Detection of Undeclared Halogen Substituted Drug Compound in a Natural Health Product

1Centre for Applied Research & Innovation, British Columbia Institute of Technology, Burnaby, BC, Canada, V5G3H2 2Department of Biology, University of British Columbia, Kelowna, BC, Canada 3Isura, 501-3292 Production Way, Burnaby, BC, V5A4R4 4Flora Research Laboratories LLC, 1000 SE M Street Unit B, Grants Pass, Oregon, USA, 97526 5Academy of Integrative Health & Medicine, 6919 La Jolla Blvd., La Jolla, CA 92037 * Corresponding Author: Paula_Brown@bcit.ca ORIGINAL RESEARCH OPEN ACCESS


Introduction
Natural Health Products (NHPs) in Canada are defined and subject to the requirements set out in the Natural Health Products Regulations and include vitamins and minerals, herbal remedies, homeopathic medicines, traditional medicines, probiotics and other products like amino acids and essential fatty acids and are restricted to oral, topical or sublingual routes of administration [1].
These products cannot be prescription products, drugs administered by puncturing the skin or any substance that is regulated under the Tobacco or the Controlled Drugs and Substances Acts of Canada [2]. Although these products are considered a subset of drugs, they are exempt from many of the requirements in the Food and Drugs Regulations unless specifically referenced in the Natural Health Products Regulations under the authority of the Natural and Non-prescription Health Products Directorate (NNHPD) [1]. All manufacturers, packagers, labellers and importers of natural health products are required to hold a site license, for which they must demonstrate that they are in compliance with the Good Manufacturing Practices (GMP) requirements set out in the regulations [1]. NHPs are required to undergo a premarket approval process and obtain a license prior to entering the market [1].
Under the existing regulatory framework post-market activities for NHPs are predominately risk-based and complaint driven. In a Health Canada report published in 2016, it recommended the use of proactive tools including inspections to balance the current paper-based licensing process [3]. While Health Canada does engage in a number of post-market activities to limit potential health risks to Canadians, the 2016 report acknowledges that challenges remain with the largely reactive approach, recommending conducting on-site inspections, more laboratory testing as part of compliance  [17]. https://doi.org/10.33211/jnhpr. 8 verification, and the need for stronger post-market authority [3].
The U-Dream line of products are licensed natural health products (NHPs) marketed in Canada by Biotrade Canada Ltd. and are sold as over the counter sleep aids (Figure 1) [17]. As of October 10 2019, the U-Dream Full Night product with Natural Health Product Number (NPN) 80070691 holds the #3 Amazon Bestsellers Rank in Medicinal Sleep Aids [18]. As of August 23 rd , 2019, this line of products is also present as a dietary supplement in the United States through amazon.com [19].
The product is purported to be an all-natural herbal medicine consisting of several botanical extracts and L-tryptophan. There are different formulations of the U-Dream product on the Canadian market with slight differences in presence and levels of declared herbal medicinal ingredients; as an example the medicinal ingredients for the U-dream Full product with NPN 80088078 and U-Dream Full with NPN 80070691 are shown in Tables 1 and 2. All U-Dream products state they are 100% natural, and products with different NPN license numbers have essentially the same external packaging color, style, and design.
The U-Dream products have found significant success in the market with many reviews touting their effectiveness as a sleep aid. However, some testimonials associated with the product are concerning such as consumers describing symptoms not typically associated with the listed ingredients, reports that the product left a metallic taste in the mouth following ingestion with others describing serious withdrawal symptoms after only short-term use [18,19]. These effects are not included in the cautions or risk warnings for the product and are more typically associated with prescription pharmaceuticals rather than an over-the-counter NHP. With concerns for the risk to public health, an investigation of the product to screen for undeclared pharmaceutical ingredients was initiated with an initial focus on known prescription drugs used to treat insomnia such as Lorazepam. Preliminary screens of the product did not detect known pharmaceutical sleep aids but more detailed analysis indicated the presence of an unknown suspicious compound. This paper describes the further analytical efforts made to determine the identity of this unknown, undeclared compound.

