CHAPTER 1 INTRODUCTION
1.3 Research aim of this dissertation
The DBS sampling technique contain numerous advantages compared to the traditional plasma sampling technique. However, due to the critical challenges associated with the DBS sampling technique, its applications in clinical practice, especially in TDM, forensic toxicology and metabolomics are still minimal. Resolving the critical challenges associated with the DBS sampling technique, would allow it to be widely utilized for various clinical fields.
In this dissertation, I will discuss the approaches which we used to mitigate the DBS associated challenges to increase its applicability in pharmaceutical and metabolomics analysis for clinical practice.
Due to the small sample volumes used for DBS, sensitivity is the major challenge in developing DBS based methods. In general, mass based LC-MS/MS platforms are commonly used for DBS based methods to improve the sensitivity and specificity of the assay. But for certain drugs with very low therapeutic concentrations, general optimization of LC-MS parameters may not be sufficient to improve the method sensitivity. Moreover, improvement of compound sensitivity in MS depends upon the amount of ionized analyte ions that enters into the detector, improvements in ionization efficiency increases the assay sensitivity. Therefore, in the second chapter we discussed the strategy that incorporates the ion-booster (IB) source to improve assay sensitivity for wide range abused drugs in DBS samples.
doi:10.6342/NTU201804215
9
The two critical factors of variation in Hct effect and spot volume associated with DBS sampling technique largely influences the applicability of the DBS sampling technique in clinical practice. Therefore, in the third and fourth chapters of this thesis we discussed resolving these issues by developing the PCI-IS assisted LC-ESI-MS/MS method for both pharmaceutical and metabolomics analysis using DBS samples. In the third chapter we developed an accurate method for the simultaneous quantification of 6 anti-HIV drugs in DBS samples. Correlated concentrations between DBS and plasma demonstrate the suitability of DBSs for clinical monitoring. In the fourth chapter we developed an analytical platform for DBS-based metabolomics studies to improve their data quality. Through developing effective and accurate methods for DBS-based metabolomics studies, we anticipate the application of the DBS sampling technique to more clinical applications.
doi:10.6342/NTU201804215
10
Chapter 2
Sensitive screening of abused drugs in dried
blood samples using ultra-high-performance
liquid chromatography-ion booster-quadrupole
time-of-flight mass spectrometry
(UHPLC-IB-QTOF-MS)
doi:10.6342/NTU201804215
11
2.1 Introduction
Drug abuse is an increasing global social burden. Most social problems are associated with drug abuse, including sexual assault, child abuse, suicide, murder, traffic accidents, and violence. The increase of drug abuse rates is a global problem. Globally, the occurrence of drug abuse including opioid, cocaine and amphetamine drugs was 5.2%
(range: 3.5 to 7.0%) in the 15 to 64-year-old population from 2009 to 2012 33. In 2013, the estimated rate of drug abuse including cocaine, hallucinogens, and nonmedical use of prescription-type sedatives in the United States was 9.4% of the population aged 12 or older 34. Therefore, it is important to have efficient and effective screening methods for drug-related crime and clinical management.
Several analytical methods have been developed to analyze abused drugs in biological samples include urine and whole blood 33, 35-38. Currently, urine is the most commonly used biological fluid for drug testing. However, adulteration of urine samples may give false positive results, and it is not possible to obtain urine samples in the case of death. Although blood samples avoid certain limitations of urine samples, the invasive collection and the storage and handling of whole blood samples limit the wide application of this sampling method. Additionally, blood collection may cause the transmission of some diseases such as HIV, hepatitis, and other blood-borne viruses 39.
The DBS sampling technique has a long history that can be traced back to 1960, but until recently this technique had not gained great attention in the bioanalysis of adults.
The DBS method spots a small amount of blood on a filter paper card followed by a drying process. The drying process increases the stability of most photosensitive compounds.40-41 Additional advantages of the DBS technique include low biohazard risk during the shipment of samples and decreased invasiveness 39, 42-43. Although the DBS technique is increasingly used in medical care due to its broad advantages, the application
doi:10.6342/NTU201804215
12
of this convenient sampling technique to abused drugs is still rare.
