Transcriptome analysis of TRPV1-positive rat dorsal root ganglion neurons

Table of contents

Zusammenfassung / Abstract

1 Introduction
1.1 Pain and nociception
1.2 TRP ion channels
1.3 TRPV1 and modulation of its sensitivity
1.4 Gene expression profiling
1.5 Methods for subpopulation isolation
1.5.1 Fluorescence-activated cell sorting (FACS) and magnetic-bead cell sorting
1.5.2 Separation using a bovine serum albumin (BSA) density gradient
1.6 Aim of this work

2 Materials and methods
2.1 Materials
2.1.2 Chemicals and solutions
2.1.3 Buffers
2.1.4 Molecular biological kits and enzymes
2.1.5 Antibodies
2.1.6 Primers
2.1.7 Other consumables
2.2 Methods
2.2.1 Preparation of dorsal root ganglion (DRG) neurons
2.2.2 Depletion of DRG neurons
2.2.3 Separation of TRPV1-positive neurons on BSA gradient purification
2.2.4 RNA isolation
2.2.5 Determination of RNA concentration
2.2.6 Assessment of RNA quality
2.2.7 Complementary DNA (cDNA) synthesis
2.2.8 Primer design for real-time PCR
2.2.9 Method for real-time PCR normalization and quantification
2.2.10 Real-time reverse-transcription PCR (qPCR)
2.2.11 Immunocytochemistry (ICC)
2.2.12 Single cell quantitative automated microscopy (QuAM)
2.2.13 Immunohistochemistry (IHC)
2.2.14 Immunoblot
2.2.15 Whole-genome gene expression analysis using Illumina RatRef-12 expression beadchip kit
2.2.16 Gene list analysis by Database for Annotation, Visualization and Integrated Discovery (DAVID)

3 Results
3.1 TRPV1 is expressed in a subpopulation of DRG neurons
3.2 Depletion approach by immunoblotting
3.3 Validation of the depletion approach by qPCR
3.4 Calcium dependent depletion of TRPV1-positive neurons
3.5 Validation of the depletion approach by quantitative immunocytochemistry
3.6 RNA quality assessment for gene expression profiling
3.7 qPCR validation for gene expression profiling
3.8 Array-based transcriptome detection
3.9 Generation of final gene lists
3.10 Validation of 14 genes by qPCR
3.11 CART expression exclusively in TRPV1-positive neurons
3.12 Gene Ontology annotation and gene clustering by DAVID
3.13 KEGG (Kyoto Encyclopedia of Genes and Genomes) pathway enrichment

4 Discussion
4.1 Isolation of the TRPV1 positive subpopulation
4.2 Concept of gene expression experiments
4.3 Validation of 14 selected candidates by qPCR
4.4 Gene enrichment analysis by the DAVID tool
4.5 KEGG pathway enrichment by DAVID tool
4.6 Summary of enrichment analysis
4.7 Outlook

5 Abbreviations

6 Supplemental data
6.1 Stringent gene list
6.2 Relaxed gene list

7 Table of figures

8 References

Zusammenfassung:

Schmerz ist ein physiologischer Schutzmechanismus, der vornehmlich in Wirbeltieren zu finden ist. Die Detektion schmerzhafter Stimuli erfolgt durch spezialisierte Neuronen: die nozizeptiven Neuronen. Nozizeptive Neuronen (auch als Nozizeptoren bezeichnet) sind sensorische Neuronen mit freien Nervenendungen im Epithel des Körpers, die schmerzauslösende Reize detektieren und an das Rückenmark weiterleiten. Über sekundäre Neuronen im dorsalen Wurzelhorn des Rückenmarks gelangt der Reiz zur Großhirnrinde und ins limbische System, wo er emotional bewertet wird. Eine komplexe subjektive Sinneswahrnehmung entsteht: Schmerz.

Es wurden erhebliche Fortschritte im Verständnis der molekularen Zusammenhänge bei der Schmerzentstehung erzielt. Eine Vielzahl von Ionenkanälen ist für die Schmerzdetektion von thermischen, chemischen oder mechanischen Stimuli verantwortlich. Ein wichtiger Bestandteil der Nozizeptoren sind dabei ihre vorwiegend für bivalente Kationen selektiven Ionenkanäle (transient receptor potential (TRP)), die zur Membrandepolarisation und damit zur Generierung von Aktionspotentialen beitragen. Zusätzlich ermöglichen sie eine molekulare und funktionelle Einteilung der Nozizeptoren. Insbesondere die TRPV1-positiven Neuronen haben eine herausragende Rolle bei der Schmerzdetektion und dem Temperaturempfinden.

Eine weitere wichtige Eigenschaft der TRP Ionenkanäle ist die Möglichkeit, ihre Aktivierung durch Sensitivierung oder Desensitivierung zu beeinflussen. Bis heute sind die molekularen Mechanismen der Schmerztransduktion jedoch nach wie vor nur unzureichend erforscht und die beteiligten Moleküle nur in Ansätzen charakterisiert. Das Wissen über das Transkriptom einzelner Subpopulationen nozizeptiver Neuronen würde die Suche nach neuen Molekülen und Mechanismen entscheidend unterstützen.

Um eine Transkriptomanalyse in TRPV1-positiven Neuronen durchzuführen, habe ich die funktionelle Eigenschaft von TRPV1 Ionenkanäle genutzt, nach Aktivierung mit spezifischen Agonisten (Capsaicin und Resiniferatoxin (RTX)) Kalzium in die Zelle einströmen zu lassen. Dies ändert die Dichte der Neuronen und ermöglicht die Abtrennung mittels Dichtegradientenzentrifugation. Dieser neuartige Ansatz zur Abreicherung erlaubt im Vergleich zu einer Kontrollgruppe (inklusive TRPV1-positiven Neuronen) die Bestimmungen von ausschließlich in TRPV1-positiven Neuronen angereicherten Transkripten.

