The Determination and Control of Genotoxic Impurities in APIs

Starting materials, reagents, intermediates, byproducts, and degraded products are often found as impurities in active pharmaceutical ingredients (APIs). Some of these known impurities are mutagens or carcinogens with the potential to cause adverse effects on the human body, even at trace levels. Control and determination of these impurities at parts-per-million or parts-per-billion levels are significant challenges for analysts. When developing synthetic routes to APIs, it is the primary responsibility of laboratory personnel to identify the stages in which impurity generation can occur. The analyst must also identify and determine genotoxic impurities (GTIs) and control them at the stages of formation (see Figure 1). The toxicologist must perform safety evaluations of high-priority compounds, known API impurities, and impurities with a high probability of occurring, and subsequently classify these compounds as genotoxic or routine impurities and propose limits. Various chromatography and spectroscopy methods help identify GTIs in an API. To detect and quantify the required levels and determine the signal-to-noise ratio, a derivatization must be performed, provided interference does not occur. This article will describe how to select methods for determining and controlling GTIs, based on their properties.

Figure 1: Representative structures of potential genotoxic impurities. (ALL FIGURES ARE COURTESY OF THE AUTHORS)

Limits for impurities

Guidelines from the International Conference on Harmonization (ICH) and EMA provide the limits for impurities in drug substances and drug products (1–3). These limits do not apply to GTIs because of their adverse affects, hence it is necessary to determine limits based on the daily dose of the drug substance. This task drains process-development resources. To overcome this problem, scientists have to identify GTIs early in process development, develop analytical methods (i.e., for quantifying the genotoxic impurity), and demonstrate the necessary synthetic process controls.

EMA guidelines. EMA guidelines classify GTIs into two categories. The first, GTIs with sufficient experimental evidence for a threshold-related mechanism, is regulated using methods outlined in ICH Q3C (R4) for Class 2 solvents (4). For the second category, GTIs without sufficient experimental evidence for a threshold-related mechanism, EMA proposes a "threshold of toxicological concern (TTC)." A TTC value of 1.5 g/day intake of a GTI is considered to be associated with an acceptable risk (see Table I) (5). The concentration limit in ppm of GTI permitted in a drug substance is the ratio of TTC in μg/day and daily dose in g/day. EMA also released a "Question and Answer" document to clarify questions that arose from its original guidance (6, 7).

Table I: FDA and EMA recommend acceptable qualification thresholds for genotoxic impurities in pharmaceuticals used in clinical studies.

The Pharmaceutical Research and Manufacturers Association's approach. The Pharmaceutical Research and Manufacturers Association (PhRMA) published a procedure for the testing, classification, qualification, and toxicological risk assessment of GTIs (8). It listed functional groups known to be involved in reactions with DNA that could be used as structural alerts. These functional groups were categorized into aromatic groups (e.g., N-hydroxyaryls, N-acylated aminoaryls, aza-aryl N-oxides, aminoaryls, and alkylated aminoaryls), alkyl and aryl groups (e.g., aldehydes, N-methylols, N-nitrosamines, nitro compounds, carbamates, epoxides, aziridines, propiolactones, propiosultones, N or S mustards, hydrazines, and azo compounds), and hetero aromatic groups (e.g., Michael-reactive acceptors, alkylesters of phosphonates or sulfonates, haloalkenes, and primary halides).

PhRMA also categorized impurities into five classes. Class 1 impurities are genotoxic (i.e., mutagenic) and carcinogenic. These impurities represent the most serious risk, and the default preference is to eliminate them by modifying the process. If this is not possible, the TTC limit can be employed as a last resort. Class 2 impurities are genotoxic, but their carcinogenic potential is not known. These impurities are to be controlled using TTC principles. Class 3 impurities contain problematic structures unrelated to the structure of the API and of unknown genotoxic potential. This group includes impurities with functional moieties that can be linked to genotoxicity based on structure. Class 4 impurities contain problematic structures related to the API. These impurities contain a potentially worrisome functional moiety shared with the parent structure. Class 5 impurities have no problematic structures, and evidence indicates the absence of genotoxicity. These compounds are to be treated as normal impurities and controlled according to the ICH guidelines. If Class 3 or 4 compounds are genotoxic or not tested, they are moved into class 2. If these are nongenotoxic, they are considered as Class 5.

