Mass spectrometry for DBS analysis - 第一部份：以液相層析質譜儀結合柱後注入內標法定量採血卡上的新型抗凝血藥物第二部分：建立液相層析質譜儀結合柱後注入

1.1 Introduction

1.1.3 Mass spectrometry for DBS analysis

Due to the small sampling volume (approximately 15±5µL per spot) and the matrix complexity, low assay sensitivity and specificity are the analytical challenges for the DBS technique [32]. In the past, immunoassays and other methods have been used for determining small molecules in DBS samples. However, high performance liquid chromatography (HPLC) coupled with UV or fluorescence detection provided favorable separation and better sensitivity for quantiﬁcation [20]. Besides relative lower sensitivity and specificity of UV detector, mass spectrometry (MS) is considered as an impressive analytical tool to quantify compounds in DBS samples [26]. Mee, Korth and Halpern are the first researchers to employ chemical ionization (CI) for quantitation of free fatty acids on the DBS in 1976 [33]. Gas chromatography coupled with MS (GC-MS) is sensitive, but it has some drawbacks such as complicated sample

doi:10.6342/NTU201803315

preparation procedures especially in derivatization steps [20]. Nowadays, liquid chromatography/tandem mass spectrometry (LC–MS/MS) has been considered as the gold standard method for the highly sensitive and selective quantitation of drugs concentrations of patients’ specimen [16]. Moreover, it has been proven that the strategy of liquid chromatography electrospray ionization mass spectrometry (LC-ESI-MS) combined with DBS is an effective and efficient platform for quantitation of therapeutic d r u g m o n i t o r i n g ( T D M ) a n d p h a r m a c o k i n e t i c s s t u d i e s [ 3 4 - 3 6 ] . 1.1.4 Challenges of using DBS sampling technique in clinical practice

In spite of many advantages by using DBS sampling technique, its clinical utility is still relatively low in general practice. Analytical challenges including sensitivity, blood volume variation and hematocrit (HCT)-based assay bias (citation).

Small-volume blood samples on DBS device leads to the reduction of sensitivity of detection. Hence, using LC-MS/MS incorporated with multiple reaction monitoring (MRM) or selective reaction monitoring (SRM) technique can increase sensitivity and selectivity and ameliorate detection problems [37, 38]. HCT, defined as volume proportion (vol%) of erythrocytes in blood sample, is variable within different age, race, gender, physical status, and nutritional condition [39, 40]. It has been reported that the HCT level can cause bias while using DBS for quantitation of drugs in blood samples [40]. The HCT different may affect blood viscosity and resulting in different spot size

doi:10.6342/NTU201803315

on cellulose-based paper substrates. Other HCT caused bias includes fluctuating extraction recovery, matrix effects, and blood-to-plasma concentration ratio of the drug [32, 41, 42]. Currently, two ways were used to cut down DBS cards; one is called the center punch method by punching a fixed dimeter center, and the other is called whole spot extraction [32, 43, 44]. Although the center punch method can control the same volume of whole blood on cards, the divergent HCT values may cause assay bias due to non-homogeneity and non-dispersion characteristics of the drug on the DBS card. Naiyu Zheng et al. had demonstrated whole spot extraction can considerably reduce HCT deviation among each samples spots, and minimized spot spreading differences caused errors [45]. Despite the elimination of HCT-related bias by using whole spot extraction, the issue of controlling spotted blood volume remains to be a dilemma because of no specified blood volume between individuals. In some laboratories, using pipet or special skills, such as Mitra micro sampling devices or pre-cut dried blood spot (PCDBS), have been performed to fix blood volume [26, 46]. However, these methods for controlling the blood volume on every DBS seem to be in conflict with the advantages of the DBS sampling (i.e., cost-effective, self-get at home, and easy collection, etc.), and may not be high-throughput for clinical use [26].

In clinical perspective, the reference ranges for TDM generally are plasma or serum concentrations ; on the contrary, drug concentrations obtained from DBS

doi:10.6342/NTU201803315

sampling represent whole blood concentrations. Therefore, investigating the concentration relationship between DBS and plasma samples is important for interpreting the drug concentrations from DBS. However, the lack of well-designed studies for bridging these two sampling ways is another urgent problem for widening application of DBS technique in clinical practice. The relationship between DBS and plasma or serum concentration could be constant, but there are still some factors (i.e.

