Identification of N-debenzisothiazole-lurasidone as an enzymatic degradation product of lurasidone

Abstract

Herein we report the confirmation of N-debenzisothiazole-lurasidone as a lurasidone degradation product associated with postmortem toxicology casework. Confirmation was realized by unambiguous synthesis and comparison of its LC–MS/MS properties to the lurasidone degradation product in postmortem blood using liquid chromatography quadrupole-time-of-flight mass spectrometry (LC–QTOF/MS). The mechanism of formation of this degradant in blood is proposed to be primarily enzymatic, given it has only been previously presumptively reported in one in vitro stability study following oxidative stress conditions. It has not been reported in other in vivo pharmacokinetic and in vitro stability studies assessing lurasidone stability toward acid, alkali, oxidation, photolysis, and heat. The detection of N-debenzisothiazole-lurasidone in postmortem casework indicates that, where possible, it should be included in toxicology screening methods targeting psychoactive compounds. Until such time that a commercially available reference standard of N-debenzisothiazole-lurasidone is available, the comprehensive accurate mass and mass spectral data of N-debenzisothiazole-lurasidone that are now available enable its inclusion as a “suspect target” in high-resolution MS screening methods.

Introduction

Lurasidone (1, Fig. 1) is an antipsychotic indicated for treating schizophrenia and bipolar depression as either monotherapy or an adjuvant [1–5]. It has been available clinically since 2010 [6]. The most common adverse side effects associated with lurasidone use are akathisia and extrapyramidal symptoms [7–10]. Lurasidone administration is contraindicated in conjunction with strong CYP3A4 inhibitors and inducers [1], and case reports have indicated that co-administration with CYP3A4 inhibitors such as fluoxetine and atazanavir may slow lurasidone metabolism and increase its effective dose [11, 12]. US poison center data and the scarcity of case reports regarding lurasidone toxicity indicate that lurasidone may be relatively safe in overdose [13–16]. In one case of acute toxicity, ingestion of 1360 mg of lurasidone, a dose 8.5 times greater than an indicated 160 mg dose [1], in combination with 5 mg clonazepam only required fluids as supportive treatment [17]. Nevertheless, potentially life-threatening side effects such as neuroleptic malignant syndrome and neutropenia have been associated with its use [18–21].

Figure 1.

Molecular structures for Lurasidone, compound 1, exo-4-aza-tricyclo[5.2.1.0*2,6*]decane-3,5-dione (exo-ATTD top right) and tert-butyl piperazine-1-carboxylate (BOC-piperazine, bottom right).

Irrespective of the seemingly low risk of toxicity associated with lurasidone use, determining its presence in death investigations may be relevant for establishing medication adherence, cause of death or the circumstances surrounding the cause of death.

The postmortem proliferation of microorganisms throughout the human body may lead to the decomposition of both biological matrices collected for analysis, such as blood, and some drugs and poisons therein [22]. This putrefaction may greatly hinder analyses, as toxicology screening methods chiefly rely on targeting known drugs, metabolites, or degradation products. Therefore, in order to effectively screen for the presence of drugs and poisons in postmortem specimens, it is necessary to understand their stability in these specimens, and, if required, include degradation products arising from microbial activity in screening methods that can serve as marker compounds of parent-drug administration.

For this purpose, our research group investigated the stability of lurasidone and found it degraded to yield a major, unknown degradation product in unpreserved ante-mortem whole blood inoculated with human fecal microbiota transplant material incubated at 37°C or stored at ambient temperature [23]. This source of microorganisms was chosen as an inoculant it contains species present in the gastrointestinal tract that may proliferate and translocate throughout the body after death. The use of postmortem material as a source of microorganisms was not practical due to local laws. Significant degradation was not observed in preserved specimens or non-inoculated controls, or in specimens stored at −20 and 4 °C, indicating a microbially mediated mechanism of degradation may be responsible [23]. Using liquid chromatography quadrupole-time-of-flight mass spectrometry (LC–QTOF/MS), the major degradation product detected in these in vitro experiments was confirmed to be present in authentic postmortem toxicology casework [24]; therefore, identification of the degradation product was warranted. On the basis of accurate mass data, mass fragmentation data, and data from acetylation experiments, its identity was proposed to be (3aR, 4S, 7R, 7aS)-2-(((1R, 2R)-2-(piperazin-1-ylmethyl)cyclohexyl)methyl)hexahydro-1H-4,7-methanoisoindole-1,3(2H)-dione, 2, Fig. 2, abbreviated to ND-LURA [24]. This proposition was further supported by consideration of the related antipsychotics risperidone and paliperidone, whose 1,2-benzisoxasole moiety is isosterically similar to the 1,2-benzisothiazole moiety present in lurasidone; both risperidone and paliperidone are known to be unstable in postmortem blood [25, 26]. Recently, ND-LURA has also been tentatively identified using LC–QTOF/MS as a degradation product of lurasidone under oxidative conditions (3% H₂O₂, 60°C, 3 hours) [27]. This is in contrast to an earlier study where ND-LURA was not reported following forced degradation of lurasidone using oxidative conditions for 24 hours at room temperature [28].

