Dopamine and Dopamine-Related Ligands Can Bind Not Only to Dopamine Receptors

The dopaminergic system is one of the most important neurotransmitter systems in the central nervous system (CNS). It acts mainly by activation of the D1-like receptor family at the target cell. Additionally, fine-tuning of the signal is achieved via pre-synaptic modulation by the D2-like receptor family. Some dopamine drugs (both agonists and antagonists) bind in addition to DRs also to α2-ARs and 5-HT receptors. Unfortunately, these compounds are often considered subtype(s) specific. Thus, it is important to consider the presence of these receptor subtypes in specific CNS areas as the function virtually elicited by one receptor type could be an effect of other—or the co-effect of multiple receptors. However, there are enough molecules with adequate specificity. In this review, we want to give an overview of the most common off-targets for established dopamine receptor ligands. To give an overall picture, we included a discussion on subtype selectivity. Molecules used as antipsychotic drugs are reviewed too. Therefore, we will summarize reported affinities and give an outline of molecules sufficiently specific for one or more subtypes (i.e., for subfamily), the presence of DR, α2-ARs, and 5-HT receptors in CNS areas, which could help avoid ambiguous results.


Introduction
The dopaminergic system is one of the most important neurotransmitter systems in the CNS. Dopamine receptors (DRs, see Abbreviations for abbreviation list) belong to G protein-coupled receptor (GPCR) family. According to their structural similarities, DRs are divided into two groups (for a review, see [1]): D 1 -like (D 1 and D 5 subtypes) and D 2 -like (D 2 , D 3 , and D 4 subtypes). The families of DRs differ in the coupling to G proteins and subsequent steps of intracellular signalization. While D 1 -like DRs activate adenylyl cyclase via G s protein, the D 2 -like family (mainly pre-synaptic D 2 DRs) inhibits adenylyl cyclase via G i protein activation. However, in detail, D 1 -like DRs activate not only adenylyl cyclase but also increase phosphoinositide metabolism [2]. Similarly, coupling with G q protein allows D 2 DRs to activate phospholipase C (see note about receptor variants below). D 1 -like receptors are characterized by non-simple interactions with various other mediators and receptor systems, which can be activity-dependent, comprise heterological oligomerization, dynamic compartmentalization of signaling components, and system integration for exquisite functional regulation (see [2] for detail). The adenylyl cyclase response is associated with the D 1 subtype, while the phosphoinositide responses may be preferentially mediated through stimulation of the D 5 receptor [2].
The genes for D 1 -like and D 2 -like families differ in the presence of introns in their coding sequence. While the D 1 -like family does not contain introns [3,4], the D 2 -like family does [5][6][7][8]. This fact allows the generation of receptor variants, "long" and "short" D 2 receptor isoforms. These two isoforms exhibit largely similar pharmacological characteristics, but their differences in G protein coupling [9] suggest different functions [10].

D 1 -like Family
D 1 -like family is the main element of the dopamine post-synaptic action (despite its pre-synaptic localization). Its members, D 1 and D 5 DRs, are pharmacologically indistinguishable. However, the affinities of D 5 DR to the agonists are up to 10 times higher than that of D 1 ones [11]. This fact could be of importance when one transmitter is supposed to have two effects-one through the high-affinity sites and the second one through the low-affinity sites in tissue expressing both subtypes. This could explain the different functions of striatal D 1 and D 5 DRs in synaptic plasticity [12]. Another difference between these two subtypes that is interesting to mention is that the D 5 dopamine receptor, unlike the D 1 subtype, is constitutively (agonist-independently) active [13]. Moreover, D 1 DRs couple preferentially to G protein heterotrimers that contain γ7 subunits [14]. D 1 DRs can also couple to another G protein, G olf (which also stimulates adenylyl cyclase) that is highly expressed in some brain areas, such as the caudate nucleus, nucleus accumbens, and olfactory tubercle. Some coupling of D 1 DR with G olf was even suggested to be preferential [15]. The generation of D 5 DR knockout mouse uncovered possible involvement of this subtype in the pathology of hypertension, as the mutant mice were hypertensive [16].

