University of Amsterdam Faculty of Science
MSc Chemistry
Analytical Sciences
Graduation research thesis
Determination of ethanol metabolites in hair
by LC-MS(/MS) and GC-MS(/MS)
byLiz Leenders
Supervisors: dr. W. F. Duvivier dhr. dr. W. T. Kok2
List of abbreviations
ORS Outer root sheath
IRS Inner root sheath
SoHT Society of Hair Testing
EtG Ethyl glucuronide
FAEEs Fatty acid ethyl esters
FA Fatty acids
TG Triglycerides
LP Lipoproteins
PL Phospholipids
SPE Solid-phase extraction
MS(/MS) (Tandem) mass spectrometry
GC Gas chromatography
LC Liquid chromatography
m/z Mass to charge
CID Collision induced dissociation
SRM Selected reaction monitoring
MRM Multiple reaction monitoring
EI Electron ionization
LOD Limit of detection
LOQ Limit of quantitation
HS-SPME Headspace solid-phase micro extraction BSTFA N,N-bis(trimethyl-silyl)trifluoro-acetamide
PFPA Pentafluoropropionic anhydride
HFBA Heptafluorobutyric anhydride
PBA Phenylboronic acid
PCI/NCI Positive/negative chemical ionization
ESI Electrospray ionization
HILIC Hydrophilic interaction liquid chromatography
MMS Matrix-matched standards
MMRS Matrix-matched recovery standards
CCα Decision limit
3
Abstract
Hair analysis is increasingly used over the past years in the detection of drugs in both forensic as clinical science. It is a sufficient and sensitive method to monitor drugs or alcohol use over a long period of time. When compared to body fluids, compounds in hair can be detected over a longer timeframe, and segmented hair analysis could give information about drug and alcohol abuse in a retrospective way. The problem in the case of ethanol is the fact that ethanol itself cannot be detected in hair due to its high volatility, only its metabolites are incorporated into hair.
In this research project, a method is developed for the detection of two ethanol metabolites in hair with LC- and GC-MS(/MS). A method for the analysis of one of the metabolites, EtG, was developed and optimized. The optimized method consisted of finely cutting 50 mg of hair, extraction with Milli-Q water in an ultrasonic bath for 4 hours followed by overnight incubation, clean-up by SPE, evaporation and reconstitution for LC-MS/MS. The method was fully validated with a mean trueness of 99.5%, recovery of 90.55%, CCα of 31.83 µg/L and CCβ of 63.65 µg/L.
The method was further tested on 2 reference and 42 real hair samples. The method was able to measure known, large concentrations accurately and is able to detect chronic excessive alcohol consumption (30 pg/mg), but the sensitivity is too low for the detection of regular alcohol consumption (7 pg/mg). Also, the initial method development for the analysis of both EtG as well as FAEEs with GC-MS(/MS) was executed. Derivatization of EtG was possible with PBA, and EtG-PBA could be detected in GC-MS/MS, however the intensity was extremely low. For the analysis of FAEEs, ethyl palmitate is extracted from the hair matrix with hexane and DMSO, followed by overnight incubation, than clean-up by SPE, evaporation and
reconstitution. These steps are similar to the optimized method for ethyl glucuronide for LC-MS/MS, so a combined sample pre-treatment could be possible in the future. The method was able to analyse ethyl palmitate with GC-MS/MS.
Samenvatting
Haar analyse wordt steeds vaker gebruikt voor het detecteren van drugs in zowel forensische als klinische scheikunde. Het is een geschikte en gevoelige methode om drugs en alcohol gebruik over een lange periode in kaart te brengen. In haar zijn moleculen langer te detecteren dan in andere humane materialen, en door een haar gesegmenteerd te analyseren kan informatie gewonnen worden op een retrospectieve manier. Het probleem in het geval van ethanol, is het feit dat ethanol niet te detecteren is in haar omdat het een hele hoge vluchtigheid heeft. Metabolieten van ethanol echter zijn wel te detecteren in haar.
In dit onderzoeksproject is een methode ontwikkeld om ethanol metabolieten in haar te detecteren aan de hand van LC- en GC-MS(/MS). Een methode voor de analyse van een van de metabolieten, EtG, was ontwikkeld en geoptimaliseerd tot het volgende: 50 mg haar wordt heel fijn geknipt, EtG wordt geëxtraheerd met water in een ultrasoon bad voor 4 uur gevolgd door incubatie gedurende de nacht, de monsters worden behandeld met SPE, gevolgd door verdamping en her-oplossing voor LC-MS/MS. De methode is vervolgens volledig
gevalideerd met een jusitheid van 99.5%, een terugvinding van 90.55%, een CCα van 31.83 µg/L en een CCβ van 63.65 µg/L.
De methode is getest aan de hand van 2 referentie en 42 echte haarmonsters. De methode was in staat om bekende, hoge concentraties van EtG nauwkeurig te meten, chronisch alcohol gebruik te detecteren (30 pg/mg), maar had helaas een te lage gevoeligheid om regelmatig alcohol gebruik te detecteren (7 pg/mg). Vervolgens is er geprobeerd om een methode te ontwikkelen voor het detecteren van zowel EtG als FAEEs met GC-MS(/MS). Het derivatiseren van EtG bleek mogelijk met PBA, en een EtG-PBA complex kom gedetecteerd worden met GC-MS/MS, echter met een extreem lage intensiteit. Voor de analyse van FAEEs is een methode gebruikt waarin ethyl palmitaat geëxtraheerd wordt uit het haar door middel van hexaan en DMSO gevolgd door incubatie gedurende de nacht, de monsters worden behandeld met SPE, gevolgd door verdamping en her-oplossing voor GC-MS/MS. Deze stappen tonen veel overeenkomsten met de geoptimaliseerde methode voor de detectie van EtG met LC-MS/MS, dus een gecombineerde monstervoorbewerking zou een mogelijkheid kunnen zijn in de toekomst. De methode was in staat om ethyl palmitaat te detecteren met GC-MS/MS.
4
Preface
The main goal of this master research project is to develop a method to detect ethanol metabolites in human hair. Hair analysis is a relatively simple, sufficient and sensitive method to monitor alcohol use over a long period of time. In the last few years, research in the field of hair analysis is rising, also within RIKILT Institute of Food Safety.
Two ethanol metabolites are studied, and a method to detect these metabolites with both LC-MS/MS and GC-LC-MS/MS was developed. Optimization and validation of the developed method should provide us with more insight in the possibilities of hair analysis for detection of alcohol (ab-)use. Chapter 1 of this thesis is an introduction into hair and hair analysis, ethanol metabolites and the techniques available for hair analysis. Also, the objectives of this project are described in this chapter. In Chapter 2, the initial method development for the analysis of one ethanol metabolite is given. Chapter 3 describes the optimization and validation of one of the methods developed. In Chapter 4 the initial method development of the other ethanol metabolite is described. Chapter 5 gives recommendations for future research, and general conclusions are stated in the last chapter, Chapter 6.
All the research described in this thesis is carried out at RIKILT Institute of Food Safety in Wageningen, under daily supervision of dr. W.F. Duvivier.
