ABSTRACT
The manufacturing process of pharmaceutical products, especially active pharmaceutical ingredients (APIs), often involves the use of reactive materials (eg, starting materials, intermediates, catalysts, and reagents) These reactive materials along with certain reactive by-products resulting from the manufacturing process or introduced from the starting materials, intermediates, catalysts, and reagents could remain at trace levels in the final pharmaceutical products Based on their structures and reactivity, some of these compounds have been classified as genotoxic impurities (GTIs)
Due to the concern for patient safety, the issue of GTIs in pharmaceutical products has attracted increasing attention from the industry1-5 as well as regulatory agencies6-11 The European Medicines Evaluation Agency9 and Food and Drug Administration11 guidelines have led to an expectation that a control strategy is developed that demonstrates or justifies control of these undesired GTIs to an acceptable level (often low parts per million) in the final pharmaceutical products as appropriate for the stage of development Analysis of GTIs in pharmaceutical
81 Introduction 255 82 Overview for Gas Chromatography-Mass Spectrometry Applications
for Genotoxic Impurity Analysis 257 821 Dilute and Shoot 261 822 Extraction 265 823 Headspace 269 824 Derivatization 270
83 Conclusion 274 References 274
products often poses significant challenges12-15 First, the control limits associated with GTIs (eg, low parts per million) are significantly lower than anything previously measured (eg, 005%) by up to three orders of magnitude Even though the implementation of staged threshold of toxicological concern (TTC) guidelines provides some degree of relief for GTI control limits for pharmaceutical products dosed less than 1 year in clinical settings, extremely high sensitivity is still required for GTI analysis16 Second, complex matrix effects may arise from in-process samples, APIs, or excipients based on the typically high concentrations at which they need to be prepared These matrix effects hinder the ability to observe and quantify genotoxic analytes and need to be overcome Third, many GTIs are highly reactive and are thus prone to degradation during the analysis, which makes quantification very complicated Fourth, the structure and physicochemical properties of certain GTIs are similar to those of the intermediate or API Thus, highly specific analysis is often desired Finally, these selective and sensitive methods must be amendable for testing of pharmaceutical products at all stages of development, including commercial manufacturing The need to conduct these trace-level analyses after the drug is on the market further challenges the analytical chemist These methods need to be rugged and robust so that they can be successfully transferred and executed at commercial manufacturing sites around the world17 Beyond method development, another challenge is successful implementation of the method at the testing site by ensuring that there are appropriate instrument resources (eg, mass spectrometers) and experienced analysts capable of troubleshooting, instrument maintenance, and conducting routine testing
To overcome these challenges for GTI analysis, many pharmaceutical companies have set up highly skilled analytical groups and developed various methodologies and practices based on gas chromatography-flame ionization detection (GC-FID),18,19 gas chromatography-electron capture detection (GC-ECD),20 gas chromatographymass spectrometry (GC-MS),21,22 high-performance liquid chromatography (HPLC)-UV,23-25 HPLC-fluorescence,26 HPLC-MS,27-34 nuclear magnetic resonance,35 and ion chromatography36 Initially, many of these methods were developed for individual GTIs and products or even specific manufacturing stages or steps37-39 As these GTI analytical methodologies and practices evolved, more and more generic analytical methodologies have emerged for the analysis of individual classes of GTIs based on their functional groups linked to genotoxicity22,28,34,40 Furthermore, very recently strategic approaches for GTI analysis have been established to accommodate fast-growing pipelines and compressed development time lines14,16,41
Among all of the analytical methodologies that have been reported for GTI analysis, GC-MS is the most widely adopted one because many of the GTIs are volatile to semivolatile and low-molecular weight compounds14,41 GC-MS is well suited for GTI analysis because this technique provides chromatographic selectivity, sensitivity, and specificity for many different classes of GTIs A typical capillary gas chromatography (GC) column can yield more than 100,000 theoretical plates,42 and a wide range of stationary phases are available to obtain appropriate chromatographic selectivity and thus resolution of GTIs from other components By operating under the selective ion monitoring (SIM) mode, electron ionization provides superior specificity and detection sensitivity in GC-MS17 For example, mass spectrometry (MS) detection could offer
