Strategies for In-House Development of Chiral Separations for Routine Analytical Applications

Transcription

1 Strategies for In-ouse Development of Chiral Separations for Routine Analytical Applications T EAS 2008 Thomas E. Beesley Dr. J. T. Lee Dave Bell

2 Strategy: Use generic screening to develop options for a variety of potential applications Quick analytical method (possibly suitable for later optimization and validation); sample solubility options Suitable opportunities for possible small scale prep Applicable to trace analysis of unwanted isomer: reversal of elution order possibilities Provides impurity profiling Applicable to LC-MS methodologies

3 Tactics Choose a set of CSPs: that are broad-based to increase chances of success for a wide range of molecular types that offer selectivity in a wide range of mobile phases for increased selectivity possibilities and sample solubilities that are complementary to each other in order to increase overall hit rate where hits (ie. partial separations) can be quickly and easily optimized to baseline separation.

4 Modern Chiral Stationary Phases Polymeric Small molecule ligands Synthetic Methacrylate Polycyclic amine-2 Natural Cellulose Amylose Proteins Copper complex-2 π-complex Crown ether Cyclodextrin-12 Macrocyclic glycopeptides-6 Astec-Supelco Phases

5 Published Statistics 53 Chiral Compounds % Positive CSP 87% AS, AD, D, J Mobile Phases No. perating Parameters % V, T, R % Combined 7 26 *If P-RSP had been added, % would have gone to 74% due to antifungal agents. *If DNP had been added in the PM, % would have gone to 90%. Ref: Evaluation of Generic Liquid Chromatography Screens for Pharmaceutical Analysis, Andersson, M.E., Aslan, D., Clarke, A., Roeraade, J. agman, G., Journal of Chromatography A, 1005 (2003) CIRALCEL and CIRALPAK are registered trademarks of Daicel Chemical Industries Ltd.

6 Complementary Method Development CIRALCEL/PAK*: Compound must be in neutral form - interaction always non-ionic Separate samples into acids, bases and neutrals (neutrals can be screened with either acids or bases) CIRBITIC: Compound must be ionized or a salt; ionic interactions are a key mechanism Same mobile phase screens are used for all samples, but can choose selective screening for acid, bases or neutrals Note: Functional group on or near stereogenic center dictates whether analyte is acid or base * Trademarks of Daicel/Chiral Technologies, West Chester,PA

7 Teicoplanin, CIRBITIC T Structure of Teicoplanin CSP sugar and alkyl chain Key sites NR C 2 ionic site NCC 3 N C C 2 Cl N A B N N C C 2 Cl N D N ionic site N 2 ionic site N C Teicoplanin Aglycone, CIRBITIC TAG N A Cl B N N C Cl N N D N 2 ionic site

8 Comparison CIRBITIC V2 vs. V Enantio-separation of Ritalin by CIRBITIC V2 (250x4.6mm) (top) and CIRBITIC V (bottom), respectively. Mobile phase is 95/5, Me/20mM N4Ac, p 4.1 and the flow rate is 1 ml/min (ambient temperature).

9 3-Point Chiral Interactions on Multi-modal CIRBITIC Phases The most plausible interaction forces in descending order of strength under different mobile phase conditions: 1. Polar Ionic Mode (ionizable compounds only) A. Ionic B. ydrogen Bonding C. Steric/π-π 2. Polar rganic/normal Phase Mode (neutral compounds) A. ydrogen Bonding B. π-π C. Steric/Dipole 3. Reversed Phase Mode (all compounds) A. Ionic B. ydrogen Bonding C. Steric/Inclusion/ydrophobic

10 Comparison Four Basic Mobile Phase Types for CIRBITIC CSPs on same column Compound is: Ionizable Neutral Neutral/Polar All types Polar Ionic mode: Methanol + Acid + Base ( , v/v/v) or Methanol + Volatile Ammonium Salt ( % v/w) Normal Phase mode: Polar + Nonpolar: Et+eptane; 70/30;v/v Polar organic mode: Polar/Nonpolar: Me; Et or ACN or combinations, eg Me/ACN Reversed Phase mode: rganic + Aqueous Buffer: ACN+TEAA or N4Ac

11 Conversion Reversed Phase to Normal Phase Using MtBE RP NP Methyl phenyl sulfoxide separation by CIRBITIC V (250x4.6mm). Top: 20/80, TF/20 mm N4N3. Bottom: 97/2/1/, MtBE/ACN/Me.

