- Email: [email protected]
Development and validation of a liquid chromatographic method for the analysis of squaric acid dibutyl ester and its impurities
Accepted Manuscript Title: Development and validation of a liquid chromatographic method for the analysis of squaric acid dibutyl ester and its impurities Authors: Marwa F. Mansour, Peixi Zhu, Ann Van Schepdael, Erwin Adams PII: DOI: Reference:
S0731-7085(17)30451-X http://dx.doi.org/doi:10.1016/j.jpba.2017.04.022 PBA 11220
To appear in:
Journal of Pharmaceutical and Biomedical Analysis
Received date: Revised date: Accepted date:
22-2-2017 13-4-2017 15-4-2017
Please cite this article as: Marwa F.Mansour, Peixi Zhu, Ann Van Schepdael, Erwin Adams, Development and validation of a liquid chromatographic method for the analysis of squaric acid dibutyl ester and its impurities, Journal of Pharmaceutical and Biomedical Analysishttp://dx.doi.org/10.1016/j.jpba.2017.04.022 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Development and validation of a liquid chromatographic method for the analysis of squaric acid dibutyl ester and its impurities Marwa F. Mansour 1,2, Peixi Zhu1, Ann Van Schepdael1, Erwin Adams1 1
KU Leuven, University of Leuven, Department of Pharmaceutical and Pharmacological Sciences,
Pharmaceutical Analysis, Leuven, Belgium 2
National Organization for Drug Control and Research, Giza, Egypt
Correspondence: Professor Erwin Adams, Farmaceutische Analyse, KU Leuven, O&N2, PB 923, Herestraat 49, B3000 Leuven, Belgium E-mail: [email protected] Fax: +32-16-323448 Tel.: +32-16-323444
Highlights
First time that a quality control LC method is described for squaric acid dibutyl ester (SADBE). Method is simple, short, selective and robust. Characterization of SADBE and its forced degradation products by LC-MS.
1
Abstract A simple, fast and selective stability indicating liquid chromatographic method has been described for the simultaneous determination of squaric acid dibutyl ester and its impurities. The chromatographic separation was achieved on a C2 column (250 mm × 4.6 mm i.d., 5 μm) using a mobile phase consisting of 0.15 % phosphoric acid – acetonitrile – methanol (30:60:10, v/v/v). Isocratic elution was performed at a flow rate of 1.0 mL min-1. The analytes were detected by UV at 252 nm. The method was validated according to the ICH guidelines and satisfactory results were obtained. The specificity of the developed method was tested using forced degradation solutions of the drug substance. Characterization of squaric acid dibutyl ester and its forced degradation products was achieved by coupling mass spectrometry (MS) to the liquid chromatographic (LC) system. The method was successfully applied for quality control purposes including assay and determination of related compounds as required by regulatory guidelines to ensure its safety and efficacy since no monograph is available in official compendia. Keywords: Squaric acid dibutyl ester; ; ; ; , quality control, liquid chromatography, forced degradation, mass spectrometry.
2
1. Introduction Since destructive and surgical treatment of warts is painful, especially for children, topical immunotherapy with contact sensitizers is encouraged [1-2]. However, some of these sensitizing compounds such as 2,4-dinitrochlorobenzene are considered as mutagenic [3-4]. Therefore, effective alternative compounds are of considerable interest. Recent studies showed that warts in children caused by the human papilloma virus could be treated safely by topical application of squaric acid dibutyl ester (SADBE). In addition, SADBE is also successfully used as topical immunomodulator for the treatment of alopecia [5]. SADBE is chemically known as (3,4-dibutoxycyclobut-3-ene-1,2-dione). It is a four membered ring compound derived from squaric acid (SA). The latter is also easily formed by hydrolysis as shown in Figure 1. Therefore, SA is considered to be a likely impurity of SADBE since it is starting material and degradation product as well. It is worth noting that SA decreases the activity of SADBE, as it is considered to be inactive as contact sensitizer [6]. In vitro results on the human skin support the assumption that the ester can combine with a protein to form an antigen while SA can not [4]. The medicinal significance of SADBE has been acknowledged since FDA approved recently its inclusion in the list of drug substances used to compound human drug products. Although SADBE plays an important role in the treatment of several patients, no monograph of it is adopted in a pharmacopoeia or no proper analytical method can be found in literature to determine its impurities. However, such methods are crucial, as it is known that impurities reduce the effectiveness of a drug and often cause allergic reactions and side effects. Only one report related to this was found describing a GC-MS method to test the presence of carcinogenic contaminants. In this research SA, one of the hydrolysis products, was detected by spectrophotometry at 480 nm after its complexation with ferric chloride [6].
3
Therefore, it was important to develop an accurate, reliable analytical method for the analysis of SADBE. LC was considered the most appropriate technique for quality control purposes. Special attention was paid to SA since it was observed that the concentration of SA as impurity in SADBE influences the therapeutic activity in the treatment of warts and alopecia [4,6]. Whereas SADBE is rather apolar, SA is usually present in solution as a polar anion (pKa values of 0.5 and 3.6). Traditionally, normal phase and ion pair reversed phase LC are powerful tools for the retention of polar compounds. However, normal phase LC suffers from limitations such as solubility problems for the analytes while ion pair systems need a long equilibration time and are mostly not compatible with LC-MS [7]. Therefore, an alternative separation mode such as hydrophilic interaction liquid chromatography (HILIC) could be considered for the separation of polar compounds that suffer from poor solubility in the organic solvents used in normal phase LC. However, HILIC requires also a rather long equilibration time and it is quite costly due to the consumption of relatively high amounts of organic solvents. Thus, if a reversed phase method could avoid SA to be eluted in the void volume, it could be interesting in terms of time, cost and simplicity of application. The aim of this paper was to develop a stability indicating method for the determination of the main compound SADBE as well as its impurities by applying reversed phase LC. The drug substance was subjected to forced degradation studies in order to induce its degradation products. These solutions were further used to optimize the chromatographic conditions and to test the robustness of the developed method. 2. Experimental 2.1. Chemicals and reagents
4
SADBE was purchased from Sigma-Aldrich (St. Louis, MO, USA). SA, methanol (HPLC grade) and formic acid were supplied by Acros Organics (Morris Plains, New Jersey, USA). Phosphoric acid (85 % m/m) was obtained from Acros Organics (Geel, Belgium). Acetonitrile (HPLC gradient grade) and hydrochloric acid were obtained from Fisher Scientific (Leicester, United Kingdom). Acetonitrile and methanol MS grade were acquired from Biosolve LTD (Valkenswaard, The Netherlands). Ammonium acetate MS grade was purchased from Sigma-Aldrich (Zwijndrecht, The Netherlands). Trifluoroacetic acid was supplied by Carl Roth GmbH (Karlsruhe, Germany). Sodium hydroxide was obtained from VWR International (Leuven, Belgium) and hydrogen peroxide 30 % was acquired from Merck (Darmstadt, Germany). Nitrogen supplied by Air Liquide (Liège, Belgium) was used as sheath and auxiliary gas for MS. Helium gas was purchased from Air Products (Brussels, Belgium). Water was produced in-house using a Milli-Q Gradient water purification system (Millipore, Bedford, MA, USA). A pharmaceutical formulation containing 1 % SADBE was prepared in acetone as a delivery system [2,4]. 2.2. Forced degradation studies Forced degradation studies of the bulk drug (1 mg mL-1 solutions) were performed in accordance with Singh et al. [8]. SADBE was subjected to stress studies involving acidic, basic and neutral conditions at 40 °C. An oxidative study was carried out in 3 % (v/v) H2O2 at room temperature. Details have been outlined in Table 1. In all cases, the duration was such that sufficient degradation products were obtained. The resultant solutions were withdrawn at appropriate times, neutralized if applicable and diluted to obtain a concentration of 100 μg mL-1 using acetonitrile – water (95:5, v/v) as solvent. 2.3. LC instrumentation and chromatographic conditions
5