Methods
Solvents and reagents for the analytical work (Formic acid, LC grade methanol, acetonitrile) were purchased from VWR (Edmonton, AB, CAN), Sigma-Aldrich (Oakville, ON, CAN) and Fisher Scientific (Ottawa, ON, CAN). U-dream capsules were obtained from a local store in Burnaby, British Columbia, Canada. Two products were obtained, U-Dream Full 450 mg, NPN 80088078, lot number UDF008 and U-Dream Full, NPN 80070691, lot number UDF004.

Mass Spectral Analysis for Lorazepam
An intact sealed capsule was removed from the product package and analyzed for the presence of lorazepam. Approximately 100 mg of powder was extracted in acetonitrile:water (1:1) using sonication followed  to find discrete chemical entities; auto MS/MS "find compound" mode.

HR-LCMS Screening for Undeclared Compounds
For each U-Dream product package, all capsules were opened and the contents, a yellow powder, pooled. 50 mg of the pooled powder was weighed into a 15-mL polypropylene tube and 1600 µL of methanol and 400 µL of water added using volumetric pipettes. Samples were mixed on a vortex mixer for 10 seconds and centrifuged at 3,000 g for 5 minutes. The supernatant was filtered through a 0.2-µm syringe filter into a microcentrifuge tube, then centrifuged at 15000 g for 5 minutes. The supernatant was then transferred to an HPLC sample vial for analysis. Samples were analyzed on a Vanquish uHPLC equipped with a Q-Exactive Orbitrap (Thermo Fisher Scientific, Waltham, MA, USA). Separation was achieved on a Kinetix Polar C18, 2.6 µm, 100 × 3.0 mm column (Phenomenex, Torrance, CA, USA) set at 40°C using a gradient (Table 3) with 0.1% formic acid in water as Mobile Phase A and 0.1% formic acid in methanol as Mobile Phase B. Injection volume was 1 µL and flow rate was 0.5 mL/min. The DAD was set to store at 215, 230, 250, and 280 nm. The Q-Exactive was operated in full scan as well as full scan product ion modes. Positive ESI was used for all analyses. Spray voltage was set at 4000 V, capillary temperature was 320°C, sheath gas at 10 units, auxiliary gas at 40 units, probe heater at 30°C and S-lens level at 50.

Isolation, Structural Elucidation and Activity Testing of Undeclared Compound
Sixteen capsules from the U-Dream Full 450 mg product, lot number UDF008, were opened and the contents pooled. Approximately 30 mg of the material was weighed into a test tube with 600 µL water and 1400 µL methanol and placed in an ultrasonic water bath for 10 minutes. The liquid portion was then filtered through a 0.2-µm syringe filter into HPLC vials. Samples were injected into an Agilent 1100 HPLC equipped with a quaternary pump, auto-sampler, column compartment, diode array detector and Gilson FC203B fraction collector. Separation was achieved with a Phenomenex Kinetex C18 5 µm, 100 × 4.6 mm column (Phenomenex, Torrance, CA, USA) using a gradient of 5% B at 0 min., 40% B at 6.0 min., 90% B at 9.3 min., 5% B at 10.0 min., with 0.1% formic water as Mobile Phase A and 0.1% formic acid in methanol as Mobile Phase B. Injection volume was 10 µL with a flow rate of 1.6 mL/min and detector set to store signals at 210, 254, 280, 330 and 510 nm. The peak eluting at 5.2 min. was collected with the fraction collector programmed to collect in a time window from 5.1 to 5.7 min. Fractions from multiple runs were pooled together and dried under nitrogen for further evaluation. A portion of the pooled isolated fraction was dissolved in chloroform-d and transferred through a glass wool filter to an NMR tube for analysis using a Bruker AVANCE 400 spectrometer equipped with Bruker Topspin software (Bruker, Ltd, Milton, ON, CAN). 1 H and two dimensional DFQ-COSY NMR experiments were conducted with chemical shift values presented in δ (ppm) and referenced to the residual solvent signal of CDCl 3 . NMR data files were processed and analyzed using Mnova 11.0.4 (Mestrelab Research, Escondido, CA, USA).
Commercial zopiclone/eszopiclone ELISA kits were purchased from Neogen (Lexington, KY, USA). The U-dream samples were analysed using these kits as per

