Development and validation of an ion-exchange chromatography method for heparin and its impurities in heparin products.

www.jss-journal.com

Title:

Page 1

Journal of Separation Science

Development and validation of an ion-exchange chromatography method for heparin and its impurities in heparin products

Author names: Sumate Thiangthuma,b, Yvan Vander Heydenc, Wolfgang Buchbergerd, Johan Viaenec, Brompoj Prutthiwanasana, Leena Suntornsuka,e* a

Department of Pharmaceutical Chemistry, Faculty of Pharmacy, Mahidol

University, Bangkok, Thailand b

Bureau of Drug and Narcotic, Department of Medical Sciences, Ministry of

Public Health, Nontaburi, Thailand c

Department of Analytical Chemistry and Pharmaceutical Technology, Center

for Pharmaceutical Research (CePhaR), Vrije Universiteit Brussel (VUB), Brussel, Belgium d

Institute of Analytical Chemistry, Johannes Kepler University Linz, Austria

e

Center of Excellence for Innovation in Drug Design and Discovery, Faculty

of Pharmacy, Mahidol University, Bangkok, Thailand Running title: Anion-exchange chromatography of heparin and its impurities Corresponding author:

Leena Suntornsuk Department of Pharmaceutical Chemistry, Faculty of Pharmacy, Mahidol University 447 Sri-Ayudhaya Rd., Rajathevee, Bangkok, 10400, Thailand Tel: 662 644 8695

Fax: 662 644 8695

E-mail: [email protected] Received: 30-Mar-2014; Revised: 03-Aug-2014; Accepted: 11-Aug-2014

This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1002/jssc.201400348.

This article is protected by copyright. All rights reserved.

www.jss-journal.com

Page 2


Keywords: anion-exchange chromatography, dermatan sulfate, experimental design and response-surface methodology, heparin, oversulfated chondroitin sulfate, Abstract An anion-exchange liquid chromatography method for the determination of heparin and its impurities (dermatan sulfate and oversulfated chondroitin sulfate) was developed using chemometric assisted optimization, including multivariate experimental design and response surface methodology. The separation of heparin, dermatan sulfate and oversulfated chondroitin sulfate (Rs above 2.0) was achieved on a Dionex RF IC IonPac AS22 column with a gradient elution of 10 to 70% of 2.5 M sodium chloride and 20 mM Tris phosphate buffer (pH 2.1) at a flow rate of 0.6 mL/min and UV detection was at 215 nm. Method validation shows good linearity (r > 0.99), acceptable precision (%relative standard deviations < 11.4%) and trueness (%recovery of 92.3–103.9%) for all analytes. The limits of detection for dermatan sulfate and oversulfated chondroitin sulfate are equivalent to 0.11% w/w (10.5 μg/mL) and 0.07% w/w (7.2 μg/mL), while the limits of quantification are 0.32% w/w (31.5 μg/mL) and 0.22% w/w (22.0 μg/mL) relative to heparin, respectively. The method is specific for heparin, dermatan sulfate and oversulfated chondroitin sulfate without interference from mobile phase and sample matrices and could be used for accurate quantitation the drug and its impurities in a single run. Applications of the method reveal contents of heparin between 90.3–97.8%. Dermatan sulfate and oversulfated chondroitin sulfate were not detected in any of the real-life samples. 1 Introduction Heparin is a highly sulfated glycosaminoglycan (GAG) with the highest negative charges density of any known biological molecule [1]. The basic structure of heparin is comprised of repeated disaccharide subunits of uronic acid and (1,4)-D-glucosamine. Heparin is widely used as an anticoagulant and anti-thrombotic agent [2,3] preventing the formation This article is protected by copyright. All rights reserved.

www.jss-journal.com

Page 3


of clots and prolonging the clotting time of blood. Pharmaceutical-grade heparin is derived from mucosal tissue of slaughtered animals, such as porcine intestine or bovine lung [4]. Since heparin is purified from animal sources, formulations of the processed heparin may contain low levels of several natural contaminants (e.g. chondroitin sulfate A, B, or C), that are not associated with adverse health effects [3]. However, problems arose in 2007–2008 when contaminated batches of heparin led to serious allergic reactions and the death of more than 100 patients from anaphylactic shock after being administered the drug [http://www.separationsnow.com/coi/cda/detail]. Collaborative studies, involving researchers from the United States Food and Drug Administration (US-FDA), industry and academia, identified oversulfated chondroitin sulfate (OSCS) as a non-native heparin contaminant [5,6]. OSCS was found in crude heparin, even after purification processes, and remained in the active pharmaceutical ingredient and final formulations. Presently, OSCS is not allowed in heparin raw material and preparations, because of its serious side effects. In addition, samples of heparin were found to contain more than 1% w/w of chondroitin sulfate B or dermatan sulfate (DS), which is a marker of poor-quality purification of the crude material. Analysis of heparin and its impurities is extremely difficult because of their highly negative charges, polydispersity, and macroheterogeneity. The assay for heparin sodium in the USP monograph recommends measurement of heparin by anti-clotting activity [7–12]. The assay cannot detect heparin impurities, including OSCS and DS; therefore, USP has continuously revised the method for identifying the impurities since 2009 [7–12]. USP 32 [7] described NMR spectroscopy and CE techniques for identification of OSCS and DS. Subsequently, the CE method was replaced with chromatographic identity using a strong anion-exchange chromatographic (SAX-HPLC) method [8–12] since SAX-HPLC provides higher efficiency with improved resolution. Identification of heparin, DS and OSCS by NMR spectroscopy is done by comparison of ratios of the chemical shifts of the analytes, whereas This article is protected by copyright. All rights reserved.

