Comparative neuropharmacology of structurally distinct non-fentanyl opioids that are appearing on recreational drug markets worldwide

Abstract

Background:

The emergence of novel synthetic opioids (NSOs) is contributing to the opioid overdose crisis. While fentanyl analogs have historically dominated the NSO market, a shift towards non-fentanyl compounds is now occurring.

Methods:

Here, we examined the neuropharmacology of structurally distinct non-fentanyl NSOs, including U-47700, isotonitazene, brorphine, and N-desethyl isotonitazene, as compared to morphine and fentanyl. Compounds were tested in vitro using opioid receptor binding assays in rat brain tissue and by monitoring forskolin-stimulated cAMP accumulation in cells expressing the human mu-opioid receptor (MOR). Compounds were administered subcutaneously to male Sprague-Dawley rats, and hot plate antinociception, catalepsy score, and body temperature changes were measured.

Results:

Receptor binding results revealed high MOR selectivity for all compounds, with MOR affinities comparable to those of morphine and fentanyl (i.e., nM). All drugs acted as full-efficacy MOR agonists in the cyclic AMP assay, but nitazene analogs had greater functional potencies (i.e., pM) compared to the other drugs (i.e., nM). When administered to rats, all compounds induced opioid-like antinociception, catalepsy, and body temperature changes, but nitazenes were the most potent. Similar to fentanyl, the nitazenes had faster onset and decline of in vivo effects when compared to morphine. In vivo potencies to induce antinociception and catalepsy (i.e., ED₅₀s) correlated with in vitro functional potencies (i.e., EC₅₀s) but not binding affinities (i.e., K_is) at MOR.

Conclusions:

Collectively, our findings indicate that non-fentanyl NSOs pose grave danger to those individuals who use opioids. Continued vigilance is needed to identify and characterize synthetic opioids as they emerge in clandestine drug markets.

1. Introduction

The United States (US) is experiencing an unprecedented opioid epidemic (CDC, 2022). While illicitly manufactured fentanyl is a major driving force (Ciccarone, 2019), the emergence of novel synthetic opioids (NSOs) has added a layer of complexity to the crisis (Prekupec et al., 2017). The number of NSOs continues to grow, with more than 70 identified by the European Monitoring Centre for Drugs and Drug Addiction (EMCDDA) and the United Nations Office on Drugs and Crime (UNODC) (EMCDDA, 2022a; UNODC, 2022). Particularly worrisome is the putative high potency of many NSOs, which portends risk for respiratory depression and death (Adamowicz and Nowak, 2022). The NSO market can be broadly categorized into fentanyl analogs (e.g., phenethylpiperidines) and non-fentanyl compounds of various chemical classes (e.g., cyclohexylbenzamides (Baumann et al., 2020; Kyei-Baffour and Lindsley, 2020), 2-benzylbenzimidazoles (Ujváry et al., 2021; Vandeputte et al., 2021), cinnamylpiperazines (Fogarty et al., 2022)). In 2017, 77% of NSOs reported to the EMCDDA were fentanyl analogs, but by 2021, all newly reported NSOs were structurally distinct from fentanyl (EMCDDA, 2022b; UNODC, 2020). The chemical structures of representative NSOs, as compared to morphine and fentanyl, are depicted in Figure 1.

Figure 1.

Chemical structures of the non-fentanyl synthetic opioids U-47700, brorphine, etonitazene, isotonitazene, and N-desethyl isotonitazene as compared to morphine and fentanyl.

Beginning in 2015, the most prevalent non-fentanyl NSO was U-47700. This cyclohexylbenzamide compound was patented in the 1970s by the Upjohn company (Szmuszkovicz, 1978) as part of a series of novel antinociceptive agents, but was never approved for clinical use (Baumann et al., 2020; Kyei-Baffour and Lindsley, 2020; Sharma et al., 2019). U-47700 garnered worldwide attention due to its association with multiple fatalities (Kyei-Baffour and Lindsley, 2020). As a result, the drug was internationally scheduled in 2017 (CND, 2017), then gradually disappeared. The number of new fentanyl analogs and U-compounds declined from 2018–2019, though various NSOs with different chemical structures transiently appeared (UNODC, 2020; Vandeputte et al., 2022d). In 2019, the NSO market took a new turn with the emergence of isotonitazene, a 2-benzylbenzimidazole opioid structurally related to the internationally scheduled compound, etonitazene (Blanckaert et al., 2020; Ujváry et al., 2021).