Mass Spectral Analysis for Lorazepam
Neither lorazepam nor lorazepam glucuronide were detected in the sample using the described method. Examination of the mass spectral acquisition analysis in Mass Hunter to find discrete chemical entities revealed the presence of a prominent peak showing a halogen isotope cluster suggesting the presence of a brominated compound (Figure 2).

HR-LCMS Screening for Undeclared Compounds
Full scan positive ESI data were acquired from each of the U-Dream products using the Q-Exactive Orbitrap. For both products lots, the compound detected in  the initial screen was identified as a prominent peak at 3.94 min in the TIC. The full scan mass spectrum of this compound is shown in Figure 3. The spectrum shows the H + -adduct at 433.0620 m/z and the associated Na + -adduct at 455.0432 m/z as well as ions at 435.0589 m/z and 457.0410 m/z. The near 1:1 ratio and mass differences between these adducts and the ions at 435.0589 m/z and 457.0410 m/z provides strong evidence for the presence of a single bromine atom in this compound. This identification is further supported by the presence of similar isotopic ratios for other in-source fragmentation products observed in the full scan mass spectrum shown in Figure 3.
The theoretical mass spectra of several known halogen-containing compounds with sedative activity, were compared against the mass spectrum of the unknown compound. It was observed that the substitution of the bromine atom in the observed mass for the unknown compound with a chlorine atom would result in a mass of 389 g/mol. As the unknown compound is represented in the mass spectra as an H + -adduct at 433.0620 m/z and an Na + -adduct at 455.0432 m/z the target mass of the analogue drug would have a mass of 388 g/mol which is consistent with the known prescription drug zopiclone (monoisotopic mass 388.1051 g/mol), shown in Figure 4 along with the suggested structure of the unknown compound.
The major fragmentation m/z values for several chlorine-containing drug compounds were obtained from MassBank [20]. The reference data was then compared to the observed fragments from the unknown, recalculated to account for the substitution of the chlorine atom with a bromine atom. In Table 4 the calculated modified reference m/z values for zopiclone fragments is compared with the observed fragments in mass spectrum of the unknown compound, following MS 2 analysis on the 433 m/z product ion.
Retrospective analysis of the full spectrum scan for the peak also showed A+2 ions at near a 1:1 ratio, corresponding to the major m/z ions suspected to contain bromine atoms, including ions at 157.9, 184.9, 290.9, 291.9, 308.9, 322.9 and 391.0.

Isolation, Structural Elucidation and Activity Testing of Undeclared Compound
The pooled fraction isolated by HPLC containing the compound of interest was analyzed by NMR and the one-dimensional 1 H spectra is shown in Figure 5.
The presence of a large water signal at 1.6 ppm and several other signals indicated the fraction was not pure, however several hallmark signals were observed that are indicative of aromatic protons, consistent with the structure proposed in Figure 4. The NMR spectrum showed a series of doublets in the range of 7.95-8.92 ppm that were assigned to protons in pyridyl and pyrazinyl rings   Table 4. Comparison of major m/z for zopiclone reported in Massbank [20], theoretical m/z for a zopiclone compound with its chlorine replaced with a bromine atom and observed m/z observed from the MS/MS analysis of the 433 m/z product ion in the U-Dream product.
( Figure 5). Observation of correlations of the 8.89 and 8.85 ppm signals as well as with the 8.52, 8.48 and 7.95 ppm signals in 1 H-1 H COSY spectrum further confirmed these assignments ( Figure 6). The isolated fraction tested positive in a commercial zopiclone/ eszopiclone ELISA kit.