www.jss-journal.com

Page 4


in SAX-HPLC, resolution values between heparin, DS and OSCS peaks are determined. Proof of the absence of OSCS by both methods must be confirmed. Additionally, measurement of DS and other galactosamine containing impurities by HPLC with amperometric detection is described in the heparin monograph under the impurities section [12]. The total hexosamine, including DS, OSCS and other impurities, should be less than 1%. Various publications also reported the analysis of heparin and its impurities by NMR spectroscopy, chromatography and CE [13–23]. SAX-HPLC provided LODs in the range of 0.02–0.03% w/w with recoveries (R) between 82.9 and 111.0% for OSCS [13,14], while the LOD and R were 0.1% w/w and 96.0–103.0%, respectively, for DS [13]. Keire et al. [14] reported that the CE method was not feasible for the quantitation of OSCS, whereas Volpi et al. [20] proposed a CE method that required sample pre-treatment, including acidic hydrolysis and derivatization, prior to analysis. Although no baseline-separation of OSCS and heparin was fully obtained, Wielgos et al. [21] presented a CE method for the determination of heparin impurities (i.e. OSCS, DS) providing LODs of 0.1 and 0.5% for OSCS and DS, respectively. The purpose of our research was to develop a simpler, more accurate and sensitive HPLC method for the simultaneous quantitation of heparin and its two major impurities, DS and OSCS. To achieve the goal, chemometrics-based techniques, including multivariate experimental design and response surface methodology were utilized for method optimization [24,25]. The use of this chemometric approach provides the experimental conditions giving the best or optimal analytical response after executing a limited number of experiments. Finally, the developed method was validated and applied to the analysis of raw materials and heparin products commercially available in Thailand. 2. Materials and methods This article is protected by copyright. All rights reserved.

www.jss-journal.com

Page 5


2.1 Chemicals and reagents Heparin sodium salt, from porcine intestinal mucosa, and DS were purchased from Sigma (St. Louis, MO, USA). OSCS was purchased from SERVA electrophoresis (Heidelberg, Germany). One batch of heparin raw material and six batches of heparin products, available in Thailand, were purchased from local manufacturers and drugstores. Acetonitrile and benzyl alcohol were from Fisher Scientific (Leicestershire, UK). Sodium dihydrogen phosphate, tris(hydroxymethyl)aminomethane (Tris), phosphoric acid, sodium chloride and sodium perchlorate were purchased from Merck (Darmstadt, Germany). 2.2 Instrumentation HPLC analyses were performed on a Dionex 680 HPLC system from Thermo Scientific Dionex (Idstein, Germany) equipped with a Dionex 680 Pump, a Dionex ASI-100 autosampler and a Dionex PDA-100 photo diode array detector. Chromeleon® software was used for data acquisition and processing. The chromatographic separations were carried out on a Dionex RF IC IonPac AS22 (250 x 4.0 mm i.d.) column (Thermo Scientific Dionex, Idstein, Germany). The AS22 stationary phase characteristics are that it consists of beads with diameter of 6.5 μm and 2000 Å particle pore size. The column was selected since it was an SAX column containing alkanol quaternary ammonium ions [26]. It was suitable for separation of heparin and its impurities, which are highly negatively charged polymers, due to the ionized sulfate and carboxylate groups in their molecules [11, 13, 14]. Other columns in the same series (e.g. IonPac AS11 and AS24) can also be used [27,28], but AS22 was employed due to its availability in this laboratory. 2.3 Standard and sample preparations A stock solution containing 40 mg/mL heparin, 2 mg/mL OSCS, and 1 mg/mL DS was prepared by dissolving accurately weighed amounts of the standards in Milli-Q water. Working solutions were prepared by diluting the stock solution with water to obtain This article is protected by copyright. All rights reserved.

www.jss-journal.com

Page 6


concentrations 3000, 5000, 10 000, 15 000 and 20 000 μg/mL for heparin; 90, 100, 120, 140 and 160 μg/mL for OSCS and 200, 250, 300, 350 and 400 μg/mL for DS. Sample blank was 5% benzyl alcohol in water since it is a commonly used preservative in heparin injection products. For sample preparation, 540 μL of heparin sodium injection (5000 IU/mL) was transferred into a 1.5 mL vial and diluted to volume with water. The final concentration was 10 mg/mL (180 IU = 1 mg). 2.4 HPLC optimization and experimental-design methodology HPLC conditions for the separation of heparin DS and OSCS were modified from that of the USP [11] and literature [13], subsequent optimization was performed by varying types of salts (NaClO4, LiCl and NaCl) and organic solvent types (ACN and MeOH). Experimental design, i.e. a central composite design (CCD) combined with multivariate linear regression (MLR) to model the response [29–31] was employed to further optimize the HPLC conditions. The effects of buffer pH and salt concentrations on resolution (Rs), retention time (tR) and peak width at half height (w0.5) were investigated and illustrated on response surface plots. The conditions from the model was finally optimized by modifications of the gradients in order to reduce analysis time. The resolution was calculated from the following equation according to USP [11]; RS = 2(tR2 – tR1)/(w1 + w2)

(1)

where tR2 and tR1 are the retention times of the two components; and w1 and w2 are the corresponding widths at the bases of the peaks obtained by extrapolating the relatively straight sides of the peaks to the baseline. Briefly, the mobile phase was Milli-Q water (A) and 20 mM Tris phosphate buffer (pH 2.1) containing 2.5 M sodium chloride (B). The buffer was adjusted to the required pH by addition of phosphoric acid. The gradient was 0–2 min 90% A with 10% B, linear gradient This article is protected by copyright. All rights reserved.

