Mass Spectrometry Imaging of Small Molecules in Zebra Finch Songbirds

(1)

Mass Spectrometry Imaging

of Small Molecules in Zebra

Finch Songbirds

Distribution of Neurotransmitters in

Different Ontogeny Stages of the Brain

Wesley Suntjens1,2, Nina Ogrinc Potočnik1, Pascal Hommelberg2, and Ron Heeren1

1

Maastricht Multimodal Molecular Imaging (M4I) Institute, Division of Imaging Mass Spectrometry, Maastricht University, Universiteitssingel 50, 6229 HX Maastricht, The Netherlands.

2

Avans University of Applied Sciences, Academy of Technology for Health and Environment (ATGM), Biology & Medical Laboratory Research, Lovensdijkstraat 61, 4818 AJ Breda, The Netherlands.

(2)

Information Page I

Molecules in Zebra Finch Songbirds

Distribution of Neurotransmitters in Different Ontogeny Stages of the Brain

Information Page

Student:

Wesley Suntjens wesley.suntjens@gmail.com

(Student No.: 2061622) Tel. +31 (0)6 53 13 9627

Bachelor Thesis:

Version 6

Date August 2nd 2016

Start Date Internship February 1st 2016 End Date Internship July 1st 2016

Internship at Company:

Maastricht Multimodal Molecular Imaging (M4I) Institute Division of Imaging Mass Spectrometry (IMS)

Maastricht University Universiteitssingel 50 6229 HX MAASTRICHT Tel. +31 (0)43 388 2222

Supervisors:

Dr. Nina Ogrinc Potočnik n.ogrinc@maastrichtuniversity.nl Prof. Dr. Ron Heeren r.heeren@maastrichtuniversity.nl

Educational Institute:

Avans University of Applied Sciences

Academy of Technology for Health and Environment (ATGM) Biology & Medical Sciences (BMO)

Lovensdijkstraat 61-63 4818 AJ BREDA Tel. +31 (0)76 52 5050

Supervisor:

Dr. Pascal Hommelberg pph.hommelberg@avans.nl

Sources Cover Page Image: http://www.mdpi.com/2218-1989/4/2/319 and http://www.dierennieuws.nl/vogels/is-het-een-koppel-of-toch-niet. Edited by Anne ten Hoeve, student Graphic Design at Regional Education Centre Gilde in Roermond, the Netherlands.

(3)

Preface II

Preface

Before you lies the thesis “Mass Spectrometry Imaging of Small Molecules in Zebra Finch

Songbirds; Distribution of Neurotransmitters in Different Ontogeny Stages of the Brain”, a study to

localize, as the title states, the distribution of neurotransmitters in different ontogeny stages of the Zebra Finch (Taeniopygia guttata) brain. It has been written to fulfil the graduation requirements of the Biology & Medical Laboratory Research program at Avans University of Applied Sciences. I was engaged in research and writing this thesis from February to July 2016 at Maastricht University, Division of Imaging Mass Spectrometry (IMS).

My supervisor at Maastricht University, Dr. Nina Ogrinc Potočnik, provided me with a wide variety of projects to choose from. I knew from the start I had a lot to learn about Mass Spectromerty, since this was never a subject in school. However, with the help I received from my supervisor, co-workers and self study I believe we retrieved some interesting results that helped me answer the research question given to me. Fortunately, both my supervisor and my tutor from Avans University, Dr. Pascal Hommelberg, were always available to answer my questions.

I would like to thank my supervisors Dr. Ogrinc Potočnik and Prof. Dr. Heeren for the opportunity to help in this research and their excellent guidance and support during this process. Also I would like to thank my tutor from Avans University Dr. Hommelberg for always supporting me in my research and answering my questions as soon as possible.

Furthermore, I would like to thank all my co-workers at the M4I institute. You all have given me wonderful time within the IMS division. I also want to thank my friends for all your interest in the research. A special thanks goes out to my parents Ton Suntjens and Monique Suntjens-Janssen and my sister Janou Suntjens: thank you for all your admiration, kind words and wise words of advice. If I ever doubted myself you made sure I stayed motivated.

I hope you enjoy reading my thesis.

Wesley Suntjens

(4)

Abstract III

Abstract

Mass Spectrometry Imaging (MSI) is a useful tool for direct in-situ analysis of tissue sections. This label-free technique has been applied in different fields due to its advantages over other imaging techniques such as, immunohistochemistry and autoradiography, by avoiding time-consuming extraction, purification or separation steps, which have the potential for producing artefacts. The MSI is capable of providing simultaneous distribution and identification of hundreds of different (unknown) compounds in a single imaging experiment (Mascini & Heeren, 2012). However, the detection of small molecules, such as neurotransmitters, by matrix-assisted laser desorption/ionization (MALDI) mass spectrometry has been difficult due to their low ionization efficiency and isobaric compounds. By derivatization, a technique to alter the chemical and physical characteristics of a sample, these neurotransmitters can be detected more easily, since the detection sensitivity increases during MALDI MS. Therefore, in order to detect the neurotransmitters in question the samples were subjected to different derivatization reagents and matrices. In order to achieve best spatial localization and coverage on tissue we also tested different application techniques such as spraying with the SunCollect or temperature controlled spraying with HTX TM-Sprayer and matrix deposition with sublimation.

In-solution a mixtures of the neurotransmitters γ-aminobutyric acid (GABA), glutamine, L-methionine, L-phenylalanine and L-tyrosine were subjected to 4-hydroxy-3-methoxycinnamaldehyde (CA) and 2,4-diphenyl-pyranylium tetrafluoroborate (DPP-TFB) derivatization. These solutions were then incubated and the two matrices were applied on a target plate prior to analysis with MALDI Synapt G2Si ETD. On-tissue derivatization was applied with both CA and DPP-TFB. The DHB matrix was applied with three different matrix applications, as previously described. For high mass resolution analysis we used the SolariX FT-ICR. One of the DPP-TFB derivatized slides was imaged without the use of any matrix, since DPP-TFB has potentially self-assisted laser desorption ionization capabilities.

The detection of small molecules was increased after applying derivatization reagents and DHB matrix in solution. After the application of matrix on tissue it became clear that both the SunCollect and Sublimation device increased the chance of delocalization of the neurotransmitters. With the HTX TM-Sprayer the derivatization reagents and matrix were applied for the high mass resolution analysis on the SolariX FT-ICR. In the research of neurotransmitter distribution in different ontogeny stages of the zebra finch brain DPP-TFB was used without the use of DHB matrix. The images showed a significant change in distribution of the neurotransmitters, although further measurements need to be done in order to show biological significance and to fully optimize this method. Within M4I multiple studies on neurodegeneration are performed that could benefit from this method. This method could potentially be used in other studies such as the distribution of neurotransmitters in brain tissue sections of healthy wild type vs. APP/PS1 mice, overexpressing amyloidosis which is a major hallmark in Alzheimer’s disease.