Die Validierung dieser neuen Methode erfolgte mit Hilfe von Western-Blot, qPCR und Immunocytochemie und zeigte eine erfolgreiche Abreicherung TRPV1-positiver Neuronen um bis zu 95%. Bei der anschließenden Genexpressionsanalyse konnten Kandidaten ermittelt werden, die hochspezifisch in TRPV1-positiven Neuronen angereichert sind. Die Analyse der Kandidaten erfolgte im Anschluss mit Datenbanken, die eine Einordnung der Gene in verschiedene Klassen ermöglichen, um Informationen über Art und Häufigkeit der angereicherten Genklassen aufzufinden.

Eine Vielzahl von Kalium- und Kalziumionenkanälen, sowie G-Protein gekoppelte Rezeptoren (inklusive Regulatoren) waren in TRPV1-positiven Neuronen angereichert. Zudem wurde das Neuropeptide Cocaine and Amphetamine regulated transcript (CART) identifiziert. Wir haben gezeigt, dass CART ausschließlich in einer sehr kleinen (<12%) Subpopulation TRPV1-positiver Neuronen exprimiert wird. Es steht in engem Zusammenhang mit der Signalweiterleitung zwischen nozizeptiven Neuronen und dem Rückenmark.

Um genauere Aussagen über molekulare Zusammenhänge zu machen, werden im Labor etablierte Methoden genutzt werden. Mit diesen wird der Einfluss von ebenfalls stark angereicherten G-Protein gekoppelten Rezeptoren wie z.B. LPAR3, MRGPRD oder OPRK1 auf Aktivierungszustände von Signalwegen überprüft. Dabei wird die Aktivierung von Transduktionsmolekülen untersucht (Proteinkinase A Aktivierung), sowie durch Kalziumeinstrommessungen der Einfluss auf die Aktivierbarkeit von Ionenkanälen analysiert werden. Das Potential zur Modulation von Signalzuständen gibt Aufschluss über die Rolle von G-Protein gekoppelten Rezeptoren bei der Schmerzentstehung und zeigt möglicherweise Konzepte für neue Schmerztherapien auf.

Abstract:

Pain is an unpleasant sensory and emotional experience and an essential component of the body’s defense system. Noxious stimuli are detected by special afferent sensory neurons called nociceptors (or nociceptive neurons), which connect peripheral tissues with the central nervous system. Noxious signals travel along the nociceptor to the dorsal horn of the spinal cord from where they projected to higher brain regions via 2nd order neurons. After evaluation in brain stem, thalamus, and cortex a complex sensory perception arises: Pain.

Extensive research has been performed to gain insight into molecular mechanisms of pain sensation and its modulation. Many of the transduction channels that convert thermal, mechanical or chemical stimuli into electrical activity are transient receptor potential (TRP) ion channels, which are widely expressed in nociceptors and play a prominent role in pain sensation. Still many transduction molecules that trigger pain have not been discovered.

To identify transcripts enriched in TRPV1-positive neurons I used the ability to activate these channels with the specific agonists capsaicin and resiniferatoxin (RTX). The resulting calcium influx changes structural properties and thus alters the density of TRPV1-positive neurons. This enables their separation via density gradient centrifugation. In comparison to a control group (including TRPV1-positive neurons) this newly developed method allows me to find predominantly expressed genes in TRPV1-positive neurons.

To validate the depletion I utilized western blot, immunocytochemistry and qPCR. The results showed effective depletion of TRPV1-postitive neurons for up to 95%.

Subsequent gene expression analysis revealed promising candidates that were highly enriched in TRPV1-positive neurons. The candidate genes were classified and clustered according to bioinformatics databases to gain information about the properties and distribution of enriched genes.

Among them several potassium- and calcium ion channels, G-protein coupled receptors (including regulatory genes), and members of the phosphatidylinositol pathway were enriched.

A strong enrichment was seen for the neuropeptide, Cocaine and Amphetamine regulated transcript (CART). Immunostaining proofed its expression in very small population (<12%) of TRPV1-positive DRG neurons. Indeed, CART plays a role in transducing signals between nociceptors and interneurons within the dorsal horn.

To get a more detailed view of candidate genes, recently established methods will be used to show influence of enriched GPCRs like LPAR3, MRGPRD, and OPRK1. For this purpose activity of signal pathways is evaluated (Proteinkinase A activity). Additionally modulation of ion channels is analyzable with calcium-sensitive dyes that make calcium influx visible.

To improve our understanding of pain sensation and modulation further analysis of underlying molecular mechanism and signaling molecules is needed.

1 Introduction

1.1 Pain and nociception

Pain and nociception| Pain is an unpleasant sensory and emotional experience and an essential component of the body’s defense system (Bonica, 1979). The strength and duration of pain is strongly driven by emotional and subjective perception and influenced by past experiences, environment interactions and many other external and internal factors.

The neuronal process by which intense thermal, mechanical, or chemical stimuli are detected and transmitted from the site of stimulation to the central nervous system (CNS) is termed nociception. Nociception is performed by specialized sensory afferent neurons, called nociceptors (or nociceptive neurons), which connect peripheral tissues with the CNS.

The function of the nociceptors, located in skin, muscle, joints and viscera is to signal the body tissue damage and thus prevent further exposure to noxious stimuli (Hucho & Levine, 2007).