D 2 -like Family
D 2 DRs are as D 1 DRs [17] localized both pre-and postsynaptically. D 2 DR has a relatively low (nanomolar) affinity for dopamine, which supports its importance as a modulatory (pre-synaptic) receptor. D 2 DR isoforms (long and short) are differently distributed and thus may possess distinct functions. The short isoform seems to serve as an autoreceptor, whereas the long isoform is primarily a post-synaptic receptor [18]. Using genetically targeted deletion of the D 2 dopamine receptor gene in mice revealed that other members of the receptor family were not affected [19] and these mutants had reduced locomotion and less coordinated movement [19]. D 3 subtype of DR appears to have similar distribution as the D 2 dopamine receptor [1]. Similar to D 2 DR, alternative splicing variants of D 3 DR were observed. These variants were hypothesized to contribute to the availability of active D 3 DRs in some psychiatric conditions [20]. This hypothesis suggests that inactive D 3 DRs affect ligand binding to the active D 3 DRs and thus influence their function.
The D 4 DR has high densities in the cerebral cortex, amygdala, hypothalamus, and pituitary [21]. In the striatum, the occurrence of the D 4 DR is much lower than the D 1 and D 2 subtypes [22].
It is necessary to mention (please see the values in this review) that binding assessed parameters (i.e., pK i s) differ from the values determined using functional studies (i.e., dose-response determined constants, pEC 50 s [28]). This is because in studies based on dose-response determined parameters; the ligand usually discards the presence of other receptors on the studied effect by a combination of pharmacological means to attribute properly the receptor involved. Another possibility is that in dose-response studies, the formation of a ligand-receptor complex with activation of G protein and further with target second messenger producer activation is more complicated than the binding of ligand to the receptor in binding studies. The interesting correlation between pK i and pEC 50 has been demonstrated for neurokinin NK 1 receptors [29]. Although this is a specific example for specific receptors and specific ligands, we can assume that a similar correlation can be found for DRs and their ligands too. As reported here, the pK i s and pEC 50 s differ for D 1 -like DR to SKF 38393. With some methodological reservation, one could construct the correlation between these values reported in [13,23,[30][31][32][33][34] in humans and rats.
A similar multitarget binding can be found for DR agonists and antagonists. This review will focus on such interactions that can broaden the physiological effects elicited by dopamine ligands in the central nervous system. Besides, these interactions could present the potential problem with results interpretation: the ligand activating more neurotransmitter receptors that have similar affinity to them can distort the conclusions made. With this point of view, this review could help with careful interpretation of the results obtained. We will focus on orthosteric binding sites only, although there are also described allosteric binding sites on D 2 DR [35]. The allosteric binding sites [36,37] and their interaction with other molecules exceed the topics of this review. The inclusion criteria were the ability to bind to other targets with pK i ≥ 7.0, pK IC50 ≥ 7.0 if the pK i for DRs is between 8 and 9. Interestingly, some papers report a surprisingly high concentration of drugs used as proof of specific dopamine subtype involvement even though the selectivity of such ligand is limited (e.g., SKF 38393 in concentration 100 µmol/L affects all dopamine receptor subtypes and also α 2C -AR). If the ligand is sufficiently specific to dopamine receptors (i.e., the affinity differs at least two orders of magnitude), then it is not reviewed here.
The interested researcher should search available databases carefully for the ligands with well-documented selectivity to specific DR subtypes and not rely on the information from the manufacturer. The specific ligand should be at least two orders of magnitude more specific for the respective DR subtype than to the others. In other words: ∆pK i s(pK i1, pK i2 ) ≥ 2. The examples of such ligands are shown in Table 1. On the other hand, the new research can bring new knowledge, and the supposed selectivity of the specific ligand could be doubted. Thus, it is necessary, before the choice of ligand, carefully check the present knowledge to avoid the use of non-specific ligands.  1,9 1 The selectivity is expressed to D 1 -like DRs. 2 Biased D 2 DR agonist [38]: it antagonizes arrestin recruitment to D 2 DR but behaves as an agonist in its capacity to induce D 2 DR signaling. 3 D 2 DR and D 3 DR selective over D 4 DR. 4 D 2 DR and D 3 DR selective. 5 The selectivity is expressed to D 2 -like DR. 6 D 3 DR selective over D 2 DR. 7 Slightly more selective to D 4 DR than to D 2 DR. 8 Selectivity D 4 DR > D 3 DR > D 2 DR. 9 Please note that this is radioligand.
