HPLC for Pharmaceutical Scientists 2007 (Part 13)

Chia sẻ: Big Big | Ngày: | Loại File: PDF | Số trang:36

Thêm vào BST

Báo xấu

115
lượt xem 12
download

Download Vui lòng tải xuống để xem tài liệu đầy đủ

Recent advances in mass spectrometry have rendered it an attractive and versatile tool in industrial and academic research laboratories. As a part of this rapid growth, a considerable body of literature has been devoted to the application of mass spectrometry in clinical studies. In concert with separation techniques such as liquid chromatography, mass spectrometry allows the rapid characterization and quantitative determination of a large array of molecules in complex mixtures. Herein, we present an overview of the above techniques accompanied with several examples of the use of liquid chromatography– tandem mass spectrometry in pharmacokinetics/drug metabolism assessment during drug...

Chủ đề:

Bình luận(0) Đăng nhập để gửi bình luận!

Lưu

Nội dung Text: HPLC for Pharmaceutical Scientists 2007 (Part 13)

13 THE ROLE OF LIQUID CHROMATOGRAPHY–MASS SPECTROMETRY IN PHARMACOKINETICS AND DRUG METABOLISM Ray Bakhtiar, Tapan K. Majumdar, and Francis L. S. Tse 13.1 INTRODUCTION Recent advances in mass spectrometry have rendered it an attractive and ver- satile tool in industrial and academic research laboratories. As a part of this rapid growth, a considerable body of literature has been devoted to the appli- cation of mass spectrometry in clinical studies. In concert with separation tech- niques such as liquid chromatography, mass spectrometry allows the rapid characterization and quantitative determination of a large array of molecules in complex mixtures. Herein, we present an overview of the above techniques accompanied with several examples of the use of liquid chromatography– tandem mass spectrometry in pharmacokinetics/drug metabolism assessment during drug development. Since the evolution of pharmaceutical research [1, 2], the stages of drug dis- covery and development have followed three predominant patterns: (i) the systematic and methodical approach by chemists to rationally design and syn- thesize a molecule to target a speciﬁc molecular system (e.g., ion channels, receptors, enzymes, DNA); (ii) the isolation and puriﬁcation of the active ingredients of medicinal plants or microorganisms to screen their spectrum of HPLC for Pharmaceutical Scientists, Edited by Yuri Kazakevich and Rosario LoBrutto Copyright © 2007 by John Wiley & Sons, Inc. 605
606 THE ROLE OF LC–MS IN PK AND DRUG METABOLISM activity using in vitro models; or (iii) the serendipitous discovery of a com- pound with a novel pharmacological action (e.g., the accidental discovery of antidepressants). Today, one of the increasingly popular and complementary approaches for drug discovery in the pharmaceutical industry is to perform massive parallel synthesis in solution or on a solid support. In addition, with the advent of functional genomics and proteomics, cell-based assays, and mol- ecular biology, a multitude of therapeutic targets have been validated [3]. With an increasing number of potential molecular targets identiﬁed through the science of functional proteomics and genomics, diverse libraries of new chemical entities (NCEs) have to be generated and evaluated. Consequently, the rapid growth of combinatorial libraries has posed a need for faster, accurate, and sensitive analytical techniques capable of large-scale high- throughput screening (HTS). Although in vitro assays do not necessarily reﬂect the complexity of the in vivo interactions, the speed and simplicity of the former have rendered them an integral part of the screening process. In recent years, the in silico and experimental modeling of pharmacokinetic/ pharmacodynamic (PK/PD) relationship have become increasingly popular [4, 5]. The integration of PK (i.e., drug dose and biological ﬂuid concentration) and PD (i.e., pharmacologic effect) provides a key determinant in under- standing the dosing regimen and therapeutic effect of a potential drug com- pound. To this end, analytical assays also play a pivotal role in deﬁning the PK/PD relation of NCEs. In many cases, both the drug concentration and PD biomarkers (vide infra) can be directly measured in peripheral ﬂuids using speciﬁc analytical techniques. Furthermore, samples generated from large-scale clinical trials along with the ambitious development timelines to get safe and efﬁcacious drugs to market warrant the use of HT bioanalysis. Numerous improvements in speed, sensitivity, and accuracy, augmented with innovations in automation in con- junction with mass spectrometry (MS) detection, have allowed for versatile and multifaceted platforms [6–8]. 13.2 IONIZATION PROCESSES Mass spectrometry (MS) is playing an increasingly visible role in the molecu- lar characterization of combinatorial libraries, natural products, drug meta- bolism and pharmacokinetics, toxicology and forensic investigations, and proteomics. Toward this end, electrospray ionization (ESI), atmospheric pres- sure chemical ionization (APCI), and atmospheric pressure photo-ionization (APPI) have proven valuable for both qualitative and quantitative screening of small molecules (e.g., pharmaceutical products) [9–14]. The utility of ESI (Figure 13-1) lies in its ability to generate ions directly from the solution phase into the gas phase. The ions are produced by appli- cation of a strong electric ﬁeld to a very ﬁne spray of the solution containing the analyte. The electric ﬁeld creates highly charged droplets whose subse-
IONIZATION PROCESSES 607 Figure 13-1. A simpliﬁed schematic of the ESI process. (Courtesy of Dr. P. Tiller.) Figure 13-2. A simpliﬁed schematic of the APCI process. (Courtesy of Dr. P. Tiller.) quent vaporization (or desolvation) results in the production of gaseous ions. The fact that ions are formed from solution has established the technique as a convenient mass detector for liquid chromatography (LC/MS) and for automated sample analysis. In addition, ESI-MS offers many tangible beneﬁts over other mass spectrometric methods including the ability to qualitatively analyze low-molecular-weight compounds, inherent soft-ionization, excellent quantitation and reproducibility, high sensitivity, and its amenability to automation. Analogous to the ESI interface, APCI (Figure 13-2), also referred to as the heated nebulizer (HN), induces little or no fragmentation to the analyte.