Two types of nociceptive pain are usually distinguished: pain emanating from the skin and deeper tissues (e.g. joints and muscle) is referred to as somatic pain while pain emanating from the internal organs is referred to as visceral pain. Somatic pain is usually well localized whereas visceral pain is not (Fein, 2010).

Impact on health care systems | The fact that about 20-25% of the world population suffers from chronic pain and only few obtain sufficient relief by currently available drugs indicates the need for further intensive studies of underlying molecular mechanisms and components (Romanovsky et al., 2009). Although there are many analgesics for the treatment of chronic pain states like COX-2 inhibitors, morphine-derivates, paracetamol and several non-steroidal anti-inflammatory drugs (NSAIDs) especially neuropathic pain in most cases is not treatable so far. Moreover many currently available drugs have severe side effects such as sedation, cognitive impairment, respiratory depression, tolerance, constipation, gastrointestinal bleeding, ulcers, myocardial infarction, stroke, ataxia, arrhythmias, nausea, fatigue and addiction (Katz and Barkin, 2010). This often severely restricts the use and thereby the relieving potential of current drugs.

To improve our understanding of pain sensation and current therapeutics as well as how pain sensitization is modulated, further analysis of underlying molecular mechanisms and signaling molecules is urgently needed.

Classification of Nociceptive Neurons | Nociceptive neurons are a heterogeneous group of neurons. Their cell bodies are located within so called dorsal root ganglia (DRG) next to the spinal cord. Their peripheral nerve endings are connected with a long bidirectional axon that conducts action potentials from the periphery to the CNS.

The connection of nociceptor presynaptic terminals with 2nd order neurons is located in distinct lamina (predominantly lamina I and outer lamina II) inside the gray matter of the spinal cord. These 2nd order neurons then transmit the signal to structures in the brain stem, thalamus, and cortex where it is integrated into the perception of actual pain (see Figure 1).

illustration not visible in this excerpt

Figure 1: Activation of nociceptors in peripheral tissue results in the sensation of pain.

Afferent functions: Injured tissue (either by noxious chemical, thermal or mechanical stimuli) and inflammatory cells release a variety of molecules, such as ATP, bradykinin, H+, nerve growth factor, prostaglandins and vascular endothelial growth factor (VEGF), that excite and/or sensitize nociceptors. Painful stimuli are detected by nociceptors and are conducted to neurons in the spinal cord which project to higher centers of the brain. Pain signals seem to ascend to the brain by at least two main spinal-cord pathways - the spinothalamic tract and the dorsal column.

Efferent functions: Nociceptor activation results in the release of neurotransmitters, such as calcitonin gene-related peptide (CGRP), endothelin, histamine, glutamate and substance P. Nociceptor activation also causes the release of prostaglandins from the peripheral terminals of sensory fibers, which can induce plasma extravasation, recruitment and activation of immune cells, and vasodilatation. Figure modified from Patrick W. Mantyh, Denis R. Clohisy, Martin Koltzenburg & Steve P. Hunt, 2002.

The nociceptor has a unique assembly which is referred to as pseudo uni-polar, meaning that they have one axon with two branches: The central branch goes to the spinal cord, where it forms synapses with other neurons. The peripheral branch goes to peripheral tissue.

The peripheral nerve terminal embedded in the tissue responds to endo-and exogenous stimuli (afferent functions) but also releases a variety of molecules that influence the local tissue environment (efferent functions). Neurogenic inflammation, for example, refers to the process whereby peripheral nerve termini release the neuropeptides such as calcitonin gene-related peptide (CGRP) and substance P in response to inflammatory mediators. Thereby, they induce among others vasodilatation and extravasation of plasma proteins, respectively (Basbaum and Jessell, 2000).

Nociceptive neurons can be divided according to biophysical criteria into three main groups:

A δ mechano-sensitive neurons - myelinated, fast conducting neurons that respond to noxious mechanical stimuli (pressure).

A δ mechanothermal neurons - lightly myelinated, fast conducting neurons that respond to noxious mechanical stimuli (pressure, touch) and to heat.

Polymodal neurons (C fibers) - unmyelinated, slowly conducting neurons that respond to a variety of stimuli.

illustration not visible in this excerpt

Figure 2: Classification of afferent nerve fibers connecting peripheral tissues with the CNS. Aβ fibers detect innocuous stimuli like warmth and soft pressure and are not involved in pain sensation. Aߜ fibers act as nociceptors for fast mechanical and heat pain. Mainly polymodal C-fiber nociceptors detecting strong mechanical and heat stimuli are of special interest for pain research because their sensory functionality is in the first place of pain sensation and is commonly altered in pathological pain states, which has great impact on health care systems worldwide (Nevena, 2007). Figure provided by J. Isensee (content adapted from Ewan St. John Smith and Gary R. Lewin, 2009).

The nociceptive neurons can be further classified according to different activation thresholds (e.g. for temperature or mechanical stimulus), mode of action, myelination degree, morphology of nerve terminals or protein expression (Meyer et al., 2008).

Because of their afferent functions and their biophysical classification nociceptors are also denoted as afferent nociceptive fibers, afferent nerve fibers, nociceptive C-fibers, or nociceptive neurons.

Nociceptors are characterized by the expression of a large number of different ion channels, receptors, neurotransmitters, and neuromodulators (Woolf, 1996). The diversity and heterogeneity among nociceptive afferents has probably contributed to the difficulty in the identification of new therapeutic agents and makes it impossible to discriminate nociceptive neurons with only single criteria regardless of which marker (e.g. anatomical, molecular, functional) is used.

The molecular tools to distinguish the heterogeneous population of nociceptors are still at a nascent stage compared to e.g. differentiation of subgroups of immunocells, which can be divided into clearly classified “clusters of differentiation (CD)”.