When using radioligand for receptor detection, one should be aware that a better option is to use an antagonist than an agonist because of stronger binding and lower possibility of dissociation of such ligand from the receptors.

So-Called Selective Dopamine Receptor Agonists
The typical problem with dopamine ligand lies in the fact that manufacturers usually declare the ligand as selective, which could be, in some cases, far from reality. This could be misleading, and it could distort the conclusions made with such a "selective" drug. In the following paragraphs, we will describe the DR agonist in which the selectivity is limited. Other ligands that are selective according to present knowledge will not be mentioned.
We can generalize that dopamine drugs (both agonists and antagonists) bind in addition to DRs also to α 2 -ARs and 5-HT receptors. Thus, it is important to consider the presence of these receptor subtypes in specific CNS areas as the function virtually elicited by one receptor type could be the effect of other-or the co-effect of multiple receptors. The presence of neurotransmitter receptors in the CNS is shown in Table 2. In addition to that, dopamine ligands often bind to H 1 histamine receptors. These receptors are present in many CNS structures [39]: cerebral cortex, hippocampal dentate gyrus, amygdaloid complex, basal forebrain, nucleus accumbens, islands of Calleja, septal nuclei, thalamus, hypothalamus (medial preoptic area, dorsomedial, ventromedial, and most posterior nuclei, including the tuberomammillary complex), nuclei of origin of most cranial nerves, and in the dorsal horn of spinal cord. Table 2. The co-presence of receptor types in specific brain areas.

CNS Area DR Presence α 2 -AR Presence 5-HT Presence
The presence of specific receptor types was referred to by [1,[40][41][42][43][44]. D 1 -like means the presence of D 1 DRs and D 5 DRs,.D 2 -like means the presence of D 2 DRs, D 3 DRs, and D 4 DRs. The presence of receptors in the cerebral cortex can be more specific to layers, part of the cortex, etc. Please see [1,[40][41][42][43][44] for detail. 1 Referenced as a presence of subtype in basal ganglia (no further specification). 2 mRNA expression only does not necessarily mean the presence of receptors binding sites. 3 Specifically in the dorsomedial hypothalamus and the paraventricular nucleus. 4 Generally in the hypothalamus, specifically in the suprachiasmatic nucleus. 5 Low autoradiography detected levels. 6 Referenced as a presence of subtype in substantia nigra (no further specification). 7 Specifically in the putamen.
As an example, we can use SKF 38393. One of the manufacturers claims that this is a prototypical D 1 -like DR selective partial agonist. The careful search for pK i values (pEC 50 values, respectively, see the discounts in Section 1.3), however, can indicate pK i = 6.41-6.8 [13,23,30] in human, pK i = 7.19 in rat [31], pEC 50 = 5.0-8.96 in human for D 1 DR [32,34], pK i = 6.91-7.0 for D 5 DR in human [23,33], and pK i = 5.16 for D 2 DR in rat [31]. These values indicate selectivity to D 1 -like DRs, but still show some effect on D 2 DR. More importantly, SKF 38393 is also bound by α 2C -AR with pKi = 7.08 [45], i.e., in the rank in which D 1 and D 5 DRs are activated. This is important in tissues in which are DRs and ARs co-expressed (see Table 2). D 1 -like DRs are present [40] together with α 2C -ARs [41] in the following brain areas: the cerebral cortex and amygdala. In general, α 2C -ARs presence is described in the basal ganglia, and D 1 DRs are abundantly present in the subthalamic nucleus and caudateputamen. The D 2 DRs (although they have a lower affinity to SKF 38393) are simultaneously present in α 2C -ARs in the substantia nigra pars compacta and the ventral tegmental area. In those brain areas, one should be careful when interpreting the results obtained with SKF 38393 as both effects on DRs and α 2 -ARs can be present. Ignoring the fact that SKF 38393 activates D 1 -like DRs and blocks α 2C -ARs could lead to misinterpretation of the results.
