Prime Time for G-Protein-Coupled Receptor Heteromers as Therapeutic Targets for CNS disorders: The Dopamine D1-D3 Receptor Heteromer

Abstract

A number of G-protein-coupled receptors (GPCRs) are currently under consideration as potential therapeutic targets for drugs acting in the central nervous system (CNS). Attempts to discover new medications have operated under the assumption that GPCRs are monomers and that a specific drug activates one single receptor coupled to one single signal transduction mechanism. In the neuronal membrane, GPCRs are now known to be arranged into homo- and hetero-oligomers; drugs acting on a single receptor within a specific heteromer context are thought to induce a particular downstream signaling. However, there is recent evidence howing that heteromer tailored drugs can be designed that display different affinities for a given receptor depending on the receptor partners contained within the heteromer. It can therefore be predicted that customized drugs targeting a specific receptor heteromer in the CNS might imporove safety and efficacy for their therapeutic targets. Finally, it will be important to identify receptor heteromers that are involved in the pathogenesis of diseases, such as the recently discovered dopamine D1-D3 receptor heteromer, which might play a key role in L-DOPA-induced dyskinesia in Parkinson’s disease.

INTRODUCTION

G-protein-coupled receptors (GPCRs) are often used as receptor targets for the discovery and design of novel molecules used for the treatment of a variety of diseases, including neurological and mental disorders. Until now, drug design was performed assuming that GPCRs were monomeric. The occurrence of GPCRs as dimers and higher-order homo- and hetero-oligomers, and the consequences of homo- and hetero-oligomerization should drive a completely new approach in drug design. With the beginning of the new century, the discovery of receptor heteromers was accompanied by evidence substantiating novel properties of the heteromer in terms of ligand recognition, signaling and receptor trafficking [1–8]. For instance, when reporting on heteromerization of μ- and δ-opioid receptors, George et al. showed that the receptor heteromer had novel ligand recognition properties and different G protein coupling properties in a heterologous system when compared with the properties of individually expressed μ- or δ-opioid receptors [9]. More recently, Rozenfeld and Devi have shown that μ-δ-opioid receptor heteromerization leads to a constitutive recruitment of beta-arrestin2 to the receptor complex, ultimately resulting in changes in the spatio-temporal regulation of ERK1/2 signaling [10]. The involvement of beta-arrestin2 is furthersupported by studies using beta-arrestin2 siRNA in cells endogenously expressing μ-δ receptor heteromers [10]. Several other examples are also given below that illustrate how receptor heteromerization changes classical views about the way neurotransmission takes place in both healthy and pathological brains. Furthermore, we will underscore the importance of finding receptor heteromers that are involved in the pathogenesis of CNS disorders. Finally, clues will be given about how to obtain heteromer-tailored drugs for CNS disorders, which may be the most appropriate choice in the near future in terms of efficacy and safety.

MARKETED DRUGS ARE ALREADY TARGETING GPCR HETEROMERS

The effects of antiparkinsonian and antipsychotic drugs depend on their ability to bind to dopamine receptors. Do these drugs already target dopamine receptor heteromers in the CNS? Although we still have little evidence indicating that targeting receptor heteromers is beneficial or detrimental for a specific therapy, what it is already clear is that compounds targeting CNS receptors are, in fact, acting on receptors forming heteromers. L-DOPA, which converts to dopamine in the brain, represents the gold-standardtherapeutic drug for Parkinson’s disease (PD). One subclass of dopamine receptors are the D1-like receptors which include the D1 and D5 receptor subtypes, that couple to Gs/olf proteins when not forming heteromers. The other subclass of dopamine receptors, the D2-like receptors, include the D2, D3 and D4 receptor subtypes, that couple to Gi/o proteins when not forming heteromers [11]. Animportant number of dopamine receptor-containing heteromershave been identified which are most probably localized in the brain.Therefore, current formulations of L-DOPA must be targeting a variety of dopamine receptor heteromers, which include D1-D2 [12, 13], D1-D3 [14, 15], adenosine A1- D1 [16], adenosine A2-D2 [17], cannabinoid CB1- D2 [18, 19], D2-somatostatin SST5 [20], D2-μ-opioid [21], D2-glutamate metabotropic mGlu5 [22], D1-histamine H3 [23] D2-H3 [24] and D1-Ghrelin Ghr1a [25] receptor heteromers. Also, recent reports indicate the existence of A2A-CB1-D2 and A2ACB1-mGlu5 receptor heteromultimers [21, 26]. Among the heteromers identified to date, the A2-D2 receptor heteromer has been already considered as a target for the treatment of PD [5, 27]. In the receptor heteromer, the agonist binding to A2A receptor decreases the dopamine affinity for D2 receptor due to an allosteric interaction in the receptor heteromer [28]. Thus, A2A receptor antagonists were suggested to be useful for the treatment of PD to counteract the endogenous adenosine-mediated inhibition of D2 receptor function and increase the therapeutic index of L-DOPA [29]. In fact, A2A receptor antagonists are under clinical evaluationfor PD therapy [30, 31]. The main underlying message is that drugs that target a specific receptor in the context of different receptor heteromers are likely to exert different actions by acting on different receptor heteromers expressed in different neurons.

