An efficient HILIC-MS/MS method for the trace level determination of three potential genotoxic impurities in aripiprazole active drug substance

A sensitive and selective hydrophilic interaction liquid chromatography-tandem mass spectrometry (HILIC-MS/MS) method was developed and validated for trace analysis of potential genotoxic impurities (PGIs): 2,3-dichloroaniline (PGI-1), bis(2-chloroethyl) amine (PGI-2), and 2-chloroethylamine (PGI-3), in aripiprazole (APZ) active drug substance. Separation of analytes was achieved on ACE HILIC–N Column (HILN-5-1046U, 100 × 4.6 mm, 5 μm) in gradient elution mode with mobile phase A [acetonitrile:ammonium formate buffer (95:5 v/v)] and mobile phase B [acetonitrile:ammonium formate buffer (50:50 v/v)] at a flow rate of 0.8 mL/min. Developed method was linear in the concentration range of 8–100 ppm for PGI-1, 11–100 ppm for PGI-2, and 12.5–100ppm for PGI-3 with R2 > 0.996. The developed method was accurate for quantification of each PGI with percent recoveries greater than 96% and RSD (%) not more than 5%. The developed method was precise for quantification of PGIs in aripiprazole with RSD (%) of not more than 4% for any of the PGIs. There was no interference of diluent peaks at the retention time of the PGIs and APZ in the method. All the PGIs and sample solutions were found to be stable at ambient laboratory temperature (25 ± 5 °C) and refrigerated condition (2–8 °C) for a period of 48 h. The developed HILIC-MS/MS method can be used for trace quantification of PGIs in aripiprazole drug in quality control laboratories of the pharmaceutical industry.


Introduction
The assessment and control of impurities in pharmaceutical products, especially genotoxic/mutagenic impurities have received considerable attention in the recent years. International Council on Harmonization (ICH) provides guidance on assessment of genotoxic impurities in pharmaceuticals (International Council for Harmonization, n.d.-a; United states food and drug administration, n.d.). Active pharmaceutical ingredients typically contain a range of residual trace-level impurities arising from the reagents, starting materials, process intermediates and by-products generated during the course of synthesis (International Council for Harmonization, n.d.-b). The ICH M7 (International Council for Harmonization, n.d.-a) guideline proposed a threshold of toxicological concern value: 1.5 μg/day. The limit for the determination of these impurities was set based on the maximum daily dose (MDD) of the active pharmaceutical ingredient (API).
Aripiprazole synthesis involves the utilization of 2, 3dichloro aniline (PGI-1), bis (2-chloroethyl) amine (PGI-2) and 2-chloro ethylamine (PGI-3). Therefore, it is quite possible that these intermediates used in the synthesis of the API can be present as impurities in the final product. Based on the structural alerts (Muller et al., 2006;Snodin, 2010), all the three impurities could be genotoxic in nature (potential genotoxic impurities, PGIs) and should be controlled at low levels according the regulatory guidelines (Robinson, 2010;McGovern & Jacobson-Kram, 2006). Structures of these three PGIs and aripiprazole are presented in Fig. 1. Based on the maximum daily dose (MDD) of aripiprazole (30 mg/day), limit of three PGIs should be controlled below 50 ppm.
Methodologies have been developed for the determination of APZ via LC-MS/MS method in rat plasma (Liang et al., 2012) and in human plasma (Lin et al., 2009;Patel et al., 2013). There are also reports available for the simultaneous determination of APZ and its major metabolites by LC-MS/MS in human serum and plasma (Song et al., 2009;Caloro et al., 2012;Ambavaram et al., 2013;Kubo et al., 2005;Djordjevic Filijovic et al., 2014). But there are no published reports on the quantification of the potential genotoxic impurities in APZ using LC-MS/MS. Hence, the present investigation is devoted to develop and validate a simple, sensitive, accurate, and precise LC-MS/MS method for the trace level quantification of PGI-1, PGI-2, and PGI-3 in APZ.