Another "selective" D 1 DR ligand is the partial agonist A68930, although also designated as sub-family selective. This compound was reported to have a similar effect on rat D 1 and D 5 DRs (pEC 50 = 6.82 and 6.6, respectively, [46]). The other data showed higher pEC 50 at D 1 DRs in the rat (pEC 50 = 8.71, when pK i = 8.8 [47]). This study also determined pK i = 6.09, and pEC 50 = 4.99 at D 2 DRs in the rat. This drug also binds to 5-HT 1A , 5-HT 2C serotonin receptors, and β 1 -ARs with pK i = 5.59, 5.0, and 5.0, respectively [47]. Although the affinity of 5-HT 1A , 5-HT 2C serotonin receptors, and β 1 -ARs is lower than D 1 -like DRs (when considering the data from [47]), the data from [46] are quite similar, and one should be cautious with the interpretation of the results obtained with this drug.
Sumanirole (PNU-95,666) is assumed as a highly selective D 2 DR full agonist, the first of its kind to be discovered [55] with D 2 DR pK i = 8.1 [56]. 5-HT 1A receptor reveals pK i = 7.14 [57] to sumanirole, which is too close to the pK i for D 2 DR and co-effect should exist. There is also agonist activity of sumanirole at human D 3 DR transfected in HEK293T cells, revealing pK i = 6.73 [58], suggesting slightly limited selectivity of sumanirole on D 2 DR. It means that 50% of D 2 DRs are occupied by approximately 8 nmol/L sumanirole and 50% of D 3 DRs are occupied by approximately 189 nmol/L sumanirole. 20 nmol/L should completely block D 2 DRs, but also 10% of D 3 DRs.

Drugs-Dopamine Receptor Agonists with Multiple Targets of Action
Usually, the drugs used in the treatment have multiple targets of action, which can be an advantage as multiple targets are affected by one drug. In the following paragraphs, we will mention the drugs that: (1) also have DRs action, (2) are declared as a drug with multiple targets. This could help in the interpretation of the effects obtained with this drug that could be erroneously attributed to one target only.
An example of such a drug is fenoldopam, which causes arterial/arteriolar vasodilation decreasing blood pressure. Fenoldopam is used for the in-hospital, short-term (up to 48 h) management of severe hypertension, including malignant hypertension. It is declared as an agonist for D 1 DRs with moderate affinity to α 2 -ARs and no significant affinity for D 2 DRs, α 1 and β-ARs, 5-HT 1 and 5-HT 2 receptors, or muscarinic receptors.
However, fenoldopam is also bound with similar affinity to D 5 DR (pK i = 9.1 for D 1 DR, pK i = 9.2 for D 5 DR, respectively) and D 2 DR (pK i = 8.5), and with lower affinity to D 4 DR (pK i = 6.8) [59]. Some data indicate pK i to D 2 DR is lower (4.89-5.89, [60]). Early evidence showed that fenoldopam had no effect on β-ARs, but had antagonistic activity on α 1 -ARs [61] (pA2 = 8.36 ± 0.21), although in some papers characterized as weak (pK i = 5.41, [62], or modest pK i = 6.82 [26]) and α 2 -ARs [63] (pK i = 7.60-7.78, [62]). Fenoldopam thus represents the typical multiple targets drug. This is a disadvantage with respect to the specific effect of receptors when aiming to determine the subtype involved in the function but could be an advantage when targeting to specific therapeutic aim (e.g., acute severe hypertension treatment).