5
Table of Contents
List of abbreviations ... 2 Abstract ... 3 Preface ... 4 1. Introduction ... 61.1 Hair and hair analysis ... 6
1.2 Detection of alcohol (ab)-use ... 11
1.3 Analysis techniques ... 13
1.4 Aim of this project ... 18
2. Initial method development for the analysis of EtG in hair by GC-MS/MS and LC-MS/MS ... 19
2.1 Introduction ... 19
2.2 Materials and methods ... 19
2.3 Results and discussion ... 22
2.3.1 Method development for LC-MS/MS ... 22
2.3.2 Method development for GC-MS/MS ... 27
2.4 Practical obstacles ... 29
3. Optimization and validation of LC-MS/MS method for EtG in hair ... 32
3.1 Introduction ... 32
3.2 Materials and methods ... 33
3.3 Results and discussion ... 35
3.3.1 Optimization ... 35
3.3.2 Validation ... 38
3.3.3 Reference and real hair samples ... 42
4. Initial method development for the analysis of FAEEs in hair with GC-MS/MS ... 46
4.1 Introduction ... 46
4.2 Materials and methods ... 46
4.3 Results ... 47
5. Recommendations for future research ... 48
5.1 Sensitivity LC-MS/MS ... 48
5.2 Derivatization of EtG for GC-MS/MS and sensitivity ... 48
5.3 Sensitivity FAEEs with GC-MS/MS ... 48
5.4 Combining sample pre-treatment steps ... 49
6. General conclusions ... 50
References ... 52
6
1. Introduction
The use of drugs is a part of human culture for decades, and attempts to ban drugs have failed multiple times in history. Drugs are used on a social or recreational base, but also in an abusing way
1. This uncontrolled usage could have physical impact on a human being, essential organs could be
damaged, and it could play a role in crimes. So, it is important to know the effects of drugs on an individual, and his or hers general usage of drugs, from both a forensic and a clinical point of view 1.
Hair, blood, urine and oral fluid are all human materials used for toxicological analysis. Traditional biomarkers are mainly determined in blood (serum or plasma). Hair as a matrix for monitoring chronic consumption is a powerful evidential tool that has been relied upon in many criminal cases during the last decade 2. Hair analysis in forensic science is increasingly important, since the
difference between hair and the other materials is its substantially longer detection window in the order of weeks to several months, compared to that of body fluids (hours to days). Also, the high stability of hair samples and the fact that it can be stored at room temperature for a long time, allows hair analysis to be performed even centuries after growth. This makes it possible to investigate chronic or past consumption of drugs in a retrospective way 1. Hair as a testing matrix
could provide a historical overview of the individual’s exposure to drugs following chronic use, but it is also possible to detect drugs after a single exposure, which is a great advantage above other human materials 2. Hair collection is a simple and non-invasive procedure, and the potential for
manipulation of the results is low. Furthermore, there is little to no biosafety risk in handling the samples, in terms of transmission of diseases 1.
1.1 Hair and hair analysis
The entire skin of the human body is covered in hair, except the outer parts of the lips, feet soles and hand palms, and some surfaces of the external genitalia 3. Hair is used worldwide for diagnostic
purposes as testing drugs 4, and for the determination of metal concentrations related to sex and
age, for instance in poisoning cases 5. Hair can also provide information about the identity or lifestyle
of its owner, moreover it can be used in diagnosing health disorders 6.
1.1.1 Structure of the human hair
Hairs are derivatives of the epidermis, so the outermost layer of the human skin. Hairs are basically thin and flexible tubes for 90% made of dead and keratinized cells. Inside the skin hair follicles are alive, they grow into the dermis and hypodermic fat 7. Macroscopic, human hairs vary among ethnic
groups and individuals in length, colour, diameter and shape 4. In the human hair two different
structures could be separated; hair follicles inside the skin and the ones in the hair shaft visible on the skin. These structures are shown in Figure 1.
As can be seen in Figure 1B, a hair shaft is made up of a medulla in the centre, surrounded by the cortex and cuticle cells. The cortex represents most of the hair fibre composition and is an important part in describing the strength, but also the colour and the texture of human hair. It consists for roughly 50 to 60 percent of macro fibrils, which consist of cylinders of micro fibrils fixed into a matrix
8.
The cuticle forms the outer part of the hair shaft, its purpose is to cover the hair completely from root to tip to protect the cortex, and to protect the inner regions of the hair 8. It consists of flat
overlapping cells 6, each cell has a thickness of approximately 0.3 – 0.5 µm. The visible length of the
cuticle is roughly 50 µm and it is made up of several substructures, for instance the epicuticle, exocuticle and endocuticle, and the cell membrane complex, as shown in Figure 1B. The medulla is only present in really thick hair, and is made up of empty space 9.
7 Figure 1: A: Formation of hair in a follicle from matrix cells on the basement membrane to the mature hair shaft. B: Structure
and constituents of the human hair shaft. Reproduced from Ref. 10.
In Figure 1A, it is shown that hair growth is induced from the follicle. The follicle is the most
important growth structure of the hair. It is a skin organ with a bulb shape, which has two histological structures; the outer root sheath (ORS) and the inner root sheath (IRS).
The first structure, the ORS, is a part of the epidermis that can be described as some sort of room for stem cells, such as keratinocyte and melanocyte cells. It surrounds the IRS 7. A part of the ORS, on the
dermal site, consists of two orthogonal layers of collagen fibre, this layer is also known as the dermal sheet 11.
The IRS consists of three layers, of which one is the cuticle layer. The cuticle layer connects to the cuticle of the hair shaft, to make sure the hair shaft stays in the skin and the follicle will be
connected. The IRS separates the hair shaft from the ORS, and it produces hormones and keratins to give strength to the growing hair shaft 7.
The hair bulb can be described as the part of the follicle that produces the hair; the cells in the bottom of the hair divide just above the papilla 12. This division induces upward movements of the
cells, they are transported to the keratogenous zone in which they are synthesized and keratinized 13.
During the keratinization part, the proteins that are formed bind together, which results in long fibres 14.
1.1.2 Hair growth cycle
A hair is developed in a dynamic and cyclic process. Many hormones and cytokines coordinate the growth cycle of a hair. The duration of this cycle is not only dependent on the place of the body the hair is growing, but also on other factors as age of the individual, its diet, or environmental
alterations such as day length 8. Hormones and cytokines are the particles in the human hair that
instruct the follicle to change, and each hair on the human body could be in a different stage of the growth cycle at a time 15.
In the hair growth cycle, three phases could be noted: the anagen or growth phase, the catagen or transitional phase, and the telogen or resting phase (Figure 2). In these phases, the hair follicle undergoes stages of fast growth and formation of the hair shaft, followed by stages in which the hair follicle growth is inactive 16,17.
8 Figure 2: Different phases of hair cycle. Reproduced from Ref. 3
In the anagen phase, the hair follicle is in the active growth phase. The follicle grows until it reaches its typical onion shape, and a hair fibre is produced, as shown in Figure 2. This phase could be divided into six small stages; the first five stages of the anagen phase are called the proanagen phase, in which the dermal papilla grows downwards into the skin and differentiates to the hair shaft and the IRS. Cells in the new hair shaft are divided rapidly, causing growth of the hair. In the sixth phase, the metanagen phase, the epithelial hair bulb, which surrounds the dermal papilla, is formed, and the hair root embeds deeply into the dermis and sub cutis 3. The new hair shaft now appears on the
surface of the skin. The anagen phase could last for 3-6 years, and 80-90% of scalp hair is in this phase 7.
After the anagen phase comes to an end, the catagen phase starts. This phase is also called the regression phase, and in the beginning of this phase the hair shaft production is completed and the cell division stops. The hair follicle undergoes regression and is fully keratinized, this results in a reduction of the follicle of about one-sixth of the diameter. During the catagen phase, a keratinized structure of the hair looking like a brush, which is called the club hair, is formed on the base of the hair shaft 3. The catagen phase last for 2-3 weeks, and only 1-2% of scalp hair is in this phase 7.
After the catagen phase the telogen phase starts. In this phase, the hair goes into rest and stops growing completely. This period can last from a few weeks for eyelashes to a few months for scalp hair 3. The root of the hair is embedded in the follicle, and approximately 10-15% of all hairs are in
the telogen phase at any given moment 7. At the end of this stage, the hair can be considered dead
and falls out, which is called the exogen phase; a few weeks later, the hair follicle starts again in the anagen phase 14.
1.1.3 Collection of a scalp hair sample
In hair analysis, scalp hair is preferred since scalp hair has the fastest growth rate, and its follicles have the highest percentage in the anagen, growing, phase. Therefore, in this paragraph some features of scalp hair will be summarized, as well as the correct procedure for sampling.
