changes). Data are shown as mean±SEM across subjects for each time point.

Figure 9
figure 9

Effect of 5IA (5 mg/kg, n=5) and cpdA (10, 20, and 30 mg/kg, n=4, 5, and 4, respectively) infusion on the time course of MABP (relative changes). The effect of control (saline) injection has already been reported in Figure 8, and has not been plotted here to improve the clarity of the figure. Data are shown as mean±SEM across subjects for each time point.

In the acute norepinephrine challenge study, a blood pressure increase similar in amplitude and shape to that elicited by nicotine was observed (+32.7±17.0%, p<0.05 at peak compared to vehicle; Figure 8). PhMRI group activation maps of norepinephrine vs vehicle did not reveal significant rCBV changes in any brain region. Similarly, a post hoc ROI analysis in nine brain structures (see Figure 1) did not evidence significant alterations in the CBV baseline attributable to the norepinephrine challenge. These results suggest that the observed functional changes are centrally mediated and not caused by the transient peripheral increase in MBP, in line with previous studies, which showed that MBP changes within the cerebral blood flow autoregulatory range do not cause significant changes in CBV (Zaharchuk et al, 1999).

DISCUSSION

The preclinical investigation of the acute effects of nicotine is often based on experimental methods that require ex vivo approaches or focal measurements of brain activity. This has limited the possibility of obtaining a comprehensive picture of nicotine's neuromodulatory activity within the brain. As a consequence, little is known of nicotine's acute effects on neuronal activity at a systems level. Functional MRI techniques allow noninvasive three-dimensional brain mapping of the functional response induced by various types of stimulus, including pharmacologically active agents. Animal phMRI research represents a promising tool for the identification and optimization of experimental paradigms aimed at determining a drug's mechanism and site of action. CBV-sensitive phMRI methods have a high sensitivity at 4.7 T magnetic fields (Mandeville et al, 1998), and have been recently applied to map pharmacologically evoked activation elicited by several neuroactive substances in rodents (Chen et al, 1997; Nguyen et al, 2000; Marota et al, 2000; Reese et al, 2000; Schwarz et al, 2004a, 2004b; Gozzi et al, 2005b).

In the present study, we have applied phMRI methods to map for the first time the acute rCBV response to nicotine in drug-naïve rats. Several lines of evidence argue strongly that the changes in rCBV described here are a direct consequence of nicotine-induced regional increase in neuronal activity and not due to unspecific peripheral effects of nicotine or a global effect on cerebral vasculature. Firstly, the pattern of activation observed with both nicotine and 5IA is region-specific, with CBV remaining unchanged in a number of brain regions. The significant attenuation of activation by the nAChR antagonist mecamylamine also supports the pharmacological specificity of the process. Furthermore, by reproducing the blood pressure response to nicotine with a systemic infusion of norepinephrine, a potent vasopressor that does not cross the blood–brain barrier (Tuor et al, 2002), no detectable rCBV changes were detected centrally, thus corroborating the idea that the observed rCBV response is not the predominant result of peripheral vascular changes. Moreover, the magnitude of the blood pressure changes elicited by nicotine and 5IA was within the range of central hemodynamic autoregulation described by Zaharchuk et al (1999). These authors have shown that, under an anesthetic protocol very similar to the one used in our study, CBF autoregulation is preserved over a wide range a values (from 50 to 140 mmHg), and that changes in blood pressure within this range produce only small and nonsignificant changes in rCBV. Finally, previous studies have shown that perivascular microapplication of nicotine on the cortical surface exerted negligible effects on pial vessels over a wide range of concentrations, consistent with the local increase in rCBV occurring as a result of increased metabolic demand and nicotine having either no or only a minor direct influence on the cerebral circulation (Edvinsson and Krause, 2002). In conclusion, converging evidence strongly supports the hypothesis that the pattern of brain activation we observed is a direct consequence of the selective action of nicotine on particular cortical and subcortical targets resulting in neural activation within specific structures. As 5IA closely reproduced the activation pattern and cardiovascular changes elicited by nicotine, and since it is known to stimulate the same receptors as nicotine, it is reasonable to assume that the observed rCBV changes are also the result of central activity of 5IA on specific neuronal populations in the brain.

