Abuse Potential of Biased Mu Opioid Receptor Agonists

Abstract

G-protein-biased mu opioid receptor (GPB-MOR) agonists constitute an emerging class of opioid analgesics. The first-in-class GPB-MOR agonist TRV130 (oliceridine) produces typical opioid-like abuse-related effects in rodents and humans. Although GPB-MOR agonists may be safer than conventional opioids on some endpoints, prevailing evidence suggests that they will retain opioid-like abuse potential.

G-protein-biased mu opioid receptor agonists: A novel type of opioid analgesic

The increasing prevalence of opioid abuse and overdose deaths has stimulated efforts to develop novel analgesics that retain effectiveness of opioids like morphine but have fewer undesirable effects in general and lower abuse liability in particular [1]. One class of compounds to emerge from this effort has been the G-protein-biased mu opioid receptor (GPB-MOR) agonists [2]. It is now well established that receptors can couple to multiple intracellular signaling pathways, and drugs can display bias in their relative potency and efficacy to activate these different pathways. In the case of MORs, the primary focus has been on development of GPB-MOR agonists to selectively activate G-protein-signaling as opposed to signaling mediated by other intracellular signaling proteins, chiefly ß-arrestin. This focus is founded on evidence to suggest that G-protein signaling mediates analgesic effects of MOR agonists, whereas ß-arrestin signaling may mediate at least some undesirable effects [3]. TRV130 (oliceridine) is a first-in-class GPB-MOR agonist that has shown analgesic effectiveness with moderate evidence for increased safety in clinical trials [2, 4], and other GPB-MOR agonists are also being developed [5–7]. The strongest evidence for improved safety with GPB-MOR agonists indicates that they may produce less respiratory depression at analgesic doses than currently available opioid analgesics [2, 6]. It has been suggested that GPB-MOR agonists may also have lower abuse potential than existing opioid analgesics, and identification of novel analgesics with reduced abuse potential is clearly a goal of current analgesic drug development [1]; however, the prevailing evidence summarized below suggests that these compounds will retain abuse liability.

Evidence from Preclinical Studies in Laboratory Animals

Enthusiasm for GPB-MOR agonists was initially stimulated by the finding that genetic knockout of the ß-arrestin-2 gene enhanced the analgesia-like effects of morphine in mice but attenuated some undesirable effects, such as the development of tolerance to those analgesia-like effects during repeated dosing [8]. The abuse-related effects of morphine in ß-arrestin-2 knockout mice were assessed using a place-conditioning procedure [9] (see Text Box 1 for description of place conditioning and other types of procedures used in abuse potential assessment). In this study, ß-arrestin-2 knockout did not attenuate rewarding effects of morphine; rather, morphine-induced conditioned place preference was enhanced in ß-arrestin-2 knockout mice. These data were interpreted to suggest that rewarding effects of morphine are not mediated by ß-arrestin-2 signaling, and by exclusion, this implicated G-protein signaling as a likely mediator of opioid reward in mice.

Text Box 1.

The abuse potential of new drugs is routinely tested using a battery of procedures in laboratory animals and humans [13, 19, 20]. Four major procedures are summarized below, followed by a summary of GBP-MOR agonist effects in these procedures.

Place Conditioning:

Subjects (mice/rats) are confined to one compartment of a multi-compartment chamber after test-drug treatment and to a different compartment after placebo. On a subsequent test day, subjects have access to both compartments. A “conditioned place preference” for the drug-paired compartment is predictive of abuse potential.

Intracranial Self-Stimulation (ICSS):

Subjects (mice/rats) with a microelectrode implanted in a brain-reward area are trained to press an operant-response lever to earn electrical brain stimulation. Drugs that increase ICSS responding are considered to produce “ICSS facilitation” predictive of abuse potential.

Drug Self-Administration:

Subjects (mice/rats/nonhuman primates/humans) can emit an operant response (e.g. pressing a response lever) to earn drug doses. Drugs that maintain more responding than placebo are considered to produce “reinforcing effects” predictive of abuse potential.

Drug Effects Questionnaire:

Subjects (humans) complete questionnaires to describe drug experiences. For example, a questionnaire might include a list of possible drug effects such as “Good Effects,” “Liking,” and “High,” and endorsement of these experiences is predictive of abuse potential.