2-Benzylbenzimidazole opioids (“nitazenes”) were originally studied in the late 1950s (Bromig, 1958; Gross and Turrian, 1957; Hunger et al., 1960a, 1960b, 1957), but were never marketed (Ujváry et al., 2021). Data from preclinical studies demonstrate that isotonitazene is a highly potent mu-opioid receptor (MOR) agonist (Hunger et al., 1960a; Vandeputte et al., 2021), and the drug has been linked to many intoxications and fatalities (Krotulski et al., 2019, 2020a). Isotonitazene dominated the NSO market for a little over a year (Vandeputte et al., 2022a), followed by its international scheduling (DEA, 2020; UNODC, 2021) and the appearance of the benzimidazolone compound, brorphine (Krotulski et al., 2020c; Verougstraete et al., 2020). First described in the scientific literature in 2018 (Kennedy et al., 2018), brorphine quickly spread into recreational markets (Krotulski et al., 2020b; Vandeputte et al., 2022a). Scheduling actions targeting brorphine (DEA, 2021) expedited its disappearance, which was followed by the emergence of several new nitazene analogs (e.g., N-pyrrolidino etonitazene) (Papsun et al., 2022; Vandeputte et al., 2022b, 2022e). With the exception of fluorofentanyl, nitazene opioids now dominate the NSO market (EMCDDA, 2022a; Krotulski et al., 2023b). The most recent nitazene detected in confiscated drug material is N-desethyl isotonitazene, a known metabolite of isotonitazene (Krotulski et al., 2023a, 2022).

Limited information is available about the biological effects of NSOs when they first appear (Baumann et al., 2018b). Most preclinical studies have focused mainly on the in vitro pharmacological characterization of distinct classes of NSOs (i.e., fentanyl analogs (Åstrand et al., 2020; Eshleman et al., 2020; Hassanien et al., 2020; Kanamori et al., 2021; Vasudevan et al., 2020), U-compounds (Baumann et al., 2020; Hsu et al., 2019; Otte et al., 2022; Vasudevan et al., 2020), or nitazenes (De Luca et al., 2022; Kanamori et al., 2022; Vandeputte et al., 2021, 2022b, 2022e; Walton et al., 2022), and larger inter-class comparisons of NSOs are limited, hampering a direct evaluation of opioid activities across chemical scaffolds. Hence, the purpose of the present study was to carry out a side-by-side comparison of the pharmacology of recent, structurally diverse NSOs, and the classic opioid agonists morphine and fentanyl. In particular, we examined the effects of U-47700, isotonitazene, and brorphine, as prototypical examples of cyclohexylbenzamide, 2-benzylbenzimidazole, and benzimidazolone opioids. The parent nitazene compound, etonitazene, and the isotonitazene metabolite, N-desethyl isotonitazene, were also included (Figure 1). In vitro radioligand binding and cyclic AMP (cAMP) functional assays were performed in rat brain tissue and in cells stably expressing human MOR, respectively. Furthermore, in vivo opioid effects in rats were studied and compared across the different compounds.

2. Materials and Methods

2.1. Drugs and chemicals

U-47700 HCl (U-47700), isotonitazene, N-desethyl isotonitazene, and brorphine HCl (brorphine) were purchased from Cayman Chemical (Ann Arbor, MI, US), whereas etonitazene and morphine sulfate (morphine) were obtained from the National Institute on Drug Abuse (NIDA), Intramural Research Program (IRP) Pharmacy (Baltimore, MD, US). Fentanyl HCl (fentanyl) was generously donated by the NIDA Drug Supply Program (Rockville, MD, US). For in vitro assays, 10 mM stock solutions were prepared in dimethyl sulfoxide (DMSO) and stored at −80° C. On the day of an experiment, aliquots of 10 mM stock solution were diluted in the appropriate assay buffer. [³H]DAMGO, [³H]DADLE, and [³H]U69,593 were purchased from Perkin Elmer (Waltham, MA, US). For in vivo studies, 1 mg/mL stock solutions were prepared in either sterile 0.9% saline (vehicle for U-47700, fentanyl, and morphine) or 10% DMSO in sterile 0.9% saline (vehicle for brorphine and the nitazenes) and stored at 4°C. Aliquots of the 1 mg/mL stock solutions were diluted in their vehicles and administered subcutaneously (s.c.) in a volume of 1 mL/kg. All other chemicals and reagents were obtained from Millipore Sigma (St Louis, MO, US), unless otherwise noted.