I.v. infusion of nicotine elicited region-specific increase in rCBV in several cortical (medial-prefrontal, cingulate, orbital, piriform, frontal and rhinal cortices) and subcortical (amygdala, caudal-ventral hippocampal structures) structures. All the brain regions identified in this study have been reported to have moderate to high density of nAChR subunit mRNA and nicotine-binding sites (Clarke et al, 1985; Wada et al, 1989; Tribollet et al, 2004). The anatomical pattern of the functional response comprises many structures that have previously been identified with different brain-mapping modalities to be acutely modulated by nicotine. Mathieu-Kia et al (1998) reported a c-Fos-like pattern strikingly similar to the activation maps reported in this study, with increased Fos immunoreactivity in terminal fields of neurons of the ventral tegmental area, such as the mPFC and the central amygdala, as well as in the cingulate, orbital, and piriform cortices. Increased c-fos immunoreactivity in these structures was also reported by other investigators (Matta et al, 1993; Kiba and Jayaraman, 1994; Schilstrom et al, 2000; Seppa et al, 2001). The present study also seems to indicate a preferential involvement of cortical structures with a rich dopamine innervation (medial prefrontal, cingulate, orbital, pirirform) compared with those (frontal and parietal) with fewer dopaminergic terminals. This finding is in close agreement with the results of previous studies with c-fos immunodetection (Mathieu-Kia et al, 1998; Schilstrom et al, 2000; Seppa et al, 2001). Using autoradiographic 2-deoxy-D-[1-14C]glucose methods, many cortical areas were analyzed and only the cingulate cortex (Grunwald et al, 1991) showed an increased local cerebral glucose utilization. Hyder et al (2000) recorded an increase in CBF in the somatosensory cortex of rats challenged with a high dose of nicotine (4 mg/kg). Of particular interest is the activation found in the rhinal cortex, since this area, together with the prefrontal and orbitrofrontal cortices and the hippocampus, serves as a key neuroanatomical substrate of recognition memory in rats (Steckler et al, 1998). The involvement of brain structures that mediate recognition memory is consistent with the results of behavioral studies that show improved working memory functions following acute treatment with nicotine in drug-naïve rats (reviewed by Levin and Simon, 1998).

Increasing evidence suggest a modulation of mesolimbic and mesocortical dopamine pathways by acute nicotine administration (Kiba and Jayaraman, 1994; Laviolette and Van der Kooy, 2004). Functional c-fos studies reported preferential activation in structures that receive dopaminergic afferents from the VTA, such as the nucleus accumbens and the central amygdala, but not in the dorsal caudate-putamen innervated by projections from the substantia nigra pars compacta (Mathieu-Kia et al, 1998). Consistent with this, we observed positive activation in the mPFC and the central amygdala. A clear signal enhancement was also found in the anterior portion of the nucleus accumbens. However, the response is not uniform across the structure, and appears to be weaker in its posterior segments. As the most anterior portion of the nucleus accumbens borders on strongly activated areas such as the orbital cortex, a signal contamination from the highly responsive neighboring structures cannot be excluded. The data presented here therefore do not establish unequivocally whether the nucleus accumbens is activated by acute nicotine under our experimental conditions, and the answer to this question must await further experiments. The presence of activation in the mPFC, the amygdala, and nucleus accumbens would be indicative of a preferential nicotine interaction with dopaminergic neurons of the ventral tegmental area as reported by Mathieu-Kia et al (1998), although a presynaptic action of nicotine on dopaminergic terminals or a direct nicotine activation of neurons in the activated structure cannot be excluded.

Notably, although previous c-fos and local cerebral glucose utilization studies (London et al, 1988a, 1988b; Grunwald et al, 1991) reported increased immunoreactivity in visual and limbic structures connected with the retina, none of these structures seem to be functionally activated in our rCBV activation maps. As the superior collicoli are involved in the representation of visual, auditory, and somatosensory stimuli (Sparks, 1986; Guitton, 1992), the lack of activation in central visual pathways may be attributed to the absence of multimodal sensorimotor information, a general consequence of the direct general suppressive effects of the anesthetics on neural transmission (Franks and Lieb, 1998; Angel, 1993). This hypothesis is corroborated by the results of Wang et al (2000), who reported an anesthesia-induced suppression of c-Fos immunoreactivity induced by noxious stimuli in different layers of the superior colliculus.

