Chiral Separation and Thermodynamic Investigation of Ezetimibe Optical Isomers on a Chiralpak IC Column

Transcription

1 Journal of Chromatographic Science, 2016, Vol. 54, No. 9, doi: /chromsci/bmw032 Advance Access Publication Date: 9 March 2016 Article Article Chiral Separation and Thermodynamic Investigation of Ezetimibe Optical Isomers on a Chiralpak IC Column Lili Sun 1, Qiaogen Zou 2, *, Ping Wei 1, and Pingkai Ouyang 1 1 College of Biotechnology and Pharmaceutical Engineer, Nanjing Tech University, 5 Xinmofan Road, Nanjing, Jiangsu Province , China, and 2 School of Pharmaceutical Sciences, Nanjing Tech University, 5 Xinmofan Road, Nanjing, Jiangsu Province , China *Author to whom correspondence should be addressed. qiaogenzou@hotmail.com Received 25 August 2013; Revised 13 November 2015 Abstract A liquid chromatographic separation method of ezetimibe optical isomers on the immobilised-type Chiralpak IC chiral stationary phase, onto which cellulose tris-(3,5-dichlorophenylcarbamate) was covalently grafted as a chiral selector, was developed under normal-phase conditions. The optimized chromatographic conditions were as follows: n-hexane isopropanol diethylamine (90:10:0.1, v/v) as mobile phase, 1.0 ml min 1 as flow rate, 256 nm as UV detection wavelength and 25 C as column temperature. Several factors concerning the enantioseparation were studied, including quality and quantity of different polar organic modifiers, additives, and flow rate and column temperature. Separation and resolution in the temperature range of C was investigated. Apparent thermodynamic parameters were calculated from the Van t Hoff plots and used to explain the strength of interactions between the optical isomers and chiral selector. The validated method is simple, rapid and robust. Four ezetimibe optical isomers can be effectively separated from each other and the resolutions are 3.5, 2.7 and 2.5 successively. The method has been demonstrated to be applicable in routine quality control of Ezetimibe. Introduction Ezetimibe (CAS , EZ), designated as 1-(4-fluorophenyl)- 3(R)-[3-(4-fluorophenyl)-3(S)-hydroxypropyl]-4(S)-(4-hydroxy phenyl)- 2-azetidinone is a cholesterol absorption inhibitor that targets uptake at the jejunal enterocyte brush border without affecting the absorption of fat-soluble nutrients (1). Ezetimibe used alone (marked as Zetia, Ezetrol or Ezedoc) or together with statins (e.g., ezetimibe/simvastatin) (2) is an effective low-density lipoprotein cholesterol lowering agent which proved to be safe and well tolerated (3). Three asymmetric carbons in the molecule give rise to eight stereoisomers, among which ezetimibe has the lowest ED 50 (4). When ezetimibe is administered, the dose is reduced while the level of toxicity is at the same time minimized. Therefore, enantioseparation of ezetimibe and its stereoisomers is very important to therapeutic efficacy and safety. During the synthesis of ezetimibe, three of the eight chemically possible stereoisomers were obtained as impurities affecting the quality of the drug. The remaining stereoisomers were not formed due to the selectivity of the synthetic route adopted. At present, the three stereoisomers obtained as impurities of ezetimibe (EZ) are 1-(4-fluorophenyl)-3 (R)-[3-(4-fluorophenyl)-3(R)-hydroxypropyl]-4(S)-(4-hydroxy phenyl)- 2-azetidinone (EZ-II), 1-(4-fluorophenyl)-3(S)-[3-(4-fluorophenyl)-3(S)- hydroxypropyl]-4(r)-(4-hydroxyphenyl)-2-azetidinone (EZ-III) and 1-(4-fluorophenyl)-3(S)-[3-(4-fluorophenyl)-3(R)-hydroxypropyl]-4 (R)-(4-hydroxyphenyl)-2-azetidinone (EZ-IV) (Figure 1). Nowadays, most of the enantioseparation can be realized by chromatography with chiral stationary phases (CSPs) (5, 6) as it is rapid, reproducible and effective. Chiralpak IC, which results in enhanced success rate of enantiomeric resolution (7, 8), is a widely used polysaccharide-based column containing cellulose tris(3,5-dichlorophenylcarbamate) as chiral selector. The chiral polymer adheres to the silica-gel matrix, making the adsorbent robust and compatible with various solvent systems. The Author Published by Oxford University Press. All rights reserved. For Permissions, please journals.permissions@oup.com 1489

