jou rn al hom epage: www.elsev ier .com/ locate /chroma
etermination of (R)-timolol in (S)-timolol maleate activeharmaceutical ingredient: Validation of a new supercritical fluidhromatography method with an established normal phase liquidhromatography method
drian Marleya,b, Damian Connollyc,∗
Allergan Pharmaceuticals, Westport, Mayo, IrelandIrish Separation Science Cluster (ISSC), National Centre for Sensor Research, Dublin City University, Glasnevin, Dublin 9, IrelandPharmaceutical and Molecular Biotechnology Research Centre (PMBRC), Department of Chemical and Life Sciences, Waterford Institute of Technology,aterford, Ireland
r t i c l e i n f o
rticle history:eceived 26 July 2013eceived in revised form 27 October 2013ccepted 4 December 2013vailable online 14 December 2013
An enantioselective supercritical fluid chromatography (SFC) method was developed and validated tomeet the current European Pharmacopoeia requirements of a limit test for the determination of S-timololmaleate enantiomeric purity in timolol maleate drug substance. The developed method is presented asan alternative to the current normal phase high performance liquid chromatography (NP-HPLC) methoddescribed in the European Pharmacopoeia (Timolol Maleate Monograph). Using a 4.6 mm × 250 mm Chi-ralcel OD-H (dp: 5 �m) column and a mobile phase of (93:7) CO2/0.1% (v/v) TEA in MeOH delivered at4.0 mL min−1 resolution of 2.0 was achieved within 5 min, representing a 3-fold reduction in run-timeand an 11-fold reduction in solvent consumption relative to the NP-HPLC method. Method robustnesswas examined by the variation of flow rate (±0.5 mL min−1), column temperature (±5 ◦C) and columnback-pressure (±10 bar) and resolution was maintained at ≥1.9 in all cases. R-timolol was resolved from
imolol maleate all potential impurities and the limit of detection was improved by increasing the sample concentrationthreefold compared to the NP-HPLC method such that the method could detect the R-timolol enan-tiomer at 0.5% (w/w) with respect to S-timolol maleate. Additional validation parameters demonstratedthat the potential of the method to be used for routine release testing of timolol maleate raw material fordrug product manufacturing in which the quantitation of R-timolol impurity in S-timolol maleate drugsubstance would be a requirement.
The chromatographic separation of enantiomers presents sig-ificant challenges in analytical chemistry. Since amino acids andarbohydrates contain chiral centres, chirality is a fundamentalharacteristic of all living organisms. All essential physiologicalrocesses display enantioselectivity, where one enantiomer inter-cts more strongly with a certain target site than the other due toifferences in its spatial configuration [1]. While it is well known
hat substantial pharmacological differences exist between enan-iomeric pharmaceuticals, it was not until after the thalidomideisaster in the 1960s that research activity increased in the field
of chirality. Today the United States Food and Drug Administra-tion (USFDA) requires enantiomeric studies to be performed onall new racemic drugs as described below [2] as do the Euro-pean and Japanese regulatory authorities. These guidelines includethe development of enantioselective identification and quantifica-tion methods for each active pharmaceutical ingredient with chiralproperties. In addition pharmacokinetic and toxicological assaysshould be executed using both pure enantiomers and with the race-mate. Furthermore, there is accumulating evidence demonstratingthe medicinal advantage of using pure enantiomers over racematesas active drug substances. As a result, numerous methods haverecently been adopted to replace existing racemates with singleenantiomeric drugs [3].
Beta adrenergic receptor blocking agents, commonly called�-blockers [4,5], are a group of drugs used to treat highblood pressure, heart failure and myocardial ischaemia dis-eases [6]. Most �-blockers are available as racemates with
Fig. 1. Molecular structure of (a) S-timolol maleate and (b) R-timolol maleate.
elatively few marketed as single-enantiomer drugs. Timo-ol maleate, (S)-(−)-1-(tert-butylamino)-3-[(4-morpholino-1,2,5-hiadiazol-3-yl)oxy]-2-propanol maleate (shown in Fig. 1), belongso the thiadiazole class of compounds and is used for the treat-
ent of essential hypertension, ocular hypertension and chronic,pen-angle glaucoma, including aphakia [7].
A 2008 review by Wang et al. discusses recent developmentsn enantioseparations of adrenergic pharmaceuticals [8]. Specif-cally, enantioseparation of timolol maleate has been previouslyeported using capillary electrophoresis and NP-HPLC. In the casef electrophoretic separations, Hedeland et al. demonstrated thenantioresolution of timolol using 1S,4R-(+) ketopinic acid as a chi-al selector in non-aqueous capillary electrophoresis (NACE) [9].esolution of 4.2 was achieved between R-timolol and S-timololithin 10.5 min and a limit of detection of 0.2% (w/w) R-timolol
n S-timolol standards was achieved using pre-concentration byransient isotachophoresis. Servais et al. investigated the use ofACE with heptakis (2,3-di-O-methyl-6-O-sulfo)-b-cyclodextrin
HDMS-b-CD) as chiral selector in combination with potassiumamphorsulfonate to achieve a resolution of 8.5 between bothimolol enantiomers within 12.5 min [10]. The same method wasubsequently validated [11] and subjected to inter-laboratorytudies [12] and robustness tests [13] and shown to permithe detection of 0.1% (w/w) R-timolol in S-timolol samples.ncertainty in measurements was shown to be concentrationependent.