The concentration limit (in ppm) of a GTI in a drug substance can be calculated based on the expected daily dose to the patient and the TTC value using the following equation:

The TTC concept should not be applied to carcinogens where adequate toxicity data are available from long-term studies and allow for a compound-specific risk assessment.

Identification methods for GTIs. GTIs can be identified by the following methods:

• Consulting a list of known genotoxins

• Determining whether an element has functional groups similar to those of known genotoxins

• Performing genotoxicity assays

• Using computer-based structure-activity relationship software programs (e.g., MultiCASE's Mcase, Accelrys's Topcat, or LHASA's DEREK)

• Performing Ames bacterial mutagenicity testing after software provides a structural alert. Results from the Ames test are considered more definitive than the in silico data.

Selection of the analytical technique

Lou et al. reported a method-development strategy for the control of GTIs (9–11). Analytical techniques can be selected by dividing GTIs into two groups based on their volatility. High-performance liquid chromatography (HPLC) with UV detection should generally be the first choice for nonvolatile GTIs because of the methods' simplicity and availability (12). However, HPLC may not offer sufficient sensitivity for certain GTIs in low-level analysis. If GTIs offer insufficient UV response, ultraperformance or ultra fast liquid chromatography (UPLC or UFLC) can be used because they have enhanced UV-detector sensitivity. When GTIs lack chromophores, an evaporative light-scattering detector (ELSD) is the alternate choice. However, ELSD is limited in sensitivity and dynamic range. Refractive index detectors and fluorescence detectors are alternate detectors used in HPLC. Because low-quantitation limit establishment is challenging, coupling HPLC or UPLC with mass spectrometers (MS) significantly improves method sensitivity and speed (13, 14). These detectors are selective and minimize issues caused by interference in the sample matrix, thus improving data quality. However, these instruments are expensive and differ from vendor to vendor. Thus, transferring a method between the developing and receiving laboratories, which may use instruments from different vendors, requires optimization of multiple instrumental parameters (9).

Volatile GTIs, meanwhile, can be quantitated by gas chromatography (GC) with a flame ionization detector (FID) (14, 15). GC–FID in direct and headspace injection modes is generally the preferred method, but this depends on the properties of GTIs and sample matrices. An electron-capture detector (ECD) can be used when GTIs consist of halogens. Nitrogen–phosphorus detectors (NPD) offer an additional tool for GTIs containing nitrogen or phosphorus atoms. However, the applications of ECD and NPD are limited. GC–MS offers the most sensitive and selective detection and reduced background noise. The method also is less prone to interferences for low-level analysis of GTIs (16, 17). If the GTIs are labile, do not possess chromophores, and have reactive functional groups, they can be derivatized to form detectable species (e.g., hydrazine derivatizes with benzaldehyde to form 1,2-dibenzylidenehydrazine) (14, 18). The derivatization reagent should be selected according to the functional groups in the analyte. Derivatization helps in stabilization, incorporation of a unique structural moiety, enhancement of fluorescence, ionization for mass detection, and volatization for GC.

Method development and validation

The analytical technique developed to determine GTIs should take into account origin, control, and suitability at detection limits. The authors developed methods to identify and quantify the genotoxic impurities in APIs, including levetiracetam, salmeterol xinafoate, and montelukast sodium. Validation was performed to assess whether each method fulfilled its intended purpose. The parameters of specificity, detection limit, quantification limit, and accuracy were also evaluated.

Reagents. Methyl methanesulfonate (MMS), ethyl methanesulfonate (EMS), and methyl p-toluenesulfonate (MTS) were purchased from Sigma–Aldrich. Dimethyl sulfoxide, methanol, and acetonitrile were supplied by TCI American. All water used in the experiment was purified by an in-house Milli-Q system (Millipore). All drug substances used for validation and testing were obtained from current projects at Neuland Laboratories and prepared in house.