Consistent with this conclusion, TRV130 produced a conditioned place preference in mice, albeit at dose higher than that producing antinociception [10]. More recent studies have begun to examine the abuse potential of GPB-MOR agonists in animal models using not only place conditioning procedures, but also procedures of intracranial self-stimulation (ICSS) and drug self-administration that are routinely used to evaluate abuse potential of opioids and other drugs (see Text Box 1). Results of abuse potential assessment with GPB-MOR agonists in these procedures are also summarized in Text Box 1. Most of these studies have been conducted with TRV130, and they generally suggest that TRV130 retains abuse potential similar to that of abused opioid analgesics. For example, Altarifi et al. [11] compared the effects of TRV130 and morphine in an ICSS procedure in rats and found that TRV130 produced abuse-related ICSS facilitation similar to that produced by morphine. Siuda et al. [2], in reviewing recent research on the development of GPB-MOR agonists for the treatment of pain, alluded to unpublished data from Trevana Inc. (the company that developed TRV130) investigating TRV130 drug self-administration. They reported that TRV130 was no different than morphine, implying that TRV130 functioned as a reinforcer with abuse potential similar to that of morphine. However, no information on the relative potencies or efficacies of the drugs as reinforcers was included. Only recently have published data become available investigating TRV130 (or any GPB-MOR agonist for that matter) as a reinforcer in a drug self-administration procedure. In that report, Zamarripa et al. [12] compared TRV130 to oxycodone, another opioid analgesic with high abuse potential. Both drugs were studied across a 30-fold range of intravenous doses in a demanding drug self-administration procedure that used a progressive-ratio schedule of reinforcement, in which the number of lever presses required to obtain drug delivery increased with each successive injection. Despite the use of this demanding procedure, TRV130 was similar to oxycodone in both reinforcing potency and effectiveness (Figure 1). In particular, both drugs maintained similar maximum numbers of injections per session, suggesting a comparable potential for abuse. These investigators also found that TRV130 and oxycodone had similar potencies and effectiveness to produce analgesia-like effects in rats, suggesting no advantage for TRV130 relative to oxycodone in potency to produce favorable analgesic effects versus unfavorable abuse-related effects.

Figure 1.

Self-administration of oxycodone and TRV130 under a progressive-ratio schedule of reinforcement in male Sprague-Dawley rats. Two-way ANOVA indicated a significant main effect of dose, but no differences between drugs at any dose. Filled symbols represent a significant difference from mean number saline infusions. Sal = saline; TD = training dose of oxycodone (0.056 mg/kg). Notably, oxycodone and TRV130 were also equipotent in the hot-plate test of thermal antinociception [ED₅₀ values (95% confidence limits) = 1.70 (1.06-2.56) and 1.25 (0.79-1.99) mg/kg for oxycodone and TRV130, respectively; data not shown in figure]. [12]

In the only study failing to observe an abuse-related effect with TRV130, Manglik et al. [5] reported that TRV130 and the other putative GPB-MOR agonist PZM21 failed to produce conditioned place preferences in mice at single doses that did produce analgesia-like behaviors. Similar findings and conclusions have been reported for another putative GPB-MOR agonist, mitragynine pseudoindoxyl [7]. However, these negative results should be interpreted with caution for several reasons. First, the experimental designs examined only one or two doses at only one time after drug administration, and place conditioning procedures are very sensitive to both dose and timing parameters [13]. Second, the conclusion by Manglik et al. [5] that TRV130 does not produce conditioned place preference in mice is not consistent with evidence discussed above for (a) enhanced morphine-induced place preference in ß-arrestin 2 knockout mice [9], (b) conditioned place preference in mice by a higher TRV130 dose [10], or (c) the robust expression of abuse-related effects by TRV130 in ICSS and drug self-administration procedures in rats [2, 11, 12]. Third, the status of PZM21 as a GPB-MOR agonist has been recently challenged [14], and the failure of mitragynine pseudoindoxyl to produce a conditioned place preference may reflect its short duration of action or its pharmacology as a mixed MOR agonist and delta opioid receptor antagonist rather than its bias for MOR-coupled G-protein signaling [7]. More generally, studies claiming an absence of abuse potential for GPB-MOR agonists have arguably focused more on the therapeutic than the abuse-related effects of GPB-MOR agonists, with the latter work relegated to limited dose and time ranges that do not constitute thorough assessments of abuse potential. It is therefore not surprising that the notion of abuse potential for GPB-MOR agonists has been treated as a “gray area” in recent discussions [1, 2], with some reports suggesting a favorable profile regarding abuse potential.