Previous functional studies using c-fos immunodetection or the deoxyglucose method also highlighted the acute effects of nicotine also in nAChR-dense areas that were not identified in the present study, such as the dorsolateral striatum (Kiba and Jayaraman, 1994; Nisell et al, 1997), various catechoalminergic regions of the brainstem (Matta et al, 1993), and thalamic and hypothalamic structures (London et al, 1988a; Mathieu-Kia et al, 1998; Seppa et al, 2001). An increase in thalamic areas was also revealed by measuring the local cerebral blood flow rats (Grunwald et al, 1991). Although differences in the dose and route of administration of nicotine exist, these discrepancies could be due to the use of the anesthetic. Our phMRI protocol includes the use of general anesthesia to avoid motion artifacts and ensure stable and reproducible animal physiology. Halothane is an anesthetic widely used in preclinical phMRI, as it has been shown to preserve cerebrovascular reactivity and cerebral blood flow autoregulatory response (Morita et al, 1977; Young et al, 1991; Lee et al, 1994). However, halothane, like most general anesthetics, exert, its activity through simultaneous interactions with some nAChR-dense neuronal structures, including the thalamus, midbrain reticular formation, and the hypothalamus (for a review, see Rudolph and Antkowiak, 2004). Also, several volatile and injectable anesthetics have been reported to inhibit in vitro various nAChRs subtypes at concentrations that might be clinically relevant (Violet et al, 1997). Therefore, complex in vivo interactions between the anesthetic and the nicotinic receptors can in principle affect the extent or distribution of the rCBV response. Nonetheless, the distribution of the rCBV changes observed in this work is congruous, reproducible, and consistent with several previously published results, thus suggesting that, if these interactions do occur, they only marginally affect the functional response to nicotinic agonists.

By using mecamylamine, a nonselective nAChR antagonist that is thought to block the nicotinic ion channels (Giniatullin et al, 2000), we were able to demonstrate the involvement of nAChRs in the observed functional response. This result is in agreement with in vivo studies in nonanesthetized animals, in which blockade of nAChRs by preadministration of mecamylamine has been shown to antagonize the behavioral and neurometabolic effects of nicotine (London et al, 1988a; Ren and Sagar, 1992; Pang et al, 1993; Kiba and Jayaraman, 1994; Marshall et al, 1997). In subsequent experiments, we sought to assess the relative contribution of different receptor subtypes to the response to nicotine by using selective nAChR agonists. Understanding the differential contribution of the diverse nAChR isoforms represents an important opportunity to develop drugs that selectively retain the cognition-enhancing properties of nicotine, while minimizing the unwanted side-effects (eg, hypothermia, nausea, cardiovascular changes, and abuse liability). The majority of the pharmacological tools available target the two most abundant nAChRs subtypes in the brain: α7 and α4β2*. These two receptor isoforms have both shown to be implicated in the functions of learning and memory and the cognitive-enhancing properties of nicotine (Rezvani and Levin, 2001; Levin et al, 2002). Recent results suggest that both receptors are implicated in the behavioral modulation of the reinforcing properties of nicotine, although with distinct roles (Wonnacott et al, 2005). Several behavioral and pharmacological studies of nicotinic receptor subtype take advantage of the availability of potent and selective nAChR antagonists such as methyllcaconitine (Ward et al, 1990) and dihydro-beta-erythroidine (Williams and Robinson, 1984). However, both compounds are competitive antagonists, and MLA has been reported to have a poor brain penetration and it is often administered directly into the brain to bypass the blood–brain barrier (Wonnacott and Sharples, 2001; Navarro et al, 2002; Van Kampen et al, 2004). The fact that these drugs interact reversibly with the nAChR channel, along with the poor PK profile of MLA, may compromise studies like the present one in which the infusion of a high-dose bolus of nicotinic agonists (eg nicotine) is required to produce detectable signal changes. Unsurprisingly, attempts to antagonize nicotine's response by using the aforementioned drugs failed to give consistent results (Gozzi A, unpublished results). To overcome these shortcomings, here we have applied a different approach and used two novel compounds that show selective agonist activity at α7 and α4β2 receptors, respectively.

