NSOs with a 2-benzylbenzimidazole core (‘nitazenes’) continue to emerge on recreational drug markets. N-Pyrrolidino metonitazene, N-pyrrolidino protonitazene, and N-desethyl etonitazene are among the most recent nitazenes identified in the U.S. (Krotulski et al. 2023). Although great strides have been made in the evaluation of nitazene structure–activity relationships (SAR) (Ujváry et al. 2021; Vandeputte et al. 2021, 2022b, e, 2023; De Luca et al. 2022; Kanamori et al. 2023; Walton et al. 2023; Malcolm et al. 2023; Glatfelter et al. 2023; Kozell et al. 2024; Tsai et al. 2024), larger side-by-side comparisons of the effect of different systematic modifications to the nitazene core structure on MOR activity are rather scarce. Specifically, our understanding of SAR for ‘ring’ substituted analogues (i.e., N-pyrrolidino and N-piperidinyl analogues) remains limited (Ujváry et al. 2021; Vandeputte et al. 2022b, e; Kozell et al. 2024; Tsai et al. 2024). For example, similar to N-pyrrolidino etonitazene (Vandeputte et al. 2022b), N-pyrrolidino metonitazene and N-pyrrolidino protonitazene were not included in the early CIBA studies, and can thus be considered truly ‘new’ synthetic opioids. To further unravel nitazene SAR, we performed in vitro pharmacological characterization of a panel of 25 differentially substituted nitazenes, including 9 previously uncharacterized analogues, focusing on three distinct positions of the 2-benzylbenzimidazole core structure (R1–3 in Fig. 1 and Table 1). In addition, complementing the in vitro pharmacological findings and demonstrating the danger posed by these substances to the users, we report the first identification and toxicological analysis of etodesnitazene, N-desethyl etonitazene, N-desethyl isotonitazene, N-pyrrolidino metonitazene, and N-pyrrolidino protonitazene in forensic cases from North America as well as the identification of N-pyrrolidino metonitazene, N-pyrrolidino protonitazene, and N-desethyl isotonitazene in cases from the U.K. Hence, in line with earlier recommendations that the opioid crisis should be dealt with in a multidisciplinary manner (Baumann et al. 2018; Morrow et al. 2019), and not solely by, e.g., isolated case reports or pharmacology-only studies, this study aimed at bridging these two areas. The latter is also particularly relevant for the different stakeholders involved, as observations in forensic casework can be supported by pharmacological data. This is a prerequisite when considering, e.g., both scheduling efforts and harm reduction strategies.