2.3.1. Non-volatile system for LC-UV The LC system from Merck-Hitachi (Tokyo, Japan) was equipped with a high pressure pump (L7100), an autosampler (L-7200), a UV/VIS (L-7400) or photodiode array (L-2450) detector and inline solvent degasser (L-7614). EZ Chrome Elite software version 3.1.6 (Merck-Hitachi) was used for data processing and acquisition. Optimum chromatographic separations were achieved on a Maxsil C2 column (250 mm × 4.6 mm i.d., 5 μm) (Phenomenex, Torrance, CA, USA) which was maintained at 25 °C. During method development, also a Polaris C18 (250 mm × 4.6 mm i.d., 5 µm) column (Agilent, Santa Clara, CA, USA) was examined. The mobile phase was pumped isocratically at a flow rate of 1.0 mL min-1. It consisted of 0.15 % phosphoric acid – acetonitrile – methanol (30:60:10, v/v/v). The injection volume was 20 μL and UV detection was performed at 252 nm. The mobile phase was degassed before use by sparging helium. Stock solutions of SADBE and SA (1 mg mL-1), working solutions (100 µg mL-1) and further dilutions hereof were prepared using acetonitrile – water (95:5, v/v) as solvent. 2.3.2. Volatile system for LC-MS An LCQ ion trap (IT) mass spectrometer from ThermoFinnigan (San Jose, CA, USA) equipped with an electrospray ionization source (ESI) was operated in the positive mode for the confirmation of the molecular mass of SADBE and in the negative mode for SA and the identification of an unknown impurity. Helium was used as the damping and collision gas. Instrument and data acquisition were performed with Xcalibur 1.2 software from ThermoFinnigan. The optimized settings for the mass spectrometer in positive mode were as follows: sheath gas flow rate: 85 arbitary units (arb); auxiliary gas flow rate: 25 arb; spray voltage: 4.0 kV; capillary temperature: 210 °C; capillary voltage: 7.0 V; tube lens offset: -30.0 V; multipole 1 offset: -2.75 V;
6
lens voltage: -24 V; multipole 2 offset: -6.00 V; RF amplitude: 400 V, peak-to-peak; scan ranges: m/z 50-800. The optimized settings for the mass spectrometer in negative mode were as follows: sheath gas flow rate: 85 arb; auxiliary gas flow rate: 25 arb; spray voltage: 4.0 kV; capillary temperature: 210 °C; capillary voltage: -5.0 V; tube lens offset: 10 V; multipole 1 offset: 3.00 V; lens voltage: 16 V; multipole 2 offset: 7.00 V; RF amplitude: 400 V, peak-to-peak; scan ranges: m/z 50-800. For LC-MSn experiments, the ions of interest were isolated in the IT with an isolation width of 2 u and the collision energy level (CEL) was set at 30 %. The separation was also realized using a Maxsil C2 column (250 mm × 4.6 mm i.d., 5 μm). As the non-volatile mobile phase described in 2.3.1 contained phosphoric acid, it could not be used for the LC-MS system. Therefore, it was replaced by 5 mM ammonium formate pH 3 and so the composition used for LC-MS was ammonium formate (pH 3; 5 mM) – acetonitrile – methanol in a ratio of 30:60:10 (v/v/v). It was pumped at a flow rate of 1.0 mL min-1. A T-piece was used to split one fifth of the mobile phase to the MS. The samples injected were the working solutions of SADBE and SA (100 µg mL-1) in acetonitrile – water (95:5, v/v). 2.4. Factorial analysis
A more in-depth study of the experimental parameters or factors will lead to more insight in the chromatographic behavior of the main compound and its impurities. This is helpful to adjust the conditions when the separation would be somewhat different, for example when the method is applied in another laboratory. The use of an experimental design is appropriate to explore the parameters in an organized way with a minimal number of experiments. It allows to quantify the importance of the parameters and to decide whether
7
they are significant or not. This way, analysts know which parameters are the most interesting to adapt or which can be considered robust. The set-up of the design, analysis of the results by means of multivariate analysis and construction of the graphical plots were performed using R-software (R foundation, Vienna, Austria). The tested experimental factors were column temperature and percentages of phosphoric acid, acetonitrile and methanol in the mobile phase. These factors might have an influence on the overall separation of the developed method. A two level fractional factorial design was applied with a number of runs equal to 2k-1 + n, where k is the number of factors and n is the number of center points [9,10]. Therefore, 11 experiments including 3 center points were performed. Several responses were selected: relative retention time for SA, relative retention time for impurity 2 (relative vs. the main peak), resolution between SA and impurity 2, resolution between impurity 2 and SADBE as well as the symmetry factors of SA, impurity 2 and SADBE. Taking each response as dependent variable y, a relationship between y and the experimental variables xi, xj,… could be generated applying multivariate regression analysis where the regression coefficients described the quantitative effect of the respective experimental variables. 3. Results and Discussion 3.1. Method development As the final goal was to provide a high quality drug for therapeutic use, a reliable analytical method for purity determination was mandatory [11]. The challenge here was that SADBE was a di-ester, while SA was a strong acid, meaning a high variation in hydrophobicity. This implied that the neutral compound SADBE would normally be quite well retained on a reversed phase column, but that SA, which was easily ionized, would show only low affinity. However, the wide variety of column chemistry in reversed phase LC nowadays could offer reasonable retention of polar as well
8
as neutral compounds. Preliminary studies were conducted using a C18 column modified to retain polar compounds (250 mm × 4.6 mm i.d., 5 µm) and a C2 column (250 mm × 4.6 mm i.d., 5 μm). Both columns retained SA using an isocratic mobile phase consisting of 30 % of 0.1 % trifluoroacetic acid and 70 % of acetonitrile. In terms of peak shape and repeatability, the C2 stationary phase gave better results than the modified C18 column since SA suffered from tailing on the latter column. So, further optimization was performed on the C2 column. The difference in polarity between SADBE and SA had also consequences for the selection of the sample solvent. To dissolve SA, water was required, but SADBE has been reported to be unstable in aqueous solutions [6]. As a consequence, the percentage of water in the solvent was kept as minimal as possible. Moreover, alcohols could inverse the rate of hydrolysis and cause transesterification [6]. So, several ratios of acetonitrile and water were tested. Finally, acetonitrile – water (95:5, v/v) was selected as solvent. 3.2. Further optimization of the method The developed method was further optimized concerning the percentage of organic modifier and the acid in the aqueous part of the mobile phase. Two impurities were found to be present in commercial SADBE samples, but the overall resolution was rather poor when the mobile phase contained 70 % of acetonitrile. Decreasing the percentage of acetonitrile to 60 % increased the total analysis time, but it did not improve the resolution. Additional experiments revealed that a mobile phase consisting of 0.1 % of trifluoroacetic acid – acetonitrile – methanol (30:60:10, v/v/v) improved the peak shape, but that the selectivity was still insufficient. In order to improve this, the pH was adjusted using as aqueous part of the mobile phase, varying concentrations of phosphate buffer pH 3, ammonium formate buffer pH 3, phosphoric acid and formic acid. In general, controlling the pH improved the peak shape of SA as well as its retention on the stationary
9
phase. For quantification purposes, 0.15 % of phosphoric acid was selected as it provided the best repeatability and resolution. A typical chromatogram is shown in Figure 2. However, phosphoric acid was not compatible with MS, so that the 5 mM ammonium formate buffer pH 3 (which showed similar, but somewhat inferior results) was used as an alternative to characterize the impurities by LC-MS. 3.3. Stress decomposition studies The objective of the international guidelines that regulate pharmaceutical products is to ensure that patients get effective and safe medicines. One of the keys to ensure both terms is monitoring the stability of the active drug substance. So, a stability indicating method is needed to identify and quantify degradation impurities. Therefore, forced degradation studies are required to provide information about potential degradation products [12]. Although forced degradation studies have been described in different international guidelines such as the International Council for Harmonization of Technical Requirements for Pharmaceuticals for Human Use (ICH), only general terms for the conditions used for forced degradation are defined [13-14]. Thus, more specific guidelines about how to conduct stress tests were followed [8]. Here, conditions were adjusted to achieve not more than 30 % of degradation products. Therefore, drastic conditions were avoided to minimize the excessive formation of degradation products which were not relevant. The degradation solutions obtained under the different stress conditions outlined in Table 1 were injected applying the final chromatographic conditions mentioned under section 2.3.1. The solutions subjected to acidic, neutral and basic conditions showed sufficient degradation products after 1 h at 40 ºC. Stress oxidation was applied at room temperature and it took about 3 h to form a reasonable amount of degradation products. In all cases SA and impurity 2 were produced. It
10
