Ac

Ma

b

a

ARRAA

KDCGC

1

aapDrcc

Acimpwiatm

TT

0d

Journal of Pharmaceutical and Biomedical Analysis 58 (2012) 27–33

Contents lists available at SciVerse ScienceDirect

Journal of Pharmaceutical and Biomedical Analysis

journa l homepage: www.e lsev ier .com/ locate / jpba

validated CE method for determining dimethylsulfate a carcinogen andhloroacetyl chloride a potential genotoxin at trace levels in drug substances

uzaffar Khana,∗, K. Jayasreea, K.V.S.R. Krishna Reddya, P.K. Dubeyb

Analytical Development, Aptuit Laurus Private Limited, ICICI Knowledge Park, Shameerpet, Hyderabad 500078, IndiaDepartment of Chemistry, Jawaharlal Nehru Technological University, Kukatpally, Hyderabad 500072, India

r t i c l e i n f o

rticle history:eceived 6 May 2011eceived in revised form 13 August 2011ccepted 18 September 2011vailable online 1 October 2011

a b s t r a c t

A simple and rapid capillary zone electrophoretic method was developed for determining dimethyl sul-fate a possible human carcinogen and mutagen and chloroacetyl chloride a potential genotoxic agent attrace levels in pharmaceutical drug substances by indirect photometric detection. A systematic screen-ing of various anionic probes was performed to obtain the best separation conditions and sensitivity.High sensitivities with low quantification and detection levels were achieved for dimethylsulfate and

eywords:imethyl sulfatehloroacetyl chlorideenotoxicapillary electrophoresis and validation

chloroacetyl chloride using a background electrolyte (BGE) containing 5 mM pyridine dicarboxylic acidas the probe ion. The method is specific, precise and accurate for the two genotoxins. The optimizedmethod was validated for specificity, precision, linearity, accuracy and stability in solution. Calibrationcurves were linear (R > 0.999) for both dimethylsulfate and chloroacetyl chloride in the range LOQ – 300%of nominal concentrations. The CE method was effectively implemented for estimating dimethylsulfateand chloroacetyl chloride in two different active pharmaceutical ingredients (APIs).

. Introduction

Dimethyl sulfate (DMS) is widely used in the pharmaceuticalnd chemical industries as a reagent for the methylation of phenols,mines and thiols. Compared to other methylating agents DMS isreferred by the industry because of its low cost and high reactivity.MS can affect the base-specific cleavage of guanine in DNA by

upturing the imidazole rings present in guanine [1]. This processan be used to determine base sequencing, cleavage on the DNAhain and other applications.

DMS is classified as a Class 2 carcinogen by Internationalgency for Research on Cancer [2] and is mutagenic, poisonous,orrosive, environmentally hazardous and volatile (presenting annhalation hazard). DMS is absorbed through the skin, mucous

embranes, and gastrointestinal tract. Delayed toxicity allowsotentially fatal exposures to occur prior to development of anyarning symptoms [3]. DMS has been tested for carcinogenicity

n rats by inhalation, subcutaneous and intravenous injection,

nd following prenatal exposure. It produced local sarcomas andumours of the nervous system [2]. DMS, in our case is used as a

ethylating agent in one of the manufacturing steps of an active

∗ Corresponding author at: Analytical Development, Aptuit Laurus Private Limited,urkapally, Shameerpet, Hyderabad, Andhra Pradesh 500078, India.el.: +91 40 230413531; fax: +91 40 23045438.

E-mail address: muzaffarkhan76@rediffmail.com (M. Khan).

731-7085/$ – see front matter © 2011 Elsevier B.V. All rights reserved.oi:10.1016/j.jpba.2011.09.019

© 2011 Elsevier B.V. All rights reserved.

pharmaceutical ingredient (API-1). With concerns rising due to theknown genotoxicity of DMS, it is necessary to demonstrate thatthe levels of DMS are within acceptable limits in API-1 through asuitable analytical method. Based on the daily dosage, a limit of50 ppm of DMS was considered in API-1 applying the threshold oftoxicological concern (TTC) concept.

Chloroacetyl chloride (CAC) is bifunctional; the acyl chlorideeasily forms esters and amides, while the other end of the moleculeis able to form other linkages, e.g. with amines. CAC is a knownreagent for acylation [4] and as a two carbon building block forcyclization reactions [5]. In our application it has been used foracylation followed by cyclization in an intermediate synthesis ofAPI-2. Although there is limited information available on the car-cinogenicity and genotoxicity of CAC, it can be categorized as astructural alert for genotoxic potential (Class 3 category as perMuller classification [10]). CAC decomposes on heating produc-ing toxic and corrosive fumes including phosgene and hydrogenchloride. A suitable analytical method should be developed todemonstrate that CAC is within acceptable limits of not more than75 ppm in API-2. This limit was arrived at, by considering the max-imum daily dosage of API-2 and TTC.