To date, only a few methods have been developed to analyze abused drugs in DBS.
Radioimmunoassays using DBS samples have been used to analyze benzoylecgonine (BZE), a main metabolite of cocaine, in newborns and child-bearing women 44-45. Cocaine and its metabolites have been analyzed in DBS by high-performance liquid chromatography fluorescence 46. One of the main drawbacks of the DBS technique is its sensitivity because only a small amount of blood is spotted on the card. The high sensitivity and selectivity of LC-ESI-MS makes the technique a perfect match for coupling with DBS analysis. Therefore, growing numbers of LC-MS methods have been developed to analyze abused drugs on DBS cards. However, most studies have only included very limited abused drugs such as cocaine, opiates and benzodiazepines 47-49. Recently, Odoardi et al. developed an LC-MS/MS method to analyze opiates, methadone, fentanyl and analogues, cocaine, amphetamines and amphetamine-like substances, but the sample extraction time was not efficient, which may hamper its application to routine screening work50.
Currently, the most commonly used MS platforms to analyze drugs of abuse in biological samples are triple quadrupole and TOF mass spectrometry 35, 51-55. The need for pre-established transition ion pairs and the decreased sensitivity with the increased number of transition ion pairs in multiple reaction monitoring mode are the major limitations of triple quadrupole mass spectrometry 56. Time-of-flight mass spectrometry provides accurate mass measurement at the millidalton (mDA) range as well as high full-scan sensitivity. The high mass accuracy provides the advantage of using exact monoisotopic masses and isotopic patterns for compound identification. This advantage provides an opportunity for extension of the screening targets. However, with respect to the identification of unknown abused drugs in biological samples, TOF-MS can give false
doi:10.6342/NTU201804215
13
positive results when simply identifying by accurate mass 57. Therefore, creating spectral information using a hybrid mass such as QTOF-MS is an effective method for the identification and confirmation of abused drugs. Recently, methods using QTOF-MS have been developed for the screening and confirmation of abused drugs in biological samples 56, 58. Considering the selectivity and sensitivity of LC-QTOF-MS, using this powerful tool will be beneficial for abused drug screening on DBS samples.
This study aims to develop a simple and efficient analytical method to screen a wide range of abused drugs in DBS samples using UHPLC-IB-QTOF-MS. The most commonly used abused drugs, including amphetamines, opioids, cocaine, benzodiazepines, ketamine, lysergic acid diethylamide (LSD) and many other new and emerging abused drugs were selected as our target analytes. Metabolites including norketamine, norephedrine, 7-aminoflunitrazepam, and nordiazepam were also included in this study. To provide sufficient sensitivity for low-concentration drugs in the small quantities of DBS samples, an IB ion source with an extra heated spray zone was applied.
IB uses a controlled vaporizer temperature to enhance ionization efficiency of the target analytes by evaporating the solvent of analyte ions even at high mobile phase flow rates.
As the sensitivity improvement using IB source depends upon the chemical properties and thermal stability of the analytes, this study evaluated the sensitivity improvement for various abused drugs. In addition, this study performed a six-month stability test to investigate the stability of abused drugs on the card. Although improving compound stability is one of the main advantages of the DBS sampling technique, there is no previous research investigating the stability of abused drugs on DBS cards, which limits its use in real-world analysis. To the best of our knowledge, this is the first study developing a simple and efficient UHPLC-IB-QTOF-MS method for a wide range of abused drugs using a DBS sampling technique. This method offers a new approach for
doi:10.6342/NTU201804215
14
abused drug control.
