|Home | About | Journals | Submit | Contact Us | Français|
More than 1 billion people around the world smoke, with 10 million cigarettes sold every minute. Cigarettes contain thousands of harmful chemicals including the psychoactive compound, nicotine. Nicotine addiction is initiated by the binding of nicotine to nicotinic acetylcholine receptors, ligand-gated cation channels activated by the endogenous neurotransmitter, acetylcholine. These receptors serve as prototypes for all ligand-gated ion channels and have been extensively studied in an attempt to elucidate their role in nicotine addiction. Many of these studies have focused on heteromeric nicotinic acetylcholine receptors containing α4 and β2 subunits and homomeric nicotinic acetylcholine receptors containing the α7 subunit, two of the most abundant subtypes expressed in the brain. Recently however, a series of linkage analyses, candidate-gene analyses and genome-wide association studies have brought attention to three other members of the nicotinic acetylcholine receptor family: the α5, α3 and β4 subunits. The genes encoding these subunits lie in a genomic cluster that contains variants associated with increased risk for several diseases including nicotine dependence and lung cancer. The underlying mechanisms for these associations have not yet been elucidated but decades of research on the nicotinic receptor gene family as well as emerging data provide insight on how these receptors may function in pathological states. Here, we review this body of work, focusing on the clustered nicotinic acetylcholine receptor genes and evaluating their role in nicotine addiction and lung cancer.
The molecular cloning of nicotinic acetylcholine receptors (nAChRs) from brain cDNA libraries in the mid-1980s was a watershed event as it opened the window to not only a molecular understanding of cholinergic signaling within the nervous system, but also to a structural understanding of how members of the Cys-loop superfamily of ligand-gated ion channels function. In addition to neuronal and muscle nAChRs, members of this family include the ionotropic receptors for glycine, 5-hydroxytryptamine (5-HT3), and γ-aminobutyric acid (GABAA and GABAC) (Dani and Bertrand, 2007; Le Novère and Changeux, 1995). As the cloning frenzy subsided, a tremendous amount of effort was put forth to understand the pharmacological and biophysical diversity of nAChRs. As a result, nAChRs are among the most well understood allosteric membrane proteins from a structural and functional point of view (Albuquerque et al., 2009). Furthermore, nAChRs have emerged as key therapeutic targets for a variety of pathologies including schizophrenia, depression, attention deficit hyperactivity disorder, Alzheimer’s disease and of course, tobacco addiction (Arneric et al., 2007; Levin and Rezvani, 2007; Romanelli et al., 2007; Taly et al., 2009). More recently, genetic studies have identified single nucleotide polymorphisms (SNPs) in the chromosomal locus encoding three nAChR genes as risk factors for 1) nicotine dependence 2) lung cancer, 3) chronic obstructive pulmonary disease, 4) alcoholism and 5) peripheral arterial disease (Amos et al., 2008; Berrettini et al., 2008; Bierut et al., 2008; Caporaso et al., 2009; Freathy et al., 2009; Hung et al., 2008; Pillai et al., 2009; Saccone et al., 2009b; Saccone et al., 2007; Sasaki et al., 2009; Schlaepfer et al., 2008; Stevens et al., 2008; Thorgeirsson et al., 2008; Weiss et al., 2008). This cluster of nAChR genes encodes the α3, α5 and β4 subunits (designated CHRNA5/A3/B4) (Boulter et al., 1990). As a result of these genetic studies, new attention has been brought to bear on the CHRNA5/A3/B4 genes and their encoded subunits. This review summarizes recent work on the clustered nAChR subunits highlighting their structure, function, and expression in a variety of pathological conditions.
Signaling through neuronal nAChRs underlies several fundamental biological processes both during development and in the adult (Albuquerque et al., 2009). In the central nervous system (CNS), presynaptic nAChRs modulate release of most classical neurotransmitters including norepinephrine, acetylcholine (ACh), glutamate and GABA (McGehee et al., 1995). Postsynaptic nAChRs are intimately involved in fast ACh-mediated synaptic transmission in addition to activity-dependent gene expression, which is critical for synaptic plasticity (Albuquerque et al., 2009; Dani and Bertrand, 2007; Hu et al., 2002; Ji et al., 2001). Within the peripheral nervous system (PNS), nAChRs mediate fast excitatory transmission in most, if not all, autonomic ganglia and are involved in modulating visceral and somatic sensory transmission (Boyd et al., 1991; Genzen et al., 2001; Hu and Li, 1997; Steen and Reeh, 1993; Sucher et al., 1990; Wang et al., 2002b). More recently, numerous studies have revealed the expression of nAChRs on non-neuronal cells such as lung, glia, keratinocytes, endothelial cells, as well as cells of the digestive and immune systems, and evidence is accumulating indicating that the receptors play crucial roles in signal transduction underlying many physiological processes outside the nervous system (Arredondo et al., 2001; Battaglioli et al., 1998; Gahring et al., 2004b; Gahring and Rogers, 2006; Kawashima and Fujii, 2003; Macklin et al., 1998; Maus et al., 1998; Nguyen et al., 2000; Spindel, 2003; Wang et al., 2001; Wessler and Kirkpatrick, 2008). The importance of nAChR-mediated signaling is reflected in the many pathologies in which cholinergic signal transduction is compromised. For example, significant alterations in nAChR expression and function have been documented in several diseases such as Alzheimer’s disease, autosomal dominant nocturnal frontal lobe epilepsy, megacystis-microcolon-intestinal hypoperistalsis syndrome, Parkinson’s disease, schizophrenia, and Tourette’s disease (De Fusco et al., 2000; Isacson et al., 2002; Lena and Changeux, 1997; Perl et al., 2003; Perry et al., 2001; Quik, 2004; Quik et al., 2007; Richardson et al., 2001; Silver et al., 2001; Steinlein et al., 1995; Teaktong et al., 2003; Whitehouse et al., 1988; Zanardi et al., 2002). In addition, nAChRs are key players in the initial steps and subsequent downstream health consequences of nicotine addiction (Kedmi et al., 2004; Laviolette and van der Kooy, 2004). Coupled to this is a growing awareness that nAChRs may directly contribute to the pathogenesis of lung cancer (Catassi et al., 2008; Egleton et al., 2008; Schuller, 2008, 2009; Song et al., 2008).