12 Method ptimization in Reversed-Phase Mode rganic modifier type and composition can dramatically effect retention and selectivity ther parameters to consider: - Temperature - Flow rate - rganic type - p Retention (min) Fluoxetine in Reversed-Phase Mode S-enantiomer R-enantiomer separation factor (alpha) Selectivity - Buffer % Me The graph shows both reversed-phase and normal-phase behavior for fluoxetine on CIRBITIC V2 as a function of methanol composition - selectivity is optimal between 65% and 70% methanol - Studies showed that methanol acted as a better organic modifier than ACN

13 Reversed Phase Solvents Factors Influencing a Separation p: CIRBITICS rganic modifier: Me, ACN Buffer type and concentration Flow rate: 0.2 to 1.0 ml/min Temperature: 0 to 50 C

14 Temperature Study During Method Development P.K. wens, et. al. J.Pharm.Bomed. Anal. 2001, 25, Courtesy of Elsevier Science. Compound: Captopril diastereoisomers Column: Chirobiotic T (4.6 mm x 250 mm) Mobile phase: 0.05% TEAA (p3.8) buffer Flow rate: 1.0 ml/min Detection: mass spectrometer in negative ion mode scanning between 210 and 220 mass units

15 Polar Ionic Mode: Ionizable analytes Composition: Dominant interactions: Me + Ac + TEA Ionic interaction, hydrogen bonding Type of CSPs: Macrocyclic glycopeptides only - CIRBITIC V2, T, R and TAG

16 Salt Effects in the Polar Ionic Mode Acid: Atrolactic acid CIRBITIC T2 Base: Mianserin CIRBITIC V2 mau /0.1% Me/N4TFA mau /0.1% Me/N4TFA min min mau /0.1% Me/N4Formate mau /0.1% Me/N4Formate min min mau /0.1% Me/N4Ac mau /0.1% Me/N4Ac min min

17 Example Importance of Acid/Base Ratios In PIM Mobile Phase 100/0.1/0.1 Peak min. Peak min. Ratio: 1:1 CIRBITIC V Me/Acid/Base Mianserin 3 C N N 100/0.15/0.05 Peak min. Peak min. Ratio: 3:1 Nitrogen on Mianserin group is N(+), C on V is (-) Note: short retention times 100/0.05/0.15 Peak min. Peak min. Ratio: 1:3 Nitrogen on Mianserin group as free amine, but C on V is fully charged: weak ionic interaction

18

19 Test Range of Racemic Switches- Representatives of Popular Chiral Drugs Basic racemates: Analgesic: CNS stimulant: Gastroprokinetic: Antihelmetic: Decongestant: Antipsychotic: Muscle relaxant: Sedative: Methadone; Nefopam; Tramadol Methylphenidate Mosapride xamniquine Pseudoephedrine Thioridazine Tolperisone Zopiclone

20 Test Range of Racemic Switches- Representatives of Popular Chiral Drugs Basic racemates continued: Bronchodilator: Antihypertensive: Antifungal: Anesthetic: Antihistaminic: Antidepressant: Albuterol; Clenbuterol; Epinephrine; Formoterol; Isoproterenol; Terbutaline Amlodipine; Lercanidipine; Metoprolol; Nicardipine; Propranolol; Sotalol Miconazole Bupivacaine Chlorpheniramine Citalopram; Fluoxetine; Mianserin; Nefopam;Sertraline; Trimipramine