NP-HPLC was first reported in 1990 by Enien et al. [14] for theuantitation of R-timolol in S-timolol maleate using a Chiracel OD-
(cellulose tris-3,5-dimethylphenylcarbamate) stationary phasend a mobile phase of hexane/isopropanol/diethylamine (95:5:0.4).he European Pharmacopoeia method [15] specifies a resolutionf 4.0 between the enantiomers using this stationary phase andobile phase system. Further optimisation of the method by Marini
t al. [16–19] permitted the separation of both enantiomers inhe presence of other potential S-timolol impurities (isotimolol,imer maleate and dimorpholinothiadiazole within 20 min byodifying the mobile phase to hexane/isopropanol/diethylamine
96.5:3.5:0.1) and the column temperature (previously unspecifiedn the European Pharmacopoeia monograph) to 23 ◦C.
SFC is currently experiencing a new renaissance (particularlyithin the pharmaceutical industry) due to the relatively recent
ommercial availability of SFC instrumentation which exhibits sig-
ificantly improved performance characteristics relative to the
nstrumentation of the early 1990s [20,21]. The potential bene-ts over conventional HPLC in terms of solvent cost reduction andaste removal are becoming ever more important to satisfy the
ogr. A 1325 (2014) 213– 220
need to generate more environmentally friendly modes of analy-sis. As a result, packed column SFC is becoming more widely usedfor chiral separations and has the potential to replace HPLC as a firstchoice technique for enantioseparations and purifications in drugdiscovery and development processes.
However, SFC has historically suffered from a serious lack ofsensitivity and has never found a place in trace analysis or reg-ulated environments [22,23]. One of the major obstacles is that(until recently) SFC instruments employing UV detection were lesssensitive compared to HPLC-UV instruments making the detectionof trace level impurities very challenging [24]. A recent reviewby De Klerk et al. reveals that SFC is becoming ever more popu-lar for enantioseparation of pharmaceuticals [25]. Gyllenhaal et al.demonstrated the use of a Hypercarb packed column using l-(+)-tartaric acid as chiral selector to resolve enantiomers of selected�-blockers [26]. Medvedovici et al. used polysaccharide stationaryphases to separate �-agonists [27]. Lui et al. used three macrocyclicglycopeptide chiral selectors and compared their performancein the separation of a range of �-blockers [28]. To the best ofour knowledge however, no report has appeared in the litera-ture describing the use of SFC for the enantioseparation of timololmaleate or indeed any active pharmaceutical ingredient used inophthalmic medications.
During the last decades, a large number of CSPs have been devel-oped to achieve chiral separation of a wide variety of racemiccompounds and enantioseparations on CSPs has thus far been oneof the most popular applications of SFC [25,29,30]. A wide range ofCSPs have been used in SFC including Pirkle type, cyclodextrins andderivatised polysaccharide-based CSPs. The polysaccharide-basedstationary phases, mainly the tris(3,5-dimethylphenylcarbamate)of amylose (Chiralpak AD) and cellulose (Chiralcel OD), have provento be two of the most successful and widely applied CSPs due tothe high number of compounds resolved in normal phase mode.Fortuitously, these polysaccharide-based stationary phases are alsoamong the most successful in chiral SFC applications. The deriva-tised polysaccharide phase, Chiralcel OD-H, incorporates a cellulosecarbamate derivative coated on silica gel as the chiral selector andhas been used to separate R-timolol and S-timolol as describedabove.
Although fundamental studies of the chiral recognition mecha-nisms on polysaccharide CSPs have been reported [31–33] a cleardescription of the chromatographic processes operating on theseCSPs is still missing. However, a clear difference in enantioselectiv-ity has been observed between SFC and LC modes with hydrogenbonding being found to play an important role in the differen-tial binding of the enantiomers to the CSPs in SFC applications[30]. Consequently, method development in chiral SFC, as in chi-ral NP-HPLC, generally relies on a systematic screening of CSPs andmobile phases, in a preferential order or in parallel based on per-sonal experience of the analyst. However, with shorter retentiontimes compared to conventional NP-HPLC, method developmentwith chiral SFC is faster without sacrificing efficiency [25]. Themobile phase parameters to be optimised are relatively limited,with only a few organic modifiers commonly used and investi-gations into the effect of pH generally omitted with CO2-basedmobile phases. As with NP-HPLC, temperature can play a significantrole in SFC method development particularly in chiral separations[34].