Figure 2.

Overview for the synthesis of ND-LURA, compound 2. 

As a commercial reference standard for ND-LURA was not available to enable conclusive confirmation of its presence in in vitro experiments and postmortem blood samples, its synthesis was required. Here we report the synthesis of ND-LURA and its confirmation as a major degradation product of lurasidone.

Experimental, synthesis overview

In order to determine whether ND-LURA is a microbial metabolite of lurasidone, its synthesis was attempted using a 7-step pathway (Fig. 2) that was based on the procedures described by Ganesh et al. [29], Ae and Fujiwara [30] and Mishra et al. [31] for the synthesis of lurasidone. Ganesh’s method, which commences with the readily available starting material 1,2-cyclohexanedicarboxylic anhydride compound 3, was followed closely for the synthesis of the cis-diester compound 4 and its epimerization into a mixture of diesters compound 5 and compound 4 (in the ratio of approximately 10:1, respectively). In subsequent synthetic steps that made use of the diester, 10:1 mixture of epimers were obtained. Part-way through the research described here it became possible to ­purchase the diol compound 6 in high epimeric purity; this commercial product was used for the synthesis of the dimesylate compound 7 and subsequent products contained only trans isomers.

The trans-cyclohexyl moiety in lurasidone has two chiral centres (analogous to those marked with asterisks in compound 5 in Fig. 2) and as lurasidone is sold as a single enantiomer, a resolution step is required in its manufacture. This can be conveniently accomplished via hydrolysis of the diester compound 5 into its diacid salt, which is then resolved using an appropriate optically active amine. However, in regard to the synthesis of ND-LURA for the purpose of identifying whether it is the microbial degradation product detected in our studies, an enantioselective synthetic sequence is not required as our routine toxicological analytical technique (LC–MS/MS) is not enantioselective. Therefore, synthesis was carried out using the methods based on those described by Ganesh et al. [29] with some modifications based on the work of Ae and Fujiwara [30] and Mishra et al. [31] but resolution of starting materials was not involved.

Published syntheses of lurasidone involve the treatment of dimesylate compound 7 with the nucleophile 3-(1-piperazinyl)-1,2-benzisothiazole to form a quarternary ammonium salt that is in turn treated with another nucleophile, exo-4-aza-tricyclo[5.2.1.0*2,6*]decane-3,5-dione (abbreviated to exo-ATTD), in order to achieve the final product. However, as the benzisothiazole moiety is absent in ND-LURA, in theory all that is required for the synthesis of ND-LURA is to treat compound 8 (which is analogous to the quaternary ammonium salt referred to above for the synthesis of lurasidone) with piperazine. Whilst the use of piperazine was successful, in the end it was decided to use tert-butyl piperazine-1-carboxylate (t-BOC-piperazine) rather than piperazine because the latter might react at both nitrogen atoms, thus forming an unwanted dimer and reducing overall yield of the desired product. Reaction between compound 8 and exo-ATTD and subsequent hydrolysis of the product, compound 9, using phosphoric acid (84%) afforded a product identified as ND-LURA. As well as ND-LURA, compound 8 and compound 9 have not been described fully in literature hitherto; consequently, full descriptions of the synthetic procedures used and detailed NMR and MS data for compound 8, compound 9 as well as ND-LURA are presented here in the Supplementary Information section.

Results and discussion

Lurasidone is sold as a single enantiomer (i.e. (3aR, 4S, 7R, 7aS)-2-{(1R, 2R)-2-[4-(1,2-benzisothiazol-3-ylpiperazin-1-ylmethyl]cyclohexylmethyl} hexahydro-4,7-methano-2H-isoindole-1,3-dione) and as a consequence ND-LURA produced in vivo is also likely to be formed as a single enantiomer (i.e. (3aR, 4S, 7R, 7aS)-2-{(1R, 2R)-2-[piperazin-1-ylmethyl]cyclohexylmethyl}hexahydro-1H-4,7-methano-2H-isoindole-1,3-dione)). As indicated above, a resolution step was not included in the synthesis carried out for the research described here, and therefore the ND-LURA produced would be a mixture of 1R, 2R and 1S, 2S enantiomers. This is not a significant limitation; however, as analytical techniques applied here, previously [23, 24] and in the majority of forensic toxicology laboratories would not be capable of resolving analytically the 2 isomers of ND-LURA. Compounds 4, 5, 6, and 7 were ­synthesized using literature methods and the identity of these synthesis products could be verified by comparison of NMR spectral data with published data.