21 Test Range of Racemic Switches- Representatives of Popular Chiral Drugs Acidic racemates: Anti-inflammatory: Ibuprofen; Ketoprofen; Naproxen Neutral racemates: Antiulcerative: meprazole Anxiolytic: Lorazepam; xazepam Pediculicide: cis-permethrin Anticoagulant: Warfarin

22 Racemic Switches: ptimized Conditions CIRBITIC Phases Compounds Mobile Phase Column k1 α Rs Albuterol 100/0.1w%, Me/N4TFA T Amlodipine 100/0.1w%, Me/N4TFA V Bupivacaine 100/0.1w%, Me/N4TFA V Citalopram 100/0.1w%, Me/N4TFA V Clenbuterol 100/0.1w%, Me/N4TFA T Fluoxetine 100/0.1w%, Me/N4TFA V Formoterol 100/0.6/0.4, Me/Ac/TEA T

23 Racemic Switches: ptimized Conditions CIRBITIC Phases con t Compounds Mobile Phase Column k1 α Rs Ibuprofen 10/90, TF/20mM NaCitrate, p6.3 V Isoproterenol 100/0.1/0.1, Me/Ac/TEA T Ketoprofen 20/80, Me/0.1% TEAA, p 6.0 R Lercanidipine 100/0.02w%, Me/ N4TFA V Lorazepam 100 % Me T Methylphenidate 100/0.1w%, Me/N4TFA V Metoprolol 100/0.1w%, Me/N4TFA T Mianserin 100/0.1w%, Me/N4Formate V

24 Racemic Switches: ptimized Conditions CIRBITIC Phases con t Compounds Mobile Phase Column k1 α Rs Mosapride 30/70, ACN/0.1% TEAA, p 4.1 V Naproxen 10/90, TF/NaCitrate,20mM, p6.3 V Nicardipine 100/0.02w%, Me/N4TFA V xamniquine 100/0.1w%, Me/N4TFA V xazepam 100% Me T cis-permethrin 99.5/0.5, ex/et TAG Propranolol 100/0.1w%, Me/N4TFA T Pseudoephedrine 100/0.1w%, Me/N4TFA T

25 Racemic Switches: ptimized Conditions CIRBITIC Phases con t Compounds Mobile Phase Column k1 α Rs Sotalol 100/0.2/0.1, Me/Ac/TEA T Terbutaline 100/0.1w%, Me/N4TFA T Thalidomide 100% Me V Tolperisone 100/0.1w%, Me/N4TFA V Trimipramine 100/0.1w%, Me/N4TFA V Warfarin 40/60, Et/0.1% TEAA, p 4.1 V

26 Racemic Switches: ptimized Conditions CYCLBND Phases Compounds Mobile Phase Column k1 α Rs Chlorpheniramine 10/90, ACN/0.1% TEAA, p4.1 β-cd Chlorthalidone 15/85, ACN/10mM N4Ac, p4.0 P-RSP Epinephrine 25/75, ACN/20mM N4Ac, p4.1 P-RSP Methadone 20/80, ACN/ 10mM N4Ac, p 3.6 P-RSP Miconazole 25/75, ACN/10mM N4Ac, p 4.0 P-RSP Nefopam 25/75, ACN/20mMN4Ac,p 4.1 DNP meprazole 100/0.4/0.1, ACN/Ac/N4 DMP Sertraline 30/70, ACN/20mM N4Ac, p4.1 P-RSP Thioridazine 20/80, ACN/0.1% TEAA, p4.1 β-cd Tramadol 20/80, ACN/20mM N4Ac, p5.5 DMP Zopiclone 95/5/0.3/0.2,ACN/Me/Ac/TEA β-cd

