Long-term effects of aripiprazole exposure on monoaminergic and glutamatergic receptor subtypes: comparison with cariprazine

Objective This study examined the chronic effects of aripiprazole and cariprazine on serotonin (5-HT1A and 5-HT2A) and glutamate (NMDA and AMPA) receptor subtypes. In addition, the effects of aripiprazole on D2 and D3 receptors were tested and compared with previously reported cariprazine data. Methods Rats received vehicle, aripiprazole (2, 5, or 15 mg/kg), or cariprazine (0.06, 0.2, or 0.6 mg/kg) for 28 days. Receptor levels were quantified using autoradiographic assays on brain sections from the medial prefrontal cortex (MPC), dorsolateral frontal cortex (DFC), nucleus accumbens (NAc), caudate-putamen medial (CPu–M), caudate-putamen lateral (CPu–L), hippocampal CA1 (HIPP–CA1) and CA3 (HIPP–CA3) regions, and the entorhinal cortex (EC). Results Similar to previous findings with cariprazine, aripiprazole upregulated D2 receptor levels in various regions; D3 receptor changes were less than those reported with cariprazine. All aripiprazole doses and higher cariprazine doses increased 5-HT1A receptors in the MPC and DFC. Higher aripiprazole and all cariprazine doses increased 5-HT1A receptors in HIPP–CA1 and HIPP–CA3. Aripiprazole decreased 5-HT2A receptors in the MPC, DFC, HIPP–CA1, and HIPP–CA3 regions. Both compounds decreased NMDA receptors and increased AMPA receptors in select brain regions. Conclusions Long-term administration of aripiprazole and cariprazine had similar effects on 5-HT1A, NMDA, and AMPA receptors. However, cariprazine more profoundly increased D3 receptors while aripiprazole selectively reduced 5-HT2A receptors. These results suggest that the unique actions of cariprazine on dopamine D3 receptors, combined with its effects on serotonin and glutamate receptor subtypes, may confer the clinical benefits, safety, and tolerability of this novel compound in schizophrenia and bipolar mania.


Introduction
Dysfunctions in dopaminergic, serotonergic, and glutamatergic neurotransmissions are believed to play an important role in the pathophysiology and symptomatology of schizophrenia and other idiopathic psychotic disorders. [1][2][3][4] In addition to their effects at dopamine receptor subtypes, atypical antipsychotics modulate serotonergic and glutamatergic neurotransmission, which may contribute to their clinical efficacy, safety, and tolerability. 1,5,6 The atypical antipsychotics cariprazine and aripiprazole are both dopamine D 2 /D 3 receptor partial agonists; however, the drugs differ in that cariprazine is D 3 receptor-preferring and has moderate affinity for serotonin 5-HT 1A receptors. 7 Moreover, cariprazine, but not aripiprazole, demonstrated high and balanced occupancy of both D 2 and D 3 receptors at pharmacologically effective and antipsychotic doses in both animals 8 and schizophrenia patients. 9 In animal behavioral models, cariprazine showed procognitive 10 and antidepressant-like effects, 11 suggesting the potential benefits of cariprazine for both the cognitive deficits and depressive symptoms of schizophrenia. These effects of cariprazine are likely to be mediated by D 3 receptors, since they were absent in transgenic knockout mice that lack expression of functional D 3 receptors. 12, 13 We recently reported that long-term treatment with cariprazine increased D 2 and D 3 receptor levels in selected rat forebrain regions, whereas other typical and atypical antipsychotics upregulated D 2 receptors with no effects on D 3 receptors. 14 In the present study, we quantified the long-term effects of aripiprazole on D 2 and D 3 receptors and compared them to findings previously obtained for cariprazine and other antipsychotics.
Previous studies in rats have shown that long-term treatments with the atypical antipsychotics olanzapine, risperidone, quetiapine, and asenapine increased serotonin 5-HT 1A receptor levels in the medial prefrontal cortex (MPC) and dorsolateral frontal cortex (DFC) while decreasing 5-HT 2A receptor levels in the same regions. 15,16 Additionally, antipsychotic exposure altered the levels of ionotropic glutamate N-methyl-D-aspartic acid (NMDA) and α-amino-3-hydroxy-5-methyl-4isoxazolepropionic acid (AMPA) receptors in the caudate putamen, nucleus accumbens, and hippocampus. [17][18][19] However, all of the tested atypical antipsychotics displayed combined dopamine D 2 /serotonin 5-HT 2A receptor antagonism, which is a common feature of atypical antipsychotic drugs. 1,5 The chronic effects of D 2 /D 3 receptor partial agonists on serotonin and glutamate receptors require further investigation. Accordingly, we investigated the effects of long-term aripiprazole and cariprazine treatment on serotonin (5-HT 1A and 5-HT 2A ) and glutamate (NMDA and AMPA) receptor subtypes in the rat forebrain.

