Competitive Collaborative and

Subscriber access provided by NEW YORK UNIV

Full Paper

Mutagenic Impurities: Precompetitive/Competitive Collaborative and Data Sharing Initiatives David P Elder, Angela White, James Harvey, Andrew Teasdale, Richard Vaughan Williams, and Elizabeth Covey-Crump Org. Process Res. Dev., Just Accepted Manuscript • DOI: 10.1021/acs.oprd.5b00128 • Publication Date (Web): 22 May 2015 Downloaded from http://pubs.acs.org on May 28, 2015

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Organic Process Research & Development is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Organic Process Research & Development

Mutagenic Impurities: Precompetitive/Competitive Collaborative and Data Sharing Initiatives Authors David P Elder‡*, Angela White‡, James Harvey‡, Andrew Teasdale∆, Richard Williams^, Elizabeth Covey-Crump^ ‡

GlaxoSmithKline R&D Ltd, Park Road, Ware, Hertfordshire, SG12 0DP,

United Kingdom ∆

AstraZeneca, Charter Way, Silk Road Business Park, Macclesfield, Cheshire SK10 2NX,

United Kingdom ^Lhasa Limited, Granary Wharf House, 2 Canal Wharf, Leeds, LS11 5PS, United Kingdom

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

TOC Graphic


Page 2 of 31

Page 3 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Abstract Pre-competitive and competitive collaborations have flourished during the development of the ICH M7 guidelines (mutagenic impurities). The majority of these interactions are ‘honest broker’ type arrangements, where an autonomous, independent third party is tasked with coordinating the activities of the consortium. Sometimes, more than one ‘honest broker’ can be involved in the activity. Lhasa has been at the centre of many of these collaborative initiatives; i.e., Derek Nexus Pharmaceutical Intermediates, Excipients, Zeneth, Mirabilis. The Product Quality Research Institute (PQRI) coordinated activities allowing a mechanistic understanding of the reaction mechanism for the formation of sulfonate esters, which after the Viracept withdrawal had been at the forefront of regulatory thinking. Other ‘not for profit’ organisations that have played a role in the advancement of the greater understanding of mutagenic impurities include the various trade organizations, i.e. the Pharmaceutical Research and Manufacturers of America (PhRMA) and the European Federation of Pharmaceutical Industry Associations (EFPIA). The other collaborative interactivity is the JDI (Just Do It). This is primarily focussed on analytical collaborative efforts, where individuals have collaborated on areas of mutual interest. To a lesser extent this has also been seen in the area of pharmaceutical intermediates, where individual companies have collaboratively pooled data on areas of common interest, e.g. acyl/sulfonyl halide false positives. The majority of the most significant JDI collaborations involve sharing feedback on the use and applicability of a second structure activity relationship system (SAR), analytical methodologies and analytical strategies. One of the key outcomes of these collaborations are peer reviewed scientific publications whereby other interested parties (including regulators) can be made aware of the findings from the consortia. Going forward, the area of mutagenic impurities enshrined under ICH M7, will continue to be a fruitful area for precompetitive collaborations of all types building on the success of the previous initiatives described.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Keywords: pre-competitive, pre-collaborative, ICH M7, mutagenic impurities, genotoxic risk assessment, honest broker, just do it, PQRI, scientific publications INTRODUCTION Precompetitive Collaborations Precompetitive collaborations are becoming increasingly common in the pharmaceutical industry 1. Woodcock2 has defined precompetitive collaborations as ‘a subset of translational research that is focussed on improving the tools and techniques needed for successful translation, and not on development of a specific product’. Welch et al.1 argued that enhanced cross-industry collaborations in this area will increase efficiency, innovation, sustainability and ultimately patient benefit. Generally, participants in collaborations will require ‘freedom-to-operate’ (FTO) within the sphere of the collaboration and want the ability to publish and otherwise disseminate the newly acquired information without significant impediments, e.g. royalty or licensing fees. There appear to be several different avenues open to potential collaborators. Just do It (JDI)/Formalized Collaborations, where the research is carried out by all of the participants and the data shared for mutual benefit. The only difference between the two is that the latter typically involves a signed agreement. Funding projects at external providers, is where the contributing pharmaceutical companies provide monies/time/effort for mutually beneficial research via an independent third party, e.g. Product Quality Research Institute (PQRI), International Society for Pharmaceutical Engineering (ISPE), Innovation and Quality Consortium (IQ) etc. The objective here is to demonstrate that the activities/findings have not been overly influenced by the pharmaceutical collaborators (i.e. independent research) and/or where the third party provides a ‘consensus voice’ on behalf of the industry collaborators, e.g. scalability of API design space from IQ consortium3.


Page 4 of 31

Page 5 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Different types of precompetitive collaborations Honest Broker is where a formal entity is created on behalf of the collaboration consortium; for example, the Consortium for the Investigation of Genotoxicity of Aromatic Amines (CIGAA)), joint venture or ‘not for profit’ corporations, e.g., Lhasa Limited, etc. Additional programs may be initiated after the inception of the initial undertaking(s). Additionally, these collaborative groups can utilize an independent thirdparty laboratory to generate data. Academic Consortia are where a group of academic laboratories collaborate with a governmental agency on areas of mutual interest and where the funding comes primarily from the governmental agency. FDA's innovative initiatives, including Critical Path4 and Advancing Regulatory Science5, emphasize that ‘bringing regulatory science into the 21st century requires the collaborative efforts of all stakeholders-including academia, industry, and other governmental agencies’.

Industry/ Academic Consortium are where a group of academic laboratories collaborate with a trade organization, e.g. European Federation of Pharmaceutical Industry Association (EFPIA), or multiple industry members, on areas of mutual interest and where the funding comes primarily from a national or supra-national agency. For example, the Innovative Medicines Initiative (IMI)6 is a joint undertaking between the European Union (EU) and the EFPIA. IMI supports ‘collaborative research projects and builds networks of industrial and academic experts in order to boost pharmaceutical innovation in Europe’.

ICH M7 One area that has benefited more than most from various different types of collaborations is the field of mutagenic impurities covered by ICH M7 guidance7. However, even prior to the publication of the EMA8 and draft FDA9 guidance on genotoxic impurities (now renamed mutagenic impurities) there were



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

flourishing cross-industry collaborative efforts to pool knowledge and understanding of this emerging topic. ICH M77 highlights anticipated considerations for both safety and quality risk management in establishing acceptable levels of mutagenic impurities that are expected to pose negligible carcinogenic risk. It outlines those recommendations for assessment and subsequent control of mutagenic impurities that may potentially carry over into the final API or drug product, taking into consideration the intended conditions of human use. ICH M77 predicates the need to perform a computational toxicology assessment using (Q)SAR approaches that are highly predictive of the outcome of a bacterial mutagenicity assay. The guideline mandates that two quantitative structure activity relationships (Q)SAR) prediction tools that are complementary with each other should be utilized. One approach should involve the use of an expert rulebased system (i.e. Derek Nexus10, Leadscope Genetox Expert Alerts11) whereas, the second (Q)SAR approach should be statistically-based (i.e. Sarah Nexus12, Leadscope Genetox Statistical QSAR13, Case ULTRA14, etc.). Therefore, the key challenges facing Industry were how to perform GRAs (genotoxic risk assessments), how to use (Q)SAR tools to predict potential (PMIs) and real mutagenic impurities (MIs). Equally important was how to analyse very low levels of these PMIs and MIs in the API and the drug product and were there any approaches that would minimise testing and focus testing only on those PMIs or MIs that would logically be anticipated to carry through to the API. Faced with a common challenge there was a huge impetus to share best practices and key learnings from ongoing development activities and regulatory intelligence. This article is aimed at summarising the most significant collaborative efforts and identifying key lessons learned. KNOWLEDGE BASED (Q)SAR APPROACHES Derek Nexus10