27 Statistics f forty racemic switch drugs tested in this study: The success rate for CIRBITIC phases is 80%. The success rate for CYCLBND phases is 40%. Eight compounds can be separated by both types (20%). Resolution range: CYCLBND Phases: 1.5 to 2.8 CIRBITIC Phases: 1.5 to 11.0

28 ptions from Results of Dual Screen meprozole (Nexium) 3 C N N S N C 3 C 3 C 3 CIRBITIC R (NP) CIRBITIC R (RP) CYCLBND I 2000 DMP (PM) R(+) 13.4 min R(+) 24.5 min S(-) 27.5 min R(+) 10.8 min S(-) 15.5 min S(-) 14.1 min 40/60: Et/eptane 30/70: Me/10mM N4Ac, p /0.4/0.1: ACN/Ac/N4

29 Accomodation diverse structures with same chiral environment: Beta-blocker Separations Simultaneous chiral separations of 4 beta blockers using CIRBITIC T (250x4.6mm in polar ionic mode. The mobile phase is 15 mm Ammonium Formate in C3, 1mL/min (ambient temperature).

30 Accomodates multiple chiral centers: Separation of 2,3-Dihydroxy-3-phenyl-propionic acid enantiomers on CIRBITIC R Racemate A Racemate B B B A A Column: CIRBITIC R, 5 μm, 25cm x 4.6mm Mobile Phase : 50:50; Me/.01% N4C2, p 4.1 Flow Rate: 1 ml/min. Temperature: Ambient. Detection: Injection: 2,3-Dihydroxy-3-phenyl-proprionic acid isomers: Racemate A (4.85 and 6.95 min.) Racemate B (5.33 and 6.29 min.) Figure courtesy DSM Fine Chemicals

31 Reverse Elution rder Reversal of elution order between CIRBITIC V2 (top) and CIRBITIC T2 (bottom) columns under the exact same condition. Mobile phase is 15mM N4 formate in Me and the flow rate is 1 ml/min. The polarimeter showed the first peak is (-) on the CIRBITIC V2 while the first peaks is (+) on the CIRBITIC T2.

32 Ion Repulsion Ionic effects showing the repulsion phenomenon between CSP (CIRBITIC T, 250x4.6mm) and the analyte (mandelic acid) in 30/70, Me/TEAA. At p 5 (0.1% TEAA), the first peak eluted before column void (3 ml). igher concentrations (0.5% TEAA) of buffer can alleviate that effect.

33 Ion Repulsion CIRBITIC T (250x4.6mm) Dansyl methionine. igher salt concentrations can reduce this response. The column void is 3.0 ml and the flow rate is 1 ml/min.

34 Dorzolamide *Cl Method Development Processed through method development screen: 3 CIRBITIC phases: V2,T,TAG 3 CYCLBND phases: B-CD,DNP,P-RSP. C 3 S S S 3 C S S S N C 3 N 2 N C 3 N 2 Dorzolamide ydrochloride 4S,6S Dorzolamide ydrochloride Related Compound A 4R,6R

35 Spectrum Column mode elution File CIRBITIC TAG RP No Retention C:\ascii\AC D_Chrom1 373.cdf CIRBITIC TAG PIM Separation C:\ascii\AC D_Chrom1 384.cdf CIRBITIC V2 RP Separation C:\ascii\AC D_Chrom1 397.cdf CIRBITIC V2 PIM Separation C:\ascii\AC D_Chrom1 404.cdf CIRBITIC T RP No Separation C:\ascii\AC D_Chrom1 416.cdf CIRBITIC T PIM No Separation C:\ascii\AC D_Chrom1 423.cdf Cyclobond I 2000 RP No Retention C:\ascii\AC D_Chrom1 435.cdf Cyclobond I 2000 PM Unknown C:\ascii\AC D_Chrom1 442.cdf Cyclobond 2000 P-RSP RP No Retention C:\ascii\AC D_Chrom1 461.cdf Cyclobond 2000 P-RSP PM Unknown C:\ascii\AC D_Chrom1 485.cdf Cyclobond 2000 DNP RP No Separation C:\ascii\AC D_Chrom1 497.cdf Cyclobond 2000 DNP PM Unknown C:\ascii\AC D_Chrom1 504.cdf