was observed that the percentage of impurity 2 was higher than SA under mild conditions. Upon applying longer degradation times than indicated in Table 1, the peak size of impurity 2 decreased while this of SA increased, indicating that impurity 2 was an intermediate product. Studies have shown that esters of SA could degrade to form SA through hydrolysis [6]. Thus, as intermediate during the degradation process and based on its retention time in the chromatogram, impurity 2 could be assumed to be squaric acid monobutyl ester, which was formed by loss of one butyl side chain from SADBE (Figure 1-B). 3.4 Confirmation of impurities by MS Depending on the amount of a drug substance that is administered per day, the identification threshold varies from 0.05 % to 0.10 % [14]. Here, LC-MS experiments were conducted to study the structure of impurities whereof the amount exceeded 0.05 %. As mentioned before already, phosphoric acid was replaced in the mobile phase by 5 mM ammonium formate buffer pH 3. First the fragmentation pattern of SADBE was studied. In positive ion mode, the protonated base peak ([M+H]+) yielded an m/z of 226.8. This ion was further fragmented with a CEL of 30 %. The protonated molecular ion could form fragment ions at m/z 170.8 and m/z 115.0 through consecutive neutral losses of two 1-butene side chain moieties (2×56 Da). Those product ions corresponded to the protonated structures B and C shown in Figure 1. No further fragmentation was noticed. In the negative mode, SADBE could not be detected as it could not lose a proton to form a deprotonated ion. Two impurities were detected above the threshold in commercial samples. The first impurity was already identified as SA by comparing its retention time with that of a reference standard of SA. It was further confirmed by MS in negative ion mode (no signal was obtained in positive mode), yielding an [M–H]- at m/z 113.1. Impurity 2 was also investigated both in the positive and negative
11
mode to confirm its structure proposed under 3.3. Similar to SA, it could not be detected in the positive mode. In the negative mode, it showed a deprotonated molecule ([M–H]-) at m/z 169.0, indicating that the molecular weight was 170. This was 56 Da more compared to SA and 56 Da less compared to SADBE. So, it could be deduced that impurity 2 was squaric acid monobutyl ester (Figure 1-B). Upon collision with a CEL of 30 % to induce fragmentation, the mass of the ion with m/z 169.0 decreased by 57 Da to form the most abundant ion at m/z 112.1 (Figure 3). In contrast with SADBE, it was formed through cleavage of the radical ion C4H9 instead of C4H8 and so it did not obey the “even-electron rule” [15]. Similar to SA, the fragment ion at m/z 112.1 might become more aromatic upon deprotonation [16]. Recent studies have shown that in some cases, structures involving aromatic rings could lose an alkyl group as radical, which could be rationalized by resonance stabilization of the product ions [17-19]. The deprotonated molecular ion also lost butyraldehyde (72 Da) to form the fragment ion at m/z 97.2. No further fragment ions were found in the spectrum. 3.5. Method validation 3.5.1. Specificity The specificity of the developed method was investigated by injecting the stress samples. All peaks were well separated and the resolution was in all cases not less than 2. Moreover, a photodiode array detector was used to scan the SADBE and all impurity peaks in the range from 200 to 400 nm. Spectra indicated that no co-elution was present. 3.5.2. Sensitivity The limits of detection (LOD) of SADBE and SA were determined as the concentration of the analyte yielding a signal-to-noise (S/N) ratio of 3. For both compounds, this concentration was equal to 0.003 µg mL-1 (0.003 %). The limit of quantification (LOQ) was determined at a S/N ratio
12
of 10 and corresponded for both SADBE and SA to a concentration of 0.01 µg mL-1 (0.01 %). The percentages are expressed taking a sample concentration of 100 µg mL-1 as 100 %. Since no considerable differences in LOD and LOQ values were noticed for SADBE and SA, it was assumed that those were of the same order for impurity 2, whereof no reference substance was available. 3.5.3. Linearity The linearity was checked by analyzing 6 concentrations covering a range from 0.01 to 120 µg mL1
for SADBE and from 0.01 to 10 µg mL-1 for SA.
Each concentration was injected in triplicate. The least squares method was used to calculate the regression equation which was y = 428553 x + 32183 for SADBE and y = 679977 x – 76867 for SA with y the peak area and x the concentration expressed in µg mL-1. The coefficients of correlation for SADBE and SA were both 0.9999. The residual plots showed a random scattering around the horizontal zero axis for both compounds. Besides, zero was included in the 95 % confidence interval of both intercepts indicating that the intercepts were statistically not significant. From the slopes of the regression equations, the relative response factor for SA vs. SADBE was derived. It was found to be 1.6. 3.5.4. Precision Precision was tested on two levels: repeatability and intermediate precision. The intraday precision (repeatability) was evaluated by analyzing six times the same solution of SADBE and SA. The relative standard deviations were 0.7 % and 0.3 % respectively. The intermediate precision was calculated over a time span of three consecutive days. Satisfactory results were also obtained here since 1.0 % was found as relative standard deviation for both
13
SADBE and SA. Obviously, the developed method was found to be precise as the results met the criteria (≤ 2 %) [20]. 3.5.5. Accuracy The accuracy was evaluated by determination of the SADBE concentration by back calculation using the regression equation. A recovery of 99.7 % (RSD = 0.7 %, n = 6) for the drug substance was obtained. These results indicated that the method was accurate. 3.5.6. Factorial analysis The four factors employed in the fractional factorial design have been outlined in Table 2. They were selected to mimic the reliability of the method during routine application. The factors were quantitative and were examined at two levels [10]. As responses, parameters such as resolution, relative retention time and symmetry factor were monitored (see also 2.4). Figure 4 shows the regression coefficient plot for resolution 1 between SA and impurity 2. None of the factors had a significant effect, except the column temperature which had a small influence. Indeed, increasing the column temperature improved somewhat resolution 1. However, in all cases resolution 1 was not less than 2.7 as illustrated in the response surface plot (Figure 5). On the other hand, acetonitrile had a negative significant effect on resolution 2 between impurity 2 and SADBE (Figure 6). However, Figure 7 shows that resolution 2 was not lower than 2 in the domain examined. Generally, this is an acceptable value, but controlling acetonitrile within these specific limits is advisable to avoid problems. Since the symmetry factor is also regularly mentioned in monographs as a system suitability parameter, it was also considered for SA, impurity 2 and SADBE. No significant influence of the tested factors was observed on the symmetry factors of SA and impurity 2. However, increasing
14
the column temperature had a slightly negative influence on the symmetry of the SADBE peak. On the other hand, the value did not exceed 1.3, a value which is still acceptable according to the European Pharmacopoeia [21]. The separation could also be evaluated by monitoring the relative retention times of SA and impurity 2. Although SA as a polar compound was maybe sensitive to changes, no significant influence was observed on the relative retention times of SA and impurity 2. It could be concluded that the small deliberate changes imposed during the factorial analysis study caused no important differences in the responses. Even if the influence was significant, the responses were still within acceptable limits. 3.5.7. Stability of the sample solutions In order to check the stability of the sample solution, a solution of SADBE (100 µg mL-1) containing also small amounts of SA and impurity 2, was kept at room temperature. The solution was injected at appropriate time intervals for 18 h. SADBE and impurity 2 showed no degradation, but the SA peak decreased by 30 % over 3 h. Experiments performed with samples cooled at 5 ºC revealed that SA was stable for at least 6 h. So, it is recommended to use freshly prepared sample solutions or a temperature controlled autosampler. Acetone (which is present in the examined pharmaceutical formulation) was found to have no influence on the stability of the test solution. 3.6. Analysis of bulk samples and of a compounded formulation For each bulk sample, the impurities were determined by injecting the respective test solution (100 µg mL-1) and a dilution hereof (1 µg mL-1) which was used as 1 % reference. Quantification was carried out by comparing the peak areas of the impurities with this of the main compound in
15
the reference solution. For the calculation of the content of SA, a correction factor of 0.6 was applied. The content of SA was also calculated using a 1 % SA reference solution. The same results were achieved, indicating that both procedures could be applied. Two batches were examined and impurities above 0.05 % were reported. The results showed that two impurities were present. SA was found to be 0.09 % and 0.06 % for batch 1 and 2, respectively. Values for impurity 2 were 4.5 % and 2.3 %, respectively. The developed method was also used to determine the content of SADBE in a pharmaceutical formulation. The content was 99.6 % with an RSD of 1.3 % (n = 3). The formulation was also found to be stable for at least 2 weeks, which is sufficient since it is usually not longer stored after reconstitution. 4. Conclusion The availability of a proper LC method to ensure the quality of drug substances and drug products is essential, but until now such a method was missing for SADBE. Forced degradation studies of SADBE combined with MS revealed the formation of SA and its monobutyl ester. All these components could be well separated using the developed method. It is the first time that an LC method has been described in literature for the determination of SADBE and its impurities. The simple and fast method was successfully validated and applied to real samples.