In this study, two distinct cell-based assays were used for the functional characterization of a series of nitazenes at MOR: a β-arrestin 2 recruitment assay, which monitors receptor-proximal (upstream) activation of the G protein-independent pathway, and a GloSensor® cAMP assay, which monitors the inhibition of forskolin-stimulated cAMP accumulation downstream of Gαi protein activation. While measuring MOR activation at different levels (upstream and further downstream) in a common cellular background (HEK 293T cells) can be considered a strength of this study, it is important to consider some inherent differences in the characteristics of the assays that were deployed (Zhang and Xie 2012; Vandeputte et al. 2022d). For example, the comparatively higher level of signal amplification in the cAMP assay, as well as the modification of MOR and co-expression of GRK2 in the βarr2 recruitment assay, are factors to be considered. Evaluation of these assay-specific parameters was beyond the scope of this study, which aimed at determining nitazene SAR using assays that have been extensively used before, by us and others (Vandeputte et al. 2022d). In this context, it is important to note that all efficacies in the cAMP assay converged around ~ 100% compared to hydromorphone, hampering a comparative assessment of the efficacies of the evaluated compounds. This ‘plateauing effect’ is a well-known characteristic of assays assessing downstream effects, such as the cAMP assay (Vandeputte et al. 2022d). The MOR-βarr2 assay, in contrast, has historically been demonstrated to yield a wide range of efficacies (Vandeputte et al. 2021, 2022c). For many of the compounds evaluated here, the Emax approached the efficacy of DAMGO. For some (isotonitazene (18), N-pyrrolidino isotonitazene (19), N-piperidinyl isotonitazene (20) and the evaluated N-desethyl analogues (6, 12, 17 and 22)), the efficacy even surpassed that of DAMGO, although in most instances the confidence intervals were still overlapping with that of DAMGO. This is in line with observations by others, some of whom have used the term MOR ‘superagonism’ to refer to MOR activation induced by several nitazenes (Malcolm et al. 2023; Tsai et al. 2024). While absolute potencies differed in both our assays, the inclusion of multiple comparator opioids in this study (A–C) allowed a comparative assessment of potencies (Vandeputte et al. 2022c) that was highly similar between both assays. This strengthens the observed nitazene SAR trends, which are discussed below. Importantly, while there are clear limitations to the translatability of in vitro studies to the in vivo situation (e.g., because bioavailability, blood–brain barrier penetration, and formation of active metabolites are not taken into account), prior studies have shown that in vitro functional assays at MOR (including the MOR-βarr2 assay employed here) can serve as a tool for predicting in vivo potency trends of NSOs (including nitazenes) in rodents (Vandeputte et al. 2022b, 2023; Glatfelter et al. 2023). Therefore, the proposed in vitro SAR trends, indicating a (very) high MOR activating potential of many of the investigated compounds, may be cautiously extrapolated to the in vivo setting (Gillis et al. 2020; Tsai et al. 2024). Related to the latter, as elaborated below, the inclusion of forensic case series (although typically confounded by the co-presence of other drugs, including opioids) allowed to corroborate the in vitro findings, by demonstrating that these substances are found at low concentrations in fatal intoxications.

Structural modifications of the hitherto identified nitazenes (Fig. 2) are confined to three specific regions of the 2-benzylbenzimidazole template (R1–3 in Fig. 1 and Table 1) (Glatfelter et al. 2023). Considering the para-benzyl position (R1), prior research has shown that the length of the para-alkoxy side chain dictates the impact on potency, with intermediate chain lengths (i.e., ethyl and isopropyl) being most optimal for MOR activation. Shorter (i.e., methyl) or longer (i.e., propyl and butyl) tails lead to less potent analogues (Vandeputte et al. 2021; Kanamori et al. 2023; Glatfelter et al. 2023; Kozell et al. 2024). Substitution of the alkoxy chain by a halogen atom or free phenol greatly decreases the MOR activity to a level more comparable to that of morphine (Hunger et al. 1960b; Vandeputte et al. 2021; Kozell et al. 2024). The in vitro potency trends of all reference nitazenes included in this study are consistent with the abovementioned findings.

Elimination of the nitro group at the 5-position of the benzimidazole ring (R2) generates so-called ‘desnitro’ analogues or ‘desnitazenes’. Previous in vitro and in vivo data showed that removal of the 5-nitro group leads to a significant reduction in potency (Hunger et al. 1960a; Vandeputte et al. 2021; Kozell et al. 2024). In line with those studies, our findings for metodesnitazene, etodesnitazene, and isotodesnitazene confirm that absence of the nitro group reduces the potency approximately 10–100-fold. The newly characterized protodesnitazene follows this trend, displaying a > 13-fold lower potency than protonitazene in both assays.

