Viewpoint pubs.acs.org/chemneuro
Heteromer Induction: An Approach to Unique Pharmacology? Philip S. Portoghese,* Eyup Akgün, and Mary M. Lunzer Department of Medicinal Chemistry, College of Pharmacy, University of Minnesota, Minneapolis, Minnesota 55455, United States ABSTRACT: It is proposed that two types of opioid receptor heteromers exist: a) those that are constitutive and b) those that are induced by bivalent ligands. Mu opioid agonists interact with constitutive MOR-DOR heteromer to mediate tolerance and dependence. Bivalent ligand, MDAN21, is devoid of these adverse effects by virtue of its DOR antagonist pharmacophore. We propose that bivalent ligands MMG22 and MCC22 induce colocalized receptors to form heteromers (MOR-mGluR5 and MOR-CCR5, respectively) that do not occur naturally, thereby promoting unique pharmacology. Heteromer induction with bivalent ligands offers a general approach to unique pharmacology that complements traditional SAR. KEYWORDS: heterodimer, opiod receptor, MOR, DOR
■
MOR-DOR: A CONSTITUTIVE OPIOID RECEPTOR HETEROMER
Since the first report in 1982 of a bivalent ligand approach that was consistent with in vitro existence opioid receptor dimers, the topic of G protein-coupled receptor (GPCR) oligomers has received increasing attention given emerging new methods for validation of physical interactions between GPCRs in vitro.1 The question in the title has been raised because of the profound differences in the in vivo pharmacology of bivalent ligands that interact with constitutive versus induced heteromers. Here we present our Viewpoint on this subject using relatively recent examples of approaches that suggest in vivo induction of heteromers by opioid bivalent ligands affords unique antihyperalgesic profiles that are superior to clinically employed monovalent ligands. Given that in vitro studies have suggested the possible existence of nearly two dozen heteromers of opioid receptors, it seems likely that some of these heteromers may be constitutive in vivo and a source of adverse effects when activated by clinically employed opioid analgesics. For example, tolerance and dependence to morphine is believed to be linked to chronic interaction with the mu opioid receptor (MOR) protomer of a MOR-DOR heteromer, inasmuch as mice devoid of functional delta opioid receptors (DOR) do not display these side effects (knockout mice, antisense, or delta opioid antagonist). MDAN21, a 21-atom homologue in a series of bivalent ligands that contain mu opioid agonist and delta opioid antagonist pharmacophores, provided support for involvement of a MOR-DOR heteromer in opioidinduced tolerance and dependence.1,2 MDAN21 and MDAN19 afforded high analgesic potency without adverse effects that were interpreted in terms of bridging MOR and DOR protomers of a MOR-DOR heteromer. Homologues with shorter linkers (spacers) produced tolerance and dependence, presumably via univalent interaction with the MOR protomer of a constitutive MOR-DOR heteromer. More recent studies have reported on the interaction between delta antagonist and mu opioid agonists (morphine, fentanyl, methadone) in HEK293 cells, mice, and rhesus monkeys, further implicating MOR-DOR as a constitutive heteromer that mediates tolerance and dependence in humans. © XXXX American Chemical Society
■
DO BIVALENT LIGANDS MMG22 AND MCC22 INDUCE MOR HETEROMERS? This question was addressed during our studies with MMG22 and its homologues.3 The MMG series contains mu opioid receptor agonist and metabotropic glutamate receptor-5 (mGluR5) antagonist pharmacophores, as mGluR5 and MOR are colocalized in central neurons and glia. The receptor target selection was based on the interaction of these receptors in cultured cells and the concept that targeting such a heteromer with a structurally optimized bivalent ligand might effectively inhibit hyperalgesia associated with the sensitization of spinal afferent neurons. Such an approach would offer a distinct advantage over morphine and other mu opioid analgesics that are known to activate spinal glia upon chronic administration, thereby inducing hyperalgesia that leads to reduced efficacy and other side effects. In inflamed (LPS) mice, optimal antihyperalgesic potency among MMG homologues was observed with a spacer length of 22 atoms,3 which is in the range reported for other opioid bivalents. Significantly, intrathecal (i.t.) MMG22 (ED50 ∼ 9 fmol/mouse) was 4400× more potent in inflamed mice relative to normal mice. The intracerebroventricular (i.c.v.) potency in inflamed mice was reduced by a factor of 43 000× relative to i.t. administration, suggesting activated glia and/or spinal neurons as targets. The critical role of MOR to the potency of MMG22 was confirmed with opioid receptor knockout mice (unpublished data). That MMG22 was >1000-fold more effective than its lower 20-atom homologue and ∼30-fold greater than its higher homologue (24 atoms), strongly suggests an optimal Received: January 4, 2017 Accepted: January 10, 2017