The current ICH guidelines describe a general concept of qual-ification of impurities in active substances (ICH Q3A (R) [6]) and

medicinal products (ICH Q3B (R2) [7]). These guidelines howeverdo not adequately address the concern for genotoxic impuritiesin drug substances and drug products. A Guideline on the Lim-its of Genotoxic Impurities [8] was subsequently issued by the

2 cal and

CcptgtottracipattrirfftoatofacisvfascsSrErtdap

ntpioishwos

cesmata

8 M. Khan et al. / Journal of Pharmaceuti

ommittee for Medicinal Products (CHMP) to overcome the shortomings of the ICH Q3A and Q3B guidelines. The CHMP guidelinerovides a general framework and practical risk based approacho deal with genotoxic impurities in drug substances. As per thisuideline, the genotoxic impurities with sufficient evidence for ahreshold-related mechanism should be addressed using methodsutlined in ICH Q3C (R3) for Class 2 solvents. For genotoxic impuri-ies without sufficient evidence for a threshold-related mechanism,he guideline proposes a policy of controlling levels to “as low aseasonably practicable” (ALARP principle). The CHMP guideline andmore recent draft guidance document on limits of genotoxic and

arcinogenic impurities [9] issued by the US Food and Drug Admin-stration recommend that an exposure level of 1.5 �g per personer day for each genotoxic impurity can be considered an accept-ble qualification threshold. Any impurity found at a level belowhis threshold generally should not need further safety qualifica-ion for genotoxicity and carcinogenicity concerns. The acceptableisk is defined as an additional cancer risk of not greater than 1n 100,000 based on a lifetime’s exposure to a genotoxic impu-ity. This approach is applied to impurities in the absence of datarom carcinogenicity long-term studies or data providing evidenceor a threshold mechanism of genotoxicity and is defined by thehreshold of toxicological concern (TTC) considering an exposuref 1.5 �g/day lifetime intake of a genotoxic impurity being associ-ted with an acceptable risk [10]. However the CHMP opines thathe genotoxic impurity limits can also vary based on differing peri-ds of exposure, while a 10-fold lower values are recommendedor high potency carcinogens such as N-nitrosoamines, aflatoxinsnd azoxy compounds the limits can be relatively relaxed for car-inogens with established dietary exposure and life threateningndication such as cancer. In accordance with this, Muller et al. [11]uggest a staged TTC; whereby the acceptable daily intake valuesary between 1.5 �g/day intake for lifetime exposure to 120 �g/dayor 28 days (or less) exposure [12]. Dobo et al. have discussed thepproach that can be made during drug development to under-tand potential mutagenic and carcinogenic risks associated withompounds used for synthesis and to understand the capability ofynthetic processes to control genotoxic impurities in the API [13].nodin provides a qualification strategy based on a review of rep-esentative compounds from structurally alerting substances [14].ven though structurally alerting compounds, particularly highlyeactive reagents introduced in the early stages of a multistep syn-hesis are unlikely to be carried over to the API, regulatory agenciesemand that carry over studies are performed to demonstrate thebsence of PGIs in the drug substance at TTC levels, considering theotential threat of genotoxins to human health.

GC–FID/MS and HPLC–UV/MS are the most widely applied tech-iques for determining genotoxic impurities at trace levels owingo their inherent high sensitivity and precision [15–17]. The samplereparation often involves extractions for enhancement of sensitiv-

ty and removal of matrix interferences. When the quantificationf a genotoxic impurity at trace levels in the API becomes daunt-ng, a higher limit for this impurity can be set in the intermediatetep where this genotoxic impurity is introduced. This limit shouldowever be supported with a scientific rationale that the impurityill either be structurally altered so that it is no longer genotoxic

r is eliminated through several steps of purification in subsequenttages of synthesis.

Capillary electrophoresis offers simplicity, high separation effi-iencies, low reagent consumption and is a cost effective andco-friendly technique, specifically for the analysis of ionizablepecies. However the application of CE is not rampant for deter-

ining trace analytes due to low sensitivity and poor precision

s compared to HPLC and GC. However in CE, the sensitivity ofrace analytes can be increased through sample stacking techniquesnd enrichment of the analytes from the sample matrix using solid

Biomedical Analysis 58 (2012) 27–33

phase extractions (SPE) and liquid–liquid extractions (LLE). The useof a MS detector coupled with CE can also enhance the sensitivityof trace analytes. The precision can be improved with the use of aninternal standard for quantification purposes.

