Using Dendritic Cell Receptors to Enhance Immunity

Using Dendritic Cell Receptors To Enhance Immunity

Jessica Li

ORCID: 0000-0003-1951-8320

Submitted in total fulfilment of the requirements of the degree of Doctor of Philosophy

May 2017

Department of Microbiology and Immunology The University of Melbourne

Abstract

Dendritic cells (DCs) are the most potent initiators of immune responses, being highly specialised for the uptake and presentation of antigens (Ag) to activate T cells. Their priming potential can be harnessed to generate stronger immune responses by targeting Ag to DCs via monoclonal antibodies (mAbs) specific for DC-expressed surface receptors. This thesis builds upon the concept of targeting DCs in two main ways: firstly, by investigating a novel method of targeting adjuvant to DCs, and secondly, by investigating how DC-targeting constructs can be used to prime and boost responses.

It was considered whether not only Ag, but also adjuvants could be targeted to DCs to improve their efficacy. A recent finding that the DC receptor DEC-205 can bind to and mediate the immunostimulatory effects of CpG oligonucleotide (ODN) adjuvants led to the hypothesis that CpG ODNs could be targeted to DCs via DEC-205 in order to enhance their potency. The interaction between DEC-205 and CpG ODNs was further characterised to determine the molecular properties of ODNs required for binding. This information was then used to enhance the DEC-205 binding capacity of a particular CpG ODN that normally only weakly binds DEC-205. Enhanced DEC- 205 binding was found to significantly improve the stimulatory capacity of this ODN, demonstrating that targeting adjuvant to DCs could be a viable method to improve adjuvant potency. Another receptor, CD14, has also been reported to bind CpG ODNs, so the potential for CD14 to act in synergy with DEC-205 was investigated. However, CD14 was not observed to mediate the uptake or stimulatory effects of CpG ODNs.

The identification of natural ligands of DEC-205 is critical for understanding its physiological function. Although ODNs are synthetic molecules, their binding to DEC-205 may signify that DEC-205 is capable of binding other types of DNA that structurally resemble ODNs. A panel of biological DNA samples was screened for DEC-205 binding. While none of the DNA samples were observed to bind DEC-205, some DNA samples were found to bind another receptor, RAGE, suggesting a role for RAGE as a detector of both pathogenic and self-DNA.

Most vaccines must be administered more than once, or “boosted”, to achieve optimal efficacy, and DC-targeted vaccines should be no exception. However, our data suggested that simply administering the same DC-targeting construct twice does not effectively boost the response. This was due to interference from the primary antibody response, which can cross-react with and neutralise a subsequently administered boosting construct. To overcome this issue, the efficacy of various heterologous prime-boost strategies designed to reduce the reactivity of the primary response against the boosting construct was assessed. Ultimately, a combination of anti-Clec9A and anti-XCR1 targeting constructs was found to induce the least cross-reactivity and strongest response after boosting.

These findings contribute to the development of better adjuvants and immunisation strategies that optimise the efficacy of DC-targeted vaccines. More broadly, they also highlight the value of understanding the underlying biological mechanisms that drive immune responses, which can then be applied to the rational design of more effective vaccines.

Declaration

This is to certify that:

(i) this thesis comprises only my original work towards the PhD except where indicated in the preface;

(ii) due acknowledgment has been made in the text to all other material used;

(iii) this thesis is fewer than 100,000 words in length exclusive of figures, tables, and references.

Jessica Li

iii

Preface

This work was conducted in the laboratories of Associate Professor Irina Caminschi at the Burnet Institute and Monash University and Professor William Heath at the University of Melbourne. This work was funded by grants from the National Health and Medical Research Council. My studies were supported by an Australian Postgraduate Award.

Some of the work presented in this thesis was collaborative, and the contribution of others is duly acknowledged in the text. The approximate proportion of work that was my original contribution is as follows:

Chapter 3: 95% Chapter 4: 70% Chapter 5: 100% Chapter 6: 100%

Acknowledgments

This PhD was only made possible through the support of many people around me, to whom I would like to express my sincerest gratitude. First and foremost, thank you to my supervisors, Irene Caminschi and Bill Heath, who took me under their wing as an Honours student, and guided me through years of ups and downs with utmost patience and dedication. I could not have asked for more inspiring and supportive mentors to foster my development both as a scientist and as a person.

To Ken Shortman, Mireille Lahoud, Meredith O’Keeffe, Jose Villadangos, Nicole La Gruta and Andrew Lew, thank you for providing valuable insight and advice on many occasions. Thank you to Mireille for providing many key reagents, and to Meredith for providing feedback on Chapter 3.

To the members of the Caminschi, Lahoud and O’Keeffe and Heath labs, thank you for assisting with both experimental and moral support. Having such friendly and open colleagues certainly made my days in the lab more enjoyable. In particular, thank you to Fatma Ahmet, for teaching me almost every experimental technique I know, taking care of every little matter to ensure the lab runs smoothly, and always being such a cheerful presence in the lab.

To my fellow students who shared the Honours and PhD journey with me, thank you for the good times and for your care and support during the tough times. I wish you all the best in your future endeavours. A special thank you to Ee Shan, Zen, Filipp and Ting, for being so good at distracting me from work, but mostly for being there for me when I needed it the most.

Finally, thank you to my parents, who have done everything possible to help me in any way they can. Your unconditional love and support has been the biggest source of encouragement and inspiration for me throughout this journey.

vii

viii

Abbreviations

Ab antibody ABTS 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) diammonium salt AF488 Alexa Fluor 488 AF647 Alexa Fluor 647 AF700 Alexa Fluor 700 Ag antigen APC allophycocyanin APC antigen presenting cell Batf3 basic leucine zipper ATF-like transcription factor 3 BDCA blood dendritic cell antigen bp base pairs BSA bovine serum albumin BSS balanced salt solution CD cluster of differentiation cDC conventional dendritic cell CFSE carboxyfluorescein succinimidyl ester CHO Chinese hamster ovary Clec9A C-type lectin domain family 9 member A CLR C-type lectin-like receptor CRD carbohydrate recognition domain CTL cytotoxic T lymphocyte CTLD C-type lectin-like domain DAMP damage-associated molecular pattern DC dendritic cell DEC DEC-205 DMSO dimethyl sulfoxide DNA deoxyribose nucleic acid ds double-stranded EDTA ethylenediaminetetraacetic acid ELISA enzyme-linked immunosorbent assay FCS fetal calf serum

ix FITC fluorescein isothiocyanate Foxp3 forkhead box P3 g gravitational acceleration constant g gram G gauge GM-CSF granulocyte-macrophage colony-stimulating factor

H2O water hDEC human DEC-205 hi high HIV human immunodeficiency virus HMGB1 high mobility group box 1 hr hour HRP horseradish peroxidase IgG immunoglobulin G IL interleukin ITAM immunoreceptor tyrosine-based activation motif ITIM immunoreceptor tyrosine- based inhibitory motif i.v. intravenous ICS intracellular cytokine staining IFN interferon Irf interferon regulatory factor kb kilobase kDa kilodalton kg kilogram L litre lo low LPS lipopolysaccharide M molar mAb monoclonal antibody mDEC mouse DEC-205 MFI mean fluorescence intensity mg milligram MHC I major histocompatibility complex class I MHC II major histocompatibility complex class I x

min minute ml millilitre mM millimolar MyD88 myeloid differentiation primary response 88 μF microfarad μg microgram μl microlitre μm micrometre μM micromolar NET neutrophil extracellular trap NF-κB nuclear factor kappa B NK cell natural killer cell ng nanogram nm nanometre nM nanomolar nmol nanomole O.D. optical density ODN oligodeoxynucleotide OVA ovalbumin PAMP pathogen-associated molecular pattern PBMC peripheral blood mononuclear cells PBS phosphate buffered saline pDC plasmacytoid dendritic cell PE phycoerythrin PI propidium iodide polyIC polyinosinic-polycytidylic acid PRR pattern recognition receptor RAGE receptor for advanced glycation end-products RNA ribonucleic acid rpm revolution per minute RT room temperature s.c. subcutaneous SEM standard error of the mean ss single-stranded

xi TCR T cell receptor Th T helper TLR Toll-like receptor TNF tumor necrosis factor Treg regulatory T cell TRIF TIR-domain-containing adapter-inducing interferon-β V volt v/v volume per volume w/v weight per volume WEHI Walter and Eliza Hall Institute ºC degrees Celsius

xii

Table of Contents

Abstract ...... i Declaration ...... iii Preface ...... v Acknowledgments ...... vii Abbreviations ...... ix Table of Contents ...... xiii List of Tables and Figures ...... xxi

Chapter 1: Introduction ...... 1 The immune system has two arms: innate & adaptive ...... 1 Dendritic cell maturation ...... 2 Antigen presentation on MHC II ...... 3 Antigen presentation on MHC I ...... 4 Activation of T cells ...... 5 T helper cells ...... 6 T follicular helper cells ...... 7 Regulatory T cells ...... 7 Cytotoxic T lymphocytes ...... 8 Activation of B cells ...... 8 Immunological memory ...... 9 Memory T cells ...... 10 Memory B cells ...... 10 Instruction of the adaptive immune response by innate immunity ...... 11 Dendritic cell subsets ...... 12 Resident DCs ...... 12

CD8+ DCs ...... 13

CD8- DCs ...... 14 Migratory DCs ...... 15 Human conventional DCs ...... 16 Unifying conventional DC subsets ...... 17 Plasmacytoid DCs ...... 18

xiii Inflammatory DCs ...... 19 Detecting danger ...... 19 Toll-like receptors ...... 20 RIG-I-like receptors ...... 21 NOD-like receptors ...... 22 cGAS-STING ...... 22 C-type lectin-like receptors ...... 23 PRR expression on DCs ...... 25 Dendritic cell vaccines ...... 26 Ex vivo pulsed DCs ...... 27 In vivo DC targeting ...... 28 Targeting DC receptors ...... 28 Targeting C-type lectin-like receptors ...... 29 DC subset-specific receptors ...... 29 Properties of targeted receptors ...... 30 DEC-205 targeting ...... 33 Clec9A targeting ...... 34 XCR1 targeting ...... 37 Requirement for adjuvant ...... 38 Choice of adjuvant ...... 40 Improving adjuvants by targeting DCs ...... 41 Optimising prime-boost with DC-targeted vaccines ...... 42 Aims ...... 43

Chapter 2: Materials and methods ...... 45 2.1. Media and buffers ...... 45 2.2. Antibodies ...... 46 2.3. Targeting constructs ...... 49 2.4. Recombinant proteins and peptides ...... 49 2.5. ODNs ...... 51 2.6. Biological DNA samples ...... 52 2.7. Other reagents ...... 53 2.8. Cell lines ...... 56 xiv

2.9. Mice ...... 57 2.10. Injections ...... 58 2.11. Instruments and software ...... 58 2.12. Statistical tests ...... 58 2.13. Cell labelling ...... 58 CFSE labelling ...... 58 PKH26 labelling ...... 59 2.14. Measuring T cell responses in vivo ...... 59 T cell enrichment ...... 59 OT-I and OT-II in vivo proliferation assay ...... 60 Tetramer staining ...... 61 In vivo CTL assay ...... 61 Intracellular cytokine staining (ICS) ...... 62 2.15. Tumour model ...... 63 B16 melanoma tumour model ...... 63 Measuring OVA expression of harvested tumours ...... 63 2.16. Staining primary DCs ...... 64 Splenic DC enrichment ...... 64 Staining DEC-205 and CD14 on DCs and macrophages ...... 64 Binding of targeting constructs to DCs and B cells ...... 65 2.17. Measuring ODN binding, uptake and stimulation ...... 65 Annealing double-stranded ODNs ...... 65 Binding of ODNs to DEC-205-CHO cells ...... 66 Binding of ODNs to DEC-205 by ELISA ...... 66 Stimulation of purified B cells in vitro ...... 67 Activation of DCs and B cells by ODNs in vivo ...... 67 Binding and uptake of ODN by peritoneal macrophages ...... 68 2.18. Measuring binding of biological DNA ...... 69 Binding of YOYO-1-labelled DNA to CHO cells ...... 69 DNA biotinylation ...... 69 Transfection of CHO cells with RAGE ...... 69 Binding of biotinylated DNA to DEC-205- or RAGE-expressing cells ...... 70

xv Binding of HMGB1, granulin and LL-37 to DNA by ELISA ...... 70 Fractionation of mouse serum ...... 70 2.19. Dead cell binding assays ...... 71 Freeze/thawing cells ...... 71 Dead cell binding to soluble protein ...... 71 Dead cell binding to cell-bound protein ...... 71 2.20. Other ELISA assays ...... 72 Measuring anti-OVA Ab titres by ELISA ...... 72 IL-6 ELISA ...... 72 IL-12 ELISA ...... 73 CD14 ELISA ...... 73 Measuring the presence of 10B4-OVA in serum by ELISA ...... 73 Measuring the anti-construct reactivity of Ab responses by ELISA ...... 74 2.21. Miscellaneous protocols ...... 74 Agarose gel electrophoresis ...... 74 Collection of plasma ...... 75 Collection of serum ...... 75

Chapter 3: Using DEC-205 to modulate CpG ODN function ...... 77 Abstract ...... 77

Introduction ...... 78 Classes of CpG oligonucleotides ...... 78 Use of CpG ODN adjuvants in humans ...... 81 TLR9 determines the potency of species-specific CpG ODNs ...... 83 A-ODNs and B-ODNs activate distinct downstream signalling pathways ...... 84 Surface receptors suggested to mediate the uptake of ODN ...... 86

Results ...... 91 3.1. Requirements for the binding of CpG ODNs to DEC-205 ...... 91 3.1.1. Examining the capacity of DEC-205 to bind phosphorothioated and phosphodiester ODNs ...... 91 3.1.2. Examining the capacity of DEC-205 to bind ss and ds ODNs ...... 92 3.1.3. ODN nucleotide bases preferred for DEC-205 binding ...... 94

xvi

3.1.4. Optimal ODN length for DEC-205 binding ...... 96 3.1.5. 14T is a DEC-205 binding motif ...... 99 3.1.6. Introduction of diester bonds disrupts DEC-205 binding of ODN ...... 101 3.2. The contribution of DEC-205 to the immunostimulatory effects of ODN ..... 102 3.2.1. Examining the capacity of 2006 and 21798 to bind DEC-205 ...... 102 3.2.2. The requirement of DEC-205 for the activation of purified B cells by 2006 or 21798 ...... 103 3.2.3. The contribution of DEC-205 to the stimulatory effects of ASO 518477 ..... 106 3.3. Utilising DEC-205 to design more potent CpG adjuvants ...... 114 3.3.1. Enhancing the DEC-205 binding of 21798 by addition of 14T ...... 114 3.3.2. Improved adjuvant activity of 14T-21798 ...... 115 3.3.3. Enhancing the DEC-205 binding of 21798 by addition of 2006 ...... 116 3.3.4. Improved adjuvant activity of 2006-21798 ...... 119 3.3.5. Comparison of 2006-21798 with standard CpG ODN adjuvants ...... 121

3.3.6. 2006-21798 does not compensate for inefficient CD8+ T cell responses induced in the absence of Clec9A-targeting of Ag ...... 124 3.3.7. The requirement of DEC-205 for the activation of purified B cells by 2006- 21798 ...... 126 3.3.8. The requirement of TLR9 for the induction of CTL by 2006-21798 ...... 128

Discussion ...... 133 Molecular properties of ODN determine their DEC-205 binding ability ...... 133 ODN properties required for efficient DEC-205 binding correlate with those required for effective immune stimulation ...... 134 Lack of DEC-205 binding may restrict the uptake of self-DNA ...... 135 Differential DEC-205 binding may control the divergent stimulatory effects of A- ODNs and B-ODNs ...... 135 2006 is dependent on DEC-205 for optimal stimulatory activity ...... 136 21798 does not require DEC-205 for stimulatory activity ...... 137 Non-CpG ASO 518477 exhibits some DEC-205-dependent stimulatory activity . 138 Enhanced DEC-205 binding improves the capacity of CpG ODNs to promote

CD8+ T cell responses ...... 139

xvii Enhanced DEC-205 binding of CpG ODNs has no effect on the stimulation of

CD4+ T cells or B cells ...... 142 Targeting DEC-205 to design more potent adjuvants ...... 143

Chapter 4: The contribution of CD14 to CpG ODN uptake and function ...... 145 Abstract ...... 145

Introduction ...... 146

Results ...... 148 4.1. Surface expression of CD14 and DEC-205 on DCs and macrophages ...... 148 4.2. The requirement of CD14 for the in vivo activation of DCs by CpG ODN .... 149 4.3. The requirement of CD14 for the induction of serum cytokines by CpG ODN ...... 151 4.4. Binding of CpG ODNs to CD14 by ELISA ...... 152 4.5. The requirement of CD14 for the binding and uptake of CpG ODN by peritoneal macrophages ...... 155 4.6. The requirement of CD14 for the binding of various classes of CpG ODN to peritoneal macrophages in the absence of DEC-205 ...... 157

Discussion ...... 159 CD14 does not play a role in the uptake of CpG ODNs for immune stimulation . 159 Lack of evidence for direct binding of CpG ODNs by CD14 ...... 160 In vivo stimulation by CpG ODN in CD14-deficient mice ...... 161 CD14 may mediate TLR9 activation, but not CpG ODN uptake ...... 162

Chapter 5: Investigating biological ligands of DEC-205 ...... 163 Abstract ...... 163

Introduction ...... 164 DEC-205 is an endocytic CLR ...... 164 DEC-205 in thymic development ...... 165 DEC-205 binds dead cells ...... 166 DEC-205 binds PAMPs ...... 168

Results ...... 170 5.1. The binding of biological DNA to DEC-205 and RAGE ...... 170

xviii

5.1.1. Binding of YOYO-1-labelled DNA to DEC-205-expressing CHO-K1 cells 170 5.1.2. Binding of biotin-labelled DNA to DEC-205- or RAGE-expressing CHO-K1 cells ...... 173 5.1.3. Binding of biotin-labelled DNA to DEC-205- or RAGE-expressing CHO-K1 cells in the presence of HMGB1, granulin and LL-37 ...... 176 5.1.4. Binding of HMGB1, granulin and LL-37 to DNA ...... 179 5.1.5. Binding of biotin-labelled DNA to DEC-205- or RAGE-expressing CHO-K1 cells in the presence of serum ...... 180 5.1.6. Binding of biotin-labelled DNA to DEC-205- or RAGE-expressing CHO-K1 cells in the presence of serum and DNA-binding cofactors ...... 183 5.1.7. Mouse serum contains a factor that enhances the binding of DNA to RAGE ...... 185 5.2. Binding properties of DEC-205 under acidic conditions ...... 188 5.2.1. Binding of ODNs to soluble DEC-205 in acidic conditions ...... 188 5.2.2. Binding of dead cells to soluble DEC-205, or Clec9A, in acidic conditions . 189 5.2.3. Binding of dead cells to cell surface-expressed DEC-205, or Clec9A, in acidic conditions ...... 191 5.2.4. Anti-DEC-205 mAbs can enhance the binding of ODNs to human DEC-205 ...... 193

Discussion ...... 195 RAGE binds various DNA of biological origin ...... 195 Serum contains a factor that promotes DNA binding to RAGE ...... 197 The effect of pH on binding of ligands to DEC-205 ...... 198

Chapter 6: Targeting DCs with heterologous prime-boost ...... 201 Abstract ...... 201

Introduction ...... 202 Targeting Ags to cDC1s to promote CTL responses ...... 202 Homologous prime-boost is not necessarily effective ...... 204 T cell competition interferes with boosting ...... 206 Anti-vector immunity interferes with boosting ...... 208 DC-targeted vaccines may encounter the same issues ...... 209

xix Results ...... 212 6.1. Comparison of the efficacy of Clec9A and XCR1 targeting ...... 212

6.1.1. Binding of Clec9A- and XCR1-targeting constructs to CD8+ DCs ...... 212 6.1.2. Capacity for Clec9A- and XCR1-targeting constructs to induce T cell responses ...... 213 6.2. Heterologous prime-boost strategies to boost immunity ...... 219

6.2.1. Homologous prime-boost with 10B4-OVA does not boost CD8+ T cell responses ...... 219 6.2.2. The primary Ab response interferes with boosting ...... 220 6.2.3. Heterologous prime-boost with targeting mAbs possessing different species backbones ...... 222 6.2.4. Heterologous prime-boost with targeting mAbs carrying different forms of Ag ...... 224 6.2.5. Heterologous prime-boost combining Clec9A- and XCR1-targeting ...... 227 6.2.6. Heterologous prime-boost combining Clec9A- and XCR1-targeting generates effective anti-tumour responses ...... 231

Discussion ...... 238 Targeting Clec9A via anti-Clec9A mAb is more effective for inducing T cell responses than targeting XCR1 via XCL1 construct ...... 238 Boosting is inhibited by the generation of a potent primary Ab response ...... 239 Heterologous prime-boost overcomes interference from primary Ab response ... 240 Properties of heterologous prime-boost using XCL1 and 10B4 constructs ...... 241 Factors influencing the efficacy of prime-boost ...... 243

Chapter 7: Final conclusions ...... 245

Bibliography ...... 251

List of Tables and Figures

Table 2.1. Conjugated antibodies...... 46 Table 2.2. Secondary antibodies...... 47 Table 2.3. Unconjugated antibodies...... 48 Table 2.4. ODN sequences used in this study...... 51

Figure 3.1. Binding of diester and thioated ODNs to DEC-205-expressing CHO-K1 cells...... 92 Figure 3.2. Binding of ss and ds ODNs to DEC-205-expressing CHO-K1 cells...... 93 Figure 3.3. ODNs with higher T content are preferred for DEC-205 binding...... 95 Figure 3.4. Optimal length of ODN for DEC-205 binding...... 99 Figure 3.5. Binding of 14T to mouse DEC-205...... 100 Figure 3.6. One diester bond disrupts DEC-205 binding...... 101 Figure 3.7. Binding of 2006 and 21798 to DEC-205...... 103 Figure 3.8. Stimulation of purified B cells by 2006, 21798 and 1668...... 105 Figure 3.9. Binding of ASO 518477 to mouse DEC-205...... 107 Figure 3.10. Activation of B cells 6 hours after ASO 518477 administration...... 109 Figure 3.11. Activation of B cells 24 hours after ASO 518477 administration...... 110 Figure 3.12. Activation of DCs 6 hours after ASO 518477 administration...... 111 Figure 3.13. Activation of DCs 24 hours after ASO 518477 administration...... 112 Figure 3.14. Expression of DEC-205 on DCs 6 and 24 hours after ASO 518477 administration...... 113 Figure 3.15. 14T-21798 binds DEC-205...... 114

Figure 3.16. 14T-21798 enhances Ag-specific CD8+ T cell responses...... 116 Figure 3.17. 2006-21798 binds DEC-205 and forms aggregates...... 118

Figure 3.18. 2006-21798 enhances Ag-specific CD8+ T cell responses and Ab responses...... 121 Figure 3.19. OT-I and OT-II proliferation in response to immunisation with anti- Clec9A-OVA plus various ODN adjuvants...... 124

Figure 3.20. CD8+ T cell responses induced in the presence of 2006-21798 are inefficient in the absence of Clec9A-targeting of Ag...... 125

xxi Figure 3.21. Stimulation of purified B cells by 2006-21798 and 2006 + 21798...... 127

Figure 3.22. 2006-21798 requires TLR9 to enhance Ag-specific CD8+ T cell responses...... 129 Figure 3.23. In vivo CTL activity induced by immunisation with anti-Clec9A-OVA plus ODN adjuvants requires TLR9...... 131 Figure 4.1. Expression of CD14 and DEC-205 on DCs and macrophages...... 148 Figure 4.2. Activation of DCs in vivo by CpG ODN does not require CD14...... 149 Figure 4.3. DEC-205, but not CD14, is required for optimal activation of DCs in vivo by CpG ODN...... 151 Figure 4.4. Production of serum cytokines in response to i.v. CpG ODN injection does not require CD14...... 152 Figure 4.5. Different recombinant CD14 constructs display inconsistent binding properties...... 154 Figure 4.6. The binding and uptake of CpG ODNs by peritoneal macrophages does not require CD14...... 156 Figure 4.7. The binding of A-, B- or C-class CpG ODN to peritoneal macrophages does not require CD14...... 158 Figure 5.1. Binding of CpG ODNs and DNA to DEC-205-expressing CHO-K1 cells...... 173 Figure 5.2. Binding of CpG ODNs and DNA to DEC-205- and RAGE-expressing CHO-K1 cells...... 175 Figure 5.3. Binding of DNA to DEC-205- and RAGE-expressing CHO-K1 cells in the presence of DNA-binding cofactors...... 178 Figure 5.4. Binding of HMGB1, granulin and LL-37 to DNA...... 180 Figure 5.5. Binding of DNA to DEC-205- and RAGE-expressing CHO-K1 cells in the presence of serum...... 182 Figure 5.6. Binding of DNA to DEC-205- and RAGE-expressing CHO-K1 cells in the presence of serum and DNA-binding cofactors...... 184 Figure 5.7. Binding of DNA to RAGE-expressing CHO-K1 cells in the presence of serum fractionated by size and treated with trypsin, EDTA or heat...... 187 Figure 5.8. Binding of CpG ODNs to DEC-205 is not enhanced by acidic pH...... 189

xxii

Figure 5.9. Clec9A, but not mouse or human DEC-205, binds to dead cells at pH 6 and pH 7...... 190 Figure 5.10. Labelled dead cells do not bind Clec9A-, mouse DEC-205-, or human DEC-205-expressing CHO-K1 cells...... 192 Figure 5.11. Certain anti-DEC-205 mAbs enhance the binding of CpG ODNs to DEC- 205...... 194 Figure 6.1. Binding of Clec9A- and XCR1-targeting constructs to DCs...... 213 Figure 6.2. OT-I and OT-II proliferation induced by immunisation with Clec9A- or XCR1-targeted OVA...... 215

Figure 6.3. CD8+ T cell responses induced by immunisation with XCR1-targeted or soluble OVA...... 216 Figure 6.4. OT-II proliferation induced by immunisation with XCR1-targeted or soluble OVA...... 217 Figure 6.5. Anti-OVA Ab responses induced by immunisation with Clec9A- or XCR1- targeted OVA...... 218

Figure 6.6. Prime-boost immunisation with 10B4-OVA fails to boost CD8+ T cell responses...... 220 Figure 6.7. Targeting construct is rapidly eliminated from the serum of pre-primed mice...... 221

Figure 6.8. Prime-boost immunisation with 10B4-OVA effectively boosts CD8+ T cell responses in μMT mice...... 222 Figure 6.9. Heterologous prime-boost immunisation with m10B4-OVA followed by

r10B4-OVA fails to boost CD8+ T cell responses...... 223 Figure 6.10. Heterologous prime-boost immunisation with r10B4-OBI followed by

m10B4-OVA fails to boost CD8+ T cell responses...... 225 Figure 6.11. Cross-reactivity of plasma from primed mice against targeting constructs...... 226 Figure 6.12. Heterologous prime-boost immunisation with r10B4-OBI and XCL1-

OVA effectively boosts CD8+ T cell responses...... 228 Figure 6.13. Cross-reactivity of plasma from primed mice against XCL1 or 10B4 targeting constructs...... 230

xxiii Figure 6.14. Prime-boost is no more effective than a single vaccination 6 days prior to tumour inoculation for preventing B16-OVA tumour growth...... 232 Figure 6.15. Prime-boost vaccination 14 days prior to tumour inoculation protects mice from B16-OVA tumour growth...... 235 Figure 6.16. Expression of OVA on the surface of tumours harvested at endpoint. .. 237

xxiv

Chapter 1: Introduction

The immune system has two arms: innate & adaptive

Our immune system, which serves to protect us from the enormous myriad of diseases one could encounter in a lifetime, is composed of a highly complex and vast network of molecules, cells and organs. The human immune system is organised into two arms, the innate and adaptive immune systems. The innate immune system is a frontline defence that reacts immediately to the presence of particular molecular patterns that signal infection or danger. Innate immune cells such as macrophages, neutrophils and dendritic cells are the first to detect and engulf pathogens, leading to their destruction via the release of toxic compounds, as well as the activation and recruitment of further immune cells and components (1).

However, the short-lived and non-specific innate response is often insufficient to completely clear a pathogen, requiring the support of the adaptive immune system. The main effectors of adaptive immunity are B and T lymphocytes, also known as B cells and T cells. Unlike innate cells, which use germline-encoded receptors to detect a restricted set of microbial components conserved in a broad range of pathogens, B cells and T cells possess receptors that undergo random mutation within each individual cell during development, such that each cell recognises a unique antigen (Ag). The total pool of cells is thus capable of recognising a highly diverse range of Ags (2). This extraordinary specificity allows the adaptive immune system to develop immunological memory against specific Ags, such that, upon subsequent exposure to the same Ag, the adaptive immune response is activated far more rapidly and potently, generating a secondary response that in many cases is able to eliminate the pathogen before it causes disease.

Although B cells and T cells are the main effectors of adaptive immunity, it is cells of the innate immune system that are responsible for the initial detection of foreign Ags or danger, and subsequently delivering Ags alongside the appropriate stimulatory

signals to initiate the activation of the adaptive immune response. One particular cell type is highly specialised for this function – the dendritic cell (DC). DCs possess many unique properties that are vital for the induction of immunity (3). Firstly, they are located ubiquitously throughout the body, continually surveying their environment for signs of danger and capturing potential Ag. Secondly, DCs are able to efficiently process captured Ags into peptides for loading on major histocompatibility complex (MHC) molecules, which can then be recognized by T cells. Finally, DCs also provide co-stimulatory molecules and soluble factors that are both necessary for T cell activation, and serve to direct the type of response generated based on the type of pathogen encountered. Thus, DCs can be considered the master controllers of adaptive immunity, and a key cell type to study for understanding how the adaptive immune system functions.

Dendritic cell maturation

Dendritic cells originate from progenitors in the bone marrow, from which they migrate to seed all organs of the body (4). In the steady state, DCs exhibit a resting or immature phenotype. In this state, they are less able to stimulate T cells, due to lower levels of MHC and co-stimulatory molecules, but are instead specialised for surveillance and Ag capture, as immature DCs have a greater capacity for the uptake of particles and microbes by phagocytosis, macropinocytosis and receptor-mediated endocytosis (4). While immature DCs synthesize large amounts of MHC II, it is predominantly sequestered in endosomal compartments, rather than on the cell surface (5, 6). In response to danger signals, such as microbial components or inflammation, DCs become activated, or mature. Mature DCs experience a transient upregulation followed by a rapid downregulation of phagocytosis and macropinocytosis (6, 7), although pinocytosis and receptor-mediated endocytosis remains intact (8-10). Mature DCs also drastically reduce the rate of turnover of MHC II on their surface, allowing the accumulation of large amounts MHC II complexes on their surface (6). These changes result in a reduced capacity to take up newly encountered Ags in favour of promoting Ag presentation. Surface expression of MHC

I also increases, but the rate of turnover appears unchanged from immature DCs (6, 11).

Importantly, maturation also results in the upregulation of co-stimulatory molecules, such as CD40, CD80 and CD86 on the DC surface (4). This is a key property of mature DCs that allows them to efficiently prime T cells. In the absence of such co- stimulatory molecules, such as on the surface of an immature DC, Ag presentation alone appears to induce tolerance. It has been proposed that DCs in their immature state promote tolerance against self Ags in the steady state, and only after activation, triggered by the presence of danger, are DCs appropriately equipped to induce T cell immunity (12). The caveat to this model is the observation that tolerogenic DCs are not necessarily always phenotypically immature and can display high levels of MHC and co-stimulatory molecules. Thus, the DC maturation state as assessed by surface markers cannot always be equated with functional DC maturation, causing some confusion in terminology (7, 13). For simplicity, here the term ‘mature DCs’ will refer to DCs that are both phenotypically and functionally mature.

Antigen presentation on MHC II

DCs possess specialised intracellular machinery that allows them to process captured Ag into peptides suitable for loading onto MHC molecules for presentation to T cells. There are two classes of MHC molecules, MHC I and MHC II, and correspondingly two pathways for Ag to be processed (14). MHC II molecules are loaded with peptides processed in the endolysosomal pathway. After internalisation, endocytosed Ag are degraded in increasingly acidic and proteolytically active vesicles, commonly referred to as early endosomes, late endosomes and lysosomes. Peptides of the appropriate length may be loaded onto MHC II molecules at any stage of the endosomal pathway, but most commonly in late endosomal or lysosomal compartments (14). MHC II molecules access these compartments by either direct translocation from the trans- Golgi network, or by endocytosis from the cell membrane (14, 15). Phagocytosed Ags can also enter the MHC II presentation pathway by fusion of the phagosome with the lysosome to form the phagolysosome. A similar process also occurs with

autophagosomes, which engulf parts of the cytosol, fusing with the lysosome, resulting in self Ag being loaded on MHC II (14, 15). Thus, MHC II molecules display both endogenous and exogenous Ag.

The constitutive expression of MHC II is restricted to only certain types of cells, known as antigen presenting cells (APCs). APCs include B cells, macrophages and DCs, and play an important role in activating CD4+ T cells through the presentation of Ag on MHC II, an event that is essential for generating adaptive immune responses, as will be discussed below. Amongst APCs, DCs are the most potent stimulators of immunity, being the only cell type capable of activating naïve T cells to induce a primary immune response (3, 16). T cell stimulation by macrophages and B cells appears to be effective only subsequent to priming by DCs, and thus seems to serve an ancillary role. In the case of B cells, this is because they only acquire the ability to stimulate T cells after themselves being licensed by interaction with activated CD4+ T cells, which requires previous stimulation by other APCs (17). Why DCs are more effective T cell stimulators than macrophages may be explained in part by differences in Ag uptake and processing. Both DCs and macrophages are efficient at capturing Ags, but the endosomal pathway in macrophages is much more destructive, rapidly degrading proteins completely into amino acids. In DCs, the slower degradation results in antigenic peptides being retained for longer in endolysomal compartments and a greater availability of peptides for presentation (18).

Antigen presentation on MHC I

Unlike MHC II, MHC I molecules are not restricted to APCs and are found on all nucleated cells. In most cells, MHC I molecules only present endogenous Ag, as the source of peptides for MHC I is the proteasome, which degrades proteins in the cytosol (14). These proteins can include any proteins produced by ribosomes in the cell, as well as defective or incompletely produced proteins known as defective ribosomal products. After proteasomal degradation, peptides are transported into the endoplasmic reticulum where they can be loaded on MHC I molecules. Therefore, in most cells, MHC I molecules display a continuous sample of the intracellular

environment of the cells. The display of Ags that are foreign, such as from an intracellular pathogen, or altered self-Ag from a tumour, signals CD8+ cytotoxic T cells to kill the affected cell. However, to become effective killers, CD8+ T cells must first be primed by an APC. This can occur without the need for the APC itself to be infected or transformed, by way of a process known as cross-presentation. Cross- presentation refers to the presentation of peptides on MHC I that are exogenous in origin, and is a function performed predominantly by DCs in vivo (19).

Two main pathways for exogenous Ag to enter the MHC I presentation pathway have been proposed (20). The first involves internalized Ags being exported to the cytosol, where they are degraded by the proteasome and are loaded on MHC I molecules in a similar manner as endogenous Ag. A second pathway has also been described in which Ags can be degraded in endosomal compartments, and then subsequently transferred to MHC I molecules within the endocytic system. The first, cytosolic pathway appears to be the dominant method employed to cross-present Ag by DCs (20, 21).

Activation of T cells

The main purpose of Ag presentation by mature DCs is to activate naïve T cells specific for that Ag. Naïve T cells circulate centrally in the blood and lymphatic systems, until they are activated by encounter with Ag-bearing DCs in secondary lymphoid organs, i.e. the lymph nodes or spleen. The widely accepted model of T cell activation involves the integration of three signals (22, 23). Signal 1 is the recognition of MHC-peptide complexes by the T cell receptor (TCR). CD8+ T cells only recognise Ags presented upon MHC I molecules, while CD4+ T cells only recognise MHC II molecules. Signal 2 is the engagement of co-stimulatory molecules on APCs with their receptors on the T cell. Classically, this refers to the engagement of CD28 on T cells by CD80 or CD86 on APCs, but can also include signalling through molecules such as 4- 1BB, inducible T cell costimulator (ICOS) and OX40 on T cells (24). Signal 3 refers to presence of cytokines produced by APCs or other innate immune cells, that augment and modulate the type of T cell response generated.

T helper cells

The activation of CD4+ T cells is a crucial event in the amplification of adaptive immune responses, because activated CD4+ T cells proliferate and differentiate into T helper cells that support the activation of effector CD8+ T cells and antibody- producing B cells (25). Depending on the cytokine environment, CD4+ T cells differentiate into one of several subsets of T helper cells, such as Th1, Th2 or Th17 cells, each subset being specialised in directing certain types of immune responses.

Th1 cells are induced by viral, bacterial and protozoan pathogens, and promote responses directed against intracellular pathogens, by inducing cytotoxic T lymphocyte (CTL) and classical macrophage responses to kill infected cells (25). Th1 differentiation is promoted by the cytokines interferon (IFN)γ, interleukin (IL)-2, IL- 12 and IL-18. Th1 cells are typically characterised by their production of IFNγ and IL- 2, but they are also known to produce tumour necrosis factor (TNF)α, lymphotoxin and granulocyte-macrophage colony-stimulating factor (GM-CSF) (26).

Th2 cells were originally thought to be a regulatory subset that dampened the inflammatory response induced by Th1 cells. While cytokines produced by Th2 cells do oppose the development of Th1 cells, and vice versa (27), it is now appreciated that Th2 cells promote a distinct immune phenotype of their own. Th2 responses are primarily directed against large extracellular helminth parasites (28), and exert many effects on non-haematopoietic cells, such as smooth muscle contraction and mucus production, that aid in expulsion of parasites from the gastrointestinal tract (25, 28). Th2 cell differentiation is driven by IL-4, and Th2 cells characteristically produce IL-4, IL-5 and IL-13 (26).

Th17 cells are pro-inflammatory, and have been reported to play a role in defence against extracellular bacteria and fungi. The differentiation of Th17 cells involves a range of cytokines including IL-23, IL-21, IL-6 and transforming growth factor (TGF)-β (25). Th17 cells are primarily known for their production of IL-17A and IL-

17F amongst other inflammatory cytokines such as IL-22, GM-CSF and TNFα, which promote the recruitment of neutrophils and other innate cells (25, 26). Recently, further subsets of Th cells have been described, such as IL-22-producing Th22 cells, which are related to Th17 cells, and IL-9-producing Th9 cells, which appear to be related to Th2 cells (26).

T follicular helper cells

Another subset of T helper cells specialises in supporting B cell differentiation and maturation in secondary lymphoid organs. T follicular helper (Tfh) cells provide signals, in the form of cytokines and cell-surface molecules such as IL-21, IL-4 and CD40L, that are necessary for the proliferation and maturation of B cells, (29, 30). Tfh are identified by their expression of various molecules including programmed cell death protein 1 (PD-1), ICOS, B cell lymphoma 6 (BCL6), B and T lymphocyte attenuator 4 (BTLA4) and the chemokine receptor CXCR5, which causes migration towards the B cell follicle, the area in secondary lymphoid organs where B cells reside (29, 30). Tfh cell differentiation is driven by various cytokines, including IL-6 and IL- 21 and surface ligands like ICOS, although Tfh cells show considerable heterogeneity in their phenotype and the signals required to generate them. This is hypothesised to reflect the need to induce B cell responses against a large variety of pathogens, in contrast to the pathogen-tailored response induced by Th1, Th2 or Th17 cells (29).

Regulatory T cells

CD4+ T cells can also differentiate into regulatory T (Treg) cells that play a distinct role in regulating immune responses to prevent autoimmunity. Treg cells are typically recognised by their constitutive high expression of CD25 and the transcription factor forkhead box P3 (Foxp3), and production of IL-10 and TGF-β, immunosuppressive cytokines that also drive the differentiation of Treg cells (31, 32). In contrast to the aforementioned Th classes, which differentiate upon activation by mature DCs, Treg cells are preferentially induced by immature DCs that do not express high levels of co-

stimulatory molecules (33). Treg cells also develop and mature in the thymus under steady state conditions, in which case they are referred to as natural Treg cells (26, 32).

Cytotoxic T lymphocytes

Activated CD8+ T cells differentiate into cytotoxic T lymphocytes (CTLs) which function to eliminate infected, cancerous, or otherwise damaged cells. CTL differentiation is promoted by several signals, including IL-12, IL-2, IL-21 and type I IFNs (34). Fully activated CTLs produce pro-inflammatory cytokines such as IFNγ and TNFα, and acquire the ability to kill cells displaying their cognate Ag on MHC I by releasing granules containing perforin and granzyme, or inducing Fas-mediated apoptosis (34, 35). CTLs have been shown to be essential for resolving many viral infections, and are also known to be key mediators of anti-tumour immunity (22, 34).

CD4+ T cell help has been reported to play a prominent role in promoting the induction of CTL responses (36-38), although the need for CD4+ T cell help can be bypassed by a high level of inflammatory signals induced by certain pathogens (39, 40). Whilst the requirement for CD4+ T cell help to induce primary CD8+ T cell responses may vary, it appears to be essential for the generation of memory CD8+ T cells. In the absence of CD4+ T cell help, CD8+ T cells can still acquire cytolytic effector function, but fail to form a memory population that can respond to restimulation (41-43). In contrast to CD4+ T cells, which appear to require continuous Ag recognition to support their expansion, CD8+ T cells can engage Ag on APCs for a short period of time, and then proliferate for multiple generations without further direct contact (44, 45).

Activation of B cells

B cells recognise Ag via their B cell receptor (BCR). Unlike the TCR, the BCR recognises conformational and linear epitopes that do not need to be presented on MHC. Thus the BCR can engage soluble Ag without the aid of an APC, although direct presentation by APCs to B cells does occur (46). However, similar to T cells, Ag

recognition alone is insufficient to generate long-lived mature B cells, which requires additional signalling in the form of help provided by activated CD4+ helper T cells, as discussed above (47). B cells contribute to the adaptive response primarily by differentiating into plasma cells that secrete large amounts of antibodies (Abs), also known as immunoglobulins (Ig), specific for their cognate Ag. In the absence of T cell help, B cells can only differentiate into short-lived plasma cells that secrete Ab that are of lower affinity. By contrast, a T-dependent B cell response involves the differentiation of high-affinity, long-lived plasma cells through a process known as the germinal centre reaction. During this process, B cells undergo class-switch recombination, to produce Ig of different classes with different functional properties, such as IgA or IgG, and affinity maturation to achieve higher specificity for their target (47). Secreted Abs function by binding their target Ag, causing neutralisation of pathogens and toxins, opsonisation to enhance phagocytosis, and destruction by effector cells such as natural killer (NK) cells or direct lysis through complement fixation (48).

Once activated, B cells can also play a role as APCs by presentation of their cognate Ag on MHC alongside co-stimulatory molecules, supporting prolonged Ag presentation to T cells in secondary lymphoid organs after the initial APCs, DCs, have already died (29, 49).

Immunological memory

Activation of T and B cells is marked by a rapid expansion of Ag-specific clones and differentiation to acquire specific effector functions. The peak of this effector response is typically around 7-15 days post-infection, after which point the response rapidly contracts, with effector cells undergoing mass apoptosis. This can be considered a negative feedback mechanism brought about by removal of danger signals and Ag by an effective response (22, 50, 51). However, a small portion of Ag-experienced cells will survive, as memory cells. Memory cells persist long-term in the host, potentially for many years, without the need for stimulation by Ag, and are homeostatically maintained through certain survival factors, such as IL-7 and IL-15 for T cells (52, 53)

and TNF superfamily members B cell activating factor (BAFF) and a proliferation- inducing ligand (APRIL) for B cells (54, 55).

Memory T cells

At the conclusion of a primary immune response, approximately 90-95% of effector T cells will die. The remaining memory T cells can be broadly categorised into three subsets – central memory (Tcm), effector memory (Tem) or tissue-resident memory (Trm) (22, 56). Tcm cells express CD62L and CCR7, receptors that mediate homing to lymph nodes via binding to their respective ligands, peripheral node addressins (PNAd) and the chemokines CCL19 and CCL21 (57). This confines Tcm cells to recirculation predominantly through the blood and secondary lymphoid organs (58). On the other hand, Tem cells express neither of these central homing markers and can be found in both lymphoid and non-lymphoid tissues (59). Tcm cells produce IL-2 and have greater proliferative capacity, while Tem cells are less proliferative but more readily produce effector cytokines such as IFNγ (56). The third subset of memory T cells, Trm cells, have only been characterised recently. Trm cells do not express CD62L or CCR7, but can be identified by expression of CD69 and CD103, although it has been observed that not all Trm cells express both markers (60, 61). The defining feature of Trm cells is the fact that, in contrast to other memory cells, they do not recirculate and remain permanently within a tissue niche, often at the original site of infection, acting as an early frontline defence against re-infection (60, 62, 63).

Memory B cells

Activation of B cells during an immune response also results in the generation of long-lived memory cells. There are two main components of B cell-mediated memory, the first being long-lived plasma cells that reside primarily in the bone marrow and continuously produce protective Ab for up to decades after the initial response (64, 65). These cells express little to no BCR on their surface and cannot be further stimulated, and thus are considered terminally differentiated (54, 64). The second form of memory B cells are found in secondary lymphoid organs and are quiescent

until restimulated by subsequent Ag encounter – in which case they rapidly divide and differentiate into Ab-secreting plasma cells (54, 64). Both types of memory B cells develop via the germinal centre reaction and therefore require T cell help during their initial activation, and produce Abs that have undergone class switching and are of high affinity (47, 54, 64). Some memory B cells can also be generated outside of the germinal centre, in which case they produce Abs that are of lower affinity for their target (54).

Memory cells are the basis of one of the fundamental features of adaptive immunity – the ability to respond more potently and rapidly to a pathogen that has been seen before, such that future occurrences of the same disease can be less severe or even entirely prevented. They are able to mount more effective responses than naïve lymphocytes for several reasons. For example, memory cells that are specific for a previously encountered Ag are present in much greater frequency than in the naïve population, and some subsets of memory cells are located in the periphery where they can encounter and respond to pathogens at an earlier stage of infection. Memory cells are also more sensitive to activation at lower doses of Ag and relatively less dependent on co-stimulation, and thus more readily transition from the resting state to fully- fledged effector cells that can eliminate the pathogen (50, 66, 67).

Instruction of the adaptive immune response by innate immunity

As illustrated above, the adaptive immune response is highly dynamic, such that assault by different types of pathogen can trigger the generation of different types of responses. Information about the type of pathogen must be transmitted from the innate cells that detect the pathogen and integrated into signals that direct the development of the adaptive immune response. This is the primary role of DCs, the cells that specialise in both pathogen detection and activation of adaptive immunity.

We have previously discussed how DCs are uniquely capable of initiating and regulating adaptive immune responses. But how are DCs capable of governing the type of adaptive immunity generated? This can be understood by considering certain

aspects of the biology of DCs. Firstly, DCs comprise a multitude of subsets with varying location and functional specialities, allowing for the regulation of different aspects of adaptive immunity by different subsets. Secondly, DCs possess an array of receptors that allow them to not only detect the presence of a pathogen, but also distinguish between different types of pathogens, and relay this information via molecular signals, such as the production of certain cytokines.

Dendritic cell subsets

The dendritic cell was first identified by Steinman and Cohn in the 1970s in mouse lymphoid organs as an adherent, nucleated cell with dendritic processes extending from the cytoplasm (68, 69). Since then, it has become clear that DCs are in fact highly heterogeneous, and are located throughout the body. Dendritic cells have been organised into various subsets based on differences in phenotypic markers, location, function, and ontogeny. A major division is between conventional DCs (cDCs), which describe the majority of DCs, and plasmacytoid DCs (pDCs), which will be described later (70). cDCs can be categorized as either resident or migratory populations, which can be further subdivided into several distinct subsets (71, 72). So far, a universal, specific surface marker for DCs has not been identified. DCs constitutively express MHC II and CD11c, but these markers are also expressed by some macrophages and other cells, and can be influenced by maturation state. Thus, positive identification of DCs generally requires a combination of surface markers including subset-specific markers (72).

Resident DCs

Resident DCs develop and reside within lymphoid tissues – the lymph nodes, spleen and thymus. Approximately 50% of DCs in lymph nodes and virtually all DCs found in the spleen and thymus are resident (13). In the steady state, resident DC are immature, and may remain so for their entire lifespan in the absence of infection or other activating signals. In response to danger signals that may arrive via the blood or

afferent lymph, resident DCs become activated and efficiently prime circulating T and B cells (13, 73).

There are two major classes of resident DCs, which differ in both their surface phenotype and functional specialisation. In the mouse, resident DCs can be divided based on the expression of CD8α, into CD8+ and CD8- subsets. Due to the expression of CD11b on CD8- but not CD8+ DCs, CD8- DCs may also be referred to as CD11b+ DCs. CD8- DCs may be further classified based on their expression of CD4 into CD4+ or CD8-CD4- double negative subsets, although most functional studies consider them as a single CD8- population (71, 74). CD8+ DCs also express high levels of DEC- 205 (CD205), CD24, C-type lectin domain family 9 member A (Clec9A) and XCR1, which are low or absent on CD8- DCs, but lack signal regulatory protein α (Sirpα), dendritic cell immunoreceptor 2 (DCIR2) and Dectin-1, which are expressed on CD8- DCs (75, 76).

CD8+ DCs

CD8+ DCs constitute approximately 20-40% of the resident DCs in spleen and lymph nodes, but are the dominant population in the thymus (70, 72). CD8+ DCs are primarily known for their superior ability at cross-presenting Ags for priming CD8+ T cell responses (77, 78). This appears to be one of their major functions in vivo, as mice specifically lacking the CD8+ DC subset are deficient in responses requiring cross- presentation (79). The greater capacity of CD8+ DCs to cross-present Ag was initially believed to be due to their superior ability to take up dead cell-associated Ags (80, 81). However, CD8+ and CD8- DCs have been shown to be equally capable of taking up both soluble and bead-associated Ags, and have a similar capacity for presentation of exogenous Ag on MHC II, while the same Ag is only cross-presented effectively by CD8+ DCs (78, 82). This suggested that enhanced cross-presentation by CD8+ DCs may instead be attributable to specialised intracellular processes that favour the cross- presentation pathway (13). For instance, CD8+ DCs express more MHC I machinery than CD8- DCs (83), limit the proteolytic activity of phagosomes, and have greater

capacity to export Ag from the endosome in the cytosol (84, 85), all of which may preferentially direct Ags into the cross-presentation pathway.

Besides their greater capacity to cross-present Ags, CD8+ DCs possess several other attributes that enhance their ability to prime CD8+ T cells, and in particular CTL responses. CD8+ DCs are the main source of IL-12, a key cytokine for polarising Th1 responses and generating CTLs (86-88). CD8+ DCs are also the only cells in the spleen to express XCR1, the receptor for XCL1, a chemokine that is produced by CD8+ T cells upon Ag recognition and is believed to facilitate the interaction of DCs with CD8+ T cells (89, 90). Overall, CD8+ DCs appear to be specialised for priming T cell responses against viral infection and intracellular pathogens.

CD8- DCs

Where CD8+ DCs excel at cross-presentation, CD8- DCs appear to be more proficient at MHC II presentation (78, 83). Under conditions of identical Ag uptake, CD8- DCs were found to be significantly more effective than CD8+ DCs at stimulating CD4+ T cells (82). Priming by CD8- DCs was shown to polarise Th2 type responses (88). CD8- DCs generally demonstrate poor cross-presentation ability, with many studies showing CD8- DCs to be inefficient at cross-presenting various soluble, cell- associated, or bead-bound Ags (91). However, CD8- DCs have been shown to cross- present immune complexes, a process that appears to be regulated by Fc receptors (77, 92). CD8- DCs have also been shown to cross-present Ags acquired upon uptake of Escherichia coli or Saccharomyces cerevisiae (81, 93) or in the presence of a saponin- based adjuvant (94). Thus it would appear that CD8- DCs can be induced to cross- present Ags in the presence of certain stimulatory components that are found in specific pathogens or adjuvants.

Migratory DCs

Migratory DCs refer to all DCs that can be found in peripheral, non-lymphoid organs. In the periphery, these cells are immature and dedicated to Ag capture and surveillance. In response to danger signals, migratory DCs traffic via afferent lymphatics to lymph nodes, where they exhibit a mature phenotype. This migration and maturation also occurs constitutively in the steady state, though at a lower rate (95, 96). In the lymph node, migratory DCs have been shown to transfer captured Ag to resident DC subsets (97), in addition to directly presenting Ag to T cells themselves (98). Migratory DCs comprise approximately 50% of DCs in the lymph node, where, in the steady state, they can be distinguished from resident DCs by their mature phenotype (13, 99).

Migratory DCs are commonly studied in the skin, where there exist three main subsets - Langerhans cells, CD103+ DCs, and CD11b+ DCs. Langerhans cells reside in the epidermis and express high levels of langerin (CD207) and DEC-205 (CD205), and are epithelial cell adhesion molecule (EpCAM)hi, CD11bint and CD103- (71, 76). Langerhans cells have been shown to play a role in priming Th17 type immune responses against skin infections such as Candida albicans, contact hypersensitivity, as well as an immunoregulatory role (76, 100).

The other two major subsets of migratory DCs are referred to as dermal DCs in the skin, due to their location in the dermis, or interstitial DCs in other organs (76). CD103+ DCs express langerin, DEC-205, Clec9A and XCR1, but lack CD8 and CD11b (76, 101). Functionally, CD103+ DCs resemble CD8+ resident DCs, being highly proficient at cross-presentation, promoting Th1 type responses and producing IL-12 (76). On the other hand, CD11b+ interstitial DCs, which lack langerin but express DEC-205 and Sirpα, correspond to the CD8- population of resident DCs, being more proficient at presentation to CD4+ T cells and inducing Th2 type responses and Tregs (76, 101). It should be noted that studies of CD11b+ DCs in peripheral organs have often been obscured by contamination with CD11b+ macrophages, and there is evidence to suggest that CD11b+ DCs are themselves a

heterogeneous population with a subset deriving from monocyte precursors rather than pre-DCs (76, 102). The division of migratory DCs into CD103+ and CD11b+ subsets has been documented in most non-lymphoid organs, although there is some variation in surface markers in different locations (102).

Human conventional DCs

While all cDCs in humans constitutively express MHC II and CD11c, as in the mouse, many of the markers commonly used to delineate DC subsets in the mouse, such as CD8 and CD103, are not useful for identifying the same subsets in humans. Instead, human cDCs have been separated into two subsets by contrasting expression of blood dendritic cell antigen 1 (BDCA1) (CD1c) and BDCA3 (CD141), which can be found in both lymphoid and non-lymphoid tissues (103, 104). Through transcriptomic and ontogenic analysis, it is now generally accepted that BDCA1+ human DCs correspond to CD11b+ DCs in the mouse, and BDCA3+ DCs are the human equivalents of CD8+ or CD103+ DCs in the mouse (103-106). The BDCA3+ subset in humans has also been shown to be functionally homologous to mouse CD8+ or CD103+ DCs, being highly efficient at cross-presentation (106-110). Interestingly, several studies have found human BDCA1+ DCs to also be capable of cross-presentation, suggesting this function may not be restricted to one subset as it is in the mouse (90, 108, 109, 111- 113).

Another population of DCs has been observed in human non-lymphoid tissues that is CD14+BDCA1-BDCA3- and is poor at cross-presentation, but can induce follicular T helper cells (103, 104). These CD14+ DCs have properties similar to monocytes, and potentially correspond to the monocyte-derived subpopulation of mouse CD11b+ DCs. Human skin also contains langerin+ Langerhans cells, similar to those found in the mouse (103, 104).

Unifying conventional DC subsets

In the mouse, CD8+ and CD103+ DCs are functionally and phenotypically very similar, despite their different anatomical locations and life cycles. The discovery that the transcription factors basic leucine zipper ATF-like transcription factor 3 (Batf3), inhibitor of DNA binding 2 (Id2) and interferon regulatory factor 8 (Irf8) specifically control the development of CD8+ resident and CD103+ migratory DCs, but do not influence CD11b+ resident or migratory DCs, provided further evidence that these subsets were fundamentally related (79, 102, 114, 115). The development of CD11b+ DCs is regulated by a distinct set of transcription factors that includes Irf4, RelB and PU.1 (107, 116-118). This led to the view that all cDCs could be categorised into two main subsets based on a common developmental origin and broadly similar functional characteristics (71, 119). It has been proposed that these subsets be named cDC1 and cDC2, for the Batf3-dependent and Irf4-dependent populations, respectively, avoiding the ambiguity associated with nomenclature based on surface markers (119). This cDC1 and cDC2 nomenclature will be adopted throughout the rest of this thesis.

While defining DC subsets based on ontogeny may be more logical and consistent, transcription factors are not particularly practical for identifying DC subsets experimentally, especially if the cells are to remain alive. Thus, there is still a need for surface markers that can specifically identify each subset. Ideally, these markers would also be able to identify homologous subsets across species. Although no markers that are unique to the cDC2 subset have been identified so far, Clec9A and XCR1 have been proposed to fulfil this need for the cDC1 subset. Clec9A is expressed at high levels by cDC1s in both human and mouse, but is also expressed at low levels by pDCs in the mouse (120-122). XCR1 is the preferred marker for designating this subset, as it appears to be exclusive to cDC1s in both human and mouse (123, 124).

Plasmacytoid DCs

Plasmacytoid DCs (pDCs) are phenotypically, ontogenically, and functionally quite distinct from cDCs. In the steady state, pDCs display a rounded morphology similar to plasma cells, from which they derive their name, and only adopt the classical dendritic morphology upon activation. pDCs reside predominantly in lymphoid organs, but can also be found in the blood and peripheral tissues (125, 126). In the mouse, pDCs are commonly identified by markers such as CD45R (B220), lymphocyte antigen 6C (Ly6C), bone marrow stromal cell antigen 2 (BST2) and sialic acid binding Ig-like lectin H (Siglec-H), and an intermediate expression of CD11c. In humans, pDCs express BDCA2, ILT7 and CD123, but are CD11c- (125, 126). In both humans and mice, the transcription factor E2-2 drives pDC development (125, 127).

The primary function of pDCs appears to be the rapid secretion of large amounts of type I IFNs in response to viral nucleic acids. The role of pDCs as APCs has been controversial, although it is generally accepted that activated pDCs do have some capacity to present Ag to T cells (125, 128). This is facilitated by the upregulation of MHC II and co-stimulatory molecule expression by pDCs upon activation. However, MHC II and co-stimulatory molecule expression is much lower on pDCs than on cDCs, and accordingly, pDCs appear to be significantly less efficient than cDCs at stimulating T cells (126, 128). Similar to cDCs, Ag presentation by pDCs can result in either T cell activation or tolerance, depending on their activation state and the context in which they were activated (126, 128). Studies in mice have indicated that pDCs have very limited ability to cross-present Ags (91, 128-131), although this has been disputed by a report suggesting that mouse pDCs can acquire the capacity to cross-present Ag upon activation (132). By contrast, effective cross-presentation of various Ag by human pDCs, even without the addition of activating factors, has been demonstrated (133-136).

Inflammatory DCs

Another class of DCs known as inflammatory DCs, or monocyte-derived DCs, appears transiently in response to inflammatory stimuli. First derived in vitro from the culture of human blood with GM-CSF and IL-4 (137, 138), in vivo inflammatory DCs differentiate from Ly6Chi monocytes in the mouse, or CD14+ equivalents in humans, at the site of inflammation (103, 139). They migrate to the LN, present Ag to activate T cells, and produce various cytokines depending on the original inflammatory stimulus (139, 140). Some studies have indicated that inflammatory DCs also have the capacity to cross-present Ags in vitro, although the contribution of this process to responses in vivo is yet to be determined (20, 21, 91, 141, 142).

Detecting danger

The ability of DCs to initiate immune responses is coupled to their ability to detect danger. The innate immune system, including DCs, has evolved to recognise molecular patterns that are either components of pathogenic agents, or induced by the presence of such agents, and are not present in the body or not revealed under normal steady-state conditions. These ‘danger signals’ are known as pathogen-associated molecular patterns (PAMPs) and damage-associated molecular patterns (DAMPs). PAMPs are typically microbial components that are essential for survival and thus are conserved over entire classes of pathogens, but are not found in the host. On the other hand, DAMPs may be molecules such as DNA that can be found in the host, but are only exposed by tissue damage and necrosis, which often occurs in conjunction with pathogen invasion. The receptors that recognise PAMPs and DAMPs are pattern recognition receptors (PRRs), germline-encoded receptors that are expressed by various innate immune cells.

Several distinct classes of PRRs have been identified, which are typically either transmembrane or cytosolic receptors. A third category of secreted PRRs also exists, although, as they are not cell-associated, they do not directly mediate the activation of

innate and adaptive cells, and their primary function appears to be promoting the binding of other receptors and cells by opsonisation (143).

Toll-like receptors

The prototypical and most widely studied class of PRRs are the Toll-like receptors (TLRs), a family of receptors originally identified by homology with the Toll protein in Drosophila melanogaster (144, 145). TLRs are type I transmembrane proteins comprising an ectodomain with leucine rich repeats that mediates PAMP recognition, a transmembrane domain, and a Toll-interleukin 1 receptor (TIR) cytoplasmic domain that mediates downstream signalling. There are currently 10 and 12 known functional members of the TLR family in humans and mice, respectively. Both species express TLR1-TLR9, but TLR10 is only functional in humans and TLR11 only in mice. TLR12 and TLR13 are present in mice but not in humans (146).

TLR1, TLR2, TLR4, TLR5, TLR6 and TLR10 are expressed on the cell surface, where they recognize PAMPs that are generally external components of microbes, such as lipopolysaccharide (LPS), a component of Gram-negative bacterial cell walls recognised by TLR4, or flagellin, the major component of bacterial flagella recognised by TLR5. TLR2 forms heterodimers with TLR1 or TLR6, and binds to a range of lipoproteins found in bacteria, fungi, parasites and viruses (146, 147). In humans, TLR10 also cooperates with TLR2 to recognise a range of microbial components that overlap with those recognised by TLR1 (148). In mice, TLR11 and TLR12 have been shown to bind profilin from Toxoplasma gondii (149, 150), while TLR13 recognises bacterial 23S ribosomal RNA (151).

TLR3, TLR7, TLR8 and TLR9 are the endosomal TLRs, all of which are located on the membrane of intracellular vesicles and recognise nucleic acids (146). TLR3 recognises double-stranded RNA (dsRNA) while TLR7 and TLR8 recognise single-stranded RNA (ssRNA). These TLRs serve to detect infection by dsRNA or ssRNA viruses, respectively. TLR3 can also be activated by the synthetic dsRNA analog, polyinosinic- polycytidylic acid (polyIC), while TLR7 and TLR8 can be activated by

imidazoquinoline derivates such as imiquimod or resiquimod (R848) (146). TLR9 recognises a specific motif in DNA, unmethylated 2’-deoxyribo(cytidine-phosphate- guanosine) (CpG), which is frequently found in viral or bacterial genomes, but is rare in mammalian DNA. TLR9 can also be activated by synthetic CpG oligonucleotides (ODNs) (146).

Activation of TLRs generally leads to the production of inflammatory cytokines and type I IFNs, with the distinct activation pathways of each TLR dependent upon which adaptor molecules are recruited. All TLRs, with the exception of TLR3, are known to recruit the adaptor protein myeloid differentiation primary response 88 (MyD88), which leads to the production of inflammatory cytokines via the activation of the transcription factor nuclear factor kappa B (NF-κB) and mitogen-activated protein kinases. TLR3 and TLR4 can utilise an alternate pathway involving the adaptor molecule TIR-domain-containing adapter-inducing interferon-β (TRIF), which results in the activation of Irf3 and NF-κB, and production of IFN as well as inflammatory cytokines (146, 152). In pDCs, the recruitment of MyD88 downstream of TLR7 and TLR9 signalling also leads to the production of type I IFNs via the activation of IRF7.

RIG-I-like receptors

Besides the endosomal TLRs, several other families of receptors are known to act as intracellular PRRs. Retinoic acid-inducible gene I (RIG-I), and the related receptors melanoma differentiation associated factor 5 (MDA5) and laboratory of genetics and physiology 2 (LGP2), form the RIG-I-like receptor (RLR) family of PRRs. RLRs detect the presence of dsRNA from RNA viruses in the cytoplasm. RIG-I detects short dsRNA or polyIC less than 1 kb in length, while MDA-5 detects longer dsRNA and polyIC. Both induce the production of type I IFN and pro-inflammatory cytokines via an N-terminal caspase activation and recruitment domain (CARD). By contrast, LGP2 lacks a CARD and cannot induce signalling on its own, and appears to be a positive regulator of the activity of both RIG-I and MDA5 (152, 153).

NOD-like receptors

The other main family of cytosolic PRRs is the NOD-like receptors (NLR). NLRs number 23 members in human and 34 in mice (152). The most well-characterised NLRs are nucleotide-binding oligomerization domain-containing protein (NOD)1 and NOD2, which recognise peptidoglycans in bacterial cell walls and therefore detect the presence of intracellular bacteria in the cytoplasm. Detection of PAMPs triggers the oligomerization of these receptors, which leads to the activation of NF-κB and MAP kinases to induce inflammatory cytokine production (152, 153).

Other members of the NLR family, such as NLR family pyrin domain containing (NLRP)1, NLRP3, NLR family CARD containing (NLRC)4 and NLR family apoptosis inhibitory protein (NAIP)5, are known to form components of the inflammasome, a multiprotein complex that is assembled in response to a variety of PAMPs, DAMPs and environmental pollutants such as silica or asbestos (152, 153). The inflammasome proteolytically activates caspase-1, which in turn cleaves inactive forms of IL-1 family cytokines into their active forms, such as IL-1β and IL-18 (152, 153). Inflammasome activation can also lead to caspase-1-dependent inflammatory cell death, or pyroptosis (154).

cGAS-STING

Cyclic GMP-AMP synthase (cGAS) and stimulator of IFN genes (STING) were recently discovered to form another intracellular PRR system that detects pathogenic DNA. STING is located in the endoplasmic reticulum and was initially shown to respond to cyclic dinucleotides that are produced by bacteria as second messengers (155). STING was subsequently shown to be a downstream adaptor for cytosolic DNA sensors, most notably for cGAS. cGAS binds to DNA and catalyses the production of cyclic GMP-AMP (cGAMP), which is structurally similar to bacterial cyclic dinucleotides. Activation of STING by cyclic dinucleotides or cGAMP leads to the production of IFN-β and related cytokines in an Irf3-dependent manner (156-158). Other cytosolic DNA sensors have been proposed to activate STING signalling,

including DNA-dependent activator of IFN-regulatory factors (DAI), gamma- interferon inducible protein 16 (IFI16) and DDX41, although, unlike cGAS, they do not appear to be essential for STING activation in response to DNA (157, 158).