16
References [1] G. Micali, M.R. Nasca, A. Tedeschi, F. Dall’Oglio, N. Pulvirenti, Use of squaric acid dibutylester for cutaneous warts in children, Pediatr. Dermatol. 17 (2000) 315-318. [2] S. Pandey, E.N. Wilmer, D.S. Morrell, Examining the efficiency and safety of squaric acid therapy for treatment of recalcitrant warts in children, Pediatr. Dermatolol. 32 (2015) 85. [3] F. Dall’Oglio, M.R. Nasca, M.L. Musumeci, G. La Torre, G. Ricciardi, C. Potenza, G. Micali, Topical immunomodulator therapy with squaric acid dibutylester is effective treatment for severe alopecia areata: Results of an open-label, paired comparison, clinical trial, J. Dermatolog. Treat. 16 (2005) 10-14. [4] E.F. Sherertz, K.B. Sloan, Percutaneous penetration of squaric acid and its esters in hairless mouse and human skin in vitro, Arch. Dermatol. Res. 280 (1988) 57-60. [5] A.K. Tiwary, D.K. Mishra, S.S. Chaudhary, Comparative study of efficacy and safety of topical squaric acid dibutylester and diphenylcyclopropenone for the treatment of Alopecia Areata, N. Am. J. Med. Sci. 6 (2016) 237-242. [6] M.G. Wilkerson, J. Henkin, J.K. Wilkin, R.G. Smith, Squaric acid and esters: Analysis for contaminants and stability in solvents, J. Am. Acad. Dermatol. 13 (1985) 229-234. [7] U.D. Neue, HPLC Columns: Theory, Technology and Practice, Wiley-VCH, New York, 1997. [8] S. Singh, M. Bakshi, Guidance on conduct of stress tests to determine inherent stability of drugs, Pharm. Technol. on-line 4 (2000) 1-14. [9] D.L. Massart, B.G.M. Vandeginste, L.M.C. Buydens, S. de Jong, P.J. Lewi, J. Smeyers-Verbeke, Handbook of Chemometrics and Qualimetrics: Part A and B, Elsevier, The Netherlands, 2003.
17
[10] G. Hanrahan, F.A. Gomez, Chemometric Methods in Capillary Electrophoresis, John Wiley & Sons, New Jersey, 2010. [11] S. Görög, Various aspects of the estimation of impurities in drugs; In: Identification and determination of impurities in drugs, Görög, S. Ed., 1st ed., Elsevier, Amsterdam, The Netherlands, 2000. [12] D. Jain, P.K. Basniwal, Forced degradation and impurity profiling: Recent trends in analytical perspectives, J. Pharm. Biomed. Anal. 86 (2013) 11-35. [13] K.K. Hotha, S.P.K. Reddy, V.K. Raju, L.K. Ravindranath, Forced degradation studies: practical approach-overview of regulatory guidance and literature for the drug products and drug substances, Int. Res. J. Pharm., 4 (2013) 78-85. [14] ICH Q3A (R2), Impurities in new drug substances, Geneva, 2006. [15] M. Karni, A. Mandelbaum, The even-electron rule, Org. Mass Spectrom. 15 (1980) 53- 64. [16] R.I. Storer, C. Aciro, L.H. Jones, Squaramides: physical properties, synthesis and applications, Chem. Soc. Review 40 (2011) 2330-2346. [17] X. Zhang, P. Zhu, H. Zhang, Z. Li, K. Jiang, M.R. Lee, The competing radical eliminations in the tandem mass spectrometry of the OH-deprotonated benzyl vanillate, J. Mass Spectrom. 50 (2015) 432-436. [18] X. Zhang, F. Li, H. Lv, Y. Wu, G. Bian, K. Jiang, On the origin of the methyl radical loss from deprotonated ferulic and isoferulic acids: electronic excitation of a transient structure, J. Am. Soc. Mass Spectrom. 24 (2013) 941-948. [19] J.H. Bowie, The fragmentations of even-electron organic negative ions, Mass Spectrom. Rev.
18
9 (1990) 349-379. [20] J. Ermer, J.H. McB. Miller, Method validation in pharmaceutical analysis, Wiley-VCH Verlag GmbH & Co. KGaA, Germany, 2008. [21] European Pharmacopoeia, 9th Edition, Council of Europe; Strasbourg, 2017. Figure Caption
19
Figr-1 O
O
O
O
H2O
O
O
O
OH
A Mw: 226
H2O
B Mw: 170
O
OH
O
OH C Mw: 114
Figure 1. Hydrolysis scheme of squaric acid dibutyl ester in the presence of water: A) Squaric acid dibutyl ester, B) Squaric acid monobutyl ester, C) Squaric acid
20
1600
Figure 2. Overlay of chromatograms: A) Acetone (10 µg/mL), B) Squaric acid (5 µg/mL), C) Squaric acid dibutyl ester (100 µg/mL) Chromatographic conditions: mobile phase: (0.15 % phosphoric acid- acetonitrile1200 methanol (30: 60: 10, v/v/v), flow rate: 1.0 mL/min, column temperature: 25 ºC, SADBE injection volume: 20 µL, UV detection: 252 nm, column: Maxsil C2 (250 mm x 1000 4.6 mm i.d., 5 μm)
1400
800
600
Impurity 2
400
SA 200
Acetone
C B
0 0
1
2
3
4
5
Minutes
21
6
7
8
9
A 10
Figure 3. [M–H] ‾ spectrum acquired for impurity 2
22
Figure 4. Regression coefficient plot obtained from the factorial analysis study of resolution 1 between SA and impurity 2.
23
Figure 5. Response surface plot showing the influence of column temperature and methanol on resolution 1 between SA and impurity 2.
24
Figure 6. Regression coefficient plot obtained from the factorial analysis study of resolution 2 between impurity 2 and SADBE.
25
Figure 7. Response surface plot showing the influence of acetonitrile and phosphoric acid on resolution 2 between impurity 2 and SADBE.
26
Table 1: Optimized stress conditions for squaric acid dibutyl ester 40 ºC
RT
Acid
Neutral
Base
Oxidation
Concentration
1 mM HCl
H2O
0.1 mM NaOH
3 % H2O2
Duration
1h
1h
1h
3h
RT: room temperature
27
Table 2: Chromatographic parameter settings applied in the fractional factorial design, corresponding to the low (-), central (0) and high (+) levels.