CpdA is a potent (EC50=14 nM) and centrally penetrant nAChR agonist with nearly full efficacy at human α7 receptors expressed in GH4 cells. This compound shows negligible activity at other major nAChR subtypes (α2β4, α3β4, α4β4, α3β2, and α4β2) and is devoid of 5-HT3 activity (De Filippi et al, 2002; O'Neill et al, 2002; Bunnelle and Decker, 2003). In this study, the evaluation of cpdA as a single challenge at four increasing doses failed to induce significant rCBV changes at any of the doses tested. In previous reports, cpdA (1–10 mg/kg i.p.) showed rapid and robust in vivo central effects as assessed by behavioral (Wadsworth et al, 2005; Cilia et al, 2005) and neurochemical studies (Tong and Dawson, 2005). Therefore, the absence of significant rCBV changes following cpdA administration is unlikely to be the result of low potency or poor CNS penetration. The magnitude of the negative blood pressure changes in rats treated with high doses (20 and 30 mg/kg) of cpdA were within the autoregulatory range identified by Zaharchuk et al (1999). Moreover, cpdA did not produce significant activation at any of the lower doses, including 5 and 10 mg/kg, at which cpdA induces robust behavioral effects (Wadsworth et al, 2005; Cilia et al, 2005), without significantly affecting the blood pressure. Therefore, peripheral effects are unlikely to have precluded detection of centrally mediated response to the compound. A further potential issue is the rapid desensitization that is characteristic of agonist-mediated responses at the α7 receptor (Seguela et al, 1993; Quick and Lester, 2002). Given that any exogenously administered drug will be present in the extracellular space of the receptor for a time period of minutes to hours, the α7 receptor may be in a predominantly desensitized state. CpdA, like other selective α7 agonists, induces rapid receptor desensitization in vitro, and probably in vivo (O'Neill et al, 2002). However, nicotine at the dose tested in this study reaches brain concentration levels that are substantially above the submicromolar range needed for effective α7 desensitization (Ghosheh et al, 2001; Quick and Lester, 2002). It is therefore conceivable that α7 receptors are also in a desensitized state following nicotine administration. Based on these considerations, it can be assumed that there was no substantial difference between the maximum steady-state currents induced by nicotine and cpdA. As discussed above, the anesthetic is another experimental factor that, in principle, could interfere with the neurometabolic cascade underlying the rCBV response. However, α7 nicotinic receptors do not appear to be significantly inhibited by clinically relevant concentrations of halothane (Violet et al, 1997; Tassonyi et al, 2002). Taken together, these data suggest that a failure of cpdA to elicit significant rCBV changes is unlikely to be due to explanations such as stronger receptor desensitization or unfavorable interactions with the anesthetic. Since cpdA did not produce a significant response at any of the doses tested, we therefore conclude that the selective pharmacological stimulation of α7 nicotinic receptors is not accompanied by hemodynamic changes. As nicotine showed a robust neurovascular response, this finding suggests that the rCBV response to nicotine is not predominantly mediated by α7 nAChRs.

The evaluation of 5IA revealed a pattern of activation showing a striking spatial correspondence with the distribution of nicotine-induced CBV changes. Functional activation was found in all the cortical and subcortical structures previously described for nicotine. Interestingly, 5IA also elicited focal activation in regions of the dorsomedial thalamus and in the hypothalamic areas. The presence of activation in these nuclei is in good agreement with the very high and uniform density of high-affinity α4β2* nicotinic receptor in the thalamus (Clarke et al, 1985). Several experimental factors can contribute to the detection of significant thalamic activation in 5IA but not in nicotine rCBV maps, and a direct comparison of the pharmacological potency of these two compounds in vivo is beyond the scope of this paper. As brain-concentration levels of 5IA following i.v. administration are not available in the literature, no conclusions can be drawn on the relative extra synaptic concentrations of the two compounds under the experimental conditions of this study. However, a higher in vivo potency with respect to nicotine would be compatible with the pharmacological profile of 5IA, as demonstrated by competitive binding studies in which 5IA showed seven-fold higher affinity than nicotine at α4β2* receptors (Saji et al, 2002).

