|Home | About | Journals | Submit | Contact Us | Français|
In the twenty-five years since the observation that chronic exposure to nicotine could regulate the number and function of high affinity nicotine binding sites in the brain there has been a major effort to link alterations in nicotinic acetylcholine receptors (nAChRs) to nicotine-induced behaviors that drive the addiction to tobacco products. Here we review the proposed roles of various nAChR subtypes in the addiction process, with emphasis on how they are regulated by nicotine and the implications for understanding the cellular neurobiology of addiction to this drug.
The tobacco plant alkaloid, nicotine, is generally agreed to be the major, if not sole, compound responsible for driving the addiction of more than one billion people (≈20 % world population), which, in turn, results in five million deaths worldwide each year . The unaided quitting rate for smokers is 3–5 % [2, 3] and, despite the availability of several nicotine replacement therapies , only about one third of people that would like to stop using tobacco products are permanently successful by the age of 60, usually after multiple failed attempts [reviewed in 5]. After nicotine enters the body, it binds to nicotinic acetylcholine receptors (nAChRs) of the central nervous system (CNS), specifically those in the brain, and initiates drug addiction [6, 7]. The persistent interaction between nicotine and nAChRs must ultimately lead to downstream plasticity at the molecular, cellular and circuit levels that then results in the behavioral desire to continue to intake nicotine.
Transmembrane nAChRs are fast-activating ligand-gated ion channels that produce membrane depolarization and cellular excitation . While much is known about the physiological role of nAChRs in the peripheral nervous system (PNS), the relevance of these receptors in CNS signaling has been somewhat obscure [for a concise historical perspective see 9]. In the PNS, it is now firmly established that the major role of nAChRs present postsynaptically at both the neuromuscular junction (NMJ) and within the autonomic ganglia, is to faithfully detect the presence of the chemical neurotransmitter, acetylcholine (ACh), thereby enabling efficient synaptic signaling. The synapses at the NMJ are designed never to fail and vertebrates would no doubt not have survived long if they were less than 100 % reliable [see discussion in 10, 11]. Exceptions to this generalization occur during development and with disease, e.g. myasthenia gravis, and usually result in significant changes in postsynaptic nAChR number [12, 13]. Autonomic ganglion synapses may be more pliable than the NMJ, as indicated by their ability to support activity dependent long-term changes in synaptic transmission . Mechanisms underlying the plasticity of nAChRs in the periphery may provide useful clues for understanding the changes in CNS receptors following chronic exposure to nicotine, particularly those in the autonomic ganglia, which share an overlapping neuronal subtype.
In the CNS the situation is more elusive, with only a few clear examples of fast synaptic transmission involving nAChRs, despite the widespread expression of these receptors and innervation by cholinergic fibers [9, 15, 16]. Coupled with a lack of anatomically defined cholinergic synapses, this has led to the postulation that nAChRs may function in part through more diffuse signaling, perhaps contributing to “volume” transmission –with receptors detecting ambient levels of ACh [17–19]. Although there is a somewhat limited knowledge of the operation of nicotinic synapses in the CNS, it is well established that nAChRs can contribute to long-lasting neuronal plasticity – including changes induced by the exogenous drug nicotine. Importantly, such plasticity likely helps condition the brain to secondary drug-related cues and/or context that make successful withdrawal from drugs like nicotine extremely difficult [20, 21], lending support to the idea that chemical addiction is a form of associative learning . Indeed, it is now appreciated that changes in synaptic efficacy as well as downstream gene regulation may provide a common molecular and cellular basis for both normal learning and addiction . The effects of nicotine on synaptic transmission and plasticity have been discussed elsewhere [24–26] and here we will focus on the mechanisms that lead to nicotine-induced alterations in nAChR number and function, changes in other classes of proteins, and their relationships to long-lasting nicotine dependence.
The diversity of nAChR subtypes, both in terms of their regional and subcellular distribution, implies that specific receptors may be localized to control cellular events that ultimately underlie a variety of discrete behaviors [16, 27–29]. Thus, in order to understand how nicotine-induced plasticity contributes to the long-term disruptions of neuronal activity in the CNS that underlie the behavioral adaptations acquired during nicotine exposure and withdrawal, it is necessary to know how these receptors operate and what specific properties, based on their subunit composition, allow them to interact with the low concentrations of nicotine in the cerebrospinal fluid (CSF) that are associated with the use of tobacco products.
