High throughput screening of pka by capillary electrophoresis and mass spectrometry (CE/MS) and long-term validation

Transcription

1 High throughput screening of pka by capillary electrophoresis and mass spectrometry (CE/MS) and long-term validation Hong Wan Lead Generation, DMPK and Physical Chemistry AstraZeneca R&D Mölndal SE Mölndal. SWEDE 5 th Physchem Forum, Stevenage, UK, June 9th

2 utline Why screening pka Advantages of CE/MS over established methods High throughput pka screening method by CE/MS Effects of buffer type and ionic strength on pka Comparison of CE/UV and CE/MS Effects of pressure on migration times/sensitivity/mobility Comparisons of measured pka and ref. and predicted values Reproducibility and accuracy and long-term validation Summary

3 Screening for pka, why? Info about change in charge of molecule Solubility QSAR descriptor Formulation Bioavailability Lipophilicity pk a (dissociation constant) Stability Metabolism Protein binding Hydrogen bonding potency Absorption Permeability

4 Is CE/MS better technique for pka than others? Sirius GLpKa well established and widely used method! pi's Gemini for routine pka ($5/cdp, $76/cpd/co-solvent) CE/UV (CombiSep), commercially available CE/MS (new technology), AZ, Mölndal, 3 published. H. Wan, A. Holmén, M. ågård, W. Lindberg, J. Chromatogr. A. 979 () 369. H. Wan, A. Holmén, Y. Wang, W. Lindberg, M. Englund, M. ågård, R. Thompson, Rapid Commun. Mass Spectrom. 7 (3) 639.

5 Advantages of CE/MS over established methods Specification CE/MS Titration/UV (D-PAS) Amount of cpd μg, or μl of mm solution - mg (titration) 3 μl of mm DMS stock (Sirius 3T) Concentration - μm ( μl) > μm (UV) Purity required o Yes Co-solvent o Yes, for poorly soluble cpds Accuracy < ±. ±.-.? Throughput capacity >5 cpds/seq.(6h) 4 min/cpd (Sirius T3) Limitation Poor ionization (ESI) Impurity? poor solubility (precipitation), Ionization group close to chromophore

6 Principle of pka determination by CE pka 5% ionization K th a = {H + }{A {HA} } pk th a = ph log m m m eff m eff +.585z I +.38α I

7 Equations used for pka calculation Graphs and table from Miller, Electrophoresis, 3, 833

8 High-throughput pka screening by CE/MS -4 different ph buffers Constant ionic strengths Pressure mau CE separation 5 5 UV non-volatile buffers (Single compound) - μm MS Volatile buffers (Pooled compounds) (min) Mobility calculation m eff = m obs m eof = L tot L V eff t obs t eof pk a evaluation (fit-pk a ) Effective mobility pka pka ph

9 Sample pooling and data analysis flow scheme Reference stand. 3 manual data analysis H. Wan, A. Holmén, Y. Wang, W. Lindberg, M. Englund, M. ågård, R. Thompson, Rapid Commun. Mass Spectrom. 7 (3) 639.

10 Current pka screening - integrated & automated assay Generate sample run list (Abase) Automated sample pooling (mass) CE/MS AZ883- A ph meff res Fitting ption: x Di-Base x x Value: Ionization: Ion. Func. Gr x B pka : 3. B x pka : 8.33 B x pka 3 : V x r : x B x Comment: x x x x ph x meff Automated migration time conversion pka evaluation pka reported to database Approved

11 CE/MS instrumentation Higher initial investment Long-term savings HPCE 3D, series LC/MSD trap SL (Agilent) Untreated fused silica capillary (5-6 cm/5 SEK) can be used as long as it works. Buffers without filtration (stock solutions can be used up to >3 years). Sheath liquid ( L) can be used for more than 6 months (99% recycling).

