Dopamine receptors are membersof the Class A G-protein coupled receptors (GPCRs) superfamily. GPCRs,also known as seven-transmembrane domain receptors (7TM receptors),are protein receptors that mediate most of the physiological responsesto many hormones, neurotransmitters, etc., and constitute the largestclass of drug targets. The dopamine receptors are further dividedinto five subtypes, included in two families. The D₁-likefamily of receptors contains D₁ and D₅, whilethe D₂-like family contains D₂, D₃, and D₄.^1,2 The dopamine receptors are associatedwith numerous neurological processes including memory, learning, motivation,pleasure, and cognition; and as such, they are familiar drug targets.² Included in the list of disorders linked to dysfunctionof dopaminergic signaling is schizophrenia,³⁻⁵ attention-deficithyperactivity disorder (ADHD),⁶⁻⁸ Parkinson’s disease,^9,10 and drug^2,11 and alcohol^12,13 dependence.Recently there has been mounting evidence linking elevations in synapticdopamine levels with the reinforcing effects of cocaine and henceits abuse potential.¹⁴⁻¹⁶

Cocaine is a powerful stimulant made from theleaves of the cocaplant and produces short-term euphoria and energy bursts. Accordingto the National Survey on Drug Use and Health (NSDUH), there are ∼2M current cocaine users in the US with young adults (18–25years old) representing the largest population of users. Unfortunately,there are no approved treatments for cocaine dependence.¹⁷ Cocaine does not directly bind to the D₁ and D₂ receptors, but rather binds to the dopaminetransporter, thereby increasing synaptic levels of dopamine and itsdownstream effects on D₁ and D₂ receptors thatenables the cocaine reinforcement effects. The involvement of D₄ receptors as another potential target for cocaine reinforcement/dependenceis due to its tissue distribution in the limbic and cortical brainregions implicated in cocaine addiction.^3,18 In fact, thedopamine D₄ receptor has been coined the “adventuregene” due to the higher novelty seeking scores in individualsgrouped by the long, polymorphic repeat region in exon III of theD4DR (L-D4DR) gene.^19,20 Although the notion of the adventuregene has been questioned,^21,22 further data has suggestedthat D4 has a role in severity of dependence.²³ While there have been numerous studies on the role of dopamine D4,the field has been hampered by the lack of selective D₄ receptor antagonists.² Herein, we reportthe discovery and characterization of a potent and selective dopamineD₄ receptor antagonist, ML398.

Recently we reportedan enantioselective synthesis of a chiralmorpholine analogue via an organocatalytic α-chlorination ofaldehydes followed by cyclization to form the morpholine scaffold.²⁴ As part of the previous report, we synthesizeda known dopamine D₄ antagonist,1, in orderto confirm the stereochemistry of the active enantiomer (Figure1).²⁵ Biological evaluationof1 revealed the (R)-enantiomer wasthe active isomer with a D₄ IC₅₀ = 180 nM andK_i = 70 nM. (R)-1 was inactive against dopamine D₁ and D₂ (>100μM) and weakly active against D₃ (IC₅₀ = 46.2 μM andK_i = 15.7 μM).The (S)-1 was inactive against D₄ (>100 μM), and the racemic-1 was lessactive (IC₅₀ = 360 nM andK_i = 140 nM). On the basis of the activity and binding of (R)-1, we embarked on a medicinal chemistrycampaign to explore the structure–activity relationship (SAR)around the benzimidazole portion of the molecule.

In an effort to evaluatea number of compounds, the racemic phenethylmorpholine was synthesized and then rapidly analogued and assayedat D₄. Active compounds would then be assayed in enantiopureform. The racemic synthesis of the initial analogues for the SAR evaluationis outlined in Scheme1 and is identical tothe previous work, except for the organocatalyst.²⁴ The racemic synthesis utilizesdl-proline as thecatalyst to obtain the racemic α-chloro aldehyde,3. Next, the alcohol was activated (Tf₂O, lutidine) anddisplaced with the amino alcohol to yield4. Next, themorpholine ring was formed via an intramolecular cyclization and thenthe benzyl protecting group was removed to yield the common intermediate,5. The benzylic analogues in Table1 were synthesized by alkylation of the morpholine nitrogen (R-Br,K₂CO₃). The amide analogues were synthesizedvia the corresponding acid chloride (DIPEA, DMF, rt, 1 h), the directaryl compounds were synthesized via a palladium catalyzed cross-coupling(Pd₂(dba)₃, XANTPHOS, Cs₂CO₃, and 1,4-dioxane), the urea analogues were made via the aryl isocyanate(THF, rt, 1 h), and the sulfonamides were made via reaction with theappropriate sulfonyl chloride (DIPEA, DMF, 40 °C, 1 h).