Instrumentation. A Shimadzu GC-2010 with headspace (HT3, Teledyne Tekmar) and autosampler (AOC-20i, Shimadzu), Perkin Elmer Clarus 600 gas chromatograph with headspace (Turbo matrix 40, Perkin Elmer) and Perkin Elmer Clarus 500 gas chromatograph with autosampler were used to determine the presence of selected GTIs in selected APIs.

Methods for determining GTIs in levetiracetam. Based on the route of synthesis for levetiracetam, the expected GTIs were 4-chloro butyryl chloride, methyl 2-bromobutyrate, and 2-bromo butyric acid. An analytical technique must determine these GTIs at 0.5 ppm. These GTIs are rare in APIs.

4-chloro butyryl chloride. 4-chloro butyryl chloride is a clear, colorless to yellow volatile liquid with a boiling point of 174 °C and a density of 1.26 g/mL. Based on the properties of the analyte, wall-coated capillary columns of various brands, phases, and dimensions were investigated. A nonpolar HP-005 column (30 m length, 0.53 mm i.d.) with a stationary phase of 5% phenyl dimethyl polysiloxane film of 5.0 μm is suitable for the determination of 4-chloro butyryl chloride at the 0.5 ppm level (see Figure 2). The column fulfills all the requirements of the method (i.e., high sensitivity and short run time). A Perkin Elmer Clarus 600 headspace auto sampler was used for method development and validation. Oven temperature was maintained at 60 °C for 5 min, and a linear thermal gradient of 10 °C/min to 240 °C was used with a final hold of 1 min with 150 °C injector temperature and 250 °C detector temperature. Helium was used as a carrier gas at constant pressure flow rate of 3.0 psi. Operation mode was splitless. Thermostat time was 20 min at 80 °C.

Figure 2: 0.5 ppm level chromatogram of 4-chloro butylryl chloride.

A stock solution of 4-chloro butyryl chloride in dimethyl sulfoxide was prepared and injected. Retention time was 8.289 min with good response. The limit of detection (LOD) and limit of quantification (LOQ) were 0.0494 ppm and 0.163 ppm, respectively. The method gave excellent precision and accuracy, even at the LOQ level.

Methyl 2-bromobutyrate. Methyl 2-bromo butyrate is a brownish liquid with a boiling point of 138 °C and density of 1.573 g/mL. A polar DB-FFAP column (30 m length, 0.53 mm i.d.) with a stationary phase of acid-modified polyethylene glycol film of 1.0 μm is suitable for the determination of methyl 2-bromo butyrate at the 0.5-ppm level (see Figure 3). A Shimadzu-2010 with autosampler was used for the method development and validation. Oven temperature was maintained at 80 °C for 5 min, and a linear thermal gradient of 5 °C/min to 150 °C was used with a final hold of 1 min at 170 °C injector temperature and 250 °C detector temperature. Helium was used as a carrier gas at constant pressure flow rate of 3.0 psi. Operation mode was split (2:1), and injection volume was 2.0 μL.

Figure 3: 0.5 ppm level chromatogram of methyl 2-bromo butyrate.

A stock solution of methyl 2-bromo butyrate in methanol was prepared and injected. Retention time was 12.223 min with good response. LOD and LOQ were 0.047 ppm and 0.156 ppm, respectively. The method gave good precision and accuracy even at the LOQ level.

2-bromo butyric acid. 2-bromo butyric acid is a colorless liquid with a boiling point of 214–217 °C and density of 1.567 g/mL. 2-bromo butyric acid was determined through derivatization with methanol to methyl-2-bromo butyrate. The method described earlier was used for the determination of 2-bromo butyric acid.

For the derivatization of 2-bromo butyric acid with methanol, 32.45 μL of 2-bromo butyric acid standard was placed in a 100-mL volumetric flask, and diluent was added up to the mark. The solution was refluxed for 4 h at 60–65 °C and then cooled to room temperature. Next, 1.0 mL of refluxed solution was added to a 100-mL volumetric flask, and diluent was added up to the mark.