Different modifications of the R3 position of the nitazene core structure are possible. The two ethyl groups of the common N,N-diethyl amine group can be substituted for a pyrrolidine or piperidine moiety, leading to the formation of ‘ring’ nitazenes. Previous in vitro studies (Vandeputte et al. 2022b, e; Tsai et al. 2024) showed that both N-pyrrolidino and N-piperidinyl substitutions of the amine side chain are well-tolerated in terms of MOR activation, with N-pyrrolidino etonitazene being equally potent and N-piperidinyl etonitazene being somewhat less potent than etonitazene. Furthermore, Kozell et al. (2024) observed similar trends across a larger panel of ‘ring’ nitazenes. In that study, N-piperidinyl etonitazene and N-piperidinyl isotonitazene displayed a slightly lower or similar potency compared to etonitazene and isotonitazene, respectively, and different N-pyrrolidino analogues (N-pyrrolidino etonitazene, N-pyrrolidino isotonitazene, N-pyrrolidino protonitazene, and N-pyrrolidino metonitazene) displayed a similar or slightly higher potency than their corresponding N,N-diethyl amine equivalents. By contrast, N-pyrrolidino metonitazene was about two times less potent than metonitazene (Kozell et al. 2024). These findings are broadly in line with our results, and further building on ‘ring’ nitazene SAR, we show that these modifications generally yield highly active drugs, with a trend of N-pyrrolidino analogues being more potent than the corresponding N-piperidinyl analogues. A possible exception to this trend was observed for analogues of 4′-OH nitazene: here, the N-piperidinyl analogue displayed a similar or slightly higher potency than the corresponding N-pyrrolidino analogue. Furthermore, with the exception of N-piperidinyl 4′-OH nitazene, piperidine analogues are generally less potent than the corresponding N,N-diethyl amine analogues. The decrease in potency was most pronounced (i.e., 20x) for N-piperidinyl metonitazene, the analogue with the shortest alkoxy side chain. Upon lengthening the alkoxy tail, the potency shift appeared to become progressively smaller, with N-piperidinyl etonitazene being > 5 times less potent and N-piperidinyl protonitazene being equipotent or slightly less potent than the corresponding N,N-diethyl amine equivalents. For N-piperidinyl isotonitazene, no potency shift compared to isotonitazene was observed with the employed assays. In contrast, pyrrolidino analogues were generally about equipotent as their N,N-diethyl amine counterparts.

4′-OH nitazene is a universal metabolite of all nitazenes with alkoxy chain modifications only (Taoussi et al. 2024; Kanamori et al. 2024; Walton et al. 2021). Similarly, based on recent findings for N-piperidinyl etonitazene where the N-piperidinyl 4′-OH nitazene metabolite was identified in authentic urine samples from two patients that took this compound (Vandeputte et al. 2022e; Berardinelli et al. 2024), it can be hypothesized that N-pyrrolidino 4′-OH nitazene and N-piperidinyl 4′-OH nitazene are common O-dealkylated metabolites of the herein evaluated ‘ring’ nitazenes. In support of this hypothesis, N-pyrrolidino 4′-OH nitazene was detected in the blood samples of two N-pyrrolidino protonitazene cases (cases 52 and 61) included in Table S6-2. While a quantitative assessment of this metabolite was not performed, its detection in only 2/41 cases involving N-pyrrolidino metonitazene and/or N-pyrrolidino protonitazene might indicate its presence at considerably lower concentrations than the parent drugs in blood, as also seen for 4′-OH nitazene (Walton et al. 2023). Combined with in vitro findings, which suggest that N-pyrrolidino 4′-OH nitazene and N-piperidinyl 4′-OH nitazene are generally less potent than the evaluated ‘ring’ nitazenes, it is likely that any potential impact of the activity of these metabolites on the overall in vivo effects of the parent ‘ring’ nitazene is limited. Conversely, in the case of N-piperidinyl metonitazene, the expected N-piperidinyl 4′-OH nitazene metabolite may be equally or slightly more active than the parent compound itself. However, dedicated studies on the in vivo human metabolism of ‘ring’ nitazenes are warranted (Berardinelli et al. 2024).

