Elucidating GPCR Functional Selectivity for Drug Development

Overview

G protein-coupled receptors (GPCRs) are the largest class of targets in drug development and the targets for 40% of available drugs. Upon their discovery, GPCRs were thought to relay extracellular signals only by associating with intracellular G proteins (GTP-binding proteins). But these receptors are now known to signal through other pathways, notably those mediated by arrestins. Arrestins are multifunctional signaling modulators first described for their role in GPCR desensitization, the termination of GPCR activity through phosphorylation. Independent of G proteins, arrestins can recognize active GPCR conformations and induce downstream signaling. Some GPCR ligands induce G protein signaling, while others induce arrestin-mediated signaling at the same receptor. Ligand bias toward particular signal transduction pathways at the same GPCR is known as functional selectivity. On September 30, 2014, the Academy's Biochemical Pharmacology Discussion Group held a symposium titled Elucidating GPCR Functional Selectivity: Novel Opportunities for Drug Development to examine ligand bias and its implications for the design of better targeted and more effective drugs.

John A. Allen, a principal scientist at Pfizer, outlined the challenges involved in this kind of research. By understanding functional selectivity, scientists can identify suitable molecules for drug development that will engage the desired pathways while avoiding adverse events. However, the process of profiling molecules with potential as therapeutics must involve multiple in vitro signaling assays early in preclinical development to reveal bias for specific downstream outcomes. Furthermore, biased ligand binding activity observed in vitro may not occur in vivo, indicating the need for in vivo readouts for distinct signaling pathways. Terry Kenakin of the University of North Carolina School of Medicine also highlighted the differences between in vitro and in vivo ligand activity, describing three common methods used to quantify biased receptor signaling and explaining why it is difficult to use these equations to predict an agonist's in vivo therapeutic response. Signaling is driven by both a ligand's ability to bind its receptor, known as affinity, and its ability to induce a cellular response upon binding, known as efficacy. Cells express varying levels of the same receptors, so each biased ligand exhibits tissue-specific behavior that depends on its affinity for its receptor. Affinity and efficacy determine how receptor availability and receptor–ligand interaction impact the functional selectivity of compounds in vivo.

Laura Bohn of Scripps Research Institute described her work to move drug candidates from in vitro studies to animal models. Bohn's research focuses on therapeutic targeting of the kappa opioid receptor (KOR), the receptor for morphine and other opioid analgesic drugs. Evidence suggests G protein signaling through this receptor is associated with pain and itch relief, while β-arrestin-mediated signaling may induce depression and dysphoria. Bohn is therefore looking for KOR ligands that selectively activate G protein signaling. After testing G protein-biased KOR agonists in mice, she showed that the compounds elicited the expected antinociceptive and anti-itch effects without negative behavioral changes. These data suggest it is possible for biased agonists discovered in vitro to be translated into therapies that improve on existing drugs.

JoAnn Trejo from the University of California, San Diego, focused on the role of receptor protein modification in regulating GPCR coupling with G protein subtypes. The PAR1 receptor for the coagulation factor thrombin is modified by N-linked glycosylation at various sites, two located on the second extracellular loop that mediates tethered ligand interaction. Trejo's group investigated the functional significance of this glycosylation by studying the signaling pathways of a mutated form of the receptor and by using pharmacological agents to block glycosylation of the endogenous receptor in cultured human endothelial cells. The researchers discovered that N-linked glycosylation of the extracellular loop does not impact receptor trafficking or surface expression, but it does drive PAR1 to selectively couple with G_α12/13 proteins. Downstream, this signaling bias has an impact on cytoskeletal responses (specifically stress fiber formation) and cellular proliferation. Trejo's presentation highlighted a novel mechanism through which active receptors can be modified to promote distinct G protein signaling pathways, allowing us to better understand how to target GPCRs to achieve therapeutic benefit.

The thrombin receptor PAR1 is glycosylated at several sites, two of which are on the surface of the second extracellular loop (ECL2). Study of a mutated receptor lacking these sites (PAR1 NA ECL2) revealed that glycosylation of ECL2 impacts G protein signaling, biasing PAR1 to couple with G_α12/13 subunits rather than with G_α. This signaling bias has an effect on downstream cellular functions, including stress fiber formation and cellular proliferation. (Image courtesy of JoAnn Trejo)