Neuronal nAChRs are transmembrane proteins that form pentameric structures assembled from a family of subunits that include α2-α10 and β2-β4 (Brejc et al., 2001; Cooper et al., 1991). Each nAChR subunit consists of an approximately 200-residue extracellular N-terminus, four transmembrane segments (designated M1–M4), a variable intracellular loop (100–200 residues) between M3 and M4, and an extracellular C-terminus (4–28-residue, Fig. 1A) (Corringer et al., 2000). The N-terminus contains the ACh-binding domain (Eisele et al., 1993) with the interface between adjacent subunits forming a hydrophobic pocket that serves as the agonist-binding site. The M2 transmembrane segment of all 5 subunits forms the conducting pore of the channel, with regions in the M1–M2 intracellular loop contributing to cation permeability and agonist binding affinities (Corringer et al., 2000; McGehee and Role, 1995). When activated by an agonist in native or heterologous expression systems, nAChRs exhibit a rapid (500 ms to peak current, [ACh] = EC50) inward current that desensitizes (Barnard et al., 1982) and is potentiated by calcium ions (Vernino et al., 1992). The functional diversity exhibited by the neuronal nAChR family is a consequence, in large part, of the differential expression of the various subunit genes leading to the incorporation of distinct subunits into mature receptors.
Much of what is known about the biophysical and pharmacological properties of nAChRs is based on studies in heterologous expression systems (McGehee and Role, 1995). These systems make use of nAChR mRNA or cRNA injected into Xenopus oocytes as well as nAChR cDNA transfected into mammalian cell lines in order to express nAChR subunits singly or in combination. The propensity for nAChR subunits to form homomeric subtypes was determined by expressing subunits singly. When expressed alone, α7, α8, α9, and α10 are able to form functional receptors (Couturier et al., 1990a; Gerzanich et al., 1994). In contrast, other α subunits require the presence of β subunits to form functional receptors (Fig. 1B).
Of the clustered receptor subunit genes, co-expression of α3 and β4 nicotinic receptor subunit cRNA results in functional receptors with a single channel conductance, γ, of ~24 pS (McGehee and Role, 1995; Nelson and Lindstrom, 1999). ACh and nicotine are less potent agonists of α3β4 nAChRs compared to the prototypic high affinity α4β2 nAChRs (Table 2). In addition, α3 can coassemble with β2 nAChR subunits and form a functional receptor (Table 2) (Boulter et al., 1987; Wada et al., 1988). Unlike α3, the α5 subunit fails to form functional receptors when co-expressed with either β4 or β2 subunits (Boulter et al., 1987; Couturier et al., 1990b). Although the α5 subunit harbors the characteristic viscinal cysteines in the large extracellular domain, a tyrosine residue (Tyr198) required for ligand binding has been substituted with an aspartic acid residue, likely explaining the lack of functionality of this subunit when expressed with either β subunit alone (Albuquerque et al., 2009; Karlin, 2002). Thus, for several years after its initial cloning, the impact of the α5 subunit on nAChR function was poorly understood. This mystery was solved when α5 was co-expressed with both α4 and β2 subunits (Ramirez-Latorre et al., 1996). ACh-evoked currents from Xenopus oocytes expressing the three subunits were distinct from ACh-evoked currents in oocytes expressing only α4 and β2 subunits. In particular, expression of α5 dramatically shifted the ACh concentration response curve to the right and significantly increased the predominant single channel conductance (α4β2 = 24 pS, α4α5β2 = 44 pS at −100 mV). Finally, cysteine mutagenesis of amino acid residues in the M2 domain of the α5 nAChR subunit conferred MTSET sensitivity to α4β2* nAChRs indicating participation of the α5 subunit M2 domain with the pore of assembled receptors (“*” denotes containing, (Lukas et al., 1999). Later studies analyzed the impact of α5 expression on α3β2 and α3β4 nAChR function. In contrast to its effects on the pharmacology of α4β2 nAChRs, α5 expression increased the sensitivity of α3β2 nAChRs to nicotine and ACh (Table 2)(Wang et al., 1996). Surprisingly, the ACh/nicotine concentration-response relationship between α3β4 and α3α5β4 nAChRs was found to not dramatically differ (Wang et al., 1996). However, α5 subunit expression did accelerate the rate of desensitization of these receptors. In addition, expression of the α5 subunit reduced channel sensitivity to agonist in α3β4 nAChRs with engineered point mutations in the M2 domain that renders receptors hypersensitive to agonist definitively illustrating that α5 nAChR subunits can co-assemble with α3β4 nAChRs (Groot-Kormelink et al., 2001). Finally, coassembly of the α5 subunit with α3β2 or α3β4 nAChRs increases the calcium permeability of the resulting receptors (Gerzanich et al., 1998) indicating that these receptors could play significant roles in the initiation of ACh-induced signaling cascades under normal and pathological condition. In addition to uncovering a functional role of the α5 nAChR subunit, these studies demonstrated that a portion of functional nAChRs in native tissue likely consist of heteropentamers containing three or more distinct subunits.
The CHRNA5/A3/B4 nAChR subunit genes are found in a tight cluster in chromosomal region 15q24–25 (Fig. 1C) (Boulter et al., 1990). Admixtures of the nAChR subunits encoded by this locus form the predominant nicotinic receptor subtypes expressed in the PNS (Conroy and Berg, 1995; Covernton et al., 1994; Flores et al., 1996; Rust et al., 1994; Vernallis et al., 1993) as well as at key sites in the CNS such as the medial-habenula (Gotti et al., 2007; Grady et al., 2009). To determine the function of the clustered nAChR subunits, knockout (KO) mice have been generated. Mice that do not express the α3 subunit usually die within a week of birth due to multi-organ dysfunction (Xu et al., 1999a). α3 KO mice develop enlarged bladders causing bladder infection, dribbling urination, and urinary stones – a phenotype resembling that of a rare human condition called megacystis-microcolon-intestinal hypoperistalsis syndrome (Xu et al., 1999a). Consistently, patients with this disease do not appear to express α3 mRNA (Richardson et al., 2001). α3 KO mice also display extreme pupil dilation and lack of pupil contraction in response to light and have retinal wave activity with altered spatiotemporal properties delaying the refinement of retinal ganglion cell dendrites (Bansal et al., 2000; Xu et al., 1999a). Bladder contraction in response to nicotine is also lost. In addition, electrophysiological characterization shows that nicotine-induced whole-cell currents are abolished in the superior cervical ganglion (SCG) of α3 KO mice.