Alzaga et al. [18] have developed a generic approach for thedetermination of alkylating agents by derivatisation followed byheadspace GC/MS. This method utilizes an in situ derivatisationprocedure with pentafluorothiophenol (PFTP) as the derivatisationagent. Methods for determining DMS by Head Space Gas Chro-matography (HS-GC) with MS detection [19] and HS-GC with FIDdetection [20] in intermediates and drug substances have beenreported. Hansen and Sheribah [21] have determined five residualalkylating impurities including alkyl chloride and alkyl bromide inbromazepam API using capillary electrophoresis with LOQ of 0.05%.To the best of our knowledge, no CE methods have been reported sofar in the literature for the determination of DMS and CAC. In thisarticle, we describe a simple and fast method for determining DMSand CAC in drug substances at trace levels by capillary electrophore-sis. This method has been successfully applied to demonstrate thatDMS and CAC in two different APIs were within acceptable regula-tory limits.

2. Materials and methods

2.1. Chemicals and reagents

Sodium hydroxide (0.1 N, CE grade) and CE grade water wereprocured from Agilent Technologies (Waldbronn, Germany). Pyri-dine 2,6-dicarboxilic acid was procured from Merck (Hohenbrunn,Germany), hexadecyl trimethyl ammonium bromide (CTAB),pyromellitic acid and paratoulene sulfonic acid were procured fromSigma–Aldrich (Steinheim, Germany). Benzoic acid and phthalicacid were procured from Rankem (New Delhi, India). DMS and CACwere procured from Spectrochem (Mumbai, India). All reagentswere of analytical grade or highest available purity. API-1 and API-2 were synthesized in Aptuit Laurus Private Limited (Hyderabad,India).

2.2. Equipment

The separation was performed on Agilent Technologies CapillaryElectrophoresis system with a built-in diode-array detector. TheAgilent ChemStation software was used for system control, dataacquisition and post-run processing. The separation was performedin a 64.5 cm (56 cm length to detector), 50-�m id, bare fused silicacapillary with extended light path having a bubble factor of 3 forenhanced sensitivity (Agilent Technologies, Waldbronn, Germany).An alignment interface, containing an optical slit matched to theinternal diameter of 150-�m, was used.

2.3. Preparation of analyte solutions and background electrolyte

About 30 mg each of DMS and CAC standards were accuratelyweighed into separate 100 mL volumetric flasks. The standardswere dissolved in 10 mL methanol by ultra sonication for 10 min.The flasks were cooled to room temperature and the volumes werethen made up to 100 mL mark with water. A 1.0 mL aliquot of thesestock solutions were further diluted independently to 100 mL each,with the diluent (methanol:water 10:90%, v/v) to obtain the DMSand CAC standard solutions. The standard solutions were filteredthrough 0.2 �m nylon syringe filters.

About 0.5 g of each API was accurately weighed into a 5 mL vol-umetric flask. A 0.5 mL aliquot of methanol was added to the flask

and sonicated for about 5 min. The flask was cooled to room tem-perature and volume was made up to 5 mL mark with the diluentand sonicated for 15 min with intermittent shaking. The samplesolutions were filtered through 0.2 �m nylon syringe filters.

al and

a11wt

2

p65dw2ar

wuaw

2

Ceva

2

apel

2

i1LttpDp

2

lastcT

2

s1tt

M. Khan et al. / Journal of Pharmaceutic

The background electrolyte (BGE) was prepared by dissolvingbout 83.0 mg of pyridine carboxylic acid and 14.5 mg of CTAB in00 mL HPCE grade water followed by ultra sonication for about0 min to facilitate complete dissolution. The pH of this solutionas then adjusted to 5.6 ± 0.1 with 1 N NaOH. The BGE was filtered

hrough 0.2 �m nylon syringe filters.

.4. Electrophoretic conditions

A 30 kV voltage was applied with negative polarity setting. Sam-les were injected hydrodynamically by pressure at 50 mbar for0 s followed by injection of a BGE plug by pressure at 50 mbar fors. The capillary cassette temperature was maintained at 20 ◦C. Theetector signals were monitored at a wavelength of 350 nm (band-idth 80 nm) with a reference wavelength of 200 nm (bandwidth

0 nm). The migration times for DMS and CAC were about 3.0 minnd 3.3 min respectively and the total run time was 5 min. The BGEeplenishment was done after every 25 injections.