Parameter
Low value (-)
Central value (0)
High value (+)
% Methanol in MP
9.5
10.0
10.5
% Acetonitrile in MP
59
60
61
0.13
0.15
0.17
23
25
27
% Phosphoric acid in MP Column temperature
MP: mobile phase
28
S0731-7085(17)30451-X http://dx.doi.org/doi:10.1016/j.jpba.2017.04.022 PBA 11220
To appear in:
Journal of Pharmaceutical and Biomedical Analysis
Received date: Revised date: Accepted date:
22-2-2017 13-4-2017 15-4-2017
Please cite this article as: Marwa F.Mansour, Peixi Zhu, Ann Van Schepdael, Erwin Adams, Development and validation of a liquid chromatographic method for the analysis of squaric acid dibutyl ester and its impurities, Journal of Pharmaceutical and Biomedical Analysishttp://dx.doi.org/10.1016/j.jpba.2017.04.022 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Development and validation of a liquid chromatographic method for the analysis of squaric acid dibutyl ester and its impurities Marwa F. Mansour 1,2, Peixi Zhu1, Ann Van Schepdael1, Erwin Adams1 1
KU Leuven, University of Leuven, Department of Pharmaceutical and Pharmacological Sciences,
Pharmaceutical Analysis, Leuven, Belgium 2
National Organization for Drug Control and Research, Giza, Egypt
Correspondence: Professor Erwin Adams, Farmaceutische Analyse, KU Leuven, O&N2, PB 923, Herestraat 49, B3000 Leuven, Belgium E-mail: [email protected] Fax: +32-16-323448 Tel.: +32-16-323444
Highlights
First time that a quality control LC method is described for squaric acid dibutyl ester (SADBE). Method is simple, short, selective and robust. Characterization of SADBE and its forced degradation products by LC-MS.
1
Abstract A simple, fast and selective stability indicating liquid chromatographic method has been described for the simultaneous determination of squaric acid dibutyl ester and its impurities. The chromatographic separation was achieved on a C2 column (250 mm × 4.6 mm i.d., 5 μm) using a mobile phase consisting of 0.15 % phosphoric acid – acetonitrile – methanol (30:60:10, v/v/v). Isocratic elution was performed at a flow rate of 1.0 mL min-1. The analytes were detected by UV at 252 nm. The method was validated according to the ICH guidelines and satisfactory results were obtained. The specificity of the developed method was tested using forced degradation solutions of the drug substance. Characterization of squaric acid dibutyl ester and its forced degradation products was achieved by coupling mass spectrometry (MS) to the liquid chromatographic (LC) system. The method was successfully applied for quality control purposes including assay and determination of related compounds as required by regulatory guidelines to ensure its safety and efficacy since no monograph is available in official compendia. Keywords: Squaric acid dibutyl ester; ; ; ; , quality control, liquid chromatography, forced degradation, mass spectrometry.
2
1. Introduction Since destructive and surgical treatment of warts is painful, especially for children, topical immunotherapy with contact sensitizers is encouraged [1-2]. However, some of these sensitizing compounds such as 2,4-dinitrochlorobenzene are considered as mutagenic [3-4]. Therefore, effective alternative compounds are of considerable interest. Recent studies showed that warts in children caused by the human papilloma virus could be treated safely by topical application of squaric acid dibutyl ester (SADBE). In addition, SADBE is also successfully used as topical immunomodulator for the treatment of alopecia [5]. SADBE is chemically known as (3,4-dibutoxycyclobut-3-ene-1,2-dione). It is a four membered ring compound derived from squaric acid (SA). The latter is also easily formed by hydrolysis as shown in Figure 1. Therefore, SA is considered to be a likely impurity of SADBE since it is starting material and degradation product as well. It is worth noting that SA decreases the activity of SADBE, as it is considered to be inactive as contact sensitizer [6]. In vitro results on the human skin support the assumption that the ester can combine with a protein to form an antigen while SA can not [4]. The medicinal significance of SADBE has been acknowledged since FDA approved recently its inclusion in the list of drug substances used to compound human drug products. Although SADBE plays an important role in the treatment of several patients, no monograph of it is adopted in a pharmacopoeia or no proper analytical method can be found in literature to determine its impurities. However, such methods are crucial, as it is known that impurities reduce the effectiveness of a drug and often cause allergic reactions and side effects. Only one report related to this was found describing a GC-MS method to test the presence of carcinogenic contaminants. In this research SA, one of the hydrolysis products, was detected by spectrophotometry at 480 nm after its complexation with ferric chloride [6].
3
Therefore, it was important to develop an accurate, reliable analytical method for the analysis of SADBE. LC was considered the most appropriate technique for quality control purposes. Special attention was paid to SA since it was observed that the concentration of SA as impurity in SADBE influences the therapeutic activity in the treatment of warts and alopecia [4,6]. Whereas SADBE is rather apolar, SA is usually present in solution as a polar anion (pKa values of 0.5 and 3.6). Traditionally, normal phase and ion pair reversed phase LC are powerful tools for the retention of polar compounds. However, normal phase LC suffers from limitations such as solubility problems for the analytes while ion pair systems need a long equilibration time and are mostly not compatible with LC-MS [7]. Therefore, an alternative separation mode such as hydrophilic interaction liquid chromatography (HILIC) could be considered for the separation of polar compounds that suffer from poor solubility in the organic solvents used in normal phase LC. However, HILIC requires also a rather long equilibration time and it is quite costly due to the consumption of relatively high amounts of organic solvents. Thus, if a reversed phase method could avoid SA to be eluted in the void volume, it could be interesting in terms of time, cost and simplicity of application. The aim of this paper was to develop a stability indicating method for the determination of the main compound SADBE as well as its impurities by applying reversed phase LC. The drug substance was subjected to forced degradation studies in order to induce its degradation products. These solutions were further used to optimize the chromatographic conditions and to test the robustness of the developed method. 2. Experimental 2.1. Chemicals and reagents
4
SADBE was purchased from Sigma-Aldrich (St. Louis, MO, USA). SA, methanol (HPLC grade) and formic acid were supplied by Acros Organics (Morris Plains, New Jersey, USA). Phosphoric acid (85 % m/m) was obtained from Acros Organics (Geel, Belgium). Acetonitrile (HPLC gradient grade) and hydrochloric acid were obtained from Fisher Scientific (Leicester, United Kingdom). Acetonitrile and methanol MS grade were acquired from Biosolve LTD (Valkenswaard, The Netherlands). Ammonium acetate MS grade was purchased from Sigma-Aldrich (Zwijndrecht, The Netherlands). Trifluoroacetic acid was supplied by Carl Roth GmbH (Karlsruhe, Germany). Sodium hydroxide was obtained from VWR International (Leuven, Belgium) and hydrogen peroxide 30 % was acquired from Merck (Darmstadt, Germany). Nitrogen supplied by Air Liquide (Liège, Belgium) was used as sheath and auxiliary gas for MS. Helium gas was purchased from Air Products (Brussels, Belgium). Water was produced in-house using a Milli-Q Gradient water purification system (Millipore, Bedford, MA, USA). A pharmaceutical formulation containing 1 % SADBE was prepared in acetone as a delivery system [2,4]. 2.2. Forced degradation studies Forced degradation studies of the bulk drug (1 mg mL-1 solutions) were performed in accordance with Singh et al. [8]. SADBE was subjected to stress studies involving acidic, basic and neutral conditions at 40 °C. An oxidative study was carried out in 3 % (v/v) H2O2 at room temperature. Details have been outlined in Table 1. In all cases, the duration was such that sufficient degradation products were obtained. The resultant solutions were withdrawn at appropriate times, neutralized if applicable and diluted to obtain a concentration of 100 μg mL-1 using acetonitrile – water (95:5, v/v) as solvent. 2.3. LC instrumentation and chromatographic conditions
5
2.3.1. Non-volatile system for LC-UV The LC system from Merck-Hitachi (Tokyo, Japan) was equipped with a high pressure pump (L7100), an autosampler (L-7200), a UV/VIS (L-7400) or photodiode array (L-2450) detector and inline solvent degasser (L-7614). EZ Chrome Elite software version 3.1.6 (Merck-Hitachi) was used for data processing and acquisition. Optimum chromatographic separations were achieved on a Maxsil C2 column (250 mm × 4.6 mm i.d., 5 μm) (Phenomenex, Torrance, CA, USA) which was maintained at 25 °C. During method development, also a Polaris C18 (250 mm × 4.6 mm i.d., 5 µm) column (Agilent, Santa Clara, CA, USA) was examined. The mobile phase was pumped isocratically at a flow rate of 1.0 mL min-1. It consisted of 0.15 % phosphoric acid – acetonitrile – methanol (30:60:10, v/v/v). The injection volume was 20 μL and UV detection was performed at 252 nm. The mobile phase was degassed before use by sparging helium. Stock solutions of SADBE and SA (1 mg mL-1), working solutions (100 µg mL-1) and further dilutions hereof were prepared using acetonitrile – water (95:5, v/v) as solvent. 2.3.2. Volatile system for LC-MS An LCQ ion trap (IT) mass spectrometer from ThermoFinnigan (San Jose, CA, USA) equipped with an electrospray ionization source (ESI) was operated in the positive mode for the confirmation of the molecular mass of SADBE and in the negative mode for SA and the identification of an unknown impurity. Helium was used as the damping and collision gas. Instrument and data acquisition were performed with Xcalibur 1.2 software from ThermoFinnigan. The optimized settings for the mass spectrometer in positive mode were as follows: sheath gas flow rate: 85 arbitary units (arb); auxiliary gas flow rate: 25 arb; spray voltage: 4.0 kV; capillary temperature: 210 °C; capillary voltage: 7.0 V; tube lens offset: -30.0 V; multipole 1 offset: -2.75 V;
