Involvement of Norepinephrine in the Control of Activity and Attentive Processes in Animal Models of Attention Deficit Hyperactivity Disorder

Functional and morphological studies in children affected by Attention Deficit Hyperactivity Disorder (ADHD) suggest a prefrontal cortex (PFc) dysfunction. This cortical region is regulated by subcortical systems including noradrenergic (NEergic), dopaminergic (DAergic), cholinergic, serotonergic, and histaminergic pathways. A wealth of data in humans and in animal models demonstrates altered dopamine (DA) regulation. Drugs that modulate norepinephrine (NE) transmission are also effective in ADHD patients, thus leading to the hypothesis of a NEergic disorder. This review covers the regulation of PFc functions by NE and the interaction between the NE and DA systems, as suggested by pharmacological, electrophysiological, morphological, and gene knock out (KO) studies. A negative feedback between NE and DA neurons emerges from KO studies because KO mice showing increased (NE transporter (NET) KO) or decreased (DBH and VMAT2 KO) NE levels are respectively associated with lower and higher DA levels. Locomotor activity can be generally predicted by the DA level, whereas sensitivity to amphetamines is by NE/DA balance. Some animal models of ADHD, such as spontaneously hypertensive rats (SHR), show alterations in the PFc and in the DA system. Evidence about a correlation between the NE system and hyper-locomotion activity in such animals has not yet been clarified. Therefore, this review also includes recent evidence on the behavioral effects of two NET blockers, reboxetine and atomoxetine, in two animal models of ADHD: SHR and Naples High Excitability rats. As these drugs modulate the DA level in the PFc, certain effects are likely to be due to a rebalanced DA system. We discuss the significance of the results for theories of ADHD and make suggestions for future experimentation.


INTRODUCTION
The grayish-blue colored small pigmented region of the fourth ventricle floor, the locus coeruleus (LC), includes a cluster of about 1600 neurons per nucleus in rats, several thousand in monkeys, and 10,000 to 15,000 in humans (Foote et al., 1983; for review see Berridge & Waterhouse, (C) 2004 Freund & Pettman, U.K. 2003). These neurons contain norepinephrine (NE), as shown by Dahlstrom andFuxe in 1964 (1964), and are the origin of widespread nerve terminals in the mammalian brain and spinal cord .
Norepinephrine neurons give rise to two major bundles, the dorsal and ventral one. The former (dorsal tegmental bundle) originates from LC neurons and terminates in many regions of the forebrain, cerebellum, and spinal cord (Cerbone & Sadile, 1994). The latter or central tegmental tract originates from diffuse NE cell groups in the pons and medulla and terminates in the hypothalamus and spinal cord (Mason, 1981).
Several drugs interfere with the NE system. The psychostimulant amphetamine and the dopamine (DA) transporter blocker, methylphenidate, inhibit LC firing (Lacroix & Ferron, 1988). Interestingly, both drugs are largely used in the treatment of a disorder characterized by hyperactivity and inattention, known as attention deficit hyperactivity disorder (ADHD). As a consequence, the possibility that ADHD represents a NEergic disorder has been recently discussed (Biederman & Spencer, 1999).
Attention deficit hyperactivity disorder is a neurodevelopmental problem diagnosed in to 3 percent of children (Castellanos& Tannock, 2002;Sergeant et al., 2003). The main features are "a persistent pattern of inattention and/or hyperactivity/impulsivity that is more frequently displayed and more severe than in individuals at a comparable level of development" (American Psychiatric Association (APA), 1994). The disorder can occur before the 7 year of age and in different environments. Different variants have been identified: hyperactive-impulsive, inattentive, and combined (APA, 2004).
The contributions of NE and DA neurotransmissions to the motor and cognitive symptoms of ADHD have been investigated in rodent and primate models and in humans as well (Castellanos et al., 1996). Dopamine  The existence of a PFc in rodents has been widely debated (Brown & Bowman, 2002;Uylings et al., 2003). In particular, the human PFc can be divided into different subregions, which share a common input represented by the mediodorsal thalamie nucleus (Uylings et al., 2003). In rats, however, the latter projects only to the medial and orbital frontal lobe, but not to dorsolateral regions. The PFc has been involved in attention, decision making (Mulder et al., 2003), temporal organization of behavior (Fuster, 2000), working memory (Goldman-Rakic, 1996), and reversal learning (McAlonan & Brown, 2003).