2 1490 Sun et al. (LC-10AT) with a UV-Vis Detector (SPD-10A). Samples were loaded via a column thermostat (AT-330; Tanjin Automatic-Science Instrument Co., Ltd, Tanjin, China). N2000 chromatography software (Zhejiang University, Zhida Information Engineering Co., Ltd, Zhejiang, China) was used to run the system and collect data. The column used was Chiralpak IC-cellullose tris-(3,5-dichlorophenylcarbamate) immobilizedon5μm silica-gel (250 mm 4.6 mm ID, 5 μm) from Daicel Chiral Technologies (China) Co., Ltd, Shanghai, China. All solvents were degassed by an ultrasonic bath (Kunshan Ultrasonic Instruments Co., Ltd, Jiangsu, China). Figure 1. Chemical structure of EZ (ezetimibe), EZ-II, EZ-III and EZ-IV. There has been some literatures reporting the isolation and structure elucidation of impurities in ezetimibe (9 12) based on HPLC, LC MS- MS, IR and NMR. Filip et al. achieved the separation of undesirable stereoisomer EZ-II(R,R,S) from ezetimibe by RP-HPLC analysis (11). Radhakrishnan and Subba Rao (12) also reported a chiral liquid chromatographic method for separation of EZ and EZ-II. However, to our knowledge, there is no published report describing the analytical method for ezetimibe and the three stereoisomers mentioned above. This article focuses on the separation for process-related stereoisomers including the EZ(R,S,S), EZ-II(R,R,S), EZ-III(S,S,R) and EZ-IV(S, R,R) based on the procedure described in the Schering Co. patent (13). The effects of quality and quantity of different polar organic modifiers, additives and flow rates on the enantioseparation were studied. Furthermore, separation and resolution in the temperature range of C were investigated. Apparent thermodynamic parameters were deduced from the Van t Hoff plots and used to explain the strength of interactions between the optical isomers and chiral selector (14, 15). Experimental Chromatography Analytical chromatographic separations were carried out on a Chiralpak IC column (250 mm 4.6 mm I.D., 5 μm). The mobile phases were composed of n-hexane isopropanol (IPA) diethylamine (DEA) (90:10:0.1, v/v), and the injection volume was 20 μl. The flow rate was set at 1.0 ml min 1, and the detection wavelength was fixed at 256 nm. The temperature was kept at 25 C except where the effect of temperature was studied. Required amounts of the sample solutes were weighed and dissolved in the mobile phase as sample solutions. The holdup time was determined from the elution of the solvent peak. The enantiomeric elution order on the columns was evaluated by analyzing nonracemic samples. Reagents n-hexane, IPA and ethanol of HPLC grade were purchased from Shanghai Lingfeng Chemical Reagent Co. Ltd, Shanghai, China. DEA, formic acid (FA) and other reagents of HPLC grade were obtained from Aladdin Reagent Co. Ltd, Shanghai, China. EZ, EZ-II, EZ-III and EZ-IV, purity >99.2%, were purchased from Nanjing Healthnice Pharmaceutical Co. Ltd, Nanjing, China. Instrumentation The HPLC experiments were performed on an HPLC system (SHI- MADZU, Kyoto, Japan) consisting of a solvent delivery pump Determination of optical isomer impurities in ezetimibe bulk drug Three batches of ezetimibe bulk drug samples (batch no , and , respectively) were analyzed according to the proposed conditions. The mobile phase was adopted as diluent to obtain final sample solutions, which contained 1mgmL 1 of ezetimibe. The content of each enantiomer was determined using external standard working solutions from pure reference compound run simultaneously. Results Stereochemistry for the crucial step of the EZ synthetic process has been specially checked as illustrated in Figure 2 (compounds from the proper synthetic route are framed). The