1.2 TRP ion channels

Beside functional studies there is the possibility to distinguish subgroups of nociceptors with antibodies against prominent ion channels, which are expressed in specific subpopulations and which have often also been connected functionally with nociception.

Many of the transduction channels that convert thermal, mechanical or chemical stimuli into electrical activity are transient receptor potential (TRP) channels, which are widely expressed in nociceptors. Therefore they play a major role in pain transduction and sensitization (Gold & Gebhart, 2010).

Today 28 mammalian transient receptor potential (TRP) ion channels are known which are subdivided into six superfamilies (see Figure 4). These channels are expressed in nearly all tissues and are important in many regulative cell functions. Studies underlined the high importance of members of this evolutionary conserved superfamily of ion channels for pain as they have been brought into context of noxious stimuli detection. For example TRPV1, TRPA1 and TRPM8 are expressed in neuronal cell types and are important for sensing different chemical stimuli, mechanical stimuli and temperature.

The majority of TRP channels are permeable for calcium ions (Ca² +). The cation influx through these channels is triggered by a wide range of natural and synthetic substances and can be modulated via intra- or extracellular mechanisms. Activation of ion channels (e.g. calcium or sodium channels) produces a generator potential, which depolarizes the cell. Depolarization eventually activates voltage- gated sodium channels (NaV) and leads to the generation of actions potentials (Gees, Colsoul, & Nilius, 2010).

illustration not visible in this excerpt

Figure 3: TRPV1, TRPA1, TRPM8. Plant products mimic physical stimuli. Capsaicin, present in hot chili peppers, menthol, produced by mint plants, and mustard oil act as agonists for TRPV1, TRPM8 and TRPA1, respectively.

Figure kindly provided by Jan Erik Siemens, Max-Delbrueck- Center for Molecular Medicine (MDC).

illustration not visible in this excerpt

Figure 4: Phylogenetic tree of the mammalian transient receptor potential (TRP) channel superfamily. TRPC (canonical), TRPM (melastatin), TRPV (vanilloid), TRPA (ankyrin), TRPP (polycystin), and TRPML (mucolipin) are the only identified subfamilies in mammals. (Nilius, Owsianik, Voets, & Peters, 2007)

TRPV1, TRPA1, and TRPM8 have different properties, which help to classify nociceptive neurons with molecular markers.

TRM8 is known as a sensor for moderate cold temperatures (~26°C) and activated by plant product menthol. Mice lacking TRPM8 gene are severely impaired in their ability to detect cold temperatures. Interestingly, these animals are deficient in many diverse aspects of cold signaling, including cool and noxious cold perception (Daniels & McKemy, 2007).

TRPA1 has a central role in the pain response to endogenous inflammatory mediators and to irritants, such as mustard oil, formaldehyde and garlic. Although several in vitro studies showed clear evidence for cold activation , in vivo results are not clearly confirmed so far (Tai, S. Zhu, & N. Zhou, 2008).

TRPV1 is commonly activated by heat (activation threshold of ~43°C), capsaicin, protons, and endovanilloids. Moreover, the TRPV1-positive neurons play an important role in pain perception and temperature sensation, which was demonstrated in studies of TRPV1-deficient mice revealing an impaired nociception and pain sensation (Caterina et al., 2000).

In vivo TRPA1 and TRPV1 are expressed in partly overlapping subpopulations, where TRM8 is exclusively expressed in TRPA1-negative populations (Hondoh et al., 2010).

1.3 TRPV1 and modulation of its sensitivity

One feature of nociceptors is their ability to acquire a more reactive phenotype often designated as sensitization. Sensitization is typically triggered as a consequence of tissue insult or inflammation and lowers the activation threshold of nociceptors (Woolf and Salter, 2000). Nociceptor sensitization may thereby lead to hyperalgesia, a condition, which is characterized by stronger and prolonged responses to painful stimuli (Woolf and Mannion, 1999). Another common result of modulation is allodynia, a state where pain is initiated by normally innocuous stimuli.

Aspects of sensitization can be linked to the modulation of conductivity properties of TRP ion channels. Due to its prominent role in pain research and its ability to be modulated by many inflammatory mediators I focus on sensitization events in TRPV1-positive neurons.

TRPV1 is a Ca² + permeable non-specific cation channel, located on peripheral sensory neurons, that serves as a molecular detector for capsaicin, protons, and endovanilloids. Moreover, its role as a heat sensor (activation threshold of ~43°C), its modulation by inflammatory agents (i.e. nerve growth factor, bradykinin, prostaglandins, etc.) and its efferent vasodilatory effect on blood vessels (by releasing CGRP) makes TRPV1 an important component of the pain pathway (Spicarova & Palecek, 2009).

Modulation of sensitivity is mainly mediated by various G-protein coupled receptors (GPCR) expressed on nociceptive neurons such as TRPV1-positive neurons. Many inflammatory mediators such as neuropeptides or plasma proteins are generated at inflammed sites, and many of them modulate the sensitivity of nociceptive sensory neurons after binding to their respective GPCRs. Indeed, many Gαq coupled receptors such as bradykinin receptor 2, prostaglandin receptor, protease activated receptor 2, histamine receptor 1, metabotropic glutamate receptors (mGluR1 and mGluR5), and receptor tyrosine kinase (RTK) receptors, are implicated in sensitization of sensory neurons via TRPV1 modulation during inflammation-induced thermal hyperalgesia (figure 5) (Woo et al., 2008).