A wide spectrum of action also reveals cabergoline which is an ergot-derived, longacting D 2 DR agonist and prolactin inhibitor. However, the D 2 DR selectivity is rather declared than it corresponds to the reality. This drug binds, besides to DRs, to other receptor proteins [50]: D 2 DRs and D 3 DRs bind this drug with similar affinity as a partial agonist (pK i = 9.0-9.2, and pK i = 9.1 for D 2 DR and D 3 DR, respectively), similar affinity reveal 5-HT 2B receptors (pK i = 8.9, full agonist) and very close affinity show 5-HT 2A and 5-HT 1D (pK i = 8.2 and pK i = 8.1, respectively for 5-HT 2A (full agonist) and 5-HT 1D receptors [partial agonist]). On the other D 2 -like DRs (D 4 DR) it also behaves as a partial agonist, but the affinity is lower (pK i = 7.3). Besides these effects cabergoline acts also as an antagonist on α 2A -AR, α 2C -AR, α 2B -AR, and α 1A -AR (with pK i = 7.9, pK i = 7.7, pK i = 7.1, and pK i = 7.1, respectively on α 2A -AR, α 2C -AR, α 2B -AR, and α 1A -AR) and as a full agonist on 5-HT 1A receptor (pK i = 7.7) [50]. One should be cautious when thinking about the D 2 DR or D 2 -like selectivity. Although about 1.5 order of magnitude difference (pK i about 9.0 for D 2 DRs), the affinity of D 1 -like receptors could still play a role in the action of cabergoline: on D 5 DR it behaves like a full agonist with pK i = 7.7, on the D 1 DR it reveals a similar type of action (full agonism), but the pK i = 6.7 is significantly lower [50]. The affinity (full agonism) of 5-HT 1B and 5-HT 2C is much lower than the affinity of other receptors (pK i = 6.3 and pK i = 6.2, respectively) [50].
The drug with declared multiple effects is apomorphine, historically used to relieve anxiety and craving in alcoholics, as an emetic, or in treating erectile dysfunction. Currently, apomorphine is used in the treatment of Parkinson's disease but should be used together with antiemetics. Contrary to its name, apomorphine does not contain morphine or its skeleton, nor does it bind to opioid receptors. It is declared as a non-selective dopamine agonist which activates both D 2 -like and, to a much lesser extent, D 1 -like receptors, an antagonist of 5-HT 2 and α-AR with high affinity. In detail, D 4 DR binds this compound as a partial agonist with pK i = 8.4 [50], rat and human D 3 DR binds this compound as a partial agonist with pK i = 7.7 [7], and with pK i = 6.1-7.6 [50], respectively. Rat and human D 2 DRs bind this compound as a partial agonist with pK i = 7.6 [7], and pK i = 5.7-7.5 [50], respectively. α 2C -AR binds this compound as an antagonist with pK i = 7.4 [50], α 2B -AR binds this compound as an antagonist with pK i = 7.2 [50], D 5 DR binds this compound as a partial agonist with pK i = 6.4-7.8 [50], 5-HT 2C receptors bind this compound as an antagonist with pK i = 7.0 [50], 5-HT 1A receptors bind this compound as a partial agonist with pK i = 6.9 [50], 5-HT 2A receptor binds this compound as an antagonist with pK i = 6.9 [50], 5-HT 2B receptor binds this compound as an antagonist with pK i = 6.9 [50], α 2A -AR binds this compound as a partial agonist with pK i = 6.9 [50]. All these values, except stated otherwise, come from human receptors.

So-Called Selective Dopamine Receptor Antagonists
An example of a drug declared as D 1 (or D 1 -like family, pK i = 8.4 for D 1 DR) selective antagonist is flupentixol [13]. However, this antagonist also affects σ 3 -receptors [72] (pK i = 8.86). In addition to that, this ligand also antagonizes the D 2 -like family (pK i = 8.82 for D 2 DR, and pK i = 8.96 for D 3 DR, respectively) [73].
Domperidone, acting peripherally, as it is extensively metabolized in the liver, and has the low central nervous system penetration, is the next example of a declared specific D 2 and D 3 DR antagonist (pK i = 7.9-8.4, and pK i = 7.1-7.6, for D 2 and D 3 DRs, respectively [73]) is also able to bind to 5-HT 3A /5-HT 3B receptors with pK IC50 = 7.0 [76].
On the other hand, another radiolabeled ligand, raclopride is specific for DR and has a similar affinity to D 2 DR (pK i = 7.77 [87]) and D 3 DR (pK i = 7.82 [87]) but do not bind significantly to D 4 DR (pK i = 5.51 [87]) and also not to D 1 DR (pK i = 4.43 [87]).

