The Role of Histone Modifications in Severe Asthma

A Thesis submitted to Imperial College London

for the degree of Doctor of Philosophy by

Peter Owen Brook

Molecular Cell Biology Airway Disease Section National Heart and Lung Institute The Faculty of Medicine Imperial College London London SW3 6LY

2 Abstract Severe asthma is an airways disease that causes restriction of airflow to the lung leading to difficulty breathing, where even high doses of medication does not alleviate the symptoms. There are studies that have demonstrated differentially expressed genes in both CD8+ T-cells and the bronchial epithelium of severe asthmatics, however the causes of the changes in gene expression have not been elucidated. Histones, proteins which store DNA, can be acetylated or methylated which changes their binding of DNA allowing for modified transcription. I hypothesise that modifying the histone modification of Bronchial Epithelium and CD8+ T-cells would be able to limit the inflammatory effects of severe asthma, that measuring changes in histone modifications would show differences between severe and non-serere asthmatics and that by using modern analysis of gene expression and histone modifications it would be possible to discover novel genes and pathways involved in asthma.

The Bromodomain and Extra terminal domain mimic JQ1 was able to suppress release of Interleukin-6 and Interleukin-8 and SGC-CBP30 the CBP/p300 inhibitor limited proliferation in bronchial epithelium. CD8+ T-cells from a small group of severe and non-severe asthmatics had gene arrays carried out and were analysed by weighted geneome coexpression network analysis which found modules containing genes involved in intracellular motility and vesicle formation that linked strongly to % predicted Forced Expiratory volume in one second. CD8+ ChIP-Seq studies of T-cell histone modifications were attempted, but not successful.

A pilot test on Severe asthmatics CD8+ T-cells treated with SGC-CBP30 showed a reduction in the release of the inflammatory mediator MIP1α and suggested that this was without affecting its mRNA production or cell viability, however higher numbers of patients would be needed to confirm this.

3 Declaration of originality

I, Peter Owen Brook declare that I wrote this thesis and the experiments and work described herein, except where appropriately referenced, was performed by myself. Information derived from the published and unpublished work of others has been acknowledged in the text and full references are given.

‘The copyright of this thesis rests with the author and is made available under a Creative Commons Attribution Non-Commercial No Derivatives license. Researchers are free to copy, distribute or transmit the thesis on the condition that they attribute it, that they do not use it for commercial purposes and that they do not alter, transform or build upon it. For any reuse or redistribution, researchers must make clear to others the license terms of this work’

4 Acknowledgements I would like to extend the sincerest thanks to my supervisors Prof. Ian Adcock, Dr Paul Lavender, Dr Andrew Durham and Dr Sharon Mumby. Firstly for giving me the opportunity to work in their fantastic groups, for the academic and intellectual support, but also for their moral support and friendship.

I would like to thank Asthma UK/MRC for generously funding my PhD.

I would also to thank Prof Fan Chung and Dr Pank Bhavsar for helping gain access to asthmatic blood and airway samples. I’d like to extend a great appreciation to Dr Audrey Kelly for guiding me though the gene array and ChIP-Seq work and to Dr Kate Heesom for carrying out the SILAC mass spectroscopy.

The group who I’ve worked with these last few years have had a massive influence on the way I work in science and have been fantastic colleagues including Dr. Christos Rossios, Dr, Coen Wiegman, Dr. Stelios Pavlidis, Dr. Scott Kuo, Elen Jazwari, Dr. Seun Ojo, Dr. Charis Michaeloudes and Dr. Mark Perry. Along with my fellow Students: Kirsty, Younis, Amir, Rebekah, Tank, Lily and Jack and the wonderful masters and undergrad students who I’ve worked with over the last four years.

Finally I have to thank Lily, my Mum and Dad, Annabel, and all my friends for helping me through when things were tough and celebrating with when things went well.

5 Table of Contents Chapter 1: Introduction ...... 20 1.1 Asthma ...... 21 1.1.1 Definition ...... 21 1.1.2 Prevalence ...... 21 1.1.3 Mechanism of asthma ...... 21 1.2 Treatments ...... 24 1.3 Severe Asthma ...... 27 1.3.1 Mechanisms and subtypes of Severe Asthma ...... 28 1.4 Bronchial Epithelial Cells ...... 32 1.5 T-Cells ...... 34 1.5.1 T helper 1 cells ...... 35 1.5.2 T helper 2 cells ...... 36 1.5.3 T helper 17 Cells ...... 37 1.5.4 Regulatory T-Cells ...... 38 1.5.5 Cytotoxic T-Cells ...... 39 1.5.6 The Role of Cytotoxic T-Cells in Asthma ...... 41 1.6 Genetics of Asthma ...... 44 1.7 Epigenetics ...... 45 1.8 Histones Modification ...... 45 1.8.1 Histone Acetylation ...... 46 1.8.2 Histone Methylation ...... 53 1.8.3 H3K4 ...... 53 1.8.4 H3K9 ...... 55 1.8.5 H3K27 ...... 55 1.9 Hypothesis and Aims ...... 60

Chapter 2: Materials & Methods ...... 61 2.1 General Methods ...... 62 2.1.1 Enzyme Linked Immunosorbent Assays (ELISA) ...... 62 2.1.2 Viability Assay ...... 62 2.2 Cell proliferation Assay ...... 63 2.3 Human Bronchial Cell Models ...... 63 2.3.1 BEAS-2B Culture ...... 63 2.3.2 Normal Human Bronchial Epithelial Cell Culture ...... 64

6 2.3.3 BEAS-2B Epigenetic Probe/Mimic Treatments ...... 64 2.3.4 Treatment of nHBE Cells with Epigenetic Probes/mimics ...... 66 2.3.5 Effect of Probes/mimics in combination with a corticosteroid in BEAS-2B Cells 66 2.3.6 SILAC Proteomics Methods ...... 66 2.3.7 SILAC labelling and Cell culture ...... 66 2.4 SILAC Methods ...... 68 2.4.1 SILAC Protein Extraction ...... 68 2.4.2 Proteomics using mass spectroscopy ...... 68 2.5 Histone Purification and Analysis ...... 70 2.5.1 Global Histone Modification analysis ...... 70 2.5.2 Histone Purification ...... 70 2.5.3 Histone H3 Post Transcriptional Modification Multiplex ELISA ...... 71 2.6 Isolation of CD8+ T-Cells from Asthmatics ...... 73 2.6.1 Patient Selection Criteria for CD8+ T-cells Collection ...... 73 2.6.2 CD8+ Isolation ...... 74 2.6.3 Ficoll-Paque Isolation ...... 74 2.6.4 CD8+ T-Cell Isolation ...... 74 2.7 CD8+ T-cells Extraction and analysis ...... 75 2.8 Chromatin Immunoprecipitation-Sequencing (ChIP-seq) ...... 80 2.8.1 Chromatin Preparation ...... 80 2.9 CD8+ T-Cell Activation Model ...... 87 2.9.1 Cell Culture Methods for CD8+ T-Cells ...... 87 2.9.2 Stimulation CD8+ T-cells ...... 87 2.9.3 Dynabead Concentration response ...... 88 2.10 Epigenetic Probe/mimic concentration responses in CD8+ T-cells .... 88 2.11 PCR and single gene CBP ChIP for CD8+ T-Cells ...... 88 2.12 Statistical analysis ...... 89

Chapter 3: Epigenetic Mechanisms in Asthmatic Bronchial Epithelium 91 3.1 Chapter Introduction ...... 92 3.2 Chapter Hypothesis and Aims ...... 93 3.3 Results ...... 94 3.3.1 Gene Expression of Histones and Histone Modifying Enzymes in Asthma 94 3.3.2 Stable Isotope Labelling with amino acids in culture (SILAC) Data ...... 101

7 3.3.3 Histone modifications in normal human bronchial epithelial cells (nHBECs) ...... 107 3.3.4 Effects of epigenetic probes on normal Human Bronchial Epithelial Cell responses ...... 109 3.4 Discussion ...... 115 3.4.1 U-BIOPRED Brushing and Bronchial biopsy Data ...... 115 3.4.2 SILAC Analysis ...... 117 3.4.3 Global Histone Modifications ...... 118 3.4.4 Epigenetic Probes ...... 118 3.5 Conclusion and Next steps ...... 119

Chapter 4: CD8+ T-Cell Transcriptome in Asthma ...... 122 4.1 Chapter Introduction ...... 123 4.2 Chapter Hypothesis and Aims ...... 124 4.3 Results ...... 125 4.3.1 Investigation of gene expression in peripheral blood mononuclear cell in the U-BIOPRED cohort ...... 125 4.3.2 CD8+ T-Cell Isolation and Gene Array ...... 127 4.4 Discussion ...... 153 4.4.1 U-BIOPRED Peripheral Mononuclear cells Gene Expression ...... 153 4.4.2 CD8+ T-cell Gene Array Experimental Design ...... 154 4.4.3 WGCNA Analysis of CD8+ T-cell Data ...... 160 4.4.4 Tsitsiou Purple Module ...... 163 4.5 Conclusion and next steps ...... 163

Chapter 5: H3K27ac in Asthmatic CD8+ T-Cells ...... 166 5.1 Chapter Introduction ...... 167 5.2 Chapter Hypothesis and Aims ...... 168 5.3 Results ...... 168 5.3.1 ChIP Quality Control ...... 168 5.3.2 H3K27ac ChIP-Seq ...... 171 5.4 Discussion ...... 180 5.5 Conclusion and next steps...... 181

Chapter 6: The Role of CBP and p300 in CD8+ T-cells of Severe Asthmatics ...... 182 6.1 Chapter Introduction ...... 183

8 6.2 Chapter Hypothesis and Aims ...... 183 6.3 Results ...... 185 6.3.1 Primary human CD8+ T-Cells ...... 185 6.3.2 RT-qPCR ...... 187 6.3.3 CREB Binding Protein ChIP ...... 190 6.4 Discussion ...... 193 6.5 Conclusion and next steps ...... 195

Chapter 7: Discussion ...... 197 7.1 General Discussion ...... 198 7.1.1 Bronchial Epithelial Epigenome Exploration ...... 198 7.1.2 CD8+ T-cell transcriptome ...... 201 7.1.3 CD8+ T-cell Histone Acetylation and Methylation ...... 204 7.1.4 Role of Vesicle creation and budding in Severe Asthma CD8+ T-cells 204 7.2 Conclusion and future studies ...... 208

Chapter 8: References ...... 210

9 List of Figures Figure 1.1 Mechanisms of Atopic Asthma ...... 23 Figure 1.2 Interactions of Brd4 ...... 51 Figure 3.1. Histone Modifications after FP treatment and IL-1β stimulation in nHBECs ...... 108 Figure 3.2. Bromosporine concentration-response on IL-1β stimulated Beas2B Cells ...... 111 Figure 3.3. JQ1(+) concentration-response on IL-1β stimulated Beas2B Cells ...... 112 Figure 3.4. SGC-CBP30 concentration-response on IL-1β stimulated Beas2B Cells ...... 113 Figure 4.1. FACS Gating Strategy for CD8+ T-cells...... 129 Figure 4.2 FACS Gating Strategy for CD8+ T-cells from Tsitsiou, 2012 ...... 130 Figure 4.3. Gene Array Read Frequency Plot for CD8+ T-cell ...... 131 Figure 4.4. Gene Array Genes Detected and PCA ...... 133 Figure 4.5. PCR/Gene Array Confirmation ...... 138 Figure 4.6. Gene Array/Tsitsiou Gene Array Comparison ...... 139 Figure 4.7. CD8+ T-cell Sample Gene Dendrogram and Trait Heatmap ..... 141 Figure 4.8. WGCNA Soft Power Selection Criteria ...... 142 Figure 4.9. Gene dendrogram and module colours of WGCNA clusters ..... 143 Figure 4.10. Clustering of module eigengenes ...... 145 Figure 4.11. WGCNA Module-trait Relationship ...... 146 Figure 4.12. WGCNA Cluster Dendrogram and Module Colours of Tsitsiou Data ...... 150 Figure 4.13. Tsitsiou Data Module-Trait Relationships ...... 151 Figure 5.1. DNA Bioanalyser read ...... 169 Figure 5.2. Sequence Length Distribution ...... 170 Figure 5.3. CAMP H3K27ac ChIP-Seq Peaks ...... 173 Figure 5.4. GGA1 H3K27ac ChIP-Seq Peaks ...... 174 Figure 5.5. CD81 H3K27ac ChIP-Seq Peaks ...... 175 Figure 5.6. FOS H3K27ac ChIP-Seq Peaks ...... 176 Figure 5.7. HIST1H1C H3K27ac ChIP-Seq Peaks ...... 177

10 Figure 5.8. DEFA1 H3K27ac ChIP-Seq Peaks ...... 178 Figure 5.9. RBM39 H3K27ac ChIP-Seq Peaks ...... 179 Figure 6.1. Concentration-Response to Epigenetic Probes of Dynabead- Stimulated Severe Asthamtic CD8+ T-cells ...... 187 Figure 6.2. RT-qPCR analysis of gene expression in CD8+ T-cells from Severe asthmatics after SGC-CBP30 and Dynabeads treatment ...... 189 Figure 6.3. CBP Binding to the FOXP3 Gene after SGC-CBP30 treatment 191 Figure 6.4. CBP Binding to the IFNγ Gene after SGC-CBP30 treatment ..... 192 Figure 6.5. CBP Binding to the CCL3 Gene after SGC-CBP30 treatment ... 192 Figure 6.6. Micrograph of CD8+ T-cell Clustering around Dynabeads ...... 195 Figure 7.1. Intracellular and extracellular traffic of CD8+ T-cells...... 206

11 List of Tables Table 1.1. Breakdown of Severe asthma phenotypes ...... 31 Table 1.2 T-cell Subtype Breakdown ...... 35 Table 2.1. List of Epigenetic Probes/mimics used ...... 65 Table 2.2. SILAC Treatment and Label for nHBECs ...... 68 Table 2.3. Histone Modification Binding Conjugated Antibodies ...... 72 Table 2.4. Components for Sybr Green RT-qPCR master mix ...... 79 Table 2.5. PCR Primers for Sybr Green ...... 79 Table 2.6. ChIP MNase Digest Buffer Components ...... 81 Table 2.7. ChIP Sonication buffer ...... 81 Table 2.8. ChIP Dialysis Buffer ...... 82 Table 2.9. Modified RIPA Buffer ...... 83 Table 2.10. ChIP Wash Buffer 1 ...... 83 Table 2.11. ChIP Wash Buffer 2 ...... 83 Table 2.12. ChIP Elution Buffer ...... 84 Table 2.13. Protease Inhibitors Cocktail 100 x stock ...... 84 Table 2.14. Calbiochem Phosphatase Inhibitors Cocktail Set II ...... 84 Table 2.15. PCR Cycling Conditions ...... 86 Table 2.16. ChIP-Seq Traces for H3K27ac ...... 87 Table 3.1 Table of Genes of interest for Bronchial Biopsies and Brushing .... 95 Table 3.2. DEGs from Bronchial Brushings Healthy vs. Severe Asthmatics . 97 Table 3.3. DEGs from Bronchial Brushings Moderate vs. Severe Asthmatics ...... 98 Table 3.4 DEGs from Bronchial Brushings Healthy vs. Moderate Asthmatics98 Table 3.5. DEGs from Bronchial Biopsies Healthy vs. Severe Asthmatics .... 99 Table 3.6. DEGs from Bronchial Biopsies Moderate vs. Severe Asthmatics 100 Table 3.7. Protein change after IL-1β stimulation in nHBECs ...... 103 Table 3.8. Histone related protein change after IL-1β stimulation in nHBECs ...... 104 Table 3.9. Protein change after FP treatment and IL-1β stimulation in nHBECs ...... 105

12 Table 3.10. Histone related protein change after FP treatment and IL-1β stimulation in nHBECs ...... 106 Table 3.11. Epigenetic probe concentration-response summary ...... 114 Table 4.1. DEG in Peripheral Mononuclear Cells – Severe Asthmatics vs. Healthy Subjects ...... 126 Table 4.2. DEG in Peripheral Mononuclear Cells – Severe vs. Moderate Asthmatics ...... 126 Table 4.3. CD8+ T-cell patient demographics...... 127 Table 4.4. Enrichment of CD8+ T-cells ...... 130 Table 4.5. Significantly DEGs of CD8+ T-cells – NSA vs SA ...... 134 Table 4.6. Large negative FC DEGs of CD8+ T-cells – NSA vs SA ...... 135 Table 4.7. Large Positive FC DEGs of CD8+ T-cells – NSA vs SA ...... 136 Table 4.8. Genes within the WGCNA Green-Yellow Module ...... 147 Table 4.9. Genes within the WGCNA Light-Green Module ...... 147 Table 4.10. Genes within the WGCNA Red Module ...... 148 Table 4.11. Genes within the WGCNA Blue Module ...... 149 Table 4.12. Genes within the Tsitsiou WGCNA Purple Module ...... 152 Table 5.1. Overrepresented sequences in ChIP-Seq ...... 170 Table 5.2. ChIP-Seq H3K27ac Promoter Peak Summary ...... 172 Table 6.1. Patients Demographics for CD8+ T-cells used for compound screening ...... 186 Table 6.2. Patient Demographics CD8+ T-cells used for SGC-CBP30 RT- qPCR screening ...... 188 Table 6.3. Patient Demographics for SGC-CBP30 treated CD8+ T-cells used for CBP ChIP ...... 191

13 Abbreviations ACD Acid Citrate Dextrose ADEPT Airways Disease Endotyping for Personalized Therapeutics AEBSF 4-(2-Aminoethyl) benzenesulfonyl fluoride hydrochloride AHR Airway Hyperresponsiveness AHS Airway Hypersensitivity AIDA axin interactor, dorsalization associated ANOVA analysis of variance AP-1 Activator protein 1 APC Antigen Presenting Cells ARMCX1 armadillo repeat containing, X-linked 2 ASM Airway Smooth Muscle ATS American Thoracic Society AUI Arbitrary Units of Intensity BCA bininchoninic acid BEBM Bronchial Epithelial Cell Basal Media BEGM Bronchial Epithelial Cell Growth Media BET bromodomain and extraterminal domain proteins BLT1 Leukotriene B4 Receptor BMI Body Mass index BPE Bovine Pituitary Extract Brd4 Bromodomain 4 BrdU Bromodeoxiuridine BSA Bovine Serum Albumin CAMP Cathelicidin Antimicrobial Peptide CANX calnexin CCL3 Chemokine (C-C motif) ligand 3 ChIP-Seq Chromatin Immunoprecipitation-Sequencing Cbp/P300 Interacting Transactivator With Glu/Asp Rich Carboxy- CITED2 Terminal Domain 2 CLCA1 chloride channel, calcium-activated, family member

14 COPD Chronic obstructive pulmonary disease CpG cysteine residues followed by guanine CBP cAMP response element-binding protein-binding protein CREB cAMP response element-binding CRISPR Clustered regularly interspaced short palindromic repeats CTLA-4 cytotoxic T lymphocyte-associated antigen 4 CS Corticosteroids CXCL8 chemokine (C-X-C motif) ligand 8 DC dendritic cells DCTN1 dynactin subunit 1 DEFA1 Defensin 1A DEG differentially expressed genes DMSO Dimethyl Sulfoxide DNA deoxyribonucleic acid DNMT1 DNA methyltransferase DST dystonin DUSP1 dual specificity phosphatase 1 EC50 Effective Concentration of a drug to give a 50% maximal response ECM Extracelluar Matrix EDTA Ethylenediaminetetraacetic acid EGF Endothelial Growth Factor EHMT2 Euchromatic Histone-Lysine N-Methyltransferase 2 ELISA Enzyme Linked Immunosorbent Assays Emax The maximum response to drug or stimulus European Network For Understanding Mechanisms Of Severe ENFUMOSA Asthma ERS European Respiratory Society EZH2 Enhancer of zeste homolog 2 FACS Fluorescence-activated cell sorting Fc Crytalising Fraction of the Antibody FC Fold Change

15 FcεRI high-affinity IgE receptor FDR false discovery rate corrected p-value FE Fold Enrichment

FEV1 Forced Expiratory Volume in 1 second FOXP3 forkhead box P3 FP Fluticasone Propionate FCS foetal calf serum FSC Forward scatter FVC Forced vital capacity GA gene array GATA3 Trans-acting T-cell-specific transcription factor 3 GCS Glucocorticosteroids GRE Glucocorticoid Response Element GGA1 Glucosidase Alpha, Acid GNB2L1 guanine nucleotide-binding protein subunit beta-2-Like 1 GR glucocorticoid receptor GWAS genome-wide association study H3K Histone 3 Lyseine HAT histone acetyltransferase HDAC histone deacetylase HEPES 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid HIV human immunodeficiency virus IC50 Inhibitory Concentration of a drug to give a 50% inhibitory response ICS Inhaled Corticosteroid IFN-γ Interferon Gamma Ig Immunoglobulin IL Interleukin Imax The maximal inhibitory response to drug or stimulus IRF Interferon regulatory factors KDSR 3-ketodihydrosphingosine reductase KO Knock-Out

16 LABA long acting β2 agonists LAR Late Asthmatic Response LTC4 Leukotriene C4 LTQ Linear Trap Quadrupole MAP3K2 mitogen-activated protein kinase kinase kinase 2 MHCI histocompatibility complex I MHCII mass histocompatibility complex 2 MIP1a Macrophage Inflammatory Protein 1a miRNA microRNA MLL Mixed Linage Leukemia MMP8 matrix metallopeptidase 8 Mnase micrococcal nuclease mRNA messengerRNA MTOC microtubule organising centre MTT (4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide NF-κB nuclear factor kappa-light-chain-enhancer of activated B cells nHBEC normal Human Bronchial Epithelium NHS National Health Service NHV Normal Healthy Volunteer NSA non-smoking asthmatics OD Optical Density OVA Ovalbumin P-TEFb positive transcription elongation factor P4HB prolyl 4-hydroxylase subunit beta PBS Phosphate-buffered saline provocative concentration of inhaled bronchodilator which causes PC20 FEV1 to fall by 20%. PCA Principle Component analysis

PGE2 Prostaglandin E2 PRR pattern recognition receptors PTPN13 Protein tyrosine phosphatase, non-receptor type 13

17 RANTES Regulated on Activation, Normal T Cell Expressed and Secreted RAST radioallergosorbent test RIPA Radioimmunoprecipitation assay RPKM Reads Per Kilobase of transcript per Million mapped reads RT Room temperature RT-qPCR Quantitative reverse transcription RU Relative Units San non-smoking severe asthmatics SEM Standard Error of the Mean serpinB2 serine peptidase inhibitor, clade B (ovalbumin), member 2 SETD7 SET Domain Containing Lysine Methyltransferase 7 SFM Serum-Free Media SGC Structural Genomics Consortium SILAC Stable isotope labelling by amino acids in cell culture SNP Single Nucelotide Polymorphism SSC Side scatter STAT4 Signal transducer and activator of transcription 4 STK16 serine/threonine kinase 16 Tc Cytotoxic T TCR T-cell Receptors TE Tris-EDTA TGF-β Transforming growth factor beta 1 TH T Helper TLR Toll like receptor TNF-α Tumour Necrosis Factor TOM Topological Overlap Matrix Tregs Regulatory T-cells TSA Trichostatin A TSLP Thymic Stromal Lymphopoietin TSS transcriptional start sites TUBA1A tubulin a

Unbiased BIOmarkers in PREDiction of respiratory disease U-BIOPRED outcomes UCSC University of Califoinia, Santa Cruz UK United Kingdom USA United States of America UTX tetratricopeptide, X VEGF Vascular endothelial growth factor VSX1 visual system homeobox 1 WGCNA Weighted Gene Co-Expression Network WHCS1 Wolf-Hirschhorn Syndrome Candidate 1

19 Chapter 1: Introduction

20 1.1 Asthma

1.1.1 Definition Asthma is a chronic airways disease marked by repeated incidences of difficulty in breathing including wheezing, breathlessness, chest tightness and coughing. In asthma these ‘attacks’ results in a reversible limited expiratory flow after which breathing returns to normal. (Global Initiative for Asthma 2016, http://ginasthma.org/).

1.1.2 Prevalence Asthma poses a huge healthcare burden on society. Approximately 5.4 million people in the UK are treated for asthma and worldwide it is estimated that >300 million people suffer from asthma, nearly 5% of the world’s population (Masoli et al. 2004). There are 65,000 hospital admissions for asthma per annum in the UK with 25,000 of these admissions for under 14s (Levy et al. 2014). Treating asthma costs the NHS £1 billion a year (Levy et al. 2014). Most of these costs are associated with patients with severe asthma, patients who do not respond to the primary treatment of inhaled corticosteroids (ICS). Accordini and colleagues estimated that severe asthma costs the EU more than €20bn per annum (Accordini et al. 2013).

1.1.3 Mechanism of asthma Many asthmatics have allergic asthma or atopic asthma, around 54% of adults (Pearce et al. 1999) and have a tendency towards having an allergic response to airborne factors that would not affect a healthy subject (Platts- Mills 2001). Initial exposure to an allergen in atopic asthmatics occurs when allergens enter the airways and are sensed and phagocytosed by dendritic cells, macrophages or neutrophils (antigen presenting cells, APCs). APCs present the allergen to CD4+ naïve T-cells. These naïve T-cells cells are induced to differentiate into TH2 Cells when their antigen specific T-cell receptor binds the major histocompatibility complex 2 (MHCII) on the APC which presents the antigen, the T-cell is co-activated by CD28 (Huang et al.

21 2002). Activated TH2 cells release Th2 cytokines defined as IL-4 (Swain et al. 1990), IL-5 (Kraft et al. 2002) and IL-13 (McKenzie et al. 1993). IL-4 induces differentiation of naïve T-Cells to TH2 cells and proliferation of T-cells propagating in a feed-forward manner but also induces B cell priming (Swain, Weinberg, English, et al. 1990; Le Gros et al. 1990). B cells release immunoglobulins when their B-cell receptors detect antigens. B-cell receptors are upregulated in the cell membrane by IL-4 and both IL-4 and IL-13 causes Ig switching to IgE and B-cell proliferation (Noelle et al. 1984; Punnonen et al. 1993). The released antibodies subsequently bind to mast cells via their Fc region interacting with the FcεRI receptor (Mallamacis et al. 1993).

Once atopy in asthma is induced repeated stimulation with the antigen will activate two processes, an early phase response and a late stage response. The early phase response is activated when mast cell bound antibodies bind their antigen causing the FcεRI receptors to cross link and initiate mast cell degranulation (Metzger et al. 1986). This happens in minutes and may last for up to two hours (Janeway 2001). Degranulating mast cells release histamine, leukotrienes, serotonin and serine proteases, after this initial release mast cells will also release prostaglanidins, leukotrienes, GM-CSF, TNFα, IL-8, IL- 14 and Il-4. Histamine and serotonin both cause bronchospasm (Leff 1982; Beretta et al. 1965). See Figure 1.1 for a full over view of the mechanisms of action of atopic asthma.

TH2 cells will be activated by APCs presenting antigens and this second allergic challenge will cause them to release IL-5. IL-5 induces eosinophil migration and as eosinophils migrate through the basement membrane and pass through the epithelium they release major basic protein and eosinophil peroxidase which damages the epithelia and causes bronchospasm (Gleich et al. 1979; Gundel et al. 1991; Flavahan et al. 1988). This process is known as the Late Asthmatic Response (LAR) and occurs between 3-8 hours after allergen challenge when high levels of eosinophilia are evoked (Demonchy et al. 1985; Cartier et al. 1982). Eosinophils also release leukotrienes such as LTC4 which is a highly potent bronchoconstrictor (Shaw et al. 1985; Adelroth

22 et al. 1986). Early studies using an anti-IL-5 monoclonal antibody (Mepolizumab, GSK) given to mild-moderate asthmatics who were not stratified according to having high sputum or blood eosinophils were able to limit eosinophilia without inhibiting the late airways response demonstrating our understanding of asthma is still not complete (Leckie et al. 2000). In patients with high levels of sputum and/or blood eosinophils, anti-IL-5 monoclonal Abs are very effective at reducing asthma exacerbations although the effect on lung function, airways hyperresponsiveness (AHR) and asthma symptoms are less pronounced (Pavord et al. 2012; Bel et al. 2014; Ortega et al. 2014).

Figure 1.1 Mechanisms of Atopic Asthma

Not all asthmatics are atopic, but non-atopic patients still suffer AHR in response to indirect stimuli such as cold air, exercise or hyperventilation or to direct stimuli such as histamine and methacholine (often used as a challenge to measure AHR) (Billen & Dupont 2008). Although atopy is absent in many patients with asthma, there is evidence that airways hyperresponsiveness, the

23 late airways response and eosinophilia are IgE-dependent (Kirby et al. 1986; Cockcroft et al. 1977; Flint et al. 1985; Demonchy et al. 1985; Cockcroft & Murdock 1987; Booij-Noord et al. 1971). Omalizumab a FCεRI Binding domain antibody suppressed early and late phase response and limited allergen activation by limiting IgE binding to mast cells (Fahy et al. 1997). In a longer study Omalizumab decreased asthma exacerbations, but not totally, showing that the role of T-cell and B-cell receptors and PAMPs on non- immune cells can trigger some reactions (Busse et al. 2001). Innate Lymphoid Cells (ILCs) which are a lymphoid cell, similar in lineage to T-cells and B-cells (Walker et al. 2013) but without antigen receptors, can be categorised in a similar manner to T-cells with Group 1 ILCs producing type 1 cytokines such as IFN-γ (Fuchs et al. 2013; Bernink et al. 2013; Klose et al. 2014) and Group 2 producing type 2 cytokines (Fort et al. 2001) such as IL-4, IL-5 and IL-13 with expression of GATA-3 (Moro et al. 2010; Price et al. 2010; Neill et al. 2010; Mjösberg et al. 2012). ILC3s resemble Th17 cells releasing IL-17A and IL-22 (Spits et al. 2013). A GWAS study suggested polymorphisms in IL-33, an activator of ILC2 cells and IL-1RL1, the IL-33 receptor, are linked to the development of asthma (Tamari et al. 2013; Grotenboer et al. 2013). Despite the fact these two receptors are on different chromosomes they are reproducibly identified in asthma GWAS studies and implicated in asthma pathogenesis (Oboki et al. 2010). IL-33 has also been shown to be released by bronchial epithelium in response to allergens (Oh et al. 2013). IL-33 and ILC2 cells have also been found to be increased in the lung lavage of asthmatics (Christianson et al. 2015). The Thymic Stromal Lymphopoietin (TSLP) receptor is found on ILCs and it’s activation enhances Type two responses by enhancing GATA3 expression in human ILC2 (Mjösberg et al. 2012). Recent data in a human allergen challenge model demonstrated positive effects of an anti-TSLP antibody (Gauvreau et al. 2014).

1.2 Treatments

24 Asthma in most patients is well controlled by treatment with bronchodilators, relievers and Inhaled Corticosteroids (ICS) controllers. Step 1 of the British Thoracic Society guidelines suggest mild intermittent asthma should be treated with short acting bronchodilator such as inhaled short-acting β2 agonists or inhaled ipratropium bromide, a muscarinic inhibitor to relieve symptoms as they occur (British Thoracic Society 2014). Step 2, for more moderate asthma (using β2 agonists three times a week), the suggested treatment is a preventative ICS which will act on the underlying inflammation rather than just the bronchoconstriction symptoms. The doses are titrated to the minimum effective dose taken twice daily. Other preventer therapies include leukotriene receptor antagonists which are useful for children under five years old who can’t use an inhaler and sodium cromoglycate, a mast cell stabiliser, with some benefit in a few adults (Chauhan & Ducharme 2012). If adults are taking high doses of ICS and still suffering from nocturnal awakenings or have no relief from asthma attacks then there are further add- on therapies to consider under Step 3 of the guidelines (British Thoracic

Society 2014). These include inhaled long acting β2 agonists (LABA) in combination with ICS, leukotriene receptor antagonists, theophylline (an anti- inflammatory xanthine) and slow release β2 agonist tablets. Combination ICS/LABA inhalers can help combat poor compliance in taking inhaled steroids and often improves outcomes (Murphy & Bender 2009). After this selection of treatments, if control is still inadequate, Step 4 begins with addition of a long acting muscarinic receptor antagonist such as tiotropium along with an ICS/LABA being of benefit (Peters et al. 2010). If the long acting muscarinic antagonist does not improve asthma symptoms or exacerbations then higher dose inhaled steroids are recommended. Finally at Step 5, if the increased inhaled steroid makes no improvement the use of continuous or frequent oral corticosteroids is recommended. Prolonged treatment with high dose oral steroids is an unsatisfactory situation as there are many side effects including glaucoma, striae, buffalo hump, high blood pressure, diabetes mellitus, hyperlipidaemia, decreased bone density, cataracts in children and steroid psychosis (Stanbury & Graham 1998).

25 However, it is often a balancing act between controlling asthma and these side effects.

A selection of monoclonal Ab treatments are now available which are able to improve outcomes in small sub-sets of asthmatic patients. Most of these are against Th2 asthma and it is important to have a specific predictive marker so that the thereapy is used in the patients most likely to respond to the treatment. The anti-IgE, Omalizumab, has significantly reduced the number of exacerbations and improved lung function in severe and moderate patients (Busse et al. 2001). Anti IL-5s such as Reslizumab and Mepolizumab show significant reductions in exacerbations and improvements in lung function in severe asthmatics with high blood eosinophilia (Ortega et al. 2014; Castro et al. 2015) with similar effects for Anti-Il-5-receptor (Bleecker et al. 2016a; Castro et al. 2014; FitzGerald et al. 2016)

Anti-IL-13 Lebrikizumab demonstrated an improvement in lung function in patients with poorly controlled asthma by ICS. The increase in FEV1 following Lebrikizumab treatment was 8.2% in patients with high periostin levels (a marker for Th2 asthma) compared with a 1.6% improvement in subjects with low periostin levels (Corren et al. 2011). Tralokinumab also shows Statistically significant reduction in exacerbations and improvements in lung function in patients with high periostin and dipeptidyl peptidase 4 (Brightling et al. 2015) However, due probably to problems with periostin as a predictive biomarker, the Phase 3 studies showed less efficacy (Noonan et al. 2013; Hanania et al. 2016). Periostin is a ligand for integrins and could be involved in other disease pathways, especially infection and also increases eosinophil binding (Bentley et al. 2014). Currently, high blood eosinophils are being used to predict Th2 asthma subjects suitable for mAb treatment – it is likely that more selective biomarkers will improve the responder:non- responder ratio. Anti-Il-4 has also reduced exacerbation and improved lung

26 function in moderate to severe patients with eosinic inflammation whilst tapering ICS therapy (Wenzel et al. 2013).

1.3 Severe Asthma

Though asthma is generally well controlled, severe asthma is a subset of asthma which has been defined since 2014 as ‘asthma that requires treatment with high dose ICS plus a second controller and/or systemic corticosteroids (CS) to prevent it from becoming "uncontrolled" or that it remains "uncontrolled" despite this therapy’ by the ERS/ATS guidelines (Chung et al. 2014).

Severe asthmatics have on average 2.5 exacerbations in 12 months as opposed to 0.4 in well-controlled mild-moderate asthmatics and these patients have higher levels of anxiety and depression (Shaw et al. 2015). As severe asthmatics have poorly controlled asthma they are more likely to become hospitalised with considerably higher healthcare and societal costs. In the Europe it has been shown that the mean costs for a patient with controlled asthma was €509 and for uncontrolled asthma was €2281 (Accordini et al. 2013). In 1992 in the US health care system annual cost for a mild patient was $1336, a moderate asthmatic $2407 and a severe asthmatic $6393 demonstrating the higher associated costs with severity of asthma and the indirect costs for all patients were double the cost of treatment (Serra-Batlles et al. 1998). Similar studies in the UK have shown that the cost of treating severe asthma accounts for 80% of the healthcare cost of asthma despite severe asthmatics only representing around 10% of asthmatics (Godard et al. 2002; Masoli et al. 2004).

Severe asthmatics are difficult to differentiate from asthmatics with poor compliance as they present with similar symptoms. Reported rates of nonadherence ranges from 30 to 70 percent (Global Initiative for Asthma 2016; Rand & Wise 1994; Bender & Bender 2005). In a study of Dutch asthmatics which selected patients based on compliance (chosen by selecting patients

27 who had 80% or greater prescription filling of ICS) only 3.6% of patients had severe asthma as opposed to “difficult to control asthma” which made up 17.4% of the total asthma population. Only 20% of the ‘difficult to control asthma’ population were adherent and had correct inhalation technique (Hekking et al. 2015). Despite this challenge for research several studies have begun to identify the mechanism of severe asthma. A 2013 paper has demonstrated it is possible to measure compliance of ICS usage using fractional exhaled Nitric Oxide measurements (FeNO) and may allow future studies and clinicians simpler ways of testing compliance (McNicholl et al. 2012).

1.3.1 Mechanisms and subtypes of Severe Asthma There are a large variety of mechanisms that have been proposed to cause severe asthma and this underpins the concept of severe asthma being heterogeneous and not a single disease. There are two clear severe asthma phenotypes, type-2 inflammation and non-type-2 inflammation. Type-2 inflammatory cytokines include IL-4, IL-5 and IL-13 and these are released from Type-2 T-helper cells, ILCs and Mast Cells among others. About 40- 50% of severe asthmatics have Th2 asthma based on blood, lung and airway eosinophilia and mast cell counts (Wenzel et al. 1999; Adelroth et al. 1986). Mild Th2 asthma is associated with allergy and IgE-mediated inflammation whereas the Th2 type inflammation in severe asthma is more complex as there is an increase in eosinophilic inflammation without necessarily a concomitant increase in serum IgE (Miranda et al. 2004). T Helper cells have been shown to be elevated in severe asthmatics with high eosinophilia (Wenzel et al. 1999) although as a group there appears little difference in the number of infiltrating cells within bronchial biopsies of severe asthmatics from the U-BIOPRED study compared to non-severe asthmatics or healthy controls (Wilson et al. 2016).

Type-2 associated severe asthma can also be divided into several phenotypes (Wenzel 2016). These include a severe early onset phenotype

28 linked to atopic conditions and high IgE, but without eosinophilia and a later onset, less atopic disease with systemic and airways eosinophilia (Miranda et al. 2004; Haldar et al. 2009). Both of these phenotypes have some degree of eosinophilia in the lung but it seems that the pathway of activation may be different. Recent studies have provided further evidence for this (Kuo et al. 2017). Studies in severe asthmatic patients treated with high dose ICS and/or OCS with high sputum and/or blood eosinophils clearly demonstrate a key role for eosinophils in severe asthma. Treatment with anti-IL-5 and anti-IL-5 receptor monoclonal antibodies decreased asthma exacerbations by nearly 50% with much less effect on lung function (Pavord et al. 2012; Haldar et al. 2009; Castro et al. 2011; Laviolette et al. 2013). Benralizumab a monocolonal antibody against the IL-5 receptor α has also shown to increase FEV1 pre- bronchodilator (Bleecker et al. 2016b).

Non-type-2 patients are often found in clusters linked to neutrophilic inflammation using cluster analysis. The pathways involved have not yet been elucidated although comorbidities such as obesity and smoking are often observed (Baines et al. 2011; Moore et al. 2014; Moore et al. 2010; Haldar et al. 2008). There is a suggestion that IL-6, a biomarker of inflammation, metabolic dysfunction and asthma severity may be associated with this subset (Peters et al. 2016). A subset of patients with increased airway neutrophilia tend to have similar blood eosinophilia and IgE levels to non-severe asthmatics (The ENFUMOSA Study 2003; Moore et al. 2007). There is also suggestion that Th17 phenotype of asthma may drive neutrophilic inflammation as IL-4 neutralisation increases IL-17 along with neutrophilia in mice. In contrast, when IL-13 and IL-17 are neutralised the mice are protected from eosinophilia and neutrophilic inflammation is ablated (Choy et al. 2015).

There are indications that some non-type-2 asthma could be driven by type-1 inflammation led by the canonical cytokine IFN-γ. Severe asthmatics as a group have increased IFN-γ expression in the airways compared to healthy or

29 mildly asthmatic subjects (Shannon et al. 2008; Voraphani et al. 2014). It has also been shown SNPs in the Th1/IL-12 genes (IL12A, IL12RB1, STAT4 and

IRF) are significantly associated with both FEV1 % predicted and asthma severity (X. Li et al. 2013).

Th17 Cells, a distinct population of IL-17A, IL-17E, IL-17F and IL-22 producing cells can mediate neutrophil migration in severe asthmatics by an IL-8-driven mechanism (Roussel et al. 2010). This is important as neutrophilic SA has been shown as a subgroup of severe asthmatics who are less responsive to CS (Wenzel et al. 1997; Green et al. 2002; Choy et al. 2015). These neutrophilic patients tended to also be low in airway eosinophilia suggesting another potential mechanism of action for severe asthma (Wenzel et al. 1997). TNF-α has been proposed as playing a role in neutrophil rich severe asthma as it increases AHR, mucus production and expression of adhesion molecules and in a synergistic manner with IL-17 it can enhance neutrophil recruitment (Green et al. 2002; Baines et al. 2011). A breakdown of the subtype markers is shown in Table 1.1. However, monoclonal Abs against both IL-17 (Brodalumab) and against TNFα (etanercept and infliximab) have either failed in asthma or produced deleterious side-effects (Busse et al. 2013; Holgate et al. 2011; Erin et al. 2006).

30 Inflammation

Phenotype Type-2 Non-Type 2 Early Onset Late Onset IgE High Present Blood Eosinophilia None Airway and Blood Blood Neutrophilia None None Airway Atopy High Present Present Smoking Not correlated Not correlated Correlated Obesity Not correlated Not correlated Correlated Table 1.1. Breakdown of Severe asthma phenotypes

Large asthma cohort studies such as the U-BIOPRED study in Europe and the Airways Disease Endotyping for Personalized Therapeutics (ADEPT) study in the USA have been used to corroborate each other (Loza et al., 2016; Lefadeau et al., JACI 2016). These studies have shown, using eight easy to measure clinical variables and cluster analysis that there are four clear groups of asthma. Cluster 1 patients are mild asthmatics not treated with ICS, well controlled, normal lung function and low airways inflammation. Cluster 2 subjects are partially-controlled, have mild airflow obstruction and severe airway hyperresponsiveness and a Th2 type phenotype. Subjects in cluster 3 are partially-controlled with mild airflow obstruction and reduced vital capacity, less bronchodilator reversibility and a non-Th2 phenotype with neutrophilia. Cluster 4 subjects are poorly controlled with high airflow obstruction and notable bronchodilator reversibility and a mixed inflammatory phenotype. In the ADEPT study these phenotypes were maintained for at least 12 months (Loza et al. 2016). These clusters are similar to those determined by Severe Asthma Research Program (Cohen et al. 2007; Moore et al. 2010; Moore et al. 2014). When studying gene set variation analysis (GSVA) signatures from the transcriptomics of bronchial biopsies and airway epithelial brushings 9 described signatures demonstrated two distinct asthma signatures associated with tissue eosinophilia, high Th2 type cytokines and a lack of CS response

31 (Kuo et al. 2016). The two non-steroid responsive groups were a group with high sub-mucosal eosinophilia with high exhaled nitric oxide, high exacerbations and high oral (CS) use. The other group had high BMI and the highest levels of eosinophilia of all groups determined. In contrast to these CS non-responders, the other groups had a much greater chance of having non- eosinophilic inflammation (Kuo et al. 2016).

1.4 Bronchial Epithelial Cells

A large number of cell types have been implicated in asthma pathogenesis. For the purposes of this thesis, I will focus on epithelial cells and T-cells only. The bronchial epithelium is frontline between environment and the body; it is both a physical barrier and possesses immunological functions (Folkerts & Nijkamp 1998). In most cases of asthma there are a variety of structural changes to the airways. Airway remodelling can consist of increased smooth muscle mass, thickening of the basement membrane, an increase of goblet cells and a breach in the integrity of the epithelium.

One of the key roles of epithelium is to form a physical barrier seperating microorganisms, gases and allergens from the body by creating tight and adherent junctions. This, combined with their ability to perform mucocillary clearance, allows epithelial cells to not only stop the transfer of foreign bodies into the body but to also actively remove them (Poynter 2012). When this boundary is broken allergens are more easily able to enter the body to trigger allergic responses, in addition intra-epithelial sensory nerves can be stimulated by the inhaled foreign bodies and carry out reflex bronchoconstriction (Folkerts & Nijkamp 1998). E-cadherin is an important molecule in the adherens junctions between epithelial cells and its expression is reduced in biopsies taken from asthmatics, suggesting a reduced boundary function in asthmatic patients (de Boer et al. 2008).

Airway epithelial cells also possess important immune and inflammatory functions. Epithelial cells have a pattern recognition receptors (PRR) on their

32 surface which can trigger the release of cytokines, chemokines and antimicrobial peptides such as interleukin 6 (IL-6) and thymic stromal lymphopoietin and β-defensins in response to cellular stress, bacterial infection or tissue damage (Ge et al. 2010; Soumelis et al. 2002; Duits et al. 2003). Epithelial cell activation is key to the recruitment of dendritic cells (DC) one of the APCs that proceed to activate the immune responses (Lambrecht & Hammad 2012). DCs are able to cross the epithelial boundary to sample the air entering the lung whilst maintaining the integrity of the boundary (Sung et al. 2006).

A common PRR that is found on epithelial cells is TLR-4 which along with being activated by LPS, a sugar structure associated with bacteria, can be activated by the allergens Der p2 and p7 from house dust mite or Fel D1 from cats demonstrating the epithelia’s direct sensitivity to allergens (Trompette et al. 2009). PRRs activate nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) which controls the expression of inflammatory genes. Constitutive activation of NF-κB in epithelial cells is enough for epithelial cells to activate DCs and to promote sensitisation to immunologically neutral allergens such as ovalbumin (OVA). Importantly, in mouse models of asthma inhibition of NF-κB reduces TH2 cell recruitment (Ather et al. 2011; Broide et al. 2005). Several NF-κB-regulated cytokines including CXCL8, RANTES and IL-6 are released from human airway epithelial cells in response to rhinoviral infection and to allergens such as the house dust mite proteases Der p1 and Der p9 (King et al. 1998; Chun et al. 2013).

The epithelium is able to attract DCs and monocytes by releasing chemokines including CCL2, CCL20 and IL-12p40 which also prime DCs to promote a TH2 type inflammation including TSLP (Walter et al. 2001; Nathan et al. 2009; Hammad et al. 2009; Ito et al. 2005). The release of TSLP in human bronchial epithelium from asthmatics is higher than from healthy subjects suggesting an epithelial-driven phenotype in asthma (Ying et al. 2005). Finally it is important to note that the epithelium isn’t exclusively inflammatory, it is

33 also able to supress DC reactivity and smooth muscle tone by releasing prostaglandin E2 (PGE2) (Nagarkar et al. 2012; Hartney et al. 2006).

Thymic Stromal Lymphopoetin (TSLP) has a key role in asthma is mainly created by the epithelium and stromal cells in the skin, gut and lungs(Kashyap et al. 2011; Redhu & Gounni 2012). It was shown to be vital to mice in developing allergic inflammation when TSLP receptors were knocked out and wild-type CD4+ T-cell transfer was able to restore allergic responses to OVA showing TSLPs role directly on T-cells (Al-Shami et al. 2005; Zhou et al. 2005). IL-33 is quickly released after epithelial damage and recruits a variety of Th2 type cells and will influence CD4+ differentiation (Komai-Koma et al. 2007)(Cohn et al. 1997). Mouse models of inflammation have been shown to IL-33 dependant in murine models of Ova induced allergy (Coyle et al. 1999; Townsend et al. 2000). IL-33 is able to activate mast cells to release IL-8, IL- 5, IL-13, IL-6, TNFα and GM-CSF (Allakhverdi et al. 2007; Moulin et al. 2007; Iikura et al. 2007). It has been posited that the detection of alleregen, virus or bacterial infection could begin the the asthmatic response to some extent by activating mast cells and eosinophils (Borish & Steinke 2011).

In summary, the epithelium plays a key role in the induction of allergic asthma via recognition of allergens and other damaging proteins and the release of important inflammatory mediators. It is also a very accessible target for specific drug delivery when using inhalers and may present a possibility for new treatments which should be investigated.