2.2 Experimental section 2.2.1 Standards and reagents
Amphetamine, alprazolam, 7-aminoflunitrazepam, amobarbital, aminorex, bromazepam, butalbital, butorphanol, 4-bromo-2,5-dimethoxyphenethylamine (2C-B), butabarbital, clonazepam, chlordiazepoxide, clobazam, dihydrocodeine, diazepam, ephedrine, estazolam, fentanyl, flurazepam, flunitrazepam (FM2), ketamine, lorazepam, lormetazepam, LSD, methamphetamine, 4-methoxyamphetamine (PMA), 3,4-methylenedioxyamphetamine (MDA), 3,4-methylenedioxymethamphetamine (MDMA), 3,4-methylenedioxy-N-ethylamphetamine (MDEA), para-methoxymethamphetamine (PMMA), meperidine, methadone, midazolam, methylephedrine, methylphenidate, norketamine, norephedrine, nitrazepam, nordiazepam, nalorphine, lorazepam, pentazocine, phentermine, prazepam, pseudoephedrine, secobarbital, triazolam, temazepam, tramadol, and zolpidem were purchased from Cerilliant (Round Rock, Texas, USA). Cocaine hydrochloride, codeine, morphine, pentobarbital, barbital, phenobarbital, and ammonium acetate were purchased from Sigma-Aldrich (St. Louis, MO, USA).
Phencyclidine (PCP) was purchased from Triage®Tox Drug Screen, Biosite, (San Diego, CA, USA). ACN and acetic acid were purchased from Merck (Darmstadt, Germany).
MeOH and DI water were purchased from Scharlau (Spain). Ammonium bicarbonate was purchased from J.T. Baker (Phillipsburg, NJ, USA). All reagents and solvents used were of LC-MS grade.
doi:10.6342/NTU201804215
15
2.2.2 UHPLC-IB-QTOF-MS
The Agilent 1290 UHPLC system consisted of a degasser and a quaternary solvent pump (Agilent Technologies, Santa Clara, CA) coupled with a Bruker maXis QTOF (Bruker, Rheinstetten, Germany) equipped with an IB source. A Poroshell EC-C18 column (2.1×100 mm, 2.7 μm, Agilent) was used for compound separation. The mobile phase was composed of 0.1% acetic acid in DI water (solvent A) and MeOH (solvent B).
The gradient profile used for positive ionization detection was as follows: 0-5 min, 2-20%
B; 5-9 min, 20-60% B; 9–13 min, 60-90%B; 13-15 min 90-95% B; and then re-equilibration of the column for 3 min. The gradient for negative ionization detection was as follows: 0-2.5 min, 25-65% B; 2.5-4 min, 65-85% B; 4-5 min, 85% B. The column re-equilibration time for both modes after each gradient was 3 min. The flow rate was maintained at 0.4 mL min-1, and the sample injection volume was 10 µL. The parameters of the mass spectrometer for both positive and negative ionization mode were as follows:
end plate offset voltage was 400 V, charging voltage was 300 V, capillary voltage was 1000 V, drying gas flow was 4 L/min, nebulizer flow was 60 psi, drying gas temperature was 200 °C, vaporizer temperature was 330 °C and sheath gas flow was 150 L/hr. The mass spectrometer was calibrated daily with 10 mM sodium formate before analysis.
Methyl stearate and hexakis were used as the reference mass for positive and negative ionization modes for mass accuracy correction, respectively.
2.2.3 Sample collection
Whole blood samples were collected from healthy volunteers without medication.
The stock standard solution (10 μg mL-1) was prepared in a mixture of MeOH / DI water (50/50, v/v), and the working standards were diluted with DI water prior to addition to the whole blood samples.
doi:10.6342/NTU201804215
16
2.2.4 Sample preparation
Twenty-five microliters of spiked whole blood were carefully spotted on a Whatman 903 card using a pipette and dried for 2 h at room temperature. After drying, a 6-mm diameter (8.9 ± 1.7 μL blood) card was manually punched from each spot and transferred into a Protein LoBind Eppendorf (Sigma-Aldrich, Hamburg, Germany), to avoid the loss of analytes onto the active surfaces of the Eppendorf. The dried spot was extracted with 200 μL of 80% ACN for 5 min at 1000 rpm using a Geno grinder (SPEX®
Sample Prep (2010), Metuchen, NJ) and subjected to centrifugation at 15,000×g for 5 min. One hundred and seventy microliters of the supernatant were evaporated under nitrogen gas and reconstituted with 150 μL of DI water. After vortexing, the samples were filtered and injected onto the LC system for analysis.