HCT value, type of cards, percentage of free form drugs, etc.) to alter slope, or intercept of the regression analysis. Thus, for conquering this problems, to explore the translation ratio (deviation %) between the concentrations obtained from DBS and plasma or serum in patients is especially important [42, 47-49].

1.1.5 Exploration of a quantitative relationship between the blood and plasma DBS samples are small volumes of whole blood from capillary vascular composed of erythrocytes, serum, platelets, white blood cells, intracellular and interstitial fluids, and the composition is quite different from plasma samples [49, 50].

Therefore, the inevitable biological discrepancies could have a great influence on measuring drug concentrations. It has been suggested that patient-related factors (i.e., HCT value and blood spot volume, etc.) and variabilities between whole blood and plasma (i.e., matrix effect and distributed ratio of drug concentration between these two compartments, etc.) are main sources that lead to discrepancies between plasma and

doi:10.6342/NTU201803315

DBS concentrations [50]. Moreover, Emmons and Rowland proposed a decision tree to interpret drug concentrations upon measuring the samples using DBS or plasma, and the considered parameters acting as the bridging strategies on the basis of blood-to-plasma drug concentration ratio, HCT constant, and unbound drug fraction in plasma [48].

Nowadays, there are several bridging studies being carried to establish the relationship of drug concentrations between DBS sample and plasma or serum, such as antiepileptic drugs, antibiotics, as well as antidepressants and anticancer drugs [25, 42, 49]. Nevertheless, the lack of NOACs data from those former studies may restrict the development of DBS in clinical monitoring. In sum, it is important to investigate the conversion factors (deviation %) and individual affecting parameters for translating DBS to plasma or serum concentrations for NOACs [42, 47-49].

1.1.6 Current methods for quantification of NOACs in DBS

The applications of DBS to analyze drug concentrations have been reported for many clinical drugs [26]. One study reported an LC-MS method for measuring the concentration of apixaban from the DBS card [45], however, using DBS sampling for TDM in clinical has not been addressed specifically in that study. The study used a pipette or a Tecan liquid handing robot to control blood volume on the DBS card. They indicated the whole spot extraction could eliminate the effect of HCT in the analysis of apixaban more effectively than the center punch method. In addition, they used

doi:10.6342/NTU201803315

liquid-liquid extraction (LLE) method to minimize the matrix effect from the phospholipid in blood. Moreover, their results showed the normalized accuracy was not affected by different HCT levels when using whole spot extraction.

The other study discussed the clinical validation of four NOACs (dabigatran, apixaban, rivaroxaban and endoxaban) in DBS card by LC−MS assay [51]. The study spotted 30 μL of whole blood by a pipette and punched the fixed size for LC-MS analysis. Although this study indicated there was no concentration differences between different DBS sites, previous study reported the accuracy of apixaban would be affected by HCT values. Their sample preparation method included a 40mins extraction by 95%

methanol followed by solid-phase extraction (SPE). The extraction recoveries for four NOACs were in the range of 24-81%. They declared no HCT effect to the quantification accuracies for four drugs. In the clinical application, they used 33 paired plasma and DBS samples collected from patients under regular NOAC therapy to obtain the DBS-to-plasma conversion factors for four drugs. The result showed high correlation between two sampling methods and the predicted and measured plasma concentrations r e v e a l e d m i n i m a l d i f f e r e n c e s ( < 2 0 % ) . The previous method used an additional equipment to control blood volume on DBS card, and used LLE combined with SPE which will increase the cost, complexity and uncertainty of the measurement. Therefore, it is critical to develop a method with

doi:10.6342/NTU201803315

simple sample preparation procedure for simultaneously estimating blood volume, correcting matrix effect and measuring drug concentrations on DBS cards.

So far there is only one study that investigates DBS and plasma concentration correlation for NOACs, but their sample number is relatively small which restricts the applications of using DBS sampling technique in clinical practice. The establishment of an effective DBS analytical method for NOACs combining with the relationship of concentrations between DBS and plasma can significantly facilitate the use of DBS in clinical assessment and fulfill NOACs therapy in personalized medicine.

1.7 Research aims

The purpose of this research is to develop a simple and accurate LC-ESI-MS coupled with postcolumn infused-internal standard (PCI-IS) method for quantification of NOACs in DBS samples. The developed method was further validated and then applied to assess the relationship between plasma and DBS samples of NOACs.