Safety concerns
All the three impurities, PGI-1, PGI-2, and PGI-3, are considered to be potentially genotoxic. Therefore, all samples were handled under fume hood. During liquid chromatographic separation runs, column flow was collected separately in hazardous waste containers and disposed using appropriate institutional safety procedures.

Instrumentation
The liquid chromatography mass spectrometric (LC-MS) system consisted of Shimadzu (Nexera X2) chromatographic separation unit (Shimadzu Corp, Kyoto, Japan) coupled with Shimadzu 8040 triple quadrupole mass analyzer (Shimadzu Corp, Kyoto, Japan). Data acquisition and integration were carried out using LC-MS Lab Solutions software (Shimadzu Corp, Kyoto, Japan). Calibrated automated micropipettes were used for preparation of all the samples during the analyses. Aqueous phase component in mobile phase was filtered through a 0.22-μm polyvinylidene difluoride (PVDF) filtration membrane while the sample solutions were filtered through 0.22 μm PVDF syringe filters (Millipore® MA, USA) before the analysis. Liquid chromatographic and mass spectrometric conditions LC method conditions Chromatographic separation was achieved on ACE HILI C-N Column (HILN-5-1046U, 100× 4.6 mm, 5 μm) in HILIC mode. Mobile phase was composed of acetonitrile: ammonium formate buffer (pH 6) (95:5 v/v) (solvent A) and acetonitrile:ammonium formate buffer (pH 6) (50:50 v/v) (solvent B). Gradient program was employed to achieve the separation between the analytes and APZ. Samples were analysed under gradient conditions at a flow rate of 0.8 mL/min with total run time of 10 min. Injection volume was 10 μL in the method. Temperature for auto sampler was set at 25 ± 5°C and column oven was set at 35°C. Complete details of the LC method conditions are presented in Table 1.

MS method conditions
For the current method, mass spectrometer was operated with electrospray ionization (

Preparation of stock solutions and sample solutions
The mass concentrations of PGI-1, PGI-2, and PGI-3 in acetonitrile amounted to 0.1 mg/mL. All the individual primary stock solutions were stored under refrigerated condition (2 to 8°C). System suitability for the developed method was established by injecting six different replicates of primary standard solutions (100 ng/mL) of each of the PGIs and determining the relative standard deviation [RSD (%)] of the obtained responses. For determining limit of detection (LOD) and limit of quantification (LOQ) values of the PGIs in the developed method, a series of standard solutions for each PGI were prepared in the concentration range of 2-30 ng/mL (≈ 1-15 ppm with respect 2 mg/mL API concentration) from their respective primary standard solutions (100 ng/mL).
Precision of the developed method was determined by performing repeatability (intra-day precision) and intermediate precision studies using the actual API samples. In the repeatability studies, six replicate samples containing 2 mg/mL of APZ were injected. Intermediate precision studies were performed by different analysts on different days using the same concentration of the API samples.
Accuracy studies of the method were performed using un-spiked and spiked samples of APIs. Triplicate samples of APZ at a 2 mg/mL concentration were prepared separately using acetonitrile as the diluent. Spiked samples were prepared by adding known volume of standard solutions of the PGIs at three different levels (LOQ, 100% and 150%) to API samples (2 mg/mL). Spiked samples were also prepared in triplicates.
For determining the robustness of the developed method, primary standard PGIs solutions of 100 ng/mL concentrations were used. The spiked samples of APZ of concentration: 2 mg/mL with three PGIs at 100% concentration levels were used in these robustness investigations. These studies were performed in triplicates.

Method validation
The developed method for the trace level quantification of PGI-1, PGI-2, and PGI-3 in APZ were systematically validated as per the analytical method validation guidance from regulatory agencies (Djordjevic Filijovic et al., 2014), International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use, n.d.]. The method was validated for selectivity, sensitivity, linearity, precision, accuracy, and robustness. Stability studies of all the PGIs were also performed to assess their stability during short-term sample processing and long-term storage.

Specificity
Specificity was determined by injecting the diluent blank, process impurities, and solvents. The interference should be less than the limit of detection (LOD).