Unlike muscle type nAChRs whose subunit composition is known and fixed, except during development , the molecular composition of neuronal nAChRs expressed in the nervous system is not completely defined [28, 31–35]. Indeed, even armed with the knowledge that only certain subunits can co-assemble, the question of just how many different native nAChR subtypes exist is difficult to answer. The mere presence of a large number of subunits [and possibly more based on sequence homology in other species, e.g. C. elegans; 36] has led some to conclude that diversity is the general rule rather than the exception [27, 28]. Based to some extent on the non-overlapping cellular distributions of nAChR subunit mRNAs in the CNS [e.g. 37], it is likely that the number of native nAChR subtypes will be restricted by specific patterns of subunit expression. Conversely, because many central neurons express multiple genes, their potential promiscuous assembly may not always limit the number of nAChRs subtypes found on single cells – and hinder determination of receptor composition . However, despite the enormity of the task, the wide range of available technologies, including high signal-to-noise ligands for radiolabeling binding sites, more recently developed selective pharmacological tools [39, 40] and genetically modified mice [28, 41], have helped define which subunits are incorporated into receptors, along with their regional and in some cases subcellular distribution. It is now clear that all subunits can participate in at least one type of nAChR present in the brain (Table 1). Below we have highlighted the essential characteristics of the major nAChRs and provided brief descriptions of their known or potential roles in the mechanisms of nicotine addiction. Of the receptor subtypes it is the α4β2* subtype which has to date served the pivotal role in nicotine addiction , perhaps because its relatively high affinity allows it to interact with the low levels of nicotine present in the CSF after tobacco smoke inhalation [see discussion in 43]. However, roles for other subtypes of nAChRs – especially those localized to the central reward pathways – are beginning to emerge. Nomenclature follows international standards, where an asterisk represents the inclusion of possible additional subunits .
Receptors containing α4 and β2 subunits represent the main population of central nAChRs, and account for the majority of high affinity binding sites labeled by [3H]nicotine [44, 45]. Indeed, the ligand specificity profiles of receptors formed from heterologously expressed α4 and β2 subunits are quite similar to high affinity [3H]nicotine sites . Moreover, genetic knockout of either the α4 or β2 subunits –aside from a small population in the interpeduncular nucleus (IPN) – eliminated all high affinity binding sites in the brain [47, 48]. However, where detailed functional comparisons of nAChRs expressed in Xenopus oocytes/mammalian cells have been made, no exact matches with native nAChRs were possible, implying that some α4β2* receptors may incorporate additional subunits, including α5 [49–51]. This suggestion is echoed in data from single cell RT-PCR studies showing co-expression of α4, β2 and α5 in cortical neurons . Although originally described as having a fixed stoichiometry , these receptors can exist in two distinct stoichiometries – a higher affinity (α4)2(β2)3 and a lower affinity (α4)3(β2)2 form – both in expression systems  and within the brain . The stoichiometry and precise subunit composition of α4β2* nAChRs is likely an important factor governing their upregulation by nicotine (see section 3).
Indeed, it was upregulation of these high affinity [3H]nicotine binding sites after chronic nicotine that began to define a central role for α4β2* nAChRs in addiction [56, 57] – a finding that is difficult to reconcile, at least at the level of the receptor, with the accompanying development of tolerance to nicotine . Following on from the initial demonstration that animals would self-administer nicotine , it was shown that the α4β2* receptor-preferring antagonist, dihydro-β-erythroidine (DHβE), reduced drug-seeking behavior when infused into the ventral tegmental area (VTA), but not after it had been introduced in the nucleus accumbens . Because DHβE is not a pure α4β2* nAChR antagonist, it was not until the development of suitable in vivo genetic manipulations that the involvement could be confirmed. In the first example, knockout of the β2 gene in mice produced parallel deficits in the activation of midbrain dopaminergic neurons and the self-administration of nicotine , which could be reversed by a viral rescue of β2 expression specifically in the VTA . These results have now been extended to α4 subunits . Moreover, in mice containing a hypersensitive α4 allele knocked-in in place of the wild-type α4 gene, it was possible to show that activation of α4 subunit-expressing nAChRs was sufficient to produce conditioned place-preference . Although α4 and β2 subunits can contribute to other receptors, e.g. those with α6 (see Table 1; ), in sum these experiments reveal a central role for α4β2* nAChRs in nicotine addiction. Evidence from brain slice electrophysiological studies is beginning to unmask how this may take place at the level of synaptic integration, including studies that show how acute nicotine-induced desensitization of putative α4β2* receptors on midbrain interneurons blocks their endogenous activation by acetylcholine and shifts the balance of inhibitory-excitatory input to the VTA dopaminergic cells in the direction of excitation [64, 65]. Although chronic nicotine may have additional regulatory effects on these particular VTA nAChRs [66; see section 3], these data imply that nAChR desensitization, along with receptor activation, likely contributes to aspects of nicotine addiction .