C-type lectin-like receptors

Although TLRs are more prominently discussed, another group of receptors, the C- type lectin-like receptor (CLR) family, are also known to act as PRRs on the cell surface. Members of the CLR family are characterized by the presence of one or more carbohydrate recognition domains (CRDs) or the structurally similar C-type lectin- like domain (CTLD), which does not necessarily bind carbohydrates (159). CLRs are a large superfamily of over 1000 proteins that have been categorized, based on structural and phylogenetic homology, into 17 groups (159). Only certain CLRs have been identified to be expressed on innate cells and have PRR function. These fall into three groups: Group II, containing a single CRD with Ca2+ and carbohydrate binding capacity, Group V, containing a single CTLD without classical Ca2+ and carbohydrate binding motifs, and Group VI, containing 8-10 CTLDs (160, 161). In addition to this grouping, receptors can also be categorized according to functionality based on the intracellular signalling motifs they possess (160, 162). CLRs can be activating, coupled to Syk signalling through an immunoreceptor tyrosine-based activation motif (ITAM) or hemITAM motif, or antagonise Syk signalling via an immunoreceptor tyrosine- based inhibitory motif (ITIM). CLRs may also have neither ITAM nor ITIM motifs, and instead engage endocytic machinery to mediate ligand uptake and intracellular trafficking without affecting cell activation (160, 163, 164).

CLRs were initially described to have a major role in anti-fungal defence, as the first activating CLR to be discovered and characterised was Dectin-1. Dectin-1 is a group V CLR with a hemITAM motif and recognises β-1,3 glucans which are present in many fungal cell walls, as well as in some plants and bacteria (165). Many other CLRs, such as Dectin-2, dendritic cell-specific intercellular adhesion molecule-grabbing nonintegrin (DC-SIGN) and macrophage inducible Ca2+-dependent lectin (Mincle), also recognise fungi, by binding to various mannan residues that are a component of

fungal cell walls (166). Dectin-2 and Mincle both induce activatory signals via an ITAM motif, while DC-SIGN lacks an ITAM or ITIM motif and instead signals through the Raf-1 kinase pathway to modulate other signalling pathways, including TLR-activated pathways (164, 166). Dectin-1, Dectin-2, and Mincle can also recognise mycobacteria, while DC-SIGN, which recognises mannose and fucose residues, has been shown to recognise mycobacteria along with a variety of other species of bacteria (161, 164, 167). DC-SIGN and Dectin-2 have also been shown to detect glycan residues on helminth pathogens (161). DC-SIGN is also well known for its ability to bind human immunodeficiency virus (HIV), as well as a range of other viruses including Measles and Dengue virus. However, this is suggested to be an adaption by the virus to exploit the host immune system to promote the dissemination of the virus (161).

DCIR (homologous to Dcir1 and Dcir2 in the mouse), Clec12A (also known as MICL or DCAL-2) and macrophage antigen H (also known as Clec12B) are examples of CLRs that contain an inhibitory ITIM motif (160, 163). Although the ligands for some of these receptors have not been well characterised, cross-linking of these ITIM- coupled receptors with antibodies has been shown to inhibit the effects of signalling downstream of activating CLRs and TLRs (160, 163). Other CLRs lack both ITAM and ITIM motifs, such as DC-SIGN, mannose receptor, langerin and DEC-205. These CLRs appear to be primarily involved in mediating endocytosis, and directing Ag to the appropriate compartments to facilitate Ag presentation, although DC-SIGN has been shown to also modulate the signalling pathways of other PRRs (160).

It is understood that many CLRs bind to self-ligands, and may have roles outside of pathogen recognition in mediating homeostatic processes such as adhesion, migration and intercellular communication (160). However, recently it has become clear that the recognition of self-ligands by CLRs can also play a role in immunity, as some CLRs act as DAMP receptors (162). For instance, Mincle recognises SAP130, a molecule released by dead cells, and induces the production of pro-inflammatory cytokines via an ITAM motif (168). Clec9A detects F-actin, which is exposed on necrotic cells, but does not appear to induce inflammation despite containing a hemITAM motif (169,

170). Instead, Clec9A has been shown to promote cross-presentation of dead cell- associated Ags (171-173). DEC-205 has recently been reported to bind keratin exposed by dead cells, but the downstream consequences of such an interaction have not been investigated (174, 175). Clec12A has also been recently identified to detect uric acid crystals released by dead cells, but as it possesses an ITIM motif, it appears to have a role in negatively regulating the inflammation induced by cell death (176).

CLRs have also been shown to recognise tumour-associated Ags. These Ags are often self-Ags that become dyregulated and display alterations to their glycosylation that can be detected by CLRs (177, 178). For example, DC-SIGN has been shown to recognise the aberrantly glycosylated self-Ag carcinoembryonic antigen (CEA) present on tumour cells, but not the normal form found on non-tumour cells (179). Although such recognition of tumours by CLR can mediate anti-tumour responses, this interaction can also promote tumour growth and immune evasion by inhibition of inflammatory signalling via modulating receptors such as DC-SIGN, mannose receptor and BDCA2 (180).

PRR expression on DCs

As discussed previously, the ability of the adaptive immune response to generate a particular type of response against different types of pathogens, hinges upon the instruction of APCs and other innate immune cells, which produce certain cytokine signatures to influence the development of the adaptive response. This in turn is triggered by the engagement of PRRs that can discriminate specific types of pathogens. Thus, the receptors expressed by innate cells can be considered a key determinant of their capacity to stimulate adaptive immunity.

This is illustrated by the fact that different subsets of DCs, which are known to preferentially promote certain types of responses, have distinct sets of PRRs. For example, CD8+ and CD103+ DCs in the mouse, collectively referred to as cDC1s, are known for their superior capacity to induce Th1 and CTL responses, and are the main subsets to express TLR3 (181). TLR3 agonists such as polyIC are known to stimulate

production of type I IFN and IL-12, key Th1-promoting cytokines, and promote cross-priming of viral Ags (182-184). cDC1s are also the main subset to express Clec9A, a CLR that promotes the cross-presentation of dead-cell associated Ag for priming of CTL responses (120-122). Similarly, pDCs are specialized for the production of type I IFN, a key anti-viral response, and express the highest levels of the viral detectors TLR7 and TLR9 amongst all DC subsets in mice (185). In humans, the paucity of cells able to be isolated from primary samples has hindered the investigation of specific subsets of DCs. However, it is known that TLR7 and TLR9 expression in humans is even more restricted to only the pDC subset, suggesting there may be a more discrete division of labour (186, 187).

On the other hand, cytosolic sensors such as RLRs and STING tend not to be restricted to certain cell types and are widely expressed by many cells, including non- immune cells. This could reflect the fact that any cell in the body could be susceptible to intracellular infection or damage, and thus would need a mechanism to detect this intrinsically and alert the immune system (188).

Dendritic cell vaccines

The key role of DCs in both initiating and instructing adaptive immune responses has naturally led to consideration of how they can be used therapeutically. Much work has been done to investigate how DCs can be used to engage the adaptive immune response to protect against disease, particularly those against which current vaccines have little effect, such as many chronic infectious diseases or cancer. Classically, vaccination involves the administration of an Ag against which an immune response is desired, often with the addition of stimulating agents known as adjuvants, which engage PRRs and induce inflammation and DC maturation, mimicking the environment in which pathogenic Ag are normally encountered. This is intended to induce an adaptive immune response against the Ag of interest, and subsequently protective memory, without the need for natural infection and disease. However, when a vaccine is administered systemically, only a small part of the dose may be captured by DCs and reach secondary lymphoid organs where T cells are primed. A

more efficient form of vaccination would be to have the Ag of interest introduced directly to DCs, ensuring the Ag is concentrated to the cells that are the most potent stimulators of adaptive immunity.

Ex vivo pulsed DCs

The initial approach of using DCs for immunotherapy involved ex vivo propagation of DCs, which were utilised to prime anti-tumour immunity. Due to the scarcity of DCs in the blood, CD34+ DC precursors or monocytes were isolated from patient blood and expanded and matured in culture in the presence of cytokines such GM-CSF and IL-4, although protocols vary between different studies (189, 190). Some studies have also directly used blood DCs, although this required repeated leukapheresis to obtain enough cells. These DCs were then loaded with Ag ex vivo and readministered to patients (189).

Over 200 clinical trials have now been conducted with ex vivo pulsed DCs cultured under various conditions and with various Ag, and although some clinical responses have been observed, overall objective response rates have been disappointingly low (191). This may be due to the recruitment of late-stage cancer patients who are likely to be in an immunosuppressed state, limiting the efficacy of immunotherapeutic strategies (190). Despite this, the first and so far only DC vaccine, Sipuleucel-T (Provenge), was approved in 2010, for the treatment of metastatic castration-resistant prostate cancer. The Sipuleucel-T vaccine involves the incubation of autologous peripheral blood mononuclear cells (PBMCs) with a fusion protein of a prostate antigen and GM-CSF, which are then re-injected into patients. This therapy increased median overall survival of patients by approximately 4 months (192-195). However, since its approval, Sipuleucel-T has not become readily available, partly due to complications surrounding its complex production and high cost (196).

In vivo DC targeting

Due to the limitations of therapies that involve pulsing DCs ex vivo with Ag, methods of targeting Ag to DC in situ have been developed. The basic premise is that Ags of interest can be delivered to DCs in vivo by attachment to a targeting construct – typically a monoclonal Ab (mAb) or a natural ligand that homes specifically to a surface receptor expressed on DCs. Ag can be attached to targeting molecules via chemical conjugation or genetic fusion, although the latter is generally preferred for higher quality and consistency. The production of such targeting constructs is far less expensive, laborious and time-consuming than ex vivo culture of DCs, and one product can be used universally, instead of being individually tailored, allowing for large-scale production (197).

In vivo targeting also has immunological advantages. Reintroduction of ex vivo pulsed DCs via intradermal injection was reported to result in only a small fraction, 2%, reaching the lymph nodes (198). Additionally, the DCs generated by culture of monocytes or CD34+ precursors with cytokines in vitro may not necessarily possess the same capacity to prime T cells as DCs that develop in situ. One of the only studies to directly compare the two methods found administration of DCs pulsed ex vivo with OVA to be far less effective at inducing OVA-specific T cell responses than immunisation with OVA targeted to DCs in vivo (199). However, the specificity of targeting has been raised as a potential issue. While using ex vivo pulsed DCs ensures that only activated DCs carry the Ag, most DC receptors are not exclusively expressed by DCs, and thus targeting these receptors can result in other cells taking up the Ag, which could affect the outcome (197).

Targeting DC receptors

A variety of receptors have been used to successfully induce adaptive immune responses by in vivo targeting, including MHC II (200, 201), CD11c (202-204), mannose receptor (205, 206), CD36 (207), and LOX-1 (208). Although all of these receptors are expressed by DC, they are also expressed by a variety of other cells,

meaning that mAb against these receptors are actually targeting Ag to multiple cell types. This may be less than ideal, as other cell types would not be as efficient as DCs at priming T cells, and may not even be APCs. Incidental Ag uptake by non-DC may even have inhibitory effects – for example, it has been shown that Ag presentation by B cells can result in tolerance (209). Several studies have shown that, even when targeting a receptor present on multiple cell types, it is the DCs that are essential for inducing immunity, as depletion of DCs resulted in severe abrogation of the response (206, 210, 211). Thus, it can be expected that targeting to receptors that are more specific to DCs would result in greater immune stimulation.

Targeting C-type lectin-like receptors

Recently, receptors of the CLR family have attracted greater attention as Ag delivery targets. Many CLRs are endocytic receptors, which is a significant advantage for targeting Ag, as they are naturally efficient at internalising bound ligands and directing them to Ag presentation pathways. Several CLRs also induce downstream signalling upon binding, which could be used to augment immune activity (160, 163). Importantly, certain CLRs exhibit expression patterns that are relatively restricted to DCs, and even certain subsets of DCs. This is advantageous for ensuring that Ag is delivered to DCs more specifically, and also allows examination of the effects of targeting Ag to specific DC subsets.

DC subset-specific receptors

As the functional specialisations of different DC subsets became clear, the possibility of targeting specific DC subsets to influence the type of adaptive immune response generated was proposed. The cDC1 subset has been shown to be more effective at stimulating CD8+ T cells via Ag presentation on MHC I, while the cDC2 subset appears to be more effective at Ag presentation to CD4+ T cells via MHC II (78, 83). In accordance with this trend, targeting Ag to the receptor DEC-205, which is predominantly expressed on cDC1s, was found to be more efficient at inducing Ag- specific CD8+ T cell responses, while targeting receptors only expressed on cDC2s,

such as DCIR-2 or Dectin-1, preferentially promoted Ag-specific CD4+ T cell responses, but weaker CD8+ T cell responses (83, 212). In another study, targeting cDC2-specific receptors F4/80-like receptor (FIRE) or C-type lectin immune receptor (CIRE/CD209a, one of the murine homologues of DC-SIGN (213)) with rat IgG mAbs in the absence of adjuvant elicited rat IgG-specific Ab responses, whereas the same protocol with DEC-205-targeting mAbs did not (214). This mirrored the trends observed in Carter et al’s study, in which only Dectin-1-targeted Ag but not DEC-205- targeted Ag was observed to induce Ab responses (212). These findings would suggest that Ag delivered to the cDC2 subset have a greater propensity to induce Ab responses owing to the more effective stimulation of CD4+ T cells, including Tfh.

However, not all DC subset-specific receptors necessarily follow this model. Clec9A is only expressed on cDC1s and not cDC2s, yet targeting of this receptor induces strong Ab responses (120, 215). Indeed, in vivo targeting of Clec9A was shown to induce potent Ag-specific CD4+ T cell responses, particularly Tfh, to a much greater degree than targeting of DEC-205 (215). This demonstrates that the expression of a receptor on particular DC subsets does not necessarily predict the immunological outcome of targeting that receptor.

Properties of targeted receptors

The DC subset targeted may be less relevant than the specific receptor utilised, as the properties of individual receptors, such as their internalisation kinetics or intracellular trafficking, may have a greater influence on the response induced. It might be expected that receptors capable of internalising greater amounts of targeted Ag would promote correspondingly greater stimulation of Ag-specific responses. Surprisingly, studies have indicated that this is not the case. Chatterjee et al. measured the uptake of fluorescently labelled anti-CD40, anti-MR or anti-DEC-205 mAbs by human DCs, and found anti-MR to be accumulated most efficiently, followed by anti-DEC-205 then anti-CD40 (216). However, when the receptors were compared for their capacity to promote cross-presentation of an attached peptide Ag, anti-CD40 was most effective, despite delivering the least amount of Ag. DEC-205 was surprisingly

inefficient at mediating cross-presentation in this study, despite promoting greater internalisation of Ag than CD40. Further analysis suggested that this was due to enhanced degradation of the attached peptide Ag in the late endosomes targeted by DEC-205. This effect was likely exacerbated by using peptide Ags, which are more susceptible to degradation. The preferential localisation of MR and CD40 in less degradative early endosomes may explain why targeting these receptors promoted greater Ag presentation. However, the discrepancy between the greater internalisation of anti-MR constructs but lower cross-presentation of Ag relative to anti-CD40 was not resolved. The authors speculated that differential uptake kinetics of the two receptors might influence their capacity to mediate cross-presentation. These findings illustrate that the Ag load delivered by a particular DC receptor is not necessarily predictive of its capacity to mediate Ag presentation, as other factors such as intracellular localisation and rate of uptake could have an impact.

A separate study investigating the impact of DC receptor properties on their capacity to mediate Ag presentation also found no correlation between Ag load delivered and efficacy of Ag presentation (217). Using a fluorescent DNA-based probe to measure the internalisation of each receptor, Reuter et al. found that CD40 mediated significantly greater internalisation than DEC-205 in mature DCs. However, delivery of Ag via DEC-205 mediated more efficient MHC I and MHC II presentation, despite the lower Ag load. Although these results support the theory that the Ag load delivered has little correlation with Ag presentation efficiency, the greater Ag presentation but lower Ag load delivered by DEC-205 relative to CD40 is in direct contrast to the trends observed by Chatterjee et al. This discrepancy could reflect differences between the DCs under investigation. Receptors expressed by human BDCA1+ or monocyte-derived DCs used in Chatterjee et al.’s study could potentially exhibit different properties to those expressed on the activated mouse CD8+ DCs used Reuter et al’s study. Notably, Reuter et al. measured the capacity for DEC-205 to mediate presentation of a protein Ag (217), which would be more resistant to degradation than a peptide. This likely circumvented the reduced presentation efficiency caused by excessive degradation of peptide Ags seen by Chatterjee et al.

The intracellular compartment targeted upon uptake has been proposed to impact the route of Ag presentation. Ag directed to early endosomes appears to be preferentially presented on MHC I, while Ag directed to late endosomes or lysosomes is preferentially presented on MHC II (218). In support of this notion, CD40, MR and CD11c have all been observed to localise to the early endosomes of DCs upon internalisation, and all are effective targets to induce CD8+ T cell responses (216, 219). Furthermore, Cohn et al. demonstrated that the relatively weak ability of BDCA1+ DCs to cross-present could be enhanced to levels comparable to BDCA3+ DCs by directing Ag to the early endosomal compartment via anti-CD40 or anti-CD11c mAbs (219).

Thus, the intracellular trafficking of targeting receptors appears to be more important than the DC subset they reside on for directing the type of response induced. However, the intracellular route does not always dictate the capacity of a receptor to promote certain presentation pathways. For example, DEC-205 enters late endosomes or lysosomes upon internalisation, yet induces potent cross-priming of CD8+ T cells against protein Ags (199, 220). CD40, which localises to early endosomes, has also been described to promote efficient MHC II presentation (216).

Beyond choosing which receptor to target, it should be noted that, even when targeting the same receptor, the use of different targeting constructs could cause variation in the generated response. For example, targeting via the anti-Clec9A mAb 7H11 in the absence of adjuvant was reported to induce tolerance and generate Foxp3+ Treg cells (221). However, in a separate study, targeting of the same receptor with another anti-Clec9A mAb, 10B4, did not generate Treg cells, and instead induced Tfh development and potent Ab responses (120, 215). This was hypothesized to be due to the presence of helper epitopes in the IgG2a backbone of 10B4 that were absent in the IgG1 backbone of 7H11 (222). In another example, it was found that a panel of mAbs specific for DC-SIGN exhibited variable capacity for inducing an activating signal upon binding (223). It has also been shown that mAbs that bind to the neck region of DC-SIGN versus mAbs that bind the carbohydrate recognition domain have distinct internalisation mechanisms and intracellular routing (224).

Therefore, it is best to assess each targeting receptor, and possibly even each targeting molecule, on its own merits. In the DC targeting field, there has been particular interest in developing targeting molecules that are efficient at promoting the cross- priming of CD8+ T cells to generate potent CTL responses. Current vaccines, while capable of inducing effective Ab responses, are typically weak inducers of T cell responses. Many diseases for which there is no effective cure or vaccine, including cancer, HIV and malaria, could potentially be controlled with CTL responses, so the development of vaccines that generate effective CTL responses is a pressing unmet need. Several in vivo DC targeting strategies have already shown great potential in this respect, one of the most highly developed of which involves the targeting of DEC-205, the prototypical DC-specific receptor, although promising candidate receptors continue to emerge.

DEC-205 targeting

DEC-205 is a CLR expressed at high levels on cDC1s and thymic endothelial cells, and at lower levels on cDC2s, T cells, B cells and granulocytes in the mouse, as well as NK cells and monocytes in humans. DEC-205 is an endocytic receptor that has been identified to bind keratin exposed by necrotic cells (175). DEC-205 appears to facilitate Ag presentation, as mAbs specific for DEC-205 were shown to be internalised into late endosomal compartments containing MHC II, and subsequently presented to CD4+ T cells with high efficiency (225, 226). The natural endocytic ability of DEC-205, coupled with its high expression on cDC1s, fuelled investigation into its potential as a receptor for the delivery of Ags to DCs. It was found that targeted delivery of Ag via DEC-205-specific mAbs resulted in the marked enhancement of presentation to both CD8+ and CD4+ T cells (220, 227). However, in the absence of adjuvant, the rapid expansion of T cells was transient, and DEC-205 targeting was observed to induce tolerance via T cell anergy or deletion. Only by co- administration of adjuvants to activate DCs, such as an agonistic anti-CD40 mAb, could sustained Ag-specific T cell responses be generated (220, 227). Utilising anti- DEC-205 mAbs to deliver the model Ag ovalbumin (OVA) with anti-CD40, potent

OVA-specific CTL responses could be generated from a naïve endogenous repertoire that protected mice from OVA-expressing tumours, or an OVA-expressing virus (199). Mahnke et al. further demonstrated that anti-tumour immunity could also be generated by targeting bona fide cancer associated-Ags, tyrosine related protein (TRP)2 and gp100, to DEC-205, which protected mice from melanoma (228). In another model, mice were protected from breast cancer after vaccination with the breast cancer Ag human epidermal growth factor receptor 2 (HER2) fused to an anti- DEC-205 mAb (229). Many studies have also demonstrated the capacity for DEC-205 targeted vaccines to induce anti-viral immunity, most notably against HIV Ags (230- 234).

These encouraging results led to the first phase I clinical trial, in which NY-ESO-1, a cancer-testis Ag overexpressed in various types of cancers, was genetically fused to a human DEC-205-targeting mAb and administered to late-stage cancer patients alongside a combination of TLR3 and TLR7/8 agonists. While most patients exhibited NY-ESO-1-specific Ab and T cell responses, only 2 out of 45 participants showed tumour regression at one point during the study, although both patients ultimately experienced continued growth of the tumour (235). This may reflect the difficulty of overcoming the immunosuppressive environment in advanced cancer patients, suggesting that future trials will likely benefit from being initiated in patients at earlier stages of disease. Furthermore, DC-targeted Ag delivery may need to be combined with checkpoint inhibitors to maximise T cell responses.

Clec9A targeting

Another potential issue with targeting DEC-205 that may have contributed to its lacklustre clinical performance is the fact that the receptor is not exclusively expressed by DCs. Especially in humans, multiple cells types express DEC-205 that do not contribute to Ag presentation, which would reduce the availability of targeted Ag for DCs (236).

Clec9A has emerged as an alternative receptor for specific targeting to cDC1s. The expression of Clec9A has been reported to be highly restricted to cDC1s, especially in humans where expression was only detected on cDC1s (120-122). In mice, lower expression is also observed on pDCs (120, 121). Clec9A, like DEC-205, is a CLR with endocytic properties, though it internalises bound ligand into early recycling endosomes (172). Clec9A is known to bind F-actin filaments that are exposed upon necrotic cell death, and promote the cross-presentation of dead cell associated-Ags (169-172). The natural function of Clec9A suggests that Ags targeted to this receptor would be efficiently cross-presented. Indeed, delivery of SIINFEKL, the immunodominant MHC I-restricted peptide of OVA, via anti-Clec9A mAbs induced potent CTL responses that protected mice from OVA-expressing melanoma tumours (121). Protection against tumour growth could also be induced by vaccination with endogenous melanoma peptides conjugated to anti-Clec9A mAbs, while none of the peptides were protective when conjugated to an isotype-control non-targeting Ab (121). Peptides from the endogenous adenocarcinoma Ag MUC1 have also been successfully targeted to Clec9A to induce anti-tumour responses that delayed tumour growth in transgenic mice expressing human MUC1 (237).

Clec9A has also been notable as an effective target for enhancing Ag presentation via MHC II to activate CD4+ T cells and induce Ab responses (120, 215, 221). Strikingly, Tfh and Ab responses could even be generated upon Clec9A targeting in the absence of adjuvant (120, 215). This was attributed to the specificity and persistence of the targeting mAb in the serum, leading to an extraordinarily high concentration of Ag associated with DCs for a prolonged period of time, a situation that has been reported to be able to stimulate immunity despite the lack of DC activation (238). However, some anti-Clec9A targeting mAbs require adjuvant to prime CD4+ T cells effectively, such as the anti-Clec9A mAb 7H11 used by Joffre et al. (221) A direct comparison of anti-Clec9A mAbs that can induce Ab responses in the absence of adjuvant, such as 10B4, with mAbs that require adjuvant, such as 7H11, found that the difference may stem from the inherent immunogenicity of different Ab isotypes. The IgG2a backbone of 10B4 appears to have more helper epitopes than the IgG1 backbone of 7H11, allowing the induction of immunity even in the absence of adjuvant (222).

Nevertheless, these studies all agree that targeting Ag to Clec9A can induce potent T and B cell responses.

In a direct comparison between DEC-205 and Clec9A targeting, delivery of OVA to either receptor in vivo induced comparable levels of MHC I presentation and OVA- specific CTL responses, while Clec9A was superior to DEC-205 at promoting MHC II presentation, CD4+ T cell expansion and Tfh generation (215). Another study examined the targeting of the HIV protein gag-p24 to various DC receptors in the presence of anti-CD40 and polyIC as adjuvants, and confirmed that targeting to DEC- 205 or Clec9A induced similar levels of gag-p24-specific IFNγ-producing CD8+ T cells. In this scenario, targeting to DEC-205 or Clec9A also induced comparable levels of gag-p24-specific IFNγ-producing CD4+ T cells (234). The discrepancy between the two studies regarding the relative ability of the receptors to induce CD4+ T cell responses may be due to differences in the experimental conditions. Lahoud et al. used CpG ODN as an adjuvant and measured primary proliferation of OVA-transgenic T cells at the peak of the response, while Idoyaga et al. used anti-CD40 and polyIC as adjuvants and measured the endogenous T cell response after in vitro restimulation with peptides 14 days after immunisation.

In contrast to these findings, Tullett et al. found a peptide Ag fused to an anti-DEC- 205 mAb to be poorly cross-presented by human DCs in vitro in comparison with Ag fused to an anti-Clec9A mAb, although both mAbs were equally efficient at presentation to CD4+ T cells (239). One possible explanation is that the outcome of targeting certain receptors may differ across species. However, previous evidence would suggest that peptide Ags in particular are inefficiently presented when targeted to DEC-205. This has been attributed to the preferential targeting of DEC-205 to late endosomal compartments, which can cause rapid complete degradation of peptide Ag (216, 219). On the other hand, these compartments may be suitable for generating MHC epitopes from larger proteins. Indeed, studies demonstrating effective presentation of DEC-205-targeted Ags have generally utilised larger Ag, such as the whole protein OVA, that are more resistant to degradation (199, 215). Thus, the nature of the Ag delivered may need to be considered when assessing the efficacy of

DEC-205 targeting. Clec9A delivers Ag to less degradative early endosomes, which could result in better preservation of peptide Ags for presentation. Indeed, effective presentation of both peptide and protein Ags targeted to Clec9A has been reported (121, 215, 240).

The Clec9A protein shares 53% sequence identity between mouse and human, and exhibits similar function in both species, facilitating the translation of these findings into humans (169, 170). Whether Clec9A targeting retains such efficacy at inducing Ag-specific responses in humans remains to be seen, although preliminary studies with human blood DCs (237, 241) and humanized mice (239) suggest that strong CD4+ and CD8+ T cell responses can be expected. Encouraging results have also been reported in non-human primates, in which a potent Ab response could be induced after immunisation with anti-Clec9A mAbs in the absence of adjuvant, mirroring results obtained in mice (222).

XCR1 targeting

One of the advantages of targeting Clec9A over DEC-205 is its highly restricted expression pattern, ensuring that Ag is delivered almost exclusively to cDC1s. XCR1 is the only other receptor identified so far that is similarly restricted to cDC1s. While Clec9A is expressed by pDCs in mice, XCR1 has not been found to be expressed on any other cell type, and like Clec9A, is also restricted to the homologous cDC1 subset in humans (89, 90, 108, 124). This discovery led to investigations into whether XCR1 is similarly a useful target for in vivo delivery of Ags to DCs. XCR1 is a chemokine receptor that binds to the chemokine XCL1, which is specifically produced by activated NK and CD8+ T cells and mediates their interaction with cDC1s during an immune response (89, 90, 242).

Using either anti-XCR1 mAbs or XCL1-Ag constructs to target XCR1 in mice has been reported to promote effective presentation to both CD8+ and CD4+ T cells, resulting in Ag-specific CTL and Ab responses that are polarised towards Th1 immunity (243, 244). In two separate studies utilising different targeting constructs

and vaccination methods, mice immunised with OVA targeted to XCR1 showed protection from the growth of OVA-expressing tumours (243, 245). Fossum et al. also demonstrated that DNA vaccination with the influenza Ag hemagglutinin (HA) fused to XCL1 could induce HA-specific CTL responses that protected mice from a lethal challenge with influenza (244). XCL1-HA vaccination also induced moderate HA- specific Ab responses, which were not protective in this influenza model, and required boosting for neutralising function to be observed. Interestingly, DNA vaccination or laser-assisted intradermal delivery of the XCL1 construct bypassed the need for adjuvant, likely because these methods generate cell damage that may provide the necessary danger signals to promote immunogenicity (244, 245).

An XCL1 homologue exists in humans, as well as a highly related variant, XCL2, that differs at two amino acids and also binds XCR1 (246). Using mice deficient for murine XCR1 but expressing human XCR1 on cross-presenting DC subsets, targeting OVA via human XCL1 or XCL2 constructs resulted in the generation of OVA-specific CTL responses, demonstrating the feasibility of this approach in the human system (243).

The ability to target XCR1 with its chemokine ligand may have advantages over targeting with mAbs. As an endogenous molecule, there is unlikely to be reactivity raised against the targeting component, enabling it to be reused for subsequent vaccinations, and XCL1 retains its chemotactic properties when conjugated to an Ag, which could aid its homing to cDC1s. XCL1 may also have distinct uptake and diffusion kinetics compared to a mAb, although the impact of these properties on targeting efficacy has not yet been examined.

Requirement for adjuvant

Under the classical paradigm of immune activation, DCs require maturation by danger signals to become immunogenic and are tolerogenic when immature. This aligns with the responses seen when targeting DEC-205 – in the absence of adjuvants or DC maturing agents, delivery of DEC-205-conjugated Ag to DCs stimulated transient T cell proliferation but ultimately resulted in the induction of tolerance (220,

227). Only in the presence of an adjuvant, such as anti-CD40 or TLR ligands, to activate DCs could immunogenic responses be induced (220, 227) (199, 215). A similar requirement for adjuvant to induce immunity instead of tolerance has been observed upon targeting to other DC receptors, such as Clec12A and DCIR2 (211, 247).

Conversely, several studies have found that targeting various receptors, including CD45 (248), CD45RA (248), CD4 (248), MHC I (249), MHC II (200), FcγRII (249), mannose receptor (206), CIRE (250), FIRE (250), DCAR1 (251) and Clec9A (120, 215), can induce Ab responses in the absence of adjuvant. Most of these studies did not confirm the maturation status of DCs, which may have been inadvertently activated by endotoxin or other contaminants present in the vaccine. However, DCs targeted by CIRE, FIRE, DCAR1 and Clec9A showed no overt signs of activation, with no detectable change in expression of co-stimulatory molecules (120, 250, 251). For Clec9A, the major potential sources of stimulatory contaminants were also ruled out when responses were not affected in mice lacking MyD88, Trif or TLR4 (120, 215).

Despite its ability to induce potent Ab responses in the absence of adjuvant, the induction of CTL responses via Clec9A targeting was still dependent on adjuvant (215). The provision of adjuvant appears to be a global requirement for generating CTL responses that applies regardless of the receptor targeted. The only apparent exception is a report asserting that CD36 targeting can elicit CTL responses in the absence of adjuvant, although this was achieved by seeding Ag-specific transgenic T cells. Moreover, the activation status of the DCs was not tested and the possibility of activation due to contaminants or triggering of the receptor was not addressed (207). So far, there is no evidence that an endogenous CTL response can be induced by DC targeting in the absence of DC activation.

It should be noted that the tolerance induced by targeting certain receptors in the absence of adjuvant may be desirable in certain circumstances, such as for the treatment of autoimmune diseases. Tolerogenic targeting of Ag to DEC-205 has been used successfully to treat disease in mouse models of type 1 diabetes (252-254),

multiple sclerosis (experimental autoimmune encephalomyelitis) (255, 256) and autoimmune arthritis (257).

Choice of adjuvant

Adjuvants play a key role in ensuring the induction of effective immune responses upon targeting of Ag to DCs. Not only do adjuvants provide an activating signal, but the varying activation pathways triggered by different adjuvants also have a significant impact on polarising the type of response generated. For instance, targeting Ag to Clec9A in the presence of adjuvants such as polyIC results in the strong induction of IFNγ-producing Th1 cells (221). Yet, in the same study, using curdlan (a Dectin-1 agonist) as the adjuvant alongside Clec9A-targeted Ag resulted in minimal induction of Th1 cells. Instead, the use of curdlan as an adjuvant preferentially polarised the induction of Th17 cells (221). This demonstrates that the adjuvant can play a significant role in determining the type of immune response generated, and to induce strong CTL responses, choosing an adjuvant that supports Th1 induction would be advantageous.

However, only a handful of adjuvants have been approved for human use, most of which have been described to induce strong Ab responses, but do not necessarily promote strong CTL responses. For example, the earliest approved and most widely used adjuvant, alum, is known to heavily skew towards Th2-type responses, promoting strong Ab responses but weak CTL responses (258, 259). While more recently approved adjuvants, such as the squalene-based oil-in-water emulsions MF59 and AS03, have been reported to induce a more mixed Th1/Th2-type response (260, 261), there remains a need to develop more potent Th1-polarising adjuvants to augment vaccine strategies that focus on generating better CTL immunity.

One promising source of new adjuvants is TLR ligands, many of which are PAMPs that naturally induce strong Th1 responses. TLR ligand derivatives such as polyICLC (TLR3 agonist), monophosphoryl lipid A (TLR4 agonist), resiquimod (TLR7/8 agonist) and CpG ODNs (TLR9 agonist) have been reported to promote Th1-type

responses when used as adjuvants against a range of diseases, although none, with the exception of monophosphoryl lipid A, have been approved for human use as yet (261, 262).

Improving adjuvants by targeting DCs

Just as the generation of immune responses against Ag can be enhanced by delivering Ag to DCs, it may be possible to develop more potent adjuvants by targeting adjuvants to DCs. This could improve the activity of adjuvants by applying a similar theory; that concentrating the immunostimulatory agent with the DCs that perform the priming of T cells is likely to improve the induction of the desired Ag-specific responses. Targeting adjuvant to DCs could also reduce non-specific stimulation of other cells, which could reduce the potential for systemic inflammation or other adverse effects.

A previous study attempted to target adjuvants to DCs by direct conjugation to anti- DEC-205 mAbs. Conjugation of CpG ODN adjuvant to an anti-DEC-205 construct carrying OVA Ag resulted in stronger CTL responses than the same dose of unconjugated CpG ODN co-administered with the anti-DEC-205-Ag construct (263). However, the enhanced killing after vaccination with the anti-DEC-205-Ag-CpG construct was found to be independent of DEC-205 targeting. In fact, CpG conjugation appeared to interfere with the binding specificity of the anti-DEC-205 construct, resulting in high levels of binding to DEC-205- DCs. This non-specific binding of the anti-DEC-205-Ag-CpG conjugate made it unclear whether the improved immunogenicity was specifically due to enhanced targeting of adjuvant to DCs.

The recent finding that DEC-205 itself is a receptor for CpG ODNs provided the opportunity to investigate a novel method of targeting CpG ODNs to DCs, by manipulating their innate capacity to bind DEC-205. DEC-205 was found to promote the uptake of CpG ODNs by DCs, and play a key role in mediating their immunostimulatory effects (264). Thus, modifications to enhance CpG ODN binding to DEC-205 may enhance their adjuvant capacity. This would be an important proof-

of-principle to demonstrate that enhanced targeting of adjuvants to DCs could be a viable method to develop more potent adjuvants.

On another note, further characterising the molecular characteristics of ODNs that promote DEC-205 binding could also contribute to the potential identification of a new ligand for DEC-205. Although ODNs are synthetic, they may structurally resemble a natural ligand of DEC-205, such as naturally occurring DNA. Identifying a new ligand of DEC-205 could provide critical insight into the physiological function of DEC-205 and its role in the immune system, which is not well defined. Further information about the biology of DEC-205 could allow a better understanding of the mechanism of action and potential side effects of therapeutics that utilise DEC-205, which is particularly important since DEC-205-targeted vaccines have already entered the clinic.

Optimising prime-boost with DC-targeted vaccines

The fundamental concept of DC-targeted vaccines has already proven to be a viable method of generating stronger immune responses than conventional vaccines. Many studies have been conducted to lay the groundwork by identifying effective targeting molecules and establishing their mechanisms of action. One targeting molecule has already progressed to clinical trials in cancer patients, while several others show promising results in mice with strong prospects for translation into humans.

For most vaccines, the primary immune response generated after a single immunisation can be greatly enhanced by subsequently administering one or more boosting immunisations. However, evidence from other fields suggests that boosting the response generated by DC-targeted vaccines may not be a straightforward task. It has been observed that, particularly for vaccine constructs that induce strong primary responses, repeated immunisation with the same construct can result in suboptimal boosting. This has been proposed to be due to the primary Ab response reacting with and neutralising the boosting construct, and can be overcome by utilising distinct vaccine constructs for the prime and boost immunisations, in what is known as a

heterologous prime-boost (265). As DC-targeted vaccines are known to induce highly potent responses after a single immunisation, it is quite likely that interference from the primary Ab response could prevent effective boosting. However, to our knowledge, the impact of this problem on the capacity to boost with DC-targeted vaccines has not been investigated.

Although it is highly encouraging that DC-targeted vaccines can induce potent responses after only one vaccination, the more realistic measure of their potential efficacy in human patients would be to examine their performance in a prime-boost scenario. Therefore, it is important to establish whether the primary response generated after DC-targeted vaccination could interfere with boosting, and investigate heterologous prime-boost strategies that could effectively overcome it. Particularly when the diseases targeted include cancer and HIV, for which there is no cure precisely because it has been difficult to raise immunity against them, developing an optimally effective prime-boost strategy for DC-targeted vaccines could help maximise the chances of success.

Aims

This thesis aims to build upon the concept of targeting DCs to enhance immune responses in two major directions - firstly, by investigating a novel method of targeting adjuvant to DCs to design more potent adjuvants, and secondly, by utilising existing DC-targeting strategies to develop more effective prime-boost regimens.

In Chapter 3, the newly uncovered function of the DC receptor DEC-205 to bind CpG ODNs will be further investigated, and based on these findings, a novel method of designing more potent CpG ODN adjuvants will be devised and tested for the first time. The molecular requirements for CpG ODNs to bind DEC-205, and the immunological consequences of this interaction will be examined, culminating in the demonstration that enhancing DEC-205 binding of CpG ODNs is a viable method of designing more potent adjuvants.

In Chapter 4, the potential for CD14, another receptor reported to bind CpG ODNs, to influence the immunostimulatory activity of CpG ODNs will be examined. CD14 could serve as another novel means of modulating CpG ODN function, and could potentially act in synergy with DEC-205.

In Chapter 5, the physiological role of DEC-205, which is not well understood, will be explored by investigating the potential for DEC-205 to bind DNA of biological origin.

In Chapter 6, the efficacy of targeting Ag to either Clec9A or XCR1, the only two methods currently described to target cDC1s specifically, will be compared. An observation that the initial immune response can be detrimental to the generation of a secondary response upon boosting will be investigated, and the capacity for heterologous combinations of DC-targeting constructs to overcome this issue and allow effective boosting will be explored.

Chapter 2: Materials and methods

2.1. Media and buffers

EDTA-BSS. Mouse tonicity balanced salt solution supplemented with EDTA, pH 7.22. Prepared by mixing 968 ml 1.68M NaCl (Univar, WA, USA), 24 mL 1.68M KCl

(Univar, WA, USA), 15 ml potassium phosphate buffer (composed of 1.68M KH2PO4

(VWR, PA, USA) and 1.12M K2HPO4 (Univar, WA, USA) at 1:4 ratio), 96 ml 1.68M HEPES (JRH Biosciences, KS, USA), and 560 ml 0.099M EDTA into 9424 ml MilliQ

H2O. Produced by WEHI, Victoria, Australia.

EDTA-BSS-2%FCS. EDTA-BSS supplemented with 2% (v/v) FCS.

FCS-EDTA. FCS supplemented with 10% (v/v) 0.1M EDTA.

RPMI. Mouse tonicity RPMI 1640 prepared by mixing RPMI 1640 powder (Gibco, NY, USA), 200 ml 1.68M HEPES (VWR, PA, USA), 1.1 g sodium pyruvate (Sigma,

MO, USA), 20 g NaHCO3 (Merck, Darmstadt, Germany), 100 U/ml pencillin and 100

µg/ml streptomycin (Sigma-Aldrich, MO, USA) in 10 L MilliQ H2O.

RPMI-2%FCS. RPMI supplemented with 2% (v/v) FCS.

RPMI-5%FCS. RPMI supplemented with 5% (v/v) FCS.

Complete media. Mouse tonicity RPMI 1640 supplemented with 1% 2- mercarptoethanol (Sigma-Aldrich, MO, USA) and 10% FCS. Produced by WEHI, Victoria, Australia.

PBS. Used for all cell-based assays. Dulbecco's phosphate-buffered saline, no calcium, no magnesium, Gibco, NY, USA.

PBS (ELISA). Used for all ELISA buffers. Mouse tonicity phosphate buffered saline, pH 7.2. Prepared as 10x concentrate (22.08 g NaH2PO4.H2O (Sigma, MO, USA) +

90.80 g NaHPO4 (Sigma, MO, USA) + 350.4 g NaCl (Sigma, MO, USA) dissolved in

MilliQ H2O to 4 L volume), diluted to 1x concentration in MilliQ H2O.

PBS-0.1%BSA. PBS supplemented with 0.1% (w/v) BSA.

PBS-0.3%BSA. PBS (ELISA) supplemented with 0.3% (w/v) BSA.

PBS-1%BSA. PBS supplemented with 1% (w/v) BSA for cell-based assays, or PBS (ELISA) supplemented with 1% (w/v) BSA for ELISA assays.

PBS-2%BSA. PBS supplemented with 2% (w/v) BSA.

Tetramer buffer. PBS supplemented with 1% (w/v) BSA and 1% (v/v) 1M sodium azide.

ELISA blocking buffer. PBS (ELISA) supplemented with 5% (w/v) milk powder, prepared fresh as needed.

ELISA wash buffer. PBS (ELISA) supplemented with 0.05% (v/v) Tween20.

2.2. Antibodies

Table 2.1. Conjugated antibodies.

Antibody conjugate Clone Source anti-CD11b APC M1/70 WEHI, Victoria, Australia anti-CD11c PE-Cy7 N418 BD Biosciences, NJ, USA anti-CD14 biotin Sa2-8 eBioscience, CA, USA anti-CD19 FITC 1D3 WEHI, Victoria, Australia anti-CD3 APC 145-2C11 eBioscience, CA, USA anti-CD4 PE GK1.5 WEHI, Victoria, Australia anti-CD40 PE FGK45.5 WEHI, Victoria, Australia anti-CD44 AF700 IM7.81 eBioscience, CA, USA anti-CD44 FITC IM7.81 WEHI, Victoria, Australia anti-CD45R Pacific Blue 14.8 BD Biosciences, NJ, USA

anti-CD8 AF647 YTS169.4 WEHI, Victoria, Australia anti-CD8 APC YTS169.4 WEHI, Victoria, Australia anti-CD8 PE YTS169.4 WEHI, Victoria, Australia anti-CD8 PE-Cy7 53.67 BD Biosciences, NJ, USA anti-CD80 PE 16-10A1 eBioscience, CA, USA anti-CD86 PE P03.1 eBioscience, CA, USA anti-F4/80 FITC F4/80 WEHI, Victoria, Australia anti-FLAG biotin 9H1 WEHI, Victoria, Australia anti-FLAG FITC 9H1 WEHI, Victoria, Australia anti-human DEC-205 FITC MG38 WEHI, Victoria, Australia anti-IFNγ AF647 XMG1.2 BD Biosciences, NJ, USA anti-Ly5.1 APC A20.1 BD Biosciences, NJ, USA anti-MHC II PE M5/114 WEHI, Victoria, Australia anti-mouse DEC-205 APC NLDC-145 WEHI, Victoria, Australia anti-mouse DEC-205 biotin NLDC-145 WEHI, Victoria, Australia anti-mouse DEC-205 PE NLDC-145 WEHI, Victoria, Australia anti-OVA biotin WEHI, Victoria, Australia anti-Vα2 PE B20.1 BD Biosciences, NJ, USA

IL-12 detection Ab C17.8 WEHI, Victoria, Australia (biotinylated)

Table 2.2. Secondary antibodies.

Antibody Source anti-mouse IgG HRP Donkey anti-mouse IgG, horseradish peroxidase conjugate. Chemicon (Millipore), MA, USA. anti-mouse IgG PE Goat anti-mouse IgG, PE conjugate. eBioscience, CA, USA. anti-mouse IgG2a biotin Goat anti-mouse IgG2a, horseradish peroxidase conjugate. Southern Biotech, AL, USA.

anti-mouse IgG2b biotin Goat anti-mouse IgG2b, horseradish peroxidase conjugate. Southern Biotech, AL, USA. anti-mouse IgG2c biotin Goat anti-mouse IgG2c, horseradish peroxidase conjugate. Southern Biotech, AL, USA. anti-rat IgG biotin Goat anti-rat IgG, biotin conjugate. Biolegend, CA, USA. streptavidin APC Streptavidin APC conjugate. BD Pharmingen, CA, USA. streptavidin HRP Streptavidin horseradish peroxidase conjugate. GE Healthcare UK, Buckinghamshire, England. streptavidin PE Streptavidin PE conjugate. BD Pharmingen, CA, USA.

Table 2.3. Unconjugated antibodies.

Antibody Clone Source anti-CD28 37.51 Functional grade, eBioscience, CA, USA. anti-FcR (anti-FcγRIII/II) 2.4G2 WEHI, Victoria, Australia anti-FLAG 9B4 Kindly provided by Dr. Lorraine O’Reilly, WEHI, Victoria, Australia. anti-human DEC-205 MMRI7 Kindly provided by Dr. Kristen Radford, Mater Medical Research Institute and University of Queensland, Australia. anti-human DEC-205 HD24 Kindly provided by Dr. Chae Gyu Park, Rockefeller University, NY, USA. (266) anti-human DEC-205 HD71 Kindly provided by Dr. Chae Gyu Park, Rockefeller University, NY, USA. (266) anti-human DEC-205 HD83 Kindly provided by Dr. Chae Gyu Park, Rockefeller University, NY, USA. (266) anti-mouse DEC-205 NLDC-145 WEHI, Victoria, Australia IL-12 capture Ab R29A5 WEHI, Victoria, Australia

2.3. Targeting constructs

Rat 10B4. Rat anti-mouse Clec9A mAb (10B4, isotype IgG2a), generated in house as previously described (120).

Rat 10B4-OVA. Rat 10B4 carrying full-length OVA on the heavy chain, generated in house by genetic fusion, as previously described (215).

Rat 10B4-OBI. Rat 10B4 carrying OBI peptide on the heavy chain, generated in house by genetic fusion. OBI peptide has sequence LESIINFEKLTELKISQAVHAAHAEINEAGREVKLPGFGDSIE containing linear OT-I, OT-II and OBI (267) epitopes of OVA in that order (epitopes are underlined).

Mouse 10B4. Mouse chimeric 10B4 mAb with rat constant regions replaced by mouse IgG1-κ constant regions.

Mouse 10B4 OVA. Mouse 10B4 carrying full-length OVA on the heavy chain, generated in house by genetic fusion.

Rat GL117. Rat IgG2a non-targeting isotype control (anti-β-galactosidase), produced in house.

Rat GL117-OVA. Rat GL117 carrying full length OVA on the heavy chain, generated in house by genetic fusion, as previously described (215).

XCL1-OVA and NIP-OVA. Vaccibody constructs, containing a human dimerization unit consisting of hinge1 and hinge4 as well as CH3 from human γ3 as previously described (244, 268), were kindly provided by Dr. Even Fossum and Professor Bjarne Bogen, K.G. Jebsen Center for Influenza Vaccine Research, Institute of Immunology, University of Oslo and Oslo University Hospital, Oslo, Norway.

2.4. Recombinant proteins and peptides

CD14-Fc. Recombinant mouse CD14 fused with C-terminal polyhistidine-tagged Fc region of human IgG1 at the C-terminus, Sino Biological, Beijing, China.

CD14-His. Recombinant mouse CD14 with C-terminal polyhistidine tag, Sino Biological, Beijing, China.

Clec9A (mouse). Mouse Clec9A ectodomain was subcloned into a pEF-Bos vector modified to contain a biotinylation consensus sequence (a peptide consensus sequence NSGLHHILDAQKMVWNHR recognised specifically by E. coli biotin holoenzyme synthetase BirA) and the FLAG epitope, as previously described (170). The resulting fusion constructs thus included (in order of N-terminus): the IL-3 signal sequence (to ensure secretion), the biotinylation consensus sequence, a FLAG-tag, and Clec9A cDNA fragment. Recombinant proteins were expressed in FreeStyle 293F cells by transient transfection using FreeStyle Max transfection reagent and FreeStyle Expression Media (Invitrogen, Victoria, Australia) and harvested 5 days post- transfection. Secreted proteins were purified by affinity chromatography using an anti-FLAG M2 agarose resin (Sigma, Castle Hill, Australia) and elution with 100 μg/ml FLAG peptide (Auspep, Victoria, Australia), and further purified by size- exclusion chromatography using a pre-packed Superdex 200 column (GE Healthcare, Rydalmere, Australia).

DEC-205 (human). The codon-optimised ectodomain of human DEC-205 (CD205; NP_002340; 1-MRT…PLG-1664) was synthesised by GeneArt in frame with a C- terminal FLAG-tag and a biotinylation consensus sequence, and subcloned into pcDNA3.1+ expression vector, as previously described (264). Recombinant proteins were expressed in FreeStyle 293F cells by transient transfection using FreeStyle Max transfection reagent and FreeStyle Expression Media (Invitrogen, Victoria, Australia) and harvested 5 days post-transfection. Secreted proteins were purified by affinity chromatography using an anti-FLAG M2 agarose resin (Sigma, Castle Hill, Australia) and elution with 100 μg/ml FLAG peptide (Auspep, Victoria, Australia), then further purified by size-exclusion chromatography using a prepacked Superose 6 column.

DEC-205 (mouse). The ectodomain of mouse DEC-205 (CD205; NP_038853.2; 1- MRT ... PLS-1665) was synthesised by GeneArt in frame with a C-terminal FLAG-tag and a biotinylation consensus sequence, as previously described (264). The ectodomain was subcloned into pcDNA3.1+ expression vector and expressed and purified in the same manner as described above for human DEC-205.

Granulin. Recombinant human granulin, Abcam, Cambridge, UK.

HMGB1. Mouse High Mobility Group Box 1 (HMGB1), protein carrier-free, eBioscience, CA, USA.

LL-37. Human LL-37 peptide with sequence H- LLGDFFRKSKEKIGKEFKRIVQRIKDFLRNLVPRTES-OH synthesized by Mimotopes, Victoria, Australia.

OT-I peptide (SIINFEKL). H-2Kb MHC I restricted epitope of OVA, residues 257- 264 (SIINFEKL), Peprotech, NJ, USA.

OT-II peptide. I-Ab MHC II restricted epitope of OVA, residues 323-339 (ISQAVHAAHAEINEAGR), Peprotech, NJ, USA.

OVA. Ovalbumin from chicken egg white, ≥98%, Sigma, MO, USA.

2.5. ODNs

ODNs with the following sequences were purchased from Geneworks (Adelaide, Australia) and reconstituted in PBS. For binding studies, ODNs were synthesized with an additional 3’ biotin. For macrophage uptake studies, 1668 was synthesized with a 5’ Cy3 tag. * indicates a phosphorothioate linkage, - indicates a phosphodiester linkage.

Table 2.4. ODN sequences used in this study.

ODN Sequence

1668 5'-T*C*C*A*T*G*A*C*G*T*T*C*C*T*G*A*T*G*C*T-3'

1668 (diester) 5'-T-C-C-A-T-G-A-C-G-T-T-C-C-T-G-A-T-G-C-T-3'

1826 5'-T*C*C*A*T*G*A*C*G*T*T*C*C*T*G*A*C*G*T*T-3'

1826 (diester) 5'-T-C-C-A-T-G-A-C-G-T-T-C-C-T-G-A-C-G-T-T-3'

2006 5'-T*C*G*T*C*G*T*T*T*T*G*T*C*G*T*T*T*T*G*T*C*G*T*T-3'

2006 (diester) 5'-T-C-G-T-C-G-T-T-T-T-G-T-C-G-T-T-T-T-G-T-C-G-T-T-3'

2216 5'-G*G*G-G-G-A-C-G-A-T-C-G-T-C*G*G*G*G*G*G-3'

2216 (diester) 5'-G-G-G-G-G-A-C-G-A-T-C-G-T-C-G-G-G-G-G-G-3'

2216 (fully 5'-G*G*G*G*G*A*C*G*A*T*C*G*T*C*G*G*G*G*G*G-3' thioated)

2395 5'-T*C*G*T*C*G*T*T*T*T*C*G*G*C*G*C*G*C*G*C*C*G-3'

6T-1668d 5'-T*T*T*T*T*T*T-C-C-A-T-G-A-C-G-T-T-C-C-T-G-A-T-G-C-T-3'

6C-1668d 5'-C*C*C*C*C*C*T-C-C-A-T-G-A-C-G-T-T-C-C-T-G-A-T-G-C-T-3'

6G-1668d 5'-G*G*G*G*G*G*T-C-C-A-T-G-A-C-G-T-T-C-C-T-G-A-T-G-C-T-3'

6A-1668d 5'-A*A*A*A*A*A*T-C-C-A-T-G-A-C-G-T-T-C-C-T-G-A-T-G-C-T-3'

20mer 5'- C*G*T*T*T*T*G*T*C*G*T*T*T*T*G*T*C*G*T*T-3'

18mer 5'- T*T*T*T*G*T*C*G*T*T*T*T*G*T*C*G*T*T-3'

14mer 5'- G*T*C*G*T*T*T*T*G*T*C*G*T*T-3'

10mer 5'- T*T*T*T*G*T*C*G*T*T-3'

14T 5'- T*T*T*T*T*T*T*T*T*T*T*T*T*T-3'

7T-7T 5'- T*T*T*T*T*T*T-T*T*T*T*T*T*T-3'

21798 5'-T*C-G*T*C-G*A*C-G*A*T*C-G*G*C*G*C-G*C*G*C*C*G-3'

14T-21798 5'- T*T*T*T*T*T*T*T*T*T*T*T*T*T*T*C-G*T*C-G*A*C-G*A*T*C- G*G*C*G*C-G*C*G*C*C*G-3'

2006-21798 5'-T*C*G*T*C*G*T*T*T*T*G*T*C*G*T*T*T*T*G*T*C*G*T*T*T*C- G*T*C-G*A*C-G*A*T*C-G*G*C*G*C-G*C*G*C*C*G-3'

The anti-sense ODN (ASO) 518477 designed to inhibit the expression of the mouse β-1,4-galactosyltransferase polypeptide 6 gene, was kindly provided by Dr. Ted Yun of Isis Pharmaceuticals (now renamed Ionis Pharmaceuticals), CA, USA. 518477 has sequence 5'-C*T*C*A*C*A*T*T*G*A*C*A*C*T*G*A*G*G-3' with “4-10-4” MOE chemistry (the first four 5’ and last four 3’ bases have an additional modification of 2’- O-methylribose).

2.6. Biological DNA samples

Bacterial DNA. Enterobacter aerogenes, Enterobacter cloacae, Pseudomonas aeruginosa and Pseudomonas fluorescens bacterial DNA kindly provided by Professor Dick Strugnell, Department of Microbiology and Immunology, University of Melbourne, Australia.

Bacterial plasmids. 238 (pJTU1238) and SKT (SK+) bacterial plasmids, previously described to possess the thioating dnd gene cluster (pJTU1238) or not (SK+) (269), kindly provided by Dr Lianrong Wang, Shanghai Jiao Tong University, Shanghai, China.

Extracellular bacterial DNA. Kindly provided by Associate Professor Cynthia Whitchurch, University of Technology Sydney, Australia.

Mitochondrial DNA. Kindly provided by Dr. Seth Masters, WEHI, Victoria, Australia.

Neutrophil extracellular traps (NETs). Kindly provided by Dr. Seth Masters, WEHI, Victoria, Australia.

Plasmodium berghei DNA. Kindly provided by Professor Tania de Koning-Ward, Deakin University, Victoria, Australia.

Plasmodium falciparum DNA. Kindly provided by Dr. Paul Gilson, Burnet Institute, Victoria, Australia.

Salmon DNA. Salmon sperm DNA, purchased from Invivogen, CA, USA, resuspended in PBS.

Schistosoma mansoni egg antigen. Kindly provided by Dr. Rachel Lundie, Monash University, Victoria, Australia.

2.7. Other reagents

ABTS. 2,2’-Azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) diammonium salt. Sigma-Aldrich, MO, USA. Prepared as 50x stock by diluting 200 mg in 7.3 mL MilliQ

H2O.

ABTS substrate. Freshly prepared for each use at the following ratio: 10 μl of 30%

H2O2 + 200 μl of 50x stock ABTS in 10 ml 0.1M citric acid.

Agarose. High purity analytical grade agarose, Promega, WI, USA.

BD Cytofix/Cytoperm. 10x stock solution provided in BD Cytofix/Cytoperm Plus Fixation/Permeabillization Kit (BD Biosciences, NJ, USA), diluted to 1x in EDTA- BSS-2%FCS for each use.

BD GolgiPlug. Provided in BD Cytofix/Cytoperm Plus Fixation/Permeabillization Kit (BD Biosciences, NJ, USA). Contains brefeldin A.

Biomag beads. Goat anti-rat IgG beads, Qiagen, Victoria, Australia.

Blue/Orange loading dye, 6x. Promega, WI, USA.

BSA. Bovine serum albumin. Sigma-Aldrich, MO, USA.

CFSE. Carboxyfluorescein succinimidyl ester, Molecular Probes OR, USA. Stock solution prepared at 2.25 mM in DMSO.

Citric acid. Sigma-Aldrich, MO, USA. Stock solution prepared by diluting to 0.1M in MilliQ H2O, adjusted to pH 4.1.

Depletion cocktail for DC enrichment. The following rat anti-mouse mAbs were diluted in EDTA-BSS-2%FCS: KT3-1.1 (anti-CD3), RA36B2 (anti-CD45R), T24/31.7 (anti-Thy-1), RB68C5 (anti-Ly6C/G) and TER119 (anti-erythrocyte) and stored in frozen aliquots. mAbs produced by WEHI, Victoria, Australia.

Depletion cocktail for T cell enrichment. The following rat anti-mouse mAbs were diluted in RPMI-2%FCS: M1/70 (anti-CD11b), TER-119 (anti-erythrocytes), M5/114 (anti-MHC II), RB6-8C5 (anti-Ly6C/G), F4/80 (anti-macrophages), plus one of either GK1.5 (anti-CD4) or 53-6.7 (anti-CD8), and stored in frozen aliquots. mAbs produced by WEHI, Victoria, Australia.

DMSO. Dimethyl sulfoxide, Sigma-Aldrich, MO, USA.

DNase/collagenase. Stock solution prepared by adding 0.1 g DNase I from bovine pancrease (Roche, Mannheim, Germany) and 0.7 g collagenase type II (Worthington Biochemicals, NJ, USA) to 100 ml RPMI.

EDTA. Ethylenediaminetetraacetic acid, Fisher Scientific, Leicestershire, UK. Prepared as 0.1M solution in MilliQ H2O, adjusted to pH 7.2.

Fc block. Anti-FcR (anti FcγRIII/II, clone 2.4G2, WEHI) diluted 1:5 in 1 mg/ml rat IgG (rat gamma globulin, Jackson Immunoresearch Labs, PA, USA).

FCS. Fetal calf serum, heat-inactivated at 56ºC for 1 hour. Gibco, NY, USA.

G418 sulfate. G418 sulfate (geneticin) 50 mg/ml, AG Scientific, CA, USA.

GelRed nucleic acid stain. GelRed nucleic acid gel stain, 10,000x in water, Biotium, CA, USA.

H2O2. Hydrogen peroxide solution 30%, puriss, stabilized, Riedel-de Haen, Hanover, Germany.

Heparin. 1 ml sterile PBS was added to sodium heparin coated tubes (158 USP units, BD, NJ, USA), vortexed, and 10 μl (for submandibular bleeds) or 20 μl (for cardiac bleeds) added to each tube used for collection of blood (one mouse per tube).

HMGB1. Mouse High Mobility Group Box 1 (HMGB1), protein carrier-free, eBioscience, CA, USA.

Hydrochloric acid. 37% hydrochloric acid, Sigma-Aldrich, MO, USA.

LPS biotin. Biotinylated lipopolysaccharide, Invivogen, CA, USA.

Mirus Label IT kit. Label IT Nucleic Acid Labeling Kit, Biotin, Mirus Bio, WI, USA.

Nycodenz. Mouse tonicity solution prepared from iohexol powder (Axis-Shield, Oslo, 3 Norway). Prepared as 0.372M stock in MilliQ H2O, adjusted to 1.077 or 1.091 g/cm at 4ºC by diluting in EDTA-BSS.

OVA-specific MHC I tetramer. H-2Kb SIINFEKL PE, produced by Dr. Jie Lin, A. Brooks Laboratory, Peter Doherty Institute, Victoria, Australia.

PCR markers. Promega, WI, USA. Contains bands at 50, 150, 300, 500, 750 and 1,000 bp.

PI. Propidium iodide, Calbiochem (EMD Millipore), MA, USA. Prepared as 100 μg/ml stock by dissolving 5 mg in 50 ml mouse tonicity PBS. Used at a final concentration of 0.5 μg/ml (1:200) in staining media for flow cytometry.

PKH26. Provided in PKH26 Red Fluorescent Cell Linker Kit for General Membrane Staining, Sigma-Aldrich, MO, USA. polyIC. Polyinosinic-polycytidylic acid. Amersham, Buckinghamshire, England.

RCRB. Red cell removal buffer containing 156 mM NH4Cl (VWR, PA, USA),

11.9 mM NaHCO3 (Merck, Darmstadt, Germany), and 0.097 mM EDTA (Sigma, MO,

USA) dissolved in MilliQ H2O. Produced by WEHI, Victoria, Australia.

Sodium azide. Sigma-Aldrich, MO, USA. Stock solution prepared by dilution to 1M in MilliQ H2O.

Tris-borate-EDTA (TBE) buffer. Prepared 0.5x buffer from 10x stock (Invitrogen, CA, USA) by diluting 1:20 in MilliQ water.

Trypsin. 0.05% trypsin 1:250 (Gibco, NY, USA) prepared in solution containing 136.9 mM NaCl (Univar, WA, USA), 5.37 mM KCl (Univar, WA, USA), 5.55 mM D-

Glucose (C6H12O6,Univar, WA, USA), 6.9 mM NaHCO3 (Merck, Darmstadt,

Germany) and 1.3 mM EDTA (Sigma, MO, USA) dissolved in MilliQ H2O. Produced by WEHI, Victoria, Australia.

Tween20. Polyethylene glycol sorbitan monolaurate. Sigma-Aldrich, MO, USA.

YOYO-1. YOYO-1 iodine 1 mM solution in DMSO, Molecular Probes, OR, USA.

2.8. Cell lines

CHO-K1 cells. Chinese hamster ovary cells. mDEC-CHO cells. Mouse DEC-205-expressing CHO-K1 cells were previously generated by transfection (electroporation) with full-length untagged mouse DEC-205, and high expressors sorted by flow cytometry. hDEC-CHO cells. Human DEC-205-expressing CHO-K1 cells were previously generated by transfection (electroporation) with full-length untagged human DEC-205, and high expressors sorted by flow cytometry.

Clec9A-CHO cells. Mouse Clec9A-expressing CHO-K1 cells were previously generated by transfection (electroporation) with full-length untagged mouse Clec9A, and high expressors sorted by flow cytometry.

RAGE-CHO cells. CHO-K1 cells transiently transfected with RAGE-mCitrine, as detailed below in “Transfection of CHO cells with RAGE”.

HEK293F cells. Freestyle 293F cells derived from human embryonic kidney cells 293 (HEK293), Invitrogen, CA, USA.

B16-OVA cells. B16-F10 mouse skin melanoma cells stably transfected with FLAG- tagged OVA.

All cell lines were maintained in RPMI-5%FCS in a 37ºC incubator with 10% CO2. mDEC-CHO, hDEC-CHO, Clec9A-CHO and B16-OVA cells were maintained in RPMI-5%FCS supplemented with G418 sulfate (0.5 mg/ml) to select for transfectants. Cells were harvested by treatment with 0.01M EDTA for 5 mins at 37ºC to detach cells from flask, washed twice, then resuspended in the appropriate media for downstream use.

2.9. Mice

C57BL/6J (B6), DEC-205-/- (B6.129P-Ly75tm1Mnz/J), CD14-/- (270), TLR9-/-, OT- I/Ly5.1 and OT-II/Ly5.1 mice were bred under specific pathogen free conditions at the Walter and Eliza Hall Institute (WEHI) or at the Burnet Institute. OT-I (H-2Kb b restricted TCR transgenic specific for OVA257-264 peptide) (271) and OT-II (I-A restricted TCR transgenic specific for OVA323-339 peptide) (272) mice crossed to Ly5.1 mice were used as a source of donor OT-I or OT-II cells that can be distinguished from recipient B6 (Ly5.2) cells by expression of Ly5.1 on all haemopoietic cells. All references to OT-I and OT-II mice refer to these OT-I/Ly5.1 and OT-II/Ly5.1 mice.

Gender- and age-matched mice (6-12 weeks) were used and handled according to the guidelines of the National Health and Medical Research Council of Australia. Experimental procedures were approved by the Animal Ethics Committee, Melbourne

Health Research Directorate. Mice were sacrificed by CO2 asphyxiation.

2.10. Injections

All injections were diluted in 100 μl PBS per mouse and delivered with 27.5G insulin syringes (Terumo Medical Corporation, NJ, USA) via the tail vein for an intravenous injection, or subcutaneously into the inguinal region for tumour inoculation.

2.11. Instruments and software

All flow cytometry data was collected on a FACSCalibur or LSR II machine (Becton Dickinson, NJ, USA). Single colour controls were included in every experiment and spectral overlap was compensated using CellQuest or FACSDiva software. Flow cytometry data analysis was performed with WEASEL software (Dr Frank Battye, WEHI, Victoria, Australia). ELISA optical density was measured with a VersaMax plate reader (Molecular Devices, CA, USA). Interpolation of ELISA data and statistical testing was performed with Prism (GraphPad Software, CA, USA).

2.12. Statistical tests

For all data, except where indicated in the Figure legend, statistical significance was determined by two-tailed unpaired Student’s t-test. ns = not significant; * 0.01 < p < 0.05; ** 0.001 < p < 0.01; *** 0.0001 < p < 0.001; **** p < 0.0001.

2.13. Cell labelling

CFSE labelling

Cell suspensions were washed in PBS-0.1%BSA and then resuspended at 1 x 107 cells/ml in PBS-0.1%BSA. 2 μl of 2.25 mM CFSE per 107 cells was added, mixed and incubated for 10 min at 37ºC. Cells were then washed twice in RPMI-2%FCS (at 4ºC) and once in PBS.