Page 6 of 31

Page 7 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Since the inception of genotoxic risk assessments (GRA), Derek Nexus10 has played a pivotal role in structure based assessments15. Derek Nexus is an expert knowledge base (Q)SAR system which contains expert knowledge rules (derived from both public and proprietary data) in toxicology, and particularly mutagenicity (within the ICH M77 context) and applies these rules to make in silico predictions about the toxicity of chemical intermediates, impurities, degradants (and potential impurities and degradants) and metabolites; usually in the absence of experimental data. Proprietary data donated by Lhasa members has been used, to a greater or less extent, in approximately 25% of the bacterial in vitro (Ames test) mutagenicity alerts in Derek Nexus10. Further, proprietary data sets are used to validate the performance of alerts for this, and other, endpoints, to provide an indication of predictive performance within the chemical space of most interest to users. The collaborations underlying these data donations have taken one of several forms 1. Members provide Lhasa with a data set to improve (or validate) alerts across a whole endpoint. This is then analysed to identify lead (mis-predicted) compounds for alert development. Whilst providing data directly is the most efficient method for alert development, in some cases this conflicts with member confidentiality policies, and in those cases work has been progressed via an on-site secondment to enable knowledge scientists from Lhasa to work behind the member’s firewall (see Secondments). Additionally, data share groups (such as Vitic Nexus16 intermediates and CIGAA17) have also been used for alert development (Cayley et al. 18). 2. Members contact Lhasa to indicate that an alert is insufficiently predictive within their chemical space, and provide the data to support an alert modification. In some cases, this includes discussions with Lhasa knowledge scientists that has led to directed testing within a class to either fill data gaps or resolve discordance within a data set. In all cases, the aim of the work is to convert sensitive, proprietary data into generic, publically-available structure-activity relationships. The level of disclosure is negotiable, with some members preferring to



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

remain completely anonymous (e.g. from alert 760 for 2-halopyridines: ‘four compounds donated by a Lhasa member which have all been reported non-mutagenic’) and others happy to disclose more information (e.g. from alert 475 for 3-aminomethyl-1,2,4-oxadiazoles ‘This alert is based on data from Hoffmann-La Roche AG and describes the Ames test activity observed for a series of 3-aminomethyl1,2,4-oxadiazole compounds’). The benefit of the work is the improved predictivity within the chemical space that is most important to members (both from Industry and Regulatory Agencies). Thus, Derek Nexus has been reported to show higher sensitivity within proprietary pharmaceutical chemical space19. Leadscope Genetox Expert Alerts11

The Leadscope Genetox Expert Alerts system11 satisfies the expert rule-based methodology requirement for ICH M77. It is based on a dictionary of structural alerts extracted from the public literature and elsewhere, and qualified using a large database of bacterial mutagenicity data (reference set). The alerts system encodes the structural reasons for activation and deactivation of the alerts and is supplemented with information on the mechanism. A prediction is made where the test chemical is unambiguously within the applicability domain of the alert system. If an alert is present and there are no deactivating fragments also present in the test chemical, then a positive classification is made; whereas the absence of any alert or the presence of a deactivating fragment will results in the test chemical predicted negative. A precision score is also generated showing the degree of confidence in the alert based on data in the reference set. Leadscope regularly conducts face-to-face meeting with users at their sites to ensure that the Leadscope expert alert’s system11 reflect proprietary chemical space. These meetings focus on both the overall performance of the system for proprietary chemicals as well as on specific alerts that can be improved based on proprietary data. Proprietary datasets are often examined under a confidentiality agreement and knowledge derived is added to the expert alerts system. Where the data is determined to be nonconfidential, these sets are transferred to Leadscope and used to improve the alerts. The data is also added


Page 8 of 31

Page 9 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


to the alert’s reference set and available to all users. Leadscope also works with regulatory agencies who have performed validations of the Leadscope Expert Alerts system against large databases and obtained 85% sensitivity and 78% negative predictivity20. In addition, all alerts along with the reasons for deactivation are validated against proprietary bacterial mutagenicity databases. This involved running the entire list of alerts including both deactivating and activating factors as structure searches over individual organization’s entire database of bacterial mutagenicity data. The results including the number of positive and negative examples for each alert are provided to Leadscope who then pools the data across all organizations. No information on individual chemicals or the data are exchanged. This information is in turn used to improve the quality of the expert alerts system. Statistically Based (Q)SAR Leadscope13 Leadscope has established a knowledge-sharing program with interested corporate sponsors to address specific (Q)SAR regulatory issues identified through discussions with sponsors and regulatory agencies. The initiative allows the use of proprietary corporate information to be investigated under confidentiality restrictions and identify potential solutions to specific predictivity issues or increase the number of compounds which can be predicted. Knowledge sharing can be in the context of improving predictivity for structurally similar compounds or more generally for improving the performance of entire alerts/(Q)SAR models. An example of such an interaction would be to improve the domain of applicability and prediction accuracy of specific classes of corporate compounds not well represented in the public domain. In such cases even the sharing of proprietary data for a single structure and allowing incorporation into the training set can be sufficient to increase the model coverage of proprietary chemistry for subsequent similar structures. The extent of data disclosure is determined on a case by case basis depending upon the sensitivity of the information provided. Full disclosure, where the structure and



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

data become incorporated into the (Q)SAR statistical models and alert-based expert system, helps to preserve model transparency but where required, compound confidentiality can be retained. Knowledge from proprietary data has also been used by Leadscope to increase the specificity for selected chemical classes. For example the specificity of a model to predict primary aromatic amines was increased by 14% with no decrease in sensitivity as a result of such data sharing with a single pharmaceutical company21. This improvement was successfully achieved by using proprietary data made available through confidentiality agreements without releasing any confidential business information. The process for knowledge sharing was through the use of structural fingerprints for several compound classes. A list of 591 chemical fingerprints based on the Leadscope fragment hierarchy, data analysis of a reference set and external knowledge containing a variety of primary aromatic amines were derived. The list of substructures includes meta-, para-, ortho-, hetero-substituted, polycyclic, as well as more complex substitution patterns. When these fingerprints were applied to a proprietary data set by the data owner(s) the result is a listing of named substructures present along with the number of positive and negative bacterial mutagenicity examples which are present. Only the results for the pre-defined substructures in the fingerprint are summarized and it is therefore possible to apply these fingerprints to a proprietary database without revealing information for individual compounds or data. This project now includes 13 pharmaceutical companies and regulatory agencies and has resulted in continued improvement in the performance around this class. This fingerprint methodology is also being applied to other chemical classes including boronic acids and alkyl halides. Sarah Nexus12 Sarah Nexus is a highly transparent statistical model for the prediction of mutagenicity. Created with input from the Food and Drug Administration (FDA) under a Research Collaboration Agreement22, the model has been derived using a unique, machine learning methodology23. When assessed by the FDA its predictive performance was found to be comparable to other widely-used commercial model systems,


Page 10 of 31

Page 11 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


with an output that was considered to provide sufficient transparency and interpretability for the qualification of pharmaceutical impurities under ICH M7.7 The overall performance of Toxtree, Leadscope Expert Alert System, and Sarah Nexus compare favorably to the most widely-used commercial model systems tested previously with the same data sets, and provide sufficient transparency and interpretability for the qualification of pharmaceutical impurities under ICH M77.