starting material EZ-SM reacts with the desired intermediate EZ-IM(S,S) to generate EZ, in which process potential impurities of final EZ are taken into account. Except EZ as main product, EZ-III(S,S,R) is possible to be detected as by-product if stereoselectivity of the reaction is reversed. The forming of EZ-V(S,S,S) and EZ-VI (R,S,R) is very inconvenient and excluded from detection due to low occurrence probability. On the other hand, under certain conditions the undesired isomers of intermediate EZ-IM(S,S) would be formed (marked here as EZ-IM-II(R,S), EZ-IM-III(S,R) and EZ-IM-IV(R,R), respectively). Each isomer of EZ-IM would theoretically produce four isomers of EZ by analogy. However, considering the limited content of EZ-IM isomers as contaminants, only the main product of each isomer was given consideration to enantioseparation (marked as EZ-II(R,R,S), EZ-III(S,S,R) and EZ-IV(S,R,R), accordingly). Summarizing, the process-related impurities of EZ to be detected were concluded to be EZ-II(R,R,S), EZ-III(S,S,R) and EZ-IV(S,R,R). EZ-III(S,S,R) was not only the by-product resulting from stereoselectivity reversion, but also the process-related impurity originating from EZ-IM-III(S,R). EZ-II(R,R,S) and EZ-IV(S,R,R) were process-related impurities originating from EZ-IM isomers, EZ-II(R,R,S) and EZ-IV (S,R,R), respectively. Method development and optimization Several pretrials were conducted on five polysaccharide type chiral columns, namely Chiralpak AD-H, Chiralpak AS-H, Chiralcel OJ-H, Chiralpak AY-H and Chiralpak IC from Daicel. A mixture solution of ezetimibe (EZ) and its stereoisomers (EZ-II, EZ-III and EZ-IV) were used during the method development. n-hexane: IPA was used as the initial mobile phase according to the instructions, and then we varied the amount of IPA accordingly. The effects of different mobile phase compositions and different amounts of the additive were studied. However, the results suggested that separation cannot be reached on Chiralpak AS-H, Chiralcel OJ-H and Chiralpak AY-H columns. Separation on the Chiralpak AD-H column was observed at first, but fine resolution between EZ and EZ-II (R 12 ) could not be

3 Chiral Separation and Thermodynamic Investigation of Ezetimibe Optical Isomers 1491 Figure 2. Stereochemistry for the crucial step of the ezetimibe synthetic process. achieved under the chromatographic condition of n-hexane IPA DEA (80:20:0.1, v/v), which was considered to perform best among all conditions conducted. Online combination of Chiralpak AD-H and AS-H columns was also tried, which resulted in no better performance. Preliminary trials performed on the Chiralpak IC column were encouraging. Four stereoisomers can be clearly separated from each other. Replacement of isopropyl alcohol by ethanol in the mobile phase was tested but not adopted. Ethanol could marginally decrease the retention time but was accompanied by poor resolutions between the stereoisomers. Instead of trifluoroacetic acid (TFA) adopted in the literature (12), DEA was chosen as additive considering that the analytes are weakly alkaline. The introduced of DEA to the mobile phase has proved to improve the resolution. In the range 90/10/0.05 to 90/ 10/0.2 of n-hexane/ipa/dea as the mobile phase, resolution between EZ-II and EZ-III (R 23 ) and resolution between EZ-III and EZ-IV (R 34 ) increased whereas resolution between EZ and EZ-II (R 12 ) decreased. Therefore, we adopted 90/10/0.1 of n-hexane/ipa/dea to keep sound value of R 12, R 23 and R 34. In the range from 0.8 to 1.2 ml min 1 of the flow rate, the separation factors (α) were almost stable while all resolutions decreased. Considering the recommended pressure for prolonging column life, the flow rate was set at 1.0 ml min 1. A typical HPLC chromatogram is given in Figure 3, where all stereoisomers achieve baseline separation. The concentration adopted is 100 μgml 1 for EZ and 50 μgml 1 for others. Effect of IPA (%) in