With this in mind, the aim of this study was to develop analternative limit test assay for the determination of S-timololmaleate enantiomeric purity and compare it to the current PharmEur NP-HPLC compendial method of analysis which uses a cel-
lulose tris(3,5-dimethylphenylcarbamate) coated silica stationaryphase [14,16–19] under normal phase conditions. The developedmethod was compared with the existing normal phase methodusing standard analytical performance criteria.
romat
2
2
4pfwwMpH(p2mAdgpoiCsoa
2
EERpcM((c
2
aAitaaSfc4uaaartg
2
a
A. Marley, D. Connolly / J. Ch
. Experimental
.1. Instrumentation
Both NP-HPLC and SFC assays were carried out on a.6 mm × 250 mm Chiralcel OD-H column packed with 5 �m silicaarticles coated with cellulose tris(3,5-dimethylphenylcarbamate)rom Daicel Limited Industries (Tokyo, Japan). The NP-HPLC assayas performed on a Waters Alliance 2695 HPLC system coupledith Waters 2487 Dual Wavelength Absorbance Detector (Waters,ilford, MA, USA). Chromatographic data was acquired and
rocessed using Waters Empower 2 software. For the optimised NP-PLC assay the mobile phase of hexane/2-propanol/diethylamine
960:40:2) was delivered at 1.0 mL min−1 at ambient column tem-erature. Injection volume was 5 �L and detection was by UV at97 nm with data acquisition at 5 Hz. The SFC method develop-ent and validation was performed on an Agilent 1260 Infinitynalytical SFC system coupled with an Aurora SFC Fusion A5 CO2elivery system equipped with a DAD detector (Agilent Technolo-ies, Santa Clara, CA, USA). Chromatographic data was acquired androcessed using Agilent Chemstation Rev B.04.03 software. For theptimised SFC assay, the mobile phase of (93:7) CO2/0.1% (v/v) TEAn MeOH was delivered isocratically at a flow rate of 4.0 mL min−1.olumn temperature was set to 40 ◦C with backpressure regulationet to 130 bar. An injection loop of 5 �L was over-filled with 15 �Lf sample and peaks were detected at 297 nm with data acquisitiont 20 Hz.
.2. Materials and reagents
Samples of S-timolol maleate and R-timolol base and Pharm.ur. specified impurities B, C, D and F were obtained from theuropean Pharmacopoeia Secretariat (Strasbourg, France). (Note:-timolol is also known as Impurity A.) N-Hexane of LC grade wasurchased from Romil (Waterbeach, Cambridge, UK). Methylenehloride, HPLC grade methanol (MeOH) and 2-propanol was fromerck (Darmstadt, Germany), and reagent grade diethylamine
DEA) and reagent grade triethylamine (TEA) from Sigma–AldrichDublin, Ireland). Food grade CO2 (99.9%, v/v, minimum) was pur-hased from BOC Ltd. (Dublin, Ireland).
.3. Solution preparation for NP-HPLC analysis
(Note: Solution preparation for the NP-HPLC assay and the SFCssay (below) is also detailed in Table 1 for the sake of clarity.)ll standard solutions were prepared (and diluted as appropriate)
n methylene chloride/2-propanol (10:30) using actinic glasswareo protect from light. Two stock standards of S-timolol maleatend one stock standard of R-timolol were prepared at 9.5 mMnd 7.6 × 10−2 mM respectively based on timolol free base. Both-timolol maleate stock standards were diluted to 9.5 × 10−2 mMor recovery and method precision studies. A racemic mixtureontaining S-timolol maleate and R-timolol at a concentration of.7 × 10−2 mM was prepared from stock standard dilutions andsed as a control resolution mixture (CRM) to assess resolutionnd to calculate a relative response factor (RRF) for R-timolol rel-tive to S-timolol maleate. Finally, a solution of S-timolol maleatend R-timolol was prepared at 9.5 × 10−2 mM and 9.5 × 10−4 mMespectively (R-timolol concentration being equivalent to 1.0% ofhe S-timolol maleate concentration). This solution was used toenerate data for comparative purposes with the SFC method.
.4. Solution preparation for SFC analysis
All standard solutions were prepared in 100% MeOH usingctinic glassware. Four stock standards of S-timolol maleate and
ogr. A 1325 (2014) 213– 220 215
three stock standards of R-timolol were prepared at 28 mM and2.3 × 10−1 mM respectively based on timolol free base. Two of theS-timolol maleate stock standards were diluted to 2.8 × 10−1 mMfor recovery and method precision studies. Standards were pre-pared containing S-timolol maleate at 2.8 × 10−1 mM and R-timololat 5.7 × 10−3 mM, 2.8 × 10−3 mM, 1.4 × 10−3 mM, 2.8 × 10−4 mM,1.4 × 10−4 mM, equivalent to 2.0, 1.0, 0.5, 0.1 and 0.05% of theS-timolol maleate concentration. These standards were used forvalidation of the SFC method. A solution containing S-timololmaleate and R-timolol at 1.4 × 10−1 mM was prepared as a CRMto assess resolution and to calculate an RRF for R-timolol relativeto S-timolol.
