5-HT6 receptors and Alzheimer's disease

During the past 20 years, the 5-HT6 receptor has received increasing attention and become a promising target for improving cognition. Several studies with structurally different compounds have shown that not only antagonists but also 5-HT6 receptor agonists improve learning and memory in animal models. A large number of publications describing the development of ligands for this receptor have come to light, and it is now quite evident that 5-HT6 receptors have great pharmaceutical potential in terms of related patents. However, 5-HT6 receptor functionality is much more complex than initially defined. According to the existing data, different cellular pathways may be activated, depending on the drug being used. This article reviews preclinical and clinical evidence of the effects that 5-HT6 receptor compounds have on cognition. In addition, the biochemical and neurochemical mechanisms of action through which 5-HT6 receptor compounds can influence cognition will be described. Overall, several 5-HT6-targeted compounds can reasonably be regarded as powerful drug candidates for the treatment of Alzheimer's disease.

shares 89% homology with the human form [6,7]. Th e 5-HT 6 receptor belongs to the G-protein-coupled receptor (GPCR) family, displaying seven transmembrane domains. Th ey are quite diff erent from all other 5-HT receptors: they are characterized by a short, third cytoplasmatic loop and a long C-terminal tail and contain one intron located in the middle of the third cytoplasmatic loop. Th e 5-HT 6 receptor has no known functional isoforms. A non-functional truncated splice variant of the 5-HT 6 receptor has been identifi ed but appears not to have any physiological signifi cance. Kohen and colleagues [6] identi fi ed a silent polymorphism at base pair 267 (C267T). Although there is evidence linking this poly mor phism to several syndromes that aff ect cognition, including dementia, AD, and schizophrenia, these fi nd ings have not always been replicated and their signi fi cance has not yet been determined. 5-HT 6 receptor expression is restricted mainly within the central nervous system (CNS). In situ hybridization and northern blot studies revealed an exclusive distribution of 5-HT 6 mRNA in the rat CNS, and the highest density was found in the olfactory tubercle, followed by the frontal and entorhinal cortices, dorsal hippocampus (that is, dentate gyrus and CA1, CA2, and CA3 regions), nucleus accumbens, and striatum. Lower levels were observed in the hypothalamus, amygdala, substantia nigra, and several diencephalic nuclei. Th ese fi ndings have been corroborated by immunolocalization and radioligandbinding studies, which showed a similar distri bution of 5-HT 6 receptor protein in the rat CNS [8,9]. Th erefore, 5-HT 6 receptors appear to be localized in brain areas involved in learning and memory processes.

5-HT 6 receptor signaling
Interestingly, it has been suggested that both 5-HT 6 receptor agonists and antagonists may have pro-cognitive activities, implying that activation and inhibition of this receptor could evoke similar responses. Th e selective 5-HT 6 receptor agonist LY-586713 caused a bell-shaped dose-response curve on hippocampal brain-derived neurotrophic factor (BDNF) mRNA expression. It also increased the Arc mRNA levels, and this eff ect was blocked by the 5-HT 6 receptor antagonist SB-271046. However, in some brain regions, the antagonist was not able to block the agonist eff ect and, in fact, induced an increase in Arc expression [10], consistent with a potential diff erential mechanism. An excellent review [11] regarding the eff ects of 5-HT 6 receptor agonists and antagonists on cognition in normal adult rats and in rodent models of psychiatric disorders, as well as data obtained from some clinical studies, suggested that agonists and antagonists are able to act on receptors located on distinct neuronal populations.
