Review pubs.acs.org/CR Genotoxic Impurities in Pharmaceutical Manufacturing: Sources, Regulations, and Mitigation † ‡ § ‡ Gyorgy Szekely,*, Miriam C. Amores de Sousa, Marco Gil, Frederico Castelo Ferreira,*, § and William Heggie*, † School of Chemical Engineering & Analytical Science, The University of Manchester, The Mill, Sackville Street, Manchester M13 9PL, United Kingdom ‡ Department of Bioengineering and Institute for Bioengineering and Biosciences (iBB), Instituto Superior Tecnico,́ Universidade de Lisboa, Avenida Rovisco Pais, 1049-001, Lisbon, Portugal § Hovione FarmaCiencia SA, R&D, Sete Casas, 2674-506, Loures, Portugal *S Supporting Information 4.1.1. Altering the Synthesis AC 4.1.2. Adjusting Reaction Conditions To Miti- gate GTI Formation AC 4.1.3. Quality by Design AE 4.2. API Purification AF 4.2.1. Purge Factors AF 4.2.2. Separation Technologies AG 5. Conclusions and Future Trends AL Associated Content AM Supporting Information AM Author Information AM Corresponding Authors AM Notes AM CONTENTS Biographies AM Acknowledgments AN 1. Introduction A References AN 2. Genotoxicity: Mechanisms, Risk and Regulation C 3. Chemical Classes of Common Genotoxic Impur- ities E 1. INTRODUCTION 3.1. Genotoxic Compounds Used as Reactants F Most pharmaceutical products are manufactured either by 3.1.1. Alkyl Halides F applying a total synthesis approach or by modifying a naturally 3.1.2. Dialkyl Sulfates I occurring product. In both cases, a wide range of reactive 3.1.3. Epoxides K reagents are used. Therefore, it is natural that low levels of such 3.1.4. Hydrazines L reagents or side products are present in the final active 3.1.5. TEMPO N pharmaceutical ingredient (API) or drug product as impurities. 3.1.6. Aromatic Amines O Such impurities may have unwanted toxicities, including 3.1.7. Boronic Acids P genotoxicity and carcinogenicity. The risk for patient’s health 3.2. Genotoxic Compounds Formed in Side caused by the presence of small molecules as impurities in APIs Reactions R has become an increasing concern of pharmaceutical 3.2.1. Sulfonate Esters and Their Precursors. companies, regulatory authorities, patients, and doctors alike. Overview R Thus, pharmaceutical regulatory agencies such as the Food and 3.2.2. Sulfonate Esters and Their Precursors Drug Administration (FDA) and the European Medicines Used in Stoichiometric Amounts S Agency (EMA) have raised concerns regarding the presence of 3.2.3. Sulfonate Esters and Their Precursors genotoxic impurities (GTIs) in APIs that could impact Used in Catalytic Amounts X negatively on human health. 3.2.4. Alkyl Halides Y There is an increasing scientific interest in this field, as is 3.2.5. Acetamide Y illustrated in Figure 1, from data obtained from the ISI Web of 3.3. Genotoxicity and Carcinogenicity of Com- Science showing the number of publication hits on mon Organic Solvents AA “genotoxicity” and on “genotoxic impurity”.1 The graph based 4. Approaches for GTI Mitigation in the Pharma- ceutical Industry AA 4.1. Chemical Synthetic Approaches AB Received: March 7, 2012 © XXXX American Chemical Society A DOI: 10.1021/cr300095f Chem. Rev. XXXX, XXX, XXX−XXX Chemical Reviews Review genotoxicity according to chemical structures. These systems follow either rule-based or quantitative structure−activity relationship models (QSAR). Rule-based systems are derived from identified mechanisms of action of chemicals in the cell genome or metabolic proteins. This approach was introduced by Miller and Miller in 19774 and followed by other authors. In spite of its mechanistic clarity, it has been criticized as being based only on single interactions and therefore failing to be comprehensive. QSAR models may use several inputs simultaneously, e.g., information on Ames test results, log P, molecule polarity and electrical distribution, and chemical substructures. The use of QSAR models is particularly useful for the prediction of the biological effect of a broad range of chemicals and new molecules with a high degree of accuracy. This subject is further explored in several studies, and examples include comparison of the use of three models for prediction of 5 ff 6 Figure 1. Importance of genotoxicity demonstrated by the increasing Ames genotoxicity and presentation of di erent case studies. number of publications on the topic, resulting from an ISI Web of QSAR models commonly used for determination of structural Science search on “genotoxicity” and “genotoxic impurity”. alerts to predict genotoxicity are the MULTICASE and the deductive estimation of risk from existing knowledge (DEREK) 7−9 on the former search shows the overall importance of the field ones. However, for numerous chemical classes, structural of genotoxicity, including chemistry, analytical methods, alerts overpredict mutagenicity