12 n-line CE/UV/MS experimental setup UV Detector interface Capillary cassette Air cooling To MS (Agilent series LC/MSD Trap) Capillary inlet cm > 75 cm MS w/o 55 cm Pressure (4-5 mbar) Courtesy from Agilent (with small modifications)

13 Sensitivity from volatile vs non-volatile buffers in CE/MS S/ μm Lidocaine Papaverine ot detectable 5 mm H4AC 5 mm H4AC 75 mm H4AC mm H4AC 5 mm sodium/phosphate 5 mm sodium/phosphate 75 mm sodium/phosphate mm sodium/phosphate on-volatile buffer decreased signals at higher concentration (ion suppression) Ion source contamination by non-volatile buffer

14 Effects of buffer type and ionic strength on pka pk th a = meff ph log mm m eff.585z I α I pka by CE/UV CE/UV (non-volatile buffer, I=.5) CE/MS (volatile buffer, I=.5) icotine Atenolol Propranolol Codein icotine e Lidocaine Pilocarpine Papaverine Tryptophan (ACD) 9.55/ 9.34/ 9./ 9.3/ 9.3/ 9.9 H H H 9.5/.34 (CE/MS, I=.5) 9.43/.3 (CE/UV, I=.).37 /.37/.3/.39/.75 I=. I=. I=.5 I=.5 I=. I= pka by CE/MS.. pka(b) pka(ha) Titration method uses physiological buffer with ionic strength at.5 M. Ionic strengths ( M) have a small effect on pka (Δ pka=.64)

15 Comparison of UV and MS (Ion trap) sensitivity Intens. [mau] x UV EF MS(EIC) Pilocarpine Papverine Lidocaine Atenolol Codeine Propranolol Sample:.5 µm Ammonium acetate, (ph=7) I=.5 Pressure: 4 mbar Time [min] UV working concentration: - μm

16 Sensitivity and reproducibility of pressure-assisted-ce/ms Total ion chromatogram (TIC) Intens. x 6. μm SES-C.D: TIC ±All x 6 x 7 μm ( mixed compounds) SES-C.D: TIC ±All SES-C3.D: TIC ±All μm Time [min] (EIC) and migration times after 3 runs Ammonium acetate (ph=7) I=.5; pressure, 4 mbar Intens. x 4. μm S/= SES-C.D: EIC 6 ±All MS x 6..5 μm (propranolol) S/=35 SES-C.D: EIC 6 ±All MS. x 7.5 μm S/=864 SES-C3.D: EIC 6 ±All MS Time [min]

17 Sensitive and selective MS detection over UV UV 5 AR-H73379? DMS (.5%) ph=3 5 μμ - PKAC36.D: UV Chromatogram,.4 nm Intens. x6.5 AR-H (M+) +MS, 7.3min 74) (#9. MS x6 6 AR-H73379 PKA-C36.D: EIC 79; 48 ±All, DMS m/z Time [min] n-line tandem UV and MS detection

18 Effects of pressure on migration times/sensitivity Time (s) Mass scan 7-66 m/z Pilocarpine+Papaverine Codeine+Propranolol mbar Atenolol Codeine Lidocaine Quinidine Pilocarpine Propranolol icotine Papaverine CE capillary sheath liquid CE sprayer tip nebulizing gas Intensity mbar Atenolol Codeine Lidocaine Quinidine Pilocarpine Propranolol icotine Papaverine Pressure reduced migration times egligible signal suppression Improved RSD of effective mobility High throughput capacity

19 Effects of pressure on effective mobility 3.5 Ibuprofen Meff( -4 cm V - s - ) pka=4.5 5 mbar 5 mbar mbar mau (A) ph=3 (c) P=5 mbar (b) P=5 mbar (a) P= mbar H Benzoic acid 3 H Ibuprofen H 4 min Lidocaine min ph Effective mobility shift around pka (an interesting observation) Mobility shift caused by pressure doesn t affect pka values. H. Wan, A. Holmén, M. ågård, W. Lindberg, J. Chromatogr. A. 979 () 369.

20 How many compounds can be pooled by CE/MS? 8 DMS (Peak height) 7 DMS (Peak area) DMS % DMS < 5% in sample Constant mobility High MS resolution More than 5 cpds/sequence μ eff Pilocarpine Lidocaine Propranolol Atenolol Papaverine. 3 4 DMS % 5

21 Comparison of measured pka and lit./predicted values Poor prediction for conjugated structures CH pk 3 a (CE/MS) (a) pk a (CE/MS) (b) F 9 9 H 8 R =.99 8 H Flumequine R = Cinoxacin Phenylbutazone pk a (literature) pk a (Cal.) ACD6 H. Wan, A. Holmén, Y. Wang, W. Lindberg, M. Englund, M. ågård, R. Thompson, Rapid Commun. Mass Spectrom. 7 (3) 639.

