Metabolite detection and elucidationFirst approach: targeted screening

The first step for detecting metabolites involved a targeted screening, in which metabolites previously described in the literature were directly searched for. For instance, Fig. 2 shows the nw-EICs and the corresponding fragmentation spectrum for CUMYL-4CN-BINACA and three of its metabolites found in rat serum, which are consistent with previously reported in vitro data. The spectra confirmed their identity, as most of the observed fragment ions matched those reported in the in vitro metabolism study (Öztürk et al. 2018).

Fig. 2

Chromatograms (left) and fragmentation spectra (right) for CUMYL-4CN-BINACA and 3 selected metabolites found in the rat serum of this in vivo study. Green ticks indicate fragment ions that match the previously reported in vitro metabolism data (Öztürk et al. 2018)

Second approach: expected biotransformations

Moving to the next approach, hydroxylation is one of the most common reactions in the metabolism of SCRAs and therefore it was included in the list of expected biotransformations. Figure 3A displays the nw-EIC corresponding to a hydroxylated metabolite of 5 F-AB-PINACA in urine (exact mass m/z 365.1989, + 16 Da shift compared to the parent compound) in the 7.8–9 min chromatogram range. The presence of multiple chromatographic peaks at this m/z suggests the formation of several hydroxylated metabolites.

Fig. 3

Identification of three hydroxylated 5 F-AB-PINACA metabolites by the expected biotransformation strategy. A nw-EIC for m/z 365.1985, corresponding to the hydroxylated metabolite in urine. B LE (left) and HE (right) spectra for the parent compound and metabolites M4, M8 and M6. Purple text and dashed lines refer to 5 F-AB-PINACA, blue to M4, orange to M8 and green to M6. Mass errors are calculated in parts per million (ppm)

As an example, Fig. 3B shows the elucidation of three hydroxylated metabolites eluting at 8.25 min (M4), 8.59 min (M6) and 8.85 min (M8). The LE spectra of these metabolites (Fig. 3B, left) revealed mass errors lower than 4 ppm. The mass shift of + 16 Da was also observed in the HE spectra for the fragment ion at m/z 320 (Fig. 3B, right).

For M4 (8.25 min), the fragment ion at m/z 249 shows the + 16 Da shift relative to the parent fragment at m/z 233, indicating hydroxylation on either the indazole ring or the fluoropentyl chain. The presence of the shared fragment at m/z 145 rules out modification on the indazole ring, confirming hydroxylation on the fluoropentyl tail.

In contrast, M8 (8.85 min) exhibits the + 16 Da shift only in the fragment ion at m/z 320, with the remaining fragments identical to those of the parent. This suggests that hydroxylation likely occurred in the methyl butanamide moiety, most plausibly within the isopropyl group.

Finally, for M6 (8.59 min), in addition to the fragment at m/z 320, only the fragment at m/z 161 was observed, showing a + 16 Da shift relative to the parent fragment at m/z 145, which corresponds to the core structure. This pattern suggests that hydroxylation took place on the indazole ring.

Third approach: common fragmentation

SCRAs containing an indazole core and a fluoropentyl tail typically produce a characteristic fragment ion at m/z 213.1022. This fragment was therefore employed to identify metabolites following the same fragmentation pathway. Figure 4A shows the nw-EIC of this fragment ion in the serum sample from a rat administered with 5 F-AMB-PINACA. Besides the peak at 11.61 min, corresponding to the parent compound, three additional peaks were observed, indicative of metabolites structurally related to the parent compound.

Fig. 4

Identification of three 5 F-AMB-PINACA metabolites by common fragmentation pathway. A nw-EIC for m/z 213.1022, corresponding to a characteristic fragment ion in serum. B LE (left) and HE (right) spectra for the parent compound and metabolites M1, M4 and M6. Purple text and dashed lines refer to 5 F-AMB-PINACA, blue to M1, orange to M4 and green to M6. Mass errors are calculated in parts per million (ppm)

As shown in Fig. 4B (right), these three metabolites share the same fragmentation spectra. The presence of the fragment ion at m/z 233, corresponding to the indazole core with the fluoropentyl tail, indicates that the biotransformation occurred in the linked group. Examination of the LE spectra (Fig. 4B, left) allowed further elucidation.

M1 (eluting at 10.75 min, m/z 350.1883, 0.9 ppm) exhibited a mass shift of -14 Da relative to the protonated molecule of 5 F-AMB-PINACA, consistent with an O-demethylation. M4 (9.63 min, m/z 366.1836, 1.9 ppm) showed a + 16 Da shift relative to M1, indicating hydroxylation of the isopropyl moiety following O-demethylation. Finally, M6 (9.07 min, m/z 394.1781, 0.8 ppm) presented a + 30 Da shift compared with 5 F-AMB-PINACA, suggesting oxidation to carboxylic acid on the isopropyl moiety.

In vivo metabolism assessment

Comprehensive information on chromatographic retention times, ionization modes, elemental compositions, accurate-mass fragmentations, and mass errors for all elucidated metabolites is shown in Tables S1-S8. Additionally, they include the relative abundance of each metabolite as well as the corresponding identification confidence level according to the classification proposed by Schymanski et al. (2014). Figures S1-S8 illustrate the proposed metabolic pathways for each of the SCRAs studied. These figures specify the matrix type in which the metabolites were detected and whether they have also been previously reported in in vitro studies (Fabregat-Safont et al. 2022; Andersson et al. 2016; Öztürk et al. 2018; Stalberga et al. 2023; Gandhi et al. 2013; Kavanagh et al. 2017; Carlier et al. 2017; Wohlfarth et al. 2015; Diao et al. 2016a, 2018).

