Sample Preparation in LC-MS Bioanalysis

List of Contributors

Gilberto Alves, PhD

CICS‐UBI – Health Sciences Research Centre

University of Beira Interior

Covilhã

Portugal

Miguel Ángel Bello‐López, PhD

Department of Analytical Chemistry

Universidad de Sevilla

Sevilla

Spain

Matthew Barfield, PhD

Research and Development

GlaxoSmithKline Pharmaceuticals

Ware

UK

Michael G. Bartlett, PhD

Department of Pharmaceutical and Biomedical Sciences

University of Georgia

Athens, GA

USA

Babak Basiri, PhD

Department of Pharmaceutical and Biomedical Sciences

University of Georgia

Athens, GA

USA

Ian A. Blair, PhD

Department of Systems Pharmacology and Translational Therapeutics

Perelman School of Medicine

University of Pennsylvania

Philadelphia, PA

USA

Chester L. Bowen, MS

Research and Development

GlaxoSmithKline Pharmaceuticals

Collegeville, PA

USA

Stacy Brown, PhD

Department of Pharmaceutical Sciences

Gatton College of Pharmacy at East Tennessee State University

Johnson City, TN

USA

Pilar Campíns‐Falcó, PhD

Química Analítica

Universitat de València

Burjassot

Spain

Jennifer Carmical, PharmD

Department of Pharmaceutical Sciences

Gatton College of Pharmacy at East Tennessee State University

Johnson City, TN

USA

Zhongzhe Cheng, PhD

School of Pharmacy

Weifang Medical University

Weifang, Shandong

China

Theo de Boer, PhD

LC‐MS Bioanalysis

Ardena Bioanalytical Laboratory (ABL)

Assen

The Netherlands

Myriam Díaz‐Álvarez, MSc

Department of Environment

Instituto Nacional de Investigación y Tecnología Agraria y Alimentaria (INIA)

Madrid

Spain

Fuyou Du, PhD

Department of Applied Chemistry

Guilin University of Technology

Guilin, Guangxi

China

Amílcar Falcão, PhD

Laboratory of Pharmacology

Faculty of Pharmacy

University of Coimbra

Coimbra

Portugal

Rut Fernández‐Torres, PhD

Department of Analytical Chemistry

Universidad de Sevilla

Sevilla

Spain

Ana Fortuna, PhD

Laboratory of Pharmacology

Faculty of Pharmacy

University of Coimbra

Coimbra

Portugal

Yunlin Fu, MS

Pharmacokinetic Sciences

Novartis Institutes for BioMedical Research

East Hanover, NJ

USA

Hong Gao, PhD

Drug Metabolism & PharmacokineticsVertex Pharmaceuticals

Boston, MA

USA

Rodrigo A. González‐Fuenzalida, PhD

Química Analítica

Universitat de València

Burjassot

Spain

Rosa Herráez‐Hernández, PhD

Química Analítica

Universitat de València

Burjassot

Spain

Bruce J. Hidy, BSc

R&D, PPD

Richmond, VA

USA

Samuel Hofbauer, BS

Department of Systems Pharmacology and Translational Therapeutics

University of Pennsylvania

Philadelphia, PA

USA

Mike (Qingtao) Huang, PhD

Clinical Pharmacology

Akros Pharma Inc.

Princeton, NJ

USA

Rand G. Jenkins, BSc (retired)

PPD

Mechanicsville, VA

USA

Allena J. Ji, PhD, NRCC, DABCC

Biomarkers & Clinical Bioanalyses‐Boston, Sanofi

Framingham, MA

USA

Wenying Jian, PhD

Janssen Research & Development, LLC

Spring House, PA

USA

Hongliang Jiang, PhD

Tongji School of Pharmacy

Huazhong University of Science and Technology

Wuhan, Hubei

China

Neus Jornet‐Martinez, PhD

Química Analítica

Universitat de València

Burjassot

Spain

Maria Kechagia, MSc

Chemistry Department

Aristotle University of Thessaloniki

Thessaloniki

Greece

Jaeah Kim, PhD

Department of Pharmaceutical and Biomedical Sciences

University of Georgia

Athens, GA

USA

Maria Kissoudi, MSc

Chemistry Department

Aristotle University of Thessaloniki

Thessaloniki

Greece

Fumin Li, PhD

R&D, PPD

Middleton, WI

USA

Ning Li, PhD

Department of Pharmaceutical Analysis

School of Pharmacy

Shenyang Pharmaceutical University

Shenyang, Liaoning

China

Wenkui Li, PhD

Pharmacokinetic Sciences

Novartis Institutes for BioMedical Research

East Hanover, NJ

USA

Ang Liu, PhD

Bioanalytical Sciences

Translational Medicine

Bristol‐Myers Squibb

Princeton, NJ

USA

Rao N.V.S. Mamidi, PhD, DABT

Janssen Research & Development, LLC.

Raritan, NJ

USA

Yan Mao, PhD

Drug Metabolism & Pharmacokinetics

Boehringer Ingelheim Pharmaceuticals, Inc.

Ridgefield, CT

USA

Antonio Martín‐Esteban, PhD

Department of Environment

Instituto Nacional de Investigación y Tecnología Agraria y Alimentaria (INIA)

Madrid

Spain

Henri Meijering, MSc

LC‐MS Bioanalysis

Ardena Bioanalytical Laboratory (ABL)

Assen

The Netherlands

Clementina Mesaros, PhD

Department of Systems Pharmacology and Translational Therapeutics

University of Pennsylvania

Philadelphia, PA

USA

Akira Namera, PhD

Department of Forensic Medicine

Graduate School of Biomedical and Health Sciences

Hiroshima University

Hiroshima

Japan

Ragu Ramanathan, PhD

Medicine Design – ADME Sciences, Pfizer, Inc

Groton, CT

USA

María Ramos‐Payán, PhD

Department of Analytical Chemistry

Universidad de Sevilla

Sevilla

Spain

Márcio Rodrigues, PhD

CICS‐UBI – Health Sciences Research Centre

University of Beira Interior

Covilhã

Portugal

Guihua Ruan, PhD

Department of Applied Chemistry

Guilin University of Technology

Guilin, Guangxi

China

Takeshi Saito, PhD

Department of Emergency and Critical Care Medicine

Tokai University School of Medicine

Isehara

Japan

Ashkan Salamatipour, BS

Department of Systems Pharmacology and Translational Therapeutics

University of Pennsylvania

Philadelphia, PA

USA

Victoria F. Samanidou, PhD

Chemistry Department

Aristotle University of Thessaloniki

Thessaloniki

Greece

Nico van de Merbel, PhD

PRA Health Sciences

Assen

The Netherlands

Cong Wei, PhD

Drug Metabolism & Pharmacokinetics, Vertex Pharmaceuticals

Boston, MA

USA

Zongyu Wei, MS

Department of Applied Chemistry

Guilin University of Technology

Guilin, Guangxi

China

Naidong Weng, PhD

Janssen Research & Development, LLC.

Spring House, PA

USA

John Williams, PhD

Drug Metabolism & Pharmacokinetics Vertex Pharmaceuticals

Boston, MAUSA

Xin Xiong, MS

Department of Pharmacy

Peking University Third Hospital

Beijing

China

Long Yuan, PhD

Bioanalytical Sciences

Bristol‐Myers Squibb

Princeton, NJ

USA

Qiulian Zeng, MS

Department of Applied Chemistry

Guilin University of Technology

Guilin, Guangxi

China

Jun Zhang, PhD

Dynamega LLC

Lake Forest, IL

USA

Dafang Zhong, PhD

Shanghai Institute of Materia Medica

Chinese Academy of Sciences

Shanghai

China

Yunting Zhu, PhD

Shanghai Institute of Materia Medica

Chinese Academy of Sciences

Shanghai

China

Preface

Sample preparation is a pivotal part of the integral LC‐MS bioanalysis, which has been heavily employed in the determination of drugs, drug metabolites, biomarkers, and other molecules of interest in various biological matrices (e.g. fluids or tissues) for decades. It has been playing an important role in a variety of human healthcare studies, ranging from drug discovery and development, therapeutic drug monitoring, to biomarker analysis. While highly sophisticated LC‐MS systems with better sensitivity and higher bioanalytical throughput have been continuously introduced, challenges that remain unchanged are the sample preparation prior to LC‐MS quantitation, for which data quality has direct impact on study conclusion.