1.5 T-Cells

T-cells are lymphocytes that express T-cell receptors (TCR) which are capable of specific antigen binding and along with CD4 or CD8 receptors which stabilise the binding of the TCR to MHCII and MHCI respectively, activate the T-Cells leading to induction of the innate and adaptive immune responses. TCRs are always co-activated by either CD4 or CD8; T helper

34 cells which have a CD4 cell marker come in multiple forms, T helper 1 (Th1), T helper 2 (Th2), T helper 17 (Th17) and regulatory T-cells (Tregs) whereas cytotoxic T-cells have a CD8+ marker. An overview of these cell types activation, key transcription factors and cytokines that induce them can be seen in Table 1.2. Cell Inducing Transcription Cytokines Inhibitory Type Cytokine Factors Produced Cytokines Th1 IL-12, IL-18, IFN-γ T-Bet, STAT1 IFNgamma IL-4, IL-10 and TGF-β Th2 IL-2, IL-4, IL-5, IL-13 GATA3, STAT6 IL-4 IFN-gamma, TGF-β Th17 IL-6, TGF-β RoRγt, STAT3 IL-17A IL-4, IL-2, IL-27, IFN-γ Treg IL-2, TGF-β FOXP3, STAT5 TGF-Beta IL-17, I-23, IL-6

Table 1.2 T-cell Subtype Breakdown 1.5.1 T helper 1 cells The role of the Th1 cells is to combat bacterial and protozoa infection. Th1 cells do this by release of IFN-γ which will cause macrophage chemotaxis, activation and phagocytosis of bacteria and IgG switching in B Cells. Th1 cells can be activated by IL-2 which promotes TCR activation and cell proliferation, these effects are mediated by STAT4 and T-bet transcription factors. IL-12 is produced by dentritic cells and macrophages and also activates the STAT4 pathway via Tyk2 and JAK2 (Sugimoto et al. 2003)

STAT1 is induced by IFN-γ and induces T-box transcription factor TBX21 (T- bet) during Th1 differentiation (Afkarian et al. 2002; Lighvani et al. 2001) which can amplify the Th1 response by subsequently up-regulating IFN-γ in a positive feedback loop. However, in toxoplasmosis-infected Stat1 knock-out (KO) mice, IFN-γ-producing CD4 cells are still found, demonstrating that STAT1 has an element of redundancy (Lieberman et al. 2003). STAT2 forms a heterodimer with STAT1 in response to type I IFNs (Au-Yeung et al. 2013).

T-bet, induced by STAT1, is vital in for inducing Th1 reactions and IFN-γ production (Szabo et al. 2000). T-bet overexpression/activation can increase IFN-γ production whilst inhibiting IL-4 production in Th2 cells demonstrating the plasticity of Th cells. It does this by up-regulating IL-12Rβ2 expression

35 which promotes Th1 cell expansion and IFN-γ production. T-bet-expressing CD4+ T-cells are dramatically reduced in human asthma and may account for increased viral exacerbations alongside dysfunctional CD8+ T-cells (Finotto et al. 2002). Knockouts of the T-bet gene, Tbx21, result in major defects in Th1 responses in vivo and vitro (Szabo et al. 2002; Mullen et al. 2002; Mullen et al. 2001).

Similarly, in mouse models Tbx21 KO mice spontaneously develop AHS (Finotto et al. 2002). As Tbx KO cells can produce normal quantities of IFN-γ under IL-4 low conditions it maybe that the main role of T-bet is to inhibit GATA3 (Usui et al. 2006). However some studies have shown that T-bet alone may not supress GATA3 (Afkarian et al. 2002).

Knockout studies have shown that T-bet acts through enhancers to allow recruitment of P-TEFb and Mediator, the super elongation complex, to inhibit the release of Th1 target genes. STAT4 is activated by IL-12 and IFN-γ and is important in Th1 responses. STAT4 activation can be downregulated by the Th2 drivers IL-4 and GATA3 (Frucht et al. 2000; Thierfelder et al. 1996; Kaplan et al. 1996; Usui et al. 2003).

1.5.2 T helper 2 cells Th2 cells are effectors against parasites and helminths. They are triggered by IL-4, IL-5 and IL-13, which they also produce (described above) to drive the recruitment of eosinophils, basophils, mast cells and B cells creating a cytotoxic environment that destroys parasites. The Th2 environment also causes bronchoconstriction and cough to speed the exhalation of helminths or parasites during infection. The intracellular processes involved in activating Th2 cells include enhanced DNA binding by the transcription factors STAT6 and GATA3 (Wan 2014).

GATA3 is the Th2 master regulator (Zhang et al. 1997; Zheng & Flavell 1997) and vital for development of CD4+ T-cells. GATA3 is expressed by naïve

36 cells and is upregulated to enable Th2 differentiation and downregulated for Th1 cell development (Ouyang et al. 1998; Zhang et al. 1997; Usui et al. 2006). This was demonstrated using a dominant negative form of GATA3 which reduced Th2 cytokine expression and blocked airway hypersensitivity (Zhang et al. 1999).

STAT6 is the signal transducer for IL-4-mediated Th2 differentiation and expansion (Kaplan et al. 1996). STAT6 mediates it’s function by increasing expression of GATA3 (Zhu et al. 2001; Kurata et al. 1999; Ouyang et al. 1998). STAT5 has two isoforms a & b and blocking either causes major defects in Th2 cell differentiation (Kagami et al. 2000; Kagami et al. 2001; Zhu et al. 2003; Cote-Sierra et al. 2004; Moriggl et al. 1999). Low levels of STAT5 signalling are needed for proliferation and survival, however, for full Th2 differentiation strong STAT5 signalling is required (Zhu et al. 2003; Cote- Sierra et al. 2004). STAT5 activation is induced by IL-2 and is important for the further production of IL-2 and of IL4 (Burchill et al. 2008; Burchill et al. 2007; Davidson et al. 2007).

1.5.3 T helper 17 Cells Th17 cells are able to promote neutrophil infiltration and activation, potentially for the clearance of fungal and bacterial infections, however may also contribute to inflammatory diseases, including asthma. They release IL-17A, IL-17F, TNFα which causes the epithelial cells to release IL-6 and IL-8 which can recruit neutrophils (Chung et al. 2001). Th17 have been shown to be increased in the blood and airways tissue of severe asthmatics (Al-Ramli et al. 2009).

STAT3 is activated by IL-6, IL-21 and IL-23 and controls Th17 cell differentiation, amplification and maintenance. STAT3 induction by IL-6 also causes the downregulation of Treg cells (Veldhoen et al. 2006; Zhou et al. 2007; Nurieva et al. 2007; Korn et al. 2007; Bettelli et al. 2006; Mangan et al.

37 2006; Yang et al. 2007; Xu et al. 2007a; Yang et al. 2008). STAT 3 also increases RAR-related orphan receptor gamma t (RORγt) expression, a DNA binding transcription factor which directs the differentiation of Th17 cells (Wei et al. 2007; Zhou et al. 2007a; Ivanov et al. 2006)

1.5.4 Regulatory T-Cells Treg cells also have a CD4+ marker on their membrane but their role is limiting any deleterious T-cell activity (Sakaguchi et al. 2007). The primary function of Treg cells is to maintain self-tolerance towards other T-cells staving off autoimmune attacks (Sakaguchi et al. 1995). They are now known, however, to have many other functions including the suppression of asthma and allergy (Zuany-Amorim et al. 2002; Akbari et al. 2002; Curotto de Lafaille et al. 2001), inducing maternal tolerance to the foetus (Aluvihare et al. 2004) and supressing T-cell activation triggered by weak stimuli (Baecher- Allan et al. 2002). There have been questions raised about how to identify Treg cells as they have many markers which are not exclusive to themselves such as CD25 and cytotoxic T lymphocyte-associated antigen 4 (CTLA-4). FOXP3 which is generally considered to be the most reliable marker (Read et al. 2000; S Read et al. 2000; McNeill et al. 2007; Fontenot et al. 2005; Khattri et al. 2003; Fontenot et al. 2003; Hori et al. 2003; Sakaguchi et al. 1995) although FOXP3 can be found in some CD8+ T-cells (Fontenot et al. 2005). Tregs supress asthmatic responses; in a 2001 study, the high IgE response and AHR seen in mice sensitised and challenged with ovalbumin (OVA) was completely ablated by adoptive transfer of CD25+ and CD25- CD4+ Treg cells from unstimulated mice (Curotto de Lafaille et al. 2001). The results show that regulatory T-cells do not prevent hyper IgE production by turning naive OVA-specific T-cells to IFN-γ producing Th1 cells as both Th1 and Th2 responses were decreased when Tregs are infused into the mice.

Tregs work in an epitope specific manner to limit eosinophilia (Zuany-Amorim et al. 2002). It was also hypothesised that presentation of bacteria in early life

38 could protect against future asthma by class switching - turning Th2 cells to Th1 cells. Studies showed that in response to certain mycobacterial strains there was a shift from Th1 to Th2 inflammation in mice (Gajewski & Fitch 1988; Herz et al. 1998; Erb et al. 1998). It was finally shown that Mycobacterium vaccae gave rise to allergen-specific Treg cells showing that Tregs also have antibacterial properties which limit Th1 and Th2 inflammation in an epitope-specific manner which limited eosinophilia (Zuany-Amorim et al. 2002). The inhibitory abilities of Treg cells was attenuated by anti-IL-10 and anti-ICOS-ligand providing a mechanism for this effector function (Akbari et al. 2002). Mast cells can induce immunotolerance in allografts and limit host rejection. They are recruited by Tregs and release IL-9 which is essential for immunotolerance and is demonstrated in mast cell KO mice where this is not seen (Lu et al. 2006).

Patients and mice that are FOXP3 deficient lack Treg cells (Wildin et al. 2001; Patel 2001; Brunkow et al. 2001) and continuous expression of FOXP3 is vital for the continued suppressive effects of Tregs (Williams & Rudensky 2007). It has been suggested that rather than FOXP3 being a master regulator it may simply amplify and fix features of Treg cells (Gavin et al. 2007).

STAT5, which is vital for Th2 differentiation, is also vital for Treg differentiation and may bind to the Foxp3 promoter to enable Foxp3 induction (Davidson et al. 2007; Yao et al. 2007).

1.5.5 Cytotoxic T-Cells Cytotoxic T (Tc) cells have a CD8 marker on their surface and as the name suggest they are cytotoxic against aberrant self-expressing cells which have either viral or cancer cell markers on their surface. The TCR on Tc cells are only able to recognise antigens bound to the major histocompatibility complex I (MHCI) which is expressed by all non-professional antigen presenting cells, unlike Th cells, which can only recognise MHCII from APCs. To fully activate CD8+ T-cells a co-stimulatory signal is needed alongside the TCR. This

39 signal can be detected by CD28 receptor which is activated by either CD80 or CD86, both of which are found on the surface of the presenting cell. Once activated the T-cells can undergo clonal expansion in response to IL-2 (Sakaguchi et al. 2007). The activation of CD8+ T-cells causes a release of stored perforins, granzymes and granulysin which can puncture holes in the targeted cells causing cell death. However, they may also induce apoptosis by expressing the Fas ligand which can bind the Fas receptor on the affected cell.

Signal transducer and activator of transcription (STAT) proteins are vital for Tc fate determination. STAT activation is not controlled by large changes in gene expression but by cytokine-controlled post-translational modifications such as phosphorylation (Levy & Darnell Jr. 2002).

T-bet is important in the early stages of CD8+ cell development (Cruz-Guilloty et al. 2009) however, T-bet is not as important in developed CD8+ cells in so far as IFN-γ production is concerned. IFN-γ production remains normal in T- bet KO mice suggesting that T-bet is not the only mediator involved in CD8+ T-cell differentiation (Szabo et al. 2002). Eomes is another T-box transcription factor which is expressed in activated CD8+ cells. Double knockouts of T-bet and Eomes in CD8+ cells do largely limit IFN-γ production and these mice fail to control certain viral infections (Pearce et al. 2003; Owen et al. 2013; Nurse et al. 2000).

CD8+ T-cells activate GATA3 when their TCR is stimulated alongside cytokine stimulation much like when the combination of these stimuli activate Th2 cells. The importance of GATA3 in CD8+ T-cell function was demonstrated in GATA3 KO CD8+ T-cells which had defective responses to viral infections (Yunqi Wang et al. 2013).

40 1.5.6 The Role of Cytotoxic T-Cells in Asthma Though CD8+ T-cells were originally designated with an anti-cancer and antiviral function there is growing evidence that they have a role in allergy and asthma. CD8 cells have Th1-like and Th2-like differentiation, Cytotoxic T-cell type 1 (Tc1) and Cytotoxic T-cell type 2 (Tc2), as well there being regulatory Tc reg and naïve Tc0 cells. Between 80-90% of circulating Tc in the blood are Tc1 and these have been defined in culture as expressing IFN-γ and no detectable IL-4, the remaining circulating cells produce both IL-4 and IFN-γ (Vukmanovic-Stejic et al. 2000). A Tc2 profile can be induced by stimulating Tc cells with IL-2 and IL-4 which evoked no IFN-γ. Tc2 cells are also induced to produce IL-13 by the presence of IL-4 producing CD4+ Th2 cells (Koya et al. 2007).

A key piece of evidence to suggest cytotoxic T-cells have a role in asthma is a study of bronchial biopsies comparing lung function decline with CD4+ and CD8+ T-cell counts over 14 years (den Otter et al. 2016). This study showed a weak correlation between the rate of lung function decline and the CD4+ cell count at the start of the study and a strong correlation between lung function decline and CD8+ T-cell counts, both at the start and end of the study. Although the lung function decline in the asthma population is generally similar to that of the healthy population, asthmatics with chronic airflow limitation have a decline in FEV1 of around 50ml a year, similar to that seen in patients with chronic obstructive pulmonary disease (COPD) (Contoli et al. 2010). This confirmed a previous study which showed a link between asthma severity as measured by hyperresponsiveness to histamine and sputum CD8+ cells in atopic asthmatics (Cho et al. 2005).

The phenotype of the CD8+ T-cells in the lung was not investigated in the den Otter study, but it can be speculated based upon previous studies and on other information in the study by den Otter and colleagues (den Otter et al. 2016). These authors demonstrated enhanced levels of granzyme B in the patients with greatest fall in lung function and interestingly, levels of granzyme

41 B were also increased in fatal asthma (Annoni et al. 2015) which suggests a classic Tc1 anti-viral response, implicating a key role for viral infection in the decline in lung function in asthmatics. Natural Killer cells also produce this and are activated by IgE via the FcγRIII receptor (Arase et al. 2003). Asthmatics have a more severe clinical response to viral (RV16) infection than non-asthmatic subjects which may reflect the fact that Tc cells remain in the asthmatic airway longer than in normal subjects (Hogan et al. 2001). These cells may be Tc1-type effector-memory cells which remain active and patrol the airways (Contoli et al. 2006; Bedoui et al. 2016; Hogan et al. 2001). Wheezing episodes in young children caused by RSV infections are also associated with the development of asthma (Jackson et al. 2008).

In mice this persistent CD8+ activation can cause lung damage and in humans persistent CD8+ activation may be associated with epithelial damage and greater immigration of immune cells (Cannon et al. 1988). Some studies, but not all, have reported that epithelial cells from asthmatic subjects produce lower levels of IFNβ and λ than healthy subjects and are more likely to have unfavourable reactions to normally mild airway viral infections. The lower production of IFN implies that the Tc1 type antiviral CD8+ T-cells which produce IFN-γ could have an important antiviral role in asthmatic airways (Baraldo et al. 2012). It is of note that bacterial infections may also be more persistent in the airways. It has also been demonstrated that macrophages cultured from sputum of severe asthmatics have impaired phagocytic function of bacteria which may lead to increased exacerbations due to higher levels of colonisation of the airways and more persistent inflammation (Liang et al. 2014).

Leukotriene B4 acting through BLT1 receptors on Tc2 cells may induce IL-13 and the production of Th2 cells both of which cell types are correlated with asthma (Dakhama et al. 2013). IgE-mediated mast cell release of LTB4 recruits CD8+ T-cells into the lung which then release IL-13 (Taube et al.

42 2004; N. Miyahara et al. 2005; Nobuaki Miyahara et al. 2005; Taube et al. 2006; Gelfand & Dakhama 2006). Tc2 type cells can also be modulated by vitamin D which has been implicated in limiting asthmatic inflammation by inhibiting IL-4-induced IL-13 production and thus limiting IgE class switching and B cell proliferation (Hejazi et al. 2016; Schedel et al. 2016).

CD8+ cells release macrophage inflammatory protein 1α (MIP1α) also known as chemokine (C-C motif) ligand 3 (CCL3) when activated (Cocchi et al. 1995). MIP1α has been shown to be strongly chemotactic for murine eosinophils and macrophages suggesting that it is able to evoke both Th1- and Th2-type responses (Oliveira et al. 2002; Reichel et al. 2009). A study has shown strong links between the mediator of Treg cells, FOXP3, and serum MIP1α. Low concentrations of MIP1α increase the stability of FOXP3 and thus the activation of Treg immunosuppression, however high concentrations lead to instability of FOXP3 which would limit Treg function (Chen et al. 2013). This suggests a mechanism by which Treg cells may limit low levels of inflammation whilst allowing high levels. If in asthma there was an incorrect ‘cut off’ for high inflammation this could explain some effects of the disease. This disease model may apply in the lungs as well as the skin where there are high levels of CD8+ T-cells in asthma.

Circulating CD8+ T-cells have been shown to have highly differentially expressed genes in severe asthmatics. Despite the fact that there are no differences in CD4+ or CD8+ cell numbers in peripheral blood when comparing healthy, severe and non-severe asthmatic subjects (Tsitsiou et al. 2012). However, the CD8+ T-cells in severe asthmatic blood demonstrated 1359 genes upregulated by more than 1.5-fold and 207 downregulated by this amount compared to genes expressed in CD8+ cells from healthy subjects. In contrast with only 40 genes were upregulated and no genes were downregulated genes in CD4+ T-cells from severe asthmatics blood compared to healthy patients. Pathway analysis indicated that CD8+ T-cells had several activated pathways including CD28 T-cell stimulation, T-cell

43 adhesion, JAK-STAT, IL-4 and IL-9 signalling. A selection of miRNAs including key mIRs for T-cell differentiation and proliferation such as miR-146a and miR-146b (Lu et al. 2006), which were also significantly reduced in severe asthmatic CD4+ and CD8+ cells.

1.6 Genetics of Asthma

Asthma twin studies and polygenic heritability estimates show a heritable component to the disease (Thomsen et al. 2010; McGeachie et al. 2013). As such genome wide association studies look for specific single nucleotide polymorphisms (SNPs) which may account for differently functioning genes and proteins. However, limited genetic links to asthma have been found. The largest study of this type was reported in 2010 with 10,365 asthmatics and 16,110 unaffected controls. Significant asthma-associated SNPs were for genes encoding IL1Rl1/IL18R1, HLA-DQ, IL-33 and ORMDL3. /GSDMB. Only ORMDL3/GSDMB was associated with early onset asthma (Moffatt et al. 2010). Asthma is however a heterogeneous disease and it would have been surprising to find one polymorphism which causes asthma since the definition of asthma used in these genetic studies does not enable sub-phenotyping. In addition, although these SNPs and putative disease-associated genes are significant on a population scale they are by no means predictive of disease in an individual. Furthermore, gene expression may be regulated at different levels beyond the DNA sequence alone.

Despite the lack of SNPs associated with asthma it is possible to see clear differences in gene expression. For example, a genome-wide profiling study of epithelial cells comparing 42 asthmatics with healthy volunteers showed that the expression of genes coding for chloride channel, calcium-activated, family member (CLCA1), periostin and serine peptidase inhibitor, clade B (ovalbumin), member 2 (serpinB2) were up-regulated in asthma and that CS decreased their expression. However, despite CS upregulating the expression of FKBP51, a high basal expression of FKBP51 was linked with a poor response to CS and could potentially indicate severe asthmatics

44 (Woodruff et al. 2007). Similarly, as discussed in the section on the role of cytotoxic T-cells in asthma above, there are large changes in gene expression in CD8+ T-cells and some changes in CD4+ T-cells which demonstrate differences in the control of gene expression in asthma. This control of gene expression could be via a variety of epigenetic mechanisms.

1.7 Epigenetics

Epigenetic factors hereitably alter the expression of genes without changing the DNA sequence. These modifications can be controlled by the environment and the genome. Examples of these factors are DNA methylation, the methylation of cytosine residues on DNA which generally inhibits local gene expression (Rastogi et al. 2013), and histone modifications which alter the affinity of histones for DNA, act as tags for the recruitment of transcriptional co-factors/repressors and alter the ability of DNA to interact with RNA polymerase II enabling changes in gene transcription. Histone modifications can both increase (methylation on specific lysine residues) or decrease (acetylation) their affinity for DNA, alter the local chromatin structure and change the accessibility of binding sites for other transcription factors resulting in either gene repression or activation (Schones & Zhao 2008).

1.8 Histones Modification

A nucleosome is an octomeric complex of histones molecules which binds a 146 base pairs of DNA (Luger et al. 1997). The octomeric complex is made up of 2 pairs of the histones H2A and H2B and two pairs of H3 and H4, the nucleosomes being tethered by a histone H1 molecule (Ransom et al. 2010). The primary function of histones is the safe storage of DNA, allowing the long DNA chains to be organised into a more condensed space (Ransom et al. 2010). Histones bind tightly to DNA to create a structure which may limit transcription (Huisinga et al. 2006). Relaxing the interaction of histones with DNA results in a more ‘open’ chromatin state. The control of histone’s affinity for DNA and for transcription factors is controlled by posttranslational

45 modifications of the histone tails. Most data is focussed on modifications of histones H3 and H4, where methylation, acetylation, phosphorylation and ubiquitination of lysine and arginine residues regulates transcriptional activity (Tessarz & Kouzarides 2014). These modifications alter the molecular charge of the histone tails, since DNA is negatively charged an increase in the positive charge of the histone tail will increase the affinity between DNA and histones to produce heterochromatin (Dumuis-Kervabon et al. 1986). It is evident that other modifications on H2A and H2B may also play a role in the regulation of specific cellular functions although these will not be discussed here (Tessarz & Kouzarides 2014).

1.8.1 Histone Acetylation Acetylation increases the negative charge of the histones and causes repulsion of DNA from the histone creating a looser chromatin structure and allows increased DNA accessibility (Dumuis-Kervabon et al. 1986; Xu et al. 2007b). The acetylated histone tags also provide binding sites for bromodomain-containing proteins that transduce the acetylation signal to the basal transcriptional complex (Yang et al. 2005). Acetylation of multiple lysines on a tail of a histone can increase acetylation such as both the histone 3 lysine 9 (H3K9) and lysine 27 (H3K27) (Dumuis-Kervabon et al. 1986; Xu et al. 2007b).

Histone acetylation is controlled by histone acetyltransferases (HATs) and histone deacetylases (HDACs) and the effects of this modification are further transduced by acetyl readers (bromodomain-containing proteins) which integrate the downstream effects on basal transcription factor recruitment and activity and subsequent gene expression. HAT activity has been shown to be increased in cross-sectional studies in asthmatic biopsies in adults and children and the HAT/HDAC ratio alters with asthma severity (Ito et al. 2002; Su et al. 2008).

H3K27ac is associated with the enhancer regions of actively expressed genes (Creyghton et al. 2010). When combined with GWAS studies epigenetic

46 surveys can demonstrate which SNPs might be most important in disease (Gerasimova et al. 2013). For example, H3K27ac along with H3K4me1 has been shown to be enriched in non-coding SNPs of asthmatic CD4+ T-cells (Gerasimova et al. 2013).

It is also key to note that even with the siRNA depletion of HATs in drosophila cells histone acetylation was often constant and sometimes higher demonstrating that the acetylome is probably more fixed than variable although this may be cell- and species-dependent (Imhof et al. 2015).

1.8.1.1 HDACs HDACs 1 and 2 are important for the repair and remodelling of the airways. During development HDAC1/2 knockout mice affect the expression of the lung-specific progenitor Sox2 (Wang Y et al., 2013). However, HDAC1/2 ablation post-natally in airway epithelial cells does not affect Sox2 expression but increases the expression of the cell cycle regulators Rb1, p21/Cdkn1a and p16/Ink4a after injury causing cell cycle stalling and the loss of epithelial membrane barrier function (Y Wang et al. 2013).

Acetylation can modulate transcription factor binding to genes and acetylation of H3K9 and H3K27 can be read by proteins such as bromodomain containing proteins such as bromodomain and extraterminal domain proteins (BET) and cAMP response element-binding protein-binding protein (CBP) and Protein wise E1A binding protein p300 (EP300/p300)(Zeng & Zhou 2002). Brd4, a BET protein, has two bromodomain regions so can bring together two proteins with acetylated lysine such as acetylated histones and acetylated p65, one of the two elements of NF-κB complex. p65 activation is further enhanced by acetylation on K310 and Brd4 binding is important for controlling NF-κB activity in a cell-dependant manner (Belkina et al. 2013). Brd proteins can increase both proliferation and inflammation (Huang et al. 2009a; S. Y. Wu & Chiang 2007). Inhibition of STAT5-activated genes by targeting Brd function limits Th2 type inflammation (Pinz et al. 2015).

Trichostatin A (TSA), a non-selective class 1/2 HDAC inhibitor isolated from fungi (Yoshidas 1990) has been shown to enhance histone acetylation and increase inflammatory responses in vitro (Ito et al., 2002) but to limit IL-17 and T-helper cell numbers and increase Treg cell activation in vivo (Hou et al. 2014). An explanation of these seemingly divergent effects has been proposed whereby heightened genome-wide histone acetylation prevents correct sensing of gene-specific enhancement resulting in a global suppression rather than activation of gene expression (Bhadury et al. 2014). It is also suggested that the Th2/Th1 balance in asthma is shifted towards Th1 type inflammation by a non-selective histone deacetylase inhibitor (HDACi) in mice as it increases acetylation at the TGFβ promoter resulting in elevated BAL TGFβ protein levels (Hou et al. 2014; Choi et al. 2014).

In mouse models of allergic rhinitis HDAC1, HDAC3 and TSLP are upregulated and H3K9 acetylation is decreased. Treatment with sodium butyrate, another HDACi, in an OVA-induced mouse model of asthma, TSLP expression and the promoter-associated H3K9 acetylation levels were restored to normal (Wang et al. 2016). In addition, the levels of OVA-specific IgE were decreased by sodium butyrate (Wang et al. 2016). However H3K18, acetylation is differentially expressed in asthmatic epithelium at the ΔNp63, EGFR and STAT6 promoters and could not be modified with TSA. This suggests that some modifications are more permanent and have other layers of control over them (Stefanowicz et al. 2015).

The reliability of the HDAC inhibitor studies in vivo have been called into question as many studies have been unable to replicate the anti-inflammatory effects of these drugs (Banerjee et al. 2012; Kim et al. 2012; Kienzler et al. 2013; Shi et al. 2012) and others which have shown them to increase NF-κB mediated inflammation (S. Y. Wu & Chiang 2007; Royce 2012; Hou et al. 2014).

48 HDACs which reverse the acetylation status of histones can also affect other non-histone proteins such as the glucocorticoid receptor (GR). GR is acetylated on lysine residues K494 and K495 within the DNA binding domain of GR. Acetylation of these residues does not affect GR nuclear import, association with DNA at GREs (Ito et al., JEM 2006). Selective knockdown of HDAC2 resulted in increased GR acetylation and attenuated the ability of GR to associate with the p65 subunit of NF-κB and thereby block inflammatory gene expression. In addition, site-directed mutagenesis of K494 and K495 reduced GR acetylation and GR suppression of NF-κB-driven inflammation is no longer affected by histone deacetylase blockade (Ito et al., JEM 2006). This is important in severe asthmatics who are CS insensitive where there is often defective HDAC2 expression (Ito et al. 2006; Marwick et al. 2009). However, Butler and colleagues were unable to detect reduced HDAC2 in severe asthma biopsies and reported that increased epithelial remodelling associated with severity correlated with enhanced HDAC1 expression (Butler et al. 2012). The mechanisms of glucocorticosteroid anti-inflammatory actions are multi-factorial and include effects on both HAT and HDAC expression and activity and there is evidence that reduced HDAC2 activity is linked to steroid insensitivity (Royce & Karagiannis 2014). Theophylline, a weak bronchodilator drug used clinically, can enhance the activity of HDAC2 and is able to improve steroid sensitivity in vitro and in vivo (Barnes 2013).

1.8.1.2 Bromodomains Proteins Brd4 has two bromodomain regions which gives it the opportunity to bind two proteins with acetylated lysine and act as a bridge (S. Y. Wu & Chiang 2007). Brd4 has been shown to bind to the p65 subunit of NF-κB, acetylated histones and p-TEFB which in turn can bind RNA Pol II to promote transcription (Zhang et al. 2012).

CXCL8 gene promoter histone acetylation status is abnormal in airway smooth muscle cells from asthmatic patients. This has been associated with a specific increase in H3K18 acetylation due to enhanced EP300 promoter

49 binding. The binding of the acetylated histone reader proteins Brd3 and Brd4 to the CXCL8 promoter was also seen and accounts for the ability of Brd inhibitors to suppress CXCL8 expression and to limit local histone acetylation status (Clifford et al. 2015).

The ability of BET inhibitors to limit proliferation has been appealing for those looking to create anti-cancer drugs however, many such compounds such as JQ1 have relatively broad effects. The discovery of Brd inhibitors or mimics resulted from a research group researching small-molecule modulators of epigenetic targets as cancer therapeutics utilising high-throughput screening platforms for bromodomain isoforms (Frye 2010; Oprea et al. 2009). The first patent was from Mitsubishi Pharmaceuticals who demonstrated that simple thienodiazepines possessed binding activity for BRD4 (Miyoshi et al. 2009). Studying the structure-activity relationships and the binding pocket of Brd4 allowed the creating of a prototype ligand racemic molecule, JQ1, which is small and highly specific for Brds. Researchers have subsequently created each enantiomer separately providing a positive effective enantiomer and a suitable inactive control (Filippakopoulos et al. 2010).

JQ1 inhibits Brd-mediated functions including those involving NF-κB and p- TEFB by competitively binding the Lysine reader region on Brds (Filippakopoulos et al. 2010). An overview of interactions can be seen in Figure 1.2. The JQ1 molecule binds competitively to acetyl-lysine recognition motifs and may inhibit Brd4 binding to bromodomains leading to reduction in asthma relevant processes including IL-1β-induced inflammation (Filippakopoulos et al. 2010; Khan et al. 2014).

50 JQ1 p65 p50 pTEFb Brd4

RNA Pol II

Figure 1.2 Interactions of Brd4 Brd4 is able to bind using its two bromodomain regions to the p65 subunit of NF-κB which and also to pTEFb by acetylated lysine. This is able to mediate the pro-transcriptional nature of NF-κB. JQ1 competes to bind Brd4s BET acetylated lysine reader region.

It has also been suggested that most clinical trials using BET inhibitors have used them at a far too high a dose to be specific or biologically relevant, as JQ1 with a reported Kd of 50nM for Brd4 was used at 2-20mM to treat breast cancer cell lines and may risk having off-target and/or toxic effects (Andrieu et al. 2016; Shu et al. 2016).

Other off-target effects of BET mimics include reactivating latent HIV virus in carriers of the virus and as such these drugs could be dangerous in clinical use in at-risk individuals and more selective drugs should be developed (Z. Li et al. 2013).

1.8.1.3 CBP/p300 CBP and p300, are two homologous proteins which act as transcriptional coactivators, histone acetylases and have bromodomains. CBP/p300 has been shown to be important in cell growth, tumorigenesis, hematopoiesis and cell growth and differentiation (Goodman & Smolik 2000) and modulate gene expression in part through the acetylation of H3K18 and H3K27 (Jin et al. 2011).

51 Inhibition of these proteins inhibits the regulatory ability of Treg cells and has been suggested as a treatment for cancers which recruit Treg cells to limit cytotoxic T-cell killing of cancerous cells (Ghosh et al. 2016). When p300 is inhibited in Foxp3+ Tregs of mice, the cells demonstrate an increased anti- tumour function (Y. Liu et al. 2013).

CBP30 selectively binds to the Bromodomain of CBP/p300 and inhibits CBP/p300 functions (Hay et al. 2014). CBP has been shown to suppress Th17 responses including the release of IL-17A. CBP30 demonstrated a more restricted effect than the pan-BET inhibitor JQ1 and may suggest that more selective BET mimics may be more suitable for clinical develpment as they may result in fewer side effects (Hammitzsch et al. 2015). This concept needs to be tested in mouse models and, most importantly, clinically.

C646 is a highly potent histone acetyl-transferase inhibitor which is selective for p300 by targeting its Lys-CoA binding pocket at low concentrations and inhibits pro-inflammatory gene expression in murine macrophage lines by inhibiting the effects of acetylation at NF-κB sites in target gene promoters (Bowers et al. 2010; Van Den Bosch et al. 2016). Cbp/p300-interacting transactivator 2 coded for by the CITED2 gene is able to limit acetylation by removing p300 from activated genes under hypoxic conditions and targeting selective histone acetylation events might be a more useful way to treat chronic inflammatory diseases in the future despite the current limitations of this approach (Yoon et al. 2011). Targeting of p300 has also been achieved using CRISPR-Cas9 but this may not be a feasible approach in man where complete ablation is not required (Hilton et al. 2015).

There is evidence for cross-talk between different histone marks. For example, in Th2 cells expressing high levels of IL-13 and possessing high levels of H3K27ac at the IL13 locus, knockout of the DNA methyltransferase 3a resulted in increased IL-13 expression and greater levels of airway

52 inflammation in a mouse allergic airway inflammation model of asthma (Yu et al. 2012).

1.8.2 Histone Methylation There are multiple residues in the histone tail which can be either mono, di or tri-methylated. The functional effect can either be an increase or decrease in transcription depending on modified residue, the level of methylation (mono, di or tri) and where on the gene or promoter the modification is located (Young et al. 2011).

The promoter and enhancer regions of genes regulate the gene’s expression and histone methylation in these regions can both increase and decrease transcription. When the regulatory regions of a gene contain both pro- transcriptional and suppressor histone modifications this is known as a bivalent gene. Inhibitory histone modifications take priority until removed and, as such, the gene is considered as primed (Simon & Lange 2008). In the primed state the transcription complex may be present, but halted and unable to produce RNA. H3K4 di-methylation is often found poised at active enhancer regions (Koche et al. 2011). One of the clearest examples of this is on a single stretch of chromosome 5 where IL-4, IL-5 and IL-13 are coded along with the DNA repair protein RAD50 (Vijayanand et al. 2012). RAD50 has H3K4me1 at its four conserved enhancer regions in its introns and the levels of this modification is increased in T-cells from asthmatic patients suggesting higher basal histone methylation and transcriptional pause occurs for this gene (Seumois et al. 2014).

1.8.3 H3K4 Histone methylation is often associated with T-cell differentiation including CD4+ T-cell differentiation (Wei et al. 2009). The Il4/Il13 locus in mouse Th2 cells is tri-methylated on H3K4 suggesting that methylation activates the transcription of these genes. However, H3K27 tri-methylation is also found in

53 the same position in Th1 cells where it is associated with transcriptional repression (Wei et al. 2009). The mixed linage leukemia gene (MLL) is an H3K4-selective histone methyltransferase which is key for maintaining H3K4 modification at the Gata3 and Il4/Il13 locus in T memory cells. However, MLL does not seem to be involved in applying H3K4 methylation in differentiating Th2 cells demonstrating that multiple methylases are required depending upon the precise state of the cell (Yamashita et al. 2006).

Chromatin immunoprecipitation-sequencing (ChIP-Seq), a method to measure the location of every histone modification on a cells genome, which when combined with GWAS in Th1, Th2 and naïve T-cells shows that there is an up-regulation of H3K4me2 at SNPs within the promoters of regulatory regions of asthma-linked genes including CCR4 and CCL5 (Seumois et al. 2014). CCR4 is linked to Th2 recruitment to the lung and CCL5 is chemotactic for T- cells (Perros et al. 2009; Chan et al. 2012).

The H3K4me2 mark is localised to enhancers during Th2-cell differentiation which enables an asthmatic phenotype to prevail (Seumois et al. 2014). In contrast, increased transcription of both IFNG and IL4 is linked with H3K4me3 marks across stretches of the gene and in the promoter region (Wei et al. 2009).

Though there are currently no licensed drugs which target histone methylation there are new tool compounds which can be used in disease models. PFI-2 is a compound which targets the histone methyltransferase SET (domain containing) (SETD7) which is specific for H3K4. SETD7 plays a role in inflammation as it has been shown to activate inflammatory gene expression by inducing H3K4 methylation at NF-κB binding sites in cancer cell lines (Barsyte-Lovejoy et al. 2014; Wang et al. 2001).

54 1.8.4 H3K9 The H3K9me3 mark is associated with gene repression when localised to gene promoters by preventing RNA Pol II binding to these sites (Zhu et al. 2012). Vascular endothelial growth factor (VEGF) is a signal protein which is hyper-secreted by airway smooth muscle (ASM) cells in asthmatic airways and may account for airway remodelling which is a cardinal feature of asthma (Yuksel et al. 2013). In asthmatic ASM the H3K9 repressive complex at the promoter of the VEGF gene is decreased allowing for the hypersecretion. In healthy patients the G9a methyltransferase is vital for the repression of VEGF release (Clifford et al. 2012).

The H3K9me3 demethylase, JMJD2D, is able to activate transcription of Mdc and Il12b in dendritic cells and macrophages by removing the H3K9 methylation mark after exposure to an external stimulus (Zhu et al. 2012; Whetstine et al. 2006). UNC0642, an inhibitor of G9a histone 3 binding may be useful for future studies, however, inhibition of G9a generally could be dangerous in vivo as it is vital for embryogenesis (F. Liu et al. 2013; Clifford et al. 2012; Tachibana et al. 2002).

1.8.5 H3K27 The location of H3K27 methylation marks within a gene can vary its functional effects. H3K27me3 marks found in the center of the gene is associated with inhibition of gene expression, whilst when found at transcriptional start sites, H3K27me3 marks are associated with bivalent genes and transcriptional pause (Young et al. 2011). Finally, when localized to the promoter region of a gene it is associated with increased transcription of that gene (Young et al. 2011).

EZH2 is a histone lysine methyltransferase which specifically catalyses the tri- methylation of H3K27 and has been associated with gene repression effects of H3K27me3 (Simon & Lange 2008). EZH2 is vital for CD4+ differentiation

55 as it methylates H3K27 to block the production of IL-4 and IL13 in Th1 cells and prevent the expression of IFNG in Th2 cells. (Yamashita et al. 2006; Tumes et al. 2013). Adoptive transfer of Th2 cells from OVA-sensitised Ezh2 KO animals to wild type mice who were subsequently given an acute OVA challenge resulted in higher levels of eosinophilia, IL-4, IL-5 and IL-13, and mucus hypersecretion compared to mice which had been transferred Th2 cells from wild type mice (Tumes et al. 2013). There was also an overproduction of IFN-γ resulting in increased Th1 type inflammation (Tumes et al. 2013).

Two H3K27-specific demethylases, JMJD3 and tetratricopeptide, X chromosome (UTX), remove these repressive marks and are important for pro-inflammatory gene regulation and Th cell development (Miller et al. 2010; Xiang et al. 2007; Satoh et al. 2010). JMJD3 knockdown in the THP-1 monocyte cell line and in macrophages increased H3K27me3 marks at the promoters of NF-κB inducible genes leading to decreases inflammatory gene expression (Xiang et al. 2007; Ishii et al. 2009). Treatment of DCs with the JMJD3 inhibitor GSK-J4 (Kruidenier et al. 2012) has been shown to specifically attenuate Treg stability and suppressive effects without affecting Th1 and Th17 cells via dendritic cell suppression (Doñas et al. 2016).

IL-4 treatment of airway epithelial cells results in demethylation of the H3K27me3 mark at the ALOX15 promoter, thus increasing ALOX15 gene expression. The demethylation is under the control of UTX and is required for this IL-4 induced activation (Han et al. 2014; Kuperman et al. 2005). ALOX15 codes for a protein which oxygenates poly-unsaturated fatty acids to synthesise inflammatory signaling mediators and its expression is positively correlated with asthma severity (Kuperman et al. 2005). Whether this effect is causal or just correlative in asthma may be studied in future using GSK-J4. GSK-J1 and J4 have been shown to prevent LPS-induced inflammation in macrophages (Kruidenier et al. 2012).

56 Histone demethylation is essential for T-cell differentiation where large changes in epigenetic states occur. However, demethylases that inhibit repression such as JMJD3 are expressed in terminally differentiated cells suggesting that they have a broader role in mediating cellular responses. JMJD3 has additional roles in targeting histone modifications and regulating chromatin remodeling such as mediating the interaction T-bet and a Brg1- containing SWI/SNF DNA remodeling complex and so tool compounds which target protein may have off-target effects (Miller et al. 2010).

DNA Methylation DNA methylation is associated with gene repression and is found on complementary pairs of cysteine residues followed by a guanine (CpG) (Law & Jacobsen 2010). DNA methylation can control cell fate, inactivation of X chromosomes and gene specific activation and silencing. DNA methyltransferase (DNMT)1 maintains methylation marks between generations of cells as it converts hemi-methylated DNA into fully methylated DNA (Hermann et al. 2004). During the production of germ cells, cell-specific methylation is removed and then re-established by DNMT3A and 3B which create specific parental gene expression profiles in a process called genetic imprinting (Law & Jacobsen 2010).

The IL-4 promoter in naïve CD4+ T-cells is transcriptionally limited by methylated CpG regions. However, dust mite exposure decreases DNA methylation at the IL4 locus in cells from asthmatic subjects but not in cells from non-asthmatic subjects. This results in enhanced IL-4 release from asthmatic cells which is correlated with the reduction in DNA methylation and demonstrates active and specific demethylation (Kwon et al. 2008; Santangelo et al. 2009). To back up the change to Th2 type CD4+, the expression of Th1 type genes and proteins such as IFNG was blocked by dust mite exposure which was associated with increased IFNG promoter methylation (Brand et al. 2012). IFNG production can be reactivated in Th2 genes by demethylating the DNA of the IFNG gene, this is mediated by

57 GATA3 and T-Bet, demonstrating the extent to which the epigenome can be plastic (Williams et al. 2013).

In mouse OVA-driven mouse asthma models, DNMT1 expression has been shown to be self-regulating. Allergic stimulation decreased DNMT1 expression in the lung which was linked to DNMT1 promoter DNA methylation (Verma et al. 2013).

FOXP3 expression is controlled by DNA methylation and the methylation status of the FOXP3 promoter is able to discriminate Tregs from activated T- cells resulting from altered FOXP3 expression (Floess et al. 2007; Baron et al. 2007; Polansky et al. 2008). FOXP3-regulated genes are then themselves hypo-methylated in human T-Reg cells. CD4+ and Tregs had comparative methylated DNA sequencing carried out. There were 2315 differentially methylated cytosine–guanosine dinucleotides shown between Tregs and CD4+ cells in 127 regions of differential methylation (RDM). The RDMs for putative FOXP3 binding sites were hypomethylated in Tregs suggesting a link between the methylation and FOXP3 binding (Zhang et al. 2013).

Cross-talk can occur between epigenetic processes such as H3K27 methylation and DNA methylation. In zebra fish, the development of the lef1 gene, a neurogenesis regulator, is controlled by a network that includes Dnmt3 and G9a and these work together to silence cell fate regulators (Rai et al. 2010). Cross-talk also occurs between RNA and DNMT1 where a section of RNA from the CEBPA locus is able to interact with DNMT1 and prevent methylation of its own gene, thus increasing its own transcription (Di Ruscio et al. 2013).

The combination of GWAS studies and DNA methylation studies has demonstrated that there is an increased risk of asthma at age 18 with high DNA methylation at asthma relevant SNPs such as those at the IL-4R gene (Zhang et al. 2014). DNA methylation at 36 loci in peripheral blood

58 eosinophils is also strongly linked to serum IgE levels (Liang et al. 2015). These loci encode important eosinophil products and phospholipid inflammatory mediators and the unknown genes are potential new targets for treatment.

DNA methylation is vital to the activation of asthmatic immune responses and modifying the DNA methylome in certain cells may be of therapeutic benefit to some patients. Anti-cancer drugs have been developed which target DNMT1- Stat3 complexes giving a level of specificity to the targeted inhibition of the DNA methylase. In this instance, it was possible to limit the differentiation of Th17 cells without causing global hypomethylation which could lead to catastrophic effects on the epigenetic control of many cells (Cheray et al. 2014). The newly synthesised tool compounds from the Cheray, 2014 paper bind to DNMT1 complexed with either HDAC1, STAT3, PCNA, CFP1 or USP7 give a wide variety of specific agents with which to explore the relationship between DNA methylation and T-cell differentiation (Robertson et al. 2000; Cheray et al. 2014).

59 1.9 Hypothesis and Aims

The bronchial airway epithelium is a relatively accessible target for highly selective epigenetic modifying drugs. By studying the mechanism(s) by which the epithelial epigenome is primed in asthma and by testing epigenome modifying tool compounds on epithelial inflammation it will be possible to determine new targets for drug development. CD8+ T-cells appear to have a critical role in the severe-asthma phenotype with highly differentially expressed transcriptome. Therefore, I hypothesise histone modifications play a key role in regulating epithelial and CD8+ T-cell dysfunction in the severe asthma phenotype and by investigating the role of histone modifications other mechanisms of disease will become apperent. To investigate this hypothesis, I will undertake the following specific aims:

• To examine the transcriptomic and epigenetic status in the asthmatic airway epithelium and study their response to epigenome modifying compounds • To investigate the transcriptomic and epigenetic status of CD8+ T-cells in severe and non-severe asthmatics and study the effect of epigenome modifying compounds • To determine the pathways by which these epigenetic mechanisms and cellular functions are controlled using unbiased gene networking analysis

60 Chapter 2: Materials & Methods

61 2.1 General Methods

2.1.1 Enzyme Linked Immunosorbent Assays (ELISA)

CXCL8, IL-6 and MIP1α levels were measured using duo-set ELISA, following manufacturer’s instructions (R&D systems, Abingdon, UK). The manufacturer states that there is < 0.5% cross reactivity observed with available related molecules in all used ELISAs. In brief, 96-well Nunc Maxisorb Plate (eBiosciences, CA, USA.) were coated overnight at room temperature (RT) with the relevant capture antibody e.g. anti-CXCL8, then blocked (to prevent non-specific binding) with 1% bovine serum albumin (BSA) in PBS for 2 hours. Samples were added in duplicate to the wells alongside an eight point standard curve. Following 2 hour incubation at RT the plate was washed with PBS-Tween from powder (Sigma Aldrich) and detection antibody in PBS was added for two hours. The plates were washed again and Streptavidin-HRP added and incubated for 20 minutes (mins) at RT. After a final wash the Substrate Reagent Pack (TMB, BD Sciences, Erembodegem, Belgium) components were mixed 1:1 and added to the plate and the colour reaction stopped with 2N H2SO4 after 20 mins. The intensity of the colour change was measured by absorbance 450-570 (Bio Tec Instruments Microplate Reader & Gen 5 Analysis Software; Winooski, VT, USA). The values of the unknown samples were interpolated using the standard curve.

2.1.2 Viability Assay Cell viability was approximated by studying mitochondrial activity and was measured using (4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), which is soluble and able to enter a cell, where it is reduced to insoluble formazan. Formazan’s purple colour can be measured and was used to determine cell viability. 1mg/ml (final concentration) of MTT solution in media was added to the samples and incubated for 30 mins at 37°C after which the MTT solution was removed from the plate. The plate was then blotted and left to dry. Once dry 50µl of DMSO (Dimethyl Sulfoxide) was

62 added to each well, the plate was manually agitated for 30s and absorbance (570nm) was measured on a photo-spectrometer (Bio Tec Instruments Microplate Reader & Gen 5 Analysis Software; Winooski, VT, USA). Background absorbance was determined by adding MTT solution to wells with no cells and the measurements are normalised to untreated cells.

2.2 Cell proliferation Assay

Cell proliferation was determined using the BrdU assay, which measures the amount of bromodeoxiuridine (BrdU) incorporated into the DNA during synthesis in proliferating cells. The amount of incorporated BrdU can subsequently be measured by colorimetric analysis. The Cell Proliferation ELISA BrdU (Colorimetric) kit (Roche, Mannheim, Germany Cat: 11 647 229 001) was used following the manufactures protocol and is explained here in brief. Cells were seeded 1x104 cells/well in 96-well plates and incubated for 48 hours. Subsequently, the BrdU (100µM final concentration) was added to each well. After 24 hours, the media was removed, and FixDent solution added to each well after 30 mins at room temperature. It was removed and the Anti-BrdU-POD antibody is added (90 mins RT). The plate was washed with wash solution and finally 100µl of substrate solution is added. After incubating for 3 mins the luminescence was read on a photospectrometer (Bio Tec Instruments Microplate Reader & Gen 5 Analysis Software; Winooski, VT, USA). Background luminescence was controlled for by removing the value of control wells without BrdU from the read value of BrdU wells.