22 Long-term validation of CE/MS pka (5-8) Scatter Plot Paracetmol Imipramine Terfenadine Clofazimine Haloperidol 4 buffers pka: 8.7 (5) 7.88/7.66 (8) Tamoxifen Amiodarone Meclizine Tamoxifen Terbutaline Phenytoin Clozapine Quinine lidocaine clonidine Cimetidine propranolol 6 5 Ethionamide Warfarin Flumequine Enalapril Papaverine HCL Methotrexate Sulfacetamide Glyburide R =.996 Phenylbutazone 4 Furosemide Diazepam 3 Antipyrine Benzocaine pka(5) buffers

23 An example of poorly soluble compound. H + pka = 8.7 H pka=8.7 Tamoxifen AZ95-95 A. Ionization not close to chromophore. Low solubility.45 μm (ph 7.4), LogD=6.5 pka (DPAS): no pka (GlpKa): co-solvent, precipitation? pka (CE/MS): 8.7 (5), 7.88/7.76 (8) (ref: 7.6) ph meff res Fitting ption: x Mono-Base x x Value: Ionization: Ion. Func. Gr. 8 B x pka : x pka : V x pka 3: x r 4 pka7.8 :.99 V x x Comment: x x x x ph x Approved meff Ref. (Karl Box et al., Anal. Chem., 3, 75, ).

24 Long-term reproducibility of QC (5-8) QC AZ_ AZ_ AZ_3 AZ_4 AZ_5 AZ_6 AZ_7 Compound Atenolol Lidocaine Papaverine icotine Propranolol Quinine Codeine Structure H H H H H 3C H H H C H 3 C H H H H CH H H pka(b/b) Mean (n=5) (3.) (4.) 8.45 SD.9.9..(.7).8.(.9).8 Ref / (3.) (4.) CE/UV (3.) (4.4) 7.97 Seven compounds as on-line QC for pka screening (single measurement) Reproducibility <. pka units

25 Examples of determination of poorly soluble compounds 8.9 H S μ eff (a) Imipramine (b) Tamoxifen (c) Hexetidine (d) Ketoconazole (e) Mebendazole (f) meprazole (g) Indomethacin (h) Sulfacetamide H 3 C 4.4 (f) meprazole (ph) (b) Tamoxifen Possible to measure pka with solubility < µm pka =5% ionization (max. mobility)/ H 3 C H 3 C H 3 C H Mobility charge/mass (c) Hexetidine

26 pka fitting example ph.5-.5/ buffers(.3-.8/4 buffers).8? H H 3 C 8.9 H.5.5 mono-base mono-base/acid H meprazole 3 (ref)? H Debrisoquin S Sulfacetamide S H mono-acid di-base meprazole Clozapine Chlorpromazine Phenylbutazone Ampicillin Mebendazole Sulfacetamide Paracetmol Clonidine Cimetidine Debrisoquin Ethionamide Antipyrine

27 Summary Pressure-assisted CE/MS for pka: High throughput screening pka from.3 to.8 Reproducibility and accuracy of pka ±. units Insensitive to compound purity, requiring minute sample Concn: (- µm) beneficial for poorly soluble cpds Independent on type of buffer, capillary length, DMS concn. Provide charge and structure/stability information (MS)

28 Acknowledgements Anders Holmén, Mats ågård, Walter Lindberg, Yudong Wang, Marie Englund, Richard Thompson, Johan Ulander H. Wan, A. Holmén, M. ågård, W. Lindberg, J. Chromatogr. A. 979 () 369. H. Wan, A. Holmén, Y. Wang, W. Lindberg, M. Friberg, M. ågård, R. Thompson, Rapid Commun. Mass Spectrom. 7 (3) 639. H. Wan, R. Thompson; Drug Discovery Today, Technologies, (5), 7. H. Wan, J. Ulander, Expert pinion on Drug Metabolism & Toxicology, (6), 389. Eva Emanuelsson (validation), Fredrik Bergström (new pka-fit template test) Eva Thorin (Activitybase/Dixy group), Michael Wirth Färdigh (automation) Zhiping You (new pka-fit program), AZ Boston Lead generation, Physical Chemistry group, AZ Mölndal