The purpose of sample preparation is not only to make the analyte(s) of interest available in sample extracts at an appropriate concentration for MS detection but also to remove interfering matrix elements (e.g. phospholipids and salts) that, if not addressed properly, can alter MS response (e.g. signal suppression). In quantitative LC‐MS bioanalysis, clean sample extracts means: (i) better chromatography, (ii) lower limit of quantification, (iii) decreased assay variability (due to reduced matrix effects), (iv) less chance of false‐positive/negative results, (v) longer column lifetime, (vi) less instrument downtime, and (vii) minimized costs in manpower and equipment maintenance, etc. In practice, the best sample preparation strategies should always be considered, evaluated, and implemented whenever possible in developing a robust quantitative LC‐MS bioanalytical method.

As a companion for the previously published Handbook of LC‐MS Bioanalysis: Best Practice, Experimental Protocols and Regulations (Li, Zhang, and Tse, 2013, Wiley), the current book is to provide a timely and comprehensive update along with representative experimental protocols on all important sample preparation techniques for quantitative LC‐MS bioanalysis of small and large molecules. The 26 chapters of the book are divided into three parts. The first part of the book is focused on not only the basic but also the contemporary sample preparation techniques in LC‐MS bioanalysis. These include Protein Precipitation, Liquid–Liquid Extraction, and Solid‐Phase Extraction (Chapter 1), Online Extraction and Column Switching (Chapter 2), Equilibrium Dialysis, Ultracentrifugation, and Ultrafiltration (Chapter 3), Phospholipid Depletion (Chapter 4), Salting‐out Assisted Liquid–Liquid Extraction (SALLE) (Chapter 5), Supported Liquid Extraction (SLE) (Chapter 6), Immunocapture (Chapter 7), Microextraction (Chapter 8), Microsampling (Chapter 9), Extraction via Nanomaterials (Chapter 10), Extraction via Molecularly Imprinted Polymers (MIP) (Chapter 11), Stir‐bar Sorptive Extraction (Chapter 12), Monolithic Spin Column Extraction (Chapter 13), Aptamer‐based Sample Preparation (Chapter 14), and Sample Extraction via Electromembranes (Chapter 15).

In Part II, the current sample preparation techniques for LC‐MS bioanalysis of biological sample matrices other than common whole blood, plasma, or serum are discussed in detail along with experimental protocols. These matrices include but are not limited to Tissues, Hair, Nail, Skins, and Bones (Chapter 16), Peripheral Blood Mononuclear Cells (Chapter 17), Urine, Cerebrospinal Fluid, Synovial Fluid, Sweat, Tears, and Aqueous Humor (Chapter 18), and Liposomal Samples (Chapter 19).

Part III of the book is focused on sample preparation for LC‐MS bioanalysis of challenging molecules. This part starts with some Key Pre‐analytical Considerations in Quantitative LC‐MS Bioanalysis (Chapter 20), which is followed by Derivatization strategies for enhancing assay sensitivities in quantitative LC‐MS bioanalysis of molecules with poor ionization efficiency (Chapter 21). Sample preparation for quantitative LC‐MS bioanalysis of Lipids is captured in Chapter 22. In Chapter 23, detailed instructions and associated stepwise protocols are provided for LC‐MS bioanalysis of peptides. Expanding from peptides, detailed instructions of sample preparation for LC‐MS bioanalysis of Proteins, Oligonucleotides, and Antibody–drug Conjugates (ADCs) are captured in Chapters 24, 25, and 26, respectively.

Our purpose in committing to this project was to provide scientists in industry, academia, and regulatory agencies with all practical tricks in extracting various analyte(s) of interest from biological samples for LC‐MS quantification according to the current health authority regulations and industry practices. In this book we are confident that we have accomplished our goal. The book represents a major undertaking which would not have been possible without the contributions of all the authors and the support of their families. We also wish to thank the terrific editorial staff at John Wiley & Sons and give a special acknowledgment to Michael Leventhal, Managing Editor; Vishnu Narayanan, Project Editor; Beryl Mesiadhas, Project Manager; S. Grace Paulin Jeeva, Production Editor; and Robert Esposito, Associate Publisher, at John Wiley & Sons, for their premier support of this project.