mobile phase The retention and selectivity vary accordingly when IPA content alters. The chromatographic data are shown in Table I. When we set 20, 15 and 10% of IPA in the mobile phase, there is no significant Figure 3. Typical HPLC chromatograms of ezetimibe spiked with stereoisomers. Column: Chiralpak IC (250 mm 4.6 mm I.D.); detection: UV at 256 nm; mobile phase: n-hexane IPA DEA (90:10:0.1, v/v); flow rate: 1 ml min 1 ; column temperature: 25 C. difference in separation factors (α). Nevertheless, when the solvent polarity increases, resolution (Rs) and retention (t R ) of all stereoisomers decreased. The effect of the content of IPA observed indicates that hydrogen bond interaction might not be pivotal to these stereoisomers for chiral recognition on the IC column. Discussion Effect of column temperature and thermodynamic parameters on enantioseparation To further optimize the chromatographic parameters, a temperature study was carried out. Considering the maximum usage temperature of the column (40 C), the effect of temperature was studied over a

4 1492 Sun et al. Table I. The Effect of IPA Content and Column Temperature on Retention and Enantioseparation of Ezetimibe Stereoisomers IPA (%) (v/v) Column temperature ( C) Retention parameters Capacity factor (k ) Separation factor (α) Resolution (Rs) k 1 k 2 k 3 k 4 α 12 α 23 α 34 R 12 R 23 R range of C (in 5 C increments). The results are shown in Table I. The Rs values increased with increasing temperature, which resulted from the reduced peak width and improved efficiency. The relationships between chromatographic data and the column temperature can be expressed by the Van t Hoff equation ln k 0 ¼ ΔHo RT þ ΔSo R þ ln Φ ¼ ΔHo RT þ ΔS ; ð1þ ln α ¼ ΔΔHo RT þ ΔΔSo R ; ð2þ where k is the retention factor, α is the separation factor, R is the gas constant, T is the temperature and Φ is the column phase ratio. ΔH and ΔS represent the differences in the enthalpy and entropy, respectively, when one enantiomer is transferred from the mobile phase to the stationary phase. ΔΔH and ΔΔS represent the differences of ΔH and ΔS, respectively, for a given pair of stereoisomers. If the Van t Hoff plots of ln k or ln α versus 1/T are linear in a temperature range, the corresponding thermodynamic parameters ΔH o, ΔS*, ΔΔH o and ΔΔS o, which may give explanation for the strengths of interactions between the stereoisomers and CSP, can be calculated from the slope and the intercept. In the plot of ln k versus 1/T, ΔH o = slope R and ΔS* = intercept R; in the plot of ln α versus 1/T, ΔΔH o = slope R and ΔΔS o = intercept R. The Van t Hoff plots of ln α versus 1/T for EZ and its three stereoisomers (II, III and IV) were proved to be linear in the temperature range of C. This indicates that the conformation of the CSP did not change obviously in the investigated range of temperatures, nor did retention mechanisms and selective interactions. Estimated parameters ΔΔH o and ΔΔS o along with regression coefficients (r 2 )are shown in Table II. For all compounds, ΔΔH o and ΔΔS o values were negative, which indicates that the separation was driven by enthalpy. Later-eluting stereoisomer had a stronger interaction with the CSP, and the molecular became more ordered when the complex formed. The separation factors (α) decreased with increasing temperature. As the absolute value of ΔΔH o was in the range of kj mol 1, we infer (16) that the forces vital to successful enantioseparation may consist of π π interactions, hydrogen bonding and steric hindrance. Moreover, π π interactions may be the vital interaction for the Table II. The Thermodynamic Parameters for the Enantioseparation of EZ Optical Isomers Compound EZ, EZ-II EZ-II, III EZ-III, IV ΔΔH o (kj mol 1 ) ΔΔS o (J mol 1 K 1 ) r conclusion obtained from the effect of the content of IPA. The π π interactions are apparently due to the close distance between aromatic rings in analytes and CSP. They also have many functional