36 Chromatographic Results PIM mode best! Dorzolamide Cl (4S,6S) Peak 1 (Rt1): min. Peak 2 (Rt2): min. 1 Dorzolamide Cl Rel. Compd. A (4R,6R) Time (min) CIRBITIC V2, 250x4.6mm I.D., 5μ particles (15024AST) 100:5 Me:2, 3.81mM N4TFA (or 0.05% N4TFA) 22 C, 0.3mL/min., UV-254nm Inj. Vol. 10 mg/ml in Me

37 CIRBITIC LC Columns in LC-MS-MS PIM mode (nonaqueous): igh enantioselectivity in 100% Me with acid/base or salts, short Rt times, works well with ESI Me/Ac/TEA; 100/0.1/0.1 or Me/0.1% N4Ac, N4TFA or N4C2 RP mode (aqueous): Compatible with ammonium acetate or formate - No problem with principle solvents used in APCI (from alcohols to hydrocarbons, DMS, DMF)

38 Column Coupling: Separation of Ritalin and Its Metabolite Column: CIRBITIC T2 (20x4.0mm) + CIRBITIC V2 (150x4.6mm) Mobile Phase: 93/7: Me/20 mm N4Ac, p 4.1 Flow Rate: 1.0 ml/min mau Ritalinic acid/t2 Ritalin/V min

39 bjective: Bioanalytical Sample Preparation - Concentrate analytes of interest - Selectively separate analytes of interest from endogenous interferences inherent with the sample - more compatible sample matrix for analytical system Available techniques: dialysis, monlithic chromatography, SFC extraction, protein precipitation, liquid/liquid extraction, SPE

40 Phospholipid & Ion-Suppression Phospholipids: primary constituent of cell membranes & documented to be a major cause of ion-suppression in the positive ion electrospray mode (+ESI) Present in extremely high concentrations in biological matrixes & represent the second largest lipid component next to triglycerides and fats

41 ow are Phospholipids Selectively Removed using ybridspe? Si Si Si Si Zr 2+ Zr + The phosphate moiety of phospholipids is a strong Lewis base (electron donor) that interacts Zr atoms coated on the silica service. Si Si Proprietary Zirconia coated Silica Note: The presence of 1% formic acid in the MeCN precipitation agent is critical because of: 1) formic acid is a stronger Lewis base than most carboxyl (-C) groups found in acidic pharmaceutical compounds but not as strong a Lewis base as the phosphate moiety found in phospholipids; and 2) the low p environment neutralizes residual silanol activity on the silica surface thereby eliminating secondary cation-exchange interaction with basic compounds of interest. The Zr atom acts as a Lewis acid (electron acceptor) because it has empty d- orbitals.

42 ff-line Precipitation Method for ybridspe 1 ml cartridge or 96-well format 2) Centrifuge 3) Transfer supernatant to ybridspe cartridge or 96- well plate Retained Phospholipids 1) Precipitate Proteins: Combine 100 µl plasma serum with 300 µl 1% formic acid in acetonitrile. Add I.S. as necessary. 4) Apply vacuuum. Phospholopids are retained on ybridspe sorbent while small molecules (e.g., pharma compounds & metabolites) pass through un-retained. 5) Resulting filtrate/eluate is free of proteins and phospholopids and ready for immediate LC-MS/MS analysis; or it can be evaporated and reconstituted as necessary prior to analysis