(5)

Samenvatting IV

Samenvatting

Massa Spectrometrie Imaging (MSI) is het handige tool voor de in-situ analyse van weefsel secties. Deze label-vrije techniek wordt toegepast in verschillende werkvelden, doordat het voordelen heeft die andere technieken, zoals immunohistochemie en autoradiografie, niet hebben door tijdrovende extractie, purificatie en scheidingstechnieken te vermeiden. Door MSI is het mogelijk om de distributie en identificatie van honderden verschillende stoffen the bepalen in één enkel imaging experiment (Mascini & Heeren, 2012). De detectie van kleine moleculen, zoals neurotransmitters, door matrix-assisted laser desorption/ionization (MALDI) massa spectrometrie is zeer lastig door lage ionizatie efficiëntie en isobarische stoffen. Door middel van derivatizatie, een techniek om de chemische en fysische eigenschappen van een sample aan te tasten, kunnen deze neurotransmitters gemakkelijker gedetecteerd worden, omdat de detectie sensitiviteit dan verhoogd wordt tijdens MALDI MS. Om deze verschillende neurotransmitters te detecteren werden de samples onderworpen aan verschillende derivatizatie reagentia en matrices. Om de beste ruimtelijke localisatie en dekking op het weefsel te krijgen hebben we ook verschillende applicatie technieken, zoals het besproeien met de SunCollect of temperatuur gecontroleerd sproeien met de HTX TM-Sprayer en matrix afzetting met sublimatie, getest.

In oplossing werden mengsels van neurotransmitters γ-aminobutyric acid (GABA), glutamine, L-methionine, L-phenylalanine en L-tyrosine gederivatizeerd met 4-hydroxy-3-methoxy-cinnamaldehyde (CA) en 2,4-diphenyl-pyranylium tetrafluoroborate (DPP-TFB). Na incubatie van deze oplossingen werden CHCA enDHB matrix toegevoegd vlak voor de analyse met de MALDI Synapt G2Si ETD. Tijdens de derivatizatie van weefsels met zowel CA als DPP-TFB. De DHB matrix was toegevoegd met de drie verschillende matrix applicaties, zoals eerder beschreven. De SolariX FT-ICR werd voor hoge massa resolutie analyse gebruikt. Een van de DPP-TFB gederivatizeerde slides werd geanalyseerd zonder matrix, omdat DPP-TFB mogelijk zelf assisterende lazer desorptie ionisatie capaciteiten heeft.

De detectie van kleine moleculen werd verhoogd na het toevoegen van de derivaitzatie reagentia en DHB matrix in oplossing. Na het toevoegen van matrix op het weefsel werd het duidelijk dat de SunCollect en het Sublimatie apparaat een verhoogde kans hadden voor delokalisatie van de neurotransmitters. Met de HTX TM-Sprayer zijn vervolgens de derivatizatie reagentia en matrix toegevoegd op het weefsel voor de hoge massa resolutie analyses. De images die gemaakt zijn met de SolariX FT-ICR laten goed de verandering in distributie van de neurotransmitters zien, maar dit kan ook komen door het diepte verschil van de weefsel secties. Er moet meer onderzoek gedaan worden om deze methode volledig te optimaliseren. Binnen M4I worden meerdere studies gedaan naar neurodegeneratie die voordeel uit deze methode kunnen halen. Deze methode zou mogelijk gebruikt kunnen worden in het onderzoek van distributie van neurotransmitters in hersenweefsel secties van gezonde wild type vs. APP/PS1 muizen die over-expressie van amyloidosis geven, wat een groot teken is in Alzheimer’s ziekte.

(6)

Table of Contents V

1 Introduction

Maastricht University consists of six major faculties, one of which is the Faculty of Health, Medicine & Life Sciences which facilitates the Maastricht Multimodal Molecular Imaging (M4I) Institute. M4I is a state-of-the art molecular imaging institute that brings together a powerful palette of high-end and innovative imaging technologies. This institute consists of two research divisions, the Division of Nanoscopy, led by Prof. Peter Peters and the Division of Imaging Mass Spectrometry, led by Prof. Ron M.A. Heeren. This bachelor thesis is written based on the BRAINPATH project researched within the Division of Imaging Mass Spectrometry. Within the project studies are performed in order to link specific molecular signatures to the development and progression of different neurodegenerative diseases such as Alzheimers, different plasticity changes during ontogeny of songbirds and ischemia post stroke.

The brain is the most complex organ in a living organism. The more we learn about it, the clearer it becomes that we know just too little about it. Mankind has been studying the brain of all kinds of species for centuries. The chosen study is dedicated to link specific molecular biomarkers to the plasticity changes during ontogeny in zebra finch songbirds. During ontogeny stages young zebra finch fledglings learn to vocalize a mating song from their father, but once matured, the song is fixed, and the young avain is not able to learn any new songs or modify the memorized songs (Brainard & Doupe, 2000). The nuclei in the brain responsible for vocal memory and reproducibility are able to “store” a template song from the parent and later reproduce and modify it. During the different ontogeny stages several speculations have been made that the distribution of certain molecules, like neurotransmitters, tends to change inside the brain. To detect these neurotransmitters Mass Spectrometry Imaging (MSI), an analytical technique, is used to unveil the molecular composition of the brain.

Unfortunately, the detection of neurotransmitters tends to be difficult with MS due to their low-ionization efficiency and isobaric compounds (Esteve, et al., 2016). By chemically derivatizing the neurotransmitters in the brain tissue the detection sensitivity and background interferences would not have a negative effect on the MS analysis. For the ionization of these neurotransmitter derivatives a soft-ionization technique is used called matrix-assisted laser desorption/ionization (MALDI). With the use of a matrix these molecules are ionized and ablated into the gas-phase once struck by a pulsed laser. The ionized molecules travel at high velocity towards a detector through a mass analyser (Marvin, et al., 2003). From the data analysed by the detector a mass spectrum is generated in which the relative abundance of the analytes are plotted against the mass-to-charge ratio (m/z).

During the analysis of the whole brain tissue section an image is generated where each individual pixel contains a mass spectrum. With Mass Spectrometry Imaging (MSI) all this spectral data is mined to generate a two-dimensional molecular map of all analysed molecules (McDonnell & Heeren, 2007). This way the neurotransmitter derivatives can be localized in the brain tissue.

The goal of this thesis was to optimize and validate the sample preparation for on-tissue derivatization of neurotransmitters by comparing different derivatization reagents and matrices, as well as multiple application systems. After this methodology was fully optimized and validated the distribution of specific neurotransmitters (γ-aminobutyric acid (GABA), glutamine, methionine,

(8)

L-Chapter 1: Introduction 2

phenylalanine and L-tyrosine) within four zebra finch brain sections at the age of 20, 40, 65 and 120 days post hatching, were analyzed with different MALDI techniques.

The thesis is divided in seven chapters. Chapter 2 includes a more detailed background on the zebra finch songbird, Mass Spectrometry Imaging and the small molecules of interest. A comprehensive description of the experimental work carried out in this project is described in Chapter 3. The results of these experiments are found in Chapter 4 paired with a discussion on how the results are related to the research question. Chapter 5 contains a final conclusion of all results. My personal recommendations for future research are stated in Chapter 6. All references throughout the thesis are mentioned in Chapter 7.

(9)

Chapter 2: Theoretical Background 3

2 Theoretical Background

2.1 The Zebra Finch Songbird

The zebra finch (Taeniopygia guttata) is a songbird that originates from Australia and Indonesia. They inhabit a wide range of grasslands and forests, usually close to water. They are typically found in open steppes with scattered bushes and trees. The zebra finch has learned to adapt to human disturbances, taking advantage of human-made watering holes and large patches of deforested land (Clayton & Birkhead, 1989).