POTENTIAL NEW DRUGS TARGETING GPCR HETEROMERS

Whereas the effect of a drug on a receptor in a specific receptor heteromer may be beneficial, the effect of the same drug on the same receptor subtype but on a different receptor heteromer could be detrimental. As a consequence, drugs tailored to target specific receptors on specific receptor heteromers might be a viable therapeutic design strategy in the near future: such drugs might show enhanced efficacy and provide a reduced side effect liability. Different approaches are being explored to selectively target receptor heteromers. One approach, for instance, involves screening compounds that target a specific receptor within the conformationpresent in the receptor heteromer. Thus, some compounds may be selective for a specific receptor but exhibit higher affinity if the receptor is in a specific heteromer context. An example was provided by Rashid et al., who found that the selective D1 receptor agonist SKF83959 also binds and activates the D2 receptor in the D1-D2 receptor heteromer [13]. Furthermore, the compound leads toincreases in intracellular calcium via Gq in cells expressing the D1-D2 receptor heteromer and only activates Gs if the D1 receptor is not forming heteromers with D2 receptor [13]. Therefore, SKF83959 can be considered as a receptor heteromer-selective compound. Another specific receptor heteromer ligand is 6’-guanidinonaltrindone, which has a high selectivity for μ-δ opioid receptor heteromers. 6’-guanidinonaltrindone induces analgesia by acting as a heteromer-selective compound and only when delivered in the spinal cord but not in the brain [32]. These findings demonstrate a proof of concept for tissue-selective drug targeting based on GPCR heteromerization. In this context, the targeting opioid heteromers could provide an approach toward the design of analgesic drugs with reduced side effects. For instance, the opioid μ-δ receptor heteromer seems to provide a better target than either μ or δ receptors alone, since blockade of the δ receptor decreases tolerance to the analgesic effects of the most widely used μ receptor agonist, morphine [33].

Another approach to obtain receptor-heteromer specific drugs is to develop bivalent ligands, i.e. compounds linked by a sizevariable spacer being able to bind simultaneously to the two receptors in a receptor heteromer. In this approach different possibilities may occur: A bivalent compound can be made of agonists or antagonists for both receptors in the receptor heterodimer or it can be made of an agonist at one receptor and antagonist at the other. With the idea of decreasing tolerance to the analgesic effects of μ-opioid receptor agonists, Daniels et al. developed bivalent compounds containing a μ-opioid receptor agonist and a δ-opioid receptor antagonist linked by a spacer of different sizes [34]. After intracerebroventricular administration in mice, these compounds acted as agonists with potencies ranging from 1.6- to 45-fold greater than morphine when tested in tail-flick assays. More interestingly, chronic intracerebroventricular administration studies revealed that bivalent ligands whose spacer was 16 atoms or longer produced less dependence than either morphine or a μ-opioid receptor monovalent control agonist. Both physical dependence and tolerance were suppressed when using spacer lengths of 19 atoms or greater [34]. These data support the receptors modulates μ-mediated tolerance and dependence. In summary, μ-δ-opioid bivalent ligands of precise sizes exhibit a higher potency than morphine and the potential to achieve analgesia without tolerance and dependence. Bivalent ligands for the A2A-D2 receptor heteromer, with and A2A antagonist and a D2 receptor agonist, have also been recently reported [35].