Inflammatory mediators such as ATP, bradykinin and trypsin potentiate TRPV1 activity in a PKC- dependent manner. Direct phosphorylation of TRPV1 by PKC has been described, which results in increased current when the channel is activated by chemicals (e.g. capsaicin) (Bhave et al., 2003). Phosphorylation of TRPV1 can be also achieved by protein kinase A (PKA), which is activated downstream of G-protein coupled receptors (e.g. after prostaglandin E2 treatment)(Mishra & Hoon, 2010). Figure 5 provides an overview of activation and sensitization in TRPV1-positive nociceptors.

Up to now, the diversity on the molecular level of nociception is poorly understood. Beside distinguishable subgroups (e.g. classified by ion channel expression) even nociceptive neurons of the same subgroup show major differences in reaction to stimuli. Furthermore, partly overlapping populations contribute to this diversity.

Signaling pathways and underlying regulatory processes build a complex network that influence reactivity and plasticity of nociceptive neurons. Knowledge of the genes expressed in TRPV1-positive subpopulation will therefore improve our understanding of processes involved in pain transduction and modulation and therefore help to find new molecules and pathways connected with altered pain states.

illustration not visible in this excerpt

Figure 5: Activation and sensitization of nociceptors. (a) Transduction can involve both direct and indirect pathways. The ion channel TRPV1 can be directly opened by increases in temperature or by chemicals released from resident (mast) and recruited (polymorphonuclear leukocyte; PMNL) immune cells, epithelial cells, Schwann cells, fibroblasts and sympathetic post- ganglionic neurons (SPGN).

(b) There are multiple points of interaction between second messenger pathways that are engaged after nociceptor activation via their corresponding GPCRs, including at the levels of signaling molecules such as Ca²⁺, effector molecules such as PKCε, and common targets, such as TRPV1 and NaV1.8 (not shown) for the pathways activated.

There is also the possibility of translocation of TRPV1 to the cell surface, which may contribute to injury-induced increases in channel activity.

(c) Sensitization of nociceptors also involves positive feedback. Activation of ion channels such as TRPV1 results in membrane depolarization and Ca² + influx through TRPV1 and voltage-gated calcium channels (VGCC). Ca2+ influx can drive the release of neuropeptides (stored in dense core vesicles; pink) and glutamate (stored in clear vesicles; yellow), both of which can drive further activation of receptors on nociceptors and release mediators from the sources described in (a).

Abbreviations: PGs, prostaglandins; OLAMs, oxidized linoleic acid metabolites; NE, norepinephrine; ER/GPR30, estrogen receptor/G protein receptor-30; 5-HT, serotonin; CaM, calmodulin; PLC, phospholipase C. DAG, diacylglycerol; IP3, inositol triphosphate; AC, adenylate cyclase; EPAC, cAMP-activated guanine exchange factor; PI3K, phosphoinositide 3-kinase;ER Kl/2, extracellular signal-regulated kinases 1 and 2; TNFR, tumor necrosis factor receptor (adapted from Gold amp; Gebhart, 2010).

1.4 Gene expression profiling

To find new molecules or pathways connected with pain sensation in TRPV1-positive subpopulation a general approach should be used which identifies all genes expressed in this subpopulation. This starts with gene expression profiling experiments, which are strongly depended on starting material. In this context the heterogeneity of tissue, cultured cells or organs plays an important role in the detection of expressed genes between two or more conditions.

In heterogeneous samples, like whole DRGs or cultured DRG neurons, it is difficult to decipher the influence of single cell populations to whole transcript amount. Dependent on cell type and function the transcript pattern can strongly deviate within a sample.

This problem is seen in nerve injury models which were developed to screen for differentially expressed gene patterns in DRG neurons after cutting nerve fibers. Because nociceptive neurons are surrounded by a huge amount of glia and other non-neuronal cells there is a strong impact of excess transcript on gene expression data. The main functions of glia cells are to surround neurons and hold them in place, to supply nutrients and oxygen, to insulate one neuron from another, and to destroy pathogens and remove dead neurons. An excessive amount of transcript derived from non-neuronal cells may obscure the detection of differences within the neurons. Most of the changes in gene expression induced by a treatment may have occurred within non neuronal cells. An assignment of genes to a defined subpopulation like TRPV1-positive neurons is impossible.

To circumvent this problem a more defined way is the isolation and separation of subpopulations by physical separation. Isolating a defined and homogenous subpopulation as starting material for the transcriptome analysis makes it possible to analyze a well-defined population. Transcripts can therefore be clearly assigned to their corresponding subpopulation.

A first approach to identify differentially expressed genes in small and large DRG neurons was conducted by Luo et al. in 1999 using laser capture microdissection of DRG neurons. It was already known that mainly small C-fiber neurons act as nociceptors and they were aiming to identify pain relevant targets. Due to technical limitations at this time, the expression of only 477 genes was analyzed with a cDNA microarray. Although a few genes showed significant differential expression between small and large neurons, clear results or even functional gain could not be obtained (Luo et al., 1999).

It is therefore necessary to find a method to separate TRPV1-positive cells from the heterogeneous population of DRG neurons.

1.5 Methods for subpopulation isolation

A number of methods for cell-cell separation are known such as fluorescence-activated cell sorting (FACS), magnetic bead cell sorting, or density gradient centrifugation. Not all of them are suitable for the purification of TRPV1-positive nociceptive neurons.

1.5.1 Fluorescence-activated cell sorting (FACS) and magnetic-bead cell sorting

FACS is routinely used by immunologists to separate and isolate fluorescently labeled populations of e.g. blood cells. The technique is based on one-cell-at-a-time analysis in a liquid stream, which is focused, scanned, and later on separated by electrostatic deflection. Cells are sorted upon the specific light scattering and fluorescent characteristics.