Wenkui Li, PhD

Wenying Jian, PhD

Yunlin Fu, MS

List of Abbreviations

2D two‐dimensional 3NPH 3‐nitrophenylhydrazine 5‐FU 5‐fluorouracil 5‐HETE 5‐hydroxyeicosatetraenoic acid AA acrylamide AA alendronic acid AAC α1‐antichymotrypsin ACE angiotensin I converting enzyme ACE automatic cartridge exchange ACN acetonitrile ADA anti‐drug antibody ADC antibody–drug conjugate ADME absorption, distribution, metabolism, and excretion ADP adenosine diphosphate ADS alkyl‐diol‐silica AFA adaptive focused acoustics AFMC aptamer‐functionalized monolithic column AFMPC aptamer‐functionalized material‐packed column AFM aptamer‐functionalized material AFOTCC aptamer‐functionalized open tubular capillary column AFSC aptamer‐functionalized spin column AG 2‐arachidonoylglycerol AGP acid glycoprotein AIBN azo(bis) isobutyronitrile AML acute myeloid leukemia AMP adenosine monophosphate APA anti‐peptide antibody APCI atmospheric pressure chemical ionization Apt‐AC aptamer‐based affinity column Apt‐AuNR aptamer‐functionalized gold nanorod Apt‐MM aptamer‐functionalized magnetic material Apt‐MNP aptamer‐functionalized magnetic nanoparticle Apt‐PANCMA aptamer‐functionalized poly(acrylonitrile‐co‐maleic acid) Apt‐PP‐fiber aptamer‐based‐polypropylene fiber Apt‐SA‐SPE aptamer‐based surface affinity solid‐phase extraction Apt‐SBSE aptamer‐functionalized stir‐bar sorptive extraction Apt‐SPE aptamer‐based solid‐phase extraction Apt‐SPME aptamer‐based solid‐phase microextraction ATP adenosine triphosphate AUC area under the curve AuNP gold nanoparticle BAL bronchoalveolar lavage BEAD bead extraction and acid dissociation BEH bridged ethylene hybrid BLQ below limit of quantification BNP B‐type natriuretic peptide BP bisphosphonate BP‐3 benzophenone‐3 BPA bisphenol A BSA bovine serum albumin BSL‐2 biosafety level‐2 BSTFA N , O ‐bis(trimethylsilyl)trifluoroacetamide CAD collision‐activated dissociation Cape capecitabine CCSHLLE counter current salting‐out homogenous liquid–liquid extraction CDA cytidine deaminase CDI carbonyl diimidazole CDR complementarity‐determining region CE capillary electrophoresis CE cholestryl oleate CHAPS 3‐([3‐cholamidopropyl]dimethylammonio)‐1‐propanesulfonate CID collision‐induced dissociation CIP chiral imprinted polymer CNBF 4‐chloro‐3,5‐dinitrobenzotrifloride CNS central nervous system CNT‐PDMS carbon nanotube–poly(dimethylsiloxane) CNT carbon nanotube COXs cyclooxygenases CPT cell preparation tube CSF cerebrospinal fluid CV coefficient of variation CZE‐C⁴D capillary zone electrophoresis with capacitively coupled contactless conductivity detection D distribution ratio D2EHPA di‐(2‐ethylhexyl)phosphoric acid DAD diode array detection DADPA diaminodipropylamine DAG diacylglycerol (1,3‐dilinoleoyl‐rac‐glycerol) DAR drug‐to‐antibody ratio DBS dried blood spot DCM dichloromethane DEHP di‐(2‐ethylhexyl) phosphate DEME dynamic electromembrane extraction DEX dextromethorphan DI direct immersion DIC diclofenac DIEA diisopropylethylamine DI‐SDME direct immersion single‐drop microextraction DLLE dispersive liquid–liquid extraction DLLME dispersive liquid–liquid microextraction DMBA dimethylbutylamine DMF N,N ‐dimethylformamide DMSO dimethyl sulfoxide DNS‐Cl dansyl chloride DOPA dihydroxyphenylalanine DOR dextrorphan DP IV dipeptidyl peptidase IV DPBS Dulbecco’s phosphate‐buffered saline DPX disposable pipette extraction DSEA dansyl sulfonamide ethyl amine D‐SPE dispersive solid‐phase extraction DTT dithiothreitol DVB divinylbencene DXR doxorubicin EA ethyl acetate EBF European Bioanalytical Forum ECAPCI electro capture atmospheric pressure chemical ionization ED equilibrium dialysis EDC·HCl 1‐ethyl‐3‐(3‐dimethylaminopropyl) carbodiimide hydrochloride EDC/NHS N ‐(3‐dimethylamnopropyl)‐ N ‐ethylcarbodiimide hydrochloride/ N ‐hydroxysuccinimide EDTA ethylenediaminetetraacetic acid EG ethylene glycol EGDMA ethylene glycol dimethacrylate EHS ethylhexyl salicylate ELISA enzyme‐linked immunosorbent assay EME electromembrane extraction EME‐DLLME electromembrane extraction dispersive liquid–liquid microextraction EME‐LDS‐USAEME electromembrane extraction low‐density solvent‐based ultrasound‐assisted emulsification electromembrane microextraction EM‐SPME electromembrane‐surrounded solid‐phase microextraction ENB 1‐ethyl‐2‐nitrobenezene EPR enhanced permeation and retention ESI electrospray ionization EtOH ethanol FA fatty acid FA formic acid FBAL α‐fluoro‐β‐alanine FBS fetal bovine serum Fc fragment crystallizable region/constant region FcRn human neonatal Fc receptor FDA Food and Drug Administration FLD fluorescence detection FLM free liquid membrane Fmoc‐Cl 9‐florenylmethoxycarbonyl chloride FNME fiber‐packed needle microextraction GAC green analytical chemistry GC gas chromatography GC‐FID gas chromatography–flame ionization detection GC‐MS gas chromatography–mass spectrometry GIP glucose‐dependent insulintropic peptide GLP‐1 glucagon‐like peptide‐1 GMA glycidylmethacrylate GnRH gonadotropin‐releasing hormone GPChos glycerophosphatidylcholines GPCs glycerophosphatidylcholines GPE gum‐phase extraction GPI glycosylphosphatidylinositol HETP height equivalent to a theoretical plate HFIP 1,1,1,3,3,3‐hexafluoro‐isopropanol HF‐LPME hollow fiber liquid‐phase microextraction HILIC hydrophilic interaction liquid chromatography HIV human immunodeficiency virus HMS homosalate HND‐G high nitrogen‐doped graphene HNE human neutrophil elastase HPIM homemade polymer inclusion membrane HPLC high‐performance liquid chromatography HRMS high‐resolution mass spectrometry HS headspace HSSBSE headspace stir‐bar sorptive extraction HS‐SDME headspace single‐drop microextraction HTLC high‐turbulence liquid chromatography IACUC Institutional Animal Care and Use Committee IAE immunoaffinity extraction IAM iodoacetamide IA‐SPE immunoaffinity solid‐phase extraction IC immunocapture iCAT isotope‐coded affinity tag ICP‐MS inductively coupled plasma mass spectrometry ID internal diameter IGF insulin‐like growth factor IgG immunoglobulin G IL‐21 interleukin‐21 IMAC immobilized metal ion affinity chromatography IPA isopropanol IS internal standard ISET integrated selective enrichment target IS‐MRM in‐source multiple reaction monitoring ISR incurred sample reanalysis ISTD internal standard ITMS ion trap mass spectrometry iTRAQ isobaric tags for relative and absolute quantification IT‐SPME in‐tube solid‐phase microextraction IUPAC International Union of Pure and Applied Chemistry IV intravenous IVT in vitro transcription IX‐SPE ion exchange‐solid‐phase extraction Kb/p blood to plasma ratio Ke/p red blood cell partition coefficient LBA ligand‐binding assay LC liquid chromatography LC‐MS liquid chromatography–mass spectrometry LC‐MS/MS liquid chromatography–tandem mass spectrometry LC‐UV/FL liquid chromatography with ultraviolet/fluorescence detection LD liquid desorption LGPChos lysoglycerophosphocholines LLE liquid–liquid extraction LLOQ lower limit of quantification LOXs lipoxygenases LPCs lyso‐phosphatidylcholines LPME liquid‐phase microextraction LSC liquid scintillation counting MA methyacrylate, methyl acrylate MA minodronic acid MAA methacrylamide MAA methacrylic acid mAb monoclonal antibody MADB poly(methacrylic acid‐3‐sulfopropyl ester potassium salt‐co‐divinylbenzene) MAG monoacylglycerol (1‐stearyl‐rac‐l glycerol) MALDI matrix‐assisted laser desorption ionization MAX mixed‐mode anion exchange MCV mean cell volume MCX mixed‐mode cation exchange MDA‐LDL malondialdehyde‐modified low‐density lipoprotein MDMA 3,4‐methylenedioxy‐ N‐ methylamphetamine MDS myelodysplastic syndromes MeOH methanol MEPS microextraction by packed sorbent MF matrix factor MI‐MSPE molecularly imprinted micro‐solid‐phase extraction MIPs molecularly imprinted polymers MISPE molecularly imprinted solid‐phase extraction MISPE‐DPE molecularly imprinted solid‐phase extraction with differential pulsed elution MISPE‐PE molecularly imprinted solid‐phase extraction with pulsed elution MIST metabolites in safety testing MIT molecular imprinting technology MLLE micro‐liquid–liquid extraction MMA methylmalonic acid MMAE monomethyl auristatin E MMST monolithic molecularly imprinted polymer sol–gel packed tip MNP magnetic nanoparticle MPB 2‐bromo‐3′‐methoxyacetophenone mPGES‐1 microsomal prostaglandin E synthase‐1 MPS 3‐methacryloyloxypropyltrimethoxysilane MRM multiple reaction monitoring mRNA messenger RNA MS mass spectrometry MS/MS tandem mass spectrometry MSP magnetic supraparticle MSPD matrix solid‐phase dispersion MSPE magnetic solid‐phase extraction MTBE methyl tert ‐butyl ether MTBSTFA N ‐(tert‐butyldimethylsilyl)‐ N ‐methyl trifluoroacetamide MW molecular weights MWCNT multiwall carbon nanotube MWCO molecular weight cutoff NAaPs nucleic acid associated proteins NAb neutralizing antibody NA nucleic acid NCEs new chemical entities NEM N ‐ethylmaleimide NHS N ‐hydroxysuccinimide NK natural killer NPOE 2‐nitrophenyloctyl ether NPPE 2‐nitrophenyl pentyl ether NPs nanoparticles NSB nonspecific binding NSE neuron‐specific enolase NTproBNP N‐terminal pro‐B‐natriuretic peptide OC octocrylene OD‐PABA ethylhexyl dimethyl p ‐aminobenzoate ODS octadecyl OH‐PAH monohydroxylated polycyclic aromatic hydrocarbon OH‐PDMS hydroxyl polydimethylsiloxane OTT open tubular trapping OxLDL oxidized low‐density lipoprotein P partition ratio PA phosphatidic acid PA polyacrylate PA‐EG poly(methyl methacrylate/ethyleneglycoldimethacrylate) Pa‐EME parallel electromembrane extraction PALME parallel artificial liquid membrane extraction PANCMA poly(acrylonitrile‐co‐maleic acid) PAR peak area ratio PBD pyrrolobenzodiazepine PBMC peripheral blood mononuclear cell PBS phosphate‐buffered saline PBST phosphate‐buffered saline with Tween‐20 PCA perchloric acid PCB polychlorinated biphenyl PCI protein C inhibitor PCs phosphatidylcholines PD pharmacodynamics PD phospholipid depletion PDMS polydimethylsiloxane PE phosphoethanolamine PEG polyethylene glycol PEME pulsed electromembrane extraction PEO polyethylene oxide PE phosphatidylethanolamine PFB pentafluorobenzyl PG phosphatidylglycerol PGs prostaglandins PHMB 4‐(hydroxymercuri)benzoate PI phosphatidylinositol PK pharmacokinetics PK/PD pharmacokinetic/pharmacodynamic PK/TK pharmacokinetic/toxicokinetic PKU phenylketonuria PLs phospholipids PMMA pentamethylated minodronic acid PMMA poly(methyl methacrylate) PMSF phenylmethylsulfonyl fluoride poly(GMA‐co‐EDMA) poly(glycidyl methacrylate‐coethylene dimethacrylate) PP polypropylene PPB plasma protein binding PPESK poly(phthalazine ether sulfone ketone) PP‐fiber porous polymer‐coated fiber PPT protein precipitation PPY polypyrrole ProGRP pro‐gastrin releasing peptide PS phosphatidylserine PTFE polytetrafluorethylene PTV programmable temperature vaporize PU polyurethane foams PUFA polyunsaturated fatty acids QC quality control QTOF quadropole time‐of‐flight QuEChERS quick, easy, cheap, effective, rugged, safe extraction method RA risedronic acid RAM restricted access material RBC red blood cell REC extraction recovery RED rapid equilibrium dialysis rhTRAIL recombinant human tumor necrosis factor‐related apoptosis‐inducing ligand RISC RNA‐induced silencing complex ROS reactive oxygen species RP reversed phase RP‐SPE reversed‐phase solid‐phase extraction RPV rilpivirine SA‐EME surfactant‐assisted electromembrane extraction SALLE salting‐out assisted liquid–liquid extraction SAX strong anion exchange SBSE stir‐bar sorptive extraction SCAP sample card and prep SCIT (+)‐( S )‐citalopram SCX strong cation exchange SDCIT (+)‐( S )‐desmethylcitalopram SDDCIT (+)‐( S )‐didesmethylcitalopram SDF stromal cell‐derived factor SDME single‐drop microextraction SDS‐PAGE sodium dodecyl sulphate–polyacrylamide gel electrophoresis SDU solvent delivery unit sEGFR soluble epidermal growth factor receptor SELEX systematic evolution of ligands by exponential enrichment SF synovial fluid SHBG sex hormone‐binding globulin SIL stable isotope labeled SIL‐IS stable isotopically labeled internal standard SiNWA silicon nanowire array SISCAPA stable isotope standards and capture by anti‐peptide antibodies SLE supported liquid extraction SLM supported liquid membrane SM sphingomyelin SPDE solid‐phase dynamic extraction SPE solid‐phase extraction SPME solid‐phase microextraction SRM selected reaction monitoring SRM single reaction monitoring SSH steroid sex hormone SWCNT single‐wall carbon nanotube TAG triacylglycerol (1,3‐dipalmitoyl,2oleoyl‐glycerol) TAHS p‐N,N,N ‐trimethylammonioanilyl N′ ‐hydroxysuccinimidyl carbamate iodide TBS tris‐buffered saline TCA trichloroacetic acid TCAFMF thermally controlled aptamer‐functionalized microfluid TCEP tris(2‐carboxyethyl)phosphine TD thermal desorption TD toxicodynamic TDU thermal desorption unit TEA triethylamine TEHP tris(2‐ethylhexyl)phosphate TEPA tetraethylenepentamine TFA trifluoroacetic acid TFC turbulent flow chromatography TFME thin‐film microextraction Tg thyroglobulin THCA 11‐nor‐9‐carboxy‐Δ ⁹ ‐tetrahydrocannabinol THF tetrahydrofuran THU tetrahydrouridine Ti titanium TK toxicokinetic TK/TD toxicokinetic/toxicodynamic TLC thin layer chromatography TMMA tetramethyl minodronic acid TMS‐DAM trimethylsilyldiazomethane TNFα tumor necrosis factor alpha TRAIL tumor necrosis factor‐related apoptosis‐inducing ligand Tris tri(hydroxymethyl)aminomethane TSV‐DEME two‐step voltage dual electromembrane extraction TXB2 thromboxane B 2 Tyr tyrosine UC ultracentrifugation UF ultrafiltration UHPLC ultra‐high‐performance liquid chromatography ULOQ upper limit of quantitation UPLC ultra performance liquid chromatography UV ultraviolet VAMS volumetric absorptive microsampling VIDB vinylimidazole–divinylbenzene VPy vinylpyridine WAX weak anion exchange WBC white blood cell WCX weak cation exchange Zr zirconium β‐NGF beta‐nerve growth factor γ‐MPTS γ‐mercaptopropyltrimethoxysilane μ‐EME micro‐electromembrane extraction μ‐SPE micro‐SPE