2.5. Solution preparation for evaluation of SFC method specificity
A solution was prepared containing all of the European Pharma-copoeia Timolol specified impurities, i.e. B, C, D, E and F. ImpurityE was generated as per the European Pharmacopoeia [35] by reac-ting timolol maleate (2 mg) with maleic acid (20 mg) in acetonitrile(10 mL). 1 mL of this solution was then evaporated under nitrogenbefore being dried at 105 ◦C for 1 h. The reagent was then reconsti-tuted with 1 mL of methanol. 200 �L of the Impurity E solution wasadded to a vial of European Pharmacopoeia system suitability mix(containing impurity B, C, D and F) containing 800 �L of methanolto generate an impurity mix solution containing all of the specifiedtimolol impurities.
3. Results and discussion
3.1. SFC method development
All SFC development work was carried out on same Chiral-cel OD-H cellulose tris(3,5-dimethylphenylcarbamate) stationaryphase as was used for the NP-HPLC assay [15,35] This stationaryphase was selected primarily so that a direct comparison couldbe made between the SFC method described here, and the previ-ously developed NP-HPLC assay. Furthermore, a recent report byKhater et al. [36] reveals that the Chiralcel OD-H stationary phasestill offers the most versatility when compared with four genericversions of the Chiralcel OD (CelluCoat, RegisCell, Lux Cellulose-1,Reprosil OM) and the immobilised version (ChiralPak IB), involvinga “unique and unequalled mechanism to achieve enantioseparation”.For these reasons, alternative CSPs were not used during methoddevelopment.
A CRM was prepared containing the R-timolol and S-timololat 1.4 × 10−1 mM which was used to assess the chromatography.The chromatography was assessed in the first instance by measur-ing resolution between the enantiomers but these initial screeningstudies using MeOH alone as the modifier resulted in very poor peakshape and correspondingly poor sensitivity as shown in Fig. 2(b)
In HPLC separations, poor peak shape due to tailing issues andin some cases, a lack of peak elution have been attributed to thepresence of so called “active sites” on the surface of the stationaryphase support [37]. For silica based columns, these active sites usu-ally consist of different types of residual silanol groups (germinal,vicinal, etc.) or metal ion contamination resulting from the station-ary phase manufacturing process. The presence of these active sitescan result in an unwanted competing secondary retention mecha-nism resulting in severe tailing. Additives are very polar substancesthat are added to the mobile phase in low concentrations and areexpected to facilitate solute elution and improve peak shapes by
covering up, adsorbing onto or even reacting with these so calledactive sites [37]. In pSFC, additives appear to have four major func-tions: (1) coverage of so called “active sites”, (2) altering the polarityof the stationary phase, (3) altering the polarity of the mobile phase
216 A. Marley, D. Connolly / J. Chromatogr. A 1325 (2014) 213– 220
Fig. 2. Separation of R-timolol and S-timolol via SFC using a mobile phase with(a) and without (b) the presence of 0.1% (v/v) TEA. Chromatographic conditions;column: 4.6 mm × 250 mm Chiralcel OD-H, 5 �m. Mobile phase: (93:7) CO2/MeOHwith 0.1% TEA for chromatogram (a) and (93:7) CO2/MeOH for chromatogram (b).Flow rate: 4.0 mL min−1. Injection volume: 5 �L. Column temperature: 40 ◦C. Back-p4
oAesbasctptfomwabNpc
TS
ressure regulation: 130 bar. Detection: 297 nm. Peak assignment: R-timolol at.7 × 10−2 mM and S-timolol at 4.7 × 10−2 mM.
r (4) suppression of ionisation or ion pair formation by solutes.lthough the functions of additives can be broadly defined, thexact role of additives in pSFC still remains unclear with furthertudy required if a mechanistic understanding of their actions is toe achieved. Due to the lack of comprehensive understanding ofll the secondary retention mechanisms that give rise to poor peakhape and tailing in pSFC, there are generally no set rules when itomes to choosing an additive for a given separation other thanhe additive being a stronger acid or stronger base than the sam-le components being separated as well as being compatible withhe choice of detector. In general, acidic additives should be usedor acidic solutes and basic additives for basic solutes [37]. The pKa
f timolol base has been reported as 9.2 (determined via potentio-etric titration in water at 25 ◦C) [38] and so triethylamine (TEA)as selected as the basic additive and added to the MeOH modifier
t 0.1% (v/v). This concentration was selected based upon reportsy Aboul-Enein et al. who selected 0.4% (v/v) diethylamine in their
P-HPLC assay, observing that the absence of base on the mobilehase resulted in no separation, whereas higher levels (1%, v/v)ould result in damage to the silica support particle [14].
able 1olution preparation for NP-HPLC and SFC analysis.
Diluent Solution Analyte
Solutions for NP-HPLC
Methylene chloride/2-propanol (10:30)
R-timolol Stock standard R-timolol
S-timolol Stock standard S-timolol
Recovery/methodprecision standards
S-timolol
CRM mixture R-timolol
S-timolol
S-timolol/R-timololspiked standards
R-timolol
S-timolol
Solutions for SFC
MeOH
R-timolol Stock standard R-timolol
S-timolol Stock standard S-timolol
Precision studies S-timolol
S-timolol/R-timololspiked standards
R-timolol
S-timolol
CRM mixture R-timolol
S-timolol
Fig. 3. Plot of retention factor (k) and resolution versus % MeOH via SFC. Chromato-graphic conditions as in Fig. 2(a).