608 THE ROLE OF LC–MS IN PK AND DRUG METABOLISM Therefore, the APCI spectrum also tends to be simpler in interpretation than the traditional electron ionization (EI), which results in extensive fragmenta- tion of the precursor ion. As a result, APCI and ESI are referred to as “soft- ionizations,” while EI is considered a “hard-ionization” technique. Generally, volatile and thermally stable compounds can be subjected to LC/APCI/MS analysis. In quantitative analysis, APCI provides a greater (i.e., in terms of lin- earity) dynamic range than ESI and it is considered rugged, easy to operate, and relatively tolerant of higher buffer concentrations (i.e., fewer matrix effects). In ESI, at about 10–5 M and higher, the ion signal becomes ﬁxed and independent of sample concentration (plateauing effect) and may exhibit non- linearity at higher concentrations. In contrast, APCI can offer a wider linear dynamic range. For example, in our laboratory (data not shown) we have rou- tinely developed reversed-phase LC/APCI/MS/MS assays ranging from 1.0 ng/mL to 10,000 ng/mL with a correlation coefﬁcient of >0.996. Further- more, APCI can accommodate ﬂow rates of up to 2.0 mL/min and is effective in the analysis of medium- and low-polarity compounds [12]. In qualitative drug metabolism studies, a combination of APCI and ESI experiments can prove valuable in distinguishing certain oxidative biotransformations (e.g., N-oxidation versus hydroxylation) [15, 16]. In contrast to ESI, APCI is not suited for the analysis of biopolymers, proteins, peptides, and thermally labile species. In the APCI process, electrons originating from a corona discharge needle ionize the analyte via a series of gas-phase ion-molecule reactions. For example, in the positive-ion mode, the energetic electrons start a sequence of reactions with the nebulizing gas (typically nitrogen), giving rise to nitrogen molecular ions. Using APCI, depending on the composition of the HPLC mobile phase, ions such as [H2O + H]+, [CH3OH + H]+, [NH3 + H]+, and/or [CH3CN + H]+ are formed via series of ion-molecule reactions with the nitro- gen molecular ions. Subsequently, additional ionization is initiated by exother- mic proton transfers from the protonated solvent ions to the neutral analyte molecules yielding [analyte + H]+, [analyte + CH3OH + H]+, [analyte + NH3 + H]+ ions, and so on. In general, metal adduct ions are observed less commonly in APCI as opposed to ESI, where they are more prevalent. Greater sensitiv- ity is attained if the solvent is polar and contains ions through the addition of an electrolyte. The desolvation process is then further enhanced by the heating element within the APCI assembly, which is maintained at 300–550°C. One of the drawbacks of APCI is its lack of compatibility with low eluent ﬂow rates. The stability of the ionization response may be poor at low rates (i.e., less than 50 µL/min). In contrast, ESI is compatible with miniaturized columns and amenable to sample-limited scenarios such as biochemical and biotechnological applica- tions. ESI can be considered a ﬂow-sensitive technique. The dimension of the primary droplets is dependent on the ﬂow rate. Therefore, by using columns with a smaller internal diameter (i.d.) and consequently lower ﬂow rates, the concentration of the analytes in the spray solution can vary and it can be
IONIZATION PROCESSES 609 considered a concentration-dependent ionization process. It is concentration- dependent in the sense that the surface charge density of the droplets in the gas phase is higher due to more effective desolvation of the droplets since lower ﬂow rates are used. The use of solvent-buffer post-column addition also allows optimization for improved analyte ion current response. Increasing the ﬂow rate increases droplet size, which decreases the yield of gas-phase ions from the charged droplets. Recently, atmospheric pressure photo-ionization (APPI) [17–19] was intro- duced as a complementary ionization technique to ESI and APCI. APPI (Figure 13-3) is now commercially available by several MS vendors such as Agilent Technologies, Applied Biosystems (Sciex), Waters (Micromass), and Thermo Electron (Finnigan) Corporations. This technique can be used to ionize an analyte that otherwise is not easily ionizable using either APCI or ESI. In APPI, to increase ionization efﬁciency, a high-intensity UV radiation source is used (i.e., a 10-eV krypton discharge lamp) in a direct or an indirect mode. In the direct mode, often a molecular ion is generated by irradiation; while in the indirect mode, a dopant is used in conjunction with the analysis. A photoionizable dopant such as acetone or toluene is employed to mediate (as dopant photo-ions) the production of ions by proton or electron transfer. The dopant is introduced to the APPI ionization chamber by a separate pump at an optimized steady ﬂow rate during analysis (e.g., 10–15% of the mobile phase ﬂow rate, post-column). A number of excellent articles have recently been published on the applicability of APPI for the analysis of small mole- cules [17–19]. UV Lamp Curtain Plate Nebulizer Gas (Gas1) Heater Quartz Tube Orifice LC Effluent Primary Dopant Ionization PhotoSpraytm Region Source Block Curtain Gas Figure 13-3. A simpliﬁed schematic of the APPI process. (Courtesy of Sciex/Applied Biosystems Corporation.)