Treatment and tissue preparation
For 28 days, 7 groups of rats (8 animals/group) received a daily intraperitoneal injection of vehicle control (distilled water, 1 ml/kg) or active treatment of aripiprazole (2, 5, or 15 mg/kg) or cariprazine (0.06, 0.20, or 0.60 mg/kg). The selected doses were active in different behavioral paradigms. 8,20 No gross effects on motor behaviors or significant changes in body weight were observed after repeated treatment with the different doses of aripiprazole or cariprazine compared with vehicle-treated animals.
At the end of the treatment period, rats were decapitated 24 hours after the last injection; their brains were quickly removed, frozen in chilled isopentane, and stored at -80°C. Brain regions assessed included the medial prefrontal cerebral cortex (mPFC), dorsolateral frontal cortex (DFC) (3.2-4.2 mm anterior to bregma), nucleus accumbens (NAc) (1.7-2.2 mm anterior to bregma), medial caudate putamen (CP-M) and lateral caudate putamen (CP-L) (0.7-1.2 mm anterior to bregma), hippocampus (HIPP), and entorhinal cortex (0.2-0.7 mm anterior to bregma). 21 The selected extrapyramidal, limbic, and cortical regions of interest mediate the cognitive, emotional, and motor behaviors that are typically disturbed in patients with schizophrenia and other psychotic disorders. 1

Receptor autoradiography and image analysis
Coronal sections (10 μm) of these samples were cut in a cryostat at -20°C, mounted on gelatin-coated microscope slides, and stored at -80°C. To minimize experimental variability, brain sections from all vehicle-and drug-treated animals used for a given receptor subtype assay were evaluated at the same time.

Dopamine receptor autoradiography
For the D 2 and D 3 assays, sections were first preincubated for 1 h at room temperature (RT). We have previously shown that this preincubation step is effective in minimizing the effects of endogenous ligands and the potential interference of residual drugs. 22

D 2 receptor binding
Sections were first preincubated for 1 h at RT in 50 mM Tris-HCl buffer (pH 7.4) containing 120 mM NaCl, 5 mM KCl, 2 mM CaCl 2 , and 1 mM MgCl 2 . Sections were incubated for 1 h at RT in the same buffer containing 1.0 nM [ 3 H]nemonapride with 0.5 µM DTG and 0.1 µM pindolol to mask sigma (σ 1,2 ) and 5-HT 1A sites, respectively. Nonspecific binding (NSB) was determined with 10 µM S(-)-sulpiride. After incubation, slides were washed twice for 5 min in ice-cold buffer, dipped in cold water, and air-dried. Radiolabeled slides and [ 3 H] standards were exposed to Hyperfilm for 4 weeks at 4°C. 23,24 D 3 receptor binding Sections were preincubated for 1 h in 50 mM Tris-HCl buffer (pH 7.4) containing 0.3 mM GTP and 40 mM NaCl (no MgCl 2 added) to minimize labeling of the highaffinity agonist binding state of D 2 receptors, and then incubated for 1 h in the same buffer containing 3 nM [ 3 H]7-OH-DPAT, with 5 µM DTG to mask sigma sites. NSB was determined with 1 µM S(-)-eticlopride. After incubation, slides were washed (2 ×3 min) in ice-cold buffer and dried. Radiolabeled slides and [ 3 H] standards were exposed to Hyperfilm for 6 weeks at 4°C. 23,24 Serotonin receptor autoradiography