2.3 Human Bronchial Cell Models

2.3.1 BEAS-2B Culture BEAS-2B cells, a human virus transformed bronchial epithelial cell line, (ATCC, VA, USA) were grown in Keratinocyte Serum-Free Media (SFM) (Life Technologies, Paisley, UK) supplemented with Supplements for Keratinocyte- SFM (Endothelial Growth Factor (EGF) 2.5µg and Bovine Pituitary Extract

63 (BPE) 25mg) (Life Technologies) and incubated at 37°C with 5% CO2. They were passaged by washing with Phosphate Buffered Saline (PBS), detached by using trypsin (Sigma-Aldrich, Dorset, UK) and resuspended at 0.25 x106/ml. BEAS-2B cells were used between passages 10 and 20.

2.3.2 Normal Human Bronchial Epithelial Cell Culture Normal human bronchial epithelial (nHBE) cells (Lonza, Cat: CC-2540, Lot: 0000259108) were grown in Bronchial Epithelial Cell Basal Media (BEBM) with Bronchial Epithelial Cell Growth Media (BEGM) SingleQuots supplements (BPE, Insulin, Hydrocortisone, GA-1000 (Gentamicin and Amphotericin), Retinoic Acid, Transferrin, Triiodothyronine, Epinephrine, human EGF)

(Lonza) and incubated at 37°C with 5% CO2. Cells were grown to 70-80% confluence before passaging by washing with 4-(2-hydroxyethyl)-1- piperazineethanesulfonic acid (HEPES)-buffered saline, detaching from the flask using trypsin/ Ethylenediaminetetraacetic acid (EDTA) for 5 mins and neutralising with Trypsin Neutralising Solution (All Lonza). Cells were resuspended at 0.25x106/ml in 25ml of BEBM and used at passages 3-7.

2.3.3 BEAS-2B Epigenetic Probe/Mimic Treatments The effects of a variety of epigenetic probes or mimics obtained from the Structural Genomics Consortium (SGC) (Oxford, UK) (Table 2.1) on BEAS-2B cell proliferation and the release of CXCL8 in response to IL-1β stimulation was studied. BEAS-2B cells were grown until 70% confluent in a 96-well plate. Probes or mimics were diluted 1 in 2. The concentrations were selected based on those suggested by the SGC. Cells were pre-treated with the probe/mimic for two hours before being stimulated with 1ng/ml IL-1β for 24 hours. CXCL8 and IL- 6 release were measured by ELISA (see Enzyme Linked Immunosorbent Assays (ELISA) section), viability was measured by MTT assay (Viability Assay) and proliferation was measured using a BrdU assay (Cell proliferation Assay).

64 Probe/mi Top Concentration Cross reactivity Notes Target mic (nM)

UNC0638 G9a (EHMT2)/ G9a IC50 = <15nM 2048 probe GLP IC50 = 19nM UNC0642 G9a N/A (EHMT2)/GLP 512 (EHMT1) Probe UNC1999 Greater than 50% inhibition EZH2/1 Probe 2048 of 4 7TM targets @ 10 µM. SGC- 40 fold selective for CBP CBP30 CBP/p300 Probe 1136 over Brd4, next most efficacious target. Bromospo Has similar thermal melt rine temperatures for BRD2, BRD4 + CBP 4096 BRD3, BRD4, BRDT, CECR2, PB1, CBP GSK343 Minimum selectivity over EZH2 Probe 1136 EZH1 is 100 times more selective. GSK2801 Melatonin (MT-1) receptor BAZ2B/A 16384 4% inhibition. GSK-J1 JMJD3 IC50 = 0.06µM Jarid1b IC50 = 0.95µM JMJD3 Probe 256 Jarid1c IC50 = 1.76µM JMJD2e IC50 = 19.6µM GSK-J4 J1 Pro-drug 256 As above JSK-J5 GSK-4 Negative NA 256 Control

JQ1(+) BRD4 Kd = 49.0nM BRD3 Kd = 82.0nM BrdT + Brd4 1024 BRD2 Kd = 128.4nM

JQ1(-) JQ1(+) Negative NA 1024 Control

PFI-2 SETD7 IC50 = 2nM PFI-2 is highly selective (>1,000-fold) for SETD7, SETD7 512 over a panel of 18 other human protein methyltransferases and DNMT1.

Table 2.1. List of Epigenetic Probes/mimics used Probe compound, reported target and top concentration used in BEAS-2B cell response studies and cross reactivity notes from Structural Genome Consortium.

2.3.4 Treatment of nHBE Cells with Epigenetic Probes/mimics The effects of JQ1(+), JQ1(-), GSK-J4, GSK-J5 and Bromosporine on normal human bronchial epithelial release of CXCL8 in response to IL-1β was measured by ELISA. nHBEs were grown until 70% confluent in a 96-well plate. Probes were diluted 1:1. Cells were pre-treated with the probe/mimics for two hours before being stimulated with 1ng/ml IL-1β for 24 hours. CXCL8 release was measured by ELISA and viability was measured by MTT assay.

2.3.5 Effect of Probes/mimics in combination with a corticosteroid in BEAS-2B Cells The effect of epigenetic probes/mimics in combination with CS on the release of CXCL8 and viability in response to IL-1β in BEAS-2B cells was studied. Probes were diluted 1:1. Cells were pre-treated with the probe and fluticasone propionate (FP) (10-8M) a common inhaled corticosteroid for two hours before being stimulated with 1ng/ml IL-1β. CXCL8 release was measured by ELISA and viability was measured by MTT.

2.3.6 SILAC Proteomics Methods Stable isotope labelling by amino acids in cell culture (SILAC) is a proteomics method which integrates atomic isotope labelled amino acids into the proteins of cultured cells and uses mass spectrometry to measure the difference in ratios of proteins between treatment groups by comparing adjacent peaks in the spectrograph. In this case between unstimulated cells, IL-1β stimulated cells and IL-1β and FP-treated cells.

2.3.7 SILAC labelling and Cell culture 3 sets of normal human epithelial cells from different patients were obtained commercially (Lonza) from non-asthmatic patients (nHBE) and cultured in

66 bronchial epithelial growth media. Cells were grown to 80% confluence in a six well plate, and before stimulation were transferred to SILAC media for 5 days with media changed daily. The SILAC media was modified RPMI-1640 media including L-Glutamine and sodium bicarbonate but without L-Arginine (R), L-Leucine (L), L-lysine (K) or phenol red (Sigma-Aldrich) and supplemented with 10% dialysed (to remove unlabelled amino acids) foetal calf serum (FSC) (Sigma-Aldrich). Specifically labelled amino acids (Cambridge Isotope Laboratories Ltd, Andover, MA, USA) were added as described in Table 2.2. The cells were then treated with FP (10-8M) and IL-1β (1ng/ml) (both Sigma-Aldrich) for 24 hours. The cell work was carried out by Dr. Andrew Durham.

67 Group Leucine Lysine Arginine Control Unlabelled Unlabelled Unlabelled 2 13 IL-1β Unlabelled H4-Lys C6-Arg (DLM-2640-PK) (CLM-2265-H-PK) 13 15 13 15 IL-1β & Unlabelled C6 N2-Lys C6 N4-Arg FP NLM-291-H-PK) (CNLM-539-H-PK) Table 2.2. SILAC Treatment and Label for nHBECs Modified isotope amino acid SILAC labelling with each treatment. Labelled amino acids used for each group and their catalogue numbers (Cambridge Isotope Laboratories) are listed.

2.4 SILAC Methods

2.4.1 SILAC Protein Extraction After stimulation, media was removed and the cells were washed with PBS and lysed with Radioimmunoprecipitation (RIPA) buffer (Sigma-Aldrich) supplemented with protease and phosphatase inhibitors (Halt Protease Inhibitor Cocktail and Halt Phosphatase Inhibitor Cocktail; Thermo Scientific, Warrington, UK). Protein content of the whole cell extracts were quantified using the bininchoninic acid (BCA) assay to measure protein mass as following manufacturers methods. (Sigma Aldrich). Cells were normalised to the same concentration of protein and SILAC proteomic analysis was carried out.

2.4.2 Proteomics using mass spectroscopy Proteomics was carried out by Dr. Kate Heesom at Bristol University. Pooled Protein samples were fractionated on 10% SDS-PAGE gels. Using a ProGest automated digestion unit (Digilab, Marlborough, MA, USA) the gel lanes were given an in-gel tryptic digestion. The peptides were fractionated using a Dionex Ultimate 3000 nanoHPLC system in line with an LTQ-Orbitrap Velos mass spectrometer (Thermo Scientific). The Acclaim PepMap C18 nano-trap column (Dionex) had the resulting peptides in 1% (vol/vol) formic acid injected into it. The formic acid peptides were washed with 0.5% (vol/vol) acetonitrile 0.1% (vol/vol) and resolved on a 250mm × 75μm Acclaim PepMap C18 reverse phase analytical column (Dionex) over a 150min organic gradient,

68 using 7 gradient segments (1-6% solvent B over 1min., 6-15% B over 58 min., 15-32%B over 58min., 32-40%B over 3min., 40-90%B over 1min., held at 90%B for 6min and then reduced to 1%B over 1min) with a flow rate of 300nl min−1. Where Solvent A was 0.1% formic acid and Solvent B was aqueous 80% acetonitrile in 0.1% formic acid. Using a nano-electrospray ionization at 2.1 kV using a stainless-steel emitter with an internal diameter of 30 μm (Thermo Scientific) and a capillary temperature of 250°C peptides were ionised. Tandem mass spectra were captured using an LTQ-Orbitrap Velos mass spectrometer controlled by Xcalibur 2.1 software (Thermo Scientific) operated in data-dependent acquisition mode. The LTQ-Orbitrap was set to analyse the survey scans at 60,000 resolution (at m/z 400) in the mass range m/z 300 to 2000 and the top six multiply charged ions in each duty cycle selected for MS/MS in the LTQ linear ion trap. Charge state filtering, where unassigned precursor ions were not selected for fragmentation, and dynamic exclusion (repeat count, 1; repeat duration, 30s; exclusion list size, 500) were used. The conditions for fragmentation in the LTQ were as follows: normalized collision energy, 40%; activation q, 0.25; activation time 10ms; and minimum ion selection intensity, 500 counts.

Using Proteome Discoverer software v1.2 (Thermo Scientific) the raw data files were processed, quantified and searched against the UniProt/SwissProt

Human database using the SEQUEST (Version 28 Revision 13) algorithm.

Peptide precursor mass tolerance was set at 10ppm, and MS/MS tolerance was set at 0.8Da. Search criteria included carbamidomethylation of cysteine (+57.0214) as a fixed modification and oxidation of methionine (+15.9949) 2 13 13 15 and appropriate SILAC labels ( H4-Lys, C6-Arg for duplex and C6 N2-Lys 13 15 and C6 N4-Arg for triplex) as variable modifications. The reverse database search option was enabled and all peptide data was filtered to satisfy false discovery rate (FDR) of 5%. The minimum cross-correlation factor (Xcorr) filter was readjusted for each individual charge state separately to optimally meet the predetermined target FDR of 5% based on the number of random

69 false positive matches from the reverse decoy database. Thus, each data set has its own parsing parameters. Quantitation was performed using a mass precision of 2ppm. After extracting each ion chromatogram several filters were used to check for interfering peaks and the presence of the expected isotope pattern using the Proteome Discoverer software. Peptides which did not pass these filters were not used in calculating the final ratio for each protein. Results are presented as protein ratios and represent the median of the raw measured peptide ratios for each protein.

2.5 Histone Purification and Analysis

2.5.1 Global Histone Modification analysis nHBE cells were grown to confluence in 6-well plates. The nHBE cells were then treated with FP (10-8M) and stimulated after 2 hours with IL-1β (1ng/ml). The concentrations were governed by concentration response curves carried out in BEAS-2B cells. Global histone modifications are investigated by isolating the histones and then running purified histones on a multiplex ELISA to give relative concentrations and changes after treatment with a CS to represent the standard anti-inflammatory asthma treatment.

2.5.2 Histone Purification Histone purification was carried out using the Active Motif Histone Purification Kit (Active Motif, Brussels, Belgium. Cat: 102385) according to the manufacturer’s instructions, which will be briefly described here. The samples were extracted by washing the adhered cells with serum free media and scraped in the presence of 0.3ml of extraction buffer. The cells were pipetted to homogenise and incubated for 2 hours at 4°C. Following this the homogenate was spun at 12,000 xg for 5 mins and then the supernatant transferred to a new tube where it had a quarter of the original volume of extraction buffer added. To purify the histones the crude histones were incubated with 0.5ml of Equilibration buffer before being added to the spin columns and centrifuged at 500 g for 3 mins at 4°C. The run through was

70 removed and 0.5ml histone wash buffer was added to the column before centrifugation. The washing step was repeated 3 times. Finally, the columns were added to new tubes and 100µL of histone elution buffer was added to the columns before being spun. The final stage was to precipitate out the histones using 4% perchloic acid overnight at 4°C. The samples were subsequently spun at 12,000 g for 1 hour at 4°C. The pellet was washed with 1ml of cold 4% perchloric acid for 5 mins at full speed and the process repeated 3 times. Finally, the pellet was washed with acetone before being repelleted. The supernatant was then removed and the pellet allowed to air dry. Finally, the pellet was resuspended in 35µL of sterile water. The concentration of the histones was quantified using a photospectrometer measuring absorbance at 230nm when diluted 1:10 in water in a Corning Costar UV-Transparent Microplates plate. An OD of 0.42 indicated a protein concentration of 1mg/ml.

2.5.3 Histone H3 Post Transcriptional Modification Multiplex ELISA Changes in histone modification was measured using the Histone H3 PTM Multiplex Kit, following the manufacturer’s instructions (Active Motif, Brussels, Belgium Cat: 103952) and the Luminex Magipix Plate Reader using xPONENT for Magipix 4.2 software. The assay works by staining the specific modifications with different colour antibodies and measuring the fluorescent intensity of each wavelength. Buffer solution containing deacetylase inhibitor, protease inhibitor and phosphatase inhibitor was used to dilute the histones to 150ng/25ul. The multiplex bead master mix was made up with 1µL of each Ab-Conjugated Beads from Table 2.3 made up to 25µL with buffer for each well.

Antibody Conjugated Beads Catalogue Number Lot Number H3 total 33116 29113002 H3K4me3 33121 11213001 H3K9me1 33118 11213001 H3K9me2 33119 11213001 H3K9me3 33120 11213001 H3K9ac 33117 29312001 H3T11ph 33128 01001001 H3K27ac 33127 01001001 H3K27me2 33124 11213001 H3K27me3 33125 11213001 H3K56ac 33126 01001001 H3 pan-acetyl 33123 11213001 Table 2.3. Histone Modification Binding Conjugated Antibodies Conjugated Antibody beads for Histone Multiplex ELISA (All Active Motif).

1:1 of master mix and diluted histone sample or positive control were added to the wells to a total of 50µl. The wells were then covered with adhesive sealant and tin foil and incubated for 1 hour on an orbital shaker. Following this the plate was washed by setting the plate on a 96-well magnet for 60 seconds and washing with wash buffer and then removing the buffer by flicking whilst the plate was still attached to the magnet. This was repeated 3 times. Next 50µL diluted biotinylated antibody was added to each well, the plate covered and incubated for one hour on an orbital shaker. The wash steps were repeated and followed by adding 50µL diluted Streptavidin-PE to each well. The plate was covered and incubated again on an orbital shaker for 30 mins. The supernatant was removed using a 96-well magnet and the beads resuspended in 100µL of wash buffer. The plate was then read on a Magpix Luminex Plate Reader (Luminex Corp. Austin, Texas, USA) following the

72 instructions on the xPONENT software and normalised against total H3 histones.

2.6 Isolation of CD8+ T-Cells from Asthmatics

2.6.1 Patient Selection Criteria for CD8+ T-cells Collection Severe and non-severe asthmatics were selected following these inclusion: • All patients had a physician diagnosis of asthma, were aged 18-60, be a non-smoker or ex-smoker with a less than 5 pack year history. • Non-severe asthmatics were well controlled with minimal symptoms while on an ICS therapy not exceeding 2000µg beclomethasone equivalent. • Severe asthmatics had to either have had treatment with continuous or near continuous (>50% of the year) oral CS or a requirement for treatment with high dose inhaled steroids. As well as this they must have at least two of the following: • Requirement for daily treatment with controller medication in addition to ICS e.g. Long acting beta-adrenoreceptor agonists, theophylline or leukotriene antagonists • Asthma symptoms requiring short acting beta-adrenoreceptor agonists on a daily or near daily basis

• Persistent airways obstructions (FEV1 <80% predicted, diurnal PEF variation >20%) • One or more emergency care visits a year • 3 or more steroid “bursts” a year • Prompt deterioration in symptoms with < 25% reduction in oral or ICS • Near fatal asthma event in the past

The exclusion criteria for all patient types were: • Current smoker or less than 3 years since smoking cessation • Less than 4 weeks since the last exacerbation • On platelet or anti-coagulant drugs • Low platelet count

73 • On steroid sparing agent or immunosuppressant drugs such as azathioprine, methotrexate and cyclosporine. • Concomitant anti-IgE therapy • Pregnant or breastfeeding • Intubation for asthma within 6 months of entry into this study.

2.6.2 CD8+ Isolation Cells were extracted using Ficoll-Paque and negative selection using magnetic beads. The protocol is below described in brief.

2.6.3 Ficoll-Paque Isolation Blood was collected from patients and mixed with 1ml ACD (Acid Citrate Dextrose Solution)/10ml blood. ACD containing blood samples were mixed with PBS (1:1) and carefully layered onto Ficoll-Paque (GE Healthcare, Buckinghamshire, UK) and spun at 400g for 30 mins at RT with acceleration set to 5 and brake rate at 2. The layer containing the mononuclear cells was removed using Pasteur pipettes. The mononuclear cell layer was spun at 200g for 10 mins and the pellet resuspended in 5ml of ice cold water to lyse red blood cells before being centrifuged, resuspended in PBS and repelleted as before. Cells were counted and 250,000 cells were put in FACS fix (2% Formaldehyde in PBS).

2.6.4 CD8+ T-Cell Isolation CD8+ T-cells were enriched by removing non-CD8+-T-cells using the CD8+ T- cell Isolation Kit (Miltenyi Biotec, Surrey, UK). The T-cell isolation kit uses negative selection by using magnetic CD4, CD15, CD16, CD19, CD34, CD36, CD56, CD123, TCRγ/δ, and CD235a (Glycophorin A) binding antibodies to pull out all non-CD8+-T-cells.

The pellet was resuspended in Isolation Buffer (PBS, 0.5% BSA, 1:10 ACD) and 10µl of CD8+ T-cell Biotin-Antibody Cocktail per 107 cells was added and

74 incubated for 10 mins between 2-8°C. Then 20µl of CD8 T-cell MicroBead Cocktail was added per 107 cells added and incubated for 15 mins at 2-8°C. Then the solution was spun at 300g for 10 mins at 4°C and resuspended in 500µl of isolation buffer. The LS MACS Columns (Miltenyi Biotec) was put into the QuadroMACS Separator (Miltenyi Biotec) a magnetic stand and the cells were washed through with buffer. Cells were counted and 250,000 cells were put in FACS fix. When the cell count was lower than 2 million, half the cells were suspended in either 20% formaldehyde to crosslink DNA to histones and then quenched with glycine for ChIP-seq analysis and the other half were suspended in RTL lysis buffer (Qiagen, West Sussex, UK) and β- mercaptoethanol (Sigma-Aldrich) for RNA extraction for transcriptomic analysis by gene arrays. When >2 million cells were obtained, 1 million cells were set aside for gene array and the remainder was used for ChIP-Seq.

2.7 CD8+ T-cells Extraction and analysis

2.7.1.1 RNA Extraction RNA was extracted from the CD8+ T-cells with the RNeasy Mini Kit (Qiagen) following manufacturer’s instructions by using a silica-membrane which binds RNA when primed with ethanol and releases it in aqueous solutions. In brief cells were lysed by spinning for 3 mins in a QIAshredder lysis columns. The supernatant from this tube was added to the top of the extraction column after being mixed with 350ul of ethanol (70%) and spun, then with RW1 buffer and then with RPE buffer. The RNA is then eluted with RNase free water. RNA quality was run on a Pico RNA Bioanalyser when being used for gene array (Agilent Technologies, Kidlington, UK as described in 2.7.1.6).

2.7.1.2 Reverse Transcription and Amplification for Gene Array For gene arrays cDNA was created from mRNA using the cDNA Ovation Pico WTA System (NuGen, CA, USA). This was carried out as per manufacturer’s instructions and is described in brief here. The generation of the first strand of

75 cDNA was by using a proprietary mix of random and oligo dT primers so that priming occurs across the whole transcript and reverse transcriptase. The first cDNA/mRNA hybrid also contains a unique SPIA-RNA tag as a priming site for the SPIA process. The DNA was purified at this stage. The generation of a DNA/RNA Hetreoduplex Double Strand cDNA was carried out by fragmentation of the mRNA in the heteroduplex to allow DNA polymerase to synthesise a second cDNA strand with a complimentary SPIA-Tag. The final SPIA amplification step used the DNA/RNA chimeric primer, DNA polymerase and RNase H. RNase H removes the RNA portion of the heteroduplex SPIA tag revealing a binding site for the SPIA primer. DNA polymerase starts at the 3’ end of the primer and displaces the existing forward strand when synthesising cDNA. The chimeric SPIA primer reproduces the SPIA tag and creates a new substrate for RNase H and starts the next round of cDNA synthesis. This was repeated for a fast and high yield amplification. The SPIA cDNA was then purified with MinElute Reaction Cleanup Kit (Qiagen) and analysed using a Nano DNA Bioanalyser in order to assess quality and quantity of RNA.

2.7.1.3 Biotinylation The process of biotinylating amplified cDNA is to create labelled binding targets for the Genome Expression BeadChips using the Encore Biotin Module (Nugen, CA, USA). The manufacture’s instructions were followed but are explained in brief here. The uracil-DNA glycosylase is used to remove uracil bases which were incorporated in the amplification step creating an abasic site on the cDNA strand. This is where the biotin moiety will bind to the strand. The sample was then purified with MinElute Spin Column (Qiagen).

2.7.1.4 Gene Array The hybridization cocktail was prepared from the fragmented biotin-labelled cDNA by using the Affymetrix Hybridization, Wash and Stain kit

(ThermoFisher, Scientific, MA, USA) and hybridised to the Affymetrix

76 GeneChip Human Gene 1.0 ST Array (ThermoFisher Scientific) for 18 hours. Once hybridisation was completed washing and staining steps were performed on an Affymetrix GeneChip Fluidics Station 450 Affymetrix GeneChip Fluidics Station 450, and arrays were scanned with an Affymetrix GeneChip 3000 7G scanner (ThermoFisher Scientific). Microarray data normalization was carried out using a robust multichip average method and quality checks using Partek Genomics Suite 6.6 (Partek GS).

Partek software was also used to carry out a principal component analysis (PCA). This process find the sources of variability in a data set then is able to present it as values that represents groups of highly variable genes. The software takes all the variables, in this case genes, and plots them against each other, ranks them into orders of variability and then assembles groups of genes with similar variability as PC1, PC2 then PC3. The PCA is then plotted on a 3 axis graph allowing to see groupings of the patients.

2.7.1.5 WGCNA Analysis of CD8+ Data The WGCNA method (Zhang et al. 2005) within the WGCNA-Package in Bioconductor R (Langfelder & Horvath 2008) was used to find groups of significant genes that move together and were correlated to clinical outcomes. In short, a network is constructed using Pearson correlations which gives weighted values for each gene, then using a Topological Overlap Matrix (TOM) a module is identified. The eigengene (1st Principal Component) of each module is used to find a single representative of each module and this can then be taken on into further analysis of module significance with patient traits which here was clinical outcomes. As using all the data from the gene arrays would be too computationally intensive only the genes with p<0.05 by ANOVA in Partek (2042 genes) from gene arrays were used although this reduced the unbiased nature of the analysis. The reason genes were chosen like this (as opposed to by fold change) is because WGCNA will be able to group by small movements in expression, whereas if the genes were selected

77 by fold change then there would only be one large network created which would not show anything other than the genes put into the analysis.

2.7.1.6 RNA/DNA Method Bioanalyser Protocol DNA and RNA quality and quantity was measured with a High Sensitivity DNA Kit or a Nano (25-500ng/µl) or Pico (50–5000pg/μL) RNA Kit (Aligent Technologies CA, USA) depending on quantity of RNA. Manufacturer’s instructions were followed, but in brief are described below. Dye and gel reagents were mixed and used to prime the Assay chip by adding to the “Gel” well, and using the chip priming station to use air pressure to force the gel into the chip. Next the marker was added to all wells including the ladder well. Then 1µl of ladder or sample was added to each well, 12 samples were run per chip. After being vortexed the chip was placed in an Aligent 2100 Bioanalyser and run on the suitable setting to produce a RNA Integrity Number (RIN) number (Schroeder et al. 2006) and RNA quantity by using electrophoresis.

2.7.1.7 CD8+ Flow Cytometry Staining For confirmation of cell isolation and purity, flow cytometry was used. Cells were stored at 500,000 cells/ml in 0.5ml of FacsFix (4% Formaldehyde in PBS) from before and after the isolation step at 4°C and kept from light. Before running the cells through FACS the cells were centrifuged at 300 xg for 10 mins, resuspended in 100µL buffer and mixed with 10 µL of both CD8-PE (Milteny Biotec, Cat: 130-098-075) and CD3-FITC (Milteny Biotec, Cat:130- 098-162) for 10 mins in a refrigerator in the dark. This allowed the FACS machine to discriminate the CD8+ CD3+ cells. The cells were washed with 1ml of buffer, spun for 10 mins at 300 xg, the supernatant was removed and cells resuspended in 200µL of buffer in a FACS tube. The cells were finally run through a FACS Canto 3 (BD Bioscience) and the data was analysed using FlowJo (Treestar, USA).

78 2.7.1.8 PCR Confirmation of differentially expressed genes Real Time quantitative Polymerase Chain Reaction (RT-qPCR) was used to confirm fold change (FC) results from the gene arrays using a QuantitTect SYBRGreen (Qiagen) and Rotor-Gene 3000 Machine (Corbett Research, Crawley, UK) and analysed using Rotor-Gene 6 Software (Corbett Research).

Component Volume 2X QuantiTect SYBR Green PCR Master Mix 10µl Forward Primer (10µM) 1µl Backward Primer (10µM) 1µl RNase-Free Water 3µl Total Master Mix reaction volume 15µl cDNA Product 5µl Table 2.4. Components for Sybr Green RT-qPCR master mix

Gene Product Forward 5’-CTTAGAGGGACAAGTGGCG-3’ 18s Backward: 5’-ACGCTGAGCCAGTCAGTGTA-3’ Forward: 5’-ACCACCACCGTGTTCAACTT-3’ DUSP1 Backward: 5’-GAGGTCGTAATGGGGCTCTG-3’ Forward: 5’-TTCCAGGAGCGAGATCCCT-3’ GAPDH Backward: 5’-CACCCATGACGAACATGGG-3’ MPO Qiagen, QuantiTect Primer Assay – Cat No: QT00067613 DEF1B Qiagen, QuantiTect Primer Assay – Cat No: QT01529234 ZFP36 Qiagen, QuantiTect Primer Assay – Cat No: QT00091357 CD81 Qiagen, QuantiTect Primer Assay – Cat No: QT00010591 GGA1 Qiagen, QuantiTect Primer Assay – Cat No: QT01018213 NDUFB8 Qiagen, QuantiTect Primer Assay – Cat No: QT00088704 TAGAP Qiagen, QuantiTect Primer Assay – Cat No: QT01029770 HIST1H1C Qiagen, QuantiTect Primer Assay – Cat No: QT00200410 Table 2.5. PCR Primers for Sybr Green

79 The total reaction volume for each PCR was 20μl. See Table 2.4 for components used for Sybr Green Master Mix and Table 2.5 for primers used. To account for variations within the protocol, gene transcript levels of 18S or GAPDH rRNA, housekeeping genes, were also quantified and used for normalisation. The reaction involved an initial heat step of 95°C for 15 mins to activate the DNA polymerase followed by 30-40 cycles of denaturation (94°C for 15 seconds), annealing (55-60°C for 30 seconds) and extension (72°C for 30 seconds) using a Rotor-Gene 3000 Machine (Corbett Research, Crawley, UK) and analysed using Rotor-Gene 6 Software (Corbett Research; Australia).

2.8 Chromatin Immunoprecipitation-Sequencing (ChIP-seq)

2.8.1 Chromatin Preparation The chromatin was cross-linked with 37% formaldehyde prior to being flash frozen in liquid nitrogen. A ChIP protocol was used for immunoprecipitating Histone 3 Lysine 27 trimethylation-associated DNA. The cells were lysed by freeze/thawing and micrococcal nuclease (MNase, 10U/µl) (New England Biolabs Inc., MA, USA) was added to cell fragments in MNase Digest Buffer (Table 2.6) to break the DNA into mono-nucleosomes. This was carried out for 5 mins at 37°C and stopped with 20mM EDTA (pH 8.0). After spinning, the supernatant was stored and the pellet was resuspended for sonication in sonication buffer (Table 2.7). Sonication took place at 80% power on a Vibracell (ProScientific; NT, USA) with a 3mm probe for 5 bursts of 20 seconds whilst on ice and with at least 60 second cooling time between each burst. The sample was spun and resuspended in Dialysis buffer (Table 2.8).

An aliquot of a sample was taken, reverse cross-linked and DNA extracted as described below before being run on a DNA Bioanalyser chip to measure yield.

80 2.8.1.1 ChIP-Seq ChIP Protocol 10% of each sample was set aside to be used as an input. The sonicated DNA was mixed with 500µl RIPA buffer (Table 2.9) with 4µl anti-trimethyl- Histone H3 (Lys27) antibody (Merck Millipore, Darmstadt, Germany; Cat: 07- 449) and 25µl Novex Dynabeads Protein G Magnetic Beads (ThermoFisher Scientific) and incubated overnight at 4°C. The samples were put on a magnetic stand to pellet beads which were washed with wash buffer 1 (Table 2.10), wash buffer 2 (Table 2.11) and Tris-EDTA (TE) buffer. The beads were then resuspended in Elution Buffer (Table 2.12) and returned to the magnetic stand where the supernatant contained the eluted DNA. The DNA prepared by reverse crosslinking and DNA extraction as detailed below was amplified using the NEBNext ChIP-Seq Library Prep Master mix set.

Compound Concentration Sucrose 0.32M Tris-Cl (pH 7.5) 50mM

MgCl2 4mM

CaCl2 1mM Protease and Phosphatase Inhibitors (Table 2.13 and Table 2.14) Table 2.6. ChIP MNase Digest Buffer Components

Compound Concentration Hepes (pH 7.5) 50mM NaCl 140mM EDTA 1mM Triton 1% SDS 0.1% Sodium deoxycholate 0.1% Table 2.7. ChIP Sonication buffer

Compound Concentration Tris-Cl (pH 7.5) 1mM EDTA 0.2mM Protease and Phosphatase Inhibitors (Table 2.13 and Table 2.14) Table 2.8. ChIP Dialysis Buffer

82 Compound Concentration NaCl 140mM Tris (pH7.5) 10mM EDTA 1mM EGTA 0.5mM TX-100 1% SDS 0.01% Sodium Deoxycholate 0.1% Protease and Phosphatase Inhibitors (Table 2.13 and Table 2.14) Table 2.9. Modified RIPA Buffer

Compound Concentration Tris.Cl (pH 8.1) 20mM NaCl 50mM EDTA 2mM TX-100 1% SDS 0.1% Table 2.10. ChIP Wash Buffer 1

Compound Concentration Tris.Cl (pH 8.1) 10mM NaCl 150mM EDTA 1mM NP40 1% Na deoxycholate 1% LiCl 250 mM Table 2.11. ChIP Wash Buffer 2

83 Compound Concentration

NaHCO3 0.1M SDS 1% Table 2.12. ChIP Elution Buffer Compound Concentration AEBSF (4-(2-Aminoethyl) benzenesulfonyl fluoride 104mM hydrochloride) Aprotinin 80μM Leupeptin 2mM Bestatin 4mM Pepstatin 1.5mM E-64 1.4mM Table 2.13. Protease Inhibitors Cocktail 100 x stock (Sigma Aldrich, Dorset, UK)

Compound Concentration Imidazole 200mM Sodium Fluoride 100mM Sodium Moybdate 115mM Sodium Orthovanadate 100mM Sodium Tartrate Dihydrate 400mM Table 2.14. Calbiochem Phosphatase Inhibitors Cocktail Set II (Merck Millipore) 100x stock components.

2.8.1.2 Reverse Crosslinking The samples were reverse cross-linked by incubating for four hours with RNase A (10 μg/μl) and NaCl at a final concentration of 250mM at 65°C. Then incubated in the presence of proteinase K (0.5 μg/μl) at 37°C for 1hr.

84 2.8.1.3 DNA Extraction DNA was extracted using phenol:chloroform:isoamyl-alcohol (25:24:1) extraction. Equal volumes of the phenol-chloroform was added to each sample and mixed by vortexing. After settling the upper aqueous layer was removed and NaCl (250mM final concentration) was added along with 40µg of glycogen and an equal volume of iso-propanol. This was mixed and DNA precipitated at -80°C for 2 hours. After thawing, the sample was centrifuged at 18,000g for 15 mins at 4°C and the supernatant was discarded. The pellet was washed with 70% ethanol, microcentrifuged as above, the supernatant removed and the pellet allowed to air dry.

2.8.1.4 DNA Clean-up and Tagging for ChIP-Seq DNA cleanup was carried out using the NEBNext DNA Library Prep MasterMix Set for Illumina (New England BioLabs Inc.) following manufacturer’s instructions. The process of the kit is to “end-repair” or blunt- end the DNA fragments by chewing back 3’ and 5’ overhangs to allow for adenylation of the 3’ tail. The adenylated or dA-tailed DNA allows adaptor ligation. Adapters were attached as barcodes to the DNA so that multiple libraries can be run at once. The adaptors are loops and following the cutting of these loops using the USER enzyme, PCR can take place.

Between each step AMPure beads are used to wash the previous reagents away. These are polystyrene beads surrounded by a layer of paramagnetic (magnetic only in a magnetic field) magnetite. The paramagnetic quality stops the beads clumping and falling out of solution. The final layer of the bead is made of carboxyl molecules which allows the DNA to reversibly bind to the beads in the presence of polyethylene glycol (PEG) and NaCl (~20% PEG, 2.5M NaCl). The immobilisation is dependant on the concentration of PEG and NaCl.

The step following adapter ligation was used to size select only mono- nucleosomes with adapter so the protocol was followed for 200bp samples.

85 PCR Enrichment of Adaptor Ligated DNA used 15µl sample, 10µl Index/Universal Primer Mix and 25µl NEBNext Q5 Hot Start HiFi PCR Master mix for a total volume of 50µl. The cycling conditions are described in Table 2.15. Following this, a further AMPure cleanup step was performed and the amplified DNA was run on an Aligent Bioanlyser. The libraries with adequate DNA were sequenced on an Illumina HiSeq using single-end 50bp reads. Dr. Audrey Kelly helped greatly with these steps and the King’s Genomic Centre carried out the sequencing.

Cycle Step Temp Time Cycles Initial Denaturation 98°C 30 Seconds 1 Denaturation 98°C 10 Seconds 4 Annelaing/Extention 65°C 75 Seconds Final Extension 65°C 5 Seconds 1 Hold 4°C ∞ - Table 2.15. PCR Cycling Conditions PCR Cycling Conditions for NEBNexDNA Library Prep Amplification

2.8.1.5 ChIP-Seq Analysis To analyse the sequence data, Galaxy (https://usegalaxy.org, Version 2.3.5.0) (Blankenberg, Kuster, et al. 2010), an open source web based platform for data intensive biomedical research, was used. The workflow had six steps. 1. The FASTQ data from the Illumina HiSeq was converted to Sanger format using the FastQ Groomer (v.1.0.4) (Blankenberg, Gordon, et al. 2010). 2. The first 5 base pairs from sequences which often contain erroneous information was removed using the Trim Sequence (Galaxy 1.0.0) tool. 3. Over represented sequences were removed using Clip (Version 1.0.1)’ 4. Alignment of sequences to the hg19 version of the reference human genome was performed using Bowtie2 (Langmead & Salzberg 2012; Langmead et al. 2009). 5. FastQC (version 0.63) was used to provide quality control outputs for the runs.

86 6. Files were converted from the Galaxy Bam files to bigWig files using deepTools2 (Ramirez et al. 2016). The fragments were normalised to fragments per kilo base per million Reads Per Kilobase of transcript per Million mapped reads (RPKM). The bigwig files were displayed on the UCSC Genome Browser v338 and shown alongside H3K27ac peaks from 7 other studies (Table 2.16).

Colour Cell Type GEO Accession Number Red GM12878 - B-Lymphocyte GSM733771 Orange Human Embryonic Stem Cells GSM733718 Green Human Skeletal Muscle Myoblasts GSM733755 Blue HUVEC (Human Umbilical Vein Endothelial Cells) GSM733691 Indigo K562- Leukaemia GSM733656 Purple Human Epidermal Keratinocytes GSM733674 Pink Human Lung Fibroblasts GSM733646 Table 2.16. ChIP-Seq Traces for H3K27ac H3K27ac ChIP-seq traces shown alongside ChIP-seq data collected.

2.9 CD8+ T-Cell Activation Model

2.9.1 Cell Culture Methods for CD8+ T-Cells CD8+ T-cells were grown in Iscove’s Modified Dulbecco’s Medium (IMDM) (ATCC) supplemented with Human Albumin 2.5μg/ml (Sigma-Aldrich), D- Mannitol 0. 5μg/ml (Sigma-Aldrich) and 20% Foetal bovine serum. The media was regularly replenished with Human IL-2 (100units/ml diluted in 100mM Acetic Acid and 5% trehalose carrier protein). Cells were maintained at between 4x105 and 1x106 cells/ml at 37°C.

2.9.2 Stimulation CD8+ T-cells Suitable stimulation protocols to induce Macrophage Inflammatory Protein 1α (MIP1α) release in TALL-104 cells were studied in 96-well plate at 8x104 cells/well. Dynabeads® CD3/CD28 Human T-Activators (Thermofisher Scientific) were used at their recommended concentration of 4x107 Beads/ml. MIP1α was measured by ELISA as described in the Enzyme Linked Immunosorbent Assays (ELISA) section.

2.9.3 Dynabead Concentration response A submaximal concentration of Dynabeads was used to stimulate CD8+ T- cells in 96-well plates at 8x105 cells/ml. 7 1:2 steps from 1,600,000 dynabeads/ml were used. MIP1α was measured by ELISA as in section 2.1.1.

2.10 Epigenetic Probe/mimic concentration responses in CD8+ T-cells

CD8+ T-cells cells (8x105 cells/ml) were treated with a range of probe/mimic concentrations for one hour (Table 2.1) in a 96-well plate before being stimulated with Dynabeads (2x105 Beads/ml) for 24 hours. The supernatants were analysed for MIP1α release and cell viability was measured with MTT.

2.11 PCR and single gene CBP ChIP for CD8+ T-Cells

Cells were plated at 8x105 cells/ml in 6 well plates and treated with SGC- CBP30 (400nM) for one hour and stimulated with Dynabeads (2x105 Beads/ml) for 2 hours for PCR and ChIP analysis or for 24 hours for ChIP analysis. For PCR, RNA was extracted as in Section 2.7.1.1. RNA concentration was determined using a NanoDrop 2000c spectrophotometer and standardised to 500ng/µL. Reverse transcription to create single stranded cDNA was performed using a high-capacity cDNA kit (Applied Biosystems, Foster City, CA, USA), following the manufacturer's instructions. qPCR was performed in a Rotor-Gene 3000 PCR machine (Corbett Research, Cambridge, UK) using a QuantiTect SYBR Green PCR kit and normalised to 18S as described in Section 2.7.1.8. The genes expression of FOXP3, TUBA1A, JUND, DEFA1A, CD81, IFNG and CCL3 (MIP1α gene) (All QuantiTect from Qiagen) were measured. Data are presented as delta FC using the delta-delta CT (2-ΔΔCT) method.

88 To perform single gene ChIP, the MIlipore EZ-ChIP kit (Cat: 17-371, Merk Millipore, Watford, UK) was used exactly as described by the manufacturer. Protein-DNA was cross-linked using 1% formaldehyde before cells were lysed and the sample sonicated using eight 15 second bursts of 40% amplitude in a Vibracell Sonicator (ProScientific) to fragment the DNA. Sonication was performed on ice with at least a 30 second break between each burst to avoid heat damage. Aliquots of sonicated material were immunoprecipitated with 5ug of CBP antibody (Thermofisher Scientific; Cat No. PA1-847). The DNA input and output was run on q-PCR using the protocol above using selected primers for the CREB region of the FOXP3 DNA sequence (5’ Primer: GGGGTATCTGCCCTCTTCTC, 3’ Primer: TCCCTTTCTGACTGGGTTTC), the CREB region of CCL3 MIP1α gene (5’ Primer: CAATTCCATCCACATGACCA, 3’ Primer: TATCAATCCCCAAACCAGGA) and a region of the IFNG (5’ Primer: TCCCATGGGTTGTGTGTTTA, 3’ Primer: AAGCACCAGGCATGAAATCT). The data is shown as input-output ratio, where the input is DNA from the samples prior to immunoprecipitation. Data is displayed as individual patients at the 2 hours and 24 hours.

2.12 Statistical analysis

Results are expressed as mean ± standard error of mean (SEM) unless otherwise stated. EC50 values were calculated from concentration response curves, and considered to be the concentration of compound that inhibits or induces the effect by 50% of the maximal response. GraphPad Prism was used for statistical analysis, analysis of variance (ANOVA) was performed by Kruskal-Wallis analysis (non-parametric) and for significant results; Dunn’s post-test was used. For matched paired non-parametric multiple comparisons Freidman’s test was used with Dunn’s multiple comparison. p-values less than 0.05 were considered statistically significant. U-BIOPRED data was analysed on Array studio (Omicssoft Corporation, Cary, NC, USA) based on general linear model and with False discovery rate (FDR) adjusted p-value. WGCNA and gene array analysis is described under the corresponding

89 methodology heading. Pathway analysis was carried out using gProfiler

(Reimand et al. 2016).

90 Chapter 3: Epigenetic Mechanisms in Asthmatic Bronchial Epithelium

91 3.1 Chapter Introduction

Histone modifications, and their readers, have the potential to effect differential changes in gene expression that occurs between severe and non- severe asthmatics. Histone acetyl-transferase (HAT) activity is increased in children with increasing severity of asthma (Su et al. 2008) and enhanced histone acetylation allows for the binding of transcription factors and histone code readers such as Brd4, which is implicated in NF-κB p65-induced stimulation of inflammation (Belkina et al. 2013; Huang et al. 2009b; S.-Y. Wu & Chiang 2007). Tool compounds such as JQ1(+) and bromosporine inhibit the function of Bromodomain readers such as Brd4 and the more selective SGC-CBP30 is able to inhibit both CBP and EP300 bromodomain readers. Epithelial cells can respond to IL-4 by demethylating the H3K27me3 tagged ALOX15 gene thus increasing IL4 transcription (Han et al. 2014). ALOX15 codes for a protein which oxygenates polyunsaturated fatty acids to synthesise potent signalling mediators and it’s expression correlates with increasing severity of asthma (Kuperman et al. 2005). ALOX15 promoter demethylation is controlled by JMJD3/UTX which allows for inhibition with GSK-J1 and GSK-J4 (Kruidenier et al. 2012).

The bronchial epithelium acts as frontline between environment and the body. Not only is it a physical barrier, but it also possesses immunological functions. Immunological functions relevant to asthma include the induction of proliferation of CD4+ T-cells by release of TSLP (Soumelis et al. 2002) and IL- 6 (Poynter 2012). The airway is also able to respond to and release IL-13; a mediator of lung repair and central cytokine in Th2-driven allergic asthma (Woodruff et al. 2009). The bronchial epithelium is also able to mount innate responses against pathogens in releasing defensins, prostaglandins and interferons (Bals & Hiemstra 2004).

Oral and inhaled CS are gold standard treatments for severe asthma, however in severe asthma there can be a relative steroid insensitivity. In steroid-sensitive asthmatics CS are able to inhibit release of inflammatory

92 mediators via mechanisms including inhibition of histone acetyl transferase expression and activity and by recruiting and/or activating histone deacetylase (HDAC), these events are less evident in severe asthmatics in response to GSC (Barnes et al. 2005).

Leverage of pre-existing ‘-omics’ data sets is a useful way to identify the most relevant histone modifications in asthma before moving to more targeted studies designed for studying histone modifications. The data sets that will be explored in this chapter are from Unbiased BIOmarkers in PREDiction of respiratory disease outcomes (U-BIOPRED) a large clinical asthma study where samples including blood, airway brushings, nasal brushings, induced sputum and lung biopsies were taken from a large cohort of severe asthmatics, non-severe asthmatics and healthy subjects. Gene array analysis of bronchial biopsies and brushings was carried out in initial studies. I also used a proteomics dataset obtained from normal human bronchial cells (HBECs) that were stimulated with IL-1β in the presence and absence of fluticasone propionate and analysed using SILAC (Stable Isotope labelling with amino acids in culture). These approaches may provide evidence for the role of some epigenetic mechanisms in asthma and in response to GCS.

3.2 Chapter Hypothesis and Aims

Epigenetic mechanisms, including histone acetylation and methylation in bronchial epithelial cells will be associated with asthmatic cell traits and responses to anti-inflammatory GCS (and proxies for it). As a result, epithelial cells from severe asthmatics will be defined by a specific pattern of histone 3 lysine 27 acetylation and tri-methylation. These epigenetic mechanisms can be altered by GCs and inhibitors of epigenetic mechanisms. This hypothesis will be tested by examining whether: • The expression of epigenetic tags or enzymes are linked to severe asthma by analysing gene arrays from asthmatics, and non-asthmatic patients bronchial brushings and biopsies to measure their expression,

93 • Glucocorticosteroids affect the expression of epigenetic tags/enzymes in proteomic examinations of bronchial epithelium and • Exposure to epigenetic probes affects inflammatory gene expression and release of inflammatory proteins in bronchial epithelial cell lines.

3.3 Results 3.3.1 Gene Expression of Histones and Histone Modifying Enzymes in Asthma 3.3.1.1 Bronchial Brushings and Biopsy Gene Arrays from U-BIOPRED Data The Unbiased BIOmarkers in PREDiction of respiratory disease outcomes (U- BIOPRED) Consortium is an international study from which large quantities of patient samples have been collected and analysed in an attempt to improve diagnosis and treatment for individual asthma patients (Kuo et al. 2016). The consortium collected and analysed bronchial biopsies from 40 non-smoking severe asthmatics (SAn), 28 non-severe, non-smoking asthmatics (NSA) and 26 normal healthy volunteers (NHV). Gene arrays were also run on bronchial brushings from 49 SAn, 36 NSA and 44 NHV. This data has been tabulated here (Table 3.2) to show genes relating to histones and the epigenetics modifying enzymes for histones that are differentially expressed ( p-value < 0.05). The genes of interest are shown in Table 3.1. Expression data of bronchial biopsies and epithelial brushing from Unbiased BIOmarkers in Prediction of REspiratory Disease outcomes (U-BIOPRED) Project is available within Gene Expression Omnibus at GSE76227.

The aim of this exploration of expression data is to search for genes which code for histones and histone modifying enzymes in asthma. If specific enzymes are shown to be differentially expressed in asthmatic patients they could be targeted in further experiments to study the role of the expressed proteins on the asthmatic phenotypes.