Th e mechanism for paradoxically similar eff ects of agonist/antagonists on cognition could be related to the existence of alternative biochemical pathways activated by 5-HT 6 receptors. Th e 5-HT 6 receptor is a GPCR that positively stimulates adenylate cyclase activity, meaning that, upon agonist activation, cAMP formation is increased. In fact, activity on adenylate cyclase confers the classic defi nition as agonist/antagonist on 5-HT 6 receptors. 5-HT 6 receptor coupling to Gαs has been widely described, but coupling of 5-HT 6 receptors to other Gα protein subunits (Gα i/o or Gα q ), using a scintillation proximity assay/antibody-immunocapture technique, has also been recently reported [12]. In addition, the coupling of 5-HT 6 receptors to Ca 2+ signaling by using a chimeric G-protein has been reported [13]. It has been reported that, with a yeast two-hybrid assay, the carboxyl-terminal region of the 5-HT 6 receptor interacts with the Fyn-tyrosine kinase, a member of the Src family of non-receptor protein-tyrosine kinases [14]. Th is same study showed that the activation of 5-HT 6 receptor activated the extracellular signal-regulated kinase1/2 via a Fyn-dependent pathway. Th ese fi ndings suggest that Fyn plays an important role in 5-HT 6 receptor-mediated signaling pathways in the CNS. In addition, improvement in learning, associated with the administration of the 5-HT 6 receptor antagonist SB-271046 in the Morris water maze learning task, is associated with increased phosphorextracellular signal-regulated kinase1/2 (pERK1/2) levels [15]. All of these data suggest that 5-HT 6 receptors activate the ERK1/2 via a Fyn-dependent pathway (Figure 2). At this point, it is worth mentioning a purported relationship between Fyn and Tau. Tau is a microtubuleassociated protein and, in a hyperphos phory lated state, a main component of neurofi brillary tangles, one of the pathologic hallmarks of AD. Most of the Tau phosphorylation sites that have been routinely characterized are serine and threonine residues, but recent reports state that Tau can be phosphorylated at tyrosine residues by kinases, including Fyn. In addition, pERK1 is one of the kinases involved in Tau phos phory lation. Th erefore, it is possible to suggest that modulation of 5HT 6 receptors might lead to increased tau phos phory lation. In other words, it is even possible to speculate that 5HT 6 receptor modulation might, in the short term, improve cognitive function (as described in the following sections) but, over a longer term, enhance the neurodegenerative processes in AD. A physical inter action between 5-HT 6 receptor and the Jun activation domain-binding protein-1 (Jab-1), using diff erent experi mental approaches, has also been described, suggesting another signal transduction pathway for these receptors [16].
However, it should be noticed that drugs that are considered to be a reference agonist/antagonist upon 6-HT 6 receptors might be regulating 5-HT 6 receptor-inde pendent events. In an investigation of the eff ects of EMD386088, a 5-HT 6 receptor agonist, on cell viability, it was found that EMD386088 potentiated cell death in diff erent cultured neuronal cell lines and that these cytotoxic eff ects, regardless of the presence of 5-HT 6 recep tors, were mediated by the downregulation of ERK1/2 activities. Furthermore, the specifi c 5-HT 6 receptor antagonist SB258585 potentiated cell death and induced an increase in the concentration of intracellular Ca 2+ , whereas EMD386088, or 5-HT, did not aff ect calcium concentration [17]. Th erefore, these compounds that have been intensively used as 5-HT 6 receptor ligands could display 5-HT 6 receptor-independent eff ects.

Neurochemical mechanisms mediating 5-HT 6 receptor functions
A postsynaptic location of 5-HT 6 receptors is expected because quantitative reverse transcription-polymerase chain reaction distribution of serotonin 5-HT 6 receptor mRNA in the CNS of rats subjected to a selective serotonergic lesion using 5,7-dihydroxytryptamine has shown that 5-HT 6 receptors are present within 5-HT projection fi elds and not in serotonergic raphe neurons [18]. Th erefore, 5-HT 6 receptors appear to be located in neurons that are not serotonergic.
It has been consistently described that the infl uence of 5-HT 6 receptors on memory is mediated, at least partially, by increased cholinergic neurotransmission. Behavioral studies have shown that 5-HT 6 receptor blockade leads to an increase in behaviors such as the number of yawns or stretches in rats. Th ese behaviors are largely dependent on the cholinergic system because they are reversed by muscarinic antagonists. Further supporting this cholinergic mediation, 5-HT 6 receptor antagonists increase acetylcholine release both in vitro [19] and in vivo [20].