Zebra finches are loud and boisterous singers. Each male has his own unique mating song that he learns when he’s a fledgling, but once matured, the song is fixed, and he is not able to learn any new songs or modify memorized songs (Williams, 2001). Female zebra finches have the ability to sing and call as well, but male zebra finches are usually more vocal than females.

The learning and vocalizing of a song in zebra finches is somewhat similar as the learning and vocalization of speech in humans. In both cases it is about experiencing the vocalizations of other individuals. Hearing the voice of others as well as one’s own is required for vocal learning. In songbirds, these two components of learning can be quite separated in time. First, during a critical period early in life, 15 to 20 days after hatching, the sensory phase of song learning is initiated. Young birds listen to the sounds of the adult songs, mostly from the father, and form a template of a song as displayed in Figure 1. During the second, sensorimotor phase of song learning, birds begin to sing the song and refine their song using auditory feedback. This means they listen to their own voice and evaluate it by comparing their voice to the internal template song. This process is repeated until the song approximates the tutor song, called song crystallization, with some improvisations of their own (Brainard & Doupe, 2000).

Figure 1: Timetable of the song learning behaviour in Zebra Finch. Fifteen to twenty days post hatch the young zebra finch males listen to their father’s vocalization of the song, called the sensory stage. During the sensorimotor stage they start to sing and modify the song themselves. At approximately ninety days after hatching the song crystallizes and cannot be modified (Brainard & Doupe, 2000).

2.2 Song System’s Major Song Nuclei and Neuronal Connections

To get a better understanding of the song learning process, specific pathways in the brain are studied. The neurotransmitters that are involved in the learning process are important to this study. In Figure 2 a schematic drawing of a sagittal section and a H&E stain with depicted areas of an adult male zebra finch brain is displayed. The anterior forebrain pathway (grey) is the pathway for learning the template song in the sensory phase. In white, the auditory lobule is shown which enables the zebra finch to edit and refine the known template song in the sensorimotor phase. The song motor

(10)

pathway (in black) is the pathway that ensures the song output or production (Amaya, et al., 2011; Pytte, et al., 2011).

Figure 2: Schematic drawing of a sagittal section and an H&E stain of an adult male zebra finch brain. The different pathways shown all contribute to the learning process, editing and crystalization of the song. In grey the anterior forebrain pathway is displayed allows the zebra finch to learn the song and store a template in the sensory phase. This pathway consists of AreaX, the lateral magnocellular nucleus of the anterior nidopallium (LMAN), and the dorsomedial division of the medial thalamus (DLM). The auditorry lobule, shown in white, enables the bird to edit and refine the template in the sensorimoter phase. It consists of Field L, the caudal medial mesopallium (CMM), and the caudal medial nidopallium (NCM). The song motor path way, in black, ensures the song output or production which includes the robust nucleus of the arcopallium (RA) and the HVC (formerly known as high vocal center) (Amaya, et al., 2011; Pytte, et al.,

2011).

Neurotransmitters are the brain chemicals that play a vital role in the nerve cell communication. They coordinate behaviour by stimulating or inhibiting impulses. Neurotransmitters are responsible for the regulation of a variety of biological processes and behaviours in an organism, such as regulating heartbeats, body temperature regulation, peristaltic movement in the gastrointestinal tract, inhalation of oxygen into respiratory organs but also cognition and motor skills (Manier, et al., 2014; Pytte, et al., 2011; Esteve, et al., 2016). In this research five specific neurotransmitters and amino acids were analyzed: γ-aminobutyric acid (GABA), L-glutamine, L-methionine, L-phenylalanine and L-tyrosine (as seen in Figure 3). These specific neurotransmitters were studied since abnormal concentrations of these neurotransmitters are linked to certain neurodegenerative diseases like migraine (D’Andrea & Leon, 2010), depression (Mitani, et al., 2006), but also various central nervous system disorders such as schizophrenia (Hirvonen & Hietala, 2011), Parkinson’s disease (Klein, et al., 2010) and Alzheimer’s disease (Lanari, et al., 2006). Once a neurodegenerative disease is developing the neurotransmitter concentration change either by up or down regulation, and therefore play a major role in brain development and aging.

GABA is an inhibitory neurotransmitter that is widely distributed throughout the brain. It is mainly responsible for mood modulation. Low levels of GABA can lead to restlessness, anxiety and irritability. GABA, as well as glutamate, is synthesized from the amino acid L-glutamine (Albrecht, et al., 2010; Bak, et al., 2006). L-methionine is an amino acid that can be converted into serotonin and epinephrine, which are responsible for the regulation of mood, appetite and sleep (serotonin) and play an important role in the fight-or-flight response (epinephrine). The primary use of L-phenylalanine in the body is in the production of L-tyrosine which is the precursor of neurotransmitters such as dopamine, epinephrine and norepinephrine. Dopamine plays an important role in reward-motivated behaviour and is known to affect learning behaviour and mood. The

(11)

general function of epinephrine, also known as adrenaline, as well as norepinephrine is to mobilize the brain and body for action. Norepinephrine also increases arousal, alertness, retrieval of memory, and focuses attention.

Figure 3: Molecular structures of GABA, L-glutamine, L-methionine, L-phenylalanine and L-tyrosine with their corresponding masses.

2.3 Mass Spectrometry

Mass Spectrometry (MS) is an analytical technique that ionizes chemical compounds and separates the ions with the application of electromagnetic fields based on their mass-to-charge ratio (m/z). The data collected from the analysis are converted to a mass spectrum. A mass spectrum is a plot in which the relative abundance of the analytes are plotted against the m/z. By varying the electromagnetic fields, ions of different mass can be detected by a detector (see Figure 4). This occurs under a high vacuum, so the accelerating ions from the sample won’t collide with gas molecules, neutrals, and contaminating non-sample ions, which can collide with sample ions and alter their paths or produce non-specific reaction products (Glish & Vachet, 2003; Scientific, sd).

Figure 4: The MALDI Synapt G2Si mass spectrometer and a schematic drawing of deflecting ions of different mass in a mass spectrometer. The ions are separated by mass and charge via electromagnetic deflection. The amount of deflection depends on the mass of the ion and the strenght of the electromagnetic field. The dark blue ions are light ions and are deflected too much by the electromagnetic field and will not reach the detector. The yellow ions are heavier ions that are not deflected sufficiently enough and neither will reach the detector. Only the ions that are alligned properly are

(12)

detected and amplified. After signal amplification, the data that is generated reports on the relative abundance of each ion based on its mass-to-charge (m/z) ratio (Scientific, sd).

There are multiple ionization methods that can be used to obtain ions from a sample. Examples of these methods are secondary ion mass spectrometry (SIMS), electrospray ionization (ESI) and MALDI. In this study matrix-assisted laser desorption/ionization (MALDI) is used. This ionization method (as shown in Figure 5) involves a laser striking a matrix of small organic molecules which is applied to analyte molecules. The matrix and analytes are vaporized, within nanoseconds, which transforms all the struck molecules into ions. With the use of electromagnetic fields these ions are guided to the mass analyser (Marvin, et al., 2003).

Figure 5: MALDI sample ionization. A laser beam strikes the analyte/matrix mixture on the target and ionizes all molecules that it hits. The generated ions are blasted through the vaccuum towards the mass analyser with use of electromagnetic fields.