The effect of the additive was immediate, with peak shape andsensitivity being dramatically improved as shown in Fig. 2(a) Elu-tion order remained unaffected but resolution of 2.82 and peakasymmetries of 0.82 for both R-timolol and S-timolol was achievedby the inclusion of the additive to the mobile phase. Although otherworkers [16–19] and the European Pharmacopoeia monograph [35]have indicted the use of diethylamine as the basic additive in NP-HPLC assays, in the work presented herein, no further efforts weremade to investigate the effect of alternative basic additives or theeffect of TEA concentration, since adequate resolution was achievedfor the purposes of a limit test.
The remainder of the development work involved performingoptimisation experiments in which flow rate and mobile phasecomposition were adjusted to reduce runtimes for increased sam-ple throughput, while maintaining USP resolution >2 and peaktailing <1.5 for both enantiomers in the CRM. Isocratic methodswere favoured over gradient methods to eliminate the necessarysystem equilibration time after each run. In order to normalise theeffect of flow rate, a plot of retention factor (k) versus % modifier wasconstructed for both enantiomers as shown in Fig. 3. As expected,retention decreased steadily as % MeOH increased with a corre-
sponding decrease in resolution at higher modifier contents. Forexample, 20% MeOH (containing 0.1% TEA) resulted in resolutionof 0.7 whereas with 10% modifier the resolution was 1.5.
Concentration (mM) % R-timolol w.r.t. S-timolol
7.6 × 10−2 mM n/a9.5 mM9.5 × 10−2 mM
4.7 × 10−2 mM4.7 × 10−2 mM9.5 × 10−4 mM 1.0%9.5 × 10−2 mM
2.3 × 10−1 mM28 mM n/a2.8 × 10−1 mM5.7 × 10−3 mM, 2.8 × 10−3 mM, 1.4 × 10−3 mM,2.8 × 10−4 mM, 1.4 × 10−4 mM,
2.0%, 1.0%, 0.5%, 0.1%, 0.05%
2.8 × 10−1 mM1.4 × 10−1 mM 50%1.4 × 10−1 mM
A. Marley, D. Connolly / J. Chromatogr. A 1325 (2014) 213– 220 217
Fig. 4. Comparison of optimised SFC separation (a) and normal phase separation(b) of timolol enantiomers. Chromatographic conditions for (a) are as given inFig. 2(a). Chromatographic conditions for (b); column: 4.6 mm × 250 mm ChiralcelO1a
As a compromise between runtime and resolution a final mobilehase composition of (93:7) CO2/0.1% TEA in MeOH at a flow ratef 4.0 mL min−1 was selected as optimum which resulted in a 5 minuntime and resolution of 2.01 between the enantiomers. Bothnantiomers eluted within a retention time window of just 0.5 mins shown in Fig. 4(a) with chromatographic efficiency of 17,500 N/mnd 16,500 N/m observed for R-timolol and S-timolol respectively.he developed SFC method is therefore three times faster thanhe optimised normal phase separation described by Marini et al.16–19] and reproduced in Fig. 4(b) for comparative purposes. Theuropean Pharmacopoeia method [35] specifies a resolution of 4.0etween the enantiomers using this stationary phase and mobilehase system. Furthermore, the optimised method is also overwice as fast as the CE method described by Hedeland et al. [9]nd three times faster than the NACE methods described by Ser-ais et al. [10] and Marini et al. [11–13] based upon the migrationime of the second enantiomer. As such, this method representsn improvement over previously published enantioseparations ofimolol maleate. (It should be noted that the injection cycle timeas <21 s and therefore did no contribute significantly to overall
untimes.) To help improve the limit of detection, timolol concen-rations were increased threefold compared to the NP-HPLC assayor the SFC method validation.
.2. SFC method validation
The SFC method was developed to replace an existing NP-HPLCssay currently in use within industry (Allergan Pharmaceuticals)nd reported by Marini et al. [16–19]. R-timolol is an unwantedmpurity in S-timolol drug substance and in the pharmaceuticalndustry the purity of an active pharmaceutical ingredient can beetermined using either a limit test or a quantitative test. The NP-PLC assay is described in the European Pharmacopoeia [35] as
limit test based upon a specification of not more than (NMT).0% of R-timolol in the presence of S-timolol. The ICH require-ents for the validation of a limit test are specificity and limit of
etection (LOD). Therefore, in the first instance these two analyt-cal performance criteria were initially evaluated to demonstrate
he applicability of the developed SFC method as a limit test andhe results are evaluated below. For a quantitative test relative to alimit test”, additional analytical performance characteristics muste evaluated, namely: accuracy, precision, LOQ, linearity, range and
Fig. 5. Chromatogram of R-timolol at the LOD of 0.5% in the presence of S-timololmaleate (b) overlaid with a MeOH blank (a). Chromatographic conditions as inFig. 2(a).
robustness [39]. Therefore, further selected validation criteria werethen studied to demonstrate the additional potential of this methodas a quantitative test for R-timolol impurity in S-timolol drug sub-stance and the results presented hereafter.