two to three orders of magnitude more sensitivity than flame ionization detection, the most commonly used GC detector4 Although electron capture detection (ECD) and MS detection provide comparable sensitivity for analytes with halogen or strong electronwithdrawing groups, the specificity of ECD is clearly limited to this subset of GTIs Furthermore, the availability of robust and low-cost GC-MS systems from several major instrument suppliers has helped to reduce operational costs and simplify the process of method transfer to commercial testing sites or third party manufacturing laboratories
This chapter aims to provide an overall review of GC-MS applications for the analysis of GTIs in pharmaceutical products Also, this chapter provides a comprehensive review of various sample preparation approaches for the analysis of GTIs by GC-MS
GC-MS methods for determining GTIs in pharmaceutical products need to demonstrate sufficient specificity, linearity, sensitivity, precision, and accuracy for their intended use Table 81 summarizes the GC-MS applications for GTI analysis reported in the literature Generally, the success of these GC-MS applications depended on (1) sample preparation and introduction, (2) chromatography, and (3) detection and quantitation14
Compared to HPLC-based applications (such as liquid chromatography-mass spectrometry [LC-MS]), sample preparation is often more critical for GC-MS as the analytes of interest need to be properly vaporized and introduced into the GC column prior to separation Thus, it is important to choose appropriate sample preparation approaches based on the physicochemical properties of target GTI analytes and sample matrixes The direct injection approach (or the so-called dilute and shoot approach) is widely used since it requires the least effort for sample handling and manipulation However, the major drawback of direct injection is the potential for matrix interference Since most direct injectors only deliver microliter quantities of solutions, the sample needs to be diluted and injected at a high concentration (eg, 5-50 mg/mL) to achieve the desired sensitivity As most pharmaceutical samples (intermediates, APIs, or drug products) are nonvolatile, direct injection of these materials often leads to a buildup of residue that can undergo thermal degradation, contaminate the inlet liner, and lead to interferences in the chromatography The application of direct injection is limited for isolated intermediates and APIs Direct injection is not applicable for analyzing drug product samples (eg, formulated solution, tablet, or suspension) because excipients in the drug product may lead to severe interferences when directly injected In addition, direct injection is typically not applicable when the sample cannot achieve sufficient solubility based on the required GTI control limit Instead, one could consider headspace injection and other approaches such as liquid-liquid extraction (LLE), solid-phase extraction (SPE), and solid-phase microextraction (SPME) because these methodologies are very effective in reducing matrix interferences from the sample and enriching the target GTIs When highly reactive GTIs are being analyzed, selective derivatization might improve the stability and volatility of the GTIs
TA B
LE 8
.1 G
C -M
S A
pp lic
at io
ns fo
r G
TI A
na ly
si s
TA B
LE 8
.1 (
C on
ti nu
ed )
G C
-M S
A pp
lic at
io ns
fo r
G TI
A na
ly si
s
Chromatography is used to resolve GTIs from diluent and other matrix peaks Typically, low-bleeding, midpolarity to low-polarity phase capillary columns, such as DB-5 and DB-624, are utilized for most of the GC-MS applications These phases have high inertness, which is beneficial for analyzing reactive GTIs The high efficiency of these columns also improves detection sensitivity
GC-MS detection is generally performed on a single-quad mass spectrometer with an electron impact source acquiring signal in the SIM mode It is essential to investigate the GTI MS fragment profile and carefully select appropriate diagnostic ions The selection of these ions is critical for achieving the desired specificity and sensitivity For less complicated sample preparation (eg, direct injection), the external standard method typically provides successful quantitation However, when extensive sample preparations are involved (eg, extraction or derivatization), it is recommended to introduce an internal standard (IS) Isotopic analogs of the desired GTI are generally preferred for ISs as they provide similar recovery and chromatographic behavior