In contrast to the α3 KO mice, α5 and β4 KO mice are both viable and lack any gross abnormalities (Wang et al., 2003; Wang et al., 2002a; Xu et al., 1999a). α5 KO mice do exhibit abnormal cardiac parasympathetic ganglionic transmission and are less sensitive to acute nicotine treatment. Loss of α5 selectively affects axonal nAChRs in the SCG while leaving somatodendritic receptors unaffected (Fischer et al., 2005). Similarly, ganglionic transmission is impaired in β4 KO mice, attenuating ileal and bladder contractile responses to nicotinic agonists. Nicotine-induced whole cell currents in the SCG of β4 KO mice are also reduced but still present, suggesting that compensation from another subunit (i.e. β2) may be occurring (Xu et al., 1999b). Consistent with this notion, nicotine-induced currents in the SCG are abolished in double β2-β4 KO mice. Moreover, double β2-β4 KO mice exhibit similar bladder and pupil dysfunction as α3 KO mice. Taken together, these studies indicate that the clustered nAChR subunits are essential for normal ganglionic function and that compensation by β2 can occur with the loss of β4.
In addition to their PNS-specific phenotypes, α3, α5 and β4 KO mice also exhibit CNS-centric abnormalities compared to WT mice. For example, α3, α5 and β4 KO animals are resistant to nicotine-induced seizures compared to their respective WT littermates and are not as sensitive to nicotine-induced inhibition of locomotion (Salas et al., 2004a; Salas et al., 2003a). β4 KO mice also appear less anxious compared to WT mice in two specific anxiety assays suggesting a role for β4* nAChRs in modulating anxiogenic stimuli (Salas et al., 2003b). These mice also have a lower core body temperature which is less responsive to modulation by acute nicotine infusion (Sack et al., 2005).
Although the initial focus of nAChR expression was in the nervous system, it is now clear that “neuronal” nAChR genes are also expressed in non-neuronal cells where they participate in a number of fundamental processes (Gahring and Rogers, 2006; Sharma and Vijayaraghavan, 2002; Spindel, 2003; Wessler and Kirkpatrick, 2008). This is certainly true for the CHRNA5/A3/B4 subunit genes. The CHRNA5/A3/B4 genes are co-expressed in many cell types, thus their clustering may reflect coordinate regulation (Table 1). This hypothesis is supported by the fact that the transcriptional activities of the promoter regions of the three genes are regulated by many of the same transcription factors (Fig. 2). However, the CHRNA5/A3/B4 genes are not always co-expressed, suggesting that independent regulation of each gene also occurs (Table 1 and Fig. 2.).
In the nervous system, the α3 subunit is highly expressed in the periphery with a more restricted expression profile in the CNS (Gotti and Clementi, 2004; Greenbaum and Lerer, 2009). In the PNS, α3 subunit expression is seen in trigeminal sensory neurons (Flores et al., 1996; Liu et al., 1998), facial motoneurons (Senba et al., 1990), retina (Feller, 2002; Moretti et al., 2004), dorsal root ganglia (Zoli et al., 1995) as well as SCG, adrenal medulla, spenopalatine and otic ganglia (Rust et al., 1994). Centrally, the α3 subunit is expressed in the brainstem (Morley, 1997), cerebellum (Hellström-Lindahl et al., 1999; Turner and Kellar, 2005), spinal cord (Hellström-Lindahl et al., 1998; Keiger et al., 2003), substantia nigra (Azam et al., 2002), medial habenula (Grady et al., 2009; Zoli et al., 1995), pineal gland (Zoli et al., 1995), hippocampus (Gahring et al., 2004a; Guan et al., 2002; Hellström-Lindahl et al., 1999; Terzano et al., 1998; Winzer-Serhan and Leslie, 1997), cortex (Guan et al., 2002; Hellström-Lindahl et al., 1999) thalamus (Perry et al., 2002; Terzano et al., 1998), ventral tegmental area (Greenbaum and Lerer, 2009; Perry et al., 2002) and interpeduncular nucleus (Grady et al., 2009; Perry et al., 2002; Winzer-Serhan and Leslie, 1997).
Outside the nervous system, the α3 subunit is expressed in human oral keratinocytes (Arredondo et al., 2005; Arredondo et al., 2001; Conti-Tronconi et al., 1994) where its expression, both mRNA and protein, is increased following exposure to nicotine (Arredondo et al., 2005; Arredondo et al., 2001; Zia et al., 2000). α3-containing nAChRs are also expressed in lymphocytes (Benhammou et al., 2000), the gastrointestinal tract (Flora et al., 2000a; Glushakov et al., 2004), vascular endothelial cells (Macklin et al., 1998; Wang et al., 2001), polymorphonuclear cells (Benhammou et al., 2000), bronchial epithelium (Maus et al., 1998; Wang et al., 2001) and O2A progenitors (Rogers et al., 2001). The α3 subunit is also expressed in lung (Improgo et al., 2010; Lam et al., 2007; Sartelet et al., 2008) and its expression increases in small cell lung carcinoma (Improgo et al., 2010).
Similar to the α3 subunit, the α5 subunit is most highly expressed in the PNS but is also expressed in several key regions of the CNS (Gotti and Clementi, 2004; Greenbaum and Lerer, 2009). Centrally, α5 is expressed primarily in the cerebellum and thalamus (Flora et al., 2000a) but is also detected in the cortex, hippocampus, brainstem, spinal cord, habenula, interpeduncular nucleus and other midbrain nuclei (Azam et al., 2002; Gahring et al., 2004a; Grady et al., 2009; Hellström-Lindahl et al., 1998; Keiger et al., 2003; Wada et al., 1990; Zoli et al., 2002). In the PNS, α5 is expressed in most autonomic ganglia (Flora et al., 2000a; Liu et al., 1998) and the retina (Moretti et al., 2004).