2.4 Tandem MS Analysis

To determine the structural information of detected molecules, the specific ions can be fragmented in a tandem mass analyser. Tandem mass spectrometry (MS/MS) offers additional information about specific ions. With this analyser, precursor ions, or parent ions, can be selected based on their m/z from the first MS and are fragmented into smaller ions, product ions, which can be measured by a detector for a more detailed analysis of these molecules as seen in Figure 6. The patterns are matched with the METLIN Metabolite Database to validate the compound (METLIN, sd). There are many different types of mass fragmentation, e.g. electron-capture dissociation (ECD), electron-transfer dissociation (ETD), electron-detachment dissociation (EDD), surface-induced dissociation (SID). In this research collision-induced dissociation (CID) is used to fragment the neurotransmitters. CID is a mass spectrometry technique to induce fragmentation of molecular ions in the gas phase. This fragmentation happens when the ions collide with neutral molecules, such as helium-, nitrogen- or argon gas (Es-Safi, et al., 2005).

(13)

Figure 6: A mass spectrum generated after MS/MS. At m/z 240 the precursor ion, the whole ion, is detected. The other peaks are product ions that are created during fragmentation. This specific ion can fragment at different locations and different product ions can be formed.

2.5 Derivatization Reagent/Matrix Application Systems

Since the neurotransmitters are hard to detect on their own derivatization is performed in order to improve the ionization efficiency of small molecules like amino acids and neurotransmitters. Derivatization is a technique in which a compound is chemically altered. Not only does it improve the ionization efficiency but also helps to separate target molecules from isobaric compounds (compounds with the exact same mass as the target molecule) by derivatizing the amino group as seen in Figure 7 and causing a mass shift in the mass spectrum. In previous studies derivatization reagents 4-hydroxy-3-methoxycinnamaldehyde (CA) and 2,4-diphenyl-pyranylium tetrafluoroborate (DPP-TFB) were used (Esteve, et al., 2016; Shariatgorji, et al., 2014). To validate their results the same derivatization reagents were used in this study.

Figure 7: Derivatization reaction. A part of the derivatization reagent is attached to the amino group of an amino acid or neurotransmitter.

There are many different systems sufficient for applying derivatization reagents and matrices for MALDI experiments on tissue. In this study the SunCollect sprayer, Sublimation device and the HTX TM-Sprayer were tested. The SunCollect sprayer by SunChrom (Figure 8a) is a small and compact bench top instrument, which uses compressed air or nitrogen gas that allows constant spray conditions while the nozzle moves across the slide containing the tissue sections. The sublimation device developed by Maastricht University (Figure 8b) uses a vacuum and a heat source to rapidly sublimates a solid substance into the gas-phase. First the matrix is dissolved in acetone and sublimated under vacuum. When the acetone evaporates the solid matrix is sublimated onto a glass slide that is suspended above the heating plate. This results in a more homogeneous distribution and

(14)

smaller crystal size of the matrix. With smaller crystal sizes The HTX TM-Sprayer from HTX Technologies (Figure 8c) is a uniquely designed machine that can heat the matrix solution to accelerate adsorption into the tissue, as well as a controlled flow of dry air or nitrogen to focus the spray and control drying time.

Figure 8: A. The SunCollect sprayer (SunChrom), B. Sublimation device (Maastricht University) and C. the HTX TM-Sprayer (HTX Technologies).

2.6 Mass Spectrometry Imaging

Mass spectrometry imaging (MSI) involves acquiring mass spectral data at predefined locations (pixels) across the tissue surface. Spectral data can be mined to produce two-dimensional molecular maps representing the spatial expression of any analyte (m/z) recorded across the surface of the tissue section (McDonnell & Heeren, 2007). The molecular specificity of mass spectrometry along with its ability to map locations of compounds provides information that is not easily acquired by other means (Norris & Caprioli, 2013). Visualization of changes in the concentrations of neurotransmitters in situ will be essential in understanding their role in various neurophysiological processes in different regions of the brain (Esteve, et al., 2016).

Figure 9: A schematic of MALDI MSI. A cell pellet, tissue section or whole body section is preparated on an glass- or ITO-slide and covered with a matrix. With the use of a laser the molecules in the tissue and the matrix are ionized and shot into the mass analyser. Here the ions are separated based on their m/z ratio. Mass spectra are aquired per

(15)

acquisition. From all these mass spectra a 2D molecular map can be generated to visualize where specific compounds in the tissue reside.

2.7 Flowchart of the Experiments

In Figure 10 the flowchart of the experiments is displayed. The brain tissue was cut into 12 µm thin slices, thaw-mounted on the conductive ITO slide and stored at -80oC prior to analysis. Derivatization reagents were tested in neurotransmitter solutions by spotting them on a MALDI target plate. This plate was analysed on the MALDI Synapt G2Si ETD to obtain a profile of the neurotransmitter derivative masses. To verify the components MS/MS was performed. A new mass spectrum was obtained showing the fragmented compounds. Once the verification of the neurotransmitters was done the on-tissue derivatization was performed. The matrix was sprayed on the tissue sections with three different application techniques (SunCollect, Sublimation and TM-Sprayer). These tissues were analysed on the MALDI Synapt and later on the SolariX FT-ICR for high mass resolution spectra.

Figure 10: Flowchart of Experiments. A) Derivatization reagents were tested on standard solutions of neurotransmitters and amino acids. B) The profiling analysis of the derivatized neurotransmitters and amino acids was performed on the MALDI Synapt G2Si ETD. C) Verification of the compounds was done with MS/MS. D) Tissue sections were cut on the Cryostat at a thickness of 12 µm. E) Derivatization reagents and matrices were applied on the tissue and analysed on the MALDI Synapt. F) For higher mass resolution the Solarix FT-ICR was used. G) Data analysis was performed with Masslynx and HDImaging for the MALDI Synapt experiments and with SCiLS Lab software for the Solarix experiments.

(16)

Chapter 3: Materials & Methods 10

3 Materials & Methods

3.1 Chemicals & Equipment

Conductive indium tin oxide (ITO)-coated glass slides were purchased from Bruker Daltonik GmbH (Bremen, Germany). Triethylamide (TEA), acetic acid (AcOH), a-cyano-4-hydroxycinnamic acid (CHCA), 2,5-Dihydroxybenozic acid (DHB), 4-hydroxy-3-methoxycinnamaldehyde (CA), and all standards of γ-aminobutyric acid (GABA), L-glutamine, L-methionine, L-phenylalanine and L-tyrosine were purchased from Sigma-Aldrich (Zwijndrecht, The Netherlands). 2,4-diphenyl-pyranylium tetrafluoroborate (DPP-TFB) was purchased from AstraZeneca (Cambridge, UK). Chemicals suitable for mass spectrometry of high analytical grade (99,5%) such as methanol (MeOH), ethanol (EtOH), chloroform, acetone and water were purchased from Biosolve B.V. (Valkenswaard, The Netherlands). The SunCollect sprayer (SunChrom, Friedrischsdorf, Germany), TM sprayer (HTX Technologies, Carrboro, NC, USA) and Sublimation device (Maastricht University, Maastricht, The Netherlands) were used for matrix application on tissue sections. To analyze the derivatized neurotransmitters in solution and in mouse- and zebra finch brain tissue test sections the MALDI Synapt G2Si ETD (Waters, Manchester, UK) and SolariX FT-ICR (Bruker Daltonik GmbH, Bremen, Germany) were used.

Four replicates of male zebra finch fresh-frozen brain tissue were collected at different time points (20, 40, 65 and 120 days post hatching). They were obtained from Bio-Imaging Lab, University of Antwerp, Belgium.