610 THE ROLE OF LC–MS IN PK AND DRUG METABOLISM 13.3 TANDEM-MASS SPECTROMETRY (MS/MS) For purposes of quantitative analysis, selected ion monitoring (SIM) and selected reaction monitoring (SRM) are two commonly utilized approaches. The latter is also referred to as multiple reaction monitoring (MRM). In both modes, considerable structural information is lost; nonetheless, these tech- niques are extremely powerful for target compound quantiﬁcation in biolog- ical matrices, if the compound of interest is known. In the SIM mode, the MS is tuned to a particular m/z window (preferably at unit resolution), which corresponds to the ion of interest (i.e., [M + H]+, or a stable adduct such as [M + X]+, where X = Na, K, NH4, etc.). SIM may require a more elaborate chromatographic separation in order to minimize interfer- ence from endogenous species. However, in the SRM approach, higher selec- tivity and sensitivity are realized. Thus, shorter chromatographic runs (faster injection cycles) and limited sample pretreatment could be tolerated without signiﬁcant loss in sensitivity. In addition, due to lack of MS/MS capability, SIM has been more commonly performed on single quadrupole MS, while SRM has been broadly adapted on triple quadrupole (Figure 13-4) and ion-trap mass spectrometers. The increase in sensitivity and selectivity of SRM stem from the ion-chromatogram (i.e., LC-MS/MS) obtained by speciﬁc precursor-to-product ion transition for an analyte of interest (Figure 13-4). Conversely, in an SIM mode, the relative background noise due to the presence of other isobaric species (i.e., ions with a same m/z as the analyte of interest) can result in a lower signal-to-noise ratio for the analyte. Due to the widespread acceptance of SRM in quantitative analysis, the remaining part of this section focuses on a description of tandem- mass spectrometry (MS/MS), which is utilized in SRM (or MRM) experiments. Tandem-mass spectrometry or collision-induced dissociation (CID) is one of the most widely used techniques for probing the structure of ions in the gas phase [20]. To this end, ease of application to various instrumental types, along with its experimental simplicity, account for the wide popularity of CID. In a typical CID experiment, a beam of ions with a speciﬁc m/z (denoted as the precursor or parent ion) is selected and collided with a neutral and nonreac- tive gas-phase target (e.g., argon, xenon, helium, nitrogen). These collisions result in subsequent fragmentation and product ions that are a direct conse- quence of dissociation of the precursor ion. Generally, the resulting fragmen- tation pattern is unique to a particular ion structure. The various CID techniques can be subdivided into categories based on the translational or collision energy of the precursor ion prior to collision with the target gas. The two main categories include low-energy CID, in the range of 1–300 eV (i.e., used in triple quadrupole and ion-trap instruments), and high-energy CID at approximately 1–25 KeV (i.e., used in guided-ion beam or sector instruments). Currently, one of the most common approaches is to perform MS/MS exper- iments on a triple quadrupole instrument. Tandem-MS experiments have been particularly popular for the qualitative and quantitative analysis of small mol-
SAMPLE PREPARATION USING AN OFF-LINE APPROACH 611 Figure 13-4. Representative MRM scans (plasma extract of a proprietary compound) using an API 5000 triple quadrupole unit (Sciex). Each panel contains a distinct MRM transition for the same compound: m/z 1021.6 → 1003.5 (left panel) and m/z 1021.6 → 971.5 (right panel). Signal-to-noise ratio is designated as S/N. Experimental conditions: ESI, positive ion mode, protein precipitation was used for sample preparation, injec- tion volume was 10 µL, the column was a C18, and dimension was 20 × 2.1 mm, using a linear gradient elution: 0 min (20% B)–6 min (90% B)–8 min (90% B), where B was 0.2% formic acid in acetonitrile and A was 0.2% formic acid in water; separation was performed at room temperature. ecules such as pharmaceutical products in biological ﬂuids [21–23]. In recent years the sensitivity and selectivity of MS/MS analysis of xenobiotics have been put to use in toxicokinetics, pharmacokinetics, metabolic, formulation, and early drug discovery studies. 13.4 SAMPLE PREPARATION USING AN OFF-LINE APPROACH One of the critical steps in qualitative and quantitative analysis is the sample preparation procedure. Sample preparation step can affect speciﬁcity, sensi- tivity, accuracy, precision, and throughput of a bioanalytical procedure. In addi- tion to development and optimization of the chemistry involved in sample processing, the use of semiautomated or fully automated protocols has been