3.2.1. Validation of the SFC method as a limit test – specificityand LOD
The limit of detection (LOD) is defined as the smallest quan-tity of the target substance (in this case, R-timolol) that can bedetected relative to baseline noise, but not accurately quantified inthe sample. Baseline noise was compared with the peak height of astandard containing 0.5% R-timolol in the presence of S-timolol asshown in Fig. 5(a). The magnitude of baseline noise was measuredin a blank chromatogram over a distance equivalent to 5 times thepeak width at half height of the R-timolol peak, centred around itsexpected retention time (noise window: 3.8 min). The LOD resultsare summarised in Table 2 and indicate that the LOD was found tobe 0.5% for R-timolol in the presence of S-timolol maleate, satis-fying the requirement for the limit test method [35] to be able todetect R-timolol at the 1.0% level of the S-timolol maleate peak.
The specificity of a method is defined as the ability to assessunequivocally the analyte in the presence of components whichmay be expected to be present. The European Pharmacopoeia spec-ifies five impurities which can potentially be present in S-timololdrug substance; Impurity A (R-Timolol), Impurity B, Impurity C,Impurity D, Impurity E and Impurity F. A commercially availableEuropean Pharmacopoeia impurities system suitability mixturecomprising S-timolol spiked with all known impurities was usedto assess the specificity of the method. (Impurity E was separatelyadded to this mixture.) Although only Impurity A (R-timolol) andImpurity E could be individually identifies by retention time, nev-ertheless Fig. 6 clearly indicates that all five impurities are baselineseparated from R- and S-timolol, demonstrating the specificity ofthe method for the determination of R-timolol in S-timolol drugsubstance.
3.2.2. Investigation of SFC method as a potential quantitativeimpurities assay3.2.2.1. Precision studies: repeatability and intermediate precision.The repeatability of the SFC method was evaluated by prepar-ing a standard containing R-timolol at 1.0% relative to S-timololmaleate concentration and injecting six times on Day #1. Interme-
diate precision was determined by preparing two separate freshpreparations of the standard on Day #2 and injecting each six times.As shown in Table 2, the relative standard deviations (% RSD) val-ues were ≤2.7% for both repeatability and intermediate precision.
218 A. Marley, D. Connolly / J. Chromatogr. A 1325 (2014) 213– 220
Table 2Analytical performance for the determination of R-timolol impurity in S-timolol maleate.
Repeatability and intermediate precisionR-timolol retention time % RSD <0.1% <0.1% <0.1% <0.1% <0.1% <0.1%R-timolol peak area % RSD 2.1% 1.8% 2.7% 2.0% 2.0% 2.7%
S/N % area
Limit of detectionR-timolol 3.4 0.52
Accuracy
R-timolol 101%
a Flow rate: 4.0 mL min−1, Column temperature: 40 ◦C, backpressure regulation: 130 bar.b n = 3.
onditi
Rb
3mais(pTnpem
FtF
c Resolution calculated as R = 2(tr1 − tr2)/(WB(1) − WB(2)).d n = 6.e Ratio of analytical performance criterion versus performance under optimum c
etention time precision for R-timolol was ≤0.1% within day andetween days.
.2.2.2. Robustness. The robustness of an analytical procedure is aeasure of its ability to remain unaffected by small, but deliber-
te changes in method parameters and provides an indication ofts reliability of the analytical method during normal usage. In thistudy, the experimental conditions that were altered were flow rate±0.5 mL min−1), column temperature (±5 ◦C) and column back-ressure (±10 bar) based both upon ICH Q2R1 recommendations.he range of variance was selected to evaluate method robust-ess in the event that a different SFC instrument (with different
erformance characteristics) was used for the method. Firstly, forach selected experimental condition, three injections of the CRMixture were made and standard chromatographic performance
ig. 6. Overlay of S-timolol standard spiked with R-timolol and all know impuri-ies (a) overlaid with a blank chromatogram (b). Chromatographic conditions as inig. 2(a).
ons.
criteria were evaluated, namely: relative retention time (RRT), res-olution, selectivity and peak symmetry. When compared with thechromatogram obtained under optimised conditions, deviationsin chromatographic performance were not significant as shownin Table 2 and in all cases, resolution was ≥1.9. Specifically, RRTchanged by only 1%, resolution by 10%, selectivity by <1% and peaksymmetry by 3%. Furthermore, for each condition investigated,six injections of an R-timolol standard (at 1.0% with respect toS-timolol) were made and changes in repeatability, if any, wereevaluated. Relative to the optimum chromatographic conditions,there was no significant effect upon repeatability as shown inTable 2, demonstrating that the method is capable of maintainingthe desired performance regardless of the robustness challengesunder investigation. The equivalency of chromatographic perfor-mance at each stress condition relative the optimum conditionswas evaluated by calculating the ratio of the means, with equiva-lency of 1.0 reported in all cases (Table 2).