A
DOI: 10.1021/acschemneuro.7b00002 ACS Chem. Neurosci. XXXX, XXX, XXX−XXX
Viewpoint
ACS Chemical Neuroscience
based on a BRET study in cultured cells that revealed colocalized receptors that fail to form heteromers can be induced to do so in the presence of an appropriate bivalent ligand.6 Thus, it is possible that the unusual pharmacologic profiles associated with MMG22 and MCC22 may arise from in vivo induction of heteromers not normally present in spinal glia. As an example, a simple model for MMG22 induction of MOR-mGluR5 heteromer could involve the following steps (Figure 1): (a) Univalent
spacer requirement for bridging of protomers in the target MOR-mGluR5 heteromer. The critical importance of the spacer to bridging was convincingly demonstrated by the ∼38 000-fold greater i.t. efficacy of MMG22 when compared to a mixture of monovalent mu agonist and mGluR5 antagonist. A dramatic illustration of enhanced efficacy was illustrated by the 3.6 millionfold greater potency relative to that of morphine in inhibiting hyperalgesia in mice with chronic bone cancer.4 Molecular modeling using the TM5,6 interface for MOR-mGluR5 is consistent with the optimal 22-atom spacer requirement of MMG22 (unpublished data). Treatment of inflamed mice with the N-methyl D-aspartate receptor (NMDAR) antagonist, MK801, totally blocked the antihyperalgesic effect of MMG22. This is consistent with the role of mGluR5 as a coreceptor of the NR2B subunit of the NMDAR, given its presence at both synaptic and extra synaptic locations. Spinal astroglia are likely targets because MMG22-induced antihyperalgesia was potently blocked by the selective astroglia inhibitor, L-α-aminoadipate (LAA), whereas the microglia-selective inhibitor, minocycline, only weakly suppressed hyperalgesia (unpublished data).
Figure 1. Illustration of the concept of MOR−mGluR5 heteromer induction by MMG22. Green circle and red square represent mGluR5 antagonist and MOR agonist pharmacophores, respectively.
binding of MMG22 to a mGluR5 oligomer; (b) dissociation of MMG22-bound to mGluR5 protomer from the oligomer; (c) equilibrium-driven association with a MOR oligomer via assistance of the free mu opioid agonist pharmacophore of MMG22; (d) dissociation of the bridged MMG22-bound oligomer from the MOR oligomer to afford the MMG22− heteromer complex. While a number of other models of induced heteromer formation are possible, the essential feature for all would be the optimal spacer length of the bivalent ligand.
Based upon our success with MMG22, a similar approach was employed to target a heteromer consisting of MOR and chemokine receptor CCR5, given evidence for MOR-CCR5 in cultured cells and the reported colocalization of these receptors on glia and neurons. Linking of mu opioid agonist and CCR5 antagonist pharmacophores with 14- to 24-atom spacers afforded homologues in the MCC series, whose 22-atom homologue (MCC22) possessed optimal potency (i.t. ED50 = 15 fmol) in LPS inflamed mice.5 MCC22 (i.t.) was 3100-fold more potent in inflamed mice relative to normal mice, suggesting activated glia as a target. The crucial role of the spacer for efficacy was indicated by the fact that i.t. MCC22 possessed ∼3600-fold greater potency than a mixture of monovalent mu agonist and CCR5 antagonist pharmacophores. Studies with MOR KO mice and the MOR-selective antagonist, β-FNA, supported the critical importance of MOR (unpublished data). The complete inhibition of MCC22-induced hyperalgesia by pretreatment of inflamed mice with the selective microglia inhibitor, minocycline, suggested microglia as a target.5 Unlike MMG22, the antihyperalgesic effect of MCC22 was weakly inhibited by the NMDAR antagonist, MK801, suggesting only minor involvement of the NMDAR as a mechanism of action (Unpublished data).