43 Extraction of 10 ng/ml Clenbuterol from Dog Plasma TIC of +MRM (2 pairs): Exp 1, from Sample 9 ( ) of wiff (Turbo Spray) Max cps. TIC of +MRM (2 pairs): Exp 1, from Sample 8 ( ) of wiff (Turbo Spray) Max cps Standard Protein PPT: 1) 100 µl spiked dog plasma µl 1% formic acid in MeCN 2) Centrifugation 3) Inj. of supernatant => 60% Abs Recov Phospholiipids Clenbuterol SPE Procedure: 9 step SPE procedure optimized for trace analysis of clenbuterol Clenbuterol Phospholiipids Time, min Time, min TIC of +MRM (2 pairs): Exp 1, from Sample 6 ( ) of wiff (Turbo Spray), SG Smoothed (10) Max cps ybridspe-ppt: 1) 100 µl spiked dog plasma µl 1% formic acid in MeCN 2) centrifuge 3) Load sup onto ybridspe 1 ml cartiridge 4) Inj. of eluate => 76% Abs Recov Clenbuterol Phospholiipids Column: Chirobiotic T 10cm X 2.1mm, 5 um Mobile Phase: 10mM Ammonium Formate/Methanol Flow: 300 ul/min Temp: 30 C Detection: MS-MRM Time, min

44 ybridspe Recovery of Representative Pharma Compounds Summary of Abs. Recovery from Plasma: Compound Name Spike Conc. Matrix Abs. Recovery Propranolol 100ng/ml Rat plasma 62% Ketoprofen 100ng/ml Rat plasma 82% Doxepin 7.5ng/ml rat plasma 87% Imipramine 7.5ng/ml rat plasma 92% Amitriptyline 7.5ng/ml rat plasma 90% xyphenbutazone 20ng/ml rat plasma 104% Ketoprofen 20ng/ml rat plasma 92% Naproxen 20ng/ml rat plasma 93% Phenylbutazone 20ng/ml rat plasma 99% Flunixin 20ng/ml rat plasma 76% Procainamide 12.5ng/ml rat plasma 42% Doxepin 12.5ng/ml rat plasma 89% Mirtazapine 12.5ng/ml rat plasma 56% Dextromethorphan 12.5ng/ml rat plasma 88% p-coumaric Acid 12.5ng/ml rat plasma 54% Clenbuterol 10ng/ml dog plasma 75% Sulfamethizone 10ng/ml dog plasma 101% Sulfacetamide 10ng/ml dog plasma 98% Sulfamerazine 10ng/ml dog plasma 94% Sulfathiazole 10ng/ml dog plasma 98% Sulfanilamide 10ng/ml dog plasma 96% Sulfadiazine 10ng/ml dog plasma 49% Summary of Abs. Recovery Solvent (no plasma): During the course of development, 12 mixes containing 5-8 unique/representative pharma compounds were prepared at a concentration600 ng/ml (diluted in 1% formic acid in MeCN:Water (75:25, v/v)) total of 66 compounds. Each test mix was passed through a 1 ml ybridspe cartridge and the eluate was collected for LC-MS analysis. Recovery was assessed by measuring the analyte response before and after ybridspe processing. An average absolute recovery of 83% observed for the 66 compounds tested. The majority of the analytes tested achieved recoveries > 80%

45 Bupivacaine in Plasma by LC/MS-ESI Column: CIRBITIC V2, 150x2.1 mm Mobile Phase: 90/10, Me/10mM N 4 Ac, p 4.1 Flow Rate: 0.2 ml/min Excellent Correlation Coefficients > Linear Range = 0.15ng-50ng/mL

46 Supelco Chiral Services Laboratory

47 Conclusions A successful high throughput screening system consisting of 4 CIRBITIC phases and 4 CYCLBND phases has been devised employing 8 mobile phases. Polar ionic mode offers excellent opportunities for LC/MS and preparative applications. A 100% hit rate for 40 divergent racemic switches was obtained with a 20% overlap and resolutions from 1.5 to LC-MS/MS technology offers excellent benefits in terms of throughput and sensitivity; but good sample prep is still required. ybrid SPE aids speed by removal of phospholipid interferences.