Discussion
Several studies performed over the last decade have shown intentional adulteration of natural health products and dietary supplements by the addition of pharmaceutical drugs [9][10][11][12][13][14]. Certain categories of products are often found to be adulterated, including products for weight loss, body building, erectile dysfunction, sleep problems, inflammatory conditions and metabolic disorders [8][9][10][11][12][13]. The complex matrices of these products can be analytically challenging for detection of the adulterants [13][14][15]. Further the use of analogs of those substances, with potentially unknown safety or pharmacological activity, not only makes detection very difficult, it renders the products themselves an unquantifiable public health risk [14,15].
The reports of effects in consumers, not typically associated with the herbal ingredients listed on the U-Dream label, prompted an investigation that began with screening for known pharmaceutical drugs. The combination of the characteristic bromine isotope pattern observed in full scan accurate mass spectrum suggested the presence of a halogenated compound and prompted further investigations. From the MS/MS analysis of the 433 m/z product ion found in the U-Dream product it was observed the aligned fragmentation pattern (Table 4) of the unknown adulterant was consistent with a brominated analogue of the pharmaceutical sedative zopiclone.
To support the putative identification from the mass spectra studies, the compound was isolated and structure elucidated using 1 H NMR. In Figure 5 we show the one-dimensional proton NMR of the unknown. A lack of resolution in the aliphatic region of the 1 H NMR spectrum made assignment of the piperazinyl protons difficult and overall the intensity of the signals was lower than expected. These observations suggest that the fraction may have contained degradation products and impurities that were likely created over the course of the isolation experiment. Despite the limitations in the proton NMR spectra obtained for the unknown, the correlations between the 8.89 and 8.85 ppm signals and the 8.52, 8.48 and 7.95 ppm signals in 1 H-1 H COSY spectrum further confirmed the assignments. As depicted in Table 5, these assignments compared very well with signals for the published 1 H NMR spectra of zopiclone [21]. Similar to the assignments proposed by Ming et al., (2007) the signals in this range, which were integrated for one proton each, were assigned to protons in the pyridyl and pyrazinyl rings. The assignment of the signals to the pyridyl protons were facilitated through the observed multiplet structure and coupling constants as depicted in Table 5.
The assignment of the unknown adulterant as a brominated analogue of zopiclone was further supported by the positive results from the commercial zopiclone/ eszopiclone ELISA kit, indicating that the adulterant is likely biologically active. This finding is particularly concerning as the natural health product tested appears to have been adulterated with an undeclared compound having chemical similarities to a known prescription  Table 5. Comparison of NMR data for the aromatic region reported for zopiclone and the analysis of the isolated fraction. Zopiclone chemical shifts were reported in Ming et al., 2007 [21]. Assignment numbers refer to the numbered atoms in Figure 3. pharmaceutical. It has been noted in the literature that synthesis of a chemically modified analogue of parent drug compounds, can be an effective means to avoid detection [15,16]. Our study demonstrates that using a routine conventional drug screening method which targets specific pharmaceutical drugs, is not effective alone in identifying adulteration. The detection of the undeclared brominated analogue within this product required analysis of the mass spectrum followed by careful evaluation of its mass ion profile. Such a detailed analysis is not typically performed on a routine basis as only the targeted compounds are actively sought for and analysts would seldom have the time to check every single peak in a product that contains multiple herbal ingredients. Furthermore, because ELISA testing of the U-Dream product itself did not produce a positive result, the presence of herbal extracts in the product may have effectively masked the detection of the adulterant with ELISA test kits.
Zopiclone itself is an active pharmaceutical agent that while chemically unrelated to benzodiazepines, binds with high affinity to benzodiazepine receptors [22]. Cautionary statements to patients concerning zopiclone use include hepatic and renal impairment, history of drug use or psychiatric illness and contraindications include myasthenia gravis, respiratory failure, severe sleep apnoea syndrome, severe hepatic impairment, pregnancy and breast-feeding [23]. The most frequently reported adverse event, as cited in the 2006 WHO 34 th Expert Committee on Drug Dependence, are bitter or metallic taste and dry mouth [23,24]. Zopiclone, like many prescriptive sleeping medications, has the potential for abuse and addiction [25]. This risk is higher in patients with a prior history of drug/alcohol abuse or a history of psychiatric illness [26]. In addition, a small number of patients in many clinical trials of zopiclone developed rebound insomnia after discontinuing the drug [25].
Given this compound's structural similarity to zopiclone and the testimonial reports describing adverse events consistent with those reported for zopiclone [18,19,23] it is likely the undeclared analogue has pharmacological activity. However this brominated analogue has not been subjected to pharmacological, toxicological or clinical investigations. "Designer drugs", often referring to the manipulation of the chemical structure of known drugs such that resulting product is structurally similar but not identical, has been a growing concern worldwide [27][28][29]. Structural analogy with a known pharmaceutical is not alone sufficient to predict function or safety [27]. Although zopiclone is a prescription drug [30,31], rather than an illegal psychoactive drug, as are often the targets of designer drugs, the suspected bromine analogue has no known safety profile and as such poses a substantial public health risk.