Functional subdivision of the PFc in rats, mainly based on lesion studies, comprises three main fields: medial (dorsal and ventral), orbitofrontal, and dorsolatcral. Lesions of the rat medial PFc impair attention and attentional set shift (Brown & Bowman, 2002;Uylings t al., 2003), reduce anxiety, and increase novelty-induced locomotor activity (Deacon et al., 2003). Lesions of the dorsomedial (anterior cingulate) PFc impair discriminative accuracy, whereas the ventromedial (infralimbic) region sustains inhibitory control (Chudasama et al., 2003). Orbitofrontal lesions impair behavioral flexibility, leading to perseverative tendencies, thus decreasing performance in various learning and memory tasks (Chudasama et al., 2003;Vafaei & Rashidy-Pour, 2004). Lesions of the dorsolateral PFc lead to shifting strategy alterations, impairing behavioral flexibility in rats. Conversely, a corresponding lesion in humans impairs working memory.
The activity of the PFc is modulated by a set of subcortical systems comprising acetylcholine-, DA-, 5-hydroxytryptamine (5-HT)-, NE-, and histaminecontaining neurons (Robbins, 1997;Uylings et al., 2003). Recently, a sixth system has been characterized, originating from the lateral hypothalamus, which releases glutamate and orexin (also called hypocretin) as neuromodulat6rs (Torrealba et al., 2003). All derive from a relatively small number of neurons with a terminal arborization that allows the innervation of the entire forebrain and spinal cord. These subcortical systems partially share target regions and influence each other in a very complex manner. In particular, NE neurons from the LC project to almost the entire neocortex, and the NE terminals are denser in layers IV and V, although the laminar distribution is less evident in rodents (Foote et al., 1987). Th is widespread input to the cortex differs from the DA input, which is restricted at the cortical level to the medial PFc (Bjorklund & Lindvall, 1984). The only cortical areas that project back to the LC are the same medial PFc in rats and the dorso-lateral PFc in primates (Amsten, 1997). Interestingly, the orexin system also interacts with both DA (Fadel & Deutch, 2002) and NE systems (Monda et al., 2004). Finally, the PFc receives input also and projects back to the reticular formation (Newman et al., 1989;Gritti et al., 1997).
Electrophysiologieal evidence shows that PFe neurons e activated during the delay period of a delayed-response trial (see Goldman-Rakie, 1996), thus suggesting its involvement in working memory functions. Recent data would also suggest a direct involvement in reward-directed learning.
In fact, neurons in the PFc increase their firing rate at the presentation, whereas they reduce the firing rate from the conditioned-stimulus presentation up to the reward delivery, aider the task has been learnt (Mulder et al., 2003). Thus, the PFc has access to reward stimuli. In fact, a loop has been anatomically and functionally described, involving the PFc, striatum, globus pallidum, and substantia nigra (the striatal output), the thalamus (ventromedial nucleus), and back to the PFc (Uylings et al., 2003). This circuit is thought to link motivationally relevant stimuli to motor areas (Kalivas et al., 1993). The electrophysiologieal data reported above rather suggest, however, that PFe neurons encode for prediction in reward delivery and transfer this information directly to the VTA, where the error in prediction is calculated (Williams et al., 2004).
Interestingly, the process of error prediction is possibly altered in ADHD, thus suggesting a direct involvement of PFc modulation in this disorder. More direct evidence of an involvement of the PFc in ADHD is the following" reduced volume of right PFc (Hynd et al., 1990;Castellanos et al., 1996;Filipek et  Moreover, the analysis of the PFc in animal models of ADHD reveals similar alterations in the PFc. In fact, rats selected over a mixed population for their impulsivity in an attentive task show a lower metabolic activation of the PFc (Barbelivien et al., 2001). In addition, the PFc of SHR compared with Wistar Kyoto (WKY) control rats shows (a) a reduced number of tyrosine hydroxylase (TH) fibers (King et al., 2000) and (b) increased NE activity in the PFc (Russell, 2000).

Modulation of PFc functions by NE and DA
The PFc receives projections from both the VTA and the LC and sends back projections to the VTA and to regions where dendrites of NE LC cells are present (Jodo et al., 1998). In fact, several data (see e.g. Jodo et al., 1998) show that the PFc is a major source of excitation of the LC.