However, the purported localization of 5-HT 6 receptors on cholinergic neurons was discarded because a selective cholinergic lesion, induced by injection of the selective immunotoxin 192-IgG-Saporin, failed to alter the density of 5-HT 6 receptor mRNA or protein expression in the deaff erentated frontal cortex [19]. Th erefore, the eff ects of 5-HT 6 receptor ligands on cholinergic neurons could be mediated by other neurotransmitter systems, such as the glutamatergic system [21]. Treatment with a 5-HT 6 receptor antagonist or atypical anti-psychotics with high affi nities for 5-HT 6 receptors, such as clozapine, enhanced glutamate levels in the frontal cortex and hippocampus. On the other hand, 5-HT 6 receptor ago nism attenuated stimulated glutamate levels elicited by high KCl treatment [22]. A recent work aimed to study the eff ect of 5-HT 6 receptor activation on glutamatergic transmission by means of whole-cell patch-clamp electrophysiological recordings from medium spiny neurons of the striatum and layer V pyramidal neurons of the prefrontal cortex. 5-HT 6 receptor activation by the novel agonist ST1936 reduced the frequency of spontaneous excitatory postsynaptic currents. 5-HT 6 receptor activation also reduced the amplitude of spontaneous excitatory postsynaptic currents recorded from medium spiny neurons, suggesting a mechanism of action involving postsynaptic 5-HT 6 receptors. Th e inhibitory eff ect of ST1936 on glutamatergic transmission was prevented by the selective 5-HT 6 receptor antagonist SB258585 [23].
It has also been shown that 5-HT 6 receptors may be expressed on GABAergic spiny neurons of the striatum. Th e co-localization of glutamic acid decarboxylase and 5-HT 6 receptors in rat cerebral cortex and hippocampus has also been demonstrated, and almost 20% of 5-HT 6like immunoreactive neurons have been shown to be GABAergic [24]. It can be suggested, on the basis of all these data regarding localization of 5-HT 6 receptors and on the basis of data from releasing experiments [22,25], that 5-HT 6 receptor agonists/antagonists modulate cholinergic or glutamatergic systems (or both) via disinhi bition of GABAergic neurons.

5-HT 6 receptor ligands
Since the initial discovery of the fi rst ligands in the late 1990s by using high-throughput screening technologies on compound libraries, a growing number of scientifi c publications and patent applications have developed [26]. Th e synthesis of 5-HT 6 receptor ligands has been very successful, and a number of highly potent ligands have been reported [27].
At the preclinical level, 5-HT 6 receptor medicinal chemistry is benefi ting from knowledge that has been acquired since the discovery of the receptor using tools such as pharmacophore modeling, three-dimensional molecular docking or structure similarity algorithms. As a result, an increasing number and diversity of novel, highly selective 5-HT 6 receptor ligands of all functional types have been reported, although the principal eff orts have been focused on antagonism. Some of these compounds have been used extensively as pharmacological tools (that is, Ro-04-6790 or SB-271046). Th e search for new 5-HT 6 receptor ligands continues. A new 5-HT 6 receptor agonist, ST1936, was recently reported. ST1936 bound to human 5-HT 6 receptors with good affi nity (K i = 28.8 nM) and behaved as a full 5-HT 6 agonist on cloned cells; it was able to increase Ca 2+ concentration and phosphorylation of Fyn kinase and regulate the activation of ERK1/2 (downstream target of Fyn kinase). Th ese eff ects were completely antagonized by 5-HT 6 receptor blockade with selective antagonists [28]. Epiminocyclohepta[b]indole analogs [29], tetracyclic tryptamines with the rigidized N-arylsulphonyl, N-arylcarbonyl, and Nbenzyl substituents [30], or conformationally restricted N(1)-arylsulfonyl-3-aminoalkoxy indoles [31] have been shown to have acceptable ADME (absorption, distri bution, metabolism, and excretion) properties, adequate brain penetration, and favorable pharmaco kinetic profi le. Using a newly devised chemocentric informatics methodo logy for drug discovery integration showed that selective estrogen receptor modulators are putative ligands of 5-HT 6 receptors [32]. Positive results in animal models of cognition have been reported for both lead compounds (that is, L-483518, Ro-4368144, BGC20-761, or E-6801) and newly synthesized ligands, further confi rming the involvement of this receptor in cognitive processes and its therapeutic potential. How ever, despite encouraging results at a preclinical level, very few 5-HT 6 receptor selective ligands (and all of them acting as antagonists) have reached the clinical phases of development for the treatment of cognitive disorders.