Method for determining GTIs in salmeterol xinafoate. Based on the route of synthesis for salmeterol xinafoate, bromo acetyl bromide and 1,6-dibromo hexane are suspected genotoxic impurities. TTC values indicate the permitted level for both of these impurities is 1500 parts per million (ppm) (see Figure 4).

Figure 4: Selected concentration chromatogram for Bromo acetyl bromide and 1,6-dibromo hexane.

Bromo acetyl bromide and 1,6-dibromo hexane are colorless liquids with boiling points of 147 °C and 243 °C, respectively. Their densities are 2.317 g/mL and 1.586 g/mL, respectively. Based on the properties of these impurities, wall-coated capillary columns of various brands with various phases and dimensions were investigated. Nonpolar DB-WAX columns (30 m length, 0.53 mm i.d., Agilent J&W) with a stationary phase of polyethylene-glycol film of 1.0 μm are suitable for the determination of bromo acetyl bromide and 1,6-dibromo hexane at these concentrations. A Perkin Elmer Clarus 500 autosampler was used for the method development and validation. Oven temperature was maintained at 40 °C for 5 min, and a linear thermal gradient of 20 °C/min to 180 °C was used. A final hold lasted 10 min with 170 °C injector temperature and 250 °C detector temperature. Helium was used as a carrier gas at a constant pressure flow rate of 3.0 psi. Operation mode was split (2:1).

Stock solutions of bromo acetyl bromide and 1,6-dibromo hexane were prepared and injected. Retention times were 1.988 min for bromo acetyl bromide and 15.659 min for 1,6-dibromo hexane. The LOD and LOQ for bromo acetyl bromide were 30.33 ppm and 96.97 ppm, respectively. The LOD and LOQ for 1, 6-dibromo hexane were 100 ppm and 320 ppm, respectively. The LOD of bromo acetyl bromide was 50 times lower than the selected concentration.

Method for determining GTIs in montelukast sodium. The catalyst sulfonyl chloride is used to synthesize montelukast sodium, thus creating the possibility for the formation of mesylates when reacted with alcohols. These mesylates are genotoxic, and permitted concentrations are at the 75-ppm level (see Figure 5). Generally, mesylates are determined by gas chromatography–mass spectrometry–head-space sampling (GC–MS–HS) at 1 ppm. Here, however, the permitted level is higher, so GC–MS–HS is unnecessary. GC–FID is suitable for the determination of these GTIs in the selected concentrations. Depending upon the route of synthesis, the expected mesylates are methyl methane sulfonate and isopropyl methane sulfonate.

Figure 5: Chromatogram for mesylates.

Methyl methane sulfonate and isopropyl methane sulphonate are colorless liquids with boiling points of 202.5 °C and 220 °C and densities of 1.3206 g/mL and 1.15 g/mL, respectively. Based on the properties of these impurities, wall-coated capillary columns of various brands with various phases and dimensions were investigated. Nonpolar DB-1 columns (30 m length, 0.53 mm i.d.) with a stationary phase of 100% and dimethyl polysiloxane film of 5.0 μm are suitable for the determination of mesylates at these concentrations. A Shimadzu-2010 autosampler was used for method development and validation. Oven temperature was maintained at 100 °C for 5 min, and a linear thermal gradient of 10 °C/min to 200 °C was used with a final hold of 5 min at 170 °C injector temperature and 250 °C detector temperature. Helium was used as a carrier gas at a constant pressure flow rate of 3.0 psi. Operation mode was split (2:1).

Stock solutions of methyl methane sulfonate and isopropyl methane sulfonate were prepared and injected. Retention times were 10.021 min and 12.904 min, respectively. The LOD and LOQ were 4.38 ppm and 1.48 ppm for methyl methane sulfonate and 14.44 ppm and 4.87 ppm for isopropyl methane sulfonate, respectively.

Control of selected GTIs in APIs

GTIs can be well controlled by the process, without affecting quality, by modifying the route of synthesis. Methods for controlling sulfonates and alkylating agents are described below.