PHARMACEUTICAL INTERMEDIATES The pharmaceutical intermediates data sharing consortium run by Lhasa Ltd., currently involves 10 pharmaceutical companies. Members share Ames mutagenicity data for common intermediates and compounds containing functional groups of interest (e.g. nitro compounds, sulfonyl chlorides, aromatic amines, aryl boronic acids etc.), for which data needs have been specifically identified, either to fill gaps in the chemical space coverage (lack of data or conflicted data in the literature, need of further testing) and/or to improve in silico mutagenicity prediction tools (knowledge development, rule writing, inclusion/exclusion patterns). The latest database release contains Ames data on 969 intermediates. Organisations continue to donate data each year according to agreed requirements and decide functional group priorities. New organisations are joining every year. The project is currently heading towards its 8th release. The Intermediates data sharing project aims to limit expensive and time consuming testing and provide a high quality dataset (mainly 5 strain Ames data without and with S9 metabolic activation performed according to GLP or GLP-like studies, with indication of the test substance purity) for the development of new or existing in silico tools. Some specific examples are provided below. Boronic Acids Collaborations can be used to quickly and efficiently obtain data and knowledge in order to rapidly answer important questions as they arise. The mutagenicity of arylboronic acids was first discussed in the public domain in 201024,25. This was closely followed by a publication26, which directly led to the



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 31

implementation of a boronic acid alert in Derek Nexus10 Nexus. Further discussions on the mutagenicity of arylboronic acids have since been published27-30, indicating the current interest in the area. Although there is a publicly available dataset consisting of 36 arylboronic acids and derivatives it is not possible to formulate a robust structure-activity relationship owing to the vast area of chemical space covered by this class and the fact that there is little mechanistic understanding. In order to rapidly increase the dataset, a consortium consisting of eight pharmaceutical companies was established. Testing via this consortium not only resulted in the dataset being rapidly increased by preventing duplicate testing, but later enabled testing to be targeted towards specific areas of chemical space. Through the consortium, a larger dataset consisting of 90 arylboronic acids and derivatives was obtained and centrally stored31, resulting in refinement of the alert in Derek Nexus10, despite the continuing lack of a clear mechanistic rationale. In addition to the bacterial mutagenicity data (Ames test), the consortia has actively driven programs to further evaluate the potential mechanism of action through the conduct of other genotoxicity in vitro studies using the data gathered in mammalian in vitro systems (rodents or human): mutagenicity (mouse lymphoma assay), clastogenicity (chromosome aberration assay, micronucleus test, comet assay) to increase knowledge / understanding. Aromatic Amines (AAs) Aromatic amines (AAs) are key intermediates of numerous pharmaceutical drugs, food additives and cosmetics; however, a significant number of this class of reactive intermediates demonstrate mutagenic activity in the Ames test. The ability to predict using (Q)SAR tools, which aromatic amines are likely to be non-mutagenic would significantly reduce redundant testing and allow re-prioritization of resources onto those intermediates that are truly mutagenic and where there are genuine human safety concerns. The screening of AAs is also critical during the candidate drug design phase given the fact that many of the potential leads take the form of amides / sulfonamides where the ‘embedded’ AA may be both a degradant and a metabolite. A (Q)SAR study highlighted the importance of electronic and steric factors


Page 13 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


for both promoting and alleviating mutagenic activity, the knowledge from which was used to improve the predictive performance of the expert system, Derek Nexus10. However, additional assessments of Derek Nexus using external test sets32, demonstrated that this additional data, in itself, was not sufficient to improve the predictivity of this in silico system. Following, a Precompetitive Workshop at the Royal Society of Chemistry33 in London, a collaborative cross-pharma project was initiated with the aim of creating a database of aromatic amine mutagenicity information that could be used to facilitate regulatory submissions and to enhance the performance of these in silico predictive models. The resulting Consortium for the Investigation of Genotoxicity of Aromatic Amines (CIGAA)17 aims to enhance the understanding and predictability of the Ames test for primary aromatic amines. This precompetitive data sharing group consists of 13 pharmaceutical companies who share data for drug intermediates and precursors, with the objective of building a database to facilitate the development of improved predictive tools. The existing database contains at the time of writing, 300 compounds with 3639 associated genetic toxicity in vitro records. Additionally, some data analysis are ongoing within the consortium challenging experimental Ames data from the CIGAA database with various (Q)SAR model predictions. Moving towards its 3rd release, the project is currently in its data collection stage17. Acyl/Sulfonyl Halides The very reactive acyl halides/sulfonyl halides have historically elicited a lot of regulatory focus. In the early stages of the implementation of the EMA8 guidance, one company had been requested to provide analytical data in support of the contention that thionyl chloride (a representative member of the sulfonyl halide class) was not present in API, despite the fact that this reagent is very reactive and had been introduced several stages up-stream from the final API. In fact, so reactive is thionyl chloride that the company found that it was extremely difficult to develop any analytical strategies to measure it within the



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 31

API. Indeed, this was the genesis for both option 4 (ICH M7) approaches and the genotoxin impurity (GTI) Purge Tool, now named Mirabilis34. Acyl/sulfonyl halides are known to react with dimethyl sulfoxide (DMSO) to form alkyl halides, i.e. halodimethyl sulphides such as chlorodimethyl sulfide (CDMS) which are reported to be mutagenic35,36 and hence it is significant that the majority of the positive Ames findings reported in the literature for this chemical class were associated with the use of DMSO as the test vehicle. All (Q)SAR predictions for this chemical class in in silico systems such as Derek Nexus10 and Leadscope Genetox Statistical QSAR13 etc. are based on this positive acyl halides/sulfonyl halide Ames data from the literature. Therefore, to investigate the influence of various solvents/vehicles (e.g. DMSO, acetone, ethanol, water, etc.) on the potential mutagenicity of acyl/sulfonyl halides, a precompetitive collaboration between Sanofi-Aventis and GSK was initiated. The collaborators collated and re-examined internally generated Ames data as well as external Ames data, and where appropriate, they re-tested the appropriate compounds in either the five strain Ames (GSK), or the screening Ames II assay (Sanofi-Aventis), using alternative solvents to DMSO. A review of the available public domain data indicated that the mutagenicity of this ‘class of compounds’ are based on numerous non-reproducible positive findings in the bacterial mutagenicity Ames assay and positive bacterial mutagenicity data on a series of compounds where formation of mutagenic CDMS in DMSO may have compromised the interpretation of the Ames data. These findings call into question the relevance of the established in silico SAR rules for this class, highlighting the need for these alerts to be re-assessed. Some in silico systems, such as Derek Nexus10, contain rules which detail that the

activity of this class should be considered a ‘false positive’ flag for mutagenicity and ideally any in silico SAR rules for mutagenicity in other systems should be likewise addressed. This highlights the necessity that (Q)SAR systems remain current and are rapidly updated to reflect newly emerging