New bare fused-silica capillaries were flushed with CE gradeater for 30 min following by the BGE for 30 min. Prior to everyse the capillary was conditioned by flushing for 10 min with waternd then with the BGE for 15 min. Between analyses, the capillaryas flushed with the BGE for 2 min.

.5. Method validation

The developed method was validated for determining DMS andAC with the assessment of specificity, precision, sensitivity, lin-arity and range, accuracy and stability of analyte solutions. Thealidation was performed keeping in mind the ICH guidelines fornalytical method validation [22].

.5.1. SpecificitySpecificity was evaluated by injecting DMS and CAC individu-

lly and spiked to the respective APIs along with all other knownrocess related impurities and solvents at specification levels. Thelectropherograms were examined for interferences of other ana-ytes with DMS and CAC in the specificity samples.

.5.2. PrecisionThe repeatability of the method was evaluated by separately

njecting replicate preparations (n = 6) of a spiked solution of API-containing DMS and spiked solution of API-2 containing CAC at

OQ levels, 50, 100, 150 and 300% of nominal analyte concentra-ions. To evaluate intermediate precision, six replicate samples ofhe APIs containing DMS and CAC at these concentrations were pre-ared and injected every day, on three different days. The %RSD forMS and CAC and their migration times were evaluated to assessrecision of the method.

.5.3. SensitivitySensitivity of the method was determined by establishing the

imit of detection (LOD) and limit of quantitation (LOQ) for DMSnd CAC. The detector response was obtained for a series of diluteolutions with known concentrations of DMS and CAC. Concentra-ions resulting in signal-to-noise ratios of about 3:1 and 10:1 wereonsidered as detection limits and quantitation limits respectively.he precision at LOQ level was assessed in terms of %RSD.

.5.4. Linearity and rangeThe linearity solutions were prepared from individual stock

olutions of DMS and CAC at nine concentration levels – LOQ, 50, 75,00, 125, 150, 175, 200 and 300% of analyte concentrations, each inriplicate. The data was subjected to linear regression analysis withhe least squares method.

Biomedical Analysis 58 (2012) 27–33 29

2.5.5. AccuracySamples of API-1 and API-2 were spiked with DMS and CAC

respectively at LOQ, 50, 100, 150 and 300% of the nominal ana-lyte concentrations. The spiking was performed in triplicate at eachlevel and the spiked samples were analyzed as per the method.Recoveries for DMS and CAC were calculated against freshly pre-pared standards. The mean percentage recoveries at each level wereused as a measure of accuracy of the method. An ANOVA test wasperformed to confirm that the recoveries were independent of thespiked concentrations.

2.5.6. Stability in analytical solutionStandard solutions of DMS and CAC were prepared in the diluent

at analyte concentrations. Each solution was analyzed immediatelyafter preparation and divided into two parts. While one part wasstored at 2–8 ◦C in a refrigerator, the other at bench top in tightlycapped volumetric flasks. The stored solutions were reanalyzedafter 24 and 48 h and the percentage recoveries of DMS and CACwere calculated against the zero hour samples.

3. Results and discussion

3.1. Method development and optimization

Dimethyl sulfate rapidly decomposes on contact with water tomethanol and methyl sulfate [23] as shown in Fig. 1a. Similarlychloroacetyl chloride reacts with water and the end products arechloroacetic acid and hydrochloric acid (Fig. 1b). Both DMS andCAC exist as uni-negative ions at the working pH of 5.6 and tend tomigrate towards the anode terminal under the influence of appliedelectrical field.

In the development and optimization trials, different anionicprobes were evaluated at 2, 5 and 10 mM concentrations for obtain-ing maximum sensitivity, peak symmetry and selectivity for DMSand CAC through indirect photometric detection. A sub-micellarconcentration of CTAB (0.4 mM) was used in all BGE systems forEOF reversal and separations were performed by applying potentialin negative polarity mode (detector end towards anode termi-nal). The peaks were found to be fronting in the BGE containing5 mM pyromellitate + 0.4 mM CTAB (pH 7.7 ± 0.1). While there waspoor separation in 5 mM p-toluene sulfonic acid + 0.4 mM CTAB(pH 6.0 ± 0.1), the DMS peak was found to be splitting in 5 mMbenzoic acid + 0.4 mM CTAB (pH 6.0 ± 0.1). Peak symmetries werepoor and a high level of background noise at the operationalwavelength was observed with 5 mM phthalate + 0.4 mM CTAB (pH6.5 ± 0.1).