3.2 Tissue Sample Preparation

The MALDI-MSI is performed on thin intact tissue sections, thus optimal sectioning conditions are essential to avoid raptures and preserve the spatial distribution of molecules in question. Sagittal sections of the brain tissue were obtained with a cutting instrument-Cryostat (Microm). The brain tissue (stored at -80oC) was carefully mounted on the sample holder using dH2O. The cryostat stage

holding and blade temperature were set to -21oC. Twelve μm thick sections were obtained and thaw mounted onto conductive ITO-coated glass slides. Immediately after sectioning the samples were stored at -80oC for later use. The tissue sectioning process was performed according to standard laboratory practice which was also described elsewhere (Chughtai & Heeren, 2010).

3.3 Optimization and Validation of the Derivatization Method

The first step in optimizing the derivatization method was performing derivatization of the neurotransmitters in solution. Standard solutions of GABA, glutamine, methionine, L-phenylalanine and L-tyrosine were prepared in 50% MeOH (all 2 mg/mL). The derivatization reagents CA and DPP-TFB (both 1 mg/mL) were prepared in 50% and 100% MeOH, respectively. For both CA and DPP-TFB derivatization a total of 50 µL of the standard solutions (10µL of each standard solution) were dissolved in 50 µL derivatization reagent and 450 µL 50% MeOH. 0.5 µL TEA was added to the DPP-TFB reaction before incubation. Both mixtures were incubated at room temperature, DPP-TFB for 30 minutes and CA for one hour. The DPP-TFB derivatization reaction was stopped with the addition of 1 µL AcOH.

Both the stock solutions for CA and DPP-TFB and the derivatized solutions were spotted on a MALDI target plate. According to Shariathorji et al. DPP-TFB derivatives undergo a self-assisted laser

(17)

desorption ionization, thus can be analyzed with MALDI-MS imaging without the assistance of matrix. Therefore, DPP-TFB stock- and derivative solutions were spotted on the MALDI target place twice, to analyze it with and without the use of matrix. For the detection on the MALDI Synapt G2Si ETD, 1 µL CHCA matrix (10 mg/mL in 70% MeOH) was applied on all spots, except for one DPP-TFB stock- and one derivative spot. This same experiment was repeated with DHB matrix (35 mg/mL in 50% MeOH) instead of CHCA.

Before determining the neurotransmitters on tissue, the exact masses of the neurotransmitters were fragmented with the use of tandem MS (MS/MS). By fragmentation the precursor ions of all five derivatized neurotransmitters were analysed on the MALDI Synapt G2Si ETD. This was done by low-energy CID (40 eV). The ionized molecules of the standard solutions collided with neutral argon molecules. Fragmentation patterns were analysed.

For the on tissue derivatization of neurotransmitters the SunCollect sprayer, HTX TM-Sprayer and the Sublimator were used to apply the derivatization reagents and matrices on tissue. On the SunCollect CA (4 mg/mL in 50% MeOH) was applied by spraying three layers on the tissue at a flow rate of 10 µL/min. Incubation of the derivatization reaction took place overnight at 37oC. For DPP-TFB (5 mg/mL in 100% MeOH) five layers were sprayed on tissue with a flow rate of 10 µL/min and were incubated overnight at room temperature as stated in the articles of Esteve, et al. and Shariatgorji, et. al. In later experiments DPP-TFB incubation was done on 37oC as well. DHB matrix (35 mg/mL in 50% MeOH) was prepared and sprayed using two layers at a flow rate of 5 µL/min followed by six layers at 10 µL/min. The same derivatization solutions were prepared for the HTX TM-Sprayer. On this particular sprayer only two layers of CA and DPP-TFB were sprayed on the tissue at 30oC with a flow rate of 0.1 mL/min. DHB (35 mg/mL in 50% MeOH) was applied on tissue at 85oC at 0.1 mL/min. Lastly, matrix application by sublimation was performed by dissolving 150 mg DHB matrix in acetone. The tissue slides were first derivatized with CA or DPP-TFB before DHB matrix was applied.

Table 1: Table of used concentrations, temperatures, amount of layers applied and flow rate for each application system.

SunCollect TM-Sprayer Sublimation

Conc. Layers FlowRate Conc. Temp. Layers FlowRate Mass

CA 4 mg/mL 3 10 µL/min 4 mg/mL 30°C 2 0.1 mL/min DPP-TFB 5 mg/mL 5 10 µL/min 5 mg/mL 30°C 2 0.1 mL/min DHB 35 mg/mL 2 +6 5 µL/min 10 µL/min 35 mg/mL 85°C 6 0.1 mL/min 150 mg CHCA 10 mg/mL 2 +6 5 µL/min 10 µL/min

3.4 MALDI Analysis of Neurotransmitter Distribution

The on-tissue derivatization of neurotransmitters in zebra finch brain tissue was detected on MALDI Synapt G2Si ETD. HDI software (Waters) was used to select the sample region on the ITO-slides. Data acquisition was done in positive ion mode in the range of m/z 100 - 500. A laser spot size of 100x100 µm was used to acquire the data from the tissue. Mass spectra were analysed in

(18)

Masslynx (Waters) and the images generates in HDImaging (Waters). The mass spectra acquired on the MALDI Synapt were normalized to their total-ion-count.

For accurate mass analysis and high mass resolution the SolariX Fourier Transform Ion Cyclotron Resonance (FT-ICR) was used. There are several differences between the MALDI Synapt and the SolariX, one is the way of ion separation. With the use of a Time-of-Flight analyser the MALDI Synapt determines the ion’s m/z via a time measurement. Lighter ions and ions with a higher charge reach the detector first while the heavier ions and less charged ions travel at a lower velocity (Amaya, et al., 2011). Instead of a ToF analyer the SolariX FT-ICR uses a magnetic field that forces the ions in a circular motion. The ions are clustered based on their mass and rotate at their own frequency. The detector plates can receive a charge as the ion packets pass close to them which gives a signal that can be transformed by performing a Fourier transform. The output data from this technique is a mass spectrum with high mass accuracy (Marshall, et al., 1998).

Figure 11: The difference between the MALDI Synapt G2Si ETD and the SolariX FT-ICR.

HDI software was used to allocate the sample region on the ITO-slides. The data was acquired with a 100x100 µm laser spot size in the mass range m/z 100 to 1000. The images were processed in SCiLS Lab (Bruker). No normalization was required for the SolariX FT-ICR spectra.

(19)

Chapter 4: Results & Discussion 13

4 Results & Discussion

4.1 Optimization and Validation of Derivatization Methodology

Prior to on-tissue experiments, the derivatization reagents were tested in-solution to maximize the detection sensitivity of the neurotransmitters (GABA, L-glutamine, L-methionine, L-phenylalanine and L-tyrosine). Mixtures of these neurotransmitters were derivatized using CA and DPP-TFB reagents, both with a concentration of 1 mg/mL. Matrices CHCA and DHB were applied in a concentration of 10 and 30 mg/mL, respectively. These matrices were chosen based on their ability to help small metabolites, such as neurotransmitters, to ablate into the gas-phase once the laser strikes the compounds, promoting ionization in the process (Norris & Caprioli, 2013).

In-solutions analysis of the derivatives was done on the MALDI Synapt G2Si ETD. The derivatization led to the formation of [M]+ ions, no [M]2+ or [M + Na]+ ions were observed during the analysis. In all of the studies this mass shift occurs (Esteve, et al., 2016; Shariatgorji, et al., 2014). In Table 2 the expected masses of CA and DPP-TFB derivatized neurotransmitters are displayed.