Sections 821 through 824 further discuss the individual GC-MS applications in Table 81 for the analysis of GTIs The sections are organized according to the sample preparation approach chosen
The direct injection approach is the most convenient injection mode for GTI analysis since it requires minimal sample handling This approach requires GTI analytes to have sufficient volatility and stability on injection and vaporization at a hot inlet As a large amount of the generally less volatile API is introduced into the GC inlet, the GC inlet may require frequent maintenance to address potential contamination issues
Ramjit and others21 developed a GC-MS method to determine methyl mesylate (MMS) and ethyl mesylate (EMS) in the free base and the bismesylate salt of DPI 201-106, a positive inotropic agent used for treating heart failure This report is one of the pioneering studies investigating residual levels of alkyl mesylates in intermediates and APIs A 100 mg sample was dissolved in 1 mL of acetonitrile with 02 wmL N,N-dimethylformamide (DMF) to improve solubility A DB-WAX column (30 m × 025 mm internal diameter [id], 025 μm film thickness) was used for separation Analyses were performed using a splitless injection mode and an injection port temperature of 200°C Quantitative GC-MS was performed in SIM mode using m/z 79 as the common fragment ion for both MMS and EMS, whereas n-propyl mesylate (PMS) was introduced as the IS The authors found 051 and 131 ppm for MMS and EMS, respectively, in the bismesylate salt and did not detect either of these alkylating agents in the free base sample
Raman and others43 reported a GC-MS method for determining two process impurities, 2-(chloromethyl)-3,4-dimethoxypyridine hydrochloride (CDP) and dimethyl sulfate (DMS), in pantoprazole sodium (PPS) The control limit for CDP and DMS was calculated at 375 ppm based on the TTC and 40 mg daily dose of PPS The PPS sample was dissolved in H2O/methanol 3:1 (v/v) and was directly injected onto a DB-624 column (60 m × 032 mm, 18 μm film thickness) Quantitation was conducted under the
SIM mode, which is mass selective and thus highly selective The practical quantitation limit (PQL) for the method was 3 ppm for each analyte, and the method gave a linear response for CDP and DMS concentrations from 3 to 45 ppm with acceptable recovery The same group also reported another GC-MS method for determining two GTIs, methyl camphorsulfonate and ethyl camphorsulfonate, in esomeprazole magnesium (EOM)44 The EOM sample was prepared at a concentration of 50 mg/mL in 1,3-dimethyl imidazolidin-2-one The separation was conducted on a DB-5 column (30 m long × 032 mm id, 10 μm film thickness) The quantitation was carried out under SIM mode The limit of detection (LOD) of method was found as 3 ppm
Sarat and others45 developed and validated a GC-MS method for the determination of methyl methanesulfonate (MMS) and ethyl methanesulfonate (EMS) at partsper-million levels in undisclosed APIs The API samples were dissolved in methanol/ chloroform at 80:20 (v/v) at 600 mg/mL and directly injected onto a DB-624 capillary column (30 m × 053 mm, 3 μm film thickness) under splitless mode The SIM mode was used to quantify both the methanesulfonate esters with the common m/z 79 fragment ion The PQL for the method was 052 ppm for methyl methanesulfonate and 054 ppm for ethyl methanesulfonate A linear response from the PQL value to 15 ppm was observed along with good recovery for the spiked API samples The authors discussed the criteria for choosing appropriate organic solvents, including purity, sufficient solubility for the samples (to achieve the desired sample concentration), compatibility with GTIs, and no interferences in chromatography As many pharmaceuticals exist in salt forms, mixed solvents provide better solubility than pure organic solvents Also, the authors demonstrated acceptable precision without the use of an IS