groups such as carbonyl, hydroxyl and amide, which can form hydrogen bonds. Steric hindrance could be induced by the half rigid tripod of the selector. It is remarkable that chiral recognition capacity of Chiralpak IC was high in comparison to other analogue toward EZ and its three stereoisomers (EZ-II, III, IV), which might be attributed to the different substituent groups in chiral selectors. More specifically, presence of electron withdrawing substituent chlorine at the 3 and 5 position of phenylcarbamate enabled the aromatic ring to be electron deficient. The π π stacking interaction between electron-rich aromatic ring in analytes (benzene ring with substituent hydroxyl comprising asymmetry carbon atom) and electron-deficient aromatic ring in CSP presumably appeared to be largely enhanced, which facilitated better resolution ability. Method validation The analytical method was validated according to International Conference on Harmonization (ICH) guidelines (17) with respect to specificity, linearity, stability, precision, accuracy and robustness. The selectivity of the analytical method was evaluated by the analysis of a solution containing EZ, EZ-II IV. The linearity of EZ and its three stereoisomers (EZ-II, III, IV) was evaluated by the analysis of six working sample solutions ranging from the LOQ to 120% of the permitted maximum level (0.1%), ranging approximately μgml 1. Reasonably good linearity patterns were observed with correlation coefficients between concentration and

5 Chiral Separation and Thermodynamic Investigation of Ezetimibe Optical Isomers 1493 Table III. The Linearity, LOD and LOQ Data Stereoisomer Range (μgml 1 ) Regression equation r 2 LOD (μg ml 1 ) LOQ (μgml 1 ) EZ A = 2.81e4c e 3 23e 3 EZ-II A = 2.79e4c e 3 24e 3 EZ-III A = 2.54e4c e 3 24e 3 EZ-IV A = 2.74e4 c e 3 18e 3 Figure 4. Typical HPLC chromatogram of optical isomer impurities in the ezetimibe bulk drug sample (1 mg ml 1 of ezetimibe). detector response (r 2 > 0.99). The LOD and LOQ were determined at a signal-to-noise ratio of 3 and 10, respectively. The linearity, LOD and LOQ data are listed in Table III. Stability of the sample solution was tested at 2 h intervals. The experimental data indicated that the solution was stable up to 8 h. The precision was tested by repeatability (intra-day) for the stereoisomer mixture at 1 μgml 1, which was analyzed in six replicates. Method precision had a relative standard deviation (RSD %) below 2.0% for repeatability (0.87% for retention time and 1.04% for peak area). Accuracy for the determination of stereoisomers was determined by preparing three drug substance samples at 80, 100 and 120 of the target concentration (1 μg ml 1 ). The apparent recovery ranged from to %. The overall percent recovery was (RSD% 1.4), (RSD% 1.7) and (RSD% 2.0) for EZ-II IV, respectively. The effects of flow rate (1.0 ± 0.1 ml min 1 ), column temperature (25 ± 2 C), amounts of the additive (0.1 ± 0.01) and mobile phase composition (changing the amounts of IPA by ±1%) on the resolution were assessed. Under all the above-mentioned varied conditions, the resolution between the four optical isomers was always greater than 2.0, illustrating the robustness of the method. Analytical application The applicability of the proposed method in chiral analysis was evaluated by the determination of enantiomers in ezetimibe bulk drugs. Figure 4 exhibits a typical chromatogram of the sample assay. Summarizing, EZ-III and IV cannot be detected in the samples whereas 0.06% of EZ-II was assayed in all three batches. All the analytical values fell within limit amount of 0.1% as required. Conclusion A simple and sensitive normal-phase enantioseparation HPLC method has been developed and validated for the quantitative determination of stereoisomers of ezetimibe on a Chiralpak IC column. The effects of quality and quantity of different polar organic modifiers, additives and flow rates on the enantioseparation