www.jss-journal.com

Page 7


to 70% B at 26 min, hold until 31 min, then linear gradient to 10% B with 90% A at 32 min and hold until 40 min before the next injection was done. A flow rate was set at 0.6 mL/min and a detection wavelength was at 215 nm. This wavelength was selected because most studies on heparin analysis, which use a UV detector, recommend the detection wavelengths of pure heparin in unfractionated form in the range of 200–220 nm [11, 13, 14 , 32–35]. These low wavelengths provided high S/Ns with low %RSDs [13]. 2.5 Method validation The optimized method was validated by a standard procedure to ensure adequate linearity,

LOD,

LOQ,

precision, accuracy,

and specificity,

following the

ICH

recommendations [36]. 2.5.1 Linearity Method linearity was determined by spiking the analytes into the sample blank, while system linearity was done in pure standard solution. Linearity was studied from two calibration sets. The first was the calibration set for heparin standard solutions in a range of 3000–25 000 μg/mL. The second was the standard mixtures of OSCS (90–160 μg/mL) and DS (200–400 μg/mL) containing fixed amounts of heparin (10 000 μg/mL). Each solution was prepared and analyzed in triplicate. According to ICH guideline [36], the linearity should be evaluated by visual inspection of a plot of signals as a function of analyte concentration or content. If there is a linear relationship, test results should be evaluated by appropriate statistical methods, for example, by calculation of a regression line (e.g. correlation coefficient, y intercept and slope of the regression line) by the method of least squares. Currently, linear regression and correlation coefficients (r), were calculated by Excel(R). Linearity of the calibration curve was initially visualized and lack of fit in linear


www.jss-journal.com

Page 8


regression for all analytes was investigated and F values were calculated. In addition, response factors were determined and their plots were established.

2.5.2 LOD and LOQ LOD and LOQ were estimated using the ICH guideline recommendations [36], which can be determined by several approaches. Among them is an approach based on the SD of the response and the slope using the following equations; LOD = 3.3 σ

(2)

LOQ = 10 σ S

(3)

S

where σ was the SD of the response and S was the slope of the calibration curve. σ may be estimated from SDs of y intercepts of regression lines [36]. 2.5.3 Precision Precision was evaluated in spiked standard solutions, at the lowest, middle and highest concentrations of the calibration curves, into the sample blank. The analysis was daily performed in triplicate and repeated for five days. Precision of the method was expressed as percent relative standard derivations (%RSDs) of repeatability (df = 10) and time-different intermediate (df = 3) precision. The repeatability (sr2) and the time-different intermediate precision (si(t)2) were estimated at each concentration level from an analysis of variance (ANOVA) table and using the follows equations: si(t)2 = sr2 + sb2

(4)

where sb2 represents the between-days variance. sb2 = (MSbetween – MSwithin)/n

(5)


www.jss-journal.com

Page 9


MSbetween and MSwithin are between-days and within-day mean squares, respectively, obtained from the ANOVA table and n is number of replicates within one day (n = 3). The repeatability is estimated from the replicated measures within one day (df = 10).

2.5.4 Trueness Trueness of the method was evaluated by the standard addition method and expressed as percent recoveries. Standard addition solutions were prepared by spiking the sample blank with heparin, OSCS and DS at similar concentrations of the calibration curves (i.e. 3000–25 000 µg/mL for heparin; 200–400 µg/mL for DS and 90–160 µg/mL for OSCS). The solutions were prepared and analyzed in triplicate. 2.5.5 Specificity Specificity of the method was tested by comparison of chromatograms from different solutions under the optimum HPLC condition. These solutions included mobile phase at the initial composition, sample blank, standard solutions of heparin (8 mg/mL), OSCS (5 mg/mL), DS (5 mg/mL), a mixture of heparin, OSCS and DS (10, 0.12 and 0.3 mg/mL, respectively) and a heparin commercial sample (10 mg/mL). The ability to separate all exicipients from heparin was assessed by their resolution between the peaks. Peaks of the analytes were identified by retention times and peak purity was verified with the UV spectra from the PDA detector. 3 Results and Discussion 3.1. HPLC optimization 3.1.1 Preliminary optimization Initial optimization for heparin and OSCS separation (DS was not included at this stage) was performed by the modified HPLC conditions from USP [11], at the Center for Pharmaceutical Research (CePhaR), Vrije Universiteit Brussel, where sodium perchlorate This article is protected by copyright. All rights reserved.