A large toolbox of fluorescently labeled antibodies allows differentiating a large number of different immunocells. This advantage could also be used for co-labeling of different neuronal (or nociceptive) markers (e.g. labeling of nociceptive neurons with TRPV1, TRPA1 and PGP9.5 antibodies coupled to different dyes) to define more precise subpopulations. Moreover the ability to sort cells on-the-fly would be perfect for subgroup isolation in concerns of speed.

Nevertheless, sorting using flow cytometry is difficult with neuronal cells because they are adherent cells and are very variable in size, shape and structural stability. Another major limiting factor is the cell number. Routinely up to 10⁶ to 10⁷ cells are commonly used in cell sorting protocols. To obtain such a large number of sensory neurons is challenging, e.g. the isolation of DRG neurons from rat normally yields only up to 3x 10⁴ cells per rat (Sergent-Tanguy, 2003).

Despite the mentioned problems another problem is the reliability and availability of suited antibodies for subgroup isolation of DRG neurons. Most of them define the classes only incomplete or with overlapping boundaries to other functional and/or molecular markers (Gold & Gebhart, 2010). Especially TRP ion channels have only three small and conserved extracellular domains. When analyzing living cells with intact cell membranes it is nearly impossible to stain them differentially. Therefore FACS is not a reliable method for subgroup isolation in our case.

Another approach uses magnetic nano particles coupled to antibodies. The experimental setup is based either on batch preparation or on filter-based flow-through techniques. In both ways unspecific binding of coupled antibodies has to be avoided (Evgeny Klyuchnikov et al., 2011). It is desirable to have a highly specific antibody raised against extracellular structures, which are accessible by the antibody. This consideration highlights the problem of extracellular structures. : A high grade of variable glycosylation of extracellular protein domains makes it difficult to get a strong interaction with antibodies (William G. Flynne, 2008) and in the case of TRPV1 there exist only three extracellular domains. Homolog regions among the TRP channels make it difficult to raise specific antibodies directed against extracellular epitopes and currently none is commercially available. Moreover, TRPV1 is routinely stored in vesicles inside the cell and is therefore not exposed to the cell surface. All in all, these disadvantages make antibody-based magnetic sorting and FACS not suitable for our needs.

A tool, which uses functional properties of TRPV1 ion channel, would circumvent the problematic use of antibodies for separation of TRPV1-positive and TRPV-1negative neurons.

1.5.2 Separation using a bovine serum albumin (BSA) density gradient

High density protein, glycerol or polysaccharide mixtures are routinely applied to separate cell suspensions according to their respective density. For example separation of peripheral blood mononuclear cells from whole blood is commonly accomplished through density gradient centrifugation using Ficoll (a hydrophilic polysaccharid). After the centrifugation step, Ficoll separates layers of blood, with lymphocytes and monocytes under a layer of plasma (Noble & Cutts, 1967).

But in addition to the separation of clearly different cell populations which differ significantly by shape, size and density determined by their morphology in vivo, there is also the possibility to separate cells using compound-induced alteration of cell densities (e.g. cytotoxic agents, chemical compounds). If the compound acts specifically on e.g. differentially expressed ion channels such as TRPV1 this allows the separation of functionally distinct subgroups out of otherwise rather similar cells (Ajiro, Th'ng, Yau, & Nishi, 2004; Drebot et al., 2008). This method circumvents the antibody problematic and makes use of the intrinsic properties of the TRPV1 ion channel.

For our purpose I therefore used the ability of capsaicin and RTX to specifically activate the TRPV1 ion channel and provoke a strong calcium influx into the cytoplasm.

Structural changes | It is believed that the entry of high concentrations of calcium seems to be a significant event in the neurotoxic process, since structural changes were reported a few minutes after capsaicin application (Marsh, 1987).

Rearrangements of F-actin structures, the stabilizing scaffold of cells, changes cell structure and stability (Gerits et al, 2007). Finally, this process leads to alterations in osmotic pressure and thus to alterations of the cell density. Therefore TRPV1-positive cells could be separated on a BSA density gradient and should accumulate on top of the BSA together with lipid-rich axon fibers and cell debris (Figure 6).

illustration not visible in this excerpt

Figure 6: Pattern after density centrifugation on a 15% BSA gradient. Only very few cells are present in MEM and BSA phase. Interphase (between BSA and MEM): cell debris, axon fibers, glia cells and TRPV1-positive cells that were not able to pass the BSA mesh; cell pellet: population of nociceptive neurons or rather TRPV1 depleted population which could pass the BSA mesh.

This BSA density centrifugation approach could therefore be used for separation of TRPV1-positive neurons without the disadvantageous use of antibodies. Unlike the direct isolation methods (FACS and magnetic-bead cell sorting) I would rather gain a TRPV1-depleted neuronal population then a separated TRPV1-fraction, mainly because a direct use of TRPV1-positive cells was not feasible due to high amount of cell debris and glia cells in the interphase.

In order to reveal genes predominantly expressed in TRPV1-positive neurons the gene expression profiling would compare non-depleted controls with depleted samples.

Intracellular calcium | A sustained increase of intracellular calcium is the main event leading to cell death after TRPV1 activation (Grant, Dubin, S.-P. Zhang, Zivin, & Z. Zhong, 2002).

A massive calcium influx caused by capsaicin or RTX quickly raises intracellular calcium concentration (usually between 100-200 nM) up to 500 nM, which mitochondria actively take up. As the mitochondrial take in calcium, ATP synthesis stops, the inner-mitochondrial membrane depolarizes, the matrix pH rises, and the binding of excess calcium to the membrane transition pore causes it to open fully and irreversibly. Once this channel opens completely a cascade of events occur that lead to the release of apoptosis inducing factor (D.-R. Conrad, 2011).