N-Pyrrolidino metodesnitazene and N-pyrrolidino etodesnitazene were included in this study to assess the effect of combining a pyrrolidine function (known to be well-tolerated in terms of MOR activation potential, cfr. supra) with removal of the 5-nitro group, known to negatively impact opioid activity. The opioid activity of N-pyrrolidino etodesnitazene was readily reported in 1960 by means of a mouse tail-flick assay (Hunger et al. 1960a), which revealed an antinociceptive potency 20 times greater than that of morphine. In agreement with these findings, our in vitro data show that N-pyrrolidino etodesnitazene is at least 8 times more potent than morphine. As anticipated, N-pyrrolidino metodesnitazene had a pronounced lower potency (i.e., ~ 100×) than N-pyrrolidino metonitazene. In line with the equipotency observed for N-pyrrolidino metonitazene and metonitazene in our assays, the potencies of N-pyrrolidino metodesnitazene and metodesnitazene are of the same order of magnitude. Similar trends were observed for N-pyrrolidino etodesnitazene. Hence, the combined impact on in vitro MOR activity of the two structural alterations (‘ring’ modification + removal of the 5-nitro group) is roughly equivalent to the cumulative effect of each individual modification, with the substantial loss of potency being attributable to the removal of the 5-nitro group.

Removal of one ethyl moiety from the N,N-diethyl amine side chain (R3) yields N-desethyl nitazenes. Prior pharmacological assessments of N-desethyl etonitazene and N-desethyl isotonitazene showed that these analogues have important MOR activity (Vandeputte et al. 2021, 2023; Walton et al. 2023; Malcolm et al. 2023; Kozell et al. 2024; Tsai et al. 2024). Further exploring these SAR, our results from both MOR-βarr2 and GloSensor® cAMP assays showed that the N-desethyl modification mostly results in a lower potency compared to the respective N,N-diethyl equivalents. Intriguingly, N-desethyl isotonitazene showed the opposite trend, with a ~ threefold increase in potency compared to isotonitazene in both assays. While this is in line with previous in vitro (Vandeputte et al. 2021, 2023; Kozell et al. 2024; Tsai et al. 2024) and in vivo (rat) findings (Walton et al. 2023; Vandeputte et al. 2023), Malcolm et al. (2023) did not observe this trend in in vitro and in vivo assays. This apparent discrepancy might be attributed to the use of different assays and/or experimental conditions.

N-Desethyl etonitazene and N-desethyl isotonitazene are metabolites of etonitazene and isotonitazene, respectively (Taoussi et al. 2024; Kanamori et al. 2024; Krotulski et al. 2020). As previously discussed for N-piperidinyl 4′-OH nitazene, the highly potent N-desethyl metabolites may theoretically contribute to the toxicity of the parent drug. Research (Walton et al. 2023) found that the highest concentration of N-desethyl isotonitazene in rats was at least 20 times lower than that of the parent molecule. Therefore, the in vivo relevance of N-desethyl metabolites may be limited, at least in rats. However, these analogues may pose significant health risks when used as standalone drugs. This is demonstrated by the 16 reported toxicological cases where N-desethyl isotonitazene was detected without isotonitazene, and one case where N-desethyl etonitazene was found without etonitazene. N-Desethyl isotonitazene is of particular concern, as it may combine high MOR binding affinity and functional potency (Walton et al. 2023; Vandeputte et al. 2023; Kozell et al. 2024) with prolonged respiratory depression in vivo (Malcolm et al. 2023). Given that, in protonitazene cases, N-desethyl protonitazene is typically only present to a minor extent, we speculate that the standalone presence of N-desethyl protonitazene in two cases from March 2023 and from June 2023 (Table S6-2/3, cases 73 and 45), in the absence of protonitazene, may indicate that also this metabolite may be available as such. However, confirmation of this scenario requires the detection of N-desethyl protonitazene without protonitazene in more toxicology cases and/or in drug material.

When considering the different modifications (R1–3) together, it is interesting to note that the ‘classic’ nitazenes differing in the alkoxy chain length (represented by the black curves in Fig. 3A–D) are generally the most potent (i.e., most left-shifted) of the respective panels. With the exception of N-pyrrolidino analogues, all other