www.jss-journal.com

Page 10


was available. The conditions were a 30 min gradient elution of 2.8–12.6% of 2.5 M NaClO4 in 20 mM NaH2PO4 buffer (pH 3.0) using a flow rate of 0.8 mL/min, an injection volume of 40 µL and a detection wavelength at 215 nm. This wavelength was suggested in the literatures [11, 13, 14, 32–35] to enhance the signal sensitivity and precision [13]. The UV spectra of DS, heparin and OSCS (Fig S1) were in good agreement with those in the literature [11, 13, 14 , 32–35]. The USP conditions provided the separation of heparin and OSCS with a resolution (Rs) of 2.5, but the peaks were broad and asymmetric. The poor peak shapes would make accurate quantification difficult. Subsequently, the USP conditions were disregarded due to the inaccessibility of sodium perchlorate in Thailand. According to the Thai Ministry of Interior Announcement: Safety Precautions on Working with Hazardous/Toxic Materials [37], sodium perchlorate is classified as a hazardous chemical that needs to be controlled under The Arms Control Act B.E. 2530 (1987) [38]. Purchasing of such chemicals has to be done only through the Ministry of Military [39]. Thus, lithium chloride or sodium chloride were considered as alternatives using the modified HPLC conditions from Trehy et al. [1]. Both lithium chloride or sodium chloride did not offer satisfactory separation for DS, heparin and OSCS. However, sodium chloride provided a better separation with some baseline drift, which was less severe than that from lithium chloride (Fig S2). A 40 min gradient elution of 5–60% of 2.5 M NaCl in 20 mM Tris phosphate buffer (pH 3.0) with a flow rate of 0.8 mL/min provided sharper peaks of heparin and OSCS with an Rs of 1.8 in 20 min. Thus, sodium chloride was selected for further optimization. The flow rate of 0.8 mL/min was reduced to 0.6 mL/min was also employed to increase the resolution. Organic solvent is usually added to the mobile phases to improve the separation efficiency in HPLC. Presently, low amounts of acetonitrile (less than 20%) enhanced the analyte signals, but a large peak of benzyl alcohol (a preservative in heparin injection This article is protected by copyright. All rights reserved.

www.jss-journal.com

Page 11


samples) was observed. The peak was broad from 4.0 to 8.0 min and overlapped with the heparin peak (Fig S3). Methanol caused too high of a back pressure in the system and could not be used. Consequently, addition of organic modifiers did not improve the separation and were excluded from the mobile phase. 3.1.2 Experimental design and final optimization So far, the HPLC conditions, using a 40 min gradient elution of 5–60% of 2.5 M NaCl in 20 mM Tris phosphate buffer (pH 3.0) with a flow rate of 0.6 mL/min, could provide some separation of heparin, DS and OSCS, but the resolution of DS/OSCS was unsatisfactory (Rs = 1.11). Thus, experimental design and response-surface methodology was employed to improve the separation of the analytes. A central composite design (CCD) was used to study the effects of pH of the mobile phase and sodium chloride concentration, and afterwards to define the optimal chromatographic conditions. Each factor was examined at five levels (–√2, –1, 0, +1, +√2). Sodium chloride concentrations were varied in the range of 0.8 to 2.5 M, and pH of mobile phase was adjusted in the range of 2.0 to 4.0, while the flow rate, injection volume and detection wavelength were fixed at 0.6 mL/min, 20 µL and 215 nm, respectively, in all experiments. Nine experiments were performed (n = 3) in a randomized sequence to minimize the influences of uncontrolled factors (e.g. temperature) that may introduce bias in retention times (tR), Rs and peak widths at half height (w0.5). Results are reported in Table 1. Experiment 9 (pH = 3.0, NaCl = 1.5 M) offered a good resolution for both DS/heparin (Rs = 1.28) and heparin/OSCS (Rs = 1.37), whereas experiment 4 (pH = 3.7, NaCl = 2.2 M) provided the shortest analysis time (< 15 min) and the smallest peak width at half height but heparin and dermatan sulfate were co-eluting. The second-order polynomial models with the y intercepts, coefficients and correlation coefficients of each model for the resolution, retention time and peak width at half height are shown in Table 2. For the polynomial models, the three-dimensional response surfaces of the response were plotted in order to This article is protected by copyright. All rights reserved.

www.jss-journal.com

Page 12


easily and precisely evaluate the chromatographic behavior of the investigated variables (Fig. 1 and Fig S4). Increasing sodium chloride concentrations in the mobile phase led to resolution enhancement of DS/heparin and heparin/OSCS and reduction of retention times and peak width at half heights of all investigated compounds. Increasing the pH also reduced the retention times limitedly, while its effect on resolution depended on the sodium chloride concentration. A clear two-factor interaction effect between pH and sodium chloride concentration can be observed, but their effects on the separation were neither straightforward nor uniform (Table 2). Based on the second-order polynomial models, fine tune optimization of the HPLC conditions was done by modification of the gradient in order to shorten the gradient run time. The condition that provided acceptable separation performances and reasonable retention times were a 26 min gradient of 10–70% of 2.5 M sodium chloride containing 20 mM Tris phosphate buffer (pH 2.1) using a flow rate of 0.6 ml/min, injection volume of 20 µL and a detection wavelength at 215 nm. The conditions provided a resolution of 2.05 for DS/heparin and of 2.70 for heparin/OSCS with retention times of 21.1, 25.2 and 30.6 min for DS, heparin and OSCS, respectively. Fig. 2A represents a typical chromatogram of DS (0.3 mg/mL), heparin (20 mg/mL) and OSCS (0.12 mg/mL) obtained from the optimized condition. These concentrations, corresponding to 1.5% DS and 0.6% OSCS in heparin, were presented to mimic a real situation, where heparin exists in a much higher amount than those of the contaminants [13,32]. The broad peak shape is likely due to the polydispersity of heparin [40] since heparin is a complex mixture with carboxylic groups. The molecule is a linear unbranched heteropolysaccharide chain with two types of sulfated disaccharide residues linked by α-1,4-glycosidic linkages. 3.2. Method validation 3.2.1. Linearity This article is protected by copyright. All rights reserved.