Additionally intracellular accumulation of ions leads to osmotic changes and activation of proteolytic enzyme processes.

1.6 Aim of this work

Peripheral sensory neurons represent a heterogeneous population, which differs in morphology, transcript/protein expression, and functionality. Knowledge about the differentially expressed proteins is of importance to understand in greater detail mechanisms such as pain sensitization. Up to now, transcriptomes have only been analyzed from whole tissues with excess of non-neuronal cells such as whole ganglia. The major aim of this work is therefore to optimize a method for subpopulation separation to then identify the transcriptome of the TRPV1-positive subpopulation.

This project is structured into several parts. First, I have to optimize a method to deplete TRPV1- expressing neurons from dissociated DRGs using agonists of TRPV1 and subsequent density-gradient purification. In particular the incubation time, spinning time, handling, and concentration of both gradient mixture and TRPV1 agonist will be optimized.

The degree of depletion will be validated on protein level (western blot), at the single cell level (immunocytochemistry), and at the transcript level (qPCR). The respective protocols have to be optimized.

Second, with the optimized protocol the separation will be performed and the transcriptome of TRPV1- positive nociceptive neurons will be determined. To achieve this, total RNA derived from TRPV1- depleted and non-depleted dissociated DRG-cells will be isolated, amplified, labeled, and hybridized onto Illumina RatRef-12 expression bead chips.

Third, the result list of differentially expressed transcripts will be validated by qPCR analysis of 14 enriched genes.

Bioinformatics resources will be used to identify enriched classes/pathways and may help to evaluate the biological significance of the findings. Subsequent analysis of the identified target genes may help to address current mechanistical questions in the field of pain research and may lead to the identification of novel markers for specific nociceptive subgroups.

2 Materials and methods

2.1 Materials

2.1.1 Laboratory equipment

illustration not visible in this excerpt

2.1.2 Chemicals and solutions

illustration not visible in this excerpt

2.1.3 Buffers

Normal goat serum blocking (NGSB) | 2% goat serum, 1% BSA, 0.1% Triton X-100, 0.05% Tween 20, 0.01 M PBS, pH 7.2, stored frozen at - 20 °C.

Blotting buffer(5x) | glycin 200 mM, Tris 250 mM, SDS 0.2 % (w/v) Laemmli buffer | Tris-HCl 25 mM, glycine 0.2 M, SDS 0.1 % (w/v)

TBS-T buffer | NaCl 137 mM, KCl 27 mM, Tris-HCL 25 mM pH 7.4, tween 20 0.1 % (v/v) TAE buffer | Tris 40 mM, Na-acetate 5 mM, EDTA 1 mM

TE Buffer | Tris 10 mM pH 8.0, EDTA 1 mM

Loading buffer (2x) | urea 8 M, Tris-HCl 15 mM pH 7.5, glycerol 8.7 % (v/v), SDS 1 % (w/v), bromophenol blue 0.004 % (w/v), 2-mercaptoethanol 143 mM

2.1.4 Molecular biological kits and enzymes

illustration not visible in this excerpt

If not otherwise mentioned kits and enzymes were used regarding manufacturers instruction.

2.1.5 Antibodies

illustration not visible in this excerpt

2.1.6 Primers

illustration not visible in this excerpt

For further information on gene description see 2.2.7 Primer design for real-time PCR.

2.1.7 Other consumables

illustration not visible in this excerpt

2.2 Methods

2.2.1 Preparation of dorsal root ganglion (DRG) neurons

Cultures of dissociated lumbar DRG were prepared from adult male Sprague Dawley rats (200-300 g; Charles River Laboratories). Rats were housed in a temperature controlled environment at 21 °C under a 12 h light / dark cycle and free access to food and water. Rats were killed by slow CO2 intoxication. The skin was removed and the area along the spine was disinfected with ethanol (70% v/v). To both sides of the spine the skin was sliced to approximately 5 mm depth with a scalpel. After removing residual muscles, flesh and fibers around the spine, six dorsal processes were visible. The first lumbar (L1) vertebra is cracked and the two dorsal subjacent DRG can be extracted carefully. For further treatment DRGs are stored in a petri-dish containing MEM. The preparation of L2 to L6 was done in the same way.

The twelve collected DRGs were now carefully desheated under a binocular microscope. This step helps to improve the purity of the culture because epithelial sheets around the DRG and residual axonal strands are removed. To further improve the dissociation of the cells the DRG were incubated in 2.5 ml MEM incl. 100 μL collagenase (5 U/ml) for 60 min at 37 °C which destroys connective tissue consisting of collagen fibers between DRG neurons and glia cells.

To separate the single cell clusters the DRG were transferred to 1.25ml MEM and triturated 10-times by slowly pipetting up and down using a firepolished siliconated Pasteur-pipette. The cell suspension is carefully stored in a new tube. Non-dissociated tissue is again triturated 10-times in fresh 1.25 ml MEM. The procedure is repeated twice with an even smaller Pasteur-pipette.

The final volume after titruation is 5 ml MEM. The suspension is divided into 2-4 tubes each with a final volume of 2.5 ml cell suspension regarding to required experimental setup.

2.2.2 Depletion of DRG neurons

In preparation for the depletion of TRPV1 positive cells from the whole cell suspension the cells are treated with specific TRPV1 agonists (e.g. capsaicin, resiniferatoxin (RTX)) at following conditions: Tube 1: DMSO control (addition of 2.5 μL DMSO, final concentration: 0.1% DMSO) Tube 2/3: depletion (addition of 2.5 μL 10 mM Capsaicin / 100 μM RTX dissolved in DMSO, f. c.: 10 μM Capsaicin / 100 nM RTX , 0.1% DMSO) In preliminary test the optimal incubation time of 30 min at 37 °C was determined.