612 THE ROLE OF LC–MS IN PK AND DRUG METABOLISM Figure 13-5. Photograph of a Biomek 2000 setup for semiautomated PPT, LLE, or SPE process in the authors’ laboratory (also see www.beckmancoulter.com and reference 99). implemented in recent years [24, 25]. The popularity of off-line sample pro- cessing in batch-mode has dramatically improved the throughput of this rate- limiting step. Generally, there are three commonly used approaches for off-line sample processing: SPE (solid-phase extraction), LLE (liquid–liquid extraction), and protein precipitation (PPT). These three methods have been successfully used in conjunction with robotics for achieving an increase in sample preparation throughput. For example, Figure 13-5 is the photograph of a Beckman’s Biomek 2000 (other models such as Biomek 3000 and Biomek FX are also applicable) for semi-automated sample preparation that can accommodate SPE, LLE, and PPT procedures. This scheme has been established for use with SPE, LLE, or PPT in a 96-well plate format to analyze pharmaceutical prod- ucts in biological matrices (e.g., whole blood, plasma, serum, and cerebral spinal ﬂuid (CSF)) in our laboratories (unpublished data). 13.4.1 SPE In the 96-well SPE format, similar to the traditional manual procedure, issues such as the nature of the bonded-phase (e.g., ion exchange, C2, C8, C18, cyano, phenyl, polymeric, strong or weak cation exchange, strong or weak anion exchange, mixed phases, etc.), solvent strength (for conditioning/washing of the phases, target analyte elution), and chemical characteristics (e.g., solubil- ity, presence of the key functional groups) of the analyte(s) need to be addressed. A general scheme for initial development of an SPE method is out-
SAMPLE PREPARATION USING AN OFF-LINE APPROACH 613 lined below. Depending on the structure of the compound (hydrophobicity and ionizable functionalities), speciﬁc steps to optimize sample recovery are needed. • Condition sample for optimum retention • Condition SPE bed with methanol • Equilibrate SPE bed with water • Load sample onto SPE bed (a) Cation exchange: wash with 2% formic acid [low pH (3)] (b) Anion exchange: wash with 50 mM NaOAc buffer [high pH (8–10)] • Wash bed with 5% methanol • Elute retained materials with an organic solvent (i.e., CH3OH, CH3CN, isopropyl alcohol, or a combination thereof) (a) Cation exchange: add 5% NH4OH to eluent (b) Anion exchange: add 2% formic acid to eluent Some of the most commonly utilized robotic modules for the 96-well SPE procedure are Tomtec Quadra (Tomtec, Hamden, CT, USA), Packard Multi- Probe (Packard Instruments, Meriden, CT, USA), Biomek (Beckman–Coulter, Fullerton, CA), and Tecan (Durham, NC, USA) units. For example, we have successfully and routinely adopted the Tomtec Quadra technology in the development and validation of several off-line SPE assays in whole blood, plasma, and urine followed by MS detection. The Packard Multi-Probe liquid handling workstation (Figure 13-6) has also shown promise for off-line SPE procedures involving plasma and serum [26–28]. In addition, this unit as well as the Tecan and Biomek systems can be pro- grammed for the initial sample (e.g., plasma) transfer step from vials to the 96-well blocks, buffer addition (if applicable), and to aliquot internal standard. The advantage of the above capabilities is a signiﬁcant reduction in time and labor for the entire sample processing procedure. Possible technical problems such as carry-over by ﬁxed-tip pipettes used to aliquot the biological ﬂuid can be alleviated by incorporation of several wash cycles or their replacement with disposable pipette tips. In addition, possible inaccurate transfer of samples from the collection tubes to the 96-well blocks due to pipette tip clogging by endogenous protein clots or lipid layers should also be considered. Speciﬁc steps such as storage of the plasma samples at −80°C and/or centrifugation at 14,000 rpm prior to sample transfer can be considered for precluding ﬁbrino- gen clot formation. 13.4.2 PPT Due to its ease of use and speed, PPT is one of the most common approaches in sample preparation in early drug discovery [25]. While PPT is fast, easy-to- apply, and applicable to a broad class of small molecules, it also suffers from
614 THE ROLE OF LC–MS IN PK AND DRUG METABOLISM Figure 13-6. Photograph (top panel) of a Packard Multi-Probe (www.perkinelmer. com) platform for semiautomated PPT, LLE, or SPE process in the authors’ labora- tory. The bottom panel shows a typical layout using the corresponding operating software package. several disadvantages. Brieﬂy, in a PPT procedure, often an equal or higher volume (e.g., 1 : 3) of acetonitrile (or sometimes methanol) is added to a sample of plasma, which contains the test article as well as an internal standard. The sample is mixed and centrifuged, resulting in the formation of a protein pellet and its corresponding supernatant. The supernatant is transferred, dried, reconstituted, or directly injected onto a LC column. Clearly, this procedure is easily amenable to automation and is applicable to a host of structurally