It is important to comment that μ-δ opioid receptor heteromer is a focus of debate, since a recent study questioned the colocalization of δ and μ opioid receptos [36], an obvious prerequisite for receptor heteromerization. Using recently generated δ opioid eceptor-eGFP knock-in mice [37], δ and μ opioid receptor cellular co-localization in dorsal root ganglia was carried out using anti-GFP and anti-μ receptor antibodies and was reported to be just ~5% of the neuronal population [36]. But given the increased level of δ opioid receptor expression in these knock-in mice [37] and the high avidity of the anti-GFP antibody as compared to the anti-μ opioid receptor antibody, it is most likely that the level of μ opioid receptor co-expressed with δ opioid receptor was underestimated in this study. Furthermore, previous studies have found that the GFP tag at the C-terminus affects the maturation of δ opioid receptors [38] and that the levels of δ opioid receptor attenuate the maturation of μ opioid receptors [39]. Finally, an important point to consider is that GPCR expression is altered during development and in pathology. Under these conditions, the level of receptor coexpression is likely to be significantly altered, making some particular GPCR heteromers attractive drug targets. The dopamine D1-D3 receptor heteromer constitutes a putative example of a pathology-involved receptor heteromer (see below).

TARGETING GPCR HETEROMERS WITH PATHOLOGICAL SIGNIFICANCE: THE DOPAMINE D1-D3 HETEROMER IN L-DOPA-INDUCED DYSKINESIA

It is not only important to find new drugs to target specific receptor heteromers but also to find new receptor heteromer targets with pathological relevance. The recently discovered D1-D3 receptor heteromer is a seminal example. Two recent studies report the ability of D1 and D3 receptors to heteromerize in mammalian transfected cells [14, 15]. The two complementary studies demonstrated the existence of multiple biochemical characteristics of the D1-D3 receptor heteromer. D1-D3 receptor heteromerization is constitutive and not regulated by exposure to dopamine receptor agonists [15]. Fiorentini et al. found that the D1 receptor in the D1-D3 receptor heteromer has 50 times more affinity for dopamine than when not forming heteromers with D3 receptors [15]. This is a very significant finding, since it is well known that, under normal conditions, striatal D1 receptors have a much lower affinity for dopamine than D2 or D3 receptors [11]. Heteromerization with D3 receptors could therefore be a mechanism by which striatal D1 receptors could increase their sensitivity to dopamine. An important additional biochemical characteristic of the D1-D3 receptor heteromer, discovered by Fiorentini et al., was an increase in the potency of dopamine in stimulating adenylyl cyclase (AC) through D1 receptors [15]. Under non-heteromerization conditions, stimulation of D1 receptors, by coupling to Gs/olf proteins, stimulates AC, whereas D3 receptor stimulation, by coupling to Gi/o proteins, inhibits AC [11]. Remarkably, there was no additive effect when dopamine co-stimulated D1 and D3 receptors in the D1-D3 receptor heteromer, but a clear potentiation of the effect of D1 receptor stimulation, which correlated with the increase in affinity of D1 receptors for dopamine [15]. Fiorentini et al. also found that heteromerization alters the pattern of D1 and D3 receptor trafficking. Thus, under non-heteromerization conditions in transfected cells, long-term exposure to agonists induces internalization of D1 but not D3 receptors, while stimulation of either D1 or D3 receptors does not induce internalization of the D1-D3 receptor heteromer. On the other hand, co-stimulation of D1 and D3 receptors induces internalization of the D1-D3 receptor heteromer [15]. Nevertheless, a recent study by Bether et al., suggests that D3 receptor stimulation might prevent internalization of the D1-D3 receptor heteromer in the striatum of L-DOPA-treated 6-hydroxydopamine-lesioned rats with abnormal involuntary movements [40], a rodent model of L-DOPA-induced dyskinesia. But, do D1-D3 receptor heteromers exist in native tissue, i.e. in th striatum? In transfected cells, we could demonstrate the existence of a putative allosteric interaction in the D1-D3 receptor heteromer, by which stimulation of the D3 receptor increases the affinity of the D1 receptor for agonists [14]. On the other hand, D1 receptor agonists did not modulate the binding characteristics of the D3 receptor [14]. This allosteric interaction was then used as a biochemical fingerprint to identify the D1-D3 receptor heteromer in membrane preparations from rat striatum, further confirming a substantial striatal D1-D3 receptor colocalization [14].