2.2. Animals and surgery

Male Sprague-Dawley rats (250–350 g) (Envigo, Frederick, MD, US) were group-housed under controlled conditions with food and water freely available. Male rats were used in these experiments to allow for comparison to our previous work, and relevant studies in literature, which used male rats. Future studies will compare effects of opioid compounds in male versus female rats. Lights were on between 7 am and 7 pm. The animal experiments were approved by the NIDA IRP Animal Care and Use Committee. All procedures followed the NIH Guide for the Care and Use of Laboratory Animals. Vivarium facilities were accredited by the Association for Assessment and Accreditation of Laboratory Animal Care. For implantation of temperature transponders, rats were briefly anesthetized in a drop jar containing a gauze pad saturated with isoflurane. Once the rat was anesthetized, a temperature transponder (14 × 2 mm, model IPTT-300, Bio Medic Data Systems, Seaford, DE, US) was delivered s.c. on the back using a pre-packaged sterile guide needle. Rats were single-housed post-operatively, monitored for 3 days, and given 1 week to recover before experiments.

2.3. In vitro methods

We used radioligand binding assays in rat brain tissue as a first step in identifying the specific opioid receptor sites recognized by the drugs. Radioligand binding assays were carried out as described (Truver et al., 2020; Vandeputte et al., 2022b). Whole rat brains minus cerebellum (BioIVT, Westbury, NY, US) were frozen at −80°C until use. Thawed brain tissue (0.5 g) was homogenized in 20 mL of ice-cold 50 mM Tris HCl (pH 7.5). The tissue suspension was centrifuged twice at 30,000 g for 15 min at 4 °C and the pellet was resuspended (100 mg/mL). Binding experiments were initiated by adding 100 μL of tissue suspension to polypropylene tubes containing 300 μL Tris buffer, 50 μL test drug, and 50 μL radioligand (final concentration 1 nM). The mixture was incubated for 1 h at room temperature. [³H]DAMGO, [³H]DADLE, or [³H]U69,593 was used to label MOR, delta- (DOR), or kappa-opioid (KOR) receptors, respectively. Assays were terminated by rapid filtration over Whatman GF/B filters using a cell harvester (Brandel Instruments, Gaithersburg, MD, US). Washed filters were transferred to 24-well plates containing Cytoscint (MP Biomedicals, Irvine, CA, US). Radioactivity was counted using a Microbeta2 counter (Perkin Elmer). Counts per minute (cpms) were normalized to percent of radioligand bound using Microsoft Excel (Redmond, WA, US), and GraphPad Prism (San Diego, CA, US) was used to compute K_i values following nonlinear regression analysis.

MOR functional activity was determined by means of the CISBIO homogenous time-resolved fluorescence (HTRF) cAMP Gi kit (Perkin Elmer), where inhibition of forskolin-stimulated cAMP accumulation is measured in FLP-FRT-HEK cells stably expressing human MOR (Cai et al., 2019). While the cAMP accumulation assays used here are not directly comparable to the radioligand binding assays carried out in rat brain tissue, we find that cell-based methods are optimal for characterizing drug potency at MOR (Cai et al., 2019; Vandeputte et al., 2021). In this endpoint bioassay, cAMP produced by cells competes with labeled cAMP (Europium donor) for binding to a cAMP d2-labeled antibody (acceptor). Hence, higher FRET signals indicate more inhibition of forskolin-stimulated (cellular) cAMP accumulation by a test compound. Briefly, cell suspensions were incubated with test drugs for 45 min, followed by incubation with forskolin (final concentration 5 μM) for 45 min at 37 °C. cAMP Eu-cryptate reagent and cAMP-d2 antibody were added according to manufacturer instructions. FRET signals were assessed by calculating the fluorescence ratio at 665 and 620 nm using a PHERAstar FSX plate reader (BMG Labtech, Ortenberg, Germany). GraphPad Prism was used to calculate potency values (EC₅₀) following nonlinear regression analysis.

2.4. In vivo methods

In vivo methods were carried out as previously described (Vandeputte et al., 2022e). On the day of an experiment, rats were moved to the laboratory in their home cages and allowed one h to acclimate. Rats (n=6/dose) received s.c. injections of either vehicle or test drug, and doses were randomly assigned. Each rat was tested twice in separate sessions at least 3 days apart to ensure drug washout. Catalepsy score, body temperature, and hot plate latency were determined prior to injection and at 15, 30, 60, 120, and 240 min thereafter. At each timepoint, catalepsy was scored by an experienced rater who identified signs of immobility, flattened body posture, and splayed limbs. Each symptom was scored as either absent (= 1) or present (= 2) during a 1 min observation period. Body temperature was measured non-invasively using a handheld reader (DAS-7006/7r, Bio Medic Data Systems) sensitive to signals from the s.c. temperature transponder. Finally, rats were placed on a hot plate apparatus (IITC Life Sciences, Woodland Hills, CA, US) set at 52°C, from which they were removed when exhibiting hind paw licking. Time spent on the hot plate was recorded, and a 45 sec cut-off was applied.