Part I

Current Sample Preparation Techniques in LC‐MS Bioanalysis

1

Basic Sample Preparation Techniques in LC‐MS Bioanalysis: Protein Precipitation, Liquid–Liquid Extraction, and Solid‐Phase Extraction

Wenkui Li, Wenying Jian, and Yunlin Fu

1.1 Introduction

Bioanalysis is a subdiscipline of analytical chemistry for the determination of xenobiotics (chemically synthesized or naturally extracted drug candidates and genetically produced biological molecules and/or their metabolites or post‐translationally modified products) and biotics (macromolecules such as proteins and DNA, small‐molecule endogenous metabolites) in biological systems. The focus of bioanalysis in the human health industry is to provide a quantitative measurement of active drug and/or its metabolite(s) and/or biomarkers for the accurate assessment of pharmacokinetics, toxicokinetics, bioequivalence, bioavailability and/or exposure–response (e.g., pharmacokinetics/pharmacodynamics, toxicokinetics/toxicodynamics) relationships in support of drug discovery and development, and post‐approval therapeutic drug monitoring (Unger et al. 2013).

Many techniques have been employed in bioanalysis, including liquid chromatography with ultraviolet/fluorescence detection (LC‐UV/FL), liquid chromatography–mass spectrometry (LC‐MS), and gas chromatography–mass spectrometry (GC‐MS). Among these techniques, liquid chromatography–tandem mass spectrometry (LC‐MS/MS) has been the most widely used and most reliable tool due to its high sensitivity and specificity. In LC‐MS bioanalysis, a high‐performance liquid chromatography (HPLC) or ultra‐high‐performance liquid chromatography (UHPLC) system is employed to separate the analyte(s) of interest from other unwanted matrix components in sample extracts based on the specific interactions between the analyte(s) of interest and the analytical LC column. The LC eluent is then introduced to a mass spectrometer for molecule‐based separation via multiple reaction monitoring (MRM), also named as selected reaction monitoring (SRM). In MRM or SRM, the specific precursor ion(s) of analyte of interest is selected for collision‐induced dissociation (CID) or collision‐activated dissociation (CAD) to generate fragment ions that are also specific to the analyte(s) of interest, and usually the most abundant fragment ion is selected for MS detection. With the specificity and sensitivity provided by both the LC separation and the MRM of the MS system, LC‐MS/MS has become one of the most suitable tools for quantitative bioanalysis (Unger et al. 2013).

The common matrix that is subjected to LC‐MS bioanalysis includes various body fluids (e.g. plasma, serum, whole blood, saliva, tears, and urine) and organ tissues (e.g. kidneys, liver, lung, skin, and brain tissue). In general, these biological samples contain abundant various endogenous components like salts, small molecules, proteins, and lipids, or exogenous components such as formulation ingredients. In contrast, the analyte(s) of interest is often at very low concentration levels, typically in low ng ml−1 concentration range and even at pg ml−1 level for highly potent molecules. The presence of abundant (typically at μg ml−1 to mg ml−1 range) endogenous or exogenous components in the biological matrix, compounded by the very low concentration of the analyte(s) of interest, is definitely a challenge for bioanalytical scientists in the field in developing and validating a robust LC‐MS bioanalytical assay (Unger et al. 2013).

In order to ensure adequate sensitivity, selectivity, and reproducibility of the LC‐MS assay method for measuring analyte(s) of interest in biological samples, sample preparation, also known as sample pretreatment or sample cleanup, is a must step. Sample preparation in LC‐MS bioanalysis is considered a pre‐analytical separation process that involves selective isolation of analyte(s) of interest from the matrix, minimization or elimination of matrix components in the extracted samples, and, if necessary, enrichment of analyte(s) to ensure achievable assay sensitivity. An ideal sample preparation method should be able to reduce matrix effect to a minimal level while maintaining a reasonable and consistent extraction recovery (REC) (e.g. 80%). However, due to the many factors affecting matrix removal and analyte recovery, developing an optimal sample preparation procedure can be difficult, tedious, and labor‐intensive, which makes it one of the most significant parts in the development of a robust LC‐MS bioanalytical method. In this chapter, we describe the commonly used sample preparation techniques in LC‐MS bioanalysis, namely protein precipitation (PPT), liquid–liquid extraction (LLE), and solid‐phase extraction (SPE), with focus on the importance of understanding the physicochemical properties of the analyte(s) of interest that determines its extractability and the need of balancing REC and assay selectivity. There are a variety of techniques that are derived from these three basic sample preparation workflows as well as emerging methodologies that utilize advanced techniques for more selective and/or efficient extraction of analyte(s) of interest. They will be covered in detail in subsequent chapters of the book and therefore will not be discussed here. In addition, although dilution‐and‐shoot (DAS) of biological samples has been used in LC‐MS bioanalysis, it has been primarily used in the area of drug discovery for the analysis of samples generated from in vitro system rather than in vivo studies, and therefore is not covered in this chapter.