In addition to the nervous system, α5 subunit expression has been detected in the gastrointestinal tract, where its expression is substantially higher than that of the α3 subunit, thymus and testis (Flora et al., 2000a; Glushakov et al., 2004). Furthermore, the α5 subunit is expressed in many of the same cell types as the α3 and β4 subunits including oral epithelium (Arredondo et al., 2005; Arredondo et al., 2001), vascular endothelial cells (Macklin et al., 1998; Wang et al., 2001), bronchial epithelium (Maus et al., 1998; Wang et al., 2001), O2A progenitors (Rogers et al., 2001) and a variety of immune cells (Wessler and Kirkpatrick, 2008).
As with the other two genes in the nAChR cluster, the β4 subunit gene is widely expressed in the PNS with more limited expression centrally (Gotti and Clementi, 2004). β4 expression is relatively high in trigeminal sensory neurons (Flores et al., 1996; Liu et al., 1998) as well as the superior cervical, dorsal root, spenopalatine and otic ganglia and sympathetic neurons (Mandelzys et al., 1994; Rust et al., 1994; Zoli et al., 1995). The β4 subunit is also expressed in the adrenal medulla (Di Angelantonio et al., 2003) with lower expression in the retina (Moretti et al., 2004). In the CNS, β4 expression is particularly high in the olfactory bulb, pineal gland, medial habenula and interpeduncular nucleus (Dineley-Miller and Patrick, 1992; Grady et al., 2009; Winzer-Serhan and Leslie, 1997) with lower expression in other thalamic nuclei, the cortex, hippocampus, spinal cord, cerebellum and midbrain (Azam et al., 2002; Gahring et al., 2004a; Hellström-Lindahl et al., 1998; Keiger et al., 2003; Perry et al., 2002; Quik et al., 2000; Turner and Kellar, 2005).
Outside the nervous system, there is again significant overlap of β4 expression with α3 and α5 expression. β4 is expressed in multiple cell types of the intestine (Flora et al., 2000a; Glushakov et al., 2004), vascular endothelial cells (Macklin et al., 1998), oral keratinocytes (Arredondo et al., 2005; Arredondo et al., 2001), polymorphonuclear cells (Benhammou et al., 2000), bronchial epithelium (Maus et al., 1998; Wang et al., 2001) and O2A progenitors (Rogers et al., 2001). Finally, β4 is co-expressed with α3 and α5 in lung (Improgo et al., 2010; Lam et al., 2007; Sartelet et al., 2008) and is also up-regulated in lung cancer (see below) (Improgo et al., 2010).
The co-expression of the CHRNA3/A5/B4 genes coupled with their genomic clustering suggests they may share common regulatory mechanisms in addition to specific regulation of each gene. Further support for this idea comes from several observations. First, nucleotide sequencing of the individual gene promoters revealed that they each lack classical CAAT and TATA boxes (Boulter et al., 1990). Instead, the promoters are GC-rich and contain several binding sites for the transcription factors, Sp1 and Sp3 (Fig. 2). Both Sp factors positively regulate transcription of each of the clustered subunit genes through multiple binding sites in each individual promoter (Bigger et al., 1996; Bigger et al., 1997; Boyd, 1996; Campos-Caro et al., 2001; Campos-Caro et al., 1999; Flora et al., 2000b; Melnikova and Gardner, 2001; Melnikova et al., 2000a; Terzano et al., 2000; Valor et al., 2002; Yang et al., 1995). Chromatin Immunoprecipitation (ChIP) experiments demonstrated Sp1 binding activity in the context of native chromatin for all three promoters (Benfante et al., 2007; Scofield et al., 2008). It is likely that Sp1 is involved in tethering the basal transcription machinery to the TATA-less nAChR subunit gene promoters (Pugh and Tjian, 1991). Second, in addition to the Sp factors, the CHRNA3/A5/B4 promoter regions can directly interact with and be trans-activated by the more spatially restricted regulatory factors Sox10 and SCIP/Tst-1/Oct-6 (Fig. 2)(Fyodorov and Deneris, 1996; Liu et al., 1999; Yang et al., 1994). Third, the mRNA levels of the CHRNA3/A5/B4 genes are coordinately up-regulated during neural development (Corriveau and Berg, 1993; Levey et al., 1995; Levey and Jacob, 1996) and coordinately down-regulated following denervation (Zhou et al., 1998). Perhaps the most compelling evidence for a coordinated regulatory scheme comes from the Deneris lab, which showed that two transcriptional regulatory elements, β43′ and conserved noncoding region 4 (CNR4), play key roles in directing expression of the clustered nAChR genes in a tissue-specific manner with β43′ being important for expression in the adrenal gland and CNR4 being critical for expression in the pineal gland and superior cervical ganglion (Fig. 2)(Xu et al., 2006). CNR4 is likely to play an important role in directing nAChR gene expression in the brain as well (Xu et al., 2006). In addition to these shared regulatory features, the CHRNA3/A5/B4 genes are subject to gene-specific regulation.
In vitro experiments have shown that the paired-like homeodomain transcription factor, PHOX2A, regulates transcription from the α3 promoter (Benfante et al., 2007). PHOX2A does not appear to bind directly to DNA, however, as the DNA-binding domain does not need to be completely intact for PHOX2A to regulate transcription from the α3 promoter (Benfante et al., 2007). Co-immunoprecipitation experiments demonstrate a physical interaction between Sp1 and PHOX2A, suggesting that PHOX2A is tethered to the α3 promoter through its interaction with Sp1, similar to the interactions of Sp1 with homeodomain transcription factors observed in other systems (Shimakura et al., 2006).