Table 2: Accurate masses of neurotransmitters, derivatized with CA or DPP-TFB and underivatized (Esteve, et al., 2016; Shariatgorji, et al., 2014).

Neurotransmitter Mass (Da) Mass after derivatization with CA (m/z)

Mass after derivatization with DPP-TFB (m/z) GABA 103.0633 264.1152 318.1488 L-glutamine 146.0691 307.1210 361.1546 L-methionine 149.0510 310.1029 364.1365 L-phenylalanine 165.0790 326.1309 380.1645 L-tyrosine 181.0739 342.1258 396.1594

In Figure 12 and Figure 13 comparisons of both the derivatization reagents and the matrices are displayed. The mass spectra of CA derivatization (Figure 12) show that in combination with CHCA matrix almost all five neurotransmitters can be detected, with the exception of L-glutamine. There is a very small peak noticeable at m/z 307, which may correspond to glutamine. The reason that L-glutamine is not detected could be caused by high signals of other compounds (McDonnell & Heeren, 2007). With the DHB matrix all neurotransmitters are detectable but it seems the background signal, compared with the CHCA spectrum, is relatively high. During the derivatization with DPP-TFB as shown in Figure 13 the ionization of the neurotransmitter derivatives in CHCA matrix was poor. GABA, at m/z 318.12, was the only detected peak out of the five neurotransmitters of interest, although detection intensity was high. With DHB the intensity of GABA was much lower, while more peaks were detected in the spectrum.

From these four spectra can be concluded that there is a significant difference between the CHCA and DHB matrices. Especially, when the neurotransmitters are derivatized with DPP-TFB it is difficult to detect any other than GABA, once CHCA is applied. The detection sensitivity of the DHB matrix is higher in DPP-TFB, since the background signals in the CA derivatized samples could interfere with further research.

(20)

Figure 12: Mass spectra of CA derivatized neurotransmitters with CHCA matrix (top spectrum) and DHB matrix (bottom spectrum) spotted on top. Comparing the two matrices it appears that detection of neurotransmitters with DHB matrix is more sufficient than with CHCA matrix, although there is much background signal.

Figure 13: Mass spectra of DPP-TFB derivatized neurotransmitters covered in CHCA matrix (top spectrum) and DHB matrix (bottom spectrum). The CHCA matrix in combination with DPP-TFB derivatization of the neurotransmitters gives a high signal for GABA but is not sufficient for the others. DHB matrix on the other hand shows more of the specific peaks of interest (with the exception of L-glutamine) while the intensity was lower compared with CHCA.

(21)

To test out the reproducibility of the previous analysis the experiments were re-executed. A few adjustments were made during the setup of this experiment. CA was spotted on the MALDI target plate with a concentration of 4 instead of 1 mg/mL and DPP-TFB with 5 instead of 1 mg/mL. The matrices applied to the derivatives were again CHCA (10 mg/mL) and DHB (15 mg/mL). One new addition was done to the matrices, namely Norharmane (7 mg/mL) prepared in MeOH.

Figure 14 contains the matrix comparison spectra of CHCA, DHB and Norharmane matrix on the CA derivatives. In all three spectra the only undetected CA derivative was L-tyrosine (342 Da). GABA gave a high signal in CHCA and DHB matrix while Norharmane was relatively low compared to the previous two. Overall the Norharmane matrix can distinguish the neurotransmitters of interest, unfortunately at low intensity. With DHB the peaks for glutamine, methionine and L-phenylalanine gave a relatively high signal.

Figure 14: Mass spectra of in-solution matrix comparison of CHCA, DHB and Norharmane on CA derivatized neurotransmitters. In these spectra all neurotransmitter derivatives were detected, with the exception of L-tyrosine. The rest were detected with a decent signal.

The comparison of the three matrices on DPP-TFB derivatives is displayed in Figure 15. Due to a high background signal not all neurotransmitters were found in the spectra. Even with a magnification of 15x it was hard to spot any other than GABA. After magnifying the spectra from m/z 350 to 400 L-phenylalanine was visible in all spectra. In both CHCA and Norharmane too low signal intensity was measured to detect the other derivatives, while in DHB at m/z 361.16 L-glutamine was detected.

(22)

Figure 15: Mass spectra of in-solution matrix comparison of CHCA, DHB and Norharmane on DPP-TFB derivatized neurotransmitters. The spectra have been magnified 15x from m/z 350 to 400 to show low abundance peaks more clearly. These results show only GABA is being able to be detected without any magnification. For the rest the only neurotransmitter that is detected in all matrices is L-phenylalanine. In DHB matrix L-glutamine was detected as well. The black line indicates a magnification of the signal after analysis. These signals were relatively low due to the high intension of the GABA peak.

When it comes to comparing CA and DPP-TFB derivatives in both cases the Norharmane matrix is not suitable for the optimal detection of these neurotransmitters. In the DPP-TFB derivatives the spectra had to be magnified 15 times from m/z 350 to 400 to the lower intensity peaks more clearly. Unfortunately, only phenylalanine is visible in all spectra with the exception of DHB where L-glutamine is detected as well. In the CA derivatives more peaks are detected and lower background signal is measured. The profiling experiment without any matrix on the DPP-TFB derivatized amino acids and neurotransmitters was performed as well, unfortunately no signal was obtained during the analysis.

Both derivatization reagents seem to detect nearly all neurotransmitters and amino acids of interest, although DPP-TFB shows lower background signals. Especially in the first test performed the intensity of the compounds is significantly higher than with CA derivatization. During the second test the results show a higher overall signal intensity for the CA derivatized than the DPP-TFB derivatized. It is yet unclear which of the derivatization reagents works optimal for the detection of neurotransmitters. In the following experiments both derivatization reagents are used to test out the effectiveness on tissue sections.

As a conclusion of this experiment DHB matrix is the matrix with the optimal results for this type of study. From the tests done with CHCA matrix low signal intensities were obtained and was not sufficient enough to test on tissue samples. The rest of the experiments described in this thesis are performed with DHB matrix unless stated otherwise.

(23)

4.2 MS/MS Analysis

MS/MS was used to verify if the compounds that were found in the tissue are the compounds of interest. Precursor, or intact ions of the metabolite of interest, and specific fragments were found as shown in the following figures. These MS/MS spectra were compared with the METLIN metabolite database, a database containing metabolite information and MS/MS data designed to facilitate metabolite identification in metabolomics (METLIN, sd). All ions found were protonated molecules [M+H]+_{, meaning the precursor ions show a gained mass of m/z 1.008 (e.g. GABA has a mass of}

103.0633, it will show around 104.0641. See Table 2 for the masses of the molecules). In Figure 16 to Figure 19 all mass spectra of the MS/MS analysis are presented for CA derivatization as well as DPP-TFB derivatization.

Figure 16: MS/MS spectrum of GABA. The precursor ion of GABA was found at m/z 104.0876, as well as fragments of the ion at m/z 85.0932 and 87.0803 while derivatized with DPP-TFB. In CA derivatized standard solutions no fragments that correspond to the METLIN database were observed.

Figure 17: MS/MS spectrum of L-glutamine. Fragments of the L-glutamine ion were found in the DPP-TFB derivatized solution. The fragments m/z 84.1159, 85.0634, 105.0742 and 130.0930 corresponded to peaks found in METLIN. Only the precursor ion was found in the CA derivatized solution at m/z 147.0899.