94 bromodomain adjacent to zinc finger domain, 1A histone cluster 1, H1d histone deacetylase 7 Bromodomain adjacent to zinc finger domain, 1B histone cluster 1, H1e histone deacetylase 9 bromodomain adjacent to zinc finger domain, 2A histone cluster 1, H2ac histone H4 transcription factor bromodomain containing 4 histone cluster 1, H2ag jumonji C domain containing histone demethylase 1 homolog D bromodomain, testis-specific histone cluster 1, H2ak Jumonji domain containing 1C E1A binding protein p300 histone cluster 1, H2ak jumonji domain containing 8 enhancer of zeste homolog 1 histone cluster 1, H2bb lysine (K)-specific demethylase 1B euchromatic histone-lysine N-methyltransferase 1 histone cluster 1, H2bj lysine (K)-specific demethylase 2A Euchromatic histone-lysine N-methyltransferase 2 histone cluster 1, H2bk lysine (K)-specific demethylase 3B H1 histone family, member N, testis-specific histone cluster 1, H3 Lysine (K)-specific demethylase 4B H1 histone family, member X Histone cluster 1, H3e lysine (K)-specific demethylase 4C H1 histone family, member X histone cluster 1, H3h lysine (K)-specific demethylase 4D H2A histone family, member B histone cluster 1, H4a lysine (K)-specific demethylase 5A H2A histone family, member V histone cluster 2, H2aa3 lysine (K)-specific demethylase 5B H2A histone family, member X histone cluster 2, H2aa4 lysine (K)-specific demethylase 5C H2A histone family, member Y Histone cluster 3, H2a lysine (K)-specific demethylase 6A H2B histone family, member M histone cluster 3, H3 lysine (K)-specific demethylase 6B H2B histone family, member X, pseudogene histone deacetylase 10 Lysine (K)-specific demethylase 8 H3 histone, family 3A /// H3 histone, family 3B (H3.3B) histone deacetylase 11 SET domain containing (lysine methyltransferase) 7 HIR histone cell cycle regulation defective homolog A histone deacetylase 2 SNF2 histone linker PHD RING helicase, E3 ubiquitin protein ligase histone cluster 1, H1c histone deacetylase 4 Table 3.1 Table of Genes of interest for Bronchial Biopsies and Brushing

3.3.1.2 Investigation of gene expression in Bronchial Brushings U- BIOPRED data The data shows more significant differences between SAn and NHV subjects (Table 3.2) than between NHV and NSA subjects (Table 3.4) and between San and NSA volunteers (Table 3.3). When comparing NHV to SAn, the histone cluster H1e is significantly decreased with a log2 FC of 1.39 and an FDR corrected p-value of 0.022 (Table 3.2). In contrast, two other histone clusters are significantly increased, H2bk and H2bj, with a log2 FC of 1.38 and

1.25 respectively (Table 3.2). H2bk is marginally increased with a log2 FC of 1.22 between NSA and SAn along with a small but significant decrease in the E1A binding protein p300 (Table 3.3). H2bj also shows a small but significant increase between NHV and NSA subjects (Table 3.4).

3.3.1.3 Investigation of gene expression in the Bronchial Biopsies U- BIOPRED data Comparing NHV and NSA subjects showed no differentially expressed histone or histone modifying genes. When comparing NHV and SAn subjects we see an increase in Histone Cluster 1, H1a (log2 FC=1.27 P-Value = 0.0264) and

Histone Cluster 1, H2bk (log2 FC=1.23 P-Value = 0.0479) (Table 3.5). We also see a decrease in expression of Cbp/p300 interacting transactivator (log2 FC= -1.24 P-Value = 0.0705, Table 3.5) which conversely is significantly decreased from moderate to severe asthmatics (log2 FC= -1.23, P-Value = 0.038, Table 3.6). We also see an increase in expression of large selection of histone related genes between NSA and SAn including; Histone Cluster1, H1a, H2bk and 3A (Table 3.6).

96 Healthy vs Severe Healthy vs Severe Healthy vs Severe Gene Symbol Gene Title asthma asthma asthma Log2 FC Raw p-value FDR -1.3849 0.0006 0.022 HIST1H1E Histone cluster 1, H1e 1.3761 0.0012 0.0316 HIST1H2BK Histone cluster 1, H2bk 1.2483 0.0024 0.0462 HIST1H2BJ Histone cluster 1, H2bj 1.1153 0.0054 0.0688 BRDT Bromodomain, testis-specific -1.1117 0.0114 0.0988 EP300 E1A binding protein p300 -1.17365 0.011925 0.094875 BAZ1B Bromodomain adjacent to zinc finger domain, 1B -1.1951 0.0141 0.1104 SHPRH SNF2 histone linker PHD RING helicase 1.0882 0.0225 0.1412 HIST3H3 Histone cluster 3, H3 -1.12695 0.0234 0.1426 KDM3B Lysine (K)-specific demethylase 3B -1.18554 0.0317 0.15394 KDM5A Lysine (K)-specific demethylase 5A -1.15355 0.0403 0.155325 H3F3A H3 histone, family 3A 1.1477 0.0415 0.195 HIST1H2AK Histone cluster 1, H2ak 1.1022 0.0495 0.20935 HDAC10 Histone deacetylase 10 Table 3.2. DEGs from Bronchial Brushings Healthy vs. Severe Asthmatics Differential expressed genes encoding histones and histone modifying enzymes in bronchial brushings of Healthy vs. Severe non- smoking Asthmatics. Highlighting histone and histone modifying genes which show a raw p<0.05. Red values indicate adjusted

p<0.05. Log2 FC – fold-change, FDR –false discovery rate corrected p-value. n=49 non-smoking severe asthmatics (SAn), n=36 non-severe, non-smoking asthmatics (NSA) and 44 normal healthy volunteers (NHV).

Moderate vs Severe asthma Moderate vs Severe asthma Moderate vs Severe asthma Gene Symbol Gene Title Log2 FC Raw p-value FDR -1.1449 0.0018 0.7383 EP300 E1A binding protein p300 1.0815 0.0392 0.8178 HIST3H3 Histone cluster 3, H3 1.2261 0.0421 0.8189 HIST1H2BK Histone cluster 1, H2bk Table 3.3. DEGs from Bronchial Brushings Moderate vs. Severe Asthmatics Moderate vs Severe Asthmatics Gene Array on Bronchial Brushings from the U-BIOPRED consortium. Highlighting histone and histone modifying genes which show raw p<0.05. Log2 FC – fold-change, FDR – false discovery rate corrected p-value.

Healthy vs. Moderate asthma Healthy vs. Moderate asthma Healthy vs. Moderate asthma Gene Symbol Gene Title Log2 FC Raw p-value FDR 1.2039 0.0132 0.3858 HIST1H2BJ Histone cluster 1, H2bj 1.1687 0.0406 0.4917 HIST1H1D Histone cluster 1, H1d

Table 3.4 DEGs from Bronchial Brushings Healthy vs. Moderate Asthmatics Healthy vs. Moderate Asthmatic Gene Array on Bronchial Brushings from the U-BIOPRED consortium. Highlighting histone and histone modifying genes which show raw p<0.05. Log2 FC – fold-change, FDR – false discovery rate corrected p-value.

98 Healthy vs Severe asthma Healthy vs Severe asthma Healthy vs Severe asthma Gene Symbol Gene Title Log2 FC Raw p-value FDR 1.1399 0.0264 0.2211 H3F3A H3 histone, family 3A -1.1604 0.035 0.2561 BAZ2B bromodomain adjacent to zinc finger domain, 2B 1.2649 0.0396 0.2734 HIST1H1A histone cluster 1, H1a -1.1818 0.04476 0.26684 EZH1 enhancer of zeste homolog 1 (Drosophila) -1.1584 0.046 0.2932 SETD7 SET domain containing (lysine methyltransferase) 7 1.2278 0.0479 0.3003 HIST1H2BK histone cluster 1, H2bk Table 3.5. DEGs from Bronchial Biopsies Healthy vs. Severe Asthmatics Healthy vs Severe Gene Array on Bronchial Biopsies from the U-BIOPRED consortium. Highlighting histone and histone modifying genes. Log2 FC – fold-change, FDR – false discovery rate corrected p-value. n=40 non-smoking severe asthmatics (SAn), n=28 non-severe, non-smoking asthmatics (NSA) and n=26 normal healthy volunteers (NHV)

Moderate vs severe asthma Moderate vs Severe asthma Moderate vs Severe asthma Gene Symbol Gene Title Log2 FC Raw p-value FDR

1.34 0.008 0.34 HIST1H1A histone cluster 1, H1a 1.29 0.011 0.37 HIST1H2BK histone cluster 1, H2bk 1.15 0.013 0.38 H3F3A H3 histone, family 3A 1.28 0.018 0.4 HIST1H2BH histone cluster 1, H2bh 1.18 0.036 0.41 H2AFY H2A histone family, member Y Cbp/p300-interacting transactivator, with -1.23 0.038 0.46 CITED2 Glu/Asp-rich carboxy-terminal domain, 2 1.19 0.042 0.48 HIST1H2BC histone cluster 1, H2bc 1.2 0.049 0.51 HIST1H2BD histone cluster 1, H2bd Table 3.6. DEGs from Bronchial Biopsies Moderate vs. Severe Asthmatics Moderate vs Severe Gene Array on Bronchial Biopsies from the U-BIOPRED consortium. Highlighting histone and histone modifying genes which show p<0.05. Log2 FC – fold-change, FDR – false discovery rate corrected p-value. n=40 non-smoking severe asthmatics (SAn), n=28 non-severe, non-smoking asthmatics (NSA) and n=26 normal healthy volunteers (NHV)

100

3.3.2 Stable Isotope Labelling with amino acids in culture (SILAC) Data

3.3.2.1 Protein Quantification of Histones and histone modifying Normal Human Bronchial Epithelial Cells (HBECs) were grown under SILAC conditions as described in 2.4. Cells were pre-treated with fluticasone propionate (FP, 10-8M) for 2hrs before being stimulated with IL-1β (1ng/ml) for 24hrs. Intracellular proteins were extracted and the effect of treatments compared by FC and p-value. The experimental work was carried out by Dr. Andrew Durham (Imperial College London) and proteomics by Dr. Heesom (Bristol University), I performed the data analysis.

IL-1β stimulation increased the isotope ratio of many proteins which reflects a change in expression levels, though SILAC is unable to measure post- transcriptional modifications such as acetylation and methylations. The top differentially expressed proteins ranked according to p-value included 3

Histone proteins; H4 (log2 FC = 92.70 p=1.41E-04), H1.5 (log2 FC = 96.40 p=4.06E-04) and H2 type 1-H (log2 FC = 81.80 p=2.20E-03) (Table 3.7). When histone and histone-related proteins were examined specifically, two additional proteins were found to be significantly increased by IL-1β: Core

Histone Macro-H2A.1 (log2 FC = 81.80 p=8.40E-03) and Histone 1.2 (log2 FC = 63.93 p=0.0108) (Table 3.8).

Pre-treatment of IL-1β-stimulated cells with FP (IL-1β+FP- versus IL-1β- treated cells) down-regulated the expression of 13 proteins (raw p-value <0.05)(Table 3.9). FP did not significantly affect the expression of any histone or epigenome modifying proteins (Table 3.10). Though there were limited changes in histone methylase or acetylase protein expression this will not necessarily reflect the activity of these enzymes which

101 were subsequently analysed by determining the total cell levels of the associated histone modifications.

102

Accession Log2 FC p-value Protein Number Q96GX5 98.13 7.80E-05 Serine/threonine-protein kinase greatwall P62805 92.70 1.41E-04 Histone H4 B4DYB4 94.00 2.58E-04 Nucleoporin NUP53 (Nucleoporin Nup35) P16401 96.40 4.06E-04 Histone H1.5 (Histone H1a) (Histone H1b) (Histone H1s-3) P29034 64.23 1.16E-03 Protein S100-A2 (CAN19) (Protein S-100L) E9PF10 78.03 1.49E-03 Nuclear pore complex protein Nup155 K7ES02 80.93 1.69E-03 Bleomycin hydrolase (Fragment) Q96KK5 81.80 2.20E-03 Histone H2A type 1-H (Histone H2A/s) B4DIT7 44.63 2.56E-03 Protein-glutamine gamma-glutamyltransferase 2 P62314 73.30 3.24E-03 Small nuclear ribonucleoprotein Sm D1 Q01469 74.37 4.20E-03 Fatty acid-binding protein, epidermal Q70UQ0 75.77 4.44E-03 Inhibitor of nuclear factor kappa-B kinase-interacting protein P37198 62.80 4.68E-03 Nuclear pore glycoprotein p62 P09382 66.57 5.14E-03 Galectin-1 P10253 60.87 5.18E-03 Lysosomal alpha-glucosidase P11586 69.10 5.18E-03 C-1-tetrahydrofolate synthase, cytoplasmic P30049 59.87 5.69E-03 ATP synthase subunit delta, mitochondrial Table 3.7. Protein change after IL-1β stimulation in nHBECs Top IL-1β (1ng/ml) induced proteins ranked according to raw p-value. Log2 FC compared to unstimulated. Specific histone proteins are shown in bold.

103

Accession Log2 FC P-Value Protein Number P62805 92.70 1.41E-04 Histone H4 P16401 96.40 4.06E-04 Histone H1.5 Q96KK5 81.80 2.20E-03 Histone H2A type 1-H O75367 51.07 8.40E-03 Core histone macro-H2A.1 P16403 63.93 0.0108 Histone H1.2 K7EMV3 62.00 0.184 Histone H3 P57053 51.73 0.184 Histone H2B type F-S P0C0S5 -237.50 0.413 Histone H2A.Z C9JBC2 11.20 0.423 Histone deacetylase 7 K7EQB9 -4.43 0.423 Histone-arginine methyltransferase CARM1 P06899 28.67 0.423 Histone H2B type 1-J Q7KZ85 -24.97 0.423 Transcription elongation factor SPT6 Q86TU7 -12.27 0.423 Histone-lysine N-methyltransferase setd3 Q9H7B4 -3300.00 0.423 Histone-lysine N-methyltransferase SMYD3 Q09028 22.73 0.457 Histone-binding protein RBBP4 P07305 10.33 0.485 Histone H1.0 B3KRS5 25.43 0.606 Histone deacetylase Q92522 5.57 0.778 Histone H1x Q8WTS6 3.43 0.835 Histone-lysine N-methyltransferase SETD7 Q13547 -9.57 0.890 Histone deacetylase 1 (HD1)

Table 3.8. Histone related protein change after IL-1β stimulation in nHBECs Effect of IL-1β (1ng/ml) on the expression of histone and histone-related proteins. Log2 FC compared to IL-1β stimulation alone. Proteins are ranked according to significance (raw p-value).

104

Accession Log2 FC p-value Protein Number P05161 -41.40 0.0042 Ubiquitin-like protein ISG15 B4DXZ1 -49.47 0.0088 Asparagine synthetase B4DLR8 -81.70 0.0118 NAD(P)H dehydrogenase Q9Y617 -20.77 0.0176 Phosphoserine aminotransferas Q9ULC3 -28.40 0.0206 Ras-related protein Rab-23 O00116 -48.10 0.0216 Alkyldihydroxyacetonephosphate synthase Q03169 -39.50 0.0220 Tumor necrosis factor alpha-induced protein 2 P47989 -56.70 0.0279 Xanthine dehydrogenase/oxidase A8KA84 -32.53 0.0350 NA Q53GA4 -37.43 0.0382 Pleckstrin homology-like domain family A member 2 Q96AZ6 -36.93 0.0390 Interferon-stimulated gene 20 kDa protein P29373 -40.70 0.0436 Cellular retinoic acid-binding protein 2 O14879 -46.30 0.0442 Interferon-induced protein with tetratricopeptide repeats 3 Table 3.9. Protein change after FP treatment and IL-1β stimulation in nHBECs -8 Effect of fluticasone propionate (10 M) pre-treatment for 2hrs on IL-1β stimulated protein expression. Log2 FC compared to IL-1β stimulation alone. Proteins are ranked according to significance (raw p-value).

105 Accession Log FC P-Value Protein Number 2 C9JBC2 -33.00 0.423 Histone deacetylase 7 (Fragment) Q9H7B4 -33.00 0.423 Histone-lysine N-methyltransferase SMYD3 Q8WTS6 -20.27 0.248 Histone-lysine N-methyltransferase SETD7 Q09028 -14.77 0.203 Histone-binding protein RBBP4 Q7KZ85 -14.50 0.423 Transcription elongation factor SPT6 P0C0S5 -14.43 0.595 Histone H2A.Z (H2A/z) Q86TU7 -11.77 0.423 Histone-lysine N-methyltransferase setd3 K7EQB9 -11.20 0.423 Histone-arginine methyltransferase CARM1 (Fragment) P07305 -9.43 0.534 Histone H1.0 B3KRS5 -5.33 0.727 Histone deacetylase (EC 3.5.1.98) Q92522 -3.60 0.802 Histone H1x P62805 -2.17 0.639 Histone H4 Histone H2B type 1-J (Histone H2B.1) (Histone H2B.r) P06899 -0.20 0.423 (H2B/r) K7EMV3 -0.17 0.943 Histone H3 Table 3.10. Histone related protein change after FP treatment and IL-1β stimulation in nHBECs Effect of fluticasone propionate (10-8M) pre-treatment for 2hrs on IL-1β stimulated histone and histone-related protein expression. Log2 FC compared to IL-1β stimulation alone. Proteins are ranked according to decrease in Log FC.

106 3.3.3 Histone modifications in normal human bronchial epithelial cells (nHBECs) To investigate the effects of stimulation and CS on global histone modifications in nHBECs, cells were stimulated with IL-1β (1ng/ml) in the presence or absence of 10-8M FP and all results were normalised to total cell histone 3 content to minimise any variability between histone quantities in each sample on a multiplex ELISA (Figure 3.1A). The effect of IL-1β and FP on cell function was demonstrated by measuring the ability of FP to supress IL-1β-induced CXCL8 expression (Figure 3.1B). There were no significant changes in total cellular histone 3 pan-acetylation (Figure 3.1C, ANOVA p=0.80), H3K9ac (Figure 3.1D, ANOVA p=0.53), H3K27ac (Figure 3.1E, ANOVA p=0.82), H3K56ac (Figure 3.1F, ANOVA p=0.57), H3K4me3 (Figure 3.1G, ANOVA p=0.73), H3K9me2 (Figure 3.1H, ANOVA p=0.50), H3K9me3 (Figure 3.1I, ANOVA p=0.76), H3K27me2 (Figure 3.1J, ANOVA p=0.59), H3K27me3 (Figure 3.1K, ANOVA p=0.87), H3T11ph (Figure 3.1H, ANOVA p=0.86) at 24hrs. The limit of quantitation for this assay is ~25 Relative Units (RU of luminance) and so only the trend for IL-1β to enhance pan-acetylation, H3K4me3, H3K9me3 and H3K27me2 have any validity. Though the investigation into global histone modifications has shown no gross changes it is possible that there will be modifications which both increase and decrease across the genome, causing a zero net sum. When studying global changes this would not show, but could account for changes in gene expression.

107

A Total Histone 3 Protein B CXCL8 (pg/ml) CHistone 3 Pan Acetylation 30 3000 80 * NS 60 20 2000 40 RU RU 10 1000 20 CXCL8 (pg/ml)

0 0 0 IL-1β - + + IL-1β - + + IL-1β - + + FP - - + FP - - + FP - - +

D H3K9ac E H3K27ac F H3K56ac 15 2.5 4

2.0 3 10 1.5 2 RU RU 1.0 RU 5 1 0.5

0 0.0 0 IL-1β - + + IL-1β - + + IL-1β - + + FP - - + FP - - + FP - - +

G H3K4me3 H H3K9me2 I H3K9me3 500 0.8 200

400 0.6 150 300 0.4 100 RU RU 200 RU 0.2 50 100

0 0.0 0 IL-1β - + + IL-1β - + IL-1β - + + FP - - + FP - - FP - - +

J H3K27me2 K H3K27me3 H H3T11ph 150 1.5 4

3 100 1.0

2 RU RU RU 50 0.5 1

0 0.0 0 IL-1β - + + IL-1β - + + IL-1β - + + FP - - + FP - - + FP - - +

Figure 3.1. Histone Modifications after FP treatment and IL-1β stimulation in nHBECs Measure of global histone modifications in normal human bronchial epithelial cells (nHBECs). RU – Relative Units. nHBECs were treated with 10-8M fluticasone propionate (FP) for two hours and then stimulated with IL-1β (1ng/ml) for 24 hours. Histones were purified, measured on a histone multiplex ELISA, and the values were normalised to total H3 and had background readings removed. The supernatant was assayed for levels of CXCL8 by ELISA. Results are expressed as mean±SEM of n=4 independent experiments, *p<0.05 compared with unstimulated cells with one way ANOVA and Dunn’s multiple comparison.

108 3.3.4 Effects of epigenetic probes on normal Human Bronchial Epithelial Cell responses To investigate the functional role of histone modifying enzymes a variety of compounds which specifically limit the ability of histone modifying enzymes to elicit downstream functional responses were used. The model was an IL-1β- stimulated model of airway inflammation using Beas2B cells. This will show the targets importance in inflammation and proliferation, key aspects of the asthmatic phenotype.

3.3.4.1 Bromodomain Targeting Probes As detailed in the Introduction (Section 1.8.1.1), bromodomain-containing proteins are able to bind to and read specific acetylated lysines resulting in activation of cellular function. These effects of histone acetylation can be attenuated by bromodomain mimics such as bromosporine, JQ1, GSK2801 & SGC-CBP30 and may demonstrate anti-inflammatory effects.

3.3.4.1.1 Bromosporine Bromosporine shows a concentration-dependant inhibition of IL-1β-induced IL-6 release which reaches significance at 568nM (20.7±4.6% stimulated release, P<0.05) with an IC50 = 61.2±1.2nM and a maximal suppression (Imax) = 93.4±2.1% (Figure 3.2A). There is also a large significant reduction in IL-1β-induced CXCL8 release at concentrations >142nM (43.6±13.1% stimulated release, P<0.05)(Figure 3.2B) with an IC50 = 74.9±1.3nm and an Imax = 98.7±0.5%. There was no significant effect on Mitochondrial activity as a proxy for viability (Figure 3.2C), however, there was a concentration- dependent decrease in cell proliferation as measured by the BrdU assay

(Figure 3.2D). This reached significance at 4096nM with an IC50 of 335±1.7nM and an Imax = 89.1±9.6% (Table 3.11).

3.3.4.1.2 JQ1

JQ1(+) significantly decreased IL-1β-induced IL-6 release with an IC50 of 152.9±1.4nM and Imax =86.9±8.1% (Figure 3.3A). There was a strong

109 negative trend towards suppression of IL-1β-induced CXCL8 release although this did not reach significance despite almost complete inhibition of CXCL8 release (Figure 3.3B). JQ1(+) had no significant effect on mitochondrial activity as a proxy for viability (Figure 3.3C) or on cell proliferation (Figure 3.3D). The negative enantiomer JQ1(-) had no significant effect on IL-1β-induced IL- 6 or CXCL8 release. However, JQ1(-) increased cell mitochondrial activity as measured by MTT assay (EC50 of 327±1.3nM and Emax =122±13.5% (Table 3.11).

3.3.4.1.3 GSK2801 & SGC-CBP30 GSK2801 no significant effect on either cell proliferation or inflammation whilst SGC-CBP30 had no significant effect on IL-1β-induced IL-6 or CXCL8 release (Figure 3.4 A & B). SGC-CBP30 also had no effect on mitochondrial activity as a proxy for viability (Figure 3.4C) although it significantly decreased cell proliferation (IC50 of 21.0±1.1nM, Imax = 100±0%) (Figure 3.4D) (Table 3.11).

3.3.4.2 Histone Methyl-transferase Inhibitors UNC0638, UNC0642, UNC1999, PFI-2 and GSK343 had no significant effects on the release of IL-1β-induced inflammatory mediators, cell proliferation or on mitochondrial activity as a proxy for viability (Table 3.11).

3.3.4.3 Lysine Demethylases Inhibitors

3.3.4.3.1 GSK-J1 GSK-J1 significantly increased IL-1β-induced IL-6 release at its highest concentrations (EC50 = 108.8±1.3nM, Emax =638±228%, P<0.05). These concentrations did not affect cell viability but significantly reduced cell proliferation (IC50 = 56.48±1.1nM, Imax = 100±0%)(Table 3.11).

3.3.4.3.2 GSK-J4 + GSK-J5 GSK-J4 had no effect on IL-1β-induced IL-6 or CXCL8 release. However it potently suppressed cell proliferation (IC50 of 5.86±1.1nM, Imax = 100±0%,

110 P<0.05) and cell viability (IC50 = 22.35±1.1nM, Imax = 77.2±12.1%, P<0.05). The only effect of GSK-J5 was to enhance IL-1β-induced IL-6 release by up to

31.4±58.9% with an EC50 of 1788±3.9nM (Table 3.11).

A B 150 150

100 100 * 50 50 * * * * * *

IL-6 Release (%) Release IL-6 *

CXCL8 Release (%) * * 0 0 0 3 4 5 6 7 8 9 0 2 3 4 5 6 7 8 9 10 11 12 10 11 12

x x Unstim Bromosporine (2 nM) Unstim Bromosporine (2 nM) C D 150 150

100 100

50 50 * Cell proliferation (%)

0 0 MTT Mitochondrial Activity (%) MTT Mitochondrial 0 6 7 8 9 0 2 3 4 5 6 7 8 9 10 11 12 10 11 12

Bromosporine (2x nM) Bromosporine (2x nM) Unstim Unstim Figure 3.2. Bromosporine concentration-response on IL-1β stimulated Beas2B Cells Concentration-dependent effect of bromosporine on IL-1β stimulated Beas2B cells. Cells were pre-treated with bromosporine (0–4096nM) for 2 hours before being stimulated with IL-1β (1ng/ml) for 24 hours. Graphs show IL-1β- induced IL-6 (A) and CXCL8 (B) release, MTT viability assay (C) and a BrdU cell proliferation assay (D). IL-6 and CXCL8 release are normalised to stimulated, MTT and Cell proliferation are normalised to unstimulated. Data are presented as mean ± SEM of n=4 independent experiments, *p<0.05 compared to IL-1β stimulated cells with one way ANOVA and Dunn’s multiple comparison.

111

A B 150 150 ** * * ** * **

100 100 IL-6 (%) IL-6

50 CXCL8 (%) 50

0 0 0 2 3 4 5 6 7 8 9 0 2 3 4 5 6 7 8 9 10 10

x x Unstim Unstim JQ1(+) (2 nM) JQ1(+) (2 nM)

150 C D 150

100 100

50 50 Cell proliferation (%)

MTT Mitchondrial Activity (%) MTT Mitchondrial 0 0 2 3 4 5 6 7 8 9 0 2 3 4 5 6 7 8 9 10 10

x x Unstim Unstim JQ1(+) (2 nM) JQ1(+) (2 nM)

Figure 3.3. JQ1(+) concentration-response on IL-1β stimulated Beas2B Cells Concentration-dependent effect of JQ1(+) on IL-1β stimulated Beas2B cells. Cells were pre-treated with JQ1(+) (0-1024nM) for 2 hours before being stimulated with IL-1β (1ng/ml) for 24 hours. Graphs show IL-1β-induced IL-6 (A) and CXCL8 (B) release, MTT viability assay (C) and a BrdU cell proliferation assay (D). IL-6 and CXCL8 release are normalised to stimulated, MTT and Cell proliferation are normalised to unstimulated. Data are presented as mean ± SEM of n=4 independent experiments, *p<0.05, **p<0.01 compared to IL-1β stimulated cells with one way ANOVA and Dunn’s multiple comparison.

112

A 150 B 250

200 100 150

100 IL-6 (%) IL-6

50 CXCL8 (%) 50

0 0 0 2 3 4 5 6 7 8 9 0 2 3 4 5 6 7 8 9 10 10 Unstim Unstim SGC-CBP30 (2x nM) SGC-CBP30 (2x nM)

C D 200 150

150 * * * * * * * * * 100

100

50 50 Cell proliferation (%)

0 0 MTT Mitochondrial Activity (%) MTT Mitochondrial 0 2 3 4 5 6 7 8 9 0 2 3 4 5 6 7 8 9 10 10

SGC-CBP30 (2x nM) Unstim SGC-CBP30 (2x nM)

Figure 3.4. SGC-CBP30 concentration-response on IL-1β stimulated Beas2B Cells Concentration-dependent effect of SGC-CBP30 on IL-1β stimulated Beas2B cells. Cells were pre-treated with SGC-CBP30 (0-1024nM) for 2 hours before being stimulated with IL-1β (1ng/ml) for 24 hours. Graphs show IL-1β- induced IL-6 (A) and CXCL8 (B) release, MTT viability assay (C) and a BrdU cell proliferation assay (D). IL-6 and CXCL8 release are normalised to stimulated, MTT and Cell proliferation are normalised to unstimulated. Data are presented as mean ± SEM of n=3 independent experiments, *p<0.05 compared to IL-1β stimulated cells with one way ANOVA and Dunn’s multiple comparison.

113 Probe Target IL-1β induced IL-6 Release IL-1β induced CXCL8 Release Effect on proliferation Effect on MTT

IC50 = 61.2±1.2nM IC50 = 74.9±1.3nM IC50 = 335±1.7nM Bromosporine BRD4 + CBP NA Imax = 93.4±2.1% Imax = 98.7±0.5% Imax = 89.1±9.6%

IC50 = 152.9±1.4nM IC50 = 50.7±1.6nM JQ1(+) BrdT + Brd4 NA NA Imax = 92.6±4.1% Imax = 86.9±8.1%

EC50 = 327±1.3nM JQ1(-) JQ1(+)Control NA NA NA Emax = 122±13.5%

IC50 = 21.0±1.1nM SGC-CBP30 CBP/p300 NA NA NA Imax = 100±0% UNC0638 G9a probe NA NA NA NA UNC0642 G9a/GLP Probe NA NA NA NA UNC1999 EZH2/1 Probe NA NA NA NA GSK343 EZH2 Probe NA NA NA NA GSK2801 BAZ2B/A NA NA NA NA

EC50 = 108.8±1.3nM IC50 = 26.87±10^25nM IC50 = 56.48±1.1nM GSK-J1 JMJD3 Probe NA Emax = 634±228% Imax = 26±50% Imax = 100±0

IC50 = 5.86±1.1nM IC50 = 22.35±1.1nM GSK-J4 J1 Pro-drug NA NA Imax = 100±0% Imax = 77.2±12.1%

EC50 = 1788±3.9nM GSK-J5 GSK-4 Control NA NA NA Emax = 31.4±58.9% PFI-2 SETD7 Probe NA NA NA NA Table 3.11. Epigenetic probe concentration-response summary Summary table of epigenetic probe effects on IL-1β treated Beas2B cell function ± = SEM.

114 115

3.4 Discussion

This chapter started by leveraging existing data sets from UBIOPRED epithelial brushings and airways biopsies to measure the differences in gene expression. The data showed there were significant changes in Histone Clusters between severe asthmatics and healthy patients. Proteomics of HBECS treated with FP significantly changed the expression of 13 proteins, however HDAC7 and SETD7 proteins with histone modifying properties had large decreases in mean log2 FC. To try to measure post-translational changes to histones after FP treatment global histone modification measurements were carried out in nHBECs, but no global changes were found. Finally, to see if epigenome modifying proteins were able to control the inflammatory profile of epithelial cells, Beas2B cells were treated with a selection of probes. JQ1+, a Brd4 inhibitor was shown to reduce IL-6 and CXCL8 release and SGC-CBP30 the selective mimetic for CBP and p300 transcription factors was shown to reduce epithelial proliferation without effecting viability.

3.4.1 U-BIOPRED Brushing and Bronchial biopsy Data The U-BIOPRED data has been collected from a large cohort of stable patients with differing severity of disease (Shaw et al. 2015) with no exogenous activation. This gives it great power to investigate changes in gene expression profiles particularly of those genes relating to histone modification and histones themselves.

Bronchial brushings will collect mainly epithelial cells rather than the mixed cellular profile seen in bronchial biopsies which includes epithelial, smooth muscle, mucus and inflammatory cells along with cartilage and fibrocytes (Labont et al. 2008). Bronchial brushings may be expected, therefore, to demonstrate greater numbers of significantly differentially expressed genes (DEGs) reflecting disease severity as the results are not contaminated or diluted by other cell types. Indeed, in this study we do see a greater number of

115 116 significant DEGS related to histones and histone modifications than we observed in the bronchial biopsies.

There were changes in histone expression and the ability to collect cells and carry out experiments on isolated cells made the bronchial epithelium a good candidate for this study. The most commonly changed set of genes in bronchial brushings appears to be the Histone Cluster 1 including genes H1E, H2BK and H2BJ from a large cluster of histone encoding genes on chromosome 6 (Albig et al. 1997). However the fold changes were low so this data must be treated caustiously. Histone H1 has a role in maintaining higher order chromatin structure whilst H2a, H2b, H3 and H4 to bind to form nucleosomes (Marion & Roux 1980; Davey et al. 2002; Luger et al. 1997). These variants of the Histone H1 have differing affinities to DNA (De et al. 2002), and have different purposes depending on the stage of development (Pan & Fan 2016). H1 isoforms can differentially inhibit DNA replication through an affinity for chromatin mediated by their modified carboxyl-terminal domains (Halmer & Gruss 1996; Sun et al. 2015). When comparing bronchial brushings of healthy subjects to severe asthmatics we see that H1E is downregulated whilst H2bk and H2bj are upregulated. H2bj is also increased in bronchial brushings of severe asthmatics compared to non-severe asthma implicating it has a role more exclusively in severe asthma. This may suggest that the chromatin structure in severe asthmatics is more open than in healthy subjects allowing for increased expression of inflammatory genes. H2bk is also increased in bronchial biopsies between healthy subjects and severe asthmatics but the change is smaller than that observed in the bronchial brushings. Epithelial cells are the most common cell type collected in biopsies and this maybe the reason that there is a similar but reduced effect.

116 117

3.4.2 SILAC Analysis SILAC analysis data was used to explore the protein expression of histone modifying proteins in epithelial cells. The SILAC data was obtained using epithelial cells from healthy subjects only and does not have the fidelity to differentiate between post-translationally modified proteins, such as phosphorylation, methylation and acetylation of histones. A limitation of this data when investigating epigenetic regulation is that we will only see increases/decreases in proteins which control the epigenetic process rather than the post-translational modifications of these proteins that regulate their activity. It was important to take into account the amount of histone protein present as it may help to understand the subtypes present after certain treatments. In addition, we only analysed the data at a single time-point (24 hours) and a time-course study may have revealed other changes.

Firstly, we observed a significant increase in Histone H1.5 protein after stimulation with IL-1β. This protein is encoded by the HIST1H1B gene whose expression was not altered in bronchial brushings between asthmatics, irrespective of severity, and healthy volunteers. This follows expectation as IL-1β is not known to be a key mediator in all asthmatics. However, recent evidence indicates that asthmatics with a neutrophilic subtype of asthma have enhanced expression of IL-1β associated with increased inflammasome activity (Simpson et al. 2014). Subsequent studies will be needed to determine any link with neutrophilic asthma. In contrast, we observed a decrease in H1e mRNA expression in severe asthma bronchial brushings. The failure to see an effect of IL-1β on the expression of this protein may indicate the need an assay with higher sensitivity or for a different cell challenge such as IL-13 or IL-33, the need to perform additional time-course studies or that the altered transcription of these proteins is not reflected in differences in protein expression.

When IL-1β stimulated cells were treated with FP no significant changes of histone or epigenetic regulating proteins were discerned. There was a very

117 118 high negative fold change for the genes encoding HDAC7, SMYD3, and SETD7, a histone deacetylase, and two H3K4 histone methyltransferase however without significance at these n numbers. Two genes involved in viral responses are also downregulated in CS treated cells; Ubiquitin-like protein ISG15 (Ea & Baltimore 2009; Yang et al. 2009) and Interferon-stimulated gene 20 kDa protein (Zhou et al. 2012).

3.4.3 Global Histone Modifications To give higher fidelity to the pre-existing SILAC data with regards histone modifications after FP treatment an investigation into global histone modifications was carried out. There was a small non-significant decrease in total H3 protein after FP treatment which reflects the non-significant decrease in histone proteins seen in the SILAC studies. There is also a contradiction between the histone modification data and the SILAC data where the SILAC data shows a log2 FC increase of 62 of Histone 3, the ELISA shows no change. Whether this is because of the low levels of collected histones in the ELISA or due to an error in reading of the protein in SILAC it is unclear. There is also no real positive control that the FP is having an effect in that the CXCL8 ELISA doesn’t show FP creating a significant decrease, just a trend downwards.

3.4.4 Epigenetic Probes A selection of drugs which target various epigenetic mechanisms were tested to see the effect they had on proliferation and release of inflammatory markers. The bromodomain inhibitors bromosporine and JQ1(+) both inhibited IL-1β-induced IL-6 and CXCL8 release from Beas2B cells. Bromosporine is a very broad spectrum bromodomain inhibitor and acts on BRD2, BRD3, BRD4, BRDT, TAF1 and CECR2 (Picaud et al. 2016) whereas JQ1(+) is more specific targeting Brd2, Brd3 and Brd4 with a 10 fold specificity over other BET proteins (Filippakopoulos et al. 2010). Despite the difference in specificity they seem to have similar efficacy in terms of inhibition indicating that BRD2, -

118 119

3 and -4 are more critical for IL-1β-induced IL-6 and CXCL8 expression. Although demonstrating potent anti-inflammatory effects, these compounds may be too non-specific to use as a treatment in patients where other options are available.

SGC-CBP30 is selective for the CBP/p300 transcription factors which also contain bromodomains but does not inhibit IL-1β-induced IL-6 or CXCL8 release from epithelial cells. However, it was able to halt epithelial cell proliferation. Previous studies indicate that SGC-CBP30 modulates regulatory T-cells (Ghosh et al. 2016) and Th17 cells (Hammitzsch et al. 2015) by specifically inhibiting CBP/p300 which has a downstream effect on the FOXP3 gene. FOXP3+ T-cells promote epithelial proliferation (Mock et al. 2014) which is the opposite to the direct effect of SGC-CBP30 in epithelial cells reported here. This may however, point to common endogenous agonist for CBP/p300 that has a role in epithelial cell proliferation which maybe important for asthma as epithelial cell proliferation contributes to airway remodelling (Cohen et al. 2007). In vivo studies in mice may help to delineate whether the effects of SGC-CBP30 on epithelial cells or Treg cells are of greater importance in severe asthma.

3.5 Conclusion and Next steps

There was no change in expression of epigenome regulating proteins in the U-BIOPRED data in severely asthmatic patients, however there was changes of Histone Cluster 1 genes. This could suggest changes in affinity to histones to DNA and account for changes in expression. The expression of SETD7, SMYD3 and HDAC7 decreased following FP treatment, but not significantly, suggesting a power issue where an increase in n number could show these to be important. However there was no significant changes in the histones or the modifications that these proteins modify at the global level. Longer time-course studies may reveal whether there is a functional consequence of these protein expression changes.

119 120

Some epigenetic probes did demonstrate changes to inflammatory protein release such as bromosporine and JQ1. Proliferation of the bronchial epithelium was attenuated by SGC-CBP30 suggesting that there is a role for these proteins in inflammation and so potentially a role in asthmatic inflammation as well. The intention was to follow up these studies with a more accurate genome- wide assessment of the differences in histone modifications, specifically H3K4 me3 because of the change in expression of SETD7 and SMYD3, H3K27me3 and H3K27ac because of their broader roles, in bronchial epithelial cells from severe and non-severe asthmatics to elucidate the detail rather than global changes. This would have been carried out by Chromatin Immuno- precipitation-Sequencing (ChIP-Seq) analysis combined with gene arrays to study the correlation between these modifications and gene expression. It would be important to follow these studies up in primary asthmatic cells, especially for the probes that showed an effect. However, due to difficulties with recruitment of patients and issues concerning primary cell culture and the need from large numbers of cells it was considered not viable to perform this part of the study. As there are a number of papers implicating CD8+ T-cells as playing a role in severe asthma and the relative ease of access of blood cells, I decided to analyse differences in gene expression profiles in these cells and the effect of epigenetic probes/mimics on CD8+ cell function and whether there were differences depending on asthma severity. The remainder of the thesis describes these analyses.

Histone molecules were shown to change expression in the U-BIOPRED datasets, however there was no significant change in the expression of other epigenome modifying proteins genes leaving me to reject this element of the hypotesis that bronchial brushings and biopsies would change expression of these proteins between Asthma disease states.

There appeared to a no large significant changes in protein expression of epigenome modifiying proteins following glucocorticosteroid treatment of

120 121 nHBE cells in the SILAC study, though there was large fold changes of some genes of interest they were not significant, perhaps indicating a power issue. Currenty I have to reject the element of the hypothesis suggesting GCSs affect the expression of epigenetic tags/enzymes in proteomic examinations of bronchial epithelium.

The effect of epigenetic probes bromosporine and JQ1 were able to affect release of inflammatory proteins from Bronchial Epithelial cell lines and may have the power to minimise asthmatic inflammation so this element of the hypothesis can be accepted.

121 122

Chapter 4: CD8+ T-Cell Transcriptome in Asthma

122 123

4.1 Chapter Introduction

Although CD4+ T-cells are considered, canonically, to be a driving force in asthma it has been demonstrated that CD8+ T-cells also play a role. CD8+ T- cells primarily target and destroy virus-infected cells and cancerous cells by recognising non-self antigens on cell membranes and when activated release oxidative and cytotoxic compounds (Jenne & Tschopp 1988), activate cell death pathways (Hildeman et al. 2003) within the infected cells, and encouraging the proliferation of other CD8+ T-cells (Whitmire et al. 2005).

CD8+ T-cells show an increased production of IL-4 and IL-13 in the peripheral blood of asthmatics (Zhao et al. 2011; MacHura et al. 2010). These Th2 cytokines promote asthmatic processes and CD8+ T-cells may continue to release these cytokines after migration to the lungs. CD8+ T-cells have also been shown to suppress IgE production in healthy volunteers but not in severe asthmatics indicating these cells fail to regulate allergic asthma (Lee et al. 1997).

There is no difference between the number of circulating CD8+ T-cells or CD4+ T-cells in severe asthmatics, non-severe asthmatics or healthy subjects (Tsitsiou et al. 2012). CD4+ T-cells from severe subjects only expressed 40 significantly upregulated genes compared to cells from healthy asthmatics and no significantly downregulated genes. In comparison, CD8+ T-cells showed 1359 upregulated genes and 207 downregulated genes in a similar comparison (Tsitsiou et al. 2012). These differences in CD8+ T-cell gene expression were not present between healthy volunteers and non-severe asthmatics, indicating that severity is strongly linked to these changes. Analysis by Tsitsou and colleagues demonstrated no clear abnormal pathways and the data may benefit from unbiased pathway analysis such as Weighted Gene Connectivity Network Analysis (WGCNA).

123 124

As described in Methods and Materials (Section 2.7.1.5), WGCNA is used to analyse large gene sets to find sets of genes whose changes in expression are linked and does not discriminate against small increases or decreases. For example, small FC that go either slightly up or slightly down a similar amount in each subject can imply a pathway(s) working together. These results can then be linked to a clinical measurement such as percent predicted of Forced expiratory volume in 1 second (FEV1 % predicted).

4.2 Chapter Hypothesis and Aims

WGCNA, in conjunction with standard bioinformatic pathways analysis, will reveal a greater description of the activated/repressed pathways in CD8+ T- cells from severe and non-severe asthmatics. Analysis of these changes in a new cohort of patients will validate the approach and show novel pathways specific for CD8+ T-cells in severe asthma.

This hypothesis will be tested by examining whether: • Analysis of pre-existing PBMC’s transcriptomics in severe asthmatics , non-severe asthmatics and healthy patients show differentially expressed genes, • There is a difference between CD8+ T-cell gene expression profiles using FDR and FC in a new cohort of severe and non-severe asthmatics, • WGCNA analysis can find networks of genes that contribute to an altered function in CD8+ T-cells, and whether • These networks can be validated in a previous study cohort which showed a change in CD8+ T-cell expression in severe asthmatics.

124 125

4.3 Results

4.3.1 Investigation of gene expression in peripheral blood mononuclear cell in the U-BIOPRED cohort

The U-BIOPRED consortium collected and analysed circulating mononuclear cells from 246 non-smoking severe asthmatics (SAn), 77 non-severe, non- smoking asthmatics (NSA) and 87 non-asthmatic healthy volunteers (NHV) using gene array, DEGs were identified by ANOVA, including covariates for RNA quality, gender, and clinical site. This data has been tabulated to show top FC and FDR (log2 FC >2, FDR p-value ≤ 0.05).

The data shows no significant differences between NHV and NSA in this strict selection criteria however, the data shows five genes which fit this criterion when comparing SAn and NHV (Table 4.1) which include all three genes shown in SAn vs. NSA (Table 4.2). The three strongly differentially expressed genes (DEGs) in both comparisons were defensin alpha 4, olfactomedin 4 and matrix metallopeptidase 8 (MMP8). Defensin alpha 4 showed a 2.34 log2

FC with a FDR of 3.38E-06 when comparing SAn to NHV and a 2.08 log2 FC (FDR = 3.70E-06) when comparing SAn and NSA. Olfactomedin 4 showed a

2.34 log2 FC (FDR = 5.05E-05) when comparing SAn to NHV and a 2.06 log2 FC (FDR = 0.0023) when comparing SAn and NSA. MMP8 showed a 2.13 log2 FC (FDR = 4.98E-05) when comparing SAn to NHV and a 2.12 log2 FC (FDR = 6.46E-05) when comparing SAn and NSA. This data demonstrates that without some level of selection for peripheral blood mononuclear cells only limited changes are detectable.

125 126

Severe Asthma vs Severe Asthma vs Severe Asthma vs Gene Gene Title Healthy log2 FC Healthy Raw p-Value Healthy FDR Symbol 2.1454 8.21E-09 3.70E-06 LTF Lactotransferrin CEACAM8 Carcinoembryonic Antigen 2.2586 6.10E-09 3.16E-06 Related Cell Adhesion Molecule 8 2.3441 7.16E-09 3.38E-06 DEFA4 Defensin Alpha 4 2.3071 4.51E-07 5.05E-05 OLFM4 Olfactomedin 4 2.1334 6.81E-09 4.98E-05 MMP8 Matrix Metallopeptidase 8 Table 4.1. DEG in Peripheral Mononuclear Cells – Severe Asthmatics vs. Healthy Subjects Differentially expressed genes (DEGs) in peripheral mononuclear cells of severe asthmatics vs healthy subjects. Genes which show a FDR P<0.05 and log2 FC of <2. FC – fold-change, FDR – adjusted p-value.

Severe vs Non-Severe Severe vs Non-Severe Severe vs Non-Severe Gene Symbol Gene Title Asthma log2 FC Asthma Raw p-value Asthma FDR 2.0751 9.73E-07 0.0005 DEFA4 Defensin Alpha 4 2.0571 1.76E-05 0.0023 OLFM4 Olfactomedin 4 2.1231 1.48E-08 6.46E-05 MMP8 Matrix Metallopeptidase 8 Table 4.2. DEG in Peripheral Mononuclear Cells – Severe vs. Moderate Asthmatics Differentially expressed genes (DEGs) in peripheral mononuclear cells of severe non-smoking asthmatics vs non-severe asthmatics. Genes which show a FDR P<0.05. log2 FC – fold-change, FDR – adjusted p-value.

126 127

4.3.2 CD8+ T-Cell Isolation and Gene Array

4.3.2.1 Patient Demographics To study the differences in severe and non-severe asthmatics CD8+ T-cell gene expression and histone modifications patients described in Table 4.3 were recruited to give blood. The demographics are slightly skewed towards

females and subjects were similar in age. FEV1 % predicted is lower in severe patients but this was not significant (Difference 7.26%, p=0.63). Furthermore, severe-asthmatics take more oral prednisolone (difference 5mg, p=0.080) and more inhaled steroids (FP equivalent mg difference 1800, p=0.066) than non- severe asthmatics although again these did not reach significance. The lack of significance may reflect the low number of patients studied in each group, an n=10 per group would show significance in Oral Corticosteroids with the current variance by Mann-Whitney Test.

Asthma severity Non-Severe Severe Difference p-value

Sex (M:F) 2:3 1:6 Age (Median, Range) 55 (39-68) 57 (40-69) 2 0.61

FEV1 % predicted 81.11 73.85 7.26 0.63 (Median, Range) (71-89) (32 - 112) FVC % predicted 90.05 95.65 5.60 0.49 (Median, Range) (78-95) (67-120)

FEV1/FVC 69.26 66.77 -2.49 0.43 (Median, Range) (65-83) (38-81)

Oral Prednisolone 0 (0-0) 5 (5-30) 5 0.080 (mg) (Median, Range) Inhaled FP equivalent 200 2000 1800 0.066 (mg) (Median, Range) (0-3200) (1000-3200) Table 4.3. CD8+ T-cell patient demographics. FEV1 (Forced expiratory volume in 1 second), FVC (Forced Vital Capacity). p- value calculated by Mann-Whitney Test

127 128

4.3.2.2 Flow Cytometry of CD8+ Enriched T-Cells. The efficacy of the enrichment of CD8+ T-cells from patient PBMCs using negative selection by magnetic beads (Section 2.6.4) was quantified using flow Cytometry (Section 2.7.1.7). Using the gating strategy demonstrated in Figure 4.1 cells were 91.0±7.7% lymphocytes (mean±SD, Table 4.4) (Figure 4.1A & B). Of these lymphocytes 65±11.5% (Table 4.4) were CD3+/CD8+ T- cells with the gating selected based on visible populations (Figure 4.1C & D). This data shows a total of 59.1±12.5% (Table 4.4, samples excluded following PCA not shown) of all cells collected were CD3+/CD8+ T-cells. These cells were used for analysis of upregulated CD8+ T-cell pathways involved in severe asthma using gene array and ChIP-Seq.