■
FUTURE APPLICATIONS OF THE HETEROMER INDUCTION CONCEPT Significantly, the pharmacologic profiles of MMG22 and MCC22 differ remarkably from those of conventional mu agonists, mGluR5 antagonist, or a mixture of mu monovalent agonist and selective antagonist ligands. As suggested by the concept (Figure 1), their unique activity profiles are derived from an induced heteromer that is dependent on the presence of a bivalent ligand that facilitates the association of MOR with mGluR5 or CCR5. Thus, induced heteromer formation of colocalized receptors offers an approach to the development of medicinal with unique profiles that would be difficult to achieve with monovalent ligands. These examples suggest that bivalent ligands containing a combination of agonist and selective GPCR antagonist pharmacophores offer an opportunity to explore other agonist/antagonist pharmacophore combinations for the design of bivalent ligands as pharmacologic tools and potential medicines. It will be interesting to see if other colocalized receptors systems also show unique profiles by heteromer induction.
■
THE CONCEPT OF HETEROMER INDUCTION AS A NEW APPROACH TO UNIQUE PHARMACOLOGIC PROFILES All of the structure−activity relationship data for the unusual pharmacological profiles of MMG22 and MCC22 are consistent with the induction of heteromers that may not be present at significant levels in glia or neurons. The possibility of induction is
■
AUTHOR INFORMATION
Funding
This research was supported by NIH Grant R01 DA030316. Notes
The authors declare no competing financial interest. B
DOI: 10.1021/acschemneuro.7b00002 ACS Chem. Neurosci. XXXX, XXX, XXX−XXX
Viewpoint
ACS Chemical Neuroscience
■ ■
ACKNOWLEDGMENTS We thank Michael Powers for creating Figure 1. REFERENCES
(1) Portoghese, P. S., Ronsisvalle, G., Larson, D. L., Yim, C. B., Sayre, L. M., and Takemori, A. E. (1982) Opioid agonist and antagonist bivalent ligands as receptor probes. Life Sci. 31, 1283−6. (2) Daniels, D. J., Lenard, N. R., Etienne, C. L., Law, P. Y., Roerig, S. C., and Portoghese, P. S. (2005) Opioid-induced tolerance and dependence in mice is modulated by the distance between pharmacophores in a bivalent ligand series. Proc. Natl. Acad. Sci. U. S. A. 102, 19208−19213. (3) Akgün, E., Javed, M. I., Lunzer, M. M., Smeester, B. A., Beitz, A. J., and Portoghese, P. S. (2013) Ligands that interact with putative MOR-mGluR5 heteromer in mice with inflammatory pain produce potent antinociception. Proc. Natl. Acad. Sci. U. S. A. 110, 11595−11599. (4) Smeester, B. A., Lunzer, M. M., Akgün, E., Beitz, A. J., and Portoghese, P. S. (2014) Targeting putative mu opioid/metabotropic glutamate receptor-5 heteromers produces potent antinociception in a chronic murine bone cancer model. Eur. J. Pharmacol. 743, 48−52. (5) Akgün, E., Javed, M. I., Lunzer, M. M., Powers, M. D., Sham, Y. Y., Watanabe, Y., and Portoghese, P. S. (2015) Inhibition of Inflammatory and Neuropathic Pain by Targeting a Mu Opioid Receptor/Chemokine Receptor5 Heteromer (MOR-CCR5). J. Med. Chem. 58, 8647−8657. (6) Zheng, Y., Akgün, E., Harikumar, K. G., Hopson, J., Powers, M. D., Lunzer, M. M., Miller, L. J., and Portoghese, P. S. (2009) Induced association of μ opioid (MOP) and type 2 cholecystokinin (CCK2) receptors by novel bivalent ligands. J. Med. Chem. 52, 247−258.
C
DOI: 10.1021/acschemneuro.7b00002 ACS Chem. Neurosci. XXXX, XXX, XXX−XXX
