Quantitative Study of Impurities in Bedaquiline Fumarate: Identification and Characterization of Its Three Degradation Products Using HPLC, LC/ESI-MS, and NMR Analyses

The study aimed to establish a high-performance liquid chromatography (HPLC) method for the quantitative analysis of the related substances of bedaquiline fumarate. Nuclear magnetic resonance and mass spectrometry were used for characterization and assay. A chromatographic method was used for separation. The conditions used were: gradient elution system composed of methanol 0.01mol/L KH2PO4 and 0.01 mol/L K2HPO4 (pH = 4.1) with a flow rate of 1 mL/min, at 224 nm as the detection wavelength. In this study, three degradation products of bedaquiline fumarate have been disclosed for the first time. The related impurities and degradation products of the drug were well separated. The method provided linear responses within the concentration range, which varied from 0.20 to 10.08 μg/mL with limits of detection of 0.10 μg/mL and limits of quantification of 0.20 μg/mL. The mean percent recovery varied between 91.64 and 105.89%. The method was validated for other parameters such as specificity, stability, and robustness. This method was validated and worked well for the impurity studies and quality control analysis of the laboratory-prepared samples of bedaquiline fumarate.


Introduction
Bedaquiline fumarate (BDQ), chemically named (1R, 2S)-1-(6-bromo-2-methoxy-3-quinolinyl)-4-(dimethylamino)-2-(1-naphthalenyl)-1-phenyl-2-butanol fumarate, is a diarylquinoline antimycobacterial drug used for the treatment of multi-drug-resistant tuberculosis in adults. 1-3 BDQ was developed by Janssen Therapeutics, approved by the Food and Drug Administration in 2012, and has been marketed in China since 2016. 4 BDQ functions on ATP synthase of Mycobacterium tuberculosis, and inhibits ATP production by binding to the F 0 region of ATP synthetase, resulting in the death of Mycobacterium tuberculosis. [5][6][7] The current synthetic route of BDQ is shown in ►Fig. 1. 8 However, analysis methods of BDQ have not been included in the pharmacopeias of any countries, and reported only in a few literatures, such as separation and characterization of its degradation impurities, 9,10 analysis of its related substances by high-performance liquid chromatography (HPLC), 11 separation of its diastereomers by reversed-phase-HPLC, 12 and determination of BDQ in human serum by liquid chromatography-tandem mass spectrometry (LC-MS/MS). 13 Therefore, studies of BDQ quality control, especially the study of BDQrelated substances, could help meet the production goals, and are also important and necessary for further development of the drug.

Abstract
The study aimed to establish a high-performance liquid chromatography (HPLC) method for the quantitative analysis of the related substances of bedaquiline fumarate. Nuclear magnetic resonance and mass spectrometry were used for characterization and assay. A chromatographic method was used for separation. The conditions used were: gradient elution system composed of methanol 0.01mol/L KH 2 PO 4 and 0.01 mol/L K 2 HPO 4 (pH ¼ 4.1) with a flow rate of 1 mL/min, at 224 nm as the detection wavelength. In this study, three degradation products of bedaquiline fumarate have been disclosed for the first time. The related impurities and degradation products of the drug were well separated. The method provided linear responses within the concentration range, which varied from 0.20 to 10.08 μg/mL with limits of detection of 0.10 μg/mL and limits of quantification of 0.20 μg/mL. The mean percent recovery varied between 91.64 and 105.89%. The method was validated for other parameters such as specificity, stability, and robustness. This method was validated and worked well for the impurity studies and quality control analysis of the laboratory-prepared samples of bedaquiline fumarate. Based on the reported synthesis route 8 and the stress forcing experiments, the impurity profiles of BDQ in bulk drugs as well as 11 potential impurities were analyzed (►Fig. 2). These included the residuals of starting materials (A1, A2), by-products (A3, A4, A6), diastereomer impurities (A7, A8), and isomer impurities (A5), as well as the degradation products (A9, A10, A11). Among them, A9, A10, and A11 were characterized and disclosed for the first time. Furthermore, the present research also developed a reversed-phase liquid chromatography (RP-LC) method to simultaneously determine BDQ and its related substances. The method was validated and applied for quality studies of BDQ drug samples.

Synthesis of Compound A3
Intermediate 3 (5.0 g) was dissolved in tetrahydrofuran (THF) (50 mL), then Pd/C (5%, 6.0 g) was added under N 2 . The mixture was heated to 60°C and hydrogen was passed into the mixture for 8 hours. After filtration and concentration, the crude product was chromatographed on silica gel eluting with methylene dichloride (DCM):methanol (MeOH) ¼ 20:1 to give A3 (2.6 g, 60.7%) as a white solid.