In CE, the Kohlraush regulating function determines theprobe displacement by the analyte [24] and the probe’s mobil-ity and optical properties must be considered. The mobility andconcentration of the probe are crucial for the separation per-formance of the method because they influence peak shapesand efficiency. The mobilities of the probes that were evaluatedare in the order: pyromellitate > phthalate > pyridine dicarboxy-late > benzoate > p-toluene sulfonate [25]. While the highly mobilepyromellitate probe is suitable for analyzing smaller fast mov-ing anions, the benzoate and p-toluene sulfonate probes aremore suitable for analyzing the lower mobility compounds suchas short chain (C4–C8) carboxylic acids. Phthalate and pyri-dine dicarboxylate probes have similar mobilities and are mostsuitable for analyzing medium mobility species. The absorp-tivity of the probe at the detection wavelength is a key

parameter influencing the method sensitivity. The benzoateand pyridine dicarboxylate probes have higher molar absorp-tivities when compared to the other probes that were tested[26].

30 M. Khan et al. / Journal of Pharmaceutical and Biomedical Analysis 58 (2012) 27–33

OS

OO

O OS

OO

O

H2O + CH3OH+

Cl

O

ClCl

O

O+ H2O2 2 + 2 HCl

a

b

ter. (b

2tdodohs[b2e

thiapuwctataaaicknTwe

3

3

rfDpt

aIarbi

The accuracy of a method expresses the closeness between thetheoretical value and the determined value and was tested in twodifferent ways. The mean recoveries for DMS and CAC at LOQ levelwere 97.4% and 92.8% respectively. At other concentration levels

Table 1Precision results.

DMS (%RSD) CAC (%RSD)

Repeatabilitya Inter-dayprecisionb

Repeatabilitya Inter-dayprecisionb

LOQ 3.8 5.3 2.4 6.250% 4.2 6.5 3.7 2.8100% 2.2 4.2 2.6 3.5

Fig. 1. (a) Reactivity of dimethylsulfate in wa

The medium mobility and high molar absorptivity of pyridine,6-dicarboxylate rendered it as the preferred probe for analyzinghe species DMS and CAC. BGEs containing 2–10 mM pyridine 2,6-icarboxylic acid + 0.4 mM CTAB (pH 5.6 ± 0.1) were evaluated forptimum separation and sensitivity. Lower concentrations of pyri-ine 2,6-dicarboxylic acid were unfavorable as broad peaks werebserved due to electromigration dispersion. On the other handigher concentrations resulted in higher noise levels, decreasedensitivity and an adverse affect on the linearity of detection27]. Good peak symmetries, high sensitivity and high resolutionetween DMS and CAC could be achieved using 5 mM pyridine,6-dicarboxylic acid + 0.4 mM CTAB in the BGE system which wasventually finalized.

The separation was evaluated at 20, 25 and 30 kV and the migra-ion times decreased with the increase in the applied voltage. Theighest available voltage on the equipment (30 kV) was chosen as

t provided higher theoretical plates and resolution since the sep-ration proceeds rapidly minimizing the effects of diffusion andeak broadening. The temperature effects on separation were eval-ated at 15, 20, 25, 30 and 35 ◦C. The migration times decreasedith the increase in temperature due to a decrease in BGE vis-

osity. However the temperature was optimized to 20 ◦C as theheoretical plates decreased and the noise increased significantlyt elevated temperatures. In indirect photometric detection, theransparent analyte species are detected as negative peaks against

high background absorbance of the probe ions. By choosingppropriate sample wavelength (where the sample has minimumbsorbance) and reference wavelength (where the probe has max-mum absorbance) positive signals can be obtained. Using theapability of the Diode Array Detector (DAD) and with the knownnowledge of probe’s absorbance, several combinations of the sig-al (sample wavelength) and reference channels were evaluated.he highest sensitivities (signal-to-noise) ratios were obtainedith the signal acquired at 350 nm (bandwidth 80 nm) at a ref-

rence wavelength of 200 nm (bandwidth 20 nm).

.2. Method validation

.2.1. SpecificitySpecificity is the ability of the method to measure the analyte

esponse in the presence of its potential impurities and other inter-erences. No interferences were observed at the migration times ofMS and CAC in the API samples that were spiked with all otherrocess related impurities and residual solvents. A specimen elec-ropherogram of the DMS and CAC standards is presented in Fig. 2a.

The residual acetic acid, a process related solvent was well sep-rated from DMS (Fig. 2b) in the real time sample analysis of API-1.n API-2, a major peak due to chloride (formed as a byproduct and

lso introduced from the manufacturing process) was also wellesolved from the CAC peak (Fig. 2c). Thus the method was found toe specific for determining DMS and CAC in presence of potential

nterferences.