AUTOMATED SAMPLE TRANSFER 615 diverse group of small molecules. However, PPT lacks speciﬁcity and selec- tivity that SPE or LLE can offer. Consequently, signiﬁcant matrix effect and ion suppression can be observed due to the presence of other endogenous mol- ecules that compete with the analyte(s) during ionization [29–38]. In addition, compounds that are highly bound to the protein can yield low sample recov- ery in the PPT procedure. PPT procedure is also more demanding on the MS interface, which requires more frequent cleanup, due to the endogenous inter- ference and contaminants. An example of PPT in drug development will be presented in the latter part of this chapter (STI571; vide infra). 13.4.3 LLE Liquid–liquid extraction is another well-established and attractive approach, which has been useful for the analysis of xenobiotics in biological ﬂuids. LLE can be designed to be highly selective yielding cleaner sample extracts. Brieﬂy, LLE is a mass transfer procedure where an aqueous sample (e.g., analyte con- taining biological ﬂuid) is in contact with an immiscible solvent that exhibits preferential selectivity toward one or more of the components in the aqueous sample (e.g., plasma or whole blood). In an SPE procedure (vide supra), a solid sorbent material such as an alkyl-bonded silica is packed into a cartridge, into a disk, or in a 96-well plate format, and it performs essentially the same func- tion as the organic solvent in LLE. This is particularly critical in minimizing ion suppression by co-eluting matrix components, when an ESI interface is used for the LC/MS analysis. Due to a different mechanism of operation, ion suppression is not a major determinant for signal loss in APCI [37–40]). The ion suppression is exacerbated in some cases, where the chromatography results in low peak capacity factors [39]. This could be attributed to co-elution with polar species that had also partitioned into the immiscible solvent and were consequently injected onto the HPLC column. Based on a series of experiments reported by King and co-workers [35], the order of ESI response suppression is PPT > SPE > LLE, where liquid–liquid extraction yields the least amount of analyte ion loss. 13.5 AUTOMATED SAMPLE TRANSFER Lastly, one of the labor-intensive steps in bioanalytical sample processing is the accurate initial transfer of plasma or whole blood from cryogenic vials (e.g., polypropylene tubes) to 96-well plates. This step is particularly laborious and time-consuming when a large number of samples (e.g., in 1000s) are sub- jected to analysis in late-stage clinical trials. The main bottleneck involves the “manual” uncapping and re-capping steps for each individual vial. In this regard, the Tomtec Corporation is in the process of ﬁnal testing and commer- cialization of the “Formatter” (Figure 13-7), which is designed to alleviate the above bottleneck. According to the vendor (and tests during a demo by one
616 THE ROLE OF LC–MS IN PK AND DRUG METABOLISM Figure 13-7. Photograph of a Tomtec “Formatter” (a prototype) designed to de-cap, aliquot, and re-cap sample vials containing biological ﬂuids (e.g., plasma, serum, whole blood). For more details see www.tomtec.com. of the authors), this unit will be able to accurately de-cap, aliquot, transfer, and re-cap automatically. In addition, the unit can pipette 10–450 µL accu- rately, has sample bar code tracking capability, and can interface with Watson- LIMS (www.tomtec.com). In a typical experiment, plasma samples are transferred from individual vials to a 96-well plate and subsequently processed by an appropriate extraction method on a Tomtec unit (PPT, SPE, LLE). Of course, steps such as vortexing and centrifugation still require a manual inter- vention; hence, sample extraction methods using a Tomtec are often referred to as “semiautomated.” 13.6 SAMPLE PROCESSING USING AN ON-LINE APPROACH In recent years, high-throughput and automated on-line sample extraction procedures have offered viable alternatives to improve efﬁciency for sample processing [41–46]. One such approach has been turbulent ﬂow LC. In turbu- lent ﬂow LC, single- or dual-column conﬁgurations have commonly been reported. Recently, a four-channel staggered injection system (e.g., Cohesive’s multiplexing Aria LX-4 system) has been reported for decreasing the MS idle time and improving productivity (cycle time) [45]. A typical dimension of the extraction column can be 50 × 1.0 mm (i.d.), although smaller lengths can also be used. In the single-column conﬁguration, a sample containing the analyte and internal standard is loaded on the extraction column at a