Raw time course data for hot plate latency and catalepsy score were normalized to percent maximum possible effect (%MPE; 45 sec for hot plate latency, and 6 for catalepsy score) using the following equation:

$$\frac{\mathit{\text{experimental measure - baseline measure }}}{\mathit{\text{maximum possible measure-baseline measure}}}x100$$

Raw time course temperature data were normalized to change from baseline for each rat (Δ temperature in °C). Normalized time course data were analyzed by two factor (dose x time) ANOVA followed by Tukey’s post hoc test using GraphPad Prism (alpha = 0.05). Dose-response curves were computed using mean hot plate latency and catalepsy score over the first 60 min post-injection, and non-linear regression was used to determine potency values (ED₅₀).

3. Results

We first examined the binding affinities of test compounds at MOR, DOR, and KOR with competition radioligand binding assays. Concentration-response curves for MOR binding are depicted in Figure 2, left panel, while K_i affinity values at MOR, DOR, and KOR are summarized in Table 1. All opioids had nanomolar affinity at MOR, with K_i values ranging from 1.53 nM (N-desethyl isotonitazene) to 24.2 nM (brorphine). Compared to fentanyl (K_i = 6.35 nM) and morphine (K_i = 5.48 nM), most test drugs had weaker (brorphine, U-47700) or comparable (etonitazene, isotonitazene) MOR affinity, whereas N-desethyl isotonitazene bound to MOR with a stronger affinity than the reference opioids (Table 1). All test drugs were MOR selective, with DOR/MOR and KOR/MOR K_i ratios > 10. For N-desethyl isotonitazene, binding affinities at KOR and DOR were not determined.

Figure 2.

Effects of opioid test drugs on inhibition of radioligand binding at MOR in rat brain tissue (left panel) and inhibition of forskolin-stimulated cAMP formation in FLP-FRT-HEK cells stably expressing human MOR (right panel). Data are shown as mean ± SEM for n = 3 binding experiments or mean ± SEM for n = 5 cAMP experiments.

Table 1.

Effects of opioid test compounds on inhibition of radioligand binding at MOR, DOR, and KOR in rat brain tissue. Opioid drugs were incubated with 1 nM [³H]DAMGO, [³H]DADLE, or [³H]U69,593 for MOR, DOR, or KOR assays, respectively. Data are expressed as mean (with 95% confidence intervals in parentheses) for n = 3 separate experiments, each performed in triplicate. N.D. is not determined.

	MOR binding Ki (nM)	DOR binding Ki (nM)	KOR binding Ki (nM)	DOR/MOR ratio	KOR/MOR ratio
U-47700	14.3 (10.6–19.3)	1112 (655–1939)	181 (114–285)	78	13
Etonitazene	8.11 (5.17–12.6)	592 (433–811)	762 (401–1480)	73	94
Isotonitazene	11.2 (6.53–19.7)	1068 (661–1756)	325 (187–571)	95	29
N-desethyl isotonitazene	1.53 (0.922–2.48)	N.D.	N.D.	N.D.	N.D.
Brorphine	24.2 (17.9–32.7)	848 (519–1404)	579 (336–1018)	35	24
Fentanyl	6.35 (4.71–8.59)	479 (287–800)	204 (120–350)	75	32
Morphine	5.48 (4.05–7.40)	228 (138–355)	74 (51–108)	42	14

Next, the ability of the test compounds to activate MOR was evaluated in a cAMP accumulation assay (Table 2; Figure 2, right panel). All test drugs were fully efficacious in inhibiting forskolin-stimulated cAMP accumulation. With an EC₅₀ value of 4.13 pM, N-desethyl isotonitazene was the most potent opioid in the study. While etonitazene (EC₅₀ = 23.1 pM) and isotonitazene (EC₅₀ = 53.6 pM) activated MOR with a ^~5–13-fold lower potency, all three nitazenes were more potent than fentanyl (EC₅₀ = 0.104 nM). The EC₅₀ values for brorphine (EC₅₀ = 2.06 nM) and U-47700 (EC₅₀ = 4.47 nM) indicated lower potency than fentanyl, but similar potency to morphine (EC₅₀ = 1.21 nM) in this assay. When comparing MOR affinity and potency trends, it is interesting to note that the high affinity of N-desethyl isotonitazene tracked with a high functional potency in vitro. At the other end of the spectrum, the opioids with the weakest MOR affinities (brorphine, U-47700) showed the most right-shifted functional potencies (Figure 2). However, in general, the rank order of K_i values was not predictive of trends in functional potencies.

Table 2.

Summary of MOR affinity (K_i), in vitro (EC₅₀) and in vivo (ED₅₀) potency measures obtained by means of radioligand binding experiments (³[H]DAMGO displacement), inhibition of forskolin-stimulated cAMP accumulation, and rat behavioral studies, respectively.