www.jss-journal.com

Page 13


The calibration curves (plots of concentrations and peak areas) of heparin, DS, and OSCS were established by triplicate injections of five different concentrations of the working standard solutions. Linear regression shows correlation coefficients of higher than 0.99 for all analytes (Table 3). Additionally, ratios of slopes between regression from method and system linearity was 0.97, 0.96 and 0.90 for heparin, DS and OSCS, respectively. The values indicated that the excipient in heparin injection samples (5% benzyl alcohol) did not interfere in the analysis and did not cause any signal enhancement or suppression, since the ratios were close to 1.0. Visual evaluation of the calibration curves showed that they were straight and statistical. Results from a lack-of-fit (Fcal) test were 0.18 for heparin, 0.02 for DS and 0.24 for OCSC. These values were smaller than the critical values: for heparin and DS, Ftab(=0.05; df1=3, df2=20)

= 3.10, and for OSCS, Ftab(=0.05; df1=3, df2=10) = 3.71. Martinez et al. [41] inferred that

testing for lack of fit of the experimental points to the regression line is an important step in linear regression. In case that lack of fit exists, SDs for both regression line coefficients are overestimated and can give rise to too large confidence intervals. F values from calculation in this work were smaller than critical values in most cases indicating that the regression was sufficient to relate the test results (peak areas) to the analyte concentrations. Additionally, response factor plots of heparin, DS and OSCS were established (Fig S5). The response factors for heparin, DS and OSCS were 12.8, 7.0 and 15.0 with the %RSDs between 3.96 and 7.69%. 3.2.2. LOD and LOQ LOD and LOQ were determined from 3.3 and 10 times, respectively, the SDs of responses [Eqs. (2) and (3)]. LODs for heparin, DS and OSCS were 800, 10.5 and 7.2 μg/mL and LOQs were 2,500, 31.5 and 22.0 μg/mL, respectively, with %RSD of less than 10.12% at LOQ levels (Table 3). The LODs for DS and OSCS were equivalent to 0.11 and 0.07%, while This article is protected by copyright. All rights reserved.

www.jss-journal.com

Page 14


the LOQs were 0.32 and 0.22 % w/w relative to heparin. The LOD of heparin from this work (800 µg/mL) was lower than that reported by Somsen et al. [32] (1100 µg /mL). It is very likely that heparin should be detectable at concentration lower than this, but most of our optimization was done for signal enhancement of DS and OSCS, not heparin. Other published papers by SAX-HPLC method reported the LOD or LOQ of OSCS and DS, but not those of heparin [13, 14]. The present LODs and LOQ are better or at least comparable to published values (Table 3). The LODs were low enough for the detection, while the estimated LOQs were precise and accurate for the quantitation of the impurities in real samples. 3.2.3. Precision Precision was tested in the sample blank spiked with standard solutions and precision data were shown as %RSDs of repeatability (within-day precision) and time-different intermediate precision. The %RSDs of repeatability were below 1.46, 4.61 and 3.38%, for heparin, DS and OSCS, respectively (Table 4). Time-different intermediate precision showed %RSDs below 3.59, 6.38 and 11.42% for heparin, DS and OSCS, respectively (Table 4). The values were comparable with reported data [13,14]. 3.2.4. Trueness Trueness of the method was evaluated from the sample blank spiked with standard solutions and results were expressed as %recovery. Percent recoveries were between 95.2– 103.6 (%RDS < 3.55%), 92.3–103.9 (%RDS < 5.32%) and 92.6–102.0% (%RDS < 3.13%), for heparin, DS and OSCS, respectively (Table 4). Percent recoveries and their %RSDs were satisfactory and comparable to reported values, indicating that the method is accurate for quantitation of heparin, DS and OSCS. 3.2.5. Specificity Specificity is a validation parameter that determines whether the developed method is fit for the intended purpose without interferences from mobile phase compositions, sample This article is protected by copyright. All rights reserved.

www.jss-journal.com

Page 15


matrices, expected contaminants, etc. Presently, mobile phase, sample blank, standard solutions of heparin, OSCS, DS, a mixture of all analytes and a commercial heparin sample, were studied. Resolution of heparin and the impurities was greater than 2.05 and no interfering peaks were observed in the chromatograms (Fig. 2B). Retention times of the analyte peaks in standard and sample solutions and their UV spectra were identical (Fig S1). Thus, the method was specific and could be utilized for analysis of heparin, DS and OSCS in samples. From the above data, the proposed method was valid and accurate for the determination of heparin and its impurities, DS and OSCS. LODs and LOQs for DS and OSCS were comparable to reported values (0.07–0.32 vs. 0.03–0.8%) [13,14]. Recoveries for the impurities of the current work were better than in the literature (92.3–103.9 vs. 96.0– 111.0 [13] and 82.9–96.0% [14]). SAX-HPLC normally provided superior analytical performance characteristics to CE and NMR techniques [32,14] because of its higher sensitivity and better reproducibility. 3.3. Application The developed and validated method was applied for the determination of heparin and its impurities in one heparin raw material sample and six heparin injection products available in Thailand. The results revealed that percent label amounts of heparin in the seven samples were between 90.3 and 97.8% with %RSDs < 3.40%. The impurities, DS and OSCS, were detected neither in the raw material nor in the heparin injection products. 4. Concluding Remarks An anion-exchange LC method has been established for the analysis of heparin and its impurities, OSCS and DS. Response-surface methodology based on central-composite design (CCD) results was used to study the effects of the important factors, pH and concentration of sodium chloride in the mobile phase, and to determine the optimal chromatographic This article is protected by copyright. All rights reserved.