Page 15 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


knowledge. This data has been shared with the Lhasa Intermediates consortium. A more comprehensive overview of these findings can be found in a separate publication37. CHEMICAL PURGING During December 2013, Lhasa initiated an industrial collaboration to develop a software tool to assess the likelihood of chemical purging34 of mutagenic impurities (MI) or potentially mutagenic impurities (PMI), without recourse to extensive analytical testing. The requirement to utilize highly reactive intermediates for synthesis has been well documented; however, as well as being chemically reactive they also tend to be biologically reactive, and can be either mutagenic or can subsequently generate mutagenic impurities. Conversely this inherent reactivity often ensures facile purging of these compounds. Consequently, they are unlikely to be present in the final API. This collaboration uses the approach of Teasdale et al.38, 39 to generate a semi-quantitative, repeatable and standardized approach to predicting the likely purge factor for these impurities at every stage within a synthetic pathway up to, and including the formation of the API. This principle is clearly defined in ICH M7 within the control section as option 4. Within this section direct reference is made to the Teasdale et al. papers38,39. The first version of the software was released to the current consortium, comprising of eight pharmaceutical companies, in December 2014. This release included a model for predicting reactivity purge factors, which was established through expert elicitation, a form of knowledge sharing, which provides a consensus view of the purging potential of classes of reactive impurities. This has allowed prioritization of experimental work and provides further supporting information. The software was developed and refined through collaboration with the consortium, using a “user-centred design” approach to aid development. In addition to this work, the consortium has been carrying out experimental work aimed at investigating the reaction kinetics of different classes of mutagenic impurities in standardized reactions, in order to facilitate a better understanding of the relative purging rates. The Mirabilis consortium will continue for a further two years with the objective of then becoming another member of Lhasa’s product portfolio in 201734.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

DEGRADATION CHEMISTRY (ZENETH) ICH M77 indicates that ‘knowledge of relevant degradation pathways can be used to help guide decisions on the selection of potential degradation products to be evaluated for mutagenicity e.g., from degradation chemistry principles, relevant stress testing studies, and development stability studies’. Zeneth40,41 is a degradation chemistry software system developed by Lhasa that can be used to identify potential degradation pathways that can subsequently be confirmed (or not) by relevant stress testing studies, and development stability studies. The system is an offshoot of the Meteor metabolite prediction software42. Five pharmaceutical companies (Amgen, Eli Lilly, GSK, Johnson & Johnson and Pfizer) have collaborated from 2007 onwards to form a joint steering committee providing requirements and guidance for Zeneth software development as well as degradation expertise for building the knowledge base. Zeneth employs a reasoning engine, together with an evolving, curated, high-quality knowledge base. This knowledge base is grounded in general degradation chemistry extracted from books and from the primary literature, the web-based Pharmaceutical Drug Degradation Database (Pharma D343), and in cerebro knowledge contributed by the Zeneth collaborators. New degradation chemistry knowledge is incorporated on an ongoing basis, with the principal focus being on frequently encountered degradation pathways. The steering group members continue to provide overall guidance for the development of the Zeneth system. In addition, any Zeneth member (not only the steering group members) can participate in knowledge sessions where proposals for the addition of new knowledge are discussed and prioritised, and expertise and experience are shared. These sessions are web conferences taking place at least semiannually and they are the principal input channel for the expansion of the knowledge base. Members have also participated in evaluation of the Zeneth system, in the form of a benchmarking study. Five companies (Amgen, Eli Lilly, GSK, Merck & Co and Pfizer) shared confidential degradation data from 27 small molecules, including data on forced, accelerated and long-term degradation with Lhasa.


Page 16 of 31

Page 17 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


The main criterion studied was the predictivity of the system (i.e. the number of experimentally observed degradants that are successfully predicted by Zeneth). The findings were recently published44 and show that the predictivity of the system has increased from 31% in 2009 to 54% in 2012 as a result of the collaborative development activities of the consortium. An additional aim of the study was to pinpoint future areas of enhancement: closing gaps in chemistry coverage, extending the scope of degradation chemistry transformations where they currently lack breadth, and combining multiple steps within a single transformation to arrive more rapidly and effectively at primary degradation products. COLLABORATIONS THROUGH SECONDMENTS Collaborations can take place via several different routes. Often, a prerequisite is that confidentiality of any proprietary information must be adequately protected. Lhasa has a deserved reputation for maintaining such confidentiality. Glowienke45 recently reported on unique collaboration between Novartis and Lhasa involving the secondment of an experienced in silico scientist to work with Novartis in Basel, Switzerland. The secondee was responsible for data mining of Novartis information to facilitate development of new/additional Derek Nexus10 alerts and validation exercises using Sarah Nexus12 versus Novartis internal test data. As a result of this work, amongst others, a new Derek Nexus10 alert for hydroxylamine derivatives was proposed, which gave a predictivity of 86% versus an internal Novartis data set. During the collaboration, Novartis scientists assessed the predictivity of using both a knowledge based system, e.g. Derek Nexus10 and a statistically based system, e.g. Case Ultra14, Leadscope13, Sarah Nexus12, in tandem. Specifically, the combination of Derek Nexus10 /Case Ultra (A7B)14, Derek Nexus10/Leadscope13 and Derek Nexus10 /Sarah Nexus12 for the assessment of 789 impurities was evaluated. Based on this assessment, Sarah Nexus12 was found to have the best potential; it has an easy to use interface with connectivity to Derek Nexus10, similar predictivity to the other systems, together with fewer ‘out of domain’ alerts compared with the other two (Q)SAR systems. PRODUCT QUALITY RESEARCH INSTITUTE (PQRI) COLLABORATIONS