It is widely accepted that D1 and D2 receptor subtypes are largely segregated in the two most populated types of striatal neurons. D1 receptors are mostly expressed by the gamma aminobutyric-acidergic(GABAergic) dynorphinergic neuron, which also expresses substance P (SP), while D2 receptors are mostly localized in the GABAergic enkephalinergic neuron [41]. D3 receptors are less abundant that D1 and D2 receptors and they are mostly colocalized with D1 receptors in the GABAergic SP-dynorphinergic neurons [42]. The biochemical characteristics of the D1-D3 receptor heteromer predicts that the main function of the striatal D3 receptor is to potentiate the D1 receptor-mediated modulation of the function of the GABAergic SP-dynorphinergic neuron [14, 15]. The evaluation of locomotor activity induced by dopamine receptor agonists in reserpinized mice is a very useful in vivo model to study the function of striatal postsynaptic D1 and D2 receptors localized in GABAergic SP-dynorphinergic and GABAergic enkephalinergic neurons without the influence of endogenous dopamine [43]. We found that in reserpinized mice a selective D3 receptor agonist (PD 128907) does not produce any locomotor activity changes on its own, but dose-dependently potentiates the locomotor activation induced by a D1 receptor agonist (SKF 39393) [14]. The effect of the D3 receptor agonist was selectively antagonized by a selective D3 receptor antagonist (ST 198) and also by genetic inactivation of the D3 receptor [14].

There is a substantial amount of preclinical data suggesting that D1 and D3 receptors in GABAergic SP-dynorphinergic neurons (which constitutes the direct striatal efferent pathway), play a keyrole in the development of L-DOPA-induced dyskinesia in PD. The working hypothesis is that dopamine denervation followed by chronic L-DOPA treatment leads to an increased D1 receptor signaling promoted by an up-regulation of D3 receptors, which produces an imbalance in basal ganglia processing due to an increased activation of the direct pathway [40, 44–48] (Fig. 1). Results from experiments in monkeys have clearly shown that the development of L-DOPA-induced dyskinesia does not correlate with changes in the density of striatal D1 (or D2) receptors, although there is an increased D1 receptor-mediated neurotransmission in the direct pathway [46, 47]. For instance, dyskinetic animals show a selective increase in D1 agonist-induced [35S]GTPgammaS binding [46, 47]. L-DOPA-induced dyskinesia in monkeys and L-DOPA-inducedbehavioral sensitization in unilateral 6-hydroxydopaminelesioned rats (another rodent model of L-DOPA-induced dyskinesia) are associated with a parallel increase in the density of striatal D3 receptors [44, 45, 47].

Fig. 1.