(24)

Figure 18: MS/MS spectrum of L-phenylalanine. No precursor ions were found in DPP-TFB or CA derivatized solutions though several fragment peaks matched with the METLIN database. In DPP-TFB m/z 77.0715, 91.0931, 103.0924 and 149.1085 are fragments that matched the database and in CA m/z 77.0715, 91.0889, 103.0924 and 131.0957.

Figure 19: MS/MS spectrum of L-tyrosine. The precursor ion was not detected in the L-tyrosine solutions, though fragments were. In DPP-TFB derivatized m/z 77.0753, 91.0931, 119.0935 and 147.0952 correspond to fragments according to the METLIN database. This was also the case for m/z 77.0715, 91.0887, 94.0794, 147.0952 and 161.1103 in the CA derivatized solution.

In all MS/MS spectra the DPP-TFB derivatized compounds are shown in the top spectra and the CA derivatized compounds in the bottom spectra. When comparing the spectra to the data in the METLIN database we can conclude that the detection of GABA, glutamine, phenylalanine and L-tyrosine with derivatization reagents CA and DPP-TFB is possible. For L-methionine no specific peaks matched with the METLIN database. Through this experiment we can say that the detection of the neurotransmitters and amino acids of interest was successful.

4.3 On-Tissue Analysis of Neurotransmitters Derivatives

For the comparative study of the two different derivatization reagents and the three different matrix application systems, on tissue derivatization was performed on mouse and zebra finch brain test tissue sections. The analysis was executed on the MALDI Synapt G2Si ETD. On-tissue derivatization was done on six zebra finch brain tissue samples. Three with CA (4 mg/mL) and three with DPP-TFB (5 mg/mL). DHB matrix was, as concluded in the previous experiments, the most useful matrix and was sprayed on all slides, one of each derivatization reagent on the SunCollect, HTX

(25)

TM-Chapter 4: Results & Discussion 19

Sprayer and by sublimation. CHCA was concluded to be insufficient for the derivatized samples. The procedures are described in detail in the Materials & Methods section.

In Figure 20 only the results of GABA and L-glutamine are included. Other neurotransmitter derivatives were not detected. The results collected for the CA derivatization and matrix application on the SunCollect show some delocalization of GABA and L-glutamine, due to wet derivatization and matrix application (GSK, 2012). As seen in Figure 20 most of the detected derivatives are located on the edges of the tissue. The location of the tissue is marked as a region of interest (yellow mark). With Sublimation, we also observe delocalization of the neurotransmitter derivatives. A possible explanation for this is while generating the images it is not possible to separate the matrix peaks from the compound peaks. Due to this problem the matrix peaks outside the tissue, which have a significant higher intensity, are causing the image to look delocalized. With this we can conclude that the matrix application does not cause the delocalization but the application of the derivatizing reagent and protocol itself. The images clearly show the GABA and L-glutamine being detected in the matrix region and not from the tissue itself. This could be due to the insufficient mass resolving power of the MALDI Synapt to resolve isobaric and matrix compounds that interfere with the analysis.

The DPP-TFB derivatization contains similar images when using the SunCollect. Poor detection and delocalization were obtained for both GABA and L-glutamine. Fortunately, with the HTX TM-Sprayer GABA was detected and located in the tissue itself and gave no delocalization. L-glutamine was not detectable in this measurement. Unfortunately, the results for HTX application on CA derivatized tissue gave no significant signal. No images were made from this data. Due to a power outage the results for the Sublimation on DPP-TFB derivatized tissue was lost and thus no data collected as well.

Figure 20: MALDI single ion image of on-tissue derivatization with CA and DPP-TFB. The yellow marks indicate the location of the tissue sections. In this figure four tissue sections are visible, two CA derivatized en two DPP-TFB

(26)

derivatized tissue sections. DHB matrix was applied with different application systems, the SunCollect, Sublimation and with the TM-Sprayer. DHB matrix applied with the SunCollect caused low detection intensity and delocalisation of the neurotransmitter derivatives in both CA and DPP-TFB derivatized tissue. Sublimation on CA derivatives gave high signals but caused, just like the SunCollect, delocalisation. A possible explanation for this is while generating the images it is not possible to separate the matrix peaks from the compound peaks. Due to this problem the matrix peaks outside the tissue, which have a significant higher intensity, are causing the image to look delocalized. Only in the DPP-TFB derivatized tissue with matrix sprayed on by the TM-sprayer the signal is localized inside the tissue. With the use of High Mass Resolution analysis it is possible to separate these matrix peaks from the compound peaks of interest.

The problem that we run into during this experiment is the delocalization of the compounds in the tissue sections, whether that is due to a too wet matrix application, the derivatization reagent itself or the imaging with the MALDI Synapt. To test if the detection capabilities of the MALDI Synapt are due to this problem the next experiment was performed on an analyser with higher mass accuracy. This way the matrix peaks that could possibly interfere with the results are excluded from the analysis.

4.4 High Mass Resolution Imaging of Derivatized Tissue Sections

After the on-tissue analysis on the MALDI Synapt G2Si ETD it was concluded to continue with high mass resolution analysis on the SolariX FT-ICR. With a higher mass resolution that is possible on the SolariX, the distinction between the analyte of interest and the matrix peak and isobaric compounds can be made. The reason for the usage of the SolariX halfway through the study is that it became available several months after the start of the research. The first experiment performed on the SolariX FT-ICR were tested first on coronal mouse brain tissue. Other studies were focussed on mouse brain tissue (Esteve, et al., 2016; Shariatgorji, et al., 2014), this way the retrieved results could be compared with the other studies. In Figure 21 a schematic drawing and an H&E stain of a coronal mouse brain section are shown. Three coronal mouse brain tissue sections were examined, one with CA and one with DPP-TFB derivatization (2 layers of 4 and 5 mg/mL respectively at 30oC), both were incubated overnight at 37oC. The next day they were sprayed with DHB (6 layers of 35 mg/mL at 85oC) on the HTX TM-Sprayer. One section was sprayed with DPP-TFB, just like the previous section, the only change applied was the DHB matrix, which was not applied at this section. Shariatgorji, et al. described in their article that DPP-TFB derivatives undergo self-assisted laser desorption ionization, meaning that derivatized compounds, like GABA, can be measured without applying a matrix (Shariatgorji, et al., 2014). In the article of Esteve, et al. DHB matrix was applied on top of the DPP-TFB derivatives (Esteve, et al., 2016). To verify this statement of Shariatgorji, et al. the section without DHB matrix was added to the experiment.

Figure 21: A schematic drawing and an H&E stain of a coronal mouse brain section. In these images the important anatomic sectors are displayed (Cryan & Holmes, 2005).

(27)

Figure 22 shows the distribution of the neurotransmitters of interest detected with the SolariX FT-ICR in coronal mouse brain tissue sections. When comparing the CA and DPP-TFB derivatized with DHB matrix the signal intensity of CA is significantly lower than in DPP-TFB. Another noticeable thing is the distribution of the neurotransmitters and amino acids in the CA derivatized section doesn’t correspond to the regions in the brain found in other studies. The DPP-TFB derivatized section shows a more localized distribution of the compounds, though comparing the results of both derivatization reagents with DHB matrix to the DPP-TFB derivatized without any matrix applied the distribution seems delocalized. Especially for GABA the section without the DHB matrix shows a very nice distribution across the tissue section without delocalization. One of the reasons of delocalization could be the application of matrix.