3.2.2.3. Accuracy. To assess the accuracy of the 1.0% R-timolol sam-ple solutions used in the validation studies for both the HPLCand SFC methods, S-timolol was used as an external bracket-ing standard. To eliminate differences in UV response betweenR-timolol and S-Timolol, relative response factors (RRF) weregenerated by calculating the differences in peak area versus con-centration for the CRM solution, since both enantiomers werepresent at the same concentration. The RRF values for both theHPLC and SFC methods are shown in Table 3. Recovery values of98% and 101% are reported for R-timolol in the NP-HPLC and SFCmethods respectively.
3.3. Analytical performance comparison: SFC versus NP-HPLC
The SFC method offered comparable performance to the NP-
HPLC method in terms of peak response, precision and detectionlimit (with increased sample concentration). A comparison of theSFC and NP-HPLC methods is outlined in Table 3. For comparativepurposes, Fig. 7 illustrates chromatograms from both methods in
A. Marley, D. Connolly / J. Chromatogr. A 1325 (2014) 213– 220 219
Fig. 7. Typical chromatograms comparing SFC (a and b) and NP-HPLC (c and d) for the anare the CRM, and (b and d) are R-timolol at 1.0% w.r.t. S-timolol. Chromatographic condit
Table 3Comparison of analytical performance between the SFC and NP-HPLC methods.
Resolution 2.0 4.8S-timolol peak area repeatabilitya 0.2% 0.4%R-timolol peak area repeatabilityb 2.1 2.5% recovery of R-timololc 101%c 98%c
S-timolol working standard concentration 0.009% (w/v) 0.003% (w/v)Analysis time per sample (min) 5 16Solvent usage per sample (mL) 1.4 16
waFe
4
opaiNST
a n = 6.b 1.0% R-timolol in S-timolol, n = 6.c 1.0% R-timolol in S-timolol.
hich the benefits of SFC over NP-HPLC in terms of analysis timend solvent consumption are obvious. To the best of our knowledge,ig. 7 represents the fastest enantiomeric separation of timololnantiomers reported to date.
. Conclusion
A method has been described in which the enantiomers of tim-lol maleate have been separated on a Chiralcel OD-H stationaryhase within 5 min, representing a 3-fold decrease in runtimend an 11-fold decrease in solvent consumption relative to the
ndustry standard, European Pharmacopoeia method based uponP-HPLC. Due to the low viscosity and high diffusivity of theFC mobile phase, a fourfold increase in flow rate was possible.he method validation parameters required for a limit test for
alysis of R-timolol in the presence of S-timolol maleate. Chromatograms (a and c)ions as in Fig. 4.
R-timolol in S-timolol maleate (specificity and detection limit)were established for the SFC method. In addition, the potentialof this method to be used for quantitation of R-timolol impuritywas investigated by evaluation of further analytical performancecriteria (robustness, precision, accuracy). Clearly the developed SFCassay demonstrates potential as a full quantitative assay, and repre-sents the fastest separation of timolol enantiomers to date, relativeto previously reported NP-HPLC or NACE-based methods. Futurework will involve the use of shorter chiral columns packed withsmaller particles (3 �m) in order to further decrease runtimes inchiral SFC. Dissolution of the samples in a more non-polar solventsuch as heptane or heptane/isopropanol mixtures compared withmethanol may also result in improved chromatographic efficiency.
Acknowledgments
The authors wish to thank the management of Allergan Pharma-ceuticals Ireland Ltd. for supporting this work. Invaluable technicalassistance from Mr. Stephen Fuller is greatly appreciated. Theauthors also acknowledge Science Foundation Ireland for fundingunder the Strategic Research Cluster Programme (Grant Number08/SRC/B1412) and Mr. Declan Murray of Agilent TechnologiesIreland for provision of the SFC instrumentation.
References
[1] W.A. Bonner, Chirality 12 (2000) 114.[2] Food & Drug Administration, FDA’s policy statement for the development of
new stereoisomeric drugs, 1992, 57 Fed. Reg. 22249.
[3] Augustian, Process Biochem. 45 (2010) 1587.[4] W.H. Frishman, Circulation 107 (2003) 117.[5] M.J. Reiter, Prog. Cardiovasc. Dis. 47 (2004) 11.[6] J. Agustian, A.H. Kamaruddin, S. Bhatia, Process Biochem. 45 (2010) 1587.[7] T.J. Zimmerman, Invest. Ophthalmol. Vis. Sci. 16 (1977) 687.