As briefly described above, the POU domain factor SCIP/Tst-1/Oct-6 has been shown to positively regulate transcription from the α3 promoter in a cell-type-specific manner (Yang et al., 1994). Similar to PHOX2A, the POU domain factor SCIP/Tst-1/Oct-6 does not require DNA binding for trans-activation of the α3 promoter (Yang et al., 1994). Deletional analysis of the SCIP/Tst-1/Oct-6 transcription factor demonstrated that only the POU domain is needed for trans-activation. Interestingly, this trans-activation does not depend on the presence of an Sp1 motif in the promoter region and is likely occurring through protein-protein interactions with the basal transcription machinery (Fyodorov and Deneris, 1996). The transcription factor Brn-3a also trans-activates the α3 promoter, while the other members of the Brn-3 family, Brn-3b and 3c, modestly repress α3 promoter activity (Milton et al., 1996). The positive regulation by Brn-3a is thought to be a result of protein-protein interaction as the α3 promoter lacks an obvious octamer or octamer-related binding site for Brn-3 factors (Milton et al., 1996).
A transcriptional enhancer has been discovered upstream of the α3 promoter in a region that overlaps with a 3′-untranslated exon of the β4 gene (McDonough and Deneris, 1997). This transcriptional enhancer consists of two identical 37-base pair repeats separated by a 6-base pair spacer. The β4 3′ enhancer acts as a cell-type-specific enhancer and is capable of enhancing transcription from the α3 promoter in neuronal-like cells, as well as in neurons in SCG cultures (McDonough et al., 2000). The enhancer contains several E-twenty six (ETS) factor-binding sites, the mutation of which dramatically decreases, but does not completely abolish, α3 promoter activity. The ETS-domain binding factor, Pet-1, has been shown to activate reporter gene transcription in a manner that is both cell type- and β4 3′ enhancer-dependent (Fyodorov et al., 1998). Taken together, these experiments suggest that Pet-1 interacts directly with the α3 promoter to activate transcription, though it likely requires additional cell-type-specific cofactors. In vivo experiments using transgenic mice showed that a larger DNA fragment between the α3 and β4 genes, containing both the β4 3′ enhancer and the α3 promoter, is capable of directing expression of a reporter gene to several areas of endogenous α3 expression in the brain (Wada et al., 1989; Yang et al., 1997). Surprisingly however, this DNA fragment did not direct reporter gene expression anywhere in the peripheral nervous system, in which the α3 gene is highly expressed (see above), suggesting that elements in this fragment may be acting as repressors or that other sequences are necessary for peripheral expression.
The presence of an intronic repressor element in the fifth intron of α3 has in fact been reported (Fuentes Medel and Gardner, 2007). The sequence of this α3 intron 5 repressor (α3I5) is highly conserved and is capable of bidirectional repressor activity in vitro. Notably, cell-type-specific repression of promoter activity was observed to be more potent in non-neuronal cell lines than in neuronal cell lines (Fuentes Medel and Gardner, 2007). These data suggest that this segment of DNA and the factors with which it interacts function to restrict expression of α3 to neuronal cell types. The protein-DNA interactions that mediate this effect have yet to be elucidated.
While regulation of the α3 subunit gene in neuronal cells has been extensively studied, the mechanisms regulating α3 expression in non-neuronal cells remain largely obscure. Recently, however, our group was the first to report that a transcription factor, the achaete-scute complex homolog-1 (ASCL1), regulates the expression of α3 and β4 and modestly of α5 in lung cancer cells (Improgo et al., 2010).
The α5 promoter region has been described in several genomic contexts including those in rodents and humans. Transcription of α5 occurs in the opposite direction as α3 and β4 (Fig. 1), suggesting that in addition to transcription factors that regulate the entire cluster, distinct mechanisms may govern α5 expression. However, apart from the regulatory factors described above that control expression of all three clustered genes, little is known about these mechanisms. SCIP/Tst-1/Oct-6 does not appear to regulate α5 though it regulates α3 and β4. Similarly, in lung cells, ASCL1 appears to regulate α3 and β4 but not α5 (Improgo et al., 2010). No other transcription factors regulating α5 expression have been reported, underscoring the need for more research effort in this area.
In addition to the Sp factors, Sox10 and SCIP/Tst-1/Oct-6, the β4 promoter is positively regulated by c-Jun (Melnikova and Gardner, 2001). Trans-activation by all of these factors is abolished when the Sp-binding site on the β4 promoter (referred to as a CA box) is mutated. Conversely, synergistic activation of the β4 promoter is observed when Sp1 is supplied in concert with Sox10, Sp3 or c-Jun (Melnikova and Gardner, 2001; Melnikova et al., 2000a). Co-immunoprecipitation experiments demonstrated that all of these factors physically interact (Melnikova et al., 2000b) and ChIP experiments confirmed that these interactions occur in the context of native chromatin (Scofield et al., 2008). These findings suggest the existence of a positively-acting multi-subunit transcriptional regulatory complex that assembles on the β4 promoter. This result is consistent with the hypothesis that Sp1 is critical for transcription from the β4 promoter and likely nucleates the regulatory complex that drives expression of the β4 gene.
Two additional transcription factors have been shown to interact with the β4 promoter, Purα and heterogeneous nuclear ribonucleoprotein K (hnRNP K)(Du et al., 1998; Du et al., 1997). These proteins interact with another motif, the CT box, located directly upstream of the CA box. hnRNP K is capable of repressing Sp factor-mediated trans-activation of the β4 promoter (Du et al., 1998) and also physically interacts with Sox10 (Melnikova et al., 2000b). Similar to hnRNP K, Purα physically interacts with Sox10 (Melnikova et al., 2000b). Moreover, Purα and hnRNP K themselves physically interact (Melnikova et al., 2000b). These proteins may participate in the multi-subunit complex described above to modulate expression of the β4 gene in the appropriate cellular context. In vitro binding experiments demonstrated that each factor binds preferentially to the opposing single strand elements of the CT box, suggesting that some local DNA helix unwinding may occur (Krecic and Swanson, 1999). Interestingly, Purα and hnRNP K have been shown to function together to negatively impact transcription of genes in other systems and the same may be occurring at the β4 promoter (Da Silva et al., 2002). In vivo experiments have also shown that a 2.3-kb fragment of the β4 promoter, containing the CA and CT boxes, is capable of directing reporter gene expression to brain regions that endogenously express β4, further supporting the importance of these elements in regulating β4 gene expression (Bruschweiler-Li et al., 2010).