For AMB-FUBINACA, a total of 8 metabolites were detected (Figure S1, Table S1). The predominant metabolite was M1, formed through O-demethylation. After O-demethylation, hydroxylation on the isopropyl group (M2, M3), lactone formation (M6) and N-dealkylation (M8) were also observed. N-methylation after O-demethylation and N-dealkylation yielded M7, a biotransformation previously reported in related studies (Fabregat-Safont et al. 2019, 2022). Additionally, oxidation to carboxylic acid on the isopropyl group produced M4 and M5.

ADB-FUBINACA generated 7 metabolites (Figure S2, Table S2), with hydroxylated (M1) and N-dealkylated (M7) metabolites being the most abundant. M1 gave rise to M2 through cyclization, to M3 through N-dealkylation and to M4 via amide hydrolysis and cyclization. Subsequent oxidation of M2 produced M5, which underwent N-dealkylation to produce M6.

Four metabolites were detected for SDB-005 (Figure S3, Table S3). O-denaphtylation yielded M1, for which, curiously, two chromatographic peaks appeared (at 10.54 and 11.0 min) even though there are no chiral or isomerization sites in the structure. Moreover, the retention times of both peaks differ significantly from the SDB-005 standard (13.56 min), indicating that neither corresponds to an in-source fragment of the parent compound. Thus, the exact origin of this phenomenon cannot be unambiguously determined based on the available data and may warrant further investigation. Subsequently, M1 gave rise to an O-glucuronide metabolite (M2) and two hydroxylated metabolites on the pentyl tail (M3, M4).

APINACA produced 7 metabolites (Figure S4, Table S4), resulting from one to three hydroxylations on the adamantyl group (M1–M4) or simultaneous hydroxylations on the adamantyl and pentyl groups (M5–M7).

For CUMYL-4CN-BINACA, 4 metabolites were identified (Figure S5, Table S5). M1, M2 and M3 arose from oxidation to carboxylic acid of the cumyl moiety, decyanation to COOH, and N-dealkylation, respectively. M4 resulted from cumyl hydroxylation of M2.

5 F-AMB-PINACA presented 22 metabolites (Figure S6, Table S6). The most abundant was the O-demethyl metabolite (M1), followed by mono- and di-hydroxylated compounds (M2-M4, M10-M12). Hydroxylations occurred mainly in the isopropyl (linked group) and fluoropentyl (tail) moieties, being also observed in the indazole ring (M19). Additional biotransformations included defluorination to OH (M9) or COOH (M8, M15), N-dealkylation (M5, M18, M19), and sulphation (M21, M22), among others.

The compound with the greatest metabolic variety was 5 F-AB-PINACA, with 33 metabolites (Figure S7, Table S7). Predominant modifications included hydroxylation on the fluoropentyl tail (M2-M5), dihydroxylation (M12-M17) and defluorination to OH (M30) or COOH (M28). The remaining metabolites arose from additional hydroxylations (M1, M6-M8), N-dealkylation (M9), amide hydrolysis (M10), or combinations thereof (M11, M18-M23). Less common biotransformations, such as epoxide formation followed by hydrolysis to a dihydrodiol (M24), were also observed.

For THJ-2201 (Figure S8, Table S8), the metabolic pathway was simpler, with only 3 metabolites. M1 was formed by defluorination to COOH, M2 via subsequent hydroxylation on the naphthalene moiety, and M3 following a second hydroxylation.

Comparison between rat in vivo metabolism and reported human data

The metabolic profiles obtained in this study were compared with the available human data to evaluate the translational relevance of the rat in vivo model. Nevertheless, human metabolic data are currently available for only five of the eight compounds investigated herein.

For AMB-FUBINACA, the detected metabolites showed a high degree of concordance with those previously reported in human samples, specifically M1 (O-demethylation), which has been described as a key biomarker for quantification in forensic blood samples (Presley et al. 2020).

Similarly, the metabolites identified for ADB-FUBINACA, namely M1 (OH-tert-butyl), M2 (OH-tert-butyl, cyclization) and M5 (OH-tert-butyl, oxidation, cyclization), were consistent with those previously detected in post-mortem blood and urine samples (Kovács et al. 2019; Kavanagh et al. 2017).

In the case of APINACA, the identification of metabolites M1 and M2, corresponding to mono- and di-hydroxylation on the adamantyl moiety, agrees with findings reported in authentic human urine (Vikingsson et al. 2015).

Furthermore, the prevalence of M2 (decyanation to carboxylic acid) for CUMYL-4CN-BINACA is in agreement with its recognition as a major human biomarker in human samples (Öztürk et al. 2018), together with the detection of M4 (decyanation and cumyl hydroxylation).

Finally, for THJ-2201, the detection of M1 (defluorination to COOH) further supports the consistency between rat and human phase I metabolic pathways (Gaunitz et al. 2019).

Although no human data are currently available for the remaining three compounds, the strong agreement observed for the other five SCRAs supports the utility of the rat in vivo model as a valuable approach for metabolic characterization. It not only confirms previously described metabolic routes but also may provide a predictive framework for the identification of biomarkers of emerging SCRAs for which human reference data are still lacking.