when they do not take into manufacturing, purification, diseases, medical aspects, genotox- account factors such as high molecular weight, hydrophilicity, icity tests, mechanism of action, assessment, and environment. high reactivity, steric hindrance, molecular symmetry, and facile 10,11 The latter search illustrates the increasing attention of industry metabolism. to GTIs, mainly related to drugs and food. On the other hand, their presence in the manufacture of APIs Compounds categorized as GTIs actually include a broad is not stochastic, since these genotoxic chemicals often have range of unrelated chemicals with very different structures and specific inherent roles in the chemical routes used in API from very different chemical families. From 4000 compounds synthesis. The presence of such chemical in the reaction is a tested, 44 molecular structures were correlated with muta- result of their introduction into the reaction in stoichiometric genicity and correlated highly with electrophilic reagents, such or catalytic amounts or as solvents, as well as their formation as as epoxides (63%), aromatic amines (49%), and primary alkyl side products. The presence of genotoxins is usually inherently monohalides (46%).2 Aromatic amines are not electrophiles, controlled during API manufacture, as several stages of but their decomposition leads to the formation of electrophilic intermediate API isolation and purification are included in reactive species such as aryl nitrenium ion. In section 3.1.6 the production process, during which most of the GTIs are, examples of aromatic amine reactants are described. These together with other impurities, removed. Additionally, many of compounds have a shared ability to react with DNA, resulting the synthetic reaction sequences initially designed for in an associated carcinogenic risk. However, from a chemical production of new drugs are often further improved through point of view, they do not have common chemical−physical optimization of reaction conditions or by substituting with properties or chemical structural elements that can contribute different reaction steps. Such improvements aim at higher to easy identification. Experimental assessment of genotoxicity yields, reaction selectivity, and more efficient use of reactants, test models, such as the Ames test, allows direct study of which results in lower amounts of unreacted compounds and genotoxicity, and the Committee for Medicinal Products for side products formed. Nevertheless, production of APIs with Human Use (CHMP) has defined GTIs as impurities that have low GTI content is a major concern for API-manufacturing been demonstrated to be genotoxic using such genotoxicity test companies. Ideal solutions consist of the simplest possible, models. The Ames test, developed in the early 1970s by Bruce robust process, using cost-effective reagents to obtain high N. Ames, is an experimental procedure to evaluate the potential product yields through selective reactions and purification carcinogenicity of chemicals, based on mutagenicity effects on steps. Development and validation of such processes in a timely Salmonella typhimurium histidine auxotrophic mutants strains. manner are important for the industry, and as such, it is It became widely used due to its simplicity, low cost, and quick important to be aware of the chemical mechanisms in which analysis without the need for animal testing. genotoxic compounds are involved, whether as reagents or As discussed in ICH Q3A and Q3B, actual impurities in API reaction side products, and of existing strategies to circumvent are the ones that exceed the reported threshold when the lot is their use or remove them from postreaction streams. released or arise, for example, as degradation product, during In addition to the Introduction and concluding remarks, this storage and distribution over the shelf life of the API, whereas review includes the following sections: potential impurities may or not actually be present in the API Section 2 provides a brief description of genotoxic but are identified as the ones that can theoretically arise during mechanisms and a risk analysis, as well as the regulatory manufacture or storage. In the particular case of GTIs, approaches taken concerning this issue. Further reviews on “potential GTIs” are the ones that have structural alerts, i.e., specific topics of risk assessment,12 toxicology,13 and functional chemical groups, for genotoxicity but have not been mechanism of action14 of such compounds can be found in experimentally assessed; note that here potentially is not related the literature. with the presence or absence of the impurity.3 In silico systems Section 3 of this review focuses on GTIs related to starting are commonly used to identify structural