1.2 Physicochemical Properties of Drugs and Their Metabolites

The first and foremost important aspect of sample preparation in LC‐MS bioanalysis is to understand and utilize the relevant physicochemical properties of the analyte(s) of interest. These include but are not limited to hydrophilicity, lipophilicity, and protolytic properties (log P, log D, and pK a, etc.). Understanding these properties can help select suitable sample preparation techniques and the associated experimental conditions for the analyte(s) of interest.

1.2.1 Hydrophilicity vs. Lipophilicity of Analyte(s)

Hydrophilicity refers to the ability of a molecule to dissolve in water and other hydrophilic solvents. The interaction of a hydrophilic or polar molecule with water and other polar molecules is more thermodynamically favorable than its interactions with oil or other hydrophobic/lipophilic molecules. A hydrophilic molecule is typically charge‐polarized and capable of hydrogen bonding. In contrast, lipophilicity refers to the ability of a molecule to dissolve in fats, oils, lipids, and nonpolar solvents such as hexane or toluene. Such nonpolar solvents are themselves lipophilic. In the field of bioanalysis, lipophilicity, hydrophobicity, polarity and nonpolarity may have been used to describe the same tendency of a molecule and these terms are often used interchangeably. There are two key parameters to help understand the hydrophilicity vs. lipophilicity of the analyte(s) of interest:

The partition coefficient, P, is the concentration ratio of a molecule, specifically for unionized molecule, in two immiscible solvents at equilibrium. When one of the solvents is water and the other is a nonpolar one (e.g. n‐octanol), the log P value is the logarithm of the concentration ratio of the unionized form of the molecule between the two immiscible phases. Log P is considered a measure of lipophilicity or hydrophobicity of a given molecule, for which the higher the log P value, the more lipophilic or hydrophobic the molecule is.

where Corganic is the concentration of the neutral form of the molecule in the water‐immiscible solvent; Caqueous is the concentration of the neutral form of the molecule in the aqueous phase.

The distribution coefficient, D, is the ratio of the sum of the concentrations of both ionized and unionized forms of the molecule between the two immiscible phases, i.e. aqueous and organic phases. Since ionization of a molecule in aqueous phase is pH dependent, the log D value is also pH dependent. The aqueous phase is adjusted to certain pH for log D measurement.

where Corganic is the concentration of the molecule (ionized and neutral) in the water‐immiscible solvent, Caqueous (ionized) is the concentration of the ionized form of the molecule in the aqueous phase, and Caqueous (neutral) is the concentration of the neutral form of the molecule in the aqueous phase. Accordingly, in LLE and to some extent SPE, the REC of a given molecule under given conditions (pH, organic solvent, volume ratio between aqueous and organic) can be predicted (Liu and Aubry 2013).

1.2.2 Protolytic Properties of Analyte(s)

Protolytic refers to a reaction or process in solution that involves transfer of a proton from one molecule to another (one of the molecules usually belonging to the solvent). Many molecules show protolytic properties and are often present in both ionic/ionized (acidic [H+ donor] or basic [H+ acceptor]) and neutral forms in aqueous solution.

In chemistry, pH is the negative logarithm of the concentration of the hydrogen ion (pH = −log10(H+ ) and is used to specify the acidity or basicity of an aqueous solution. Solutions with a pH <7 are considered acidic while solutions with a pH >7 are considered basic. Pure water is neutral at pH 7 (25 °C). pK a is the negative base‐10 logarithm of the acid dissociation constant (K a) of a solution (pK a = −log10 K a) and is also defined as the pH where a molecule exists as 50% ionized and 50% unionized. pK a is a property of a given molecule that tells us how acidic or basic it is. The lower the pK a a molecule is, the stronger the acid it is. For example, the pK a value of acetic acid is 4.8, while the pK a value of a stronger acid, lactic acid is 3.8. The pK a value of a given molecule can be related to its charge state in solution under a given pH. In this regard, the pH of an aqueous solvent has a great impact on degree of ionization of a given analyte and the choice of the sample preparation method to be employed in LC‐MS bioanalysis. An acid analyte in an acid solution will generally not ionize. In contrast, it will ionize in a basic solution. Similarly, a basic analyte will generally not ionize in basic solution but will ionize in acidic solution. When pH is equal to pK a, 50% of the analytes are in ionized form and 50% in unionized (neutral) form. However, when pH is <K a , almost 100% of the acidic analytes are unionized but basic analytes are almost 100% ionized; when pH is >>pK a , almost 100% of the acidic analytes are ionized but basic analytes are almost 100% unionized.

Accordingly, in LLE and SPE with reverse‐phased (RP) stationary phase, the best REC can be obtained at a pH at which most of the analytes are not charged, i.e. neutral form. In contrast, in SPE with ion‐exchange stationary phase, the best REC can be obtained when the analyte molecules are all ionized (charged) to interact with the charged stationary phase (Liu and Aubry 2013).

1.3 Pre‐analytical Variables of Analyte(s) of Interest in Biological Matrix

LC‐MS bioanalysis is dealing with a variety of biological matrices such as plasma, serum, whole blood, urine, CSF, tissue homogenates, etc. The composition and complexity (e.g. pH, nature, and concentration of proteins, lipids, and salts) of these matrices is significantly different from one to the other. Even for the same matrix, it can be largely different from one subject to the other, depending on the age, gender, disease stage, medication, and other factors of the subject. Understanding the above can generally be helpful in developing an overall sample preparation strategy. However, each analyte is considered a unique entity in LC‐MS bioanalysis. In addition to understanding the physicochemical properties and others discussed in Section 1.2 and complexity of the matrix discussed above, understanding some analyte‐specific variables, including stability, possible nonspecific binding, protein binding and/or blood‐to‐plasma ratio, and red blood cell partition can be pivotal in defining and/or optimizing a specific sample preparation method for LC‐MS bioanalysis of the analyte(s) of interest.

1.3.1 Stability

Stability is an important pre‐analytical variable for quantitative LC‐MS bioanalysis of drug molecules and/or their metabolites in biological matrices. Instability of an analyte in any stage of the bioanalysis process, including sample collection, processing, storage, extraction, and LC‐MS analysis, can result in under‐ or over‐estimation of analyte exposure if an adequate preventive procedure is not in place (Li et al. 2011). Therefore, one should carefully examine the structural characteristics and the in vivo and/or in vitro biotransformation of any analyte of interest prior to developing a specific LC‐MS bioanalysis method. Previous internal and/or external (i.e. literature) experiences with structurally similar compounds can be very useful for a general understanding of the potential instability alert. The possible instability of some analyte(s) that contains biologically or chemically labile moieties, e.g. thiol, ester, or catechol, etc., can be readily predicted and estimated for effectiveness of specific precautions throughout the entire method development, validation, and sample analysis process. However, instability may not be readily predictable for many analyte(s) unless necessary stability assessment has been conducted using QC and/or incurred samples against freshly prepared calibration standards. For example, an unstable conjugated metabolite could convert to parent drug leading to overestimation of parent drug concentration, a phenomenon that can only be estimated using incurred samples when the putative conjugated metabolite is not available. Instability of an unstable analyte may occur at one or several points of the bioanalytical process, i.e. sample collection, processing, storage, extraction, and LC‐MS/MS analysis, etc. Some general strategies should be formulated with incorporation of one or more specific guidance (e.g. addition of enzyme inhibitors, pH modifiers, or antioxidants, etc.) in developing a robust LC‐MS/MS quantitative bioanalytical method for the unstable molecules. For more details on stability strategies one can refer to Strategies in quantitative LC‐MS/MS analysis of unstable small molecules in biological matrices (Li et al. 2011, 2013a).