α7 subunit-containing receptors provide a molecular explanation for the large number of [125I]α–bungarotoxin (αBTX) binding sites found in the brain [68, 69] and the pharmacological sensitivity of fast activating nAChRs in areas such as the hippocampus [70, 71]. Although α7 subunits readily form homomeric receptors in Xenopus oocytes [but see 72], it is unclear whether they exist in this form in the brain . Brain receptors purified using αBTX as a probe only contain α7 subunits [74, 75], and mice null for the α7 gene have a complete loss of both αBTX binding sites in the CNS and functional αBTX-sensitive nicotinic responses in hippocampal neurons . However, α7 can co-assemble with other α and β subunits  resulting in receptors present in situ with a lower sensitivity to αBTX  conceivably missed in the above αBTX purification experiments. The α7* nAChR conductance is unique amongst nAChRs; typified by very fast activation and desensitization kinetics and a high calcium permeability [68, 69, 71, 79]. No generalized role for these receptors in central synaptic transmission can be ascribed; in the CNS they are present and function at postsynaptic [80, 81] and presynaptic sites [for review see 82].
As discussed above, α4β2* nAChRs have an essential role in nicotine self-administration (section 2.1.1), but recent data show that this is not likely to be sufficient to account for all aspects of addiction to nicotine. Some of the most convincing evidence that other nAChR subtypes are involved comes from further analysis of β2 knockout mice . In these animals, nicotine can act though methyllycaconitine (MLA)-sensitive α7* nAChRs to normalize aberrant neurochemical and behavioral deficits. Changeux and colleagues conclude that both and α4β2* and α7* receptors are necessary to produce the homeostatic neuroadaptations that develop in the presence of nicotine . However, with respect to a direct role in establishing drug-seeking behavior, it appears that α7 subunit-containing receptors are not required . Using both pharmacological and genetic approaches it was concluded that α7* nAChRs do not support conditioned place preference , although high doses of MLA were shown to reduce nicotine self-administration , implying a possible α7* nAChR contribution to the reinforcing properties of nicotine. Precipitation of somatic withdrawal symptoms using the non-specific blocker, mecamylamine, was attenuated in α7 knockout mice [86; but see 85], however, MLA-induced withdrawal was unaffected in the absence of the α7 gene, urging caution and implying that non-α7 MLA-sensitive nAChRs may be also involved in withdrawal from nicotine .
At the synaptic level, α7* nAChR receptors may contribute to plasticity of glutamate synapses, including those impinging onto dopaminergic neurons in the VTA. Here nicotine triggers release of glutamate, which under the appropriate conditions can increase synapse efficacy for prolonged periods of time, both in vitro  and in vivo . These persistent effects of α7* nAChR stimulation may be representative of a more generalized cognitive role for this receptor subtype  – especially given the findings of α7* receptor-induced synaptic plasticity in the hippocampus . Moreover, its potential role in information processing is highlighted in schizophrenic subjects, who smoke at four times that of the general population, possibly in part to control sensory gating dysfunction caused by an altered α7 gene .
The initial limited expression of β4 subunits in the CNS  supported the idea that β4 subunit-containing nAChRs served a dominant function only in peripheral ganglia . However, closer analysis of β4 distribution revealed that it was often associated with the α3 gene [93–95], and analysis in regions that co-expressed these two subunits, e.g., medial habenula (MHb), locus coeruleus, and superior colliculus, revealed receptors with functional properties and a likely minimal molecular composition including α3 and β4 subunits . Like α4β2* nAChRs, these native receptors may also incorporate α5 [51, 97–99], a finding that, if true, makes sense as all three subunits are present within a single gene cluster . The β3 subunit may also be included, further increasing the diversity of these nAChRs . The lower affinity of these receptors for nicotine  prevents them from being detected using [3H]nicotine, but the higher affinity, and relatively non-selective agonist, epibatidine, strongly labels these sites. Nicotine displaces radiolabeled [3H] or [125I]epibatidine with high and low components corresponding to the α4β2* and α3β4* nAChRs, respectively [102–104]. As expected, these receptors remain in the β2 subunit knockout mouse, in which they can be further subdivided into additional classes [103, 105]. Unlike its dominant role in ganglionic transmission [106, 107], the synaptic function of the α3β4* nAChR within the CNS has not been established.