SAMPLE PROCESSING USING AN ON-LINE APPROACH 617 Figure 13-8. Simpliﬁed schematic of two-valve/two-column system (built in-house) used in the turbulent ﬂow LC experiment [47]. The ﬁrst column (e.g., Cyclone P, 50 × 0.5 mm) is used for the extraction step. In our laboratory, this design has been used in conjunction with chiral, standard narrow bore (e.g., 50 × 2 mm), and monolithic ana- lytical columns. high linear velocity (e.g., 5.0 mL/min). The analyte is retained via rapid diffu- sion into the packing, while other matrix components are washed into waste using an aqueous mobile phase. Subsequently, the analyte is eluted by a step or linear gradient and is detected by the mass spectrometer. In the dual-column conﬁguration, a standard analytical column (e.g., C18 or C8; 50- × 4.6-mm i.d.) is placed after the extraction column to improve chromatographic separation and sample cleanup. In our laboratory, we have successfully validated and applied the dual-column (Figures 13-8 and 13-9) conﬁguration to perform racemic reversed-phase as well as chiral LC-MS/MS analysis. In the latter assay, we replaced the second column by one containing a chiral stationary phase. A full account of the assay optimization and validation of an on-line achiral–chiral column conﬁguration has been reported else- where [47]. Direct sample injection has also been accomplished by using on-line C18 (4- mm i.d.) guard cartridges for cytochrome inhibition studies, capillary SPE, C18- alkyl-diol-silica restricted access phase, and PROSPEKTTM SPE modules. A more recent instrument by Spark–Holland (Figure 13-10, the Symbiosis unit) consists of Shimadzu HPLC pumps, “conditioned stacker,” an autosampler housing, and a “high-pressure dispenser” SPE station [48, 49]. This unit can be used as a stand-alone autosampler (LC mode) and/or an on-line sample- processing module (XLC mode) with a mass spectrometer. The companion software is embedded in the Sciex’s Analyst® 1.4.1 and can be easily used within the sample batch design and as part of its acquisition method. The vendor also provides 96-well plate-screening cartridges containing HySphere SPE resins such as C18, C8, C2, CN, ion exchange, and so on, in a 12 × 8 format (dimension: 2 × 10 mm).
618 THE ROLE OF LC–MS IN PK AND DRUG METABOLISM Figure 13-9. Conﬁguration of a dual channel on-line sample extraction, the CTC Trio Valve system (LEAP Technologies, www.leaptec.com; also see reference 100 for tech- nical details), which is commercially available. AC and PC signify analytical and pro- cessing columns, respectively. Figure 13-10. Photograph of a Spark–Holland system, which can be used in the off-line (LC) and on-line SPE (XLC) modes. This unit contains an autosampler, two Shimadzu pumps, a degasser, and SPE capability (for more details see www.sparkholland.com).
MATRIX EFFECT AND ION SUPPRESSION 619 A caveat for all direct sample injection assays is an understanding of the analyte chemical stability in the biological ﬂuid during the analysis period. Nonetheless, an increasingly growing body of literature is suggestive that direct injection of post-dose biological ﬂuids for quantiﬁcation purposes has become a routine and efﬁcient procedure. 13.7 MATRIX EFFECT AND ION SUPPRESSION Ion suppression or so-called matrix effect is a common problem in atmos- pheric pressure ionization (API) mass spectrometry [29–40]. There are differ- ences in opinion as to the amount of ion suppression that is acceptable for an analytical method. In some laboratories, ion suppression in an analytical method is not accept- able, but in other laboratories it is acceptable if there is no signiﬁcant effect on the validity of the analytical data with appropriate quality controls. The phenomenon of ion suppression results in reduction of signal intensity. Con- sequently, the lower limit of quantiﬁcation (LLOQ) for highly sensitive bio- analytical methods could be difﬁcult to achieve. The problem can be further complicated by (a) differential ion suppression due to intersubject variability and (b) the use of blank bio-matrix with varying lot numbers (i.e., control blood obtained from different patients/subjects) in preparing the calibrators. Differences in ion suppression between the analytes and structurally different internal standards may also be problematic. This issue can be mitigated by the use of stable-isotope-labeled (SIL) analogues as internal standards. The extent of ion suppression is dependent on the methods for sample preparation and chromatographic separation. The supernatant produced by the PPT method is most likely to cause an ion suppression in ESI due to the lack of selectivity of PPT procedure (co-eluting endogenous compounds such as lipids, phospho- lipids, fatty acids, etc., that affect the ESI droplet desolvation process; for more details see reference 38 and technical notes on www.tandemlabs.com). The extracts obtained from solid-phase extraction (SPE) and liquid–liquid extrac- tion (LLE) are relatively cleaner. Further studies on the molecular identities of co-eluting endogenous compounds, leading to ion suppression, are required to clearly delineate their contribution. One of the widely used methods to qualitatively assess the matrix effect consists of post-column addition of analytes to the LC-eluent ﬂowing from the column to the ESI interface of the mass spectrometer (Figure 13-11) [35]. Brieﬂy, an analyte and the internal standard (IS) dissolved in the same LC eluent are infused (e.g., ﬂow rate 10 µL/min) using a syringe pump, through a “tee-mixer,” located between the column eluent (e.g., ﬂow rate 200 µL/min) and the ESI interface of the mass spectrometer. An extract (using LLE or SPE) or supernatant (if PPT is used) from an analyte-free matrix, such as blank or control plasma, is injected via the autosampler, while the test article and the internal standard are introduced, post-column, to the MS ionization