www.jss-journal.com

Page 16


conditions. The simple gradient elution by 2.5 M sodium chloride in 20 mM Tris phosphate buffer (pH 2.1) was efficient for the separation of the analytes. All analytical performance characteristics meet the requirements and the method is suitable for the intended purposes. The LODs at below 0.1% w/w relative to heparin could be achieved. In conclusion, this method is economical, fast, simple and reliable for the identification and quantification of heparin and its impurities for the QC of both heparin raw material and formulations. Acknowledgements Financial support from the Thailand Research Fund through the Royal Golden Jubilee Ph.D. Program (Grant No. PHD/0285/2549) to Sumate Thiangthum and Leena Suntornsuk and from the Office of the High Education Commission and Mahidol University under the National Research Universities Initiative are gratefully acknowledged. References [1] Liu, H., Zhang, Z., Linhardt, R.J., Nat. Prod. Rep. 2009, 26, 313–321. [2] Lepor, N.E. Rev. Cardiovas. Med. 2007, 8, S9-S17. [3] Fischer, K.G. Hemodial. Int. 2007, 11, 178–189. [4] Linhardt, R.J., Gunay, N.S., Sem. Thromb. Hem. 1999, 3, 5–16. [5] Guerrini, M., Beccati, D., Shriver, Z., Naggi, A., Viswanathan, K., Bisio, A., et al., Nat. Biotechnol. 2008, 26, 669–675. .Author: et al. is not allowed. Please list all author names [6] Kishimoto, T.K., Viswanathan, K., Ganguly, T., Elankumaran, S., Smith, S., Pelzer, K., N. Engl. J. Med. 2008, 358, 2457–2467. [7] The United States Pharmacopeial Convention, The United States Pharmacopeia 32 The National Formulary 27, Rockville 2009, pp. 2552–2554. [8] The United States Pharmacopeial Convention, The United States Pharmacopeia 33 The National Formulary 28, Rockville 2010, pp. R915-R920. This article is protected by copyright. All rights reserved.

www.jss-journal.com

Page 17


[9] The United States Pharmacopeial Convention, The United States Pharmacopeia 34 The National Formulary 29, Rockville 2011, pp. 3036–3040.

[10] The United States Pharmacopeial Convention, The United States Pharmacopeia 35 The National Formulary 30, Rockville 2012, pp. 3403–3407. [11] The United States Pharmacopeial Convention, The United States Pharmacopeia 36 The National Formulary 31, Rockville 2013, pp. 2552–2554. [12] The United States Pharmacopeial Convention, The United States Pharmacopeia 37 The National Formulary 32, Rockville 2014, pp. 3222–3228. [13] Trehy, M.L., Reepmeyer, J.C., Kolinski, R.E., Westenberger, B.J., Buhse, L.F., J. Pharm. Biomed. Anal. 2009, 49, 670–673. [14] Keire, D.A., Trehy, M.L., Reepmeyer, J.C., Kolinski, R.E., Ye, W., Dunn, J., Westenberger, B.J., Buhse, L.F., J. Pharm. Biomed. Anal. 2010, 51, 921–926. [15] Beyer, T., Diehl, B., Randel, G., Humpfer, E., Schafer, H., Spraul, M., Schollmayer, C., Holzgrabe, U., J. Pharm. Biomed. Anal. 2008, 48, 13–19. [16] Griffin, C.C., Linhardt, R.J., Van Gorp, C.L., Toida, T. R., Hileman, E., Schubert, R.L., Brown, S.E., Carbohydr. Res. 1995, 27, 183–197. [17] Imanari, T., Toida T., Koshiishi, I., Toyoda, H., J. Chromatogr. A. 1996, 720, 275–293. [18] Patel, R.P., Narkowicz, C., Hutchinson, J.P., Hilder, E.F., Jacobson, G.A., J. Pharm. Biomed. Anal. 2008, 46, 30–35. [19] Turnbull, J.E., Meth. Mol. Biol. 2001, 171, 141–147. [20] Volpi, N., Maccari, F., Linhardt, R.J., Anal. Biochem. 2009, 388, 140–145. [21] Wielgos, T., Havel, K., Ivanova, N., Weinberger, R., J. Pharm. Biomed. Anal. 2009, 49, 319–326.


www.jss-journal.com

Page 18


[22] Wang, B., Buhse, L.F., Al-Hakim, A., Boyne Ii, M.T., Keire, D.A., J. Pharm. Biomed. Anal. 2012, 67–68, 42–50. [23] Korir, A.K., Larive, C.K., Anal. Bioanal. Chem. 2009, 393, 155–169. [24] Perrin, C., Vargas, M.G., Vander Heyden, Y., Maftouh, M., Massart, D.L., J. Chromatogr. A 2000, 883, 249–265 [25] Dejaegher, B., Vander Heyden, Y., J. Pharm. Biomed. Anal. 2011, 56, 141–158. [26] Product manual for the IonPac AG 22 guard column and IonPacAS 22 analytical column, Document no. 065119, Dionex Cooperation 2008, Revision 4, September 2008. [27] Product manual for the IonPac AG 11 guard column and IonPacAS 11 analytical column, Document no. 034791, Dionex Cooperation 2008, Revision 12, 7 April 2009. [28] Product manual for the IonPac AG 22 guard column and IonPacAS 22 analytical column, Document no. 065119, Dionex Cooperation 2008, Revision 4, September 2008 [29] Dejaegher, B., Vander Heyden, Y., LC–GC Eur. 2009, 22, 256–261. [30] Dejaegher, B., Vander Heyden, Y., LC–GC Eur. 2009, 22, 583–585. [31] Vander Heyden, Y., LC–GC Eur. 2006, 19, 469–475. [32] Somsen, G.W., Tak, Y.H., Torano, S.J., Jongen, P.M.J.M., de Jong, G. J., Chromatogr. A 2009, 1216, 4107–4112. [33] Lima, M.A., Rudd, T.R., de Farias, E.H.C., Ebner, L.F., Gesteira, T.F., de Souza, L.M., et al., PLoSOne, 2011, 6, 1–8. .Author: et al. is not allowed. Please list all author names [34] Nieduszynzki, I. A., Atkin, E.D.T., Biochem. J. 1973, 135, 729–733. [35] Kobyakov, V.V., Ovsepyan, A.M., Frolova, N.A., et al Khim.-farm.Zh, 1987, 7, 130– 134.Author: et al. is not allowed. Please list all author names [36] European Medicines Agency, ICH Topic Q 2 (R1) Validation of Analytical Procedures, Text and Methodology, London, 1995. This article is protected by copyright. All rights reserved.