6
lens voltage: -24 V; multipole 2 offset: -6.00 V; RF amplitude: 400 V, peak-to-peak; scan ranges: m/z 50-800. The optimized settings for the mass spectrometer in negative mode were as follows: sheath gas flow rate: 85 arb; auxiliary gas flow rate: 25 arb; spray voltage: 4.0 kV; capillary temperature: 210 °C; capillary voltage: -5.0 V; tube lens offset: 10 V; multipole 1 offset: 3.00 V; lens voltage: 16 V; multipole 2 offset: 7.00 V; RF amplitude: 400 V, peak-to-peak; scan ranges: m/z 50-800. For LC-MSn experiments, the ions of interest were isolated in the IT with an isolation width of 2 u and the collision energy level (CEL) was set at 30 %. The separation was also realized using a Maxsil C2 column (250 mm × 4.6 mm i.d., 5 μm). As the non-volatile mobile phase described in 2.3.1 contained phosphoric acid, it could not be used for the LC-MS system. Therefore, it was replaced by 5 mM ammonium formate pH 3 and so the composition used for LC-MS was ammonium formate (pH 3; 5 mM) – acetonitrile – methanol in a ratio of 30:60:10 (v/v/v). It was pumped at a flow rate of 1.0 mL min-1. A T-piece was used to split one fifth of the mobile phase to the MS. The samples injected were the working solutions of SADBE and SA (100 µg mL-1) in acetonitrile – water (95:5, v/v). 2.4. Factorial analysis
A more in-depth study of the experimental parameters or factors will lead to more insight in the chromatographic behavior of the main compound and its impurities. This is helpful to adjust the conditions when the separation would be somewhat different, for example when the method is applied in another laboratory. The use of an experimental design is appropriate to explore the parameters in an organized way with a minimal number of experiments. It allows to quantify the importance of the parameters and to decide whether
7
they are significant or not. This way, analysts know which parameters are the most interesting to adapt or which can be considered robust. The set-up of the design, analysis of the results by means of multivariate analysis and construction of the graphical plots were performed using R-software (R foundation, Vienna, Austria). The tested experimental factors were column temperature and percentages of phosphoric acid, acetonitrile and methanol in the mobile phase. These factors might have an influence on the overall separation of the developed method. A two level fractional factorial design was applied with a number of runs equal to 2k-1 + n, where k is the number of factors and n is the number of center points [9,10]. Therefore, 11 experiments including 3 center points were performed. Several responses were selected: relative retention time for SA, relative retention time for impurity 2 (relative vs. the main peak), resolution between SA and impurity 2, resolution between impurity 2 and SADBE as well as the symmetry factors of SA, impurity 2 and SADBE. Taking each response as dependent variable y, a relationship between y and the experimental variables xi, xj,… could be generated applying multivariate regression analysis where the regression coefficients described the quantitative effect of the respective experimental variables. 3. Results and Discussion 3.1. Method development As the final goal was to provide a high quality drug for therapeutic use, a reliable analytical method for purity determination was mandatory [11]. The challenge here was that SADBE was a di-ester, while SA was a strong acid, meaning a high variation in hydrophobicity. This implied that the neutral compound SADBE would normally be quite well retained on a reversed phase column, but that SA, which was easily ionized, would show only low affinity. However, the wide variety of column chemistry in reversed phase LC nowadays could offer reasonable retention of polar as well
8
as neutral compounds. Preliminary studies were conducted using a C18 column modified to retain polar compounds (250 mm × 4.6 mm i.d., 5 µm) and a C2 column (250 mm × 4.6 mm i.d., 5 μm). Both columns retained SA using an isocratic mobile phase consisting of 30 % of 0.1 % trifluoroacetic acid and 70 % of acetonitrile. In terms of peak shape and repeatability, the C2 stationary phase gave better results than the modified C18 column since SA suffered from tailing on the latter column. So, further optimization was performed on the C2 column. The difference in polarity between SADBE and SA had also consequences for the selection of the sample solvent. To dissolve SA, water was required, but SADBE has been reported to be unstable in aqueous solutions [6]. As a consequence, the percentage of water in the solvent was kept as minimal as possible. Moreover, alcohols could inverse the rate of hydrolysis and cause transesterification [6]. So, several ratios of acetonitrile and water were tested. Finally, acetonitrile – water (95:5, v/v) was selected as solvent. 3.2. Further optimization of the method The developed method was further optimized concerning the percentage of organic modifier and the acid in the aqueous part of the mobile phase. Two impurities were found to be present in commercial SADBE samples, but the overall resolution was rather poor when the mobile phase contained 70 % of acetonitrile. Decreasing the percentage of acetonitrile to 60 % increased the total analysis time, but it did not improve the resolution. Additional experiments revealed that a mobile phase consisting of 0.1 % of trifluoroacetic acid – acetonitrile – methanol (30:60:10, v/v/v) improved the peak shape, but that the selectivity was still insufficient. In order to improve this, the pH was adjusted using as aqueous part of the mobile phase, varying concentrations of phosphate buffer pH 3, ammonium formate buffer pH 3, phosphoric acid and formic acid. In general, controlling the pH improved the peak shape of SA as well as its retention on the stationary
9
phase. For quantification purposes, 0.15 % of phosphoric acid was selected as it provided the best repeatability and resolution. A typical chromatogram is shown in Figure 2. However, phosphoric acid was not compatible with MS, so that the 5 mM ammonium formate buffer pH 3 (which showed similar, but somewhat inferior results) was used as an alternative to characterize the impurities by LC-MS. 3.3. Stress decomposition studies The objective of the international guidelines that regulate pharmaceutical products is to ensure that patients get effective and safe medicines. One of the keys to ensure both terms is monitoring the stability of the active drug substance. So, a stability indicating method is needed to identify and quantify degradation impurities. Therefore, forced degradation studies are required to provide information about potential degradation products [12]. Although forced degradation studies have been described in different international guidelines such as the International Council for Harmonization of Technical Requirements for Pharmaceuticals for Human Use (ICH), only general terms for the conditions used for forced degradation are defined [13-14]. Thus, more specific guidelines about how to conduct stress tests were followed [8]. Here, conditions were adjusted to achieve not more than 30 % of degradation products. Therefore, drastic conditions were avoided to minimize the excessive formation of degradation products which were not relevant. The degradation solutions obtained under the different stress conditions outlined in Table 1 were injected applying the final chromatographic conditions mentioned under section 2.3.1. The solutions subjected to acidic, neutral and basic conditions showed sufficient degradation products after 1 h at 40 ºC. Stress oxidation was applied at room temperature and it took about 3 h to form a reasonable amount of degradation products. In all cases SA and impurity 2 were produced. It
10