These enrichment scores are relatively poor because this FACS protocol was not correctly compensated to account for the interference between the FITC and the PE flurochrome. In previous studies using a FACS protocol with flurochromes which do not interfere and using the same isolation method as carried out above the purity of CD8+ T-cells was consistently around ~90%. An example of a correctly stained cells in a FACS plot is given in Figure 4.2.

128 129

Figure 4.1. FACS Gating Strategy for CD8+ T-cells. (A) shows gating for lymphocytes and monocytes pre-enrichment with (B) demonstrating the same gating post-enrichment. The cells from the Lymphocyte gate pre- (C) and post-enrichment (D).

129 130

Figure 4.2 FACS Gating Strategy for CD8+ T-cells from Tsitsiou, 2012 (A) shows gating for lymphocytes and monocytes post-enrichment with (B) demonstrating the CD3/CD8 Stains.

Lymphocyte count CD3+/CD8+ Count Total CD3+/CD8+ Severity (%/SD) (%/SD) Cells (%/SD) NS 1 95.9 82 78.6 NS 2 93 56.4 52.5 NS 3 93.2 78.9 73.5 S 1 93 78 72.5 S 2 94.3 61.2 57.7 S 3 76.8 61.3 47.1 S 4 79 56.7 44.8 S 5 99.7 51 50.8 S 6 93.8 59.1 55.4 All 91 / 7.7 65 / 11.47 59.1 / 12.47 Non-Severe 94 / 1.6 72.4 / 13 68.2 / 8 Severe 89.4 / 9.3 61.2 / 9.1 54.7 / 10

Table 4.4. Enrichment of CD8+ T-cells Enrichment of CD8+ T-cells as counted by gating demonstrated in Figure 4.1A and B for lymphocyte count and CD3+/CD8+ as demonstrated in Figure 4.1C and D. Data are shown for individual samples and for pooled data.

130 131

4.3.2.3 Gene Array RNA was extracted and reverse transcription and amplification was carried out as described in Methods Sections 2.7.1.1 to 2.7.1.4. The arrays were normalised using a robust multichip average method within Partek software. Figure 4.3 shows that the distribution of read intensity across all the arrays and between groups have been suitable normalised as the peaks overlay.

Figure 4.3. Gene Array Read Frequency Plot for CD8+ T-cell Read frequency Plot for CD8+ T-cell blood samples run on gene array post- normalisation. The data shows Non-Severe (Red, n = 3), the excluded Non- Severe (Blue, n= 2) and Severe (Green, n = 6) comparing Frequency of reads of each value against each value read.

131 132

4.3.2.4 Principal Component Analysis Principal Component Analysis (PCA) is method to simplify multidimensional data to show its most variable elements in a few metrics and can be shown as a single three-dimensional image. Partek software takes all the variables (genes) and plots them against each other, ranks the genes by the amount they change and assembles the genes into groups of similar variability. The 3 most variable groups are plotted by asthma severity on the PCA graph (Figure 4.4). The top 3 PCA functions account for 53.2% of the gene variability.

On the graph three outliers can be seen, while there was considerably fewer genes counted in two of them, it was decided to exclude all three as probable technical failures. This left array data from 6 SA and 3 NSA for further analysis.

Once the data was selected Partek was used to draw up a list of genes showing high fold-change (FC), significant p-value and significant FDR value.

132 133

Figure 4.4. Gene Array Genes Detected and PCA Genes detected for CD8+ T-cell samples (A) by gene array and principal component analysis (B) plotting the same samples. Red shows non-severe asthmatic, Green shows severe asthmatic (one non-severe sample is obscured by a severe sample). Excluded outliers have red circles.

133 134

4.3.2.5 Differential Gene Expression There were 7 differentially expressed genes by FDR between severe and non- severe asthmatics with a log2 FC of x>1.5 or x<-1.5. These genes are shown in Table 4.5. There are also some non-significant genes (by FDR) with high negative log2 FC which are shown in Table 4.6. Genes in Table 4.7 are those with a high log2 FC. These are paired with the gene ontology outputs from gProfiler to show related pathways derived from the literature. The data shows significantly differentially expressed genes were associated with congenital disorder of glycosylation and immunodeficiency and the genes with high log2 FC included pathways such as response to bacteria.

Gene Symbol log2 FC FDR p-value HIST1H1C -1.528 3.14E-02 MAN2B2 1.790 3.81E-02 SSB 1.573 4.01E-02 GNB2L1 -1.655 4.28E-02 LOC440280 1.702 4.32E-02 TUBA1A 1.653 4.71E-02 TAF15 2.130 4.73E-02 Table 4.5. Significantly DEGs of CD8+ T-cells – NSA vs SA Differentially expressed genes (DEGs) of CD8+ T-cells from non-severe and severe asthmatics as calculated by Partek FDR p-value 0.05. FC>1.5, FC<- 1.5 FDR – false discovery rate, log2 FC – fold-change.

134 135

Gene Name log2 FC FDR p-value NDUFB8 -2.812 0.132 GGA1 -2.652 0.350 SFRS6 -2.632 0.180 PSME4 -2.398 0.150 LOC641814 -2.339 0.181 LOC644250 -2.330 0.287 CD81 -2.307 0.140 SNHG3 -2.241 0.279 LOC100130053 -2.237 0.222 F2R -2.208 0.238 LOC100132119 -2.116 0.120 BLVRB -2.114 0.468 ALG8 -2.107 0.409 CNN2 -2.094 0.276 LOC100129553 -2.077 0.303 LOC651483 -2.057 0.104 TRIM16L -2.048 0.254 MX1 -2.047 0.247 C14ORF85 -2.047 0.263 LOC440027 -2.038 0.226 LOC645899 -2.016 0.221 Table 4.6. Large negative FC DEGs of CD8+ T-cells – NSA vs SA Log2 fold-change (FC) <-2 differentially expressed genes of CD8+ T-cells from non-severe and severe asthmatics as calculated by Partek. FDR – false discovery rate adjusted p-value.

135 136

Column ID log2 FC FDR p-value Column ID log2 FC FDR p-value DEFA1B 4.677 0.305 SNORD46 2.298 0.203 MPO 4.213 0.258 SLC9A4 2.264 0.074 JUN 4.129 0.085 C9ORF41 2.223 0.198 ZFP36 3.639 0.078 MIR1974 2.200 0.339 FAM177A1 3.122 0.155 SAT1 2.182 0.209 DEFA3 3.084 0.465 C9ORF80 2.181 0.228 DEFA1 3.061 0.466 STK36 2.161 0.133 C6ORF48 2.979 0.316 C6ORF111 2.157 0.139 CAMP 2.803 0.336 TAF15 2.131 0.047 JUND 2.778 0.054 DAZAP2 2.122 0.092 BMS1P5 2.726 0.254 RPL18 2.113 0.235 DDIT4 2.694 0.085 ARHGAP21 2.107 0.197 HDC 2.652 0.455 CDC2L2 2.093 0.141 PRR14 2.577 0.075 HNRNPA3 2.084 0.302 EGR1 2.553 0.403 ZNF91 2.057 0.127 GCA 2.543 0.330 FOS 2.051 0.274 DUSP1 2.483 0.145 PARP1 2.031 0.143 CTSG 2.481 0.372 CBX5 2.020 0.087 RSL24D1 2.392 0.107 OAZ1 2.010 0.075 SDF2L1 2.378 0.071 SRRM2 2.006 0.075 LASP1 2.001 0.096

Table 4.7. Large Positive FC DEGs of CD8+ T-cells – NSA vs SA Log2 FC>2 differentially expressed genes (DEGs) of CD8+ T-cells from Non- Severe and Severe asthmatics as calculated by Partek. FDR – false discovery rate adjusted p-value.

136 137

4.3.2.6 PCR Confirmation PCR was performed on several genes highlighted as being of interest by the gene array for confirmation of large positive fold change genes including those which represent key groups including anti-bacterial genes (MPO and DEFA1B), corticosteroid reactive genes (DUSP1) and one gene from the group involving mRNA stablility (ZFP36). There was a good correlation between the gene array data and the RT-qPCR data for almost all genes studied (Figure 4.5). The only gene which trended differently was MPO (Figure 4.5A&B). DEFA1B (Figure 4.5C&D), ZFP36 (Figure 4.5E&F) and DUSP1 (Figure 4.5G&H) were all expressed similarly according to array and PCR data.

4.3.2.7 Cross-Study Confirmation Comparing the data from the Tsitsiou 2012 paper, stored online at GEO (GSE31773), and the gene arrays performed here, there was a strong correlation in expression levels between genes such as HIST1H1C (Figure 4.6A&B), ZFP36 (Figure 4.6C&D), DUSP1 (Figure 4.6E&F) & CD81 (Figure 4.6G&H), but not GGA1(Figure 4.6H&I).

137 138

A B MPO PCR MPO GA 10 12 10 * * 109 10 108 ct Δ AUI 107 8 106

105 6 Non-Severe Severe Non-Severe Severe

C D DEFA1B GA DEFA1B PCR 14 1014

1013 12

12 10 10 ct AUI Δ 1011 8 1010

109 6 Non-Severe Severe Non-Severe Severe

E F ZFP36 PCR ZFP36 GA 108 11

7 ** 10 10 106 9

ct 105 Δ AUI 8 104 7 103

102 6 Non-Severe Severe Non-Severe Severe

G H DUSP1 GA DUSP1 PCR 10 4000 * * 3000 9 300

200

ct 8 AUI Δ 100 7 0

-100 6 Non-Severe Severe Non-Severe Severe

Figure 4.5. PCR/Gene Array Confirmation PCR results compared with gene array data for MPO (A & B), DEFA1B (C & D), ZFP36 (E & F) and DUSP1 (G & H). * = p<0.05, ** = p<0.01 by Mann- Whitney Test. GA – gene array. AUI = Arbitrary Units of Intensity

138 139

A C E G I HIST1H1C GA ZFP36 GA DUSP1 GA CD81 GA GGA1 GA 7.5 11 10 10 15 * ** * 10 9 9 7.0 10 **** 9 8 8 AUI AUI AUI AUI AUI 8 7 6.5 5 7 7 6

6.0 6 6 5 0 Non-Severe Severe Non-Severe Severe Non-Severe Severe Non-Severe Severe Non-Severe Severe

B HIST1 ET GA D ZFP36 ET GA F DUSP1 ET GA H CD81 ET GA J GGA1 ET GA 10 11 11 12 10 **** **** 8 10 11 *** 10 ** 9 6 9 10 9 AUI AUI AUI AUI 4 8 AUI 9 8 8 2 7 8

0 6 7 7 7 Non-Severe Severe Non-Severe Severe Non-Severe Severe Non-Severe Severe Non-Severe Severe

Figure 4.6. Gene Array/Tsitsiou Gene Array Comparison CD8+ T-cell gene array confirmation of non-severe and severe asthmatics from Tsitsiou Dataset (Bottom panel, ET) downloaded from GEO (GSE31773) compared to CD8+ T-cell data collected in this thesis (Top panel). HIST1 (A & B), ZFP36 (C & D), DUSP1 (E & F) CD81 (G & H) and GGA1 (I & J). * = p<0.05, ** = p<0.01 by Mann-Whitney Test. GA – gene array. AUI = Arbitrary Units of Intensity

139 140

4.3.2.8 WGCNA WGCNA is an unbiased method used to find groups (modules) of significant genes which change together and correlate with clinical outcomes. This analysis was performed using methods described by Zhang and colleagues (Zhang et al. 2005).

Firstly, in a test to find outliers; the subjects were hierarchically clustered based on gene expression of genes with raw p-value < 0.05 as shown in Figure 4.7. The clustering does not distinguish the severity of disease and does not show any distinct outliers. There are three branches which break off to show two non-severe asthmatics in one branch. One branch just contains severe asthmatics, the other shows severe asthmatics with one non-severe asthmatic. The heat map in Figure 4.7 shows the intensity of clinical features which are measured including treatment, airway measurements and general clinical measurements which were used in the WGCNA analysis.

To minimize noise from gene array data, the Pearson correlations created in the network analysis were raised to the power β otherwise known as the soft power value. The soft power value chooses an appropriate level of connectedness between the genes to find those moving together. To select the soft power value we use the Scale Free Topology criterion (Zhang et al. 2005). Different values of β are plotted against Scale Free Topology criterion and used to select the soft power value to take forward as in Figure 4.8. Figure 4.8A shows the Scale Free Topology Model, which begins to saturate at a value of 12 so this was used as the soft power value for the Topological Overlap Matrix. Median (Figure 4.8B) and mean (Figure 4.8C) connectivity are non-zero and max connectivity (Figure 4.8D) is also appropriate.

Following the soft threshold setting the data was applied to the Topological Overlap Matrix (TOM), modules were created and assigned a random colour name to aid referencing groups as demonstrated in Figure 4.9.

140 141

Figure 4.7. CD8+ T-cell Sample Gene Dendrogram and Trait Heatmap CD8+ T-cell sample dendrogram by hierarchical clustering of gene expression and trait heat map for severe (S) and non-severe asthmatics (NS). The greater intensity in value is reflected by the darker red colour. Grey colouring represents missing values. BMI = Body Mass Index, FP = Fluticasone Propionate, FEV1 = Forced Expiratory Volume after 1 second, FVC = Forced vital capacity RAST = radioallergosorbent test.

141 142

Figure 4.8. WGCNA Soft Power Selection Criteria Selected β or soft power value using free scale topology model fit (A), median connectivity (B), mean connectivity (C) and max connectivity (D). The plotted numbers show the average number of connections of each gene.

142 143

Figure 4.9. Gene dendrogram and module colours of WGCNA clusters

The eigengenes were clustered using hierarchical clustering. As none were found to be similar no modules were merged (Figure 4.10). Finally, the correlation between module expression and patient traits was measured and shown in Figure 4.11. The green-yellow module shows high positive correlation to severity (r=0.95, p=7x10-06) whilst only showing a small correlation with limited significance to oral prednisolone usage (r=0.33, p=0.4). This indicates that the green-yellow module can discriminate prednisolone usage from asthma severity (Table 4.8). FEV1 % predicted was positively correlated with the light-green module (r=0.72, p=0.03) and strongly negatively correlated with the red (r=-0.86, P=0.002) and blue (-0.97, px10-07) modules. The 47 genes from the green-yellow module (Table 4.8), the 43 genes from the light-green module (Table 4.9), the 78 genes from the red

143 144 module (Table 4.10) and the 143 genes from the blue module (Table 4.11) were run in gProfiler to determine if there are any known relationships between these genes. gProfiler gene ontology analysis from the blue module showed nodes of high connectivity which included IL17A, ZNF24, BID, PDE9A, BOX12, KDM4D and EDN. The green-yellow module’s gene ontology networks key nodes were TUBA1A, DST, DCTN1, VSX1, GNBL2, STK16 and ARMCX2. The red module’s gene ontology network included KCNIP2, FGFR4 and IL10 as key nodes. The light-green module’s gene ontology network showed WHSC1, PTPN13 and CPEB1 to be key nodes. Pathway analysis indicated that the Green-Yellow module was associated with a number of pathways including “Neuropathy, Distal Hereditary Motor Type VIIB”, “Perry Syndrome, Parkinsonism with Alveolar Hypoventilation and mental depression” and Adenine Phosphoribosyltransferase Deficency”. In contrast, the gene lists within the light green, red and blue modules did not show significant associations with any currently defined pathways.

144 145

Figure 4.10. Clustering of module eigengenes Clustering of module eigengenes showing no similarity between groups and therefore no combining of modules.

145 146

Figure 4.11. WGCNA Module-trait Relationship Module-trait relationship (R-value, (p-value)). WGCNA modules are shown are shown by their colour. In the chart red shows high positive correlation and blue shows high negative correlation with white as no correlation. BMI = Body Mass Index, FP = Fluticasone Propionate, FEV1 = Forced Expiratory Volume after 1 second, FVC = Forced vital capacity, RAST = radioallergosorbent test.

146 147

ANKLE1 GNB2L1 NUP205 APCDD1 HIST1H1C OR13D1 APOL2 HOXA11AS PCDHB10 APRT HS.565001 RIMS2 ARFGEF2 HS.567088 RUVBL1 ARMCX2 HSDL2 SEC22C BAG4 KIAA1160 SLC23A2 C1QTNF4 LANCL1 SLC35E2 CBWD7 LDB3 SMYD3 CNTROB LMCD1 SSB CTAG1B LOC100128025 STK16 DCTN1 LOC100130795 SUCLG2 DDX23 LOC730083 SVEP1 DEPDC1 LOC731479 TAF15 DGKK M6PRBP1 TUBA1A DLGAP4 MAN2B2 UBE2V1 DST MED27 VSX1 FAM90A5 MGC70870 ZFP2 FGD6 NFKBIZ ZNF654 Table 4.8. Genes within the WGCNA Green-Yellow Module (57 genes)

AIM1L GCK HS.580515 MSGN1 AP4S1 GCNT3 HS.582211 MT1E BCR GGTLC1 KIAA0895L OR5K3 C1QTNF9B HBG1 LOC100128918 PTPN13 C2ORF42 HNF4A LOC100131693 PYCR1 C3ORF35 HOXB7 LOC100131970 RAB7B CPEB1 HS.538525 LOC100134333 RS1 CRTAP HS.556255 LOC100134414 TMEM184B DNAJC5B HS.562488 LOC729744 TMEM19 FBN2 HS.576771 MANEA TTF2 HS.579070 MIR96 WHSC1

Table 4.9. Genes within the WGCNA Light-Green Module (43 genes)

147 148

ACSM3 FGFR4 IL18 MIR555 ADAMTSL1 FOLR1 KCNIP2 NCCRP1 AOC2 FSCB LCN6 OR1L1 ARL16 GABPB2 LILRB1 OR5F1 C10ORF31 GLRB LOC100008588 OR6C6 C21ORF24 GRK1 LOC100130168 OR6Y1 C21ORF55 HS.122053 LOC100130445 PCDHGA8 CCBE1 HS.130178 LOC100132154 PDE4C CDKN2AIPNL HS.282556 LOC100132727 PDXDC2 COQ7 HS.537603 LOC100133649 POU3F4 CRCP HS.539981 LOC127099 PPA2 CYLN2 HS.544884 LOC728857 PPT2 DEFB106A HS.546089 LOC729090 RPL7L1 DFNA5 HS.569711 LOC730281 SLC34A3 EDN2 HS.570636 LPPR5 USP49 EIF2AK4 HS.578850 LRP8 WDYHV1 FBXO33 HSPC268 LRRC45 ZBTB10 FCGR2B IL10 MGC16703 ZNF430 ZNF84

Table 4.10. Genes within the WGCNA Red Module (78 genes)

148 149

ABCC13 COLQ HS.555115 LOC100132391 MTIF2 REPS2 ABHD14B DBX1 HS.557218 LOC100132593 NDRG4 RREB1 ACSM2B DHRS2 HS.562585 LOC100132839 NEUROG3 RRP7B ADAM11 EDG7 HS.564949 LOC100132891 NRF1 SCAND3 ADC EDN1 HS.569985 LOC100133465 NUDCD3 SLC25A29 ALG10B EGFL8 HS.573047 LOC100133893 OR10W1 SLC8A1 ANKRD28 FAM101B HS.574784 LOC100133990 OR2T34 SMOC1 AP1G1 FAM90A13 IGDCC4 LOC100134423 OR51I2 SNRNP40 APOBEC3F FANCA IL17A LOC202134 OVOL2 SOX12 ATXN7 FKBP9L IL1F7 LOC728969 PAEP ST8SIA5 BID FLJ44124 IMAA LOC729756 PDE9A STARD13 BOLL GLYBP JSRP1 LOC730066 PEX11A STK35 C10ORF73 GNAI1 KDM4D LOC730159 PFTK2 TBXAS1 C14ORF65 GP2 KIAA0895 LOC732215 PLGLB1 TLR10 C1ORF158 GPR173 KRT18P17 LRRC37A4 PLSCR4 TNXA C5ORF42 H2BFWT KRT71 LRRC41 POLD2 TTC22 C9ORF125 HIST1H3E LHX4 MGC57359 POMT1 ZBTB34 C9ORF82 HS.124623 LIFR MIOS POU2AF1 ZDHHC11 CARD8 HS.148844 LOC100128904 MIR1252 PPIL4 ZFP41 CASC2 HS.511717 LOC100129489 MIR186 PSAT1 ZHX2 CGB HS.537922 LOC100129758 MIR548Q PTBP2 ZIK1 CLUAP1 HS.545028 LOC100129946 MIR592 RAB19 ZNF24 CLUL1 HS.552087 LOC100130979 MT1IP RANBP6 ZNF280C CNN3 HS.552427 LOC100131530 MTERFD2 RAP2B

Table 4.11. Genes within the WGCNA Blue Module (143 genes)

149 150

4.3.2.9 WGCNA of Tsitsiou Data An identical analysis was performed on the Tsitsiou data comparing non- severe and severe asthmatics. When clustering module eigenvalues some colours were merged as shown in Figure 4.12. There were fewer patient traits published in the Tsitsou paper giving a more limited module-trait relationship (Figure 4.13). However, the purple module (Figure 4.13) had a significant link to FEV1 % predicted (r=0.59, p=0.04) and an equally strong negative correlation for severity (r=-0.79, p = 0.001) and prednisolone usage (r=-0.77, p=0.002). The purple module contains 56 genes (Table 4.12) and gProfiler gene ontology shows that KDSR, QKI, MATR3, AIDA, P4HB, CANX, STAT1 and MAP3K2 are key nodes. Pathways as suggested by gProfiler included the endomembrane system where 23 of the selected 46 genes gprofiler could identify were involved to some extent, vesicles, intracellular vesicles, cytoplasmic vesicles, pigment granule and melanosomes.

Figure 4.12. WGCNA Cluster Dendrogram and Module Colours of Tsitsiou Data

150 151

Tsitsiou, 2012 Dataset downloaded from GEO (GSE31773) WGCNA gene dendrogram and module colours pre- and post-merger of modules.

Figure 4.13. Tsitsiou Data Module-Trait Relationships Tsitsiou dataset downloaded from GEO (GSE31773). Module-trait relationship (R-value, (p-value)). In the chart red shows high positive correlation and blue shows high negative correlation with white as no correlation. FEV1 = Forced Expiratory Volume after 1 second, FVC = Forced vital capacity, PC20 = provocative concentration of inhaled bronchodilator which causes FEV1 to fall by 20%.

151 152

LACTB ANKRD20A1 PTPN11 VDAC1 ITGB1 ARF4 CLIP1 ZBED2 ITGB1 VAMP3 MAP3K7 AIDA DNAJB9 ZFR MAP3K7 CES3 CRTAP GDE1 SH3GL3 YIPF5 METTL8 FAM20B VAMP3 NAA50 QKI PCYOX1 IPO5 EXOC5 MATR3 CLN5 VDAC1 TRIM4 KDSR RAB1A DLAT NGRN CTD-2076M15.1 APLP2 H2AFY HIBADH RP1-142L7.9 ADH5 ADAM10 YIPF5 P4HB CANX ITGA6 RSPRY1 RAC1 SRI TOR1AIP1 GLDN BOD1L1 FRG1B STAT1 HIBADH Table 4.12. Genes within the Tsitsiou WGCNA Purple Module (56 genes)

152 153

4.4 Discussion

This chapter has utilised data from the U-BIOPRED study looking at the geneomics of peripheral blood mononuclear cells, which contains T-cells and shows a decrease in CEACAM8 in severe asthmatics. CD8+ T-cells collected from severe and non-severe asthmatics shown that severe asthmatics have a significantly increased expression of TUBA1A and WGCNA shows there are groups of genes involved in cytoskeletal activity and vesicle budding. It is also interesting to note that there are high fold change, but not statistically significant changes in expression of anti-microbial genes in CD8+ T-cells which have also been shown in another similar data set, and this group of genes has also been selected by WGCNA to be important in severe asthma. This study has low power and there should be concerns about the purity of the cells as the FACS was incorrectly carried out, however the comparison to previous studies helps to demonstrate that the expression sets are similar and that a repeat of this study in a larger group may well help back up the results.

4.4.1 U-BIOPRED Peripheral Mononuclear cells Gene Expression The DEG analysis from this blood PBMC study shows downregulation of the genes which encode defensin alpha 4 (Lehrer & Ganz 1992), MMP8 (Ugarte- Gil et al. 2013) and lactoferrin (Farnaud & Evans 2003). In addition, the expression of CEACAM8 was also down-regulated. The protein CEACAM8 regulates cell adhesion and activation of human eosinophils (Yoon et al. 2007) and decreases Toll Like Receptor 2 (TLR2) triggered immune responses minimizing pro-inflammatory immune responses in the airway (Singer et al. 2014). The decrease in CEACAM8 mRNA expression in severe asthmatics is one way in which severe asthmatics might have a more limited response to treatment, but the decrease in the antibacterial genes is most likely due to CS treatment. It is of note that CEACAM8 mRNA expression was previously reported as up-regulated in CD4+ T-cells and CD8+ T-cells in severe asthmatics (Tsitsiou et al. 2012). The differences in gene expression seen here and with the Tsitsiou study may reflect the fact that U-BIOPRED

153 154 dataset was not deconvoluted to account for cell type differences that occur in the blood of asthmatic patients with different disease severity. Future studies may analyse the expression of these genes to reflect the individual contribution of various peripheral blood cell types such as monocytes, neutrophils, T-cells and B-cells make. In further studies it would also be of interest to carry out proteomics to investigate if the gene expression carries through into protein production.

4.4.2 CD8+ T-cell Gene Array Experimental Design By comparing severe and non-severe subsets we were able to suggest new aberrant pathways which could be further investigated in larger studies and may eventually be able to understand severe asthma. Previous cluster analysis focussed on clinical variables and a few biochemical or cellular measures have shown asthma to be highly heterogeneous (Moore et al. 2010).

The low number of patient samples used here (n=3 non-severe asthma and n=6 for severe asthma after exclusions for outliers) have given this study lower power. It would take an n=10 to differentiate the groups by dose of Oral Corticosteroids with the current variance and this seems like a good starting place for further studies. It would be interesting to expand the groups to include differentiated cell subtypes from severe and non-severe asthmatics, but this would necessitate even larger patient numbers. Generally the PCR confirmed what was found in the gene array. The differences between the MPO PCR confirmation and the MPO gene array could be dues to differences between the sequences examined by PCR and the GA. A primer specifically matching the sequence on the gene array should have been used rather than a stock MPO primer from Qiagen which has a propriatory design and cannot be compared.

154 155

A number of genes were elucidated using WGCNA which were able to differentiate between the effects of the high dose of steroids that severe asthmatics take and the disease itself. The green-yellow module (Figure 4.11) which without being significantly linked to oral prednisolone (r=0.33, p=0.4) or inhaled FP equivalent (r=0.3, p=0.4) was strongly linked to severity (r=0.95, p=7e-06). This demonstrates that there are some genes which are differentially expressed in severe asthmatics independent of steroid usage. A limitation of the study was that no objective measure such as cortisol levels or plasma prednisolone metabolites were measured and so we cannot discount compliance being an issue. As such, the differentiation between the groups may reflect a true subgroup or an effect of not using drugs. However, these subjects have been treated at the Royal Brompton Hospital; Severe asthma clinic for many years, are not well controlled and have steroid side- effects. Another potential limitation of the study was the fact that we did not achieve 100% cell purity. Flow Cytometry showed that 90% of cells were lymphocytes and 65% of these lymphocytes were CD3+/CD8+ cells. The low purity could be a concern, as the FACS was incorrectly controlled and cannot be repeated, however a prior study using the same method showed ~90% purity in all samples (Tsitsiou et al. 2012). Alongside this, the same previous study had similar expression to this data. Despite this a repeat of this study with proper FACS analysis, protein expression studies and higher n-numbers would help to confirm or deny these next points.

4.4.2.1 Cytoskeleton Genes The cytoskeleton plays a vital role in T-cell activation. After TCR activation f- actin is polymerised and creates an ‘immunological synapse’. Inhibition of actin polymerisation in T-cells inhibits T-cell activation (Valitutti et al. 1995; Grakoui et al. 1999).

Tubulin is part of the microtubule organising centre (MTOC) for cell-cell interactions between antigen presenting cells (APC) and T-cells (Kupfer &

155 156

Dennert 1984; Kupfer et al. 1987) which regulates the secretion of effector molecules and targeted receptor endocytosis (Sancho et al. 2002). Whilst interacting with the MTOC in APCs, the cytoskeleton and Golgi apparatus reorients beneath the contact site (Kuhné et al. 2003; Kuhn & Poenie 2002; Saito & Yokosuka 2006; Kupfer & Dennert 1984; Kupfer et al. 1987), In addition, microtubules direct IL-2, IL-4 and IL-5 stores to the surface of T-cells when they are associated with APCs. (Kupfer et al. 1991; Reichert et al. 2001).

TUBA1A mRNA which codes for tubulin α, one of the two main components of tubulin, is significantly upregulated in severe asthmatics CD8+ T-cells. Conversely, several genes involved with actin filament support including CD81, CNN2, which encodes Calponin 2 and GGA1 which encodes golgi associated, gamma adapter ear containing ARF binding protein 1, were had high negative fold-change suggesting an imbalance between tubulin and actin filaments if this was found to be significant in future studies.

CD81 is important for T-cell activation and myotubule maintenance (Charrin et al. 2013). In T-cells CD81 forms a link with CD4 or CD8 (Imai et al. 1995) to provide a co-stimulatory signal with CD3 (Todd et al. 1996; Tachibana & Hemler 1999). It is also important for T-cell maturation in mice as when mice fetal thymus organ cultures were treated with CD81 antibodies they were prevented from differentiating from CD4-CD8- cells into CD4+ cells (Boismenu et al. 1996). CD81 also inhibited αβ T-cell maturation while sparing γδ cells. CD81 matures double positive CD4+/CD8+ to single positive CD4+ or CD8+ (Cibotti et al. 1997). CD81 deficient mice express normal numbers and subsets of T-cells but have diminished antibody responses to protein antigens. With a large increase in CD81 in T-cells you would expect a greater reaction to stimuli. CD81s role with the cytoskeleton is demonstrated most clearly in B cells were CD19 movement to cell depends upon CD81s presence where they form a complex with Leu13 to increase B cell activation (Levy 2014).

156 157

Calponin 2 is a calcium binding protein which also binds actin, calmodulin, troponin C, and tropomyosin, and functions in the structural organization of actin filaments which in T-cells are vital for formation of lamellipodium structure which increases the interactive APC-T-cell binding surface area. Golgi-Associated, Gamma Adaptin Ear Containing, ARF Binding Protein 1 is important for Golgi traffic and therefore release of active proteins in promoting cell death. In future studies it would be vital to investigate the proteomics to understand whether these proteins expression are upregulated similarly to the genes.

4.4.2.2 RNA Modifying Genes and AP-1 SSB and TAF15 genes were significantly increased more than 1.5-fold in severe asthmatics and represent the RNA Binding Motif Containing (RBM) family of genes (http://www.genenames.org/cgi-bin/genefamilies/set/725) which are linked by high homology. SSB and TAF15 have roles in microRNA functions, RNA protection and in regulation of post-transcriptional modifications of RNA. These modifications of RNA may play a part in the changed expression profile of severe asthma.

SSB produces Sjogren syndrome antigen B otherwise known as ‘Lupus La protein’ which binds to UUUOH 3’ to prevent exonuclease digestion of the RNA which protects RNA from degradation (Stefano 1984). TAF15 encodes TATA-Box Binding Protein Associated Factor 15 which is one of many proteins that RNA polymerase II requires for gene transcription.

The up-regulation and translation of RBM genes could well be linked to the non-significant but high fold-change up-regulation of the FOS, JUN & JUND genes. Both Jun and JunD form a heterodimer with Fos to form AP-1 which has previously been reported to be implicated in relative steroid insensitivity in peripheral blood cells in patients with severe asthma (Lane et al. 1998).

157 158

AP-1 is a transcription factor which has been shown to regulate differentiation, proliferation and apoptosis of cells which is key to the T-cell receptor activation. T-cell receptor activation is context and state dependant in its effects and can be responsible for burst like clonal proliferation, memory cell differentiation, release of cytokines and cytotoxic molecules and to induce apoptosis (Wirth et al. 2010; Ahmed & Gray 1996; Kaech & Cui 2012).

BACH2 is a transcriptional repressor which is recruited to enhancers and limits TCR-driven genes by attenuating the availability of AP-1 (Roychoudhuri + et al. 2016) and promotes differentiation of Foxp3 Treg cells to prevent lethal inflammation. Furthermore, in TH1 cells AP-1 activation in conjunction with IL- 4 stimulation of the JAK-STAT pathway can activate further IL4 production in a feed-forward manner to induce TH2 differentiation following Gata3 expression (Ouyang et al. 2000; Macián et al. 2000). Gata3 can then drive epigenetic changes at the TH2 cytokine cluster.

4.4.2.3 Antimicrobial Properties of Cytotoxic T-cells Looking at the highest FC genes (Table 4.6) there is an insignificant trend that suggests that anti-bacterial genes are broadly up-regulated in severe asthmatics, particularly genes such as DEFA1B/A3/A1, CAMP, MPO and CTSG. Though there is a potential that these cells were contaminated with other cells in the enrichment stage of their extraction, the gene DEFA4 was also shown to be upregulated in this cell type in severe asthma in Tsitsiou et al. 2012. These genes code for the proteins neutrophil defensin 1 (and other defensins), which possess a membrane pore-forming antibacterial mechanism to puncture the bacterial cell membrane (Sahl et al. 2005); cathelicidin, an antimicrobial peptide with a similar mechanism of action to defensins (Kościuczuk et al. 2012); myeloperoxidase, a peroxidase enzyme which produces hypohalous acids which carry out antibacterial activity (Klebanoff et al. 1984) and cathepsin G, a serine protease (Thomas et al. 2014) and broad- spectrum anti-bacterial protein able to target both gram positive and negative bacteria (Shafer et al. 1991)

158 159

These up-regulated antibacterial genes are linked by gene ontology analysis (Section 4.3.2.5) to “response to bacterium” and “disruption of cells of other organisms” pathways. Although CD8+ aren’t commonly cited in the literature as defensin producing cells it is not an entirely novel finding. α-defensins were found in CD8+ T-cells of HIV infected patients, and it was suggested by the authors that it was due to both intrinsic production and uptake from co- cultured cells (D’Agostino et al. 2009). Proteomics, ELISAs or western blots would be able to show for sure whether this upregulation has been translated into protein.

It is interesting that genes that may lead to an increased ability to mount a response to bacterial infections were found in cells from severe asthmatics, suggesting that asthmatics code for a more aggressive cytotoxic T-cell phenotype. It would be interesting to see if this is reflected in defensin release and cellular function as well as expression. Although the role of cytotoxic T- cells is not to respond to bacteria (except intracellular bacteria), CAMP, MPO and CTSG are all genes that can be used to respond to viral infections which may provide a link between the fact that CD8+ T-cells remain in the airways for longer following viral infection (Hogan et al. 2001) and that severe persistent asthma is associated with frequent bacterial and viral exacerbations.

4.4.2.4 Miscellaneous Genes

HIST1H1C (-1.528 log2 FC and corrected p-value 3.14E-02) is the most significantly changed gene and in a proteomics study was shown to be one of the most abundant proteins in mouse cytotoxic T-cells (Hukelmann et al. 2016). The role of Histone H1 proteins is to create higher order chromatin structures and condensing DNA into heterochromatin. This change could equally be a result of stage in the cell cycle or possibly epigenetic changes in levels of packing of the DNA.

159 160

Finally, dual specificity phosphatase 1 (DUSP1) has been purported to be a mediator for ICS and is induced by CS (Abraham & Clark 2006; Clark & Lasa 2003; Clark 2003) and has been implicated in relative CS insensitivity. In this study it was found to have a high fold change, but non-significant increase, confirmed in larger stuies it would be interesting to see if severe asthmatics who weren’t currently taking Oral Corticosteroids also had dysregulated expression of this gene to rule out interference from (Shah et al. 2016).

4.4.3 WGCNA Analysis of CD8+ T-cell Data Using WGCNA with a selected input of genes with raw p-value <0.05 allows for a semi-unbiased selection of pathways associated with various clinical traits. Employing a completely unbiased analysis (i.e. running all expressed genes) was considered too computationally intensive. When analysing these data sets it is important to note that the WGCNA doesn’t take into consideration the direction of change of the genes in it’s output, just that the genes changed are related. These results are discussed below.

4.4.3.1 Green-Yellow Module The Green-Yellow module showed a high positive correlation to asthma severity (r=0.95, p=7x10-06) with only a small non-significant correlation with oral prednisolone usage (r=0.33, p=0.4). This indicates that this group of genes are the ones most likely to be associated to FEV1 % predicted rather than the effects of oral prednisolone or ICS. Using gene ontology, key nodes were found including the genes coding for tubulin α (TUBA1A), dystonin (DST), dynactin subunit 1 (DCTN1), guanine nucleotide-binding protein subunit beta-2-Like 1 (GNB2L1), serine/threonine kinase 16 (STK16), visual system homeobox 1 (VSX1) and armadillo repeat containing, X-linked 2 (ARMCX1) from pathways including “Neuropathy, Distal Hereditary Motor Type VIIB”, “Perry Syndrome, Parkinsonism with Alveolour Hypoventilation and mental depression” and Adenine Phosphoribosyltransferase Deficency”.

160 161

TUBA1A, DST, DCTN1 and STK16 are involved in cell motility and internal localisation. TUBA1A codes for one of the key proteins in tubulin formation (Sancho et al. 2002) and dystonin is required for anchoring intermediate filaments to the actin cytoskeleton and so is key for cell-cell adhesion junction plaques. The three post-transcriptionally created isoforms of DCN have different roles: one is actin binding, one is localised to the nuclear envelope and one to the plasma membrane (Jefferson et al. 2006; Young et al. 2006). It was not possible in this study to distinguish which isoform was present from the gene array data. Dynactin 1 is a subunit of dynactin which can activate ER-to-Golgi transport in a microtubule-independent manner and is involved in vesicle budding (Verissimo et al. 2015). Serine/Threonine Kinase 16 has also been shown to be involved in vesicle budding (In et al. 2014). GNB2L1 codes a protein that is required to allow membrane localisation by VANGL2 (Van Gogh-Like Protein 2) and inhibits Wnt activity (Li et al. 2011).

The two other genes that were determined by this method have not previously been associated with asthma and may be novel mechanisms or candidate targets for novel severe asthma therapies. VSX1 has a well described role in binding to red/green visual pigment in the eye and its protein family also contains deltaEF1. DeltaEF1 knock-out mice have skeletal and T-cell abnormalities with reduced CD4-CD8+ levels relative to CD4+CD8- cells (Higashi et al. 1997). ARMCX2 is involved in development and maintenance of tissue integrity through an unknown mechanism and its role in severe asthma should be examined. It has also been shown to be released in extracellular vesicles, from Colorectal cancer, mesenchymal stem cells and in urine, however its role is unknown (Hong et al. 2009; Fraser et al. 2013; Bruno et al. 2009). For both of these genes it would be interesting to investigate their changes in protein to see if this is simply a gene change or whether they may begin having a practical role.

161 162

4.4.3.2 Light-Green, Red and Blue Module The light green module was found to be positively correlated with a higher

FEV1 % predicted. The Light-Green modules most highly linked genes show a wider variety of pathways that may work together including a paralog of EHMT1 (H3K9 methylase) and a probable histone-lysine N-methyltransferase (Stec et al. 1998), Wolf-Hirschhorn Syndrome Candidate 1 (WHSC1). Protein tyrosine phosphatase, non-receptor type 13 (PTPN13) is another relatively unstudied gene although PTP family members tend to regulate cell growth, differentiation and mitotic cycle (Goebel-Goody et al. 2012; Paul & Lombroso 2003; Denu & Dixon 1998; Zhang et al. 2010). Finally cytoplasmic polyadenylation element binding protein 1 regulates mRNA translation and 3’ untranslated region and may play a role in cell proliferation (Sousa Martins et al. 2016). This gene list did not provide any specific pathway based on the current literature and the role of these genes require further investigation and examination of expression in other chronic inflammatory diseases. However, the role of H3K9 methylation in T-cell function and in severe asthma deserves further study.

The Red and Blue modules are both inversely correlated with FEV1 % predicted and like the light green module have a more disparate selection of genes with no clear connection. The red modules key nodes contain KCNIP2 a potassium voltage-gated channel interacting protein 2 which is a calcium regulator for potassium voltage-gated channels, FGFR4 (fibroblast growth receptor 4), a tyrosine-protein kinase that acts as cell-surface receptor for fibroblast growth factors (Loo et al. 2000) and IL-10 which has anti- inflammatory effects by down regulating the expression of Th1 cytokines (Sher et al. 1991; Zhang et al. 2011). The blue module also contains a variety of genes including the proinflammatory cytokine IL17A, phosphodiesterase PDE9A, the sequence specific DNA binding transaction factor ZNF24, the lysine demethylase KDM4D and endothelin 1 without a clear link between their activation and often with contradictory functions. For example, in the red module IL17 is pro-NF-κB whereas IL10 is inhibitory towards NF-κB.

162 163

4.4.4 Tsitsiou Purple Module The data from Tsitsiou 2012 showed a weak correlation of the purple module with FEV1 % predicted. g:Profiler showed a strong link for many of the genes in vesicular transport and other intracellular transport proteins a few of these genes were SRI which codes Sorcin, which induces movement from the cytoplasm to sarcoplasmic reticulum in response to elevated calcium levels in muscle, RAC1 which codes for Ras-Related C3 Botulinum Toxin Substrate 1 invovled in cytoskeletal reorganization, ADAM10 whose protein is involved intracellular adhesion, RAB1A which codes a GTPase which controls vesicle traffic to the Golgi apparatus and many others with related function. Though the genes were not identical to those found in the patients that had been collected specifically for this study, the key pathway is strongly linked. This demonstrates this phenotype is reproducible in other data sets and warrants further future work.

4.5 Conclusion and next steps

These gene arrays may show that CD8+ T-cells in severe asthma have changes in cytoskeleton activation compared to non-severe asthmatics. This is hinted at by gross decreases in CD81, calponin 2 and golgi adapter ear containing ARF binding protein 1 and significant increases in tubulin α alongside changes in the pathways of vesicle formation and internal cell localisation as shown by WGCNA. The important genes include DST, DCTN1 and GNB2L1, STK16 and VSX1. It is also possible that a new protein involved in vesicle formation has been elucidated, ARMCX2. Currently ARMCX2 has been associated in two cancer papers with a focus on Cisplatin resistance in ovarian cancer and having tumour suppressive roles in Fragile X syndrome patients (Zeller et al. 2012; Rosales-Reynoso et al. 2010), a change in protein expression of Armcx2 in the ECM of damaged liver tissue (Klaas et al. 2016) and a role in murine gonadal development (Smith et al. 2005) with out any clear suggestion of it’s mechanism.

163 164

Activation of T-cells invokes cytoskeleton mechanisms which bind virally infected or cancerous cells, bind to other T-cells when activated and are then required in order to release their cytotoxic payloads from vesicles. This may suggest that this canonically anti-viral, anti-cancerous cell type is involved in asthma in ways other than those previously understood, or that viral infections in severe asthmatics are dealt with in a different manner by CD8+ T-cells. Functional analysis of the anti-viral capacity of animals with restricted and inducible changes in the expression of these proteins will provide further information.

There are also substantial FC increase in antibacterial genes in CD8+ T-cells and U-BIOPRED peripheral mono-nuclear cells in severe asthma. These include up-regulation of defensins (DEFA4 and DEFA1B) and metalloproteases (MMP9 and MPO) and suggests an enhanced anti-bacterial response. The fact that this change is also seen in all mononucleocytes in a much bigger cohort suggests that this may not be simply a CD8+ T-cell response in severe asthmatics, or that it is such a big change in CD8+ T-cells that it is discernable through the noise of the other cell types. Deconvolution of the U-BIOPRED data may give further evidence for abnormalities in the innate, and adaptive, immune systems in severe asthma.

AP1 activation induces the release of defensin from epithelial cells (Wehkamp et al. 2004). I have provided evidence of increased mRNA expression of the AP1 components FOS and JUN in these cells and it is possible that AP-1 drives the increased expression of defensins that we see in CD8+ T-cells in severe asthma, however studies on protein will have to be carried out to confirm this.

The big question to address is whether these changes are as a result of an overall asthmatic phenotype or whether these changes in CD8+ T-cells are driving a distinct asthmatic phenotype. It is also unclear how the peripheral blood CD8+ cells interact with other cell types in the lungs and in the blood

164 165 and does this alter in severe asthma. Bronchoalveolar lavage would provide means to investigate this further with the possibility of using single cell – omics.

The results presented in this chapter provide some evidence that epigenetic changes may be important in severe asthma. Differences in the expression of epigenetic modifying enzymes along with those of histone H1 indicate a link between these changes and epigenetic regulation of the expression of the DEGs reported here. In the next two chapters the effect of epigenome modifying compounds will be tested on CD8+ T-cells and the level of coverage of histone modifications across the genome will be measured using ChIP-Seq.

Though it is not possible to fully confirm the hypothesis with the low numbers of patients samples and even lower numbers after exclusions I was able to investigate the CD8+ T-cell and use the methodology to see suggestions of novel genes involved in the CD8+ T-cells using WGCNA. There were small numbers of significantly changed genes in the severe asthmatics compared to non-severe asthmatics and pathways for these genes were also validated in a previous study. PMBCs transcriptome from severe and non-severe asthmatics and healthy patients did show few differentially expressed genes.

165 166

Chapter 5: H3K27ac in Asthmatic CD8+ T-Cells

166 167

5.1 Chapter Introduction

H3K27 tri-methylation (H3K27me3) has been shown to inhibit expression when found across the body of a gene, cause transcriptional pause when bound to transcriptional start sites (TSS) and it is associated with increased transcription when on the promoter region (Young et al. 2011). This profile was discovered using ChIP-Sequencing, a method to study the level of coverage of the genome with a particular protein. A conference abstract has reported that H3K27me3 demethylation is vital for activation of CD4+ T-cells (LaMere 2015). This study indicated that the inhibitory effect of H3K29ac in the body of genes was abrogated when both naïve and memory T-cells were activated. This variety of transcriptional control and its clear role in the activation of other T-cells makes it an ideal histone modification to study in relation to the modified transcriptome in CD8+ T-cells.

The role of H3K27me3 in transcriptional pause suggests that it can mark bivalent genes which also contain pro-transcriptional histone modifications in the TSS. H3K27ac is a pro-transcriptional mark and has also been shown to be involved in the mechanism of transcriptional pause (Nimura & Kaneda 2015). Increased HAT transferase activity is shown to in asthmatic biopsies of the lung and studies suggest that HAT:HDAC ratios are related to asthma severity (Ito et al. 2002; Su et al. 2008). H3K27ac is associated with enhancer regions of actively expressed genes (Creyghton et al. 2010).

Analysis of the combination of H3K27me3 and H3K27ac marks would begin to give a picture of the CD8+ T-cell epigenome and how they regulate the DEGs that are seen in CD8+ T-cells between severe and non-severe asthmatics.

167 168

5.2 Chapter Hypothesis and Aims

The difference between asthma severity and gene expression of select genes and pathways maybe linked to differences in H3K27me3 and H3K27ac in CD8+ T-cells.

This chapter, therefore, set out to examine the following aims:

• Perform an analysis of H3K27ac and H3K27me3 ChIP-Seq on the same samples as the gene expression data, • Create a dataset with both gene expression and percentage coverage of histone modifications to show the strength of the relationship between expression and H3K27ac and H3K27me3 in severe and non- severe asthmatics.

5.3 Results

CD8+ T-cells were isolated at the same time and in the same manner as the previous chapter and from the same patients. Samples were snap frozen as the samples were taken over the course of 6 months so that the processing could be carried out at the same time to minimise batch effects. However the effects of snap freezing the cells may have had an effect on the on the stability of the DNA and is not part of the standard ChIP-Seq protocol where the DNA would be taken up to the point of being ready to be sequenced before storing.