) Reactivity of chloroacetyl chloride in water.

3.2.2. PrecisionThe repeatability and inter-day precision of the method was

determined in terms of %RSD for migration times and recoveriesof DMS and CAC in the spiked samples. The overall RSD of migra-tion times was not more than 2.2% for DMS and not more than 2.0%for CAC. Similarly the overall RSD of recoveries of DMS and CACwere not more than 6.5% and 6.2% respectively (Table 1).

3.2.3. SensitivityThe LOD and LOQ were found to be 0.3 �g/mL and 1.0 �g/mL

respectively for both DMS and CAC corresponding to 3 ppm and10 ppm with respect to the sample concentrations. These resultsemphasize that the method is sensitive enough for determiningDMS and CAC at trace levels in real time samples considering thepermissible levels of these toxic impurities (Fig. 3).

3.2.4. Linearity and rangeThe detector response linearity to varying analyte concentra-

tions was established by analyzing standard solutions at ninedifferent concentrations ranging from LOQ to 300% of nominalanalyte concentration. Linearity curves (Area vs Conc.) were plot-ted for DMS and CAC and the data was subjected to regressionanalysis. Linear relationships confirm that the test results aredirectly proportional to the concentrations. The linear equationof regression for DMS was y = 9.9398x − 0.8980 with a correla-tion coefficient (R) of 0.9990. Similarly the regression equation forCAC was y = 15.9874x − 0.9366 with a correlation coefficient (R) of0.9992.

The range of a method is the interval in which it has a suitablelevel of precision, accuracy and linearity. From the results of val-idation tests that were performed, the range for this method wasLOQ to 300% of the nominal analyte concentration.

3.2.5. Accuracy

150% 3.7 4.9 4.7 4.1300% 2.5 3.3 2.9 5.3

a n = 6 determinations.b n = 6 determinations each on three different days.

M. Khan et al. / Journal of Pharmaceutical and Biomedical Analysis 58 (2012) 27–33 31

min0.5 1 1.5 2 2.5 3 3.5 4 4.5

mAU

-30

-25

-20

-15

-10

-5

0

5Chloroacetyl chloride

Chloride

min0.5 1 1.5 2 2.5 3 3.5 4 4.5

mAU

-30

-25

-20

-15

-10 Dimethyl sulfate

Acetic acid

min0.5 1 1.5 2 2.5 3 3.5 4 4.5

mAU

-30

-25

-20

-15

-10

-5Dimethyl sulfate

Chloroacetyl chloride

a

b

c

Fig. 2. Specimen electropherograms of (a) DMS and CAC standards, (b) API-1 spiked with dimethylsulfate and (c) API-2 spiked with chloroacetyl chloride.

min0.5 1 1.5 2 2.5 3 3.5 4 4.5

mAU

-35

-30

-25

-20

-15

-10

Blank solution

LOQ Solution

LOD Solution

Dimethylsulfate Chloroacetyl chloride

Fig. 3. Overlaid electropherograms of Blank, LOD and LOQ solutions.

32 M. Khan et al. / Journal of Pharmaceutical and Biomedical Analysis 58 (2012) 27–33

Table 2Accuracy results.

Accuracy level Dimethyl sulfate spiked to API-1a Chloroacetyl chloride spiked to API-2a

Amt. spiked(ng)

Amt. recovered(ng)

%Meanrecovery ± SD

Amt. spiked(ng)

Amt. recovered(ng)

%Meanrecovery ± SD

LOQ 1007.8 981.8 97.4 ± 7.87 981.2 910.9 92.8 ± 2.5450% 1511.7 1441.1 95.3 ± 7.07 1471.8 1379.9 93.8 ± 2.69100% 3023.4 2929.2 96.9 ± 6.36 2943.4 2793.4 94.9 ± 3.52150% 4535.1 4434.3 97.8 ± 2.07300% 9070.2 8726.8 96.2 ± 3.60

a n = 3 determinations.

Table 3Content of DMS in API-1 and CAC in API-2.

Dimethyl sulfate Chloroacetyl chloride

Lot # of API-1 DMS contenta (ppm)(limit = 50 ppm)

Lot # of API-2 CAC contenta (ppm)(limit = 75 ppm)

001 <10 001 17002 11 002 ND003 15 003 ND004 10 004 13005 <10 005 ND

N

t9

eavtcfiad

3

esae

3

1ect

4

gdosThTisTG

[

[

[

[

[

[

[

006 ND 006 22

D, not detected.a n = 3 determinations.

he recoveries for DMS and CAC ranged between 95.3–97.8% and0.7–97.0% respectively (Table 2).