Strikingly, DA does not modulate NE release in the PFc (Vanderschuren et al., 1999), whereas NE can modulate DA release (Bymaster et al., 2002). In fact, the excitation of LC neurons increases the release of DA in the PFc, but this effect does not occur through a direct excitation of DA neurons or terminals (Kawahara et al., 2001). A mechanism responsible in part for this effect is represented by the reuptake of DA by the NE transporter (NET) (Carboni et al., 1990). As a consequence, the NE terminals in the PFc would regulate the resting levels of DA through the NET. This offers an interpretation of the increase NE and DA levels in the PFc after treatment with drugs that block the NET, such as reboxetine and atomoxetine (Bymaster et al., 2002). In fact, the effects of psychostimulants such as D-amphetamine on DA outflow can be partially controlled by NE receptors (Darracq et al., 1998). The NE and DA systems, however, can also interact at the nuclei of origin and in other target sites.

Dopamine and norepinephrine interactions
Direct connections have been found between the VTA and the LC. In fact, DA neurons in the VTA express a-2c NE receptors (Lee et al., 1998), and there is a tonic release of NE in the VTA (Reith et al., 1997). Indeed, morphological evidence supports a direct projection from the LC to the VTA (Phillipson, 1979). In addition, direct projections from the VTA to the LC exist (Swanson, 1982) and are able to excite NE neurons (Deutch et al., 1986).
The nucleus accumbens is the target of DA neurons from the VTA and NE neurons from the nucleus tractus solifarius (NST) (Berridge & Waterhouse, 2003). Thus, local injection of adrenoceptor agonists modulates DA release (Yavich et al., 1997), and, conversely, DA modulates the release ofNE (Vanderschuren et al., 1999). These interactions at the nuclei of origin and target sites and the DA reuptake by the NET make it difficult to disentangle the behavioral effects of NE and DA drugs. In particular, inferences about the pathogenesis of ADHD based on the efficacy of drugs acting on the NE system should be advanced with caution. To this aim, the behavioral effects of genetic manipulation of the NE system and the changes in the NE system evidenced in animal models of ADHD are discussed.   (Hirsch, 1998) (Kobayashi, 1995;,  ( Thomas, 1997) (Kobayashi, 1994) (Takahashi, 1997) (Xu, 2000) (Zwart, 2001) (Gogos, 1998) (Cases, 1998) (Carson, 2002) (Drouin, 2002) (Zuscik, 2000) Impaired response to noxious stimuli (Tanoue, 2002) a 2a-a 2c a 2c++ KO knockout, DA dopamine, NE norepinephrine -: no changes; >: increased; <: decreased (Schramm, 2001;Lahdesmaki, 2002; Davies 2003) (Sallinen, 1999) (Sallinen, 1999) Phox2a controls NEergic traits during development, possibly through an interaction with Mash-l, a gene transiently expressed in the central nervous system. In fact, the targeted mutation of Mash-1 also causes the absence of central and peripheral NE neurons (Hirseh et al., 1998). Knockout (KO) mice have been useful for dissecting the cascade that controls the NE phenotype.

Synthesis and clearance
Norepinephrine is synthesized from L-dihydroxyphenylalanine (L-DOPA), which is first converted to DA by the enzyme tyrosine hydroxylase (TH) and then to NE, through the action of the enzyme DA-beta-hydroxylase. Therefore, perturbations in TH (see below, TH KO) will affect the synthesis of both NE and DA. In fact, TH KO mice lack both DA and NE (Kobayashi et al., 1995;Zhou, 1995a). The animals survive embryogenesis but die at 3 to 4 weeks of age from severe hypoactivity and hypophagia. The synthesis of NE, however, can be normalized by the transgenic expression of TH under the control of the DA-b-hydroxylase promoter, thus generating pure DA KO mice. Such mice survive, displaying normal NE synthesis, but make no DA (Zhou, 1995a) and are severely hypoactive. Unfortunately, no data are available about their sensitivity to amphetamines.
Interestingly, the induction of a mutation in the TH gene, which leads to a 40% reduction in TH activity, leads to a lowered learning performance and memory formation. In these animals, spatial learning and LTP are normal, and the performance in behavioral tasks can be restored by stimulating NE activity (Kobayashi et al., 2000). Therefore, low levels of NE reduce motivation and performance in learning tasks.