Li and Sluggett46 reported a direct injection GC-MS method for determining the trace level of carbonic acid chloromethyl tetrahydro-pyran-4-yl ester (CCMTHP) in a β-lactam API The authors evaluated several analytical techniques including LC-MS, GC-FID, GC-ECD, and GC-MS during their method development and found that GC-MS provided the best detection sensitivity About 25 mg of the API sample was dissolved in 5 mL of acetonitrile The separation was conducted on a DB-5ms column (30 m long × 025 mm id, with 10 μm film thickness) The authors evaluated the impact of splitless versus split injection, API matrix effects, liner type, and injection temperature The β-lactam API undergoes thermal degradation, and one of the resulting degradants, tetrahydropyranol (THP), interferes the analysis for CCMTHP Li applied a 10:1 split ratio to limit the amount of nonvolatile species introduced to the GC column but maintain acceptable sensitivity To reduce background noise, quantitation was conducted in the SIM mode using m/z 49, which is not observed from the THP degradant (Figure 81) To minimize carryover and contamination, Li set the inlet temperature to 250°C The method showed a linear response from 10 to 1000 ppm with a limit of quantitation (LOQ) at 10 ppm The spiked recovery at 10 ppm was 82%
Ramakrishna and others47 developed a GC-MS method for the identification and determination of two GTIs (MMS and EMS) in imatinib mesylate (INM) A total of 100 mg of pure INM solid (or tablet powder equivalent to 100 mg of INM) was weighed in a 10 mL volumetric flask and dissolved with n-hexane The separation of MMS and EMS was obtained on a DB-1 capillary column (30 m × 025 mm id, 025 μm film thickness) An injection volume of 2 µL was found to provide acceptable sensitivity and peak shape Data were collected under the scan mode with an m/z
range of 35-130 amu The authors validated the method per International Conference on Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use guidelines For both the mesylate esters, the LOD and LOQ values were found to be 03 and 10 µg/mL, respectively Further, the response was linear within the range of 1-15 µg/mL for both compounds The method specificity was tested by performing recovery studies of MMS and EMS at 1, 10, and 15 µg/mL with three different batches of INM tablets The spiked recovery was between 98% and 100% with no observed interference from the excipients
The direct injection technique requires the introduction of concentrated sample solutions to the GC inlet, which often leads to contamination David and others48 reported a two-dimensional capillary GC method using Deans switching to address the contamination issue The authors chose an HP-5ms column (30 m × 025 mm id, with 025 μm
film thickness) for the first-dimension separation followed by a DB-17ms column (30 m × 025 mm id, with 025 mm film thickness) as the second-dimension separation for heart-cut fractions from the first dimension that contain the GTIs The second capillary column was installed in a low-thermal-mass oven, which allowed for independent temperature-programmed analysis Quantitation was conducted under SIM mode This two dimensional GC (2D-GC) setup significantly reduced the amount of undesired materials (such as APIs, solvents, and derivatization agents) introduced onto the second column or into the MS detector In this fashion, the contamination for the GC column/system was reduced, which helped to improve the method’s robustness The authors used the 2D-GC-MS system to analyze two classes of GTIs, that is, Michael reactive acceptor GTIs (cinnamonitrile and 3-ethoxy-2-cyclohexenone) and haloalcohols [2-bromoethanol, 2-iodoethanol, 4-chloro-1-butanol, 2-(2-chloroethoxy)ethanol, and 11-bromo-1-undecanol] Using the direct injection 2D-GC-MS approach, good resolution and high sensitivity for cinnamonitrile and 3-ethoxy-2-cyclohexenone were achieved with a LOD of 001 and 005 ppm (microgram per gram API), respectively (Figure 82) Derivatization with N,O-bis(trimethylsilyl)trifluoroacetamide (BSTFA)
helped to overcome the low volatility of the haloalcohols Good linearity and sensitivity were observed with LODs all lower than 05 ppm
Although direct injection is very convenient, various extraction techniques, including LLE, SPE, SPME, or liquid-phase microextraction (LPME), are often considered, especially when sample solubility is limited or when there are severe interferences from the sample matrix Although certain extraction techniques are labor intensive and often raise concern for recovery, they can also be effective in enriching GTI analytes and reducing the matrix effects