Evidence from Studies in Humans

In the only published study to evaluate abuse-related effects of a GPB-MOR agonist in humans, the subjective effects of intravenous TRV130 and morphine were compared across a range of time points in healthy male subjects using a Drug Effects Questionnaire [4] (see Text Box 1). Both drugs produced similar effects on endpoints such as “Good Effects,” “Liking”, and “High” that are often predictive of abuse potential. Thus, these initial results in humans agree with results of most studies in laboratory animals in suggesting that TRV130 retains opioid-like abuse potential.

Conclusions and future perspectives

The development of GPB-MOR agonists is founded on the hypothesis that MOR-coupled G-protein signaling mediates therapeutically useful analgesic effects of MOR agonists, whereas ß-arrestin signaling mediates at least some undesirable opioid effects. Data from genetically altered mice suggest that ß-arrestin signaling does not mediate opioid reward, and by exclusion, this implicates G-protein signaling as a likely contributor to opioid abuse potential. Consistent with this conclusion, the most thoroughly evaluated GPB-MOR agonist to date has been TRV130, and studies in mice, rats and humans suggest that this compound displays abuse potential similar to that of existing opioid analgesics such as morphine and oxycodone. Limited data from mouse place-conditioning studies suggest low abuse potential for the other putative GPB-MOR agonists PZM21 and mitragyinine pseudoindoxyl, but more extensive testing with these compounds is needed to confirm or refute their status as GPB-MOR agonists and to clarify both their expression of abuse-related effects and the role of G-protein signaling bias as a determinant of those effects. Newer GPB-MOR agonists are also being actively developed [6]. These newer compounds have higher degrees of bias than TRV130 for G-protein vs. ß-arrestin signaling, and abuse potential of these compounds has yet to be evaluated. However, prevailing evidence suggests that MOR-coupled G-protein signaling mediates opioid reward, that the first-in-class GPB-MOR agonist TRV130 displays abuse potential similar to that of existing opioid analgesics, and that abuse potential of newer GPB-MOR agonists warrants careful assessment. It is certainly premature to conclude that GPB-MOR agonists as a class are likely to have lower abuse potential than currently available opioid analgesics. As a final point, GPB-MOR agonists represent only one of several strategies to develop effective analgesics with reduced abuse potential and other undesirable effects. These other strategies include development of MOR agonists with slow brain entry to minimize the “rush” common with highly abused drugs [15], pH sensitivity to focus effects to injured and acidified tissue while reducing effects at other sites such as brain areas involved in abuse [16], mixed action at both MORs and other targets that might enhance analgesic effects but reduce side effects [17], and selectivity for MOR variants that might preferentially contribute to analgesia but not to abuse potential [18]. As future research evaluates and compares the relative effectiveness of these and other novel opioids to produce analgesia or other therapeutic benefits while reducing risk of abuse potential and other undesirable effects, it will become critically important to communicate and address all positive and negative findings in a thorough manner.

Abuse Potential Assessment Procedure [Reference]	Species	Drug (Doses, Route)	Abuse-Related Effect?
Subjective Effects [4]	Human	TRV130 (1.5-4.5 mg IV)	Yes
Intracranial Self-Stimulation [11]	Rat	TRV130 (0.1-3.2 mg/kg SC)	Yes
Drug Self-Administration [2]	Rat	TRV130 (Not Reported)	Yes
Drug Self-Administration [12]	Rat	TRV130 (0.01-0.32 mg/kg IV)	Yes
Place Conditioning [10]	Mouse	TRV130 (1.25, 10 mg/kg IP)	Yes
Place Conditioning [5]	Mouse	TRV130 (1.2 mg/kg SC)	No
Place Conditioning [5]	Mouse	PZM21 (20 mg/kg SC)	No
Place Conditioning [7]	Mouse	mitragynine pseudoindoxyl (1.3, 3 mg/kg IP)	No