Sulfonates. Sulfonate ester formation is dramatically reduced at low amounts of water, or when acid is partially neutralized by substoichiometric concentrations of the weak base, 2,6 lutidine, to mimic conversion of a basic API to a salt. In the presence of a slight excess of base, ester formation was drastically reduced. These findings, particularly those involving an excess of base, are compelling and provide a scientific understanding to allow for the design of process conditions to minimize and control sulfonate formation (19).

Alkylating agents. Effective control of the alkylation process was possible because dimethyl sulfonate reagents barely reacted with nylon at room temperature (i.e., 25 °C). At 100 °C, the reaction could be stopped rapidly by immersing the tube in an ice bath. The optimum incubation times for alkylation with dimethyl sulfonate and diethyl sulfonate were 3 and 10 min, respectively. Dimethyl sulfonate, a less toxic reagent than dimethyl sulfonate, caused less damage to the nylon tubes and produced more chemically stable O-alkylated derivatives (20).

Conclusion

Identification and control of genotoxins in a synthetic process are challenging because of GTIs' evolving nature and variable points of entry. Hence, synthetic routes should be screened for the identification of structural elements that cause genotoxicity. Based on EMA guidelines, the TTC value of GTIs is 1.5 μg/day, but the value varies according to daily dose and length of use. Bearing in mind the sensitivity requirements of respective impurities, the authors developed an analytical matrix using GC methods. The analytical methods described here fulfilled the regulatory guidelines.

V. Gangadhar* is an executive in the Analytical Research Department, Y. Pardha Saradhi is a manager in the Analytical Research Department, and R. Rajavikram is a project trainee, all at Neuland Laboratories, Temple St., Bonthapalli Village, Jinnaram Mandal, Medak District 502 313 (A.P.), India, tel. +91 08458 392629, dhar.ctr@gmail.com.

*To whom all correspondence should be addressed.

References

1. ICH, S2 (R1) Guideline on Genotoxicity Testing and Data Interpretation for Pharmaceuticals Intended for Human Use (2008).

2. T. McGovern and D.J. Kram, Trends Analyt. Chem. 25 (8), 790–795 (2006).

3. ICH, Q3A (R2) Guideline on Impurities in New Drug Substances (2006).

4. ICH, Q3C(R4) Guideline for Residual Solvents (2009).

5. R. Kroes et al., Food Chem. Toxicol. 42 (1), 65–83 (2004).

6. EMA, Guideline on the Limits of Genotoxic Impurities (London, January 2007).

7. EMA, Guideline on the Limits of Genotoxic Impurities (February 2009).

8. L. Muller et al., Regul. Toxicol. Pharmacol. 44 (3), 198–211 (2006).

9. M. Sun, D.Q. Liu, and A.S. Kord, Org. Process Res. Dev. 14 (4), 977–985 (2010).

10. D.Q. Liu, M. Sun, and A.S. Kord, J. Pharm. Biomed. Anal. 51 (5), 999–1014 (2010).

11. D.Q. Liu et al., J. Pharm. Biomed. Anal. 50 (2), 144–150 (2009).

12. N.V.V.S.S. Raman et al., J. Pharm. Biomed. Anal. 48 (1), 227–230 (2008).

13. D.P. Elder et al., J. Pharm. Sci. 99 (7), 2948–2961 (2010).

14. L. Bai et al., J. Chromatogr. A 1217 (3), 302–306 (2010).

15. E.J. Delaney, Regul. Toxicol. Pharmacol. 49 (2), 107–124 (2007).

16. K. Ramakrishna, J. Pharm. Biomed. Anal. 46 (4), 780–783 (2008).

17. R. Alzaga et al., J. Pharm. Biomed. Anal. 45 (3), 472–479 (2007).

18. N.V.V.S.S. Raman et al., Talanta 77 (5), 1869–1872 (2009).

19. A. Teasdale, Org. Process Res. Dev. 13 (3), 429–433 (2009).

20. F.N. Onyezili, Analyst 114 (1), 789–791 (1989).