The growth rate of scalp hair is reported as 1 cm per month, continuously for 3-5 years while the hairs are in the anagen phase 18. After the anagen phase, the hair growth stops and is followed by a
short catagen phase and a telogen phase of approximately 200 months. In this phase, the old hair is shed, and after the telogen phase the new hair starts to grow from the same follicle during the next anagen phase.
A scalp hair fibre is approximately 60 – 80 µm in diameter. It consists of a layer of flat and imbricated scales that point from root to tip of the hair 8.
9 sample must be cut from the posterior vertex region of the head (Figure 3); this is the region with less variation in the growth rate, and it must be cut as close as possible to the scalp 19. Locks of hair
with the thickness of a pencil must be collected, with a diameter of about 0.3 – 0.5 cm, and the root end must be clearly indicated 19. Two locks of hair must be collected to allow initial testing, followed
by confirmatory or re-testing of the sample if necessary 3. To obtain a timeline of drug use, the hair
can be cut into segments starting from the hair root, usually 1 – 3 cm pieces, and these segments are analysed individually, as shown in Figure 3.
Figure 3: Left: Vertex region of the scalp. Right: Retrospective timeline of drugs in hair. Reproduced from Ref. 1 1.1.4 Hair analysis
As a consequence of the advantages of hair analysis mentioned in the first part of this chapter, it is routinely used in forensic toxicological investigations, and it could be used in both post-mortem as alive cases. It is for instance used in drug-related crimes and deaths or child protection cases, and to monitor drug misuse in workplace drug testing or drug rehabilitation programs. The use of hair analysis is not limited to human samples, it is also used in veterinary control 1. More and more
laboratories offer hair testing, and the Society of Hair Testing (SoHT) recommends official guidelines for drug testing in hair 20. Instrument sensitivity and method validity are continuously improving, but
the main challenge is interpretation of the analytical results in evidence cases 1. This challenge is due
to many factors that affect the amount of drugs present in hair, its degradation over time and the different routes of incorporation.
Researchers were able to detect drugs in hair centuries after growth, for instance arsenic in hair collected from Napoleon Bonaparte 21 and cocaine in the hair of Peruvian mummies 22. However, the
concentrations of drugs in hair decrease over time because of natural wash out 23, and the stability is
dependent on the physicochemical properties of the hair 24.
1.1.5 Challenges in hair analysis
Contamination is the most important issue in hair analysis; the fact that analytical results could be interpreted as false negatives or false positives should be avoided 1. Alcohol markers are present in
several commercially available hair products. Traces of these compounds were detected in hair samples of people who frequently used the products 25. Decontamination of the hair sample before
analysis is thus of great importance in order to avoid false positives.
Weather conditions, such as exposure to sunlight, could reduce the concentration of several drugs in hair on a daily basis 26. Also, hair treatment could cause a reduction of the concentrations.
Shampooing was shown to not significantly affect the concentration of drugs in hair 27,28, but
blow-drying, curling or straightening can damage or destroy the cuticle, which provides routes for contamination, and it can be responsible for removal of incorporated drugs. Harsher cosmetic treatments could cause even more damage 19,27,29; all literature studies agree on the fact that
10 Bleaching created the largest decrease in drug concentrations. These factors influencing the
concentration, or the incorporation of drugs into the growing hair, need to be fully understand by the analyst, and the results of the hair analysis needs to be fully interpret in order to make the correct conclusion in criminal cases 1.
1.1.6 Mechanisms of drug incorporation
Up to the present day, the exact mechanism for the incorporation of drugs into hair is not fully understood, as well as the factors that could influence their stability. Drugs inside the body are believed to incorporate into the human hair in times of increased activity in metabolism and cell division during the anagen growth phase 31. However, there are three different ways for the
incorporation of drugs in hair. These three routes are shown in Figure 4, the first one is directly from the blood supply, the second route is from sebum or sweat bathing the hair, and the last one is from external contamination 32,33. It is unclear which route contributes to what extent to the drugs
incorporation in hair, but what is known is the fact that it varies from drug to drug 34.
Figure 4: Incorporation routes into the hair follicle. Reproduced from Ref. 1
Scientists proposed several models for the incorporation of drugs, trying to explain the different drugs profiles in human hair. The first model, which is the most simplistic one, describes the diffusion of drugs as a passive process, in which the diffusion is directly from the blood supply to the hair follicle. The diffusion process is passive, which means that the concentration of the drugs in hair are expected to be correlated to the drug concentrations in blood at the time of the analysis 35. However,
the metabolic profiles in hair and blood are different; in blood, detection of the parent drug is less common when compared to its primary metabolites, whilst in hair the parent drug is more commonly detected.
To explain the incorporation of drugs in growing hair in an endogenous way, the ‘biochemical concept’ was described by Baumgartner et al. 36. In this model, the ratio of the parent drug to its
metabolites in hair is explained. Other subjects in this model are an explanation of the
physiochemical properties of drugs and their dependency on the incorporation, incorporation of a drug in hair that is not pigmented and the dependency of drugs on hair pigmentation. Using this model, it became clear that the affinity and binding capacity of hair is different for each drug, including a different binding mechanism to the hair matrix 37. Another factor influencing the
11 incorporation or binding of the drugs to human hair is the lipophilicity; molecules that are neutral and lipophilic can penetrate the cell membrane easier than others, which results in a higher incorporation rate 38.
Examining all existing drugs is beyond the scope of this thesis, so in this thesis the focus will be on ethanol metabolites incorporated into hair. Different analytical techniques and approaches will be tested and compared, and factors influencing the incorporation of the metabolites in hair are examined.
1.2 Detection of alcohol (ab)-use
1.2.1 Ethanol metabolites in hair
For decades, the clinical and forensic research field has been searching for suitable markers for alcohol consumption. Due to its high volatility, alcohol detection in hair is difficult. However, metabolites of ethanol are detectable in hair 39.
An overview of possible ethanol markers in hair is given in Scheme 1. Ethanol markers could be direct or indirect. The direct markers are the markers still containing the carbon atoms of ethanol, the indirect ones are a product of a pathological change caused by ethanol in the metabolism. Overall, direct markers are preferable, because of the fact that indirect markers could also be formed via other pathologic pathways. The direct markers in hair are metabolites of ethanol which still contain the C2H5-group, or products which were formed during a reaction of acetaldehyde with physiologic
molecules 40.
In this thesis, the focus will therefore be on direct markers. The ethanol markers most studied in literature are ethyl glucuronide (EtG) and fatty acid ethyl esters (FAEEs). They have the advantage that they contain the ethyl group of ethanol which is a great indication that alcohol was responsible for the positive result 1. These markers will be discussed below in more detail including mechanisms
of drug incorporation in hair. The probability to find these ethanol markers in hair will be discussed regarding their formation and concentrations in hair, and some of their chemical properties will be discussed.
12 1.2.2 Ethyl glucuronide (EtG)
In 1952, EtG was described by Kamil et al. as a minor metabolite of ethanol 41. Investigation of EtG as
important marker of alcohol consumption was mainly performed by Skopp et al. 42. EtG is a direct
metabolite of ethanol, and is a water-soluble and non-volatile compound with high stability 43. EtG is
more sensitive and specific than most ethanol metabolites, because of the fact that EtG is only detectable in human hair after ethanol intake 44. In Scheme 2, the formation of EtG is shown. Ethanol
is conjugated with glucuronic acid via Uridine 5'-diphospho-glucuronosyl transferase
(UDP-glucuronosyl transferase). Glucuronic acid is located in the endoplasmic reticulum, in liver cells and in small amounts in for instance the lungs. Glucuronic acid is formed by an enzymatic hydroxylation in the presence of UDP-glucuronosyl transferase, and is necessary in the glucuronidation step of ethanol 45. The molecular formula of EtG is C
8H14O7, its monoisotopic molecular weight is 222.074
g/mol. The melting point or decomposition temperature is approximately 150 °C 1.