1.3.2 Nonspecific Binding

Some matrix, e.g. urine or CSF, does not normally contain proteins and lipids that are present at ~8% in whole blood, plasma, or serum (Lentner 1981). The lack of protein and lipids in these samples can be associated with issues of nonspecific binding or container surface adsorption of drug molecules, especially those lipophilic/hydrophobic and those with high affinity to proteins (high protein binding), in LC‐MS bioanalysis (Li et al. 2010). The nonspecific binding or container surface adsorption is often evidenced by the unusually low REC of the analyte(s) of interest and/or nonlinearity of the calibration curves or highly variable QC sample results. Unfortunately, the issue is often overlooked during the early stage of bioanalytical method development, especially when assay sensitivity is not an issue and/or both the calibration standards and QC samples are prepared in the same fashion, i.e. freshly spiked, daily prepared, or pre‐pooled. In the latter cases, the unexpected low recovery of analyte would often be interpreted due to matrix effect or signal suppression because the problem, if any, may be masked by a similar degree of LC‐MS signal loss of the analyte for both the calibration standards and QC samples that are prepared in the same fashion. In some cases, the issue might not be realized until after many failed feasibility runs. Failure to promptly assess and adequately address this issue would result in underestimated urine/CSF drug concentrations. Details on how to assess and address the issues due to nonspecific binding or container surface adsorption can be found in some good research and review articles (Fu et al. 2011; Ji 2013; Ji et al. 2010; Li et al. 2010). Some detailed strategies can be found in Chapter 18 of this book.

1.3.3 Protein Binding

Depending on the affinity of drug molecules to plasma protein, a portion of the drug molecules may become bound to plasma proteins, with the remainder being unbound. Therefore, a given drug molecule generally exists in two forms in blood, plasma or serum: bound and unbound. If the protein binding is reversible, then a chemical equilibrium will exist between the bound and unbound states of the drug molecules, such that: Protein + drug ⇌ Protein–drug complex. Since albumin, a plasma protein, is alkalotic, acidic and neutral drug molecules will primarily bind to albumin. If albumin becomes saturated, then these molecules will bind to lipoprotein. In contrast, basic drug molecules will primarily bind to the acidic alpha‐1 acid glycoprotein in whole blood, plasma, or serum.

In general, a LC‐MS bioanalytical method is employed to measure the total analyte concentrations in the intended study sample matrix unless otherwise specified. Therefore, release of the analyte of interest from plasma proteins by interrupting the protein binding is considered a key step for good and consistent REC, regardless of choice of the sample preparation method and protein binding rate of the analyte. In this regard, the biological samples are generally treated with acid (e.g. acetic acid or formic acid), base (e.g. ammonium hydroxide), buffer, or organic to free the analyte from the plasma proteins prior to further processing (de Boer and Wieling 2013).

1.3.4 Blood‐to‐plasma Ratio and Red Blood Cell Partition

The blood to plasma ratio (Kb/p) of a drug is the ratio of its concentration in whole blood (containing both red blood cells, RBCs, and plasma) to the corresponding value in plasma, namely C B/C P. In contrast, the red blood cell partition coefficient (Ke/p) is the ratio of the drug concentration in the RBCs (i.e. not including plasma) to its concentration in plasma, namely C RBC/C P. In practice, when working with whole blood samples, the intended method of sample preparation should be capable of releasing the analyte from the RBCs (Brockman et al. 2007).

Temperature‐dependent re‐equilibration of drug molecules between RBCs and plasma is a known phenomenon (Dell 2004). Ideally, if plasma is the matrix of choice for bioanalysis, blood samples should be centrifuged to separate the plasma as soon as they are collected. Extended stay of the collected blood samples at 4 °C may be an issue if temperature‐dependent re‐equilibration occurs to the analyte(s) of interest. As a result of temperature‐dependent re‐equilibration, the analyte concentration in plasma obtained from blood that have been left on wet ice (~4 °C) for some time may no longer be the original one. For example, dehydronorketamine, a ketamine major metabolite, is stable in plasma under all normal conditions. Due to re‐equilibration of compound into blood cells at 4 °C with time, a significant decrease in its plasma concentration was observed over time. In contrast, no change was seen in the concentrations of the plasma samples prepared from the blood that has been stored at ambient temperature. This phenomenon can be overcome by centrifuging the blood immediately after collection, so that the actual in vivo plasma analyte concentration can be maintained and determined (Hijazi et al. 2001). In the case where the collected blood samples have somehow stayed on wet ice for a period of time, the blood samples can be incubated at 37 °C for ~30 minutes to resume the equilibrium between RBCs and plasma for the analyte of interest (to mimic in vivo condition) prior to centrifugation of the blood for plasma (unpublished data).

1.4 Most Commonly Used Sample Preparation Methods in LC‐MS Bioanalysis

The most commonly employed sample preparation methods in LC‐MS bioanalysis generally include PPT, LLE, and SPE. Many other contemporary sample preparation methods, such as online extraction and column switching, phospholipids depletion, salting‐out assisted LLE, supported liquid extraction (SLE), immunoextraction, microextraction, sample preparation via nanomaterials, molecularly imprinted polymers (MIP), aptamers or electromembranes, and stir‐bar sorptive extraction, etc. are captured in other chapters of this book.

1.4.1 Protein Precipitation (PPT)

The common biological matrix, such as blood, plasma, or serum contains ~8% (w/w) of proteins as mentioned earlier. Direct injection of these samples onto the LC‐MS system is generally not a good option for LC‐MS bioanalysis. This is because the proteins in these samples could readily precipitate as a result of contact with organic solvents and/or buffers in the mobile phase. As a result of PPT in the LC system, the performance of the LC column will rapidly deteriorate unless the majority of the protein in the samples is removed.

Proteins are large biological molecules composed of amino acids that are linked to each other by peptide bonds. Under normal physiological conditions, a soluble protein has one or more peptide chains in a folded conformation. Within this conformation, the majority of hydrophobic amino acid residues are toward inside while the charged or hydrophilic amino acid residues are toward outside. In the inside of the conformation, the peptide chains are folded together mainly through hydrophobic interactions between hydrophobic amino acid residues; other interactions such as hydrogen bonds, salt bridges, or disulfide bonds also contribute to the folded structure of proteins. On the outside of the conformation, the charged or polar surface residues interact with the biological environment, where water forms a solvation layer that surrounds the protein. The formation of the solvation layer weakens the ionic interactions between proteins and decreases the likelihood of aggregation (Li and Bartlett 2014).

PPT is a simple, quick, and convenient sample preparation technique in LC‐MS bioanalysis. In this process, a small volume of blood, plasma, serum, tissue homogenate, or other aqueous matrices is mixed with a certain volume of protein precipitant. When proteins in sample matrix/solution come in contact with precipitant, the conformation of the proteins is altered due to the interaction. This results in aggregation and precipitation of the proteins. As a result of conformation changes of the proteins, the analyte(s) of interest that are bound to the proteins are released into and stay in the solution. Upon centrifugation and/or filtration, the precipitated proteins are separated from the analyte(s) containing supernatant.

The equipment required for PPT is relatively simple. It includes low‐cost and disposable centrifuge tubes or 96‐well plates and a centrifuge. A bioanalytical scientist can find many published methods ready for use as the starting point. If method development and/or optimization are necessary, it is relatively simple and straightforward due to the simple nature of the technique. Briefly, optimization of a PPT method includes selection of precipitant and the amount of precipitant along with centrifugation and/or filtration.

A significant advantage of PPT is its high recovery as compared to other techniques, e.g. LLE and SPE. Since only proteins are hypothetically removed from the sample matrix by the method, small‐molecule analyte(s) should remain in the solution and this yields a theoretical recovery of 100%. Such an advantage of PPT has made it very popular in the bioanalytical community.

Common protein precipitants include: (i) water‐miscible organic solvents, (ii) acids, (iii) metal ions, or (iv) salts. Among these precipitants, the water‐miscible organic solvents and acids are the most common ones.

1.4.1.1 Water‐miscible Organic Solvents

Common water‐miscible organic solvents include acetonitrile, acetone, ethanol, and methanol and they are 100% miscible in aqueous solution. When these solvents are added to blood, plasma, serum, or tissue homogenate, they quickly displace water molecules of the solvation layer of the proteins in the sample matrix. When the solvation layer becomes thinner and thinner, proteins get closer and closer to each other via attractive electrostatic or dipole interactions, leading to the aggregation. The protein aggregates grow by diffusive additions of other protein molecules and eventually reach a critical size for precipitation, forming protein sediments or floccules in the mixture of samples and precipitant (Li and Bartlett 2014).