PKH26 labelling

PKH26 Red Fluorescent Cell Linker Kit (Sigma-Aldrich, MO, USA) was used to label cells. 12.5 x 106 cells were washed in PBS, centrifuged at 700g for 7 min and the supernatant discarded. Cells were resuspended in 125 μl Diluent C, then 0.5 μl PKH26 dye diluted in 125 μl Diluent C added and immediately mixed by pipetting. Suspension was mixed continuously for 3 min at RT. Staining was stopped by adding 500 μl FCS and incubating for 1 min, then cells were centrifuged at 400g for 10 min at 20ºC. Supernatant was removed and cells washed in 10 ml complete media twice.

2.14. Measuring T cell responses in vivo

T cell enrichment

Lymph nodes and spleens harvested from donor OT-I or OT-II mice were pressed through a metal sieve, and then washed in RPMI-2%FCS. Lymph node cells were washed a further 2 times in RPMI-2%FCS. Meanwhile, splenocytes were resuspended in 5 ml 1.091 g/cm3 Nycodenz, then overlayed onto 5 ml of 1.091 g/cm3 Nycodenz. 1 ml FCS-EDTA was then layered on top and density separation performed by centrifugation at 1700g for 10 min at 4ºC with slow deceleration. The upper ~6 ml of low-density cells was collected and washed in RPMI-2%FCS. Both lymph node and spleen cells were then resuspended in depletion cocktail containing the following rat anti-mouse mAbs: M1/70, anti-CD11b; TER-119, anti-erythrocytes; M5/114, anti- MHC II; RB6-8C5, anti-Ly6C/G; F4/80, anti-macrophages; plus either GK1.5, anti- CD4 to enrich for OT-I cells or 53-6.7, anti-CD8 to enrich for OT-II cells. Cells were incubated with cocktail (10 μl per 106 cells) for 30 min at 4ºC, then diluted to 9 ml with EDTA-BSS-2%FCS, underlayed with 1 ml FCS-EDTA and centrifuged at 700g for 7 min. Cells were then resuspended in 400-600 μl EDTA-BSS-2%FCS and mixed with pre-washed anti-rat IgG magnetic Biomag beads at 10 beads/cell. Cells were incubated with beads for 25 min at 4ºC with continuous rotation on a roller mixer. The mixture was then diluted with EDTA-BSS-2%FCS to 2 ml and placed on a magnet. After allowing the bead-bound cells to attach to the magnet, the T cell- enriched supernatant was recovered. If the number of cells recovered from lymph

nodes alone was insufficient, the required number of spleen cells was pooled with lymph node cells. A small aliquot of cells was stained with anti-Vα2 PE, washed, then resuspended in EDTA-BSS-2%FCS with PI to determine purity by flow cytometry (typically 90-95% for lymph nodes, 70-90% for spleen). Cells were then diluted in PBS to the desired concentration of live OT-I or OT-II cells (Vα2+PI-) for injection. Where indicated, enriched T cells were labelled with CFSE (as in “CFSE labelling”) and viability determined by cell count with a haemocytometer prior to dilution in PBS to the desired concentration for injection.

OT-I and OT-II in vivo proliferation assay

B6 mice were adoptively transferred i.v. with 5 x 104 OT-I or OT-II T cells enriched as described in “T cell enrichment”. OT-I or OT-II cells expressing Ly5.1 enriched from OT-I/Ly5.1 or OT-II/Ly5.1 mice were used in all assays. One day later, mice were immunised i.v. as specified, and spleens harvested 6 days later. Each spleen was pressed through a metal sieve, lysed of red blood cells with RCRB, and then washed in RPMI-2%FCS. Cells were resuspended in 3 ml EDTA-BSS-2%FCS and 200 μl taken for staining. Cells were pelleted and stained in 50 μl of cocktail containing anti-Ly5.1 APC and anti-CD8 PE or anti-CD4 PE (30 min at 4ºC in EDTA-BSS-2%FCS). Cells were then washed and resuspended in EDTA-BSS-2%FCS containing PI. The number of live (PI-) OT-I (Ly5.1+CD8+) or OT-II (Ly5.1+CD4+) cells as a proportion of live endogenous CD8+ (Ly5.1-CD8+) or CD4+ (Ly5.1-CD4+) cells was enumerated by flow cytometry.

In a variation, used in Figures 6.2 and 6.4, 5 x 105 CFSE-labelled OT-I or OTII cells were transferred i.v. into B6 mice 1 day prior to immunisation. Spleens were harvested 3 days (OT-I) or 4 days (OT-II) post-immunisation, and processed and stained as above. The number of live (PI-) OT-I (Ly5.1+CD8+) or OT-II (Ly5.1+CD4+) cells that had undergone at least one division (CFSE fluorescence lower than undivided control) as a proportion of live endogenous CD8+ (Ly5.1-CD8+) or CD4+ (Ly5.1-CD4+) cells was enumerated by flow cytometry.

Tetramer staining

Mice were immunised i.v. as specified, and spleens harvested 6 days later. Each spleen was pressed through a metal sieve, lysed of red blood cells with RCRB, and then washed in tetramer buffer (PBS-1%BSA-1%sodium azide). Cells were resuspended in 4 ml tetramer buffer, and 100 μl of each sample taken for staining. Cells were pelleted and stained with 50 μl of OVA-specific MHC I tetramer (H-2Kb-SIINFEKL) PE (45 min at 37ºC in tetramer buffer), then 50 μl of 2x concentrated anti-CD8 APC and anti-CD44 FITC or anti-CD44 AF700 was added and further incubated (30 min at 4ºC in tetramer buffer). Cells were then washed in tetramer buffer and resuspended in tetramer buffer containing PI. The number of live (PI-) CD8+tetramer+ cells as a proportion of total live CD8+ cells was determined by flow cytometry.

In vivo CTL assay

Mice were immunised i.v. as specified, then 6 days later injected i.v. with target cells. Target cells were prepared as previously described (273). Briefly, lymph nodes and spleens of donor mice were pressed through a metal sieve and lysed of red blood cells with RCRB. Cells were then washed in RPMI-2%FCS, resuspended in complete media (1 ml/mouse harvested) and divided equally into two parts. OT-I peptide (SIINFEKL, 1 μg/ml) was added to one part, while the other received no peptide. Both suspensions were incubated for 1 hour at 37ºC, then washed twice in RPMI-2%FCS and once in PBS-0.1%BSA. Both suspensions were then resuspended in 5 ml PBS-0.1%BSA and labelled with 5 μl of 2.25 nM CFSE (peptide-pulsed, CFSEhi) or 0.5 μl of 2.25 nM CFSE (unpulsed, CFSElo) for 10 min at 37ºC. Cells were then washed twice in RPMI- 2%FCS and once in PBS, and then equal numbers of each population pooled and resuspended in PBS to a final concentration of 100 x 106 pooled cells/ml. Mice were injected i.v. with 10 x 106 pooled cells in 100 μl. Spleens from recipient mice were harvested 18-24 hr after transfer of targets. Each spleen was pressed through a metal sieve, lysed of red blood cells with RCRB, and then washed in RPMI-2%FCS. Cells were resuspended in 3 ml EDTA-BSS-2%FCS and 100 μl taken for analysis. After addition of PI, the ratio of live (PI-) CFSElo to CFSEhi cells was determined by flow

cytometry. The percentage of OVA-specific killing was calculated by [1 − (r unimmunised/r immunised)] × 100 where r = CFSElo/CFSEhi for each sample (273).

Intracellular cytokine staining (ICS)

Mice were immunised i.v. as indicated and spleens harvested 6 days later. Each spleen was pressed through a metal sieve, lysed of red blood cells with RCRB, washed in RPMI-2%FCS and then resuspended in 2 ml complete media. Intracellular cytokine staining was performed with reagents from the BD Cytofix/Cytoperm Plus Fixation/Permeabillization Kit (BD Biosciences, NJ, USA). 100 μl of each sample was transferred to a round-bottom 96 well plate and restimulated with OT-I peptide (SIINFEKL) (final concentration 1 μg/ml), OT-II peptide (final concentration 1 μg/ml) and anti-CD28 (final concentration 2 μg/ml) diluted in complete media to a total volume of 200 μl per well. After 1 hour incubation in a 37ºC incubator with 10%

CO2, 20 μl of BD GolgiPlug pre-diluted 1:100 in complete media was added to each well (final dilution 1:1000). Cells were then incubated at 37ºC with 10% CO2 for a further 5 hours. Cells were then centrifuged at 700g for 7 min, the supernatant discarded and cells washed twice with EDTA-BSS-2%FCS. Cells were then incubated with 50 μl of a cocktail containing anti-FcR, anti-CD4 PE, anti-CD8 PE-Cy7 and anti- CD44 AF700 (30 min at 4ºC in EDTA-BSS-2%FCS). Cells were then washed twice with EDTA-BSS-2%FCS and incubated in 100 μl BD Cytofix/Cytoperm solution (20 min at 4ºC). Cells were then washed twice in 1x BD Perm/Wash buffer (made from 10x stock diluted in EDTA-BSS-2%FCS) and resuspended in 50 μl of anti-IFNγ AF647 diluted in 1x BD Perm/Wash buffer. After overnight incubation (4ºC in the dark), cells were washed once in 1x BD Perm/Wash buffer and resuspended in EDTA- BSS-2%FCS. The proportion of CD8+CD44+ cells producing IFNγ was determined by flow cytometry.

2.15. Tumour model

B16 melanoma tumour model

Mice were immunised as specified, then inoculated with B16-OVA tumours 6 days or 14 days after the final immunisation. Pre-confluent B16-OVA cells were harvested and washed twice, then diluted in PBS and 105 cells injected s.c. into the inguinal region. Mice were monitored daily and tumour size, once palpable, measured daily with calipers. Tumour area was determined as length x width, and mice were killed when area reached 100 mm2.

Measuring OVA expression of harvested tumours

B16-OVA tumours were harvested when tumour area reached 100 mm2. Tumours were chopped into fine fragments with scissors and then digested in 3 ml RPMI- 2%FCS with 0.5 ml DNase/collagenase (final concentration of 0.14 mg/ml DNase and 1 mg/ml collagenase) for 20 min at RT with constant pipetting using a transfer pipette. 350 μl of 0.1M EDTA was then added for a further 5 min with constant pipetting to disrupt cell clusters. Samples were filtered through a metal sieve to remove undigested fragments, underlayed with 1 ml FCS-EDTA and then centrifuged at 700g for 7 min. The supernatant was discarded and the pellet resuspended in 4 ml 1.091 g/cm3 Nycodenz, then overlayed onto 4 ml of 1.091 g/cm3 Nycodenz. 1 ml FCS- EDTA was then layered on top and density separation performed by centrifugation at 1700g for 10 min at 4ºC with slow deceleration. The upper ~5 ml of low-density cells was collected, passed through a 70 μm cell strainer, then washed in EDTA-BSS- 2%FCS. Cells (105) were then stained with anti-FLAG biotin (30 min at 4ºC in EDTA- BSS-2%FCS) to detect the FLAG-tagged OVA, then washed and stained with streptavidin PE (30 min at 4ºC in EDTA-BSS-2%FCS). Cells were then washed and resuspended in EDTA-BSS-2%FCS containing PI, and the PE MFI of live (PI-) cells was determined by flow cytometry.

2.16. Staining primary DCs

Splenic DC enrichment

Splenic DCs were enriched as previously described (74, 274). Briefly, spleens were chopped into fine fragments with scissors and then digested in 6 ml RPMI-2%FCS with 1 ml DNase/collagenase (final concentration of 0.14 mg/ml DNase and 1 mg/ml collagenase) for 20 min at RT with constant pipetting using a transfer pipette. 700 μl of 0.1M EDTA was then added for a further 5 min with constant pipetting to disrupt cell clusters. Samples were filtered through a metal sieve to remove undigested fragments, underlayed with 1 ml FCS-EDTA and then centrifuged at 700g for 7 min. The supernatant was discarded and the pellet resuspended in 5 ml 1.077 g/cm3 Nycodenz, then overlayed onto 5 ml of 1.077 g/cm3 Nycodenz. 1 ml FCS-EDTA was then layered on top and density separation performed by centrifugation at 1700g for 10 min at 4ºC with slow deceleration. The upper ~6 ml of low-density cells was collected and washed in EDTA-BSS-2%FCS, then resuspended in a depletion cocktail containing the following rat anti-mouse mAbs: KT3-1.1, anti-CD3; T24/31.7, anti- Thy1; TER119, anti-erythrocytes; RA36B2, anti-CD45R; and RB6-8C5, anti-Ly6C/G. Cells were incubated with cocktail (10 μl per 106 cells) for 30 min at 4ºC, then diluted to 9 ml with EDTA-BSS-2%FCS, underlayed with 1 ml FCS-EDTA and centrifuged at 700g for 7 min. Cells were then resuspended in 400-600 μl EDTA-BSS-2%FCS and mixed with pre-washed anti-rat IgG magnetic Biomag beads at 10 beads/cell. Cells were incubated with beads for 20 min at 4ºC with continuous rotation on a roller mixer. The mixture was then diluted with EDTA-BSS-2%FCS to 2 ml and placed on a magnet. After allowing the bead-bound cells to attach to the magnet, the DC-enriched supernatant was recovered.

Staining DEC-205 and CD14 on DCs and macrophages

DCs were isolated from spleens as described in “Splenic DC enrichment”. Peritoneal macrophages were obtained by flushing the peritoneum of mice with RPMI-5%FCS. Cells were pre-incubated with 25 μl Fc block (rat IgG and anti-FcR mAb; 10 min at 4ºC), then 25 μl of 2x concentrated anti-DEC-205 biotin, anti-CD14 biotin or isotype

control (IgG2a) biotin was added and further incubated (30 min at 4ºC in EDTA-BSS- 2%FCS). Cells were then washed and bound biotinylated mAbs detected by staining with streptavidin PE, in a cocktail with either: anti-CD11c PE-Cy7, anti-CD8 APC and anti-CD45R Pacific Blue for DCs; or anti-CD11b APC and anti-F4/80 FITC for macrophages (30 min at 4ºC in EDTA-BSS-2%FCS). Cells were then washed and resuspended in EDTA-BSS-2%FCS containing PI, and the PE MFI of live (PI-) cells was determined by flow cytometry.

Binding of targeting constructs to DCs and B cells

Spleens from B6 mice were partially enriched for DCs by DNase/collagenase digestion, EDTA treatment and density separation with a 1.077 g/cm3 Nycodenz gradient as described in “Splenic DC enrichment”. Light density cells were washed in EDTA-BSS-2%FCS, then pre-incubated in 25 μl Fc block (rat IgG and anti-FcR mAb; 10 min at 4ºC) before 25 μl of 2x concentrated XCL1-OVA, NIP-OVA or 10B4-OVA (diluted in EDTA-BSS-2%FCS), or 25 μl of EDTA-BSS-2%FCS alone, was added and further incubated (30 min at 4ºC). Cells were washed and incubated with anti-OVA biotin (30 min at 4ºC in EDTA-BSS-2%FCS), then washed and incubated with a cocktail containing streptavidin PE, anti-CD11c PE-Cy7, anti-CD8 APC and anti- CD45R Pacific Blue (30 min at 4ºC in EDTA-BSS-2%FCS). Cells were then washed and resuspended in EDTA-BSS-2%FCS containing PI. The PE MFI of live (PI-) B cells (CD11c-CD45R+), CD8+ DCs (CD11c+CD8+) and CD8- DCs (CD11c+CD8-) was determined by flow cytometry.

2.17. Measuring ODN binding, uptake and stimulation

Annealing double-stranded ODNs

0.12 nmol of biotinylated 1668 or 2006 was incubated with 1.2 nmol of the complementary strand (fully diester), or an irrelevant non-complementary strand (the forward sequence of 1668, fully diester), in 20 μl of PBS for 10 min at 60ºC, then

allowed to cool to RT. This mixture was then serially diluted in PBS-2%BSA to the indicated concentrations of biotinylated ODN for binding studies.

Binding of ODNs to DEC-205-CHO cells

CHO-K1 cells, mDEC-CHO or hDEC-CHO cells (105 cells) were incubated with the indicated concentrations of biotinylated ODNs (1 hour at 4ºC in PBS-2%BSA). EDTA-containing media was not used due to the potential for EDTA to disrupt DEC- 205 binding. Cells were then washed and incubated with streptavidin PE (30 min at 4ºC in PBS-2%BSA) to detect bound ODN, plus either anti-mouse DEC-205 APC or anti-human DEC-205 FITC to detect expression of the respective DEC-205 molecules. In Figure 3.3D, human DEC-205 expression was instead detected by co-staining with mouse anti-human DEC-205 mAb followed by anti-mouse Ig PE secondary, and ODN binding was detected by streptavidin APC. Following secondary Ab incubation, cells were washed and resuspended in PBS-2%BSA containing PI. Mean fluorescence intensity (MFI) of live (PI-) DEC-205hi cells was determined by flow cytometry. In Figure 5.11, CHO cells were first pre-incubated in 50 μl of the indicated anti-DEC-205 mAbs (20 min at 4ºC in PBS-2%BSA), before 50 μl of 2x concentrated biotinylated ODNs was added and further incubated (30 min at 4ºC in PBS-2%BSA). Cells were then washed and incubated with streptavidin PE (without anti-DEC-205 co-stain; 30 min at 4ºC in PBS-2%BSA), and then washed and resuspended in PBS-2%BSA containing PI for analysis by flow cytometry.

Binding of ODNs to DEC-205 by ELISA

ELISA plates (96 well, Costar, Cambridge, UK) were coated with anti-FLAG (9B4), anti-human DEC-205 (MMRI7) or anti-mouse DEC-205 (NLDC-145) capture Ab at the specified concentrations (4ºC overnight in PBS). Plates were then washed 4 times with ELISA wash buffer (PBS-0.05%Tween20) and incubated with 5 μg/ml mouse or human DEC-205 (4ºC overnight in PBS-1%BSA). Alternatively, if no capture Ab was used, plates were directly coated with 5 μg/ml mouse or human DEC-205 (4ºC overnight in PBS) as the first step. After coating, plates were washed and incubated

with serial dilutions of biotinylated ODN added in duplicate (2 hours at RT in PBS- 1%BSA), then washed and incubated with streptavidin HRP (2 hours at RT in PBS- 1%BSA). Plates were then washed and incubated with ABTS substrate (1.5 hours at RT in the dark) before measuring optical density (O.D.) at 405-490nm. In Figure 5.8, buffers used for the ODN and streptavidin HRP incubation steps and all wash steps were adjusted to pH 6 or pH 5 with hydrochloric acid.

Stimulation of purified B cells in vitro

Spleens were harvested from B6 or DEC-205-/- mice in RPMI-2%FCS, pressed through a metal sieve and centrifuged for 7 min at 700g. The supernatant was discarded and the pellet resuspended in 2 ml 1.091 g/cm3 Nycodenz, then overlayed onto 2 ml of 1.091 g/cm3 Nycodenz. 0.5 ml FCS-EDTA was then layered on top and density separation performed by centrifugation at 1700g for 10 min at 4ºC with slow deceleration. The upper ~2.5 ml of low-density cells was collected and washed in EDTA-BSS-2%FCS, then stained with anti-CD19 FITC and anti-CD3 APC (30 min at 4ºC in EDTA-BSS-2%FCS). Cells were then washed, filtered through a 70 μm cell strainer and resuspended in EDTA-BSS-2%FCS containing PI to a final concentration of approximately 10 x 106 cells/ml. CD19+CD3-PI- live B cells were purified by fluorescence activated cell sorting on a BD FACSAria. Cells were counted and used immediately for stimulation assays. Purified B cells (105 cells/well) were incubated with the indicated concentrations of CpG ODN in a round-bottom 96 well plate in complete media at 37ºC with 10% CO2. After 24 hours, supernatants were collected and frozen for analysis by ELISA, while cell pellets were stained with anti-CD40, CD80, CD86, or MHC II PE (30 min at 4ºC in EDTA-BSS-2%FCS). Cells were then washed and resuspended in EDTA-BSS-2%FCS containing PI. The PE MFI of live (PI- ) cells was determined by flow cytometry.

Activation of DCs and B cells by ODNs in vivo

Two mice per group were injected i.v. with ODNs as specified, and spleens were harvested at specified time points. Spleens were pooled from both mice and DCs

isolated as described in “Splenic DC enrichment”. After digestion with DNase/collagenase and treatment with EDTA, approximately 1/7th of the sample was set aside as splenocytes for B cell analysis, and the remainder was used to isolate DCs. The splenocytes intended for B cell analysis were resuspended in 2 ml 1.091 g/cm3 Nycodenz and overlayed onto 2 ml of 1.091 g/cm3 Nycodenz. 0.5 ml FCS-EDTA was then layered on top and density separation performed by centrifugation at 1700g for 10 min at 4ºC with slow deceleration. The upper ~2.5 ml of low-density cells was collected and washed in EDTA-BSS-2%FCS. Both splenocytes (2 x 106 cells/well) and enriched DCs (1 x 106 cells/well) were stained with anti-CD40, CD80, CD86, MHC II, or DEC-205 PE (30 min at 4ºC in EDTA-BSS-2%FCS). Splenocytes were co-stained with anti-CD45R Pacific Blue to identify B cells, while enriched DCs were co-stained with anti-CD11c PE-Cy7 and anti-CD8 AF647. Cells were then washed and resuspended in EDTA-BSS-2%FCS containing PI. The PE MFI of live (PI-) B cells (CD45R+), CD8+ DCs (CD11c+CD8+) and CD8- DCs (CD11c+CD8-) was determined by flow cytometry.

Binding and uptake of ODN by peritoneal macrophages

Peritoneal macrophages were obtained by flushing the peritoneum of mice with RPMI-5%FCS. Cells (105/well) were washed in EDTA-BSS-2%FCS and incubated with serial dilutions of Cy3- or biotin-labelled ODNs or LPS biotin (in duplicate) in either EDTA-BSS-2%FCS at 4ºC for 30 min or in complete media at 37ºC for 1 hour. Cells were then washed in EDTA-BSS-2%FCS and stained with streptavidin PE (omitted for samples with Cy3-labelled ODNs) and anti-CD11b APC (30 min at 4ºC in EDTA-BSS-2%FCS). Cells were then washed and resuspended in EDTA-BSS- 2%FCS containing PI, and the Cy3 or PE MFI of live (PI-) CD11b+ macrophages was determined by flow cytometry.

2.18. Measuring binding of biological DNA

Binding of YOYO-1-labelled DNA to CHO cells

DNA samples were added at a final concentration of 20 μg/ml to a solution of YOYO- 1 dye (stock 1 mM) pre-diluted 1:1000 in PBS. DNA and dye was incubated for 30 min at 4ºC, then 50 μl of labelled DNA solution was added to 50 μl of CHO-K1 or mDEC-CHO cells (105 cells) in PBS-2%BSA and incubated for 10 min at 4ºC. Cells were then washed in PBS-2%BSA and resuspended in PBS-2%BSA containing PI. The YOYO-1 fluorescence of live (PI-) cells was determined by flow cytometry.

DNA biotinylation

DNA samples were labelled with biotin using the Mirus Label IT kit (Mirus Bio, WI, USA) according to manufacturer’s instructions. DNA (5 μg) was mixed with 5 μl 10x Labelling Buffer A and 5 μl Label IT reagent diluted to a total volume of 50 μl in MilliQ water, and incubated at 37ºC for 1 hour. The labelled sample was then purified using the provided G50 Microspin Purification Columns.

Transfection of CHO cells with RAGE

RAGE-mCitrine plasmid (275) (kindly provided by Dr. Damien Bertheloot and Professor Eicke Latz, University of Bonn, Germany) was amplified in high efficiency competent cells of the E. coli strain JM109 (Promega, WI, USA) and isolated using a QIAprep spin maxiprep kit (Qiagen, Victoria, Australia) according to manufacturer’s instructions. Pre-confluent CHO-K1 cells (107) were harvested and washed twice in RPMI, then resuspended in 500 μl RPMI. Cells were added to a Gene Pulser 0.4 cm electrode gap electroporation cuvette (Bio-Rad, CA, USA) along with 20 μg of RAGE- mCitrine plasmid and gently mixed and incubated at 4ºC for 10 min. Cells were then electroporated at 270V and 950 μF with a Gene Pulser II electroporation system (Bio- Rad, CA, USA), and rested for 10 min at 4ºC. 1 ml of RPMI-5%FCS was added and the suspension overlayed onto 1 ml of FCS in a new tube. Cells were centrifuged at 700g for 7 min, the supernatant removed, and the cells resuspended in RPMI-5%FCS

into a tissue culture flask and placed in a 37ºC incubator with 10% CO2. These RAGE- CHO cells were used in binding assays approximately 24 hours later.

Binding of biotinylated DNA to DEC-205- or RAGE-expressing cells

CHO-K1 cells, mDEC-CHO cells, or RAGE-CHO cells (1 day post-transfection), were incubated with biotin-labelled DNA (labelled as described in “DNA biotinylation”) at the indicated concentrations (1 hour at 4ºC in PBS-2%BSA). Cells were then washed and incubated with streptavidin PE to detect bound DNA (30 min at 4ºC in PBS- 2%BSA). mDEC-CHO cells were co-stained with anti-mouse DEC-205 APC. Cells were then washed and resuspended in PBS-2%BSA containing PI, and the PE MFI of live (PI-) DEC-205hi or mCitrinehi cells was determined by flow cytometry. In variations of this experiment, incubation of cells with DNA was performed in PBS containing additional factors such as HMGB1, granulin, LL-37, EDTA, or mouse serum at the specified concentrations.

Binding of HMGB1, granulin and LL-37 to DNA by ELISA

ELISA plates (96 well, Costar, Cambridge, UK) were coated with HMGB1, granulin, LL-37, OVA or mouse DEC-205 in duplicate at the specified concentrations (4ºC overnight in PBS). Plates were washed and blocked with PBS-1%BSA (30 min at RT), then incubated with biotinylated CpG ODN 1668 or DNA (2 hours at RT in PBS- 1%BSA). Plates were washed and incubated with streptavidin HRP (2 hours at RT in PBS-1%BSA), then washed and incubated with ABTS substrate (1.5 hours at RT in the dark) before measuring optical density (O.D.) at 405-490nm.

Fractionation of mouse serum

Mouse serum used in binding studies was collected from naïve B6 animals. To separate serum into fractions by size, serum was centrifuged at 4000g in a 3 kDa cut- off centrifugal filter unit (Millipore, MA, USA). The flow-through was used as the less than 3 kDa fraction, while the concentrate was diluted with PBS to the original

volume then used as the greater than 3 kDa fraction. For treatment with trypsin, 100 μl of serum or serum fraction was mixed with 100 μl of 0.5 mg/ml trypsin overnight on a 37ºC shaker. For heat inactivation, serum or serum fractions were diluted to the final concentration of 10% in PBS, then heated to 90ºC for 5 min, before being cooled to 4ºC and used for staining.

2.19. Dead cell binding assays

Freeze/thawing cells

CHO-K1 or HEK293F cells were freeze/thawed by four cycles of freezing cell pellets in dry ice then thawing in a 37ºC water bath. Some samples, as indicated, were further fragmented after freeze/thawing by sonication for approximately 30 seconds.

Dead cell binding to soluble protein

Live or dead (freeze/thawed) cells (5 x 105) were washed in PBS-1%BSA and incubated with 10 μg/ml of soluble FLAG-tagged mouse DEC-205, human DEC-205 or mouse Clec9A (20 min at RT in PBS-1%BSA), then washed and incubated with anti-FLAG FITC (30 min at 4ºC in PBS-1%BSA). Cells were then washed and resuspended in PBS-1%BSA containing PI and FITC fluorescence was determined by flow cytometry. For pH 6 samples, PBS-1%BSA was adjusted to pH 6 with hydrochloric acid and used for all steps from the first wash to the final suspension. For pH 7 samples, normal PBS-1%BSA media was used.

Dead cell binding to cell-bound protein

Live CHO-K1, Clec9A-CHO, mDEC-CHO and hDEC-CHO cells, and dead cells (HEK293F cells labelled with CFSE, PKH26 or unlabelled, then freeze/thawed or freeze/thawed and sonicated) were resuspended at 107 cells/ml in PBS at pH 7 or pH 6. Dead cells (50 μl, 5 x 105) were incubated with CHO-K1, Clec9A-CHO, mDEC-

CHO or hDEC-CHO cells (50 μl, 5 x 105) and 0.5 μl PI (20 min at RT). CFSE or PKH26 fluorescence of CHO cells was determined by flow cytometry.

2.20. Other ELISA assays

Measuring anti-OVA Ab titres by ELISA

ELISA plates (96 well, Costar, Cambridge, UK) were coated with 10 μg/ml OVA (4ºC overnight in PBS). Plates were washed 4 times in ELISA wash buffer and incubated with serially diluted plasma samples added in duplicate (4ºC overnight in ELISA blocking buffer (5% milk powder in PBS)). Plates were then washed and incubated with anti-mouse IgG HRP (4ºC overnight in ELISA blocking buffer), then washed and incubated with ABTS substrate (1.5 hours at RT in the dark), before measuring optical density (O.D.) at 405-490nm. Ab endpoint titre was defined as the highest dilution to produce an O.D. reading above background. A positive control was included on each plate to monitor plate-to-plate reproducibility.

IL-6 ELISA

IL-6 concentration in culture supernatants or serum samples was determined using the Mouse IL-6 ELISA Ready-SET-Go! kit (eBioscience, CA, USA), according to manufacturer’s instructions. Briefly, ELISA plates (96 well, Costar, Cambridge, UK) were coated with capture Ab (4ºC overnight in supplied coating buffer), then plates were washed 4 times with ELISA wash buffer and blocked with supplied diluent buffer (1 hour at RT). Plates were then incubated with supernatants or serum (used either neat or diluted with the supplied diluent buffer) and the supplied IL-6 standard, added in duplicate (4ºC overnight). Plates were then washed and incubated with detection Ab followed by avidin HRP according to manufacturer’s instructions. Plates were then washed and incubated with ABTS substrate (1.5 hours at RT in the dark), before measuring optical density (O.D.) at 405-490nm. IL-6 concentration was interpolated from the standard curve using asymmetric sigmoidal non-linear regression on log transformed (X=log(X)) data in Prism (GraphPad Software, CA, USA).

IL-12 ELISA

ELISA plates (96 well, Costar, Cambridge, UK) were coated with 0.7 μg/ml anti-IL-12 capture Ab (R29A5; 4ºC overnight in PBS), then washed 4 times with ELISA wash buffer and blocked with PBS-0.3%BSA (2 hours at RT). Plates were then incubated with serum and IL-12 standard, added in duplicate (4ºC overnight in PBS-0.3%BSA). Plates were washed and incubated with anti-IL-12 biotin detection Ab (C17.8; 4 hours at RT in PBS-0.3%BSA), then washed and incubated with streptavidin HRP (2 hours at RT in PBS-0.3%BSA). Plates were then washed and incubated with ABTS substrate (1.5 hours at RT in the dark) before measuring optical density (O.D.) at 405-490nm. IL-12 concentration was determined by calculating the slope of the linear region of the standard curve in Microsoft Excel (Microsoft Corporation, WA, USA).

CD14 ELISA

ELISA plates (96 well, Costar, Cambridge, UK) were coated with a fixed amount or serial dilutions of CD14-Fc, CD14-His, mouse DEC-205 or no protein, added in duplicate (4ºC overnight in PBS). Plates were washed and blocked with PBS-1%BSA (30 min or 1 hour at RT), then incubated with biotinylated CpG ODN 1668, LPS biotin or anti-CD14 biotin (2 hours at RT in PBS-1%BSA). Plates were washed and incubated with streptavidin HRP (1 hour at RT in PBS-1%BSA), then washed and incubated with ABTS substrate (1.5 hours at RT in the dark) before measuring optical density (O.D.) at 405-490nm.

Measuring the presence of 10B4-OVA in serum by ELISA

ELISA plates (96 well, Costar, Cambridge, UK) were coated with 0.1 μg/ml Clec9A (4ºC overnight in PBS). Plates were washed and blocked with PBS-1%BSA (30 min at RT), then incubated with serially diluted serum or purified 10B4-OVA standard, added in duplicate (2 hours at RT in PBS-1%BSA). Plates were washed and incubated with anti-rat IgG biotin (1 hour at RT in PBS-1%BSA), then washed and incubated

with streptavidin HRP (1 hour at RT in PBS-1%BSA). Plates were then washed and incubated with ABTS substrate (1.5 hours at RT in the dark) before measuring optical density (O.D.) at 405-490nm. 10B4-OVA concentration was interpolated from the standard curve using asymmetric sigmoidal non-linear regression on log transformed (X=log(X)) data in Prism (GraphPad Software, CA, USA).

Measuring the anti-construct reactivity of Ab responses by ELISA

ELISA plates (96 well, Costar, Cambridge, UK) were coated with the specified concentrations of OVA, mouse 10B4, rat 10B4, GL117, rat 10B4-OBI or XCL1-OVA (4ºC overnight in PBS). Plates were then washed 4 times in ELISA wash buffer and incubated with serially diluted plasma samples from immunised mice, added in duplicate (4ºC overnight in ELISA blocking buffer). Plates were washed and incubated with anti-mouse IgG HRP (4ºC overnight in ELISA blocking buffer). Alternatively, as indicated in the Figure legend, plates were incubated with a cocktail of anti-mouse IgG2a HRP, anti-mouse IgG2b HRP and anti-mouse IgG2c HRP (4ºC overnight in ELISA blocking buffer). Plates were then washed and incubated with ABTS substrate (1.5 hours at RT in the dark) before measuring optical density (O.D.) at 405-490nm.

2.21. Miscellaneous protocols

Agarose gel electrophoresis

5 g of agarose was dissolved in 100 ml of 0.5x Tris-borate-EDTA (TBE) buffer and 10 μl of GelRed nucleic acid stain added before pouring the gel and allowing to set. 1 μl (1 nmol) of each ODN was mixed with 1 μl 6x Blue/Orange loading dye and 4 μl water and loaded into the gel well. 5 μl of PCR markers was included in a separate lane. The gel was run at 100V for approximately 1 hour in 0.5x TBE buffer, and bands visualised with a UV Gel Documentation System (Fisher Biotec, WA, Australia)

Collection of plasma

Cardiac or submandibular bleeds were collected in heparin, centrifuged at 5000 rpm in a benchtop microfuge (Heraeus Biofuge Pico, Thermo Scientific, MA, USA) for 10 min and the supernatant (plasma) collected and frozen for ELISA analysis.

Collection of serum

Cardiac bleeds were collected (without heparin) and allowed to clot overnight at 4ºC before removing the clot. Samples were then centrifuged at 5000 rpm in a benchtop microfuge (Heraeus Biofuge Pico, Thermo Scientific, MA, USA) for 10 min and the supernatant (serum) collected and frozen for ELISA analysis.

Chapter 3: Using DEC-205 to modulate CpG ODN function

Abstract

CpG oligonucleotides (ODNs) are potent Th1-promoting adjuvants and agonists of TLR9. Recent work from our lab has discovered DEC-205 to be a surface receptor that mediates the binding and uptake of CpG ODNs in APCs. Importantly, interaction with DEC-205 was found to be essential for the optimal stimulatory activity of CpG ODNs. Building upon these findings, in this study we sought to further investigate the nature of the interaction between CpG ODNs and DEC-205 and its implications for modulating the stimulatory activity of CpG ODNs. The molecular properties of ODNs necessary for binding to DEC-205 were examined by comparing the DEC-205- binding capacity of various types of ODNs. We found that a linear phosphorothioated ODN backbone is necessary for binding to DEC-205, and that more efficient binding was seen when ODNs had more thymidine residues and were at least 14 bases in length. Next, the contribution of DEC-205 to the stimulatory activity of various ODNs (identified by the numbers 2006, 21798 and 518477) was investigated by comparing responses in wild-type and DEC-205-deficient mice. We found DEC-205 to play a significant role in the stimulatory activity of CpG ODN 2006, but not 21798, a CpG ODN with a weak capacity to bind DEC-205. Interestingly, DEC-205 also appears to have a slight influence on the stimulatory activity of a non-CpG ODN, 518477. Finally, we demonstrated that the ability of DEC-205 to mediate the stimulatory effects of CpG ODNs can be utilised to generate stronger adjuvants. We show that improving the capacity of 21798 to bind DEC-205 significantly enhances its potency as an adjuvant for the stimulation of CTL responses.

Introduction

Classes of CpG oligonucleotides

Synthetic oligodeoxynucleotides (ODNs) containing unmethylated cytosine guanosine motifs (CpG), often referred to as CpG ODNs or simply CpG, are a Toll- like receptor 9 (TLR9) agonist commonly used as an adjuvant. As the name suggests, ODNs consist of a short sequence of deoxyribonucleotides, which can be linked by either phosphodiester (diester) bonds, as in normal mammalian DNA, or phosphorothioate (thioate) bonds, or a combination of the two. The phosphorothioate modification, which replaces a non-bridging oxygen atom with sulphur, is routinely used to improve the stability of synthetic ODNs, as it confers nuclease resistance, and a much greater in vivo half-life (276-278). Thioate ODNs are known to exhibit greater non-specific binding to a variety of cell types (279). Notably, besides their use as adjuvants, thioated ODNs are also used in gene therapy approaches as anti-sense ODNs (ASOs) designed to inhibit the expression of specific genes (280).

The stimulatory activity of CpG ODNs is dependent on the presence of the unmethylated CpG motif, as activity is lost if the CpG dinucleotide is reversed to GpC, or if the cytosine is methylated (281, 282). Mammalian DNA, which is non- stimulatory when tested in the same assays, has a lower percentage of CG dinucleotides, and most of them have a methylated cytosine (283). This led to the suggestion that unmethylated CpG motifs are recognised by the immune system as a foreign PAMP. This is further supported by the fact that stimulatory CpG motifs were first identified in bacterial DNA extracts (284). While unmethylated CpG motifs are not completely absent from the mammalian genome, it is believed that other mechanisms prevent reactivity against self-DNA. Such mechanisms include the presence of inhibitory sequences in mammalian DNA (285), and the localisation of TLR9 in endosomes, where access to self-DNA should be limited under normal conditions (286).

Synthetic CpG ODNs are categorised into classes based on certain structural features, which are associated with particular immunostimulatory properties. Class A ODNs (A-ODNs, also known as D-ODNs) are composed of at least one CpG motif contained within a diester palindromic sequence flanked by thioate poly-G regions at both 3’ and 5’ ends. The hybrid diester-thioate backbone is based on studies that showed the stimulatory activity of A-ODN is strongest if the CpG motif and flanking bases are diester, but activity is also enhanced if the end regions are thioate instead of diester, presumably by protecting the ODN from nuclease degradation (287). A-ODNs characteristically stimulate high levels of type I IFN production from pDCs, although they are relatively poor at inducing B cell and DC maturation (287-289). Stimulation by A-ODNs also leads to the activation of NK cells, Th1 CD4+ T cells and CD8+ T cells, which is mediated by the IFNα secreted from pDCs (288, 290, 291). A-ODNs are known to form higher order aggregates, due to the poly-G flanking regions that form G-tetrads, as well as the central palindrome. The aggregated structure of A-ODNs appears to be central to their stimulatory activity, as removing the poly-G tails, disrupting G-tetrad formation with 7-deaza guanosine or removal of the central palindrome impairs their capacity to stimulate IFN production (287, 288).

Class B ODNs (B-ODNs, also known as K-type ODNs) are single-stranded, monomeric molecules consisting of one or more CpG motifs in a fully thioated backbone. In contrast to A-ODN, they do not induce high levels of type I IFN production or NK cell activation, but instead are more efficient at inducing B cell and DC activation and the production of pro-inflammatory cytokines such as IL-12 and TNFα (287, 288, 292-294). B-ODNs are also particularly potent at stimulating B cells to proliferate and secrete high levels of IL-6 and IgM (287, 289). As B-ODNs are linear and do not form secondary structures, the influence of the primary nucleotide sequence on ODN potency can be more readily examined. Through comparison of various B-ODN sequences, the optimal stimulatory CpG motif was determined to be purine-purine-C-G-pyrimidine-pyrimidine (281). ODNs containing the motif GACGTT were found to be most effective at activating mouse B cells (281). This optimal mouse motif is used in B-ODNs commonly considered to be optimally stimulatory in mouse, such as 1668 and 1826. For human cells, the optimal

stimulatory motif was instead GTCGTT (295). However, in contrast with mouse, CpG ODNs with only one optimal motif were ineffective at stimulating human peripheral blood mononuclear cells (PBMCs). Instead, optimal activity appeared to require two or three CpG motifs. The CpG motifs were more effective if proceeded by TC on the 5’ side, and each spaced apart with TT rather than directly next to each other. CpG ODN 2006, which incorporated all of these features, was found to be the most optimal ODN for the stimulation of human cells (295).

Class C ODNs (C-ODNs), have properties that are intermediate between A- and B- ODN. They are fully thioated, similar to B-ODN, but are capable of forming dimers or stem-loop structures due to the inclusion of a palindromic sequence (296, 297). The stimulatory effects of C-ODN also incorporate features from both A- and B-ODN. Like A-ODNs, C-ODNs stimulate high levels of IFNα production from pDCs, though to a lesser degree than A-ODNs. But unlike A-ODNs, C-ODNs are also capable of stimulating B cell maturation, proliferation and IL-6 secretion in a similar manner to B-ODNs (296-298). It has been suggested that C-ODNs are even more potent stimulators of B cells in PBMCs than B-ODNs, due to a combination of direct and indirect stimulation of B cells mediated by the activation of pDCs (296). Similar to A- ODNs, the ability of C-ODNs to induce type I IFN production is dependent on their secondary structure. Reducing the capacity of C-ODN to form dimers, by removal or shortening of the palindrome, significantly impaired their ability to induce IFNα (298). On the other hand, extending the length of the palindrome resulted in a corresponding increase in IFNα stimulating ability, but did not appear to affect B cell stimulating capacity (299).

Recently, a fourth class known as P-ODNs has been described. P-ODNs are similar to C-ODNs in that they stimulate IFNα production like A-ODNs, whilst promoting B cell and DC activation like B-ODNs (300). However, P-ODNs have certain structural modifications that appear to enhance their potency compared to C-ODNs. Where C- ODNs have one palindromic sequence that allows the formation of dimers, P-ODNs have 2 independent palindromes that allow the formation of both dimers and larger multimeric structures. They also have diester linkages at each CpG residue, which was

found to be more stimulatory than a fully thioated backbone (300). While P-ODNs appeared comparable to C-ODNs at promoting B cell proliferation and activation, they induced greater IFNα production from human PBMCs than all other classes of ODNs (300). The most potent P-ODN identified in this study, 21798, was shown to also be effective in vivo, inducing significantly higher plasma levels of IL-6, IP-10 and IFNα than C-ODN after subcutaneous injection in mice (300).

Use of CpG ODN adjuvants in humans

In mice, multiple immune cell types including B cells, cDCs, pDCs, macrophages and monocytes express TLR9 and thus can respond to CpG ODN stimulation (284). By contrast, the expression of TLR9 is far more limited in humans, being restricted to B cells and pDCs (186). Therefore, CpG ODNs can only directly stimulate B cells and pDCs in humans, though this could lead to the indirect activation of other cell types mediated by, for example, the secretion of IFNα (186, 301, 302). This biological difference could theoretically cause differences in responsiveness to CpG ODNs between mice and humans. Despite this, CpG ODNs have been shown to effectively stimulate the human immune system in clinical trials.

The ability of CpG ODNs to promote strong Th1 and CTL responses, which common adjuvants such as alum fail to do, has driven significant interest in their potential use as vaccine adjuvants (303-305). The development of CpG ODN adjuvants for use in humans have largely focused on B-ODNs, as A-ODNs form highly heterogeneous aggregated structures, which, while necessary for their function, is less suitable for producing a consistent therapeutic product (297). The most common CpG ODN chosen for use in clinical trials is 2006, based on the work of Hartmann and Krieg that established 2006 as the most potent stimulator of human cells (295).

The predominant use of CpG ODN in current clinical trials is for cancer therapies, where their propensity for promoting Th1 immunity and the induction of CTL responses is highly desirable. Peptide vaccines targeting a wide range of cancers, including melanoma, lung cancer, B cell lymphoma, esophageal cancer and various

cancers expressing NY-ESO-1, have been trialled with CpG ODN as adjuvant (306, 307). Overall, while many of these studies showed promising enhancement of immune parameters such as Ag-specific T cell and Ab responses, this often did not translate to significant clinical benefit (306-308). This is possibly due to the immunosuppressed state of later-stage cancer patients, which hampers vaccines that rely on the generation of effective immune responses. Strategies to bypass this immunosuppression, such as combination with other therapeutic agents, or treatment earlier in the course of disease, may be required to see significantly improved clinical responses.

Numerous trials have also utilised CpG ODNs as an adjuvant to enhance immune responses against a variety of infectious agents, including leishmania, malaria, anthrax, measles, influenza, herpes simplex virus, hepatitis B virus and tetanus (306, 307, 309). In contrast to cancer trials, these vaccines tend to be far more effective at inducing protective immunity, and while no CpG ODN adjuvanted vaccines have been licensed for human use as yet, recent results suggest it may only be a matter of time. For example, in a Phase III trial of a vaccine against hepatitis B surface Ag, the use of CpG ODN adjuvant caused more rapid induction of significantly greater protective titres compared with the currently licensed hepatitis B vaccine, Engerix-B, which is adjuvanted with alum. Moreover, these improved responses were achieved with only 2 immunisations of the CpG ODN-adjuvanted vaccine compared with 3 doses of Engerix-B (310, 311).

A common concern with most adjuvants is the potential for toxicity, particularly as their systemic administration can cause broad non-specific activation and inflammation. CpG ODNs appear to be well tolerated, with no severe adverse events reported in trials. The most commonly reported adverse effects are localised inflammatory or flu-like symptoms, which typically resolve within a few days (306, 307, 309). Another concern is that administration of CpG ODNs could trigger anti- DNA immune responses that could lead to autoimmunity. Multiple studies have shown that healthy human volunteers treated with CpG ODNs show no signs of autoimmunity, and most show no change in serum auto-Ab levels (306). However, a

higher rate of autoimmune responses detected in cancer patients undergoing treatment with CpG ODNs cautions that some patient groups may be more susceptible, and should be closely monitored (306, 307, 309).

TLR9 determines the potency of species-specific CpG ODNs

Studies to define the optimal stimulatory motifs and resultant immune effects of CpG ODNs have certainly been essential to enable the use of these agents as adjuvants in humans. To support further development of CpG ODNs as therapeutic agents, it is also important to investigate the biochemical mechanisms that underpin these findings. Understanding the factors that govern ODN potency could allow the design of improved adjuvants with greater stimulatory effects and less off-target toxicity.

In 2000, Hemmi et al. first identified that TLR9 was the receptor for CpG ODN, by observing a lack of responsiveness to CpG ODN stimulation in TLR9-deficient mice (312). While this report only examined one class of CpG ODN (B-ODN), subsequently all classes of CpG ODNs were found to require TLR9 to exert their stimulatory effects (298, 313). Naturally, this discovery led to investigations of whether the previously observed requirements for potent CpG ODN activity were related to TLR9 binding. Indeed, TLR9 was found to mediate the species-specific effects of CpG ODN. By comparing human optimized 2006 with mouse optimised 1668, it was found that cells transfected with human TLR9 were more responsive to 2006, and less to 1668, while cells transfected with mouse TLR9 were more responsive to 1668, and less to 2006. This paralleled the preferential stimulation of human PBMCs with 2006 and murine splenocytes with 1668 (314). Furthermore, by stepwise substitution of bases to convert 1668 to become more similar to 2006, progressively enhanced activation of human TLR9 was seen, with a concurrent decrease in activation of mouse TLR9 (314). It should be noted that these species-specific effects have only been observed with fully thioated ODNs. When synthesised with natural diester backbones, the ODNs 1668 and 2006 appear to have a similar capacity to stimulate both mouse and human cells (315).

A-ODNs and B-ODNs activate distinct downstream signalling pathways

It was then considered whether the distinct stimulatory properties of different classes of ODN may also relate to differences in their interaction with TLR9. The differential effects of A-ODNs and B-ODNs were found to stem from the triggering of distinct pathways downstream of TLR9 activation. The induction of IFNα secretion from pDCs upon stimulation with A-ODN was found to be critically dependent on the activation of IRF7 downstream of TLR9 (316). On the other hand, stimulation by B- ODNs is independent of IRF7, but instead involves the activation of mediators such as NF-κB and IRF5 to promote inflammatory cytokine production (316, 317). Further investigations have revealed that the propensity for A-ODNs or B-ODNs to activate divergent downstream signalling pathways is due to their differential intracellular trafficking, which is in turn regulated by their structural properties.

Honda et al. studied the movements of A-ODN and B-ODN in DCs using dextran- FITC to identify early endosomes and LysoTracker to identify lysosomes (318). Within 90 minutes of incubation with pDCs, B-ODNs were found in lysosomes but not in early endosomes. Strikingly, A-ODNs showed the opposite pattern, localising in early endosomes but not lysosomes. A-ODNs were also found to co-localise with IRF7 in pDCs, but not in cDCs. This aligns with the capacity of pDCs, but not cDCs, to produce type I IFN in response to A-ODNs (316). Interestingly, in cDCs, both A- ODNs and B-ODNs were localised to lysosomes, suggesting that retention within early endosomes is necessary to activate IRF7 and induce type I IFN production. In support of this hypothesis, A-ODNs manipulated to enter early endosomes in cDCs by formation of complexes with 1,2-dioleoyloxy-3-trimethylammonium-propane (DOTAP) resulted in IRF7 activation and IFNα secretion. Furthermore, even B-ODNs in complex with DOTAP could induce high levels of IFNα secretion from pDCs, but not from cDCs. Consistent with the proposed model, B-ODN complexed with DOTAP localised to early endosomes only in pDCs but not in cDCs. These findings showed that the capacity of CpG ODNs to induce activation of the IRF7 pathway and

type I IFN secretion is dependent on their localisation within early endosomes, which is differentially regulated between different classes of ODN and in different cell types.

It was further determined that the differential intracellular localisation of A-ODNs and B-ODNs is regulated by their structural properties, specifically, whether they are single-stranded (ss) or form multimers. While normal ss B-ODNs localise to late endosomes and lysosomes, B-ODNs complexed with polymyxin B (PMXB) to form larger particles were no longer found in late endosomal compartments and instead were observed in early endosomes. Similarly, when A-ODNs were forced to assume a single-stranded structure by heating and flash cooling, or inhibition of G-tetrad formation, they were no longer retained in early endosomes and instead localised to late endosomes (319). Notably, the changes in intracellular location caused corresponding changes to the stimulatory outcome. B-ODNs in multimeric complexes with PMXB could induce a similar level of IFNα production from pDCs as A-ODNs, but had a corresponding loss in their ability to stimulate DC maturation, a hallmark of normal ss B-ODNs. Conversely, ss A-ODNs were more potent at promoting DC maturation, but lost their IFNα-stimulating ability, which could be reversed by incorporation into PMXB complexes. The impact of intracellular localisation on the stimulatory effects of CpG ODN was further confirmed with C- ODNs, to exclude the potential for artefacts caused by the manipulation of ODN properties. C-ODNs are normally found in both early endosomes and lysosomes, consistent with their ability to induce the immune effects associated with both A- ODNs and B-ODNs. By encapsulating C-ODNs in liposomes that release their cargo below pH 5.75, C-ODNs could be excluded from early endosomes and restricted to late endosomes and lysosomes. Doing so abolished IFNα production, but the capacity to induce DC maturation remained intact (319).

These studies have established that the structural properties of CpG ODNs play a critical role in determining their immunostimulatory activity, by controlling their intracellular localisation. However, the mechanism responsible for directing CpG ODNs with certain structural features into particular intracellular compartments is less clear. One potential explanation is that endocytic receptors that mediate the

uptake of CpG ODNs differentially recognise specific ODN structures. CpG ODNs were initially believed to be taken up non-specifically, but the capacity to distinguish between different classes of ODN implies the involvement of specific receptors. The identification of receptors involved in regulating the activity of different CpG ODNs would not only provide a more complete understanding of how CpG ODNs exert their stimulatory effects, but could also provide a means of modulating these effects by manipulating the receptor-ODN interaction.

Surface receptors suggested to mediate the uptake of ODN

Several receptors have been proposed to mediate uptake of CpG ODNs, most being members of the scavenger receptor (SR) family. SRs are a very diverse family of receptors that are predominantly expressed on macrophages, the prototypical member of which is SR-A (320). SR-A was observed to bind A-ODNs in a sequence- independent manner mediated by their poly-G tails (321, 322). However, SR-A binding does not appear to be the dominant method of CpG ODN uptake in macrophages (323, 324), and studies in SR-A-deficient mice have indicated that it is not required for stimulation of macrophages by CpG ODNs or bacterial DNA (325). Another scavenger receptor, MARCO or SR-A2, appeared to enhance the production of IL-12 from macrophages stimulated by CpG ODN (326). However, MARCO deficiency did not impact CpG ODN uptake by macrophages, so the exact mechanism by which MARCO influences the activity of CpG ODN is not known (326).

CXCL16 is another SR that was found to bind A-ODN, enhancing its uptake and stimulation in cells expressing TLR9 (327). As for other SRs, this binding was mediated by the poly-G tails of A-ODN, so minimal binding to B-ODN was observed. CXCL16 transfected into HEK293 cells was found to co-localise with A-ODNs, but not B-ODNs. While B-ODN could activate NF-κB signalling in HEK293 cells transfected with TLR9 alone, A-ODN could only elicit a response if co-transfected with TLR9 and CXCL16, demonstrating a role for CXCL16 in mediating the stimulatory activity of A-ODNs. Correlation between CXCL16 expression and responsiveness to A-ODNs was also demonstrated in primary human immune cells. B

cells were uniformly CXCL16 negative and did not respond to A-ODN stimulation, whereas the 20-45% of pDC that were found to express surface CXCL16 were capable of producing IFNα in response to A-ODN. CXCL16-negative pDCs did not produce IFNα in response to A-ODN, although they were equally efficient as the CXCL16- positive population at producing TNFα in response to B-ODN, providing strong evidence that CXCL16 specifically mediates the stimulatory activity of A-ODNs (327). Further investigations showed that the stimulation of pDCs by A-ODNs could be inhibited with oxidised low density lipoprotein (ox-LDL), a SR ligand, and enhanced by treatments that increased CXCL16 surface expression (328).

While CXCL16 and other SRs may mediate the stimulatory effects of A-ODNs, they do not bind B-ODNs. As B-ODNs are far more commonly used in the clinic, it is of great interest to identify receptors that may mediate the activity of B-ODNs. Several receptors have been implicated in the binding of B-ODN, including mannose receptor (MR), killer immunoglobulin–like receptor (KIR) KIR3DL2, receptor for advanced glycation end-products (RAGE), CD14 and DEC-205.

MR was found to play a role in the uptake and intracellular trafficking of B-ODNs in macrophages. However, this was only observed in wild-derived mouse strains, as MR did not appear to impact CpG ODN uptake in macrophages from inbred laboratory strains such as the classical C57BL/6 mice (329). Thus, the role of MR in the uptake of CpG ODNs appears to be variable between different mouse strains, and it is not known whether MR interacts with CpG ODNs in humans.

KIR3DL2 was demonstrated to mediate the uptake of B-ODNs and C-ODNs in human NK cells (330). While only KIR3DL2+ NK cells and not KIR3DL2- NK cells were found to produce TNFα and IFNγ in response to C-ODN stimulation, whether this was specifically due to KIR3DL2-mediated uptake, rather than other properties that may differ between the two subsets, was not clearly addressed. Furthermore, resting NK cells are not directly activated by CpG ODNs and only gain the capacity to respond to CpG ODN stimulation upon activation by pro-inflammatory cytokines such as IL-12 or IL-8 (331). This is likely because resting NK cells express negligible

levels of TLR9, which possibly increase upon activation (186, 331). Therefore, it is not clear how much the direct uptake of CpG ODNs by NK cells may contribute to their stimulatory activity in vivo.

RAGE was shown to specifically bind A-ODNs, B-ODNs and C- ODNs, with slightly higher affinity for A-ODNs and C-ODNs than B-ODNs (275). Binding of CpG ODNs to RAGE appeared to promote their uptake for the activation of TLR9, as HEK293 cells transfected with both TLR9 and RAGE showed greater uptake of CpG ODNs and NF-κB activation compared with cells transfected with TLR9 alone. However, a limitation of this study is that the interaction between RAGE and CpG was only examined in transfectant cell lines, and not in immune cells endogenously expressing TLR9. Although RAGE deficiency in mice was shown to reduce inflammatory responses in the lung to intranasally administered CpG ODN, this was attributed to the activity of RAGE on lung epithelial cells, which express much higher levels of RAGE than any other cell type. It remains to be seen whether RAGE significantly affects the response of immune cells to CpG ODN stimulation.

A report investigating the capacity of CD14 to mediate the binding and uptake of CpG ODNs was significant for its demonstration of a functional consequence of CD14- mediated CpG ODN uptake in APCs (332). Bone marrow-derived DCs and macrophages from CD14-deficient mice showed significantly impaired responses to in vitro B-ODN stimulation. In vivo, CD14-deficient mice exhibited decreased cytokine production and recruitment of neutrophils to the peritoneal cavity following intraperitoneal injection of B-ODN. The role of CD14 in mediating the response to CpG ODNs will be addressed in detail in Chapter 4.

Another receptor proposed to mediate the activity of B-ODNs in APCs is DEC-205. A previous study published by our group showed that DEC-205 mediates the uptake of B-ODNs in primary mouse B cells and DCs (264). DEC-205 could also bind C-ODNs, but had limited ability to bind A-ODNs. In the absence of DEC-205, the activation of both B cells and DCs by B-ODNs was significantly impaired. DEC-205-deficient mice also produced significantly less serum IL-6 and IL-12 in response to B-ODN

administration, and were less efficient at generating Ag-specific CTL responses after immunisation with B-ODN as adjuvant. Importantly, human DEC-205 was also shown to bind B-ODNs with high affinity, implying that these findings could be translated into humans.

An interesting characteristic of DEC-205 is its predominant expression by immune cells, including a high level on DCs and lower levels on B cells, the two main cell types that are directly activated by CpG ODNs. Thus, the uptake of CpG ODNs by DEC-205 would theoretically target CpG ODNs more specifically to the cell types upon which they exert their stimulatory effects, and reduce the amount taken up by irrelevant cell types. This could explain the significant degree to which the systemic stimulatory effects of CpG ODNs were seen to be dependent on DEC-205 (264). We hypothesize that improving the DEC-205-binding capacity of CpG ODNs could enhance their stimulatory potential, by enhancing delivery to and activation of APCs. This Chapter will further investigate the correlation between DEC-205-binding and the stimulatory activity of CpG ODNs, and explore whether improving the DEC-205-binding capacity of CpG ODNs is a viable strategy for generating more potent adjuvants.

First, the molecular properties of CpG ODNs that confer DEC-205 binding will be characterised. The impact of ODN length, sequence, and structure on their ability to bind both human and mouse DEC-205 will be examined. This will allow us to better predict what types of ODNs bind to DEC-205, and what types of modifications could be made to modulate DEC-205 binding.

Next, the consequences of DEC-205 binding on the immunostimulatory capacity of CpG ODNs will be investigated. To determine whether DEC-205 binding impacts the activity of all CpG ODNs, as was reported for 1668 (264), the capacity for various ODNs to stimulate wild-type or DEC-205-deficient APCs will be compared.

Finally, the potential application of these findings to the rational design of more potent CpG ODN adjuvants will be explored. The accumulated knowledge presented in this Chapter and by Lahoud et al. (264) suggests a correlation between DEC-205

binding and CpG ODN stimulatory potency. We will investigate whether modifications to enhance the DEC-205-binding capacity of CpG ODNs can improve their potency as adjuvants. We aim to demonstrate that this novel method of modulating the activity of ODNs could be a viable strategy to design stronger adjuvants.

Results

3.1. Requirements for the binding of CpG ODNs to DEC-205

3.1.1. Examining the capacity of DEC-205 to bind phosphorothioated and phosphodiester ODNs

Our previous data indicated that not all CpG ODNs bound DEC-205 equivalently. B- ODNs were found to have strong binding to both human and mouse DEC-205, while A-ODNs bound poorly (264). A key difference between these two classes of ODNs is that the DNA backbone of B-ODNs is fully thioated, while the backbone of A-ODNs is partially thioate and partially diester. Thus, we sought to determine whether a thioated ODN backbone is necessary to mediate DEC-205 binding. To test this, we examined the binding of either fully thioated or fully diester versions of the ODNs 1668, 2006 and 1826 to mouse or human DEC-205 expressed on the surface of CHO- K1 cells. All three thioated ODN showed strong binding to both mouse and human DEC-205 (Figure 3.1). By contrast, ODNs with diester backbones showed no binding above background. Notably, background staining of the thioate ODNs was higher than that of the diester ODNs. This was not unexpected, as enhanced non-specific binding is a well-documented attribute of thioated ODN (279).

Figure 3.1. Binding of diester and thioated ODNs to DEC-205-expressing CHO-K1 cells. CHO-K1 cells (CHO) or CHO-K1 cells expressing (A) mouse DEC-205 (mDEC-CHO) or (B) human DEC-205 (hDEC-CHO) were incubated with (A) 0.12 nmol/ml or (B) 0.6 nmol/ml of biotinylated fully diester or fully thioated ODNs. Bound ODN was detected with streptavidin PE and fluorescence analysed by flow cytometry. Representative histograms of at least 2 experiments are shown.

3.1.2. Examining the capacity of DEC-205 to bind ss and ds ODNs

Our previous data indicated that, of the three ODN classes tested, B-ODNs bound DEC-205 most efficiently. A-ODNs bound relatively poorly and C-ODNs bound to an intermediate degree (264). Structurally, B-ODNs are linear single-stranded (ss) DNA molecules, whereas both A- and C-ODNs are known to form double-stranded (ds) DNA, and in the case of A-ODN, can even form larger aggregated structures. Thus, we examined whether DEC-205 has a greater capacity to bind ss ODNs compared with ds ODNs. We compared the binding of biotin-tagged B-ODNs to cell surface- expressed DEC-205 either in their normal ss state, or after annealing with the

untagged complementary strand to create ds ODNs. The complementary strands used were diester, not thioated, to ensure that they would not have efficient DEC-205 binding themselves that could inhibit binding of the tagged B-ODN. Strikingly, the ability of B-ODNs to bind both mouse and human DEC-205 was completely abolished after pre-incubation with the complementary strand (Figure 3.2). B-ODNs incubated with a non-complementary strand (the forward sequence of 1668) showed no impairment to their DEC-205-binding ability. These results indicate that ds ODNs are unable to bind DEC-205 efficiently.

A mouse DEC-205 B mouse DEC-205

300 400 1668 2006 1668 + noncomp 2006 + noncomp 1668 + comp 2006 + comp 300 200

200 MFI MFI 100 100

0 0 120 24 4.8 120 24 4.8 1668-biotin (nM) 2006-biotin (nM)

C human DEC-205 D human DEC-205

50 80 1668 2006 1668 + noncomp 2006 + noncomp 40 1668 + comp 60 2006 + comp 30 40 20 MFI MFI

10 20

0 0 -10 120 24 4.8 120 24 4.8 1668-biotin (nM) 2006-biotin (nM)

Figure 3.2. Binding of ss and ds ODNs to DEC-205-expressing CHO-K1 cells. CHO-K1 cells expressing (A, B) mouse DEC-205 or (C, D) human DEC-205 were incubated with biotinylated 1668 or 2006 at the indicated concentrations, or with biotinylated 1668 or 2006 pre-incubated with an untagged non-complementary (+ noncomp) or complementary (+ comp) strand. Bound biotinylated ODN was detected with streptavidin PE, and mean fluorescence intensity (MFI) measured by flow cytometry. Bars represent mean MFI (minus background) ± SEM of cumulative data from 2-3 experiments.

3.1.3. ODN nucleotide bases preferred for DEC-205 binding

Our previous data indicated that not all B-ODNs bind to DEC-205 equivalently, e.g. 2006 was observed to bind human DEC-205 with greater efficiency than 1668 (264). As all B-ODNs are ss and fully thioated, the only distinguishing differences between 2006 and 1668 are their length and nucleotide sequence. We investigated whether DEC-205 preferentially binds certain nucleotide bases over others. We examined whether the DEC-205-binding ability of fully diester 1668 (1668d), which was previously shown to be very poor (Figure 3.1), could be augmented by the 5’ addition of 6mer homopolymers of each of the standard DNA bases – guanosine (6G), adenosine (6A), thymine (6T) and cytosine (6C). The 6mer-1668d ODNs were compared for their ability to bind mouse or human DEC-205 expressed on the surface of CHO K1 cells.

Mouse DEC-205 appeared to bind 6T-1668d most efficiently (Figure 3.3A). 6C-1668d also bound, but to a lesser degree, while 6G-1668d and 6A-1668d showed only minimal mouse DEC-205-specific binding, even at the highest dose. All four 6mer- 1668 ODNs had comparatively weaker binding to human DEC-205 (Figure 3.3B). At the same concentrations of ODN that showed binding to mouse DEC-205, no human DEC-205-specific binding was observed (data not shown). Only by staining with an 8 times higher concentration of ODN could some human DEC-205-specific binding of 6T-1668d be seen (Figure 3.3B). Even at these very high concentrations, none of the other 6mer-1668d ODNs showed specific binding to human DEC-205. Some ODNs, particularly 6G-1668d, showed high levels of binding to CHO cells that was not mediated by mouse or human DEC-205, as there was little to no increase in binding to DEC-205 transfectants over parental CHO cells.

Figure 3.3. ODNs with higher T content are preferred for DEC-205 binding. CHO-K1 cells (CHO) or CHO-K1 cells expressing (A) mouse DEC-205 (mDEC- CHO) or (B) human DEC-205 (hDEC-CHO) were incubated with biotinylated ODNs at the indicated concentrations. Bound ODN was detected with streptavidin PE and fluorescence analysed by flow cytometry. “bg” indicates the background fluorescence of streptavidin PE binding to CHO-K1 cells in the absence of ODN. The background

fluorescence of CHO-K1 cells expressing mouse or human DEC-205 was identical and not shown for clarity. Representative histograms of 2-3 experiments are shown.

3.1.4. Optimal ODN length for DEC-205 binding

Previous studies have determined that there is a minimal length of ODN required for optimal TLR9 stimulatory activity (287). We considered whether DEC-205 also has a minimum ODN length required for binding. To investigate this, 2006, a 24mer B- ODN, was used as a starting point. 2006 was chosen as it exhibits strong binding to both mouse and human DEC-205, is composed primarily of T, the preferred base for DEC-205 binding, and is the longest of the 3 B-ODNs used in this study. By removing nucleotides from the 5’ end of 2006 we created 20mer, 18mer, 14mer and 10mer ODNs, which were tested for their ability to bind mouse or human DEC-205 by ELISA. The 20mer, 18mer and 14mer ODNs appeared to bind to mouse DEC-205 just as effectively as full length 2006 (24mer). By contrast, the 10mer showed almost a complete lack of binding to mouse DEC-205, with only slight binding above background detected at high concentrations (Figure 3.4A). The requirements for binding to human DEC-205 appeared to be more stringent. Even the removal of 4 nucleotides from the original 2006 24mer resulted in significantly reduced binding. Successively shorter ODNs exhibited further reduced binding, and the 14mer and 10mer appeared to have minimal binding (Figure 3.4B).