Scheme 2: Formation of ethyl glucuronide by conjugation of UDP-glucuronic acid and ethanol.
The incorporation rate of EtG in hair is low, because the molecule is an acid; EtG is a polar molecule with a pKa of 3.21. As a consequence, EtG is only detectable in hair in very small amounts and thus requires very powerful analytical techniques for detection 46. In order to establish universal
interpretation of analytical results for EtG in hair, the SoHT has set some cut-off values for
international use in 2010. A value of 30 pg/mg or more suggests chronic excessive consumption of alcohol, and a second cut-off value was established in the consensus of 2012 and has not change since; a value of ≥ 7 mg/mg strongly suggest repeated alcohol consumption. Another rule to remember is the fact that a positive result for EtG in hair overrules a negative result for FAEEs 18.
1.2.3 Fatty acid ethyl esters (FAEEs)
In Scheme 3, the formation of FAEEs is shown. As can be seen, ethanol undergoes esterification with free endogenous fatty acids (FA), triglycerides (TG), lipoproteins (LP) and phospholipids (PL) to form FAEEs in the presence of two enzymes, FAEE synthase and acyl-co-enzyme A(CoA)/ethanol O-acyl-transferase. FAEE synthase can be found in organs most likely to be damaged by alcohol
consumption; the pancreas and the liver 47. The FAEEs most likely to be formed are ethyl palmitate
and oleate 48, however, all esters are stable and can be found throughout the body 49.
In 1956 the process of the biotransformation of ethanol was described, in which FAEEs were shown to be products of this reaction 50. However, until 2001 nobody described FAEEs as possible markers
13 Scheme 3: Formation of fatty acid ethyl esters (FAEEs). FA: fatty acids. TG: triglycerides. LP: lipoproteins. PL: phospholipids. Preliminary studies revealed the fact that FAEEs can be found in human hair samples without exposure to alcohol, which results in false positive results. The exact reason is unclear, but it can be due to some physiological and pathological processes 1. Because ethanol is a by-product of some
physiological metabolism reactions in the human body, it is possible that some FAEEs could be detected inside the human body without actual alcohol intake 43.
In order to interpret the results correctly, the SoHT also established off values for FAEEs. The cut-off values were based on the sum of the four major FAEEs, which are ethyl myristate, palmitate, stearate and oleate. In 2016, the last consensus of the SoHT, a new cut-off value was established; the sum of the four FAEEs is no longer necessary, only the FAEE most likely to be formed, which is ethyl palmitate, is measured in the new consensus. Repeated alcohol consumption could be assumed with a cut-off concentration of 0.12 ng/mg of ethyl palmitate in the first 0-3 cm hair, or 0.15 ng/mg in the first 0-6 cm hair. A cut-off concentration of 0.35 ng/mg of ethyl palmitate in the 0-3 cm hair segment, and 0.45 ng/mg in the 0-6 cm segment indicates chronic excessive alcohol consumption 52.
1.3 Analysis techniques
After hair sample collection, the samples need further pre-treatment in terms of extraction of the targeted compounds from the matrix. Hair samples are cut into small fragments of typically 1 to 3 mm, in order to obtain a larger surface area. Attentively, the samples could be pulverized.
Compounds can be extracted from the hair matrix with organic solvents or aqueous solutions. Which specific solvent to use is dependent on the physicochemical properties of the compound 1. After
extraction, the sample needs a clean-up and concentration step like solid phase extraction (SPE), after which the sample can be analysed. Analytical methods for the determination of EtG and FAEEs in hair were described in several papers 53-56, and often these methods only differ in used
instruments or chromatographic parameters measured. Mass spectrometry (MS) has been widely accepted as the detection method of choice, and is therefore used in most hair analysis methods. Before compounds can be detected using MS, they most often are separated using gas
chromatography (GC) or liquid chromatography (LC). Analytical techniques used in hair analysis should contain some important properties; they must be suitable for explicit identification and quantitation of the drugs. Chromatographic techniques are powerful screening methods, by their ability to cover a wide range of analytes in one run 1. Due to the fact that hair contains low
concentrations of drugs and the sample sizes are small, developing a general method for drug screening in hair is difficult. This is mainly because every compound or drug class needs a specific
14 sample preparation. Therefore, the procedures described for drugs screening are typically developed for the analysis of a limited number of drugs at the same time 57,58.
1.3.1 Mass spectrometry (MS)
In hair analysis, the mass spectrometric analyser most commonly used is the quadrupole analyser. Movements of molecules in oscillating fields are stable, and a quadrupole analyser makes use of this fact to separate ions according to their mass to charge (m/z) ratios. Quadrupole analysers are made up of four rods of circular section, as shown in Figure 5. The rods must be perfectly parallel 59.
Figure 5: Schematic overview of a quadrupole analyser. Reproduced from Ref 60.
A positive ion entering the space between the rods will be drawn towards a negative rod, and the other way around. If the potential changes from sign before the ion comes in contact with the rod and discharges itself, the ion will change direction. An ion without resonance will not reach the detector in the end, an ion with resonance does 59.
Figure 6: Diagram of a triple quadrupole instrument. Reproduced from Ref. 61
In hair analysis, a triple quadrupole spectrometer is often used. In this spectrometer, three
quadrupoles are next to each other, as shown in Figure 6. The first and last quadrupole (Quad 1 and Quad 3) are mass analysers. The quadrupole in the centre, Quad 2, is used as a collision cell in which the dissociation takes place. In Quad 1 the precursor ion is selected and afterwards dissociated into product ions in Quad 2, followed by a mass scan in Quad 3 61.
In tandem mass spectrometry (MS/MS), a product ion scan could be used to determine the fragment ions of a compound. The product ion scan, also called daughter scan, selects a precursor or parent ion of a specific m/z ratio, and afterwards the product or daughter ions are determined based on collision-induced dissociation (CID) 59, as shown in Figure 7A.
15 Figure 7: Product ion scan (A) and selective reaction monitoring (B) in MS/MS. Reproduced from Ref. 59
Usually, for sensitive and targeted analysis, the detection mode selected reaction monitoring (SRM) is used. In this mode, ions with a specific m/z ratio are selected and isolated through the collision cell, as shown in Figure 7B. After CID only fragment ions with a certain m/z ratio are lead into the detector. This way, the ions selected by the first mass analyser are only detected if they produce this fragment ion with a certain m/z ratio, by a selected reaction. This means that no time is wasted on data that is not relevant for the analysis, so the sensitivity comes to a maximum 59.
SRM can also be performed by selecting multiple fragment ions before analysis, this technique is known as multiple reaction monitoring (MRM).
1.3.2 Gas chromatography – mass spectrometry (GC-MS)
One of the methods frequently used in hair analysis is gas chromatography coupled to mass spectrometry (GC-MS). GC makes use of a capillary column, this is one of the reasons it is a method with a high resolution 73. A GC-MS electron ionization (EI) mass spectrum has a really high specificity;
especially if SRM mode is used in the measurement. High specificity can also be achieved by using internal standards that are deuterated 73. All these features combined make GC-MS a very specific
and sensitive analysis technique for the detection of many drugs and its metabolites. If the chromatogram is divided into several time windows with different SRM masses, a large number of compounds can be measured at the same time in one run. Usually, the limit of detection (LOD) of GC-EI/MS is around 0.03 ng/mg for most drugs 1.
If GC-MS/MS is used, the sensitivity and specificity will be increased when compared to GC-MS. These advantages lead to the fact that GC-MS/MS is used more often nowadays in hair analysis 62-65.
The sensitivity of the method can also be increased by making use of headspace solid-phase micro extraction (HS-SPME), in which a derivatized sample is cleaned up by headspace extraction instead of regular SPE, increasing the analysed amount of sample by accumulation on the fibre 66.