620 THE ROLE OF LC–MS IN PK AND DRUG METABOLISM Figure 13-11. The post-column infusion experiment commonly used for the qualitative assessment of ion suppression and originally reported by King and co-workers [35] (MP signiﬁes mobile phase). source at a stable and continuous ﬂow using an infusion pump. Since the test article and the internal standard are introduced to the MS at a constant ﬂow, a steady ion response is obtained as a function of time. If there is an ion sup- pression, a drop in the MS ion signal is observed upon the injection of the extract obtained from a control plasma (or any other biological matrix). The infusion LC-MS/MS (MRM mode) extracted ion chromatograms of both the internal standard and the analyte are shown in Figure 13-12 for an off-line PPT plasma supernatant injection. In this example there is no signiﬁcant ion suppression from 3–10 min. If there is a signiﬁcant ion suppression, modiﬁca- tion to a more selective (cleaner) sample preparation, adoption of a more elab- orate chromatographic condition to separate ion suppressing agents from the analyte(s) of interest, and/or use of a stable label internal standard are rec- ommended [29–40]. Figure 13-13 illustrates the signiﬁcance of the choice of sample preparation on signal-to-noise ratio of an investigational compound, where an OASIS®-HLB sorbent (hydrophilic–lipophilic balanced co-polymer) (denoted as HLB in Figure 13-13) SPE and a strong anion-exchange SPE (MAX) (denoted as anion ex in Figure 13-13) yielded higher ion intensity than PPT.The count per second (CPS) signiﬁes the detector (an electron multiplier) response. More efﬁcient extraction was obtained with the anion-exchange SPE than with C18 SPE or using the PPT mode. 13.8 REGULATORY REQUIREMENTS FOR LC/MS METHOD VALIDATION Validation of quantitative LC/MS methods used in the determination of small- molecule drugs and/or their metabolites in biological ﬂuids is of paramount
REGULATORY REQUIREMENTS FOR LC/MS METHOD VALIDATION 621 Figure 13-12. Representative MRM scans obtained using the ion-suppression infusion experiment, developed by King and co-workers [35]. A signiﬁcant ion matrix effect is observed between 0.5 and 1 min using control rat plasma. The sample preparation was PPT. Note that this experiment needs to be performed “prior” to the method devel- opment and validation, so necessary changes to the sample preparation protocol and chromatographic method are made. Reprinted with permission from [101]. importance and a key determinant in obtaining reliable pharmacokinetics (PK) information. A properly developed and validated LC/MS method may often be used throughout the process of a drug’s evaluation lifecycle. These could include early discovery PK studies, preclinical toxicology studies (e.g., dose-proportionality studies), salt/formulation selection (pharmaceutical research and development), clinical PK studies, and post-marketing surveil- lance (Figure 13-14) [50–53]. Hence, intra- and interlaboratory speciﬁcity, accuracy, precision, and ruggedness have to be established [54]. In order to bridge some of the regulatory ﬁlings (e.g., within the United States, Japan, and Europe), the US Food and Drug Administration (FDA) Center for Drug Evaluation and Research (CDER), the International Conference on Harmonization (ICH), Japan’s regulatory agency, and the European Community have devised common as well as distinct requirements for bioanalytical method validation [55]. To this end, it is imperative to strictly follow these requirements during preclinical toxicology (e.g., toxicokinetics) and Phases I, II (a and b), and III clinical trials as well as during all the post- marketing PK studies (Figure 13-15). Moreover, in introduction of generic or new formulations and/or to estab- lish bioequivalence (BE) between two products, certain guidelines are fol- lowed to compare the systematic exposure of the test article to that of a
622 THE ROLE OF LC–MS IN PK AND DRUG METABOLISM Figure 13-13. Comparison of choice of sample preparation on MRM signal intensity of an investigation compound. The injection volume was 40 µL. The count per second (CPS) signiﬁes the detector (an electron multiplier) response. Protein precipitation (PPT), hydrophilic–lipophilic balanced co-polymer-based SPE (Oasis HLB- co- polymer of styrene, divinylbenzene and n-vinylpirrolidone monomers; the hydrophilic refers to the NVP monomer, and the lipophilic refers to the SDVB monomers), and strong anion exchange SPE (Max) (all in 96-well plate format) were used in control rat plasma (unpublished data). Figure 13-14. A generic ﬂow chart for the process of drug discovery and develop- ment [102].