Drugs-Dopamine Receptor Antagonists with Multiple Targets of Action
Similar to agonists, there are some drugs used in the treatment of psychiatric/neurological disorders with multiple targets action. One of them is blonanserin, an atypical antipsychotic for the treatment of schizophrenia [94]. The spectrum of targets is relatively close, but in addition to D 2 DRs (pK i = 9.9 [95]) it also antagonize the action on 5-HT 2A receptors (pK i = 9.1 [95]) and on D 3 DRs (pK i = 6.3 [96]). Blonanserin has a low affinity [97] for 5-HT 2C , α 1 -ARs, histamine H 1 , and M 1 muscarinic receptors but displays a relatively high affinity for 5-HT 6 receptors (pK i = 7.93) [97].
Trifluoperazine, a typical antipsychotic drug, binds to D 2 DR as an antagonist with pK i = 8.9-9.0 [67], to 5-HT 2A receptor as an antagonist with pK i = 7.9 [67], to D 4 DR as an antagonist with pK i = 7.4 [108], and to H 1 histamine receptor as an antagonist with pK i = 7.2 [67].
Loxapine is a typical antipsychotic drug that binds to a wide spectrum of targets: H 1 histamine receptor, where it acts as an antagonist with pK i = 8.2 [67], D 2 DR, where it acts as an antagonist with pK i = 7.9-8.3 [67], D 4 DR, where it acts as an antagonist with pK i = 8.1, 5-HT 2A receptor [108], where it acts as an inverse agonist with pK i = 8.1 [67], 5-HT 2C receptor, where it acts as an inverse agonist with pK i = 7.8-8.0 [67], D 3 DR, where it acts as an antagonist with pK i = 7.7 [118], 5-HT 6 receptor, where it acts as an inverse agonist with pK i = 7.4-7.6 [107], 5-HT 7 receptor, where it acts as an antagonist with pK i = 6.8-7.4 [98].
Promazine, a phenothiazine antipsychotic, binds not only to D 2 DR and D 3 DR (pK i = 6.5 and 6.8, respectively [119]) but also with similar, although not very high, affinity to H 1 histamine receptors (pK i = 5.9 [120]).

Discussion
The first thing that should be discussed is the similarity in the amino acid binding pocket of DRs with α 2 -ARs and 5-HT receptors. It is possible to deduce this statement from apparently similar affinities (pK i s) for dopamine as given in the Introduction. This is given by the similarity of neurotransmitter structures: noradrenaline, adrenaline, dopamine, and serotonin (see Figure 1). However, as mentioned above, the main role plays in the relationship between specific G protein-coupled receptors, i.e., the sequence homology in the binding pocket between dopamine, serotonin receptors, and adrenoceptors. These homologies have been well documented for the second extracellular loop, as discussed in [121].
A second fact that implies the similarities in binding pocket/amino acid homology is that other ligands that bind to the similar amino acid residues in DRs as dopamine would also affect 5-HT receptors and α 2 -ARs. The examples of such ligands were listed above both for agonists and antagonists.
In general, the length, organization, and amino acid homology in the D 1 -like DR subfamily is quite high [122]. This is the reason for so far not synthesizing specific agonists to D 5 DR (see below). The D 1 -like DRs have a shorter third intracellular loop and a longer carboxy-terminus compared to the D 2 -like DR subtypes [122]. The third intracellular loop and carboxy-terminus are not structures responsible for binding. The third intracellular loop and a carboxy-terminus play a role in the G protein binding. The receptor regions responsible for binding are transmembrane zones. More precisely, the predicted binding site of dopamine in D 2 DR is located in the top third of the 7-TM barrel involving TM domains 3-6 [123]. These authors also divided dopamine ligands into two groups according to their binding properties: first, clozapine-like bulky antagonists; and second, ligands with two aromatic or ring moieties connected by a flexible linker with a protonated amine group as in haloperidol [123]. The first group occupies the region between TM3, TM4, TM5, and TM6 (the agonist binding pocket), and the second group occupies the region between TM2, TM3, TM6, and TM7, with minimal contact with TM4 and TM5 [123]. The binding pocket of D 1 DR is slightly different comprising TM6, extracellular loop 2, TM5, and TM3 [121]. A second fact that implies the similarities in binding pocket/amino acid homology is that other ligands that bind to the similar amino acid residues in DRs as dopamine would also affect 5-HT receptors and α2-ARs. The examples of such ligands were listed above both for agonists and antagonists.