were studied. The effect of the column temperature (15 35 C) on the separation implied that the enantioseparation of all compounds was enthalpy driven, and the separation factors decreased with the rise in temperature. Results indicate that π π interaction is the vital driving force that causes chiral recognition of ezetimibe stereoisomers on the Chiralpak IC column. References 1. Phan, B.A., Dayspring, T.D., Toth, P.P.; Ezetimibe therapy: mechanism of action and clinical update; Vascular Health and Risk Management, (2012); 8: Rosen, J.B., Jimenez, J.G., Pirags, V., Vides, H., Hanson, M.E., Massaad, R., et al.; A comparison of efficacy and safety of an ezetimibe/simvastatin combination compared with other intensified lipid-lowering treatment strategies in diabetic patients with symptomatic cardiovascular disease; Diabetes and Vascular Disease Research, (2013); 10(3): Kostapanos, M.S., Elisaf, M.S., Mikhailidis, D.P.; Ezetimibe a new approach in hypercholesterolemia management; Pharmacological Reports, (2012); 64(4): Rosenblum, S.B., Huynh, T., Afonso, A. Jr, Davis, H.R., Yumibe, N., Clader, J.W., et al.; Discovery of 1-(4-fluorophenyl)-(3R)-[3-(4-fluorophenyl)-(3S)-hydroxypropyl] -(4S)-(4-hydroxyphenyl)-2-azetidinone (SCH 58235): a designed, potent, orally active inhibitor of cholesterol absorption; Journal of Medicinal Chemistry, (1998); 41: Xiao, Y., Ng, S., Tan, T.T., Wang, Y.; Recent development of cyclodextrin chiral stationary phases and their applications in chromatography; Journal of Chromatography A, (2012); 1269: Peng, G.M., Wu, S.Q., Fang, Z.L., Zhang, W.G., Zhang, Z.B., Fan, J., et al.; Preparation and chiral separation of a novel immobilized cellulose-based chiral stationary phase in high-performance liquid chromatography; Journal of Chromatographic Science, (2012); 50(6): Zhang, T., Nguyen, D., Franco, P., Isobe, Y., Michishita, T., Murakami, T.; Cellulose tris(3,5-dichlorophenylcarbamate) immobilised on silica: a novel chiral stationary phase for resolution of enantiomers; Journal of Pharmaceutical and Biomedical Analysis, (2008); 46: Ashok, S., Varma, M.S., Swaminathan, S.; A validated LC method for the determination of the enantiomeric purity of aliskiren hemifumarate in bulk drug samples; Journal of Chromatographic Science, (2012); 50(9): Gajjar, A.K., Shah, V.D.; Isolation and structure elucidation of major alkaline degradant of Ezetimibe; Journal of Pharmaceutical and Biomedical Analysis, (2011); 55(1): Raman, B., Sharma, B.A., Butala, R., Ghugare, P.D., Kumar, A.; Structural elucidation of a process-related impurity in ezetimibe by LC/MS/MS and NMR; Journal of Pharmaceutical and Biomedical Analysis, (2010); 52 (1): Filip, K., Bankowski, K., Sidoryk, K., Zagrodzka, J., Laszcz, M., Trzcinska, K., et al.; Physicochemical characterization of ezetimibe and its impurities; Journal of Molecular Structure, (2011); 991: Radhakrishnanand, P., Subba Rao, D.V.; Avalidated LC method for the enantiomeric separation of Ezetimibe in bulk drug and pharmaceutical dosage forms; Indian Drugs, (2009); 46: Tiruvettipuram, K.T., Xiaoyong, F., Chou-Hong, T., Timothy, L.M., John, S.C., Cesar, C.; Process for the synthesis of azetidinones; (2001); Patent US Rao, R.N., Kumar, K.N., Kumar, B.S.; Liquid chromatographic separation and thermodynamic investigation of stereoisomers of darunavir on

6 1494 Sun et al. Chiralpak AD-H column; Journal of Separation Science, (2012); 35(20): He, X., Lin, R., He, H., Sun, M., Xiao, D.; Chiral separation of ketoprofen on a chirobiotic T column and its chiral recognition mechanisms; Chromatographia, (2012); 75(23 24): Küsters, E., Loux, V., Schmid, E.; Enantiomeric separation of chiral sulphoxides: screening of cellulose-based sorbents with particular reference to cellulose tribenzoate; Journal of Chromatography A, (1994); 666(1-2): Validation of analytical procedures: text and methodology, International Conference on Harmonization (ICH) Guidelines Q2 (R1) 2005.