In vitro, 5IA features exceptional selectivity for the α4β2* subtype when compared to α7, α3β4*, and muscle type nAChRs (Mukhin et al, 2000). Recent results have shown that 5IA is also a potent activator of α6β2* receptors in vitro (Mogg et al, 2002), and based on this finding it is now generally referred to as β2*-selective nicotinic receptor ligand (Liu et al, 2003). Thanks to its high-affinity, low nonspecific binding, and lack of toxicity in mammals, 5IA has been characterized as a tool for noninvasive in vivo studies of β2* nicotinic receptor (Vaupel et al, 1998, 2005). Since the affinity of 5IA toward α7 nAChRs is only 1/25 000 of that towards α4β2, it is reasonable to assume that the compound has negligible activity at α7 receptor in vivo, and that any residual activity at α7 receptors would be largely outweighed by the predominant action at β2* receptors. In line with this, β2-knockout mice showed total loss of 5IA binding in all brain regions of Mukhin et al (2000). Taken together, these data suggest that the rCBV changes produced by 5IA are mediated by β2* nAChRs. Considering the lack of phMRI response to cpdA, and the high spatial overlap between the activation profile of nicotine and 5IA, this also suggests that the functional response to nicotine is predominantly mediated by β2* nAChRs. Consequently, in addition to having a crucial role in mediating some key aspect of the addictive and acute cognitive properties of nicotine (Levin, 2002; Picciotto, 2003; Kauer, 2005), β2-containing nAChRs appear to be also implicated in the modulation of the acute hemodynamic response to nicotine. Considering that 5IA shows significant potency only for α4β2* and α6β2*, and that the most abundant β2-containing receptors in mammalian brain are α4β2* (Flores et al, 1992), it is likely that the major specific signal produced by 5IA and nicotine would reflect a response mediated by α4β2* nAChRs. In line with this hypothesis, α6β2* nAChRs are present in low concentration and show limited distribution in the brain (Picciotto et al, 2001), with a focal localization at dopaminergic terminals, such as the medial habenula, the locus coruleus, and the substantia nigra pars compacta, that did not show activation in this study.

To the best of our knowledge, no functional neuroimaging study on the acute effects of nicotine in drug-naïve individuals has been published to date. MRI methods have been used to investigate aspects of the neuroplastic changes induced by repeated nicotine exposure nicotine in two recent rodent studies (Shoaib et al, 2004; Calderan et al, 2005). Clinical studies using positron emission tomography and fMRI in smokers, or abstinent smokers, have also been published (reviewed by McClernon and Gilbert, 2004). However, repeated exposure to nicotine induces complex and long-lasting changes in the biochemistry of the brain, including a substantial upregulation of nicotinic-binding sites (Benwell et al, 1988; Buisson and Bertrand, 2002). Moreover, neuroimaging studies in abstinent individuals do not allow an easy separation of the acute effects of the drug from the functional correlates of craving (Hommer, 1999). Taken together, these aspects limit the possibility of a direct comparison between the findings of this study and the results of previously published neuroimaging studies. In this respect, the data presented here represent a novel and valuable preclinical benchmark for future clinical investigations of the functional effects of selective nicotinic agents by using fMRI methods.

In conclusion, we have applied phMRI methods to map the functional response to nicotine in drug-naïve rats. Nicotine produced a distinct and reproducible pattern of activation including cortical and subcortical structures involved in the mediation of the acute cognitive and behavioral effects of nicotine. The pharmacological specificity of the effect was demonstrated by acute pretreatment with the nAChR ion-channel-blocking agent mecamylamine, which suppressed the rCBV response to nicotine. After acute challenge with a selective CNS penetrant α7 nAChR agonist, we did not observe a detectable CBV response, while experiments with a selective β2* nAChR agonist elicited a pattern of activation very similar to that of nicotine. Taken together, these results suggest that the hemodynamic response to nicotine in drug-naïve individuals is predominantly mediated by β2-containing nAChR isoforms presumably belonging to the α4β2* subtype. As nicotinic agents possess well-established cognitive-enhancing properties, the results of this study may help the identification of the neuronal substrates mediating aspects of cognitive and attentional functions.