Synthesis of Compound A4
n-BuLi (in n-hexane, 20 mL, 2 mol/L) was diluted with THF (12 mL), and cooled to À76°C under N 2 . A solution of A1 (10.0 g) in THF (24 mL) was added. The mixture was stirred for 1 hour, followed by addition of a solution of A2 (8.0 g) in THF (12 mL). The reaction was performed at À70°C to À75°C for 5 hours, and quenched with acetic acid. The organic phase was separated and concentrated. The crude product was chromatographed on silica gel (DCM:MeOH ¼ 20:1) to give A4 (1.3 g, 8.9%) as a white solid.

Synthesis of Compounds A7 and A8
Lithium diisopropylamide (in THF, 20 mL, 2 mol/L) was diluted with THF (12 mL), and cooled to À70°C. A solution of A1 (10.0 g) in THF (24 mL) was added. The mixture was stirred for 1 hour, followed by addition of a solution of Z1 (8.0 g) in THF (12 mL). The reaction was stirred for 2 hours and quenched with acetic acid. The organic phase was separated and concentrated to obtain a residue, which was purified by column chromatography (DCM:MeOH ¼ 20:1) to give A7 (1.1 g, 10.1%) and A8 (1.8 g, 16.2%).

Synthesis of Compound A9
A2 (5.0 g), dissolved in water, and NaOH aqueous solution (20 mL, 2 mol/L) was stirred at room temperature for 2 hours. The pH of the aqueous solution was adjusted to pH 7 with concentrated hydrochloric acid, followed by extraction with ethyl acetate. The organic phase was separated and concentrated. The crude product was purified by column chromatography (DCM:MeOH ¼ 20:1) to obtain A9 (150 mg, 3.7%).

Structure Elucidation
Nuclear magnetic resonance (NMR) and mass spectrometry (MS) spectra data were used for structure analysis. NMR spectra were obtained using a Bruker 400 MHz spectrometer (AC, AVANCE, Bruker Daltonics GmbH, Fäl-landen, Switzerland) with chloroform-d (CDCl 3 ) as a solvent, and tetramethylsilane (TMS) as an internal standard. Qualitative analysis was performed by LC-MS using an Agilent 6420 Triple Quad LC-MSD equipped with an ESI (electrospray ionization) source, which was used in positive ionization mode in full scan and product scan. The spray voltage was set at 4 kV, and the capillary temperature was maintained at 350°C. Nitrogen was used for nebulization. During full-scan MS analysis, the spectra were recorded in the range of m/z 100-1,100. The fragmentor was set at 135 eV. Analytical data were acquired using the Masshunter software.

Quantitative Analysis of BDQ and Its Impurities
Analyses were conducted using Waters series system equipped with a quaternary pump, a vacuum degasser, an autosampler, a column heater-cooler, and a DAD detector. The data were acquired and processed using the

Sample Preparation
The diluent is a solution of acetonitrile, water, and trifluoroacetic acid, with a volume ratio of 50:50:0.1. For the related substances' assay, the test concentrations of BDQ were 0.5 mg/mL. The system suitability solution was prepared by dissolving BDQ and 11 impurities in the diluent to reach the concentration 0.5 mg/mL for BDQ and 0.5 μg/mL for each impurity.
Forced degradation solutions were analyzed under different stress conditions as per ICH (International Council of Harmonization) requirements. For oxidation degradation, BDQ (1.0 mg/mL) was dissolved in acetonitrile:water (v: v ¼ 1:1), then 10% H 2 O 2 (aq) was added in the ratio of 5:1 (v: v). The mixture was kept at 60°C for 48 hours. For basic degradation, BDQ (1.0 mg/mL) was dissolved in acetonitrile: water (v:v ¼ 1:1), then 0.1 mol/L NaOH was added in the ratio of 5:1 (v:v). The mixture was kept at room temperature for 1 hour, and then neutralized with 0.1 mol/L HCl. For acidic degradation, BDQ (1.0 mg/mL) was dissolved in acetonitrile: water (v:v ¼ 1:1), then 1 mol/L HCl was added in the ratio of 5:1 (v:v). The mixture was stored at 60°C for 12 hours, and then neutralized with 1 mol/L NaOH. For thermal degradation, a solution of 1.0 mg/mL BDQ in acetonitrile was kept at 80°C for 48 hours. For photolytic degradation, a solution of 1.0 mg/mL BDQ in acetonitrile was treated with the illumination of 4,500 lux AE 500 lux and kept at room temperature for 7 days.

Method Validation for Quantitative Analysis
The developed stability-indicating LC method was validated as per ICH guidelines with respect to limits of detection (LODs) and limits of quantification (LOQs), linearity and range, stability, precision, accuracy, and robustness. The linearity test was performed at concentrations ranging from LOQ to 10.0 μg/mL. The precision was evaluated as repeatability and intermediate precision. The accuracy was evaluated at three concentration levels (50, 100, and 150% of the impurities' limit concentration). The percentage recoveries were calculated from the calibration curve. The robustness was evaluated by deliberately altering chromatographic conditions, including batches of C18 column, flow rate, column temperature, wavelength, and mobile phase. The resolution between BDQ and its impurities was assessed.