In part due to their scarcity, little attention has be paid to the role of non-α4β2* heteromeric nAChR subtypes with respect to nicotine addiction. However, two sets of independent results link the β4 subunit to mechanisms of withdrawal from nicotine (and other drugs) and provide circumstantial support that the habenulo-interpeduncular axis may represent an additional component of the “reward center”, especially given its ability to support intracranial self-stimulation . First, somatic signs of withdrawal are reduced in β4, but not β2, null mice . Second, the natural alkaloid, ibogaine, and a synthetic derivative, 18-methoxycoronaridine, known antagonists of the α3β4* receptor subtype , when injected into the MHb or IPN can reduce sensitization , self-administration  and withdrawal symptoms  to morphine. That these compounds also attenuate nicotine-conditioned place preference  may point to a more generalized role for habenulo-interpeduncular α3β4* nAChRs in addiction.
Although both β3 and α6 subunits can be incorporated into heteromeric receptors in heterologous expression systems [101, 115], the contribution of these subunits to native nAChRs in the CNS was until recently unclear. The function of these two subunits in the brain is perhaps best observed after removal of the majority of high affinity α4-subunit containing binding sites: [125I]epibatidine binding in the presence of cytisine (to mask high affinity nAChRs) leaves a population of sites with a similar distribution to α3 subunit mRNA, as predicted . Some of these “α3”-receptor sites represent α3β4* nAChRs, like those in the MHb [96, 105], whereas others may contain different gene members. This idea was confirmed using α-conotoxin MII (α-MII), a nicotinic antagonist toxin with originally described selectivity for α3β2 over α3β4 subunit-containing receptors , which displaced some of the remaining [125I]epibatidine binding . Moreover, [125I]α-MII binding is lost in either β2  or β3  null mice, suggesting initially that this receptor may be comprised of α3, β2 and β3 subunits . However, not least because many [125I]α-MII sites are found in the α6 subunit-rich striatum [119, 120], attention became focused on the contribution of this previously obscure subunit [see 121]. Thus, it is likely that α-MII sensitive presynaptic nAChRs on dopaminergic terminals [117, 122] do not contain α3 subunits, but rather express an α6β2β3* nAChR subtype [123, 124], a view supported by the lack of effect of α3 subunit knockout on [125I]α-MII binding in most areas of the brain .
The release of dopamine in the striatum is controlled by presynaptically localized α6β2β3* (in addition to α4β2*) nAChRs and, because of their additionally highly restricted distribution in the reward centers of the brain, these receptors are strongly implicated in mechanisms of nicotine addiction [123, 124, 126, 127]. Under normal conditions, the nAChRs on dopaminergic terminals are activated by endogenously released ACh from intrinsic striatal cholinergic neurons and this activity appears to be critical for the normal maintenance of dopaminergic transmission . Thus, “chronic” nicotine-induced desensitization of these presynaptic nAChRs (although the precise nature of the subtype(s) affected is unknown) can produce a profound decrease in dopamine release  and potentially interfere with other known mechanisms of synaptic plasticity in this region [for review see 129]. Recently, it has been established that α6* nAChRs specifically within the VTA are an absolute requirement for nicotine self-administration in mice , and coupled with similar data for α4 and β2 subunits (see section 2.1.1), permits the suggestion that regulation of dopamine release by presynaptic α4α6β2* nAChRs on VTA neuron terminals  may be critically important in establishing addiction to nicotine .