REGULATORY REQUIREMENTS FOR LC/MS METHOD VALIDATION 623 Figure 13-15. A simpliﬁed/generic representation of clinical trials involving Phases I (safety and tolerability in healthy participants and/or patients), II (to evaluate effec- tiveness of the drug in patients), and III (to perform expanded controlled and uncon- trolled trials and gather beneﬁt–risk data). About 70–90% of drugs entering Phase III studies successfully complete this phase of evaluation. For more details see the glos- sary at www.clinicaltrials.gov. reference drug [56, 57]. The latter is a key part of the abbreviated new drug application (ANDA). New formulations could include sustained-release prod- ucts that exhibit a lag time in quantiﬁable plasma concentration. An extensive discussion of the above topics is beyond the scope of this chapter; nonethe- less, we will summarize some of the key points in the remaining segments of this section. Some of the key validation characteristics include accuracy, precision, speciﬁcity, detection limit, quantiﬁcation limit, dynamic range, linearity, matrix effect, sample, recovery, sample stability, and overall ruggedness of the assay. An evaluation of these requirements is an obligatory step in conducing reli- able PK studies under the good laboratory practice guidelines (GLP, 21 CFR Part 58 issued by FDA). In addition, software validation pertaining to a regu- lated analytical laboratory involves installation qualiﬁcation (IQ), operational qualiﬁcation (OQ), and performance qualiﬁcation (PQ), which are used to deﬁne and demonstrate process consistency and validity. The ﬁrst step or IQ ensures that the system and all its components (including calibration) are installed correctly. The OQ involves testing critical parameters, the system’s
624 THE ROLE OF LC–MS IN PK AND DRUG METABOLISM baseline setup, and variables. Lastly, the PQ is the ﬁnal phase of validation where it examines the system to perform over long periods of time within the predetermined accepted tolerance. In the PQ phase, often the individual com- ponents of the system are not tested; instead, the system is treated as a whole (for more details see www.fda.gov; General Principles of Software Validation; Final Guidance for Industry and FDA Staff, January 11, 2002). The PQ is usually performed prior to the analysis of the samples. To establish uniformity, each GLP lab has up-to-date standard operating procedures (SOPs), which all the involved bio-analytical scientists strictly follow, to ensure quality control (QC) and quality assurance (QA) compliance. Calibration curves consisting of single blank (drug free + internal standard), double blank (drug and internal standard free), and a minimum of six to eight nonzero calibrators covering the expected dynamic range are often used. QCs obtained from multiple lots of the bio-ﬂuid (e.g., serum, whole blood, plasma, urine, saliva, etc.) are prepared for short- and long-term storage. In addition, stability tests relating to 3-cycles of freeze–thaw, autosampler, bench top, freezer storage, and stock solution (neat) are established. In summary, proper validation and its documentation in a GLP, GCP (good clinical practice), and GMP (good manufacturing practice) setting, instrument qualiﬁcation records, submission of audited reports, and archiving of electronic records (21 CFR Part 11, issued by FDA) are strictly followed in order to achieve successful regulatory ﬁling. The following two case studies, Ritalin® [58] and GleevecTM [59, 60], from our laboratory are representative applica- tions of GLP compliance validated methods that were submitted for world- wide regulatory ﬁling. 13.9 RITALIN®: AN APPLICATION OF ENANTIOSELECTIVE LC-MS/MS Currently there is a trend toward the synthesis and large-scale production of a single active enantiomer in the pharmaceutical industry [61–63]. In addition, in some cases a racemic drug formulation may contain an enantiomer that will be more potent (pharmacologically active) than the other enantiomer(s). For example, carvedilol, a drug that interacts with adrenoceptors, has one chiral center yielding two enantiomers. The (−)-enantiomer is a potent beta- receptor blocker while the (+)-enantiomer is about 100-fold weaker at the beta-receptor. Ketamine is an intravenous anesthetic where the (+)- enantiomer is more potent and less toxic than the (−)-enantiomer. Further- more, the possibility of in vivo chiral inversion—that is, prochiral → chiral, chiral → nonchiral, chiral → diastereoisomer, and chiral → chiral transforma- tions—could create critical issues in the interpretation of the metabolism and pharmacokinetics of the drug. Therefore, selective analytical methods for sep- arations of enantionmers and diastereomers, where applicable, are inherently important.