The work presented by Bohn and Trejo demonstrated that GPCR ligands can selectively induce signaling pathways in vivo and that this biased GPCR activity can result from the ligand itself or be induced by modifications to the receptor. Jonathan A. Javitch of Columbia University and New York State Psychiatric Institute took a structural approach to the understanding of functional selectivity by discussing how GPCR signaling is biased by the location and orientation of the receptor–ligand interaction. Classical approaches to GPCR drug development targeted the orthosteric binding site, the primary active binding site for endogenous ligands. But a major focus of current research is allosteric binding at alternative sites on the receptor. Allosteric modulators typically bind with lower affinity but can be highly selective among receptor subtypes, because unlike orthosteric ligands they do not target evolutionarily conserved binding regions. Javitch discussed his collaborative work to characterize bitopic ligands, which are linked molecules that simultaneously bind orthosteric and allosteric sites. He described the role of bitopic ligands in the study of GPCR dimerization, the association of two receptors to form one functional unit. Focusing on the dopamine receptor, Javitch designed a functional complementation assay to probe for communication among GPCR protomers. His data show that bitopic ligand binding to one protomer can cause conformational changes that modulate the binding of another ligand to a second protomer, demonstrating the complexity of GPCR structural relationships and cross-talk between signaling pathways.

Bryan L. Roth of the University of North Carolina School of Medicine continued the discussion of GPCR structure by giving an overview of his work on serotonin receptors. Serotonin receptor agonists are used in the treatment of migraines and Parkinson's disease but can cause valvular heart disease. Ergotamine in particular displays extreme bias toward arrestin-mediated signaling at one of the serotonin receptors, which may play a role in the severe cardiac side effect. Roth found differences in the structural motifs and ligand orientations of the binding sites of biased and unbiased serotonin receptors. These data help explain the striking difference in downstream signaling. His studies emphasize the importance of molecular structure in determining biased agonism.

In the keynote presentation, Robert J. Lefkowitz from Duke University Medical Center gave a historical overview of the β-arrestin protein family. Beginning with the discovery of the arrestins as mediators of receptor desensitization, Lefkowitz described the years of research that contributed to our understanding of the arrestins as multifunctional adapter proteins and signaling modulators. Building on the previous talks, he gave several more examples of the therapeutic benefits of biased agonism and pointed out the differences between ligand bias (differential signaling due to ligand-specific receptor conformations) and system bias (differential signaling due to unequal transducer protein coupling). Research into GPCR functional selectivity has demonstrated that agonist-mediated signaling bias is probably a result of both mechanisms. To fully understand signaling bias, we need to analyze high-resolution crystal structures of the ternary (receptor–ligand–transducer) complexes involved. Lefkowitz shared his research on the crystal structure of β-arrestin 1 bound to an activating GPCR phosphopeptide. His work highlights how crystallography and 3D modeling can elucidate arrestin signaling interactions.

The day ended with two talks focused on the clinical implications of GPCR ligand bias. Marc G. Caron from Duke University Medical Center presented his work on dopamine receptor signaling. In mouse models with either β-arrestin 2 or the downstream signaling protein GSK3β conditionally deleted in neurons, Caron's group has demonstrated that β-arrestin-dependent signaling downstream of the dopamine D2 receptor is a key mechanism of action for anti-psychotic drugs. Dopamine signaling has various consequences in different parts of the brain, so D2 receptor ligands that are functionally selective for anti-psychotic pathways may serve as more effective medications with fewer side effects.

Jonathan D. Violin, cofounder of Trevena Inc., presented clinical data that supports the idea that biased ligands make better drugs. For the treatment of acute heart failure, Trevena has developed a compound called TRV027 that targets imbalances in the renin–angiotensin system by directly binding the angiotensin II receptor. The ligand blocks G protein signaling while enhancing β-arrestin recruitment, leading to increased cardiac function. Violin also discussed the development of an opioid analgesic (TRV130) that selectively induces G protein-coupled signaling at the kappa opioid receptor. TRV130 exploits the same mechanism as the drug candidates Bohn is investigating to achieve better pain management while preventing side effects. The drugs Violin described show promise in early clinical trials, lending support to hopes that understanding GPCR functional selectivity can lead to improved health care.

TRV027 is a novel β-arrestin-biased ligand of the angiotensin II receptor that may find use as a more effective treatment for acute heart failure. By blocking G protein signaling, this compound reduces blood pressure in rats with the same efficacy as the full angiotensin receptor blocker (ARB) telmisartan. Rather than simultaneously blocking other beneficial pathways associated with this receptor, TRV027 enhances β-arrestin-mediated signaling, improving cardiac contractility and function. (Image courtesy of Jonathan D. Violin)

Use the tabs above to find multimedia from this event.

Presentations available from:
John A. Allen, PhD (Pfizer)
Marc G. Caron, PhD (Duke University Medical Center)
Jonathan A. Javitch, MD, PhD (Columbia University; New York State Psychiatric Institute)
Terry Kenakin, PhD (University of North Carolina School of Medicine)
Bryan L. Roth, MD, PhD (University of North Carolina School of Medicine)
JoAnn Trejo, PhD (University of California, San Diego)
Jonathan D. Violin, PhD (Trevena Inc.)

The Biochemical Pharmacology Discussion Group is proudly supported by

Mission Partner support for the Frontiers of Science program provided by