While the above sections have discussed the most abundant subtypes present in brain it remains an incomplete list. It is worth mentioning other nAChR classes that may be more restricted in their distribution, but none-the-less have important roles. Obvious members of this group are receptors comprised of α9 / α10 subunits, found only in sensory epithelia [131, 132], where they mediate transmission of efferent inhibitory information via release of ACh onto hair cells in the cochlea – apparent as a postsynaptic hyperpolarization triggered by activation of potassium channels via calcium flux through the nAChR channel [133, 134]. Although the only remaining “homomeric” subunit, α8, is absent in mammals, it is likely to have important function in chick retina/brain, possibly in combination with α7 [135–137]. The discovery that most [125I]α-MII binding sites represent α6β2β3*-nAChRs  has led to a decreased emphasis on α3β2* receptors in the brain. However, within the habenulo-fasciculus retroflexus tract, [125I]α-MII are almost exclusively dependent on the α3 subunit , implying that the α3β2 subunit combination exists either alone or as one of the interfaces of an α3β4* nAChR [96, 104]. Neurons in the IPN have been known for some time to express a channel functionally distinct from those in the MHb . Recent work suggests that this receptor is a second class of β4 subunit-containing receptor that may include the α2 and/or potentially the α4 subunit [35, 105]; α2 subunits exist almost exclusively in the IPN [37, 139]. Essentially nothing is known about the functional role of this highly restricted receptor subtype. Other mixed interface nAChRs may exist with up to four distinct subunits .
As discussed in the preceding sections, pharmacological and genetic manipulations indicate that specific nAChR subtypes/subunits are likely to be involved in various aspects of addiction to nicotine. Thus, it is highly likely that individual differences in nAChR alleles may help explain variability in the predisposition to smoking in the human population. Recent studies have identified single nucleotide polymorphisms (SNPs) within the α3-α5-β4 gene cluster  located on chromosome 15 that correlate with the intake of nicotine [141, 142], implying that genetic factors at the level of nAChRs may predispose certain individuals to nicotine addiction. In particular, one SNP in the presumed cytoplasmic domain of the α5 gene produces an amino acid switch and a decrease in the functional expression of heterologously expressed α4β2α5 nAChRs  allowing the authors speculate about cellular mechanisms that may increase the risk of transitioning from casual smoking to addiction. Other researchers have found that, in addition to the number of cigarettes smoked, there is a genetic association of the same gene cluster with certain smoking-related diseases such as lung cancer and peripheral artery disease , although it is unclear at present whether this is due to individual patterns of smoking.
In addition to the nAChR gene cluster, there is some evidence that other receptors, including the β3 nAChR, may be genetically linked to nicotine addiction . On the other hand there is no apparent correlation between α3-α5-β4 genes and smoking cessation , although β2 subunits may be implicated . These latter findings are somewhat puzzling and worthy of further investigation given the known relationship between β4* receptor subtype(s) and somatic withdrawal symptoms in rats  discussed above.
Animals are highly capable of adapting to different sets of conditions so that they can continue to perform optimally. Following this rule, chronic exposure to certain drugs will initiate mechanisms that act to restore normal brain function in the constant presence of the chemical. That is, the system has undergone some sort of plastic change, which persists after removal of the drug, and presumably contributes/underlies both drug seeking and a withdrawal state. For certain drugs, e.g., opiates such as morphine, this type of homeostatic mechanism at the level of the drug-receptor interaction has been useful for explaining the molecular and cellular processes that underlie behavioral addiction . However, it is unlikely that this mechanism alone accounts fully for addiction, and other mechanisms including direct and lasting interference with normal reward systems, and the induction of learned associations involving contextual cues are also of central importance [20, 22, 23]. The behavioral aspects of nicotine addiction have been dealt with effectively elsewhere , and a central goal here is to formulate models that can account for the cellular consequences of exposure to nicotine.
Nicotine acts at the ACh binding site of nAChRs, and thus any model of nicotine-induced plasticity must begin with the interaction of nicotine and its receptors. Continuous exposure to nicotine through its interactions with nAChRs will alter the pattern of ongoing endogenous neuronal activity in the brain, which is then restored to normal in the chronic presence of the drug, i.e., homeostatic neuroadaptation [see 149]. In the case of nAChRs, based on data collected over the last twenty-five years, a common starting point invokes that the minimal homeostatic changes occur in the number and/or function of the nAChRs that bind to nicotine. Obviously, the situation becomes complicated with multiple nAChR subtypes, their cellular distribution within neuronal networks, and their downstream consequences, which themselves could drive additional compensation.