Acetonitrile, acetone, ethanol, and methanol are all good precipitant with PPT efficiency in the order of acetonitrile > acetone > ethanol > methanol. Among these organic solvents, the most commonly used one is acetonitrile. It is worth noting that PPT efficiency via an organic solvent also depends on the volume of the organic solvent added. With acetonitrile as an example, when 0.2 ml of acetonitrile is added to 1.0 ml of plasma, only 13.4% of proteins are removed. With the volume of acetonitrile increased to 0.4, 0.6, 0.8, 1.0, 1.5, 2.0, 3.0, and 4.0 ml, the percentage (%) of protein removed is increased, respectively, to 14.8, 45.8, 88.1, 97.2, 99.4, 99.7, 99.8, and 99.8% (Blanchard 1981). It is apparent that by adding 2.0 ml of acetonitrile to 1.0 ml of plasma for PPT, the resulting supernatant is basically protein‐free.

Although pure organic solvents can perfectly fit the purpose of PPT, it is a common practice to add a small volume of acids (e.g. formic acid, acetic acid) (Li et al. 2014) or bases (e.g. ammonium hydroxide) (Li et al. 2007b) to the sample matrix, followed by proper mixing to interrupt protein binding and/or change the charge states of the analyte by changing the pH of the sample matrix. The acids or bases can also be added to the organic solvent to prepare a PPT solution. The addition of a small volume of acids or bases can help improve REC. On the other hand, internal standards (IS) can be added to the organic protein precipitants. This allows for the addition of the IS at the same time as for protein precipitant. This practice not only improves the throughput by combining two steps together but also enhances the precision of the assay by pipetting a relatively large volume of IS working solution instead of spiking a relatively small volume. In addition, for certain analytes with high protein binding, it is recommended to add IS in aqueous solution to plasma sample first. This allows for additional time of equilibrium before addition of an organic precipitant. By doing so, the IS may better mimic the extraction of the analyte and therefore improve the assay accuracy and reproducibility (unpublished data).

As the strongest precipitant, use of acetonitrile can lead to the formation of very solid protein precipitates of large particle sizes. Being relatively weaker protein precipitants, the use of ethanol or methanol can lead to the formation of looser flocculent precipitates. However, there is a concern that using 100% acetonitrile as the protein precipitant may not give the highest recovery. Due to the very quick PPT using acetonitrile, some analyte(s) with high plasma protein binding might coprecipitate with plasma proteins if the analyte(s) are not disassociated with the proteins during the process of protein precipitation. In a study with no acids or base added to the sample matrix prior to PPT using acetonitrile, an increasing percentage of methanol (10, 20, 30, 40, 50, 60, and 70%) was added to acetonitrile for PPT prior to direct analysis via hydrophilic chromatography (HILIC)‐MS/MS for atenolol in human plasma. The analysis of the samples showed that use of acetonitrile containing 10% methanol yielded a sample extract with LC‐MS signal higher than that obtained using acetonitrile alone (Li et al. 2005).

1.4.1.2 Acids

The commonly used acids for PPT are trichloroacetic acid (TCA) (5–15%, TCA) and perchloric acid (PCA) (6%). Both reagents are highly efficient in precipitating proteins in the sample matrix (Hee et al. 2017). It is understood that protein denaturing is a key in PPT by TCA and PCA. Upon addition of the acids, the pH of the solution/sample matrix is greatly lowered and the protein conformation is drastically altered, resulting in the aggregation of proteins. Addition of 0.2 ml of TCA (10%) to 1 ml of the matrix is capable of removing >99% of the proteins in the plasma samples. One advantage of using acids for PPT is that the resulting supernatant (of filtrate) after centrifugation is still highly aqueous and the supernatant can be directly injected onto the LC‐MS/MS system. The drawback of this approach is that the resulting supernatant is strongly acidic with pH < 2.0, which might be a concern of analyte instability in the sample extract (Li and Bartlett 2014; Pedersen‐Bjergaard et al. 2015).

Other than the composition and pH of the organic protein precipitants discussed above, temperature is another factor that should be considered for efficient PPT. Ice‐cold organic solvents are generally more efficient in precipitating proteins in the sample matrix. This approach is particularly useful when dealing with unstable analyte(s) (Li et al. 2011).

In the PPT process, upon addition of protein precipitant (discussed above), IS (in working solution or prepared in protein precipitant solution), and pH modifier (can be added to the protein precipitant), adequate vortex‐mixing has to be applied to the sample mixture. Most proteins in the sample mixture are expected to precipitate. The next step is a physical process to separate precipitated proteins from the supernatant. Usually 5–10 minutes of centrifugation at ~1500 × g is sufficient to settle the protein precipitates. The higher centrifugation force and the longer centrifugation time are applied, the more solid the protein pellet will be formed in the bottom of the sample preparation device (individual tubes or 96‐well plate). This makes it easier to transfer the supernatant without disturbing the pellet.

Since at least two volumes of organic solvents are commonly employed in PPT and the resulting supernatant generally has a high organic composition, direct LC‐MS/MS analysis of the supernatant using reversed‐phase columns may not be feasible. If assay sensitivity is permitted, dilution of the supernatant with aqueous solution (e.g. water) should be applied to lower the concentrations of organic solvents in the final sample extract prior to LC‐MS/MS analysis. Otherwise, the resulting supernatant should be subjected to evaporation step, at which the organic solvents in the supernatant are evaporated under a stream of nitrogen and resulting sample residues are reconstituted in aqueous solutions (with proper pH, organic and/or acid concentrations) or starting mobile phase. The volume of the reconstitution solution used should be adjusted to ensure suitable assay sensitivity on the intended LC‐MS/MS platform. In contrast, depending on the polarity and other physicochemical properties of the analyte(s) of interest, the resulting supernatant obtained from PPT may be directly subjected to analysis on HILIC‐MS/MS that employs mobile phases with high organic composition. This high organic composition is, in general, compatible with sample extract (supernatant) (Jian et al. 2010b; Li et al. 2005).

A representative protocol of PPT with an organic solvent is shown below.

A representative example of PPT protocol (for the analysis of a steroid 11β‐hydroxylase inhibitor, in human plasma by Li et al. 2014).

A 0.100 ml aliquot of blank plasma (for matrix blanks, zero samples [blank matrix mixed with the IS only] or fresh calibration standards), QC samples, or unknown samples was added to a 2‐ml, 96‐well plate.

To the standard samples, 0.100 ml of the respective standard working solution (in 50% aqueous methanol, v/v) was added.

To the matrix blanks, zero samples, QC samples, and unknown samples, a 0.100 ml aliquot of 50% methanol in water (v/v) was added.

This was followed by brief vortex‐mixing.

A 25 μl aliquot of the IS working solution in 50% methanol in water (v/v) was added to each well except for the matrix blank, to which a 25 μl aliquot of 50% methanol in water (v/v) was added.

After adding 25 μl of 40% formic acid in water (v/v) to each well, the plate was vortex‐mixed briefly.

To each well, 0.500 ml of acetonitrile/ethanol/acetic acid (90/10/0.1, v/v/v) was added and the plate was covered and vortex‐mixed for ~15 minutes on a pulse vortex‐mixer with a motor speed setting of about 50 units.

The sample plate was centrifuged at ~3500 rpm (~2000 × g) for ~10 minutes at ~10 °C.

The supernatants (0.500 ml) were transferred via a TomTec Quadra 96 system (Hamden, CT, USA) to the corresponding wells in a 1‐ml 96‐well plate, evaporated to dryness under a stream of nitrogen at approximately 45 °C and the sample residues reconstituted with 100 μl of reconstitution solution (30% methanol in water, v/v).

After brief vortex‐mixing and centrifugation at ~3500 rpm (~2000 × g) for approximately five minutes remove potential un‐dissolved contents in the sample, a 10 μl aliquot was injected onto the LC‐MS/MS system.

Conventional PPT involves multisteps of manipulation, including centrifugation and transfer of the resulting supernatant. The overall throughput is low. Therefore, it is not ideal in support of analysis of a large number of study samples. One of the recent approaches to overcoming such a disadvantage is the use of membrane‐based PPT filter plates, such as Unifilter® from Whatman, Strata® Impact™ from Phenomenex, Isolute® PPT+ from Biotage, and Sirocco™ from Waters. By using PPT plates, separation of the supernatant can be carried out by filtration without the need of supernatant transfer.