Sensitivity
Limit of detection (LOD) and limit of quantification (LOQ) of each PGIs were determined by injecting different standard solutions of PGI-1, PGI-2, and PGI-3 as described in the "Preparation of stock solutions and sample solutions" section, to obtain a signal-to-noise (S/ N) ratio greater than or equal to 3:1 and 10:1, respectively.

Linearity
Calibration curves were constructed for each PGI by plotting the response against the respective concentration of PGIs. Unweighted regression analysis was performed to fit the calibration curves of each PGI and to determine the regression equations. Statistical parameters like slope, intercept, and regression coefficient (R 2 ) were determined to analyze the goodness of fit and validity of regression.

Precision
In repeatability as well as in the intermediate precision studies, RSD (%) values of the responses of each PGI (at 100% level) were obtained by injecting six replicate samples containing 2 mg/mL of API. RSD (%) values of PGI responses from the six replicates should be less than 10% to confirm the repeatability of the method.
To establish intermediate precision of the developed method, RSD (%) values of each PGI were determined by varying different analysts on different days. RSD (%) values should be less than 20% to indicate that the developed method was precise.

Accuracy
Accuracy of the method was determined in terms of percentage recovery and RSD (%) values for each of the PGI by comparing the responses obtained from un-spiked and spiked samples at three different levels ranging from LOQ, 100% and 150% of the specification limit of 50 ppm. Percentage recovery values should be within 100 ± 20% while the RSD (%) values should be less than 10% to establish that the method was accurate.

Robustness
Robustness of the developed method was determined by deliberately changing mobile phase flow rate (± 20% to the set value of 0.8 mL/min) and cone voltage (± 20% to the set value of 15 V). System suitability and mean percentage recovery of the spiked samples were determined by changing the method parameters as described above. RSD (%) values obtained from system suitability should be ≤ 10% and mean percentage recovery should be within 100 ± 20% for each PGI.

Solution Stability
Stability studies were performed using secondary intermediate stock solution (100 ng/mL) of each PGI and spiked samples of APZ (2 mg/mL) with PGIs at 100% concentration levels up to 48 h at ambient laboratory temperature (25 ± 5°C) and refrigerated condition (2-8°C). The percent recoveries of primary standard solutions of each PGI and spiked samples subjected to stability studies were calculated by comparing against the freshly prepared primary standard solutions (100 ng/mL) of each PGI.

Optimization of chromatographic and mass spectrometric method conditions
Considering the small molecular size and polar nature of the analytes under investigation, hydrophilic interaction liquid chromatography (HILIC) was applied. In this study, different stationary phases such as Inertsil HILIC, Luna HILIC, Zic HILIC, and ACE HILIC-N columns with varied polarity were evaluated with LC-MS compatible aqueous mobile phases such as ammonium formate and ammonium acetate buffers (in variable ratios) to optimize the separation of PGIs from the API and also to obtain better sensitivity in their MS response. Both isocratic and gradient elution modes were evaluated for the optimal resolution between analytes and API. Better peak shapes and resolution of PGIs and APZ were obtained with ACE HILIC-N Column (100 mm × 4.6 mm × 5 μm) in gradient elution mode using mobile phase A consisting of acetonitrile:ammonium formate buffer (95: 5 v/v) and mobile phase B comprising of acetonitrile:ammonium formate buffer (50:50 v/v).
Mass spectrometer (MS) conditions were optimized in positive ionization mode. For the quantification of the analytes, multiple reaction monitoring (MRM) mode was employed. Various MS parameters such as heating block temperature, desolvation line (DL) temperature, and drying gas flow were studied to achieve optimum response for all the PGIs in MRM mode. Fragmentation patterns for PGI-1, PGI-2, and PGI-3 are presented in Fig. 2. Optimized chromatographic and mass spectrometric method conditions for qualification and quantification of PGI-1, PGI-2, and PGI-3 are presented in Table 1.