5-HT 2A receptor binding
Sections were preincubated for 60 min at RT in 50 mM Tris-HCl buffer (pH 7.7), and then incubated for 1 h at RT in fresh buffer containing 3.0 nM [ 3 H]ketanserin, 1 μM prazosin (to block α 1 -adrenoceptors), and 100 nM tetrabenazine (to block monoaminergic nerve terminal sites). NSB was determined with 1 μM methysergide. After incubation, slides were washed (2 × 30 min) in ice-cold buffer, dipped in cold water, and air-dried. Radiolabeled slides and calibrated [ 3 H] standards were exposed to Hyperfilm for 5 weeks at 4°C. 15 Glutamate receptor autoradiography NMDA receptor binding Sections were preincubated for 60 min in 50 mM Tris-HCl buffer (pH 7.4) to remove endogenous ligand and then incubated for 150 min at RT in buffer containing 10 nM [ 3 H]MK-801, 100 μM L-glutamate, 100 μM glycine, 1 mM EDTA, and 75 μM spermine to enhance the binding of [ 3 H]MK-801 to NMDA receptors. NSB was determined with 20 μM ketamine. After incubation, slides were washed in ice-cold buffer (2 × 20 min) and then air-dried. Radiolabeled slides and calibrated [ 3 H] standards were exposed to Hyperfilm for 4 weeks at 4°C. 17,18 AMPA receptor binding Sections were preincubated for 60 min in 50 mM Tris-HCl buffer (pH 7.3) at 4°C and then incubated in fresh buffer containing 20 nM [ 3 H]AMPA for 60 min at 4°C. 18 NSB was determined with 20 μM unlabeled CNQX. After incubation, slides were washed (3 × 15 sec) in ice-cold buffer and air-dried. Radiolabeled slides and [ 3 H] standards were exposed to Hyperfilm for 3 weeks at 4°C. 17,18 Image analysis Biomax MR films were developed and fixed in Kodak D-19 for 5 min at RT; images were quantified using an MCID image analyzer and MCID-M4 Imaging Research software (St. Catharines, Ontario, Canada). A calibration curve was generated from [ 3 H]-microscale standards that were exposed along with tissue sections. Regions of interest were outlined and their optical density measured, with the left and right sides of two contiguous sections representing total binding and matching samples representing nonspecific binding. Four determinations were averaged for each subject sample. Optical density of sampled regions was measured, and the amount of ligand bound was calculated as nCi/mg tissue and expressed as fmol bound/mg tissue.

Statistical analysis
Receptor binding data were analyzed first for overall effects of drug versus vehicle for all receptor types and brain regions using multiple regression modeling methods. Density measures were logarithmically transformed to achieve more Gaussian-like distributions before regression modeling. Model goodness-of-fit was checked using partial residual plot methods. Since individual brain specimens provided receptor-binding data for several brain regions, resulting in incomplete independence across observations, we used robust standard error estimates to adjust for this clustering effect. This estimation method permits relaxation of the assumption of independence of all observations and requires only that the observations be independent across specimens. 25 Estimates of interaction effects were employed for post-hoc tests of drug effects for specific receptors and brain regions, with adjustment of p values obtained from the regression analyses estimating these multiple comparisons by the standard method of Sidák. 26 Data are presented as means ± standard error of the mean (SEM). Comparisons were considered significant at p < 0.05 (two-tailed test).

Long-term effects of aripiprazole and cariprazine on NMDA receptors
Long-term administration of higher doses of aripiprazole  Figure 3B).   Figure 4B).

Long-term effects of aripiprazole and cariprazine on dopamine D 2 /D 3 receptors
Chronic aripiprazole treatment upregulated D 2 receptor levels in different brain regions, with a maximum increase similar to that previously described for cariprazine and other antipyschotics (  27 ). Aripiprazole also increased D 3 receptor levels in selective brain regions; however, these effects were less pronounced (cariprazine had~80-fold or greater potency for D 3 receptors) and were produced in fewer brain regions