1.2. Material and Methods

1.2.1 Chemicals and reagents

Dabigatran (purity 95%), apixaban (purity 98%), [¹³C, D3]-apixaban (purity 98.3%) and rivaroxaban (purity 99.5%) were purchased from Toronto Research Chemicals (Toronto, ON, Canada). [¹³C6]-dabigatran (purity 97%) and [¹³C6]-rivaroxaban (purity

doi:10.6342/NTU201803315

98%) were purchased from Alsachim (IllkirchGraffenstaden, France). Their chemical structures are shown in Figure 1.1 (p.35). LC-MS- grade acetonitrile (ACN) were obtained from J.T. Baker (Phillipsburg, NJ, USA). Mass-grade methanol was purchased from ScharlauChemie (Sentmenat, Barcelona, Spain). MS-grade formic acid solution (99%) was obtained from Sigma-Aldrich (St. Louis, MO, USA). Whatman 903 Protein Saver Cards (Maidstone, UK) were used for DBS sampling, and the paper is manufactured by 100% pure cotton liners without wet-strength additives. The 6 mm puncher, zip-lock bags and desiccants were purchased from local store.

1.2.2 UHPLC-ESI-MS system

The LC separations were performed using an Agilent 1290 UHPLC system coupled to a binary solvent pump, an autosampler, a sample reservoir, a column oven, and an Agilent 1260 quaternary solvent pump was used for the postcolumn infusion of the PCI-IS (Figure 1.2, p.35). The mass spectrometer was an Agilent 6460 triple quadrupole system (Agilent Technologies, Waldbronn, Germany). A Kinetex reversed phase core-shell C18 column (2.1 × 50 mm, 2.6 µm, 100 Å , Phenomenex, Torrance, CA, USA) was used for the separation. The mobile phase consisted of 0.1% formic acid in water (solvent A) and 0.1% formic acid in acetonitrile (solvent B). The flow rate was set at 0.35 min/mL. The gradient profile started with 1% B for 0.5 min, then changed to 15% B in 0.1 min and remained for 0.9 min, then increased to 100%B in 0.6 min and

doi:10.6342/NTU201803315

stayed for 1.4 min. Finally, the column was re-equilibrated to 1% for 1.5 min until the next injection. The total run time was 5 min. The temperature of sample reservoir was maintained at 4 , and the column oven was set at 40 . The injection volume was 3 µL.

The positive electrospray ionization mode was utilized with the following parameters: 350°C dry gas temperature, 10 L min⁻¹dry gas ﬂow rate, 45 psi nebulizer pressure, 350°C sheath gas temperature, 11 L min⁻¹ sheath gas ﬂow rate, 3500 V capillary voltage, and 500 V nozzle voltage. The MS acquisition was executed using the multiple reaction monitoring (MRM) mode and the optimized parameters for dabigatran, apixaban, rivaroxaban and the internal standards were summarized in Table 1.1 (p.42).

Two transitions were selected for each compound as quantifier and qualifier.

[¹³C6]-dabigatran, [¹³C, D3]-apixaban and [¹³C6]-rivaroxaban used as the PCI-IS were all dissolved in ACN at 15 ng mL^-1and introduced into the ESI interface at a rate of 0.1 mL min^-1.

1.2.3 Standard solutions and calibration standards

Stock solutions of dabigatran,[¹³C6]-dabigatran each at 1.00 mg/mL were prepared in methanol/water(9:1) containing 0.1% formic acid. Apixaban and [¹³C, D3]-apixaban were prepared in methanol at 1.00 mg/mL and 0.5 mg/mL, respectively.

Rivaroxaban and [¹³C6]-rivaroxaban were prepared in acetonitrile/water (9:1) at 1.00

doi:10.6342/NTU201803315

mg/mL and 0.1 mg/mL, respectively. The mix working solution of dabigatran, apixaban and rivaroxaban were prepared at 100 µg/mL by dilution of the stock solution with methanol. Calibration standards were prepared by diluting working solutions with methanol in sequence at the following concentrations：0.125, 0.5, 0.625, 1.25, 2.5, 10, 20 µg/mL for rivaroxaban and 0.25, 0.5, 1.25, 2.5, 10, 20 µg/mL for dabigatran and apixaban respectively. All of the solutions were stored at -20 .