[8] Z. Wang, J. Ouyang, W.R.G. Baeyens, J. Chromatogr. B 862 (2008) 1.[9] Y. Hedeland, J. Lehtinen, C. Pettersson, J. Chromatogr. A 1141 (2007) 287.10] A.C. Servais, M. Fillet, P. Chiap, W. Dewé, P. Hubert, J. Crommen, Electrophoresis
25 (2004) 2701.11] R.D. Marini, A.C. Servais, E. Rozet, P. Chiap, B. Boulanger, S. Rudaz, J. Crommen,
P. Hubert, M. Fillet, J. Chromatogr. A 1120 (2006) 102.12] R.D. Marini, C. Groom, F.R. Doucet, J. Hawari, Y. Bitar, U. Holzgrabe, R. Gotti, J.
Schappler, S. Rudaz, J.L. Veuthey, R. Mol, G.W. Somsen, G.J. de Jong, P.T.T. Ha, J.Zhang, A. Van Schepdael, J. Hoogmartens, W. Briône, A. Ceccato, B. Boulanger,D. Mangelings, Y. Vander Heyden, W. Van Ael, I. Jimidar, M. Pedrini, A.C. Servais,M. Fillet, J. Crommen, E. Rozet, P. Hubert, Electrophoresis 27 (2006) 2386.
13] R.D. Marini, E. Rozet, Y. Vander Heyden, E. Ziemons, B. Boulanger, A. Bouklouze,A.C. Servais, M. Fillet, J. Crommen, P. Hubert, J. Pharm. Biomed. Anal. 44 (2007)640.
14] H.Y. Aboul-Enein, M.R. Islam, J. Chromatogr. 511 (1990) 109.15] European Pharmacopoeia, 4th edition, Addendum 2002, Council of Europe,
Strasbourg.16] R.D. Marini, P. Chiap, B. Boulanger, W. Dewe, P. Hubert, J. Crommen, J. Sep. Sci.
26 (2003) 809.17] R.D. Marini, N. Matthijs, Y. Vander Heyden, J. Smeyers-Verbeke, P. Dehouck,
J. Hoogmartens, P. Silvestre, A. Ceccato, Ph. Goedert, J. Saevels, C. Herbots, G.Caliaro, R. Herraez-Hernandez, J. Verdu-Andres, P. Campıns-falco, W. Van deWauw, J. De Beer, B. Boulanger, P. Chiap, J. Crommen, Ph. Hubert, Anal. Chim.
Acta 546 (2005) 182.
18] R.D. Marini, B. Boulanger, Y. Vander Heyden, P. Chiap, J. Crommen, P. Hubert,Anal. Chim. Acta 531 (2005) 131.
19] R.D. Marini, P. Chiap, B. Boulanger, S. Rudaz, E. Rozet, J. Crommen, P. Hubert,Talanta 68 (2006) 1166.
[[
[[
ogr. A 1325 (2014) 213– 220
20] K. De Klerck, D. Mangelings, Y. Vander Heyden, LC–GC Europe (May (Suppl.))(2013) 24.
21] R.L. Stevenson, Am. Lab. 45 (2013) 36.22] T.A. Berger, B.K. Berger, J. Chromatogr. A 1218 (2011) 2320.23] R. Helmy, M. Biba, J. Zang, B. Mao, K. Fogelman, V. Vlacho, P. Hosek, C.J. Welch,
Chirality 19 (2007) 787.24] Z. Wang, H. Zhang, O. Liu, B. Donovan, J. Chromatogr. A 1218 (2011)
2311.25] K. De Klerck, D. Mangelings, Y. Vander Heyden, J. Pharm. Biomed. Anal. 69
(2012) 77.26] O. Gyllenhaal, A. Karlsson, J. Biomed. Biophys. Methods 54 (2002) 169.27] A. Medvedovici, P. Sandra, L. Toribio, F. David, J. Chromatogr. A 785 (1997)
159.28] Y. Liu, A. Berthod, C.R. Mitchell, T.L. Xiao, B. Zhang, D.W. Armstrong, J. Chro-
matogr. A 978 (2002) 185.29] K.L. Williams, L.C. Sander, J. Chromatogr. A 785 (1997) 149.30] H. Nelander, S. Andersson, K. Öhlén, J. Chromatogr. A 1218 (2011) 9397.31] C. Roussel, B. Bonnet, A. Piederriere, C. Suteu, Chirality 13 (2001) 56.32] R.B. Kasat, N.H.L. Wang, E.I. Franses, J. Chromatogr. A 1190 (2008) 110.33] S. Ma, S. Shen, H. Lee, M. Eriksson, X. Zeng, J. Xu, K. Fandrick, N. Yee, C.
Senanayake, N. Grinberg, J. Chromatogr. A 1216 (2009) 3784.34] R.W. Stringham, J.A. Blackwell, Anal. Chem. 68 (1996) 2179.35] European Pharmacopoeia (Version 7.5): Timolol Maleate monograph.
36] S. Khater, Y. Zhang, C. West, J. Chromatogr. A 1303 (2013) 83.37] M. Ventura, B. Murphy, W. Goetzinger, J. Chromatogr. A 1220 (2012)
147.38] K. Kontturi, L. Murtomaki, J. Pharm. Sci. (US) 81 (1992) 970.39] United States Pharmacopeia No. 31-NF-26, 2008 (Chapter 621).
