Evaluation of G protein bias and β-arrestin 2 signaling in opioid-induced respiratory depression

Abstract

Respiratory depression is a potentially fatal side effect of opioid analgesics and a major limitation to their use. G protein-biased opioid agonists have been proposed as “safer” analgesics with less respiratory depression. These agonists are biased to activate G proteins rather than β-arrestin signaling. Respiratory depression has been shown to correlate with both G protein bias and intrinsic efficacy, and recent work has refuted the role of β-arrestin signaling in opioid-induced respiratory depression. In addition, there is substantial evidence that G proteins do, in fact, mediate respiratory depression by actions in respiratory-controlling brainstem neurons. Based on these studies, we provide the perspective that protection from respiratory depression displayed by newly developed G protein-biased agonists is due to factors other than G protein versus arrestin bias.

Opioid analgesics are the pharmacological gold standard to treat pain. However, abuse liability and potentially lethal respiratory depression limit their clinical safety and therapeutic potential. For example, a recent study found that nearly half of hospitalized patients experienced opioid-induced respiratory depression (1). Moreover, the incidence of opioid abuse and overdose are increasing and constitute a public health crisis (2).

Opioid-induced respiratory depression is due to activation of mu-opioid receptors (MORs) (3), which are expressed on respiratory control neurons in the brainstem. Conditional deletion of MORs from neurons in two of these respiratory control areas—the Kölliker-Fuse nucleus or preBötzinger Complex—attenuates opioid-induced respiratory depression (4). MORs inhibit these neurons by activating G protein-coupled inwardly rectifying potassium (GIRK) conductance (5, 6) and inhibiting neurotransmitter release (7).

Currently, only the opioid antagonist naloxone (Narcan) is used to reverse opioid-induced respiratory depression. Although this is lifesaving, it reverses all opioid effects and leads to withdrawal and uncontrolled pain. Therefore, several alternate strategies are being developed to prevent or avoid respiratory depression while maintaining analgesia, including: 1) immunopharmacotherapies (vaccines and monoclonal antibodies) (8), 2) nonopioid respiratory stimulants (e.g., ampakines, serotonin agonists, intranasal leptin, or orexin agonists) (9), and 3) opioid analgesics with minimal respiratory depression. It is this last point that we would like to address more fully.

Affinity, efficacy, and target are characteristics that distinguish drug-receptor interactions from each other. Through the decades, our understanding of these concepts has greatly expanded. R. P. Stephenson (10) first defined agonist efficacy as linear and postulated that an agonist produced uniform activation at the receptor. Present understanding is that efficacy can be pluridimensional, meaning that ligands can have many receptor-based efficacies. Furthermore, agonists can be “functionally selective” or “biased” and stabilize different receptor conformations that preferentially activate certain effectors (11). Substantial research has aimed at understanding the potential of biased MOR agonists and will be reviewed below.

MORs are members of the G protein-coupled receptor (GPCR) superfamily. Stimulation of MORs activates heterotrimeric G proteins to inhibit adenylyl cyclase, modulate ion channel activity, and activate second-messenger signaling cascades (Fig. 1). MOR activity is regulated by a highly conserved process partly involving G protein-coupled receptor kinases (GRKs)—or other second-messenger kinases such as protein kinase C (PKC)—which phosphorylate the intracellular loops and carboxy-terminal tail of MORs and facilitate β-arrestin binding (12). Through this mechanism, MOR signaling is attenuated, MOR desensitization and internalization is initiated, and β-arrestin 2 signaling—independent of G proteins—is activated (13).

Figure 1.

Mu opioid receptor cellular signaling pathways involved in respiratory depression. Agonist binding to mu-opioid receptor (MOR) activates heterotrimeric G protein subunits: alpha (α), beta (β), and gamma (γ). Gβγ inhibits voltage-gated calcium channels (VGCC) and activates G protein-coupled inwardly rectifying potassium channels (GIRK). β-arrestin 2 binds to MOR with carboxy-terminal phosphorylation sites (P). GIRK, but likely not β-arrestin 2, activity can lead to respiratory depression.

β-arrestin 2 knockout mice exhibit impaired MOR desensitization and, consequently, increased morphine analgesia with very little morphine tolerance (14). Subsequent studies have confirmed the critical role β-arrestin 2 plays in MOR desensitization and development of tolerance (15). These initial findings sparked an interest in the regulatory role of β-arrestin 2 in MOR signaling and opioid pharmacology. The finding that β-arrestin 2 knockout mice had attenuated morphine-induced respiratory suppression kindled the flame (16). From this study, stemmed the hypothesis that β-arrestin 2 signaling—more so than G protein signaling—was responsible for deleterious opioid effects, including respiratory depression.