www.jss-journal.com

Page 19


[37] Thai Ministry of Interior Announcement: Safety Precautions on Working with Hazardous/Toxic Materials, B.E. 2543 (in Thai). [38] The Arms Control Act B.E. 2530 (in Thai). [39] Thai Ministry of Military Announcement: Trading of Hazardous/Toxic Materials, B.E. 2556 (in Thai). [40] Beni, S., Limtiaco, J.F.K., Larive, C.K., Anal, Bioanal. Chem., 2011, 399, 527–539. [41] Martinez, A., Riu, J., Rius, F.X. Chemo. Intel. Lab. Sys., 2000, 54, 61–73.


www.jss-journal.com

Page 20


Fig. 1: 3-D plots of the response surface for resolution (Rs), retention time (tR), and peak width at half height (w0.5) of heparin, originating from the CCD experiments.


www.jss-journal.com

Page 21


Fig. 2: Chromatograms of A) a standard mixture containing heparin (20 mg/mL), DS (0.3 mg/mL) and OSCS (0.12 mg/mL) and B) (A) mobile phase (starting composition), (B) sample blank (5% benzyl alcohol in water), (C) DS solution (3 mg/mL), (D) mixture of DS (5 mg/mL) and OSCS (5 mg/mL), (E) heparin solution (8 mg/mL), (F) mixture of DS (0.3 mg/mL), heparin (10 mg/mL), and OSCS 0.12 mg/mL), and (G) sample solution (heparin sodium injection: 10 mg/mL heparin). HPLC conditions: gradient elution of 10–70% of 2.5 M sodium chloride with 20 mM Tris phosphate buffer (pH 2.1) in 26 min on a Dionex RF IC IonPac AS22 column at a flow rate of 0.6 mL/min, injection volume of 20 μL and UV detection at 215 nm.


www.jss-journal.com

Page 22


Table 1: Two-factor central composite design and the experimental results No.

Factor pH

Conc

Rs

Rs

Response

NaCl(M)

DS/Hep

Hep/OSCS

tR (min)

w0.5 (min)

DS

Hep

OSCS

DS

Hep

OSCS

1

2.3

0.8

0.00

1.12

27.41

27.41

29.83

1.80

1.80

1.06

2

2.3

2.2

1.45

1.08

13.57

16.04

17.48

0.94

1.02

0.52

3

3.7

0.8

0.92

0.00

27.57

29.88

29.88

0.60

1.44

1.44

4

3.7

2.2

0.00

1.27

13.53

13.53

14.78

0.87

0.87

0.48

5

2.0

1.5

1.54

1.04

17.35

21.52

23.57

1.41

1.57

0.72

6

4.0

1.5

0.00

1.37

17.47

17.47

19.38

1.24

1.24

0.46

7

3.0

0.8

0.00

1.00

27.42

27.42

30.04

2.10

2.10

0.69

8

3.0

2.5

0.82

1.03

12.50

13.67

14.85

0.30

0.73

0.26

9

3.0

1.5

1.28

1.37

17.45

19.96

22.17

1.01

1.25

0.89

Rs = resolution, tR = retention time, w0.5 = peak width at half height, DS = dermatan sulfate, Hep = heparin, OSCS = oversulfated chondroitin sulfate


www.jss-journal.com

Page 23


Table 2: Estimated coefficients and correlation coefficients for regression models on the results from the two-factor central composite design Model

Rs

tR (min)

Coefficient DS/Hep Hep/OSCS

w0.5 (min)

DS

Hep

OSCS

DS

Hep

OSCS

b0

1.075

1.463

17.47

19.57

22.12

1.312

1.402

0.709

b1

-0.338

-0.059

0.04

-0.72

-1.07

-0.188

-0.121

-0.003

b2

0.279

0.259

-7.13

-6.76

-6.83

-0.356

-0.421

-0.311

b11

-0.134

-0.153

0.07

0.07

-0.32

-0.025

-0.030

0.007

b22

-0.389

-0.399

2.77

1.85

1.18

-0.169

-0.026

0.031

b12

-0.591

0.326

-0.05

-1.24

-0.69

0.283

0.054

-0.106

r

0.905

0.879

0.999

0.991

0.998

0.828

0.946

0.815

b0 = intercept, b1 = coefficient for pH, b2 = coefficient for NaCl concentration, b12 = twofactor interact coefficient, b11 and b22 = quadratic terms, Rs = resolution, tR = retention time, w0.5 = peak width at half height, DS = dermatan sulfate, Hep = heparin OSCS = oversulfated chondroitin sulfate


www.jss-journal.com

Page 24


Table 3 Linearity, LOD and LOQ data Analyte

Range

Method linearity

System linearity

(μg/mL) Heparin

Current work

3,000-25,000 y = 13071.8x -1689 (r = 0.9997)

