Lumateperone is a compound in clinical development as a treatment for schizophrenia, bipolar depression, agitation associated with dementia and other neuropsychiatric and neurodegenerative diseases. This novel compound exhibits high affinity in vitro binding to dopamine D2 receptors and occupies striatal dopamine D2 receptors in human brain at therapeutic doses. At dopamine D2 receptors lumateperone displays dual properties, acting both as a functional post-synaptic antagonist and pre-synaptic partial agonist. To further characterize these unique receptor interactions, we compared the activity of lumateperone relative to three known D2 partial agonists of differing intrinsic efficacy, aripiprazole, brexpiprazole, and bifeprunox, in cell-based functional assays that permit characterization of compounds of differing intrinsic efficacy (e.g. full agonism to neutral antagonism). Lumateperone had no demonstrable agonist activity in CHO cells expressing recombinant human D2L (or D2S) receptors, in assays where other known partial agonists displayed different degrees of agonist activity.

Available drugs for treatment of schizophrenia are largely based on antagonism of dopamine D2 receptors. Unalloyed D2 antagonism, however, is associated with significant limiting side effects including acute motoric disturbances and chronic irreversible dyskinesia (i.e. tardive dyskinesia). These side effects are ameliorated in second-generation antipsychotics, of which most combine antagonism of serotonin (5-HT) 2A receptors with D2 antagonism. The two current antipsychotic drugs, aripiprazole and brexpiprazole, are D2 receptor partial agonists^[1][2][3][4]. These compounds have high affinity for D2 receptors but low intrinsic efficacy, meaning that they have a reduced ability to stimulate functional activity of the receptor relative to the endogenous ligand or full agonist drugs^[5]. Recently, Li et al. have described lumateperone, a candidate antipsychotic that has potent 5-HT2A antagonism and potent activity at D2 receptors^[6]. Although lumateperone occupies striatal dopamine D2 receptors in human brain at therapeutic doses^[7], and in vivo in mouse brain it demonstrates postsynaptic antagonism, at the same doses it demonstrates presynaptic partial agonism^[8]. To study this unique dual behavior better, we compared the intrinsic efficacy of lumateperone to the known partial agonists, aripiprazole, brexpiprazole, and bifeprunox, in assays of dopamine D2 receptor function, using two naturally occurring variants of the human D2 receptor in recombinant cells.

To characterize the G-protein-mediated functional activity of lumateperone at the D2L and D2S dopamine receptors under conditions sensitive to D2 partial agonists.

To evaluate the agonist or antagonist activity of compounds, we measured the ability of these compounds to either inhibit forskolin-stimulated cAMP accumulation or reverse the inhibition produced by 30 nM of quinpirole in a Chinese Hamster Ovary (CHO) cell line expressing recombinant human dopamine D2L or D2S receptors. D2 receptors in this cell line couple with Gαi/o protein subunits to suppress adenylate cyclase activity. In antagonist assays using the D2-L receptor (Fig. 1A), lumateperone potently (IC50 = 32 nM) and completely antagonized the activity of quinpirole, with an IC50 in the same range as its published affinity for D2 receptors in a ligand-displacement assay (32 nM; Li et al., 2014). We compared this to haloperidol (IC50 = 0.33 nM), brexpiprazole (1.68 nM) and aripiprazole (3.32 nM). In keeping with their partial agonist character, brexpiprazole and aripiprazole only partially antagonized the activity of quinpirole, reaching ~70% and ~45% maximum blockade, respectively.