	MOR binding Ki (nM)	MOR potency EC50 (nM)	Antinociception ED50 (mg/kg)	Catalepsy ED50 (mg/kg)
U-47700	14.3 (10.6–19.3)	4.47 (3.86–5.16)	0.405 (0.342–0.481)	0.832 (0.644–1.07)
Etonitazene	8.11 (5.17–12.6)	0.0231 (0.0198–0.0270)	0.00174 (0.00147–0.00206)	0.00380 (0.00331–0.00438)
Isotonitazene	11.2 (6.53–19.7)	0.0536 (0.0448–0.0638)	0.00704 (0.00562–0.00872)	0.0134 (0.0115–0.0157)
N-desethyl isotonitazene	1.53 (0.922–2.48)	0.00413 (0.00327–0.00518)	0.00298 (0.00234–0.00371)	0.00696 (0.00581–0.00829)
Brorphine	24.2 (17.9–32.7)	2.06 (1.77–2.40)	0.0858 (0.0752–0.0972)	0.112 (0.0855–0.146)
Fentanyl	6.35 (4.71–8.59)	0.104 (0.0872–0.125)	0.0209 (0.0192–0.0227)	0.0359 (0.0273–0.0467)
Morphine	5.48 (4.05–7.40)	1.21 (1.06–1.38)	4.10 (3.22–5.24)	8.75 (7.53–10.1)

We next examined the in vivo effects of the test compounds when administered s.c. to male rats. All of the drugs induced antinociception, catalepsy, and body temperature changes, but potencies varied widely. Representative time course data are shown for the least potent drug morphine (Figure 3), the most potent drug etonitazene (Figure 4), and fentanyl (Figure 5). Figure 3 demonstrates that the prototypical opioid agonist morphine induced dose-related increases in hot plate latency (F[4, 150] = 178.0; p<0.0001), and catalepsy scores (F[4, 150] = 23.11; p<0.0001), along with significant biphasic effects on body temperature (F[4, 150] = 21.49; p<0.0001). For the latter measure, low drug doses induced slight hyperthermia, while higher doses evoked a more substantial hypothermia. Similar to morphine, etonitazene induced dose-related effects on hot plate latency (F[4, 150] = 171.9; p<0.0001), catalepsy scores (F[4, 150] = 19.62; p<0.0001), and body temperature (F[4, 150] = 37.39; p<0.0001) (see Figure 4). A comparison between the time course data for morphine and etonitazene revealed two key differences. First, the morphine doses required to induce pharmacodynamic effects were approximately 1000-fold greater (i.e., mg/kg range) when compared to the doses of etonitazene (i.e., μg/kg range). Secondly, the effects of morphine displayed a much slower time course when compared to the effects of etonitazene. Figure 3 shows that antinociceptive effects of morphine were maximal at 30–60 min post-injection, whereas Figure 4 shows that antinociceptive effects of etonitazene were maximal at 15 min post-injection. Moreover, the maximal pharmacodynamic effects of morphine were sustained for at least 120 min post-injection, while effects of etonitazene were declining rapidly by this time point. Figure 5 demonstrates that fentanyl induced dose-related effects on hot plate latency (F[4, 150] = 259.7; p<0.0001), catalepsy scores (F[4, 150] = 53.49; p<0.0001), and body temperature (F[4, 150] = 37.61; p<0.0001). Like etonitazene, pharmacodynamic effects of fentanyl were characterized by a rapid onset and decline. In general, all of the NSOs tested here had a more rapid onset and decay of effects when compared to morphine (data not shown).

Figure 3.

Time course effects of s.c. morphine on hot plate latency (left panel), catalepsy score (center panel), and body temperature (right panel) in male rats. Hot plate and catalepsy data are % maximum possible effect (%MPE), whereas temperature data are change from baseline. Data are expressed as mean ± SEM for n = 6 rats per dose group. Solid symbols indicate significant changes compared to vehicle control at a given timepoint (p<0.05, Tukey’s post hoc).

Figure 4.

Time course effects of s.c. etonitazene on hot plate latency (left panel), catalepsy score (center panel), and body temperature (right panel) in male rats. Hot plate and catalepsy data are % maximum possible effect (%MPE), whereas temperature data are change from baseline. Data are expressed as mean ± SEM for n = 6 rats per dose group. Solid symbols indicate significant changes compared to vehicle control at a given timepoint (p<0.05, Tukey’s post hoc).

Figure 5.

Time course effects of s.c. fentanyl on hot plate latency (left panel), catalepsy score (center panel), and body temperature (right panel) in male rats. Hot plate and catalepsy data are % maximum possible effect (%MPE), whereas temperature data are change from baseline. Data are expressed as mean ± SEM for n = 6 rats per dose group. Solid symbols indicate significant changes compared to vehicle control at a given timepoint (p<0.05, Tukey’s post hoc).