1.2.4 Samples collection

The collections of human DBS and plasma samples were approved by the Research Ethics Committee of National Taiwan University Hospital (NTUH REC No:

201706110RIFB). The fixed 15 µL whole blood was spotted onto DBS card just after collecting the blood in EDTA-tubes. The rest of the blood sample was centrifuged to obtain plasma and then stored at -80°C until analysis. The DBS cards were dried at room temperature in the dark for 2 hours followed by placing in double-layers zip lock bag with a lime desiccant and then stored in sealed box at −20 °C until analysis.

1.2.5 DBS sample preparation

The working solution containing dabigatran, apixaban and rivaroxaban was prepared fresh on the day of analysis by dilution the stock solution with methanol. The blood samples were obtained through peripheral vein via syringe collection with EDTA tubes to prevent coagulation. To prepare spiked blood for DBS samples, 4 μL of

doi:10.6342/NTU201803315

corresponding concentrations of NOACs working solution was spiked into 196 μL of whole blood and then allowed to equilibrate for 30 min at RT on the bench before spotting onto DBS cards [52]. Fifteen microliters of spiked whole blood were spotted accurately onto a Whatman 903 card by a pipette and dried at room temperature in the dark overnight (Figure 1.3 (A), p.36) [53, 54].

For the volume experiment, seven different volumes of blood samples were spotted onto DBS cards by pipetting with different blood volumes (10, 15, 20, 25, 30, 35μL) to test method accuracy (Figure 1.3 (B), p.36).

In terms of preparing samples with different hematocrit concentration levels, the whole blood samples were centrifuged at 15000 rcf for 10 min at 4 to separate the plasma from the red blood cells. The upper plasma layer and the lower red blood cell layer were then mixed or removed with different amounts to obtain artiﬁcial blood samples with specific hematocrit values (25, 35, 45, 55, 65%) (Figure 1.3 (C), p.36).

1.2.6 Dried blood spot extraction

The 6-mm diameter puncher was cleaned with 75% alcohol before utilization.

After drying, each DBS card was manually cut by puncher for several times and transferred into a clean 2.0 mL Eppendorf tube. 300 L of water containing 0.1% formic

acid was added to each tube and then shacked with a Geno/Grinder 2010 (SPEX^® Sample Prep, Metuchen, NJ) for 2 min at 1000 rpm. A 700 µL aliquot of acetonitrile

doi:10.6342/NTU201803315

was added into the water extracted then extracted with Geno/Grinder 2010 for another 3 min. After centrifugation at 18000 rcf for 5 min, 800μL of supernatant was removed to another 2.0 mL Eppendorf tube and evaporated by using EYELA CVE-200D Centrifugal Evaporator (TOKYO RIKAKIKAI CO., Tokyo, JP) until dryness. The residue was reconstituted with methanol followed by shaking with Geno/Grinder for 2 min at 1000 rpm and sonicating for 10 min. Finally, the supernatant was filtered through a 0.2-μm PP membrane filter (RC-4, Sartorius, Göttingen, Germany) after centrifugation at 18000rcf for 5 min at 4℃ and then analyzed by UHPLC-ESI-MS (Figure 1.3 (D), p.36).

1.2.7 Method validation

Method validation was implemented according to the USFDA guidelines (Guidance for Industry: Bioanalytical Method Validation). The validation included selectivity, range of calibration curves, lower limit of quantification (LLOQ), lower limit of detection (LOD), accuracy, precision, recovery, matrix effect, carryover effect, and stability. In addition, DBS-speciﬁc parameters such as hematocrit were investigated according to the European Bioanalytical Forum (EBF).

1.2.7.1. Calibrators, LLOQ and LOD

The working solutions containing dabigatran, apixaban and rivaroxaban for preparing calibration curves were generated by series dilution of the stock solution

doi:10.6342/NTU201803315

(100μg mL⁻¹) with methanol. Aliquot of working solution ranging from 20 to 0.125 μg

mL⁻¹ was spiked to blank whole blood and then spotted on DBS cards to generate the calibration curve and each concentration was analyzed for five replicates. Calibration curves for dabigatran and apixaban were assessed with six-point levels at following concentrations ： 10, 20, 50, 100, 400 and 800 ng mL⁻¹. Calibration curves for rivaroxaban was assessed with seven-point levels at following concentrations：5, 20, 25, 50, 100, 400 and 800 ng mL ⁻¹.The three calibration curves were plotted by using linear regression of the area ratio of target analytes to PCI-IS versus the target analyte concentration and only apixaban was obtained with 1/x weighting factor. The limit of detection (LOD) and lower limits of quantification (LLOQ) were defined as the signal to noise (S/N) ratio of 3:1 and 10:1, respectively.