Nicotine addiction is responsible for approximately 5 million deaths per year worldwide and is the most preventable cause of death in the USA (CDC, 2008). Nicotine addiction involves a series of events starting with the initial use of cigarettes, the transition from experimental smoking to regular smoking, and finally dependence on nicotine (Bierut, 2009b). Nicotine dependence is characterized by heavier smoking, early morning smoking, tolerance and withdrawal. Withdrawal symptoms account for the high incidence of relapse in people attempting to quit smoking (Kenny and Markou, 2001).
At the molecular level, nicotine addiction is initiated by the binding of nicotine to nAChRs. This interaction results in increased dopamine (DA) release in the DAergic mesolimbic and mesocortical circuits (Dani and De Biasi, 2001; Dani and Heinemann, 1996), a phenomenon widely associated with reward or reinforcement and drugs of abuse. DAergic projection neurons in the mesocorticolimbic pathway originate in the ventral tegmentum area (VTA) and project to the nucleus accumbens in the ventral striatum and the prefrontal cortex (Laviolette and van der Kooy, 2004). Several nAChR subtypes are robustly expressed in DAergic neurons both at the level of the soma, as well as at presynaptic terminals. Although, of the clustered nicotinic receptor subunit genes, α5 has been clearly identified as a component of nAChRs at DAergic terminals (Grady et al., 2007). Much of what is known regarding nAChR subtypes that are involved in nicotine addiction stems from studies on genetically engineered KO or knock-in mice (Champtiaux and Changeux, 2004). Mice that lack expression of β2* nAChRs fail to self-administer nicotine and exhibit impaired tolerance to nicotine (Maskos et al., 2005; McCallum et al., 2006; Picciotto et al., 1998). In addition, the predominant partner of the β2 nAChR subunit, α4 has also been implicated in nicotine dependence (Marubio et al., 2003; Tapper et al., 2007; Tapper et al., 2004). Selective activation of α4* nAChRs by low doses of nicotine in mice expressing α4* nAChRs hypersensitive to agonist is sufficient for nicotine reward/reinforcement, sensitization, and tolerance (Tapper et al., 2004). More recently, it was found that mice that do not express α6* nAChRs fail to self-administer nicotine (Pons et al., 2008). Expression of the α6 subunit has also been observed in the striatum and about half of α6* nAChRs co-assemble with α4 to form receptors with the highest affinity for ACh in the striatum (Champtiaux et al., 2002; Salminen et al., 2007). Because of their high expression in the reward pathway and high affinity for agonist, α4β2* nAChRs have been studied extensively in the context of nicotine addiction.
In recent years, linkage analyses, candidate-gene analyses, and large-scale genome-wide association studies (GWAS) have been used to screen hundreds of thousands of SNPs across thousands of individuals in search of genetic variants associated with nicotine dependence (Baker et al., 2009; Berrettini et al., 2008; Bierut, 2009a; Bierut et al., 2008; Caporaso et al., 2009; Chen et al., 2009a; Freathy et al., 2009; Greenbaum and Lerer, 2009; Grucza et al., 2008; Le Marchand et al., 2008; Saccone et al., 2009a; Saccone et al., 2009b; Saccone et al., 2007; Schlaepfer et al., 2008; Schwartz et al., 2009; Spitz et al., 2008; Stevens et al., 2008; Thorgeirsson et al., 2008; Vink et al., 2009; Wang et al., 2009a; Weiss et al., 2008). This has led to the identification of multiple SNPs in chromosome 15q24–25, a region that contains the CHRNA5/A3/B4 gene cluster as well as the iron-responsive element binding protein, a putative protein of unknown function, LOC123688, and the α4 proteasome subunit protein. Of particular interest is the non-synonymous SNP, rs16969968, found in exon5 of the α5 gene. This polymorphism changes an aspartic acid residue into asparagine at position 398 (D398N) in the second intracellular loop of α5. Individuals with one copy of the risk variant have a 1.3-fold increased risk for nicotine dependence while individuals with two copies have almost a 2-fold increase in risk. This SNP appears to influence different aspects of nicotine dependence, as it has been associated with increased risk for heavy smoking as well as increased sensitivity to the pleasurable effects of nicotine (Sherva et al., 2008; Stevens et al., 2008). In addition, the presence of this SNP may also influence the biophysical properties of α5* nAChRs. When expressed in HEK293T cells, α4β2α5D398 nAChRs exhibited a greater maximal response to nicotine compared to α4β2N398 nAChRs as measured by calcium imaging (Bierut et al., 2008). The association of this particular SNP has been identified in several studies though the majority of them have involved populations of European descent. In populations of African and Asian descent, the risk allele of rs16969968 is rare but is still associated with nicotine dependence (Saccone et al., 2009b; Wu et al., 2009).
How the clustered nAChRs function in the development of addiction is still unknown though recent studies in nAChR KO mice have revealed a potential role for these subunits in nicotine withdrawal. Withdrawal symptoms can be divided into two classes: somatic or physical symptoms, which likely involve both the PNS and CNS; and affective symptoms, which are centrally-mediated and are associated with hypoactivity of DAergic neurons in the reward circuit (Fung et al., 1996). As in humans, rodents chronically exposed to nicotine exhibit both somatic and affective withdrawal behaviors upon cessation. Somatic withdrawal symptoms in rodents are characterized by increased licking, grooming, shaking and scratching (Damaj et al., 2003; Kenny and Markou, 2001; Malin et al., 1994). Rodent affective symptoms include increased anxiety, hypolocomotion, and conditioned place aversion to nicotinic receptor antagonists (Damaj et al., 2003; Jackson et al., 2008a; Suzuki et al., 1996). The initiation and expression of withdrawal is dependent on neuronal nAChRs based on the fact that 1) symptoms can be alleviated by nicotine treatment and 2) symptoms can be precipitated by administration of nicotinic receptor antagonists during chronic nicotine exposure. Mecamylamine, a noncompetitive nicotinic antagonist, precipitates somatic withdrawal symptoms as well as affective symptoms including hypolocomotion, place aversion and anxiety, when injected i.p. in nicotine-dependent mice or rats (Damaj et al., 2003; Malin et al., 1994; Suzuki et al., 1999; Suzuki et al., 1996). The more selective high affinity nAChR competitive antagonist, dihydro-β–erythroidine (DHβE) (Harvey and Luetje, 1996; Harvey et al., 1996), can also precipitate both somatic and affective symptoms suggesting a role for high affinity nAChRs in withdrawal (Malin et al., 1998). Methyllycaconitine (MLA), a relatively selective α7 antagonist (Alkondon et al., 1992; Palma et al., 1996; Ward et al., 1990)(but see (Klink et al., 2001)), precipitates mild somatic, but few affective withdrawal symptoms (Markou and Paterson, 2001). Finally, somatic withdrawal symptoms can be precipitated by hexamethonium, a nicotinic antagonist that does not cross the blood brain barrier suggesting that peripheral nAChRs also play a role in these symptoms (Hildebrand et al., 1997) although similar symptoms can be elicited by injection of antagonist directly into the brain (Salas et al., 2009). Conversely, affective symptoms are not blocked by i.p. injection of hexamethonium indicating that they are mediated by central nAChRs (Hildebrand et al., 1997; Watkins et al., 2000).