The linearity of the method in estimating the recoveries wasvaluated by performing a one-way hierarchical analysis of vari-tion (ANOVA) within the range to separate out the systematicariability due to the sample preparation, injection and integra-ion. The recoveries were found to be independent of the spikedoncentrations (ANOVA, p > 0.05) for both the target analytes. Thesendings suggest that the described method represents a valuablend accurate tool for the analysis of DMS and CAC in pharmaceuticalrug substances.

.2.6. Stability in analytical solutionNo remarkable variations were observed in percentage recov-

ries of DMS and CAC in the initial samples and samples that weretored at 2–8 ◦C and at room temperature for 24 and 48 h. Thus thenalytes were found to be stable for at least 48 h when stored atither 2–8 ◦C or at room temperature.

.2.7. Analysis of real time samplesThe developed method was applied for determining DMS in API-

and CAC in API-2. Six consecutive commercial scale batches ofach API were analyzed in triplicate. The results concluded that theontent of DMS in API-1 and the content of CAC in API-2 were withinhe acceptable and safe limits as presented in Table 3.

. Conclusions

A CE method for determining DMS and CAC, two potentiallyenotoxic impurities in drug substances by indirect photometricetection was developed and validated. The method is sensitive,ffers simplicity and does not include laborious steps of derivati-ation as is usually the case for compounds lacking chromophores.he method is very specific, cost effective and eco-friendly; andas been successfully applied for real time sample analyses.his method has been extended for determining other genotoxic

mpurities such as diethylsulfate and diisopropylsulfate in drugubstances, and the validation of this application is under study.hough genotoxins are mostly determined using HPLC–UV/MS andS–MS techniques for their proven ruggedness and sensitivity, the

[

4415.4 4003.0 90.7 ± 1.368830.8 8567.4 97.0 ± 1.68

use of Capillary electrophoresis for estimating genotoxins can com-plement the existing approach for analysis and also open newhorizons.

Acknowledgements

The authors wish to thank the management of Aptuit Laurus Pvt.Ltd. for supporting this work. The authors also wish to acknowledgethe encouragement provided by Dr. Satyanarayana C. and Dr. SrihariRaju K., Aptuit Laurus Pvt. Ltd. throughout this work.

References

[1] A. Streitwieser, C.H. Heathcock, E.M. Kosower, Introduction to Organic Chem-istry, Prentice-Hall Inc., 1992, p. 1169.

[2] IARC monographs, Dimethylsulfate, 71 (1999) 575–588.[3] J.C.R. Rippey, M.I. Stallwood, Nine cases of accidental exposure to dimethyl

sulphate—a potential chemical weapon, Emerg. Med. J. 22 (2005) 878–879.[4] M.M. Krayushkin, V.N. Yarovenko, S.L. Semenov, I.V. Zavarzin, A.V. Ignatenko,

A.Y. Martynkin, B.M. Uzhinov, Synthesis of photochromic 1,2-dihetaryletheneusing regioselective acylation of thienopyrroles, Org. Lett. 4 (2002) 3879–3881.

[5] G.R. Brown, A.J. Foubister, Unambiguous synthesis of 3-aryloxymethyl- mor-pholine hydrochlorides without ring enlargement side reactions, J. Chem. Soc.,Perkin Trans. 1 (1989) 1401–1403.

[6] ICH Q3A (R), Impurities in New Drug Substances, February 2002, http://www.ICH.org/.

[7] ICH Q3B (R2), Impurities in New Drug Products, July 2006, http://www.ICH.org/.[8] Guideline on the Limits of Genotoxic Impurities, Committee for Medici-

nal Products (CHMP), European Medicines Agency, London, 28 June 2006(CPMP/SWP/5199/02, EMEA/CHMP/QWP/251344/2006).

[9] Guidance for Industry, Genotoxic and Carcinogenic Impurities in Drug Sub-stances and Products: Recommended Approaches, US Department of Healthand Human Services, Food and Drug Administration, Centre for Drug Evaluationand Research (CDER), 2008, December.

10] R. Kroes, A.G. Renwick, M. Cheeseman, J. Kleiner, I. Mangelsdorf, A. Piersma, B.Schilter, J. Sclatter, F. van Schothorst, J.G. Vos, G. Wurtzen, Food Chem. Toxicol.42 (2004) 65–83.