After this initial screen by ELISA, we sought to confirm whether similar binding requirements would be observed for membrane-bound DEC-205, a form that is more relevant to the biology of DEC-205 in vivo. We examined the binding of the different length ODNs to CHO-K1 cells expressing mouse or human DEC-205. Similar to the soluble protein, mouse DEC-205 expressed on the surface of CHO-K1 cells bound the 20mer and 18mer equally as efficiently as the full-length 24mer (Figure 3.4C). The 14mer also bound effectively, with a slight reduction in binding compared with 2006 noticeable at lower concentrations. The 10mer appeared to bind comparably to the 24mer at the highest concentration of 600 nM, but a strong decrease in binding was apparent at lower concentrations. The pattern of binding by human DEC-205 was also

conserved between the membrane-bound and soluble forms. ODNs shorter than the 24mer 2006 showed reduced binding to human DEC-205-expressing cells, proportional to the reduction in length (Figure 3.4D). In accordance with the ELISA, the 14mer and 10mer showed only minimal binding to human DEC-205-expressing cells.

A mouse DEC-205 B human DEC-205

2.0 0.8 24mer 24mer 20mer 20mer 18mer 1.5 18mer 0.6 14mer 14mer 10mer 10mer 1.0 0.4 O.D. O.D.

0.5 0.2

0.0 0.0 0.001 0.01 0.1 1 10 100 1000 10000 0.001 0.01 0.1 1 10 100 1000 10000 ODN (nM) ODN (nM) C

Figure 3.4. Optimal length of ODN for DEC-205 binding. (A, B) ELISA plates were coated with soluble FLAG-tagged (A) mouse DEC-205 (5 μg/ml) captured by anti- DEC-205 mAb (5 μg/ml) or (B) human DEC-205 (5 μg/ml) captured by anti-FLAG mAb (5 μg/ml). Plates were incubated with biotinylated ODN at the indicated concentrations, then washed and incubated with streptavidin HRP and visualised with ABTS substrate. Shown is the mean optical density measured at 405-490nm of duplicate samples. Representative data of 2-3 experiments are shown. (C, D) CHO-K1 cells (CHO) or CHO-K1 cells expressing (C) mouse DEC-205 (mDEC-CHO) or (D) human DEC-205 (hDEC-CHO) were incubated with biotinylated ODNs at the indicated concentrations. Bound ODN was detected with streptavidin PE or APC and fluorescence analysed by flow cytometry. “bg” indicates the background fluorescence of streptavidin PE or APC binding to CHO-K1 cells in the absence of ODN. The background fluorescence of CHO-K1 cells expressing mouse or human DEC-205 was identical and not shown for clarity. Representative histograms of 2-3 experiments are shown.

3.1.5. 14T is a DEC-205 binding motif

The above results suggest that, for both mouse and human DEC-205, ODNs with higher T content bind better. For mouse DEC-205, a 14mer was the minimal ODN length required for efficient binding. Thus, we hypothesized that an ODN with a length of 14 bases composed entirely of T (14T) may be an optimal motif for mouse DEC-205 binding. We examined the binding of 14T to both soluble and cell-bound mouse DEC-205 in parallel with 2006 and the 14mer fragment of 2006. 14T, the 14mer fragment and 2006 all appeared to have similar ability to bind soluble mouse DEC-205 by ELISA (Figure 3.5A). All three ODNs also bound to CHO-K1 cells expressing mouse DEC-205 comparably at higher ODN concentrations, but at lower concentrations, the 14mer fragment showed a slight decrease in binding, while 2006 and 14T retained effective binding (Figure 3.5B). This suggests that 14T does indeed have a high affinity for mouse DEC-205, comparable to 2006, and confirms that, for ODNs of the same length, higher T content promotes greater DEC-205 binding. We did not test the binding of 14T to human DEC-205, as it is predicted to have weak

binding affinity, given that 14mer ODNs were found to bind human DEC-205 very poorly (Figure 3.4).

A 20.8.13 m 14T 2.5 2006 14mer 2.0 14T

1.5 O.D. 1.0

0.5

0.0 0.01 0.1 1 10 100 1000 ODN (nM)

Figure 3.5. Binding of 14T to mouse DEC-205. (A) ELISA plates were coated with soluble mouse DEC-205 (5 μg/ml) captured by anti-DEC-205 mAb (5 μg/ml). Plates were incubated with biotinylated ODN at the indicated concentrations, then washed and incubated with streptavidin HRP and visualised with ABTS substrate. Shown is the mean optical density measured at 405-490nm of duplicate samples. Representative data of 3 experiments are shown. (B) CHO-K1 cells (CHO) or CHO-K1 cells expressing mouse DEC-205 (mDEC-CHO) were incubated with biotinylated ODNs at the indicated concentrations. Bound ODN was detected with streptavidin PE and fluorescence analysed by flow cytometry. “bg” indicates the background fluorescence

100

of streptavidin PE binding to CHO-K1 cells in the absence of ODN. The background fluorescence of CHO-K1 cells expressing mouse DEC-205 was identical and not shown for clarity. Representative histograms of 4 experiments are shown.

3.1.6. Introduction of diester bonds disrupts DEC-205 binding of ODN

Having established the properties of ODNs that confer optimal DEC-205 binding, we next considered the types of modifications to ODNs that could abrogate DEC-205 binding. Since diester ODNs are ineffective at binding DEC-205 (Figure 3.1), we investigated whether disrupting the fully thioated backbone of 14T with even a single diester bond would significantly impair DEC-205 binding. Indeed, in comparison to fully thioated 14T, if the central thioate bond was changed to diester in 7T-7T, a severely reduced ability to bind mouse DEC-205 by ELISA was seen (Figure 3.6).

7/11/13

1.0 14T 7T-7T 0.8

0.6 O.D. 0.4

0.2

0.0 0.1 1 10 100 1000 10000 100000 ODN (nM)

Figure 3.6. One diester bond disrupts DEC-205 binding. ELISA plates were coated with soluble FLAG-tagged mouse DEC-205 (5 μg/ml) captured by anti-FLAG mAb (5 μg/ml). Plates were incubated with biotinylated ODN at the indicated concentrations, then washed and incubated with streptavidin HRP and visualised with ABTS substrate. Shown is the mean optical density measured at 405-490nm of duplicate samples. Representative data of 2 experiments are shown.

101

3.2. The contribution of DEC-205 to the immunostimulatory effects of ODN

3.2.1. Examining the capacity of 2006 and 21798 to bind DEC-205

We have previously shown that the capacity of 1668 to bind DEC-205 has a functional role in mediating its immunostimulatory effects (264). We sought to extend these studies by determining whether other ODNs with strong DEC-205 binding affinity display a similar dependence on DEC-205 for optimal stimulatory activity. To this end, we chose to study 2006, which binds mouse DEC-205 as efficiently as 1668, and binds human DEC-205 with even greater affinity than 1668. It was also of interest to identify stimulatory ODNs that have weak DEC-205 binding, and therefore may induce stimulatory effects independent of DEC-205. The recently described P-ODN, 21798, is an ideal candidate, as it has potent stimulatory effects, but its structural features, including the presence of ds aggregates and diester bonds, would predict that it has poor DEC-205 binding capacity.

Firstly, we compared the capacity of 2006 and 21798 to bind mouse and human DEC- 205 by ELISA. As previously reported (264), 2006 exhibited strong binding to both mouse and human DEC-205. By contrast, 21798 showed minimal binding to either mouse or human DEC-205 at the same concentrations (Figure 3.7A). Although not shown here, in experiments that will be presented later, some very weak binding to mouse and human DEC-205 was observed at very high concentrations of 21798 (over 1000 nM).

102

18.4.13

2.5 mouse DEC-205 binding: 2006 2.0 21798

1.5 human DEC-205 binding: 2006 21798 O.D. 1.0

0.5

0.0

0.1 1 10 100 1000 ODN (nM)

Figure 3.7. Binding of 2006 and 21798 to DEC-205. ELISA plates were coated with soluble FLAG-tagged mouse or human DEC-205 (5 μg/ml) captured by anti-FLAG mAb (10 μg/ml). Plates were incubated with biotinylated ODN at the indicated concentrations, then washed and incubated with streptavidin HRP and visualised with ABTS substrate. Shown is the mean optical density measured at 405-490nm of duplicate samples. Representative data of 5 independent experiments.

3.2.2. The requirement of DEC-205 for the activation of purified B cells by 2006 or 21798

In our previous study of 1668, the requirement for DEC-205 was readily apparent by examining the direct activation of purified primary B cells (264). Thus, we used the same assay to determine the impact of DEC-205 on the stimulatory activity of 2006 and 21798. B6 or DEC-205-/- sorted B cells were stimulated with graded doses of 2006, 21798 or 1668. Twenty-four hours later, the upregulation of co-stimulatory markers was measured by flow cytometry. All 3 ODNs induced the upregulation of CD40, CD80, CD86 and MHC II on the surface of B6 B cells (Figure 3.8A and B). The activation of B cells by 21798 appeared to be independent of DEC-205, as DEC-205-/- B cells responded comparably to B6. By contrast, DEC-205-/- B cells had significantly impaired responses to 2006 stimulation, showing minimal activation except at the highest dose tested (100 nM). The stimulation of B cells by 1668 was also DEC-205-

103

dependent, as previously reported, although at higher doses, DEC-205 became redundant (Figure 3.8A).

The B cell culture supernatants were collected and assayed for IL-6 production by ELISA. Again, in accordance with our previous data, the stimulation of IL-6 production by 1668 was dependent on DEC-205 (Figure 3.8C). DEC-205-/- B cells produced minimal IL-6 except at the highest dose (100 nM) of 1668, in contrast to B6 B cells, which produced high levels of IL-6 when stimulated with 1668 at a range of concentrations (25-100 nM). Even at the highest dose of 2006, only a very small amount of IL-6 (62.1 ± 14.5 pg/ml) was produced by B6 B cells, and no IL-6 above background was detected from DEC-205-/- cells (Figure 3.8C). 21798 stimulation induced more production of IL-6 than 2006, but again, only at higher concentrations (Figure 3.8C). DEC-205-/- cells appeared to produce slightly less IL-6 than B6 cells in response to 21798 stimulation, but this was not statistically significant.

Taken together, these results indicate that DEC-205 is involved in the direct activation of B cells by both 2006 and 1668, but not by 21798. However, only 1668 could stimulate high levels of of IL-6 production from B cells at the concentrations tested, and this was dependent on DEC-205, in line with previous reports.

104

A CD40 CD80 CD86 MHC II

1000 B6 80 250 2000 800 DEC-205-/- 200 * * 60 1500 * 600 150 * 2006 40 1000 MFI MFI MFI 400 MFI 100 20 500 200 50

0 0 0 0 10 100 10 100 10 100 10 100 ODN (nM) ODN (nM) ODN (nM) ODN (nM)

CD40 CD80 CD86 MHC II

1000 B6 80 250 2000 800 DEC-205-/- 200 60 1500 600 150 21798 40 1000 MFI MFI MFI MFI 400 100 20 500 200 50

0 0 0 0 10 100 10 100 10 100 10 100 ODN (nM) ODN (nM) ODN (nM) ODN (nM)

CD40 CD80 CD86 MHC II

1500 B6 80 250 2000 DEC-205-/- 200 60 1500 1000 150 1668 40 1000 MFI MFI MFI MFI 100 500 20 500 50

0 0 0 0 10 100 10 100 10 100 10 100 ODN (nM) ODN (nM) ODN (nM) ODN (nM) B CD40 CD80 CD86 MHC II

B6 no stim 2006 B6 + CpG ODN DEC-/- + CpG ODN

21798

0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 C 2006 21798 1668 600 600 600 B6 DEC-205-/- 400 400 400

200 200 200 IL-6 (pg/mL) IL-6 (pg/mL) IL-6 (pg/mL)

0 0 0 10 100 10 100 10 100 ODN (nM) ODN (nM) ODN (nM)

Figure 3.8. Stimulation of purified B cells by 2006, 21798 and 1668. Purified B cells (CD19+CD3-) isolated from the spleens of B6 or DEC-205-/- mice were stimulated with graded doses of the indicated ODNs for 24 hours at 37ºC in complete media. Cells were then examined for the expression of CD40, CD80, CD86 and MHC II by flow cytometry. (A) MFI of CD40, CD80, CD86 and MHC II expression on B6 or DEC-205-/- B cells stimulated with graded doses of the indicated ODNs. Dashed line

105

indicates MFI of unstimulated B6 cells, dotted line indicates MFI of unstimulated DEC-205-/- cells. Mean ± SEM of cumulative data from 2 independent experiments is presented for 2006 and 21798 stimulation. 1668 stimulation was performed only once. Statistical significance determined by unpaired t-tests, corrected for multiple comparisons by the Holm-Sidak method. * indicates significant difference between B6 and DEC-205-/-. (B) Representative histograms of CD40, CD80, CD86 and MHC II expression on B6 or DEC-205-/- (DEC-/-) B cells stimulated with 50 nM of the indicated ODNs (+ CpG ODN). The baseline expression of each marker was equivalent on unstimulated B6 and DEC-205-/- B cells; only unstimulated B6 cells (B6 no stim) are shown for clarity. Experiments were performed twice. (C) Supernatants were collected after 24-hour stimulation with the indicated ODNs and the concentration of IL-6 measured by ELISA. Mean ± SEM of cumulative data from 2 independent experiments is presented for 2006 and 21798 stimulation. 1668 stimulation was performed only once.

3.2.3. The contribution of DEC-205 to the stimulatory effects of ASO 518477

DEC-205 was found to play a role in mediating the immunostimulatory activity of both CpG ODNs 1668 and 2006. While it is well established that the stimulatory activity of these ODNs requires the CpG motif (281, 284, 333), binding to DEC-205 does not require this motif (264). Interestingly, even in the absence of CpG motifs, fully thioated non-CpG ODNs can still exhibit immunostimulatory activity, though less potently than CpG ODNs (334-336). This led us to consider whether DEC-205 is also involved in mediating the stimulatory properties of non-CpG ODNs.

ASO 518477 is a commercial, fully thioated anti-sense ODN (ASO) designed by Isis Pharmaceuticals to inhibit the expression of a target gene, β-1,4- galactosyltransferase polypeptide 6, which is implicated in the pathogenesis of multiple sclerosis (337). However, a shortcoming of this technology is the potential for ASOs to induce unwanted immunostimulatory side effects. Indeed, 518477 has been reported to have stimulatory activity, despite attempts to minimise this by avoiding the inclusion of

106

CpG motifs and the addition of 2'-O-methylribose modification to the first and last four bases (personal communication, Dr. Ted Yun, Isis Pharmaceuticals). We investigated whether the stimulatory effects of this non-CpG ODN were mediated by DEC-205. If so, modifications to inhibit DEC-205 binding could be a viable method of removing these unwanted stimulatory effects.

First, we examined the capacity of 518477 to bind DEC-205 by ELISA. We found that 518477 was able to bind mouse DEC-205, but to a lesser degree than the positive control 2006 ODN (Figure 3.9).

2.0 2006 518477 1.5

1.0 O.D.

0.5

0.0 0.1 1 10 100 1000 10000 ODN (ng/ml)

Figure 3.9. Binding of ASO 518477 to mouse DEC-205. ELISA plates were coated with soluble FLAG-tagged mouse DEC-205 (5 μg/ml) captured by anti-FLAG mAb (5 μg/ml). Plates were incubated with biotinylated ODN at the indicated concentrations, then washed and incubated with streptavidin HRP and visualised with ABTS substrate. Shown is the mean optical density measured at 405-490nm of duplicate samples. Representative data of two independent experiments.

We next evaluated immune activation after 518477 treatment in mice by examining the upregulation of co-stimulatory markers on B cells and DCs harvested either 6 or 24 hours post-i.v. injection. We compared B6 and DEC-205-/- mice in parallel to determine whether DEC-205 contributes to these stimulatory effects.

107

At 6 hrs post-injection, B6 B cells showed little difference in surface CD40 and CD80 levels, but a slight increase in CD86 and MHC II (Figure 3.10A and B). The upregulation of MHC II appeared to be marginally lower in DEC-205-/- mice (Figure 3.10B). By 24 hours, B6 B cells had more enhanced expression of CD40, CD80, CD86 and MHC II (Figure 3.11A and B). At this time point, there was a significant reduction in the levels of CD40, CD86 and MHC II in DEC-205-/- B cells (Figure 3.11A and B).

On the other hand, DCs from B6 mice showed strong upregulation of CD86 and MHC II on both CD8+ and CD8- DC populations by 6 hours after 518477 administration (Figure 3.12A and B). There was a small increase in CD40 levels, but little change in CD80, similar to B cells. At 24 hours post-injection, the pattern of upregulation was less distinct. Both CD8+ and CD8- DCs from B6 mice no longer exhibited a single clear peak of expression, particularly of CD86 and MHC II (Figure 3.13A). Instead, the expression pattern appeared bimodal, with only a portion of the CD8+ or CD8- DCs showing strong upregulation above baseline.

This could be a result of changes in the spleen that occur 24 hours after activation. For instance, cells can die, migrate, or alter their surface phenotype in response to the inflammatory environment. Indeed, we observed lower numbers of CD11c+CD8+ cells in the spleens of mice 24 hours post-injection of 518477, which was not seen at 6 hours (data not shown). Furthermore, we have consistently observed that CD8- DCs upregulate DEC-205 expression upon stimulation by ODNs. However, CD8- DCs do not uniformly upregulate DEC-205 upon stimulation by 518477, instead appearing to have bimodal expression with a subpopulation exhibiting much higher DEC-205 expression (Figure 3.14). This suggests that the CD8- DCs are a heterogeneous population, which could also contribute to the non-uniform response of these DCs to 518477 stimulation. Nevertheless, a significant reduction in the activation of DCs from DEC-205-/- mice 24 hours after 518477 treatment was evident in both CD8+ and CD8- DC populations (Figure 3.13A and B).

108

The requirement for TLR9 was assessed in parallel by injecting TLR9-/- mice with 518477, and measuring the activation of B cells and DCs after 24 hours. No activation of either B cells or DCs was seen, indicating that the stimulatory effects of 518477 are dependent on TLR9 (Figure 3.11, 3.13).

A CD40 CD80 CD86 MHC II B6 no stim B6 + 518477 DEC-/- + 518477

2 3 4 5 2 3 4 5 2 3 4 5 2 3 4 5 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 B

CD40 CD80 CD86 MHC II ns ns ns 30 15 80 100 * 80 10 60 20 60 5 40 40 10 20 0 20 %MFI increase %MFI increase %MFI increase %MFI increase 0 0 -5 0 -20 B6 DEC-205-/- B6 DEC-205-/- B6 DEC-205-/- B6 DEC-205-/-

Figure 3.10. Activation of B cells 6 hours after ASO 518477 administration. Two B6 or DEC-205-/- mice were injected i.v. with 200 nmol 518477, or left unimmunised. Six hours later, spleens were harvested and splenocytes from both mice pooled for examining the expression of CD40, CD80, CD86 and MHC II on the B cell population (CD45R+) by flow cytometry. (A) Representative histograms of CD40, CD80, CD86 and MHC II expression on B6 or DEC-205-/- (DEC-/-) B cells from mice injected with 518477 (+ 518477). The baseline expression of each marker was equivalent on unstimulated B6 and DEC-205-/- B cells; only unstimulated B6 cells (B6 no stim) are shown for clarity. Experiments were performed 4 times. (B) Percentage increase in MFI of marker expression of 518477-injected versus unimmunised mice. Bars represent mean ± SEM of 4 independent experiments. Each point represents data from one experiment. Statistical significance determined by one-tailed paired t-test, ns = not significant (p > 0.05).

109

A CD40 CD80 CD86 MHC II B6 no stim B6 + 518477 DEC-/- + 518477

2 3 4 5 2 3 4 5 2 3 4 5 2 3 4 5 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 CD40 CD80 CD86 MHC II B6 no stim TLR9-/- + 518477

2 3 4 5 2 3 4 5 2 3 4 5 2 3 4 5 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 B

CD40 CD80 CD86 MHC II ns * ** * 100 200 200 500

80 150 150 400 60 300 100 100 40 200 20 50 50 100 %MFI increase %MFI increase %MFI increase %MFI increase 0 0 0 0 B6 DEC-205-/- TLR9-/- B6 DEC-205-/- TLR9-/- B6 DEC-205-/- TLR9-/- B6 DEC-205-/- TLR9-/-

Figure 3.11. Activation of B cells 24 hours after ASO 518477 administration. Two B6, DEC-205-/-, or TLR9-/- mice were injected i.v. with 200 nmol 518477, or left unimmunised. Twenty-four hours later, spleens were harvested and splenocytes from both mice pooled for examining the expression of CD40, CD80, CD86 and MHC II on the B cell population (CD45R+) by flow cytometry. (A) Representative histograms of CD40, CD80, CD86 and MHC II expression on B6 or DEC-205-/- (DEC-/-) B cells from mice injected with 518477 (+ 518477). The baseline expression of each marker was equivalent on unstimulated B6 and DEC-205-/- B cells; only unstimulated B6 cells (B6 no stim) are shown for clarity. Experiments comparing B6 and DEC-205-/- mice or B6 and TLR9-/- mice were performed thrice and twice, respectively. (B) Percentage increase in MFI of marker expression of 518477-injected versus unimmunised mice. Bars represent mean ± SEM of 3 (B6, DEC-205-/-) or 2 (TLR9-/-) independent experiments. Each point represents data from one experiment. Statistical significance determined by one-tailed paired t-test, * p < 0.05, ** p < 0.01, ns = not significant (p > 0.05).

110

A CD40 CD80 CD86 MHC II

CD8+ B6 no stim DCs B6 + 518477 DEC-/- + 518477 CD8- DCs

2 3 4 5 2 3 4 5 2 3 4 5 2 3 4 5 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 B

CD40 CD80 CD86 MHC II 80 50 100 250

40 80 200 CD8+ 60 30 60 150 DCs 40 20 40 100 20

%MFI increase 10 20 50 %MFI increase %MFI increase %MFI increase

0 0 0 0 B6 DEC-205-/- B6 DEC-205-/- B6 DEC-205-/- B6 DEC-205-/-

CD40 CD80 CD86 MHC II 150 30 400 250

200 20 300 100 CD8- 150 DCs 10 200 100 50 0 100 50 %MFI increase %MFI increase %MFI increase %MFI increase

0 -10 0 0 B6 DEC-205-/- B6 DEC-205-/- B6 DEC-205-/- B6 DEC-205-/-

Figure 3.12. Activation of DCs 6 hours after ASO 518477 administration. Two B6 or DEC-205-/- mice were injected i.v. with 200 nmol 518477, or left unimmunised. Six hours later, spleens were harvested and pooled, DCs isolated, and the expression of CD40, CD80, CD86 and MHC II on the CD8+ and CD8- DC populations (CD11c+CD8+; CD11c+CD8-) examined by flow cytometry. (A) Representative histograms of CD40, CD80, CD86 and MHC II expression on B6 or DEC-205-/- (DEC-/-) DCs from mice injected with 518477 (+ 518477). The baseline expression of each marker was equivalent on unstimulated B6 and DEC-205-/- B cells; only unstimulated B6 cells (B6 no stim) are shown for clarity. Experiments were performed 4 times. (B) Percentage increase in MFI of marker expression of 518477-injected versus unimmunised mice. Bars represent mean ± SEM of 4 independent experiments. Each point represents data from one experiment.

111

A CD40 CD80 CD86 MHC II

CD8+ B6 no stim DCs B6 + 518477 DEC-/- + 518477 CD8- DCs

2 3 4 5 2 3 4 5 2 3 4 5 2 3 4 5 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 CD40 CD80 CD86 MHC II

CD8+ B6 no stim DCs TLR9-/- + 518477

CD8- DCs

2 3 4 5 2 3 4 5 2 3 4 5 2 3 4 5 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 B

CD40 CD80 CD86 MHC II ns 200 * 150 100 * 100 **

150 100 CD8+ 100 50 50

DCs 50 50 0 0 0 %MFI increase %MFI increase %MFI increase %MFI increase 0 -50 B6 DEC-205-/-TLR9-/- B6 DEC-205-/-TLR9-/- B6 DEC-205-/-TLR9-/- B6 DEC-205-/-TLR9-/-

CD40 CD80 CD86 MHC II ns ns 500 150 800 * 60 ** 400 600 40 CD8- 300 100 400 DCs 200 20 50 100 200 0 %MFI increase %MFI increase %MFI increase 0 %MFI increase 0 0 -100 -20 B6 DEC-205-/-TLR9-/- B6 DEC-205-/-TLR9-/- B6 DEC-205-/-TLR9-/- B6 DEC-205-/-TLR9-/-

Figure 3.13. Activation of DCs 24 hours after ASO 518477 administration. Two B6, DEC-205-/-, or TLR9-/- mice were injected i.v. with 200 nmol 518477, or left unimmunised. Twenty-four hours later, spleens were harvested and pooled, DCs isolated, and the expression of CD40, CD80, CD86 and MHC II on the CD8+ and CD8- DC populations (CD11c+CD8+; CD11c+CD8-) examined by flow cytometry. (A) Representative histograms of CD40, CD80, CD86 and MHC II expression on B6, DEC-205-/- (DEC-/-) or TLR9-/- DCs from mice injected with 518477 (+ 518477). The baseline expression of each marker was equivalent on unstimulated B6 and DEC-205- /- B cells; only unstimulated B6 cells (B6 no stim) are shown for clarity. Experiments comparing B6 and DEC-205-/- mice or B6 and TLR9-/- mice were performed thrice and twice, respectively. (B) Percentage increase in MFI of marker expression of

112

518477-injected versus unimmunised mice. Bars represent mean ± SEM of 3 (B6, DEC-205-/-) or 2 (TLR9-/-) independent experiments. Each point represents data from one experiment. Statistical significance determined by one-tailed paired t-test, * p < 0.05, ** p < 0.01, ns = not significant (p > 0.05).

CD8+ DCs CD8- DCs

B6 no stim 6hrs B6 + 518477

24hrs

2 3 4 5 2 3 4 5 10 10 10 10 10 10 10 10

Figure 3.14. Expression of DEC-205 on DCs 6 and 24 hours after ASO 518477 administration. Two B6 mice were injected i.v. with 200 nmol 518477 (+ 518477), or left unimmunised (no stim). 6 and 24 hours later, spleens were harvested and pooled, DCs isolated, and the expression of DEC-205 on the CD8+ and CD8- DC populations (CD11c+CD8+; CD11c+CD8-) examined by flow cytometry. Representative histograms of 4 (6hrs) and 3 (24hrs) independent experiments.

113

3.3. Utilising DEC-205 to design more potent CpG adjuvants

3.3.1. Enhancing the DEC-205 binding of 21798 by addition of 14T

Our data indicate that the stimulatory activity of B-ODNs is decreased in the absence of DEC-205. This would suggest that modifications to reduce their DEC-205-binding capacity would also impair their stimulatory activity. Following this logic, it is plausible that modifications to increase the DEC-205 binding capacity of ODNs would lead to enhanced stimulatory activity.

To test this hypothesis, we used 21798 as an example of a stimulatory CpG ODN with poor DEC-205 binding. We examined whether the DEC-205 binding of 21798 could be augmented by addition of 14T, an ODN demonstrated to have strong DEC-205 binding, to the 5’ end. The binding of this new ODN (14T-21798) to mouse DEC-205 was determined by ELISA. 14T-21798 showed a marked improvement in binding capacity compared with 21798, binding to mouse DEC-205 almost as efficiently as 14T (Figure 3.15).

2.0

14T 1.5 21798 14T-21798

1.0 O.D.

0.5

0.0 0.01 0.1 1 10 100 1000 10000 ODN (nM)

Figure 3.15. 14T-21798 binds DEC-205. Plates coated with soluble FLAG-tagged mouse DEC-205 (5 μg/ml) captured by anti-FLAG mAb (5 μg/ml) were incubated with biotinylated ODN at the indicated concentrations, then washed and incubated with streptavidin HRP and visualised with ABTS substrate. Shown is the mean optical density measured at 405-490nm of duplicate samples. This experiment was performed once.

114

3.3.2. Improved adjuvant activity of 14T-21798

Having confirmed that the 5’ addition of 14T to 21798 enhances its DEC-205-binding capacity, we next examined whether this enhanced binding resulted an increase in immunostimulatory capability. Previous studies have shown that immunisation of mice with Ags targeted to the DC receptor Clec9A can induce strong Ag-specific CTL responses, but only in the presence of adjuvant (120, 121). We used this model to evaluate the potency of 14T-21798 as an adjuvant, by immunising mice with the model Ag ovalbumin (OVA) delivered via an anti-Clec9A targeting construct (10B4- OVA). We compared the immune responses induced by anti-Clec9A-OVA co- administered with either 14T-21798 or an equivalent amount of the component ODNs that were not joined (14T + 21798), in B6 and DEC-205-/- mice. We chose to use minimal amounts of ODN (2 nmol) as adjuvant, as differences between B6 and DEC-205-/- mice might become more apparent under limiting conditions.

Under these immunisation conditions, 14T-21798 was seen to strongly enhance the induction of OVA-specific CD8+ T cell responses, as measured by staining with OVA- specific MHC I (Kb-SIINFEKL) tetramers, and these responses were abrogated in DEC-205-/- mice (Figure 3.16). Similarly, when 14T was not physically joined to 21798, the stimulatory capacity of the ODN mixture (14T + 21798) was reduced and appeared to be DEC-205-independent. These data strongly support the notion that an improved capacity to bind DEC-205 enhances the stimulatory capacity of the 14T- 21798 conjugate ODN.

115

** 1.5 * *

1.0 * ns

% tetramer+ 0.5

0.0 naive no adj 14T + 14T-21798 21798

Figure 3.16. 14T-21798 enhances Ag-specific CD8+ T cell responses. B6 (filled bars) or DEC-205-/- (open bars) mice were left unimmunised (naïve) or immunised i.v. with 0.5 μg anti-Clec9A-OVA either in the absence of adjuvant (no adj), or plus 2 nmol 14T and 2 nmol 21798 (14T + 21798), or plus 2 nmol of 14T-21798. Six days later, spleens were harvested and stained with OVA-specific tetramers. The percentage of tetramer+ cells as a proportion of total CD8+ cells in the spleen is shown. Bars represent mean ± SEM of 2 independent experiments. Each point represents 1 mouse.

3.3.3. Enhancing the DEC-205 binding of 21798 by addition of 2006

The responses to 14T-21798 were an average of 2 times higher than the responses to the separated ODNs. We explored whether using another stimulatory CpG ODN instead of the non-stimulatory 14T as the DEC-205 targeting motif could produce even greater responses. We chose to use 2006, as this ODN had demonstrated strong binding to both mouse and human DEC-205 (264). Although 2006 lacks the optimal CpG motif to activate mouse TLR9, we (Figure 3.8) and others (298, 338) have

116

demonstrated that this ODN still retains some capacity to stimulate mouse cells. Thus, the addition of 2006 may facilitate DEC-205 binding as well as providing some stimulatory capacity. Importantly, since 2006 is optimally stimulatory on human cells and is already in use in human clinical trials, incorporating this ODN would also facilitate potential translation of this technology to humans.

Firstly, we confirmed that the 5’ addition of 2006 to 21798 to generate 2006-21798 results in enhanced DEC-205 binding. Indeed, 2006-21798 showed enhanced binding to both mouse and human DEC-205 by ELISA compared with 21798, and was comparable to 2006 (Figure 3.17A and B).

21798 is known to form multimeric aggregates due to its double palindrome (300). For other ODNs, the presence of aggregates has been shown to be required for optimal stimulatory activity and uptake by certain receptors (297, 321, 327, 339, 340). While inhibiting the aggregation of 21798 was not seen to affect its stimulatory activity in vitro (300), it could potentially impact its uptake and efficacy in vivo. Therefore, we investigated whether the 5’ addition of 2006, which does not have the ability to aggregate, would interfere with this property.

The extent of aggregation of 2006, 21798 and 2006-21798 was analysed by agarose gel electrophoresis (Figure 3.17C). 2006 appeared as a single distinct band, as expected, whereas 21798 appeared as a large smear indicating the formation of larger structures at a range of sizes. 2006-21798 also formed aggregates as evidenced by a smear that covered a similar size range as 21798. Interestingly, although fluorescence was seen in a continuous band similar to 21798, 2006-21798 also displayed several brighter bands within the smear that could indicate the ODN preferentially forms aggregates of particular sizes.

117

A C mouse DEC-205 2.0

2006 1.5 21798 2006-21798

1.0 O.D.

0.5

0.0 0.01 0.1 1 10 100 1000 10000

ODN (nM) B human DEC-205 0.8

2006 0.6 21798 2006-21798

0.4 O.D.

0.2

0.0 0.1 1 10 100 1000 10000

ODN (nM)

Figure 3.17. 2006-21798 binds DEC-205 and forms aggregates. ELISA plates were coated with soluble FLAG-tagged (A) mouse DEC-205 (5 μg/ml) captured with anti- FLAG mAb (5 μg/ml) or (B) human DEC-205 (5 μg/ml) without capture. Plates were incubated with biotinylated ODN at the indicated concentrations, then washed and incubated with streptavidin HRP and visualised with ABTS substrate. Shown is the mean optical density measured at 405-490nm of duplicate samples. Representative data of 5 independent experiments. (C) 1 nmol of each of 2006, 21798 and 2006-21798 were electrophoresed through a 5% agarose gel. Representative data of 2 independent experiments.

118

3.3.4. Improved adjuvant activity of 2006-21798

We investigated the potency of 2006-21798 as an adjuvant using the same experimental set-up as in Section 3.3.2. 2006-21798 induced approximately 4 times higher OVA-specific CD8+ T cell responses in B6 mice compared with the separated ODNs (2006 + 21798) (Figure 3.18A). As was seen for 14T-21798, the enhanced responses to 2006-21798 were significantly reduced in DEC-205-/- mice. These results indicate that the 5’ addition of 2006 to 21798 is more effective than the addition of 14T for promoting CD8+ T cell responses, most likely due to the additional stimulatory CpG motifs present in 2006.

Immunisation with the anti-Clec9A targeting construct, 10B4, has previously been shown to induce strong Ab responses, even in the absence of adjuvant (120). We investigated whether the addition of 2006-21798 as an adjuvant would be able to enhance the induction of these Ab responses. Mice immunised as above were bled on day 6 post-immunisation, and the production of OVA-specific Ab in plasma was determined by ELISA assay. Both 2006-21798 and the separated 2006 + 21798 appeared to induce higher anti-OVA Ab titres compared with mice immunised in the absence of adjuvant (which, for Clec9A-targeting, still induces strong Ab responses), though low numbers in the latter group prevented this from reaching statistical significance (Figure 3.18B). This suggests that both adjuvant treatments accelerate the induction of Ab responses at the early time point of 6 days post-immunisation. However, 2006-21798 was not significantly better than 2006 + 21798 in this respect. Thus, there appears to be little advantage to 5’ addition of 2006 to 21798, compared with co-injection of the separated ODNs, for generating Ag-specific Ab responses. Both adjuvant treatments were slightly less effective in DEC-205-/- mice, though this was not significant for 2006 + 21798.

119

A 5 **** *** 4

3 ** ns 2 % tetramer+

0 naive no adj 2006 + 2006-21798 21798

B * ns ns ns * 105

104

103 anti-OVA Ab titre 102

101 naive no adj 2006 + 2006-21798 21798

120

Figure 3.18. 2006-21798 enhances Ag-specific CD8+ T cell responses and Ab responses. B6 (filled symbols) or DEC-205-/- (open symbols) mice were left unimmunised (naïve) or immunised i.v. with 0.5 μg anti-Clec9A-OVA either in the absence of adjuvant (no adj), or plus 2 nmol 2006 and 2 nmol 21798 (2006 + 21798), or plus 2 nmol of 2006-21798. (A) Six days post-immunisation, spleens were harvested and stained with OVA-specific tetramers. The percentage of tetramer+ cells as a proportion of total CD8+ cells in the spleen is shown. Bars represent mean ± SEM of cumulative data from 3 independent experiments. Each point represents 1 mouse. (B) Bleeds were collected 6 days post-immunisation, and plasma assayed for OVA reactivity by ELISA. Cumulative data of 2 independent experiments is shown, each point represents 1 mouse.

3.3.5. Comparison of 2006-21798 with standard CpG ODN adjuvants

It was of interest to compare the stimulatory capacity of 2006-21798 to standard CpG ODNs. As 2006-21798 is a conjugate of two ODNs, it is approximately twice the length of other CpG ODNs used in this study, which range from 20-24mers. Thus, comparing equimolar amounts of ODNs would result in mice injected with 2006- 21798 receiving greater amounts of DNA by mass, despite receiving the same number of molecules. To avoid a potential bias in favour of the conjugate, it was decided to compare equivalents amounts of ODNs by mass. This would equalise the amount of DNA received by each group, although it could also introduce a potential disadvantage to the 2006-21798-injected group, as they would receive approximately half the number of molecules. These conditions made it more stringent to demonstrate greater activity of 2006-21798 over other ODNs.

The adjuvant capacity of the various ODNs was compared by transferring 50,000 OVA-specific CD8+ T cells (OT-I) into B6 mice, which were then immunised with anti-Clec9A-OVA plus 2006-21798, or the equivalent amount by mass of 1668, 2006, 21798, or the separated 2006 + 21798. Adjuvant efficacy was assessed by enumerating the expansion of OT-I cells in the spleen 6 days post-immunisation. 2006-21798 induced significantly greater OT-I proliferation than 1668, 2006 or 21798 (Figure

121

3.19A). 1668 and 2006 were comparable to each other, while 21798 was approximately twice as effective, inducing almost as much OT-I proliferation as 2006-21798. 2006- 21798 was significantly more effective than 2006 + 21798 (Figure 3.19A), in accordance with the results seen when measuring the endogenous CD8+ T cell response (Figure 3.18A). Interestingly, the response induced by the mixture of 2006 + 21798 was lower than that induced by the equivalent mass of purely 21798, and was comparable to 2006 (Figure 3.19A). This may indicate that co-injection of 2006 and 21798 can cause inhibition of their activity instead of a summative effect. However, it should be noted that the amount of 21798 administered in the mixture (14.7 μg) is approximately half of that given in the immunisation with 21798 alone (30 μg). Although the mixture additionally contains 15.3 μg of 2006, as 2006 is a weaker adjuvant than 21798, this would not be expected to compensate for a full dose of pure 21798.

Using this adoptive transfer system, we also investigated the ability of the ODNs to promote CD4+ T cell responses, by transferring OVA-specific CD4+ T cells (OT-II) 1 day prior to immunisation. Unlike in the OT-I assay, 2006 and 21798 appeared comparable in their capacity to promote OT-II proliferation (Figure 3.19B). 2006- 21798 and 2006 + 21798 were comparable to each other, and appeared to have little advantage over the single ODNs. Only 2006-21798 appeared to be significantly different to 21798, though the level of OT-II proliferation induced was only slightly higher. These results suggest that, in contrast to the stimulation of CD8+ T cells, enhanced DEC-205 binding has little impact on the stimulation of CD4+ T cells.

122 **** * **** ****

pool (excl3) 30ug OTI/10^6 endo CD8 A **** **** * 5×105 **** *** 4×105

3×105 endo CD8+

2×105

1×105 OT-I cells/10 OT-I

naive 1668 2006 no adj 21798 2006 + pool OTII/10^6 endo CD4 21798 2006-21798 B ns ** 6×104 ns ns

ns 4×104 endo CD4+

2×104 OT-II cells/10 OT-II

naive 2006 no adj 21798 2006 + 21798 2006-21798

123

Figure 3.19. OT-I and OT-II proliferation in response to immunisation with anti- Clec9A-OVA plus various ODN adjuvants. 50,000 (A) OT-I or (B) OT-II T cells were transferred into B6 mice 1 day prior to i.v immunisation with 1 μg anti-Clec9A- OVA plus 30 μg of the indicated ODNs. “Naïve” mice received no immunisation after T cell transfer, and “no adj” mice were immunised with 1 μg anti-Clec9A-OVA without adjuvant. The 2006 + 21798 group were immunised with 15.3 μg of 2006 and 14.7 μg of 21798, to maintain the same ratio of each ODN as in 2006-21798. Six days post-immunisation, OT-I or OT-II cells in the spleen were enumerated by flow cytometry, and represented here as the number per 106 endogenous (endo) CD8+ or CD4+ T cells, respectively. Bars represent mean ± SEM of cumulative data from 2-3 independent experiments, with the exception of “no adj” in (A) and (B), and “naïve” in (B), which were performed once. Each point represents 1 mouse.

3.3.6. 2006-21798 does not compensate for inefficient CD8+ T cell responses induced in the absence of Clec9A-targeting of Ag

It has previously been demonstrated that Ag delivered via a Clec9A-targeting mAb induces far greater CD8+ T cell responses than Ag delivered via a non-targeting isotype control mAb, even in the presence of adjuvants such as anti-CD40, LPS, 1668 CpG ODN or polyIC (121, 215). We investigated whether the strong CD8+ T cell responses induced in the presence of 2006-21798 adjuvant were similarly dependent on effective Ag delivery to DCs, or if the potent immunostimulatory activity of 2006- 21798 was sufficient to bypass this requirement. We compared the immune responses of mice immunised with OVA fused to either an anti-Clec9A or isotype control mAb in the presence of 2006-21798. By tetramer staining 6 days post-immunisation, we observed the induction of OVA-specific CD8+ T cell responses in mice immunised with anti-Clec9A-OVA, but no responses above background were induced in mice immunised with isotype-OVA despite the addition of 2006-21798 as adjuvant (Figure 3.20A). In another experiment, mice were adoptively transferred with 50,000 OT-I cells 1 day prior to immunisation. The same trends were observed; only mice immunised with anti-Clec9A-OVA, but not isotype-OVA, showed evidence of OT-I proliferation 6 days post-immunisation (Figure 3.20B). This indicated that the strong

124

Ag-specific T cell responses induced in the presence of 2006-21798 were still contingent upon effective delivery of Ag to DCs.

A ** 2.5 ns ***

2.0

1.5

1.0 % tetramer+

0.5

0.0 naive iso-OVA 9A-OVA +2006-21798 +2006-21798

B * ns 6×105 ***

4×105 endo CD8+ 6

2×105 OT-I cells/10 OT-I

0 naive iso-OVA 9A-OVA +2006-21798 +2006-21798

Figure 3.20. CD8+ T cell responses induced in the presence of 2006-21798 are inefficient in the absence of Clec9A-targeting of Ag. (A) B6 mice were left unimmunised (naïve) or immunised i.v. with 1 μg anti-Clec9A-OVA (9A-OVA) or isotype control mAb-OVA (iso-OVA) plus 5 nmol of 2006-21798. Six days later,

125

spleens were harvested and stained with OVA-specific tetramers. The percentage of tetramer+ cells as a proportion of total CD8+ cells in the spleen is shown. Bars represent mean ± SEM of cumulative data from 2 independent experiments. Each point represents 1 mouse. (B) 50,000 OT-I T cells were transferred into B6 mice 1 day prior to i.v. immunisation with 1 μg anti-Clec9A-OVA (9A-OVA) or isotype control mAb-OVA (iso-OVA) plus 30 μg of 2006-21798. “Naïve” mice received no immunisation after T cell transfer. Six days post-immunisation, OT-I cells in the spleen were enumerated by flow cytometry, and represented here as the number per 106 endogenous (endo) CD8+ T cells. Bars represent mean ± SEM of cumulative data from 2 independent experiments. Each point represents 1 mouse.

3.3.7. The requirement of DEC-205 for the activation of purified B cells by 2006-21798

In Section 3.2.2. it was observed that the ODNs 2006 and 21798 produce distinct effects on B cells after in vitro stimulation (Figure 3.8). Both ODNs induced upregulation of co-stimulatory markers, but this was only dependent on DEC-205 for 2006, not 21798. Stimulation with 21798 also induced IL-6 production, while 2006 induced little to no IL-6. We tested the combined 2006-21798 in the same assays to determine if it would exhibit properties more similar to the 2006 or the 21798 component. 2006-21798 appeared more similar to 21798 in both cases. Stimulation of purified B cells by 2006-21798 induced upregulation of co-stimulatory markers that was not significantly reduced in DEC-205-/- (Figure 3.21A and B). 2006-21798 also induced a similar level of IL-6 production as 21798, and again this was not reduced in DEC-205-/- (Figure 3.21C). When the cells were treated with the separated ODNs (2006 + 21798), the same effects as with the linked 2006-21798 were observed. Thus, the direct effects of 2006-21798 on B cells seem to be caused by the additive effects of the individual component ODNs, with no apparent benefit of enhanced DEC-205 binding.

126

A CD40 CD80 CD86 MHC II

1000 B6 80 250 2000 800 DEC-205-/- 200 60 1500 600 150 2006-21798 40 1000 MFI MFI MFI 400 MFI 100 20 500 200 50

0 0 0 0 10 100 10 100 10 100 10 100 ODN (nM) ODN (nM) ODN (nM) ODN (nM)

CD40 CD80 CD86 MHC II

1000 B6 80 250 2000 800 DEC-205-/- 200 60 1500 2006 + 600 150 40 1000 MFI MFI MFI 21798 400 100 MFI 20 500 200 50

0 0 0 0 10 100 10 100 10 100 10 100 ODN (nM) ODN (nM) ODN (nM) ODN (nM) B CD40 CD80 CD86 MHC II

B6 no stim 2006-21798 B6 + CpG ODN DEC-/- + CpG ODN 2006 + 21798

0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 C 2006-21798 2006 + 21798 600 600 B6 DEC-205-/- 400 400

200 200 IL-6 (pg/mL) IL-6 (pg/mL)

0 0 1 10 100 1 10 100 ODN conc (nM) ODN conc (nM)

Figure 3.21. Stimulation of purified B cells by 2006-21798 and 2006 + 21798. Purified B cells (CD19+CD3-) isolated from the spleens of B6 or DEC-205-/- mice were stimulated with graded doses of the indicated ODNs for 24 hours at 37ºC in complete media. Cells were then examined for the expression of CD40, CD80, CD86 and MHC II by flow cytometry. (A) MFI of CD40, CD80, CD86 and MHC II expression on B6 or DEC-205-/- B cells stimulated with graded doses of the indicated ODNs. Dashed line indicates MFI of unstimulated B6 cells, dotted line indicates MFI of unstimulated DEC-205-/- cells. Mean ± SEM of 2 independent experiments are presented. (B) Representative histograms of CD40, CD80, CD86 and MHC II expression on B6 or DEC-205-/- (DEC-/-) B cells stimulated with 50 nM of the

127

indicated ODNs (+ CpG ODN). The baseline expression of each marker was equivalent on unstimulated B6 and DEC-205-/- B cells; only unstimulated B6 cells (B6 no stim) are shown for clarity. Experiments were performed twice. (C) Supernatants were collected after 24-hour stimulation with the indicated ODNs and the concentration of IL-6 measured by ELISA. Mean ± SEM of 2 independent experiments are presented.

3.3.8. The requirement of TLR9 for the induction of CTL by 2006- 21798

As 2006 and 21798 are known to act via TLR9 (298, our unpublished observations) we investigated whether the activity of 2006-21798 is also dependent on TLR9. In an initial experiment where the OVA-specific CD8+ T cell response of mice 6 days post- immunisation was measured by tetramer staining, we saw no apparent response in TLR9-/- mice immunised with anti-Clec9A-OVA plus 2006-21798 or 2006 + 21798 (Figure 3.22). However, the response induced by 2006 + 21798 in B6 mice was so close to the background seen in the absence of adjuvant that it was difficult to discern if there was a significant decrease in TLR9-/- mice. Therefore, we opted to use a more sensitive assay to examine the CD8+ T cell response of TLR9-/- mice.

128

B6 DEC-205-/- 3 TLR9-/-

2 % tetramer+

2006 + 2006-21798 21798

Figure 3.22. 2006-21798 requires TLR9 to enhance Ag-specific CD8+ T cell responses. B6, DEC-205-/-, or TLR9-/- mice were immunised i.v. with 0.5 μg anti- Clec9A-OVA plus 2 nmol 2006 and 2 nmol 21798 (2006 + 21798), or 0.5 μg anti- Clec9A-OVA plus 2 nmol of 2006-21798. Six days later, spleens were harvested and stained with OVA-specific tetramers. The percentage of tetramer+ cells as a proportion of total CD8+ cells in the spleen is shown. Dotted line indicates background seen in B6 mice immunised with 0.5 μg anti-Clec9A-OVA in the absence of adjuvant. Bars represent mean ± SEM of 3 mice from 1 experiment.

OVA-specific CTL responses were examined by measuring the in vivo killing of CFSE-labelled OVA-peptide (SIINFEKL)-pulsed target cells transferred into mice 6 days after immunisation with anti-Clec9A-OVA plus 1668, 2006 + 21798, or 2006- 21798 as adjuvant. In B6 mice, the doses of ODN used were sufficient to cause 70-95% killing of target cells (Figure 3.23A). In TLR9-/- mice, for all 3 ODNs tested, the same dose was significantly less effective, resulting in only 8-20% killing. This indicated that 2006-21798, the separated 2006 + 21798, and the single 1668, all require TLR9 to promote effective CTL responses.

129

Residual killing above background levels in TLR9-/- mice was induced by all three ODNs, even 1668, a well-characterised CpG ODN that is known to act via TLR9 (312, 313). We investigated whether this was due to an unexpected ability of the ODNs to induce TLR9-independent stimulatory effects, or if the anti-Clec9A-OVA Ag construct itself was capable of inducing low levels of CTL without adjuvant. Mice were immunised as above with anti-Clec9A-OVA in the absence of adjuvant, and indeed, a low level of OVA-specific killing was observed, which was unaffected by the absence of TLR9 (Figure 3.23B). The CTL response to anti-Clec9A-OVA in the absence of adjuvant was comparable to that seen in mice immunised with CpG ODN adjuvant but in the absence of TLR9 (compare Figure 3.23A). This suggests that the ODNs do not have TLR9-independent activity, and that any residual CTL response in TLR9-/- mice is caused by the anti-Clec9A-OVA Ag construct.

130

pool 3exp A **** **** **** 100

B6 75 TLR9-/-

50 % killing 25

naive 1668 2006 + 2006-21798 21798 pool 2 exp 10B4 B **** ns 100

B6 75 TLR9-/-

50 % killing 25

naive 1668 no adj

Figure 3.23. In vivo CTL activity induced by immunisation with anti-Clec9A-OVA plus ODN adjuvants requires TLR9. B6 or TLR9-/- mice were immunised i.v. with 1 μg anti-Clec9A-OVA plus 30 μg of the indicated ODNs. Six days later, mice were

131

injected i.v. with OVA-peptide (SIINFEKL)-coated target cells, and the percentage killing of target cells determined by flow cytometry 18 hours later. Cumulative data of 2-3 independent experiments is shown, each point represents 1 mouse.

132

Discussion

Molecular properties of ODN determine their DEC-205 binding ability

Our investigations have established several molecular properties of ODNs that are required for DEC-205 binding. For instance, it is necessary for ODNs to be fully thioated and single-stranded for effective binding (Figure 3.1 and 3.2). These characteristics are precisely those that define the B class of ODNs, which aligns with our previous observation that B-ODNs bind DEC-205 most efficiently (264). Our findings also explain the variable DEC-205 binding capacity of different ODN classes. A-ODNs, which form aggregates and have a central diester region, bind DEC-205 very poorly. C-ODNs exhibit intermediate DEC-205 binding, and structurally they can also be considered intermediate between A-ODN and B-ODN, as they are fully thioate but can form dimers or stem-loop structures due to the presence of a palindromic sequence. While C-ODNs still retain some DEC-205 binding ability, the dimeric ODN used in our study, formed by hybridisation of a B-ODN with the complementary strand, showed no DEC-205 binding at all (Figure 3.2). This discrepancy can be explained by the structure of C-ODNs. Unlike A-ODNs, the C- ODN palindrome is located at one end of the ODN, such that the other end does not participate in dimerisation and remains single-stranded. It is possibly these single- stranded overhangs that mediate binding of C-ODN to DEC-205.

The sequence composition and length of ODNs also influences their DEC-205 binding capacity. We found that both human and mouse DEC-205 preferentially bind ODN sequences containing Ts, with mouse DEC-205 also showing a lower but substantial level of binding to ODN sequences containing Cs, suggesting a preference for pyrimidines (T or C) over purines (G or A) (Figure 3.3). For mouse DEC-205, 14 bases appeared to be the minimum length of ODN required for efficient binding, although the 14mer ODN was unable to bind human DEC-205 efficiently, for which the optimal length appears to be at least 21 bases or longer (Figure 3.4). These findings may explain why 2006 was previously seen to bind human DEC-205 more efficiently than 1668 despite both being B-ODNs (264). 1668 is only 20 bases in length, which

133

our results would predict is shorter than the optimal length for human DEC-205 binding, while 2006 is both longer, being 24 bases in length, and contains a higher proportion of T residues. These results also highlight that, while for other ODN attributes mouse and human DEC-205 seem to have similar requirements, human DEC-205 appears to have more stringent requirements for ODN length. It would be interesting to evaluate the binding of longer ODNs to human DEC-205 in future studies to determine the minimum length requirement.

ODN properties required for efficient DEC-205 binding correlate with those required for effective immune stimulation

The molecular properties of ODNs that we have here determined to be optimal for DEC-205 binding are closely aligned with those that have previously been identified to be optimal for the immunostimulatory activity of ODNs. Previous studies indicated that 14 bases is the minimum length of ODN required for optimal stimulation of mouse cells (287), while 21 bases appears to be the minimum required for the stimulation of human cells (334, 341), which correlates with the length restrictions for binding to DEC-205 that our data demonstrate (Figure 3.4). Our data show that DEC- 205 preferentially binds Ts over other bases, and evidence in the literature suggests that Ts are also preferred for optimal ODN stimulatory activity. For instance, the ODN sequence found to stimulate human cells most efficiently, 2006, consists primarily of T residues outside of its three CpG motifs (295). Furthermore, studies have shown that substitution of T residues with other bases appears to reduce stimulatory activity, while greater T content enhances ODN stimulatory ability (292, 341). Even in the absence of CpG motifs, poly-T ODNs at least 20 bases in length have been shown to have weak stimulatory activity, which was not observed for ODNs of similar length composed of other bases (334, 342). It should be noted that the stimulatory activity of CpG-free poly-T ODNs has only been demonstrated in in vitro assays, and in our in vivo assays, we saw no detectable stimulatory activity of 14T ODN when injected into mice (data not shown).

134

Overall, there appears to be a close correlation between properties that confer optimal DEC-205 binding and properties that enhance the stimulatory activity of ODNs. This would suggest that certain ODN features known to confer greater immunostimulatory activity, such as their length and structure, could be influencing their activity by mediating more effective binding and uptake by DEC-205. This notion was further investigated by examining the impact of DEC-205 binding on the stimulatory capacity of ODNs in both in vitro and in vivo assays, as discussed below.

Lack of DEC-205 binding may restrict the uptake of self-DNA

ODNs that are either fully diester or double stranded are unable to bind DEC-205. This strongly suggests that normal mammalian DNA, which possesses both of these properties, is not recognised by DEC-205. It is well established that bacterial DNA is many orders of magnitude more stimulatory than vertebrate DNA (284, 343, 344). This was often attributed to the fact that bacterial DNA has a higher proportion of unmethylated CpG motifs that activate TLR9, while vertebrate DNA has a lower CpG content and the C is typically methylated (283, 345). Studies have also suggested that vertebrate DNA contains inhibitory sequences that actively inhibit the stimulation of TLR9 (285, 346, 347). However, vertebrate DNA has been shown to be stimulatory if artificially introduced into endosomes (348-350). Thus, to prevent autoimmune reactions to self-DNA, it might be expected that mechanisms exist to exclude self- DNA from the endosome, a hypothesis that has been suggested by previous studies (349, 351, 352). Our findings implicate a role for DEC-205 in the regulation of this process. The inability of DEC-205 to bind ODNs with structural properties matching that of mammalian DNA could potentially be a mechanism by which uptake of self- DNA is restricted.

Differential DEC-205 binding may control the divergent stimulatory effects of A-ODNs and B-ODNs

Previous studies have determined that the divergent immune effects induced by B- ODNs or A-ODNs are due to differences in the endosomal location where ODNs

135

interact with TLR9 (318, 319). B-ODNs are found in late endosomes and lysosomes, while A-ODNs are instead retained in early endosomes. This in turn appears to be regulated by the ss versus aggregated nature of the ODNs (319). Interestingly, our findings indicate that the same structural attributes also determine the strong and weak DEC-205-binding capacity, respectively, of B-ODNs and A-ODNs. In conjunction with the reported ability of DEC-205 to deliver ligands to late endosomal compartments (225), these data could imply that the selective binding of DEC-205 to B-ODN but not A-ODN is part of the mechanism that ensures B-ODNs, but not A- ODNs, are delivered to late endosomal compartments to produce the stimulatory effects characteristic of B-ODNs. Another receptor, CXCL16, has previously been reported to mediate the uptake of A-ODN, but not B-ODN (327), which would align with a hypothesis that ODN class-specific stimulatory effects originate from uptake via ODN class-specific receptors.

2006 is dependent on DEC-205 for optimal stimulatory activity

Our previous study demonstrated that the immunostimulatory effects of the B-ODN 1668 are significantly abrogated in DEC-205-deficient mice (264). It was expected that 2006, another B-ODN that binds to DEC-205 highly efficiently, would exhibit a similar dependency on DEC-205. Indeed, the activation of a purified B cell population by 2006 was dependent on the presence of DEC-205 (Figure 3.8). In this assay, 1668 induced far greater activation of B6 B cells than 2006 at the same concentrations. In fact, the plateau in responsiveness of B6 cells to 1668 at higher doses would suggest that these doses are beyond saturation for this system. Over-saturating amounts of 1668 could allow uptake by non-specific or low-affinity receptors, which could explain the DEC-205-independent activity observed at these high doses (Figure 3.8A). Despite inducing B cell activation, 2006 stimulation seemed unable to induce IL-6 production from B6 B cells at the same doses that elicited strong IL-6 production upon 1668 stimulation (Figure 3.8C). The stimulation of B cells to produce IL-6 is a hallmark feature of B-ODNs, and 2006 has been shown on numerous occasions to induce high levels of this cytokine from human PMBCs (296, 298). The lack of stimulation of mouse B cells in our assay may reflect an inefficiency of the human optimal CpG

136

motifs in 2006 to stimulate mouse B cells. A similar phenomenon has been observed in previous studies (314, 353). However, seemingly in contradiction to this conclusion, 1668 and 2006 were found to be equivalent in their ability to enhance OT- I T cell responses when delivered as an adjuvant to mice (Figure 3.19A). This could suggest that 2006 is only less effective at directly activating murine B cells, but this does not significantly impact its capacity to promote T cell responses in vivo.

21798 does not require DEC-205 for stimulatory activity

Our study of 21798 demonstrated that a CpG ODN with poor DEC-205 binding has correspondingly little dependence on DEC-205 for its immunostimulatory activity, supporting a trend suggesting that the propensity for stimulatory ODNs to act via DEC-205 can be predicted by their DEC-205 binding capacity (Figure 3.8). Importantly, this also demonstrates that ODNs capable of efficient B cell stimulation do not necessarily act via DEC-205. A correlation between DEC-205 binding and the capacity of ODNs to stimulate B cells might have been inferred from the fact that B- ODNs, strong B cell activators, bind DEC-205, while A-ODNs, which do not effectively induce B cell activation, also do not bind DEC-205. However, 21798 is capable of stimulating B cells as efficiently as 2006 (Figure 3.8) and other B-ODNs (300), yet does not require DEC-205 for optimal activity. This indicates that DEC-205 is not the only receptor to mediate the uptake of ODNs by B cells. Further investigations to determine the uptake mechanism of ODNs such as 21798 and A- ODNs that do not bind DEC-205 may lead to the identification of other receptors involved in mediating the uptake and activity of stimulatory ODNs. CXCL16 is a potential candidate, having previously been demonstrated to mediate the uptake of A- ODN (327). Although Gursel et al.’s study did not detect CXCL16 on B cells, expression of CXCL16 on B cells has been demonstrated in other reports (354, 355). Whether CXCL16 can bind to 21798 is currently unknown.

Notably, the stimulation of B6 B cells by 21798 was comparable to 2006, which was far less effective than 1668 at the same doses (Figure 3.8A and B). As the original report of the potency of 21798 utilised human cells (300), this may indicate that the

137

stimulatory motifs in 21798, like 2006, are relatively inefficient in mice. However, 21798 was found to be more effective than 1668 at stimulating OT-I proliferation when delivered as an adjuvant alongside targeted Ag (Figure 3.19A). This could indicate that, while the capacity of 21798 to stimulate purified mouse B cells is lower than 1668, this may not affect its ability to promote mouse T cell responses in vivo. Whether the hierarchy of ODN potency demonstrated by in vitro assays with specific cell populations accurately reflects their activity in vivo is an important point to consider, particularly as the adjuvant capacity of ODNs has historically been judged by such in vitro assays. Our data would suggest the potency of ODNs in one assay – for example, in vitro B cell stimulation – is not necessarily predictive of their effects as an adjuvant in vivo, where the outcome is influenced by a combination of many factors, including the cytokine milieu and the interaction between different cell types.

Non-CpG ASO 518477 exhibits some DEC-205-dependent stimulatory activity

The anti-sense ODN 518477 induces the unintended side effect of immune activation when administered at high doses, despite lacking CpG motifs. The structural attributes of 518477 – fully thioate, single-stranded, 18 bases in length – suggested there was a high probability it would bind DEC-205. Indeed, 518477 bound DEC-205, though less effectively than 2006 (Figure 3.9). This prompted us to investigate whether, like other DEC-205-binding ODNs, the stimulatory activity of 518477 is mediated by DEC-205. If so, the immunostimulatory effects of 518477 could be dampened by interfering with its DEC-205 binding. Within 6 hours of 518477 administration, both B cells and DCs showed signs of activation such as modest upregulation of CD86 and MHC II expression, although these early effects appeared to be independent of DEC-205 (Figure 3.10 and 3.12). At the later time point of 24 hours, reduced activation of both DCs and B cells in DEC-205-/- mice was seen (Figure 3.11 and 3.13), implying that DEC-205 may eventually contribute to the immune effects of 518477.

138

It remains to be seen if the role of DEC-205 at the later stage of activation by 518477 is significant enough that employing methods to reduce DEC-205 binding will substantially reduce the side effects of this ASO. However, this is unlikely to completely remove the immunostimulatory effects of 518477, as the immediate direct activation of APCs does not appear to require DEC-205. Nevertheless, should other ASOs with undesirable stimulatory side effects be identified to rely on DEC-205 to a greater degree, strategies to reduce DEC-205 binding could prove to be beneficial.

Interestingly, the immune effects of 518477 were completely abrogated in TLR9 deficient mice. There have been contrasting reports on whether the immune responses induced by ASOs are TLR9-dependent or not (335, 336), and so far it is unclear what attributes of ASOs determine whether their stimulatory activity is TLR9- dependent. In any case, the dependence of 518477 upon TLR9 for its immunostimulatory activity clearly demonstrates that, not only the uptake of ODNs, but also specifically the stimulation of TLR9, can occur in a DEC-205-independent manner.

Enhanced DEC-205 binding improves the capacity of CpG ODNs to promote CD8+ T cell responses

Our data strongly suggest that, for ODNs that are capable of binding DEC-205, there is a positive correlation between DEC-205 binding and ODN potency. DEC-205 binding is mainly observed in the B-class of ODNs, and while DEC-205 may not be the only receptor responsible for their uptake and delivery to TLR9-containing endosomes, it certainly plays a role in mediating their activity. A significantly diminished response to two different B-ODNs was seen in DEC-205-/- mice (Figure 3.8)(264), implying that modifications to these ODN that reduce DEC-205 binding would reduce their activity. Theoretically, the opposite should also be true. Enhanced DEC-205 binding should improve the stimulatory capacity of ODN, potentially even those that do not normally bind DEC-205. This led us to investigate whether ODN adjuvant potency can be improved by enhancing their DEC-205 binding capacity.

139

Our investigations have revealed specific rules about DEC-205 binding that can be altered in ODN to enhance or decrease DEC binding. For instance, introducing even one diester bond to interrupt an otherwise thioate backbone is enough to severely inhibit binding (Figure 3.6). On the other hand, if DEC-205 binding is desired, diester bonds and sequences that cause aggregation such as palindromes and multiple G runs should be avoided. Another, more versatile method of improving DEC-205 binding would be to add a thioate ODN sequence known to have strong DEC-205 binding, to act as a DEC-205-targeting motif. Naturally, this motif could be a CpG ODN such as 1668 or 2006 with demonstrated strong DEC-205 binding. We have also identified a non-stimulatory sequence that binds DEC-205 just as efficiently, 14T, which can act as a DEC-205 binding motif without adding extra stimulatory motifs. This could be a useful tool to study the effect of DEC-205 binding on the activity of particular ODNs, without altering their stimulatory effects.

We have shown that the addition of 14T or 2006 as a DEC-205 targeting motif to the 5’ end of 21798 resulted in improved DEC-205 binding (Figure 3.15 and 3.17), which led to a striking improvement in its potency as an adjuvant. Compared to an equivalent amount of a mixture of the two separate component ODNs, the combined 14T-21798 and 2006-21798 ODNs were able to induce greater Ag-specific CD8+ T cell responses when administered with Clec9A-targeted Ag (Figure 3.16 and 3.18A). The 2006-21798 combined ODN was found to be more effective than 14T-21798, which was not surprising as it contains additional stimulatory CpG motifs.

Interestingly, while the response to 2006-21798 was significantly reduced in DEC-205- /- mice, it was still greater than the response to the mixed 2006 + 21798 in DEC-205-/- mice (Figure 3.18A). We have shown that the stimulatory effects of 21798 are largely DEC-205-independent. It is possible that our new combined 2006-21798 ODN, apart from an improved capacity to stimulate immunity in a DEC-205-dependent manner, may also have an improved ability to activate these DEC-205-independent mechanisms. This could explain the greater response to 2006-21798 than the separated 2006 + 21798 even in the absence of DEC-205.

140

Another explanation for this observation is that co-injection of the 2 component ODNs could cause inhibition of their activity, resulting in lower responses than would be observed if equivalent amounts of only one type of ODN was used. Indeed, in another experiment in which CD8+ T cell responses were measured by proliferation of transferred OT-I cells, we found the mixed 2006 + 21798 to be less effective than an equivalent amount, by mass, of purely 21798 (Figure 3.19A). It is possible that this is caused by competition between 2006 and 21798 ODNs for uptake via the same receptors, although this seems highly unlikely at the low doses used for immunisation, which we have determined to be below saturation (data not shown).

It has previously been observed that co-administration of an A-ODN with a B-ODN can result in lower responses than administration of only one ODN (289, 356). It appears that activation of the stimulatory pathway associated with A-ODNs prevents effective activation of the distinct pathway triggered by B-ODNs, and vice versa. Indeed, studies have shown that high doses of B-ODNs can specifically inhibit the induction of type I interferons by A-ODNs, without affecting the production of other inflammatory cytokines such as TNFα and IL-12p40 (357, 358). This process appears to involve inhibition of pathways specific for the first-wave of induction of type I interferon upon ODN stimulation (357, 359), although indirect inhibition by paracrine activity of IL-10 induced by B-ODN stimulation has also been proposed (358).