DMS is an alkylating reagent commonly used in organic syntheses and pharmaceutical manufacturing processes However, in-process testing for DMS is challenging because of its reactivity and complex matrix effects Zheng and others49 reported a GC-MS method for determining DMS in a water-soluble API intermediate, which was a methylsulfate salt DMS and an IS, d6-DMS, were extracted from the matrix with methyl tert-butyl ether (MTBE) GC separation was conducted on a DB-624 column (30 m long × 032 mm id, 18 μm film thickness) MS detection was performed on a single-quad Agilent mass spectrophotometric detector (MSD) equipped with an electron impact source while the MSD signal was acquired in SIM mode This GC-MS method showed a linear response for DMS from 10 to 60 ppm The practical quantitation limit for DMS was 10 ppm, and the practical detection limit was 03 ppm The relative standard derivation for the analyte response was 01% for six injections of a working standard equivalent to 186 ppm of DMS The spike recovery ranged from 1021% to 1085% for a sample of API intermediate spiked with 8 ppm of DMS The DMS level was 59 ppm, which was much lower than the targeted control limit in the isolated intermediate Also, it suggests that the quenching process and subsequent work-up are efficient for removing the unreacted DMS
During method development, Zheng and others49 screened several solvents for LLE MTBE was selected because it provided a clean GC-MS background and a higher recovery for DMS To compensate for the loss of DMS during the LLE procedure and injection, the authors investigated the use of three structurally similar analogs including two isotopic analogs, [13C2]-DMS and d6-DMS, as ISs (Figure 83) The two isotopic analogs, [13C2]-DMS and d6-DMS, showed similar electron ionization-mass spectrometry (EI-MS) spectra, recovery, and chromatographic retention as DMS, which made them more suitable for use as the IS However, since [13C2]-DMS and d6-DMS coeluted with DMS, the SIM ions for quantitation must be chosen carefully to avoid those overlapping ions in the EI-MS spectra of the analyte and IS The quantitation limit for DMS when using [d6]-DMS as an IS is five times better than using [13C2]-DMS as the IS This is because of the relative abundance of the quantitative ions chosen for each IS In the DMS EI-MS spectrum, the m/z 95 ion (SIM ion for d6-DMS) is much more abundant than m/z 125 (SIM ion for [13C2]-DMS) (Figure 83a)
Nassar and others50 reported a GC-MS method to investigate the formation and rate of hydrolysis of EMS in the BMS-214662 mesylate drug substance and its parenteral formulation The BMS-214662 mesylate drug substance was manufactured, and one step involved crystallization in ethanol Also, the parenteral formulation contains
300 mg/mL of ethanol in a pH 4 buffer The concern is that a reaction between residual methane sulfonic acid (MSA) and ethanol would form EMS In this study, the formulation samples were first spiked with isopropyl mesylate (IMS) as an IS and were then extracted with chloroform After centrifugation, the chloroform layer was injected into a DB-WAX column (30 m long × 025 mm id, with 025 μm film thickness) for analysis Quantitation was conducted under SIM mode using m/z 79 for EMS and m/z 123 for IMS (IS) The LOQ was 50 ppb (vs BMS-214662 mesylate API) The authors detected EMS in five batches of BMS-214662 mesylate API ranging from 02 to 08 ppm EMS levels in the parenteral formulation showed no significant increase after storage at 25°C for 18 weeks or at 60°C for 6 weeks The authors concluded that EMS formation between ethanol and MSA may not occur in the BMS-214662 formulation under the storage conditions
Wollein and Schramek51 reported a direct injection GC-MS method for the simultaneous determination of MMS, EMS, IMS, methyl besylate (MBS), and ethyl besylate (EBS) in finished drug products Generally, fine powder or solid dosage forms of the finished drug product containing around 25 mg of API were extracted with n-hexane spiked with methyl tosylate (MTS) as an IS Then, 1 µL of supernatant was immediately analyzed by GC-MS with a Restek Rxi 5-Sil MS (30 m × 032 mm, 05 μm film thickness) capillary column The SIM parameters were set at m/z 65, 80, and 95 for MMS; m/z 79, 97, and 109 for EMS; m/z 43, 79, and 123 for IMS; m/z 77, 141, and 172 for MBS; m/z 51, 77, and 141 for EBS; and m/z 91, 155, and 186 for MTS, although only one ion was used for quantitation Good sensitivity and linearity (R2 value ≥ 09998) were observed for concentrations between 001 and 133 µg/mL Spike recoveries for MMS, EMS, and IMS were greater than 71% Spike recoveries for MBS and EBS were greater than 94% MMS, EMS, and IMS were not detected in two lots of bromocriptine mesylate and two lots of doxazosin mesylate tested The authors suggested the use of n-hexane as the LLE solvent due to poor solubility of the API The choice of n-hexane helped to maintain GC-MS detection sensitivity by avoiding matrix interferences Additionally, n-hexane provides better stability for these alkyl sulfonates as solvents with higher polarity or nucleophilic properties were known to degrade the analytes of interest The disadvantage of using n-hexane was that liquid dosage forms could not be analyzed