One of the most enduring aspects of nAChR plasticity has been the discovery that chronic in vivo nicotine treatment (equivalent to persistent tobacco smoke inhalation) causes a marked increase in the number (change in Bmax but not Kd) of [3H]nicotine (α4β2*) binding sites in the CNS [45, 56, 57, 150]. The increase is long-lasting, but not permanent , and was also observed in the postmortem human brains of smokers . Because nicotine is an agonist, upregulation of the receptors appears to contradict the simplest of homeostatic mechanisms that predict downregulation as a compensatory measure, leading to the introduction of the desensitization hypothesis .
Nicotine, like most other agonists, including ACh, does not simply activate nAChRs, but ultimately induces desensitized states of the receptor . In practical terms this means that nicotine acts in a different manner from ACh. ACh is rapidly terminated (in milliseconds) by hydrolysis and keeps nAChRs at the synapse in an activatable state . Nicotine has a long half-life ≈1 hr  essentially keeping it present in the CSF permanently given repetitive intake . The resulting accumulation of desensitized receptors at synapses would effectively depress ongoing transmitter action – homeostasis would then predict receptor upregulation [56, 57, 156; see Fig. 1]. Assuming that the nAChR can detect and respond to less ACh activation, this simple scenario predicts that antagonists should also increase nAChR number, but the results from such studies have been mixed [156–160]. However, assuming antagonists do not induce desensitized receptor conformations [although see 161], it has been argued that agonist-induced desensitization is an absolute requirement for upregulation, and that receptor blockade per se is insufficient . Two lines of evidence support this view. First, in cells isolated from their synaptic input, nicotine treatment still produces upregulation , i.e., upregulation is cell-autonomous . Second, upregulation can, in some cases (e.g., when surface nAChRs are considered), be directly correlated to occupation by nicotine of high-affinity desensitized states . However, while the desensitization model gains extra credibility because the concentrations of nicotine in the CNS resulting from tobacco smoke inhalation are submicromolar  and will tend to favor interaction with desensitized receptor conformations [153, 166], alternate mechanisms have been proposed in which nicotine can interact with other receptor conformations. Recently, the concept of desensitization-induced upregulation has been challenged using mice expressing hypersensitive α4 subunit-containing nAChRs. In these animals, ultra low concentrations of nicotine, which interact only with the mutant α4* receptors, cause functional upregulation and induce drug-seeking behavior . Because the sensitivity of agonist for receptor desensitization is only marginally affected compared to a marked shift in the ability of agonist to open the mutant ion channels , it seems reasonable to suggest that activation (not desensitization) of native α4β2* nAChRs can drive both the molecular and behavioral aspects of drug addiction. However, this explanation does not satisfactorily account for upregulation observed in the absence of channel opening [see below; 163].
One of the more controversial aspects of upregulation has concerned the source of the additional surface receptors. The factors governing the regulation of receptor assembly, membrane targeting and recycling are not fully understood [168, 169], and it is therefore not surprising that the mechanism of nicotine-induced regulation of nAChR function is also unclear. Efforts have also been hampered by artificial expression systems, in which the majority of nAChRs reside in intracellular pools, possibly because factors necessary for surface membrane trafficking and stability are missing from these systems . Nevertheless, cumulative data from experiments involving both native and over-expressed nAChRs have provided potential explanations for upregulation. There is general agreement that nicotine does not induce changes in the levels of α4 and β2 mRNA transcripts [158, 170], and as such post-transcriptional mechanisms including increased translation , assembly , and trafficking from an intracellular pool , or decreased turnover [158, 174] have all been proposed to account for an increased surface expression. One problem with the desensitization theory of upregulation is that the concentration of nicotine required for upregulation is not optimal for a pure interaction with the desensitized state of the receptor . Indeed it is generally found to be intermediate between the measured “affinities” for the main activatable and desensitized states of surface nAChRs [160, 164], seemingly implying that nicotine must be interacting with some unknown receptor conformation . Resolving the nature of this state may therefore shed light on the precise mechanism of receptor upregulation. Because nicotine can readily cross membranes , it could have access to assembly intermediate forms of the nAChR, which conceivably could have different binding “affinities” for agonists compared to mature surface receptors. Although non-membrane permeable quaternary amines, e.g., tetramethylammonium, are as efficacious as nicotine in producing upregulation , implying that surface receptors are the trigger [see 164], it has been suggested that nicotine and quaternary amines, including ACh, act as chaperones to promote assembly and maturation of nAChRs within the endoplasmic reticulum [174, 176, 177]. Importantly, if these assembly intermediates (dimers and tetramers) have a lower “affinity” for interaction with agonist than the desensitized conformations of mature surface nAChRs, then this may explain why higher concentrations of nicotine are needed for upregulation than predicted directly from desensitization binding measurements.