The co-administration of a B-ODN and P-ODN as in our mixed 2006 + 21798 could also cause a similar inhibitory phenomenon. 2006 + 21798 would be an ideal control for the activity of 2006-21798 if not for this potential inhibitory effect. Nevertheless, the improved stimulatory capacity caused by enhanced targeting of 21798 to DEC-205 can also be clearly demonstrated by the improved responses generated by 2006-21798 compared with an equivalent amount, by mass, of the original 21798 ODN (Figure 3.19A).

141

Enhanced DEC-205 binding of CpG ODNs has no effect on the stimulation of CD4+ T cells or B cells

While 2006-21798 displayed an enhanced ability to stimulate CD8+ T cell responses compared with the single 2006 or 21798 ODNs, or a mixture of the two, it did not significantly improve CD4+ T cell responses (Figure 3.19B). This may be because targeting Ag via anti-Clec9A mAbs is known to promote strong CD4+ T cell responses even in the absence of adjuvant (215, 360), such that small differences in adjuvant potency may be more difficult to observe. Furthermore, the stimulatory effects of CpG ODNs are known to promote a Th1 environment that preferentially enhances CD8+ T cell responses (284). Additionally, DEC-205 expression is highest on CD8+ DC and CD103+ DCs, collectively knows as cDC1s, which are specialised for the cross-presentation of Ag and production of cytokines such as IL-12 to activate CD8+ T cells (75). Thus, it would be expected that increasing the DEC-205 binding capacity of ODNs, as we have for 2006-21798, would result in enhanced delivery to predominantly cDC1s, resulting in greater activation of CD8+ T cells.

Increased DEC-205 binding also appeared to have little effect on the stimulation of B cells. There was no obvious advantage of 2006-21798 over the mixed 2006 + 21798 or the original 21798 in their ability to activate purified B cells, and the stimulation of B cells by 2006-21798 was not significantly diminished in DEC-205-/- mice (Figure 3.21). Systemic antibody responses also did not seem to be strongly enhanced. While mice immunised with 2006-21798 alongside targeted Ag appeared to have higher Ab titres than mice immunised with targeted Ag alone, 2006-21798 was not any more effective than the 2006 + 21798 mixture (Figure 3.18B). It should be noted that, since targeting Ag to Clec9A can induce strong Ab responses even in the absence of adjuvant, the advantage of more potent adjuvants may be more difficult to discern in this system.

It should also be considered that these Ab titres were measured at day 6 post- immunisation, at a relatively early stage of the response. Thus, it is unclear whether the peak titres induced in the presence of adjuvant would also be higher than those

142

induced by targeted Ag alone, or if the results seen at day 6 merely represent the more rapid induction of a response that reaches a similar maximal level. Similarly, we cannot conclude whether the slightly lower titres seen in DEC-205-/- at this early time point reflects slower kinetics or an ultimately lower response. Serum measurements at later time points, such as 2 or 4 weeks post-immunisation, would be required to thoroughly investigate the kinetics of the Ab response induced in the presence of 2006-21798.

Targeting DEC-205 to design more potent adjuvants

In this chapter, we have examined the molecular requirements for the interaction between DEC-205 and ODN and investigated the importance of this interaction for the stimulatory function of ODNs. We then demonstrated how this information could be utilised to design more potent CpG ODN adjuvants. With the goal of being able to translate our findings to practical benefits in the clinic, we focused our study on ODNs that are most relevant to humans – 2006, the main stimulatory ODN currently used in human clinical trials, and 21798, a relatively new ODN reported to be most stimulatory of the known ODN classes. Our work has shown that modulating the DEC-205 binding of ODNs is a powerful tool to control their stimulatory activity, and can be accomplished in a relatively straightforward manner. Hopefully, this will facilitate the rational design of ODNs with more defined immune effects, which can be used in humans to promote optimal adjuvant activity while avoiding unexpected side effects.

Targeting DEC-205 is not a novel concept, and a large body of work has already demonstrated that delivering Ag to DEC-205 using mAbs can promote strong immune responses. However, the targeting mAb itself does not activate APCs and co- administration of adjuvant is required for the induction of immunity (220, 227). Our study is the first to describe a novel method of targeting adjuvant to DEC-205 by using 2006 as a targeting motif. This offers the possibility that 2006 could also be used as a means of targeting Ag to DEC-205, with the advantage of simultaneously delivering a potent adjuvant. In fact, immunisation with CpG ODN-Ag conjugates

143

has previously been demonstrated to be significantly more efficient at inducing immune responses than a mixture of the two separate components (361-363). This was hypothesized to be due to preferential uptake of the Ag into TLR9-containing endosomes in APCs, which could mediate more efficient Ag presentation, along with the simultaneous delivery of an activation signal via the CpG component (364). Our findings indicate that the conjugation of CpG ODN to Ag would likely promote uptake specifically via DEC-205. Therefore, it is possible that the increased efficacy of CpG ODN-Ag conjugates may in fact be due to enhanced targeting of Ag to DEC-205, which is known to promote the presentation and cross-presentation of Ag to prime strong T cell responses. Altogether, these findings suggest that using CpG ODN motifs to combine the targeting of Ag and adjuvant to DEC-205 in a single molecule could prove to be a viable method of developing more potent vaccines.

Our results suggest that DEC-205 is not the only receptor responsible for the uptake of ODNs for immune stimulation. Future work to identify and characterise other receptors that mediate ODN activity, and may even act in synergy with DEC-205, would further refine our understanding of ODN function, and potentially provide us with more tools to control the immune effects of ODNs. The investigation of one such potentially synergistic CpG ODN receptor, CD14, is the subject of Chapter 4.

144

Chapter 4: The contribution of CD14 to CpG ODN uptake and function

Abstract

DEC-205 has been shown to mediate binding and uptake of CpG ODNs, a process that is required for their optimal stimulatory activity. However, some responsiveness to CpG ODNs is still observed in DEC-205-/- mice, suggesting that other receptors can also mediate the uptake of CpG ODNs for immune stimulation. Several receptors have been reported to have CpG ODN binding properties, including the receptor CD14. CD14-/- mice were reported to have similar deficiencies in response to CpG ODN stimulation as DEC-205-/- mice. In light of the overlapping expression patterns of DEC-205 and CD14 on key cell types such as DCs, CD14 appeared to be a strong candidate for a CpG ODN-binding receptor that may cooperate with DEC-205 synergistically. We sought to investigate whether this was the case by assessing the responses of DEC-205-/-, CD14-/- and DEC-205-/-CD14-/- mice to CpG ODN stimulation. In contrast to the previous report, we found no effect of CD14 deficiency on the response to CpG ODNs, in either the presence or absence of DEC-205. We also did not observe a role of CD14 in the direct binding or uptake of CpG ODN.

145

Introduction

As demonstrated in Chapter 3, and previously by Lahoud et al. (264), DEC-205 is involved in mediating the immunostimulatory activity of CpG ODNs, specifically, B- ODNs such as 1668 and 2006. However, DEC-205-deficient mice are not completely incapable of responding to CpG ODN stimulation, suggesting that DEC-205 is not the only receptor capable of facilitating CpG ODN uptake. Several receptors have been shown to bind CpG ODNs, such as RAGE, KIR, MR, SR, but for most of these receptors, the significance of this interaction on the in vivo response to CpG ODNs has not been shown (reviewed in Chapter 3 Introduction).

One exception is CD14, a glycosylphosphatidylinositol-anchored, membrane- associated protein, as mice lacking this receptor were shown to have significantly impaired responses to CpG ODN. Baumann et al. found that bone-marrow derived DCs (BMDC) and macrophages (BMDM) from CD14-/- mice produced less IL-6 in response to CpG ODN stimulation than cells from B6 mice (332). This was recapitulated in an in vivo setting by intraperitoneal injection of B6 or CD14-/- mice with CpG ODN. Four hours post-injection, CD14-/- mice showed significantly reduced levels of the cytokines IL-6, keratinocyte chemoattractant (KC) and IL-1β in the peritoneal lavage compared with B6. This in turn appeared to limit the infiltration of neutrophils into the peritoneal cavity in CD14-/- mice. CD14 was shown to mediate the uptake of CpG ODN by peritoneal macrophages, and direct interaction of CpG ODNs with CD14 was inferred from ELISA binding assays.

As CD14 is known to mediate the stimulatory effects of LPS (365, 366), care must be taken to verify that any observed CD14-mediated effects are not caused by LPS contamination. Baumann et al. excluded the possibility of LPS contamination through extensive testing (332). A Limulus amebocyte lysate test showed no detectable levels of LPS in their CpG ODN reagents, and addition of an LPS-inhibiting agent had no effect on the production of IL-6 in response to CpG ODN stimulation. Similarly, TLR4-deficient macrophages had no defect in their response to CpG ODN

146

stimulation. Collectively, these experiments appeared to discount any contribution of LPS to the observed effects of CpG ODN in their assays.

The diminished response to CpG ODN in the absence of CD14 was highly reminiscent of the impaired response we saw in the absence of DEC-205. Therefore, we investigated whether CD14 could be the receptor responsible for the uptake of CpG ODN in the absence of DEC-205, and account for the residual response seen in DEC-205-/- mice. Additionally, we considered the potential for the two receptors to act synergistically in the same cells to promote the activity of CpG ODNs. To this end, the binding, uptake and stimulatory effects of CpG ODN in the absence of CD14, or both DEC-205 and CD14, was examined.

147

Results

4.1. Surface expression of CD14 and DEC-205 on DCs and macrophages

We have previously shown that DEC-205 is required for the optimal CpG-mediated activation of APCs, in particular the CD8+ DC subset, which expresses the highest levels of DEC-205. However, CpG ODN can still activate these cells, although to a significantly lower degree, in DEC-205-/- mice (264). To investigate the possibility that CD14 mediates the activation of these cells in the absence of DEC-205, we first assessed whether the two receptors were expressed in the same cells.

CD14 was expressed at very high levels on peritoneal macrophages, and at lower levels on CD8+ DCs, CD8- DCs and pDCs (Figure 4.1). DEC-205 was barely detectable on peritoneal macrophages and pDCs, but was highly expressed on CD8+ DCs, and at lower levels on CD8- DCs. Thus, CD14 and DEC-205 have overlapping expression on cDCs, but distinct expression on pDCs and macrophages.

Figure 4.1. Expression of CD14 and DEC-205 on DCs and macrophages. Splenic DCs (CD8+ DCs, CD11chiCD8+CD45R-; CD8- DCs, CD11chiCD8-CD45R-; pDCs, CD11cintCD45R+) and peritoneal macrophages (CD11bhiF4/80+) were stained with anti-DEC-205 or anti-CD14 (black), or isotype control mAb (grey). Representative data of 3 independent experiments are presented. Experiments performed by Fatma Ahmet, figure generated by Irina Caminschi.

148

4.2. The requirement of CD14 for the in vivo activation of DCs by CpG ODN

Baumann et al. demonstrated that CD14 is necessary for optimal responses of BMDC and BMDMs to CpG ODN stimulation in culture (332). We attempted to extrapolate these findings to an in vivo situation, by intravenous injection of B6 or CD14-/- mice with 5 nmol of CpG ODN and measuring the activation of CD8+ DCs 5 hours later. In B6 mice, CpG ODN treatment caused significant activation of DCs, as measured by the upregulation of co-stimulatory markers CD40, CD80, CD86 and MHC II (Figure 4.2). Surprisingly, CD14-/- DCs were activated to a similar degree as B6, with no apparent difference in the upregulation of any of these markers.

A CD40 CD80 CD86 MHC II

B6 uninjected CD14-/-

CpG ODN B6 CD14-/- injected

2 3 4 5 2 3 4 5 2 3 4 5 2 3 4 5 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 B CD40 CD80 CD86 MHC II 250 ns 120 ns 300 ns 200 ns 200 100 150 80 150 200 60 100 100 40 100 50 50

%MFI increase 20 %MFI increase %MFI increase %MFI increase 0 0 0 0 B6 CD14-/- B6 CD14-/- B6 CD14-/- B6 CD14-/-

Figure 4.2. Activation of DCs in vivo by CpG ODN does not require CD14. Two B6 or CD14-/- mice were injected i.v. with 5 nmol of CpG ODN 1668, or left uninjected. Five hours later, spleens were harvested and pooled, DCs isolated, and the expression of CD40, CD80, CD86 and MHC II on CD8+ DCs examined by flow cytometry. (A) Representative histograms of CD40, CD80, CD86 and MHC II expression on CD8+ DCs from uninjected or CpG ODN injected mice. Experiments were performed 5 times. (B) Percentage increase in MFI of CD40, CD80, CD86 and MHC II on CD8+ DCs from CpG ODN injected versus uninjected mice. Bars represent mean ± SEM of 5 independent experiments. Each point represents data from one experiment.

149

It is possible that the presence of DEC-205 in CD14-/- mice may compensate for a potentially subtle deficiency in uptake caused by a lack of CD14. If this were to be the case, we would expect any contribution of CD14 to CpG ODN uptake to be more readily observed in the absence of DEC-205. Thus, we compared the ability of DEC- 205-/- and DEC-205-/-CD14-/- double knockout mice to respond to CpG ODN. DEC- 205 deficiency resulted in significantly reduced upregulation of CD40 and CD80 by CD8+ DCs after CpG ODN treatment (Figure 4.3), reaffirming our previous data (264). This was more apparent at the higher dose of 20 nmol of CpG ODN (Figure 4.3B vs C). However, the additional loss of CD14 in DEC-205-/-CD14-/- mice did not further diminish the response. DEC-205-/-CD14-/- DCs were activated equally well as DEC-205-/- DCs after in vivo stimulation with CpG ODN.

A 20 nmol CD40 CD80 CD86 MHC II B6 uninjected DEC-/- DEC/CD14-/-

B6 CpG ODN DEC-/- injected DEC/CD14-/-

2 3 4 5 2 3 4 5 2 3 4 5 2 3 4 5 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 B 5 nmol

CD40 CD80 CD86 MHC II ns ns 250 * ns 100 * ns 300 ns ns 200

200 80 150 200 150 60 100 100 40 100 50 20 50 %MFI increase %MFI increase %MFI increase %MFI increase 0 0 0 0 B6 DEC-/- DEC-/- B6 DEC-/- DEC-/- B6 DEC-/- DEC-/- B6 DEC-/- DEC-/- CD14-/- CD14-/- CD14-/- CD14-/- C 20 nmol

CD40 CD80 CD86 MHC II

300 * ns 200 * ns 300 * ns 400 ns ns

150 300 200 200 100 200 100 100 50 100 %MFI increase %MFI increase %MFI increase %MFI increase 0 0 0 0 B6 DEC-/- DEC-/- B6 DEC-/- DEC-/- B6 DEC-/- DEC-/- B6 DEC-/- DEC-/- CD14-/- CD14-/- CD14-/- CD14-/-

150

Figure 4.3. DEC-205, but not CD14, is required for optimal activation of DCs in vivo by CpG ODN. Two B6, DEC-205-/- (DEC-/-) or DEC-205-/-CD14-/- (DEC/CD14-/-) mice were injected i.v. with (A, C) 20 nmol or (B) 5 nmol of CpG ODN 1668, or left uninjected. Five hours later, spleens were harvested and pooled, DCs isolated, and the expression of CD40, CD80, CD86 and MHC II on CD8+ DCs examined by flow cytometry. (A) Representative histograms of CD40, CD80, CD86 and MHC II expression on CD8+ DCs from uninjected or CpG ODN injected mice. Experiments were performed twice. (B, C) Percentage increase in MFI of CD40, CD80, CD86 and MHC II on CD8+ DCs from CpG ODN injected versus uninjected mice. B6 data shown in (B) is from same experiments presented in Figure 4.2. Bars represent mean ± SEM of (B) 5 or (C) 2 independent experiments. Each point represents data from one experiment.

4.3. The requirement of CD14 for the induction of serum cytokines by CpG ODN

CD14-/- mice were previously reported to have a diminished capacity to produce cytokines in response to CpG ODN treatment (332). In that study, reduced levels of IL-6, KC and IL-1β were observed in the peritoneal lavage of CD14-/- mice after intraperitoneal injection of CpG ODN. In our previous work with DEC-205-/- mice, a similar defect in serum cytokine production was observed after intravenous administration of CpG ODN (264). Thus, we sought to investigate whether the loss of both DEC-205 and CD14 in mice would result in a cumulative decrease in serum cytokine production.

DEC-205-/- mice showed significantly impaired production of IL-6 and IL-12 in the serum, in accordance with our previous work (Figure 4.4A and B). However, the additional loss of CD14 in DEC-205-/-CD14-/- mice caused no further reduction in cytokine production. Furthermore, even CD14-/- mice showed no decreased production of cytokines in response to CpG ODN compared with B6 (Figure 4.4A and

151

B). Thus, regardless of the presence of DEC-205, the absence of CD14 appeared to have no effect on the induction of serum cytokines by CpG ODN.

Figure 4.4. Production of serum cytokines in response to i.v. CpG ODN injection does not require CD14. B6, CD14-/-, DEC-205-/- or DEC-205-/-CD14-/- mice were injected i.v. with 15 nmol of CpG ODN 1668. Serum samples were collected 3 hours later and the concentration of (A) IL-6 and (B) IL-12 measured by ELISA. Mean ± SEM of pooled data from multiple experiments are presented, each point represents one mouse. Experiments performed by Fatma Ahmet and Irina Caminschi, figure generated by Irina Caminschi.

4.4. Binding of CpG ODNs to CD14 by ELISA

Our data so far indicate that mice lacking CD14 have perfectly intact responses to CpG ODN stimulation, both in terms of the in vivo activation of DCs and the systemic production of cytokines. This seemed to conflict with the findings of Baumann et al., where CD14 played a non-redundant role in the uptake of CpG ODNs (332). Since the first requirement of an uptake receptor is to bind the target molecule, we sought to confirm whether CD14 directly interacts with CpG ODNs.

152

Baumann et al. showed binding of CD14 recombinant protein to CpG ODN-coated plates in an ELISA assay, and further demonstrated the inhibition of this interaction by competition with LPS, a known CD14 ligand (332). We obtained two different soluble CD14 constructs to reproduce their observation, one tagged with Fc (CD14- Fc) as was used in their report, and another tagged with His (CD14-His). ELISA plates coated with CD14-Fc, CD14-His or DEC-205 overnight were washed and incubated with biotin-labelled CpG ODN, or LPS as a positive control. CpG ODNs bound to DEC-205, as expected, but also to the CD14-Fc construct (Figure 4.5A). However, no binding of CpG to the CD14-His construct was observed (Figure 4.5A). The CD14- His construct also appeared to have markedly weaker binding to LPS than the CD14- Fc construct, only slightly better than the non-specific binding of LPS to DEC-205 or in the absence of protein (Figure 4.5B). This led us to consider whether the lack of binding of CD14-His to CpG ODN may be due to a smaller amount of this construct being bound to the plate. However, an anti-CD14 mAb detected equivalent amounts of both constructs bound to the plate (Figure 4.5C). This raised concern about the integrity of the proteins, and whether they accurately reflect the binding properties of the native protein.

153

A CpG binding B LPS binding 1.2 1.2

1.0 1.0

0.8 0.8

0.6 0.6 O.D. O.D. 0.4 0.4

0.2 0.2

0.0 0.0

CD14-Fc CD14-His DEC-205 CD14-Fc CD14-His DEC-205no protein C Anti-CD14 detection

2.5 CD14-Fc 2.0 CD14-His DEC-205 1.5

O.D. 1.0

0.5

0.0 1 10 100 1000 10000 protein (ng/ml)

Figure 4.5. Different recombinant CD14 constructs display inconsistent binding properties. ELISA plates were coated with CD14-Fc (A, B: 3 μg/ml), CD14-His (A, B: 3 μg/ml), mouse DEC-205 (A, B: 10 μg/ml), or no protein. Plates were incubated with (A) biotinylated CpG ODN 1668 (1 ug/ml) (B) LPS biotin (1 μg/ml) or (C) anti-CD14 biotin (0.25 μg/ml), then washed and incubated with streptavidin HRP and visualised with ABTS substrate. Shown is the mean optical density measured at 405-490nm of duplicate samples. Experiments were performed (A, B) twice and (C) once, representative data is shown. Experiment shown in (A) was performed by Irina Caminschi.

154

4.5. The requirement of CD14 for the binding and uptake of CpG ODN by peritoneal macrophages

The inconsistent results obtained with two different soluble CD14 constructs raised doubts over the validity of this approach to verify the binding of CpG ODN. This prompted us to examine the capacity of endogenously expressed CD14 to bind CpG ODN, which is more likely to be representative of what occurs in vivo. Peritoneal macrophages endogenously express very high levels of CD14, so we examined the binding of CpG ODN to peritoneal macrophages obtained from B6 or CD14-/- mice. Both B6 and CD14-/- macrophages appeared equivalent in their ability to bind CpG ODN after 30 min incubation at 4ºC (Figure 4.6A). A similar experiment was performed at 37ºC in complete media, to allow active uptake of CpG ODN. After 1 hour, considerably more CpG ODN was observed on macrophages than at 4ºC, but again, there was no difference in uptake between B6 and CD14-/- cells (Figure 4.6B). As a control, the binding of LPS to macrophages was tested in parallel, which was entirely abolished in the absence of CD14, as expected (Figure 4.6C).

155

A B

CpG ODN binding 4ºC CpG ODN uptake 37ºC 80 B6 400 B6 CD14-/- CD14-/- 60 300 I I F F 40 200 M M

20 100

0 0 0 10 100 1000 0 10 100 1000 1668-Cy3 (nM) 1668-Cy3 (nM)

C LPS binding 4ºC 1500 B6 CD14-/-

1000 I F M 500

0 0 10 100 1000 LPS-biotin (ng/ml)

Figure 4.6. The binding and uptake of CpG ODNs by peritoneal macrophages does not require CD14. Macrophages obtained from peritoneal lavage of B6 or CD14-/- mice were incubated with the indicated doses of CpG ODN 1668-Cy3 or LPS-biotin for (A, C) 30 min at 4ºC in EDTA-BSS-2%FCS or (B) 1 hour at 37ºC in complete media. Fluorescence corresponding to bound CpG ODN or LPS (detected by streptavidin PE) of macrophages (CD11b+) was determined by flow cytometry. Shown is mean MFI ± SEM of 4 replicates pooled from 2 independent experiments.

156

4.6. The requirement of CD14 for the binding of various classes of CpG ODN to peritoneal macrophages in the absence of DEC-205

Although peritoneal macrophages do not express appreciable levels of DEC-205 (Figure 4.1), it is conceivable that other cells in the peritoneal wash could express DEC-205 and compete for uptake of CpG ODN. To avoid this potentially confounding factor, we examined the binding of CpG ODN to peritoneal macrophages from DEC-205-/- or DEC-205-/-CD14-/- mice. We tested a panel of CpG ODNs, including two B-ODNs – 1668, which was used in all previous experiments in this Chapter, and 1826, which was used for the investigations in Baumann et al. (332) and differs from 1668 by only 2 bases. We also tested representatives of other ODN classes, namely, the A-ODN 2216, and C-ODN 2395. For all these ODNs, we saw no difference in binding to DEC-205-/- or DEC-205-/-CD14-/- macrophages, indicating no role of CD14 (Figure 4.7). Once again, LPS binding was entirely dependent on the presence of CD14.

157

1668 1826 1500 800 DEC-/- DEC-/- DEC/CD14-/- DEC/CD14-/- 600 1000 I I F F 400 M M 500 200

0 0 0 10 100 1000 0 10 100 1000 CpG ODN (nM) CpG ODN (nM)

2006 2216 600 1500 DEC-/- DEC-/- DEC/CD14-/- DEC/CD14-/- 400 1000 I I F F M M 200 500

0 0 0 10 100 1000 0 10 100 1000 CpG ODN (nM) CpG ODN (nM)

2395 LPS 1000 1500 DEC-/- DEC-/- 800 DEC/CD14-/- DEC/CD14-/- 1000

I 600 I F F M M 400 500 200

0 0 1 10 100 1000 1 10 100 1000 CpG ODN (nM) LPS (ng/ml)

Figure 4.7. The binding of A-, B- or C-class CpG ODN to peritoneal macrophages does not require CD14. Macrophages obtained from peritoneal lavage of DEC-205-/- (DEC-/-) or DEC-205-/-CD14-/- (DEC/CD14-/-) mice were incubated with the indicated doses of biotinylated CpG ODN for 30min at 4ºC in EDTA-BSS-2%FCS. Bound CpG ODN was detected by streptavidin PE and fluorescence of macrophages (CD11b+) determined by flow cytometry. Shown is mean MFI ± SEM of pooled duplicates from 2-3 independent experiments. LPS control experiment was performed once.

158

Discussion

CD14 does not play a role in the uptake of CpG ODNs for immune stimulation

Contrary to our expectations, our results revealed no apparent role of CD14 for the uptake or immunostimulatory effects of CpG ODN. Compared to B6, CD14-/- mice showed no defect in the production of cytokines or activation of DCs in response to CpG ODN stimulation. Similarly, while the absence of DEC-205 significantly reduced responses to CpG ODN, as previously reported (264), the additional loss of CD14 in DEC-205-/-CD14-/- deficient mice had no exacerbating effect. Furthermore, we did not observe a role for CD14 in the binding or uptake of CpG ODNs by peritoneal macrophages.

These findings appear to be in direct conflict with the report by Baumann et al. (332). However, several differences in experimental design should be taken into consideration. Baumann et al. observed reduced stimulatory effect of CpG ODNs in the absence of CD14 by measuring cytokine production from BMDC or BMDM incubated with with CpG ODN for 6 hours in culture. We did not see a similar impact of CD14 deficiency on the activation of DCs in vivo (Figure 4.2, 4.3). This may speak to the phenotypic and functional differences between the splenic resident DCs that were examined in our assay, and in vitro derived DCs generated in the presence of GM-CSF, which are more closely related to inflammatory DCs, and also contain a population of macrophages (367).

Another difference is that their studies utilised the B-ODN 1826, whereas we used 1668. However, these two ODNs have the same structural properties and differ only at 2 base residues, which are highly unlikely to cause major differences in uptake or binding to CD14. Indeed, 1668 and 1826 behaved identically when tested in parallel for binding to peritoneal macrophages and neither required CD14 (Figure 4.7).

159

Lack of evidence for direct binding of CpG ODNs by CD14

Our data dispute the claim that CD14 plays a role in the uptake of CpG ODN. The binding of CpG ODN to endogenous CD14 on peritoneal macrophages is likely the best method of measuring the impact of CD14 in a biologically relevant setting. However, the data presented by Baumann detected minimal differences in the amount of CpG ODN taken up by B6 or CD14-/- macrophages after 30 minutes of incubation, and only a slight decrease in CD14-/- macrophages after 2 hours (332). It was not indicated whether this difference was statistically significant. They did observe a significantly decreased amount of CpG ODN uptake by CD14-/- macrophages after 18 hours of incubation, but this seems to be an unreasonably lengthy amount of time to allow for a molecular interaction to occur, given the half-life of thioated ODNs in serum is only 30-60 min (277, 368, 369). In our bindings assays we measured the binding or uptake of CpG ODNs by peritoneal macrophages within this 30-60 min time frame, and found no role for CD14 in this process.

Baumann et al. provided evidence that CD14 directly binds CpG ODNs with an ELISA assay that detected binding of a soluble CD14-Fc construct to CpG ODN bound to a plate. However, their experimental design did not include an irrelevant protein control for CD14 to establish the specificity of binding of plate-bound ODN, or controls to exclude the possibility of non-specific binding of the CD14 construct to the plate. The addition of LPS was observed to inhibit the binding of CD14, which was interpreted as competition with CpG ODN for overlapping binding sites, although it is possible that LPS binding to CD14 would also inhibit any non-specific binding of CD14. Furthermore, our own experiments with CD14 recombinant constructs indicated that the proteins could display artefactual binding properties, and may not be reliable substitutes for the native protein (Figure 4.5). Thus, it is difficult to draw firm conclusions about the binding of CD14 to CpG ODN by studying recombinant proteins. The binding studies with peritoneal macrophages are more representative of what would occur in vivo, and showed no role for CD14 in our experiments (Figure 4.6, 4.7), and at best only a mild impact of CD14 by Baumann et al.

160

In vivo stimulation by CpG ODN in CD14-deficient mice

Both this study, and that of Baumann et al., examined the production of cytokines in vivo after CpG ODN, but reported conflicting data on the impact of CD14-deficiency. One confounding factor that hinders direct comparison is that our study involved the intravenous injection of CpG ODN as a sole agent, whereas Baumann et al. injected mice intraperitoneally and included 20 mg of D-galactosamine (DGALN) (332). In our hands, intraperitoneal injection of CpG ODN alone was insufficient to induce detectable levels of cytokines in the peritoneal lavage (data not shown). It is possible that DGALN, an inflammatory agent typically used to induce hepatic injury (370), was included to enhance the production of cytokines to detectable levels. However, this introduces a new variable, as it is possible that CD14-deficiency impacts the response to DGALN.

Alternatively, these divergent results may indicate that CD14 plays a role in the response of cells to CpG ODN in the peritoneum specifically, which is not detected by systemic assays looking at the spleen or blood. This is plausible as macrophages are a dominant cell population in the peritoneal cavity (371, 372) and they express high levels of CD14, but not DEC-205. Potentially, CD14 is involved in the uptake of CpG ODN by this particular cell type because they lack DEC-205, whereas a similar role for CD14 is not seen in DCs, which do express DEC-205.

However, this hypothesis is challenged by results that suggest that peritoneal macrophages do not appear to require CD14 for the uptake of CpG ODN. Even the findings of Baumann showed that the uptake of CpG ODNs was largely intact in CD14-/- macrophages after 30 min incubation. Indeed, in conjunction with our data, there appears to be little evidence that CD14 directly interacts with CpG ODN at all, which would be a prerequisite for uptake.

161

CD14 may mediate TLR9 activation, but not CpG ODN uptake

It is possible that CD14 is involved in mediating the response to CpG ODN, despite not binding directly, by modulating the activation of pathways downstream of CpG ODN uptake. Though this has not been specifically tested for TLR9, CD14 was suggested to interact in such a manner with TLR7.

Interaction proteomics and co-immunoprecipitation data presented by Baumann et al. identified CD14 to be an interaction partner of all four endosomal TLRs (TLR3, 7, 8 and 9) (332). Postulating a role for CD14 in the sensing of pathogenic DNA, Baumann et al. infected B6 or CD14-/- peritoneal macrophages with the TLR7- activating viruses VSV or influenza. Though they found no contribution of CD14 in the uptake of these viruses, infected CD14-/- macrophages appeared to have reduced capacity to produce IL-6 and IFNα/β. This suggested that CD14 functions as a cofactor of TLR7 stimulation independent of ligand uptake. It is possible that CD14 acts in a similar capacity as a cofactor for the stimulation of TLR9, rather than as an uptake receptor for its ligand CpG ODN. This could explain the reduced response to CpG ODN seen by Baumann et al. in CD14-/- mice, despite our evidence suggesting that CD14 does not mediate its uptake.

Even if this were to be the case, the present data suggest that any potential role of CD14 as a cofactor of TLR9 appears to only affect the stimulation of peritoneal macrophages, and not DCs or the broader systemic response induced by i.v injection of CpG ODN. We conclude that, while CD14 may be seen to mediate the stimulation of TLR9 by CpG in certain situations or in specific cell types, it appears to have minimal contribution to the overall systemic response to CpG ODN.

162

Chapter 5: Investigating biological ligands of DEC-205

Abstract

For many years, the receptor DEC-205 has been useful for studying the biology of DCs, and in the past decade has become one of the key receptors utilised for harnessing the therapeutic potential of DCs with in vivo targeting strategies. However, the natural binding partners of DEC-205, and how they contribute to the ability of DEC-205 to induce immune responses, have not been well characterised. The finding that DEC-205 binds CpG ODNs to mediate their immunostimulatory effects may be an important clue towards unravelling the physiological function of this receptor. Although CpG ODNs are synthetic and would not be encountered naturally, it is possible that DEC-205 can bind pathogenic DNA that may structurally resemble synthetic phosphorothioated ODN. We investigated whether various naturally occurring DNA samples from pathogenic and non-pathogenic sources could bind DEC-205. Although no binding of any of the tested DNA to DEC-205 could be seen, we detected binding to another receptor, RAGE, recently identified to bind CpG ODNs. We also examined whether DNA binding could be enhanced in the presence of various cofactors. Mouse serum was found to significantly enhance the binding of several DNA samples to RAGE, but it had no effect on DNA binding to DEC-205, which was negligible. An acidic environment has previously been described to promote DEC-205 binding to dead cells, but we were unable to replicate these findings or show a similar effect on binding of CpG ODNs to DEC-205.

163

Introduction

Since its identification in 1986 with the mAb NLDC-145 (373), relatively little progress has been made towards defining the role of DEC-205 in the immune system. Although early studies with mAbs determined DEC-205 to be an endocytic receptor involved in Ag presentation, how DEC-205 may contribute to the induction of immunity or maintenance of homeostasis under normal conditions is not well understood. Particularly as immunotherapeutic strategies utilising DEC-205 to deliver Ags to APCs enter the clinic, it is vitally important to understand the natural biology of this receptor. This would allow better control over the immune outcomes of such therapeutic strategies, and hopefully reduce the chance of unexpected side effects. A better understanding of DEC-205 and its interaction with other cells or molecules could also potentially allow the development of more sophisticated methods of utilising DEC-205 to generate the desired immunological outcome, an example of which is presented in Chapter 3.

DEC-205 is an endocytic CLR

DEC-205, also known as CD205, is a C-type lectin-like receptor consisting of a signal peptide, cysteine-rich domain (CysR), a fibronectin type II domain (FNII), 10 C-type lectin-like domains (CTLDs), a transmembrane domain and a short cytoplasmic tail (225). DEC-205 was initially identified on mouse dendritic cells and cortical thymic endothelial cells (373, 374) but can also be detected at lower levels on the surface of B cells, T cells and granulocytes (375, 376). The same cell types express DEC-205 in human tissues, although additional staining on monocytes and NK cells in human blood has been observed (377). Notably, DEC-205 expression is distinct between different subsets of DCs; CD8+ and CD103+ DCs in the mouse, collectively known as cDC1s, express high levels of DEC-205, while cDC2s and pDCs express significantly lower amounts (71, 72, 378). This distinction is less clear in humans, with high levels of DEC-205 expression observed on both cDC1 and cDC2 human DCs, and a lower level on human pDCs (239). Langerhans cells and monocyte-derived DCs are also known to express high levels of DEC-205 (72, 76).

164

DEC-205 belongs to the mannose receptor family, which are unique amongst CLRs for possessing multiple CTLDs. Other members of this family include the mannose receptor (MR), phospholipase A2 receptor (PLA2R), and Endo180, all of which share the same domain structure consisting of an N-terminal cysteine-rich region, a fibronectin type II domain, 8 CTLDs (or 10 in the case of DEC-205), and a transmembrane region followed by a short cytoplasmic domain (379). Although these receptors have been described to recognise diverse ligands, both of pathogenic and self origin, a common feature is the endocytosis and recycling of the receptor (379).

DEC-205 is no exception, and its endocytic function was first demonstrated when anti-DEC-205 mAbs were found to be rapidly endocytosed within 5 minutes of incubation with DCs (225). However, in contrast to the other MR family members that recycle via early endosomes, uptake via DEC-205 appears to deliver ligands to late endosomes and lysosomes before recycling back to the surface (225, 226). These endosomes were confirmed to be MHC II containing, LAMP-1 positive late endosomes, which promoted efficient presentation of epitopes to primed T cells (226). Interestingly, mutations in the cytoplasmic tail of DEC-205 that impaired its ability to target late endosomes also reduced its capacity to present Ag (226). Thus, it was hypothesized that DEC-205 has a specialised role in delivering Ag to late endosomal compartments for presentation via MHC II. It has subsequently been shown that Ag delivered to DEC-205 is also effectively cross-presented via MHC I (199, 220).

DEC-205 in thymic development

The abundant expression of DEC-205 on cortical thymic endothelial cells (cTECs), combined with its capacity to promote presentation of Ag to T cells, led to speculation that DEC-205 may be involved in the process of T cell development in the thymus. cTECs mediate the positive selection of developing T cells in the thymus by presenting self Ag-MHC complexes (380). DEC-205 is also expressed by the majority of thymic DCs (381, 382), which have been reported to contribute to negative selection of T cells (380, 383). However, a study in which mice were injected with anti-DEC-205 mAb

165

from birth revealed no apparent role for DEC-205 in T cell development or function (384). T cells from anti-DEC-205 injected mice were present in lymphoid organs in the expected proportions and showed no defect in response to stimulation by concanavalin A or phytohaemagglutinin. Additionally, cell suspensions obtained from the lymphoid organs of injected mice appeared fully capable of stimulating primed T cells (384). Nevertheless, the authors did not rule out the possibility that the anti- DEC-205 mAb used may bind to a region of DEC-205 that does not impair its function.

The role of DEC-205 in T cell development was more definitively addressed with the used of DEC-205-deficient (DEC-205-/-) mice (385). In these mice, the thymus appeared to develop normally, with confocal microscopy confirming normal segregation of the cortical and medullary regions. Medullary and cortical thymic endothelial cells appeared in normal proportions and expressed normal levels of MHC I and II, suggesting no defect in their Ag presentation capacity. Thymocytes in various stages of development, including double negative, double positive, and CD8+ or CD4+ single positive thymocytes, were all found in normal numbers in the thymus, which was also reflected in the normal number of peripheral T cells in the spleen. Foxp3+CD4+ Tregs were also observed in normal numbers in the thymus and spleen, and no overt autoimmunity in DEC-205-/- mice was reported. No apparent skewing of the TCR repertoire was seen, altogether indicating that T development and selection proceeds normally in the absence of DEC-205 (385).

DEC-205 binds dead cells

The ability of DEC-205 to internalise and present exogenous Ags, particularly on DCs, may instead point to a potential role as a pattern recognition receptor for DAMPs or PAMPs. Several studies have suggested that DEC-205 can bind apoptotic or necrotic cells, although whether this is a homeostatic mechanism to promote clearance and tolerance against self-Ags, or an immune mechanism to detect damage-associated cell death has not yet been discerned. The work of Small and Kraal showed that a DEC- 205-expressing thymic stromal cell line preferentially bound apoptotic cells in vitro,

166

and that this interaction could be blocked with an anti-DEC-205 mAb (386). In another study, segments of DEC-205 ectodomain were cloned into human IgG fusion proteins and tested for binding to dexamethasone, anti-Fas or heat-induced apoptotic or necrotic cells. Two of these constructs, CTLD3-4 and CTLD9-10, were found to bind dead cells, albeit with low affinity, but not live cells (387).

More recently, Cao et al. also demonstrated that human DEC-205 binds apoptotic or necrotic cells, but only under acidic conditions, and not at pH 7.4 (174). This selective binding at acidic pH correlated with a conformational change in DEC-205. Under acidic conditions, DEC-205 formed a two-ringed structure with the cysteine-rich domain (CysR) interacting with CTLD3 and the fibronectin type II domain (FNII) interacting with CTLD6, and a tail comprising CTLD7-10. At basic pH, CysR dissociated from CTLD3 and FNII from CTLD6, opening both rings and causing DEC-205 to adopt a linear extended conformation. A truncated mutant of DEC-205 comprising only the small ring (CysR, FNII, CTLD1-3) also bound to dead cells under acidic conditions and not at pH7.4, while another mutant comprising just the larger ring (FNII, CTLD1-6) showed no binding in either condition. Another DEC-205 construct containing a mutation at histidine 129 of CysR, proposed to be part of the pH-sensitive interface with CTLD3, was also unable to bind dead cells in either condition. Taken together, these data imply that the pH-sensitive formation of the small ring, mediated by interaction between CysR and CTLD3, is required for dead cell binding (174). Subsequently, the specific ligand detected on dead cells by DEC- 205 at pH 6 was identified as keratin (175). The authors hypothesized that an acidic environment is required for binding to be observed because DEC-205 may facilitate the clearance of dying cells that become acidic during apoptosis or necrosis (174, 175). Another possibility is that DEC-205 recognises keratin after internalisation into acidic endosomal compartments. As several pathogens are known to interact with keratin (388), uptake of keratin by DEC-205 may promote the processing and presentation of associated pathogenic Ags. The impact of the interaction between keratin and DEC- 205 on either of these proposed functions is yet to be examined.

167

The work of Cao et al. highlighted the importance of the structural integrity of DEC- 205 constructs to appropriately examine their function. Removing the tertiary structure of DEC-205 by using smaller ectodomain fragments may have caused the weak binding reported by Shrimpton et al., which could even be an artefact of the IgG fusion constructs. Therefore, the best assessment of DEC-205 binding should be made with DEC-205 in its native format, such as on the surface of a cell. Interestingly, a study investigating the uptake of apoptotic material by DCs in vivo found no role for DEC-205 in this process. Injected apoptotic cells were selectively taken up by CD8+ DCs and not CD8- DCs, but the uptake was not impaired in DEC-205-/- mice (80). Thus, whether the binding of dead cells to DEC-205 occurs in vivo and has a significant physiological role remains to be confirmed.

DEC-205 binds PAMPs

DEC-205 has also been suggested to act as a receptor for PAMPs and facilitate the presentation of Ags derived from pathogens. It has been shown that DEC-205 can bind to plasminogen activator (PLA) expressed by the Gram-negative bacteria Yersinia pestis, promoting its uptake into APCs and subsequently dissemination from the lungs (389). DEC-205 has also been reported to facilitate the entry of HIV into renal tubular cells that lack typical HIV-binding receptors (390, 391). While the uptake of these pathogens via DEC-205 may indeed promote their dissemination, the impact of this interaction on the severity of the infection or the magnitude of the immune response has not been examined. It should be considered that uptake via DEC-205 may not only be the hijacking of an endocytic system by the pathogen, but could also aid the host in inducing an adaptive immune response against the pathogen. The immunological significance of DEC-205 binding such pathogens remains to be determined.

Recently, work from our own group has determined that CpG ODNs are another ligand for DEC-205 (264). A synthetic DNA commonly used as an adjuvant, CpG ODNs are thought to mimic PAMPs found in pathogenic DNA. Importantly, binding to DEC-205 was shown to facilitate the immunostimulatory activity of CpG ODNs.

168

These findings are the first indication that DEC-205 may detect PAMPs derived from foreign pathogens in order to promote the induction of immunity. However, CpG ODNs cannot be the ligand recognised by DEC-205 during an infection, as they are not naturally occurring molecules. Identification of a natural ligand of DEC-205 would be key for confirming its role as a PAMP receptor. Nevertheless, the ability of DEC-205 to bind CpG ODNs may stem from its capacity to bind a structurally related molecule. Our findings showed that DEC-205 binds ODNs with thioate backbones most effectively, a chemical modification that is not known to occur in mammalian DNA, but has been reported in several bacterial genomes (269, 392). Coupled with the fact that the immunostimulatory CpG sequences that led to the design of CpG ODNs were originally identified in bacterial DNA (281), we hypothesize that DEC-205 may have evolved to recognize pathogenic DNA. In this chapter, the potential for various types of DNA to bind DEC-205 will be investigated with the aim of identifying a biological ligand of DEC-205 that may be of immunological consequence. We will further investigate whether certain conditions alter the binding properties of DEC- 205, such as the presence of DNA-binding cofactors, or an acidic environment, as has previously been reported for the binding of dead cells to DEC-205 (174).

169

Results

5.1. The binding of biological DNA to DEC-205 and RAGE

5.1.1. Binding of YOYO-1-labelled DNA to DEC-205-expressing CHO- K1 cells

The identification of the natural ligands of DEC-205 would be invaluable for understanding the biological function of this receptor, which has so far not been well defined. Based on our previous work demonstrating the binding of CpG ODNs to DEC-205, we hypothesize that DEC-205 may be able to recognize a structurally related molecule, such as pathogenic DNA.

To test this hypothesis, we examined the capacity of cell surface-expressed DEC-205 to bind various DNA samples of biological origin. Samples were labelled with the DNA stain YOYO-1, then incubated with mouse DEC-205-expressing CHO-K1 cells, and the YOYO-1 fluorescence of the cells analysed by flow cytometry. To verify the capacity of this assay to detect binding, we firstly examined synthetic ODNs with varying ability to bind DEC-205 as previously defined (264). We labelled biotinylated ODNs with YOYO-1 dye, and determined their binding to DEC-205-expressing CHO cells by YOYO-1 fluorescence. The presence of the biotin tag allowed binding to be confirmed independently of YOYO-1. As expected, 1668 and 2395 showed clear binding to mouse DEC-205, while 2216 also bound but less effectively (Figure 5.1A). By altering the backbone of the partially thioated 2216 to be either fully thioated or fully diester, we could enhance or reduce DEC-205 binding, respectively. Thus, we confirmed that this assay was an effective method of detecting the binding of DNA to DEC-205 and that it confirmed the findings of other binding assays.

We then examined the binding of DNA originating from bacteria, eukaryotic parasites (Plasmodium falciparum, Schistosoma mansoni) or host DNA that is released in association with infection or cell death (mitochondrial DNA, neutrophil extracellular traps (NETs)). As our previous results suggested that DEC-205 binds DNA with a thioated backbone, we also examined the binding of plasmids from bacteria that

170

naturally thioate their own DNA (thioating) or bacteria that do not (non-thioating). We also examined the binding of salmon sperm DNA as an example of non- mammalian but non-pathogenic DNA. For all bacterial samples, and salmon DNA, no binding above background to either untransfected or DEC-205-transfected CHO-K1 cells was observed (Figure 5.1B). Slight binding above background was observed for P. falciparum, S. mansoni, mitochondrial DNA and NETs, but this did not appear to be mediated by DEC-205, as there was little difference between the untransfected and DEC-205-transfected cells. As a positive control, a sample incubated with YOYO-1- labelled fully thioated 2216 was included in each experiment.

Although the positive control 2216 confirmed that the cells and the assay were functional, it did not control for the fact that the YOYO-1 labelling reaction is unique for each DNA sample. Ideally, it would be necessary to confirm that the YOYO-1 labelling reaction for each DNA was successful with a positive control receptor known to bind all DNA.

171

A 2216 1668 bg CHO +ODN DEC +ODN

0 1 2 3 4 0 1 2 3 4 10 10 10 10 10 10 10 10 10 10 2216 (fully diester) 2395

0 1 2 3 4 0 1 2 3 4 10 10 10 10 10 10 10 10 10 10 2216 (fully thioated)

0 1 2 3 4 10 10 10 10 10 YOYO-1

bacterial DNA extracellular (P. fluorescens) bacterial DNA salmon DNA

bg CHO +ODN DEC +ODN

0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 bacterial DNA (E. aerogenes) P. falciparum DNA mitochondrial DNA

0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 bacterial plasmid S. mansoni neutrophil (non-thioating) egg antigen extracellular traps

0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 bacterial plasmid (thioating) 2216 (fully thioated)

0 1 2 3 4 0 1 2 3 4 10 10 10 10 10 10 10 10 10 10 YOYO-1

172

Figure 5.1. Binding of CpG ODNs and DNA to DEC-205-expressing CHO-K1 cells. CHO-K1 cells (CHO) or CHO-K1 cells expressing mouse DEC-205 (DEC) were incubated with (A) 10 μg/ml of biotinylated ODNs or (B) 10 μg/ml of DNA samples labelled with YOYO-1 dye. YOYO-1 fluorescence of cells was determined by flow cytometry. “bg” indicates background fluorescence of CHO-K1 cells in the absence of DNA. The background fluorescence of CHO-K1 cells expressing mouse DEC-205 was identical and not shown for clarity. Representative histograms of two (ODNs, salmon DNA, bacterial plasmids) or one (bacterial DNAs, extracellular DNA, P. falciparum DNA, S. mansoni egg antigen, mitochondrial DNA, NETs) experiments are shown. As data shown was obtained in multiple separate experiments, an example of fully thioated 2216, which was included as a control in every experiment, is shown.

5.1.2. Binding of biotin-labelled DNA to DEC-205- or RAGE- expressing CHO-K1 cells

We saw no specific binding of any of the tested DNA samples to DEC-205, although we also lacked a positive control to confirm that negative staining was not due to inefficient labelling with YOYO-1. Receptor for advanced glycation end-products (RAGE) is a transmembrane receptor that has recently been described to bind DNA and RNA in a sequence- and backbone- independent manner mediated by charge interactions (275). The capacity of RAGE to bind a wide range of different types of nucleic acid molecules made it a good candidate for a positive control in our screening of various types of DNA. We examined the binding of various DNA samples to CHO- K1 cells transiently transfected with RAGE in parallel with CHO-K1 cells stably expressing mouse DEC-205. As the RAGE expression plasmid contained an mCitrine fluorescent tag that would interfere with the detection of YOYO-1 staining, for this assay the DNA samples were instead labelled with biotin, and bound DNA detected with streptavidin PE.

As expected, we observed binding of fully thioated CpG ODN 1668 to both DEC-205- and RAGE-expressing cells (Figure 5.2A). Fully diester 1668 was unable to bind DEC- 205, in accordance with our previous data, but showed marginal binding above

173

background to RAGE. Salmon DNA, mitochondrial DNA and NETs showed no apparent binding to either DEC-205 or RAGE above the level of binding seen on untransfected CHO-K1 cells (Figure 5.2A). On the other hand, mammalian DNA (from the human embryonic kidney cell line HEK293F) (Figure 5.2A), six different bacterial DNA samples and DNA from both Plasmodium falciparum and Plasmodium berghei all showed enhanced binding to RAGE-expressing cells (Figure 5.2B). Extracellular bacterial DNA showed marginal binding to RAGE. However, none of the DNA samples tested displayed any binding to DEC-205 above the level of binding observed on untransfected cells. It should be noted that, for samples with no or minimal binding to both RAGE and DEC-205 (salmon DNA, mitochondrial DNA, extracellular DNA and NETs), it cannot be ruled out that binding may not have been detected due to unsuccessful biotin labelling of the DNA. For salmon DNA and NETs, this possibility was later excluded when binding to RAGE was seen under certain conditions in later experiments.

174

1668 thioated salmon DNA mitochondrial DNA

bg CHO +DNA DEC +DNA RAGE +DNA

0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 neutrophil 1668 diester mammalian DNA extracellular traps

0 1 2 3 4 2 3 4 5 2 3 4 5 10 10 10 10 10 10 10 10 10 10 10 10 10 SA-PE

bacterial DNA bacterial DNA extracellular (P. fluorescens) (Ps.aeruginosa) bacterial DNA

bg CHO +DNA DEC +DNA RAGE +DNA

0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 bacterial DNA bacterial plasmid (E. aerogenes) (non-thioating) P. falciparum DNA

0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 bacterial DNA bacterial plasmid (E. cloacae) (thioating) P. berghei DNA

0 1 2 3 4 0 1 2 3 4 2 3 4 5 10 10 10 10 10 10 10 10 10 10 10 10 10 10 SA-PE

Figure 5.2. Binding of CpG ODNs and DNA to DEC-205- and RAGE-expressing CHO-K1 cells. CHO-K1 cells (CHO) or CHO-K1 cells expressing mouse DEC-205 (DEC) or RAGE were incubated with 6 μg/ml of biotinylated ODNs or DNA samples labelled with biotin. Bound DNA was detected with streptavidin PE and fluorescence analysed by flow cytometry. “bg” indicates the background fluorescence of streptavidin PE binding to CHO-K1 cells in the absence of DNA. The background fluorescence of CHO-K1 cells expressing mouse DEC-205 or RAGE was identical and

175

not shown for clarity. Representative histograms of at least 2 independent experiments are shown, with the exception of mitochondrial DNA, E. cloacae and Ps. aeruginosa DNA, which were performed once.

5.1.3. Binding of biotin-labelled DNA to DEC-205- or RAGE- expressing CHO-K1 cells in the presence of HMGB1, granulin and LL- 37

The homeostatic turnover of cells causes frequent cell death, resulting in the release of nucleic acids into the extracellular space and the blood, a process that may be exacerbated by injury or infection (393). DNA in the serum, which may include pathogenic DNA during an infection, is likely to associate with DNA-binding factors, some of which, such as high mobility group box 1 (HMGB1) (394), LL-37 (348) and granulin (395), have been described to enhance the uptake of DNA by immune cells.

Thus, we considered whether the lack of binding of the various DNA samples to DEC- 205 might be due to a lack of cofactors that may be necessary for binding. To address this question, we compared the binding of biotin-labelled bacterial DNA, mammalian DNA, or fully-thioated 1668 to DEC-205- or RAGE-expressing CHO-K1 cells in media with or without HMGB1, granulin or LL-37.

The addition of HMGB1 did not appear to enhance the binding of bacterial DNA or 1668 to RAGE, but did seem to cause a slight increase in the binding of mammalian DNA to RAGE (Figure 5.3A). HMGB1 also seemed to cause a slight increase in non- specific binding of mammalian DNA to untransfected cells, but no further increase above this level was observed on DEC-205-expressing cells. Granulin did not cause any apparent increase in the binding of any of the DNA to either RAGE or DEC-205 (Figure 5.3B). At low doses of LL-37 (5 μg/ml), we observed a substantial increase in the binding of both bacterial and mammalian DNA to cells expressing RAGE, DEC- 205, or untransfected cells, though again, there was no difference between DEC-205- expressing and untransfected cells (Figure 5.3C). Higher doses of LL-37 (15 μg/ml) caused further enhanced binding, though significantly enhanced non-specific binding

176

to untransfected cells was also seen, which made it difficult to determine whether there was any RAGE-specific enhancement of binding. Interestingly, LL-37 appeared to have the opposite effect on 1668, with less binding observed at higher doses. LL-37 appeared to affect the binding of 1668 to DEC-205 more than to RAGE, with the higher dose (15 μg/ml) resulting in no binding of 1668 to DEC-205 above the background seen on untransfected cells. We noted that cells treated with LL-37 at 15 μg/ml appeared to have higher staining of PI, indicating increased permeability of cell membranes, while the even higher dose of 50 μg/ml caused a significant amount of cell death (data not shown).

177

A bacterial DNA mammalian DNA 1668 thioated

bg CHO +DNA BSA DEC +DNA RAGE +DNA

BSA + HMGB1

2 3 4 5 2 3 4 5 2 3 4 5 10 10 10 10 10 10 10 10 10 10 10 10

B bacterial DNA mammalian DNA 1668 thioated

bg CHO +DNA BSA DEC +DNA RAGE +DNA

BSA + granulin

2 3 4 5 2 3 4 5 2 3 4 5 10 10 10 10 10 10 10 10 10 10 10 10

C bacterial DNA mammalian DNA 1668 thioated

bg CHO +DNA BSA DEC +DNA RAGE +DNA

BSA + LL-37 (5)

BSA + LL-37 (15)

2 3 4 5 2 3 4 5 2 3 4 5 10 10 10 10 10 10 10 10 10 10 10 10 SA-PE

Figure 5.3. Binding of DNA to DEC-205- and RAGE-expressing CHO-K1 cells in the presence of DNA-binding cofactors. CHO-K1 cells (CHO) or CHO-K1 cells expressing mouse DEC-205 (DEC) or RAGE were incubated with 3 μg/ml of bacterial (P. fluorescens) DNA, mammalian (HEK293F) DNA or 1668 labelled with biotin in PBS containing 2%BSA (BSA) with or without (A) HMGB1 (3 μg/ml), (B) granulin (2 μg/ml) or (C) LL-37 (5 or 15 μg/ml as indicated). Bound DNA was detected with streptavidin PE and fluorescence analysed by flow cytometry. “bg” indicates the background fluorescence of streptavidin PE binding to CHO-K1 cells in the absence of DNA. The background fluorescence of CHO-K1 cells expressing mouse DEC-205 or RAGE was identical and not shown for clarity. Representative histograms of 2

178

independent experiments are shown, with the exception of 1668 plus HMGB1, 1668 plus LL-37 (5 & 15 μg/ml), and bacterial and mammalian DNA plus LL-37 (5 μg/ml), which were performed once.

5.1.4. Binding of HMGB1, granulin and LL-37 to DNA

We observed enhanced binding of bacterial and mammalian DNA in the presence of HMGB1 and LL-37, but not granulin. To examine whether these differential effects were due to differences in DNA binding capacity, the ability of the three factors to bind biotin-labelled DNA was compared by ELISA. HMGB1 showed strong binding to 1668, bacterial and mammalian DNA (Figure 5.4). LL-37 also bound all three types of DNA, but showed a preference for binding 1668. By contrast, granulin showed strong binding to 1668, but negligible binding to either bacterial or mammalian DNA. This could explain why granulin had no effect on the binding of bacterial or mammalian DNA in the previous assay. An irrelevant protein, OVA, showed no binding to any of the DNA samples, while DEC-205 bound 1668, but not bacterial or mammalian DNA, reconfirming our earlier results with DEC-205-expressing CHO- K1 cells.

Interestingly, despite the strong binding of all three DNA-binding factors to CpG ODN 1668, none of them appeared to enhance the binding of 1668 to RAGE- or DEC-205-expressing CHO cells, with LL-37 even appearing to cause some inhibition (Figure 5.3). This would suggest that the binding of these factors to 1668 does not contribute to its uptake. Alternatively, the binding of 1668 to RAGE and DEC-205 may be efficient enough that the addition of binding factors does not cause a detectable increase in binding.

179

3.0 DNA (µg/ml)

2.5 1668 thioated (3) 1668 thioated (1) bacterial (3) 2.0 bacterial (1) mammalian (3) 1.5 mammalian (1) O.D.

1.0

0.5

0.0 HMGB1 granulin LL-37 OVA DEC-205

Figure 5.4. Binding of HMGB1, granulin and LL-37 to DNA. ELISA plates coated with HMGB1 (1 μg/ml), granulin (5 μg/ml), LL-37 (5 μg/ml), OVA (10 μg/ml) or mouse DEC-205 (5 μg/ml) were incubated with biotin-labelled ODN or DNA at the indicated concentrations. Plate were then washed and incubated with streptavidin HRP and visualised with ABTS substrate. Shown is the mean optical density measured at 405-490nm of duplicate samples, minus the background optical density of samples containing no DNA. Representative data of 3 experiments are shown.

5.1.5. Binding of biotin-labelled DNA to DEC-205- or RAGE- expressing CHO-K1 cells in the presence of serum

It is possible that serum contains factors other than HMGB1, LL-37 or granulin that could enhance the binding of DNA to DEC-205 or RAGE. We examined the capacity of whole mouse serum to promote DNA binding by comparing the binding of biotin- labelled DNA to DEC-205- or RAGE-expressing CHO-K1 cells in either normal media (PBS-2%BSA) or PBS containing 10% mouse serum.

We observed similar trends for the binding of DNA to DEC-205- or RAGE-expressing cells in normal media as in previous experiments (Figure 5.5). The addition of mouse serum did not appear to significantly alter the capacity of either RAGE or DEC-205 to

180

bind fully thioated 1668. None of the DNA samples were found to bind DEC-205 in normal media, as seen previously, and this was unchanged by the presence of serum. However, intriguingly, the binding of mammalian DNA as well as DNA of pathogenic origin, including bacterial and P. berghei DNA, to RAGE was greatly enhanced in the presence of serum (Figure 5.5). Salmon DNA and NETs showed minimal binding to RAGE in normal media, but binding was noticeably enhanced in the presence of serum. The binding of extracellular DNA to RAGE remained minimal in either the presence or absence of serum, although, again, we could not rule out that this may be due to unsuccessful biotinylation of this particular DNA. Serum also appeared to slightly enhance the non-specific binding of bacterial plasmids, E. aerogenes and P. berghei DNA to untransfected cells.

181

1668 thioated salmon DNA mammalian DNA

bg CHO +DNA BSA DEC +DNA RAGE +DNA

serum

2 3 4 5 2 3 4 5 2 3 4 5 10 10 10 10 10 10 10 10 10 10 10 10 bacterial DNA bacterial plasmid neutrophil (P. fluorescens) (non-thioating) extracellular traps

BSA

serum

2 3 4 5 2 3 4 5 2 3 4 5 10 10 10 10 10 10 10 10 10 10 10 10 bacterial DNA bacterial plasmid extracellular (E. aerogenes) (thioating) bacterial DNA

BSA

serum

2 3 4 5 2 3 4 5 0 1 2 3 4 10 10 10 10 10 10 10 10 10 10 10 10 10

P. berghei DNA

BSA

serum

2 3 4 5 10 10 10 10 SA-PE

Figure 5.5. Binding of DNA to DEC-205- and RAGE-expressing CHO-K1 cells in the presence of serum. CHO-K1 cells (CHO) or CHO-K1 cells expressing mouse DEC-205 (DEC) or RAGE were incubated with 3 μg/ml of DNA samples labelled with biotin in PBS containing either 2%BSA (BSA) or 10% mouse serum (serum). Bound DNA was detected with streptavidin PE and fluorescence analysed by flow cytometry. “bg” indicates the background fluorescence of streptavidin PE binding to CHO-K1 cells in the absence of DNA. The background fluorescence of CHO-K1 cells expressing mouse DEC-205 or RAGE was identical and not shown for clarity.

182

Representative histograms of at least 2 independent experiments are shown, with the exception of extracellular DNA, which was performed once.

5.1.6. Binding of biotin-labelled DNA to DEC-205- or RAGE- expressing CHO-K1 cells in the presence of serum and DNA-binding cofactors

Both HMGB1 and LL-37 appeared to enhance the binding of DNA to RAGE, and thus could potentially be the factors present in serum that enhance the binding of DNA in serum-containing media. At the concentrations tested, neither HMGB1 nor LL-37 appeared to enhance DNA binding to the same magnitude induced by the addition of serum. This could indicate that either the serum contained higher concentrations of HMGB1 and/or LL-37 than we tested, or that there are other factor(s) present in serum that enhance DNA binding to RAGE. Should the latter scenario be the case, it may be possible to further augment the binding of DNA by addition of HMGB1 or LL-37 to serum-containing media. However, in the presence of serum, none of the three factors, HMGB1, granulin or LL-37, caused appreciable increases in the binding of DNA to RAGE (Figure 5.6). Only LL-37 caused an increase in non-specific binding to untransfected or DEC-205-expressing cells, particularly at the higher concentration (15 μg/ml). Interestingly, the level of non-specific binding caused by LL-37 in the presence of serum was substantially reduced in comparison to that caused by LL-37 in media without serum (compare 3.6C to 3.3C). This suggests that serum inhibits the activity of LL-37 by acting either upon its ability to bind DNA, or its capacity to promote cellular uptake.

183

A bacterial DNA mammalian DNA

bg CHO +DNA serum DEC +DNA RAGE +DNA

serum + HMGB1

2 3 4 5 2 3 4 5 10 10 10 10 10 10 10 10

B bacterial DNA mammalian DNA bg CHO +DNA serum DEC +DNA RAGE +DNA

serum + granulin

2 3 4 5 2 3 4 5 10 10 10 10 10 10 10 10

C bacterial DNA mammalian DNA

bg CHO +DNA serum DEC +DNA RAGE +DNA

serum + LL-37 (5)

serum + LL-37 (15)

2 3 4 5 2 3 4 5 10 10 10 10 10 10 10 10 SA-PE

Figure 5.6. Binding of DNA to DEC-205- and RAGE-expressing CHO-K1 cells in the presence of serum and DNA-binding cofactors. CHO-K1 cells (CHO) or CHO- K1 cells expressing mouse DEC-205 (DEC) or RAGE were incubated with 3 μg/ml of bacterial (P. fluorescens) or mammalian (HEK293F) DNA labelled with biotin in PBS containing 10% mouse serum (serum) plus HMGB1 (3 μg/ml), granulin (2 μg/ml) or LL-37 (5 or 15 μg/ml as indicated). Bound DNA was detected with streptavidin PE and fluorescence analysed by flow cytometry. “bg” indicates the background fluorescence of streptavidin PE binding to CHO-K1 cells in the absence of DNA. The background fluorescence of CHO-K1 cells expressing mouse DEC-205 or RAGE was identical and not shown for clarity. Representative histograms of 2 independent experiments are shown, with the exception of LL-37 at 5 μg/ml, which was performed once.

184

5.1.7. Mouse serum contains a factor that enhances the binding of DNA to RAGE

To aid in potentially identifying the factors present in serum that enhance DNA binding, we investigated their biophysical properties. As serum appeared to have the greatest effect on the binding of bacterial DNA to RAGE, we used the binding of E. aerogenes to RAGE-expressing cells in media containing serum as a benchmark for full potency of the putative binding factors. We then investigated whether this enhanced binding would be maintained in media containing serum subjected to certain conditions.

Firstly, we sought to determine the size of the binding factors by fractionating the serum with a 3 kDa cut-off centrifugal filter. The ability of each fraction to promote E. aerogenes DNA binding to RAGE-expressing cells was compared with whole serum. It appeared that both greater than and less than 3 kDa fractions were able to promote DNA binding to RAGE, to a degree comparable with whole serum (Figure 5.7A, untreated). This would suggest that multiple factors, both larger and smaller than 3 kDa, are contributing to the effect. Alternatively, there may be only one type of DNA binding factor that is smaller than 3 kDa, but not all pass through the 3 kDa filter, potentially due to inclusion in larger complexes.

Next, we determined whether the binding factors present in serum could be inactivated by heat. Media containing whole serum, or the greater than or less than 3 kDa fractions, was heated to 90ºC for 5 min, then allowed to cool to 4ºC before being used to examine the binding of bacterial DNA to RAGE-expressing cells. Heat inactivation did not abolish the ability of serum, whole or either fraction, to promote DNA binding to RAGE (Figure 5.7A).

We then investigated whether the activity of the binding factors in serum could be disrupted by digestion with trypsin, or treatment with ethylenediaminetetraacetic acid (EDTA), an agent that chelates metal ions such as Mg2+ and Ca2+, which interferes

185

with the structure and function of many proteins. Our initial results indicated that serum treated with trypsin had reduced ability to promote the binding of DNA to RAGE (Figure 5.7B). However, it was possible that the trypsin could also cause reduced binding by degrading the receptor RAGE on the surface of the CHO-K1 cells. Thus, the trypsin-treated serum was heat-inactivated to destroy the activity of trypsin prior to incubation with cells and DNA. After heat-inactivation, it became apparent that treatment with trypsin was not able to inhibit the binding of DNA to RAGE in the presence of whole serum, or in serum from either the greater than or less than 3 kDa fraction (Figure 5.7B). In fact, the binding of DNA in media containing the greater than 3 kDa fraction appeared to be slightly increased after trypsin treatment. Unexpectedly, we even saw enhanced binding of DNA to RAGE in normal PBS- 2%BSA media with heat-inactivated trypsin in the absence of serum. This would suggest that the trypsin reagent itself contained factors that enhance DNA binding, which confounds our interpretation of whether trypsin digestion affects the DNA binding factors present in serum. Treatment of whole serum with EDTA did not appear to affect its ability to promote DNA binding (Figure 5.7C).

Our results show that the factors in serum promoting the binding of DNA to RAGE are resistant to heat and smaller than 3 kDa, and thus are highly unlikely to be a protein or a molecule that requires tertiary structure to function. It is also possible that multiple factors exist, some of which are larger than 3 kDa, and are also resistant to heat. A factor that augments the binding of DNA also appears to be present in the trypsin reagent, that is itself resistant to heat and trypsin inactivation, although it is unknown whether this could be the same factor found in serum.