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

With the adoption of EMA guidelines8 for genotoxic impurities in 2007, the potential for sulfonate ester residues being present in APIs as a consequence of side reactions occurring during salt formation (involving sulfonic acids and alcoholic solvents), was a significant concern amongst regulatory authorities. This was heightened by the necessity to withdraw Viracept from the market as a consequence of contamination by mutagenic bi-products46. In parallel, the Quality by Design (QbD) guidelines (ICH Q847, Q948, Q1049 and the emerging Q1150) were encouraging industry to embrace the development of predictive scientific understanding to facilitate appropriate control strategies during the development phases of API processes. Therefore, during 2007, PQRI were approached and endorsed a proposal from six pharmaceutical companies (AZ, Pfizer, Schering-Plough, GSK, Wyeth and F. Hoffman-La Roche) to engage the services of chemical kinetics expert (Reaction Science Consulting LLC) and an independent, academic analytical laboratory (Research Institute for Chromatography, Kortrijk, Belgium) to commission a detailed study of the kinetics/dynamics of sulfonate ester formation and degradation. The proposal was aimed at providing mechanistic knowledge, developing sensitive and robust analytical methodology and establishing appropriate kinetic models using model compounds. This was with the intention of understanding the risks of sulfonate ester formation occurring during salt formation, and to facilitate the development of appropriate control strategies- if required. The outcome of this PQRI collaboration was a mechanistic understanding of the reaction mechanism using the methanol / methane sulfonic acid model system that employed isotopically labelled , O18, methanol51, combined with more comprehensive kinetic understanding of the reaction involving various alkyl alcohols (methanol, ethanol and iso-propanol), together with representative alkyl and aryl sulfonic acids and the role of water in the competing hydrolytic reaction52, a description of the supporting analytical methodology53, and specific conditions to eliminate formation54 (but equally germane to any strong acid used in conjunction with an alcoholic solvent). Finally, a review article was published articulating the biopharmaceutical, processing and pharmaceutical advantages for the continuing use of


Page 18 of 31

Page 19 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


sulfonic acid salts55. This allowed industry to adopt a risk-based approach, based on this full scientific understanding (QbD), rather than employ exhaustive analytical testing (Quality by Testing – QbT) as was the historical case. Regulators could subsequently endorse this scientific, risk-based approach, with the full knowledge that it was underpinned by significant, independent collaborative research. TRADE ASSOCIATIONS Both EfPIA and PhRMA trade associations had active genotoxic (mutagenic) impurity quality working groups (QWG) and safety working parties (SWP) that were designed to review and feedback on the emerging ICH M77 guidance. In addition, PhRMA formed several sub-groups to look at specific aspects of genotoxic impurities. Such groups played a crucial role in the development of guidance, including the introduction of the purge tool into the final step 4, ICH M77 guidance. EFPIA and PhRMA are also jointly collaborating together with members of the ICH M7 EWG to produce an addendum table to the guideline containing specific limits for key mutagenic carcinogens analogous to limits for solvents contained with ICH Q3C56. CROSS INDUSTRY SURVEYS Another way to collaboratively share data is to share performance metrics. For instance, in order to establish the suitability of any proposed new in silico methodologies (for a particular use) cross - industry surveys of EfPIA/PhRMA member companies can be initiated. Dobo et al.57 recently reported on the findings on a transnational inter-company survey (involving 8 companies) that showed an increased useage of (Q)SAR-based assessments. They reported that many publications in this area have historically focused on the capability of these (Q)SAR models (either individually or synergystically), to accurately predict the mutagenic effects of proprietary chemicals in the Ames assay. Typically, these investigations take significant numbers of compounds together with (Q)SAR-based tools to predict their likely mutagenic activity. However, there were no reports on the associated expert assessments that are also required to interpret the alert. As such, this does not reflect



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 31

how these assessments are typically performed across the global pharmaceutical industry. The authors were concerned that a structural assessment was viewed as being adequate to conclude that an impurity was non-mutagenic. This ‘hypothesis’ was investigated using an inter-company survey to assess companies’ individual success rate for predicting mutagenicity. The survey found that the Negative Predictive Value (NPV) of these (Q)SAR in silico approaches were 94%. However, when expert knowledge was introduced into this process, the NPV was increased to 99%. The authors commented on the importance of expert interpretation of such in silico predictions and they suggested that the use of multiple computational models (as required by ICH M77) is not a significant factor in the success of these approaches, with respect to NPV. Sutter et al. 58 also recently published a trans-national survey of the different (Q)SAR methodologies used by the pharmaceutical industry. The predictive value of these different (Q)SAR methodologies were also assessed. The authors reported that most pharmaceutical companies used an expert rule-based system as their standard methodology10,11, yielding negative predictivity values of ≥78% across all participating companies (14 in total). A further enhancement (>90%) was typically achieved by the use of an additional expert review and/or a second (Q)SAR methodology (typically statistically based12,13,14). However, in the second case, an additional expert review was also necessary, as conflicting results were often obtained. Based on all of the available data, the authors concluded that an expert rule-based system supplemented by either expert knowledge or a second statistically based (Q)SAR system is an acceptable approach. Total transparency of the assessment process (e.g. methods, results, arguments of weight-of-evidence approach) was achieved by full data sharing collaborative initiatives and in the future, the use of standard reporting approaches will enable regulators to fully understand the submitted in silico analyses. The authors concluded that this approach was appropriate for regulatory submissions aligned with ICH M77. ANALYTICAL COLLABORATIONS The majority of the public domain literature on analytical methods in this field is comprised of publications on the determination of solitary mutagenic impurities encountered during the development of


Page 21 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


company-specific API synthetic routes. Some representative examples are for MMS (methylmethane sulfonic acid) and EMS (ethylmethane sulfonic acid) in Lopinavir and Ritonavir59, TEMPO (2,2,6,6tetramethylpiperidin-1-yl)oxidanyl) in filibuvir60 and 4-dimethylaminopyridine in glucocorticoids61. There are a limited number of reports on generic methods for specific classes of alerting structures, which could be potentially useful for other pharmaceutical companies. Some representative examples include novel approaches for alkylating agents62,63, generic methods for hydrazine64, alkyl sulfonates and dialky sulphates65, alkyl sulfonates66, and DMS (dimethylsulfide)67. Several companies shared their analytical strategies to develop sensitive, selective analytical methods for mutagenic impurities68,69. Finally, there are a smaller number of method reviews (again, typically focussed on specific classes of alerting structures), that have been undertaken in a collaborative manner. These include sulfonyl esters70, alky halides71, hydrazines72 and epoxides73. There are reports of other types of open collaboration, for instance, an individual company collaborating with a specialized analytical laboratory; e.g. Pfizer Analytical Research Centre, University of Ghent, Belgium74 and the Research Institute for Chromatography, Kortrijk, Belgium75 or collaborations between Genentech and University of Toledo, Ohio, US76. CONCLUSIONS The area of mutagenic impurities has proved to be a fruitful one for various different types of precompetitive and competitive collaborations. The majority of these interactions are ‘honest broker’ type arrangements, where an autonomous, independent third party is tasked with coordinating the activities of the consortium. Sometimes, more than one ‘honest broker’ can be involved in the activity. Thus for example, for the collaborations centred on aromatic amines, the RSC was initially involved and having brought the various parties together they handed on responsibilities to Lhasa Ltd. The resulting CIGAA (Consortium for the Investigation of Genotoxicity of Aromatic Amines) was then formed. Indeed, Lhasa has been at the centre of many of these collaborative initiatives: Derek Nexus Pharmaceutical