Colon and Richoll52 reported an SPME GC-MS technique to determine methyl and ethyl esters of methanesulfonic, benzenesulfonic, and p-toluenesulfonic acids in APIs as a limit test (not more than 5 ppm) Typically, the APIs (which were mesylate, besylate, or tosylate salts) were dissolved in a 47 pH phosphate buffer and were then extracted using a PDMS/DVB fiber installed on an SPME device under constant agitation and at room temperature All samples were analyzed using GC-MS with an SPME injection insert, and the inlet held at 230°C operated in the pulsed splitless mode The analytes were resolved with a DB-1701 column (30 m × 025 mm, 10 μm film thickness) and quantified with SIM detection (Figure 84) The authors optimized the SPME fiber coating, extraction time, and buffer pH The validated limit test method provided acceptable reproducibility (relative standard deviation < 6% for five replicate sample preparations), linearity with or without API present (R2 value > 09), and spike recovery (within ±10% spike level) The authors demonstrated successful application of the method in the analysis of four batches of APIs with different structures Finally, the authors suggested that SPE and LPME could provide viable alternatives to the SPME methodology in cases where the API has limited solubility in aqueous media
Very recently, Ho and others53 explored the feasibility of using polymeric ionic liquids (PILs) as a new type of SPME sorbent coatings for the analysis of alkyl halides, aromatics halides, epoxides, and aromatic amines The separation of all analytes was conducted on an HP-5ms capillary column (30 m × 032 mm id, 025 μm film thickness) A glucaminium-based PIL, N,N-didecyl-N-methyl-D-glucaminium poly(2-methylacrylic acid 2-[1-(3-propylamino)-vinylamino]-ethyl ester), showed good sensitivity for long-chain aliphatic alkyl halides A poly(1-vinyl-3-propylphenylimidazolium) chloride PIL coating showed good selectivity for larger aliphatic/ aromatic analytes Excellent recovery was observed for both PIL-based coatings with limits of detection ranging from low parts-per-billion to mid parts-per-trillion levels
Garcia and others24 reported a GC-MS method for the determination of 2-chloroethanol (2-CE), a well-known alkylating agent, in cloperastine fendizoate The cloperastine fendizoate sample was dissolved in methanol (basified with ammonia, pH = 90) and loaded onto a strong anion exchange (SAX)-SPE cartridge to remove the fendizoate The eluent was subjected to GC-MS analysis using a VF-23ms capillary column (30 m × 025 mm, 025 μm film thickness) Quantitation of 2-CE was conducted under SIM mode for m/z 80 (the molecular ion of 2-CE) The most abundant fragment ion m/z 49 was not selected for quantitation due to matrix interferences The method gave good specificity and a 17 ppm LOD This GC-MS method was successfully applied for the determination of 2-CE in five different batches of cloperastine fendizoate
Martano and others54 developed and validated a GC-MS method for the simultaneous determination of four prominent volatile cleavage products (CPs) of β-carotene, a precursor of vitamin A, in cell culture media Typically, these CPs, that is, β-ionone (β-IO), cyclocitral (CC), dihydroactinidiolide (DHA), and 1,1,6-trimethyltetraline (TMT), were extracted with a Strata Phenyl SPE column and then eluted with 10:90 (v/v) tetrahydrofuran/n-hexane The SPE eluent was subjected to GC-MS analysis using a DB-20 (WAX) column (30 m × 025 mm, 050 μm film thickness) The quantitation was conducted under SIM mode using linalool and methylisoeugenol as ISs The method was linear with a LOD of 530 ng/mL The method also gave reasonable recoveries (717%– 957% at 10 μg/mL) and precision (<20%, intraday, n = 5; <48%, interday, n = 15)
Headspace GC is a common technique in pharmaceutical analysis that has been widely utilized to determine residual organic solvents remaining in pharmaceuticals As many GTIs are volatile species and the APIs are generally less volatile, headspace GC-MS offers certain advantages over direct injection Headspace injection often leads to reduced sample contamination as the nonvolatile API sample matrix remains in the vial Furthermore, a large amount of sample can be placed in the vial, which allows for higher sensitivity without the need for higher sample solubility as in direct injection