On the other hand, DA beta-hydroxylase (DBH) KO mice, which selectively lack NE (Thomas & Palmiter, 1997), do not show alterations in locomotor activity or learning, although they swim more slowly and show more rapid extinction. These animals show characteristic somatic problems such as ptosis, reduced body mass, hypotension, increased embryonic mortality, and deficient thermogenesis (Carson & Roberson, 2002). Interestingly, DA levels are elevated in most tissues (Thomas et al., 1998). The mice are hypersensitive to the behavioral effects of amphetamine (Weinshenker et al., 2002) and susceptible to seizure-inducing stimuli (Szot et al., 1999).
Reboxetine and other antidepressants have no effect in DBH KO mice (Cryan et al., 2001), which confirms that all the indirect effects ofthese drugs necessitate the release of NE. Unfortunately, no data are available about the effectiveness of reboxetine or other NET blockers in DA KO mice, which could disentangle the NE-dependent effects from the DA-dependent effects.
Conversely, the hyperexpression of DBH in transgenic mice is not accompanied by changes in NE levels (Kobayashi et al., 1994). Therefore, it is possible that the hypolocomotor phenotype of TH KO mice is, rather, due to the changes in DA levels or to the lack of compensatory DA changes in DBH mice.
The vesicular monoamine transporter (VMAT2) stores monoamines, and particularly NE, from the neuronal cytoplasm into vesicles. VMAT2 KO mice die by the second week after birth and show increased mortality, possibly from arrhythmias (Takahashi et al., 1997). Heart rate, body temperature, NET, DA transporter (DAT), adrenergic, and DAergic receptors are normal, but DA levels are increased, as is the sensitivity to amphetamines.
Clearance for NE is accomplished by the NE reuptake system or NET, by monoamine oxidases (MAO-A and MAO-B), and by catechol-O-methyltransferase (COMT) enzymes. Although the NET is the predominant transporter in neuronal tissue, a second process termed uptake-2 is responsible for the removal of NE, accomplished by the protein extraneuronal monoamine transporter (EMT) or the organic cation transporter 3 (OCT-3 induce the compensatory hyperDAergic system, normal locomotor activity but hyperresponsivity to amphetamines. On the contrary, lesions that increase the intracellular NE level (NET KO mice) also decrease the DA level, accompanied by low locomotor activity and hypersensitivity to amphetamines. Finally, lesions that decrease both DA and NE levels (TH KO mice) reduce locomotor activity.
Thus, the NE and DA systems are regulated in a negative feedback manner. The levels of DA correlate with locomotor activity, whereas the responsivity to amphetamines is the outcome of a balance between the two systems. In fact, selective and extensive depletion of NE in neonatal rats by 6-OHDA in presence of a DAT inhibitor leads to increased expression of DA, motor hyperactivity, and distractability (Raskin et al., 1983;Carli et al., 1983).

Receptors
The NE receptors have been classified in three classes: alpha receptors, which are at postsynaptic sites, alpha2 receptors, both pre-and postsynaptic, and three different beta receptors, which are mainly postsynaptic but can also act presynaptically to facilitate NE release (O'Donnell, 1993;Murugaiah & O'Donnell, 1995). A fourth beta receptors isoform has been hypothesized recently (Oostendorp et al., 2000), The fourth group, beta receptors, represents an important interface between the sympathetic nervous system and the cardiovascular system (the so called betaadrenergic axis; see Naga Prasad et al., 2001). As far as beta receptor KO mice are concerned, they are not included in this review because most of their alterations pertain to the heart. Three a subtypes (ata, (l, lb, l, ld) and four a2receptor subtypes (tX2.d) are described. These are also expressed on glial cells, and their expression is different according to the brain region, a being expressed mainly in the ventral part of the brain, ab in the thalamus, amygdala and raphe nuclei and aid in the cortex and hippocampus (see also Tanoue et al., 2002). Here, synapses are present on both pyramidal and nonpyramidal neurons, on which the a2a receptor is the predominant NE receptor isoform (Aoki et al., 1998). Mice lacking the aab, ab, aid, a2a, a2b, a2e adrenergic receptor subtypes and a mouse line containing a point mutation in the a2a receptor have been developed.
Moreover, lines overexpressing the aa, ab, tXd receptors under the control of different promoters have been studied (Sallinen et al., 1999;Tanoue et al., 2002;Drouin et al., 2002). Only one subset of these lines has been behaviorally tested. None of the al subtypes is required for embryologic development, as these KO mice do not show gross abnormalities.