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Intermediates, Excipients, Zeneth, Mirabilis, and scientific secondments (Novartis/Lhasa). Other ‘not for profit’ organisations that have played a role in the advancement of the greater understanding of mutagenic impurities include the Product Quality Research Institute (PQRI), PhRMA and EfPIA (the latter two are also trade organisations). PQRI coordinated activities allowing a mechanistic understanding of the reaction mechanism for the formation of sulfonate esters. Critically, this allowed conditions to be identified that during salt formation reactions involving APIs, sulfonate acids and alcoholic solvents eliminate the risk of sulfonate ester formation. The PQRI collaboration also involved a second ‘honest broker’, an independent laboratory (Research Institute for Chromatography, Kortrijk, Belgium) was commissioned to develop and generate all of the analytical data. The use of independent laboratories also features heavily in other analytical collaborations in this area, e.g. the Pfizer Analytical Research Centre, University of Ghent, Belgium and the University of Toledo, Ohio, US. The other collaborative interactivity is the JDI (Just Do It). This is primarily focussed on analytical collaborative efforts, where individuals have collaborated on areas of mutual interest. To a lesser extent this has also been seen in the area of pharmaceutical intermediates, where individual companies have collaboratively pooled data on areas of common interest, e.g. acyl/sulfonyl halide false positives. One of the key outcomes of these collaborations are peer reviewed scientific publications whereby other interested parties (including regulators) can be made aware of the findings from the consortia. This is particularly germane for computational toxicology assessment using (Q)SAR approaches. The ICH M7 guideline mandates that two (Q)SAR prediction tools that are complementary with each other should be utilized; i.e. a knowledge based tool (e.g., Derek Nexus, Leadscope Genetox Expert Alerts) and a statistically based tool (e.g.,Sarah Nexus Leadscope Genetox Statistical QSAR, CASEUltra, etc.). For the former, there are well established processes for Derek Nexus to increase its predictive power via the Lhasa consortia. In contrast, the statistically based tools require rapid publication of ‘new knowledge’ to allow the software to accurately predict these new alerts. In addition, to scientific publications, this interaction needs to involve collaborative efforts between industry and ‘for profit’ organisations (e.g.,


Page 22 of 31

Page 23 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Leadscope, CASE Ultra, etc.). This had led to the evolution of clear processes for sharing confidential data. The majority of the most significant Just Do It (JDI) collaborations involve sharing feedback on the use and applicability of a second (Q)SAR system and in the area of analytical methodologies and analytical strategies. Going forward, the area of mutagenic impurities enshrined under ICH M7, will continue to be a fruitful area for precompetitive collaborations of all types building on the success of the previous initiatives described. ORGANIZATIONS FACILITATING PRECOMPETITIVE COLLABORATIONS IN ICH M7 SPACE EFPIA:

The European Federation of Pharmaceutical Industry Associations (EFPIA) represents

the pharmaceutical industry within Europe. EFPIA is the leading voice of the 33 European national pharmaceutical industry associations as well as 40 leading companies undertaking R&D and manufacture within Europe of human medicinal products. EFPIA is fully engaged with the EU and, through its member organisations, with the Member States, in the discussions on the formulation, revision and implementation of EU regulation and legislation covering medicinal products and their development, manufacture and subsequent sale.

IQ:

The innovation and quality consortium is a not-for-profit organization of pharmaceutical and

biotechnology companies with the mission of advancing science-based and scientifically-driven standards and regulations for pharmaceutical and biotechnology products worldwide.

ISPE: The International Society for Pharmaceutical Engineering, is the world's largest not-for-profit association serving its Members by leading scientific, technical and regulatory advancement throughout the entire pharmaceutical lifecycle. ISPE is committed to the advancement of the educational and technical efficiency of its members through forums for the exchange of ideas and practical experience.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Lhasa Limited:

Lhasa is UK based not-for-profit organization and educational charity

specializing in chemical informatics, e.g. toxicology and metabolism prediction. Lhasa’s software is collaboratively developed with consortium members and as such they are perfectly suited for the creation and management of collaborative groups facilitating data sharing, without concerns about compromising the disclosure of any confidential, proprietary information. PhRMA:

The Pharmaceutical Research and Manufacturers of America, represents the US’ leading

biopharmaceutical researchers and biotechnology companies. PhRMA's mission is to conduct an effective advocacy strategy for public policies that boost discovery of important new medicinal products by pharmaceutical and biotechnology research companies. PQRI: The Product Quality Research Institute (PQRI) is a non-profit consortium of different organizations effectively working together to generate and share timely, significant information that supports the advancement of drug product quality and development. PQRI provides a unique environment to conduct research, share information, and propose new approaches to pharmaceutical companies, regulators, and national and international standard setting organizations. RSC: The Royal Society of Chemistry (UK) is a not-for-profit organisation representing the professional interests of chemists within the UK and world-wide. RSC is focussed on shaping the future of the chemical sciences, for the benefits of science and humanity. RSC claim that they can ‘facilitate collaborations with chemical scientists more comprehensively and effectively than any other organisation in the world’. AUTHOR INFORMATION * E-mail: [email protected] ACKNOWLEDGEMENTS


Page 24 of 31

Page 25 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


The authors would like to recognize the valuable input that Martin Ott, Anne-Laure Werner, Maria Pellizzaro, Susanne Stalford from Lhasa Limited, and Glenn Myatt from the Leadscope Corporation, have made during the review of this manuscript. REFERENCES 1. Welch, C.J.; Hawkins, J.M.; Tom, J. Org. Process Res. Dev. 2014, 18, 481-487. 2. Woodcock, J. Clin. Pharmacol. Ther. 2010, 87, 521-523. 3. Thomson, N.M.; Seibert, K.D.; Tummala, S.; Bordewekar, S.; Kiesman, W.F.; Irdam, E.A.; Phenix, B.; Kumke. D. Org. Process Res. Dev. 2014. DOI: 10.1021/op500187u 4. FDA Critical Path Initiative. http://www.fda.gov/ScienceResearch/SpecialTopics/CriticalPathInitiative/. Accessed on 14th February 2015. 5. FDA Advancing Regulatory Science Initiative. http://www.fda.gov/ScienceResearch/SpecialTopics/RegulatoryScience/default.htm?utm_campaign= Goo. Accessed on 14th February 2015. 6. Innovative Medicines Initiative. http://www.imi.europa.eu/. Accessed on 14th February 2015. 7. ICH M7. Step 4, 23rd June 2014. http://www.ich.org/fileadmin/Public_Web_Site/ICH_Products/Guidelines/Multidisciplinary/M7/M7_ Step_4.pdf. Accessed on 14th February 2015. 8. EMA. 28th June 2006. CPMP/SWP/5199/02, EMEA/CHMP/QWP/251344/2006. http://www.ema.europa.eu/docs/en_GB/document_library/Scientific_guideline/2009/09/WC5000029 03.pdf. Accessed on 14th February 2015. 9. FDA. U.S. Department of Health and Human Services, Food and Drug Administration, Center for Drug Evaluation and Research (CDER), December 2008, Pharmacology and Toxicology. http://www.fda.gov/downloads/Drugs/.../Guidances/ucm079235.pdf. Accessed on 14th February 2015.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 31

10. Derek Nexus. http://www.lhasalimited.org/products/derek-nexus.htm. Accessed on 14th February 2015. 11. Leadscope http://www.leadscope.com/white_papers/ICHM7-WhitePaper-0314.pdf. Accessed on 14th February 2015.