was observed that the percentage of impurity 2 was higher than SA under mild conditions. Upon applying longer degradation times than indicated in Table 1, the peak size of impurity 2 decreased while this of SA increased, indicating that impurity 2 was an intermediate product. Studies have shown that esters of SA could degrade to form SA through hydrolysis [6]. Thus, as intermediate during the degradation process and based on its retention time in the chromatogram, impurity 2 could be assumed to be squaric acid monobutyl ester, which was formed by loss of one butyl side chain from SADBE (Figure 1-B). 3.4 Confirmation of impurities by MS Depending on the amount of a drug substance that is administered per day, the identification threshold varies from 0.05 % to 0.10 % [14]. Here, LC-MS experiments were conducted to study the structure of impurities whereof the amount exceeded 0.05 %. As mentioned before already, phosphoric acid was replaced in the mobile phase by 5 mM ammonium formate buffer pH 3. First the fragmentation pattern of SADBE was studied. In positive ion mode, the protonated base peak ([M+H]+) yielded an m/z of 226.8. This ion was further fragmented with a CEL of 30 %. The protonated molecular ion could form fragment ions at m/z 170.8 and m/z 115.0 through consecutive neutral losses of two 1-butene side chain moieties (2×56 Da). Those product ions corresponded to the protonated structures B and C shown in Figure 1. No further fragmentation was noticed. In the negative mode, SADBE could not be detected as it could not lose a proton to form a deprotonated ion. Two impurities were detected above the threshold in commercial samples. The first impurity was already identified as SA by comparing its retention time with that of a reference standard of SA. It was further confirmed by MS in negative ion mode (no signal was obtained in positive mode), yielding an [M–H]- at m/z 113.1. Impurity 2 was also investigated both in the positive and negative
11
mode to confirm its structure proposed under 3.3. Similar to SA, it could not be detected in the positive mode. In the negative mode, it showed a deprotonated molecule ([M–H]-) at m/z 169.0, indicating that the molecular weight was 170. This was 56 Da more compared to SA and 56 Da less compared to SADBE. So, it could be deduced that impurity 2 was squaric acid monobutyl ester (Figure 1-B). Upon collision with a CEL of 30 % to induce fragmentation, the mass of the ion with m/z 169.0 decreased by 57 Da to form the most abundant ion at m/z 112.1 (Figure 3). In contrast with SADBE, it was formed through cleavage of the radical ion C4H9 instead of C4H8 and so it did not obey the “even-electron rule” [15]. Similar to SA, the fragment ion at m/z 112.1 might become more aromatic upon deprotonation [16]. Recent studies have shown that in some cases, structures involving aromatic rings could lose an alkyl group as radical, which could be rationalized by resonance stabilization of the product ions [17-19]. The deprotonated molecular ion also lost butyraldehyde (72 Da) to form the fragment ion at m/z 97.2. No further fragment ions were found in the spectrum. 3.5. Method validation 3.5.1. Specificity The specificity of the developed method was investigated by injecting the stress samples. All peaks were well separated and the resolution was in all cases not less than 2. Moreover, a photodiode array detector was used to scan the SADBE and all impurity peaks in the range from 200 to 400 nm. Spectra indicated that no co-elution was present. 3.5.2. Sensitivity The limits of detection (LOD) of SADBE and SA were determined as the concentration of the analyte yielding a signal-to-noise (S/N) ratio of 3. For both compounds, this concentration was equal to 0.003 µg mL-1 (0.003 %). The limit of quantification (LOQ) was determined at a S/N ratio
12
of 10 and corresponded for both SADBE and SA to a concentration of 0.01 µg mL-1 (0.01 %). The percentages are expressed taking a sample concentration of 100 µg mL-1 as 100 %. Since no considerable differences in LOD and LOQ values were noticed for SADBE and SA, it was assumed that those were of the same order for impurity 2, whereof no reference substance was available. 3.5.3. Linearity The linearity was checked by analyzing 6 concentrations covering a range from 0.01 to 120 µg mL1
for SADBE and from 0.01 to 10 µg mL-1 for SA.
Each concentration was injected in triplicate. The least squares method was used to calculate the regression equation which was y = 428553 x + 32183 for SADBE and y = 679977 x – 76867 for SA with y the peak area and x the concentration expressed in µg mL-1. The coefficients of correlation for SADBE and SA were both 0.9999. The residual plots showed a random scattering around the horizontal zero axis for both compounds. Besides, zero was included in the 95 % confidence interval of both intercepts indicating that the intercepts were statistically not significant. From the slopes of the regression equations, the relative response factor for SA vs. SADBE was derived. It was found to be 1.6. 3.5.4. Precision Precision was tested on two levels: repeatability and intermediate precision. The intraday precision (repeatability) was evaluated by analyzing six times the same solution of SADBE and SA. The relative standard deviations were 0.7 % and 0.3 % respectively. The intermediate precision was calculated over a time span of three consecutive days. Satisfactory results were also obtained here since 1.0 % was found as relative standard deviation for both
13
SADBE and SA. Obviously, the developed method was found to be precise as the results met the criteria (≤ 2 %) [20]. 3.5.5. Accuracy The accuracy was evaluated by determination of the SADBE concentration by back calculation using the regression equation. A recovery of 99.7 % (RSD = 0.7 %, n = 6) for the drug substance was obtained. These results indicated that the method was accurate. 3.5.6. Factorial analysis The four factors employed in the fractional factorial design have been outlined in Table 2. They were selected to mimic the reliability of the method during routine application. The factors were quantitative and were examined at two levels [10]. As responses, parameters such as resolution, relative retention time and symmetry factor were monitored (see also 2.4). Figure 4 shows the regression coefficient plot for resolution 1 between SA and impurity 2. None of the factors had a significant effect, except the column temperature which had a small influence. Indeed, increasing the column temperature improved somewhat resolution 1. However, in all cases resolution 1 was not less than 2.7 as illustrated in the response surface plot (Figure 5). On the other hand, acetonitrile had a negative significant effect on resolution 2 between impurity 2 and SADBE (Figure 6). However, Figure 7 shows that resolution 2 was not lower than 2 in the domain examined. Generally, this is an acceptable value, but controlling acetonitrile within these specific limits is advisable to avoid problems. Since the symmetry factor is also regularly mentioned in monographs as a system suitability parameter, it was also considered for SA, impurity 2 and SADBE. No significant influence of the tested factors was observed on the symmetry factors of SA and impurity 2. However, increasing
14
the column temperature had a slightly negative influence on the symmetry of the SADBE peak. On the other hand, the value did not exceed 1.3, a value which is still acceptable according to the European Pharmacopoeia [21]. The separation could also be evaluated by monitoring the relative retention times of SA and impurity 2. Although SA as a polar compound was maybe sensitive to changes, no significant influence was observed on the relative retention times of SA and impurity 2. It could be concluded that the small deliberate changes imposed during the factorial analysis study caused no important differences in the responses. Even if the influence was significant, the responses were still within acceptable limits. 3.5.7. Stability of the sample solutions In order to check the stability of the sample solution, a solution of SADBE (100 µg mL-1) containing also small amounts of SA and impurity 2, was kept at room temperature. The solution was injected at appropriate time intervals for 18 h. SADBE and impurity 2 showed no degradation, but the SA peak decreased by 30 % over 3 h. Experiments performed with samples cooled at 5 ºC revealed that SA was stable for at least 6 h. So, it is recommended to use freshly prepared sample solutions or a temperature controlled autosampler. Acetone (which is present in the examined pharmaceutical formulation) was found to have no influence on the stability of the test solution. 3.6. Analysis of bulk samples and of a compounded formulation For each bulk sample, the impurities were determined by injecting the respective test solution (100 µg mL-1) and a dilution hereof (1 µg mL-1) which was used as 1 % reference. Quantification was carried out by comparing the peak areas of the impurities with this of the main compound in
15
the reference solution. For the calculation of the content of SA, a correction factor of 0.6 was applied. The content of SA was also calculated using a 1 % SA reference solution. The same results were achieved, indicating that both procedures could be applied. Two batches were examined and impurities above 0.05 % were reported. The results showed that two impurities were present. SA was found to be 0.09 % and 0.06 % for batch 1 and 2, respectively. Values for impurity 2 were 4.5 % and 2.3 %, respectively. The developed method was also used to determine the content of SADBE in a pharmaceutical formulation. The content was 99.6 % with an RSD of 1.3 % (n = 3). The formulation was also found to be stable for at least 2 weeks, which is sufficient since it is usually not longer stored after reconstitution. 4. Conclusion The availability of a proper LC method to ensure the quality of drug substances and drug products is essential, but until now such a method was missing for SADBE. Forced degradation studies of SADBE combined with MS revealed the formation of SA and its monobutyl ester. All these components could be well separated using the developed method. It is the first time that an LC method has been described in literature for the determination of SADBE and its impurities. The simple and fast method was successfully validated and applied to real samples.