1.3.3 Analysis of ethyl glucuronide using GC-MS
To perform GC-MS, it is essential that the compound of interest is volatile enough and stable at very high temperatures. If a compound contains free amino (NH2), hydroxyl (OH), or carboxyl (COOH)
groups, these groups are able to form hydrogen bonds inside the compound so that the volatility goes down and it does not interact with, for example, the column packing 67. These chemical
16 properties are reasons why it is necessary to derivatize the compound before performing GC-MS. Derivatization reactions used for GC-MS fall into three general reaction types namely; (i) alkylation of which the general process is esterification, (ii) acylation and (iii) silylation 68. Acylation and silylation
are the two processes used in derivatization of EtG samples. Silylation
N,N-bis(trimethyl-silyl)trifluoro-acetamide (BSTFA) is a derivatization agent used often for
derivatization of EtG samples. BSTFA is a silyl reagent, which reacts with both hydroxyl as carboxyl groups in a compound, in order to form trimethylsilyl ethers and trimethylsilyl esters respectively. BSTFA contains a trifluoroacetyl group, which makes a reaction with BSTFA fast and complete. The derivatives formed with BSTFA have high volatility and can be easily separated 69. In Scheme 4, the
derivatization of EtG with BSTFA is shown.
Scheme 4: Derivatization of EtG with BSTFA.
Due to the high volatility of derivatives formed with BSTFA, they will elute early, and the derivatives of BSTFA result in a low detector noise and pollution, since BSTFA is able to derivatize all OH-groups of EtG 70. No water must come in contact with the samples, as this will lead to hydrolysis of BSTFA,
which leads to the fact that none of the targeted analytes will undergo derivatization. Acylation
Another type of derivatization used often for EtG samples is acylation. It is a type of reaction in which an acyl group is introduced to an organic compound. In case of a carboxyl group, the reaction
involves the introduction of the acyl group and loss of the OH - group 71. The two most often used
acylation derivatization agents are fluorinated anhydrides such as pentafluoropropionic anhydride (PFPA) and heptafluorobutyric anhydride (HFBA). Fluorinated anhydrides such as PFPA and HFBA could react with amines, alcohols and phenols, producing stable derivatives with high volatility 72.
Scheme 5 and 6 show the derivatization of EtG with PFPA and HFBA, respectively.
17 Scheme 6: Derivatisation of EtG with HFBA.
Using GC-MS in chemical ionization mode, both positive and negative (GC-PCI/MS and GC-NCI/MS), could increase the sensitivity of the method to an LOD between 0.2 and 15 pg/mg for compounds as benzodiazepines 73, or derivatized,perfluorinated carboxylic acid anhydrides compounds such as EtG 74. Overall, for EtG, NCI delivers better results, but the consequence is the fact that the specificity will
be less due to missing fragmentation.
1.3.4 Analysis of fatty acid ethyl esters using GC-MS
FAEEs also need some sample pre-treatment before performing GC-MS, in the forms of SPE or headspace solid-phase micro-extraction (HS-SPME). GC-EI/MS is the method most often used in the analysis of FAEEs in hair, leading to an LOD as low as 0.01 ng/mg for the sum of all four FAEEs, though nowadays only ethyl palmitate is used. In Scheme 5, the formation of fragment ions of FAEEs is shown. Molecular ions are used as qualifiers, the quantifiers are the fragment ions from McLafferty rearrangement (m/z 88) or from β-cleavage (m/z 101) 1, as shown in Scheme 7.
Scheme 7: Formation of fragment ions of fatty acid ethyl esters (FAEEs). 1.3.5 Liquid chromatography – mass spectrometry (LC-MS)
Since GC-MS is dependent on derivatization of the compound to avoid issues with volatility and stability, a promising alternative is liquid chromatography – mass spectrometry (LC-MS) 1. However,
MS is still not used as often as GC-MS because of the fact that it is a very expensive technique. LC-MS has a lower chromatographic resolution compared to GC-LC-MS for several compound classes. This is the reason why LC is only possible if coupled to tandem MS/MS in hair analysis, in order to reach a high sensitivity 1. Liquid chromatography is normally used for compounds that are not volatile and
18 1.3.6 Analysis of ethyl glucuronide using LC-MS
Several publications describe the use of LC-MS/MS in hair analysis, including the detection of several drugs as methadone and metabolites 75, benzodiazepines 73, neuroleptics 76, sildenafil 77 and EtG78.
EtG is an anion, so the most often used ionization technique is electrospray ionization (ESI) in negative ion mode. The transitions most often used are m/z 221 – 75 as quantifier and m/z 221 – 85 as qualifier. If reversed phase columns are used, EtG is eluted in the first part of the chromatogram, with large interfering signals and matrix effects 79. Further optimization of the chromatography
showed that a high water content in the mobile phase was necessary to retain EtG, as well as post column addition of acetonitrile to enhance a low ESI yield 80. It was found that these difficulties can
be solved by making use of other columns, such as hydrophilic interaction liquid chromatography (HILIC) columns 79,81, or silica columns with a mobile phase containing a high content of acetonitrile 82. Many papers perform hair analysis by directly injecting the hair extract on the column, but in
order to avoid false negative results caused by ion suppression, a clean-up by SPE is recommended 1.
1.4 Aim of this project
Hair analysis is increasingly used over the past years in the detection of drugs in both forensic as clinical science. The detection window of hair is longer when compared to body fluids, and if a hair is analysed in a segmented way, it could give information about drug and alcohol abuse in a
retrospective way, so it is a sufficient and sensitive method to monitor drugs or alcohol use over a long period of time. The problem in the case of ethanol is the fact that ethanol itself cannot be detected in hair due to its high volatility, only its metabolites are incorporated into hair.
In this project, a method is developed for the detection of two ethanol metabolites in hair with LC- and GC-MS(/MS). A method for the analysis of one of the metabolites, EtG, was developed, optimized and validated. Validation of the method is important to make sure the method has a sufficient limit of detection and quantification (LOD and LOQ), accuracy and repeatability. Therefore, a validation protocol following the quality system of RIKILT is used, in which the used method is evaluated.
Also, the initial method development for the analysis of both EtG as well as FAEEs with GC-MS(/MS) was executed. Analysis of both metabolites at the same time is not possible, since the clean-up is different for each of them. It was tried to combine some steps of the analysis, so that it takes less time when compared to a separate analysis of each metabolite.
19
2. Initial method development for the analysis of EtG in hair by
GC-MS/MS and LC-GC-MS/MS
2.1 Introduction
First step in this project was to develop a method for the detection of EtG with GC-MS/MS and LC-MS/MS.
EtG is a compound which contains three hydroxyl (OH) and one carboxyl (COOH) groups; these functional groups are able to form hydrogen bonds inside a compound, and could affect its volatility and the tendency to interact with, for example, column packing material 82. These chemical
properties are reasons why it is necessary to derivatize the EtG samples before performing GC-MS/MS. As already mentioned in the introduction, acylation and silylation are the two processes used in derivatisation of EtG samples. Three derivatization agents could be used for these processes, namely BSTFA for the silylation and HFBA and PFPA for the acylation. To develop a method for the analysis of EtG with GC-MS/MS, first the derivatization step of the GC-MS/MS method needs to be optimized. The three derivatization agents discussed in the introduction were tested under different conditions, in order to find the optimal derivatization process.
Also, another derivatization agent, phenylboronic acid, was tested. Derivatization with phenylboronic acid (PBA) is done often in analysis that include glucose-like compounds, since it derivatizes two OH-groups at once 82. In Scheme 8, the derivatization of EtG with PBA is shown.