The synthesis of filibuvir, a hepatitis C virus polymerase inhibitor candidate, involves the use of 2,2,6,6-tetramethyl-1-piperidinyloxy (TEMPO) in the second to last step TEMPO in filibuvir needs to be controlled to 4 or 8 ppm depending on the clinical study duration per the staged TTC Strohmeyer and Sluggett55 developed and validated a headspace GC-MS method for detecting low levels of TEMPO in filibuvir Headspace injection was chosen for analysis because TEMPO is volatile and headspace injection would avoid the contamination of injector and column with the less volatile API The filibuvir sample was dissolved in acetonitrile/water 1:1 (v/v) at a concentration of 100 mg/mL The headspace vial was equilibrated at 60°C for 15 minutes Separation was performed on a VF-624ms capillary column (30 m × 032 mm id, 18 μm film thickness) or a DB-35ms column (15 m × 025 mm id, 025 μm film thickness) Quantitation was conducted in SIM mode using m/z 156 (M+∙) and m/z 142 (M+–CH2∙) This “fit-for-purpose” method demonstrated good specificity for the detection of TEMPO with filibuvir present and exhibited acceptable linearity over a range of 04-60 µg/mL (equivalent to 4-60 ppm vs filibuvir)
with LOQ and LOD values at 4 and 2 ppm, respectively The method also showed appropriate repeatability and spike recovery A total of 13 batches of filibuvir were subjected to analysis and TEMPO was found to be undetectable (≤2 ppm) By conducting a purge study, the authors demonstrated that the API synthetic process can reject TEMPO well below the desired levels
Teasdale and others56 investigated the mechanism for the formation of methyl mesylate from methanol and MSA in an API process The reaction mixtures were analyzed using a static headspace GC-MS method A 20 mL headspace vial containing 10 μL of MSA and 100 μL of methanol was heated at 78°C for 2 hours and then equilibrated at 105°C for 15 minutes A total of 1 mL of headspace gas was injected with 10:1 split ratio Separations were conducted on a DB-VRX column (60 m × 025 mm, 14 μm film thickness) Detection was performed in scan mode (scan range: 10−300 m/z) The authors were able to gain valuable insight into the reaction profile for these GTIs
Since many GTIs are highly reactive, they are prone to degradation during analysis However, these reactive GTIs can be derivatized to provide a stable compound more amendable to analysis The specificity of GTI analysis may even be enhanced by choosing derivatization reagents that provide a selective reaction with the target GTIs Furthermore, the resulting derivatized compounds are often less polar, more volatile, and readily analyzed by GC-MS
Lee and others40 reported a derivatization headspace GC-MS method to analyze alkylating agents, including methyl, ethyl, and isopropyl methanesulfonate esters and dimethyl sulfonate In this study, the basic drug substance sample (a methanesulfonate salt crystallized from methanol and ethanol) was dissolved in water at 100 mg/mL and then the derivatization reagent sodium thiocyanate was added immediately prior to sealing the headspace vial Derivatization was carried out at 85°C, which gave a mixture of alkylthiocyanates and alkylisothiocyanates These alkylthiocyanate and alkylisothiocyanate derivatives are volatile and readily analyzed by headspace GC (Figure 85a) The separation was conducted on a SGE 50QC3/BPX5 05 column (50 m long × 032 mm id, with a film thickness of 05 μm) Quantitation of alkylthiocyanates and alkylisothiocyanates was conducted under SIM mode using their molecular ions The authors found that derivatization with 10% sodium thiocyanate and heating at 85°C for 20 minutes provided the optimum conversion, that is, 100% for methyl methanesulfonate and 74% for ethyl methanesulfonate, respectively The authors also found that the interference from residual alcohols at less than 1% was almost negligible The method showed linear responses for concentrations from 002 to 05 μg/mL and used an IS Detection limits were found to be 002 μg/mL for both methyl and ethyl esters, and 005 μg/mL for the isopropyl esters The spike recoveries for the methyl and ethyl mesylate esters were 108% and 94%, respectively Certain limitations for this method were also noted First, the reaction of alkyl mesylates with thiocyanate is complicated due to isomerization (Figure 86) Second, the use of sodium thiocyanate for derivatization
requires the samples to be highly soluble in aqueous solutions As a consequence, partial hydrolysis of these alkylating agents was inevitable, especially for IMS The authors suggests that replacing the aqueous solution with polar aprotic solvents like DMF or 1,3-dimethyl-2-imidazolidinone would improve solubility and avoid the hydrolytic degradation of the alkyl mesylates