2.2.5 Validation
Qualitative validation tests, including selectivity, limit of detection (LOD), precision, extraction recovery, matrix effect, carryover, and stability, were performed on pre-spiked DBS samples over the suggested therapeutic concentrations. The therapeutic levels listed in Table 2.5.2 were provided by the literature 59.
The selectivity test was performed using six whole blood samples by spotting onto the cards. Aliquots of each drug-free DBS sample were prepared by the sample preparation method described previously. The six drug-free DBS samples were used to check endogenous interference at the same retention time and exact mass of the abused drugs and their metabolites. The LOD of each analyte was signal-to-noise ratio equaled 3 (S/N=3). The intra-day precision and inter-day precision were determined in whole blood samples fortified at low concentration that is limit of quantification (S/N=10) and high concentrations (20 x times to the low concentration) (n=9) before spotting on the cards.
doi:10.6342/NTU201804215
17
Both intra-day and inter-day precision of the analyte retention time (RT) and the signal intensity were expressed as CV (%). The extraction recovery of all analytes was determined by preparing pre-spike DBS samples using drug-free whole blood samples fortified at low and high concentrations before spotting on to the cards and post-spike DBS samples using drug-free whole blood samples spotting on to the cards. After extraction, post-spike DBS samples were fortified with the same concentrations as the pre-spike samples after extraction. The extraction recovery (%) was calculated from the peak area ratio of compounds from pre-spike and post-spike samples, multiplied by 100.
The matrix effect was evaluated in drug-free DBS samples fortified at low and high concentrations after extraction and standard samples with the same concentrations. The matrix effect (%) was calculated from the peak area ratio of compounds spiked in the matrix to the standard compounds, multiplied by 100. The stability test was performed on whole blood samples fortified at 0.2 µg mL-1. The fortified samples were spotted on the DBS card and allowed to dry overnight at room temperature (rt). The dried DBS samples were stored in sealed bags containing desiccant to avoid humidity and contamination at rt and -80 °C for six months to evaluate the compound stability on DBS cards. In addition, the DBS samples for the one-month stability test were prepared by the same protocol one month before the end of the six months and stored at rt and -80 °C. After six months, all the samples, including freshly spotted DBS samples, were prepared using the sample preparation method described previously. The stability of the compounds was evaluated by comparing the analyte responses obtained after one or six months of storage with the responses from freshly prepared samples.
doi:10.6342/NTU201804215
18
2.3 Results and discussion
2.3.1 Sample preparation method development
To obtain the highest extraction efficiency for all 57 compounds without compromising the simplicity of the extraction process, the extraction solvent type and extraction time were investigated. Since the studied drugs were ionic compounds, aqueous buffers including ammonium acetate (pH 6), ammonium bicarbonate (pH 6) and 0.1% acetic acid in 50% MeOH were used as the extraction solution. Although these methods could provide satisfactory recoveries (average of 80% for all compounds), the major problem with the aqueous buffer extractions was their inability to denature the protein since the hemoglobin protein was co-extracted along with the abused drugs. To remove the residual protein and minimize its effect on the LC-MS system, a solid-phase extraction (SPE) procedure should be included following the solvent extraction 47, which increases the cost and complexity of the extraction method. To avoid this problem, we tested organic solvents such as 80% ACN for extraction. The results showed that 80%
ACN provided an average of 80% recovery for all the compounds, and 80% ACN was therefore selected as the extraction solvent (Fig. 2.6.1).