Mice KO for alb show increased reaction to novelty and novelty-induced hyperactivity (Spreng et al., 2001) although other reports failed to confirm this evidence (Drouin et al., 2002). Moreover, KO mice do not display changes in the DA system, but the levels of NE in the PFc are lower than those in wild-type mice. The locomotion-stimulating effects and the increase in DA levels induced by d-amphetamine-morphine are decreased in these animals (Drouin et al., 2002;Auclair et al., 2002). This reduction might depend on a lower efflux of DA after stimulation (Battaglia et al., 2003).
Conversely, mice overexpressing the tlb receptor have an age-dependent loss in horizontal deambulation and a reduced number of rearings, which can improve after treatment with adrenergic blockers or with L-DOPA (Zuscik et al., 2000). Moreover, this phenotype is accompanied by seizures and neurodegeneration, and NE plasma levels are reduced.
Mice lacking the tld receptor show an impaired response to noxious stimuli (Tanoue et al., 2002). Mice lacking the a2a receptor show normal locomotor activity (Schramm et al., 2001) but a reduced number of rearings in open field (Lahdesmaki et al., 2002).The resting levels of NE are normal, but the NE turnover is increased (Lahdesmaki et al., 2002), and the NE LC neurons are hypertrophic (Davies et al., 2003). Anxiety tests in these animals suggest that a2a is stress protecting (Schramm et al., 2001).
Targeted inactivation ofthe a2c receptor leads to normal or increased locomotor activity, and supersensitivity to amphetamine (Sallinen et al., 1999). Moreover, these mice exhibit lower DA turnover.
Conversely, transgenic mice overexpressing t2c receptors show normal or reduced activity and increased stress susceptibility. Anxiety tests in these animals suggest that 2e mediates stress susceptibility (Schramm et al., 2001).
Interestingly, methylphenidate, a drug that blocks the DA transporter and is widely used in ADHD pharmacotherapy, decreases the collected reward in an operant conditioning task in normal mice, but has opposite effect in 2 receptor KO mice (Ihalainen et al., 2001). This is in agreement with the increase in spontaneous locomotor activity induced by amphetamine in these animals.
It is plausible that the alteration in DA metabolism can explain the different effect of MPH (see also the companion paper by Viggiano et al., 2004 this issue).
In summary, both ab and 2 receptors control locomotor activity; KO mice for these receptors show increased activity, whereas transgenic overexpressing mice show decreased activity. Their influence on DA and NE levels is different. The alb KO mice do not show changes in DA level but have a lower NE level than the wild-type, whereas a2 KO mice show a lower DA turnover. The sensitivity to amphetamines is again reflected by the balance between DA and NE levels.

NOREPINEPHRINE SYSTEM CHANGES IN BEHAVIORAL MODELS OF ADHD
Many data are available about the NE system in an animal model of ADHD, the SHR. These mice are hyperactive and impulsive in comparison with their WKY controls. The evidence from LC neurochemistry does not show an alteration in NE levels (Kaehler et al., 2000), although the turnover can be elevated. The firing rate of these neurons is normal (Conti et al., 1997) or lower (Olpe et al., 1985). The TH levels are increased in the medulla (Kumai et al., 1996(Kumai et al., , 2003 and around vessels (Gradin et al., 2003) but decreased in the PFc (King et al., 2000). Recent data also report lower amounts of TH in a bounded postnatal period (Leo et al., 2003). The NE levels are normal in the solitary tract (Dev & Philip, 1996) but elevated in the median preoptic area (Tanaka et al., 2002), striatum and brainstem (Howes et al., 1984). The release of NE per varicosity might be increased (Jimenez-Altayo et al., 2003). The {X receptors in the cortex of SHR are increased (Hellstrand & Engel, 1980), a datum that contrasts with the hypoactivity of lb over-expression in mice. Some of these alterations might explain the elevated blood pressure in these .animals, but do not correlate with the hyperactivity (Reja et al., 2002).