11] L. Muller, R.J. Mauthe, C.M. Riley, M.M. Andino, D. de Antonis, C. Beels, J.DeGeorge, A.G.M. De Knaep, D. Ellison, J.A. Fagerland, R. Frank, B. Fritschel,S. Galloway, E. Harpur, C.D.N. Humfrey, A.S. Jacks, N. Jagota, J. Mackinnon,G. Mohan, D.K. Ness, M.R. O’Donovan, M.D. Smith, G. Vudathala, L. Yotti, Arationale for determining, testing, and controlling specific impurities in phar-maceuticals that possess potential for genotoxicity, Regul. Toxicol. Pharmacol.44 (2006) 198–211.

12] D.P. Elder, A. Teasdale, A.M. Lipczynski, Control and analysis of alkyl esters ofalkyl and aryl sulfonic acids in novel active pharmaceutical ingredients (APIs),J. Pharm. Biomed. Anal. 46 (2008) 1–8.

13] K.L. Dobo, N. Greene, M.O. Cyr, S. Caron, W.W. Ku, The application of structure-based assessment to support safety and chemistry diligence to managegenotoxic impurities in active pharmaceutical ingredients during drug devel-opment, Regul. Toxicol. Pharmacol. 44 (2006) 282–293.

14] D.J. Snodin, Genotoxic impurities: from structural alerts to qualification, Org.Process Res. Dev. 14 (2010) 960–976.

15] Q. Yang, B.P. Haney, A. Vaux, D.A. Riley, L. Heidrich, P. He, P. Mason, A. Tehim,L.E. Fisher, H. Maag, N.G. Anderson, Controlling the genotoxins ethyl chlorideand methyl chloride formed during the preparation of amine hydrochloridesalts from solutions of ethanol and methanol, Org. Process Res. Dev. 13 (2009)786–791.

16] D.P. Elder, A.M. Lipczynski, A. Teasdale, Control and analysis of alkyl and benzylhalides and other related reactive organohalides as potential genotoxic impu-

rities in active pharmaceutical ingredients (APIs), J. Pharm. Biomed. Anal. 48(2008) 497–507.

17] D.P. Elder, D. Snodin, A. Teasdale, Analytical approaches for the detection ofepoxides and hydroperoxides in active pharmaceutical ingredients, drug prod-ucts and herbals, J. Pharm. Biomed. Anal. 51 (2010) 1015–1023.

al and

[

[

[

[

[

[

[

[

M. Khan et al. / Journal of Pharmaceutic

18] R. Alzaga, R.W. Ryan, K. Taylor-Worth, A.M. Lipczynski, R. Szucs, P. Sandra,A generic approach for the determination of residues of alkylating agentsin active pharmaceutical ingredients by in situ derivatization–headspace-gas chromatography–mass spectrometry, J. Pharm. Biomed. Anal. 45 (2007)472–479.

19] J. Zheng, W.A. Pritts, S. Zhang, S. Wittenberger, Determination of low ppmlevels of dimethyl sulfate in an aqueous soluble API intermediate usingliquid–liquid extraction and GC–MS, J. Pharm. Biomed. Anal. 50 (2009)1054–1059.

20] G.F. Deng, T.W. Yao, Determination of dimethyl sulphate residual in granisetron

hydrochloride by headspace gas chromatography, Zhejiang Da Xue Xue Bao YiXue Ban 37 (2008) 156–158.

21] S.H. Hansen, Z.A. Sheribah, Comparison of CZE, MEKC, MEEKC and non-aqueouscapillary electrophoresis for the determination of impurities in bromazepam,J. Pharm. Biomed. Anal. 39 (2005) 322–327.

[

[

Biomedical Analysis 58 (2012) 27–33 33

22] ICH Q2R1, Validation of Analytical Procedures: Text and Methodology, 1995,June, http://www.ICH.org/.

23] B.H. Mathison, M.L. Taylor, M.S. Bogdanffy, Dimethyl sulfate uptake and methy-lation of DNA in rat respiratory tissues following acute inhalation, Fundam.Appl. Toxicol. 28 (1995) 255–263.

24] F. Foret, L. Krivánková, P. Bocek, Capillary Zone Electrophoresis, VCH Publishers,New York, 1993, pp. 27–35.

25] P. Doble, M. Macka, P.R. Haddad, Design of background electrolytes for indirectdetection of anions by capillary electrophoresis, Trends Analyt. Chem. 19 (2000)10–17.

26] M. Macka, C. Johns, P. Doble, P.R. Haddad, K.D. Altria, Indirect photometricdetection in CE using buffered electrolytes. Part I. Principles, LCGC 19 (January)(2001), www.chromatographyonline.com.

27] X. Xu, Th.W. Kok, H. Poppe, Noise and baseline disturbances in indirect UVdetection in capillary electrophoresis, J. Chromatogr. A 786 (1997) 333–345.