The first set of analogues weresubstituted benzyl analogues (6a–6h), which all showed varying degreesof dopamine D₄ receptor inhibition (51–98% at 10μM) (Table1). Interestingly, the pyridylanalogues (2-, 3-, and 4-pyridylmethyl) were all inactive (6i–6k). Introduction of an amide group (thus changingthe linker group from a methylene to a carbonyl) was not toleratedand led to inactive compounds (6l). Deletion of the linkergroup (direct arylation of the morpholine) was also not tolerated(6m and6n). Lastly, the urea analogue (6o) and sulfonamides (6p) were also inactive,potentially suggesting the need for a flexible or rotatable bond betweenthe morpholine and the right-hand substituent (SeeSupplemental Table 1 for full SAR).

Next, the activebenzyl analogues were progressed into the IC₅₀ andK_i determinations, and clearSAR trends emerged (Table2). First, orthosubstituents were less tolerated than themeta- andpara-substituents (even the small fluorine analogues) (6d and6h). Second, thepara-fluoro (6f) and unsubstituted benzyl (6g) were less active. Lastly, the most potent compounds contained ameta- orpara-trifluoromethoxybenzyl group(6a,6e), apara-methoxybenzylgroup (6b), or apara-chlorobenzyl group(6c). On the basis of these results, both enantiomersof6a–c were synthesized using thepreviously published route.²⁴ Both the(R)- and (S)-enantiomers were evaluated,and as previously determined, the (S)-enantiomerswere inactive. All three of the compounds tested were potent antagonistsof the dopamine D₄ receptor with high binding affinities((R)-6a, IC₅₀ = 230 nM,K_i = 65 nM; (R)-6b, IC₅₀ = 100 nM,K_i = 28 nM;(R)-6c, IC₅₀ = 130 nM,K_i = 36 nM). On the basis of these results, wefurther profiled (R)-6b and (R)-6c for their selectivity against the otherdopamine receptors, and (R)-6c was inactive(>20 μM) against all of the dopamine receptors tested. Thus,on the basis of potency, binding, and selectivity profile, (R)-6c was declared an MLPCN probe (ML398).

ML398 was further profiled in a battery of Tier 1in vitro DMPK assays (Table3). Theintrinsic clearancewas assessed in hepatic microsomes (rat and human), and ML398 wasshown to be unstable to oxidative metabolism and predicted to displayhigh clearance in both species. In addition, using an equilibriumdialysis approach, the protein binding of ML398 was evaluated, andit was shown to have good free fraction in both species (6.1% in humanand 3.9% in rat). ML398 was evaluated for its inhibition of the cytochromeP450 (CYP) enzymes using a cocktail approach in human liver microsomesas a screen for potential drug–drug interaction liability.ML398 displayed no significant activity against the panel of CYPs.ML398 was also profiled in anin vivo tissue distributionstudy (plasma and brain levels). Because of the predicted high clearance,ML398 was dosed at a single dose via intraperitoneal (IP) route ofadministration with a suspension formulation and then evaluated ata single time point. ML398 readily crosses the blood–brainbarrier with a B/P ratio of ∼2 and total brain concentrationsof ∼1 μM. Lastly, ML398 was tested using EuroFins LeadProfiling screen (radioligand binding assay panel of 68 GPCRs, ionchannels, and transporters screened at 10 μM). ML398 was foundto not significantly interact with 63 of the 68 assays performed (<50%binding at 10 μM) (SeeSupporting Information). ML398 did have activity against five targets (adrenergic, α₁A (77%); histamine, H₁ (93%); sigma, σ₁ (99%); dopamine transporter, DAT (72%); norepinephrine transporter,NET (68%).

Having identified a potent, selective, and brain penetrantD₄ antagonist, we wanted to test its ability to reversehyperlocomotioninduced by cocaine. It is believed that cocaine increases locomotoractivity by increasing synaptic concentrations of dopamine by blockingdopamine reuptake or by enhancing the release of dopamine. This, inturn, increases stimulation of postsynaptic dopamine receptors. Theincreased locomotor activity can be blocked by selective dopamineantagonists as well as haloperidol.²⁷ Thecocaine-induced hyperlocomotion assay is used to establish PK/PD relationship.The effects of both (R)-1 and ML398were evaluated in this assay, and the results are shown in Figure2. Cocaine was shown to significantly induce hyperactivityin rats, which is characterized by an increase in the number of beambreaks using the SmartFrame open field activity chambers. Both testcompounds were dosed via IP administration at doses of 3 mg/kg and10 mg/kg. Although (R)-1 showed a lineartrend of reversal, the data was not significant. However, ML398 didshow a statistically significant reversal of cocaine-induced hyperlocomotionat the highest dose tested (10 mg/kg).

In conclusion, we have identified a new, potent, and selectivedopamine D₄ antagonist based on a chiral morpholine scaffold.As many of the previous D₄ antagonists contain a piperidinemoiety, the reduced basicity of the morpholine may help contributeto the unprecedented selectivity. The SAR studies showed that thebenzylic substitution is optimal as the amides, ureas, and arylationanalogues were all inactive. In addition, the (R)-enantiomerwas confirmed as the active isomer within this series. ML398 is >100-foldselective versus the other dopamine receptors and is highly brainpenetrant. ML398 also was shown to reverse cocaine-induced hyperlocomotionin rats. Further optimization studies are ongoing in an effort todiscover an improved molecular probe for biological study as welland development of a radioligand for the D₄ receptor.