Conclusions
The NHP regulations provide a framework for high quality, safe and efficacious products to access the market in Canada. It is the responsibility of the manufacturer to assure traceability and transparency in their supply chain and establish verifiable compliance with GMP. The NHP U-Dream was suspected of containing an undeclared pharmaceutical based on testimonials of consumers. The experimental mass spectral data indicated the presence of an unknown bromine containing compound in the U-Dream product tested. Based on the fragmentation pattern the unknown was putatively identified as an analogue of zopiclone, whereby the chlorine atom had been substituted with bromine. This assignment was further corroborated by the NMR spectra and positive result from the commercial zopiclone/eszopiclone ELISA kit. The structural characteristics and consumer reviews of the product suggest that this undeclared, unknown compound has pharmacological activity, however to what extent is unknown. Of more considerable concern is the lack of any known safety profile by which to assess consumer risk. This study illustrates the importance of careful evaluation of analytical data in order to detect, not only undeclared adulterants, but unknown chemical entities and highlights the need for active monitoring and surveillance of potentially high-risk products post market entry.

List of Abbreviations
NHPs: natural health products NNHPD: natural and non-prescription health products directorate NPN: natural health product number HPLC: high performance liquid chromatography Q-ToF: quadrupole Time-of-Flight MS/MS: tandem mass spectrometry H 2 O: water MeCN: acetonitrile DAD: diode array detector ESI: electro-spray ionization NMR: nuclear magnetic resonance COSY: correlation spectroscopy CDCl 3 : deuterated chloroform ELISA: enzyme linked immunosorbent assay HR-LCMS: high resolution liquid chromatography mass spectrometry GMP: good manufacturing practice https://doi.org/10.33211/jnhpr.8

Conflicts of Interest
The authors declare that they have no conflicts of interest.

Authors' Contributions
PNB: made contributions to the design of the study and substantial contributions to the data analysis and interpretation, drafted the manuscript including critically important intellectual content, gave final approval of the version to be published and as corresponding author agrees to be accountable for all aspects of the work and ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. https://orcid.org/0000-0003-4886-7001 MC: contributed to study design and planning, assisted with the collection and analysis of data, and manuscript preparation. Significantly contributed to the data interpretation, and gave final approval of the version to be published. https://orcid.org/0000-0003-3180-3263 CC: made substantial contributions to the design of the study, the collection of data as well as interpretation and analysis of the data, revised the manuscript, and gave final approval of the version to be published.
SK: made substantial contributions to the collection of data, drafting of protocols, and gave final approval for the version to be published.
YSR: made substantial contributions to the design, acquisition, analysis, and interpretation of data, drafting of protocol, and gave final approval for the version to be published.
JN-K: made contributions to the collection of data as well as interpretation and analysis of the data, revised the manuscript including critically important intellectual content,, and gave final approval of the version to be published.
MLH: made contributions to study design, the interpretation of the data, critically revised the manuscript, made important intellectual contributions, and gave final approval of the version to be published.