This assay can also be run in agonist mode. Agents such as aripiprazole, bifeprunox, and brexpiprazole act as partial agonists, partially suppressing adenylate cyclase activity^[9]. Depending on receptor abundance and efficiency of receptor coupling to second messengers in the recombinant system, the intrinsic efficacy for aripiprazole can range from 20 to 90% that of dopamine^[2]. Figure 1B summarizes the result of a direct comparison of lumateperone with dopamine (full agonist), bifeprunox (partial agonist with high intrinsic efficacy),^[10] aripiprazole (partial agonist with medium intrinsic activity)^[11], and brexpiprazole (partial agonist with low intrinsic activity)^[3] in the D2L-receptor-induced suppression of forskolin-stimulated adenylate cyclase activity. Bifeprunox reached 77% activity relative to quinpirole, with an EC50 of 4.26 nM (Fig. 1B and C). Aripiprazole reached 37% activity, with an EC50 of 4.43 nM. Brexpiprazole has less intrinsic activity than aripiprazole at D2 receptors, with a maximum effect of 11% and an EC50 of 1.32 nM. Lumateperone (up to 3 µM) had no demonstrable agonist activity in this assay.

In vivo, D2L receptors are abundantly found in postsynaptic dopamine-responsive cells. D2S, a splice variant of this receptor lacking exon 6 encoding the third intracellular loop, is most abundant on the presynaptic cells, and has been hypothesized to be the usual form functioning as the presynaptic autoreceptor (reviewed in^[12]). Therefore, we asked whether the partial agonist activity of lumateperone would be revealed in assays using this receptor. In CHO cells expressing recombinant human D2S receptors, however, this was not the case (Fig. 1C and D). Once again, while aripiprazole and brexpiprazole acted as functional partial agonists, lumateperone showed only antagonist activity. This may reflect the fact that the G-protein repertoire of CHO cells does not replicate the presynaptic receptor cellular milieu. Prior work with many cell lines expressing recombinant D2 receptors has demonstrated that the cellular milieu, rather than the splice variant, is the major influence determining functional readout from D2 receptors^[2].

We used whole cell based assays to detect D2-receptor-dependent suppression of adenylate cyclase, a functional mimic of postsynaptic D2 receptors, to compare the activity of lumateperone to older, partial agonist, atypical antipsychotics. In multiple assays that revealed partial agonist activity of existing drugs, lumateperone acted as an antagonist. Other approaches may be needed to reveal the presynaptic factors that underlie the functional properties in vivo of this compound.

It has been documented that differences in receptor coupling to second messenger partners in test systems with recombinant receptors can influence the ability to detect partial agonism. Full understanding of the activity of this compound may require selective assay of presynaptic receptors in their authentic cellular milieu, which is beyond the scope of this report.

The activity of lumateperone in this assay could be interpreted as partial agonist activity of extremely low intrinsic efficacy.

Our working hypothesis is that presynaptic D2 autoreceptors exist in a unique receptor complex that we cannot replicate in CHO cells. Further work could include experiments assaying functional activity of a selective population of brain-derived presynaptic neurons, or genetic and biochemical approaches designed to reveal the nearest-neighbors of the D2 receptor in these cells.

Chemicals

Research chemicals were supplied by Sigma-Aldrich unless otherwise specified. Quinpirole (#1061), aripiprazole (#5584), and forskolin (#1099) were purchased from Tocris (www.tocris.com). Bifeprunox was supplied by Toronto Research Chemicals (B383200; www.trc-canada.com/). Lumateperone was synthesized as described in Li et al.(Li, Zhang et al. 2014). All other reagents were the highest quality available.

D2R cell culture

Cells grown prior to the test in media without antibiotic were detached by gentle flushing with PBS-EDTA (0.5 mM EDTA), recovered by centrifugation and resuspended in assay buffer (5 mM KCl, 1.25 mM MgSO4, 124 mM NaCl, 25 mM HEPES, 13.3 mM Glucose, 1.25 mM KH2PO4, 1.45 mM CaCl2, 0.5 g/L protease-free BSA, supplemented with 1 mM IBMX). For antagonist and agonist tests, 12 µL of cells (2,500 cells/well) were assayed as described in figure legends in the wells of a 384 well white plate.

This study was funded by Intra-Cellular Therapies, Inc.

Not Applicable.

No fraudulence is committed in performing these experiments or during processing of the data. We understand that in the case of fraudulence, the study can be retracted by ScienceMatters.