1.2.7.2. Extraction recovery and matrix effect

Recoveries and matrix effects were measured in triplicate for three NOACs at lower limit of quantification, low, medium, and high levels mentioned in section 2.7.1.

Recoveries were calculated by comparing the peak areas of the pre-spiked samples (NOACs were added to blank whole blood before extraction) with the peak area of the

post-spiked samples (NOACs were added to the processed blank samples). The matrix eﬀect was measured by dividing the peak area of the post-spiked sample by the peak

area of the pure standard solution at low, medium, and high NOACs concentrations (n=

doi:10.6342/NTU201803315

6).

1.2.7.3. Selectivity

Selectivity was determined by analyzing blank DBS samples from seven different drug-free whole bloods under the optimized UHPLC-MS/MS conditions and extraction conditions.

1.2.7.4. Accuracy and precision

To evaluate accuracy and precision of the developed method, four concentrations of NOACs were spiked into whole blood to obtain LLOQ, low, medium, and high concentrations and then spotted to DBS cards. Concentrations for dabigatran and apixaban were 10, 25, 200, and 600ng mL⁻¹. Concentrations for rivaroxaban were 5, 10, 200, and 600ng mL⁻¹. Intra-day precision was determined by analyzing the spiked DBS samples for five times within the same day. For inter-day precision, three different samples per concentration were analyzed at three separate days. The results for intra-day and inter-day precision were showed as RSD (%). Intra- and inter-day accuracies were evaluated by calculating the recoveries at the four concentrations. The acceptance levels for precision and accuracy were ≤15% and 85–115% respectively, expect for the LLOQ level. The acceptance levels for precision and accuracy at LLOQ

doi:10.6342/NTU201803315

were ≤20% within 80–120%, respectively.

1.2.7.5. Stability

The stabilities test concentrations were 25 and 600ng mL⁻¹for dabigatran and apixaban, and they were 10 and 600ng mL⁻¹ for rivaroxaban. Aliquots of working solutions were added into the whole blood for preparing DBS samples. After drying overnight at room temperature in the dark, the dried DBS samples were stored in

double-layers zip lock bags with a desiccant [20, 54]. The long term stabilities were evaluated at 25 ^◦C, 4 ^◦C and −20 ^◦C for 7 and 30 days. Fresh prepared DBS and stability samples were extracted and calculated for the recoveries. Three replicates of low and high concentration QC samples were tested and the acceptance criteria at each level should be within 15%.

1.2.7.6. Clinical application

269 DBS and plasma paired-samples for three NOACs were collected from the National Taiwan University Hospital, and these samples were used to establish the correlation for NOACs between DBS and plasma concentrations. This study was approved by the Research Ethics Committee of the National Taiwan University Hospital (registration number, 201706110RIFB), and all patients participated in this study signed informed consent prior to enrollment. To prepare paired DBS and plasma samples, fifteen microliters of blood samples were spotted on DBS cards and the rest of the blood

doi:10.6342/NTU201803315

samples were centrifuged at 3000 rcf for 15 min to obtain plasma samples. The DBS samples were stored at -20°C and the plasma samples were stored at −80 °C until use.

1.2.8 Data analysis

The Deming regression was generated by sigmaPlot software version 13 and statistical parameters were calculated using Excel 2010. Data obtained from the Agilent triple quadruple were converted into xls format, and processed by R software. Medcalc software version 12.3.0.0 (http:// www.medcalc.org) was used for Bland and Altman comparison plot.

1.3 Results and discussion 1.3.1 Method development

1.3.1.1 Optimization of the extraction solvent

We selected 15µL whole blood as the sampling volume due to its universality in clinical use. In addition, whole spot-cut DBS approach was performed to effectively eliminate the effects in the analysis of NOACs from different HCT levels, spot volume, and different punch sites on the same spot. To optimize the extraction recovery and develop a simple analytical method, the extraction solvent types and extraction time were discussed below. On the basis of previous studies, methanol and acetonitrile were

在文檔中第一部份：以液相層析質譜儀結合柱後注入內標法定量採血卡上的新型抗凝血藥物第二部分：建立液相層析質譜儀結合柱後注入內標法定量磷脂醯乙醇胺以測量採血卡上的血比容 (頁 21-0)