In order to directly compare the in vivo potencies of the various drugs, mean effect data over the first h of the test sessions were used to construct dose-response curves for increases in hot plate latency and catalepsy score. Specifically, dose-response curves were calculated using non-linear regression analysis of log agonist concentration vs mean effect. The dose-response comparison data are depicted in Figure 6, whereas corresponding ED₅₀ values are given in Table 2. The rank order of in vivo potencies was etonitazene > N-desethyl isotonitazene > isotonitazene > fentanyl > brorphine > U-47700 > morphine. Hence, with the exception of N-desethyl isotonitazene and morphine, the rank order of in vivo potencies was broadly in line with in vitro potency trends. Nitazenes were generally around 3- to 12-times more potent than fentanyl, whereas brorphine and U-47700 were approximately 3- and 20-fold less potent than fentanyl in vivo. U-47700, the least potent NSO in these assays, was still 10 times more potent in vivo than morphine. For all of the compounds, higher doses were needed to induce catalepsy as compared to antinociceptive effects, the ratio between both ED₅₀ values being generally consistent (within 1.31 – 2.20).

Figure 6.

Dose-effect curves for hot plate latency (left panel) and catalepsy scores (right panel) induced by s.c. administration of test drugs in male rats. Data represent drug effects measured over the first 60 min of the experimental sessions and are expressed as mean ± SEM of the % maximum possible effect (%MPE), for n = 6 rats per dose group.

One key aim of the present study was to examine the relationships between in vitro and in vivo measures of opioid activity for the drugs tested. Figure 7 shows correlations between the potencies to induce antinociception or catalepsy (ED₅₀s) versus MOR affinities (K_is; left panels) or MOR functional potencies (EC₅₀s; right panels). In vivo ED₅₀ values for antinociception were positively correlated with MOR functional potencies (p<0.017) but not with MOR affinities. Similarly, ED₅₀ values for catalepsy scores were positively correlated with MOR functional potencies (p<0.025) but not with MOR affinities.

Figure 7.

Correlation plots investigating the relationship between potency to induce antinociception or catalepsy in the rat with MOR binding affinity (K_i values; left panels) or MOR functional potencies (EC₅₀ values; right panels). Pearson correlation coefficients (r) were calculated with GraphPad Prism 9.

4. Discussion

The opioid crisis in the US is being driven by illicitly manufactured fentanyl, but the emergence of structurally diverse NSOs is complicating an already grim situation (UNODC, 2020). Characterizing the biological effects of emerging NSOs is critical for informing stakeholders about the health risks associated with these compounds (Baumann et al., 2018a). While clear strides have been made in this area (Vandeputte et al., 2022d), most studies have focused on the structure-activity relationships (SARs) of NSOs belonging to a single chemical class. As a result, direct comparisons of larger panels of structurally dissimilar NSOs are rather scarce, and the use of different reference compounds and experimental techniques make it difficult to compare data across studies. In the current work, we evaluated a panel of five synthetic opioids (U-47700, etonitazene, isotonitazene, N-desethyl isotonitazene, and brorphine), belonging to three different chemical classes (cyclohexylbenzamides, 2-benzylbenzimidazoles, and benzimidazolones). Importantly, the specific compounds studied have infiltrated recreational drug markets worldwide and are associated with hundreds of overdose fatalities (Baumann et al., 2020; Krotulski et al., 2020a, 2020b). Fentanyl and morphine were included as standard opioid reference agonists.

Our radioligand binding data confirm that all of the test compounds display nanomolar affinity at MOR. Given the key role of MOR in mediating both therapeutic and adverse effects of opioids (Charbogne et al., 2014), it is not surprising that the NSOs were at least 10-fold selective for MOR over DOR and KOR. The present MOR affinity results for U-47700 (K_i = 14.3 nM) agree with previous work using similar methods (Truver et al., 2020), but other groups using different methods have reported a broad range of MOR affinity values (K_i range ^~0.91–57 nM) (Baumann et al., 2018b; Janowsky, 2016; Kyei-Baffour and Lindsley, 2020; Loew et al., 1988). Despite the differences in K_i for U-47700 across studies, the high MOR selectivity of the compound (MOR >> KOR > DOR) is consistently shown. For brorphine, the MOR affinity reported here (K_i = 24.2 nM) is comparable to earlier findings (K_i = 35.4 nM) obtained with the same assay format (Vandeputte et al., 2022c). For etonitazene, the high MOR selectivity we found is consistent with earlier results, but the MOR affinity values for etonitazene reported in the literature vary considerably (K_i range 0.00042–0.41 nM) (Emmerson, 1994; Moolten et al., 1993; Toll et al., 1998; Ujváry et al., 2021, 2021; Zernig et al., 1995). For isotonitazene, our MOR affinity (K_i = 11.2 nM) is consistent with the data of Walton et al. (Walton et al., 2022) but weaker than that found by others (De Luca et al., 2022). Finally, for N-desethyl isotonitazene (K_i = 1.53 nM), our results agree with those of Walton et al. (Walton et al., 2022). The dramatic differences in literature K_i values reported for specific opioid compounds illustrate the difficulties in comparing radioligand binding results across laboratories, and highlight the importance of conducting inter-class comparisons of NSOs under similar assay conditions.