186

A whole serum serum >3kDa serum <3kDa bg untreated CHO +DNA RAGE +DNA

+ heat

2 3 4 5 2 3 4 5 2 3 4 5 10 10 10 10 10 10 10 10 10 10 10 10

B whole serum serum >3kDa serum <3kDa BSA

untreated

+ trypsin

+ trypsin + heat

2 3 4 5 2 3 4 5 2 3 4 5 2 3 4 5 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10

C whole serum

untreated

+ EDTA

2 3 4 5 10 10 10 10 SA-PE

Figure 5.7. Binding of DNA to RAGE-expressing CHO-K1 cells in the presence of serum fractionated by size and treated with trypsin, EDTA or heat. CHO-K1 cells (CHO) or CHO-K1 cells expressing RAGE were incubated with 3 μg/ml of bacterial (E. aerogenes) DNA labelled with biotin in PBS containing 10% mouse serum (whole serum), or 10% serum from greater than (>3 kDa) or less than (<3 kDa) 3 kDa fractions separated by centrifugal filtration. Prior to staining with DNA, serum- containing media was left untreated, or heat inactivated at 90ºC for 5 min, or treated with trypsin (250 μg/ml overnight at 37ºC with or without subsequent heat inactivation at 90ºC for 5 min) or EDTA (0.01M added directly to staining media). Heated media was allowed to cool to 4ºC before proceeding with staining. Bound DNA was detected with streptavidin PE and fluorescence analysed by flow cytometry. “bg” indicates the background fluorescence of streptavidin PE binding to CHO-K1 cells in the absence of DNA. The background fluorescence of CHO-K1 cells

187

expressing RAGE was identical and not shown for clarity. Representative histograms of 3 independent experiments are shown.

5.2. Binding properties of DEC-205 under acidic conditions

5.2.1. Binding of ODNs to soluble DEC-205 in acidic conditions

Our results have so far revealed no binding of biological DNA samples to DEC-205. We considered the possibility that DEC-205 requires a cofactor to promote the binding of DNA, but the addition of known DNA-binding cofactors HMGB1, granulin and LL-37, or of mouse serum, did not appear to have this effect. Recently, a study by Cao et al. provided compelling evidence that DEC-205 binds dead cells, and implied that this binding may have been missed in previous studies because it does not occur at the physiological pH 7 at which most assays are conducted, but rather only at pH 6 (174). This was due to a conformational change in DEC-205 that requires an acidic environment. Thus, we sought to investigate whether pH is a factor that influences the binding of DEC-205 to other ligands, such as CpG ODNs, and if an acidic pH is required to observe binding of biological DNA to DEC-205. We firstly sought to confirm whether pH affects the binding of CpG ODNs that have previously been shown to have strong or weak binding to DEC-205 at pH 7.

We examined the binding of various CpG ODNs to DEC-205 by ELISA conducted under different pH conditions. B-ODN 1668 and A-ODN 2216 were found to bind efficiently only at pH 7 (and only at high concentrations for 2216) (Figure 5.8). At pH 6 or pH 5, binding appeared to be almost completely abolished. 1668 modified to have a completely diester backbone, or P-ODN 21798, both of which are known to have weak DEC-205 binding under normal conditions, showed no greater binding at pH 6 or pH 5.

188

1668 1668 diester 0.8 0.8 pH 7 pH 7 0.6 pH 6 0.6 pH 6 pH 5 pH 5

0.4 0.4 O.D. O.D.

0.2 0.2

0.0 0.0 101 102 103 104 102 103 104 105 ODN (ng/ml) ODN (ng/ml)

2216 21798 2.0 0.8 pH 7 pH 7 1.5 pH 6 0.6 pH 6 pH 5 pH 5

1.0 0.4 O.D. O.D.

0.5 0.2

0.0 0.0 102 103 104 105 102 103 104 105 ODN (ng/ml) ODN (ng/ml)

Figure 5.8. Binding of CpG ODNs to DEC-205 is not enhanced by acidic pH. ELISA plates were coated with soluble mouse DEC-205 (5 μg/ml), captured by anti- DEC-205 mAb (5 μg/ml) in pH 7 PBS. Plates were incubated with biotinylated ODN at the indicated concentrations, diluted in buffer adjusted to pH 7, 6, or 5 as indicated. Plates were then washed and incubated with streptavidin HRP and visualised with ABTS substrate. Both washing steps and streptavidin HRP incubation were performed in buffers adjusted to the indicated pH. Shown is the mean optical density measured at 405-490nm of duplicate samples. Dotted line represents background seen in the absence of ODN. Representative data of 2 experiments are shown.

5.2.2. Binding of dead cells to soluble DEC-205, or Clec9A, in acidic conditions

It is possible that the requirements for binding differ between human DEC-205, which was used in the study by Cao et al., and mouse DEC-205 that was used in our ELISA. To address this issue, we compared mouse and human DEC-205 in parallel for their

189

ability to bind dead cells at pH 6, using the same experimental design described by Cao et al.

Freeze/thawed CHO-K1 or HEK293F cells were incubated with recombinant soluble mouse DEC-205, human DEC-205 or mouse Clec9A. Each protein possessed a FLAG tag and thus binding could be detected by staining with an anti-FLAG mAb. Clec9A is known to bind dead cells (169, 170) and accordingly we detected strong binding of Clec9A to dead cells at both pH 6 and pH 7, but not to live cells. However, we saw no binding of mouse or human DEC-205 to dead cells above background levels, at either pH 6 or pH 7 (Figure 5.9).

dead CHO dead 293F bg pH 7 mDEC-205 hDEC-205 mClec9A 0 1 2 3 4 0 1 2 3 4 10 10 10 10 10 10 10 10 10 10

pH 6

0 1 2 3 4 0 1 2 3 4 10 10 10 10 10 10 10 10 10 10

live CHO live 293F

pH 6

0 1 2 3 4 0 1 2 3 4 10 10 10 10 10 10 10 10 10 10 a-FLAG FITC

Figure 5.9. Clec9A, but not mouse or human DEC-205, binds to dead cells at pH 6 and pH 7. CHO-K1 or HEK293F cells were freeze/thawed (dead) or kept at 4ºC (live) prior to incubation with soluble FLAG-tagged mouse Clec9A, mouse DEC-205 or human DEC-205 for 20 min at room temperate in media at pH 6 or pH 7. Bound proteins were detected with anti-FLAG FITC and fluorescence analysed by flow cytometry. “bg” indicates background fluorescence of cells incubated with no protein. Representative histograms of 2 independent experiments are shown.

190

5.2.3. Binding of dead cells to cell surface-expressed DEC-205, or Clec9A, in acidic conditions

The lack of binding of human DEC-205 to dead cells at pH 6 appears to directly contradict the findings of Cao et al. The only apparent difference between the experiment presented here and the published study is in the different recombinant human DEC-205 constructs used. Our human DEC-205 protein is produced with a C- terminal FLAG tag, whereas the human DEC-205 used by Cao et al. had a C-terminal His tag and was a GFP fusion protein, with which they could detect binding without additional staining. Any of these additional tags and modifications could potentially alter the natural structure of DEC-205 and cause binding artefacts. To circumvent this issue, we measured the binding of dead cells to cell-bound DEC-205 expressed by mammalian cells, which theoretically more closely resemble the native form of DEC- 205 than do any of the soluble constructs.

To measure the binding of dead cells, HEK293F cells were labelled with either PKH26 or CFSE dyes, then killed by freeze/thawing, or freeze/thawing followed by sonication to create cell fragments which may help to expose intracellular components that bind DEC-205. These dead cell preparations were then incubated with CHO-K1 cells expressing mouse Clec9A, mouse DEC-205, or human DEC-205. None of the transfectant cells were associated with a greater amount of dye than the untransfected parental CHO-K1 cells, indicating that none of the proteins enhanced binding to dead cells or dead cell fragments in this assay (Figure 5.10). However, given that even the positive control Clec9A was not seen to promote binding of dead cells, this assay appears to have a sensitivity issue and requires optimisation to adequately detect binding.

191

CHO pH 7 mDEC-CHO hDEC-CHO Clec9A-CHO 2 3 4 5 10 10 10 10

pH 6

2 3 4 5 10 10 10 10 PKH26 B + f/t 293F + f/t & sonicated 293F no 293F CHO unlabelled mDEC-CHO hDEC-CHO Clec9A-CHO 0 1 2 3 4 10 10 10 10 10 CFSE labelled

0 1 2 3 4 0 1 2 3 4 10 10 10 10 10 10 10 10 10 10 CFSE

Figure 5.10. Labelled dead cells do not bind Clec9A-, mouse DEC-205-, or human DEC-205-expressing CHO-K1 cells. (A) PKH26-labelled HEK293F cells were freeze/thawed and sonicated, then incubated with CHO-K1 cells, either untransfected (CHO) or expressing mouse Clec9A (Clec9A-CHO), mouse DEC-205 (mDEC-CHO) or human DEC-205 (hDEC-CHO), for 20 min at RT in media at pH 6 or pH 7. Cells were washed and analysed by flow cytometry for PKH26 fluorescence. One experiment was performed. (B) CFSE-labelled or unlabelled HEK293F cells were freeze/thawed (f/t) or freeze/thawed and sonicated (f/t & sonicated), then incubated with CHO-K1 cells as in (A), in media at pH 6. Cells were washed and analysed by flow cytometry for CFSE fluorescence. One experiment was performed.

192

5.2.4. Anti-DEC-205 mAbs can enhance the binding of ODNs to human DEC-205

While these investigations were not able to adequately confirm that pH affects the binding of DEC-205 to ligands such as dead cells or CpG ODNs, in other experiments we saw some evidence to support the proposition put forth by Cao et al. that conformational changes in DEC-205 can influence its binding capabilities. While testing the binding of several mAbs that had been raised against different domains of DEC-205, we came across an unusual finding. To our surprise, pre-incubation of human DEC-205 expressing cells with 3 of these mAbs caused enhanced binding of CpG ODNs (Figure 5.11B). This could indicate that the binding of these mAbs alters the conformation of DEC-205, allowing better binding to CpG ODNs. We did not see the same effect for mouse DEC-205. While these 3 mAbs were also able to bind mouse DEC-205 expressing cells, they did not affect their ability to bind CpG ODNs (Figure 5.11A).

Notably, not all mAbs enhanced binding of ODNs to human DEC-205 to the same degree. MMRI7 did not appear to affect ODN binding, and HD24 caused only a slight increase in ODN binding, whereas HD71 and HD83 greatly enhanced ODN binding (Figure 5.11B). This did not appear to correlate with the region of DEC-205 recognised by the mAbs, as both MMRI7 and HD83 recognise CTLD1-2, despite their disparate effects. Similarly, HD71 recognises CTLD3-4, but enhances ODN binding comparably to HD83. HD24 recognises the cysteine-rich fibronectin neck region, distinct from the other mAbs.

193

CHO +ODN A mouse DEC-205 DEC +ODN DEC +mAb +ODN HD24 HD71 HD83 MMRI7

1668 200ng/ml

1668 40ng/ml

0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10

2006 200ng/ml

2006 40ng/ml

0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 ODN binding

CHO +ODN B human DEC-205 DEC +ODN DEC +mAb +ODN HD24 HD71 HD83 MMRI7

1668 200ng/ml

1668 40ng/ml

0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10

2006 200ng/ml

2006 40ng/ml

0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 ODN binding

Figure 5.11. Certain anti-DEC-205 mAbs enhance the binding of CpG ODNs to DEC-205. CHO-K1 cells (CHO) or CHO-K1 cells expressing (A) mouse or (B) human DEC-205 (DEC) were incubated with 10 μg/ml of the anti-DEC-205 mAbs HD24, HD71, HD83 or MMRI7 for 20 min, following by addition of biotinylated ODNs at the indicated concentrations for a further 30 min (+mAb +ODN). In parallel, samples were incubated with ODN without preincubation with mAbs (+ODN). Bound ODN was detected with streptavidin PE and fluorescence analysed by flow cytometry. Representative histograms of 2-3 experiments are shown.

194

Discussion

RAGE binds various DNA of biological origin

The possibility that DEC-205 can act as a PRR to detect nucleic acids was explored by examining its capacity to bind various DNA samples that could potentially act as PAMPs or DAMPs, such DNA from bacterial and parasitic pathogens and NETs. However, none of the tested samples were seen to bind DEC-205, even in the presence of DNA-binding cofactors such as HMGB1, LL-37, granulin, or whole mouse serum. As we were unable to quantitate the efficiency of biotin labelling of the DNA, it is possible that inefficient labelling could have caused a lack of detectable binding to DEC-205. However, all samples, with the exception of mitochondrial DNA and extracellular DNA, were observed to bind to RAGE, confirming that they were detectable in our assay. It could also be considered that the concentration of DNA that yields positive binding to RAGE may be too low to observe binding to DEC-205. To more comprehensively determine whether these samples bind to DEC-205, it could be beneficial to test binding at a range of DNA concentrations.

While we did not detect binding to DEC-205, several DNA samples, including bacterial DNA from four different species (P. fluorescens, E. aerogenes, E. cloacae, Ps. aeruginosa), bacterial plasmid DNA, DNA from P. falciparum and P. berghei, and mammalian DNA isolated from HEK293F cells showed specific binding to another receptor, RAGE. Thus, while our investigations could provide no further insight into the potential physiological role of DEC-205, they may have interesting implications for our understanding of RAGE biology.

RAGE is a cell surface receptor of the immunoglobulin superfamily that is expressed at low levels in most tissues, and at particularly high levels on lung epithelial cells. RAGE is also expressed by various immune cells, including lymphocytes, neutrophils, macrophages and dendritic cells (396, 397). Although full-length RAGE is a membrane-bound receptor, various isoforms exist, including a soluble form that can act as an antagonist (396, 398).

195

RAGE has been described to bind a diverse range of ligands. Most are endogenous proteins considered to be DAMPs associated with inflammation, including advanced glycation end-products (AGEs), for which the receptor is named, amyloid-β, HMGB1, heat shock protein (HSP)70, and members of the S100/calgranulin family (396, 398). Ligation of RAGE activates multiple signalling pathways to promote inflammatory responses, including the NF-κB and AP-1 pathways, and may cooperate with TLR signalling (396, 399, 400). RAGE expression is upregulated at sites of inflammation, and this has been associated with a range of pathological conditions, such as atherosclerosis, arthritis, Alzheimer’s disease, sepsis and cancer (397, 398).

RAGE has also been reported to bind the β2 integrin Mac-1 to facilitate leukocyte recruitment and adhesion (401-403). This suggests that RAGE may also play a role in the initiation of adaptive immune responses. Indeed, it has been shown that RAGE- deficient T cells have an impaired capacity to proliferate in response to both nominal and autoantigens, and a decreased propensity to differentiate into Th1 effectors (404). This was corroborated in another study in which RAGE-deficient mice, or wild-type mice treated with a RAGE antagonist, showed reduced rejection of allotype grafts. This was attributed to a deficiency in T cell activation and Th1 differentiation (405).

More recently, RAGE has also been shown to bind PAMPs such as LPS and CpG ODNs. RAGE was shown to directly bind LPS by surface plasmon resonance, and promoted the LPS-induced production of TNFα from peritoneal macrophages (406). The capacity for RAGE to mediate the activity of LPS in vivo was demonstrated with a model of LPS-mediated septic shock. Mice deficient in RAGE exhibited attenuated symptoms of septic shock, including reduced serum levels of TNFα, IL-6 and HMGB1, resulting in significantly longer survival (406). RAGE has also been shown to directly bind CpG ODNs, and similarly mediates their immunostimulatory effects (275, 407). pDC or bone marrow cells derived from RAGE-deficient mice showed significantly impaired ability to produce IFNα upon stimulation with CpG ODN (407). Additionally, CpG ODN-induced inflammation in the lung was found to be significantly reduced in RAGE-deficient mice (275).

196

Sirois et al. determined that RAGE in fact binds both DNA and RNA via charge interactions with the sugar-phosphate backbone in a sequence-independent manner (275). RAGE was found to bind to a variety of nucleic acids, including different classes of CpG ODNs, base-free deoxyribose backbones with either phosphodiester or phosphorothioate linkages, and both double and single-stranded ODNs. The capacity of RAGE to bind such a variety of ODNs prompted our interest in testing its binding to various DNA samples of biological origin. Our data provide the first evidence that, in addition to synthetic nucleic acids, RAGE is able to bind natural DNA of pathogenic origin. This finding adds to a growing body of evidence to suggest that RAGE may have an underappreciated role as a detector of foreign pathogens, in addition to its well-known role as a receptor for endogenous DAMPs.

Serum contains a factor that promotes DNA binding to RAGE

Our investigations have revealed that mouse serum contains a factor that potently enhances the binding of DNA to RAGE. Our preliminary efforts to characterise this putative DNA-binding factor indicated that it is smaller than 3 kDa and resistant to heat-inactivation. A peptide smaller than approximately 27 amino acids that does not rely on tertiary structure for its function would fit this profile. This eliminates HMGB1, granulin and LL-37, as HMGB1 and granulin are significantly larger proteins, and LL-37, while 37 amino acids in length, has been shown to act like a ~30 kDa protein due to oligomerisation (408). Alternatively, the factor could be a different type of macromolecule, such as a lipid. Enhanced binding of DNA to RAGE could also be induced by the serum fraction that was larger than 3 kDa, implying that either the small factor can exist in larger complexes, or that multiple factors of different sizes are contributing to the effect. Identification of such serum factors that specifically promote the binding of DNA to RAGE would contribute to a greater understanding of how RAGE can act as a PRR and stimulate inflammation and immunity, and may even serve as therapeutic targets to reduce the inflammatory effects of RAGE. Further detailed investigations will be required to properly characterise and identify these serum factors.

197

The effect of pH on binding of ligands to DEC-205

While the panel of DNA tested in this study was by no means exhaustive and does not exclude the possibility that DEC-205 is able to bind a yet-to-be-identified nucleic acid ligand, it is also possible that DEC-205 has specific external requirements for binding to occur. The potential for soluble DNA-binding cofactors to promote DNA binding to DEC-205 was examined, but the same binding factors that enhanced DNA binding to RAGE were found to cause no improvement in binding to DEC-205. A previous study by Cao et al. has suggested that an acidic environment is necessary for DEC-205 to adopt a conformation that permits the binding of dead cells (174). We examined whether a similar requirement exists for the binding of nucleic acids to DEC-205, which may explain the lack of binding seen in our assays conducted at pH 7. An initial experiment to examine the binding of CpG ODNs to DEC-205 by ELISA under various pH conditions found that increasingly acidic pH actually caused significantly reduced binding. However, it cannot be ruled out the inefficient binding seen at lower pH is due to disruption of other components of the ELISA, such as the binding of the detecting reagent (streptavidin) or the activity of the HRP enzyme.

We then sought to confirm the efficacy of the assay described by Cao et al. for detecting the binding of dead cells at pH 6. Following an identical experimental protocol, we saw strong binding of Clec9A to dead cells, but surprisingly, no binding of human or mouse DEC-205. This seemed to directly contradict the findings of Cao et al. (174) The only apparent difference between our experiment and theirs is the different recombinant DEC-205 constructs used – the protein of Cao et al. was fused with a C-terminal His tag and GFP to detect binding without additional staining, whereas the protein used in our study had a C-terminal FLAG tag and binding was detected with an anti-FLAG antibody. It is possible that these various modifications to the protein may have altered its binding properties, so we attempted to examine the binding of DEC-205 in a state more akin to its native form by expression on CHO-K1 cells. Unfortunately, we were unable to devise an assay capable of detecting the interaction of transfectant CHO-K1 cells with dead cells, as even our positive control

198

Clec9A-transfected cells showed no more binding of dead cells than untransfected controls. Thus, we were unable to adequately confirm whether pH affects the binding properties of DEC-205 expressed on the cell surface.

Nonetheless, there is probably some merit to the hypothesis that the tertiary structure of DEC-205, which Cao et al. determined to be pH sensitive (174), is important for binding. Amongst a panel of mAbs raised against different domains of DEC-205, several appeared to enhance the binding of CpG ODNs to CHO-K1 cells expressing human DEC-205. This may be due to the binding of certain mAbs altering the conformational structure of DEC-205 to promote better binding of CpG ODNs. Further investigations will be required to confirm whether these mAbs do affect the conformation of DEC-205, which would provide the first evidence that the conformational structure of DEC-205 influences its capacity to bind other ligands besides dead cells.

199

Chapter 6: Targeting DCs with heterologous prime-boost

Abstract

The ability to target Ags of interest directly to DCs in vivo via DC-specific receptors has emerged as a promising method to generate potent Ag-specific immune responses. Of particular interest is the capacity for DC-targeted vaccines to generate strong T cell responses, including CTL, which is difficult to achieve with conventional vaccines. This is also the reason for a focus on targeting CD8+ and CD103+ DCs, collectively referred to as cDC1s, which are specialised in cross-presentation of Ag for the priming of CD8+ T cells. Clec9A and XCR1 are the only two known receptors almost exclusively expressed by cDC1s, and both have been used to target Ag for the induction of CTL responses with great success. We directly compared the efficacy of Clec9A-targeting (via the anti-Clec9A mAb 10B4) with XCR1-targeting (via its natural ligand XCL1) for inducing T cell responses, and found Clec9A targeting to be the superior strategy. However, our preliminary data suggested that the potency of Clec9A targeting might be a double-edged sword, preventing effective boosting upon a second administration of the same construct. Thus, we investigated the mechanisms that could be responsible for inhibiting the boost and trialled various heterologous prime-boost strategies to overcome this problem. We found that combining different forms of Clec9A-targeting constructs was insufficient to prevent the cross-reactivity of the primary Ab response from binding and neutralising the boosting construct. Only by a combination of priming with XCR1-targeting and boosting with Clec9A- targeting could effective secondary responses be generated.

201

Introduction

Targeting Ags to cDC1s to promote CTL responses

The unique ability of DCs to initiate adaptive immune responses has led to the development of various strategies utilising DCs to enhance immune responses against Ags of interest. In particular, the in vivo targeting of Ags to DC-specific receptors has been demonstrated to be a potent method of generating stronger Ag-specific T cell and B cell responses (236, 409). The capacity of DC-targeting techniques to induce T cell responses, especially CTL responses, is of particular interest for the development of more effective vaccines, as conventional vaccines typically fail to induce potent CTL immunity (410). As cDC1s are known to be the key initiators of CTL immunity due to their specialised capacity for cross-presentation (79), strategies to target Ags to this subset have been actively pursued.

One of the earliest methods of delivering Ag to DCs in vivo involved targeting to DEC-205, which is highly expressed by cDC1s. This has been demonstrated to be an effective method of generating both CD8+ and CD4+ T cell responses (199, 220). However, the broader expression of DEC-205 on other cells, such as T cells, B cells and monocytes (375, 377), may potentially have a negative impact on its efficacy by reducing the amount of Ag delivered to DCs. Thus, receptors more specifically expressed by cDC1s were sought as potential targets for Ag delivery.

Clec9A was identified to have highly restricted expression on cDC1s, with no expression detected on any other cells except for a small amount on pDCs (120-122). Targeting Ag to Clec9A was found to also be an effective method of inducing strong CD8+ and CD4+ T cell responses. Side-by-side comparisons determined Clec9A- targeting to be comparable with DEC-205-targeting for promoting the induction of Ag-specific CTL responses (215, 234). However, Clec9A-targeting was found to promote greater Tfh differentiation and induction of Ab responses than DEC-205- targeting (215). Notably, it has been shown that different anti-Clec9A targeting mAbs do not necessarily produce the same outcomes. The anti-Clec9A mAb developed in

202

our lab, 10B4, and the anti-Clec9A mAb utilised by Caetano Reis e Sousa’s group, 7H11, were both shown, in respective studies, to induce effective CTL responses when administered with adjuvant (121, 215). However, the administration of 7H11 in the absence of adjuvant was found to lead to immune tolerance, whereas immunisation with 10B4 in the absence of adjuvant instead promoted strong Ab responses with no overt induction of tolerance (120, 221). As discussed in greater detail below, the ability for 10B4 to promote immunity in the absence of adjuvant was attributed to the presence of immunogenic epitopes in its rat IgG2a backbone that were absent from the rat IgG1 backbone of 7H11 (222). The enhanced immunogenic features of 10B4 led us to select this mAb as representative of the most potent method of delivering Ag via Clec9A in the investigations presented in this Chapter.

Another receptor specifically expressed by cDC1s is XCR1. Unlike Clec9A, this receptor has not been detected on any other cell type, and has been proposed as a definitive marker for the cDC1 subset across species (123, 124). Recent studies have shown that targeting Ag to XCR1 also promotes the induction of CD8+ and CD4+ T cell responses. However, unlike Clec9A- or DEC-205-targeting, this can be achieved using the natural ligand of XCR1, XCL1, as the targeting component. Two distinct XCL1 fusion constructs have been independently demonstrated to be effective vectors for the delivery of Ag to cDC1s. Hartung et al. used a direct fusion of XCL1 protein with the Ag (243), while Fossum et al. utilised vaccibodies consisting of XCL1 and Ag fused with a dimerization unit composed of human Ig hinge and Ch3 regions (244). Both constructs effectively induced Ag-specific T cell responses when administered i.v. The XCL1-Ag fusion construct designed by Fossum et al. was further shown to be capable of inducing Ag-specific T cell responses in vivo without additional adjuvant if delivered via DNA vaccination with electroporation or laser-assisted intradermal delivery (244, 245).

Importantly, both Clec9A and XCR1 have human homologues that exhibit similar expression and function as their murine counterparts (90, 108, 169, 170). This allows strategies that target either of these receptors to be more readily translated into humans. Indeed, preliminary data has been encouraging. Clec9A targeting has been

203

shown to generate strong immune responses in both non-human primates and a humanised mouse model (222, 239), while targeting human XCR1 expressed in mice with human XCL1 constructs has also been demonstrated to induce effective responses (243). The evidence so far suggests that both targeting strategies could be considered promising candidates for development into a human vaccine. However, the efficacy of targeting via XCL1 constructs relative to other DC-targeting strategies is currently unknown. A direct comparison of targeting via XCL1 constructs with targeting Clec9A would be important to establish which may be the more effective strategy to focus on developing for human use.

Homologous prime-boost is not necessarily effective

Most studies evaluating the efficacy of DC-targeting strategies have focused on the primary response induced after a single immunisation. It should be expected that the optimal Ag delivery construct is that which induces the strongest immune response after a single immunisation. However, this is not necessarily the case when considering a prime-boost scenario, which is typically used to elicit long-lasting immunity after vaccination. It has been noted during the development of vaccines for diseases such as malaria and HIV, that highly immunogenic constructs can actually be detrimental to attempts to boost the response with a second immunisation. This has been exemplified by the development of vaccination regimes using recombinant viral vectors.

Recombinant viral vectors, most commonly replication-defective poxviruses or adenoviruses engineered to carry genes encoding an Ag of interest, have been used as potent vaccine delivery systems. The use of viruses to deliver the encoded Ag allows the expression of Ag intracellularly, mimicking a natural infection, which is particularly effective for inducing CTL responses, as well as Ab responses (411, 412). The superior potency of recombinant viral vectors over other delivery systems such as DNA vaccination has been attributed to the presence of numerous PAMPs within the viral vector that serve as intrinsic adjuvant. However, these determinants can also

204

raise a strong anti-vector response, which can be problematic when attempting to boost the response.

A study by Li et al. was one of the first to demonstrate that homologous prime- boosting, that is, immunisation twice with the same viral vector vaccine, generates less effective CD8+ T cell responses than heterologous prime-boosting with two distinct viral vectors (413). Mice primed with recombinant influenza virus and boosted 3 weeks later with recombinant vaccinia virus, both vectors carrying genes encoding distinct malaria Ags, showed significantly greater protection against malaria infection compared with mice that received two doses of the same viral vector (413). Subsequent studies supported a similar trend for enhanced immunogenicity with heterologous prime-boost vaccination. Gilbert et al. investigated the efficacy of adenovirus (Ad) or modified vaccinia virus Ankara (MVA) viral vector vaccines carrying malaria Ags (414). Using heterologous combinations of the two vectors, by priming with Ad and boosting with MVA, or vice versa, was found to be significantly more effective than homologous prime-boost by immunisation with the same construct twice. Significantly greater numbers of Ag-specific IFNγ-producing CD8+ T cells were induced after heterologous prime-boost, which correlated with enhanced protection against malaria challenge (414).

A similar advantage of heterologous prime-boosting is seen when DNA vaccination is combined with recombinant viral vector boost. DNA vaccination involves immunisation with plasmid DNA encoding specific genes that can be taken up and translated by host cells into Ags of interest (415, 416). This is an effective method of priming Ag-specific T cell responses, and as no vector-specific immunity is induced, it can be administered multiple times to the same host. However, the immunogenicity of DNA vaccination even after multiple vaccinations is relatively low in humans, prompting investigators to consider using viral vectors to boost the response. In comparing various combinations of DNA and MVA-vectored vaccines containing malarial Ags, it was found that homologous boosting with DNA followed by DNA vaccination or MVA followed by MVA did not protect mice against malaria infection (417). Instead, only DNA vaccination following by MVA boosting conferred

205

significant protection, which correlated with significantly greater induction of Ag- specific CD8+ T cell responses. The order of immunisation was critical, as priming with MVA and boosting with DNA did not confer protection. The success of DNA prime-viral vector boost regimes has led to their widespread use in the development of various vaccines, including vaccines against Ebola (418), tuberculosis (419) and HIV (420).

The improved responsiveness to heterologous prime-boost vaccination has also been shown to extend to humans, which has significant implications for the design of human vaccine trials. In one example, human volunteers were vaccinated with malaria Ags delivered by various combinations of DNA vaccination or MVA vectors. The optimal schedule was found to be a heterologous prime-boost involving three immunisations with DNA vaccination followed by one boost with MVA, each administered three weeks apart. Although three immunisations with DNA could induce a low level of Ag-specific T cell responses, responses were significantly increased after administration of a boost with MVA. However, volunteers who received further boosts with MVA produced lower responses (421). This suggests that MVA can only effectively boost an individual that has not previously been immunised with MVA.

T cell competition interferes with boosting

Why does heterologous prime-boosting induce enhanced Ag-specific responses compared with homologous prime-boosting? One proposed theory relates to a possible T cell immunodominance effect. Immunisation with viral vectors is known to induce strong immunity against components of the vector itself as well as the inserted Ag. An investigation of the proportion of T cells responding to a vaccinia virus vector after immunisation revealed that, at the peak of the response, approximately 25% of CD8+ T cells and 5% of CD4+ T cells in the spleen responded to in vitro restimulation with vaccinia-infected cells. However, the proportion of cells responsive to the inserted Ag epitope was approximately 20- to 30-fold less, indicating that the response against the vector may be many times greater than the response against the inserted

206

Ag (422). Thus, while immunisation with a viral vector can induce potent immune responses, these may be predominantly raised against components of the vector, at the expense of responses against the inserted Ag of interest. Boosting with the same vector would further bias towards vector-specific responses over inserted Ag-specific responses. It has been hypothesized that boosting with an alternative vector, that has only the Ag of interest in common with the priming construct, is more effective because it would preferentially promote Ag-specific responses without boosting the anti-vector response from the initial immunisation (265, 423). A similar mechanism could explain the observed requirement for DNA priming to precede viral vector boosting for successful boosting. Priming with DNA does not generate vector-specific responses, inducing a primary response that is specifically directed against the Ag of interest. This specific response can then be boosted by administration of the more potent but less specific viral vector. In the opposite scenario, a primary response induced by a viral vector may be dominantly directed against the vector, resulting in a relatively weaker Ag-specific response that is poorly boosted by the less immunogenic DNA vaccination (265, 423).

A study by de Mare et al. provided evidence to support this theory (424). Homologous prime-boosting with Semliki Forest virus (SFV) vector expressing the E6 and E7 Ag from human papillomavirus induces effective anti-E7 CTL responses. Transfer of anti-SFV neutralising antibodies prior to immunisation did not appear to inhibit the anti-E7 CTL response, despite a reduction in transgene expression. However, the anti- E7 CTL response could be inhibited by priming with an SFV vector expressing an irrelevant protein prior to the boost with SFV-E6E7. This suggested that T cell- mediated immunity generated by the initial prime was inhibiting the CTL response induced by the boost. One possible mechanism by which CTL may inhibit is by the specific killing of SFV-infected cells upon boosting, although this was deemed unlikely as the reduction in transgene expression observed could be entirely accounted for by the presence of anti-SFV neutralising Abs. Instead, the authors proposed that T cell competition for limiting resources such as APCs and cytokines was inhibiting the Ag- specific response. Specifically, priming of vector-specific T cells but not Ag-specific T cells during the initial immunisation could cause a competitive advantage for vector-

207

specific T cells over Ag-specific T cells during the boost. Thus, the authors considered whether the T cell response could be skewed more towards Ag-specific T cells by including the Ag of interest in the priming immunisation. Indeed, additional co- administration of the E7 protein in virosomes during the priming with SFV expressing an irrelevant protein could restore the level of CTL activity after SFV-E6E7 boost to that seen after homologous prime-boosting with SFV-E6E7 (424). Thus, in this scenario, it appears generating T cell responses predominantly of a different specificity during priming can inhibit boosting by competing with Ag-specific T cells.

Anti-vector immunity interferes with boosting

Another factor that could interfere with the efficacy of homologous prime-boost is the direct interference of pre-existing immunity to the boosting construct. Immunisation with a viral vector can induce anti-vector responses, including neutralising Abs, that may specifically recognise and inhibit the activity of the same vector when administered as a boost. Using a distinct vector as the boost would bypass the reactivity of any primary response against the vector. This is a likely explanation for why a single boost with MVA administered in McConkey et al.’s study effectively enhanced responses, but a second boost with the same construct did not (421). The efficacy of priming with DNA prior to boosting with a viral vector could also be due to the lack of vector-specific responses raised during DNA vaccination, resulting in less interference with the boost than would occur during homologous prime-boost with a viral vector.

In the case of recombinant viral vectors, pre-existing vector-specific immunity can also arise due to natural infection with viruses that are common human pathogens, such as adenoviruses. Thus, the impact of pre-existing vector-specific immunity on the efficacy of boosting with viral vectors has been extensively studied. Yang et al. showed that pre-priming of mice with adenovirus (Ad) vector completely abolished the capacity of a homologous Ad prime-boost vaccination to induce CTL and Ab responses against the inserted vaccine Ag, but did not affect the induction of responses by MVA-vectored prime-boost vaccination. Similarly, pre-priming with

208

MVA abolished the immunogenicity of MVA prime-boost without interfering with the potency of Ad prime-boost vaccination (425). Further investigations established that both Ab and CD8+ T cell vector-specific responses could contribute to this interference. Adoptive transfer of either serum or CD8+ T cells from Ad pre- immunised mice was found to inhibit the induction of immunity upon immunisation with Ad-vectored vaccine (426).

A similar phenomenon has been demonstrated in non-human primates. Pre-exposure of Rhesus macaques to an adenovirus human serotype 5 vector (AdHu5) expressing an irrelevant Ag was found to severely impair the capacity of subsequently administered AdHu5-vectored vaccines to induce Ab responses against the inserted Ag of interest. However, pre-existing immunity against AdHu5 vectors did not interfere with the efficacy of immunisation with adenovirus vectors of a different serotype (427). Overall, these findings indicate that vector-specific immunity can severely impair the induction of immune responses to vaccines utilising the same vector, but this impairment can be bypassed by use of a heterologous vector. This is analogous to the use of distinct vectors during heterologous prime-boosting, to overcome the vector-specific immunity that can interfere with boosting if the same construct is used.

DC-targeted vaccines may encounter the same issues

The lessons learned from recombinant viral vector vaccines could well apply to new generation vaccine constructs, including the DC-targeted vaccines that are the focus of this Chapter. Whether it is due to the dominance of vector-specific T cells or the presence of vector-neutralising immunity, the main impediments to homologous boosting appear to derive from the induction of immunity directed against the vector. However, the immunogenicity of the vector itself may be a key determinant of the potency of certain vaccines. Indeed, a recent study by our group suggested that the potency of targeting Ag to DCs via anti-Clec9A mAbs may derive at least in part from the immunogenicity of the targeting mAb component (222). Amongst a panel of anti- Clec9A mAbs, some (10B4, 42D2) were found to be immunogenic in the absence of

209

adjuvant, as measured by the induction of anti-mAb Ab responses, whereas others (397, 7H11, 1F6) were not immunogenic in the absence of adjuvant. Although the inefficient priming by 397 was likely due to its poor binding to CD8+ DCs, all four of the other mAbs (10B4, 7H11, 42D2 and 1F6) exhibited similar capacity to bind Clec9A on DCs, and similar levels of persistence in the serum after i.v. immunisation (222). Instead, the potency of mAbs 10B4 and 42D2 was attributed to the presence of immunogenic epitopes in their IgG backbones that were absent from 7H11 and 1F6. 10B4 is an IgG2a-κ, 42D2 is an IgG1-λ and the less potent 7H11 and 1F6 are IgG1-κ mAbs. When the relative immunogenicity of these different mAb isotopes was compared by administering non-targeting mAbs in the presence of adjuvant, only IgG2a-κ and IgG1-λ mAbs could induce anti-mAb responses, while IgG1-κ mAbs could not induce any detectable response (222). Thus, the enhanced potency of 10B4 and 42D2 likely stems from the inherent immunogenicity of the mAb backbone. This also implies that 10B4 and 42D2 mAbs are also more likely to induce strong anti-mAb responses as well as Ag-specific responses upon immunisation.

The capacity for targeting mAbs to induce responses against the vector is reminiscent of the vector-specific responses induced by recombinant viral vectors. Thus, similar issues encountered by viral vector vaccines could arise. It is possible that the potent immune responses generated after a single immunisation with DC-targeted vaccines may not be amenable to homologous boost by administration of the same targeting construct. Indeed, preliminary data from our group indicates that homologous prime- boost with two doses of 10B4-OVA is less effective than a single immunisation with 10B4-OVA. Thus, we sought to further investigate the mechanisms that inhibit homologous prime-boosting in this scenario, with the aim of developing an effective method of boosting the potent immune response generated by DC-targeted vaccines. This would be vital for designing vaccine regimens that maximise the potential of this novel vaccine technology, to offer a greater chance of success against diseases that have long evaded effective vaccination or treatment.

In this Chapter, we will firstly determine the more potent method of targeting Ag specifically to cDC1s between the only two currently described methods – targeting

210

Clec9A with anti-Clec9A mAbs, and targeting XCR1 via XCL1 constructs. The more effective strategy will then be used to examine whether a potent primary immune response induced by DC targeting interferes with the capacity to boost. Various heterologous combinations of DC-targeting vaccines will be assessed to determine both an effective strategy for boosting, and gain insight into the mechanisms that impair the development of a response after boosting. These findings will contribute to the development of DC-targeted vaccine strategies that can induce more potent, long- lasting T cell immunity, a feat that is not readily achievable with conventional vaccines. Our investigation of the factors dictating the efficacy of prime-boost may also provide information that could be applied to the design of vaccine schedules in general.

211

Results

6.1. Comparison of the efficacy of Clec9A and XCR1 targeting

6.1.1. Binding of Clec9A- and XCR1-targeting constructs to CD8+ DCs

Targeting Ag to Clec9A via anti-Clec9A mAbs or to XCR1 via XCL1 fusion constructs are two distinct strategies for delivering Ag to CD8+ and CD103+ DCs, collectively known as cDC1s, that have both been shown, in independent reports, to generate strong Ag-specific T cell responses (121, 215, 243, 244). We directly compared these two strategies to determine which is the more effective method of enhancing Ag- specific immune responses. We evaluated the relative efficacy of Clec9A-targeting using the anti-Clec9A mAb 10B4 to the targeting of XCR1 via the XCL1 vaccibody fusion construct described by Fossum et al (244).

Firstly, the integrity of each construct, bearing the model Ag OVA, and specificity for CD8+ DCs was confirmed by staining primary mouse splenic DCs. As expected, both 10B4-OVA and XCL1-OVA constructs bound specifically to CD8+ DCs, but not CD8- DCs or B cells (Figure 6.1). NIP-OVA is a control fusion construct with identical structure to the XCL1-OVA construct, except that the XCL1 targeting unit is replaced by the hapten 5-iodo-4-hydroxy-3-nitrophenacetyl (NIP)-specific single chain fragment variable (scFv) region from the B1-8 mAb (268). Thus, NIP-OVA is intended to be a non-targeting control for XCL1-OVA. However, the NIP-OVA construct showed an unexpectedly high level of specific binding to CD8+ DCs, although approximately 2-fold lower than that of XCL1-OVA (Figure 6.1).

212

CD8+ DC CD8- DC B cells

XCL1

NIP

10B4

2 3 4 5 2 3 4 5 2 3 4 5 10 10 10 10 10 10 10 10 10 10 10 10 PE

Figure 6.1. Binding of Clec9A- and XCR1-targeting constructs to DCs. Splenocytes (low density fraction prepared as described in “Binding of targeting constructs to DCs and B cells” in Chapter 2) were incubated with Fc block (rat IgG and anti-FcR mAb), then incubated with 10 μg/ml of XCL1-OVA, NIP-OVA or 10B4-OVA as indicated (black), or media only (grey). Binding was detected with anti-OVA biotin, followed by streptavidin PE, and fluorescence analysed by flow cytometry. CD8+ DCs (CD11c+CD8+), CD8- DCs (CD11c+CD8-) and B cells (CD11c-CD45R+) were identified by co-staining with anti-CD11c, anti-CD8 and anti-CD45R mAbs. Representative histograms of 2 independent experiments are presented.

6.1.2. Capacity for Clec9A- and XCR1-targeting constructs to induce T cell responses

The comparative ability of the 10B4 and XCL1 targeting constructs to deliver the model Ag OVA for presentation to OVA-specific CD8+ and CD4+ T cells in vivo was then assessed. CFSE-labelled OT-I or OT-II T cells were transferred into mice 1 day prior to immunisation with 10B4-OVA or XCL1-OVA, and the numbers of proliferating OT-I or OT-II cells in the spleen enumerated 3 or 4 days later, respectively. GL117-OVA, a non-targeting isotype control of 10B4, and NIP-OVA, a previously described non-targeting control of the XCL1 construct (268), were also

213

included. All constructs were delivered at doses calculated to be carrying an equivalent amount of OVA.

When administered in the absence of adjuvant, 10B4-OVA appeared to induce significantly greater OT-I and OT-II proliferation than any other construct (Figure 6.2A and C). XCL1-OVA, NIP-OVA and GL117-OVA all induced much weaker responses and were comparable to each other. The same trends were observed when polyIC was co-administered as adjuvant. While the responses to all four constructs were increased, 10B4-OVA remained the most effective by a significant margin (Figure 6.2B and D).

A OT-I proliferation B OT-I proliferation **** * *** * 5×104 3×105 ns ns * ns 4×104

2×105 3×104 endo CD8+ endo CD8+

6 6

2×104 1×105 OT-I cells/10 OT-I OT-I cells/10 OT-I 1×104

0 0 10B4 GL117 XCL1 NIP 10B4 GL117 XCL1 NIP + polyIC C OT-II proliferation D OT-II proliferation

**** **** 5 1.5×105 **** 3×10 ****

ns ns ns ns

1.0×105 2×105 endo CD4+ endo CD4+

6 6

5.0×104 1×105 OT-II cells/10 OT-II OT-II cells/10 OT-II

0.0 0 10B4 GL117 XCL1 NIP 10B4 GL117 XCL1 NIP + polyIC

214

Figure 6.2. OT-I and OT-II proliferation induced by immunisation with Clec9A- or XCR1-targeted OVA. 500,000 CFSE-labelled (A, B) OT-I or (C, D) OT-II cells were transferred into mice 1 day prior to i.v. immunisation with equimolar amounts of OVA delivered via 2 μg 10B4-OVA, 2 μg GL117-OVA, 1.15 μg XCL1-OVA or 1.15 μg NIP-OVA with (A, C) no adjuvant or (B, D) 50 μg polyIC. Divided OT-I or OT-II cells (CFSElo) in the spleen were enumerated by flow cytometry three or four days post-immunisation, respectively. The number of divided OT-I or OT-II cells per 106 endogenous (endo) CD8+ or CD4+ T cells, respectively, is presented. Cumulative data from at least 2 independent experiments is presented, with the exception of GL117 and NIP groups in (A), which was performed once. Each point represents one mouse.

In these experiments, the XCL1-OVA construct appeared to have little advantage over the non-targeting control construct, NIP-OVA. However, NIP-OVA appears to be a poor control as it can also target CD8+ DCs (Figure 6.1). Nevertheless, XCL1-OVA did not appear to have much advantage over GL117-OVA either, a non-targeting isotype control for 10B4 that does not bind CD8+ DCs and has relatively minimal activity (120, 215). This raised the question of whether there is any benefit at all to XCR1 targeting compared with non-targeted Ag delivery.

To address this, we compared the efficacy of immunisation with XCL1-OVA to equivalent or higher amounts of soluble OVA, in the presence of polyIC as adjuvant. Consistent with previous reports (243-245), immunisation with XCL1-OVA was significantly more effective than free OVA, inducing more OVA-specific IFNγ- producing CD8+ T cells than an equimolar amount of OVA (Figure 6.3). Even a 2.25 μg dose of OVA delivered via XCL1-OVA was found to be more effective than a 9 times higher dose of free OVA. As only one experiment was performed, these results did not reach statistical significance, but they were sufficient to confirm the capacity of the XCL1-OVA construct to promote the induction of CD8+ T cell responses.

215

tfigexp 150430 IFNg 6

4 + γ

% IFN % 2

0 equivalent dose of OVA (ug) 0.75 2.25 2.25 6.75 20.25 XCL1-OVA OVA

Figure 6.3. CD8+ T cell responses induced by immunisation with XCR1-targeted or soluble OVA. Mice were immunised i.v. with 1.15 μg XCL1-OVA (equivalent to 0.75 μg OVA), 3.45 μg XCL1-OVA (equivalent to 2.25 μg OVA), or the indicated doses of OVA, in the presence of 50 μg polyIC. Six days later, spleens were harvested and restimulated with OVA peptides and anti-CD28 for 6 hours, the final 5 hours in the presence of brefeldin A. Cells were then fixed, permeabilised and stained for intracellular IFNγ. The percentage of IFNγ+ cells as a proportion of total CD8+ cells in the spleen is shown. Experiment was performed once, each point represents 1 mouse.

We further confirmed that delivery of Ag to XCR1 could enhance CD4+ T cell responses. We found XCL1-OVA immunisation to induce significantly greater proliferation of transferred OT-II cells than a 10 times higher dose of free OVA, both with and without the addition of adjuvant (Figure 6.4).

216

A B * * 4 5 1×10 ** 2.5×10 *

8×103 2.0×105

3 5

6×10 endo CD4+ 1.5×10 endo CD4+

6 6

4×103 1.0×105

2×103 5.0×104 OT-II cells/10 OT-II OT-II cells/10 OT-II

0 0.0 XCL1-OVA OVA (3x) OVA (10x) XCL1-OVA OVA (3x) OVA (10x)

+ polyIC

Figure 6.4. OT-II proliferation induced by immunisation with XCR1-targeted or soluble OVA. 500,000 CFSE-labelled OT-II cells were transferred into mice 1 day prior to i.v. immunisation with 1.15 μg XCL1-OVA or the equivalent of 3 times more (2.45 μg) or 10 times more (7.5 μg) OVA with (A) no adjuvant or (B) 50 μg polyIC. Divided OT-II cells (CFSElo) in the spleen were enumerated by flow cytometry four days post-immunisation. The number of divided OT-II cells per 106 endogenous (endo) CD4+ T cells from 2 independent experiments is presented. Each point represents one mouse.

Both Clec9A- and XCR1-targeting have previously been demonstrated to promote strong Ab responses, so we examined whether one strategy may be more effective than the other at promoting Ag-specific Ab production. In a preliminary experiment, plasma collected from mice at various time points after immunisation with 10B4- OVA or XCL1-OVA was analysed for anti-OVA Ab levels by ELISA. When co- administered with polyIC as an adjuvant, 10B4-OVA and XCL1-OVA both induced strong anti-OVA titres. Although mice immunised with XCL1-OVA appeared to have lower Ab titres at day 6 post-immunisation, by day 13 and day 20 they appeared to have comparable or even slightly higher anti-OVA titres than 10B4-OVA immunised mice (Figure 6.5).

10B4 has previously been observed to possess an unusual ability to generate Ab responses even in the absence of adjuvant (120, 215). It has been reported that DNA

217

vaccination with XCL1 constructs can also induce Ab responses in the absence of adjuvant (244). However, this may be due to factors such as cell damage induced by the electroporation that accompanies DNA vaccination producing inflammatory signals that bypass the need for adjuvant. We examined whether intravenous administration of the XCL1 construct in the absence of such inflammatory signals can also induce Ab responses without additional adjuvant. We found mice immunised with XCL1-OVA in the absence of adjuvant had minimal anti-OVA titres 20 days post-immunisation, in stark contrast to the strong anti-OVA titres induced by 10B4- OVA without adjuvant (Figure 6.5).

tfig pool 12_13_19/8/16tfig exp13 26/6/15 anti OVA ±adj only 106

105

104

103 anti-OVA Ab titre 102

101 day 20 6 13 20 20 6 13 20

10B4 10B4 + polyIC XCL1 XCL1 + polyIC

Figure 6.5. Anti-OVA Ab responses induced by immunisation with Clec9A- or XCR1-targeted OVA. Mice were immunised i.v. with equimolar amounts of OVA delivered via 2 μg 10B4-OVA or 1.15 μg XCL1-OVA, in the presence or absence of 50 μg polyIC as indicated. 6, 13 and 20 days post-immunisation, plasma samples were collected and anti-OVA Ab endpoint titres measured by ELISA. Cumulative data from 2 independent experiments is presented, with the exception of 10B4 + polyIC groups, which was performed once. Each point represents one mouse.

218

Altogether, these results suggest that targeting Ag via XCL1 fusion constructs is just as effective as targeting Clec9A with 10B4 for generating Ab responses in the presence of adjuvant, but targeting Clec9A appears to be the superior method of delivering Ag to stimulate CD8+ and CD4+ T cell responses. Thus, for the following investigations regarding the induction of CD8+ T cell responses, we chose to focus on Clec9A targeting via 10B4.

6.2. Heterologous prime-boost strategies to boost immunity

6.2.1. Homologous prime-boost with 10B4-OVA does not boost CD8+ T cell responses

Targeting Ag to Clec9A is well-established to be an effective method of generating more potent T cell responses, particularly CTL responses, which could prove to be critically important for the development of vaccines against diseases for which there currently is no cure. Vaccines typically require multiple doses in order to boost the immune response for optimal efficacy, and it is likely that DC-targeted vaccines will be no exception. We attempted to evaluate the degree to which the immune response can be boosted after a second immunisation of 10B4-OVA two weeks after the primary immunisation. To our surprise, mice that had been immunised twice with 10B4-OVA had lower levels of OVA-specific IFNγ-producing CD8+ T cells measured 6 days after the boost, compared with mice measured 6 days after receiving only a single immunisation (Figure 6.6). This apparent reduction in T cell responses after boosting might be expected if the boosting had failed, since mice that received a single immunisation were measured at the peak of CD8+ T cell expansion (day 6), whereas the prime-boost group were assessed 20 days after the primary immunisation (6 days after the failed boost) when the primary response is declining into memory.

219

#20 exp 160830 CD8 IFNg 2.5

2.0

+ 1.5 γ

% IFN % 1.0

0.5

0.0 d0 ------10B4 d14 --- 10B4 10B4

Figure 6.6. Prime-boost immunisation with 10B4-OVA fails to boost CD8+ T cell responses. (A) Mice were immunised i.v. with 1 μg 10B4-OVA plus 50 μg polyIC twice (day 0 and day 14) or once (day 14). Six days after the boost (day 20), spleens were harvested and restimulated with OVA peptides and anti-CD28 for 6 hours, the final 5 hours in the presence of brefeldin A. Cells were then fixed, permeabilised and stained for intracellular IFNγ. The percentage of IFNγ+ cells as a proportion of total CD8+ cells in the spleen is shown. One representative experiment of 3 independent experiments is presented. Bars represent mean ± SEM, each point represents 1 mouse.

6.2.2. The primary Ab response interferes with boosting

The inability of a second immunisation to boost the response could be due to interference from the primary response generated by the initial immunisation. It is possible that the primary immunisation generates an Ab response that recognises and neutralises the targeting construct upon a second administration, preventing it from targeting to DCs efficiently and stimulating a secondary response. To test this hypothesis, we measured the persistence of the targeting construct in the serum of mice after primary and secondary immunisations. In unprimed mice 10B4-OVA could clearly be detected in the serum for at least 2 days after administration. In stark contrast, mice that had been primed with 10B4-OVA two weeks previously showed no

220

trace of 10B4-OVA in the serum even within 1 hour after receiving their second administration of 10B4-OVA (Figure 6.7).

Interp ng/ml 150514 anti-rat 500

unprimed 400 primed

300

200 10B4-OVA (ng/ml) 100

0 1 24 48 Time (hrs post-boost)

Figure 6.7. Targeting construct is rapidly eliminated from the serum of pre- primed mice. Mice (n=4) were immunised i.v. with 2 μg 10B4-OVA plus 50 μg polyIC (primed), or left unimmunised (unprimed). Two weeks later, both primed and unprimed groups were immunised i.v. with 2 μg 10B4-OVA plus 50 μg polyIC, and serum samples collected 1 hour, 1 day and 2 days later. 10B4-OVA present in the serum was captured by an ELISA coated with Clec9A (0.1 μg/ml) and detected with anti-rat IgG biotin followed by streptavidin HRP, and visualised with ABTS substrate. The mean optical density of duplicate samples was measured at 405-490nm, and the concentration of 10B4-OVA calculated from a standard curve. The mean ± SEM of four biological replicates is presented. Representative data of 2 independent experiments is shown.

To further confirm that the primary Ab response is responsible for inhibiting the boost, we examined whether effective boosting of T cell responses could be achieved in the absence of Ab responses. μMT mice have a disruption in the IgM heavy chain gene that prevents the development of mature B cells, and thus do not generate Ab responses (428). In these mice, the same prime-boost immunisation schedule that

221

failed to boost responses in B6 mice induced stronger responses after the second immunisation compared with the primary response (Figure 6.8). This suggests that bypassing the interference from the primary Ab response would allow effective boosting. uMT pool1_3 CD8 IFNg

4 ***

3 + γ 2 % IFN %

0 d0 ------10B4 d14 --- 10B4 10B4

Figure 6.8. Prime-boost immunisation with 10B4-OVA effectively boosts CD8+ T cell responses in μMT mice. μMT mice were immunised i.v. with 1 μg 10B4-OVA plus 50 μg polyIC twice (day 0 and day 14) or once (day 14). Six days after the boost (day 20), spleens were harvested and restimulated with OVA peptides and anti-CD28 for 6 hours, the final 5 hours in the presence of brefeldin A. Cells were then fixed, permeabilised and stained for intracellular IFNγ. The percentage of IFNγ+ cells as a proportion of total CD8+ cells in the spleen is shown. Bars represent mean ± SEM of cumulative data from 3 independent experiments. Each point represents 1 mouse.

6.2.3. Heterologous prime-boost with targeting mAbs possessing different species backbones

The 10B4 targeting mAb used to generate all the data so far in this study was raised in Wistar rats and possesses a rat IgG2a backbone (120). Our lab has also developed a mousinised version of this mAb (m10B4) that possesses the same variable regions,

222

allowing targeting to Clec9A, but the constant regions of mouse IgG1. We theorised that a primary response generated by m10B4 would have reduced reactivity against rat 10B4 (r10B4), and thus would cause less interference with the boosting of the response using r10B4 constructs. To test this, mice were primed with m10B4-OVA, then boosted with r10B4-OVA, and the OVA-specific CD8+ T cell response generated 6 days later compared with mice that received only a single immunisation of r10B4- OVA. However, this heterologous prime-boost regimen was unable to induce effective boosting, as responses measured after the boost were much lower than the primary effector response measured 6 days after a single immunisation (Figure 6.9). This analysis was only examined in a single experiment, as once it was clear that this strategy did not allow effective boosting, we proceeded to investigate other methods that could further reduce22/5/14 the d21 cross tetramer-reactivity between the prime and boost constructs.

1.5

1.0

% tetramer+ 0.5

0.0 d0 ------m10B4 d14 --- r10B4 r10B4

Figure 6.9. Heterologous prime-boost immunisation with m10B4-OVA followed by r10B4-OVA fails to boost CD8+ T cell responses. Mice were immunised i.v. with 1 μg m10B4-OVA on day 0 then boosted i.v. with 1 μg r10B4-OVA on day 14, or immunised once i.v. with 1 μg r10B4-OVA at day 14. All injections were co- administered with 5 nmol CpG ODN 1668 as adjuvant. Six days after the boost (day 20), the percentage of OVA-specific CD8+ cells in the blood was determined by tetramer staining. The percentage of tetramer+ cells as a proportion of total CD8+

223

cells in the blood is shown. This experiment was performed once. Bars represent mean ± SEM, each point represents 1 mouse.

6.2.4. Heterologous prime-boost with targeting mAbs carrying different forms of Ag

We next considered whether the reactivity of the primary response against the boosting construct could be further reduced by utilising a different form of the Ag. OBI is a polypeptide containing the linear MHC I-restricted and MHC II-restricted epitopes of OVA recognised by OT-I and OT-II cells, respectively, as well as the linear B cell OBI epitope of OVA (267). Previous data from our lab has indicated that mice immunised with OBI raise Ab responses that do not recognise whole OVA effectively, but are still capable of generating strong T cell responses (data not shown). Thus, mice primed with a construct carrying OBI may be less likely to recognise and inhibit a boosting construct carrying whole OVA. To investigate whether effective boosting of CD8+ T cell responses could be achieved by using both a different mAb backbone and a different form of the Ag in the boosting construct, mice were primed with r10B4- OBI and boosted two weeks later with m10B4-OVA. However, as before, no boosting was seen in comparison to the primary effector response (Figure 6.10). In the previous experiment (Figure 6.9), a different Th1-promoting adjuvant, CpG ODN 1668, was used instead of the usual polyIC. To consider the possibility that particular adjuvants could be more proficient at promoting boosting, we tested 1668 and polyIC in parallel in this experiment. We also tested the new 2006-21798 CpG ODN adjuvant shown in Chapter 3 to have enhanced adjuvant activity. However, we observed no boosting after the second immunisation regardless of which adjuvant was used (Figure 6.10).

224

2/7/14 d20 tetramer stain (bleeds) 4

2 % tetramer+

0 d0 ------r10B4-OBI --- r10B4-OBI --- r10B4-OBI d14 --- m10B4-OVA m10B4-OVA m10B4-OVA m10B4-OVA m10B4-OVA m10B4-OVA

+ CpG ODN 1668 + polyIC + CpG ODN 2006-21798

naive CpG CpG polyI:C polyI:C combo combo Figure 6.10. Heterologousprime primeboost-boost immunisationprime boost with r10B4prime-OBI followedboost by m10B4-OVA fails to boost CD8+ T cell responses. Mice were immunised i.v. with 1 μg r10B4-OBI on day 0 then boosted i.v. with 1 μg m10B4-OVA on day 14, or immunised once i.v. with 1 μg m10B4-OVA at day 14. All injections were co- administered with 5 nmol CpG ODN 1668 or 50 μg polyIC as indicated. Six days after the boost (day 20), the percentage of OVA-specific CD8+ cells in the blood was determined by tetramer staining. The percentage of tetramer+ cells as a proportion of total CD8+ cells in the blood is shown. This experiment was performed once. Bars represent mean ± SEM, each point represents 1 mouse.

It appeared that, despite the differences between the primary and secondary constructs, we still could not effectively boost the response. Thus, we sought to measure the residual cross-reactivity between the constructs to assess whether this could account for the poor capacity to boost. To determine the degree to which the primary Ab response interacted with the boosting constructs, the cross-reactivity of plasma from primed mice against the various constructs was measured by ELISA. Firstly, we confirmed that mice immunised with the OBI construct, which contained only the minimal OVA B cell epitope, generated poor Ab responses against whole OVA (Figure 6.11A). However, mice primed with rat 10B4-OBI showed strong

225

reactivity against mouse 10B4, recognising mouse 10B4 almost as efficiently as rat 10B4 (Figure 6.11B). Further investigation uncovered a striking revelation. GL117 is a rat IgG2a isotype control mAb for 10B4 that possesses the same constant regions and only differs at the variable regions. Plasma from mice immunised with mouse 10B4 was found to bind strongly to rat 10B4, but had no detectable affinity for rat GL117. This suggests that immunisation with mouse or rat 10B4 raises Ab responses that are predominantly specific for the variable regions of 10B4. This would explain why responses raised against mouse or rat 10B4, which share the same variable regions, are strongly cross-reactive against both constructs. Thus, heterologous prime-boosting with mouse and rat 10B4 constructs is insufficient to bypass the inhibition caused by cross-reactivity of the primary response against the boosting construct.

A Anti-OVA reactivity 2.0 primed with: 1.5 r10B4-OVA r10B4-OBI

1.0 O.D.

0.5

0.0 101 102 103 104 105 106 Plasma dilution

B Primed with: r10B4-OBI C Primed with: m10B4-OVA

4 3 reactive against: reactive against: r10B4 r10B4 3 m10B4 rGL117 2

2 O.D. O.D. 1 1

0 0 102 103 104 105 102 103 104 105 Plasma dilution Plasma dilution

Figure 6.11. Cross-reactivity of plasma from primed mice against targeting constructs. Plasma was collected from mice 14 days after i.v. immunisation with r10B4-OBI, r10B4-OVA or m10B4-OVA and assayed for reactivity against OVA or various constructs by ELISA. All injections were co-administered with 50 μg polyIC as

226

adjuvant. (A) ELISA plates coated with soluble OVA (10 μg/ml) were incubated with serially diluted plasma samples from mice primed with r10B4-OVA or r10B4-OBI. Bound anti-OVA Abs were detected with anti-mouse IgG HRP and visualised with ABTS substrate. The optical density of duplicate samples was measured at 405-490nm, and the mean ± SEM of four biological replicates is presented. (B) ELISA plates coated with r10B4 or m10B4 (1 μg/ml) were incubated with serially diluted plasma from mice primed with r10B4-OBI. Bound Abs were detected with anti-mouse IgG2a/b/c HRP (to avoid detecting m10B4 coated on the plate, which is an IgG1) and visualised with ABTS substrate. The optical density of duplicate samples was measured at 405-490nm, and the mean ± SEM of four biological replicates is presented. (C) ELISA plates coated with r10B4 or rat GL117 (rGL117) (1.5 μg/ml) were incubated with serially diluted plasma from mice primed with m10B4-OVA. Bound Abs were detected with anti- mouse IgG HRP and visualised with ABTS substrate. The optical density of duplicate samples was measured at 405-490nm, and the mean ± SEM of four biological replicates is presented

6.2.5. Heterologous prime-boost combining Clec9A- and XCR1- targeting

Ag can also be targeted to the cDC1 subset, including CD8+ and CD103+ DCs, without using mAbs, by delivery to XCR1 via XCL1 fusion constructs. Although this method appears to be less effective than Ag delivery via 10B4, as demonstrated by our results in Section 6.1, it is nonetheless capable of enhancing CD8+ T cell responses. As XCL1 is a chemokine with a completely unrelated biochemical structure to mAbs such as 10B4, we expected Ab responses raised against 10B4 to be minimally reactive against XCL1, and vice versa. Thus, a heterologous prime-boosting strategy combining 10B4 and XCL1 targeting may be able to overcome the issues caused by interference from the primary response.

We investigated whether effective boosting could be achieved in mice primed with r10B4-OBI and boosted with XCL1-OVA, or vice versa. We found mice that were first immunised with XCL1-OVA, and then boosted with r10B4-OBI two weeks later, had

227

significantly greater numbers of IFNγ-producing CD8+ T cells 6 days after the boost than mice that received only the first or only the second immunisation (Figure 6.12). The reverse protocol – priming with r10B4-OBI and boosting with XCL1-OVA – also induced significantly higher CD8+ T cells than a single immunisation with XCL1- OVA 6 days prior to the readout. However, in this case, mice that received both immunisations seemed to have comparable responses to mice that received only the first immunisation of rat 10B4-OBI 20 days prior to the readout. This suggests that the order in which the two constructs are administered affects the capacity for boosting. Administering 10B4-OBI two weeks after XCL1-OVA appears to boost the response, but giving XCL1-OVA after 10B4-OBI priming does not. These results also confirm that targeting via 10B4 constructs induces far stronger responses compared to XCL1 constructs. Significantly greater responses were induced by r10B4-OBI compared with XCL1-OVA both 6 days and 20 days post-immunisation (Figure 6.12).

pool #11_13 CD8 IFNg **** 20 **** **** ns ** *** 15 + γ 10 % IFN %

0 d0 --- XCL1-OVA --- XCL1-OVA r10B4-OBI --- r10B4-OBI d14 ------r10B4-OBI r10B4-OBI --- XCL1-OVA XCL1-OVA

Figure 6.12. Heterologous prime-boost immunisation with r10B4-OBI and XCL1-

naive + OVA effectively boosts CD8 ---/OBIT cell responses. MiceOBI/--- were immunised i.v. at day 0 XCL1/------/XCL1 XCL1/OBI OBI/XCL1 and day 14 with 2 μg r10B4-OBI or 1.15 μg XCL1-OVA, or left unimmunised, as indicated. All injections were co-administered with 50 μg polyIC as adjuvant. Six days

228

after the boost (day 20), spleens were harvested and restimulated with OVA peptides and anti-CD28 for 6 hours, the final 5 hours in the presence of brefeldin A. Cells were then fixed, permeabilised and stained for intracellular IFNγ. The percentage of IFNγ+ cells as a proportion of total CD8+ cells in the spleen is shown. Bars represent mean ± SEM of cumulative data from 3 independent experiments. Each point represents 1 mouse.

We next verified whether the ability of XCL1 and 10B4 constructs to promote effective boosting when used in combination is due to their hypothesized lack of cross-reactivity. Indeed, plasma from mice primed with XCL1-OVA had no detectable reactivity to 10B4-OBI (Figure 6.13A), while plasma from mice primed with 10B4- OBI showed minimal reactivity to XCL1-OVA (Figure 6.13B). Interestingly, plasma from 10B4-OVA-primed mice was strongly reactive to both 10B4-OBI and XCL1- OVA, demonstrating that the presence of either the same targeting mAb or the same Ag was sufficient to induce cross-reactivity against another construct (Figure 6.13A and B).

229

A Reactivity against: 10B4-OBI

3 primed with: 10B4-OBI 10B4-OVA 2 XCL1-OVA O.D. 1

0 101 102 103 104 105 Plasma dilution

B Reactivity against: XCL1-OVA

3 primed with: XCL1-OVA 10B4-OVA 2 10B4-OBI O.D. 1

0 101 102 103 104 105 Plasma dilution

Figure 6.13. Cross-reactivity of plasma from primed mice against XCL1 or 10B4 targeting constructs. ELISA plates coated with (A) r10B4-OBI (1.5 μg/ml) or (B) XCL1-OVA (1.5 μg/ml) were incubated with serially diluted plasma samples collected from mice 20 days after i.v. immunisation with 2 μg r10B4-OBI, 2 μg r10B4-OVA or 1.15 μg XCL1-OVA. All injections were co-administered with 50 μg polyIC as adjuvant. Bound Abs were detected with anti-mouse IgG HRP and visualised with ABTS substrate. The optical density of duplicate samples was measured at 405-490nm, and the mean ± SEM of four biological replicates is presented.