5.3.1 ChIP Quality Control H3K27ac ChIP was successfully carried out on 3 samples from those described in chapter 4: non-severe NS1, NS4 (excluded from the gene array analysis due to a poor gene count in the PCA), and Severe S6. No inputs or H3K27me3 enrichment for the ChIP-seq were run due to failures to isolate adequate quantities of immunoprecipitated DNA. This may be due to a failure

168 169 of the ChIP, DNA amplification or isolation although the step at which this failure occurred cannot be elucidated. For example, there was no DNA present in the DNA Bioanalyser reads from the inputs, H3K27me3 as demonstrated in Figure 5.1A, and B shows a plot from a successful preparation. As previously mentioned this may have been due to the storage prior to processing.

A B

Figure 5.1. DNA Bioanalyser read DNA Bioanalyser reads showing florescent units against base pair lengths. (A)Two peaks at 100 and 10380 are the ladder used to calibrate reads and (B) an idealised plot of the sample if properly processed should look like.

Of the three samples run the number of reads that were completed was 19,440,407, 14,603,369 and 4,421,840 for NS1, NS4 and S6 respectively. The low read count for S6 may have been the cause for the sparse peaks when viewing the data (Top row in Figure 5.3 - Figure 5.9). However, in general large peaks that were found in other libraries were also present in this data. As such, data from the severe reads can be relied on when a peak is shown, however the lack of a peak may not correlate to a genuine H3K27ac sparse region. As such, care must be taken when analysing this data. We are also able to use FastQC to show that there were some reads which sequenced through the index adapters for the genes (Table 5.1). These were removed by the pipeline. The sequence length distribution (Figure 5.2) shows that read lengths were consistent at ~46bp across the reads.

169 170

Possible Patient Sequence Count Percentage Source Non-Severe GGAAGAGCACACGT TruSeq Adapter, 4 CTGAACTCCAGTCAC Index 16 559319 3.83 CCGTCCATCTCGTAT (97% over 46bp) GC Severe GGAAGAGCACACGT TruSeq Adapter, 6 CTGAACTCCAGTCAC Index 18 57007 1.29 GTCCGCACATCTCGT (97% over 36bp) AT

Table 5.1. Overrepresented sequences in ChIP-Seq Overrepresented sequences from Non-Severe 4 and Severe 6 taken from FastQC report

Non-Severe 1 Non-Severe 4

Severe 8 Sequence Length Distribution

Figure 5.2. Sequence Length Distribution Sequence Length Distribution showing number of reads against sequence length of H3K27ac ChIP from NS1, NS4 and S8. Showing the average read of 46BP.

170 171

5.3.2 H3K27ac ChIP-Seq The only comparison of peak size free of potential error between severe and non-severe asthmatic subjects is that peaks higher in the severe asthmatics are genuinely higher than those of non-severe patients. Peaks higher in non- severe asthmatics could simply be a result of a deficit in reads in the severe subject. Table 5.2 shows the peak value of genes of interest selected from Chapter 4. Of the 20 genes selected only three, CAMP, GGA1 and CD81 were shown to be more highly enriched in H3K27ac in severe asthmatics compared to non-severe asthmatic subjects as demonstrated in their UCSC genome browser reads (CAMP, Figure 5.3; GGA1, Figure 5.4; CD81, Figure 5.5). Of these three genes, the enrichment of H3K27ac was only similar to that in other cell types for GGA1. The H3K27ac peaks at the FOS (Figure 5.6), ZFP36, CNN2 and HIST1H1C’s (Figure 5.7) promoters was lower in severe asthmatics compared to NS asthmatics. This may be a genuine effect but can’t be distinguished from an effect resulting from the lack of reads (Table 5.2). DEFA1 (Figure 5.8), JUND, NDUFB8, MAN2B2, TAF15, TUBA1A and TRIM16L all have absent H3K27ac promoter peaks in the severe patient. In contrast, DUSP1, HDC, JUN, MPO, RBM39 (Figure 5.9) and MX1 all have equal sized H3K27ac promoter peaks between severe and non-severe patients.

171 172

Peak Size at Gene DEG FDR Gene Promoter Comparison Expression in p-value S8 NS1 NS4 Severe asthma <0.05 CAMP 10 5 5 Higher Peak in Severe Upregulated CD81 37 0 5 Higher Peak in Severe Downregulated GGA1 50 40 36 Higher Peak in Severe Downregulated FOS 74 163 116 Lower Peak in Severe Upregulated ZFP36 10 35 35 Lower Peak in Severe Upregulated CNN2 20 30 18 Lower Peak in Severe Downregulated HIST1H1C 15 119 108 Lower Peak in Severe Downregulated * DEFA 0 7 5 Absent Severe Peak Upregulated JUND 0 22 30 Absent Severe Peak Upregulated * NDUFB8 0 50 35 Absent Severe Peak Upregulated MAN2B2 0 25 15 Absent Severe Peak Upregulated * TAF15 0 61 50 Absent Severe Peak Upregulated * TUBA1A 0 5 0 Absent Severe Peak Upregulated * TRIM16L 0 35 16 Absent Severe Peak Downregulated DUSP1 35 35 35 Severe=Non-Severe Upregulated HDC 30 30 30 Severe=Non-Severe Upregulated JUN 46 30 45 Severe=Non-Severe Upregulated MPO 10 23 10 Severe=Non-Severe Upregulated RBM39 62 63 66 Severe=Non-Severe Upregulated * MX1 15 36 15 Severe=Non-Severe Downregulated

Table 5.2. ChIP-Seq H3K27ac Promoter Peak Summary Summary of Promoter peaks sizes and their relation to gene expression of important genes to high-fold change genes and significantly changed genes. Numbers taken from UCSC genome browser.

172 173

Scale 1 kb hg19 chr3: 48,264,500 48,265,000 48,265,500 48,266,000 48,266,500 48,267,000 48,267,500 20 -

Non-Severe 1 i

0 _ 20 -

Non-Severe 4 T

0 _ 20 -

Severe 8 GC

0 _ 20 -

Layered H3K27Ac

0 _ CAMP Figure 5.3. CAMP H3K27ac ChIP-Seq Peaks CAMP centred H3K27ac reads from CD8+ T-cells as shown in UCSC genome browser of individual patients; Severe Asthmatic 6 and Non-Severe Asthmatics 1 and 4 rated on a relative peak height. Beneath Non-Severe 1 is a plot showing H3K27ac peaks from previously reported studies and beneath that purported transcription factor binding sites. The peak of interest is aligned at the gene start site.

173 174

Scale 10 kb hg19 chr22: 38,006,000 38,007,000 38,008,000 38,009,000 38,010,000 38,011,000 38,012,000 38,013,000 38,014,000 38,015,000 38,016,000 38,017,000 38,018,000 38,019,000 38,020,000 38,021,000 38,022,000 38,023,000 38,024,000 38,025,000 38,026,000 38,027,000 38,028,000 38,029,000 56 -

Non-Severe 1 i

0 _ 56 -

Non-Severe 4 T

0 _ 56 -

Severe 8 GC

0 _ 56 -

Layered H3K27Ac

0 _ GGA1 GGA1 GGA1 GGA1 GGA1 Figure 5.4. GGA1 H3K27ac ChIP-Seq Peaks GGA1 centred H3K27ac reads from CD8+ T-cells as shown in UCSC genome browser of individual patients; Severe Asthmatic 6 and Non-Severe Asthmatics 1 and 4 rated on a relative peak height. Beneath Non-Severe 1 is a plot showing H3K27ac peaks from previously reported studies and beneath that purported transcription factor binding sites. The peak of interest is aligned at the gene start site.

174 175

Scale 10 kb hg19 chr11: 2,395,000 2,396,000 2,397,000 2,398,000 2,399,000 2,400,000 2,401,000 2,402,000 2,403,000 2,404,000 2,405,000 2,406,000 2,407,000 2,408,000 2,409,000 2,410,000 2,411,000 2,412,000 2,413,000 2,414,000 2,415,000 2,416,000 2,417,000 2,418,000 2,419,000 2,420,000 2,421,000 2,422,000 2,423,000 56 -

Non-Severe 1 i

0 _ 56 -

Non-Severe 4 T

0 _ 56 -

Severe 8 GC

0 _ 56 -

Layered H3K27Ac

0 _ BC019904 TSSC4 CD81 TSSC4 TSSC4 Figure 5.5. CD81 H3K27ac ChIP-Seq Peaks CD81 centred H3K27ac reads from CD8+ T-cells as shown in UCSC genome browser of individual patients; Severe Asthmatic 6 and Non-Severe Asthmatics 1 and 4 rated on a relative peak height. Beneath Non-Severe 1 is a plot showing H3K27ac peaks from previously reported studies and beneath that purported transcription factor binding sites. The peak of interest is aligned at the gene start site.

175 176

Scale 2 kb hg19 chr14: 75,745,000 75,745,500 75,746,000 75,746,500 75,747,000 75,747,500 75,748,000 75,748,500 75,749,000 75,749,500 190 -

Non-Severe 1 i

0 _ 190 -

Non-Severe 4 T

0 _ 190 -

Severe 8 GC

0 _ 190 -

Layered H3K27Ac

0 _ FOS FOS FOS Figure 5.6. FOS H3K27ac ChIP-Seq Peaks FOS centred H3K27ac reads from CD8+ T-cells as shown in UCSC genome browser of individual patients; Severe Asthmatic 6 and Non- Severe Asthmatics 1 and 4 rated on a relative peak height. Beneath Non-Severe 1 is a plot showing H3K27ac peaks from previously reported studies and beneath that purported transcription factor binding sites. The peak of interest is aligned at the gene start site.

176 177

Scale 2 kb hg19 chr6: 26,052,000 26,052,500 26,053,000 26,053,500 26,054,000 26,054,500 26,055,000 26,055,500 26,056,000 26,056,500 26,057,000 26,057,500 26,058,000 26,058,500 26,059,000 150 -

Non-Severe 1 i

0 _ 150 -

Non-Severe4 T

0 _ 150 -

Severe 8 GC

0 _ 150 -

Layered H3K27Ac

0 _ HIST1H1C Figure 5.7. HIST1H1C H3K27ac ChIP-Seq Peaks HIST1H1C centred H3K27ac reads from CD8+ T-cells as shown in UCSC genome browser of individual patients; Severe Asthmatic 6 and Non-Severe Asthmatics 1 and 4 rated on a relative peak height. Beneath Non-Severe 1 is a plot showing H3K27ac peaks from previously reported studies and beneath that purported transcription factor binding sites. The peak of interest is aligned at the gene start site.

177 178

Scale 1 kb hg19 chr8: 6,835,000 6,835,500 6,836,000 6,836,500 6,837,000 6,837,500 6,838,000 35 -

Non-Severe 1 i

0 _ 35 -

Non-Severe 4 T

0 _ 35 -

Severe 8 GC

0 _ 20 -

Layered H3K27Ac

0 _ DEFA1 Figure 5.8. DEFA1 H3K27ac ChIP-Seq Peaks DEFA1, DEFA3 and DEFA1B centred H3K27ac reads from CD8+ T-cells as shown in UCSC genome browser of individual patients; Severe Asthmatic 6 and Non-Severe Asthmatics 1 and 4 rated on a relative peak height. Beneath Non-Severe 1 is a plot showing H3K27ac peaks from previously reported studies and beneath that purported transcription factor binding sites. The peak of interest is aligned at the gene start site.

178 179

Scale 20 kb hg19 chr20: 34,285,000 34,290,000 34,295,000 34,300,000 34,305,000 34,310,000 34,315,000 34,320,000 34,325,000 34,330,000 34,335,000 70 -

Non-Severe 1 i

0 _ 70 -

Non-Severe 4 T

0 _ 70 -

Severe 8 GC

0 _ 70 -

Layered H3K27Ac

0 _ NFS1 RBM39 NFS1 RBM39 NFS1 RBM39 NFS1 RBM39 NFS1 RBM39 ROMO1 RBM39 ROMO1 RBM39 RBM39 RBM39 RBM39 Figure 5.9. RBM39 H3K27ac ChIP-Seq Peaks RMB39 centred H3K27ac reads from CD8+ T-cells as shown in UCSC genome browser of individual patients; Severe Asthmatic 6 and Non-Severe Asthmatics 1 and 4 rated on a relative peak height. Beneath Non-Severe 1 is a plot showing H3K27ac peaks from previously reported studies and beneath that purported transcription factor binding sites. The peak of interest is aligned at the gene start site.

179 180

5.4 Discussion

An increase in H3K27ac at the TSS normally associates with an increased expression of a gene (Creyghton et al. 2010) and this is demonstrated by CAMP (Figure 5.3). CAMP gene expression is increased in severe asthma cells and the enrichment of the H3K27ac mark suggests that it is controlled by the expected mechanism of enhanced histone acetylation. However, both GGA1 (Figure 5.4) and CD81 (Figure 5.5) had an increased peak of H3K27ac but a reduced gene expression demonstrated by gene array. This suggests that GGA1 and CD81 could be regulated by bivalent regions at the TSS which require both H3K27ac and H3K27me3 histone modifications, or other regulatory marks, to act in concert to enable gene expression to occur.

The CAMP gene encodes the cathelicidin antimicrobial peptide whose activation, therefore, is likely to be controlled by histone acetylation status. This enrichment of H3K27ac could also account for the increased expression of the DEFA genes (Figure 5.8) which are contained in one locus. However, the H3K27ac mark at the TSS has a gap in the severe asthma subject in this region. It has been shown previously that β-defensins and cathelicidin are controlled by acetylation in epithelial cells during infection (Shin et al. 2010; Chakraborty et al. 2008). Specific ChIP analysis may confirm whether this is the case here. HDC and MPO also show an equal enrichment at their promoters between the severe and non-severe patients, again the enrichment at the TSS in the severe asthmatic subject could have been underestimated by the ChIP-Seq.

CNN2 and HIST1H1C (Figure 5.7) have lower H3K27ac peaks in the severe asthmatic cells and the expression of both these genes are downregulated. This result is expected although this data cannot be relied upon since both FOS and ZFP36, which also have reduced H3K27ac peaks, are up-regulated at the gene level in these cells.

180 181

5.5 Conclusion and next steps.

This experiment needs to be repeated, not only is the n=1 in severe and n=2 non-severe insufficient for drawing broader conclusions about severe asthma, the read count is also far too low. The lack of H3K27me3 data prevented the elucidation of the role of this mark in the regulation of gene expression in human primary CD8+ T-cell and the location of bivalent genes. Ideally, a repeat experiment should be performed using much greater subject numbers that would allow analysis of subtypes of the disease. Greater cell numbers will provide more RNA and cDNA for analysis and allow the use of input DNA to assess the level of enrichment more accurately.

If all patients had been correctly analysed the data set would have allowed the relationship between the expression of every gene and its associated histone acetylation and tri-methylation score to be assessed. This would have enabled me to determine on a global genetic scale, the genes which are differentially regulated between severe and non-severe asthmatics by alterations in histone modifications.

Although the levels of acetylation and methylation were not adequately measured using ChIP-seq, in the next chapter I use epigenetic mimics to help understand the mechanisms that control gene and protein expression in CD8+ T-cells.

It was not possible to address the hypothesis for this chapter due to the low n- numbers due to poor processing protocols. In future attempts samples should be processed before storing. It may be sensible to compare single patients to their expression rather than the averages across the cells as a general merthodology if enough patients work efficiently with both ChIP-Seq and Gene Array.

181 182

Chapter 6: The Role of CBP and p300 in CD8+ T-cells of Severe Asthmatics

182 183

6.1 Chapter Introduction

MIP1α/CCL3 is a marker of CD8+ T-cell activation and is strongly chemotactic for eosinophils and macrophages suggesting a role in both Th1 and Th2 phenotypes (Cocchi et al. 1995; Oliveira et al. 2002; Reichel et al. 2009). Median levels of MIP1α have been shown to be increased in asthmatic sputum (Daldegan et al. 2005).

Th2 type inflammatory inhibition has been associated with Brd-linked acetylation (Pinz et al. 2015). This suggests that BET mimics/probes such as bromosporine, JQ1 or SGC-CBP30 may be able to reduce inflammation in CD8+ T-cells as has been demonstrated with JQ1 in LPS- and IL-1β-induced inflammation in mouse macrophages and human epithelial cells respectively (Filippakopoulos et al. 2010; Khan et al. 2014). Histone methylation is associated with CD4+ T-cell differentiation and H3K4 tri-methylation is found in Th2 cells at the IL4/IL13 locus demonstrating an activating property (Wei et al. 2009). This indicates that using methylated lysine mimics such as GSK-J1 and GSK-J4 may also limit Tc2 cell activation. There are many methylases which influence T-cell transcriptional activation, both positively and negatively. For example G9a, which targets H3K9 methylation, inhibits the release of VEGF which modulates basement membrane remodelling, EZH2 catalyses H3K27 trimethylation and can block production of IL-4 and IL-13 whilst SETD7, which specifically targets H3K4, modulates gene activation at NF-κB binding sites (Barsyte-Lovejoy et al. 2014; Wang et al. 2001; Rai et al. 2010; Yamashita et al. 2006; Tumes et al. 2013). Mimics such as UNC0642, GSK343, and PFI-2 block G9a, EZH2/1 and SETD7 actions respectively.

6.2 Chapter Hypothesis and Aims

CD8+ T-cells from severe asthmatics will have a novel mechanism of control to activate release of inflammatory cytokines that can be investigated by using epigenetic probes/mimics. The probes will have differing effects on the

183 184 function of CD8+ T-cells and bronchial epithelial cells. To investigate this hypothesis I will: • Measure the inhibitory effects of epigenetic mimics on Dynabead induced release of MIP1α from CD8+ T-cells of severe asthmatics and • Investigate the mechanism of action of mimics which reduce inflammation by carrying out ChIP and PCRs of key genes.

184 185

6.3 Results

6.3.1 Primary human CD8+ T-Cells

6.3.1.1 Effect of epigenetic probes/mimics on dynabead-induced MIP1α release from CD8+ T-cells isolated from severe asthmatics Cells from two patients (Subject 1 and 2, Table 6.1) were used for most of the screening experiments but a third subject (Subject 3, Table 6.1) was also used for the SGC-CBP30 studies to enable the use of statistics. The doses and activators were selected using a TALL-104 model cell line. The severe asthmatics were as defined in Section 2.6.1. Two of three patients had been on a course of oral prednisolone in the previous year and all were on >800mg inhaled FP or equivalent and were atopic (Table 6.1).

Most compounds (Figure 6.1: UNC0638 (B), UNC0642 (C), JQ1(-) (E), Bromosporine (F), GSK-J1 (G), GSK-J4 (H), UNC1999 (I), GSK-343 (J)) had no apperent effect on the dynabead-stimulated release of MIP1α from severe asthmatic cells. JQ1(+) had a variable concentration-dependant response (Figure 6.1D) reaching a maximal effect at 320nM of 30.6±20.4 inhibition. If the n-number had been increased it is likely that it would have been a significant increase. SGC-CBP30 significantly inhibited MIP1α release at the highest dose from stimulated CD8+ T-cells with a maximal response at 1600nM which reduced

MIP1α output to 10.1±6.2% of the maximal release with an IC50 = 210nM. MTT showed no significant decrease in mitchondiral activity suggesting acceptable viability.

185 186

ICS (FP Prednisolone FEV1 FEV1 FVC FEV1 # Age Atopy Severity equivalent, (mgs) (L) % Pred (L) :FVC mgs) 1 31 + Severe 20mg 1000 3.3 104 3.8 84 2 courses in past 2 48 + Severe 800 2.67 95 3.35 90 year 3 29 + Severe 0 800 2.67 82 3.6 75

Table 6.1. Patients Demographics for CD8+ T-cells used for compound screening Demographics of severe asthma volunteers whose peripheral CD8+ T-cells were used to screen compounds. Cells from patient 1 were used for all epigenetic mimics, cells from patient 2 were used for all studies except those for GSK-J5 and UNC1215. Patient 3 was only used for studies using SGC- CBP30. ICS = inhaled corticosteroid, FP = fluticasone propionate, FEV1 = forced expiratory flow in 1 second, FVC = forced vital capacity, % pred = percent predicted.

186 187

A B C 150 150 150

100 * 100 100 (% Max) (% Max)

(% Max)

α α α 50 50 50 MIP1 MIP1 MIP1

0 0 0 Dynabeads - + + + + + + + Dynabeads - + + + + + + + Dynabeads - + + + + + + + 0 0 0 0 50 0 0 100 200 400 800 125 250 500 75 1600 62.5 1000 150 300 600 31.25 37.5 1200 D SGC-CBP30 (nM) E UNC0638 (nM) F UNC0642 (nM) 150 150 150

100 100 100 (% Max) (% Max)

(% Max)

α α α 50 50 50 MIP1 MIP1 MIP1

0 0 0 Dynabeads - + + + + + + + Dynabeads - + + + + + + + Dynabeads - + + + + + + + 0 0 0 0 20 40 80 10 20 40 80 0 0 160 320 640 160 320 17 35 71 142 284 568

JQ1(+) (nM) JQ1(-) (nM) Bromosporine (nM) G 150 H 150

100 100 (% Max) (% Max)

α α 50 50 MIP1 MIP1

0 0 Dynabeads - + + + + + + + Dynabeads - + + + + + + 0 0 0 0 12 24 48 96 12 24 48 96 192 192

GSK-J1 (nM) GSK-J4 (nM)

I 200 J 200

150 150 (% Max) (% Max)

100 100 α α

50 50 MIP1 MIP1

0 0 Dynabeads - + + + + + + + Dynabeads - + + + + + + + 0 0 0 0 44 89 26 52 178 357 714 104 208 416 833 1428 UNC1999 (nM) GSK 343 (nM) Figure 6.1. Concentration-Response to Epigenetic Probes of Dynabead- Stimulated Severe Asthamtic CD8+ T-cells Concentration response of SGC-CBP30 (A), UNC0638 (B), UNC0642 (C), JQ1(+) (D), JQ1(-) (E), bromosporine (F), GSK-J1 (G) GSK-J4 (H), UNC1999 (I), GSK343 (J) and UNC1215(K) 24 hours post stimulation of peripheral blood CD8+ T-cells (2x105/ml) with dynabeads (2x105 beads/ml). n=2 (except A which has n=3). * = p<0.05 Freidman’s test (ANOVA non-parametric matched pairs) MTT showed no decrease in viability.

6.3.2 RT-qPCR In separate experiments, CD8+ T-cells from two severe asthmatics (Table 6.2) were stimulated with dynabeads in the presence and absence of SGC- CBP30 and PCR analysis was performed on FOXP3, TUBA1A, JUND,

187 188

DEFA1B, CD81, MIP1α and IFNγ mRNAs. TUBA1A, JUND, DEFA1B, CD81 were selected based on their changes in the severe asthma patients from Chapter 4. FOXP3 was selected as it is T-cell differentiation control gene which is regulated by CBP/p300, MIP1α on account that its protein output is inhibited by SCG-CBP30 and IFNγ as a marker for Tc1 type activation. JUND (Figure 6.2B) expression was decreased after stimulation (ΔCT, 3.06 ± 4.33 x106 v’s 2.58 ± 3.6 x105) and further attenuated with SGC-CBP30 pre- treatment (1.06 ±1.48 x104). The expression of FOXP3 (Error! Reference source not found.A), IFNγ (Error! Reference source not found.C), CD81 (Error! Reference source not found.D) and MIP1α (Error! Reference source not found.E) mRNAs all decreased with dynabeads but were unaffected by treatment with SGC-CBP30 (FOXP3 ΔCt, unstimulated = 4.17 ± 5.90 x1015, dynabeads = 5.64 ± 7.96 x108, dynabeads/SGC-CBP30 = 2.18 x108± 3.07 x106; CD81 ΔCt, unstimulated = 2.88 ± 4.07 x108, dynabeads = 5.01 ± 8.39 x106, dynabeads/SGC-CBP30 = 5.38 ± 7.61 x106; MIP1α ΔCt, unstimulated = 2.41 ± 1.58 x1010, dynabeads = 2.23 ± 1.4x108, dynabeads/SGC-CBP30 = 2.30 ± 2.04 x108; IFNγ ΔCt, unstimulated = 9.02 ± 4.38 x1010, dynabeads = 6.86 ± 7.55 x106, dynabeads/SGC-CBP30 = 6.08 ± 6.74 x106). TUBA1A mRNA expression increased after dynabeads stimulation (ΔCt, 1.14 ± 1.61 x10-3 v’s 8.67 ± 0.12 x10-2) and further increased 10-fold with SGC- CBP30 (ΔCt, 0.11 ± 0.15). Finally, DEFA1B expression increased post-dynabead stimulation (ΔCt, 9.54 ± 13.50 v’s 13.93 ± 19.70) which was reversed by pre-treatment with SGC- CBP30 (ΔCt, 3.55 ± 5.028).

Prednisolone FP (mg) FEV1 FEV1 % FEV1: Age Atopy Severity FVC (mgs) Equivalent (L) predicted FVC 47 + Severe 0 1000 2.87 72.5 3.92 68 68 - Severe 0 500 2.03 83 2.7 87

Table 6.2. Patient Demographics CD8+ T-cells used for SGC-CBP30 RT- qPCR screening

188 189

Demographics of severe asthmatics whose peripheral CD8+ T-cells were stimulated with dynabeads in the presence and absence of SGC-CBP30 and analysed using RT-qPCR.

A B 1020 107 106 15 10 105 CT CT

Δ 4 Δ 10 1010 103

JUND JUND 2 FOXP3 FOXP3 105 10 101 100 100

Dynabeads Dynabeads Unstimulated Unstimulated

Dynabeads + CBP30 Dynabeads + CBP30 C 1015 D 109 108 107 1010 106 CT CT

Δ 5

Δ 10

γ 104

IFN 5 3

10 CD81 10 102 101 100 100

Dynabeads Dynabeads Unstimulated Unstimulated E Dynabeads + CBP30 F Dynabeads + CBP30 1015 15

1010 CT 10 Δ CT Δ

CCL3 10 5 DEFA1A

100 0

Dynabeads Dynabeads Unstimulated Unstimulated

Dynabeads + CBP30 Dynabeads + CBP30 G 1

0.1 CT Δ 0.01

TUBA1A TUBA1A 0.001

0.0001

Dynabeads Unstimulated

Dynabeads + CBP30 Figure 6.2. RT-qPCR analysis of gene expression in CD8+ T-cells from Severe asthmatics after SGC-CBP30 and Dynabeads treatment Gene expression by RT-qPCR of FOXP3 (A), JUND (B), IFNγ (C), CD81 (D), CCL3 (E), DEFA1A (F) and TUBA1A (G) from Severe Asthmatics’ (Table 6.2)

189 190

Peripheral CD8+ T-cells after two hours stimulation with 2x105/ml dynabeads and pre-treated for one hour with 400nM SGC-CBP30. n=2.

6.3.3 CREB Binding Protein ChIP CD8+ T-cells were isolated from 2 severe asthmatics (Table 6.3), stimulated with dynabeads in the presence and absence of SGC-CBP30 with the same number of cells in each well for each patient. CBP immunoprecipitation followed by PCR analysis of FOXP3 (Figure 6.3), IFNγ (Figure 6.4) and MIP1α (Figure 6.5) promoters was performed to investigate whether CBP was associated with regulatory regions of these key genes. CBP binding to the FOXP3 promoter was markedly decreased by dynabead stimulation in both subjects (Fold Enrichment (FE), the Input:Output ratio, 1.079 to 0.145 in severe subject 1 and 5.540 to 0.062 in severe subject 2). The effect of CBP30, however, was variable with a further decrease in severe subject 1 (FE = 0.008) and an increase in severe subject 2 (FE = 1.050). IFNγ showed minimal CBP enrichment after dynabead stimulation (3.84x10-3 v’s 9.38x10-3) in severe subject 1 but this binding was completely ablated in the presence of CBP30. No Ct score was observed from the PCR run for dynabeads/CBP30. In severe subject 2, CBP promoter binding at the IFNγ promoter decreased after dynabeads (FE, 0.345 v’s 1.45x10-3) and further decreased in the presence of dynabeads/CBP30 (FE, 2.1x10-5). CBP binding to the MIP1α promoter of severe asthma subject 1 showed limited change between unstimulated and dynabeads-stimulated samples (FE, 7.57x10-3 v’s 8.89x10-4) in severe subject 1 but showed no binding after dynabeads/CBP30 exposure with no detectable Ct score from the PCR run. In contrast, CBP MIP1α promoter binding in severe subject 2 decreased after dynabeads stimulation (FE, 0.135 v’s 1.07x10-3) and a slight increase in CBP binding with CBP30 (2.889x10-3).

190 191

Prednisalone FP (mg) FEV1 FEV1 % FEV1: # Age Atopy Severity FVC (mgs) Equivilent (L) predicted FVC 0 (3 courses a Severe 1 56 - Severe 500 2.56 78 3 79 year) Severe 2 29 + Severe 0 1000 2.25 69 3.55 63 Table 6.3. Patient Demographics for SGC-CBP30 treated CD8+ T-cells used for CBP ChIP Demographics of subjects whose peripheral CD8+ T-cells were stimulated with dynabeads (2x105/ml) in the presence and absence of SGC-CBP30 (400nM) and CBP binding analysed using PCR after 2 hours of stimulation.

A Severe 1 B Severe 2 10 10

1 1

0.1

0.1 (Input:Output) 0.01 (Input:Output) Fold Enrichemnt FOXP3 Enrichemnt Fold Fold Enrichemnt FOXP3 Enrichemnt Fold 0.001 0.01

Unstim Unstim Dynabeads Dynabeads

Dynabeads + CBP30 Dynabeads + CBP30

Figure 6.3. CBP Binding to the FOXP3 Gene after SGC-CBP30 treatment CBP Binding to the FOXP3 gene promoter in peripheral blood CD8+ cells from 2 severe asthmatic subjects (Table 6.3) 2 hours after stimulation with dynabeads (2x105/ml) and pre-treated for one hour with 400nM SGC-CBP30.

191 192

A Severe 1 B Severe 2 1 1 γ γ 0.1 IFN IFN 0.1 0.01

0.001 0.01 (Output:Input) (Output:Input) 0.0001 Fold Enrichment Fold Enrichment 0.001 0.00001

Unstim Unstim Dynabeads Dynabeads

Dynabeads + CBP30

Figure 6.4. CBP Binding to the IFNγ Gene after SGC-CBP30 treatment CBP Binding to the IFNγ gene promoter in peripheral blood CD8+ cells from 2 severe asthmatic subjects (Table 6.3) 2 hours hours after stimulation with dynabeads (2x105/ml) and pre-treated for one hour with 400nM SGC-CBP30. Severe 1’s SGC-CBP30 treatment did not cross threshold in PCR.

A Severe 1 B Severe 2 1 1 ) ) 0.1 0.1

0.01

0.01 Output:Input Output:Input ( ( 0.001 Fold Enrichment CCL3 Fold Enrichment CCL3 0.0001 0.001

Unstim Unstim Dynabeads

Dynabeads + CBP30 Dynabeads + CBP30

Figure 6.5. CBP Binding to the CCL3 Gene after SGC-CBP30 treatment CBP Binding to the MIP1α gene promoter in peripheral blood CD8+ cells from 2 severe asthamtic subjects (Table 6.3) 2 hours hours after stimulation with dynabeads (2x105/ml) and pre-treated for one hour with 400nM SGC-CBP30. Severe 1’s dynabead stimulation alone did not cross threshold in PCR.

192 193

6.4 Discussion

SGC-CBP30, the p300/CBP specific bromodomain inhibitor (Hay et al. 2014), reduced anti-CD3/CD28-stimulated MIP1α release in a concentration- dependant manner in CD8+ T-cells (Error! Reference source not found. & Figure 6.1). This demonstrates the potential to expand the power of this study by recruiting more patients. If these findings were to remain valid there may be potential for the therapeutic use of compounds with similar targets since reducing the inflammatory response of CD8+ T-cells in the asthmatic airways could reduce bronchoconstriction and airway remodelling, particularly post- infection (Hogan et al. 2001).

The MIP1α gene expression showed a small decrease following dynabeads stimulation, but no difference between SGC-CBP30 stimulation with dynabeads and dynabeads alone. Considering the data in other cell types it was surprising that CBP30 did not decrease anti-CD3/CD28 dynabead- induced MIP1α mRNA expression (Figure 6.2E) and it is unclear from the current ChIP analysis (n=2) whether CBP protein binding to the MIP1α promoter is affected by the CBP30 mimic. Increasing the number of subjects studied will begin to clarify these issues. One possible alternative explanation is that p300/CBP has a role in MIP1α translation and/or release from vesicles rather than on gene expression per se. p300, which is closely linked to CBP in both homology and function, has a role in macrophage autophagy and directly acetylates several core proteins including microtubule-association protein 1A/1B light chain (LC3) (Lee & Finkel 2009). p300 inhibition has also been shown to increase the activity of acetyl transferase α-tubulin N-acetyl transferase 1 (αTAT1) which hyper- acetylates tubulin (Mackeh et al. 2014). This could suggest a role for CBP/p300 in microtubule release of vesicles containing MIP1α rather than direct transcriptional roles. It will be important to look at other proteins released by CD8+ T-cells to investigate whether there is inhibition of the

193 194 release of other inflammatory cytokines and cell puncturing apparatus such as IFNγ. The expression of Ifng expression was suppressed after dynabead stimulation (Figure 6.2B) and was unchanged after CBP30 stimulation. It has been shown previously that CBP30 has no effect on IFNγ in Th17 cells (Hammitzsch et al. 2015) and so it would be interesting to see its level of release in these CD8 T-cells and in other T-cell subtypes. JQ1+ has a similar effect to CBP30 on release of MIP1α if after increasing the n-number of these studies it maintained this effect it would be interesting to see the mechanism of action that it was able to carry this out by.

As there was a change in the release of MIP1α and high serum MIP1α has been linked to decreased FOXP3 stability, I explored the possibility that FOXP3 instability is associated with a reduction of its expression. This data demonstrated that FOXP3 was downregulated when stimulated with dynabeads suggesting that it could be cell activation which reduces FOXP3 expression. This would be sensible as activation of Tc cells will reduce the Treg function of these cells which is under FOXP3 control.

TUBA1A mRNA expression (Figure 6.2E) increases after stimulation of CD8 cells by dynabeads whereas the expression of CD81 (Figure 6.2D) decreases and CBP30 has no effect on the expression of either. The dynabead stimulation has enhanced a set of traits that were previously shown in severe asthma compared to non-severe asthma (Section 4.4.2.1). Dynabeads also demonstrated an enhancement of the previously demonstrated severe asthma phenotype with DEFA1A expression similar in Section 4.4.2.3. DEFA1A is the only mRNA investigated whose expression was increased following dynabead stimulation and was decreased, but not significantly, with SGC-CBP30. This suggests that CBP30 could counter the differential expression of this gene as reported in CD8+ cells from severe asthmatics as studied in Chapter 4. It will be important to determine the global effects of this mimic on gene expression profiles from severe asthma cells to obtain a full understanding of potential benefits and side-effects of CBP30.

194 195

The increase in tubulin mRNA expression could suggests a move towards the CD8+ cells adopting more cell-cell adhesion characteristics if this manifested in protein expression. An increase in cell-cell adhesion could be observed under the microscope following dynabead stimulation (Figure 6.6

Figure 6.6. Micrograph of CD8+ T-cell Clustering around Dynabeads CD8+ T-cells (clear) bind to each other and associate with dynabeads (dark) (x400, no stain).

The expression of JUND, which has been previously reported to have a role in steroid insensitivity (Lane et al. 1998), was downregulated after dynabead stimulation and further down-regulated by CBP30. This is in contrast to the elevated expression of JUND seen by Lane and colleagues examining total PBMCs but further studies will need to be conducted to determine whether there is differential expression in patients with different asthma severity. The suppression of JUND by CBP30 suggests that it may be effective in subjects/cells where it is overexpressed.

6.5 Conclusion and next steps

Dynabead activation of CD8+ T-cells has differential effects on the genes studied here. They enhance or inhibit the expression of DEFA1A, CD81 and TUBA1A in similar ways to that seen in the cells from severe asthma subjects in the previous chapter.

This preliminary data, using low n-numbers, demonstrates that CBP30 limited the release of MIP1α without affecting the total expression of MIP1α/CCL3

195 196 mRNA. p300 has a role in acetylation of tubulin and microtubule acetylation, which has been mostly studied in autophagy (Lee & Finkel 2009). This effect of p300, rather than an effect on acetylated histone interactions with basal transcription factors, may account for the suppression of the release of MIP1α-containing vesicles.

It is possible that the mechanism involved in the release of MIP1α is specifically different from the cytokine release of the the bronchial epithelium as in chapter 3. SGC-CBP30 is able to have a significant effect on MIP1α release at high doses, where as this did not happen in bronchial epithelium suggesting a cell type specific effect of this compound. SGC-CBP30 did have a dose responsive effect and higher n-numbers may have shown this to be significant at much lower doses. JQ1+ seemed to also have a similar effect which may be possible to ascertain if the n-number was increased, however this may work by a different mechanism as it also had this effect on the epithelium. I was unable to elucidate the specifics of the mechanism which inhibited the release of MIP1α, as the n-numbers were so low, however it is possible that the effects of SGC-CBP30 operate via a non-mRNA route, perhaps an effect on Tubulin acetylation. To further study this it could be possible to co- immunoprecipitate CBP and Tubulin to demonstrate an association once the n-numbers were increased to confirm that this infact the case.

196 197

Chapter 7: Discussion

197 198

7.1 General Discussion

Using a combination of differential gene expression analysis from published U-BIOPRED data on bronchial brushing cells, proteomics analysis by SILAC of FP actions on healthy human airway epithlelial cells and pharmacologic analysis of epigenetic mimics, I was able to demonstrate small mRNA changes in histone H1 and H2B expression in severe asthma and that the BET mimic JQ1 was able to suppress IL8 release without affecting cell proliferation or viability in eptithelail cells. In contrast, SGC-CBP30, that targets CBP/p300, prevented epithelial cell proliferation without affecting inflammation. Interestingly, CITED2 mRNA, which encodes CBP/p300 interacting protein, was overexpressed in bronchial biopsies but not airway structural cells of severe asthmatics. In peripheral blood CD8+ T-cells SGC- CBP30 was able to suppress the release of the inflammatory mediator MIP1α which was also affected by JQ1, but not bromosporin without affecting CCR3 mRNA levels. Gene expression analysis of CD8+ T-cells from severe asthmatics and non-severe asthmatics suggested the importance of RNA modifying genes, the pro-inflammatory AP-1 complex and antimicrobial genes. Using WGCNA it was possible to identify a module that was associated with asthma severity (as determined by FEV1 % predicted) but independent of steroid use. This module contained genes linked to alterations in histone methylation and with intracellular organelle movement such as

TUBA1A. Other modules were related to FEV1 % predicted which contained interesting genes such as IL-10, IL-17 as well as others with little known interactive function. Further higher power analysis may define the pathways involved in selective CD8+ gene expression linked to asthma severity which may provide novel pathways for drug discovery.

7.1.1 Bronchial Epithelial Epigenome Exploration The hypothesis that epigenetic status and global transcriptome of the bronchial epithelium in severe asthma could help to discover novel pathways was first tested by using pre-existing expression and proteomic data sets.

198 199

The U-BIOPRED bronchial brushings showed significant differences in the gene expression of histone cluster 1, H1E and H2BK between severe asthmatic subjects and healthy controls which was not observed in mild/moderate asthmatics (Section 3.3.1.1). These histones have differing affinities for DNA (Halmer & Gruss 1996; Sun et al. 2015) and are involved in the regulation of gene expression. The data suggests that the differing affinity may affect the global response to perturbation by exogenous stimuli that occurs in cells from severe asthmatics. No changes in the gene expression of histone modifying enzymes were shown, but this does not mean that their activity or protein expression was unchanged as they could be regulated at the post-translational level, for example phosphorylation of HDAC7 alters its nucleocytoplasmic shuttling properties (Gao et al. 2006). The effects of the anti-inflammatory asthma treatment fluticasone propionate (FP) was also studied in primary human bronchial epithelial cells. FP effects on changes in the incorporation of amino acids, reflecting de novo protein synthesis, was measured by SILAC and downstream proteomic analysis by mass spectroscopy was performed. The low power SILAC data hinted at the idea that FP decreased the amount of HDAC7 and of SETD7 and SMYD3, two H3K4 methyltransferases, in cells from healthy non-asthmatic subjects. HDACs are associated with modulating inflammatory gene expression and other cellular functions (Elsharkawy et al. 2010; Ashburner et al. 2001) and HDAC2 is involved in mediating CS responsiveness (Barnes et al. 2004).

HDAC7 is been shown to be involved in the trans differentiation of Pre-B Cells into macrophages and suppresses functions of macrophages including immune and inflammatory responses (Barneda-Zahonero et al. 2013), as FP reduces HDAC7 there is potential for it to be a mechanism of FP to reduce inflammation.

The role of H3K4 methylases in regulating the inflammatory response and its control by CS is currently unclear, however an increase in the expression of these specific enzymes may suggest that altered methylation may be

199 200 important in CS function. H3K4 methylation is implicated in the control of gene expression (Vijayanand et al. 2012) and may therefore be a key target of CS (Jangani et al. 2014).

To investigate changes in bronchial epithelial histone modifications a multiplex ELISA of a variety of histone modifications was carried out. This demonstrated that there were no global changes in levels of most histone modifications at baseline and that FP had no effect on whole cell acetylation status. This may reflect the fact that acetylation effects regulated by FP are selective for responsive genes or that these modifications in healthy patients are relatively stable with high levels of redundancy. This redundancy is demonstrated by the siRNA depletion of HATs in drosophila cells, which has no effect and that the acetylome is probably more fixed than transient (Imhof et al. 2015).

The next step which should have been followed was a full histone modification map of H3K4 methylation, H3K27ac and H3K27me3 and expression studies in the same bronchial epithelial samples. This would have enabled me to differentiate the effects of FP on specific histone marks in a gene- and promoter/enhancer-specific manner. These gene-specific effects on histone status are more likely to reflect FP actions than alterations in global methylation. This approach may also have defined differences in the cellular methylome between severe and non-severe asthmatics and the extent to which these modifications controlled the changes in gene expression identified in the U-BIOPRED samples. The data from U-BIOPRED has identified differences in gene expression between patients and is large enough that patient/sample sub-setting is viable. It would have been an ideal opportunity to explore differences in histone modifications in a large set of severe asthmatic subjects using ChIP-seq analysis. However, the pace of sample collection and cell culture issues resulted in this approach not being viable to complete over the course of this

200 201

PhD. Future studied should prioritise this type of analysis to improve our understanding of the bronchial epithelium’s relationship with severe asthma.

Of the many selective epigenome mimics used to analyse cell function, two broad spectrum bromodomain readers inhibited the release of CXCL8 and IL- 6 - bromosporine and JQ1(+). JQ1(+)s anti-inflammatory effects have been demonstrated in other cell types (Magistri et al. 2016; Meng et al. 2014). The more specific BET inhibitor, CBP30/p300, had no effect on the release of inflammatory mediators from bronchial epithelial cells measured in this study, but did have an effect on proliferation which is relevant to their mechanism of action which will be discussed later.

7.1.2 CD8+ T-cell transcriptome Exploring the CD8+ T-cell transcriptome in severe asthma was fruitful in determining genes and more importantly novel pathways linked with severe asthma.

This thesis highlighted the differential expression of a group of RNA modifying proteins, namely SSB and TAF15. TAF15 modifies the transcription of RNA and SSB inhibits exonuclease digestion of RNA. This could give them the ability to increase the life span of RNAs which are differentially expressed in these cells. Further study of these proteins may shed light on the cause of the differential expression of the other genes and ncRNAs seen in these cells in severe asthma. Fos, Jun & JunD which are components of the inflammation- and proliferation- linked transcription factor AP-1 were all highly increased in severe asthmatic CD8+ T-cells. This suggests that these cells are highly activated which is also indicated by the enhanced expression of another group of upregulated antimicrobial genes DEFA1B/A3/A1, CAMP, MPO and CTSG. Why these genes should be activated to defend against infectious agents in stable subjects with severe disease is interesting and may reflect a sub-clinical effect

201 202 of microbiome or virome dysregulation. It would be interesting to measure the release and internal storage of these proteins to investigate whether this is simply a change in mRNA expression or whether there is translation occurring to enact this phenotype. Future studies should investigate the basal and dynabead-stimulated release of analytes from CD8+ T-cells from healthy, non-severe and severe asthmatics using unbiased proteomics or SomaLogic analysis to confirm this transcriptomic profile at the protein level.

Interestingly when the PMBC transcriptome in the U-BIOPRED dataset was analysed, the expression of DEFA4, LTF, MMP8 and CEACAM8, genes representing anti-microbial proteins, were found to be decreased in severe asthmatics. This decrease in global expression of anti-microbial genes suggests that the increase observed in isolated CD8+ T-cells is selective and is in direct contrast to what other cells in the blood are producing. This may reflect a selective recruitment of these cells to the airways from the blood compartment in severe asthma. In this respect, it would be extremely useful to examine the expression of anti-microbial genes from isolated or microdissected CD8+ T-cells from airway biopsies to determine whether the same activation profile is occurring in the airways. However, this may be technically difficult as CD8+ counts are low in bronchial biopsies.

The combination of WGNCA, an unbiased pathway discovery method, and gene ontology, based on existing publications, was a useful way to find key genes driving CD8+ dysfunction in a semi-biased fashion. The light-green module positively correlated to asthma severity and the gene ontology identified EHMT1, a H3K9 methylase, and another histone-lysine N- methyltransferase, WHSC1 as being important in severe asthmatic CD8+ T- cells. Although the EHMT1/2 inhibitor (UNC0642) had no effect on the differentially expressed release of MIP1α from severe asthmatic CD8+ T-cell it is important to examine the effect of this inhibitor on whole cell transcriptomics or on the release of a wider range of analytes using SomaLogics for example. This emphasises the fact that although gene ontology is critical for drawing

202 203 out differentially regulated pathways it cannot usually select a single gene target. Whether there are specific local changes in histone methylation status in CD8+ T-cell could be investigated by ChIP-Seq.

WGCNA analysis also identified 2 modules, Red and Blue that were associated with FEV1 % predicted. These modules highlighted several genes including IL17A, ZNF24, BID, PDE9A, BOX12, KDM4D, EDN, KCNIP2, FGFR4 and IL10. Their functional interaction in CD8+ cells requires further elucidation.

Many genes are determined as differentially expressed in disease and health but they are often ignored, as their functions are unclear or unknown. These genes are often ignored during analysis in favour of genes which have greater currently available literature. These unknown genes are termed the “ignorome” (Pandey et al. 2014) and it is increasingly recognised as a critical problem. In neuroscience, the top 5% of genes are discussed in 70% of the relevant literature whereas 20% of the genes discovered have no associated literature whatsoever. Often, the only relationship between the number of publications relating to a specific gene is its date of discovery, where older genes are more highly published (Pandey et al. 2014). A similar study has been carried out in mouse asthma models using meta-analysis of data from well phenotyped models with gene arrays carried out on similar platforms (Riba et al. 2016). The mouse asthma ignorome contained many protein coding genes which are highly differentially expressed in mouse asthma models, but have no published function or relationship with the models. Though the large increases of data which further existing work, the demonstration of the ignoromes of asthma will provide novel areas of investigation in future work.

The WGCNA analysis of the Tsitsiou data (Tsitsiou et al. 2012) produced many fewer modules despite using the same algorithm. In addition, they showed a weak correlation with FEV1 % predicted unlike the Red and Blue

203 204 modules obtained in the current dataset. There were however strong links between the modules suggesting intracellular trafficking is a key process in defineing the asthmatics CD8+ T-cells phenotype The low sample sizes and heterogeneous nature of asthma will always be a problem when compared to the scale of the U-BIOPRED data. The size of the U-BIOPRED dataset enables subsetting which is essential for the identification of biomarkers that are essential for the initial steps to enable the concept of personalised medicine and the identification of highly selected treatments for identified patients to be realised. In addition, the detailed demographic details available in the U-BIOPRED dataset enables better discrimination of the WGCNA modules between the studies.

7.1.3 CD8+ T-cell Histone Acetylation and Methylation The study to compare histone acetylation and methylation and to link this to gene expression profiles was inconclusive due to technical issues with sample preparation. Performing H3K27ac and H3K27me3 ChIP-Seq on the same samples as the gene expression data would have been the most important element of the current thesis. Linking gene expression with histone modification marks across the genome would have clarified the roles and importance of these modifications. These important experiments should be repeated.