Previous studies used ultra-sonication for DBS sample extraction.10, 12. Our preliminary study showed that the optimal extraction time when using ultra-sonication was 30 mins. To speed up the sample preparation process, a Geno grinder offering a fast extraction rate was tested for extraction of abused drugs from the DBS cards. Extraction times using the Geno grinder were investigated, including 1 min, 3 min and 5 min. Fig.
2.6.2 shows the comparison of all the extraction times, and a 5-min extraction time provided recoveries higher than 80% for all the compounds, with CV lower than 20%. A 5-min extraction time was therefore chosen as the optimized extraction time. Compared with previous DBS methods 47-48, 50, this study greatly shortens the sample preparation
doi:10.6342/NTU201804215
19
time.
2.3.2 LC-QTOF-MS method development 2.3.2.1 LC method optimization
To provide efficient screening for a wide range of abused drugs, a Poroshell column, which provides high column efficiency, was employed in this study. A gradient profile using 0.1% acetic acid in DI and MeOH was developed for both positive and negative mode for separation of all analytes. Since structural isomers could not be differentiated using their accurate mass, the gradient profile was designed to provide sufficient separation for three sets of isomers, namely, ephedrine / pseudoephedrine / PMA (m/z 166.1226), clobazam / temazepam (m/z 301.0738) and pentobarbital / amobarbital (m/z 225.1245). Using the optimized gradient profile from the positive mode, the structural isomers of ephedrine / pseudoephedrine / PMA and clobazem / temazepam could be well separated within 15 min. However, pentobarbital and amobarbital could not be separated and were therefore expressed as “pentobarbital / amobarbital” in the identification results. The chromatograms obtained from the DBS samples under the developed conditions from both positive and negative mode are shown in Fig. 2.6.3.
2.3.2.2 Mass parameter optimization
One of the main challenges for DBS sample analysis is the high sensitivity requirement due to the limited quantity of analyte on the card (typically <30 µL). An IB source containing an additional heated spray zone was used to improve the detection sensitivity. Compared to an ESI source, the main drawback of using an IB source is that it is not suitable for thermally unstable compounds and the sensitivity improvement depends upon chemical characteristics of the analytes 60. To examine the limitations of
doi:10.6342/NTU201804215
20
the IB source, analyte responses from both ESI and IB sources was compared to select which provided better sensitivity for the greater number of abused drugs.
Mass parameters for both ESI and IB were optimized to achieve the best detection sensitivity. It was found that ESI source parameters including drying gas, drying temperature, and nebulizer gas greatly affected the ionization of analytes, and tuning parameters including hexapole RF, collision RF, transfer time, and pre-pulse storage, which affects the movement of ions through the channels of QTOF, also significantly affected signal intensities. However, the sensitivity improvement after optimizing these parameters was not sufficient for the compounds with lower therapeutic concentrations, which might be due to insufficient solvation and ion formation of analytes when using ESI as the ionization source.
The working principal of the IB source is similar to the ESI source; however, it contains a heated vaporizer, the charge-HV transfer tube and the additional use of nitrogen as a sheath gas to improve the evaporation and ionization of analytes. In this study, approximately 83% of the analytes showed an enhanced response with the initial conditions of the IB source (Fig. 2.6.4). IB source ionization parameters, including end plate off-set, capillary voltage, charging voltage, nebulizer gas, drying gas, drying temperature and vaporizer temperature (VT), were further optimized, and VT had the most significant effect on the signal intensity of all the parameters (Fig. 2.6.5). After optimizing all the parameters, all the analytes, when including both positive mode and negative mode, showed improvements ranging from 1.5 to 14-fold as compared to the ESI source. With this improvement, approximately 95% of the compounds could be detected at their lower therapeutic levels, which makes the method more suitable for toxic and forensic drug screening.
Our results revealed that combining the IB source with QTOF detection provided
doi:10.6342/NTU201804215
21
sensitive results with lower sample volumes such as that of the DBS sampling technique
sensitive results with lower sample volumes such as that of the DBS sampling technique