Chronic nicotine-treated α5 and β4 KO mice display significantly milder mecamylamine-precipitated somatic withdrawal symptoms compared to WT mice (Salas et al., 2004b; Salas et al., 2009). In the CNS, α5 and β4* nAChRs are robustly expressed in the habenulo-interpeduncular pathway and direct infusion of mecamylamine into either the medial habenula or the interpeduncular nucleus can precipitate withdrawal in mice dependent on nicotine suggesting that the nAChRs in this pathway are critical for the expression of nicotine withdrawal (Salas et al., 2009). However, chronic nicotine treated α5 KO mice still become anxious during withdrawal suggesting that α5* nAChRs may not be involved in affective symptoms (Jackson et al., 2008b).
To date, the role of either α5* or β4* nAChRs in nicotine reward and reinforcement is unclear. Nicotine conditions a place preference in α5 KO mice indicating that expression of α5* nAChRs may not be necessary for the rewarding properties of the drug (Jackson et al., 2010). However, these animals not only exhibit a preference for nicotine at similar doses as WT mice, but also at high doses that normally fail to condition a place preference. Additional analysis of the reinforcing properties of nicotine in KO mice that do not express α5 or β4 nAChRs and/or in knock-in mice that express the α5 polymorphism identified in the GWAS should yield valuable insight into the molecular mechanisms underlying nicotine addiction.
Lung cancer is the leading cause of cancer-related deaths for both men and women worldwide (ACS, 2009). In terms of incidence, lung cancer is second only to prostate cancer in men and breast cancer in women. Lung cancer can be divided into two main histological types: small cell lung carcinoma (SCLC) and non-small cell lung carcinoma (NSCLC). The latter can be further divided into adenocarcinoma, squamous cell, bronchioalveolar and large cell lung carcinoma.
Tobacco intake is the main risk factor associated with lung cancer. The scientific link between tobacco and lung cancer has been firmly established since the 1950s (Proctor, 2001). This linkage is not surprising given 1) the addictive effects of nicotine that promote continued use of tobacco despite health consequences and 2) the presence of at least 55 carcinogens in tobacco, including nicotine metabolites such as 4-(methylnitrosamino)-1-(3-pyridyl)-butanone (NNK) and N-nitrosonornicotine (NNN)(Hecht, 1999; Shields, 2002). NNK and NNN form DNA adducts causing mutations that initiate cancer (Hecht and Hoffmann, 1988). Interestingly, these nitrosamines also serve as ligands for nAChRs (Schuller and Orloff, 1998) and along with nicotine and ACh, have been shown to trigger a variety of cancer-related processes as described below.
In 1989, Schuller’s group first showed that nicotine and its metabolites stimulates the growth of lung cancer cells (Schuller, 1989). Shortly thereafter, Minna’s group reported that nicotine causes apoptotic inhibition (Maneckjee and Minna, 1990). These two groups and many others thereafter have reported the expression of nAChRs in both normal and malignant lung cells, supporting the notion that nicotine stimulates cancer-related processes in a receptor-mediated fashion (Maneckjee and Minna, 1990; Maus et al., 1998; Sartelet et al., 2008; Schuller, 1989; Song et al., 2003; Wang et al., 2001). Moreover, studies have shown that specific nAChR subunits are over-expressed in lung cancer tissue (Lam et al., 2007). In particular, the CHRNA5/A3/B4 gene cluster is over-expressed in SCLC and this over-expression appears to be regulated by ASCL1 (Improgo et al., 2010). ASCL1 is a basic helix-loop-helix transcription factor important in the initiation and development of SCLC (Ball et al., 1993; Jiang et al., 2009; Linnoila et al., 2000; Osada et al., 2005). Regulation of the CHRNA5/A3/B5 locus by ASCL1 suggests a mechanism by which over-expression of ASCL1 in SCLC leads to a corresponding increase in expression of the clustered nAChR subunits, thereby potentiating the proliferative and pro-survival effects of nicotine and other nAChR ligands (Improgo et al., 2010).
Different groups have shown that various nAChR ligands activate distinct cancer-signaling pathways via the α3β2, α3β4, α4β2 and α7 nAChR subtypes (Schuller, 2009). ACh stimulates cell proliferation by acting as an autocrine growth factor in SCLC, a phenomenon that can be blocked by the non-specific nAChR antagonist, mecamylamine (Song et al., 2003). Nicotine also up-regulates the expression of growth factors and their cognate receptors (Conti-Fine et al., 2000). NNK causes cell proliferation by activating the serine/threonine kinase RAF1 and mitogen-activated protein kinase, leading to the activation of the oncogenic transcription factor c-myc, a process that can be blocked by the α7-selective antagonist, α-bungarotoxin (Jull et al., 2001). In addition, nicotine and NNK play a role in apoptotic inhibition. Nicotine activates the anti-apoptotic protein BCL-2 and inactivates the proapoptotic proteins Bad and Bax (Jin et al., 2004; Mai et al., 2003; Xin and Deng, 2005). Nicotine and NNK also activate the Akt/protein kinase B pathway, inhibiting apoptosis and causing tumorigenesis (West et al., 2003). Activation of Akt by nicotine appears to depend on nAChRs containing α3 or α4 whereas NNK activation depends on α7 nAChRs. Finally, nicotine acts as a pro-angiogenic agent, promoting endothelial cell migration, proliferation, tube formation and survival, an effect antagonized by α-bungarotoxin (Cooke and Ghebremariam, 2008; Heeschen et al., 2001). Along with its metabolite, cotinine, nicotine also up-regulates the expression of the vascular endothelial growth factor in endothelial cells (Conklin et al., 2002).