Scheme of basal ganglia function and a proposed mechanism of L-DOPA-induced dyskinesia in Parkinson’s disease. The GABAergic striatal efferent neurons are subdivided into enkephalinergic and substance P-dynorphinergic neurons, which give rise to two striatal efferent systems that connect the striatum with the output structures of the basal ganglia: the substantia nigra pars reticulata and the entopeduncular nucleus (internal segment of the globus pallidus in primates). These are called “direct” and “indirect” efferent pathways. The direct pathway consists of substance P-dynorphinergic GABAergic neurons, which directly connect the striatum with the output structures. The indirect pathway originates with enkephalinergic GABAergic neurons, which connect the striatum with the globus pallidus (external segment of the globus pallidus in primates). GABAergic neurons of the globus pallidus connect with glutamatergic neurons of the subthalamic nucleus, which connect to the basal ganglia output structures. Motor activation results from the counterbalanced influence of the direct and indirect pathways on the neuronal activity in the output structures, since stimulation of the direct pathway results in motor activation and stimulation of the indirect pathway produces motor inhibition. (a) Under normal conditions dopamine induces motor activation by simultaneously activating the direct pathway (acting on stimulatory D1-like receptors localized in substance P-dynorphinergic neurons) and depressing the indirect pathway (acting on inhibitory D2–like receptors localized in enkephalinergic neurons). (b) After dopamine denervation (as in PD), there is an opposite effect, motor depression, due to increased and decreased activation of the indirect and direct pathway, respectively. (c) After dopamine denervation, L-DOPA treatment induces a normalization of the activity of the indirect pathway, but, also leads to an increased expression of D3 receptors and D1-D3 receptor heteromers in the dynorphinergic neurons, which results in an abnormal increase in the activity of the direct pathway. This can be a main pathogenetic mechanism of L-DOPAinduced dyskinesia. DA: dopaminergic neuron; GABA: GABAergic neuron; GLU: glutamatergic neuron; EPN: entopeduncular nucleus; GP: globus pallidus; SNc: substantia nigra pars compacta; SNr: substantia nigra pars reticulata; STN: subthalamic nucleus; THAL: thalamus.

The above-mentioned experimental results can be best explained by an increased formation of D1-D3 receptor heteromers in the GABAergic SP-dynorphinergic neurons after dopamine denervation followed by chronic treatment with L-DOPA (Fig. 1). First, an upregulation of D3 receptors without a parallel upregulation of D1 receptors would lead to an increase in thefraction of D1 receptors forming heteromers with D3 receptors. Second, and as mentioned before, D3 receptor stimulation seems to prevent the internalization of D1-D3 receptor heteromers [40]. Third, the biochemical characteristics of the D1-D3 receptor heteromer would perfectly explain an increase in D1 receptor signaling that is not secondary to D1 receptor upregulation. Recent in vivo functional neuroimaging studies in rodent and monkey models of L-DOPA-induced dyskinesia also support this hypothesis. Thus, administration of a D3 receptor agonist induced a significant increase in the relative cerebral blood volume in the striatum of dyskinetic animals but not in naïve or parkinsonian animals, an effect qualitatively identical to that produced by D1 receptor agonists [48]. Significantly, treatment with selective partial D3 receptor agonists or antagonists attenuates L-DOPA-induced dyskinesia in MPTP-treated monkeys [49, 50]. The D1-D3 receptor heteromer therefore emerges as a target for L-DOPA induced dyskinesias in patients with PD.

CONCLUSIONS AND FUTURE REMARKS

Now that the concept of receptor heteromerization is beginning to be accepted, it will become more obvious that receptor heteromers play important roles in CNS function and dysfunctionand that they can be used as targets for therapeutic interventions in neuropsychiatric disorders. We are beginning to obtain clear preclinical evidence for the therapeutic potential of heteromer-selective compounds and in the present review we have given some clues about the approaches to obtain receptor-heteromer specific drugs. We have also underlined the need to find receptor heteromers that are involved in the pathogenesis of diseases. We have provided an example of the newly discovered D1-D3 receptor heteromer that nicely exemplifies the points just raised. The experimental data strongly suggest that this heteromer is involved in the basal ganglia dysfunction that underlies L-DOPA-induced dyskinesia in PD. The next step should be using the D1-D3 heteromer as a whole protein target against which to direct drug discovery efforts. This will first require a clear understanding of the biochemical and biophysical characteristics of the receptor heteromer. The initial step to identifying a selective ligand for the heteromer could involve screening of dopaminergic compounds in mammalian cells transfected with both D1 and D3 receptors. Medicinal chemistry directed by initial chemical hits could then be directed at the optimization of molecules with selectivity for the heteromer. Such a selective research tool would enable further understanding of the biochemical pharmacology of D1-D3 heteromers that could lead ultimately to the discovery of drug-like molecules with potential in the therapeutic management/preventionof L-DOPA-induced dyskinesia.