Scheme 8: Derivatization of EtG with PBA
Since GC-MS/MS is really dependent on these derivatization steps, another promising and upcoming technique in the analysis of EtG is LC-MS/MS. To avoid false positives in the analysis, the sample can be washed before analysis. Since decontamination is not a big issue in case of alcohol use, in this research the samples are not washed before analysis. To avoid false negatives the samples are usually cleaned-up by SPE, which is also done in this research 1. Different SPE methods were tested
and compared, as well as the reconstitution solvents and different LC-MS/MS systems.
2.2 Materials and methods
2.2.1 Hair samples
The analysis of EtG in hair was performed on blank hair samples provided by children or abstaining adults.
2.2.2 Chemicals and reagents
Ethyl glucuronide and ethyl glucuronide-d5 were purchased from Cerilliant (USA). Milli-Q water was obtained from a Millipore system of Merck (Germany). SPE Oasis MAX 3 cc cartridges were
20 (Sweden). HFBA, PFPA, PBA, ammonium formate and sodium sulfate were purchased from Sigma-Aldrich (Germany). BSTFA was purchased from Grace Davison, Discovery Science (USA). Methanol, n-hexane, ethyl acetate, heptane, isooctane, acetone, HCl and acetonitrile were purchased from Actu-All chemicals (Netherlands). Formic acid and 25% ammonia solution were purchased from Merck (Germany).
2.2.3 Sample preparation and clean-up procedure for LC-MS
Approximately 50 mg of the blank hair samples were exactly weighed and cut into pieces (1-2 mm) with scissors. The hair samples were then spiked with different concentrations of the standard solution of EtG, which was evaporated to air for 30 min. 20 µL of internal standard solution (EtG-d5, 10 ppm) and 2 mL of Milli-Q water were added, and the extraction was performed by 4h
ultrasonication at 50 °C. After centrifugation for 7 min at 4000 rpm, three different clean-up methods were performed:
(i) Solid phase extraction using Oasis MAX 3cc cartridges (Waters, Ireland) were conditioned with 2 mL of methanol and 2 mL of Milli-Q water. The extract was transferred on the SPE column and the columns were dried for 5 min. The samples were washed with 2 mL of Milli-Q/5% ammonia solution and 2 mL of methanol, and dried for 10 min. EtG was eluted with 2 mL of methanol/2% formic acid in glass tubes and evaporated to dryness under a nitrogen stream at approximately 50 °C.
(ii) Solid phase extraction using Oasis MAX 3cc cartridges (Waters, Ireland) were conditioned with 2 mL of methanol and 2 mL of Milli-Q water. The extract was transferred on the SPE column and the columns were dried for 5 min. The samples were washed with 2 mL of n-hexane, and dried for 10 min. EtG was eluted with 2 mL of methanol/2% formic acid in glass tubes and evaporated to dryness under a nitrogen stream at approximately 50 °C.
(iii) Solid phase extraction using Isolute NH2 (aminopropyl) cartridges (Biotage, Sweden) were
conditioned with 3 mL of methanol, 3 mL of Milli-Q water and 3 mL of acetonitrile. Care was taken to ensure that the columns did not dry between these conditioning steps. The extract was transferred on the SPE column and the columns were dried for 5 min. The samples were washed with 3 mL of Milli-Q/5% ammonia solution, and dried for 15 min under a strong vacuum. EtG was eluted with 3 mL of methanol/2% ammonia solution in glass tubes and evaporated to dryness under a nitrogen stream at approximately 50 °C.
Three reconstitution solvents were tested, namely methanol, acetonitrile and mobile phase A (see below). All residues were reconstituted in 200 µL and 10 µL were injected for measurement by LC-MS.
2.2.4 Instrumentation for LC-MS/MS
Liquid-chromatography was performed on a Waters Acquity UPLC system, and a Zorbax Eclipse XDB-C8 column (5 µm, 4.6 x 150 mm) with the mobile phases A = 5mM NH4-formate and 0.01% formic
acid in Milli-Q water and B = acetonitrile, with the following gradient: 0-3 min 100% solvent A, linear to 90% B in 2 min for 3 min, linear to 100% A in 0.10 min for 2 min. The flow rate was 0.5 mL/min. The detection was performed on Micromass Quattro Ultima Pt, Waters Xevo TQ-S and AB-SCIEX QTRAP-6500 mass spectrometer instruments with ESI in negative mode using the following transitions for EtG: m/z 221 – 95, 221 – 85, 221 – 75, 221 – 57, and for EtG-d5: m/z 226 – 85, 226 – 75.
2.2.5 Method calibration
A standard calibration curve was obtained by preparing EtG solutions in mobile phase A in the following concentrations: 0, 1, 2.5, 5, 12.5, 25, 50, 125, 250, 500, 1250, 2500, 5000 and 9000 pg/µL and 100 µL internal standard (EtG-d5, 10 ppm).
21 2.2.6 Derivatization procedure for GC-MS
Two derivatization procedures were tested, one procedure for derivatization with PFPA, HFBA and BSTFA, and a second procedure for derivatization with phenylboronic acid (PBA).
(i) An EtG solution of 10.000 pg/µL (10 ppm) was prepared in methanol, and 50 µL were evaporated under a nitrogen stream at 60 °C in MS vials. In order to optimize the derivatization of EtG, several derivatization agents were tested. Standard solutions of these agents were added to the vials as mentioned in Table 2. The derivatization time and temperature varied according to Table 2. 2 µL of some of the samples were injected into the GC-MS/MS after the derivatization, others were evaporated under a nitrogen stream at 60 °C and reconstituted as mentioned in the table, before injection of 2 µL into the GC-MS/MS.
# Derivatization agent Derivatization temp./time
Reconstitution Reconstitution solvent
1 50 µL HFBA/acetone (1:4) 1 h, 60°C + 50 µL isooctane
2 50 µL BSTFA/acetone (1:4) 1 h, 60°C + 50 µL isooctane
3 20 µL ethyl acetate/BSTFA (1:1) 20 min, 80°C - -
4 50 µL BSTFA/methanol (1:4) 30 min, 80°C + 50 µL isooctane
5 100 µL PFPA 30 min, 80°C + 50 µL heptane
6 100 µL PFPA/methanol (1:5) 30 min, 80°C + 50 µL heptane
7 100 µL PFPA/methanol (1:5) 30 min, 80°C + 50 µL ethyl acetate
8 20 µL ethyl acetate/HFBA (1:1) 20 min, 80°C - -
9 20 µL ethyl acetate/PFPA (1:1) 20 min, 80°C - -
10 20 µL ethyl acetate/HFBA (1:1) 1 h, 80°C - -
11 20 µL ethyl acetate/PFPA (1:1) 1 h, 80°C - -
12 20 µL ethyl acetate/BSTFA (1:1) 1 h, 80°C - -
13 20 µL ethyl acetate/HFBA (1:1) 20 min, 80°C + 50 µL isooctane
14 20 µL ethyl acetate/PFPA (1:1) 20 min, 80°C + 50 µL isooctane
15 20 µL ethyl acetate/BSTFA (1:1) 20 min, 80°C + 50 µL isooctane
(ii) A PBA solution was prepared by weighing 1g of PBA in a 5 mL vial, 4 mL of an acetone/Milli-Q water mixture (19/1, v/v) was added and the mixture was shaken vigorously. A sodium sulfate solution was prepared by weighing 0.25g of sodium sulfate in a 5 mL vial, 5 mL of Milli-Q water was added and the mixture was placed in an ultrasonic bath to ensure the reagent was completely dissolved. EtG solutions of 10.000 pg/µL (10 ppm) were prepared in methanol, and 500 µL was put in a 5 mL vial. 2 mL of the sodium sulfate solution was added, followed by addition of 250 µL of the PBA solution. The mixture was vortexed for 10s and incubated in an ultrasonic bath at room temperature for 5min. 250 µL of a HCL (37%)/Milli-Q water solution (1/9, v/v) was added. The PBA derivatives of EtG were extracted by addition of 1 mL of n-heptane, vortexing the mixture for 10s and transfer of the upper layer to an empty glass test tube. The extraction was repeated with 1 mL of n-heptane and the two extracts were combined and evaporated to dryness under a nitrogen stream at 50 °C. The residue was dissolved in 400 µL of n-heptane and the supernatant was transferred to an empty GC-MS vial, 2µL were injected in GC-GC-MS/GC-MS in both full scan as well as SIM mode.