In general, the length, organization, and amino acid homology in the D1-like DR subfamily is quite high [122]. This is the reason for so far not synthesizing specific agonists to D5 DR (see below). The D1-like DRs have a shorter third intracellular loop and a longer carboxy-terminus compared to the D2-like DR subtypes [122]. The third intracellular loop and carboxy-terminus are not structures responsible for binding. The third intracellular loop and a carboxy-terminus play a role in the G protein binding. The receptor regions responsible for binding are transmembrane zones. More precisely, the predicted binding site of dopamine in D2 DR is located in the top third of the 7-TM barrel involving TM domains 3-6 [123]. These authors also divided dopamine ligands into two groups according to their binding properties: first, clozapine-like bulky antagonists; and second, ligands with two aromatic or ring moieties connected by a flexible linker with a protonated amine group as in haloperidol [123]. The first group occupies the region between TM3, TM4, D 3 DR and D 2 DR subtypes have substantial amino acid sequence homology [122]. The main aim of this review is to show that drugs declared by manufacturers as specific could be, in some cases, able to bind to other targets than to DRs. This can produce ambiguous results. Importantly, there are enough ligands with sufficient specificity for DR subtypes (see Table 1). The interested researcher should search available databases carefully for the ligands with well-documented selectivity to specific DR subtypes and no to rely on the information from the manufacturer. Nevertheless, one can experience different values for the same compound. As reported here, the affinities of 5-HT 1A , 5-HT 2C serotonin receptors, and β 1 -ARs to A68930 are similar to those of D 1 -like DRs (when considering the data from [47]), but the data from [46] are quite similar. Another example reported here is SKF 38393. The pK i values differ according to specific references in humans [13,23,30], which also vary from this value in rats. This can originate from different experimental conditions (temperature, incubation time, tissue, cell culture properties, and others). In such a case, one should be cautious with the selection of this compound for subtype determination or interpretation of results obtained with this drug in the literature. If possible, it is recommended to avoid such ligands.
However, the nature of drug properties reviewed here could be more complex. One should also consider the anatomical relationship between the terminals that release dopamine and other receptors-this concerns both 5-HT receptors and α 2 -ARs. Dopamine terminals are frequently localized in tight contact with other axons configuring a triad-a configuration in which a neuron is connected to both the pre-synaptic element and post-synaptic (usually dendritic) target. Triads are common in the hippocampus, striatum, and medial frontal cortex (for a review, see [124]). These triads can contain both dopamine and serotonin or adrenergic terminals. The first point on how the interaction between DRs and 5-HT receptors can occur is the formation of the heteroreceptor complexes of D 2 DR and 5-HT 2A receptors [125]. The heterocomplexes could explain the effects of atypical antipsychotic drugs [125]. One of the possible mechanisms is based on blocking the allosteric enhancement of D 2 DR protomer signaling by 5-HT 2A receptor protomer activation. Another mechanism by which dopamine can interact with serotonin is the release of L-DOPA as a "false (or substitute)" neurotransmitter in the serotonin synapse [126]. "False neurotransmitter" is considered as an ectopic neurotransmitter in a neuron, which replaces the normal neurotransmitter in storage vesicles. When it is the case of L-DOPA it is then able to increase the dopamine levels as L-DOPA is a dopamine precursor. Moreover, dopamine can also act as a "false neurotransmitter" in noradrenergic neurons [126].
Another aspect is given by the presence (although sometimes doubted in dopaminergic synapse) of volume transmission [127][128][129]. This type of connection allows the spreading of the neurotransmitter to a higher distance (more than 10 µm in comparison to 30-40 nm in classical synapse), affecting 200 other dopamine synapses instead of only one post-synaptic membrane in the classical synapse. This can further be the factor of cross action of dopamine.
On the other hand, we cannot consider this a problem; this is most probably the physiological role of the transmitter.
It can be deduced from Table 1 that a D 5 DR agonist does not exist to date and that the selectivity of the antagonist comprises the other member of D 1 -like family-D 1 DR. However, specific agonists (A77636, SKF-81297, and SKF-83959) exist for D 1 DR. Thus, it is possible to distinguish between D 1 DR and D 5 DR using the D 1 DR agonists.