While binding experiments provide valuable data about relative affinities at opioid receptor subtypes, such methods do not give information about functional effects of opioid ligands (Pottie and Stove, 2022). Therefore, we examined functional potency of test compounds in a bioassay measuring inhibition of cAMP formation. It is important to note that cAMP assays carried out in cells transfected with human MOR are not directly comparable to the radioligand bindings assays carried out in rat brain tissue. Nevertheless, the cAMP inhibition assay described here provides key information about in vitro drug potencies. We found that all test compounds fully inhibited forskolin-stimulated cAMP accumulation, with picomolar to nanomolar potencies. Notably, we observed a much wider range of functional potencies when compared to binding affinities (i.e., all K_is in the nM range, with overlapping confidence intervals). With an EC₅₀ value of 4.47 nM, U-47700 was the least potent opioid tested, and the drug was 4- and 40-fold less potent compared to morphine and fentanyl. Using different techniques, we and others (e.g., (Otte et al., 2022; Vasudevan et al., 2020; Vasudevan and Stove, 2020)) also reported a lower in vitro potency for U-47700 compared to fentanyl. Brorphine was twice as potent as U-47700 in the current study, but still about 20-times less potent than fentanyl. This finding contrasts with data from others who found a higher potency for brorphine compared to fentanyl (Grafinger et al., 2021). Consistent with prior findings (De Luca et al., 2022; Kanamori et al., 2022; Vandeputte et al., 2021), all of the evaluated nitazenes were more potent than fentanyl (2- to 25-times), and much more potent than morphine (23- to 290-times). Our results from the cAMP assay confirm the known SAR for 2-benzylbenzimidazole opioids (Hunger et al., 1960a; Vandeputte et al., 2021), with etonitazene being more potent than isotonitazene. Our data further support the literature showing that N-desethyl isotonitazene is more potent in vitro than isotonitazene itself (Vandeputte et al., 2021).

Given the limitations of in vitro binding and functional assays in predicting in vivo opioid activity (Baumann et al., 2018a), we examined pharmacodynamic effects of the various drugs in male rats. We found that all test opioids induce antinociceptive, cataleptic, and body temperature effects qualitatively similar to the effects of morphine and fentanyl. Importantly, all of the NSOs were much more potent than morphine (10- to 100-times), and the nitazene compounds were even more potent than fentanyl (3- to 12-times) in vivo. Our time course evaluation demonstrated that pharmacodynamic effects of fentanyl and the NSOs display a fast onset and decline when compared to the effects of morphine. In particular, fentanyl and the nitazene compounds display rapid effects that could be related to high lipophilicity and brain penetration of the compounds. However, it must be mentioned that our experiments employed the s.c. route of administration in rats, and this route differs from those typically used by humans (i.e., oral, intranasal, and intravenous). The drug doses needed to induce catalepsy in our experiments were somewhat higher than those needed for antinociception, as shown previously for morphine (Pöyhiä and Kalso, 1992; Taracha et al., 2009) and other NSOs (Truver et al., 2020; Vandeputte et al., 2022b, 2022e; Walton et al., 2022). Most studies examining opioid-induced catalepsy in rats have utilized the “bar test”, whereas we opted to use a behavioral scoring method that is sensitive to observable changes in mobility and posture. One limitation of our catalepsy scoring method is that it provides no information about the level of muscular rigidity that is often associated with high-dose opioid administration. The general rank order of in vivo potencies was etonitazene > N-desethyl isotonitazene > isotonitazene > fentanyl > brorphine > U-47700 > morphine. While brorphine was slightly less potent than fentanyl, in vivo potencies for U-47700 were about 20-fold lower than those of fentanyl (but still higher than those of morphine). For U-47700, the antinociceptive ED₅₀ that we observed (0.405 mg/kg) agrees with prior studies in mice (Baumann et al., 2018b; Cheney et al., 1985) and rats (Truver et al., 2020). To the best of our knowledge, the only prior in vivo study evaluating brorphine assessed discriminative stimulus effects in morphine-trained rats (Gatch, 2020). In that study, the ED₅₀ value of brorphine (0.16 mg/kg) was in the same range as the potency for catalepsy shown here.