7.1.4 Role of Vesicle creation and budding in Severe Asthma CD8+ T-cells Intracellular localisation is vital for the T-cell activation including APC-T-cell signalling, the directional release of cytotoxic proteins by relocation of golgi apparatus and for directional cytokine release (Kuhné et al. 2003; Kuhn & Poenie 2002; Saito & Yokosuka 2006; Kupfer & Dennert 1984; Kupfer et al. 1987; Kupfer et al. 1991; Reichert et al. 2001). The expression of tubulin, which is vital for cell linking and intracellular movement of vesicles was

204 205 increased in CD8+ T-cells in severe asthma (Kupfer & Dennert 1984; Kupfer et al. 1987). This increased expression suggests a potential over activation of these processes. However, the expression of other genes such as those for CNN2 involved in cell-cell binding, GGA1 involved in linking the Golgi apparatus to vesicles by intermediate filaments and CD81, important for T-cell activation and myotubule maintenance had a non-significant decrease in fold change (Charrin et al. 2013; Hines et al. 2014; Hirst et al. 2000). This dissonance suggests that this pathway is highly modulated in severe asthma and potentially plays a key role in disease – selective modulation of the expression of these proteins along with the analysis of protein expression is critical to define the functional effect of these gene expression changes. Alongside this classical expression analysis, the WGCNA yellow-green module, which differentiated disease and oral steroid use, showed that TUBA1A, the tubulin gene, along with DST, DCTN1, VSX1, GNBL2, STK16 and ARMCX2 which are also linked with this pathway, were differentially regulated as a group.

The interrelationship of these genes is shown in Error! Reference source not found. DST, which codes for dystonin, anchors intermediate filaments to hemidesmosomes, the linkers which bind to the extracellular matrix, collagen and DNCT1 (Koster et al. 2003; Liu et al. 2003). DCTN1 codes for a subunit of dynein, a protein whose role is to move vesicles along microtubules which are made of tubulin (Kiyomitsu & Cheeseman 2012). GNBL2 encodes RACK1 which regulates integrin-mediated adhesion, protrusion and chemotactic cell migration, which are all key processes in cell-cell interaction and release of cytotoxic proteins (Cox et al. 2003). STK16 codes for a serine threonine kinase which regulates, alongside MAL2, the creation of robust lysosomes and sorting of soluble cargo in the trans-Golgi network. However, VSX1 and ARMCX2, do not have a clear role in these pathways. VSX1’s has been reported to regulate the development of the eye and ARMCX2 is found in extracellular vesicles but its precise function has not been elucidated. Further experiments to determine the role of these 2

205 206 proteins in CD8+ T-cell function in severe asthma are necessary and doing this could be a step in the right direction in terms of overcoming the ignorome (Hong et al. 2009; Fraser et al. 2013; Bruno et al. 2009).

Figure 7.1. Intracellular and extracellular traffic of CD8+ T-cells. This diagram has a focus on genes found in gene array to be significantly changed in Severe Asthma (Increased in orange, decreased in blue) and from the Green-Yellow Module (Green).

206 207

The changes in the expression of genes which have roles in the activity of the cytoskeleton and intracellular trafficking can be seen in two independent data sets when analysed with WGCNA with respect to FEV1 % predicted and therefore it would be immensely beneficial to study these proteins in healthy and severely asthmatic CD8+ T-cells, firstly using proteomics to gauge whether the production of these proteins correlates to baseline mRNA measurements and whether they are altered with cell activation. In addition, elucidating the sub-cellular localisation of ARMCX2 using confocal microscopy might aid the understanding of its function. This approach will also enable the determination of tubulin expression changes and in functionality in relation to microtubule formation in CD8+ T-cells from severe asthmatics.

The importance of tubulin in CD8+ T-cell function was suggested by p300/CBP inhibitor, SGC-CBP30 inhibited the release of MIP1α, but may not have been acting specifically on transcription. This may indicate a role for tubulin acetylation. SGC-CBP30 had no effect on MIP1α mRNA but decreased its release from cells. The relationship between the p300/CBP and tubulin could be studied using a co-immunoprecipitation between CBP and tubulin, measuring the acetylation status of tubulin and measuring tubulin- alphaTAT1 acetylase activity which is shown to be upregulated when CBP is inhibited and is able to change tubulin binding to other proteins (Mackeh et al. 2014; Howes et al. 2014).

In the first results chapter, SGC-CBP30 caused inhibition of proliferation but not viability of HBECs from healthy subjects. Although not investigated, this effect may also reflect an action of SGC-CBP30 on tubulin acetylation as this is essential for nuclear division (Morrissette & Sibley 2002).

It will also be important to study the mechanisms by which SGC-CBP30 attenuates the release of other analytes/inflammatory mediators to determine whether the specificity of SGC-CBP30 is broad or very selective. This can be done by running SILAC on both cell media and on the cells themselves.

207 208

Alternatively SomaLogic analysis of >1300 mediators may be performed. Correlation with the effect of SGC-CBP30 on gene expression profiling using next generation sequencing or gene arrays will aid determination of whether tubulin acetylation is the most important target of SGC-CBP30. Studies of whether MIP1α-containing vesicles are produced and remain intracellular in response to CBP/p300 inhibition will also be important.

This thesis stated that epigenetic mechanisms play a key role in regulating epithelial and CD8+ T-cell function in the severe asthma phenotype, though this may still be the case it will require further investigation. It has, however, taken what was thought to be an epigenetic modifying compound (SGC- CBP30) and suggested that it has another role in severe asthmatic CD8+ T- cell activation pathway involved in the release of MIP1α.

Pathways associated with intracellular transport and of antimicrobial responses are substantially changed in the CD8+ T-cells of severe asthmatics and should be extended to analysis of the proteome. Alongside this, the over expression of genes which encode for antimicrobial proteins could have deleterious effect if translated and released in the lungs potentially causing lung remodelling and untreatable hypersensitivity that is present in severe asthma.

7.2 Conclusion and future studies

To answer my hypothesis I would suggest that there are some some targets for which epigenetic modifying compounds such as SGC-CBP30 and JQ1+ which are able to modulate bronchial epithelial and CD8+ T-cell release of cytokines and proliferation which may be relevant to the underlying disease of severe asthma.

208 209

In this thesis there has been a lower power than hoped and the efficacy of the experiments haven’t all provided useful results in the positive or negative and as such I can’t answer the question of whether histone modifcations play a key role in regulating epithelial and CD8+ T-cell dysfunction in the severe asthma phenotype. However, the methodology of WGCNA has demonstrated the potential to further investigate the disease, with the possibility of being a method that will help targeted exploration of the ignorome and may have presented novel genes from antibacterial products to vesicle budding and release that with further investigation could be shown to be important in CD8+ T-cell function in severe asthma.

In the future some of these attemped experiments in Bronchial Epithelium and CD8+ T-cells with ChIP-Seq and gene array should be repeated, with a clearer plan on how patient recruitment could be improved, and with the knowledge acquired during this PhD to carry out these experiments in a manner that will yield more substantial results.

209 210

Chapter 8: References

210 211

Abraham, S.M. & Clark, A.R., 2006. Dual-specificity phosphatase 1: a critical regulator of innate immune responses. Biochemical Society Transactions, 34(Pt 6), pp.1018–1023. Accordini, S. et al., 2013. The Cost of Persistent Asthma in Europe: An International Population-Based Study in Adults. International Archives of Allergy and Immunology, 160(1), pp.93–101. Adelroth, E. et al., 1986. Airway responsiveness to leukotrienes C4 and D4 and to methacholine in patients with asthma and normal controls. The New England journal of medicine, 315(8), pp.480–4. Afkarian, M. et al., 2002. T-bet is a STAT1-induced regulator of IL-12R expression in naïve CD4+ T cells. Nature immunology, 3(6), pp.549–57. Ahmed, R. & Gray, D., 1996. Immunological memory and protective immunity: understanding their relation. Science (New York, N.Y.), 272(5258), pp.54–60. Akbari, O. et al., 2002. Antigen-specific regulatory T cells develop via the ICOS–ICOS-ligand pathway and inhibit allergen-induced airway hyperreactivity. Nature Medicine, 8(9), pp.1024–1032. Al-Ramli, W. et al., 2009. TH17-associated cytokines (IL-17A and IL-17F) in severe asthma. Journal of Allergy and Clinical Immunology, 123(5), pp.1185–1187. Al-Shami, A. et al., 2005. A role for TSLP in the development of inflammation in an asthma model. The Journal of experimental medicine, 202(6), pp.829–839. Albig, W. et al., 1997. Human histone gene organization: nonregular arrangement within a large cluster. Genomics, 40(2), pp.314–322. Allakhverdi, Z. et al., 2007. Cutting edge: The ST2 ligand IL-33 potently activates and drives maturation of human mast cells. Journal of immunology (Baltimore, Md. : 1950), 179(4), pp.2051–2054. Aluvihare, V.R., Kallikourdis, M. & Betz, A.G., 2004. Regulatory T cells mediate maternal tolerance to the fetus. Nature Immunology, 5(3), pp.266–271. Andrieu, G., Belkina, A.C. & Denis, G. V., 2016. Clinical trials for BET

211 212

inhibitors run ahead of the science. Drug Discovery Today: Technologies, xxx(xx), pp.1–6. Annoni, R. et al., 2015. Increased expression of granzymes A and B in fatal asthma. European Respiratory Journal, 45(5), pp.1485–1488. Arase, N. et al., 2003. IgE-mediated activation of NK cells through Fc gamma RIII. Journal of immunology (Baltimore, Md. : 1950), 170(6), pp.3054– 3058. Ashburner, B.P., Westerheide, S.D. & Baldwin, A.S., 2001. The p65 (RelA) Subunit of NF- B Interacts with the Histone Deacetylase (HDAC) Corepressors HDAC1 and HDAC2 To Negatively Regulate Gene Expression. Molecular and Cellular Biology, 21(20), pp.7065–7077. Ather, J.L. et al., 2011. Airway epithelial NF-kB activation promotes allergic sensitization to an innocuous inhaled antigen. American Journal of Respiratory Cell and Molecular Biology, 44(5), pp.631–638. Au-Yeung, N., Mandhana, R. & Horvath, C.M., 2013. Transcriptional regulation by STAT1 and STAT2 in the interferon JAK-STAT pathway. Jak-Stat, 2(3), p.e23931. Baecher-Allan, C., Viglietta, V. & Hafler, D. a, 2002. Inhibition of human CD4(+)CD25(+high) regulatory T cell function. Journal of immunology (Baltimore, Md. : 1950), 169(11), pp.6210–6217. Baines, K.J. et al., 2011. Transcriptional phenotypes of asthma defined by gene expression profiling of induced sputum samples. Journal of Allergy and Clinical Immunology, 127(1), p.153–160.e9. Bals, R. & Hiemstra, P.S., 2004. Innate immunity in the lung: How epithelial cells fight against respiratory pathogens. European Respiratory Journal, 23(2), pp.327–333. Banerjee, A. et al., 2012. Trichostatin A abrogates airway constriction, but not inflammation, in murine and human asthma models. American Journal of Respiratory Cell and Molecular Biology, 46(2), pp.132–138. Baraldo, S. et al., 2012. Deficient antiviral immune responses in childhood: Distinct roles of atopy and asthma. Journal of Allergy and Clinical Immunology, 130(6), pp.1307–1314.

212 213

Barneda-Zahonero, B. et al., 2013. HDAC7 Is a Repressor of Myeloid Genes Whose Downregulation Is Required for Transdifferentiation of Pre-B Cells into Macrophages. PLoS Genetics, 9(5). Barnes, P.J., 2013. Theophylline. American Journal of Respiratory and Critical Care Medicine, 188(5), pp.901–906. Barnes, P.J., Adcock, I.M. & Ito, K., 2005. Histone acetylation and deacetylation: importance in inflammatory lung diseases. The European respiratory journal, 25(3), pp.552–63. Barnes, P.J., Ito, K. & Adcock, I.M., 2004. Corticosteroid resistance in chronic obstructive pulmonary disease: Inactivation of histone deacetylase. Lancet, 363(9410), pp.731–733. Baron, U. et al., 2007. DNA demethylation in the human FOXP3 locus discriminates regulatory T cells from activated FOXP3+ conventional T cells. European Journal of Immunology, 37(9), pp.2378–2389. Barsyte-Lovejoy, D. et al., 2014. (R)-PFI-2 is a potent and selective inhibitor of SETD7 methyltransferase activity in cells. Proceedings of the National Academy of Sciences of the United States of America, 111(35), pp.12853–8. Bedoui, S. et al., 2016. Parallels and differences between innate and adaptive lymphocytes. Nature immunology, 17(5), pp.490–4. Bel, E.H. et al., 2014. Oral glucocorticoid-sparing effect of mepolizumab in eosinophilic asthma. Nejm, 371(13), pp.1189–97. Belkina, A.C., Nikolajczyk, B.S. & Denis, G. V, 2013. BET protein function is required for inflammation: Brd2 genetic disruption and BET inhibitor JQ1 impair mouse macrophage inflammatory responses. Journal of immunology (Baltimore, Md. : 1950), 190(7), pp.3670–8. Bender, B.G. & Bender, S.E., 2005. Patient-identified barriers to asthma treatment adherence: responses to interviews, focus groups, and questionnaires. Immunology and allergy clinics of North America, 25(1), pp.107–30. Bentley, J.K. et al., 2014. Periostin is required for maximal airways inflammation and hyperresponsiveness in mice. The Journal of allergy

213 214

and clinical immunology, 134(6), pp.1433–1442. Beretta, C. et al., 1965. Antagonism of 5-hydroxytryptamine-induced bronchospasm in guinea-pigs by 8β-carbobenzyloxy-aminomethyl-1- methyl-10α-ergoline. Journal of Pharmacy and Pharmacology, 17(7), pp.423–428. Bernink, J.H. et al., 2013. Human type 1 innate lymphoid cells accumulate in inflamed mucosal tissues. Nature immunology, 14(3), pp.221–9. Bettelli, E. et al., 2006. Reciprocal developmental pathways for the generation of pathogenic effector TH17 and regulatory T cells. Nature, 441(7090), pp.235–238. Bhadury, J. et al., 2014. BET and HDAC inhibitors induce similar genes and biological effects and synergize to kill in Myc-induced murine lymphoma. Proceedings of the National Academy of Sciences of the United States of America, 111(26), pp.E2721-30. Billen, A. & Dupont, L., 2008. Exercise induced bronchoconstriction and sports. Postgrad Med J, 84, pp.512–517. Blankenberg, D., Kuster, G. Von, et al., 2010. Galaxy: A web-based genome analysis tool for experimentalists. Current Protocols in Molecular Biology, (SUPPL. 89), pp.1–21. Blankenberg, D., Gordon, A., et al., 2010. Manipulation of FASTQ data with galaxy. Bioinformatics, 26(14), pp.1783–1785. Bleecker, E.R. et al., 2016a. Efficacy and safety of benralizumab for patients with severe asthma uncontrolled with high-dosage inhaled corticosteroids and long-acting β2-agonists (SIROCCO): a randomised, multicentre, placebo-controlled phase 3 trial. The Lancet, 388(10056), pp.2115–2127. Bleecker, E.R. et al., 2016b. Efficacy and safety of benralizumab for patients with severe asthma uncontrolled with high-dosage inhaled corticosteroids and long-acting β2-agonists (SIROCCO): a randomised, multicentre, placebo-controlled phase 3 trial. The Lancet, 388(10056), pp.2115–2127. de Boer, W.I. et al., 2008. Altered expression of epithelial junctional proteins in atopic asthma: possible role in inflammation. Canadian Journal of Physiology and Pharmacology, 86(3), pp.105–112.

214 215

Boismenu, R. et al., 1996. A role for cd81 in early t-cell development. Science, 271, pp.198–200. Booij-Noord, H., Orie, N.G.M. & de Vries, K., 1971. Immediate and late bronchial obstructive reactions to inhalation of house dust and protective effects of disodium cromoglycate and prednisolone. The Journal of Allergy and Clinical Immunology, 48(6), pp.344–354. Borish, L. & Steinke, J.W., 2011. Interleukin-33 in Asthma: How Big of a Role Does It Play? Current allergy and asthma reports, 11(1), pp.7–11. Van Den Bosch, T. et al., 2016. The histone acetyltransferase p300 inhibitor C646 reduces pro-inflammatory gene expression and inhibits histone deacetylases. Biochemical Pharmacology, 102, pp.130–140. Bowers, E.M. et al., 2010. Virtual Ligand Screening of the p300/CBP Histone Acetyltransferase: Identification of a Selective Small Molecule Inhibitor. Chemistry & Biology, 17(5), pp.471–482. Brand, S. et al., 2012. DNA methylation of TH1/TH2 cytokine genes affects sensitization and progress of experimental asthma. The Journal of allergy and clinical immunology, 129(6), p.1602–10.e6. Brightling, C.E. et al., 2015. Efficacy and safety of tralokinumab in patients with severe uncontrolled asthma: a randomised, double-blind, placebo- controlled, phase 2b trial. The Lancet Respiratory Medicine, 3(9), pp.692–701. British Thoracic Society, 2014. British guideline on the management of asthma., Broide, D.H. et al., 2005. Allergen-induced peribronchial fibrosis and mucus production mediated by IκB kinase β -dependent genes in airway epithelium. Proceedings of the National Academy of Sciences of United States of America, 102(49), pp.17723–17728. Brunkow, M.E. et al., 2001. Disruption of a new forkhead/winged-helix protein, scurfin, results in the fatal lymphoproliferative disorder of the scurfy mouse. Nature Genetics, 27(1), pp.68–73. Bruno, S. et al., 2009. Mesenchymal Stem Cell-Derived Microvesicles Protect Against Acute Tubular Injury. Journal of the American Society of

215 216

Nephrology, 20(5), pp.1053–1067. Burchill, M. a et al., 2007. IL-2 receptor beta-dependent STAT5 activation is required for the development of Foxp3+ regulatory T cells. Journal of immunology (Baltimore, Md. : 1950), 178(1), pp.280–90. Burchill, M. a et al., 2008. Linked T cell receptor and cytokine signaling govern the development of the regulatory T cell repertoire. Immunity, 28(1), pp.112–121. Busse, W. et al., 2001. Omalizumab, anti-IgE recombinant humanized monoclonal antibody, for the treatment of severe allergic asthma. Journal of Allergy and Clinical Immunology, 108(2), pp.184–190. Busse, W.W. et al., 2013. Randomized, double-blind, placebo-controlled study of brodalumab, a human anti-IL-17 receptor monoclonal antibody, in moderate to severe asthma. Am.J.Respir.Crit Care Med., 188(1535–4970 (Electronic)), pp.1294–1302. Butler, C. a et al., 2012. Glucocorticoid receptor β and histone deacetylase 1 and 2 expression in the airways of severe asthma. Thorax, 67(5), pp.392–8. Cannon, M.J., Openshaw, P.J. & Askonas, B. a, 1988. Cytotoxic T cells clear virus but augment lung pathology in mice infected with respiratory syncytial virus. The Journal of experimental medicine, 168(3), pp.1163–8. Cartier, A. et al., 1982. Allergen-induced increase in bronchial responsiveness to histamine: relationship to the late asthmatic response and change in airway caliber. The Journal of Allergy and Clinical Immunology, 70(3), pp.170–177. Castro, M. et al., 2014. Benralizumab, an anti-interleukin 5 receptor α monoclonal antibody, versus placebo for uncontrolled eosinophilic asthma: a phase 2b randomised dose-ranging study. The Lancet Respiratory Medicine, 2(11), pp.879–890. Castro, M. et al., 2015. Reslizumab for inadequately controlled asthma with elevated blood eosinophil counts: results from two multicentre, parallel, double-blind, randomised, placebo-controlled, phase 3 trials. The Lancet Respiratory Medicine, 3(5), pp.355–366.

216 217

Castro, M. et al., 2011. Reslizumab for poorly controlled, eosinophilic asthma: A randomized, placebo-controlled study. American Journal of Respiratory and Critical Care Medicine, 184(10), pp.1125–1132. Chakraborty, K. et al., 2008. Bacterial exotoxins downregulate cathelicidin (hCAP-18/LL-37) and human ??-defensin 1 (HBD-1) expression in the intestinal epithelial cells. Cellular Microbiology, 10(12), pp.2520–2537. Chan, O. et al., 2012. The chemokine CCL5 regulates glucose uptake and AMP kinase signaling in activated T cells to facilitate chemotaxis. The Journal of biological chemistry, 287(35), pp.29406–16. Charrin, S. et al., 2013. Normal muscle regeneration requires tight control of muscle cell fusion by tetraspanins CD9 and CD81. Nature communications, 4, p.1674. Chauhan, B.F. & Ducharme, F.M., 2012. Anti-leukotriene agents compared to inhaled corticosteroids in the management of recurrent and / or chronic asthma in adults and children. Cochrane Database Syst Rev, 5. Chen, L. et al., 2013. mTORC2-PKBα/Akt1 Serine 473 phosphorylation axis is essential for regulation of FOXP3 Stability by chemokine CCL3 in psoriasis. The Journal of investigative dermatology, 133(2), pp.418–28. Cheray, M. et al., 2014. Specific inhibition of DNMT1/CFP1 reduces cancer phenotypes and enhances chemotherapy effectiveness. Epigenomics, 6(3), pp.267–275. Cho, S.H. et al., 2005. Increased interleukin-4, interleukin-5, and interferon- [gamma] in airway CD4+ and CD8+ T cells in atopic asthma. American Journal of Respiratory and Critical Care Medicine, 171(3), pp.224–230. Choi, S.W. et al., 2014. Histone deacetylase inhibition regulates inflammation and enhances Tregs after allogeneic hematopoietic cell transplantation in humans. Blood. Choy, D.F. et al., 2015. TH2 and TH17 inflammatory pathways are reciprocally regulated in asthma. Science Translational Medicine, 7(301), p.301ra129-301ra129. Christianson, C.A. et al., 2015. Persistence of asthma requires multiple feedback circuits involving type 2 innate lymphoid cells and IL-33. Journal

217 218

of Allergy and Clinical Immunology, 136(1), p.59–68.e14. Chun, Y.H. et al., 2013. Rhinovirus-infected epithelial cells produce more IL-8 and rantes compared with other respiratory viruses. Allergy, Asthma and Immunology Research, 5(4), pp.216–223. Chung, K.F. et al., 2001. IL-17-induced cytokine release in human bronchial epithelial cells in vitro : role of mitogen-activated protein ( MAP ) kinases. British journal of pharmacology, 8, pp.200–206. Chung, K.F. et al., 2014. International ERS/ATS guidelines on definition, evaluation and treatment of severe asthma. The European respiratory journal, 43(2), pp.343–73. Cibotti, R. et al., 1997. Surface molecules that drive T cell development in vitro in the absence of thymic epithelium and in the absence of lineage- specific signals. Immunity, 6(3), pp.245–255. Clark, A.R., 2003. MECHANISMS OF STEROID ACTION AND RESISTANCE IN INFLAMMATION MAP kinase phosphatase 1 : a novel mediator of biological effects of glucocorticoids ? Journal of endocrinology, 178(Newton 2000), pp.5–12. Clark, A.R. & Lasa, M., 2003. Crosstalk between glucocorticoids and mitogen- activated protein kinase signalling pathways. Current Opinion in Pharmacology, 3(4), pp.404–411. Clifford, R.L. et al., 2012. Abnormal Histone Methylation is Responsible for Increased VEGF165a Secretion from Airway Smooth Muscle Cells in Asthma. Journal of Immunology, 189(2), pp.819–831. Clifford, R.L. et al., 2015. CXCL8 Histone H3 acetylation is dysfunctional in airway smooth muscle in asthma: regulation by BET. American Journal of Physiology - Lung Cellular and Molecular Physiology, p.ajplung.00021.2015. Cocchi, F. et al., 1995. Identification of RANTES, MIP-1alpha, and MIP-1beta as the Major HIV-Suppressive Factors Produced by CD8+ T Cells. Science, 270(5243), pp.1811–1815. Cockcroft, D.W. et al., 1977. Allergen-induced increase in non-allergic bronchial reactivity. Clinical Allergy, 7(6), pp.503–513.

218 219

Cockcroft, D.W. & Murdock, K.Y., 1987. Comparative effects of inhaled salbutamol, sodium cromoglycate, and beclomethasone dipropionate on allergen-induced early asthmatic responses, late asthmatic responses, and increased bronchial responsiveness to histamine. The Journal of Allergy and Clinical Immunology, 79(5), pp.734–740. Cohen, L. et al., 2007. Epithelial cell proliferation contributes to airway remodeling in severe asthma. Am J Respir Crit Care Med, 176(2), pp.138–145. Cohn, L. et al., 1997. Induction of Airway Mucus Production By T Helper 2 (Th2) Cells: A Critical Role For Interleukin 4 In Cell Recruitment But Not Mucus Production . The Journal of Experimental Medicine, 186(10), pp.1737–1747. Contoli, M. et al., 2010. Fixed airflow obstruction due to asthma or chronic obstructive pulmonary disease: 5-year follow-up. Journal of Allergy and Clinical Immunology, 125(4), pp.830–837. Contoli, M. et al., 2006. Role of deficient type III interferon-λ production in asthma exacerbations TL - 12. Nature Medicine, 12 VN-r(9), pp.1023– 1026. Corren, J. et al., 2011. Lebrikizumab Treatment in Adults with Asthma. New England Journal of Medicine, 365(12), pp.1088–1098. Cote-Sierra, J. et al., 2004. Interleukin 2 plays a central role in Th2 differentiation. Proceedings of the National Academy of Sciences of the United States of America, 101(11), pp.3880–5. Cox, E.A. et al., 2003. RACK1 regulates integrin-mediated adhesion, protrusion, and chemotactic cell migration via its Src-binding site. Molecular biology of the cell, 14(2), pp.658–69. Coyle, A.J. et al., 1999. Crucial Role of the Interleukin 1 Receptor Family Member T1/St2 in T Helper Cell Type 2–Mediated Lung Mucosal Immune Responses. The Journal of Experimental Medicine, 190(7), pp.895–902. Creyghton, M.P. et al., 2010. Histone H3K27ac separates active from poised enhancers and predicts developmental state. Proceedings of the National Academy of Sciences of the United States of America, 107(50),

219 220

pp.21931–6. Cruz-Guilloty, F. et al., 2009. Runx3 and T-box proteins cooperate to establish the transcriptional program of effector CTLs. The Journal of experimental medicine, 206(1), pp.51–9. Curotto de Lafaille, M.A. et al., 2001. Hyper immunoglobulin E response in mice with monoclonal populations of B and T lymphocytes. The Journal of experimental medicine, 194(9), pp.1349–59. D’Agostino, C. et al., 2009. In vivo release of alpha-defensins in plasma, neutrophils and CD8 T-lymphocytes of patients with HIV infection. Curr HIV Res, 7(6), pp.650–655. Dakhama, A. et al., 2013. IL-13-producing BLT1-positive CD8 cells are increased in asthma and are associated with airway obstruction. Allergy: European Journal of Allergy and Clinical Immunology, 68(5), pp.666–673. Daldegan, M.B., Teixeira, M.M. & Talvani, A., 2005. Concentration of CCL11, CXCL8 and TNF-alpha in sputum and plasma of patients undergoing asthma or chronic obstructive pulmonary disease exacerbation. Brazilian Journal of Medical and Biological Research, 38(9), pp.1359–65. Davey, C.A. et al., 2002. Solvent mediated interactions in the structure of the nucleosome core particle at 1.9 A resolution. Journal of Molecular Biology, 319(5), pp.1097–1113. Davidson, T.S. et al., 2007. Cutting Edge: IL-2 is essential for TGF-beta- mediated induction of Foxp3+ T regulatory cells. Journal of immunology (Baltimore, Md : 1950), 178(7), pp.4022–4026. De, S. et al., 2002. Histone H1 variants differentially inhibit DNA replication through an affinity for chromatin mediated by their carboxyl-terminal domains. Gene, 292(1–2), pp.173–181. Demonchy, J.G.R. et al., 1985. Bronchoalveolar Eosinophilia During Allergen- Induced Late Asthmatic Reactions. American Review of Respiratory Disease, 131(3), pp.373–376. Denu, J.M. & Dixon, J.E., 1998. Protein tyrosine phosphatases: mechanisms of catalysis and regulation. Curr Opin Chem Biol, 2(5), pp.633–641. Doñas, C. et al., 2016. The histone demethylase inhibitor GSK-J4 limits

220 221

inflammation through the induction of a tolerogenic phenotype on DCs. Journal of Autoimmunity, pp.1–13. Duits, L.A. et al., 2003. Rhinovirus increases human [beta]-defensin-2 and -3 mRNA expression in cultured bronchial epithelial cells. FEMS Immunology and Medical Microbiology, 38(1), pp.59–64. Dumuis-Kervabon, a et al., 1986. A chromatin core particle obtained by selective cleavage of histones by clostripain. The EMBO journal, 5(7), pp.1735–1742. Ea, C. & Baltimore, D., 2009. Regulation of NF-kappaB activity through lysine monomethylation of p65. Proceedings of the National Academy of Sciences of the United States of America, 106(45), pp.18972–18977. Elsharkawy, A.M. et al., 2010. The NF-κB p50:p50:HDAC-1 repressor complex orchestrates transcriptional inhibition of multiple pro- inflammatory genes. Journal of Hepatology, 53(3), pp.519–527. Erb, K.J. et al., 1998. Infection of mice with Mycobacterium bovis-Bacillus Calmette-Guérin (BCG) suppresses allergen-induced airway eosinophilia. The Journal of experimental medicine, 187(4), pp.561–9. Erin, E.M. et al., 2006. The effects of a monoclonal antibody directed against tumor necrosis factor-?? in asthma (Retraction in: American Journal of Respiratory and Critical Care Medicine (2011) 183:3 (418)). American Journal of Respiratory and Critical Care Medicine, 174(7), pp.753–762. Fahy, J. V et al., 1997. The Effect of an Anti-lgE Monoclonal Antibody on the Early- and Late-Phase Responses to Allergen Inhalation in Asthmatic Subjects. Am J Respir Crit Care Med, 155, pp.1828–1834. Farnaud, S. & Evans, R.W., 2003. Lactoferrin - A multifunctional protein with antimicrobial properties. Molecular Immunology, 40(7), pp.395–405. Filippakopoulos, P. et al., 2010. Selective inhibition of BET bromodomains. Nature, 468(7327), pp.1067–73. Finotto, S. et al., 2002. Development of spontaneous airway changes consistent with human asthma in mice lacking T-bet. Science (New York, N.Y.), 295(5553), pp.336–338. FitzGerald, J.M. et al., 2016. Benralizumab, an anti-interleukin-5 receptor

221 222

α monoclonal antibody, as add-on treatment for patients with severe, uncontrolled, eosinophilic asthma (CALIMA): a randomised, double-blind, placebo-controlled phase 3 trial. The Lancet, 388(10056), pp.2128–2141. Flavahan, N.A. et al., 1988. Human eosinophil major basic protein causes hyperreactivity of respiratory smooth muscle. Role of the epithelium. Am Rev Respir Dis, 138(3), pp.685–688. Flint, K.C. et al., 1985. Bronchoalveolar mast cells in extrinsic asthma: a mechanism for the initiation of antigen specific bronchoconstriction. British medical journal (Clinical research ed.), 291(6500), pp.923–6. Floess, S. et al., 2007. Epigenetic control of the foxp3 locus in regulatory T cells. PLoS Biology, 5(2), pp.0169–0178. Folkerts, G. & Nijkamp, F.P., 1998. Airway epithelium: more than just a barrier! Trends in pharmacological sciences, 19(8), pp.334–41. Fontenot, J.D. et al., 2005. Regulatory T cell lineage specification by the forkhead transcription factor Foxp3. Immunity, 22(3), pp.329–341. Fontenot, J.D., Gavin, M. a & Rudensky, A.Y., 2003. Foxp3 programs the development and function of CD4+CD25+ regulatory T cells. Nature immunology, 4(4), pp.330–6. Fort, M.M. et al., 2001. IL-25 Induces IL-4, IL-5, and IL-13 and Th2-associated pathologies in vivo. Immunity, 15(6), pp.985–995. Fraser, K.B. et al., 2013. LRRK2 secretion in exosomes is regulated by 14-3- 3. Human Molecular Genetics, 22(24), pp.4988–5000. Frucht, D.M. et al., 2000. Stat4 is expressed in activated peripheral blood monocytes, dendritic cells, and macrophages at sites of Th1-mediated inflammation. Journal of immunology (Baltimore, Md. : 1950), 164(9), pp.4659–64. Frye, S. V, 2010. The art of the chemical probe. Nature Chemical Biology, 6(3), pp.159–161. Fuchs, A. et al., 2013. Intraepithelial type 1 innate lymphoid cells are a unique subset of il-12- and il-15-responsive ifn-??-producing cells. Immunity, 38(4), pp.769–781.

222 223

Gajewski, T.F. & Fitch, F.W., 1988. Anti-proliferative effect of IFN-gamma in immune regulation. I. IFN-gamma inhibits the proliferation of Th2 but not Th1 murine helper T lymphocyte clones. J Immunol, 140(12), pp.4245– 4252. Gao, C. et al., 2006. CRM1 mediates nuclear export of HDAC7 independently of HDAC7 phosphorylation and association with 14-3-3s. FEBS Letters, 580(21), pp.5096–5104. Gauvreau, G.M. et al., 2014. Effects of an anti-TSLP antibody on allergen- induced asthmatic responses. The New England journal of medicine, 370(22), pp.2102–10. Gavin, M.A. et al., 2007. Foxp3-dependent programme of regulatory T-cell differentiation. Nature, 445(7129), pp.771–775. Ge, Q. et al., 2010. TGFβ1 induces IL-6 and inhibits IL-8 release in human bronchial epithelial cells: the role of Smad2/3. Journal of cellular physiology, 225(3), pp.846–54. Gelfand, E.W. & Dakhama, A., 2006. CD8+ T lymphocytes and leukotriene B4: Novel interactions in the persistence and progression of asthma. Journal of Allergy and Clinical Immunology, 117(3), pp.577–582. Gerasimova, A. et al., 2013. Predicting cell types and genetic variations contributing to disease by combining GWAS and epigenetic data. PloS one, 8(1), p.e54359. Ghosh, S. et al., 2016. Regulatory T Cell Modulation by CBP/EP300 Bromodomain Inhibition. Journal of Biological Chemistry, 291(15), p.jbc.M115.708560. Gleich, G.J. et al., 1979. Cytotoxic properties of the eosinophil major basic protein. Journal of immunology (Baltimore, Md. : 1950), 123(6), pp.2925– 7. Global Initiative for Asthma, 2016. Global Strategy For Asthma Management and Prevention, Godard, P. et al., 2002. Costs of asthma are correlated with severity: a 1-yr prospective study. European Respiratory Journal, 19(1), pp.61–67. Goebel-Goody, S.M. et al., 2012. Genetic manipulation of STEP reverses

223 224

behavioral abnormalities in a fragile X syndrome mouse model. Genes, Brain and Behavior, 11(5), pp.586–600. Goodman, R.H. & Smolik, S., 2000. REVIEW CBP / p300 in cell growth , transformation , and development. , (1997), pp.1553–1577. Grakoui, a et al., 1999. The immunological synapse: a molecular machine controlling T cell activation. Science (New York, N.Y.), 285(5425), pp.221–227. Green, R.H. et al., 2002. Analysis of induced sputum in adults with asthma: identification of subgroup with isolated sputum neutrophilia and poor response to inhaled corticosteroids. Thorax, 57(10), pp.875–9. Le Gros, G. et al., 1990. Generation of interleukin 4 (IL-4)-producing cells in vivo and in vitro: IL-2 and IL-4 are required for in vitro generation of IL-4- producing cells. Journal of Experimental Medicine, 172(3), pp.921–929. Grotenboer, N.S. et al., 2013. Decoding asthma: Translating genetic variation in IL33 and IL1RL1 into disease pathophysiology. Journal of Allergy and Clinical Immunology, 131(3), p.856–865.e9. Gundel, R.H., Letts, L.G. & Gleich, G.J., 1991. Human eosinophil major basic protein induces airway constriction and airway hyperresponsiveness in primates. Journal of Clinical Investigation, 87(4), pp.1470–1473. Haldar, P. et al., 2008. Cluster Analysis and Clinical Asthma Phenotypes. American Journal of Respiratory and Critical Care Medicine, 178(3), pp.218–224. Haldar, P. et al., 2009. Mepolizumab and Exacerbations of Refractory Eosinophilic Asthma. New England Journal of Medicine, 360(10), pp.973–984. Halmer, L. & Gruss, C., 1996. Effects of cell cycle dependent histone H1 phosphorylation on chromatin structure and chromatin replication. Nucleic Acids Research, 24(8), pp.1420–1427. Hammad, H. et al., 2009. House dust mite allergen induces asthma via Toll- like receptor 4 triggering of airway structural cells. Nature medicine, 15(4), pp.410–6. Hammitzsch, A. et al., 2015. CBP30, a selective CBP/p300 bromodomain

224 225

inhibitor, suppresses human Th17 responses. Proceedings of the National Academy of Sciences, p.201501956. Han, H. et al., 2014. Interleukin-4-mediated 15-lipoxygenase-1 transactivation requires UTX recruitment and H3K27me3 demethylation at the promoter in A549 cells. PloS one, 9(1), p.e85085. Hanania, N.A. et al., 2016. Efficacy and safety of lebrikizumab in patients with uncontrolled asthma (LAVOLTA I and LAVOLTA II): replicate, phase 3, randomised, double-blind, placebo-controlled trials. The Lancet Respiratory Medicine, 4(10), pp.781–796. Hartney, J.M. et al., 2006. Prostaglandin E2 protects lower airways against bronchoconstriction. American journal of physiology. Lung cellular and molecular physiology, 290(1), pp.L105-13. Hay, D.A. et al., 2014. Discovery and optimization of small-molecule ligands for the CBP/p300 bromodomains. Journal of the American Chemical Society, 136(26), pp.9308–9319. Hejazi, M.E., Modarresi-Ghazani, F. & Entezari-Maleki, T., 2016. A review of Vitamin D effects on common respiratory diseases: Asthma, chronic obstructive pulmonary disease, and tuberculosis. Journal of research in pharmacy practice, 5(1), pp.7–15. Hekking, P.P.W. et al., 2015. The prevalence of severe refractory asthma. Journal of Allergy and Clinical Immunology, 135(4), pp.896–902. Hermann, A., Goyal, R. & Jeltsch, A., 2004. The Dnmt1 DNA-(cytosine-C5)- methyltransferase methylates DNA processively with high preference for hemimethylated target sites. The Journal of biological chemistry, 279(46), pp.48350–9. Herz, U. et al., 1998. BCG infection suppresses allergic sensitization and development of increased airway reactivity in an animal model. The Journal of allergy and clinical immunology, 102(5), pp.867–74. Higashi, Y. et al., 1997. Impairment of T cell development in deltaEF1 mutant mice. The Journal of experimental medicine, 185(8), pp.1467–1479. Hildeman, D. a et al., 2003. T cell apoptosis and reactive oxygen species. The Journal of clinical investigation, 111(5), pp.575–581.

225 226

Hilton, I.B. et al., 2015. Epigenome editing by a CRISPR-Cas9-based acetyltransferase activates genes from promoters and enhancers. Nat Biotechnol, 33(5), pp.510–517. Hines, P.C. et al., 2014. A novel role of h2-calponin in regulating whole blood thrombosis and platelet adhesion during physiologic flow. Physiological reports, 2(12), pp.1–11. Hirst, J. et al., 2000. A family of proteins with -adaptin and VHS domains that facilitate trafficking between the trans-golgi network and the vacuole/lysosome. Journal of Cell Biology, 149(1), pp.67–79. Hogan, R.J. et al., 2001. Activated antigen-specific CD8+ T cells persist in the lungs following recovery from respiratory virus infections. Journal of immunology, 166, pp.1813–1822. Holgate, S.T. et al., 2011. Efficacy and safety of etanercept in moderate-to- severe asthma: A randomised, controlled trial. European Respiratory Journal, 37(6), pp.1352–1359. Hong, B.S. et al., 2009. Colorectal cancer cell-derived microvesicles are enriched in cell cycle-related mRNAs that promote proliferation of endothelial cells. BMC genomics, 10, p.556. Hori, S., Nomura, T. & Sakaguchi, S., 2003. Control of regulatory T cell development by the transcription factor Foxp3. Science (New York, N.Y.), 299(5609), pp.1057–61. Hou, X. et al., 2014. Histone deacetylase inhibitor regulates the balance of Th17/Treg in allergic asthma. The clinical respiratory journal, (1), pp.1–9. Howes, S.C. et al., 2014. Effects of tubulin acetylation and tubulin acetyltransferase binding on microtubule structure. Molecular biology of the cell, 25(2), pp.257–66. Huang, B. et al., 2009a. Brd4 coactivates transcriptional activation of NF- kappaB via specific binding to acetylated RelA. Molecular and cellular biology, 29, pp.1375–1387. Huang, B. et al., 2009b. Brd4 coactivates transcriptional activation of NF- kappaB via specific binding to acetylated RelA. Molecular and cellular biology, 29(5), pp.1375–87.

226 227

Huang, J. et al., 2002. CD28 plays a critical role in the segregation of PKC theta within the immunologic synapse. Proceedings of the National Academy of Sciences of the United States of America, 99(14), pp.9369– 9373. Huisinga, K.L., Brower-Toland, B. & Elgin, S.C.R., 2006. The contradictory definitions of heterochromatin: transcription and silencing. Chromosoma, 115(2), pp.110–22. Hukelmann, J.L. et al., 2016. The cytotoxic T cell proteome and its shaping by the kinase mTOR. Nat Immunol, 17(1), pp.104–112. Iikura, M. et al., 2007. IL-33 can promote survival, adhesion and cytokine production in human mast cells. Laboratory investigation; a journal of technical methods and pathology, 87(10), pp.971–978. Imai, T. et al., 1995. Molecular Analyses of the Association of CD4 with Two Members of the Transmembrane 4 Superfamily, CD81 and CD82. The Amerlcan Assoclation of Immunologists, 155(3), pp.1229–1239. Imhof, A. et al., 2015. Global and Specific Responses of the Histone Acetylome to Systematic Perturbation Resource Global and Specific Responses of the Histone Acetylome to Systematic Perturbation. , pp.559–571. In, J.G. et al., 2014. Serine/threonine kinase 16 and MAL2 regulate constitutive secretion of soluble cargo in hepatic cells. Biochemical Journal, 463(2), pp.201–213. Ishii, M. et al., 2009. Epigenetic regulation of the alternatively activated macrophage phenotype. Blood, 114(15), pp.3244–54. Ito, K. et al., 2002. Expression and activity of histone deacetylases in human asthmatic airways. American Journal of Respiratory and Critical Care Medicine, 166, pp.392–396. Ito, K. et al., 2006. Histone deacetylase 2-mediated deacetylation of the glucocorticoid receptor enables NF-kappaB suppression. The Journal of experimental medicine, 203(1), pp.7–13. Ito, T. et al., 2005. TSLP-activated dendritic cells induce an inflammatory T helper type 2 cell response through OX40 ligand. The Journal of

227 228

experimental medicine, 202(9), pp.1213–23. Ivanov, I.I. et al., 2006. The orphan nuclear receptor RORgammat directs the differentiation program of proinflammatory IL-17+ T helper cells. Cell, 126(6), pp.1121–1133. Jackson, D.J. et al., 2008. Wheezing rhinovirus illnesses in early life predict asthma development in high-risk children. American Journal of Respiratory and Critical Care Medicine, 178(7), pp.667–672. Janeway, C., 2001. Immunobiology: the immune system in health and disease 5th ed., New York: Garland. Jangani, M. et al., 2014. The methyltransferase WBSCR22/Merm1 enhances glucocorticoid receptor function and is regulated in lung inflammation and cancer. Journal of Biological Chemistry, 289(13), pp.8931–8946. Jefferson, J.J., Leung, C.L. & Liem, R.K.H., 2006. Dissecting the sequence specific functions of alternative N-terminal isoforms of mouse bullous pemphigoid antigen 1. Experimental Cell Research, 312(15), pp.2712– 2725. Jenne, D.E. & Tschopp, J., 1988. Granzymes, a family of serine proteases released from granules of cytolytic T lymphocytes upon T cell receptor stimulation. Immunological reviews, 103(103), pp.53–71. Jin, Q. et al., 2011. Distinct roles of GCN5/PCAF-mediated H3K9ac and CBP/p300-mediated H3K18/27ac in nuclear receptor transactivation. The EMBO journal, 30(2), pp.249–62. Kaech, S.M. & Cui, W., 2012. Transcriptional control of effector and memory CD8+ T cell differentiation. Nature Reviews Immunology, 12(11), pp.749– 761. Kagami, S. et al., 2001. Stat5a regulates T helper cell differentiation by several distinct mechanisms. Blood, 97(8), pp.2358–2365. Kagami, S. i et al., 2000. Both stat5a and stat5b are required for antigen- induced eosinophil and T-cell recruitment into the tissue. Blood, 95(4), pp.1370–7. Kaplan, M.H. et al., 1996. Impaired IL-12 responses and enhanced development of Th2 cells in Stat4-deficient mice. Nature, 382(6587),

228 229

pp.174–177. Kaplan, M.H. et al., 1996. Stat6 is required for mediating responses to IL-4 and for the development of Th2 cells. Immunity, 4(3), pp.313–319. Kashyap, M. et al., 2011. Thymic stromal lymphopoietin is produced by dendritic cells. Journal of immunology (Baltimore, Md. : 1950), 187(3), pp.1207–1211. Khan, Y.M. et al., 2014. Brd4 is essential for IL-1B-induced inflammation in human airway epithelial cells. PLoS ONE, 9(4), pp.1–17. Khattri, R. et al., 2003. An essential role for Scurfin in CD4+CD25+ T regulatory cells. Nature Immunology, 4(4), pp.337–342. Kienzler, A.-K. et al., 2013. Inhibition of human B-cell development into plasmablasts by histone deacetylase inhibitor valproic acid. Journal of Allergy and Clinical Immunology, 131(6), pp.1695–1699. Kim, Y. et al., 2012. Histone deacetylase 3 mediates allergic skin inflammation by regulating expression of MCP1 protein. Journal of Biological Chemistry, 287(31), pp.25844–25859. King, C. et al., 1998. Dust mite proteolytic allergens induce cytokine release from cultured airway epithelium. Journal of immunology (Baltimore, Md. : 1950), 161(7), pp.3645–3651. Kirby, J.G. et al., 1986. Asthmatic responses to inhalation of anti-human IgE. Clinical Allergy, 16, pp.191–194. Kiyomitsu, T. & Cheeseman, I.M., 2012. Chromosome and spindle-pole- derived signals generate an intrinsic code for spindle position and orientation. Nature Cell Biology, 14(3), pp.311–317. Klaas, M. et al., 2016. The alterations in the extracellular matrix composition guide the repair of damaged liver tissue. Scientific Reports, 6, p.27398. Klebanoff, S.J., Waltersdorph, A.M. & Rosen, H., 1984. Antimicrobial activity of myeloperoxidase. Methods Enzymol, 105, pp.399–403. Klose, C.S.N. et al., 2014. Differentiation of type 1 ILCs from a common progenitor to all helper-like innate lymphoid cell lineages. Cell, 157(2), pp.340–356. Koche, R.P. et al., 2011. Reprogramming factor expression induces rapid and

229 230

widespread targeted chromatin remodeling. Cell Stem Cell, 8(1), pp.96– 105. Komai-Koma, M. et al., 2007. IL-33 is a chemoattractant for human Th2 cells. European journal of immunology, 37(10), pp.2779–2786. Korn, T. et al., 2007. IL-21 initiates an alternative pathway to induce proinflammatory T(H)17 cells. Nature, 448(7152), pp.484–487. Kościuczuk, E.M. et al., 2012. Cathelicidins: family of antimicrobial peptides. A review. Molecular Biology Reports, pp.1–14. Koster, J. et al., 2003. Analysis of the interactions between BP180, BP230, plectin and the integrin alpha6beta4 important for hemidesmosome assembly. Journal of Cell Science, 116(Pt 2), pp.387–399. Koya, T. et al., 2007. CD8(+) T cell-mediated airway hyperresponsiveness and inflammation is dependent on CD4(+)IL-4(+) T cells. The Journal of Immunology, 179(5), pp.2787–2796. Kraft, M. et al., 2002. Mycoplasma pneumoniae and Chlamydia pneumoniae in asthma: Effect of clarithromycin. Chest, 121(6), pp.1782–1788. Kruidenier, L. et al., 2012. A selective jumonji H3K27 demethylase inhibitor modulates the proinflammatory macrophage response. Nature, 488(7411), pp.404–8. Kuhn, J.R. & Poenie, M., 2002. Dynamic polarization of the microtubule cytoskeleton during CTL-mediated killing. Immunity, 16(1), pp.111–121. Kuhné, M.R. et al., 2003. Linker for activation of T cells, zeta-associated protein-70, and Src homology 2 domain-containing leukocyte protein-76 are required for TCR-induced microtubule-organizing center polarization. Journal of immunology (Baltimore, Md. : 1950), 171(2), pp.860–6. Kuo, C.-H.S. et al., 2017. Th2 and non-Th2 molecular phenotypes of asthma using sputum transcriptomics in UBIOPRED, In Press. European Respiratory Journal. Kuo, C.S. et al., 2016. A Transcriptome-driven Analysis of Epithelial Brushings and Bronchial Biopsies to Define Asthma Phenotypes in U- BIOPRED. American Journal of Respiratory and Critical Care Medicine, p.rccm.201512-2452OC.