Method validation Specificity
The total ion current (TIC) chromatograms showing the separation between PGIs and APZ are presented in Fig. 3. Typical MS chromatograms of blank, standard PGI (at 50 ppm), and spiked sample of APZ with PGIs at their corresponding m/z transition in MRM mode are presented in Fig. 4. As shown in Fig. 4a, no interfering peaks were seen at retention times of PGIs and APZ in the obtained MS chromatogram. The areas of diluent blank samples at the m/z transitions of each PGI were well within the acceptance limits when compared with the areas of PGIs at their LOQ levels.

Linearity
The calibration curves for PGI-1 (range 8-100 ppm), PGI-2 (range 11-100 ppm), and PGI-3 (range 12.5-100 ppm) were constructed with peak area on vertical axis and concentration of analyte on horizontal axis. From linear regression analysis, slope, intercept, and R 2 values were determined and results are summarized in Table 2. In the linear regression analysis for each of the PGIs, the R 2 values are above 0.99 and acceptable RSD (%) over the concentration ranges confirm that the developed method for the three PGIs is linear. Overlay of regression study linearity plot of PGI-1, PGI-2, and PGI-3 are depicted in Fig. 6.

Precision
The results obtained from precision studies are given in Table 2. Cumulative RSD (%) values for the results obtained from analyst variation (analyst 1 and analyst 2) were within the acceptance criteria of ≤ 10%. RSD (%) for each PGI content from six preparations in the precision and intermediate precision experiments were within the acceptance criteria of ≤ 10.0% indicating that the developed method is precise for the trace level quantification of PGI-1, PGI-2, and PGI-3.

Accuracy
The results obtained from accuracy studies are summarized in Table 2. Percentage recovery values of each of the PGIs, at all the three levels, were found to be more than 95% with RSD (%) less than 4%. These results indicate that the developed method was accurate as the percentage recovery and %RSD values were found to be well within the acceptable limits of 100 ± 20% and ≤ 10%, respectively.

Robustness
Robustness of the method was evaluated for mobile phase flow rate (0.7 and 1.0 mL/min) and cone voltage (12 V and 18 V) and the results obtained are presented in Table 2. RSD (%) from six replicate injections of system suitability solution of all the PGIs for the change in mobile phase flow rate and cone voltage were less than 5.0%. Percentage recovery values of triplicate spiked sample injections of all the PGIs for the change in mobile phase flow rate and cone voltage were more than 93%. The results obtained for robustness studies were found to be well within the acceptance criteria of not more than 10% (RSD) and within 100 ± 20% for percentage recovery.

Solution stability
Short-term solution stability for primary standard solution (100 ng/mL) of each PGI and spiked samples of APZ (2 mg/mL) with PGIs at 100% concentration levels were studied up to 48 h at ambient laboratory temperature (25 ± 5°C) and refrigerated condition (2-8°C). The percent recoveries of each PGI in primary standard solution and spiked samples were calculated by comparing against the freshly prepared primary standard solutions of respective PGI standard. No significant difference (< 10% to the initial result) was observed in the stability of the samples when stored at ambient laboratory temperature (25 ± 5°C) and refrigerated condition (2-8°C) for a period of 48 h.

Application
The methodology developed was applied to the aripiprazole drug product available in the market. The sample solutions were prepared by ensuring the final concentration should remain 2.0 mg/mL in line with aripiprazole drug. The results show the absence of any of the PGI impurities in the said drug formulations. Hence, known amounts of PGI impurities were added to the drug solutions and recovery studies were made. The results are  presented in Table 3. The % of recovery falls between 96.4 and 98.2%. Hence, the method developed can be successfully applied to assay the level of 3 PGI impurities in drug formulations.

Conclusions
LC-MS/MS method was developed and validated for the trace level quantification of 2,3-dichloroaniline (PGI-1), bis(2-chloroethyl) amine (PGI-2), and 2-chloroethylamine (PGI-3) in aripiprazole which are potential genotoxic impurities. Resolution was achieved by employing the HILIC chromatographic method. Validation results substantiate that the developed method was selective, sensitive, accurate, and precise for the quantification of the PGIs in APZ. Standard solutions of all the PGIs were found to be stable for 48 h at ambient and refrigerated conditions. Hence, the developed method can be used by pharmaceutical industries for trace quantification of the PGIs in aripiprazole API and formulations.