An essential component of the homeostatic model is that upregulation should represent an increase in function – to compensate for the blocked and/or desensitized receptors (see Fig. 1). Here the story becomes more complicated because initial studies showed that nicotine exposure caused functional downregulation, at least in the case of α4β2* nAChRs [158, 178], a finding that may partially be explained in retrospect by the slow release of intracellularly accumulated nicotine and re-desensitization of surface nAChRs . In contrast, several studies have shown that an increase in the function of expressed α4β2 nAChRs accompanies the change in number [e.g., 165, 174]. Such parallel changes in number and function are easy to interpret, but other studies imply more complex relationships between function and number of upregulated receptors . Vallejo et al. argue that surface receptor number is unaffected by nicotine treatment, but rather there is an allosteric-induced shift in pre-existing nAChRs to a higher affinity conformational state . Other studies also imply a shift in the affinity of receptors following upregulation, but attribute this change to increased expression of an altered α4β2 stoichiometry . Interestingly, native α4β2* nAChRs with a more complex subunit composition containing the α5 subunit do not appear to undergo upregulation in vivo . As a whole these experiments support every aspect of the homeostatic upregulation theory, but with critical differences with respect to mechanism. An all-inclusive but less than satisfying interpretation is that multiple modes of nicotine-induced modulation can occur, with exact conditions determining the prevailing form. Nevertheless, Nashmi and Lester have recently offered a thermodynamic solution that reconciles many of the inconsistencies described in the preceding sections . They propose that ligands acting at the binding site of nAChRs (including agonists like nicotine and high affinity competitive antagonists such as DHβE) will stabilize high affinity conformational receptor states (e.g. desensitized) that, due to slow kinetics, will outlast the binding of ligand, i.e., after chronic nicotine exposure a greater fraction of the receptors will be receptive to agonist and as a consequence receptor function will be enhanced. Furthermore, because these ligands bind at the αβ interface, this argument can be extended to the stabilization of the multimeric nAChR during intracellular assembly .
It has been accepted that, despite quantitative differences, there is widespread upregulation of high affinity binding sites across multiple brain regions during chronic nicotine exposure [170, 182, 183]. New experiments have now re-quantified upregulation specifically for functional α4 subunit-containing receptors by measuring fluorescence from mice expressing an α4-YFP knock-in construct. While other CNS nuclei were affected, these data highlight particularly significant increases in α4* nAChR function in GABAergic neurons of the VTA and substantia nigra (but not in the principle dopaminergic cells), as well as in the perforant path of the hippocampus .
The effects of chronic nicotine on non-α4β2* nAChRs are less than clear. Midbrain dopaminergic neurons terminate in the striatum, where the release of dopamine can be directly affected by two types of presynaptic nAChRs, α4β2* and α6β2β3* [123, 124, 126, 127]. Although in vivo chronic nicotine experiments have largely focused on upregulation of α4β2* nAChRs , work in striatum and other areas of DA axon termination indicated that a different receptor subtype may be functionally upregulated by nicotinic ligands [184, 185; but see 186]. Making use of α-MII (see section 2.1.4), it has been shown that α6 subunit-containing receptors are markedly increased in number compared with α4β2* receptors following chronic nicotine treatment [187; but see 188], in agreement with a preferential upregulation of α4β2* nAChRs on GABAergic cells in this region . In addition mRNAs for α6 and β3 subunits are increased in the substantia nigra , implying a possible transcriptional regulation, in contrast to α4β2* nAChRs . While these data support functional upregulation of striatal α6* receptors, other groups have shown that chronic nicotine produces a slight depression of agonist-stimulated release [56, 190]. It is possible that treatment paradigms or species differences underlie these disparate results. Indeed, experiments from Walsh et al. have shown that different nAChR subtypes, including those containing α6, when expressed in heterologous cell systems, can undergo upregulation, although the nicotine concentration-and time-dependence varies with receptor type .