2.2.7 Instrumentation for GC-MS/MS
A Varian CP-3800 gas chromatography – 1200L Quadrupole MS/MS system operating in EI mode was used for analysis. Compounds were separated on a fused silica column (Agilent, DB-35MS) with a (35%-phenyl)-methylpolysiloxane stationary phase (30 m length x 0.25 mm I.D. x 0.25 µm film thickness). The carrier gas was helium with a constant flow of 1 mL/min. The injector temperature was 250 °C, the initial column oven temperature of 75 °C was kept for 1 min and subsequently
22 increased at a rate of 20°C/min to 300 °C. The GC-MS/MS analysis was performed in full scan mode, as well as SIM mode for derivatives with BSTFA and PBA. The precursor ions of EtG (261) and EtG-d5 (266) were selected for BSTFA, before fragmentation in the collision cell at an energy of 10.0 eV, giving product ions of 143 for both EtG and EtG-d5. The precursor ions of EtG (308) and EtG-d5 (313) were selected for PBA.
2.2.8 Data processing
Data was processed using the software MassLynx V4.1 and MS Workstation V6.9.2.
2.3 Results and discussion
2.3.1 Method development for LC-MS/MS
Comparison of clean-up methods
To begin with, three different clean-up methods described in literature 83-85 were tested, as explained
in Chapter 2.2. In the method in which Isolute NH2 (aminopropyl) cartridges were used, no EtG was
detected in all chromatograms. The SPE cartridges are anion exchange cartridges, which should work for an acidic compound as EtG, so the problem has to originate elsewhere. A possibility is that all sample is already removed from the column by the washing solvent, which was a Milli-Q/5% ammonia solution. Unfortunately, the washing solvents were not collected in glass tubes, so this needs some further research later on in the project.
Clean-up method 1 and 2 made use of the same SPE cartridges, only the washing steps were different. The suspicious part about method 1 is the fact that one of the washing solvents is
methanol, in which many compounds dissolve easily. Therefore, during the analysis, also the washing step with methanol was collected in glass tubes, evaporated under a nitrogen stream and
reconstituted. In Figure 8, the chromatograms of a 5000 pg/µL sample treated with method 1, both of the washing step with methanol as of the reconstitution, are shown. The intensities of the peaks are in the right corner.
Figure 8: Chromatograms of EtG samples in a concentration of 5000 pg/µL with clean-up method 1. A: washing step with methanol. B: reconstituted samples. Chromatograms were recorded with a Micromass Quattro Ultima Pt
23 As can be seen in the figure, the clean-up method does not work properly. In the washing-step (Figure 8A) EtG is detected, the amount of EtG that gets lost in the analysis is even bigger than the amount detected after reconstitution (Figure 8B), which was the case for all concentrations measured.
In Figure 9 the chromatograms of the clean-up method 2 are shown, with the intensity of the signal in the right corner. As can be seen in this Figure, this method shows slightly better results, but the intensities still don’t show a relation as expected. For example, the intensity of the peak for the 2000 pg/µL sample (Figure 9D) should be twice as high as the 1000 pg/µL sample (Figure 9C), but the intensities are almost the same; even the signal for the 1000 pg/µL signal is slightly higher. Note that the chromatograms of 5000 and 10000 pg/µL (Figure 9E and 9F) are recorded on a different day, which could be the reason why the chromatograms are slightly offset. Still it is unusual, since EtG normally elutes after 2.2 min, not after 2.5 min. Therefore, it can be said the method does not work properly, since there is a variance in peak elution times and intensities of the peaks in the
chromatogram.
Figure 9: Chromatograms of EtG samples in different concentrations with clean-up method 2. Chromatograms were recorded with a Micromass Quattro Ultima Pt mass-spectrometer.
Overall, none of the clean-up methods worked perfectly. Method 3 was not able to detect EtG signals, the Oasis MAX 3cc SPE cartridges did show some EtG signals in the chromatograms, but the results were not as expected. Overall, method 2 in total showed the best results, but all methods need further improvement in order to work sufficiently. Also, the intensities of the peaks are really low. For this reason, after these experiments, the MS/MS-spectrometer was replaced by two other, more sensitive machines, namely Waters Xevo TQ-S and SCIEX QTRAP-6500.
24 Comparison of reconstitution solvents
The EtG samples were reconstituted in methanol, acetonitrile and mobile phase A in order to compare peak shapes and intensity of the peaks. In Figure 10, the chromatograms of these samples are shown, with the intensities of the peaks in the right corner.
Figure 10: Chromatograms of EtG samples reconstituted in 200 µL of respectively A: acetonitrile, B: mobile phase A and C: methanol. Chromatograms were recorded with a Waters Xevo TQ-S mass spectrometer.
As can be seen in the figure, acetonitrile (Figure 10A) is not sufficient as reconstitution solvent, since no peak appeared in the chromatogram. This is not as expected, since acetonitrile is also a polar solvent such as methanol and water (mobile phase A). A solution could be the fact that the samples in acetronitrile needed a longer time to reconstitute as in methanol and water (mobile phase A), since the reconstitution times were all equal it could be that EtG did not have time to dissolve in the acetonitrile. Both methanol (Figure 10C) and mobile phase A (Figure 10B) are sufficient as
reconstitution solvents, however the peak shape of the sample reconstituted in methanol is tailing. The peak of the sample in mobile phase A (Figure 10B) is sharper, and the intensity of the peak is slightly higher, though in the same order of magnitude. Since this peak looked really sufficient, it was decided to continue with mobile phase A as reconstitution solvent.
Different amounts of reconstitution solvents were also tested, 100, 200 and 400 µL, but the
intensities and shapes of the peaks remained the same. Based on these observations, mobile phase A was chosen as reconstitution solvent, in an amount of 200 µL, to make sure several measurements could be performed with LC-MS/MS.
Calibration
In Figure 11A, the calibration curves for EtG are shown. As can be seen, the coefficient of
determination R2 for the method with Waters Xevo TQ-S is 0.9755, the R2 for the method with SCIEX
25 Figure 11: Calibration curves of EtG corrected with EtG-d5, recorded with blue: SCIEX QTRAP-6500 mass spectrometer, orange: Xevo TQ-S mass spectrometer. A: calibration curve 0 to 9 ppm. B: calibration curve zoomed in, 0 to 0.0125 ppm. In Figure 11B, the same calibration curves are shown, but zoomed in on the smaller concentrations of EtG. As can be seen, the values for R2 become a little smaller, since the first measurements (0 to
0.0025 ppm) give quite similar results, especially when a Xevo TQ-S mass spectrometer is used. However, since the R2 value and peak areas are still higher with the SCIEX QTRAP-6500 mass
spectrometer, it was chosen to use this machine in further experiments. Control of evaporation-step in sample preparation and clean-up
Since all clean-up methods with SPE did not work properly, this experiment was done without SPE. The same sample preparation steps as described in Chapter 2.2.3 were followed. After
centrifugation, 100 µL of the supernatant was collected into vials for LC-MS/MS. The rest of the supernatant was transferred to glass tubes and evaporated under a nitrogen stream at 40 °C, and reconstituted in 200 µL of mobile phase A. The resulting peak areas are shown in Figure 12.
Figure 12: Peak areas of EtG for 100µl supernatant (blue) and evaporation and reconstitution in 200 µl mobile phase A (orange).
As can be seen in the Figure, the peak areas for the method in which 100 µL of the supernatant was collected and immediately measured are much smaller than for the method with evaporation and reconstitution in 200 µL of mobile phase A. This could be due to the fact that the samples in 100µL supernatant, in which the hair matrix is still present in the tubes while collecting the sample, are less concentrated than the reconstituted samples. This leads to the fact that evaporation after clean-up is a necessary step in the analysis.