In practice, the 96‐well PPT plate is placed on top of a 96‐well collection plate. The PPT plate contains filters in the bottom of each well with a typical pore size of 0.2 μm. Upon centrifugation or application of vacuum, the supernatant from PPT flows through the filter into the collection plate, while the pellets are retained on the filter. By this approach, significant reduction in the sample preparation time (four‐fold, from 4–5 hours to 1–1.5 hours) was reported as compared with the manual PPT method (96 samples each) (Kocan et al. 2006). Furthermore, the use of PPT plates yields cleaner sample extracts with a higher solvent recovery.

Issues related to solvent leaking with PPT plates have been reported. Among the PPT solvents tested, acetonitrile showed higher leaking than methanol. It has been also suggested that the precipitating solvent, filter material, pore size, centrifugation speed or vacuum strength, nonspecific binding of analyte to the plate, and matrix effects (ion suppression) should be concurrently considered when choosing a PPT plate. Table 1.1 summarizes commercially available PPT plates with respective technical details (Kole et al. 2011).

Table 1.1 Commercially available protein precipitation plate/tubes.

Source: Modified from Kole et al. (2011). Reproduced with permission of John Wiley & Sons.

Matrix effect is considered a major issue related to PPT. This is because, upon PPT, all other matrix components, including phospholipids, are present in the sample extract. Phospholipids are an important class of biological compounds containing one or more phosphate groups. Their molecular structures have two major functional group regions, i.e. (i) a polar head group which includes an ionizable organic phosphate moiety as well as other polar groups of various types and (ii) one or two long‐chain fatty acid ester groups which are hydrophobic. The highly ionic nature of phospholipids makes them responsible for influencing the ionization of the analyte(s) of interest and desolvation of the LC effluent droplets in the electrospray MS source, causing significant matrix effects.

Noticeably, some of the recently developed PPT plates are packed with materials that specifically retain phospholipids and, therefore, have been increasingly employed to remove the abundant phospholipids along with proteins in LC‐MS bioanalysis. These plates include but are not limited to Captiva™ ND Lipids from Agilent, Ostro™ from Waters, Phree™ from Phenomenex, and HybridSPE® from Sigma. By using these plates, the resulting sample extracts are not only free of proteins, but also free of or with significantly reduced phospholipids. The sample extracts are readily available for the subsequent LC‐MS analysis. In the case of using a HybridSPE PPT plate, once PPT occurs after the addition of acetonitrile (containing 1% formic acid), the sample extract passes through the HybridSPE packed bed. The packed bed consists of proprietary zirconia‐coated silica particles. The zirconia sites exhibit Lewis acid (electron acceptor) properties that interact strongly with Lewis bases (electron donor). As mentioned above, phospholipids consist of zwitterionic phosphonate moieties (the polar head group) and a large hydrophobic tail (two fatty acyl groups that are hydrophobic). The phosphate group(s) on the phospholipids acts as a very strong Lewis base that interact strongly with zirconia atoms functionalized on the particle surface. Formic acid or other acids is a critical modifier in PPT via phospholipid removal PPT plate approach to improve the recovery of analytes of interest, particularly those acidic. The acid plays a critical role in preventing analyte retention without affecting phospholipids’ retention on the packed bed (Ahmad et al. 2012).

Each of the commercially available PPT plates comes with standardized procedures recommended by the providers. The procedures include direct PPT on the plates (Pucci et al. 2009) or conventional PPT followed by loading the resulting supernatant onto the PPT plates (Asimakopoulos and Thomaidis 2015; Jiang et al. 2011; Zeng et al. 2010).

A representative example of PPT protocol using Sirocco PPT plate.

A 200 μl aliquot of acetonitrile was added to each well of a Sirocco PPT plate with a 1‐ml 96‐well collection plate underneath.

A 25 μl aliquot of IS working solution in acetonitrile : water (50/50, v/v) was added to each well except for the blanks, to which a 25 μl aliquot of acetonitrile : water (50/50, v/v) was added.

A 25 μl aliquot of blank, zero, standard, or QC was added to the assigned well. The plate was covered and vortex‐mixed for one minute on a pulse‐vortex‐mixer and then centrifuged for five minutes at ambient temperature.

The filtrate in the collection plate was evaporated to dryness at a temperature of 45 °C under a flow of nitrogen. A 200 μl aliquot of reconstitution solution consisting of acetonitrile in water (10/90, v/v) with 0.1% formic acid was added to each well and the plate was vortex‐mixed for five minutes.

A 10 μl aliquot of the reconstituted extract was injected onto the LC‐MS/MS system.

1.4.2 Liquid–Liquid Extraction (LLE)

LLE is another common sample preparation technique that has been widely used in LC‐MS bioanalysis. The method involves the extraction of the analyte(s) of interest or unwanted interference components from one liquid phase (e.g. biological samples) to another immiscible liquid phase (e.g. organic solvent), resulting in sample clean‐up.

In LLE, biological samples (plasma, serum, whole blood, and urine or tissue homogenate) are commonly mixed with additives (buffer, acids, or bases) to ensure efficient extraction of the target molecules. This is followed by addition of IS working solution and an organic solvent (extraction solvent), which is immiscible with water. Then the two‐immiscible phase mixture in tubes/wells is shaken or vortex‐mixed for a certain period of time to mix the sample and the organic solvent, during which the target molecules are transferred from the aqueous phase to the organic phase or vice versa. This is followed by centrifugation for phase separation. After centrifugation, the phase containing the target molecules can be collected for further processing and analysis (Liu and Aubry 2013).

1.4.2.1 Mechanism of LLE and Extraction Recovery

The mechanism of LLE can be explained using a simple phrase like dissolves like (Li and Bartlett 2014). A solute can be dissolved best in a solvent that has a similar polarity to itself. Nonpolar compounds have higher solubility in organic solvents than in water. In contrast, ionic or polar compounds have higher solubility in aqueous solutions than in organic solvents. When two immiscible solvents are present in one system, the solute currently being dissolved in the solvent with less solubility will diffuse across the liquid–liquid interphase of the two immiscible solvents to enter the one in which the solute solubility is higher. When analyte molecules are extracted from the aqueous phase into the immiscible organic phase or vice versa in LLE, interactions take place between the analyte molecules and the solvent molecules. The predominant interactions in LLE are the following (Li and Bartlett 2014; Pedersen‐Bjergaard et al. 2015):

Hydrophobic interactions: The interactions between the nonpolar analyte(s) and nonpolar organic solvent or solvent mixture in a polar (aqueous, usually water) solvent. Hydrophobic interactions are nonpolar attractive interactions between hydrocarbon moieties and related nonpolar molecular elements.

Dispersion interactions: The interactions between a relatively nonpolar electron‐rich molecule and a polar (or charged) molecule. It is generally a weak attractive interaction. The polar molecules in the system cause electron density change in the nonpolar molecules by making them slightly polar. In other words, the polar molecules induce dipole moment in the nonpolar but electron‐rich molecules.

Dipole interactions: The interactions between two molecules with permanent dipole moment due to attractive electrostatic forces, by which the positive end of one molecule is attracted by the negative end of another molecule. Both components have high permanent dipole moments.

Hydrogen bonding interactions: Interactions between the hydrogen atom that is part of a polar bond (hydrogen‐bonding donor) and an electronegative atom with a lone pair of electrons such as O and N (hydrogen‐bonding acceptor). Hydrogen bonding interaction is another type of dipole interaction.

As discussed in Section 1.2.1, partition ratio (P) of a given analyte in two immiscible phases can be expressed as follows: P = [C Organic]/[C Aqueous], where [C Organic] is the concentration of the un‐ionized analyte in the organic phase (e.g., n‐octanol) while [C Aqueous] is the concentration of the un‐ionized analyte in the aqueous phase (e.g., water) at equilibrium.

The P value is constant for a given analyte under given conditions of the two‐phase (aqueous vs. immiscible organic solvent) system. In LLE, equilibrium of the analyte(s) of interest between the two immiscible phases is normally established at the end of the extraction. The higher the P value of the analyte(s) of interest, the more efficiently it is extracted from the aqueous samples (e.g. plasma) into the organic solvent.

For neutral analyte(s), the REC can be maximized by optimizing the type of extraction solvent. The pH value of the sample matrix does not affect the extraction efficiency. However, for acidic or basic analyte(s), the situation is different. Both acidic and basic molecules tend to dissociate in aqueous