There also appears to be a concentration-dependence to the nicotine-induced upregulation of various non-α4β2* nAChRs observed in other studies. As a general rule, higher concentrations of nicotine are necessary - reflecting the lower affinity of the activatable and desensitized states of these receptors . Some α3 subunit-containing receptors are upregulated depending on the expression system [191, 192], but more importantly native α3* nAChRs in the habenulo-interpeduncular axis, as characterized by a component of [125I]epibatidine binding, are not altered by chronic exposure to nicotine in vivo , possibly because the CSF concentration of nicotine is too low. Likewise, the very low affinity α7* nAChRs/αBTX binding sites are upregulated (but only at elevated doses of nicotine compared to those needed to increase α4β2* nAChRs) both in vitro  and in vivo .
In addition to the commonly reported changes in nAChRs, chronic in vivo nicotine treatment also leads to changes in the expression of non-cholinergic targets within reward-related structures (Table 2). Many of these changes are in opposing directions depending on the structure, but given that the different structures may have unique roles in the addiction process , different drug administration paradigms can differentially effect gene expression , and alterations of these targets by nicotine can occur in multiple ways (see below), it perhaps should not be surprising that there are varied effects. The effects observed are common to most drugs of abuse and include changes in ligand-gated ion channels, the dopamine system, neuropeptides, transcription factors and growth factors . Additionally, changes in members of the two-pore domain family of potassium channels [197, 198] and L-type Ca2+ channels [199, 200] have been found following nicotine treatment (Table 2), and while changes in the expression of K+ channels and Ca2+ channels in response to other drugs of abuse may not have been reported in the literature as of yet, the function of voltage–gated channels that control the excitability of neurons within the prefrontal cortex, nucleus accumbens, and hippocampus have been shown to be altered [201–207]. Because nicotine has the ability to alter the induction of synaptic plasticity [24–26] it may influence changes in gene expression that typically occur downstream of plasticity induction leading to changes in certain molecular targets. In addition to influencing gene expression by altering plasticity, nAChRs may directly flux Ca2+ [32, 208–210], which is the principle mediator of plasticity induced gene expression . Nicotine may also play an even more direct role by accumulating in cells [174–177] and directly interacting with the ubiquitin proteosome system, thereby influencing the degradation of proteins .
The data reviewed here show that chronic exposure to nicotine can induce changes in its own receptors in addition to causing altered expression of many other cellular targets. In particular there is general agreement that high-affinity nicotine binding sites can be functionally upregulated and although the precise mechanism by which this occurs is not fully resolved it most likely involves nicotine entry into the cell and interference with the assembly process. Based on this evidence we are forced to conclude that upregulation of α4β2* nAChRs does not represent a direct homeostatic response to persistent receptor desensitization, but may in fact be a fortuitous event resulting from the ability of nicotine to cross membranes, accumulate in cells, and bind to nAChR assembly intermediates –effectively all thermodynamics, as discussed by Nashmi and Lester . Do we have to further conclude that the upregulation of nAChRs is an unfortunate epiphenomenon with no causal role in addiction to nicotine? Based on genetic and pharmacological elimination of either and/or both β2 and α7 subunit-containing receptors, Besson et al. found it difficult to incorporate the upregulation of high affinity sites into a homeostatic model in which the “opposite” action of nicotine at these two receptor classes offset both neurochemical and behavioral effects . Conversely, other researchers have incorporated the upregulation of these receptors into models of sensitization and tolerance to nicotine. It has been proposed that the transient upregulation of nAChRs increases the excitation of dopamine neurons in the VTA by permitting long-term potentiation of synaptic transmission. Together these changes result in enhanced dopamine release, which underlies behavioral sensitization . Conversely, focusing on the mechanism of tolerance to nicotine in midbrain neurons, Nashmi et al. argue that selective upregulation of α4* nAChRs on GABAergic cells coupled with chronic stimulation by nicotine may lead to a decreased output of dopaminergic neurons . Ultimately it will be important to test the validity of these models and find methods of probing directly whether upregulation of nAChRs is a necessary step in the development of behaviors that lead to nicotine addiction. There will be a similar challenge with respect to determining which of those changes in other nAChR subtypes and other cellular processes are a direct consequence of the presence of nicotine, compensatory secondary adaptations, or driven by reward associated behaviors.
Supported by PHS grant NIDA DA11940
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.