12. Sarah Nexus. http://www.lhasalimited.org/products/sarah-nexus.htm Accessed on 14th February 2015. 13. Leadscope Model Applier. http://www.leadscope.com/faq/Leadscope_Model_Applier_and_the_ICH_M7_FAQs_April2014.pdf. Accessed on 14th February 2015. 14. Case ULTRA. http://www.multicase.com/case-ultra. Accessed on 14th February 2015. 15. Dobo, K. L.; Greene, N.; Cyr, M.O.; Caron, S.; Ku, W.W. Regul. Toxicol. Pharmacol. 2006, 44, 2582-293. 16. Vitic Nexus. http://www.lhasalimited.org/products/vitic-nexus.htm. Accessed on 14th February 2015. 17. CIGAAhttp://www.Lhasalimited.org/research-and-collaboration/collaboration.htm. Accessed on 30th November, 2014. 18. Cayley A.; Marchant C.; Williams, R. Mutagenesis 2009, 24(6), 541-542. 19. Hillebrecht, A.; Muster, W.; Brigo, A.; Kansy, M.; Weiser, T.; Singer, T. Chem. Res. Toxicol. 2001, 24, 843-854. 20. Stavitskaya, L.; Minnier, B.L.; Kruhlak, N.M. Abstract 2273b, Society of Toxicology, 53rd Annual Meeting, Phoenix, Arizona, 23-27th March 2014. http://www.lhasalimited.org/Public/Library/Sarah Library/Sarah - Posters/Benchmarking Assessment of Salmonella Mutagenicity QSAR Models for Potential Use Under ICH M7.pdf. Accessed on 7th April 2015 21. http://www.leadscope.com/publications/ACT_2014Ensuring_regulatory_acceptable_QSAR_and_proprietary_chemical_space.pdf. February, 2015.


Accessed on 15th

Page 27 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


22. Barber, C.; Hanser, T.; Kruhlak, N.L.; Stavitskaya, L.; Vessey, J.D. Abstract 2262, Society of Toxicology,

53rd

Annual Meeting,

Phoenix,

Arizona,

23-27th

March

2014.

http://www.lhasalimited.org/Public/Library/Sarah Library/Sarah - Posters/Establishing best practice for use of statistical model for ICH M7.pdf. Accessed on 7th April 2015

23. Hanser, T.; Barber, C.; Rosser, E.; Vessey, J.D.; Webb, S.J.; Werner, S. J. Cheminf. 2014, 6, 1-21. 24. Ellis, P.; Kenyon, M.; Cheung, J.; Ackerman, J.; Caron, S.; Dobo, K. Mutagenesis 2010, 25, 657. 25. O’Donovan, M. R.; Mee, C. D. Mutagenesis 2010, 25, 658 26. O’Donovan, M. R.; Mee, C. D.; Fenner, S.; Teasdale, A.; Phillips, D. H. Mutat. Res. 2011, 724, 1-6. 27. Ciaravino, V.; Plattner, J.; Chanda, S. Environ. Mol. Mutagen. 2013, 54, 338-346. 28. Stannard, L.; Giddings, A.; Lad, A.; Munoz-Muriedas, J.; Harvey, J.; Kenny, J. Mutagenesis 2012, 27, 807-808. 29. Pellizzaro, M. L.; Williams, R. V. Environ. Mol. Mutagen. 2013, 54, S53. 30. Scott, H.; Walmsley, R. M. Mutat. Res. In press http://dx.doi.org/10.1016/j.mrgentox.2014.12.002. 31. Gibbins, S. Bio-IT World Conference & Expo. 2012. http://www.lhasalimited.org/Public/Library/Vitic%20Library/Vitic%20%20Posters/The%20benefits%20of%20sharing%20toxicological%20data%20in%20a%20centralised%20 database.pdf. Accessed on 18th May 2015.

32. Naven, R.T. In: Abstracts of the 35th Annual Meeting of the United Kingdom Environmental Mutagen Society, 16th–18th July 2012 at Swansea University, UK. http://www.ivtn.ru/cmtpi-2007/pdf/Abstract%20Book.pdf. Accessed on 15th January 2015.

33. RSC. 2011. http://www.google.co.uk/url?sa=t&rct=j&q=&esrc=s&frm=1&source=web&cd=8&ved=0CEYQFjA H&url=http%3A%2F%2Fwww.ukpharmsci.org%2F2013resourcepack%2FPresentations%2FDownlo



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ads.asp%3FIDNumber%3D14&ei=3Be5VJtLIvwUpnXgdAJ&usg=AFQjCNFWIbh1Qqfr9HTJYEebZvE3rStNLQ. Accessed on 15th January 34. 2015. Lhasa Limited Mirabilis. http://www.lhasalimited.org/products/mirabilis.htm. Accessed on 15th February 2015. 35. Boyle R.E. J. Org. Chem. 1966, 31, 3880–3882. 36. Thea, S.; Cevasco, G. J. Org. Chem. 1988, 53, 4121-4122. 37. Amberg, A.; Harvey, J.S.; Czich, A.; Kauffmann, H-M.; Spikl, H-P.; Robinson, S.; White, A.; Elder,

D.P. Org. Process Res. Dev. Submitted for publication. 38. Teasdale, A.; Fenner, S.; Ray, A.; Ford, A.; Philips, A. Org. Process Res. Dev. 2010, 14, 943-945. 39. A Teasdale, D Elder, Sou-Jen Chang et al, Org. Process Res. Dev. 2013, 17 (2), pp 221–2

40. Lhasa Limited, Zeneth. http://www.lhasalimited.org/products/zeneth.htm. Accessed on 15th February 2015. 41. Parenty, A.D.C.; Button, W.G.; Ott, M.A. Mol. Pharmaceutics 2013, 10, 2962-2974. 42. Marchant, C.A.; Briggs, K.A.; Long, A. Toxicol. Mech. Methods. 2008, 18, 177-187. 43. Pharma D3. URL: d3.arxspan.com. Last accessed 12 January 2015. 44. Kleinman, M.H.; Baertschi, S.W.; Alsante, K.M.; Reed, D.L.; Mowery, M.D.; Shimanovich, R.; Foti, C.; Smith, W.K.; Reynolds, D.W.; Nefliu, M.; Ott, M.A. Mol. Pharmaceutics 2014, 11, 4179-4188.

45. Glowienke, S.; Parenty, A.; Hartmann, A. http://www.Lhasa limited.org/Public/Library/Lhasa %20Library/Lhasa Ltd. %20%20Presentations/Scientific%20Collaborations%20through%20Secondments.pdf. Accessed on 02nd December 2014. 46. Elder, D.P.; Snodin, D. J. Pharm. Pharmac. 2009, 61, 1-10.