26 The peak areas for all samples are according to the concentrations of EtG in the sample, they become higher with higher concentrations. However, also the blank sample with a concentration of 0 ppm gave a peak at 2.1min, as can be seen in Figure 13. Since all samples, including the blank one, contain the internal standard EtG-d5, the internal standard was examined to check for contamination with EtG.
Figure 13: Chromatogram of blank hair sample with internal standard EtG-d5. Chromatogram was recorded with a SCIEX QTRAP-6500 mass spectrometer.
Testing internal standard EtG-d5 and evaluation of the blank hair samples
In order to examine for contamination, a 10.000 pg/μL (10 ppm) solution of the internal standard in mobile phase A was analysed.
As can be seen in Figure 14, a chromatogram of a 10.000 pg/μL (10 ppm) internal standard solution EtG-d5 shows a peak of EtG between 2.1 and 2.2 min. The intensities are quite low, in the same range as was seen in Figure 13 of a blank hair sample containing only the internal standard solution EtG-d5, so it is possible that the internal standard EtG-d5 contains a small amount of EtG, leading to these peaks in the chromatogram.
Figure 14: Chromatograms of 10 ppm EtG-d5 in methanol. A: MRM of 2 channels (EtG-d5). B: MRM of 4 channels (EtG). Chromatograms were recorded with a SCIEX QTRAP-6500 mass spectrometer.
Another possibility is the fact that the blank hair samples are in fact not really blank for EtG. Another possibility could be the fact that the blank hair samples are not blank for EtG. To test this, the measurements are repeated with cow hair, which should be blank for EtG. In Figure 15, the chromatograms for the blank cow hair samples are shown.
27 Figure 15: Chromatogram of a blank cow hair sample. Chromatogram was recorded with a SCIEX QTRAP-6500 mass
spectrometer.
As can be seen in the Figure, the blank cow hair sample shows a small peak at 2.1min, which could be originating from EtG. The intensity of the peak is in the same range as was seen in Figures 13 and 14. This leads to the fact that the internal standard, EtG-d5, contains a small amount of EtG. However, because it is such a small amount, it will still be used in further experiments, but it will be kept in mind while interpreting the results that the results are in fact slightly lower than shown in the chromatogram.
2.3.2 Method development for GC-MS/MS
Since the LC-MS method for detection of EtG in hair did not work sufficiently enough, next step in the project was to develop a method to analyse EtG in hair with GC-MS/MS.
At first, reaction 1 to 7 in Table 2 were performed and tested. Unfortunately, analysis of the derivatization products did not show any signals for EtG. Reaction number 3 showed some peaks in the chromatogram, but the spectra belonging to these peaks were not originating from EtG. Since some signal was visible in reaction number 3, this reaction was investigated further with reactions 8 to 15. The three derivatization reagents were tested in a mixture with ethyl acetate, and the temperature/time varied. The last three reactions (13-15) were evaporated after derivatization and reconstituted in isooctane. Unfortunately, none of the reactions showed any signal for EtG, including a repetition of reaction number 3 on a different day.
Figure 16: Chromatogram of boldenone derivatized with HFBA. Chromatogram was recorded with a Varian mass spectrometer.
To check if the derivatization method used works, a control sample was also derivatized at the same time. This control sample contained boldenone in acetone (1:4), and was derivatized with HFBA. This control sample did show a peak in the GC-MS chromatogram recorded in full-scan mode as shown in Figure 16, so there must be another reason why the derivatization was not working properly. The
28 fact that none of the derivatization reactions worked, could be due to the fact that there might be some water present in the samples or the solvents used; in this case the derivatization agent is hydrolysed, so no derivatization takes places at all.
Lastly, another derivatization method was tested, method (ii) in Chapter 2.2.6. In Figure 17, the chromatogram and spectrum of this derivatization, recorded in SIM mode in the range from 305-315 m/z, is shown.
Figure 17: Chromatogram (A) of 10 ppm EtG solution derivatized with PBA, and spectrum (B) of EtG-PBA. Chromatogram was recorded with a Varian mass spectrometer.
As can be seen in this Figure, several peaks show up. The first peak, eluting at 4min., could be originating from the EtG-PBA complex, since this complex has got an m/z ratio of 308.
When derivatizing with PBA, PBA can also react with itself to form a large complex, as shown in Figure 18. Its mass, an m/z ratio of 311, shows a peak eluting at 5min.
Figure 18: Chromatogram (A) of 10 ppm EtG solution derivatized with PBA, and spectrum (B) of PBA complex formed during derivatization. Chromatogram was recorded with a Varian mass spectrometer.
In Figure 19A, the full chromatogram is once again shown, with the filtered chromatogram for m/z 311 in Figure 19B, and the filtered chromatogram for m/z 308 in Figure 19C. As can be seen, the intensity in Figure 19C is extremely low, and the largest signal in the chromatogram originated from m/z 311, which originates from the PBA complex formed by a reaction of PBA with itself. Since this signal is much larger than the signal originating from EtG-PBA, it is possible that PBA reacts quickly with itself, leaving less PBA for derivatization of EtG.
29 Figure 19: (A) Full chromatogram and chromatogram of (B) m/z 311 of the PBA complex and (C) m/z 308 of EtG-PBA. Due to time limitations, unfortunately the derivatization of EtG was not investigated any further, but the derivatization of EtG with PBA is promising.
2.4 Practical obstacles
2.4.1 Evaluation of the transitions used in the LC-MS/MS method
In Figure 20, the mass spectrum of EtG with its transitions is shown. In the LC-MS/MS measurements four transitions for EtG were used, namely m/z 221 – 95, 221 – 85, 221 – 75 and 221 – 57, because they showed the highest intensity peaks in the mass spectrum, as shown in the daughter ion scan chromatogram in Attachment 1.
30 In the EtG measurements, the signal originating from the transition m/z 221 – 95 and 221 – 57 usually does not contribute to the signal as much as the transitions m/z 221 – 85 and 221 – 75, as shown in Figure 21.
Figure 21: TIC and chromatograms of four transitions of EtG (10 ppm). Chromatograms were recorded with a Waters Xevo TQ-S mass spectrometer.
To check if a peak has sufficient selectivity, some measurements could be done according to literature 86. Unfortunately, the article describes a method only for positive electrospray ionization
(ESI) mode, while these measurements are performed in negative ion mode. However, to give an idea of the ‘identification power’ of the method used, the selectivity of the transitions used is calculated using the method described in the article.
The selectivity of a method could be measured according to the probability that a compound shows the same characteristics in LC-MS/MS as the compound of interest. This probability is called the probability of interference, P(I), and can be calculated with the following formula:
𝑃(𝐼) = 𝑃(𝑀𝑝𝑟) ∗ max (𝑃(𝑀𝑝𝑑1), 𝑃(𝑀𝑛𝑙1)) ∗ max (𝑃(𝑀𝑝𝑑2), 𝑃(𝑀𝑛𝑙2)) ∗ 𝑃(𝑅𝑇)
In this formula P(Mpr) is the probability of the precursor ion to occur, P(Mpd) the probability of the
product ions of transition 1 and 2 to occur, P(Mnl) the probability of a neutral loss of transitions 1 and
2 to occur, and P(RT) the probability of a specific retention time to occur in the chromatographic system. In case of a high-resolution reversed phase LC system, P(RT) = 0.2. Max[P(Mpd), P(Mnl)]
represents the maximum value of either P(Mpd) or P(Mnl) in a worst-case scenario. All parameters can
be determined according to mathematical models present in the paper 86.
P(I) is calculated, and compared to a criterion previously calculated in the paper to determine if a method is sufficient or insufficient. If P(I) < 2*10-7, the method is sufficiently selective.