230 231

Kuperman, D. a et al., 2005. Dissecting asthma using focused transgenic modeling and functional genomics. The Journal of allergy and clinical immunology, 116(2), pp.305–11. Kupfer, A. & Dennert, G., 1984. Reorientation of the microtubule-organizing center and the Golgi apparatus in cloned cytotoxic lymphocytes triggered by binding to lysable target cells. Journal of immunology (Baltimore, Md : 1950), 133(5), pp.2762–2766. Kupfer, A., Swain, S.L. & Singer, S.J., 1987. The specific direct interaction of helper T cells and antigen-presenting B cells. II. Reorientation of the microtubule organizing center and reorganization of the membrane- associated cytoskeleton inside the bound helper T cells. The Journal of experimental medicine, 165(6), pp.1565–80. Kupfer, a, Mosmann, T.R. & Kupfer, H., 1991. Polarized expression of cytokines in cell conjugates of helper T cells and splenic B cells. Proceedings of the National Academy of Sciences of the United States of America, 88(February), pp.775–779. Kurata, H. et al., 1999. Ectopic expression of activated Stat6 induces the expression of Th2-specific cytokines and transcription factors in developing Th1 cells. Immunity, 11(6), pp.677–688. Kwon, N.-H. et al., 2008. DNA methylation and the expression of IL-4 and IFN-gamma promoter genes in patients with bronchial asthma. Journal of clinical immunology, 28(2), pp.139–46. Labont, I. et al., 2008. Quality of bronchial biopsies for morphology study and cell sampling: A comparison of asthmatic and healthy subjects. Canadian Respiratory Journal, 15(8), pp.431–435. Lambrecht, B.N. & Hammad, H., 2012. The airway epithelium in asthma. Nature medicine, 18(5), pp.684–92. LaMere, S., 2015. Genome-wide profiling of H3K4me3 and H3K27me3 dynamics reveals fundamental epigenetic influences upon CD4 T cell activation (IRC7P.423). The Journal of Immunology, 194 (1 Sup(128.4). Lane, S.J. et al., 1998. Corticosteroid-resistant bronchial asthma is associated with increased c-fos expression in monocytes and T lymphocytes.

231 232

Journal of Clinical Investigation, 102(12), pp.2156–2164. Langfelder, P. & Horvath, S., 2008. WGCNA: an R package for weighted correlation network analysis. BMC bioinformatics, 9, p.559. Langmead, B. et al., 2009. Ultrafast and memory-efficient alignment of short DNA sequences to the human genome. Genome Biology, 10(3), p.R25. Langmead, B. & Salzberg, S.L., 2012. Fast gapped-read alignment with Bowtie 2. Nature Methods, 9(4), pp.357–359. Laviolette, M. et al., 2013. Effects of benralizumab on airway eosinophils in asthmatic patients with sputum eosinophilia. Journal of Allergy and Clinical Immunology, 132(5), p.1086–1096.e5. Law, J. a & Jacobsen, S.E., 2010. Establishing, maintaining and modifying DNA methylation patterns in plants and animals. Nature reviews. Genetics, 11(3), pp.204–20. Leckie, M.J. et al., 2000. Effects of an interleukin-5 blocking monoclonal antibody on eosinophils, airway hyper-responsiveness, and the late asthmatic response. Lancet, 356(SUPPL.), pp.2144–2148. Lee, D.J. et al., 1997. Inhibition of IgE antibody formation by plasmid DNA immunization is mediated by both CD4+ and CD8+ T cells. International archives of allergy and immunology, 113(1–3), pp.227–230. Lee, I.H. & Finkel, T., 2009. Regulation of autophagy by the p300 acetyltransferase. Journal of Biological Chemistry, 284(10), pp.6322– 6328. Leff, A., 1982. Pathogenesis of asthma. Neurophysiology and pharmacology of bronchospasm. Chest, 81(2), pp.224–9. Lehrer, R.I. & Ganz, T., 1992. Defensins: endogenous antibiotic peptides from human leukocytes. Ciba Foundation symposium, 171, pp.276-290-293. Levy, D.E. & Darnell Jr., J.E., 2002. Stats: transcriptional control and biological impact. Nat Rev Mol Cell Biol, 3(9), pp.651–662. Levy, D.M.L. et al., 2014. Why asthma still kills The National Review of Asthma Deaths (NRAD), Levy, S., 2014. Function of the tetraspanin molecule CD81 in B and T cells. Immunologic Research, 58(2–3), pp.179–185.

232 233

Li, S. et al., 2011. Rack1 is required for Vangl2 membrane localization and planar cell polarity signaling while attenuating canonical Wnt activity. Proceedings of the National Academy of Sciences of the United States of America, 108(6), pp.2264–2269. Li, X. et al., 2013. Genome-wide association study identifies TH1 pathway genes associated with lung function in asthmatic patients. The Journal of allergy and clinical immunology, 132(2), p.313–20.e15. Li, Z. et al., 2013. The BET bromodomain inhibitor JQ1 activates HIV latency through antagonizing Brd4 inhibition of Tat-transactivation. Nucleic Acids Research, 41(1), pp.277–287. Liang, L. et al., 2015. An epigenome-wide association study of total serum immunoglobulin E concentration. Nature. Liang, Z. et al., 2014. Impaired macrophage phagocytosis of bacteria in severe asthma. Respiratory research, 15(1), p.72. Lieberman, L. a. et al., 2003. STAT1 Plays a Critical Role in the Regulation of Antimicrobial Effector Mechanisms, but Not in the Development of Th1- Type Responses during Toxoplasmosis. The Journal of Immunology, 172(18), pp.457–463. Lighvani, a a et al., 2001. T-bet is rapidly induced by interferon-gamma in lymphoid and myeloid cells. Proceedings of the National Academy of Sciences of the United States of America, 98(26), pp.15137–15142. Liu, F. et al., 2013. Discovery of an in vivo chemical probe of the lysine methyltransferases G9a and GLP. Journal of medicinal chemistry, 56(21), pp.8931–42. Liu, J.J. et al., 2003. BPAG1n4 is essential for retrograde axonal transport in sensory neurons. Journal of Cell Biology, 163(2), pp.223–229. Liu, Y. et al., 2013. Inhibition of p300 impairs Foxp3+ T regulatory cell function and promotes antitumor immunity. Nature Medicine, 19(9), pp.1173– 1177. Loo, B.B. et al., 2000. Production and characterization of the extracellular domain of recombinant human fibroblast growth factor receptor 4. The international journal of biochemistry & cell biology, 32(5), pp.489–97.

233 234

Loza, M.J. et al., 2016. Validated and longitudinally stable asthma phenotypes based on cluster analysis of the ADEPT study. Respiratory Research, 17(1), p.165. Lu, L.-F. et al., 2006. Mast cells are essential intermediaries in regulatory T- cell tolerance. Nature, 442(7106), pp.997–1002. Luger, K., Mäder, A. & Richmond, R., 1997. Crystal structure of the nucleosome core particle at 2.8 Å resolution. Nature, 7, pp.251–260. MacHura, E. et al., 2010. Cytokine production by peripheral blood CD4+ and CD8+ T cells in atopic childhood asthma. Clinical and Developmental Immunology, 2010. Macián, F., García-Rodríguez, C. & Rao, a, 2000. Gene expression elicited by NFAT in the presence or absence of cooperative recruitment of Fos and Jun. The EMBO journal, 19(17), pp.4783–4795. Mackeh, R. et al., 2014. Reactive oxygen species, amp-Activated protein kinase, and the transcription cofactor p300 regulate α-Tubulin acetyltransferase-1 (αtat-1/mec-17)-Dependent microtubule hyperacetylation during cell stress. Journal of Biological Chemistry, 289(17), pp.11816–11828. Magistri, M. et al., 2016. The BET-bromodomain inhibitor JQ1 reduces inflammation and tau phosphorylation at Ser396 in the brain of the 3xTg model of Alzheimer’s disease. Current Alzheimer research, pp.985–995. Mallamacis, M.A. et al., 1993. Identification of Sites on the Human FceRIcr Subunit Which Are Involved in Binding Human and Rat IgE. The Journal of biological chemistry, pp.22076–22083. Mangan, P.R. et al., 2006. Transforming growth factor-beta induces development of the T(H)17 lineage. Nature, 441(7090), pp.231–234. Marion, C. & Roux, B., 1980. Influence of histone H1 on chromatin structure. Biochemical and Biophysical Research Communications, 94(2), pp.535– 541. Marwick, J. a et al., 2009. Inhibition of PI3Kdelta restores glucocorticoid function in smoking-induced airway inflammation in mice. American journal of respiratory and critical care medicine, 179(7), pp.542–8.

234 235

Masoli, M. et al., 2004. The global burden of asthma: Executive summary of the GINA Dissemination Committee Report. Allergy: European Journal of Allergy and Clinical Immunology, 59(5), pp.469–478. McGeachie, M.J. et al., 2013. Polygenic heritability estimates in pharmacogenetics: focus on asthma and related phenotypes. Pharmacogenetics and genomics, 23(6), pp.324–8. McKenzie, A.N. et al., 1993. Interleukin 13, a T-cell-derived cytokine that regulates human monocyte and B-cell function. Proceedings of the National Academy of Sciences of the United States of America, 90(8), pp.3735–9. McNeill, A., Spittle, E. & Bäckström, B.T., 2007. Partial depletion of CD69low- expressing natural regulatory T cells with the anti-CD25 monoclonal antibody PC61. Scandinavian Journal of Immunology, 65(1), pp.63–69. McNicholl, D.M. et al., 2012. The utility of fractional exhaled nitric oxide suppression in the identification of nonadherence in difficult asthma. American Journal of Respiratory and Critical Care Medicine, 186(11), pp.1102–1108. Meng, S. et al., 2014. BET Inhibitor JQ1 Blocks Inflammation and Bone Destruction. Journal of dental research, 93(7), pp.657–62. Metzger, H. et al., 1986. The Receptor with High Affinity for Immunoglobulin E. Annual Review of Immunology, 4(1), pp.419–470. Miller, S., Mohn, S. & Weinmann, A., 2010. Jmjd3 and UTX play a demethylase-independent role in chromatin remodeling to regulate T-box family member- dependent gene expression. Molecular cell, 40(4), pp.594–605. Miranda, C. et al., 2004. Distinguishing severe asthma phenotypes: role of age at onset and eosinophilic inflammation. The Journal of allergy and clinical immunology, 113(1), pp.101–8. Miyahara, N. et al., 2005. Leukotriene B4 Receptor-1 Is Essential for Allergen- Mediated Recruitment of CD8+ T Cells and Airway Hyperresponsiveness. The Journal of Immunology, 174(8), pp.4979–4984. Miyahara, N. et al., 2005. Requirement for leukotriene B4 receptor 1 in

235 236

allergen-induced airway hyperresponsiveness. American Journal of Respiratory and Critical Care Medicine, 172(2), pp.161–167. Miyoshi, S. et al., 2009. (WO2009084693) ANTITUMOR AGENT. Mjösberg, J. et al., 2012. The Transcription Factor GATA3 Is Essential for the Function of Human Type 2 Innate Lymphoid Cells. Immunity, 37(4), pp.649–659. Mock, J.R. et al., 2014. Foxp3+ regulatory T cells promote lung epithelial proliferation. Mucosal immunology, 7(6), pp.1440–51. Moffatt, M. et al., 2010. A large-scale, consortium-based genomewide association study of asthma. New England Journal of Medicine, 363(13), pp.1211–1221. Moore, W. et al., 2007. Characterization of the severe asthma phenotype by the national heart, lung, and blood institute’s severe asthma research program. Journal of Allergy and Clinical Immunology, 119(2), pp.405– 413. Moore, W.C. et al., 2010. Identification of asthma phenotypes using cluster analysis in the Severe Asthma Research Program. American journal of respiratory and critical care medicine, 181(4), pp.315–23. Moore, W.C. et al., 2014. Sputum neutrophil counts are associated with more severe asthma phenotypes using cluster analysis. Journal of Allergy and Clinical Immunology, 133(6), p.1557–1563.e5. Moriggl, R. et al., 1999. Stat5 is required for IL-2-induced cell cycle progression of peripheral T cells. Immunity, 10(2), pp.249–59. Moro, K. et al., 2010. Innate production of T(H)2 cytokines by adipose tissue- associated c-Kit(+)Sca-1(+) lymphoid cells. Nature, 463(7280), pp.540–4. Morrissette, N.S. & Sibley, L.D., 2002. Disruption of microtubules uncouples budding and nuclear division in Toxoplasma gondii. Journal of cell science, 115(Pt 5), pp.1017–1025. Moulin, D. et al., 2007. Interleukin (IL)-33 induces the release of pro- inflammatory mediators by mast cells. Cytokine, 40(3), pp.216–225. Mullen, A.C. et al., 2002. Hlx is induced by and genetically interacts with T-bet to promote heritable T(H)1 gene induction. Nature immunology, 3(7),

236 237

pp.652–658. Mullen, a C. et al., 2001. Role of T-bet in commitment of TH1 cells before IL- 12-dependent selection. Science, 292(5523), pp.1907–1910. Murphy, K.R. & Bender, B.G., 2009. Treatment of moderate to severe asthma : patient perspectives on combination inhaler therapy and implications for adherence. J Asthma Allergy, 2, pp.63–72. Nagarkar, D.R. et al., 2012. Airway epithelial cells activate T(H)2 cytokine production in mast cells through IL-1 and thymic stromal lymphopoietin. The Journal of allergy and clinical immunology, pp.20–22. Nathan, A.T. et al., 2009. Innate immune responses of airway epithelium to house dust mite are mediated through [beta]-glucan-dependent pathways. Journal of Allergy and Clinical Immunology, 123(3), pp.612– 618. Neill, D.R. et al., 2010. Nuocytes represent a new innate effector leukocyte that mediates type-2 immunity. Nature, 464(7293), pp.1367–70. Nimura, K. & Kaneda, Y., 2015. Elucidating the mechanisms of transcription regulation during heart development by next-generation sequencing. Journal of Human Genetics, 61(April), pp.1–8. Noelle, R. et al., 1984. Increased expression of la antigens on resting B cells. Proc.Natl.Acad.Sci.U.S.A., 81(October), pp.6149–6153. Noonan, M. et al., 2013. Dose-ranging study of lebrikizumab in asthmatic patients not receiving inhaled steroids. Journal of Allergy and Clinical Immunology, 132(3), p.567–574.e12. Nurieva, R. et al., 2007. Essential autocrine regulation by IL-21 in the generation of inflammatory T cells. Nature, 448(7152), pp.480–3. Nurse, B. et al., 2000. PBMCs from both atopic asthmatic and nonatopic children show a T(H)2 cytokine response to house dust mite allergen. Journal of Allergy and Clinical Immunology, 106, pp.84–91. Oboki, K. et al., 2010. IL-33 and IL-33 receptors in host defense and diseases. Allergology international : official journal of the Japanese Society of Allergology, 59(2), pp.143–60. Oh, K. et al., 2013. Airway epithelial cells initiate the allergen response

237 238

through transglutaminase 2 by inducing IL-33 expression and a subsequent Th2 response. Resiratory Research, 14(35), pp.1–9. Oliveira, S.H.P. et al., 2002. Increased responsiveness of murine eosinophils to MIP-1beta (CCL4) and TCA-3 (CCL1) is mediated by their specific receptors, CCR5 and CCR8. Journal of leukocyte biology, 71(6), pp.1019–1025. Oprea, T.I. et al., 2009. A crowdsourcing evaluation of the NIH chemical probes. Nature Chemical Biology, 5(7), pp.441–447. Ortega, H.G. et al., 2014. Mepolizumab Treatment in Patients with Severe Eosinophilic Asthma. New England Journal of Medicine, 371(13), pp.1198–1207. den Otter, I. et al., 2016. Lung function decline in asthma patients with elevated bronchial CD8, CD4 and CD3 cells. European Respiratory Journal, p.ERJ-01525-2015. Ouyang, W. et al., 1998. Inhibition of Th1 development mediated by GATA-3 through an IL-4-independent mechanism. Immunity, 9(5), pp.745–755. Ouyang, W. et al., 2000. Stat6-independent GATA-3 autoactivation directs IL- 4-independent Th2 development and commitment. Immunity, 12(1), pp.27–37. Owen, J.A. et al., 2013. Kuby Immunology, New York: W. H. Freeman and Company. Pan, C. & Fan, Y., 2016. Role of H1 linker histones in mammalian development and stem cell differentiation. Biochimica et Biophysica Acta - Gene Regulatory Mechanisms, 1859(3), pp.496–509. Pandey, A.K. et al., 2014. Functionally Enigmatic Genes: A Case Study of the Brain Ignorome C. A. Ouzounis, ed. PLoS ONE, 9(2), p.e88889. Patel, D.D., 2001. Escape from tolerance in the human X-linked autoimmunity-allergic disregulation syndrome and the Scurfy mouse. Journal of Clinical Investigation, 107(2), pp.155–157. Paul, S. & Lombroso, P.J., 2003. Receptor and nonreceptor protein tyrosine phosphatases in the nervous system. Cellular and Molecular Life Sciences, 60(11), pp.2465–2482.

238 239

Pavord, I.D. et al., 2012. Mepolizumab for severe eosinophilic asthma (DREAM): A multicentre, double-blind, placebo-controlled trial. The Lancet, 380(9842), pp.651–659. Pearce, E.L. et al., 2003. Control of effector CD8+ T cell function by the transcription factor Eomesodermin. Science (New York, N.Y.), 302(5647), pp.1041–1043. Pearce, N., Pekkanen, J. & Beasley, R., 1999. How much asthma is really attributable to atopy? Topic collections How much asthma is really attributable to atopy? Thorax, 54, pp.268–272. Perros, F. et al., 2009. Blockade of CCR4 in a humanized model of asthma reveals a critical role for DC-derived CCL17 and CCL22 in attracting Th2 cells and inducing airway inflammation. Allergy, 64(7), pp.995–1002. Peters, M.C. et al., 2016. Plasma interleukin-6 concentrations, metabolic dysfunction, and asthma severity: A cross-sectional analysis of two cohorts. The Lancet Respiratory Medicine, 4(7), pp.574–584. Peters, S.P. et al., 2010. Tiotropium bromide step-up therapy for adults with uncontrolled asthma. The New England journal of medicine, 363(18), pp.1715–26. Picaud, S. et al., 2016. Promiscuous targeting of bromodomains by Bromosporine identifies BET proteins as master regulators of primary transcription response in leukemia. Science Advances, p.in press. Pinz, S. et al., 2015. Deacetylase inhibitors repress STAT5-mediated transcription by interfering with bromodomain and extra-terminal (BET) protein function. Nucleic Acids Research, pp.1–22. Platts-Mills, T.A.E., 2001. The role of immunoglobulin E in allergy and asthma. American Journal of Respiratory and Critical Care Medicine, 164(8 II). Polansky, J.K. et al., 2008. DNA methylation controls Foxp3 gene expression. European Journal of Immunology, 38(6), pp.1654–1663. Poynter, M., 2012. Airway Epithelial Regulation of Allergic Sensitization in Asthma. Pulmonary pharmacology & therapeutics, 25(6), pp.438–446. Price, A.E. et al., 2010. Systemically dispersed innate IL-13-expressing cells in type 2 immunity. Proceedings of the National Academy of Sciences,

239 240

107(25), pp.11489–11494. Punnonen, J. et al., 1993. Interleukin 13 induces interleukin 4-independent IgG4 and IgE synthesis and CD23 expression by human B cells. Proceedings of the National Academy of Sciences of the United States of America, 90(8), pp.3730–4. Rai, K. et al., 2010. Dnmt3 and G9a cooperate for tissue-specific development in zebrafish. The Journal of biological chemistry, 285(6), pp.4110–21. Ramirez, F. et al., 2016. deepTools2: a next generation web server for deep- sequencing data analysis. Nucleic acids research, 44(April), pp.160–165. Rand, C.S. & Wise, R.A., 1994. Measuring adherence to asthma medication regimens. American Journal of Respiratory and Critical Care Medicine, 149(2 II), pp.569–576. Ransom, M., Dennehey, B. & Tyler, J., 2010. Chaperoning histones during DNA replication and repair. Cell, 140(2), pp.183–195. Rastogi, D., Suzuki, M. & Greally, J.M., 2013. Differential epigenome-wide DNA methylation patterns in childhood obesity-associated asthma. Scientific reports, 3(Table 1), p.2164. Read, S., Malmstrom, V. & Powrie, F., 2000. Cytotoxic T lymphocyte- associated antigen 4 plays an essential role in the function of CD25(+)CD4(+) regulatory cells that control intestinal inflammation. J.Exp.Med., 192(0022–1007 (Print)), pp.295–302. Read, S., Malmström, V. & Powrie, F., 2000. Cytotoxic T Lymphocyte– Associated Antigen 4 Plays an Essential Role in the Function of Cd25 + Cd4 + Regulatory Cells That Control Intestinal Inflammation. The Journal of Experimental Medicine, 192(2), pp.295–302. Redhu, N.S. & Gounni, A.S., 2012. Function and mechanisms of TSLP/TSLPR complex in asthma and COPD. Clinical and experimental allergy : journal of the British Society for Allergy and Clinical Immunology, 42(7), pp.994–1005. Reichel, C.A. et al., 2009. Ccl2 and Ccl3 mediate neutrophil recruitment via induction of protein synthesis and generation of lipid mediators. Arteriosclerosis, Thrombosis, and Vascular Biology, 29(11), pp.1787–

240 241

1793. Reichert, P. et al., 2001. Cutting edge: in vivo identification of TCR redistribution and polarized IL-2 production by naive CD4 T cells. J Immunol, 166(7), pp.4278–4281. Reimand, J. et al., 2016. g:Profiler—a web server for functional interpretation of gene lists (2016 update). Nucleic Acids Research , pp.1–7. Riba, M. et al., 2016. Revealing the acute asthma ignorome: characterization and validation of uninvestigated gene networks. Scientific reports, 6(April), p.24647. Robertson, K., Ait-Si-Ali, S. & Yokochi, T., 2000. DNMT1 forms a complex with Rb, E2F1 and HDAC1 and represses transcription from E2F- responsive promoters. Nature genetics, 25(July). Rosales-Reynoso, M.A. et al., 2010. Gene expression profiling identifies WNT7A as a possible candidate gene for decreased cancer risk in fragile X syndrome patients. Arch Med Res, 41(2), p.110–118 e2. Roussel, L. et al., 2010. IL-17 promotes p38 MAPK-dependent endothelial activation enhancing neutrophil recruitment to sites of inflammation. J Immunol, 184(8), pp.4531–4537. Royce, S., 2012. Histone deacetylase inhibitors: can we consider potent anti- neoplastic agents for the treatment of asthma? Annals of Clinical & Laboratory Science, 42(3), pp.338–345. Royce, S.G. & Karagiannis, T.C., 2014. Histone deacetylases and their inhibitors: new implications for asthma and chronic respiratory conditions. Current opinion in allergy and clinical immunology, 14(1), pp.44–8. Roychoudhuri, R. et al., 2016. BACH2 regulates CD8+ T cell differentiation by controlling access of AP-1 factors to enhancers. Nature Immunology, 17(February 2015), pp.1–12. Di Ruscio, A. et al., 2013. DNMT1-interacting RNAs block gene-specific DNA methylation. Nature, 503(7476), pp.371–6. Sahl, H.-G. et al., 2005. Mammalian defensins: structures and mechanism of antibiotic activity. Journal of Leukocyte Biology, 77(4), pp.466–475. Saito, T. & Yokosuka, T., 2006. Immunological synapse and microclusters:

241 242

the site for recognition and activation of T cells. Current Opinion in Immunology, 18(3), pp.305–313. Sakaguchi, S. et al., 1995. Immunologic self-tolerance maintained by activated T cells expressing IL-2 receptor alpha-chains (CD25). Breakdown of a single mechanism of self-tolerance causes various autoimmune diseases. J Immunol, 155(3), pp.1151–1164. Sakaguchi, S., Wing, K. & Miyara, M., 2007. Regulatory T cells - A brief history and perspective. European Journal of Immunology, 37(SUPPL. 1). Sancho, D. et al., 2002. Regulation of microtubule-organizing center orientation and actomyosin cytoskeleton rearrangement during immune interactions. Immunol Rev, 189(2), pp.84–97. Santangelo, S. et al., 2009. Chromatin structure and DNA methylation of the IL-4 gene in human T(H)2 cells. Chromosome research : an international journal on the molecular, supramolecular and evolutionary aspects of chromosome biology, 17(4), pp.485–496. Satoh, T. et al., 2010. The Jmjd3-Irf4 axis regulates M2 macrophage polarization and host responses against helminth infection. Nature immunology, 11(10), pp.936–44. Schedel, M. et al., 2016. 1,25D3 prevents CD8(+)Tc2 skewing and asthma development through VDR binding changes to the Cyp11a1 promoter. Nature communications, 7, p.10213. Schones, D.E. & Zhao, K., 2008. Genome-wide approaches to studying chromatin modifications. Nature reviews. Genetics, 9(3), pp.179–91. Schroeder, A. et al., 2006. The RIN: an RNA integrity number for assigning integrity values to RNA measurements. BMC molecular biology, 7(1), p.3. Serra-Batlles, J. et al., 1998. Costs of asthma according to the degree of severity. European Respiratory Journal, 12(6), pp.1322–1326. Seumois, G. et al., 2014. Epigenomic analysis of primary human T cells reveals enhancers associated with TH2 memory cell differentiation and asthma susceptibility. Nature immunology, 15(8). Shafer, W.M. et al., 1991. Human lysosomal cathepsin G and granzyme B share a functionally conserved broad spectrum antibacterial peptide.

242 243

Journal of Biological Chemistry, 266(1), pp.112–116. Shah, S. et al., 2016. DUSP1 maintains IRF1 and leads to increased expression of IRF1-dependent genes: A mechanism promoting glucocorticoid-insensitivity. Journal of Biological Chemistry, 291(41), p.jbc.M116.728964. Shannon, J. et al., 2008. Differences in airway cytokine profile in severe asthma compared to moderate asthma. Chest, 133(2), pp.420–426. Shaw, D.E. et al., 2015. Clinical and inflammatory characteristics of the European U-BIOPRED adult severe asthma cohort. European Respiratory Journal, 46(5), pp.1308–1321. Shaw, R.J. et al., 1985. Activated human eosinophils generate SRS-A leukotrienes following IgG-dependent stimulation. Nature, 316(6024), pp.150–152. Sher, a et al., 1991. Production of IL-10 by CD4+ T lymphocytes correlates with down-regulation of Th1 cytokine synthesis in helminth infection. Journal of immunology (Baltimore, Md. : 1950), 147(8), pp.2713–2716. Shi, Y.L. et al., 2012. Histone deacetylases inhibitor Trichostatin A ameliorates DNFB-induced allergic contact dermatitis and reduces epidermal Langerhans cells in mice. Journal of Dermatological Science, 68(2), pp.99–107. Shin, J.E. et al., 2010. Treponema denticola suppresses expression of human {beta}-defensin-3 in gingival epithelial cells through inhibition of the toll- like receptor 2 axis. Infect Immun, 78(2), pp.672–679. Shu, S. et al., 2016. Response and resistance to BET bromodomain inhibitors in triple-negative breast cancer. Nature, 529(7586), pp.1–24. Simon, J. a & Lange, C. a, 2008. Roles of the EZH2 histone methyltransferase in cancer epigenetics. Mutation research, 647(1–2), pp.21–9. Simpson, J.L. et al., 2014. Elevated expression of the NLRP3 inflammasome in neutrophilic asthma. European Respiratory Journal, 43(4), pp.1067– 1076. Singer, B.B. et al., 2014. Soluble CEACAM8 interacts with CEACAM1 inhibiting TLR2-triggered immune responses. PLoS ONE, 9(4).

243 244

Smith, C.A., McClive, P.J. & Sinclair, A.H., 2005. Temporal and spatial expression profile of the novel armadillo-related gene,Alex2, during testicular differentiation in the mouse embryo. Developmental Dynamics, 233(1), pp.188–193. Soumelis, V. et al., 2002. Human epithelial cells trigger dendritic cell mediated allergic inflammation by producing TSLP. Nature immunology, 3(7), pp.673–80. Sousa Martins, J.P. et al., 2016. DAZL and CPEB1 regulate mRNA translation synergistically during oocyte maturation. Journal of Cell Science, 129(6), pp.1271–1282. Spits, H. et al., 2013. Innate lymphoid cells--a proposal for uniform nomenclature. Nature reviews. Immunology, 13(2), pp.145–9. Stanbury, R.M. & Graham, E.M., 1998. Systemic corticosteroid therapy — side effects and their management. Br J Opthalmol, 82, pp.704–708. Stec, I. et al., 1998. WHSC1, a 90 kb SET domain-containing gene, expressed in early development and homologous to a Drosophila dysmorphy gene maps in the Wolf-Hirschhorn syndrome critical region and is fused to IgH in t(4;14) multiple myeloma. Human Molecular Genetics, 7(7), pp.1071–1082. Stefano, J.E., 1984. Purified lupus antigen la recognizes an oligouridylate stretch common to the 3’ termini of RNA polymerase III transcripts. Cell, 36(1), pp.145–154. Stefanowicz, D. et al., 2015. Elevated H3K18 acetylation in airway epithelial cells of asthmatic subjects. Respiratory Research, 16(1), p.95. Su, R.-C. et al., 2008. Epigenetic regulation of established human type 1 versus type 2 cytokine responses. The Journal of allergy and clinical immunology, 121, p.57–63.e3. Sugimoto, N. et al., 2003. Differential requirements for JAK2 and TYK2 in T cell proliferation and IFN- q production induced by IL-12 alone or together with IL-18. , pp.243–251. Sun, J. et al., 2015. Histone H1-mediated epigenetic regulation controls germline stem cell self-renewal by modulating H4K16 acetylation. Nature

244 245

Communications, pp.1–10. Sung, S.-S.J. et al., 2006. A Major Lung CD103 ( E)- 7 Integrin-Positive Epithelial Dendritic Cell Population Expressing Langerin and Tight Junction Proteins. The Journal of Immunology, 176(4), pp.2161–2172. Swain, S.L., Weinberg, A.D., English, M., et al., 1990. IL-4 directs the development of Th2-like helper effectors. Journal of Immunology, 145(11), pp.3796–3806. Swain, S.L., Weinberg, A.D. & English, M., 1990. CD4+ T cell subsets. Lymphokine secretion of memory cells and of effector cells that develop from precursors in vitro. Journal of immunology (Baltimore, Md. : 1950), 144(5), pp.1788–99. Szabo, S.J. et al., 2000. A novel transcription factor, T-bet, directs Th1 lineage commitment. Cell, 100(6), pp.655–669. Szabo, S.J. et al., 2002. Distinct effects of T-bet in TH1 lineage commitment and IFN-gamma production in CD4 and CD8 T cells. Science (New York, N.Y.), 295(5553), pp.338–42. Tachibana, I. & Hemler, M.E., 1999. Role of transmembrane 4 superfamily (TM4SF) proteins CD9 and CD81 in muscle cell fusion and myotube maintenance. Journal of Cell Biology, 146(4), pp.893–904. Tachibana, M. et al., 2002. G9a histone methyltransferase plays a dominant role in euchromatic histone H3 lysine 9 methylation and is essential for early embryogenesis. Genes and Development, 16, pp.1779–1791. Tamari, M., ShotaTanaka & Hirota, T., 2013. Genome-Wide Association Studies of Allergic Diseases. Allergology International, 62(1), pp.21–28. Taube, C. et al., 2004. Mast cells, Fc epsilon RI, and IL-13 are required for development of airway hyperresponsiveness after aerosolized allergen exposure in the absence of adjuvant. Journal of Immunology, 172(10), pp.6398–6406. Taube, C. et al., 2006. The Leukotriene B4 Receptor (BLT1) Is Required for Effector CD8+ T Cell-Mediated, Mast Cell-Dependent Airway Hyperresponsiveness. The Journal of Immunology, 176(5), pp.3157– 3164.

245 246

Tessarz, P. & Kouzarides, T., 2014. Histone core modifications regulating nucleosome structure and dynamics. Nature reviews. Molecular cell biology, 15(11), pp.703–708. The ENFUMOSA Study, 2003. The ENFUMOSA cross-sectional European multicentre study of the clinical phenotype of chronic severe asthma. European Respiratory Journal, 22(3), pp.470–477. Thierfelder, W.E. et al., 1996. Requirement for Stat4 in interleukin-12- mediated responses of natural killer and T cells. Nature, 382(6587), pp.171–174. Thomas, M.P. et al., 2014. Leukocyte Protease Binding to Nucleic Acids Promotes Nuclear Localization and Cleavage of Nucleic Acid Binding Proteins. The Journal of Immunology, 192(11), pp.5390–5397. Thomsen, S.F. et al., 2010. Estimates of asthma heritability in a large twin sample. Clinical and Experimental Allergy, 40, pp.1054–1061. Todd, S C Lipps, S G Crisa, L Salomon, D.R.T.C.D. et al., 1996. CD81 Expressed on Human Thymocytes Mediates Integrin Activation and Interleukin 2-depedent Proliferation. J. Exp. Med, 184(November), pp.2055–2060. Townsend, M.J. et al., 2000. T1/St2-Deficient Mice Demonstrate the Importance of T1/St2 in Developing Primary T Helper Cell Type 2 Responses. The Journal of Experimental Medicine, 191(6), pp.1069– 1076. Trompette, A. et al., 2009. Allergenicity resulting from functional mimicry of a Toll-like receptor complex protein. Nature, 457(7229), pp.585–588. Tsitsiou, E. et al., 2012. Transcriptome analysis shows activation of circulating CD8+ T cells in patients with severe asthma. The Journal of allergy and clinical immunology, 129(1), pp.95–103. Tumes, D.J. et al., 2013. The polycomb protein Ezh2 regulates differentiation and plasticity of CD4(+) T helper type 1 and type 2 cells. Immunity, 39(5), pp.819–32. Ugarte-Gil, C.A. et al., 2013. Induced Sputum MMP-1, -3 & -8 Concentrations during Treatment of Tuberculosis. PLoS ONE, 8(4), pp.2–9.

246 247

Usui, T. et al., 2003. GATA-3 suppresses Th1 development by downregulation of Stat4 and not through effects on IL-12Rβ2 chain or T- bet. Immunity, 18(3), pp.415–428. Usui, T. et al., 2006. T-bet regulates Th1 responses through essential effects on GATA-3 function rather than on IFNG gene acetylation and transcription. The Journal of experimental medicine, 203(3), pp.755–66. Valitutti, S. et al., 1995. Sustained signaling leading to T cell activation results from prolonged T cell receptor occupancy. Role of T cell actin cytoskeleton. The Journal of experimental medicine, 181(2), pp.577–84. Veldhoen, M. et al., 2006. TGF[beta] in the context of an inflammatory cytokine milieu supports de novo differentiation of IL-17-producing T cells. Immunity, 24(2), pp.179–189. Verissimo, F. et al., 2015. A microtubule-independent role of p150glued in secretory cargo concentration at endoplasmic reticulum exit sites. Journal of cell science, pp.4160–4170. Verma, M., Chattopadhyay, B.D. & Paul, B.N., 2013. Epigenetic regulation of DNMT1 gene in mouse model of asthma disease. Molecular biology reports, 40(3), pp.2357–68. Vijayanand, P. et al., 2012. Interleukin-4 Production by Follicular Helper T Cells Requires the Conserved Il4 Enhancer Hypersensitivity Site V. Immunity, 36(2), pp.175–187. Voraphani, N. et al., 2014. An airway epithelial iNOS-DUOX2-thyroid peroxidase metabolome drives Th1/Th2 nitrative stress in human severe asthma. Mucosal immunology, 7(5), pp.1175–85. Vukmanovic-Stejic, M. et al., 2000. Human Tc1 and Tc2/Tc0 CD8 T-cell clones display distinct cell surface and functional phenotypes. Blood, 95(1), pp.231–240. Walker, J.A., Barlow, J.L. & Mckenzie, A.N.J., 2013. Innate lymphoid cells — how did we miss them? Nature Reviews Immunology, 13(2), pp.75–87. Walter, M.J. et al., 2001. Interleukin 12 p40 production by barrier epithelial cells during airway inflammation. The Journal of experimental medicine, 193(3), pp.339–51.

247 248

Wan, Y.Y., 2014. GATA3: A master of many trades in immune regulation. Trends in Immunology, 35(6), pp.233–242. Wang, H. et al., 2001. Purification and functional characterization of a histone H3-lysine 4-specific methyltransferase. Molecular Cell, 8, pp.1207–1217. Wang, J. et al., 2016. Therapeutic Effect of Histone Deacetylase Inhibitor, Sodium Butyrate, on Allergic Rhinitis In Vivo. DNA and Cell Biology, 35(4), pp.203–208. Wang, Y. et al., 2013. Development and regeneration of Sox2+ endoderm progenitors are regulated by a Hdac1/2-Bmp4/Rb1 regulatory pathway. Developmental cell, 24(4), pp.345–358. Wang, Y. et al., 2013. GATA-3 controls the maintenance and proliferation of T cells downstream of TCR and cytokine signaling. Nature immunology, 14(7), pp.714–22. Wehkamp, J. et al., 2004. NF-kB- and AP-1-Mediated Induction of Human Beta Defensin-2 in Intestinal Epithelial Cells by Escherichia coli Nissle 1917: a Novel Effect of a Probiotic Bacterium. Infection and Immunity, 72(10), pp.5750–5758. Wei, G. et al., 2009. Global mapping of H3K4me3 and H3K27me3 reveals specificity and plasticity in lineage fate determination of differentiating CD4+ T cells. Immunity, 30(1), pp.155–67. Wei, L. et al., 2007. IL-21 is produced by Th17 cells and drives IL-17 production in a STAT3-dependent manner. The Journal of biological chemistry, 282(48), pp.34605–34610. Wenzel, S. et al., 2013. Dupilumab in persistent asthma with elevated eosinophil levels. The New England journal of medicine, 368(26), pp.2455–66. Wenzel, S.E. et al., 1997. Bronchoscopic evaluation of severe asthma: Persistent inflammation associated with high dose glucocorticoids. American Journal of Respiratory and Critical Care Medicine, 156(3 I), pp.737–743. Wenzel, S.E., 2016. Emergence of Biomolecular Pathways to Define Novel Asthma Phenotypes. Type-2 Immunity and Beyond. American journal of

248 249

respiratory cell and molecular biology, 55(1), pp.1–4. Wenzel, S.E. et al., 1999. Evidence That Severe Asthma Can Be Divided Pathologically into Two Inflammatory Subtypes with Distinct Physiologic and Clinical Characteristics. American Journal of Respiratory and Critical Care Medicine, 160(3), pp.1001–1008. Whetstine, J.R. et al., 2006. Reversal of histone lysine trimethylation by the JMJD2 family of histone demethylases. Cell, 125(3), pp.467–81. Whitmire, J.K., Tan, J.T. & Whitton, J.L., 2005. Interferon-gamma acts directly on CD8+ T cells to increase their abundance during virus infection. The Journal of experimental medicine, 201(7), pp.1053–9. Wildin, R.S. et al., 2001. X-linked neonatal diabetes mellitus, enteropathy and endocrinopathy syndrome is the human equivalent of mouse scurfy. Nature Genetics, 27(1), pp.18–20. Williams, C.L. et al., 2013. STAT4 and T-bet are required for the plasticity of IFN-γ expression across Th2 ontogeny and influence changes in Ifng promoter DNA methylation1. Journal of Immunology, 191(2), pp.678–687. Williams, L.M. & Rudensky, A.Y., 2007. Maintenance of the Foxp3-dependent developmental program in mature regulatory T cells requires continued expression of Foxp3. Nature immunology, 8(3), pp.277–284. Wilson, S.J. et al., 2016. Severe asthma exists despite suppressed tissue inflammation: findings of the U-BIOPRED study. European Respiratory Journal, (Imi), pp.1–13. Wirth, T.C. et al., 2010. Repetitive Antigen Stimulation Induces Stepwise Transcriptome Diversification but Preserves a Core Signature of Memory CD8+ T Cell Differentiation. Immunity, 33(1), pp.128–140. Woodruff, P.G. et al., 2007. Genome-wide profiling identifies epithelial cell genes associated with asthma and with treatment response to corticosteroids. Proceedings of the National Academy of Sciences of the United States of America, 104(40), pp.1558–63. Woodruff, P.G. et al., 2009. T-helper type 2-driven inflammation defines major subphenotypes of asthma. American journal of respiratory and critical care medicine, 180(5), pp.388–95.

249 250

Wu, S.-Y. & Chiang, C.-M., 2007. The double bromodomain-containing chromatin adaptor Brd4 and transcriptional regulation. The Journal of biological chemistry, 282(18), pp.13141–5. Wu, S.Y. & Chiang, C.M., 2007. The double bromodomain-containing chromatin adaptor Brd4 and transcriptional regulation. Journal of Biological Chemistry, 282, pp.13141–13145. Xiang, Y. et al., 2007. JMJD3 is a histone H3K27 demethylase. Cell research, 17(10), pp.850–7. Xu, L. et al., 2007. Cutting edge: regulatory T cells induce CD4+CD25-Foxp3- T cells or are self-induced to become Th17 cells in the absence of exogenous TGF-beta. Journal of immunology (Baltimore, Md. : 1950), 178(11), pp.6725–9. Xu, W., Parmigiani, R. & Marks, P., 2007. Histone deacetylase inhibitors: molecular mechanisms of action. Oncogene, 26, pp.5541–5552. Yahata, T. et al., 2000. The MSG1 Non-DNA-binding Transactivator Binds to the p300 / CBP Coactivators , Enhancing Their Functional Link to the Smad Transcription Factors. The Journal of biological chemistry, 275(12), pp.8825–8834. Yamashita, M. et al., 2006. Crucial Role of MLL for the Maintenance of Memory T Helper Type 2 Cell Responses. Immunity, 24(5), pp.611–622. Yang, X.-D. et al., 2009. Negative regulation of NF-kappaB action by Set9- mediated lysine methylation of the RelA subunit. The EMBO journal, 28(8), pp.1055–66. Yang, X.O. et al., 2008. Molecular antagonism and plasticity of regulatory and inflammatory T cell programs. Immunity, 29(1), pp.44–56. Yang, X.O. et al., 2007. STAT3 regulates cytokine-mediated generation of inflammatory helper T cells. Journal of Biological Chemistry, 282(13), pp.9358–9363. Yang, Z. et al., 2005. Recruitment of P-TEFb for stimulation of transcriptional elongation by the bromodomain protein Brd4. Molecular Cell, 19, pp.535– 545. Yao, Z. et al., 2007. Nonredundant roles for Stat5a/b in directly regulating

250 251

Foxp. Blood, 109(10), pp.4368–4375. Ying, S. et al., 2005. Thymic Stromal Lymphopoietin Expression Is Increased in Asthmatic Airways and Correlates with Expression of Th2-Attracting Chemokines and Disease Severity. The Journal of Immunology, 174(12), pp.8183–8190. Yoon, H. et al., 2011. CITED2 controls the hypoxic signaling by snatching p300 from the two distinct activation domains of HIF-1?? Biochimica et Biophysica Acta - Molecular Cell Research, 1813(12), pp.2008–2016. Yoon, J., Terada, A. & Kita, H., 2007. CD66b Regulates Adhesion and Activation of Human Eosinophils. The Journal of Immunology, 179(12), pp.8454–8462. Yoshidas, M., 1990. Potent and Specific Inhibition of Mammalian Histone Deacetylase Both in Vivo and in Vitro by Trichostatin A *. October, 265(28), pp.17174–17179. Young, K.G., Pinheiro, B. & Kothary, R., 2006. A Bpag1 isoform involved in cytoskeletal organization surrounding the nucleus. Experimental Cell Research, 312(2), pp.121–134. Young, M.D. et al., 2011. ChIP-seq analysis reveals distinct H3K27me3 profiles that correlate with transcriptional activity. Nucleic acids research, 39(17), pp.7415–27. Yu, Q. et al., 2012. DNA methyltransferase 3a limits the expression of interleukin-13 in T helper 2 cells and allergic airway inflammation. Proceedings of the National Academy of Sciences, 109, pp.541–546. Yuksel, H. et al., 2013. Role of vascular endothelial growth factor antagonism on airway remodeling in asthma. Annals of Allergy, Asthma & Immunology, 110(3), pp.150–155. Zeller, C. et al., 2012. Candidate DNA methylation drivers of acquired cisplatin resistance in ovarian cancer identified by methylome and expression profiling. Oncogene, 31(42), pp.4567–76. Zeng, L. & Zhou, M.M., 2002. Bromodomain: An acetyl-lysine binding domain. FEBS Letters, 513, pp.124–128. Zhang, B. et al., 2005. A General Framework for Weighted Gene Co-

251 252

Expression Network Analysis A General Framework for Weighted Gene Co- Expression Network Analysis. , 4(1). Zhang, D.H. et al., 1999. Inhibition of allergic inflammation in a murine model of asthma by expression of a dominant-negative mutant of GATA-3. Immunity, 11(4), pp.473–482. Zhang, D.H. et al., 1997. Transcription factor GATA-3 is differentially expressed in murine Th1 and Th2 cells and controls Th2-specific expression of the interleukin-5 gene. Journal of Biological Chemistry, 272(34), pp.21597–21603. Zhang, H. et al., 2014. The interplay of DNA methylation over time with Th2 pathway genetic variants on asthma risk and temporal asthma transition. Clinical epigenetics, 6, p.8. Zhang, W. et al., 2012. Bromodomain-containing protein 4 (BRD4) regulates RNA polymerase II serine 2 phosphorylation in human CD4+ T cells. The Journal of biological chemistry, 287(51), pp.43137–55. Zhang, W.-S. et al., 2011. Neuromedin B and its receptor induce labor onset and are associated with the RELA (NFKB P65)/IL6 pathway in pregnant mice. Biology of reproduction, 84(1), pp.113–7. Zhang, Y. et al., 2010. Genetic reduction of striatal-enriched tyrosine phosphatase (STEP) reverses cognitive and cellular deficits in an Alzheimer’s disease mouse model. Proceedings of the National Academy of Sciences, 107(44), pp.19014–19019. Zhang, Y., Maksimovic, J. & Naselli, G., 2013. Genome-wide DNA methylation analysis identifies hypomethylated genes regulated by FOXP3 in human regulatory T cells. Blood, 122(16), pp.2823–2836. Zhao, Y., Yang, J. & Gao, Y.-D., 2011. Altered expressions of helper T cell (Th)1, Th2, and Th17 cytokines in CD8(+) and γδ T cells in patients with allergic asthma. The Journal of asthma : official journal of the Association for the Care of Asthma, 48(5), pp.429–436. Zheng, W. & Flavell, R. a, 1997. The transcription factor GATA-3 is necessary and sufficient for Th2 cytokine gene expression in CD4 T cells. Cell, 89(4), pp.587–596.

252 253

Zhou, B. et al., 2005. Thymic stromal lymphopoietin as a key initiator of allergic airway inflammation in mice. Nature immunology, 6(10), pp.1047– 1053. Zhou, L. et al., 2007a. IL-6 programs T(H)-17 cell differentiation by promoting sequential engagement of the IL-21 and IL-23 pathways. Nature immunology, 8(9), pp.967–974. Zhou, L. et al., 2007b. IL-6 programs T(H)-17 cell differentiation by promoting sequential engagement of the IL-21 and IL-23 pathways. Nature immunology, 8(9), pp.967–974. Zhou, Z. et al., 2012. Antiviral Activities of ISG20. Virology, 409(2), pp.175– 188. Zhu, J. et al., 2003. Stat5 activation plays a critical role in Th2 differentiation. Immunity, 19(5), pp.739–748. Zhu, J. et al., 2001. Stat6 is necessary and sufficient for IL-4's role in Th2 differentiation and cell expansion. J Immunol, 166(12), pp.7276– 7281. Zhu, Y., van Essen, D. & Saccani, S., 2012. Cell-type-specific control of enhancer activity by H3K9 trimethylation. Molecular cell, 46(4), pp.408– 23. Zuany-Amorim, C. et al., 2002. Suppression of airway eosinophilia by killed Mycobacterium vaccae-induced allergen-specific regulatory T-cells. Nature medicine, 8(6), pp.625–629.

253