Cont ntrolling lling p potent ntia ial g l geno notoxic xic imp impur urit itie ies s enc ncount untered d dur uring ing A API sy synt nthe hesis sis Utilis ilisin ing g expe pert k knowle owledge dge of of c chemic ical pr l prope opertie ies to o manag anage ri e risk Dr Michael Burns Senior Scientist Michael.Burns@lhasalimited.org
Outline • Background PMIs in synthesis Impurity carry-over workflow Purge calculations • Mirabilis Origins Workflow for ICH M7 Theoretical case study • Ongoing Mirabilis developments 2
Background • The threat posed by (potential) mutagenic impurities, (P)MIs, in drug substances arises; for example, from the use of reagents such as alkylating agents within the synthesis • What makes them useful reagents in synthesis, high reactivity, is often what makes them (P)MIs • Virtually all syntheses will involve the use of mutagenic or potentially mutagenic reagents or possess potential risk arising from a (P)MI formed in the process • Any synthetic drug therefore may have a latent (P)MI-related risk. 3
Impurity Carry-over Workflow API synthesis Implement Plan Reactive Knowledge of High utility functionality Efficient syntheses physicochemical properties Mutagenic? In-silico toxicity prediction Mutagenic Teasdale et al ’s ICH M7 regulations In-vitro toxicity Non-mutagenic scoring approach Option 1 or 2 to purge prediction Option 3 or 4 Mutagenic Not Purged Assess likelihood of Test for impurity impurity persisting Analytical challenge Purged Time consuming Option 3 Expensive Testing 4 unnecessary
Purge Factor Calculation – Basic Principles • The following key factors were defined in order to assess the potential carry-over of a (P)MI: Reactivity, solubility, volatility and any additional physical process designed to eliminate impurities e.g. chromatography. • A score is assigned on the basis of the physicochemical properties of the ( P)MI relative to the process conditions These are then simply multiplied together to determine a ‘purge factor’ (for each stage). • The overall purge factor is the product of the factors for individual stages. • Predicted purge is then compared to required purge (this being based on the safety limit and initial level introduced into the process). 5
Original Purge Prediction Scoring System • The original scoring system was built on basic principles – referred to as a ‘paper’ assessment because its not automated (manual calculation via spreadsheet) Reactivity shown to have largest effect Other factors especially solubility would also influence purging Physicochemical parameters Purge factor Reactivity Highly reactive = 100 Moderately reactive = 10 Low reactivity/unreactive = 1 Solubility Freely soluble = 10 Moderately soluble = 3 Sparingly soluble = 1 Volatility Boiling point >20 °C below that of the reaction/ process solvent = 10 Boiling point within ±20 °C of that of the reaction/process solvent = 3 Boiling point >20 °C above that of the reaction/ process solvent = 1 Ionisability Ionisation potential of GTI significantly different from that of the desired product Physical processes: chromatography Chromatography: 10−100 based on extent of separation Physical processes: e.g. other scavenger Evaluated on an individual basis. resins • Scoring system originally designed to be conservative On validation this was experimentally observed It was decided that this should be retained rather than seeking absolute parity Urquhart et al recently demonstrated the approach for Atovaquone 6 Urquhart et al . Regul. Toxicol. Pharmacol., 2018, 99, 22-32
Paper Assessment – Case Study – AZD9056 O N OH H H N N 3-aminopropan-1-ol O Cl O Cl AZD9056 Aldehyde AZD9056 Imine /Pt H 2 N OH H N AZD9056 Free Base Cl O Isopropyl chloride HCl in IPA (by-product) Cl N OH N Cl H N H N O Cl O Cl .HCl AZD 9056 Chloride (minor by-product) AZD9056 HCl MeOH / water 7 pure
Paper Assessment – Case Study – AZD9056 O N Cl H H N N O Cl O Cl Cl Predicted: 10,000 Predicted: 3 Predicted: 10,000 Measured: 112,000 Measured: 10 Measured: 38,500 (Solubility, Reactivity) (Solubility) (Solubility, Volatility) Comparison of the overall predictions with the experimental results shows that the predicted purge factor is in good correlation with each impurity Under-predicts the purge capacity of the process • Clearly demonstrates risk of carry over to be low • Predictions indicated where formation of an impurity needed to be regulated • through process control, rather than relying on the ability of the process to eliminate it. 8