EFFECTS OF ARIPIPRAZOLE ON SEROTONIN AND GLUTAMATE RECEPTORS 489
than previously described for cariprazine (Table 1; Choi et al. 27 ). In particular, aripiprazole produced no effect at any dose tested on D 3 receptors in the shell of the NAc, whereas cariprazine significantly increased these receptors in the same brain region. Increases in D 3 receptor expression in the shell of the NAc has been proposed to be a common neurobiological mechanism of antidepressant treatments. 28 These findings suggest that cariprazine is more potent than aripiprazole in upregulating D 3 receptor subtypes, and that D 3 receptors play a more prominent role in mediating the actions of cariprazine compared to aripiprazole. Moreover, the unique upregulation of dopamine D 3 receptors in areas related to antidepressant action may offer therapeutic benefits against various mood disorders, including depression and the negative symptoms of schizophrenia.  Figure 1A). This finding is consistent with the effects of the atypical antipsychotics olanzapine, risperidone, quetiapine, and asenapine on the same receptor in the same brain region. 15,16,29 Both aripiprazole and cariprazine have substantial affinity for 5-HT 1A receptors (Ki = 5.6 and 2.6 nM, respectively) and display partial agonist activity against this receptor subtype. 7,30 Observed increases in cortical 5-HT 1A receptors by both drugs may reflect direct blockade of 5-HT 1A receptors or  secondary changes resulting from more potent effects on dopamine D 2 and D 3 receptors. 7,14,30 Earlier studies reported that 5-HT 1A receptors are more highly expressed in cortical and limbic brain regions than in the basal ganglia. 31,32 In the frontal cortex, 5-HT 1A receptors are located on neocortical glutamatergic pyramidal neurons that are typically implicated in different cognitive functions. 33 Additional studies have suggested that 5-HT 1A partial agonism can stimulate cortical dopamine release [34][35][36] and minimize the incidence of antipsychotic-induced extrapyramidal side effects. 37,38 5-HT 1A receptors have increasingly become attractive targets for novel agents to treat schizophrenia and other psychiatric disorders. 39,40 Aripiprazole-and cariprazine-induced increases in cortical 5-HT 1A receptors further validate the hypothesis that the 5-HT 1A receptor may contribute, at least in part, to the beneficial therapeutic effects of these two agents as well as other dissimilar atypical antipsychotic drugs. 15 Repeated administration of aripiprazole (5 and 15 mg/kg) and cariprazine (0.06, 0.2 or 0.6 mg/kg) dose-dependently increased binding of the 5-HT 1A radioligand [ 3 H]8-OH-DPAT in hippocampal CA1 and CA3 regions. Asenapine is the only atypical antipsychotic that shares this effect, as other atypical antipsychotics, including olanzapine, risperidone, and quetiapine, did not alter 5-HT 1A receptors in rat hippocampus. 15,16 These findings suggest that aripiprazole and cariprazine can influence hippocampal 5-HT 1A receptors and consequently normalize the increased hippocampal activity postulated to occur in schizophrenic patients and ameliorate the psychotic symptoms of the disease. 41 Long-term effects of aripiprazole and cariprazine on serotonin 5-HT 2A receptors