Alzaga and others22 developed another derivatization headspace GC method for the simultaneous analysis of several common alkylating agents, including mesylates, besylates, tosylates, and sulfates To accommodate APIs or intermediates with different solubilities and polarities, the authors used a dimethyl sulfoxide (DMSO)/water mixture (1:1, v/v) as the sample diluent After adding the deuterated ISs and pentafluorothiophenol, the derivatization was conducted at 105°C for 15 minutes (Figure 85b) The reaction of the investigated alkylating agents with pentafluorothiophenol was less complex compared to reactions using sodium thiocyanate40 The resulting derivatives from pentafluorothiophenol were analyzed by headspace GC using a RH-624 column (30 m long × 032 mm id, with 18 μm film thickness) The EI-MS signal was monitored under SIM mode using the corresponding molecular ions A standard addition method with ISs was adapted for quantification to compensate for the matrix effect The introduction of deuterated analogs as ISs improved the method’s accuracy and precision by compensating for variables in the derivatization and extraction/injection processes In addition, the deuterated ISs reduce the opportunity for reporting false negative results Compared to an earlier report by Lee and others,40 this method showed better throughput, selectivity, analyte stability, and detection sensitivity This procedure was successfully applied to different pharmaceutical matrixes and is particularly suitable for routine analysis with high sample throughput
As part of industrial collaborative efforts to understand the formation mechanisms and kinetics of sulfonate esters, Jacq and others57 adopted the derivatization headspace GC-MS approach used by Alzaga and others22 and developed an automated procedure to monitor the formation of ethyl mesylate from reaction mixtures containing ethanol and MSA The authors combined a liquid-handling robot with a static headspace module for automatic sample preparation The EMS formed was analyzed after derivatization with pentafluorothiophenol using headspace GC-MS (Figure 87) Linear response was obtained for EMS concentrations between 5 and 500 µg/mL The limit of detection and limit of quantitation for EMS were 05 and 10 μg/mL, respectively Excellent linearity, repeatability, and robustness were obtained, allowing the system to be used in several kinetic studies56,58
Sun and others59 reported an in situ derivatization headspace GC-MS method for the determination of hydrazine, a known genotoxic compound, in drug substances at low parts-per-million levels The authors chose acetone or acetone-d6 as the derivatization reagent for hydrazine because the resulting acetone azine with a moderate boiling point of 133°C is suitable for headspace GC-MS analysis Note that acetone-d6 was considered as an alternative reagent when acetone was used as a solvent in the API-manufacturing process In this study, the derivatization reagent was prepared by dissolving 05 g of benzoic acid and 05 mL of acetone or
acetone-d6 in 10 mL of N-methyl-2-pyrrolidone (NMP) since NMP was found to be a universal solvent to dissolve the tested APIs After mixing 10 µg of the API sample with 100 µL of the derivatization reagent and 10 µL of the diluent of 01% ethylenediaminetetraacetic acid (EDTA) in a 10 mL headspace vial, derivatization was conducted by incubating the vial at 100°C for 10 minutes The derivatization reaction was driven to completion during headspace incubation The resulting acetone azine products were analyzed using an Agilent DB-624 column (25 m long × 02 mm id, with 112 μm film thickness) and were quantified by SIM mode (m/z 112 for the acetone derivative and m/z 124 and m/z 106 for the acetone-d6 derivative) The method showed a linear response for concentrations ranging from 01 to 10 ppm and a LOQ of 01 ppm The authors investigated recoveries of hydrazine from different APIs with various reactive functional groups such as ketones, primary amines, or even Michael acceptors and from APIs with different salt forms, including free bases, HCl salts, and fumarate salts The spike recovery at the 1 ppm level was found between 79% and 117% This generic method has been utilized to determine hydrazine in APIs at GlaxoSmithKline, Pennsylvania, during the release and process development
GC-MS is one of the most widely adopted analytical methodologies for the analysis of GTIs in pharmaceutical products GC-MS is capable of quantifying a broad range of GTIs in intermediates, APIs, and drug products Appropriate sample preparation is critical to achieve adequate specificity, sensitivity, and robustness and may even be beneficial by reducing complex matrix effects Based on these attributes, GC-MS will remain a powerful tool in drug development to ensure that GTIs are appropriately controlled in drug substances and products for years to come