The aforementioned candidate-gene analyses and GWAS have also implicated the CHRNA5/A3/B4 gene cluster in lung cancer. The same SNP associated with nicotine dependence was found to increase the risk for lung cancer (Amos et al., 2008; Hung et al., 2008; Thorgeirsson et al., 2008). In particular, Hung and colleagues analyzed approximately 317,000 SNPs in thousands of patients and controls and found that the non-synonymous SNP, rs16969968, is robustly associated with lung cancer. This association was not affected by smoking status and was observed even in never-smoker groups, suggesting that the association is not due to the influence of nicotine dependence. Furthermore, no increased risk for other smoking-related cancers such as head and neck cancers was observed, indicating that the association with lung cancer is direct (Hung et al., 2008). Two other SNPs, in exon 5 and the 3′-untranslated region of the α3 gene, were also found in these studies to be associated with lung cancer (Amos et al., 2008; Hung et al., 2008; Thorgeirsson et al., 2008).
Aside from nicotine addiction and lung cancer, the CHRNA5/A3/B4 locus has also been associated with alcoholism. SNPs in this locus are associated with alcohol dependence and the level of response to alcohol (Chen et al., 2009b; Wang et al., 2009b). Furthermore, the variants appear to influence age at initiation of both tobacco and alcohol use (Schlaepfer et al., 2008). Since alcohol and nicotine are usually co-abused, these associations may provide a genetic mechanism for this co-morbidity (Greenbaum and Lerer, 2009). The risk allele of the non-synonymous SNP, rs16969968, also appears to be protective for cocaine dependence, suggesting a more general involvement of the CHRNA5/A3/B4 gene cluster in the actions of drugs of abuse (Grucza et al., 2008).
Other smoking-related diseases are influenced by variants in the CHRNA5/A3/B4 locus. Two SNPs in this region are associated with chronic obstructive pulmonary disease, a serious lung disease characterized by chronic inflammation and progressive destruction of lung tissues (Pillai et al., 2009). Interestingly, nicotine has been shown to promote the development of SCLC-like tumors in a rodent model of chronic obstructive pulmonary disease (Schuller et al., 1995). Another study showed that a variant influencing the risk for nicotine dependence and lung cancer also increases the risk for peripheral arterial disease, a condition characterized by obstruction of arteries outside the heart and brain (Thorgeirsson et al., 2008). The association of variants in the CHRNA5/A3/B4 gene cluster on other smoking-related diseases may be an effect of nicotine dependence, suggesting a gene-environment interaction, or may represent a direct effect, indicating pleiotropy of this locus (Bierut, 2009a; Thorgeirsson et al., 2008).
Tobacco use continues to be a major global health threat, underscoring the need to understand the mechanisms leading to nicotine dependence and smoking-related diseases such as lung cancer. Nicotine activation of nAChRs in the CNS initiates nicotine dependence. However, it is clear that nAChRs are also expressed outside the nervous system and are likely involved in numerous smoking-related pathologies. Although most studies to date have implicated the high affinity α4β2* nAChRs in nicotine dependence, less is known regarding the role of α3α5β4* nAChRs in this disease (Picciotto et al., 1998; Tapper et al., 2004). However, the recent flurry of GWAS implicating the CHRNA5/A3/B4 gene cluster in nicotine dependence lends greater credence to the involvement of these hitherto understudied nAChR subunits in nicotine addiction (Berrettini et al., 2008; Bierut et al., 2008; Portugal and Gould, 2008; Saccone et al., 2009a; Saccone et al., 2007; Schlaepfer et al., 2008). That results of both hypothesis-free GWAS and hypothesis-driven candidate-gene studies converged on this gene cluster is particularly compelling. These studies underscore the importance of elucidating the mechanism by which α3α5β4* nAChRs may modulate nicotine-mediated behaviors. In particular, future studies should examine nicotine self-administration in the existing mouse nAChR KO models as well as in knock-in mice expressing the polymorphisms identified in the GWAS.
Even more provoking are similar studies implicating this gene cluster in lung cancer (Amos et al., 2008; Hung et al., 2008; Thorgeirsson et al., 2008). It is still somewhat controversial, however, as to whether the association with lung cancer is direct or merely a by-product of the strength of the addiction or smoking behavior (Chanock and Hunter, 2008). As discussed in this Review, several lines of evidence support the hypothesis of a direct role of nAChRs in lung cancer. In particular, a number of pharmacology-based studies have demonstrated that nAChR activation can initiate a variety of signal transduction cascades associated with cancer (Schuller, 2009). However, it is important to acknowledge that the currently available pharmacological reagents used in these studies, while selective, are not specific, thus care must be taken when interpreting the results. This is a particularly important point with respect to the CHRNA5/A3/B4 cluster, for which the availability of selective ligands is extremely limited. Thus, design of highly specific nAChR ligands is of prime importance. Alternatively, studies using RNA interference and knockout/knock-in mouse models should help elucidate the function of specific nAChR subtypes in lung cancer.
Determining the role of the CHRNA5/A3/B4 cluster in nicotine dependence and lung cancer is important for the rational design of therapeutic interventions. In particular, the use of nAChR ligands as smoking cessation therapies have to take into consideration the possible effects of these drugs on lung cancer.
Work in the authors’ laboratories is supported in part by grant R01NS030243 (PDG) and R01AA017656 (ART) from the National Institutes of Health.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.