The most potent opioids tested in our experiments belong to the class of 2-benzylbenzimidazole opioids. In the original studies, the potency of isotonitazene was reportedly 500-fold greater than that of morphine (Hunger et al., 1960a), and our findings are consistent with these prior findings. Since the emergence of isotonitazene on recreational drug markets, various groups have evaluated its in vivo profile in rodents, and antinociceptive potencies reported across different laboratories agree reasonably well (ED₅₀ range ^~2–20 μg/kg) (De Luca et al., 2022; Lee et al., 2022; Walton et al., 2022). Hence, our results (ED₅₀ = 7 μg/kg) are in line with existing findings. N-desethyl isotonitazene, a known isotonitazene metabolite that is formed in both rats (Walton et al., 2022) and humans (Krotulski et al., 2020a; Walton et al., 2021), was twice as potent as isotonitazene. Of particular concern is the recent emergence of N-desethyl isotonitazene as a stand-alone drug on recreational NSO markets (cfr. infra) (Krotulski et al., 2023a, 2022).

For all test drugs, transient hyperthermia was observed at low doses, but substantial hypothermia was induced at the highest dose. The same temperature effect profile has been shown for other opioids (Adler et al., 1988; Rawls and Benamar, 2011) and NSOs (Bergh et al., 2019; Truver et al., 2020; Vandeputte et al., 2022b, 2022e). Here, we used hypothermia as a proxy for adverse effects of opioids (Truver et al., 2020; Walton et al., 2022; Wong et al., 2017). While more research is needed to investigate the relationships between opioid-induced therapeutic (i.e., antinociception) and adverse (i.e., respiratory depression) effects in humans, the summed findings in rodents indicate that several NSOs are more potent than fentanyl and morphine in producing hypothermia and potentially other adverse events.

An important goal of the present study was to examine the relationships between in vitro and in vivo endpoints for the NSOs tested. Prior studies show that in vitro parameters (e.g., binding affinity and functional potency) are not always related to each other, and may not accurately predict in vivo opioid potency (Baumann et al., 2018a; Vandeputte et al., 2022b, 2022e, 2022c; Volpe et al., 2011). As a case in point, fentanyl and morphine displayed similar MOR K_i values in our experiments, yet fentanyl was 10- and 200-times more potent than morphine in vitro and in vivo. The greater potency of fentanyl is likely related to more efficient coupling to MOR-associated transduction machinery at the molecular level (Ricarte et al., 2021; Zhuang et al., 2022) and better brain penetration at the organismic level (e.g., see (Schmid et al., 2017)). We found that in vivo potencies to induce antinociception and catalepsy (ED₅₀s) were positively and significantly correlated with in vitro functional potencies (EC₅₀s) but not affinity values (K_is) at MOR. It is noteworthy that the nitazene compounds displayed extraordinary potency in the cAMP accumulation assay, suggesting these compounds are capable of highly efficient coupling to G protein-mediated processes, even at low MOR occupancy. To the best of our knowledge, the current study provides the most comprehensive pharmacological profiling of N-desethyl isotonitazene. The fact that this compound combines high MOR binding affinity and functional potency, suggests that N-desethyl isotonitazene could be extremely dangerous to humans who are inadvertently exposed to the drug. The ultra-high in vivo potency of N-desethyl isotonitazene further suggests that overdose reversal could require multiple or higher doses of naloxone (Moss and Carlo, 2019; Rzasa Lynn and Galinkin, 2018).

To summarize, we provide a comprehensive preclinical characterization of structurally distinct NSOs using in vitro and in vivo pharmacological techniques. The in vivo potency ranking of the compounds included in this study reveals an alarming trend: while U-47700, one of the first non-fentanyl NSOs, was 20-fold less potent than fentanyl, ‘newer’ NSOs like the nitazene analogs have comparable or even higher potency than fentanyl. While this is not necessarily true for every newly emerging nitazene or NSO, the trend of increased potencies over time presents great risks to people who use opioids, as well as challenges to law enforcement personnel, clinical toxicology laboratories, and public health officials. Hence, continued vigilance combined with a multidisciplinary response will be needed as NSOs continue to appear on recreational drug markets worldwide.

Highlights.

• Non-fentanyl synthetic opioids are potent MOR agonists in vitro

• Non-fentanyl synthetic opioids elicit antinociception and catalepsy in rats

• Isotonitazene and its N-desethyl metabolite are more potent than fentanyl

• In vitro drug potency at MOR is correlated with in vivo potency to induce opioid-like effects

• Non-fentanyl synthetic opioids, especially “nitazene” analogs, may pose serious risks to humans