After stimulation with the specific TRPV1 agonist, cells were spun down 5min, 100g and the MEM was discarded. To prevent cell-cell adhesion that could disturb the separation process on the BSA gradient the cells were incubated with 2.5 ml Hank’s BSS/EDTA (0.025%, pH 8) / Trypsin (245 U/ml) solution for 4 min at 37 °C. Trypsin, an endopeptidase, destroys cell adhesion proteins and EDTA captures free Ca² + ions that support cell adhesion proteins. The process is stopped by addition of 100 μL MgSO4 (10 mM).

2.2.3 Separation of TRPV1-positive neurons on BSA gradient purification

The cell suspension was now ready for density separation on a bovine serum albumin (BSA) gradient (15%, w/v). The high amount of protein in the BSA solution has a defined density. Cells are separated corresponding to their own density which varies dependent on cell content and tyep. Cells with low density will be retained and only cells with higher density than the BSA solution can pass them. Myelinated axonal debris and glia cells will not pass the BSA gradient and are separated from neuronal cell types. For this purpose the cells are carefully loaded on top of the BSA gradient and centrifuged for 10 min (120g).

Figure 7: Density centrifugation on a 15% BSA gradient. Only very few cells are present in MEM and BSA phase. Interphase: cell debris, axon fibers, glia cells and TRPV1-positive cells that were not able to pass the BSA mesh (only in capsaicin and RTX treated conditions); cell pellet: population of nociceptive neurons or rather TRPV1 depleted population which could pass the BSA mesh.

The supernatant (mainly interphase with debris etc.) can be discarded or collected for further analysis. If collected, the interphase was carefully removed, diluted 1:10 with MEM and again centrifuged to get a new pellet consisting of debris, axon fibers, neurons with lower density than BSA and glia cells. The pellet could be resuspended in neurobasal A media and plated on μClear 96-well plates if desired.

In order to support neuron attachment on plastic surfaces μClear 96-well plates were coated with polyornithine (0,1 mg/ml)-laminin (5 μg/ml) for 1h at room temperature and were air-dried at least 15 min before plating. μClear micro plates have clear bottoms of ultra-thin polystyrene films. They have extended wavelength ranges down to 340 nm, increased sensitivity, and reduced background which makes them an optimal tool for immunofluorescent analysis of plated cells.

To prepare the culture medium, neurobasal A medium was supplemented with 2 % (v/v) B27, L- glutamine (0.5 mM), L-glutamate (25 μM) and penicillin/streptomycin (100 U/ml).

To validate the depletion of TRPV1-positive neurons the initial cell pellet after BSA gradient centrifugation was used in the following experimental setups:

The cell pellet was resuspended in 1 ml neurobasal A media and used for quantitative immunocytochemistry. For this purpose the 100 μL of cell suspension (depending on experimental setup between 200-400 cells per well) were plated on a 96-well μClear plate.

Principle and procedure of immunocytochemistry is described in chapter 2.2.11.

Alternatively, residual supernatant above the cell pellet was removed and the tube was freezed in liquid nitrogen and stored at -80 °C for RNA preparation. Isolated total RNA was used for RNA bead arrays or for cDNA synthesis and respectively for qPCR analysis.

qPCR procedure is described in the following chapters (2.2.4-2.2.10).

2.2.4 RNA isolation

RNA isolation was performed with NucleoSpin RNA/Protein kits from Macherey-Nagel according to the manufactures instructions. Frozen cell pellets were lysed in 350 μL buffer RP1 with 3.5 μL ß- mercaptoethanol and filtered. RNA was bound to RNA-binding silica columns in presence of chaotropic salts. Chaotropes destabilize hydrogen bonds, van der Waals forces, and hydrophobic interactions and the association of nucleic acids with water is disrupted providing binding conditions for the silica column. Moreover ethanol supports the binding conditions. DNA was digested by DNase incubation for 15min directly on the column, columns were washed to get rid of residual salts and ethanol, and the RNA was eluted with 40-60 μL RNase-free water.

2.2.5 Determination of RNA concentration

RNA quantity was measured with a UV photometer (Nanodrop™) (1 μL used). In principal the Nanodrop™ uses the feature of the purine and pyrimidine bases of RNA to absorb UV light of 260nm wavelength. Using the Lambert-Beer-law the concentration of RNA can be calculated.

A = e ' c - d, whereas A= absorption, ߝ= molar extinction coefficient, which is 25 μL / μgήcm for ssRNA, c= concentration of nucleic acid in μg/ml, d= path length of bulb in cm.

RNA purity can be estimated using absorption of light at 260 and 280 nm (A260/280). High absorbance at 280nm is commonly caused by proteins and ideally this ratio should be close to 2 for high quality nucleic acid. The absorbance at 230nm gives evidence of contaminant salt or solvents. The ratio (A230/260) should be close to 2.0 - 2.2.

Due to omnipresent RNase on lab ware, dust, skin and pipettes RNA samples should always be stored on ice or at -20°C for short term storage. Long term storage can be accomplished at -80°C.

2.2.6 Assessment of RNA quality

In order to get a more reliable output for RNA quality then using absorbance ratios I utilized Agilents Bioanalyzer to check total RNA for potential degradation. For both qPCR and gene expression analysis the RNA quality has to be as good as possible to get reproducible and reliable results. The Agilent 2100 Bioanalyzer is a microfluidics-based platform for sizing, quantification and quality control of RNA. It combines capillary electrophoresis separation and dye incorporation to analyze RNA quality and quantity.

[...]