230

6.2.6. Heterologous prime-boost combining Clec9A- and XCR1- targeting generates effective anti-tumour responses

Having demonstrated that priming with XCL1-OVA and boosting with 10B4-OBI is an effective method of boosting CD8+ T cell responses, we next investigated whether these enhanced responses would confer greater immunity against cancer, one of the key diseases against which DC-targeted vaccines are currently being developed. Due to the challenges of generating immunity under conditions of tolerance and immunosuppression that are common in cancer patients, effective prime-boost vaccination is most likely required. Mice were immunised as before with XCL1-OVA followed by 10B4-OBI boost 2 weeks later, or with only one of the immunisations, or left unimmunised. At day 20, mice were then injected subcutaneously (s.c.) with OVA-expressing B16 melanoma cells (B16-OVA), which results in the growth of a solid tumour at the site of injection. The rate of growth of the tumours was monitored, and mice killed when the tumour reached an area of 100 mm2.

Unimmunised mice had a median survival time of 11 days, which did not seem to be enhanced by immunisation with XCL1-OVA 20 days prior to tumour inoculation. A single immunisation of 10B4-OBI 6 days prior to tumour injection resulted in a median survival of 20 days, while prime-boosted mice had a median survival of 18 days, although none of the groups were statistically significantly different from the unimmunised group, likely due to the spread within the groups (Figure 6.14A). The time to emergence of tumours between the different groups followed similar trends, though mice that received the prime-boost or a single immunisation of 10B4-OBI appeared to have delayed tumour emergence compared with mice the received a single immunisation of XCL1-OVA (Figure 6.14B). The differences between the groups were more apparent if tumour size at a particular time point was compared. At day 14 post-tumour injection, none of the mice that received a single immunisation of 10B4-OBI or the prime-boost regimen had palpable tumour growth, whereas the unimmunised mice and mice that received a single immunisation of XCL1-OVA all had sizeable tumours, bar one unimmunized mouse that appears to be an outlier (Figure 6.14C). Immunisation with a single dose of XCL1-OVA did appear to provide

231

some protection as these mice tended to have smaller tumours than the unimmunized mice at the same time point, although this trend was not significant.

A Survival 100 d0 d14 — — XCL1-OVA — — r10B4-OBI 50 XCL1-OVA r10B4-OBI % survival

0 0 10 20 30 40 days post-tumour injection

B Tumour emergence 100 d0 d14 — — XCL1-OVA — — r10B4-OBI * 50 * XCL1-OVA r10B4-OBI % tumour-free

0 0 10 20 30 days post-tumour injection

C Day 14 tumour area *

150 * ns )

2 ns 100

50 tumour area (mm

0 d0 --- XCL1-OVA --- XCL1-OVA d14 ------r10B4-OBI r10B4-OBI

Figure 6.14. Prime-boost is no more effective than a single vaccination 6 days prior to tumour inoculation for preventing B16-OVA tumour growth. Mice were immunised i.v. at day 0 and day 14 with 2 μg r10B4-OBI or 1.15 μg XCL1-OVA, or left

232

unimmunised, as indicated. All injections were co-administered with 50 μg polyIC as adjuvant. Six days after the boost (day 20), mice were injected s.c. with 105 B16-OVA cells, and tumour size measured daily with calipers. Mice were killed as an endpoint when tumours reached 100 mm2. This experiment was performed once. (A) Survival rate of tumour-bearing mice. A log-rank (Mantel-Cox) test determined none of the curves to be statistically significantly different. (B) Rate of emergence of tumours. Significance between curves was tested with a log-rank (Mantel-Cox) test, * p < 0.05. (C) Tumour area measured at day 14 post-tumour injection. Statistical significance determined by two-tailed unpaired Student’s t-test.

Due to the timing of the immunisations and the subsequent tumour challenge, the above experiment was not optimal for making certain comparisons between the groups. For example, mice that received a single immunisation of 10B4-OBI would be expected to be in the effector phase of the response when administered with tumour cells 6 days later, while mice that received a single immunisation of XCL1-OVA 20 days prior to tumour administration would likely have experienced contraction of the effector response. To reduce the impact of this confounding factor, we immunised mice as before, but waited until 14 days after the last immunisation to challenge with tumours. At this time point, it was expected that all groups would have undergone contraction of the effector response and begun the transition into the memory phase.

With this schedule, the advantage of prime-boost vaccination was far more apparent. Mice that received prime-boost vaccination had a median survival of 24 days, significantly longer than any of the groups that received a single immunisation, which had a median survival of 18-19 days (Figure 6.15A). All of the singly immunised groups showed significantly greater protection than the unimmunised group, which had a median survival of 15 days. The time of tumour emergence followed the same trends (Figure 6.15B). Comparing tumour area 14 days post-tumour challenge revealed that all of the groups that received a single immunisation tended to have smaller tumours than the unimmunised group, while the prime-boost group was entirely tumour-free (Figure 6.15C).

233

It should be noted that growth rates of the tumours should not be directly compared between independent experiments, as the growth of the tumour can vary between experiments. For example, unimmunised mice in the first experiment had a median survival of 11 days, but 15 days in the second experiment. This may stem from variables such as the accuracy of cell counting or the health and confluency of the tumour cell line when harvested for injection. Despite this variation, the trends between the groups have held consistent in independent experiments.

234

A Survival 100 d0 d14 — — XCL1-OVA — * 50 — XCL1-OVA * — r10B4-OBI % survival XCL1-OVA r10B4-OBI

0 0 10 20 30 days post-tumour injection

B Tumour emergence 100 d0 d14 — — XCL1-OVA — * 50 — XCL1-OVA * — r10B4-OBI

% tumour-free XCL1-OVA r10B4-OBI

0 0 5 10 15 20 25 days post-tumour injection

C Day 14 tumour area

ns 150 ** * * ns **** ) 2 100

50 tumour area (mm

0 d0 --- XCL1-OVA ------XCL1-OVA d14 ------XCL1-OVA r10B4-OBI r10B4-OBI

Figure 6.15. Prime-boost vaccination 14 days prior to tumour inoculation protects mice from B16-OVA tumour growth. Mice were immunised i.v. at day 0 and day 14 with 2 μg r10B4-OBI or 1.15 μg XCL1-OVA, or left unimmunised, as indicated. All injections were co-administered with 50 μg polyIC as adjuvant. 14 days after the boost (day 28), mice were injected s.c. with 105 B16-OVA cells, and tumour size measured

235

daily with calipers. Mice were killed as an endpoint when tumours reached 100 mm2. Representative data of 2 independent experiments is presented. (A) Survival rate of tumour-bearing mice. Significance between curves was tested with a log-rank (Mantel-Cox) test, * p < 0.05. (B) Rate of emergence of tumours. Significance between curves was tested with a log-rank (Mantel-Cox) test, * p < 0.05. (C) Tumour area measured at day 14 post-tumour injection. Statistical significance determined by two- tailed unpaired Student’s t-test.

It is known that an anti-tumour immune response can cause the downregulation of immunogenic Ags by tumour cells as a form of immune escape. This has been observed to occur as a consequence of both endogenous anti-tumour responses and immunotherapy (429, 430). Thus, we investigated whether the anti-tumour responses generated by our prime-boost vaccination would cause tumours to downregulate the immunogenic Ag, in our case, OVA. In a preliminary investigation to determine whether any of the vaccines could cause downregulation of OVA expression, we measured surface OVA expression on tumours retrieved when mice were killed as an endpoint. Although there was great variability within the groups, there appeared to be a trend towards lower OVA expression on tumours harvested from mice that received the prime-boost vaccination (Figure 6.16). This suggests that the stronger immune response induced by prime-boost vaccination may have put greater selective pressure on tumours to downregulate the immunogenic Ag.

236 Fold change vs bg

** 40 *

Fold change MFI 10

0 d0 --- XCL1-OVA ------XCL1-OVA d14 ------XCL1-OVA r10B4-OBI r10B4-OBI

Figure 6.16. Expression of OVA on the surface of tumours harvested at endpoint. naïve ---/OBI XCL1/------/XCL1 Mice were immunised i.v. at day 0 and day 14 withXCL1/OBI 2 μg r10B4-OBI or 1.15 μg XCL1- OVA, or left unimmunised, as indicated. All injections were co-administered with 50 μg polyIC as adjuvant. 14 days after the boost (day 28), mice were injected s.c. with 105 B16-OVA cells. Mice were killed when tumours reached 100 mm2 in area, and tumours (which express FLAG-tagged OVA) harvested and stained with anti-FLAG biotin, followed by streptavidin PE, then analysed by flow cytometry. Bars represent mean ± SEM fold change MFI versus secondary only background.

237

Discussion

Targeting Clec9A via anti-Clec9A mAb is more effective for inducing T cell responses than targeting XCR1 via XCL1 construct

CD8+ and CD103+ DCs, collectively referred to as cDC1s, are known to be the main subset of DCs that promotes cross-presentation of Ag to CD8+ T cells to induce CTL responses (77-79). The success of targeting Ag to Clec9A for generating enhanced CTL responses has been attributed to its highly selective expression on cDC1s, allowing concentrated delivery of Ag to this subset. Recently, targeting Ag to another cDC1-restricted receptor, XCR1, via its natural ligand XCL1, has emerged as another effective method of enhancing CTL responses (243-245). We compared these two targeting strategies to determine which is more efficient at promoting Ag presentation to T cells, and found targeting Clec9A with the anti-Clec9A mAb 10B4 to induce significantly greater proliferation of both CD8+ and CD4+ Ag-specific T cells than targeting XCR1 via XCL1 fusion constructs (Figure 6.2).

It should be noted that this is not necessarily reflective of the relative capacity of Ag delivery to the different receptors to induce T cell responses, but rather a comparison of the specific targeting strategies employed. It is possible that XCL1 has lower affinity for its receptor or a shorter in vivo half-life than a mAb. Indeed, this has been suggested by the work of Hartung et al., in which delivery of OVA via an anti-XCR1 mAb was approximately 10-20 times more efficient than delivery via XCL1 at promoting OT-I and OT-II cell proliferation in vivo (243). It is unknown whether the increased efficiency of Ag delivery via anti-XCR1 mAbs could induce comparable responses to Ag delivery via anti-Clec9A mAbs.

Nevertheless, we chose to focus our investigation on targeting via XCL1 fusion constructs, as have other studies, because the ability to target XCR1 with its natural ligand instead of mAbs could afford some unique advantages. This is particularly relevant because there are currently no other demonstrated means of targeting Ag to DC receptors such as Clec9A and DEC-205 without the use of mAbs. Using XCL1

238

instead of mAbs could engage the natural XCR1-XCL1 chemokine axis that promotes the interaction of CD8+ T cells with DCs. XCL1 is also a significantly smaller molecule than a mAb (~10 kDa vs ~150 kDa), which could enhance tissue penetration. Finally, as XCL1 is naturally produced in the body during immune activation, there should be minimal reactivity of the immune system against XCL1, focusing the response on the delivered Ag. This latter property could allow for multiple immunisations with XCL1 targeting constructs without generating immunity against the vector. Indeed, the ability to delivery Ag to cDC1s with XCL1 instead of mAbs proved to be critical for designing a successful heterologous prime-boost regimen, as discussed below.

Boosting is inhibited by the generation of a potent primary Ab response

Our results indicated that the primary Ab response generated after a single immunisation with 10B4-OVA can severely inhibit the induction of immunity after boosting with the same construct. While administration of 10B4-OVA to naive mice resulted in the construct being detectable in the blood for up to 2 days post- immunisation, the same construct was completely undetectable even within 1 hour after administration to mice that had previously been immunised with 10B4-OVA two weeks prior (Figure 6.7). This suggests that the presence of anti-10B4-OVA Abs can rapidly neutralise and clear the construct. Alternatively, the 10B4-OVA may still be present but may be blocked from binding Clec9A, and thus would not be detectable in our ELISA assay. Preventing the construct from binding Clec9A would destroy its capacity to target DCs, severely impairing its immunogenicity. A third possibility is that 10B4-OVA was not detected due to bound anti-10B4-OVA Abs obstructing the binding of the anti-rat secondary reagent in our assay. However, we did not detect any bound mouse Ab with an anti-mouse secondary (data not shown), ruling out this possibility. Whatever the mechanism, the presence of pre-existing anti-10B4-OVA immunity appears to significantly impair the capacity of subsequently administered 10B4-OVA to boost the immune response. Further evidence that the primary Ab response is responsible for inhibiting the boost was suggested by the fact that

239

homologous prime-boost can enhance CD8+ T cell responses in μMT mice that lack Ab responses (Figure 6.8).

Heterologous prime-boost overcomes interference from primary Ab response

A rational strategy to overcome the interference from the primary response would be to boost with a heterologous construct that the primary Ab response is less likely to react with. We attempted heterologous prime-boost by priming with mouse 10B4 and boosting with rat 10B4, such that the backbones of the targeting mAbs used in each immunisation were largely dissimilar and shared only the variable regions. We also trialled altering the Ag delivered be a linear epitope, OBI, during the priming immunisation, which does not induce strong B cell responses against the whole OVA Ag delivered by the boost. However, neither of these approaches could successfully boost the response (Figure 6.9, 6.10). According to our original hypothesis, this would suggest that the distinct constructs used for boosting could still be recognised and neutralised by the primary Ab response. Indeed, this was confirmed by the strong cross-reactivity of plasma from mice immunised with mouse 10B4 against rat 10B4 and vice versa (Figure 6.11). The minimal reduction in reactivity despite changing the species backbone of the mAbs was surprising. However, this was clarified to be due to the Ab response being predominantly directed against the variable region of the targeting mAb, which is shared by both mouse and rat 10B4. Thus any heterologous combination of mAbs that share the 10B4 variable region is unlikely to permit effective boosting.

A potential solution would be to utilise other anti-Clec9A mAbs with distinct variable and constant regions, further minimising the potential for cross-reactivity. Our lab is currently undertaking preliminary investigations in this direction, which have not been presented here. Another solution is to deliver Ag to the same cells, cDC1s, via targeting constructs that do not utilise mAbs. This approach is feasible due to the discovery that XCL1 constructs can be used to target Ag to XCR1 on cDC1s. XCL1, as a chemokine, has a completely distinct structure from a mAb, and therefore cross-

240

reactivity from a response against a mAb is highly unlikely. Indeed, we found heterologous prime-boosting with an XCL1-OVA prime and 10B4-OBI boost (XCL1 prime-10B4 boost) to effectively boost CD8+ T cell responses to approximately 2-fold greater than that induced by a primary immunisation with 10B4-OBI (Figure 6.12). Interestingly, the reverse schedule of priming with 10B4-OBI and boosting with XCL1-OVA (10B4 prime-XCL1 boost) did not appear to enhance responses, which were comparable to that induced after receiving only the first immunisation. We then verified that Ab responses raised after XCL1-OVA prime were minimally reactive against 10B4-OBI, and vice versa, in accordance with our theory that the cross- reactivity of the primary response against the boosting construct is largely to blame for ineffective boosting (Figure 6.13). Notably, we have demonstrated that priming with XCL1-OVA induces strong anti-OVA Ab responses in the presence of adjuvant, comparable to that induced by 10B4-OVA (Figure 6.5). This indicates that the lack of interference with the 10B4-OBI boosting construct after XCL1-OVA priming is not due to XCL1-OVA inducing weak Ab responses, but rather due to the generation of an Ab response that does not effectively recognise the 10B4-OBI boosting construct.

Properties of heterologous prime-boost using XCL1 and 10B4 constructs

It is curious that our heterologous prime-boost regimen is only effective if XCL1- OVA is used for the prime, not the boost, given that either way minimises cross- reactivity from the primary Ab response. Perhaps a similar reasoning for the efficacy of DNA vaccination as a prime but not a boost during heterologous prime-boost vaccination may apply in this situation. DNA vaccination is believed to be more useful as a priming agent because it does not induce any vector-specific responses, inducing a primary T cell response that is predominantly directed against the Ag of interest. This more specific response is then effectively boosted by the use of agents such as viral vectors, which are more immunogenic but can also promote the induction of vector-specific responses. This order of immunisation avoids the induction of vector- specific responses during the primary immunisation, which could be detrimental to the generation of Ag-specific responses due to the potential for T cell competition

241

(265, 423). Immunisation with XCL1 could be considered analogous to DNA vaccination in that it is likely to induce less vector-specific immunity than 10B4, being a natural product of the immune response, and is also less immunogenic than 10B4. Thus, the advantages of priming with DNA vaccination before boosting with viral vectors may also apply to priming with XCL1 then boosting with 10B4. Further investigations into the impact of T cell competition in this system would be required before this theory can be confirmed.

We have demonstrated that XCL1 prime-10B4 boost immunisation is significantly more effective than a single immunisation at protecting mice from OVA-expressing tumours inoculated 14 days after the boost (Figure 6.15). Interestingly, prime-boost did not appear to have an advantage over a single immunisation when mice were challenged with tumours 6 days after the boost (Figure 6.14). Thus, the advantage of prime-boost immunisation seems to be more apparent after the peak of the effector response, when the T cell response is expected to have undergone contraction and begun the formation of memory. We speculate that, while our prime-boost vaccination can enhance the response during the effector phase (Figure 6.12), this effect may be amplified during the memory phase, potentially due to the formation of a larger or more effective pool of memory cells. Although we have yet to examine the memory response in detail, a greater capacity for prime-boost vaccination to enhance memory responses would be particularly relevant for prophylactic vaccines designed to elicit long-lasting immunity.

While this study used a prophylactic tumour model as a proof-of-principle for prophylactic vaccines, it should be considered that a cancer vaccine designed for humans could only be administered therapeutically, that is, after the onset of the tumour. Designing a vaccine that is effective in a therapeutic setting may prove to be more challenging, as the tumour already has a head start and the immune response could be developing in an immunocompromised environment. Further work to investigate whether a heterologous prime-boost approach is also effective therapeutically would be of value to extend these findings to the design of cancer vaccines.

242

Factors influencing the efficacy of prime-boost

Some studies have suggested that allowing a longer period of time between prime and boost immunisations generates stronger responses after boosting. Repeated DNA immunisation of mice against malaria resulted in greater immunogenicity the longer the interval between prime and boost, from 2 weeks up to 12 weeks. This correlated with protection from malaria challenge observed in mice boosted 8 or 12 weeks after the prime, but not in mice that were boosted earlier (431). In another example, DNA vaccination prime followed by recombinant poxvirus boost in Rhesus macaques was found to be slightly more effective at reducing malaria parasitemia if the immunisations were administered 21 weeks apart instead of 7 weeks apart (432). However, other studies have found shorter intervals to be more effective. McConkey et al. found that DNA vaccination prime followed by MVA boost was more effective if the interval between prime and boost was 3 weeks rather than 8 weeks (421). Thus it appears that the optimal prime-boost interval may differ between different vaccine regimens. It is currently unclear why administering the boost at different intervals may impact its efficacy. It is possible that, in situations where the primary immune response can interfere with boosting, allowing a longer period of time for the initial response to wane may improve boosting. In our system, we have determined that the primary Ab response has a significant inhibitory effect on the capacity to boost with a similar construct. However, previous data would suggest that the Ab titre induced after 10B4 immunisation can remain high for at least 35 weeks (433). Thus, within this timeframe, it is unlikely that allowing a longer period of time to lapse after the prime would significantly reduce the reactivity of the primary response against the boosting construct. It is possible that varying the interval between prime and boost could impact the outcome for other reasons, but this was not evaluated in our study.

The failure of homologous prime-boost in our system does not preclude that homologous prime-boost with DC-targeted vaccines can be effective in certain situations. For instance, 10B4 has previously been used to immunise mice against HIV Ags with a homologous prime-boost schedule. Mice were boosted 4 weeks after receiving the first immunisation, and 2 weeks later, the induction of Ag-specific IFNγ-

243

producing CD4+ T cells was evident (234). In another example, splenocytes from mice immunised 2 weeks apart with three doses of the tumour peptide MUC1 conjugated to a different anti-Clec9A mAb (7H11) produced more IFNγ upon peptide-specific restimulation than splenocytes from mice that received only one dose (237). This disparity with the inability of our experiments to demonstrate boosting with homologous doses of 10B4-OVA may reflect differences in the dosage, timing, targeting Ab, Ag, adjuvants and experimental readouts. However, it is possible that these previous examples of successful homologous prime-boost strategies did experience interference from the primary Ab response, but due to certain conditions, this inhibition did not completely prevent boosting. If this were to be the case, heterologous prime-boost strategies may be able to further enhance vaccine potency by reducing any potential interference from the primary response. Therefore, even in situations where homologous prime-boost is effective, the potential benefit of heterologous prime-boost should be examined in an effort to optimise vaccine design.

The work presented here is the first to clearly demonstrate that prime-boost immunisation with DC-targeted vaccines can suffer from inhibition mediated by the primary Ab response, and that using heterologous prime-boost strategies can be an effective method of enhancing the efficacy of the boost. It is likely that the interference of the primary response is exaggerated in our system due to the particularly strong Ab response that is generated by priming with 10B4 constructs. This further emphasizes the idea that enhancing the potency of vaccine constructs, through techniques such as directly targeting DCs, can cause issues with boosting. In the endeavour to create more potent vaccines, the capacity for primary immunisation to interfere with subsequent attempts to boost may be amplified. Therefore, assessing the potential for interference from the primary response and identifying heterologous prime-boost strategies that could be used to overcome it should be a critical component of the design of future vaccines. The integration of continued efforts to improve vaccine construct potency with prime-boost regimens that optimise their efficacy will afford us a better chance of achieving protection against diseases for which current vaccines are ineffective.

244

Chapter 7: Final conclusions

The increasing understanding of how DCs play such a critical role in generating immune responses has led to the development of novel techniques that utilise DCs for therapeutic benefit. One of the most striking advances has been the ability to induce potent immune responses against Ag delivered to DCs in vivo via mAbs or other constructs that target DC-specific receptors. In contrast to conventional vaccines, which can induce strong Ab responses but generally poor CTL responses, DC- targeted vaccines have the potential to induce strong CTL responses, which could be key for developing vaccines or therapeutics against diseases such as HIV, malaria and cancer.

Currently, the most effective methods of generating CTL by targeting Ag to DC involve targeting the receptors DEC-205 and Clec9A. This is likely related to the predominant expression of these receptors on the cDC1 subset in both mice and humans, which are specialised for the cross-presentation of Ag and production of cytokines such as IL-12 that promote Th1 differentiation and CTL induction. However, the delivery of Ag to DCs alone is insufficient to induce CTL responses, which requires co-administration of an effective adjuvant to activate the DCs. Thus, developing adjuvants with greater capacity to promote CTL responses could be an effective strategy to improve the efficacy of DC-targeted vaccines, or indeed vaccines in general.

We have demonstrated that the concept of targeting Ag to DCs can be extended to the targeting of adjuvant to DCs in order to improve their immunostimulatory activity. This was achieved by utilising the innate capacity for DEC-205 to mediate the binding and uptake of CpG ODN adjuvants. 21798 is a stimulatory CpG ODN with weak DEC-205 binding. Conjugation of 21798 to ODNs with strong DEC-205 binding, such as 14T or 2006, was shown to significantly improve its capacity to promote CD8+ T cell responses when delivered as an adjuvant alongside Ag targeted to Clec9A. This

245

enhanced response was abolished in mice lacking DEC-205, strongly suggesting that improved targeting to DEC-205 enhances the potency of CpG ODN adjuvants. This demonstrated that improving the targeting of adjuvants to DCs is a viable method of enhancing their potency.

Our findings showed that CpG ODNs retain some stimulatory effects even in the absence of DEC-205, suggesting that other receptors may be mediating their uptake in the absence of DEC-205. Identifying these receptors could provide another means of modulating the activity of CpG ODN adjuvants. CD14 was a potential candidate, having been previously been shown to mediate the uptake and stimulatory activity of CpG ODNs (332). However, our results concluded that CD14 is not involved in mediating the uptake or stimulatory activity of CpG ODNs.

DEC-205 has proven to be a useful receptor for targeting Ag, and now adjuvant, to DCs. Despite this, relatively little is known about its biological function, although DEC-205 has recently been identified to recognise keratin exposed by necrotic cells (175). It was considered whether the binding of synthetic ODNs to DEC-205 could indicate that DEC-205 can also bind naturally occurring nucleic acids. However, in a panel of various biological DNA samples including pathogenic and mammalian DNA, none were observed to bind DEC-205. Interestingly, several DNA samples were found to bind RAGE, suggesting a novel role for RAGE as a DNA-sensing receptor.

Establishing more effective methods of delivering Ag and adjuvant is certainly instrumental for developing more potent vaccines, and DC-targeting appears to be one of the most promising techniques to achieve this so far. As vaccine efficacy is typically amplified by boosting the response with multiple immunisations, the performance of DC-targeting constructs in a prime-boost scenario was evaluated. Interestingly, after priming with an anti-Clec9A targeting construct, responses could not be boosted with a second dose of the same construct. The inefficiency of homologous prime-boost was found to be caused by the primary Ab response reacting with and neutralising the targeting construct delivered in the boosting immunisation. To overcome this issue, various heterologous prime-boost strategies combining

246

different targeting constructs for the prime and boost were trialled. Ultimately, a significant reduction in cross-reactivity could only be achieved using a combination of Clec9A-targeting mAb and an XCR1-targeting construct comprised of the chemokine ligand XCL1, which allowed effective boosting of the CTL response. These results suggested that the stronger the response induced by a vaccine construct, the more likely that interference from the primary Ab response can reduce the capacity to boost the response. Thus, as more potent vaccines are pursued, strategies to overcome this issue may be necessary to optimise their efficacy.

This thesis investigated several means of augmenting the immune response generated by targeting DCs: from developing more potent adjuvants by targeting adjuvant to DCs, to assessing effective methods to deliver DC-targeted vaccines in a prime-boost regimen. An important consideration for this work is the fact that strategies that enhance immune responses in mice are irrelevant if the underlying mechanisms do not apply in humans. For this reason, the CpG ODNs 2006 and 21798 were chosen to examine the effect of improved DEC-205 targeting on ODN adjuvant activity. Both 2006 and 21798 have been demonstrated to be potent stimulators of human cells (295, 300), which suggests that the 2006-21798 combined ODN we have here shown to have enhanced potency in mouse models should also promote effective responses in humans.

A caveat to this assumption is that the expression of TLR9, the activating receptor required for CpG ODN activity, is distinct between mouse and human. In mice, TLR9 is expressed by cDCs, pDCs, B cells, monocytes and macrophages, while in humans, TLR9 is only expressed by pDCs and B cells (186, 284). The lack of TLR9 on cDCs in humans has called into question whether the potent Th1-polarising effects of CpG ODN observed in mice are retained in humans, as cDC1s are the main source of IL- 12, as well as the main cell population responsible for cross-priming CD8+ T cells (77, 79). However, results from clinical trials suggest that CpG ODNs adjuvants can promote CTL responses in humans (306, 308). It could be speculated that cDCs in humans, while lacking TLR9, can be indirectly activated via the secretion of high levels of type I IFNs by pDCs. It is also possible that pDCs themselves can cross-prime

247

CD8+ T cells upon activation by CpG ODN, as human pDCs have been shown to efficiently cross-present Ag, in contrast to mouse pDCs which appear to have more limited cross-presentation capacity (91, 128, 136).

Another scenario could be inferred from the recent finding by See et al. that the markers commonly used to identify human pDCs are also expressed by a population of cDC precursors (434). These cDC precursors, which would have been included in previous analyses of human “pDCs”, were shown to produce IL-12 and TNFα in response to CpG ODN stimulation (434). This suggests that human TLR9 expression is not as restricted to pDCs as previously believed, and that CpG ODN adjuvants may directly stimulate the production of Th1-promoting cytokines by cDC precursors.

Interestingly, DEC-205 expression appears to be very low on mouse pDCs (Figure 4.1), but is expressed at relatively high levels on human pDCs (435). In light of the findings of See et al., these DEC-205-expressing human pDCs may have included cDC precursors. This suggests that targeting adjuvant to DEC-205, which preferentially targets cDC1s in mice, may preferentially target pDCs (and possibly cDCs precursors) in humans. Whether the preferential targeting of adjuvant to distinct DC subsets across species could lead to different immune outcomes remains to be seen.

The homologous expression and function of the DC receptors Clec9A and XCR1 in both mouse and humans is a significant advantage for the translation of Clec9A- and XCR1-targeting strategies into human vaccines. The outcome and requirements for inducing immune responses with anti-Clec9A and anti-XCR1 targeting constructs observed in studies with mice are likely to be conserved in humans. This is also a key motivation for focusing our investigation of the optimal heterologous prime-boost regimens with DC-targeted vaccines on Clec9A- and XCR1-targeting strategies. It should be considered that the cross-reactivity of the murine Ab response against mAb constructs containing rat regions may be stronger than would be observed in humans, particularly as vaccines designed for humans use fully humanised mAbs. Nevertheless, the fundamental notion that cross-reactivity of the primary Ab response can interfere

248

with boosting still applies in humans, and is a key issue to consider when designing vaccination regimens.

In conclusion, these studies exemplify how a greater understanding of the biology of DCs and the mechanisms that drive immune responses can be translated into therapeutic benefit. The discovery that DEC-205 can mediate the immunostimulatory activity of CpG ODNs, and the examination of the molecular requirements for ODNs to bind DEC-205, allowed the development of improved ODNs adjuvants by enhancing their DEC-205 binding. In a similar vein, determining that the inefficacy of homologous prime-boost is caused by cross-reactivity of the primary Ab response against the boosting construct, allowed the design of prime-boost strategies to specifically overcome this issue. Even the initial concept of targeting DCs to enhance immunity arose from recognising the pivotal role that DCs play in initiating immune responses. While our studies focused on improving the induction of immunity with DC-targeted vaccines, as they are currently one of the most promising methods of generating potent CTL responses, these insights could contribute to the design of better adjuvants and vaccination regimens for any vaccine. It is anticipated that further studies to understand DCs, their receptors, and their role in the intricate biological network that is our immune system can lead to greater advances in vaccines and therapeutic strategies against the diseases for which there currently is no cure.

249

Bibliography

1. Alberts, B., A. Johnson, J. Lewis, M. Raff, K. Roberts, and P. Walter. 2002. Innate Immunity. In Molecular Biology of the Cell, 4th edition. Garland Science, New York. 2. Alberts, B., A. Johnson, J. Lewis, M. Raff, K. Roberts, and P. Walter. 2002. Lymphocytes and the Cellular Basis of Adaptive Immunity. In Molecular Biology of the Cell, 4th edition. Garland Science, New York. 3. Banchereau, J., and R. M. Steinman. 1998. Dendritic cells and the control of immunity. Nature 392: 245-252. 4. Banchereau, J., F. Briere, C. Caux, J. Davoust, S. Lebecque, Y. J. Liu, B. Pulendran, and K. Palucka. 2000. Immunobiology of dendritic cells. Annu. Rev. Immunol. 18: 767-811. 5. Mellman, I., and R. M. Steinman. 2001. Dendritic cells: specialized and regulated antigen processing machines. Cell 106: 255-258. 6. Villadangos, J. A., P. Schnorrer, and N. S. Wilson. 2005. Control of MHC class II antigen presentation in dendritic cells: a balance between creative and destructive forces. Immunol. Rev. 207: 191-205. 7. Reis e Sousa, C. 2006. Dendritic cells in a mature age. Nature reviews. Immunology 6: 476-483. 8. Garrett, W. S., L. M. Chen, R. Kroschewski, M. Ebersold, S. Turley, S. Trombetta, J. E. Galan, and I. Mellman. 2000. Developmental control of endocytosis in dendritic cells by Cdc42. Cell 102: 325-334. 9. West, M. A., A. R. Prescott, E. L. Eskelinen, A. J. Ridley, and C. Watts. 2000. Rac is required for constitutive macropinocytosis by dendritic cells but does not control its downregulation. Curr. Biol. 10: 839-848. 10. Platt, C. D., J. K. Ma, C. Chalouni, M. Ebersold, H. Bou-Reslan, R. A. Carano, I. Mellman, and L. Delamarre. 2010. Mature dendritic cells use endocytic receptors to capture and present antigens. Proc. Natl. Acad. Sci. U. S. A. 107: 4287-4292. 11. Delamarre, L., H. Holcombe, and I. Mellman. 2003. Presentation of exogenous antigens on major histocompatibility complex (MHC) class I and MHC class II molecules is differentially regulated during dendritic cell maturation. The Journal of experimental medicine 198: 111-122. 12. Steinman, R. M., and M. C. Nussenzweig. 2002. Avoiding horror autotoxicus: the importance of dendritic cells in peripheral T cell tolerance. Proc. Natl. Acad. Sci. U. S. A. 99: 351-358. 13. Villadangos, J. A., and P. Schnorrer. 2007. Intrinsic and cooperative antigen- presenting functions of dendritic-cell subsets in vivo. Nature reviews. Immunology 7: 543-555.

251

14. Blum, J. S., P. A. Wearsch, and P. Cresswell. 2013. Pathways of antigen processing. Annu. Rev. Immunol. 31: 443-473. 15. Roche, P. A., and K. Furuta. 2015. The ins and outs of MHC class II-mediated antigen processing and presentation. Nature reviews. Immunology 15: 203-216. 16. Steinman, R. M., B. Gutchinov, M. D. Witmer, and M. C. Nussenzweig. 1983. Dendritic cells are the principal stimulators of the primary mixed leukocyte reaction in mice. The Journal of experimental medicine 157: 613-627. 17. Rodriguez-Pinto, D. 2005. B cells as antigen presenting cells. Cell. Immunol. 238: 67-75. 18. Delamarre, L., M. Pack, H. Chang, I. Mellman, and E. S. Trombetta. 2005. Differential lysosomal proteolysis in antigen-presenting cells determines antigen fate. Science 307: 1630-1634. 19. Jung, S., D. Unutmaz, P. Wong, G. Sano, K. De los Santos, T. Sparwasser, S. Wu, S. Vuthoori, K. Ko, F. Zavala, E. G. Pamer, D. R. Littman, and R. A. Lang. 2002. In vivo depletion of CD11c+ dendritic cells abrogates priming of CD8+ T cells by exogenous cell-associated antigens. Immunity 17: 211-220. 20. Joffre, O. P., E. Segura, A. Savina, and S. Amigorena. 2012. Cross-presentation by dendritic cells. Nature reviews. Immunology 12: 557-569. 21. Segura, E., A. L. Albiston, I. P. Wicks, S. Y. Chai, and J. A. Villadangos. 2009. Different cross-presentation pathways in steady-state and inflammatory dendritic cells. Proc. Natl. Acad. Sci. U. S. A. 106: 20377-20381. 22. Pennock, N. D., J. T. White, E. W. Cross, E. E. Cheney, B. A. Tamburini, and R. M. Kedl. 2013. T cell responses: naive to memory and everything in between. Advances in physiology education 37: 273-283. 23. Murphy, K. 2012. Janeway's Immunobiology, 8th edition. Garland Science, New York. 24. Smith-Garvin, J. E., G. A. Koretzky, and M. S. Jordan. 2009. T cell activation. Annu. Rev. Immunol. 27: 591-619. 25. Walsh, K. P., and K. H. Mills. 2013. Dendritic cells and other innate determinants of T helper cell polarisation. Trends in immunology 34: 521-530. 26. Raphael, I., S. Nalawade, T. N. Eagar, and T. G. Forsthuber. 2015. T cell subsets and their signature cytokines in autoimmune and inflammatory diseases. Cytokine 74: 5-17. 27. Fishman, M. A., and A. S. Perelson. 1994. Th1/Th2 cross regulation. J. Theor. Biol. 170: 25-56. 28. Wynn, T. A. 2015. Type 2 cytokines: mechanisms and therapeutic strategies. Nature reviews. Immunology 15: 271-282. 29. Crotty, S. 2014. T follicular helper cell differentiation, function, and roles in disease. Immunity 41: 529-542. 30. Vinuesa, C. G., M. A. Linterman, D. Yu, and I. C. MacLennan. 2016. Follicular Helper T Cells. Annu. Rev. Immunol. 34: 335-368.

252

31. Josefowicz, S. Z., L. F. Lu, and A. Y. Rudensky. 2012. Regulatory T cells: mechanisms of differentiation and function. Annu. Rev. Immunol. 30: 531-564. 32. Workman, C. J., A. L. Szymczak-Workman, L. W. Collison, M. R. Pillai, and D. A. Vignali. 2009. The development and function of regulatory T cells. Cell. Mol. Life Sci. 66: 2603-2622. 33. Jonuleit, H., E. Schmitt, G. Schuler, J. Knop, and A. H. Enk. 2000. Induction of interleukin 10-producing, nonproliferating CD4(+) T cells with regulatory properties by repetitive stimulation with allogeneic immature human dendritic cells. The Journal of experimental medicine 192: 1213-1222. 34. Zhang, N., and M. J. Bevan. 2011. CD8(+) T cells: foot soldiers of the immune system. Immunity 35: 161-168. 35. Broere, F., S. G. Apasov, M. V. Sitkovsky, and W. van Eden. 2011. T cell subsets and T cell-mediated immunity. In Principles of Immunopharmacology: 3rd revised and extended edition. F. P. Nijkamp, and M. J. Parnham, eds. Birkhäuser Basel, Basel. 15-27. 36. Bennett, S. R., F. R. Carbone, F. Karamalis, J. F. Miller, and W. R. Heath. 1997. Induction of a CD8+ cytotoxic T lymphocyte response by cross-priming requires cognate CD4+ T cell help. The Journal of experimental medicine 186: 65-70. 37. Bennett, S. R., F. R. Carbone, F. Karamalis, R. A. Flavell, J. F. Miller, and W. R. Heath. 1998. Help for cytotoxic-T-cell responses is mediated by CD40 signalling. Nature 393: 478-480. 38. Schoenberger, S. P., R. E. Toes, E. I. van der Voort, R. Offringa, and C. J. Melief. 1998. T-cell help for cytotoxic T lymphocytes is mediated by CD40- CD40L interactions. Nature 393: 480-483. 39. Williams, M. A., and M. J. Bevan. 2007. Effector and memory CTL differentiation. Annu. Rev. Immunol. 25: 171-192. 40. Wiesel, M., and A. Oxenius. 2012. From crucial to negligible: functional CD8(+) T-cell responses and their dependence on CD4(+) T-cell help. Eur. J. Immunol. 42: 1080-1088. 41. Janssen, E. M., E. E. Lemmens, T. Wolfe, U. Christen, M. G. von Herrath, and S. P. Schoenberger. 2003. CD4+ T cells are required for secondary expansion and memory in CD8+ T lymphocytes. Nature 421: 852-856. 42. Shedlock, D. J., and H. Shen. 2003. Requirement for CD4 T cell help in generating functional CD8 T cell memory. Science 300: 337-339. 43. Sun, J. C., and M. J. Bevan. 2003. Defective CD8 T cell memory following acute infection without CD4 T cell help. Science 300: 339-342. 44. Obst, R., H. M. van Santen, D. Mathis, and C. Benoist. 2005. Antigen persistence is required throughout the expansion phase of a CD4(+) T cell response. The Journal of experimental medicine 201: 1555-1565. 45. Bevan, M. J., and P. J. Fink. 2001. The CD8 response on autopilot. Nature immunology 2: 381-382.

253

46. Batista, F. D., and N. E. Harwood. 2009. The who, how and where of antigen presentation to B cells. Nature reviews. Immunology 9: 15-27. 47. Nutt, S. L., P. D. Hodgkin, D. M. Tarlinton, and L. M. Corcoran. 2015. The generation of antibody-secreting plasma cells. Nature reviews. Immunology 15: 160-171. 48. Schroeder, H. W., Jr., and L. Cavacini. 2010. Structure and function of immunoglobulins. J. Allergy Clin. Immunol. 125: S41-52. 49. Chen, X., and P. E. Jensen. 2008. The role of B lymphocytes as antigen- presenting cells. Arch. Immunol. Ther. Exp. (Warsz). 56: 77-83. 50. Ahmed, R., and D. Gray. 1996. Immunological memory and protective immunity: understanding their relation. Science 272: 54-60. 51. Janeway, C., P. Travers, M. Walport, and M. Shlomchik. 2001. The course of the adaptive response to infection. In Immunobiology: The Immune System in Health and Disease. 5th edition. Garland Science, New York. 52. Obar, J. J., and L. Lefrancois. 2010. Memory CD8+ T cell differentiation. Ann. N. Y. Acad. Sci. 1183: 251-266. 53. MacLeod, M. K., J. W. Kappler, and P. Marrack. 2010. Memory CD4 T cells: generation, reactivation and re-assignment. Immunology 130: 10-15. 54. Kurosaki, T., K. Kometani, and W. Ise. 2015. Memory B cells. Nature reviews. Immunology 15: 149-159. 55. Chu, V. T., A. Beller, T. T. Nguyen, G. Steinhauser, and C. Berek. 2011. The long-term survival of plasma cells. Scand. J. Immunol. 73: 508-511. 56. Mueller, S. N., T. Gebhardt, F. R. Carbone, and W. R. Heath. 2013. Memory T cell subsets, migration patterns, and tissue residence. Annu. Rev. Immunol. 31: 137-161. 57. Nolz, J. C., G. R. Starbeck-Miller, and J. T. Harty. 2011. Naive, effector and memory CD8 T-cell trafficking: parallels and distinctions. Immunotherapy 3: 1223-1233. 58. Sallusto, F., D. Lenig, R. Forster, M. Lipp, and A. Lanzavecchia. 1999. Two subsets of memory T lymphocytes with distinct homing potentials and effector functions. Nature 401: 708-712. 59. Masopust, D., V. Vezys, A. L. Marzo, and L. Lefrancois. 2001. Preferential localization of effector memory cells in nonlymphoid tissue. Science 291: 2413- 2417. 60. Mueller, S. N., and L. K. Mackay. 2016. Tissue-resident memory T cells: local specialists in immune defence. Nature reviews. Immunology 16: 79-89. 61. Steinert, E. M., J. M. Schenkel, K. A. Fraser, L. K. Beura, L. S. Manlove, B. Z. Igyarto, P. J. Southern, and D. Masopust. 2015. Quantifying Memory CD8 T Cells Reveals Regionalization of Immunosurveillance. Cell 161: 737-749. 62. Carbone, F. R. 2015. Tissue-Resident Memory T Cells and Fixed Immune Surveillance in Nonlymphoid Organs. J. Immunol. 195: 17-22.

254

63. Schenkel, J. M., and D. Masopust. 2014. Tissue-resident memory T cells. Immunity 41: 886-897. 64. Kalia, V., S. Sarkar, T. S. Gourley, B. T. Rouse, and R. Ahmed. 2006. Differentiation of memory B and T cells. Curr. Opin. Immunol. 18: 255-264. 65. Slifka, M. K., M. Matloubian, and R. Ahmed. 1995. Bone marrow is a major site of long-term antibody production after acute viral infection. J. Virol. 69: 1895-1902. 66. Berard, M., and D. F. Tough. 2002. Qualitative differences between naive and memory T cells. Immunology 106: 127-138. 67. Veiga-Fernandes, H., U. Walter, C. Bourgeois, A. McLean, and B. Rocha. 2000. Response of naive and memory CD8+ T cells to antigen stimulation in vivo. Nature immunology 1: 47-53. 68. Steinman, R. M., and Z. A. Cohn. 1973. Identification of a novel cell type in peripheral lymphoid organs of mice. I. Morphology, quantitation, tissue distribution. The Journal of experimental medicine 137: 1142-1162. 69. Steinman, R. M., and Z. A. Cohn. 1974. Identification of a novel cell type in peripheral lymphoid organs of mice. II. Functional properties in vitro. The Journal of experimental medicine 139: 380-397. 70. Shortman, K., and Y. J. Liu. 2002. Mouse and human dendritic cell subtypes. Nature reviews. Immunology 2: 151-161. 71. Heath, W. R., and F. R. Carbone. 2009. Dendritic cell subsets in primary and secondary T cell responses at body surfaces. Nature immunology 10: 1237- 1244. 72. Merad, M., P. Sathe, J. Helft, J. Miller, and A. Mortha. 2013. The dendritic cell lineage: ontogeny and function of dendritic cells and their subsets in the steady state and the inflamed setting. Annu. Rev. Immunol. 31: 563-604. 73. Wilson, N. S., D. El-Sukkari, G. T. Belz, C. M. Smith, R. J. Steptoe, W. R. Heath, K. Shortman, and J. A. Villadangos. 2003. Most lymphoid organ dendritic cell types are phenotypically and functionally immature. Blood 102: 2187-2194. 74. Vremec, D., J. Pooley, H. Hochrein, L. Wu, and K. Shortman. 2000. CD4 and CD8 expression by dendritic cell subtypes in mouse thymus and spleen. J. Immunol. 164: 2978-2986. 75. Shortman, K., and W. R. Heath. 2010. The CD8+ dendritic cell subset. Immunol. Rev. 234: 18-31. 76. Heath, W. R., and F. R. Carbone. 2013. The skin-resident and migratory immune system in steady state and memory: innate lymphocytes, dendritic cells and T cells. Nature immunology 14: 978-985. 77. den Haan, J. M., S. M. Lehar, and M. J. Bevan. 2000. CD8(+) but not CD8(-) dendritic cells cross-prime cytotoxic T cells in vivo. The Journal of experimental medicine 192: 1685-1696.

255

78. Pooley, J. L., W. R. Heath, and K. Shortman. 2001. Cutting edge: intravenous soluble antigen is presented to CD4 T cells by CD8- dendritic cells, but cross- presented to CD8 T cells by CD8+ dendritic cells. J. Immunol. 166: 5327-5330. 79. Hildner, K., B. T. Edelson, W. E. Purtha, M. Diamond, H. Matsushita, M. Kohyama, B. Calderon, B. U. Schraml, E. R. Unanue, M. S. Diamond, R. D. Schreiber, T. L. Murphy, and K. M. Murphy. 2008. Batf3 deficiency reveals a critical role for CD8alpha+ dendritic cells in cytotoxic T cell immunity. Science 322: 1097-1100. 80. Iyoda, T., S. Shimoyama, K. Liu, Y. Omatsu, Y. Akiyama, Y. Maeda, K. Takahara, R. M. Steinman, and K. Inaba. 2002. The CD8+ dendritic cell subset selectively endocytoses dying cells in culture and in vivo. The Journal of experimental medicine 195: 1289-1302. 81. Schulz, O., and C. Reis e Sousa. 2002. Cross-presentation of cell-associated antigens by CD8alpha+ dendritic cells is attributable to their ability to internalize dead cells. Immunology 107: 183-189. 82. Schnorrer, P., G. M. Behrens, N. S. Wilson, J. L. Pooley, C. M. Smith, D. El- Sukkari, G. Davey, F. Kupresanin, M. Li, E. Maraskovsky, G. T. Belz, F. R. Carbone, K. Shortman, W. R. Heath, and J. A. Villadangos. 2006. The dominant role of CD8+ dendritic cells in cross-presentation is not dictated by antigen capture. Proc. Natl. Acad. Sci. U. S. A. 103: 10729-10734. 83. Dudziak, D., A. O. Kamphorst, G. F. Heidkamp, V. R. Buchholz, C. Trumpfheller, S. Yamazaki, C. Cheong, K. Liu, H. W. Lee, C. G. Park, R. M. Steinman, and M. C. Nussenzweig. 2007. Differential antigen processing by dendritic cell subsets in vivo. Science 315: 107-111. 84. Savina, A., A. Peres, I. Cebrian, N. Carmo, C. Moita, N. Hacohen, L. F. Moita, and S. Amigorena. 2009. The small GTPase Rac2 controls phagosomal alkalinization and antigen crosspresentation selectively in CD8(+) dendritic cells. Immunity 30: 544-555. 85. Lin, M. L., Y. Zhan, A. I. Proietto, S. Prato, L. Wu, W. R. Heath, J. A. Villadangos, and A. M. Lew. 2008. Selective suicide of cross-presenting CD8+ dendritic cells by cytochrome c injection shows functional heterogeneity within this subset. Proc. Natl. Acad. Sci. U. S. A. 105: 3029-3034. 86. Hochrein, H., K. Shortman, D. Vremec, B. Scott, P. Hertzog, and M. O'Keeffe. 2001. Differential production of IL-12, IFN-alpha, and IFN-gamma by mouse dendritic cell subsets. J. Immunol. 166: 5448-5455. 87. Reis e Sousa, C., S. Hieny, T. Scharton-Kersten, D. Jankovic, H. Charest, R. N. Germain, and A. Sher. 1997. In vivo microbial stimulation induces rapid CD40 ligand-independent production of interleukin 12 by dendritic cells and their redistribution to T cell areas. The Journal of experimental medicine 186: 1819- 1829. 88. Maldonado-Lopez, R., T. De Smedt, P. Michel, J. Godfroid, B. Pajak, C. Heirman, K. Thielemans, O. Leo, J. Urbain, and M. Moser. 1999. CD8alpha+ and CD8alpha- subclasses of dendritic cells direct the development of distinct T helper cells in vivo. The Journal of experimental medicine 189: 587-592.

256

89. Dorner, B. G., M. B. Dorner, X. Zhou, C. Opitz, A. Mora, S. Guttler, A. Hutloff, H. W. Mages, K. Ranke, M. Schaefer, R. S. Jack, V. Henn, and R. A. Kroczek. 2009. Selective expression of the chemokine receptor XCR1 on cross- presenting dendritic cells determines cooperation with CD8+ T cells. Immunity 31: 823-833. 90. Crozat, K., R. Guiton, V. Contreras, V. Feuillet, C. A. Dutertre, E. Ventre, T. P. Vu Manh, T. Baranek, A. K. Storset, J. Marvel, P. Boudinot, A. Hosmalin, I. Schwartz-Cornil, and M. Dalod. 2010. The XC chemokine receptor 1 is a conserved selective marker of mammalian cells homologous to mouse CD8alpha+ dendritic cells. The Journal of experimental medicine 207: 1283- 1292. 91. Nierkens, S., J. Tel, E. Janssen, and G. J. Adema. 2013. Antigen cross- presentation by dendritic cell subsets: one general or all sergeants? Trends in immunology 34: 361-370. 92. Baker, K., S. W. Qiao, T. T. Kuo, V. G. Aveson, B. Platzer, J. T. Andersen, I. Sandlie, Z. Chen, C. de Haar, W. I. Lencer, E. Fiebiger, and R. S. Blumberg. 2011. Neonatal Fc receptor for IgG (FcRn) regulates cross-presentation of IgG immune complexes by CD8-CD11b+ dendritic cells. Proc. Natl. Acad. Sci. U. S. A. 108: 9927-9932. 93. Backer, R., F. van Leeuwen, G. Kraal, and J. M. den Haan. 2008. CD8- dendritic cells preferentially cross-present Saccharomyces cerevisiae antigens. Eur. J. Immunol. 38: 370-380. 94. Wilson, N. S., B. Yang, A. B. Morelli, S. Koernig, A. Yang, S. Loeser, D. Airey, L. Provan, P. Hass, H. Braley, S. Couto, D. Drane, J. Boyle, G. T. Belz, A. Ashkenazi, and E. Maraskovsky. 2012. ISCOMATRIX vaccines mediate CD8+ T-cell cross-priming by a MyD88-dependent signaling pathway. Immunol. Cell Biol. 90: 540-552. 95. Wilson, N. S., L. J. Young, F. Kupresanin, S. H. Naik, D. Vremec, W. R. Heath, S. Akira, K. Shortman, J. Boyle, E. Maraskovsky, G. T. Belz, and J. A. Villadangos. 2008. Normal proportion and expression of maturation markers in migratory dendritic cells in the absence of germs or Toll-like receptor signaling. Immunol. Cell Biol. 86: 200-205. 96. Randolph, G. J., J. Ochando, and S. Partida-Sanchez. 2008. Migration of dendritic cell subsets and their precursors. Annu. Rev. Immunol. 26: 293-316. 97. Allan, R. S., J. Waithman, S. Bedoui, C. M. Jones, J. A. Villadangos, Y. Zhan, A. M. Lew, K. Shortman, W. R. Heath, and F. R. Carbone. 2006. Migratory dendritic cells transfer antigen to a lymph node-resident dendritic cell population for efficient CTL priming. Immunity 25: 153-162. 98. Bedoui, S., P. G. Whitney, J. Waithman, L. Eidsmo, L. Wakim, I. Caminschi, R. S. Allan, M. Wojtasiak, K. Shortman, F. R. Carbone, A. G. Brooks, and W. R. Heath. 2009. Cross-presentation of viral and self antigens by skin-derived CD103+ dendritic cells. Nature immunology 10: 488-495.

257

99. Villadangos, J. A., and W. R. Heath. 2005. Life cycle, migration and antigen presenting functions of spleen and lymph node dendritic cells: limitations of the Langerhans cells paradigm. Semin. Immunol. 17: 262-272. 100. Igyarto, B. Z., K. Haley, D. Ortner, A. Bobr, M. Gerami-Nejad, B. T. Edelson, S. M. Zurawski, B. Malissen, G. Zurawski, J. Berman, and D. H. Kaplan. 2011. Skin-resident murine dendritic cell subsets promote distinct and opposing antigen-specific T helper cell responses. Immunity 35: 260-272. 101. Malissen, B., S. Tamoutounour, and S. Henri. 2014. The origins and functions of dendritic cells and macrophages in the skin. Nature reviews. Immunology 14: 417-428. 102. Ginhoux, F., K. Liu, J. Helft, M. Bogunovic, M. Greter, D. Hashimoto, J. Price, N. Yin, J. Bromberg, S. A. Lira, E. R. Stanley, M. Nussenzweig, and M. Merad. 2009. The origin and development of nonlymphoid tissue CD103+ DCs. The Journal of experimental medicine 206: 3115-3130. 103. Collin, M., N. McGovern, and M. Haniffa. 2013. Human dendritic cell subsets. Immunology 140: 22-30. 104. Segura, E. 2016. Review of Mouse and Human Dendritic Cell Subsets. Methods Mol. Biol. 1423: 3-15. 105. Robbins, S. H., T. Walzer, D. Dembele, C. Thibault, A. Defays, G. Bessou, H. Xu, E. Vivier, M. Sellars, P. Pierre, F. R. Sharp, S. Chan, P. Kastner, and M. Dalod. 2008. Novel insights into the relationships between dendritic cell subsets in human and mouse revealed by genome-wide expression profiling. Genome biology 9: R17. 106. Haniffa, M., A. Shin, V. Bigley, N. McGovern, P. Teo, P. See, P. S. Wasan, X. N. Wang, F. Malinarich, B. Malleret, A. Larbi, P. Tan, H. Zhao, M. Poidinger, S. Pagan, S. Cookson, R. Dickinson, I. Dimmick, R. F. Jarrett, L. Renia, J. Tam, C. Song, J. Connolly, J. K. Chan, A. Gehring, A. Bertoletti, M. Collin, and F. Ginhoux. 2012. Human tissues contain CD141hi cross-presenting dendritic cells with functional homology to mouse CD103+ nonlymphoid dendritic cells. Immunity 37: 60-73. 107. Schlitzer, A., N. McGovern, P. Teo, T. Zelante, K. Atarashi, D. Low, A. W. Ho, P. See, A. Shin, P. S. Wasan, G. Hoeffel, B. Malleret, A. Heiseke, S. Chew, L. Jardine, H. A. Purvis, C. M. Hilkens, J. Tam, M. Poidinger, E. R. Stanley, A. B. Krug, L. Renia, B. Sivasankar, L. G. Ng, M. Collin, P. Ricciardi-Castagnoli, K. Honda, M. Haniffa, and F. Ginhoux. 2013. IRF4 transcription factor- dependent CD11b+ dendritic cells in human and mouse control mucosal IL- 17 cytokine responses. Immunity 38: 970-983. 108. Bachem, A., S. Guttler, E. Hartung, F. Ebstein, M. Schaefer, A. Tannert, A. Salama, K. Movassaghi, C. Opitz, H. W. Mages, V. Henn, P. M. Kloetzel, S. Gurka, and R. A. Kroczek. 2010. Superior antigen cross-presentation and XCR1 expression define human CD11c+CD141+ cells as homologues of mouse CD8+ dendritic cells. The Journal of experimental medicine 207: 1273- 1281.

258

109. Jongbloed, S. L., A. J. Kassianos, K. J. McDonald, G. J. Clark, X. Ju, C. E. Angel, C. J. Chen, P. R. Dunbar, R. B. Wadley, V. Jeet, A. J. Vulink, D. N. Hart, and K. J. Radford. 2010. Human CD141+ (BDCA-3)+ dendritic cells (DCs) represent a unique myeloid DC subset that cross-presents necrotic cell antigens. The Journal of experimental medicine 207: 1247-1260. 110. Poulin, L. F., M. Salio, E. Griessinger, F. Anjos-Afonso, L. Craciun, J. L. Chen, A. M. Keller, O. Joffre, S. Zelenay, E. Nye, A. Le Moine, F. Faure, V. Donckier, D. Sancho, V. Cerundolo, D. Bonnet, and C. Reis e Sousa. 2010. Characterization of human DNGR-1+ BDCA3+ leukocytes as putative equivalents of mouse CD8alpha+ dendritic cells. The Journal of experimental medicine 207: 1261-1271. 111. Mittag, D., A. I. Proietto, T. Loudovaris, S. I. Mannering, D. Vremec, K. Shortman, L. Wu, and L. C. Harrison. 2011. Human dendritic cell subsets from spleen and blood are similar in phenotype and function but modified by donor health status. J. Immunol. 186: 6207-6217. 112. Segura, E., J. Valladeau-Guilemond, M. H. Donnadieu, X. Sastre-Garau, V. Soumelis, and S. Amigorena. 2012. Characterization of resident and migratory dendritic cells in human lymph nodes. The Journal of experimental medicine 209: 653-660. 113. Segura, E., M. Durand, and S. Amigorena. 2013. Similar antigen cross- presentation capacity and phagocytic functions in all freshly isolated human lymphoid organ-resident dendritic cells. The Journal of experimental medicine 210: 1035-1047. 114. Schiavoni, G., F. Mattei, P. Sestili, P. Borghi, M. Venditti, H. C. Morse, 3rd, F. Belardelli, and L. Gabriele. 2002. ICSBP is essential for the development of mouse type I interferon-producing cells and for the generation and activation of CD8alpha(+) dendritic cells. The Journal of experimental medicine 196: 1415-1425. 115. Hacker, C., R. D. Kirsch, X. S. Ju, T. Hieronymus, T. C. Gust, C. Kuhl, T. Jorgas, S. M. Kurz, S. Rose-John, Y. Yokota, and M. Zenke. 2003. Transcriptional profiling identifies Id2 function in dendritic cell development. Nature immunology 4: 380-386. 116. Suzuki, S., K. Honma, T. Matsuyama, K. Suzuki, K. Toriyama, I. Akitoyo, K. Yamamoto, T. Suematsu, M. Nakamura, K. Yui, and A. Kumatori. 2004. Critical roles of interferon regulatory factor 4 in CD11bhighCD8alpha- dendritic cell development. Proc. Natl. Acad. Sci. U. S. A. 101: 8981-8986. 117. Wu, L., A. D'Amico, K. D. Winkel, M. Suter, D. Lo, and K. Shortman. 1998. RelB is essential for the development of myeloid-related CD8alpha- dendritic cells but not of lymphoid-related CD8alpha+ dendritic cells. Immunity 9: 839- 847. 118. Guerriero, A., P. B. Langmuir, L. M. Spain, and E. W. Scott. 2000. PU.1 is required for myeloid-derived but not lymphoid-derived dendritic cells. Blood 95: 879-885.

259

119. Guilliams, M., F. Ginhoux, C. Jakubzick, S. H. Naik, N. Onai, B. U. Schraml, E. Segura, R. Tussiwand, and S. Yona. 2014. Dendritic cells, monocytes and macrophages: a unified nomenclature based on ontogeny. Nature reviews. Immunology 14: 571-578. 120. Caminschi, I., A. I. Proietto, F. Ahmet, S. Kitsoulis, J. Shin Teh, J. C. Lo, A. Rizzitelli, L. Wu, D. Vremec, S. L. van Dommelen, I. K. Campbell, E. Maraskovsky, H. Braley, G. M. Davey, P. Mottram, N. van de Velde, K. Jensen, A. M. Lew, M. D. Wright, W. R. Heath, K. Shortman, and M. H. Lahoud. 2008. The dendritic cell subtype-restricted C-type lectin Clec9A is a target for vaccine enhancement. Blood 112: 3264-3273. 121. Sancho, D., D. Mourao-Sa, O. P. Joffre, O. Schulz, N. C. Rogers, D. J. Pennington, J. R. Carlyle, and C. Reis e Sousa. 2008. Tumor therapy in mice via antigen targeting to a novel, DC-restricted C-type lectin. J. Clin. Invest. 118: 2098-2110. 122. Huysamen, C., J. A. Willment, K. M. Dennehy, and G. D. Brown. 2008. CLEC9A is a novel activation C-type lectin-like receptor expressed on BDCA3+ dendritic cells and a subset of monocytes. J. Biol. Chem. 283: 16693- 16701. 123. Bachem, A., E. Hartung, S. Guttler, A. Mora, X. Zhou, A. Hegemann, M. Plantinga, E. Mazzini, P. Stoitzner, S. Gurka, V. Henn, H. W. Mages, and R. A. Kroczek. 2012. Expression of XCR1 Characterizes the Batf3-Dependent Lineage of Dendritic Cells Capable of Antigen Cross-Presentation. Frontiers in immunology 3: 214. 124. Crozat, K., S. Tamoutounour, T. P. Vu Manh, E. Fossum, H. Luche, L. Ardouin, M. Guilliams, H. Azukizawa, B. Bogen, B. Malissen, S. Henri, and M. Dalod. 2011. Cutting edge: expression of XCR1 defines mouse lymphoid-tissue resident and migratory dendritic cells of the CD8alpha+ type. J. Immunol. 187: 4411-4415. 125. Swiecki, M., and M. Colonna. 2015. The multifaceted biology of plasmacytoid dendritic cells. Nature reviews. Immunology 15: 471-485. 126. Reizis, B., A. Bunin, H. S. Ghosh, K. L. Lewis, and V. Sisirak. 2011. Plasmacytoid dendritic cells: recent progress and open questions. Annu. Rev. Immunol. 29: 163-183. 127. Cisse, B., M. L. Caton, M. Lehner, T. Maeda, S. Scheu, R. Locksley, D. Holmberg, C. Zweier, N. S. den Hollander, S. G. Kant, W. Holter, A. Rauch, Y. Zhuang, and B. Reizis. 2008. Transcription factor E2-2 is an essential and specific regulator of plasmacytoid dendritic cell development. Cell 135: 37-48. 128. Villadangos, J. A., and L. Young. 2008. Antigen-presentation properties of plasmacytoid dendritic cells. Immunity 29: 352-361. 129. Smith, C. M., G. T. Belz, N. S. Wilson, J. A. Villadangos, K. Shortman, F. R. Carbone, and W. R. Heath. 2003. Cutting edge: conventional CD8 alpha+ dendritic cells are preferentially involved in CTL priming after footpad infection with herpes simplex virus-1. J. Immunol. 170: 4437-4440.

260

130. Belz, G. T., C. M. Smith, D. Eichner, K. Shortman, G. Karupiah, F. R. Carbone, and W. R. Heath. 2004. Cutting edge: conventional CD8 alpha+ dendritic cells are generally involved in priming CTL immunity to viruses. J. Immunol. 172: 1996-2000. 131. Belz, G. T., K. Shortman, M. J. Bevan, and W. R. Heath. 2005. CD8alpha+ dendritic cells selectively present MHC class I-restricted noncytolytic viral and intracellular bacterial antigens in vivo. J. Immunol. 175: 196-200. 132. Mouries, J., G. Moron, G. Schlecht, N. Escriou, G. Dadaglio, and C. Leclerc. 2008. Plasmacytoid dendritic cells efficiently cross-prime naive T cells in vivo after TLR activation. Blood 112: 3713-3722. 133. Hoeffel, G., A. C. Ripoche, D. Matheoud, M. Nascimbeni, N. Escriou, P. Lebon, F. Heshmati, J. G. Guillet, M. Gannage, S. Caillat-Zucman, N. Casartelli, O. Schwartz, H. De la Salle, D. Hanau, A. Hosmalin, and C. Maranon. 2007. Antigen crosspresentation by human plasmacytoid dendritic cells. Immunity 27: 481-492. 134. Di Pucchio, T., B. Chatterjee, A. Smed-Sorensen, S. Clayton, A. Palazzo, M. Montes, Y. Xue, I. Mellman, J. Banchereau, and J. E. Connolly. 2008. Direct proteasome-independent cross-presentation of viral antigen by plasmacytoid dendritic cells on major histocompatibility complex class I. Nature immunology 9: 551-557. 135. Lui, G., O. Manches, J. Angel, J. P. Molens, L. Chaperot, and J. Plumas. 2009. Plasmacytoid dendritic cells capture and cross-present viral antigens from influenza-virus exposed cells. PloS one 4: e7111. 136. Tel, J., G. Schreibelt, S. P. Sittig, T. S. Mathan, S. I. Buschow, L. J. Cruz, A. J. Lambeck, C. G. Figdor, and I. J. de Vries. 2013. Human plasmacytoid dendritic cells efficiently cross-present exogenous Ags to CD8+ T cells despite lower Ag uptake than myeloid dendritic cell subsets. Blood 121: 459-467. 137. Romani, N., S. Gruner, D. Brang, E. Kampgen, A. Lenz, B. Trockenbacher, G. Konwalinka, P. O. Fritsch, R. M. Steinman, and G. Schuler. 1994. Proliferating dendritic cell progenitors in human blood. The Journal of experimental medicine 180: 83-93. 138. Sallusto, F., and A. Lanzavecchia. 1994. Efficient presentation of soluble antigen by cultured human dendritic cells is maintained by granulocyte/macrophage colony-stimulating factor plus interleukin 4 and downregulated by tumor necrosis factor alpha. The Journal of experimental medicine 179: 1109-1118. 139. Segura, E., and S. Amigorena. 2013. Inflammatory dendritic cells in mice and humans. Trends in immunology 34: 440-445. 140. Hespel, C., and M. Moser. 2012. Role of inflammatory dendritic cells in innate and adaptive immunity. Eur. J. Immunol. 42: 2535-2543. 141. Cheong, C., I. Matos, J. H. Choi, D. B. Dandamudi, E. Shrestha, M. P. Longhi, K. L. Jeffrey, R. M. Anthony, C. Kluger, G. Nchinda, H. Koh, A. Rodriguez, J. Idoyaga, M. Pack, K. Velinzon, C. G. Park, and R. M. Steinman. 2010.