Th e development of a positron emission tomography (PET) radioligand for imaging 5-HT 6 receptors in the brain would, for the fi rst time, enable in vivo imaging of this target along with assessment of its involvement in disease pathophysiology. Based on the aforementioned, the develop ment of N-[3,5-dichloro-2-(methoxy)phenyl]-4-(methoxy)-3-(1-piperazinyl)benzenesulfonamide (SB399885), a selective and high-affi nity (pK i = 9.11) 5-HT 6 antagonist radiolabeled with carbon-11 by O-methylation of the corresponding desmethyl analog with [ 11 C] MeOT, has been described. PET studies with [ 11 C] SB399885 in baboons showed fast uptake followed by rapid clearance in the brain. Poor brain entry and inconsistent brain uptake of [ 11 C]SB399885 compared with known 5-HT 6 receptor distribution limit its useful ness [33]. Recently, the development of GSK215083 (Glaxo-SmithKline, Uxbridge, Middlesex, UK) has been reported. Th is compound was radiolabeled with 11 C via methy lation. Th e in vivo properties of 11 C-GSK215083 have been evaluated in pigs, non-human primates, and human subjects. 11 C-GSK215083 readily entered the brain in all three species, leading to a heterogeneous distribution (striatum > cortex > cerebellum) that is consis tent with reported 5-HT 6 receptor densities and distribution deter mined by tissue-section autoradio graphy in preclinical species and humans [34].

Experimental approaches to the role of 5-HT 6 receptors in cognition
Following the discovery of 5-HT 6 receptor ligands with good brain penetration, a growing body of preclinical evidence has supported the use of 5-HT6 receptor antagonism for treating cognitive dysfunction. In two excellent reviews, Meneses and colleagues [4] (2011) and Fone [11] (2008) described the eff ects of 5-HT 6 receptor agonists and antagonists on cognition. Th e fi rst indirect evidence of 5-HT 6 receptor involvement in memory was obtained by using antisense oligonucleotides. A few years later, pharmacological blockade of 5-HT 6 receptor was shown to produce promnesic or antiamnesic eff ects (or both) in a number of memory tasks, including water maze, passive avoidance, autoshaping, fear conditioning, novel object recognition, or social memory [35]. Further support came from studies based on how learning paradigms decrease 5-HT 6 receptor expression [15,36], whereas 5-HT 6 receptor overexpression of 5-HT 6 recep tors in the striatum, achieved by targeted gene delivery, led to cognition impairments in a reward-based instru mental learning task, a striatum-dependent learning model [37]. Diff erent 5-HT 6 receptor antagonists have been reported to be active in the novel object discrimination test in rats and to improve water maze retention, even in aged rats [38], although failing to alter acquisition of spatial learning. In senescent mice, the eff ects of 5-HT 6 receptor blockade with SB-271046 were assessed in the novel object recognition test for evaluating recognition memory (a component of episodic-like memory) and in spon taneous alternation task in the T-maze for evaluating working memory. It was found that defi cits in consolidation of both non-spatial recognition memory and working memory performances were reversed by 5-HT 6 receptor blockade [39].
One of the more consistent fi ndings regarding the involvement of 5-HT 6 receptors in memory is the ability of the 5-HT 6 receptor antagonist to reverse a scopolamine-induced cognitive defi cit in the Morris water maze or novel object recognition test [40]. Th is fi nding would be in line with the hypothesis that 5-HT 6 receptor functions are mediated, at least partially, by a modulation of the cholinergic neurotransmission. In an extensive study regarding the eff ects of the 5-HT6 receptor antagonist SB-271046 in mice presenting a scopolamineinduced cholinergic disruption of memory, it was found that SB-271046 was able to reverse the scopolamineinduced defi cits in working memory and to reverse the defi cits of acquisition and retrieval of aversive learning, whereas scopolamine-induced defi cits in episodic-like memory (acquisition and retrieval) were partially counteracted by 5-HT 6 receptor blockade. However, SB-271046 alone failed to aff ect working memory, recognition memory, and aversive learning performances [39], but it appears that 5-HT 6 receptor blockade is more consistently eff ective in alleviating memory defi cits than increasing memory in normally functioning animals [41]. Interestingly, a combined treatment of SB-271046 with an acetylcholinesterase inhibitor produced an additive increase in passive avoidance and signifi cantly reversed scopolamine-induced amnesic eff ects [41]. Similarly, this combined administration of subthreshold doses of two novel selective 5-HT 6 antagonists, compounds CMP X and CMP Y, with the acetylcholinesterase inhibitor donepezil (Aricept®; Eisai, Tokyo, Japan) (approved for symptomatic treatment of AD) enhanced memory perfor mance in young Wistar rats with cognitive defi cits induced by scopolamine [40]. Th is suggests that the administration of 5-HT 6 receptor antagonists with acetylcholinesterase inhibitors has potentially additiveenhancing eff ects on cognition.