However, the synthesis objective (i.e. ND-LURA) and compounds 8 and 9 were novel and the structures of these products required appropriate characterization and confirmation. High-resolution MS showed that the putative ND-LURA product had a monoisotopic mass of 360.2633, which was 3.6 ppm from the calculated monoisotopic mass (360.2646). Although four other compounds have monoisotopic masses within 5 ppm of 360.2633 (i.e. C₁₉H₃₂N₆O, C₁₈H₃₆N₂O_5, C₁₇H₃₀N_9, C₄H₂₈N₁₈O₂), none of these could be rational products arising from the starting materials and synthetic methods used. It has been determined [27, 32, 33] that lurasidone is prone to decomposition under either basic, photolytic, or oxidative conditions and as a result there is the possibility that ND-LURA may also be susceptible under these conditions. However, ND-LURA degradation products analogous to those reported for lurasidone would not display a [M + H]⁺ peak at m/z 360.2633. Additionally, the photolytic rearrangement takes place at the benzisothiazole moiety of lurasidone and therefore is not a concern in the synthesis of ND-LURA.

The assignment of structure to the final product of the synthetic sequence (i.e. ND-LURA) relied heavily upon its NMR spectral data, especially in comparison with data obtained from lurasidone (extracted from Latuda tablets, see Figure S21, Supplementary Information), data published by Silvermann [34] (who carried out synthesis of partially deuterated lurasidone) and NMR data (see Supplementary Information Figures S1–S20 and Table 1) acquired from starting materials during the course of the synthetic work. In addition, NMR spectral data for lurasidone are also provided by Wang et al. [34], Gamboa-Arancibia [35] and Siddig [27]. However, data from the former were collected using lurasidone HCl salt and those provided by the latter two were collected using DMSO as solvent. The data by Gamboa-Arancibia [35] and Siddig [27] were not used for structural assignment here, but the ¹³C assignments from Wang et al. [34] were particularly valuable, as were chemical shifts for the cyclohexyl protons relative to one another rather than in an absolute sense due to the perturbing influence of the protonated nitrogen atom.

Table 1.

¹H and ¹³C NMR signal assignment for ND-LURA

C Environment	13C (δ, ppm)	H Environment	1H (δ, ppm)
1	46.20	HG	2.85 (s, 4H)
2	55.00	HL	2.37 (s, 4H)
3	64.23	HA & HB	2.48 & 2.11 (AMX, 2H, J = 14.4, 9.8, 6.8 Hz)
4	37.51	HI	1.25 (m, 6H)a
5	30.83	HQ	1.83 (m, 1H) & 0.97 (m, 2H)a
6	25.53, 25.13	HO & HP	1.64 (m, 4H)a & 1.25 (6H) or 1.11 (m, 2H)a
7
8	29.90	HN	1.48 (2H) & 0.97 (2H)
9	40.62	HM	1.48 (m, 2H)a
10	42.80	HH & HF	3.87 & 3.28 (AMX, 2H, J = 13.1, 10.5, 3.9 Hz)
11	39.84, 39.81	HC	2.58 (s, 2H)
12	48.78, 48.73	HE	2.69 (s, 2H)
13	28.24, 28.18	HK	1.64 (m, 4H)a & 1.25 (m, 6H)a
14	33.40	HD & HJ	1.25 (6H) & 1.11 (2H)
–	–	N-H	1.25 (m, 6H)a
C=O	179.5	–	–

^aComplex multiplets, refer to discussion.

A sample of synthesized ND-LURA was analyzed using LC–QTOF/MS (analytical conditions used are described in Supplementary Information and reference [36]) contemporaneously with an in vitro experimental LURA degradation sample containing the compound proposed to be ND-LURA, which has previously been detected in two postmortem blood samples [24]. The precursor mass (measured m/z [M + H]⁺ = 360.2647, 0.28 ppm to theoretical) MS/MS spectra (see Supplementary Information) and retention time data (see Supplementary Information) for synthesized ND-LURA were in agreement with those observed for the degradation product (retention time difference 0.031 mins, measured m/z [M + H]⁺ = 360.2644, −0.56 ppm) and products detected in the two postmortem blood samples.

Thus, the identity of the major postmortem lurasidone degradation product was confirmed to be ND-LURA, making this report the first time this compound has been definitively reported in literature.

With the high-resolution MS results presented here and earlier [24], ND-LURA can now be included in toxicology screening methods to indicate cases where lurasidone quantitation may be unreliable or those cases where the parent drug has completely degraded.

Conclusion

ND-LURA was unambiguously synthesized and characterized by nuclear magnetic resonance (NMR) spectroscopy and high-resolution MS. Through LC–QTOF/MS, ND-LURA was confirmed to be the major degradation product of lurasidone previously reported in in vitro degradation experiments and postmortem toxicology casework. Although ND-LURA is not available as a commercial reference standard, comprehensive accurate mass and mass spectral data are now available to enable its inclusion in high-resolution MS screening methods.

Author contributions

Kyle Farrell (Formal analysis, Investigation, Methodology, Writing—original draft, Writing—review & editing), Tristan Fraser (Conceptualization, Formal analysis, Investigation, Methodology, Writing—original draft), Danielle Butzbach (Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Supervision, Writing—original draft, Writing—review & editing), Jared W. Castle (Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Writing—original draft, Writing—review & editing), and Paul Kirkbride (Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Writing—original draft, Writing—review & editing)

Supplementary data

Supplementary data is available at Journal of Analytical Toxicology online.

Funding

None declared.

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.