Optimization of Chromatographic Conditions for Quantitative Analysis
A strong chromogenic group has been found in the structure of BDQ and the related substances, which is suitable for a UV detector. The detection wavelength of 224 nm was selected as suitable by taking into account the absorption intensity. Since pK a values of BDQ and the related substances are different, and most of them are weakly alkaline and sensitive to pH changes, thus, an acidic mobile phase containing salt was selected to provide a buffer environment. Besides, time gradient with mobile phase, pH, and organic modifiers have been studied to acquire a better resolution and peak shape among the impurities. Based on these results, a HPLC method was established (►Fig. 4).

Results of the Forced Degradation
Chromatograms of the stress studies are indicated in ►Fig. 5. Our data showed that BDQ was stable under stress conditions such as photolytic and thermal conditions. Under acidic conditions, A10 was detected; under basic conditions, A9 was detected; and under oxidation conditions, A11 was detected. DAD analysis suggested a good peak purity of BDQ among all the samples. The samples had a mass balance between 90 and 110% of BDQ. The assay result of BDQ was unaffected by the presence of the related compounds, confirming its stability during the LC method.

Results of Method Validation
The determination method of the impurities of BDQ was validated systematically and separately. The results of linearity, LOD, LOQ, precision, stability, and accuracy are summarized in ►Table 2.
Our data suggested that LODs of BDQ and A1-A11 were approximately 0.1 µg/mL, and their LOQs were approximately 0.2 µg/mL. The test for linearity of BDQ and its impurities was completed by injecting a series of solutions with concentrations of 0.04, 0.08, 0.1, 0.4, 0.8, and 2% of the sample, and the peak Pharmaceutical Fronts Vol. 5 No. 1/2023 © 2023. The Author(s). area was recorded. As shown in ►Table 2, strong linear relationships were obtained with r values ranging between 0.9996 and 0.9999. The precision study (%RSD) values of A1-A11 were all below 5.0%, confirming the high level of precision of the developed LC method. Besides, the stability result showed a 72-hour stability of the system suitability solution at room temperature. The accuracy of the method was evaluated in triplicate at three different concentration levels. As shown in ►Table 3, mean recoveries of all the impurities ranged from 90.0 to 110.0%. Through deliberately altering chromatographic conditions (e.g., batches of C18 column, detection wavelength, flow rate, pH, and column temperature), resolution between impurities and BDQ was always greater than 1.5, illustrating the robustness of the method.  Abbreviations: BDQ, bedaquiline fumarate; LOD, limit of detection; LOQ, limit of quantification.

Application
The detection of related substances was performed on different batches of intermediates and the active pharmaceutical ingredient (API). Our data showed that the detected amount of impurity A3 in BDQ sample is 0.16% (►Table 4), lower than that of impurity A3 in reference listed drug. Thus, the quality of this batch of API is qualified.

Discussion
In this study, the degradation mechanism of BDQ was postulated based on structural elucidation of degradation impurities and liquid-phase result of degradation solution. The hydrogen on the benzyl group of BDQ is active, and it is easily captured by free radicals under alkaline conditions, resulting in autooxidation and then dehydration to form A1 and A2 impurities. On the tertiary amine group of A2, there are lone pair electrons that capture reactive oxygen to form nitrogen oxides, followed by cope elimination reaction, and A9 was obtained. The acid degradation mechanism of BDQ mainly depends on "acid-catalyzed autoxidation." BDQ possesses lone pair electrons on the nitrogen atoms of tertiary amines, which capture reactive oxygen to form nitrogenoxygen complexes under acidic conditions, and then dealkylation reaction to produce A10 under the catalysis of an acid. In oxidative degradation experiments, lone pair electrons on nitrogen atoms of tertiary amines in BDQ capture reactive oxygen to form nitrogen-oxygen complexes A11.
The generation mechanism of BDQ forced degradation products is shown in ►Fig. 6.

Conclusion
In this article, five degradation impurities of BDQ were determined by forced degradation test. The generation mechanism of degradation impurities was also analyzed. The structures of three BDQ forced degradation impurities were characterized by NMR and MS analyses. At the same time, a reversed-phase HPLC method was established, which could simultaneously control 11 impurities in the preparation process of BDQ bulk  drug. Particularly, several pairs of diastereomer and isomer impurities with similar structures were even effectively separated by optimizing the pH of mobile phase and improving other chromatographic conditions, such as gradient. The method was validated and proved to be suitable for quality monitoring of laboratory-prepared samples and bulk samples during batch release and stability studies.

Conflicts of Interest
There are no conflicts to declare. Quantitative Study of Impurities in Bedaquiline Fumarate Zhang et al. e54