Lu AE58054, a 5-HT 6 receptor antagonist, reversed cognitive impairment induced by subchronic phencycli dine in a novel object recognition test in rats [42]. Ro 04-6790 also reversed impairment in learning consolidation produced by the NMDA receptor antagonist MK-801, and the antagonist PRX-07034 restored the impairment of novel object recognition in the social isolation rearing model, both of which showed behavioral changes that resemble the core defects observed in schizophrenia [11]. SB271046 has also been shown to reverse memory disturbances in experimental models of stress-related psychiatric disorders that have been associated with an impairment of the hypothalamic-pituitary-adrenal axis reactivity [43].
In contrast to the works cited above, those by Russell and Dias [44] and Lindner and colleagues [45] failed to detect any eff ects of Ro 04-6790 or SB-271046 upon acquisition of an autoshaping task, scopolamine-induced defi cits in contextual fear conditioning, or retention of a water maze task. In the same way, two selective 5-HT 6 receptor antagonists, Ro-4368554 and SB-258585, showed diff erential eff ects on cognition, depending on the paradigm that was used [46]. Both compounds showed cognition-enhancing eff ects in object recogni tion, whereas only SB-258585 was able to prevent the scopolamine-induced defi cit in the Morris water maze test. Neither Ro-4368554 nor SB-258585 prevented scopolamine-induced impairment in contextual fear conditioning. Similarly, both compounds were ineff ective on MK-801-induced defi cits in contextual fear condition ing and spatial working memory. In addition, Fone [11], Kendall and colleagues [47], and Meneses and colleagues [4] reported that selective 5-HT 6 receptor agonists appear to restore memory impairments in the novel object discrimi nation paradigm. More intriguing were the results obtained when combining non-active doses of the 5-HT 6 receptor agonist E-6801 and the 5-HT 6 receptor antagonist SB-271046, which produced an im prove ment in novel object discrimination. In addition, E-6801, alone and at a non-active dose, was able to syner gistically improve the activity of non-active doses of donezepil (an acetylcholinesterase inhibitor) and meman tine (an NMDA receptor antagonist) [47]. Th us, both 5-HT 6 receptor agonist and antagonist compounds show pro-cognitive activity in preclinical studies, although the explanation for their paradoxically analogous eff ect is still not clear.

5-HT 6 receptors and Alzheimer's disease
Signifi cant reductions in 5-HT 6 receptor density in cortical areas of patients with AD have been found, although the reductions in 5-HT 6 receptor density were unrelated to cognitive status before death [48]. Since 5-HT 6 receptor blockade induces acetylcholine release, reductions in 5-HT 6 receptors may represent an eff ort to restore acetylcholine levels in a deteriorated cholinergic system. In addition, it has been reported that a dysregulation of 5-HT 6 receptor activation by 5-HT in the temporal cortex may be related to behavioral symptoms in AD [49]. In this sense, preclinical data suggest a possible role for 5-HT 6 receptors in depression and anxiety. Two selective 5-HT 6 antagonists (SB-399885 and SB-271046) and donepezil (an acetylcholinesterase inhibi tor) were evaluated in the rat forced swimming test because this test is known to identify drugs with antidepressant activity. Systemic administration of the 5-HT 6 receptor antagonist produced a signifi cant reduction in the immobility time in the rat forced swimming test, with a similar profi le in terms of 5-HT 6 receptor occupancy, measured by binding assay. Th ese data suggest that 5-HT 6 antagonists, at doses corresponding to those that occupy central 5-HT 6 receptors, could have an antidepressive eff ect in humans. Th is may diff erentiate 5-HT 6 antagonists from acetylcholinesterase inhibitors with respect to mood control in the symptomatic treatment of AD [50]. However, once again, the results of pharma cological studies are equivocal since both blockade and stimulation of 5-HT 6 receptors may evoke antidepressantand anxiolytic-like eff ects.