DS

OSCS

a

200-400

90-160

LOD (µg/mL)

y = 13487.6x – 4280 800

Reported value

Current work

Reported value

1,100b

2,500

-

(r = 0.9997)

y = 18.6x – 2835

y = 19.4x – 266

(r = 0.9985)

(r = 0.9999)

y = 73.9 x – 5383

y = 81.6x – 6182

(r = 0.9981)

(r = 0.9998)

(%RSD = 3.22) 10.5 (0.11%)

b

c

260 , 23

31.5 (0.32%)

157c

(%RSD = 6.37) 7.2 (0.07%)

19b, 6.2c, 4.3d

22 (0.22%)

23c, 16.1d

(%RSD = 10.12)

DS = dermatan sulfate, OSCS = oversulfated chondroitin sulfate, LOD = limit of detection, LOQ = limit of quantitation, %RSD = percent

relative standard deviation at LOQs, numbers in parentheses are %w/w relative to heparin b

c

LOQ (µg/mL, n = 3)

by CE method [32]

by SAX-HPLC method [13]

d

by SAX-HPLC method [14]


www.jss-journal.com

Page 25


Table 4: Precision and trueness dataa Analyte

Concentration (μg/mL)

Repeatability (%RSD, df = 10)

Heparin

3,000 5,000 10,000 15,000 20,000

DS

200 250 300 350 400

% Recovery (n = 3)

1.46 0.55 0.80

Time-different intermediate precision (%RSD, df = 3) 3.37 3.59 3.20

96.2-103.6 (3.55) 98.3-102.3 (1.60) 97.6-103.3 (2.62) 95.2-100.6 (2.01) 95.3-100.7 (2.16)

-

4.61 2.09 3.56

6.38 5.95 3.35

93.6-103.9 (5.32) 97.0-102.0 (2.73) 96.3-99.4 (1.65) 95.1-101.0 (3.21) 92.3-101.1 (4.60)

95.6-103.0 (0.6-32.4) b

25

Reported value

www.jss-journal.com

OSCS

Page 26

90 100 120 140 160

3.38 2.53 2.32


11.42 5.80 4.21

a

99.1-100.5 (0.70) 99.9-102.0 (1.11) 98.8-101.1 (1.36) 97.2-100.1 91.54) 92.6-98.1 (3.13)

94.5-110.8 (0.9-27.5)b 88.4-96.1 (0.7-5.9)c 70.8-78.0 (9.4-18.6)d

DS = dermatan sulfate, OSCS = oversulfated chondroitin sulfate, %RSD = percent relative standard deviation, df = degree of freedom Trueness is represented as %recovery, DS = dermatan sulfate, OSCS = oversulfated chondroitin sulfate, numbers in parenthesis represent % RSDs b by SAX-HPLC method [13] c by SAX-HPLC method [14] d by NMR method [14]

26

Development and validation of a simple and sensitive size-exclusion chromatography method for quantitative determination of heparin in pharmaceuticals.

Isolation and characterization of sulfated disaccharides from the deamination products of porcine heparin (alpha-heparin) and whale heparin (omega-heparin) and a comparison of the deamination products.

An improved method for bonding heparin to intravascular cannulae.

A new method for the determination of N-sulfate in heparin and its analogs.

A catalytic role for heparin. Evidence for a ternary complex of heparin cofactor thrombin and heparin.

Development and Validation of an Affinity Chromatography-Protein G Method for IgG Quantification.

Quality by design (QbD) based development and validation of an HPLC method for amiodarone hydrochloride and its impurities in the drug substance.

Isolation, Identification, and Characterisation of Degradation Products and the Development and Validation of a Stability-Indicating Method for the Estimation of Impurities in the Tolterodine Tartrate Formulation.

Method to detect contaminants in heparin using radical depolymerization and liquid chromatography-mass spectrometry.

Validation of an alternative microbiological method for tissue products.

Heparin cofactor activity measured with an amidolytic method.

MS method for simultaneous determination of three phenol impurities in ritonavir.

Development and Validation of a Stability-Indicating RP-UPLC Method for the Estimation of Impurities in Cinacalcet Hydrochloride API and its Formulation.

Development and validation of a reversed-phase HPLC method for simultaneous analysis of butylhydroxyanisol, simvastatin and its impurities in tablet dosage forms.

Anticoagulant activity of heparin: separation of high-activity and low-activity heparin species by affinity chromatography on immobilized antithrombin.

Development and validation of an ultra-performance liquid chromatography method for simultaneous analysis of 20 antihistaminics in dietary supplements.

Pharmacokinetics of heparin and low molecular weight heparin.

Quality by Design approach in the development of a solvent-modified micellar electrokinetic chromatography method: finding the design space for the determination of amitriptyline and its impurities.

Pharmacokinetic modeling of heparin and its clinical implications.

Development of a photoreactive probe-based system for detecting heparin.

Characterization of process-related impurities including forced degradation products of alogliptin benzoate and the development of the corresponding reversed-phase high-performance liquid chromatography method.

MS method for determination of p-phenylenediamine and its metabolites in blood samples.

A conductimetric method for the determination of sulphate and carboxyl groups in heparin and other mucopolysaccharides.

Development and validation of a new stability indicating reversed phase liquid chromatographic method for the determination of prednisolone acetate and impurities in an ophthalmic suspension.