16
References [1] G. Micali, M.R. Nasca, A. Tedeschi, F. Dall’Oglio, N. Pulvirenti, Use of squaric acid dibutylester for cutaneous warts in children, Pediatr. Dermatol. 17 (2000) 315-318. [2] S. Pandey, E.N. Wilmer, D.S. Morrell, Examining the efficiency and safety of squaric acid therapy for treatment of recalcitrant warts in children, Pediatr. Dermatolol. 32 (2015) 85. [3] F. Dall’Oglio, M.R. Nasca, M.L. Musumeci, G. La Torre, G. Ricciardi, C. Potenza, G. Micali, Topical immunomodulator therapy with squaric acid dibutylester is effective treatment for severe alopecia areata: Results of an open-label, paired comparison, clinical trial, J. Dermatolog. Treat. 16 (2005) 10-14. [4] E.F. Sherertz, K.B. Sloan, Percutaneous penetration of squaric acid and its esters in hairless mouse and human skin in vitro, Arch. Dermatol. Res. 280 (1988) 57-60. [5] A.K. Tiwary, D.K. Mishra, S.S. Chaudhary, Comparative study of efficacy and safety of topical squaric acid dibutylester and diphenylcyclopropenone for the treatment of Alopecia Areata, N. Am. J. Med. Sci. 6 (2016) 237-242. [6] M.G. Wilkerson, J. Henkin, J.K. Wilkin, R.G. Smith, Squaric acid and esters: Analysis for contaminants and stability in solvents, J. Am. Acad. Dermatol. 13 (1985) 229-234. [7] U.D. Neue, HPLC Columns: Theory, Technology and Practice, Wiley-VCH, New York, 1997. [8] S. Singh, M. Bakshi, Guidance on conduct of stress tests to determine inherent stability of drugs, Pharm. Technol. on-line 4 (2000) 1-14. [9] D.L. Massart, B.G.M. Vandeginste, L.M.C. Buydens, S. de Jong, P.J. Lewi, J. Smeyers-Verbeke, Handbook of Chemometrics and Qualimetrics: Part A and B, Elsevier, The Netherlands, 2003.
17
[10] G. Hanrahan, F.A. Gomez, Chemometric Methods in Capillary Electrophoresis, John Wiley & Sons, New Jersey, 2010. [11] S. Görög, Various aspects of the estimation of impurities in drugs; In: Identification and determination of impurities in drugs, Görög, S. Ed., 1st ed., Elsevier, Amsterdam, The Netherlands, 2000. [12] D. Jain, P.K. Basniwal, Forced degradation and impurity profiling: Recent trends in analytical perspectives, J. Pharm. Biomed. Anal. 86 (2013) 11-35. [13] K.K. Hotha, S.P.K. Reddy, V.K. Raju, L.K. Ravindranath, Forced degradation studies: practical approach-overview of regulatory guidance and literature for the drug products and drug substances, Int. Res. J. Pharm., 4 (2013) 78-85. [14] ICH Q3A (R2), Impurities in new drug substances, Geneva, 2006. [15] M. Karni, A. Mandelbaum, The even-electron rule, Org. Mass Spectrom. 15 (1980) 53- 64. [16] R.I. Storer, C. Aciro, L.H. Jones, Squaramides: physical properties, synthesis and applications, Chem. Soc. Review 40 (2011) 2330-2346. [17] X. Zhang, P. Zhu, H. Zhang, Z. Li, K. Jiang, M.R. Lee, The competing radical eliminations in the tandem mass spectrometry of the OH-deprotonated benzyl vanillate, J. Mass Spectrom. 50 (2015) 432-436. [18] X. Zhang, F. Li, H. Lv, Y. Wu, G. Bian, K. Jiang, On the origin of the methyl radical loss from deprotonated ferulic and isoferulic acids: electronic excitation of a transient structure, J. Am. Soc. Mass Spectrom. 24 (2013) 941-948. [19] J.H. Bowie, The fragmentations of even-electron organic negative ions, Mass Spectrom. Rev.
18
9 (1990) 349-379. [20] J. Ermer, J.H. McB. Miller, Method validation in pharmaceutical analysis, Wiley-VCH Verlag GmbH & Co. KGaA, Germany, 2008. [21] European Pharmacopoeia, 9th Edition, Council of Europe; Strasbourg, 2017. Figure Caption
19
Figr-1 O
O
O
O
H2O
O
O
O
OH
A Mw: 226
H2O
B Mw: 170
O
OH
O
OH C Mw: 114
Figure 1. Hydrolysis scheme of squaric acid dibutyl ester in the presence of water: A) Squaric acid dibutyl ester, B) Squaric acid monobutyl ester, C) Squaric acid
20
1600
Figure 2. Overlay of chromatograms: A) Acetone (10 µg/mL), B) Squaric acid (5 µg/mL), C) Squaric acid dibutyl ester (100 µg/mL) Chromatographic conditions: mobile phase: (0.15 % phosphoric acid- acetonitrile1200 methanol (30: 60: 10, v/v/v), flow rate: 1.0 mL/min, column temperature: 25 ºC, SADBE injection volume: 20 µL, UV detection: 252 nm, column: Maxsil C2 (250 mm x 1000 4.6 mm i.d., 5 μm)
1400
800
600
Impurity 2
400
SA 200
Acetone
C B
0 0
1
2
3
4
5
Minutes
21
6
7
8
9
A 10
Figure 3. [M–H] ‾ spectrum acquired for impurity 2
22
Figure 4. Regression coefficient plot obtained from the factorial analysis study of resolution 1 between SA and impurity 2.
23
Figure 5. Response surface plot showing the influence of column temperature and methanol on resolution 1 between SA and impurity 2.
24
Figure 6. Regression coefficient plot obtained from the factorial analysis study of resolution 2 between impurity 2 and SADBE.
25
Figure 7. Response surface plot showing the influence of acetonitrile and phosphoric acid on resolution 2 between impurity 2 and SADBE.
26
Table 1: Optimized stress conditions for squaric acid dibutyl ester 40 ºC
RT
Acid
Neutral
Base
Oxidation
Concentration
1 mM HCl
H2O
0.1 mM NaOH
3 % H2O2
Duration
1h
1h
1h
3h
RT: room temperature
27
Table 2: Chromatographic parameter settings applied in the fractional factorial design, corresponding to the low (-), central (0) and high (+) levels.
Parameter
Low value (-)
Central value (0)
High value (+)
% Methanol in MP
9.5
10.0
10.5
% Acetonitrile in MP
59
60
61
0.13
0.15
0.17
23
25
27
% Phosphoric acid in MP Column temperature
MP: mobile phase
28