A number of 5-HT 6 receptor antagonists have successfully undergone phase I clinical studies (healthy volunteers) and some have been evaluated in clinical phase II studies (patients) for the treatment of AD [51]. Two of these compounds appear to be showing positive results. Two phase II trials using SB-742457 (GlaxoSmithKline) have recently been completed in subjects with mild-tomoderate AD. Th e fi rst was a dose-ranging trial comparing SB-742457 with placebo, and the second was an exploratory study with SB-742457 and donepezil arms. Overall, these studies demonstrated that SB-742457 is well tolerated in patients with AD. SB-742457 produced an improvement in both cognitive and global function in AD as assessed by ADAS-cog (Alzheimer's Disease Assessment Scale-cognitive sub scale) and CIBIC+ (Clinician's Interview-Based Impres sion of Change-plus Caregiver Input), respectively [52]. Other clinical phase II studies are being performed, either alone or as add-on therapy with the acetylcholine esterase inhibitor, donepezil. Th is is the case for Lu-AE-58054 (SGS-518; Lundbeck, Copenhagen, Denmark) or PF-05212365 (SAM-531; Pfi zer Inc, New York, NY, USA). Other compounds that are in diff erent phases of clinical trials are SUVN-502 (Suven Life Sciences Ltd., Hyderabad, India) or AVN-322 (Avineuro Pharma ceuticals, San Diego, CA, USA) or PRX-07034 (Epix Pharmaceuticals, Lexington, MA, USA). In any case, treatment with 5-HT6 receptor antagonists provides symptomatic treatment that might improve cognition, perhaps via modulating neurotransmitter-related mechanisms.
Besides these selective compounds, dimebon (latrepirdine, also known as dimebolin), originally developed as an antihistamine drug, is worth mentioning. Th is compound shows good affi nity for 5-HT 6 receptors (k i = 34 nM). Dimebon received widespread publicity as a potential therapy for AD following a very positive phase 2 study [53]. However, a more recent multinational phase 3 study showed no improvements [54].

Concluding remarks
Since the discovery of 5-HT 6 receptor in 1993 and subsequent development of selective antagonists, a growing number of studies support the use of serotonin 5-HT 6 receptor antagonism as a promising mechanism for treating cognitive dysfunction. Over the past 20 years, several studies with structurally diff erent compounds have shown that not only antagonists but also 5-HT 6 receptor agonists improve learning and memory in animal models. In addition, the potential therapeutic use of 5-HT 6 receptor ligands in mood disorders associated with AD, such as anxiety, depression, or schizophrenia, has been reported. Th erefore, ligands acting on 5-HT 6 receptors are attracting attention as potential candidates for the treatment of AD. However, the full charac terization of the functional profi le of the 5-HT 6 receptor is still pending.
Currently, 5-HT 6 receptors have obvious pharma ceutical potential in terms of related patents. Several 5-HT 6targeted compounds, mainly antagonists, are regarded as powerful drug candidates for the treatment of a range of neuropathological disorders, including AD [26]. However, the failure of compounds such as dimebolin points to the hypothesis that the crucial point regarding compounds acting on 5-HT 6 receptors is the intracellular pathways activated after the interaction of the compound with the receptor. Th erefore, perhaps it is a question not only of developing an agonist or antagonist with good affi nity but also of developing compounds able to activate the necessary mechanisms for the pro-cognitive eff ects. It is expected that, in the near future, the drug discovery process will benefi t from the complexity of functional responses associated with 5-HT 6 receptors and that new molecules will enter in the scenario of treating AD.

Competing interests
The author declares that she has no competing interests.