The data presemed in Table 2 raise the question about why ADHD subjects respond to pharmacological agents that modify the NE system. In order to address this point, we will show data about the biochemical and behavioral effects of NET blockers and revise these effects from the view-point of an alteration in the attentional networks.  (Wong et al., 2000). In contrast, the affinity of the psychostimulant methylphenidate is higher for the DAT than for the NET (Bymaster et al., 2002). Reboxetine and atomoxetine exert a similar action on the levels of NE and DA in the PFc, but differ as regards the effects on 5-HT levels. In fact, both reboxetine and atomoxetine enhance DA output in the PFc, but not in the accumbens (Kuczenski & Segal, 1995;Wong et al., 2000). Possibly this effect is dependent on a blockade of DA reuptake by the NET in the PFc, a mechanism absent in the striatum, where NE terminals are few (Carboni et al., 2003). It is possible, however, that indirect mechanisms intervene, as 2-adrenergic antagonists also increase DA release in terminal fields (Shi et al., 2000;Linner et al., 2001).
The increased DA level is accompanied by an increased burst firing but not an average firing frequency of DA cells in the VTA (Linner et al., 2001). Similar changes might be responsible for altered reward perception, with a mechanism similar to methylphenidate (see also companion paper (Viggiano et al., 2004 this issue).
Similarly, the blockade of NET by these drugs increases the levels of NE (Kuczenski & Segal, 1995). High doses lead to a reduction of the firing rate of LC NE cells and 5-HT cells in a dosedependent relation (Wong et al., 2000), an effect similar to that of other stimulants (Aston-Jones & Bloom, 1981). Interestingly, a2-adrenergic antagonists and methylphenidate (Kuczenski & Segal, 2001) also enhance NE release (Thomas, 1991).
Similar studies with atomoxetine show that specific PFc modulators have a different sensibility to these drugs. In fact, NE is already increased at low doses (0.3 mg/kg), whereas DA is increased at medium doses (1 mg/kg) and 5-HT at high doses (3mg/kg) (Bymaster et al., 2002). The maximal release occurs hour after injection. Interestingly, the effects on DA are evident only in the PFc, as no changes can be detected in the striatum.

Behavior
Nowadays, low-doses of AMPH-Iike stimulants and methylphenidate are the main pharmacological treatment of ADHD. These drugs ameliorate the symptoms of inattentiveness, hyperactivity, and impulsivity. Unfortunately, about 10 to 30 percent of ADHD children are either non-responders or intolerant to psychostimulant therapy (Barkley, 1977), which may reflect the heterogeneous neural substrates of ADHD.
Therefore, the net effects were different if we consider the number of rearings, the horizontal exploratory activity, or the duration of rearing episodes. The number of rearing was decreased, but only in NHE and WKY rats, whereas SHR and NRB rats did not show any change. In SHR and WKY rats, reboxetine reduced horizontal activity and prolonged scanning time. In contrast, in NHE and NRB rats reboxetine increased rearing durations in the control line, whereas it decreased the rearing frequency in NHE rats. Similarly, atomoxetine did not modify the horizontal and vertical activity of NRB rats, whereas it reduced the number of rearings (-27%) and the traveled distance (-17%), without effects on scanning durations in NHE rats (see Table 3). The data suggest different activity of the NE system in the SHR versus NHE lines. In fact, SHR and NHE rats have different responses, reflecting different neural substrates. The differential sensitivity of these two lines to reboxetine could be due to a different involvement of the DA and NE systems in these animals. Although the effects of the NET blockade are similar in both the hyperactive SHR and the control rats, the main effects are evident in NHE rats, but not in their NRB controls.
The overall results would suggest that the NE system is similar in SHR and WKY rats, whereas differences exist between NHE and NRB rats. Alternatively, the interaction between the NE and DA systems is altered in NHE rats. These conclusions are in agreement with previous suggestions that an alteration of the NE system in adult SHR rats is indeed linked to the hypertensive phenotype but not to hyperactivity.
As both the NE and the DA systems are involved in arousal, attention, and cognitive functions, the drugs may act on two different pathways, thus changing the level of motivation (Wong et al., 2000). In fact, the LC-NE system is sensitive to novel environmental stimuli. An enhanced NE neurotransmission in a novel environment (see Lt maze) decreases attention to an individual object, increasing the scanning of the environment (Berridge & Waterhouse, 2003). Therefore, the lower frequency of rearings and the increase of their duration in NHE rats may indicate a lower firing rate of LC neurons, whereas the changes in SHR and WKY rats can be explained by the involvement of other systems, such as DA. The data lead to the suggestion that the positive effects of NEergic drugs may be due to an indirect action on the DA system. Nonetheless, further research is warranted by these findings to disentangle this intricate issue.