The goal of developing a MOR G protein-biased agonist devoid of respiratory depression has led to the development of several novel compounds (17–19). The first potential drug candidate was oliceridine (TRV-130). Preclinical studies demonstrated less respiratory depression and a threefold preference for G protein signaling compared with morphine and fentanyl (17). However, a 2018 FDA brief summarizing phase 3 clinical trials concluded that oliceridine and morphine caused comparable respiratory depression (20). Recently, reanalysis of pharmacokinetics and pharmacodynamics data from these clinical trials found oliceridine to have a safer profile and lower probability of causing respiratory depression compared with morphine (21). With the recent FDA approval of oliceridine, further testing of its safety profile can be performed. Subsequent drug discoveries had similar challenges. PZM21, another biased-MOR agonist, was developed computationally and found to activate G protein signaling, recruit minimal β-arrestin 2, and have limited respiratory depression (18). However, it was later reported to cause respiratory depression similar to morphine (22).

These G protein-biased drugs, despite the controversy, illustrate a correlation between ligands that preferentially activate G proteins and cause limited respiratory depression (19). However, a causal relationship between G protein bias and limited respiratory depression has not been established. An alternate hypothesis is that attenuated respiratory depression is due to low intrinsic efficacy (ability of a drug-receptor interaction to produce a maximal effect) rather than G protein bias (22, 23). For example, the clinically used agonist buprenorphine has a low intrinsic efficacy and produces significantly less respiratory depression (24, 25).

The absence of mechanistic evidence for how β-arrestin 2 signaling could mediate opioid-induced respiratory depression has led several groups to reexamine the β-arrestin 2 hypothesis. The first study used mice expressing phosphorylation-deficient, G protein-biased MORs that were generated by mutating MOR carboxyl-terminal residues from serine and threonine to alanine (15). This prevented β-arrestin 2 recruitment and subsequent signaling, rendering the receptor G protein biased. Interestingly, phosphorylation-deficient, G protein-biased MOR mice still exhibited respiratory depression, suggesting that G protein signaling contributes to opioid-induced side effects (15). This conclusion was substantiated by several laboratories that were unable to replicate an attenuation of morphine or fentanyl-induced respiratory depression in β-arrestin 2 knockout mice (26, 27). They concluded that their findings “do not support the original suggestion that β-arrestin 2 signaling plays a key role in opioid-induced respiratory depression and call into question the concept of developing G protein-biased MOR agonists as a strategy for the development of safer opioid analgesic drugs” (27). And most recently, He et al. (25) demonstrated that genetic and pharmacological manipulations that enhance MOR-mediated β-arrestin 2 signaling do not influence respiratory depression (but in fact reduces analgesic tolerance), corroborating preexisting evidence. This suggests that a balanced MOR agonist with low intrinsic efficacy may be “safest” because it can retain both physiological G protein function and arrestin-mediated receptor regulation and tolerance development. In addition, He et al. (25) found that 129SvJ mice are less sensitive to morphine-induced respiratory depression than C57BL6/J mice, illustrating that background strain matters in opioid-induced respiratory depression measurements. Since the original β-arrestin 2 knockout mice were on a mixed background of C57BL6/129SvJ, a difference in background could account for the different phenotypes observed by various research groups.

Furthermore, mechanistic evidence demonstrates that G protein signaling can, in fact, mediate opioid-induced respiratory depression (Fig. 1). Activation of MORs stimulates heterotrimeric G proteins, initiating the separation of the Gα-subunit from the Gβγ-subunits for subsequent G protein-signaling transduction. Gβγ-subunits activate GIRK channels, thereby decreasing neuronal excitability. Kölliker-Fuse and preBötzinger Complex neurons, which are critical to opioid-induced respiratory depression, are hyperpolarized by MOR-mediated activation of GIRK conductance (4–6). MOR activation of GIRK channels in the Kölliker-Fuse is similar in wild-type and phosphorylation-deficient G protein-biased MOR mice, illustrating that GIRK channel activation is G protein, not β-arrestin 2 mediated (15). Furthermore, GIRK channel knockout mice experience significantly less opioid-induced respiratory depression than wild-type mice (28).

Considering the mechanistic evidence that G proteins can mediate respiratory depression and the existence of multiple contradictory studies for β-arrestin 2-dependent respiratory depression, we question the utility of future studies to develop G protein-biased MOR agonists as safer analgesics on the basis of G protein versus β-arrestin 2 bias alone. Instead, other factors, such as intrinsic efficacy, receptor selectivity, or bias among G protein subunits or other downstream effectors should be considered. Despite this, the first agonist to come out of this line of work, oliceridine, was recently FDA approved and has been shown to have a better safety profile than morphine in humans (21). Whether this is due to biased signaling or other factors such as intrinsic efficacy remains an active debate (29, 30).

GRANTS

This work was supported by National Institutes of Health Grants R00DA038069 and R01DA047978. J. T. Bateman was supported by NIH Grant F31DA053798.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

J.T.B. and E.S.L. drafted manuscript; J.T.B. and E.S.L. edited and revised manuscript; J.T.B. and E.S.L. approved final version of manuscript.