Is this simply about avoiding analytical testing? R NO 2 • 3 Step reaction OH N + N - Starting material contains an aromatic N -oxide O Stage 1 Alcohol converted to alkyl halide R Coupled to a thiol Cl N Oxidation step H K + N - Stage 2 S • Only final step isolated N R' Impurity is un-reactive/highly soluble/non- H R N volatile S N R' N No purge predicted in steps 1 and 2 Stage 3 • Spiking experiment 3000ppm K + Only reduced to 2000ppm - R O N S N R' N Purge calculations showed that control at starting material is required. 9
Mirabilis • Mirabilis established to take basic principles of paper- based predictions and augment them • Key concepts: Use of an in silico template allows for greater consistency in terms of how predictions are structured and reported ( reproducibility ) Predictions by Mirabilis are informed from a knowledge base , the basis of these is clearly visible ( objectivity ) Knowledge management & pre-competitive knowledge sharing ( supports further development ) 10
Origins of Mirabilis Knowledge Base • Common alerting impurity types and popular chemical transformations identified Cell by the Mirabilis consortium 15 Impurity types 58 Transformation types • Each impurity type is analysed against every transformation to assess potential purge, generating a reactivity matrix • A consortium collaboration exercise resulted in an ‘expert elicitation’ for each reactivity purge factor • Lhasa scientists subsequently augment the ‘cells’ with scientific comments including; references, examples and supporting data 11
Mirabilis workflow Establish API Identify acceptable intake from ICH M7 treatment length (e.g. >1 – 10 years = 10 μ g/day) Calculate purge ratio to establish Calculate purge required to control option 4 Define API daily dose suitability and reach acceptable intake level regulatory requirements Define initial impurity level Repeat for each stage of the synthesis Identify impurity Start Synthetic Identify purge Return relevant purge Review reactivity purge scheme relationship and supporting info assignment Identify transformation Combine purges to Combine all stage achieve total purge per purges to achieve total stage predicted purge Add purges and Add all unit operations justification for each (e.g. Work-up, extraction, filtration) unit operation 12
Theoretical Case Study - Imatinib H N N N O Cl N N Br Br N N N Br NH 2 + NH Cl N Et 3 N, DCM K 2 CO 3 , DMF Pd 2 dba 3 , XantPhos, HN HN N NH 2 Stage 1 Stage 2 NaO t Bu, dioxane, N Cl O O t BuOH HN N Stage 3 O Imatinib • Imatinib is an anti-cancer drug with a maximum daily dose of 800mg for up to 3 years • Hopkin et al published a 3-stage synthesis to Imatinib Additional steps include basic work up/extraction (stages 1 and 3), precipitation and wash (stage 2), and column chromatography (stage 3) • Six structures in the synthesis need to be analysed for potential ICM M7 control* • ICH M7 control limit of 10 μ g for any PMIs based on dose and duration of treatment * ICH M7 does not actually apply to anti–cancer drugs 13 Hopkin et al . Org. Biomol. Chem., 2013, 11, 1822-1839. http://www.glivec.com/dosing/
Which structures are PMIs? 14
Purge Assessment in Mirabilis 15
16
17
18
19
20
21
22
23
O Cl Cl Br NH 2 Br Cl HN O 24
Regulatory Requirements O Cl Br Br Cl HN NH 2 Cl O PR = 1.25 x 10 6 PR = 12.5 PR = < 1 25 Barber et al, Regul. Toxicol. Pharmacol., 2017, 90, 22-28
Impurity Carry-over Conclusion API synthesis O Cl Br Implement Plan Reactive Knowledge of NH 2 functionality physicochemical properties Cl Subject to evidence package Mutagenic? In-silico toxicity prediction Mutagenic Teasdale et al ’s ICH M7 regulations In-vitro toxicity scoring approach to purge prediction Option 3 or 4 Assess likelihood of Test for impurity impurity persisting Purged Testing 26 unnecessary
Impurity Carry-over Conclusion API synthesis Br Implement Cl Plan Reactive HN Knowledge of functionality physicochemical properties O Mutagenic? In-silico toxicity prediction Mutagenic Teasdale et al ’s ICH M7 regulations In-vitro toxicity Non-mutagenic scoring approach Option 1 or 2 to purge prediction Option 3 or 4 Mutagenic Not Purged Assess likelihood of Test for impurity impurity persisting Testing 27 unnecessary
Recommend
More recommend