Long-term effects of aripiprazole and cariprazine on serotonin 5-HT 1A receptors
Long-term treatment with all three aripiprazole doses significantly and equipotently decreased 5-HT 2A receptor binding in the MPC and DFC regions. Aripiprazole has high affinity for 5-HT 2A receptors (Ki = 8.7 nM), 30 and the decrease in 5-HT 2A binding may reflect direct blockade of 5-HT 2A receptors by the different doses of aripiprazole. We have previously reported that long-term administration of several atypical antipsychotics-including clozapine, olanzapine, risperidone, quetiapine, and asenapine-but not the typical antipsychotic haloperidol, resulted in decreased 5-HT 2A receptor labeling in rat frontal cerebral cortex. 15,22,42,43 5-HT 2A receptor downregulation may minimize the higher incidence of extrapyramidal side effects associated with typical antipsychotic drug treatment. 1,44,45 Aripiprazole (5 and 15 mg/kg) also decreased 5-HT 2A receptors in the hippocampal CA1 and CA3 regions. Such downregulation may further contribute to the stabilization of dopamine neurotransmission in the limbic system and, consequently, to the improvement of affective and emotional behaviors typically disturbed in schizophrenia patients. 1 In contrast, cariprazine failed to alter 5-HT 2A receptor binding in all the forebrain regions examined. The lack of effect of cariprazine on 5-HT 2A receptors may result from its lower in vitro affinity to these receptors, resulting in insufficient occupancy of 5-HT 2A receptors to the level required to induce receptor changes. These findings suggest that unlike several of the atypical antipsychotics, which downregulate cortical 5-HT 2A receptors, 15,22,42,43 5-HT 2A receptors are less likely to contribute to the mechanism of action of cariprazine in vivo.
Long-term effects of aripiprazole and cariprazine on glutamate NMDA receptors Continuous treatment with the higher doses of aripiprazole (5 and 15 mg/kg) and cariprazine (0.2 and 0.6 mg/kg) significantly decreased binding of [ 3 H]MK-801 to NMDA receptors in the NAc and the medial and lateral CPu. These effects were similar to those previously reported for the atypical antipsychotics clozapine, olanzapine, risperidone, and asenapine, and the experimental atypical antipsychotic JL13, but not the typical agent haloperidol. [17][18][19]46 Aripiprazole-and cariprazine-induced reductions in striatal NMDA receptors may result from indirect interactions with dopamine and serotonin systems, both of which modulate glutamatergic neurotransmission. 47,48 Such reductions in striatal NMDA receptors may reflect posttranscriptional modifications, since dissimilar atypical antipsychotics did not alter expression of mRNA-encoding NMDA-forming subunits in rat striatum. 49,50 We have demonstrated previously that acute and subchronic administration of cariprazine, as well as other atypical antipsychotics, reverses both NMDA receptor antagonist-mediated hypermobility 8 and PCPinduced cognitive and social deficits in animal models, 10,12,51 indicating an ability to normalize abnormal glutamatergic transmission; however, the potential neurochemical and behavioral consequences of longterm aripiprazole-and cariprazine-induced reductions in striatal NMDA receptors are not clear. Studies reported that stimulation of NMDA receptors promoted typical antipsychotic-induced catalepsy, an animal model predictive of extrapyramidal side effects. 52 In contrast, blockade of NMDA receptors attenuated the cataleptic effects of typical antipsychotic drugs. 53 Downregulation of striatal NMDA receptors by aripiprazole, cariprazine, and several dissimilar atypical antipsychotic drugs, 18,22 as well as the subsequent modification of glutamate neurotransmission in the basal ganglia, may contribute, at least in part, to the benign extrapyramidal profile of atypical antipsychotic agents. 45 Repeated treatment with aripiprazole and cariprazine reduced NMDA receptor binding in the hippocampal CA1 and CA3 regions. These effects are similar to the observed reduction in hippocampal NMDA receptors after repeated administration of the atypical antipsychotics olanzapine and risperidone. 18 Reductions in hippocampal NMDA receptors may normalize the abnormalities in glutamatergic neurotransmission postulated to occur in the hippocampus of schizophrenia patients, 54 and thus subsequently improve their psychotic symptoms. 55 Long-term effects of aripiprazole and cariprazine on glutamate AMPA receptors The three doses of aripiprazole and the higher doses of cariprazine (0.2 and 0.6 mg/kg) dose-dependently increased AMPA receptors in the hippocampal CA1 and CA3 regions. These findings are in agreement with the atypical antipsychotic asenapine, which induced increases in hippocampal AMPA receptors. 19 Such increases in AMPA receptors may result from the actions of both aripiprazole and cariprazine on D 2 receptors, which are coexpressed with AMPA in rat brains. 56 Increases in hippocampal AMPA receptors may further contribute to the improvement of psychotic symptoms in schizophrenia patients. Interestingly, long-term administration of aripiprazole and cariprazine produced opposite effects on NMDA versus AMPA receptors, suggesting that ionotropic glutamate receptors may respond differently to repeated treatment with partial D 2 /D 3 receptor agonists.

Conclusions
Long-term treatment with aripiprazole induced increases in D 3 receptors in OT and IC. In contrast, repeated treatment with cariprazine produced more profound increases in D 3 receptors in OT and IC, and extended its effect to induce increases in D 3 receptors in the shell of the NAc. These findings are in agreement with the potent affinity of cariprazine at D 3 receptors, and provide a distinction in the action mechanisms of cariprazine versus aripiprazole in selectively targeting D 3 receptors as leading sites that mediate the beneficial actions of cariprazine.
Repeated treatment with aripiprazole and cariprazine induced similar changes in 5-HT 1A , NMDA, and AMPA receptors in different rat forebrain regions. In contrast, aripiprazole downregulated 5-HT 2A receptors in the frontal cortex, an effect not seen with cariprazine. Accordingly, 5-HT 2A receptors may contribute more to the molecular actions of aripiprazole and less to the actions of cariprazine. In conclusion, the combined effects of cariprazine on serotonin and glutamate receptors, together with its unique actions on dopamine D 3 receptors, may confer the clinical benefits, safety, and tolerability of cariprazine in improved treatment of schizophrenia and bipolar mania.

Disclosures
Yong Kee Choi hereby states that he has no conflicts of interest to declare. Nika Adham is an employee of Allergan. Béla Kiss is an employee of Gedeon Richter Plc. István Gyertyán was an employee of Gedeon Richter Plc at the time of the study. Frank I. Tarazi has received research grants from Forest Laboratories LLC, H. Lundbeck A/S, and Shire Plc.