Page 28 of 31

Page 29 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


47. ICH

Q8

(R2).

Step

4,

August

2009.

http://www.ich.org/fileadmin/Public_Web_Site/ICH_Products/Guidelines/Quality/Q8_R1/Step4/Q8_ R2_Guideline.pdf. Accessed on 15th February 2015. 48. ICH Q9. Step 4, November 2005. http://www.ich.org/fileadmin/Public_Web_Site/ICH_Products/Guidelines/Quality/Q9/Step4/Q9_Gui deline.pdf. Accessed on 15th February 2015. 49. ICH Q10. Step 4, June 2008. http://www.ich.org/fileadmin/Public_Web_Site/ICH_Products/Guidelines/Quality/Q10/Step4/Q10_G uideline.pdf. Accessed on 15th February 2015. 50. ICH Q11. Step 3, September 2011. http://www.ema.europa.eu/docs/en_GB/document_library/Scientific_guideline/2011/06/WC5001076 36.pdf. Accessed on 15th February 2015. 51. Teasdale, A.; Eyley, S.C.; Delaney, E.; Jacq, K.; Taylor-Worth, ⊥K.; Lipczynski, A.; Van Reif.; Elder, D.P.; Facchine, K.L.; Golec, S.; Schulte Oestrich, R.; Sandra, P.; David, F. Org. Process Res. Dev. 2009, 13, 429-433. 52. Teasdale, A.; Delaney, E.J.; Eyley, S.C.; Jacq, K.; Taylor-Worth, K.; Lipczynski, A.; Hoffmann, W.; Van Reif.; Elder, D.P.; Facchine, K.L.; Golec, S.; Schulte Oestrich, R.; Sandra, P.; David, F. Org. Process Res. Dev. 2010, 14, 999-1007 53. Jacq, K.; Delaney, E.J.; Teasdale, A.; Eyley, S.; Taylor-Worth, K.; Lipczynski, A.; Van Reif.; Elder, D.P.; Facchine, K.L.; Golec, S.; Schulte Oestrich, R.; Sandra, P.; David, F. J. Pharm. Biomed. Anal. 2008, 48, 1339-1344. 54. Elder,D.P.; Facchine, K.L.; Levy, J.N.; Parsons, R.; Ridge, D.; Semo, L.; Teasdale, A. Org. Process Res. Dev. 2012, 16, 1707-1710. 55. Elder, D.P.; Delaney, E.J.; Teasdale, A.; Eyley, S.; Van Reif.; Jacq, K.; Facchine, K.L.; Schulte Oestrich, R.; Sandra, P.; David, F. Review: J. Pharm. Sci. 2010, 99, 2948-2961.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

56. ICH Q3C(R5). Step 5. 15th May 2014. EMA/CHMP/ICH/82260/2006. http://www.ema.europa.eu/docs/en_GB/document_library/Scientific_guideline/2011/03/WC5001042 58.pdf. Accessed on 26th February 2015. 57. Dobo, K.L.; Greene, N.; Fred, C.; Glowienke, S.; Harvey, J.S.; Hasselgreen, C.; Jolly, R.; Kenyon, M.O.; Munzer, J.U.B.; Muster, W.; Neft, R.; Reddy, M.V.; White, A.T.; Weiner, S. Regul. Toxicol. Pharmacol. 2012, 62, 449-445. 58. Sutter, A.; Amberg, A.; Boyer, S.; Brigo, A.; Contrera, J.F.; Custer, L.L.; Dobo, K.L.; Gervais, V.; Glowienke, S.; van Gompel, J.; Greene, N.; Muster, W.; Nicolette, J.; Reddy, M.V.; Thybaud, V.; Vock, E.; White, A.T.; Müller, L. Regul. Toxicol. Pharmacol. 2013, 67, 39-52. 59. Kakadiya, P.R.; Reddy, B.P.; Singh, V.; Ganguly, S.; Chandrashekhar, T.G.; Singh, D.K. J. Pharm. Biomed. Anal. 2011, 55, 379-384. 60. Strohmeyer , H.E.; Sluggett, G.W. Determination and control of TEMPO, a potentially genotoxic free radical reagent used in the synthesis of filibuvir. J. Pharm. Biomed. Anal. 2012, 62, 216-219. 61. Székely, G.; Henriquez, B.; Gil, M.; Ramos, A.; Alvarez, C. J. Pharm. Biomed. Anal. 2012, 70, 251258. 62. Van Wijk, A.M.; Niederländer, H.A.G.; Siebum, A.H.G.; Vervaart, M.A.T.; de Jong, G.J. J. Pharm. Biomed. Anal. 2013, 74, 133-140. 63. Cappiello, A.; Famiglini, G.; Palma, P.; Termopoli, V.; Trufelli, H. J. Chromatogr. A. 2012, 1255, 286-290. 64. Sun, M.; Bai, L.; Liu, D.Q. J. Pharm. Biomed. Anal. 2009, 49, 529-533. 65. An, J.; Sun, M.; Bai, L.; Chen, T.; Liu, D.Q.; Kord, A. J. Pharm. Biomed. Anal. 2008, 48, 1006-1010. 66. Guo, T.; Shi, Y.; Zheng, L.; Feng, F.; Zheng, F.; Liu, W. J. Chromatogr. A. 2014, 1355, 73-79. 67. Grinberg, N.; Albu, F.; Fandrick, K.; Lorgulescu, E.; Medvedovici, A. J. Pharm. Biomed. Anal. 2013, 75, 1-6. 68. Raman, R.V.V.S.S.; Prasad, A.V.S.S.; Reddy, K.R. J. Pharm. Biomed. Anal. 2011, 55, 662-667. 69. Liu, D. Q.; Sun, M.; Kord, A. J. Pharm. Biomed. Anal. 2010, 51, 999-1014.


Page 30 of 31

Page 31 of 31

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


70. Elder, D.P.; Teasdale, A.; Lipczynski, A.M. J. Pharm. Biomed. Anal. 2008, 46, 1-8. 71. Elder, D.P.; Lipczynski, A.M.; Teasdale, A. J. Pharm. Biomed. Anal. 2008, 48, 497-507. 72. Elder, D.P.; Snodin, D.; Teasdale, A. J. Pharm. Biomed. Anal. 2011, 54, 900-910. 73. Elder, D.P.; Snodin, D.; Teasdale, A. J. Pharm. Biomed. Anal. 2010, 51, 1015-1023. 74. Alzaga, R.; Ryan, R.W.; Taylor-Worth, K.; Lipczynski, A.M. Szucs, R.; Sandra, P. J. Pharm. Biomed. Anal. 2007, 45, 472-479. 75. Vanhoenacker, G.; Dumont, E.; David, F.; Baker, A.; Sandra, P. J. Chromatogr. A. 2009, 1216, 35633570. 76. Ho, T.D.; Yehl, P.M.; Chetwyn, N.P.; Wang, J.; Anderson, J.L. J. Chromatogr. A. 2014, 1361, 217228.


Competitive Collaborative and

Recommend Documents