The GABA distribution is located in the bottom of the tissue section in and around the hypothalamus. The hypothalamus plays a major role in mood and survival regulation, which involves eating and mating (Swain, et al., 1996). In the studies of Shariatgorji and Esteve similar results were obtained (Esteve, et al., 2016; Shariatgorji, et al., 2014). L-glutamine and L-methionine distributions are located mostly in the cerebral cortex, thalamus and the two hippocampi of the mouse brain. The cerebral cortex is the outer layer of neuronal tissue and plays a key role in memory, perception and consciousness (Shipp, 2007). The function of the thalamus is to send sensory and motor signals to the cerebral cortex (Hazlett, et al., 1999). The hippocampus has a key role in memory functions (Jarrard, 1993).

Figure 22: Comparison of on-tissue derivatization with CA and DPP-TFB with DHB matrix and DPP-TFB without DHB matrix in coronal mouse brain tissue analysed on the SolariX FT-ICR. The CA derivatives are poorly imaged compared to result of the DPP-TFB derivatives. The distribution of the compounds is delocalized and do not correspond to the regions in the brain that were expected. Comparing the DPP-TFB with and without DHB matrix shows a significant increase in detection for GABA, L-glutamine and L-methionine. Unfortunately the detection of L-phenylalanine and L-tyrosine is decreased. The reason for the decrease of signal is unknown.

Interestingly, the distribution of L-glutamine and L-methionine differ when applying matrix on top of the tissue. While the analysis without matrix works very well for GABA, glutamine and

(28)

L-Chapter 4: Results & Discussion 22

methionine the detection of L-phenylalanine and L-tyrosine decreases significantly. The reason for this is yet unknown. Since GABA and L-glutamine are the main neurotransmitters of interest and they show the least delocalization we have decided to test the protocol of DPP-TFB without the matrix, where low delocalization was observed, on the tissue sections in different ontogeny stages.

Four zebra finch brain tissue sections were placed on the same ITO slide at different ages, 20, 40, 65 and 120 days after hatching. DPP-TFB was applied with the HTX TM-Sprayer (two layers, 5 mg/mL at 30oC) and incubated overnight at 37oC and analysed with SolariX FT-ICR. In Figure 23 and Figure 24 a schematic drawing, H&E staining and a MS image of the sagittal tissue section with anatomic locations of important section of the zebra finch brain are displayed.

Figure 23: A schematic drawing and a H&E stain of a sagittal tissue section with anatomic locations of important sectors of the zebra finch brain. The images are obtained from the Histological Atlas Browser of the Zebra Finch Expression Brain Atlas (ZEBrA, sd).

Figure 24: MALDI image of the selected ion m/z =844.6. the image clearly depicts the different song control nuclei and different brain regions. Here the locations of Area X, the Striatum, LMAN, HVC and many more are shownand others in a brain 120 days post hatching.

(29)

Images were generated from the selected m/z values after analysis as shown in Figure 25 to Figure 27. The distribution of GABA (m/z 318.15587), L-glutamine (m/z 361.16252) and L-methionine (m/z 364.14808) were detected in all four tissue sections. L-phenylalanine and L-tyrosine were excluded from this figure since low signal was observed. There is a noticeable difference in neurotransmitter distribution but further analysis has to be made in order to confirm this is due to biological changes in the brain.

In Figure 25 sixteen DPP-TFB derivatized sagittal zebra finch tissue sections are displayed. On the three top rows no matrix is applied, on the bottom row DHB matrix is applied. The distribution in the sections with DHB matrix is similar to the ones without any matrix sprayed on top, yet the signals in the sections are low. Without the use of matrix the images seem much clearer and a better distribution of GABA can be observed. GABA is especially localized in the mid-bottom part of the brain where the striatum (S), area X, the nucleus spiriformis medialis (SpM) and dorso-lateral nucleus of the posterior thalamus (DLP) are located. However, changes in distribution were observed in the different ontogeny stages. At 20 days post hatching the highest intensity was found in area X and Striatum (the darkest red spot) followed by the DLP and SpM areas. The older the zebra finch male gets, the more the area of GABA distribution is shrinking. Once the bird stops learning and modifying his mating song (around 60 dph) the most intense distribution of GABA is located only in the striatum.

Figure 25: Distribution of GABA in different ontogeny stages of the zebra finch brain. H&E stain on the right is for the orientation of the tissue sections. In this image a total of sixteen brain tissue sections are shown, all derivatized with DPP-TFB. The three top rows are without any matrix applied and the bottom row is with DHB matrix sprayed on top. GABA is especially localized in the mid-bottom part of the brain where the Striatum, Area X, SpM and DLP are located.

Figure 26 displays the distribution of L-glutamine in the different ontogeny stages. In this figure the tissue sections with DHB matrix are excluded since they gave images with very low intensity. At the age of 20 days the distribution of L-glutamine is mostly located in the cerebellum and spinal vestibular nuclei. The intensity of L-glutamine increases while the zebra finch grows older. Once the zebra finch male reaches the age of 60 days the intensity of L-glutamine in the cerebellum has

(30)

increased significantly. In the spinal vestibular nuclei and the SpM and DLP in intensity increases as well.

Figure 26: Distribution of L-glutamine in different ontogeny stages of the zebra finch brain. In this image only DPP-TFB derivatized sections without the use of a matrix are shown. With matrix nearly no signal was obtained from the tissue. In the sections above the distribution of L-glutamine appears to be located in the Cerebellum and spinal vestibular nuclei of the zebra finch brain.

The distribution of L-methionine, shown in Figure 27, changes in a similar way as L-glutamine. At the age of 20 days the distribution is mostly located in the cerebellum and in the spinal vestibular nuclei and increases over time. By the time the zebra finch reaches the age of 60 days the intensity of L-methionine is slightly increased in both the cerebellum and spinal vestibular nuclei.

Figure 27: Distribution of L-methionine in different ontogeny stages of the zebra finch brain. In this image only DPP-TFB derivatized sections without the use of a matrix are shown. The distribution of L-methionine is mainly localized in the cerebellum and spinal vestibular nuclei.

(31)

In the research of Shariatgorji and Esteve similar results were obtained about the GABA distribution, although they did not perform analysis on the different ontogeny stages of the zebra finch brain. In both reports the distribution was located abundantly in the hypothalamus and basal forebrain, where the striatum, area X and the DLP are located (Amaya, et al., 2011). L-glutamine was not researched in both studies. However, as stated earlier, L-glutamine is responsible for the production of glutamate. This neurotransmitter was found to be localized in cortex regions, such as the outer layer of the cerebellum, according to both studies (Esteve, et al., 2016; Shariatgorji, et al., 2014). The distribution of L-phenylalanine and L-tyrosine is still unclear. With more research it should be possible to localize the compounds in the zebra finch brain and study the changes in different ontogeny stages.

Although we obtained some clear results, more research needs to be done on the depth of cutting in the brain. The idea is to do more experiments on the zebra finch brains and study the distribution throughout the whole brain. At this moment only a small portion of the brain is analysed. By cutting at different depths in the brain and combining all this data a 3D image model of the entire brain can be generated. With this technology more accurate conclusions can be made on the changes in neurotransmitter distribution throughout the different ontogeny stages.