Specific subtypes in the D 2 -like family can be distinguished using specific antagonists for D 2 DR (pipotiazine, ML321), D 3 DR (S33084, SB 277011-A, (+)-S-14297), and D 4 DR (sonepiprazole, L745870, A-381393, L741742, ML398). One should also consider the presence of off-targets (Table 2) when evaluating the role of specific dopamine receptors, as some receptors have a lower affinity to relatively selective ligand, but if the density of off-target receptors is much higher than DR that the proportion of the binding could be shifted.
Even though the attribution of a drug to be DR agonist/antagonist can also be the result of the side effect on another receptor. Thus, some drugs can primarily bind to other receptors and also reveal dopaminergic action. Examples of such drugs are some antipsychotics listed above (bromocriptine acting mainly at 5-HT receptors [50], risperidone acting mainly at 5-HT receptors [67,98], quetiapine which is H 1 histamine receptor antagonist [67], sertindole which has a high affinity to 5-HT receptors [67], and loxapine acting on H 1 histamine receptors [67]). Other drugs that could bind to DRs as to "second target" are muscarinic receptor agonists AC-260584, 77-LH-28-1, and LY-593039, which bind similarly to M 1 muscarinic receptors and to D 2 DRs [130]. Another group of drugs binds primarily to 5-HT receptors. An example of such a drug is 8-OH-DPAT (the binding of related 7-OH-DPAT is mentioned above), which is used in the tritiated form as a radioligand for 5-HT receptors. [ 3 H]8-OH-DPAT binds to 5-HT 1A receptors with high affinity (pK i = 9.33 [131]). The affinity of 5-HT 1B receptors is lower (pK i = 6.25 [132]) and corresponds to the affinity to DR (pK i = 7.07 [133]). Another compound acting on 5-HT receptors and with similar binding to DRs is iloperidone, an atypical antipsychotic drug. This compound binds to 5-HT 1A , 5-HT 6 , and 5-HT 7 receptors with pK i = 6.8-7.7 [134,135] and to D 2 DR with pK i = 7.0 [136]. Another atypical antipsychotic drug zotepine has antagonistic activity at 5-HT receptors (5-HT 1D pK i = 9.3 [103], 5-HT 2A pK i = 8.6 [103]) and on D 2 DR (pK i = 8.0 [103]), D3 DR (pK i = 8.2 [103]), D4 DR (pK i = 7.4 [103]). Besides that, zotepine also binds to H 1 histamine receptors (pK i = 9.2 [103]) and to 5-HT 6 and to 5-HT 7 with pK i = 8.9, and pK i = 8.8, respectively [98]. These examples just illustrate the complexity of the cross bunding between drugs suggested to be selective to specific receptors. The number of such interactions would increase with the increase in our knowledge on this topic. This review can also help with the interpretation of results obtained with antipsychotic drugs as it critically reviews the real binding to different targets, and the reader can compare the affinities of specific target molecules to these ligands. In Table 2 it is possible to find the presence of other receptors (subtypes of α 2 -ARs and 5-HT receptors) that can help the interpretation of data obtained with antipsychotic drugs.
We can conclude that one should be very cautious when selecting the DR ligand with the aim to determine the role of a specific DR subtype in studied CNS function. This review can help in such selection. Not only the selectivity but also the presence of typical off-targets to dopamine ligands (subtypes of α 2 -ARs and 5-HT receptors) should be considered, and finally, the new research can bring new knowledge, and the supposed selectivity of the specific ligand could be doubted.

Conflicts of Interest:
The author declares no conflict of interest.

Abbreviations
List of abbreviations and a short explanation of the terms used.

Abbreviation Explanation DR(s)
Dopamine receptor(s) ARs Adrenoceptors 5-HT Serotonin TM Transmembrane zone pK i The negative logarithm of the K i value (the molar concentration of the competing ligand that would occupy 50% of the receptors) pK D The negative logarithm of K D value (the equilibrium dissociation constant represents the concentration of radioligand occupying half of the maximum receptor population) pA 2 The measure of the potency of an antagonist, negative logarithm of the molar concentration of an antagonist that would produce a two-fold shift in the concentration-response curve for an agonist pEC 50 The negative logarithm of EC 50 value (the molar concentration of an agonist that produces 50% of the maximum possible response for that agonist). This value can vary when comparing different activation pathways