261

Microbial stimulation fully differentiates monocytes to DC-SIGN/CD209(+) dendritic cells for immune T cell areas. Cell 143: 416-429. 142. Gutierrez-Martinez, E., R. Planes, G. Anselmi, M. Reynolds, S. Menezes, A. C. Adiko, L. Saveanu, and P. Guermonprez. 2015. Cross-Presentation of Cell- Associated Antigens by MHC Class I in Dendritic Cell Subsets. Frontiers in immunology 6: 363. 143. Medzhitov, R. 2001. Toll-like receptors and innate immunity. Nature reviews. Immunology 1: 135-145. 144. Lemaitre, B., E. Nicolas, L. Michaut, J. M. Reichhart, and J. A. Hoffmann. 1996. The dorsoventral regulatory gene cassette spatzle/Toll/cactus controls the potent antifungal response in Drosophila adults. Cell 86: 973-983. 145. Medzhitov, R., P. Preston-Hurlburt, and C. A. Janeway, Jr. 1997. A human homologue of the Drosophila Toll protein signals activation of adaptive immunity. Nature 388: 394-397. 146. Kawai, T., and S. Akira. 2010. The role of pattern-recognition receptors in innate immunity: update on Toll-like receptors. Nature immunology 11: 373- 384. 147. Akira, S., S. Uematsu, and O. Takeuchi. 2006. Pathogen recognition and innate immunity. Cell 124: 783-801. 148. Guan, Y., D. R. Ranoa, S. Jiang, S. K. Mutha, X. Li, J. Baudry, and R. I. Tapping. 2010. Human TLRs 10 and 1 share common mechanisms of innate immune sensing but not signaling. J. Immunol. 184: 5094-5103. 149. Yarovinsky, F., D. Zhang, J. F. Andersen, G. L. Bannenberg, C. N. Serhan, M. S. Hayden, S. Hieny, F. S. Sutterwala, R. A. Flavell, S. Ghosh, and A. Sher. 2005. TLR11 activation of dendritic cells by a protozoan profilin-like protein. Science 308: 1626-1629. 150. Koblansky, A. A., D. Jankovic, H. Oh, S. Hieny, W. Sungnak, R. Mathur, M. S. Hayden, S. Akira, A. Sher, and S. Ghosh. 2013. Recognition of profilin by Toll- like receptor 12 is critical for host resistance to Toxoplasma gondii. Immunity 38: 119-130. 151. Oldenburg, M., A. Kruger, R. Ferstl, A. Kaufmann, G. Nees, A. Sigmund, B. Bathke, H. Lauterbach, M. Suter, S. Dreher, U. Koedel, S. Akira, T. Kawai, J. Buer, H. Wagner, S. Bauer, H. Hochrein, and C. J. Kirschning. 2012. TLR13 recognizes bacterial 23S rRNA devoid of erythromycin resistance-forming modification. Science 337: 1111-1115. 152. Kumar, H., T. Kawai, and S. Akira. 2011. Pathogen recognition by the innate immune system. Int. Rev. Immunol. 30: 16-34. 153. Kumagai, Y., and S. Akira. 2010. Identification and functions of pattern- recognition receptors. J. Allergy Clin. Immunol. 125: 985-992. 154. Rathinam, V. A., S. K. Vanaja, and K. A. Fitzgerald. 2012. Regulation of inflammasome signaling. Nature immunology 13: 333-342.

262

155. Burdette, D. L., K. M. Monroe, K. Sotelo-Troha, J. S. Iwig, B. Eckert, M. Hyodo, Y. Hayakawa, and R. E. Vance. 2011. STING is a direct innate immune sensor of cyclic di-GMP. Nature 478: 515-518. 156. Tanaka, Y., and Z. J. Chen. 2012. STING specifies IRF3 phosphorylation by TBK1 in the cytosolic DNA signaling pathway. Science signaling 5: ra20. 157. Wu, J., and Z. J. Chen. 2014. Innate immune sensing and signaling of cytosolic nucleic acids. Annu. Rev. Immunol. 32: 461-488. 158. Bhat, N., and K. A. Fitzgerald. 2014. Recognition of cytosolic DNA by cGAS and other STING-dependent sensors. Eur. J. Immunol. 44: 634-640. 159. Zelensky, A. N., and J. E. Gready. 2005. The C-type lectin-like domain superfamily. The FEBS journal 272: 6179-6217. 160. Sancho, D., and C. Reis e Sousa. 2012. Signaling by myeloid C-type lectin receptors in immunity and homeostasis. Annu. Rev. Immunol. 30: 491-529. 161. Hoving, J. C., G. J. Wilson, and G. D. Brown. 2014. Signalling C-type lectin receptors, microbial recognition and immunity. Cellular microbiology 16: 185- 194. 162. Sancho, D., and C. Reis e Sousa. 2013. Sensing of cell death by myeloid C-type lectin receptors. Curr. Opin. Immunol. 25: 46-52. 163. Osorio, F., and C. Reis e Sousa. 2011. Myeloid C-type lectin receptors in pathogen recognition and host defense. Immunity 34: 651-664. 164. Geijtenbeek, T. B., and S. I. Gringhuis. 2009. Signalling through C-type lectin receptors: shaping immune responses. Nature reviews. Immunology 9: 465-479. 165. Brown, G. D. 2006. Dectin-1: a signalling non-TLR pattern-recognition receptor. Nature reviews. Immunology 6: 33-43. 166. Hardison, S. E., and G. D. Brown. 2012. C-type lectin receptors orchestrate antifungal immunity. Nature immunology 13: 817-822. 167. Dambuza, I. M., and G. D. Brown. 2015. C-type lectins in immunity: recent developments. Curr. Opin. Immunol. 32: 21-27. 168. Yamasaki, S., E. Ishikawa, M. Sakuma, H. Hara, K. Ogata, and T. Saito. 2008. Mincle is an ITAM-coupled activating receptor that senses damaged cells. Nature immunology 9: 1179-1188. 169. Ahrens, S., S. Zelenay, D. Sancho, P. Hanc, S. Kjaer, C. Feest, G. Fletcher, C. Durkin, A. Postigo, M. Skehel, F. Batista, B. Thompson, M. Way, C. Reis e Sousa, and O. Schulz. 2012. F-actin is an evolutionarily conserved damage- associated molecular pattern recognized by DNGR-1, a receptor for dead cells. Immunity 36: 635-645. 170. Zhang, J. G., P. E. Czabotar, A. N. Policheni, I. Caminschi, S. S. Wan, S. Kitsoulis, K. M. Tullett, A. Y. Robin, R. Brammananth, M. F. van Delft, J. Lu, L. A. O'Reilly, E. C. Josefsson, B. T. Kile, W. J. Chin, J. D. Mintern, M. A. Olshina, W. Wong, J. Baum, M. D. Wright, D. C. Huang, N. Mohandas, R. L. Coppel, P. M. Colman, N. A. Nicola, K. Shortman, and M. H. Lahoud. 2012.

263

The dendritic cell receptor Clec9A binds damaged cells via exposed actin filaments. Immunity 36: 646-657. 171. Sancho, D., O. P. Joffre, A. M. Keller, N. C. Rogers, D. Martinez, P. Hernanz- Falcon, I. Rosewell, and C. Reis e Sousa. 2009. Identification of a dendritic cell receptor that couples sensing of necrosis to immunity. Nature 458: 899-903. 172. Zelenay, S., A. M. Keller, P. G. Whitney, B. U. Schraml, S. Deddouche, N. C. Rogers, O. Schulz, D. Sancho, and C. Reis e Sousa. 2012. The dendritic cell receptor DNGR-1 controls endocytic handling of necrotic cell antigens to favor cross-priming of CTLs in virus-infected mice. J. Clin. Invest. 122: 1615- 1627. 173. Iborra, S., H. M. Izquierdo, M. Martinez-Lopez, N. Blanco-Menendez, C. Reis e Sousa, and D. Sancho. 2012. The DC receptor DNGR-1 mediates cross- priming of CTLs during vaccinia virus infection in mice. J. Clin. Invest. 122: 1628-1643. 174. Cao, L., X. Shi, H. Chang, Q. Zhang, and Y. He. 2015. pH-Dependent recognition of apoptotic and necrotic cells by the human dendritic cell receptor DEC205. Proc. Natl. Acad. Sci. U. S. A. 112: 7237-7242. 175. Cao, L., H. Chang, X. Shi, C. Peng, and Y. He. 2016. Keratin mediates the recognition of apoptotic and necrotic cells through dendritic cell receptor DEC205/CD205. Proc. Natl. Acad. Sci. U. S. A. 113: 13438-13443. 176. Neumann, K., M. Castineiras-Vilarino, U. Hockendorf, N. Hannesschlager, S. Lemeer, D. Kupka, S. Meyermann, M. Lech, H. J. Anders, B. Kuster, D. H. Busch, A. Gewies, R. Naumann, O. Gross, and J. Ruland. 2014. Clec12a is an inhibitory receptor for uric acid crystals that regulates inflammation in response to cell death. Immunity 40: 389-399. 177. van Kooyk, Y. 2008. C-type lectins on dendritic cells: key modulators for the induction of immune responses. Biochem. Soc. Trans. 36: 1478-1481. 178. van Kooyk, Y., and G. A. Rabinovich. 2008. Protein-glycan interactions in the control of innate and adaptive immune responses. Nature immunology 9: 593- 601. 179. van Gisbergen, K. P., C. A. Aarnoudse, G. A. Meijer, T. B. Geijtenbeek, and Y. van Kooyk. 2005. Dendritic cells recognize tumor-specific glycosylation of carcinoembryonic antigen on colorectal cancer cells through dendritic cell- specific intercellular adhesion molecule-3-grabbing nonintegrin. Cancer Res. 65: 5935-5944. 180. Yan, H., T. Kamiya, P. Suabjakyong, and N. M. Tsuji. 2015. Targeting C-Type Lectin Receptors for Cancer Immunity. Frontiers in immunology 6: 408. 181. Edwards, A. D., S. S. Diebold, E. M. Slack, H. Tomizawa, H. Hemmi, T. Kaisho, S. Akira, and C. Reis e Sousa. 2003. Toll-like receptor expression in murine DC subsets: lack of TLR7 expression by CD8 alpha+ DC correlates with unresponsiveness to imidazoquinolines. Eur. J. Immunol. 33: 827-833. 182. Bohnenkamp, H. R., K. T. Papazisis, J. M. Burchell, and J. Taylor- Papadimitriou. 2007. Synergism of Toll-like receptor-induced interleukin-

264

12p70 secretion by monocyte-derived dendritic cells is mediated through p38 MAPK and lowers the threshold of T-helper cell type 1 responses. Cell. Immunol. 247: 72-84. 183. Longhi, M. P., C. Trumpfheller, J. Idoyaga, M. Caskey, I. Matos, C. Kluger, A. M. Salazar, M. Colonna, and R. M. Steinman. 2009. Dendritic cells require a systemic type I interferon response to mature and induce CD4+ Th1 immunity with poly IC as adjuvant. The Journal of experimental medicine 206: 1589-1602. 184. Schulz, O., S. S. Diebold, M. Chen, T. I. Naslund, M. A. Nolte, L. Alexopoulou, Y. T. Azuma, R. A. Flavell, P. Liljestrom, and C. Reis e Sousa. 2005. Toll-like receptor 3 promotes cross-priming to virus-infected cells. Nature 433: 887- 892. 185. Dalod, M., R. Chelbi, B. Malissen, and T. Lawrence. 2014. Dendritic cell maturation: functional specialization through signaling specificity and transcriptional programming. EMBO J. 33: 1104-1116. 186. Hornung, V., S. Rothenfusser, S. Britsch, A. Krug, B. Jahrsdorfer, T. Giese, S. Endres, and G. Hartmann. 2002. Quantitative expression of toll-like receptor 1-10 mRNA in cellular subsets of human peripheral blood mononuclear cells and sensitivity to CpG oligodeoxynucleotides. J. Immunol. 168: 4531-4537. 187. Schreibelt, G., J. Tel, K. H. Sliepen, D. Benitez-Ribas, C. G. Figdor, G. J. Adema, and I. J. de Vries. 2010. Toll-like receptor expression and function in human dendritic cell subsets: implications for dendritic cell-based anti-cancer immunotherapy. Cancer Immunol. Immunother. 59: 1573-1582. 188. Iwasaki, A., and R. Medzhitov. 2010. Regulation of adaptive immunity by the innate immune system. Science 327: 291-295. 189. Lesterhuis, W. J., I. J. de Vries, G. J. Adema, and C. J. Punt. 2004. Dendritic cell-based vaccines in cancer immunotherapy: an update on clinical and immunological results. Ann. Oncol. 15 Suppl 4: iv145-151. 190. Lesterhuis, W. J., E. H. Aarntzen, I. J. De Vries, D. H. Schuurhuis, C. G. Figdor, G. J. Adema, and C. J. Punt. 2008. Dendritic cell vaccines in melanoma: from promise to proof? Crit. Rev. Oncol. Hematol. 66: 118-134. 191. Anguille, S., E. L. Smits, E. Lion, V. F. van Tendeloo, and Z. N. Berneman. 2014. Clinical use of dendritic cells for cancer therapy. The Lancet. Oncology 15: e257-267. 192. Small, E. J., P. F. Schellhammer, C. S. Higano, C. H. Redfern, J. J. Nemunaitis, F. H. Valone, S. S. Verjee, L. A. Jones, and R. M. Hershberg. 2006. Placebo- controlled phase III trial of immunologic therapy with sipuleucel-T (APC8015) in patients with metastatic, asymptomatic hormone refractory prostate cancer. J. Clin. Oncol. 24: 3089-3094. 193. Higano, C. S., P. F. Schellhammer, E. J. Small, P. A. Burch, J. Nemunaitis, L. Yuh, N. Provost, and M. W. Frohlich. 2009. Integrated data from 2 randomized, double-blind, placebo-controlled, phase 3 trials of active cellular

265

immunotherapy with sipuleucel-T in advanced prostate cancer. Cancer 115: 3670-3679. 194. Kantoff, P. W., C. S. Higano, N. D. Shore, E. R. Berger, E. J. Small, D. F. Penson, C. H. Redfern, A. C. Ferrari, R. Dreicer, R. B. Sims, Y. Xu, M. W. Frohlich, P. F. Schellhammer, and I. S. Investigators. 2010. Sipuleucel-T immunotherapy for castration-resistant prostate cancer. N. Engl. J. Med. 363: 411-422. 195. Beer, T. M., G. T. Bernstein, J. M. Corman, L. M. Glode, S. J. Hall, W. L. Poll, P. F. Schellhammer, L. A. Jones, Y. Xu, J. W. Kylstra, and M. W. Frohlich. 2011. Randomized trial of autologous cellular immunotherapy with sipuleucel- T in androgen-dependent prostate cancer. Clin. Cancer Res. 17: 4558-4567. 196. Bol, K. F., G. Schreibelt, W. R. Gerritsen, I. J. de Vries, and C. G. Figdor. 2016. Dendritic Cell-Based Immunotherapy: State of the Art and Beyond. Clin. Cancer Res. 22: 1897-1906. 197. Tacken, P. J., I. J. de Vries, R. Torensma, and C. G. Figdor. 2007. Dendritic-cell immunotherapy: from ex vivo loading to in vivo targeting. Nature reviews. Immunology 7: 790-802. 198. De Vries, I. J., D. J. Krooshoop, N. M. Scharenborg, W. J. Lesterhuis, J. H. Diepstra, G. N. Van Muijen, S. P. Strijk, T. J. Ruers, O. C. Boerman, W. J. Oyen, G. J. Adema, C. J. Punt, and C. G. Figdor. 2003. Effective migration of antigen-pulsed dendritic cells to lymph nodes in melanoma patients is determined by their maturation state. Cancer Res. 63: 12-17. 199. Bonifaz, L. C., D. P. Bonnyay, A. Charalambous, D. I. Darguste, S. Fujii, H. Soares, M. K. Brimnes, B. Moltedo, T. M. Moran, and R. M. Steinman. 2004. In vivo targeting of antigens to maturing dendritic cells via the DEC-205 receptor improves T cell vaccination. The Journal of experimental medicine 199: 815-824. 200. Carayanniotis, G., and B. H. Barber. 1987. Adjuvant-free IgG responses induced with antigen coupled to antibodies against class II MHC. Nature 327: 59-61. 201. Dickgreber, N., P. Stoitzner, Y. Bai, K. M. Price, K. J. Farrand, K. Manning, C. E. Angel, P. R. Dunbar, F. Ronchese, J. D. Fraser, B. T. Backstrom, and I. F. Hermans. 2009. Targeting antigen to MHC class II molecules promotes efficient cross-presentation and enhances immunotherapy. J. Immunol. 182: 1260-1269. 202. Wang, H., M. N. Griffiths, D. R. Burton, and P. Ghazal. 2000. Rapid antibody responses by low-dose, single-step, dendritic cell-targeted immunization. Proc. Natl. Acad. Sci. U. S. A. 97: 847-852. 203. van Broekhoven, C. L., C. R. Parish, C. Demangel, W. J. Britton, and J. G. Altin. 2004. Targeting dendritic cells with antigen-containing liposomes: a highly effective procedure for induction of antitumor immunity and for tumor immunotherapy. Cancer Res. 64: 4357-4365.

266

204. Wei, H., S. Wang, D. Zhang, S. Hou, W. Qian, B. Li, H. Guo, G. Kou, J. He, H. Wang, and Y. Guo. 2009. Targeted delivery of tumor antigens to activated dendritic cells via CD11c molecules induces potent antitumor immunity in mice. Clin. Cancer Res. 15: 4612-4621. 205. Ramakrishna, V., J. F. Treml, L. Vitale, J. E. Connolly, T. O'Neill, P. A. Smith, C. L. Jones, L. Z. He, J. Goldstein, P. K. Wallace, T. Keler, and M. J. Endres. 2004. Mannose receptor targeting of tumor antigen pmel17 to human dendritic cells directs anti-melanoma T cell responses via multiple HLA molecules. J. Immunol. 172: 2845-2852. 206. He, L. Z., A. Crocker, J. Lee, J. Mendoza-Ramirez, X. T. Wang, L. A. Vitale, T. O'Neill, C. Petromilli, H. F. Zhang, J. Lopez, D. Rohrer, T. Keler, and R. Clynes. 2007. Antigenic targeting of the human mannose receptor induces tumor immunity. J. Immunol. 178: 6259-6267. 207. Tagliani, E., P. Guermonprez, J. Sepulveda, M. Lopez-Bravo, C. Ardavin, S. Amigorena, F. Benvenuti, and O. R. Burrone. 2008. Selection of an antibody library identifies a pathway to induce immunity by targeting CD36 on steady- state CD8 alpha+ dendritic cells. J. Immunol. 180: 3201-3209. 208. Delneste, Y., G. Magistrelli, J. Gauchat, J. Haeuw, J. Aubry, K. Nakamura, N. Kawakami-Honda, L. Goetsch, T. Sawamura, J. Bonnefoy, and P. Jeannin. 2002. Involvement of LOX-1 in dendritic cell-mediated antigen cross- presentation. Immunity 17: 353-362. 209. Bennett, S. R., F. R. Carbone, T. Toy, J. F. Miller, and W. R. Heath. 1998. B cells directly tolerize CD8(+) T cells. The Journal of experimental medicine 188: 1977-1983. 210. Nchinda, G., J. Kuroiwa, M. Oks, C. Trumpfheller, C. G. Park, Y. Huang, D. Hannaman, S. J. Schlesinger, O. Mizenina, M. C. Nussenzweig, K. Uberla, and R. M. Steinman. 2008. The efficacy of DNA vaccination is enhanced in mice by targeting the encoded protein to dendritic cells. J. Clin. Invest. 118: 1427-1436. 211. Lahoud, M. H., A. I. Proietto, F. Ahmet, S. Kitsoulis, L. Eidsmo, L. Wu, P. Sathe, S. Pietersz, H. W. Chang, I. D. Walker, E. Maraskovsky, H. Braley, A. M. Lew, M. D. Wright, W. R. Heath, K. Shortman, and I. Caminschi. 2009. The C- type lectin Clec12A present on mouse and human dendritic cells can serve as a target for antigen delivery and enhancement of antibody responses. J. Immunol. 182: 7587-7594. 212. Carter, R. W., C. Thompson, D. M. Reid, S. Y. Wong, and D. F. Tough. 2006. Preferential induction of CD4+ T cell responses through in vivo targeting of antigen to dendritic cell-associated C-type lectin-1. J. Immunol. 177: 2276- 2284. 213. Garcia-Vallejo, J. J., and Y. van Kooyk. 2013. The physiological role of DC- SIGN: a tale of mice and men. Trends in immunology 34: 482-486. 214. Corbett, A. J., I. Caminschi, B. S. McKenzie, J. L. Brady, M. D. Wright, P. L. Mottram, P. M. Hogarth, A. N. Hodder, Y. Zhan, D. M. Tarlinton, K. Shortman, and A. M. Lew. 2005. Antigen delivery via two molecules on the

267

CD8- dendritic cell subset induces humoral immunity in the absence of conventional "danger". Eur. J. Immunol. 35: 2815-2825. 215. Lahoud, M. H., F. Ahmet, S. Kitsoulis, S. S. Wan, D. Vremec, C. N. Lee, B. Phipson, W. Shi, G. K. Smyth, A. M. Lew, Y. Kato, S. N. Mueller, G. M. Davey, W. R. Heath, K. Shortman, and I. Caminschi. 2011. Targeting antigen to mouse dendritic cells via Clec9A induces potent CD4 T cell responses biased toward a follicular helper phenotype. J. Immunol. 187: 842-850. 216. Chatterjee, B., A. Smed-Sorensen, L. Cohn, C. Chalouni, R. Vandlen, B. C. Lee, J. Widger, T. Keler, L. Delamarre, and I. Mellman. 2012. Internalization and endosomal degradation of receptor-bound antigens regulate the efficiency of cross presentation by human dendritic cells. Blood 120: 2011-2020. 217. Reuter, A., S. E. Panozza, C. Macri, C. Dumont, J. Li, H. Liu, E. Segura, J. Vega-Ramos, N. Gupta, I. Caminschi, J. A. Villadangos, A. P. Johnston, and J. D. Mintern. 2015. Criteria for dendritic cell receptor selection for efficient antibody-targeted vaccination. J. Immunol. 194: 2696-2705. 218. Burgdorf, S., A. Kautz, V. Bohnert, P. A. Knolle, and C. Kurts. 2007. Distinct pathways of antigen uptake and intracellular routing in CD4 and CD8 T cell activation. Science 316: 612-616. 219. Cohn, L., B. Chatterjee, F. Esselborn, A. Smed-Sorensen, N. Nakamura, C. Chalouni, B. C. Lee, R. Vandlen, T. Keler, P. Lauer, D. Brockstedt, I. Mellman, and L. Delamarre. 2013. Antigen delivery to early endosomes eliminates the superiority of human blood BDCA3+ dendritic cells at cross presentation. The Journal of experimental medicine 210: 1049-1063. 220. Bonifaz, L., D. Bonnyay, K. Mahnke, M. Rivera, M. C. Nussenzweig, and R. M. Steinman. 2002. Efficient targeting of protein antigen to the dendritic cell receptor DEC-205 in the steady state leads to antigen presentation on major histocompatibility complex class I products and peripheral CD8+ T cell tolerance. The Journal of experimental medicine 196: 1627-1638. 221. Joffre, O. P., D. Sancho, S. Zelenay, A. M. Keller, and C. Reis e Sousa. 2010. Efficient and versatile manipulation of the peripheral CD4+ T-cell compartment by antigen targeting to DNGR-1/CLEC9A. Eur. J. Immunol. 40: 1255-1265. 222. Li, J., F. Ahmet, L. C. Sullivan, A. G. Brooks, S. J. Kent, R. De Rose, A. M. Salazar, C. Reis e Sousa, K. Shortman, M. H. Lahoud, W. R. Heath, and I. Caminschi. 2015. Antibodies targeting Clec9A promote strong humoral immunity without adjuvant in mice and non-human primates. Eur. J. Immunol. 45: 854-864. 223. Hodges, A., K. Sharrocks, M. Edelmann, D. Baban, A. Moris, O. Schwartz, H. Drakesmith, K. Davies, B. Kessler, A. McMichael, and A. Simmons. 2007. Activation of the lectin DC-SIGN induces an immature dendritic cell phenotype triggering Rho-GTPase activity required for HIV-1 replication. Nature immunology 8: 569-577. 224. Tacken, P. J., W. Ginter, L. Berod, L. J. Cruz, B. Joosten, T. Sparwasser, C. G. Figdor, and A. Cambi. 2011. Targeting DC-SIGN via its neck region leads to

268

prolonged antigen residence in early endosomes, delayed lysosomal degradation, and cross-presentation. Blood 118: 4111-4119. 225. Jiang, W., W. J. Swiggard, C. Heufler, M. Peng, A. Mirza, R. M. Steinman, and M. C. Nussenzweig. 1995. The receptor DEC-205 expressed by dendritic cells and thymic epithelial cells is involved in antigen processing. Nature 375: 151- 155. 226. Mahnke, K., M. Guo, S. Lee, H. Sepulveda, S. L. Swain, M. Nussenzweig, and R. M. Steinman. 2000. The dendritic cell receptor for endocytosis, DEC-205, can recycle and enhance antigen presentation via major histocompatibility complex class II-positive lysosomal compartments. J. Cell Biol. 151: 673-684. 227. Hawiger, D., K. Inaba, Y. Dorsett, M. Guo, K. Mahnke, M. Rivera, J. V. Ravetch, R. M. Steinman, and M. C. Nussenzweig. 2001. Dendritic cells induce peripheral T cell unresponsiveness under steady state conditions in vivo. The Journal of experimental medicine 194: 769-779. 228. Mahnke, K., Y. Qian, S. Fondel, J. Brueck, C. Becker, and A. H. Enk. 2005. Targeting of antigens to activated dendritic cells in vivo cures metastatic melanoma in mice. Cancer Res. 65: 7007-7012. 229. Wang, B., N. Zaidi, L. Z. He, L. Zhang, J. M. Kuroiwa, T. Keler, and R. M. Steinman. 2012. Targeting of the non-mutated tumor antigen HER2/neu to mature dendritic cells induces an integrated immune response that protects against breast cancer in mice. Breast cancer research : BCR 14: R39. 230. Trumpfheller, C., J. S. Finke, C. B. Lopez, T. M. Moran, B. Moltedo, H. Soares, Y. Huang, S. J. Schlesinger, C. G. Park, M. C. Nussenzweig, A. Granelli- Piperno, and R. M. Steinman. 2006. Intensified and protective CD4+ T cell immunity in mice with anti-dendritic cell HIV gag fusion antibody vaccine. The Journal of experimental medicine 203: 607-617. 231. Bozzacco, L., C. Trumpfheller, F. P. Siegal, S. Mehandru, M. Markowitz, M. Carrington, M. C. Nussenzweig, A. G. Piperno, and R. M. Steinman. 2007. DEC-205 receptor on dendritic cells mediates presentation of HIV gag protein to CD8+ T cells in a spectrum of human MHC I haplotypes. Proc. Natl. Acad. Sci. U. S. A. 104: 1289-1294. 232. Cheong, C., J. H. Choi, L. Vitale, L. Z. He, C. Trumpfheller, L. Bozzacco, Y. Do, G. Nchinda, S. H. Park, D. B. Dandamudi, E. Shrestha, M. Pack, H. W. Lee, T. Keler, R. M. Steinman, and C. G. Park. 2010. Improved cellular and humoral immune responses in vivo following targeting of HIV Gag to dendritic cells within human anti-human DEC205 monoclonal antibody. Blood 116: 3828-3838. 233. Flynn, B. J., K. Kastenmuller, U. Wille-Reece, G. D. Tomaras, M. Alam, R. W. Lindsay, A. M. Salazar, B. Perdiguero, C. E. Gomez, R. Wagner, M. Esteban, C. G. Park, C. Trumpfheller, T. Keler, G. Pantaleo, R. M. Steinman, and R. Seder. 2011. Immunization with HIV Gag targeted to dendritic cells followed by recombinant New York vaccinia virus induces robust T-cell immunity in nonhuman primates. Proc. Natl. Acad. Sci. U. S. A. 108: 7131-7136.

269

234. Idoyaga, J., A. Lubkin, C. Fiorese, M. H. Lahoud, I. Caminschi, Y. Huang, A. Rodriguez, B. E. Clausen, C. G. Park, C. Trumpfheller, and R. M. Steinman. 2011. Comparable T helper 1 (Th1) and CD8 T-cell immunity by targeting HIV gag p24 to CD8 dendritic cells within antibodies to Langerin, DEC205, and Clec9A. Proc. Natl. Acad. Sci. U. S. A. 108: 2384-2389. 235. Dhodapkar, M. V., M. Sznol, B. Zhao, D. Wang, R. D. Carvajal, M. L. Keohan, E. Chuang, R. E. Sanborn, J. Lutzky, J. Powderly, H. Kluger, S. Tejwani, J. Green, V. Ramakrishna, A. Crocker, L. Vitale, M. Yellin, T. Davis, and T. Keler. 2014. Induction of antigen-specific immunity with a vaccine targeting NY-ESO-1 to the dendritic cell receptor DEC-205. Science translational medicine 6: 232ra251. 236. Caminschi, I., and K. Shortman. 2012. Boosting antibody responses by targeting antigens to dendritic cells. Trends in immunology 33: 71-77. 237. Picco, G., R. Beatson, J. Taylor-Papadimitriou, and J. M. Burchell. 2014. Targeting DNGR-1 (CLEC9A) with antibody/MUC1 peptide conjugates as a vaccine for carcinomas. Eur. J. Immunol. 44: 1947-1955. 238. Obst, R., H. M. van Santen, R. Melamed, A. O. Kamphorst, C. Benoist, and D. Mathis. 2007. Sustained antigen presentation can promote an immunogenic T cell response, like dendritic cell activation. Proc. Natl. Acad. Sci. U. S. A. 104: 15460-15465. 239. Tullett, K. M., I. M. Leal Rojas, Y. Minoda, P. S. Tan, J. G. Zhang, C. Smith, R. Khanna, K. Shortman, I. Caminschi, M. H. Lahoud, and K. J. Radford. 2016. Targeting CLEC9A delivers antigen to human CD141+ DC for CD4+ and CD8+T cell recognition. JCI insight 1: e87102. 240. Fernandez-Ruiz, D., W. Y. Ng, L. E. Holz, J. Z. Ma, A. Zaid, Y. C. Wong, L. S. Lau, V. Mollard, A. Cozijnsen, N. Collins, J. Li, G. M. Davey, Y. Kato, S. Devi, R. Skandari, M. Pauley, J. H. Manton, D. I. Godfrey, A. Braun, S. S. Tay, P. S. Tan, D. G. Bowen, F. Koch-Nolte, B. Rissiek, F. R. Carbone, B. S. Crabb, M. Lahoud, I. A. Cockburn, S. N. Mueller, P. Bertolino, G. I. McFadden, I. Caminschi, and W. R. Heath. 2016. Liver-Resident Memory CD8+ T Cells Form a Front-Line Defense against Malaria Liver-Stage Infection. Immunity 45: 889-902. 241. Schreibelt, G., L. J. Klinkenberg, L. J. Cruz, P. J. Tacken, J. Tel, M. Kreutz, G. J. Adema, G. D. Brown, C. G. Figdor, and I. J. de Vries. 2012. The C-type lectin receptor CLEC9A mediates antigen uptake and (cross-)presentation by human blood BDCA3+ myeloid dendritic cells. Blood 119: 2284-2292. 242. Yamazaki, C., R. Miyamoto, K. Hoshino, Y. Fukuda, I. Sasaki, M. Saito, H. Ishiguchi, T. Yano, T. Sugiyama, H. Hemmi, T. Tanaka, E. Hamada, T. Hirashima, R. Amakawa, S. Fukuhara, S. Nomura, T. Ito, and T. Kaisho. 2010. Conservation of a chemokine system, XCR1 and its ligand, XCL1, between human and mice. Biochem. Biophys. Res. Commun. 397: 756-761. 243. Hartung, E., M. Becker, A. Bachem, N. Reeg, A. Jakel, A. Hutloff, H. Weber, C. Weise, C. Giesecke, V. Henn, S. Gurka, K. Anastassiadis, H. W. Mages, and R. A. Kroczek. 2015. Induction of potent CD8 T cell cytotoxicity by specific

270

targeting of antigen to cross-presenting dendritic cells in vivo via murine or human XCR1. J. Immunol. 194: 1069-1079. 244. Fossum, E., G. Grodeland, D. Terhorst, A. A. Tveita, E. Vikse, S. Mjaaland, S. Henri, B. Malissen, and B. Bogen. 2015. Vaccine molecules targeting Xcr1 on cross-presenting DCs induce protective CD8+ T-cell responses against influenza virus. Eur. J. Immunol. 45: 624-635. 245. Terhorst, D., E. Fossum, A. Baranska, S. Tamoutounour, C. Malosse, M. Garbani, R. Braun, E. Lechat, R. Crameri, B. Bogen, S. Henri, and B. Malissen. 2015. Laser-assisted intradermal delivery of adjuvant-free vaccines targeting XCR1+ dendritic cells induces potent antitumoral responses. J. Immunol. 194: 5895-5902. 246. Yoshida, T., T. Imai, S. Takagi, M. Nishimura, I. Ishikawa, T. Yaoi, and O. Yoshie. 1996. Structure and expression of two highly related genes encoding SCM-1/human lymphotactin. FEBS Lett. 395: 82-88. 247. Finkelman, F. D., A. Lees, R. Birnbaum, W. C. Gause, and S. C. Morris. 1996. Dendritic cells can present antigen in vivo in a tolerogenic or immunogenic fashion. J. Immunol. 157: 1406-1414. 248. Skea, D. L., and B. H. Barber. 1993. Studies of the adjuvant-independent antibody response to immunotargeting. Target structure dependence, isotype distribution, and induction of long term memory. J. Immunol. 151: 3557-3568. 249. Snider, D. P., A. Kaubisch, and D. M. Segal. 1990. Enhanced antigen immunogenicity induced by bispecific antibodies. The Journal of experimental medicine 171: 1957-1963. 250. Corbett, A. J., I. Caminschi, B. S. McKenzie, J. L. Brady, M. D. Wright, P. L. Mottram, P. M. Hogarth, A. N. Hodder, Y. Zhan, D. M. Tarlinton, K. Shortman, and A. M. Lew. 2005. Antigen delivery via two molecules on the CD8- dendritic cell subset induces humoral immunity in the absence of conventional "danger". Eur. J. Immunol. 35: 2815-2825. 251. Kaden, S. A., S. Kurig, K. Vasters, K. Hofmann, K. S. Zaenker, J. Schmitz, and G. Winkels. 2009. Enhanced dendritic cell-induced immune responses mediated by the novel C-type lectin receptor mDCAR1. J. Immunol. 183: 5069-5078. 252. Bruder, D., A. M. Westendorf, W. Hansen, S. Prettin, A. D. Gruber, Y. Qian, H. von Boehmer, K. Mahnke, and J. Buer. 2005. On the edge of autoimmunity: T-cell stimulation by steady-state dendritic cells prevents autoimmune diabetes. Diabetes 54: 3395-3401. 253. Mukhopadhaya, A., T. Hanafusa, I. Jarchum, Y. G. Chen, Y. Iwai, D. V. Serreze, R. M. Steinman, K. V. Tarbell, and T. P. DiLorenzo. 2008. Selective delivery of beta cell antigen to dendritic cells in vivo leads to deletion and tolerance of autoreactive CD8+ T cells in NOD mice. Proc. Natl. Acad. Sci. U. S. A. 105: 6374-6379. 254. Petzold, C., J. Riewaldt, T. Koenig, S. Schallenberg, and K. Kretschmer. 2010. Dendritic cell-targeted pancreatic beta-cell antigen leads to conversion of self-

271

reactive CD4(+) T cells into regulatory T cells and promotes immunotolerance in NOD mice. The review of diabetic studies : RDS 7: 47-61. 255. Hawiger, D., R. F. Masilamani, E. Bettelli, V. K. Kuchroo, and M. C. Nussenzweig. 2004. Immunological unresponsiveness characterized by increased expression of CD5 on peripheral T cells induced by dendritic cells in vivo. Immunity 20: 695-705. 256. Ring, S., M. Maas, D. M. Nettelbeck, A. H. Enk, and K. Mahnke. 2013. Targeting of autoantigens to DEC205(+) dendritic cells in vivo suppresses experimental allergic encephalomyelitis in mice. J. Immunol. 191: 2938-2947. 257. Spiering, R., B. Margry, C. Keijzer, C. Petzold, A. Hoek, J. Wagenaar-Hilbers, R. van der Zee, W. van Eden, K. Kretschmer, and F. Broere. 2015. DEC205+ Dendritic Cell-Targeted Tolerogenic Vaccination Promotes Immune Tolerance in Experimental Autoimmune Arthritis. J. Immunol. 194: 4804- 4813. 258. Grun, J. L., and P. H. Maurer. 1989. Different T helper cell subsets elicited in mice utilizing two different adjuvant vehicles: the role of endogenous interleukin 1 in proliferative responses. Cell. Immunol. 121: 134-145. 259. Gupta, R. K. 1998. Aluminum compounds as vaccine adjuvants. Advanced drug delivery reviews 32: 155-172. 260. Lee, S., and M. T. Nguyen. 2015. Recent advances of vaccine adjuvants for infectious diseases. Immune network 15: 51-57. 261. Temizoz, B., E. Kuroda, and K. J. Ishii. 2016. Vaccine adjuvants as potential cancer immunotherapeutics. Int. Immunol. 28: 329-338. 262. Toussi, D. N., and P. Massari. 2014. Immune Adjuvant Effect of Molecularly- defined Toll-Like Receptor Ligands. Vaccines 2: 323-353. 263. Kreutz, M., B. Giquel, Q. Hu, R. Abuknesha, S. Uematsu, S. Akira, F. O. Nestle, and S. S. Diebold. 2012. Antibody-antigen-adjuvant conjugates enable co- delivery of antigen and adjuvant to dendritic cells in cis but only have partial targeting specificity. PloS one 7: e40208. 264. Lahoud, M. H., F. Ahmet, J. G. Zhang, S. Meuter, A. N. Policheni, S. Kitsoulis, C. N. Lee, M. O'Keeffe, L. C. Sullivan, A. G. Brooks, R. Berry, J. Rossjohn, J. D. Mintern, J. Vega-Ramos, J. A. Villadangos, N. A. Nicola, M. C. Nussenzweig, K. J. Stacey, K. Shortman, W. R. Heath, and I. Caminschi. 2012. DEC-205 is a cell surface receptor for CpG oligonucleotides. Proc. Natl. Acad. Sci. U. S. A. 109: 16270-16275. 265. Woodland, D. L. 2004. Jump-starting the immune system: prime-boosting comes of age. Trends in immunology 25: 98-104. 266. Park, C. G., A. Rodriguez, H. Ueta, H. Lee, M. Pack, K. Matsuno, and R. M. Steinman. 2012. Generation of anti-human DEC205/CD205 monoclonal antibodies that recognize epitopes conserved in different mammals. J. Immunol. Methods 377: 15-22. 267. Dougan, S. K., S. Ogata, C. C. Hu, G. M. Grotenbreg, E. Guillen, R. Jaenisch, and H. L. Ploegh. 2012. IgG1+ ovalbumin-specific B-cell transnuclear mice

272

show class switch recombination in rare allelically included B cells. Proc. Natl. Acad. Sci. U. S. A. 109: 13739-13744. 268. Fredriksen, A. B., I. Sandlie, and B. Bogen. 2006. DNA vaccines increase immunogenicity of idiotypic tumor antigen by targeting novel fusion proteins to antigen-presenting cells. Molecular therapy : the journal of the American Society of Gene Therapy 13: 776-785. 269. Wang, L., S. Chen, T. Xu, K. Taghizadeh, J. S. Wishnok, X. Zhou, D. You, Z. Deng, and P. C. Dedon. 2007. Phosphorothioation of DNA in bacteria by dnd genes. Nature chemical biology 3: 709-710. 270. Moore, K. J., L. P. Andersson, R. R. Ingalls, B. G. Monks, R. Li, M. A. Arnaout, D. T. Golenbock, and M. W. Freeman. 2000. Divergent response to LPS and bacteria in CD14-deficient murine macrophages. J. Immunol. 165: 4272-4280. 271. Hogquist, K. A., S. C. Jameson, W. R. Heath, J. L. Howard, M. J. Bevan, and F. R. Carbone. 1994. T cell receptor antagonist peptides induce positive selection. Cell 76: 17-27. 272. Barnden, M. J., J. Allison, W. R. Heath, and F. R. Carbone. 1998. Defective TCR expression in transgenic mice constructed using cDNA-based alpha- and beta-chain genes under the control of heterologous regulatory elements. Immunol. Cell Biol. 76: 34-40. 273. White, P. J., F. Anastasopoulos, J. E. Church, C. Y. Kuo, B. J. Boyd, P. L. Hickey, L. S. Tu, P. Burns, A. M. Lew, W. R. Heath, G. M. Davey, and C. W. Pouton. 2008. Generic construction of single component particles that elicit humoural and cellular immune responses without the need for adjuvants. Vaccine 26: 6824-6831. 274. Vremec, D., M. O'Keeffe, H. Hochrein, M. Fuchsberger, I. Caminschi, M. Lahoud, and K. Shortman. 2007. Production of interferons by dendritic cells, plasmacytoid cells, natural killer cells, and interferon-producing killer dendritic cells. Blood 109: 1165-1173. 275. Sirois, C. M., T. Jin, A. L. Miller, D. Bertheloot, H. Nakamura, G. L. Horvath, A. Mian, J. Jiang, J. Schrum, L. Bossaller, K. Pelka, N. Garbi, Y. Brewah, J. Tian, C. Chang, P. S. Chowdhury, G. P. Sims, R. Kolbeck, A. J. Coyle, A. A. Humbles, T. S. Xiao, and E. Latz. 2013. RAGE is a nucleic acid receptor that promotes inflammatory responses to DNA. The Journal of experimental medicine 210: 2447-2463. 276. Eckstein, F., and G. Gish. 1989. Phosphorothioates in molecular biology. Trends Biochem. Sci. 14: 97-100. 277. Farman, C. A., and D. J. Kornbrust. 2003. Oligodeoxynucleotide studies in primates: antisense and immune stimulatory indications. Toxicol. Pathol. 31 Suppl: 119-122. 278. Sands, H., L. J. Gorey-Feret, A. J. Cocuzza, F. W. Hobbs, D. Chidester, and G. L. Trainor. 1994. Biodistribution and metabolism of internally 3H-labeled oligonucleotides. I. Comparison of a phosphodiester and a phosphorothioate. Mol. Pharmacol. 45: 932-943.

273

279. Zhao, Q., S. Matson, C. J. Herrera, E. Fisher, H. Yu, and A. M. Krieg. 1993. Comparison of cellular binding and uptake of antisense phosphodiester, phosphorothioate, and mixed phosphorothioate and methylphosphonate oligonucleotides. Antisense Res. Dev. 3: 53-66. 280. Crooke, S. T., S. Wang, T. A. Vickers, W. Shen, and X. H. Liang. 2017. Cellular uptake and trafficking of antisense oligonucleotides. Nat. Biotechnol. 35: 230- 237. 281. Krieg, A. M., A. K. Yi, S. Matson, T. J. Waldschmidt, G. A. Bishop, R. Teasdale, G. A. Koretzky, and D. M. Klinman. 1995. CpG motifs in bacterial DNA trigger direct B-cell activation. Nature 374: 546-549. 282. Klinman, D. M., G. Yamshchikov, and Y. Ishigatsubo. 1997. Contribution of CpG motifs to the immunogenicity of DNA vaccines. J. Immunol. 158: 3635- 3639. 283. Bird, A. P. 1987. CpG islands as gene markers in the vertebrate nucleus. Trends in genetics : TIG 3: 342-347. 284. Krieg, A. M. 2002. CpG motifs in bacterial DNA and their immune effects. Annu. Rev. Immunol. 20: 709-760. 285. Stacey, K. J., G. R. Young, F. Clark, D. P. Sester, T. L. Roberts, S. Naik, M. J. Sweet, and D. A. Hume. 2003. The molecular basis for the lack of immunostimulatory activity of vertebrate DNA. J. Immunol. 170: 3614-3620. 286. Vollmer, J. 2006. TLR9 in health and disease. Int. Rev. Immunol. 25: 155-181. 287. Verthelyi, D., K. J. Ishii, M. Gursel, F. Takeshita, and D. M. Klinman. 2001. Human peripheral blood cells differentially recognize and respond to two distinct CPG motifs. J. Immunol. 166: 2372-2377. 288. Krug, A., S. Rothenfusser, V. Hornung, B. Jahrsdorfer, S. Blackwell, Z. K. Ballas, S. Endres, A. M. Krieg, and G. Hartmann. 2001. Identification of CpG oligonucleotide sequences with high induction of IFN-alpha/beta in plasmacytoid dendritic cells. Eur. J. Immunol. 31: 2154-2163. 289. Gursel, M., D. Verthelyi, I. Gursel, K. J. Ishii, and D. M. Klinman. 2002. Differential and competitive activation of human immune cells by distinct classes of CpG oligodeoxynucleotide. J. Leukoc. Biol. 71: 813-820. 290. Ballas, Z. K., W. L. Rasmussen, and A. M. Krieg. 1996. Induction of NK activity in murine and human cells by CpG motifs in oligodeoxynucleotides and bacterial DNA. J. Immunol. 157: 1840-1845. 291. Cho, H. J., T. Hayashi, S. K. Datta, K. Takabayashi, J. H. Van Uden, A. Horner, M. Corr, and E. Raz. 2002. IFN-alpha beta promote priming of antigen- specific CD8+ and CD4+ T lymphocytes by immunostimulatory DNA-based vaccines. J. Immunol. 168: 4907-4913. 292. Hartmann, G., and A. M. Krieg. 2000. Mechanism and function of a newly identified CpG DNA motif in human primary B cells. J. Immunol. 164: 944- 953.

274

293. Jakob, T., P. S. Walker, A. M. Krieg, M. C. Udey, and J. C. Vogel. 1998. Activation of cutaneous dendritic cells by CpG-containing oligodeoxynucleotides: a role for dendritic cells in the augmentation of Th1 responses by immunostimulatory DNA. J. Immunol. 161: 3042-3049. 294. Sparwasser, T., E. S. Koch, R. M. Vabulas, K. Heeg, G. B. Lipford, J. W. Ellwart, and H. Wagner. 1998. Bacterial DNA and immunostimulatory CpG oligonucleotides trigger maturation and activation of murine dendritic cells. Eur. J. Immunol. 28: 2045-2054. 295. Hartmann, G., R. D. Weeratna, Z. K. Ballas, P. Payette, S. Blackwell, I. Suparto, W. L. Rasmussen, M. Waldschmidt, D. Sajuthi, R. H. Purcell, H. L. Davis, and A. M. Krieg. 2000. Delineation of a CpG phosphorothioate oligodeoxynucleotide for activating primate immune responses in vitro and in vivo. J. Immunol. 164: 1617-1624. 296. Hartmann, G., J. Battiany, H. Poeck, M. Wagner, M. Kerkmann, N. Lubenow, S. Rothenfusser, and S. Endres. 2003. Rational design of new CpG oligonucleotides that combine B cell activation with high IFN-alpha induction in plasmacytoid dendritic cells. Eur. J. Immunol. 33: 1633-1641. 297. Marshall, J. D., K. Fearon, C. Abbate, S. Subramanian, P. Yee, J. Gregorio, R. L. Coffman, and G. Van Nest. 2003. Identification of a novel CpG DNA class and motif that optimally stimulate B cell and plasmacytoid dendritic cell functions. J. Leukoc. Biol. 73: 781-792. 298. Vollmer, J., R. Weeratna, P. Payette, M. Jurk, C. Schetter, M. Laucht, T. Wader, S. Tluk, M. Liu, H. L. Davis, and A. M. Krieg. 2004. Characterization of three CpG oligodeoxynucleotide classes with distinct immunostimulatory activities. Eur. J. Immunol. 34: 251-262. 299. Marshall, J. D., K. L. Fearon, D. Higgins, E. M. Hessel, H. Kanzler, C. Abbate, P. Yee, J. Gregorio, T. D. Cruz, J. O. Lizcano, A. Zolotorev, H. M. McClure, K. M. Brasky, K. K. Murthy, R. L. Coffman, and G. V. Nest. 2005. Superior activity of the type C class of ISS in vitro and in vivo across multiple species. DNA Cell Biol. 24: 63-72. 300. Samulowitz, U., M. Weber, R. Weeratna, E. Uhlmann, B. Noll, A. M. Krieg, and J. Vollmer. 2010. A novel class of immune-stimulatory CpG oligodeoxynucleotides unifies high potency in type I interferon induction with preferred structural properties. Oligonucleotides 20: 93-101. 301. Kadowaki, N., S. Ho, S. Antonenko, R. W. Malefyt, R. A. Kastelein, F. Bazan, and Y. J. Liu. 2001. Subsets of human dendritic cell precursors express different toll-like receptors and respond to different microbial antigens. The Journal of experimental medicine 194: 863-869. 302. Bauer, M., V. Redecke, J. W. Ellwart, B. Scherer, J. P. Kremer, H. Wagner, and G. B. Lipford. 2001. Bacterial CpG-DNA triggers activation and maturation of human CD11c-, CD123+ dendritic cells. J. Immunol. 166: 5000-5007. 303. Roman, M., E. Martin-Orozco, J. S. Goodman, M. D. Nguyen, Y. Sato, A. Ronaghy, R. S. Kornbluth, D. D. Richman, D. A. Carson, and E. Raz. 1997.

275

Immunostimulatory DNA sequences function as T helper-1-promoting adjuvants. Nat. Med. 3: 849-854. 304. Chu, R. S., O. S. Targoni, A. M. Krieg, P. V. Lehmann, and C. V. Harding. 1997. CpG oligodeoxynucleotides act as adjuvants that switch on T helper 1 (Th1) immunity. The Journal of experimental medicine 186: 1623-1631. 305. Vabulas, R. M., H. Pircher, G. B. Lipford, H. Hacker, and H. Wagner. 2000. CpG-DNA activates in vivo T cell epitope presenting dendritic cells to trigger protective antiviral cytotoxic T cell responses. J. Immunol. 164: 2372-2378. 306. Scheiermann, J., and D. M. Klinman. 2014. Clinical evaluation of CpG oligonucleotides as adjuvants for vaccines targeting infectious diseases and cancer. Vaccine 32: 6377-6389. 307. Krieg, A. M. 2012. CpG still rocks! Update on an accidental drug. Nucleic acid therapeutics 22: 77-89. 308. Shirota, H., D. Tross, and D. M. Klinman. 2015. CpG Oligonucleotides as Cancer Vaccine Adjuvants. Vaccines 3: 390-407. 309. Bode, C., G. Zhao, F. Steinhagen, T. Kinjo, and D. M. Klinman. 2011. CpG DNA as a vaccine adjuvant. Expert review of vaccines 10: 499-511. 310. Halperin, S. A., B. Ward, C. Cooper, G. Predy, F. Diaz-Mitoma, M. Dionne, J. Embree, A. McGeer, P. Zickler, K. H. Moltz, R. Martz, I. Meyer, S. McNeil, J. M. Langley, E. Martins, W. L. Heyward, and J. T. Martin. 2012. Comparison of safety and immunogenicity of two doses of investigational hepatitis B virus surface antigen co-administered with an immunostimulatory phosphorothioate oligodeoxyribonucleotide and three doses of a licensed hepatitis B vaccine in healthy adults 18-55 years of age. Vaccine 30: 2556-2563. 311. Eng, N. F., N. Bhardwaj, R. Mulligan, and F. Diaz-Mitoma. 2013. The potential of 1018 ISS adjuvant in hepatitis B vaccines: HEPLISAV review. Human vaccines & immunotherapeutics 9: 1661-1672. 312. Hemmi, H., O. Takeuchi, T. Kawai, T. Kaisho, S. Sato, H. Sanjo, M. Matsumoto, K. Hoshino, H. Wagner, K. Takeda, and S. Akira. 2000. A Toll- like receptor recognizes bacterial DNA. Nature 408: 740-745. 313. Hemmi, H., T. Kaisho, K. Takeda, and S. Akira. 2003. The roles of Toll-like receptor 9, MyD88, and DNA-dependent protein kinase catalytic subunit in the effects of two distinct CpG DNAs on dendritic cell subsets. J. Immunol. 170: 3059-3064. 314. Bauer, S., C. J. Kirschning, H. Hacker, V. Redecke, S. Hausmann, S. Akira, H. Wagner, and G. B. Lipford. 2001. Human TLR9 confers responsiveness to bacterial DNA via species-specific CpG motif recognition. Proc. Natl. Acad. Sci. U. S. A. 98: 9237-9242. 315. Roberts, T. L., M. J. Sweet, D. A. Hume, and K. J. Stacey. 2005. Cutting edge: species-specific TLR9-mediated recognition of CpG and non-CpG phosphorothioate-modified oligonucleotides. J. Immunol. 174: 605-608. 316. Honda, K., H. Yanai, H. Negishi, M. Asagiri, M. Sato, T. Mizutani, N. Shimada, Y. Ohba, A. Takaoka, N. Yoshida, and T. Taniguchi. 2005. IRF-7 is

276

the master regulator of type-I interferon-dependent immune responses. Nature 434: 772-777. 317. Takaoka, A., H. Yanai, S. Kondo, G. Duncan, H. Negishi, T. Mizutani, S. Kano, K. Honda, Y. Ohba, T. W. Mak, and T. Taniguchi. 2005. Integral role of IRF-5 in the gene induction programme activated by Toll-like receptors. Nature 434: 243-249. 318. Honda, K., Y. Ohba, H. Yanai, H. Negishi, T. Mizutani, A. Takaoka, C. Taya, and T. Taniguchi. 2005. Spatiotemporal regulation of MyD88-IRF-7 signalling for robust type-I interferon induction. Nature 434: 1035-1040. 319. Guiducci, C., G. Ott, J. H. Chan, E. Damon, C. Calacsan, T. Matray, K. D. Lee, R. L. Coffman, and F. J. Barrat. 2006. Properties regulating the nature of the plasmacytoid dendritic cell response to Toll-like receptor 9 activation. The Journal of experimental medicine 203: 1999-2008. 320. Kelley, J. L., T. R. Ozment, C. Li, J. B. Schweitzer, and D. L. Williams. 2014. Scavenger receptor-A (CD204): a two-edged sword in health and disease. Crit. Rev. Immunol. 34: 241-261. 321. Pearson, A. M., A. Rich, and M. Krieger. 1993. Polynucleotide binding to macrophage scavenger receptors depends on the formation of base-quartet- stabilized four-stranded helices. J. Biol. Chem. 268: 3546-3554. 322. Suzuki, K., T. Doi, T. Imanishi, T. Kodama, and T. Tanaka. 1999. Oligonucleotide aggregates bind to the macrophage scavenger receptor. Eur. J. Biochem. 260: 855-860. 323. Takakura, Y., T. Takagi, M. Hashiguchi, M. Nishikawa, F. Yamashita, T. Doi, T. Imanishi, H. Suzuki, T. Kodama, and M. Hashida. 1999. Characterization of plasmid DNA binding and uptake by peritoneal macrophages from class A scavenger receptor knockout mice. Pharm. Res. 16: 503-508. 324. Butler, M., R. M. Crooke, M. J. Graham, K. M. Lemonidis, M. Lougheed, S. F. Murray, D. Witchell, U. Steinbrecher, and C. F. Bennett. 2000. Phosphorothioate oligodeoxynucleotides distribute similarly in class A scavenger receptor knockout and wild-type mice. J. Pharmacol. Exp. Ther. 292: 489-496. 325. Zhu, F. G., C. F. Reich, and D. S. Pisetsky. 2001. The role of the macrophage scavenger receptor in immune stimulation by bacterial DNA and synthetic oligonucleotides. Immunology 103: 226-234. 326. Jozefowski, S., T. H. Sulahian, M. Arredouani, and L. Kobzik. 2006. Role of scavenger receptor MARCO in macrophage responses to CpG oligodeoxynucleotides. J. Leukoc. Biol. 80: 870-879. 327. Gursel, M., I. Gursel, H. S. Mostowski, and D. M. Klinman. 2006. CXCL16 influences the nature and specificity of CpG-induced immune activation. J. Immunol. 177: 1575-1580. 328. Gursel, M., D. M. Klinman, and I. Gursel. 2013. Plasmacytoid dendritic cell response to CpG ODN correlates with CXCL16 expression and is inhibited by ox-LDL. Mediators Inflamm. 2013: 312590.

277

329. Moseman, A. P., E. A. Moseman, S. Schworer, I. Smirnova, T. Volkova, U. von Andrian, and A. Poltorak. 2013. Mannose Receptor 1 Mediates Cellular Uptake and Endosomal Delivery of CpG-Motif Containing Oligodeoxynucleotides. J. Immunol. 191: 5615-5624. 330. Sivori, S., M. Falco, S. Carlomagno, E. Romeo, C. Soldani, A. Bensussan, A. Viola, L. Moretta, and A. Moretta. 2010. A novel KIR-associated function: evidence that CpG DNA uptake and shuttling to early endosomes is mediated by KIR3DL2. Blood 116: 1637-1647. 331. Sivori, S., M. Falco, M. Della Chiesa, S. Carlomagno, M. Vitale, L. Moretta, and A. Moretta. 2004. CpG and double-stranded RNA trigger human NK cells by Toll-like receptors: induction of cytokine release and cytotoxicity against tumors and dendritic cells. Proc. Natl. Acad. Sci. U. S. A. 101: 10116-10121. 332. Baumann, C. L., I. M. Aspalter, O. Sharif, A. Pichlmair, S. Bluml, F. Grebien, M. Bruckner, P. Pasierbek, K. Aumayr, M. Planyavsky, K. L. Bennett, J. Colinge, S. Knapp, and G. Superti-Furga. 2010. CD14 is a coreceptor of Toll- like receptors 7 and 9. The Journal of experimental medicine 207: 2689-2701. 333. Hartmann, G., and A. M. Krieg. 2000. Mechanism and function of a newly identified CpG DNA motif in human primary B cells. J. Immunol. 164: 944- 953. 334. Vollmer, J., A. Janosch, M. Laucht, Z. K. Ballas, C. Schetter, and A. M. Krieg. 2002. Highly immunostimulatory CpG-free oligodeoxynucleotides for activation of human leukocytes. Antisense Nucleic Acid Drug Dev. 12: 165-175. 335. Vollmer, J., R. D. Weeratna, M. Jurk, U. Samulowitz, M. J. McCluskie, P. Payette, H. L. Davis, C. Schetter, and A. M. Krieg. 2004. Oligodeoxynucleotides lacking CpG dinucleotides mediate Toll-like receptor 9 dependent T helper type 2 biased immune stimulation. Immunology 113: 212- 223. 336. Senn, J. J., S. Burel, and S. P. Henry. 2005. Non-CpG-containing antisense 2'- methoxyethyl oligonucleotides activate a proinflammatory response independent of Toll-like receptor 9 or myeloid differentiation factor 88. J. Pharmacol. Exp. Ther. 314: 972-979. 337. Mayo, L., S. A. Trauger, M. Blain, M. Nadeau, B. Patel, J. I. Alvarez, I. D. Mascanfroni, A. Yeste, P. Kivisakk, K. Kallas, B. Ellezam, R. Bakshi, A. Prat, J. P. Antel, H. L. Weiner, and F. J. Quintana. 2014. Regulation of astrocyte activation by glycolipids drives chronic CNS inflammation. Nat. Med. 20: 1147-1156. 338. Liu, Y., X. Luo, C. Yang, S. Yu, and H. Xu. 2011. Three CpG oligodeoxynucleotide classes differentially enhance antigen-specific humoral and cellular immune responses in mice. Vaccine 29: 5778-5784. 339. Kerkmann, M., L. T. Costa, C. Richter, S. Rothenfusser, J. Battiany, V. Hornung, J. Johnson, S. Englert, T. Ketterer, W. Heckl, S. Thalhammer, S. Endres, and G. Hartmann. 2005. Spontaneous formation of nucleic acid-based nanoparticles is responsible for high interferon-alpha induction by CpG-A in plasmacytoid dendritic cells. J. Biol. Chem. 280: 8086-8093.

278

340. Lee, S. W., M. K. Song, K. H. Baek, Y. Park, J. K. Kim, C. H. Lee, H. K. Cheong, C. Cheong, and Y. C. Sung. 2000. Effects of a hexameric deoxyriboguanosine run conjugation into CpG oligodeoxynucleotides on their immunostimulatory potentials. J. Immunol. 165: 3631-3639. 341. Pohar, J., A. Kuznik Krajnik, R. Jerala, and M. Bencina. 2015. Minimal sequence requirements for oligodeoxyribonucleotides activating human TLR9. J. Immunol. 194: 3901-3908. 342. Liang, H., Y. Nishioka, C. F. Reich, D. S. Pisetsky, and P. E. Lipsky. 1996. Activation of human B cells by phosphorothioate oligodeoxynucleotides. J. Clin. Invest. 98: 1119-1129. 343. Messina, J. P., G. S. Gilkeson, and D. S. Pisetsky. 1991. Stimulation of in vitro murine lymphocyte proliferation by bacterial DNA. J. Immunol. 147: 1759- 1764. 344. Yamamoto, S., T. Yamamoto, S. Shimada, E. Kuramoto, O. Yano, T. Kataoka, and T. Tokunaga. 1992. DNA from bacteria, but not from vertebrates, induces interferons, activates natural killer cells and inhibits tumor growth. Microbiol. Immunol. 36: 983-997. 345. Bird, A. 2002. DNA methylation patterns and epigenetic memory. Genes Dev. 16: 6-21. 346. Chen, Y., P. Lenert, R. Weeratna, M. McCluskie, T. Wu, H. L. Davis, and A. M. Krieg. 2001. Identification of methylated CpG motifs as inhibitors of the immune stimulatory CpG motifs. Gene Ther. 8: 1024-1032. 347. Pisetsky, D. S., and C. F. Reich. 2000. Inhibition of murine macrophage IL-12 production by natural and synthetic DNA. Clin. Immunol. 96: 198-204. 348. Lande, R., J. Gregorio, V. Facchinetti, B. Chatterjee, Y. H. Wang, B. Homey, W. Cao, Y. H. Wang, B. Su, F. O. Nestle, T. Zal, I. Mellman, J. M. Schroder, Y. J. Liu, and M. Gilliet. 2007. Plasmacytoid dendritic cells sense self-DNA coupled with antimicrobial peptide. Nature 449: 564-569. 349. Yasuda, K., P. Yu, C. J. Kirschning, B. Schlatter, F. Schmitz, A. Heit, S. Bauer, H. Hochrein, and H. Wagner. 2005. Endosomal translocation of vertebrate DNA activates dendritic cells via TLR9-dependent and -independent pathways. J. Immunol. 174: 6129-6136. 350. Yasuda, K., M. Rutz, B. Schlatter, J. Metzger, P. B. Luppa, F. Schmitz, T. Haas, A. Heit, S. Bauer, and H. Wagner. 2006. CpG motif-independent activation of TLR9 upon endosomal translocation of "natural" phosphodiester DNA. Eur. J. Immunol. 36: 431-436. 351. Haas, T., J. Metzger, F. Schmitz, A. Heit, T. Muller, E. Latz, and H. Wagner. 2008. The DNA sugar backbone 2' deoxyribose determines toll-like receptor 9 activation. Immunity 28: 315-323. 352. Barton, G. M., J. C. Kagan, and R. Medzhitov. 2006. Intracellular localization of Toll-like receptor 9 prevents recognition of self DNA but facilitates access to viral DNA. Nature immunology 7: 49-56.

279

353. Pohar, J., D. Lainscek, R. Fukui, C. Yamamoto, K. Miyake, R. Jerala, and M. Bencina. 2015. Species-Specific Minimal Sequence Motif for Oligodeoxyribonucleotides Activating Mouse TLR9. J. Immunol. 195: 4396- 4405. 354. Wilbanks, A., S. C. Zondlo, K. Murphy, S. Mak, D. Soler, P. Langdon, D. P. Andrew, L. Wu, and M. Briskin. 2001. Expression cloning of the STRL33/BONZO/TYMSTRligand reveals elements of CC, CXC, and CX3C chemokines. J. Immunol. 166: 5145-5154. 355. Lougaris, V., M. Baronio, M. Vitali, G. Tampella, A. Soresina, R. Badolato, and A. Plebani. 2012. B cell responses to CpG correlate with CXCL16 expression levels in common variable immunodeficiency. TheScientificWorldJournal 2012: 960219. 356. Klinman, D. M., F. Takeshita, I. Gursel, C. Leifer, K. J. Ishii, D. Verthelyi, and M. Gursel. 2002. CpG DNA: recognition by and activation of monocytes. Microbes and infection 4: 897-901. 357. Liu, Y. C., R. C. Gray, G. A. Hardy, J. Kuchtey, D. W. Abbott, S. N. Emancipator, and C. V. Harding. 2010. CpG-B oligodeoxynucleotides inhibit TLR-dependent and -independent induction of type I IFN in dendritic cells. J. Immunol. 184: 3367-3376. 358. Waibler, Z., M. Anzaghe, A. Konur, S. Akira, W. Muller, and U. Kalinke. 2008. Excessive CpG 1668 stimulation triggers IL-10 production by cDC that inhibits IFN-alpha responses by pDC. Eur. J. Immunol. 38: 3127-3137. 359. Kim, S., V. Kaiser, E. Beier, M. Bechheim, M. Guenthner-Biller, A. Ablasser, M. Berger, S. Endres, G. Hartmann, and V. Hornung. 2014. Self-priming determines high type I IFN production by plasmacytoid dendritic cells. Eur. J. Immunol. 44: 807-818. 360. Kato, Y., A. Zaid, G. M. Davey, S. N. Mueller, S. L. Nutt, D. Zotos, D. M. Tarlinton, K. Shortman, M. H. Lahoud, W. R. Heath, and I. Caminschi. 2015. Targeting Antigen to Clec9A Primes Follicular Th Cell Memory Responses Capable of Robust Recall. J. Immunol. 195: 1006-1014. 361. Cho, H. J., K. Takabayashi, P. M. Cheng, M. D. Nguyen, M. Corr, S. Tuck, and E. Raz. 2000. Immunostimulatory DNA-based vaccines induce cytotoxic lymphocyte activity by a T-helper cell-independent mechanism. Nat. Biotechnol. 18: 509-514. 362. Maurer, T., A. Heit, H. Hochrein, F. Ampenberger, M. O'Keeffe, S. Bauer, G. B. Lipford, R. M. Vabulas, and H. Wagner. 2002. CpG-DNA aided cross- presentation of soluble antigens by dendritic cells. Eur. J. Immunol. 32: 2356- 2364. 363. Heit, A., T. Maurer, H. Hochrein, S. Bauer, K. M. Huster, D. H. Busch, and H. Wagner. 2003. Cutting edge: Toll-like receptor 9 expression is not required for CpG DNA-aided cross-presentation of DNA-conjugated antigens but essential for cross-priming of CD8 T cells. J. Immunol. 170: 2802-2805.

280

364. Wagner, H. 2009. The immunogenicity of CpG-antigen conjugates. Advanced drug delivery reviews 61: 243-247. 365. Wright, S. D., R. A. Ramos, P. S. Tobias, R. J. Ulevitch, and J. C. Mathison. 1990. CD14, a receptor for complexes of lipopolysaccharide (LPS) and LPS binding protein. Science 249: 1431-1433. 366. Zanoni, I., and F. Granucci. 2013. Role of CD14 in host protection against infections and in metabolism regulation. Frontiers in cellular and infection microbiology 3: 32. 367. Helft, J., J. Bottcher, P. Chakravarty, S. Zelenay, J. Huotari, B. U. Schraml, D. Goubau, and C. Reis e Sousa. 2015. GM-CSF Mouse Bone Marrow Cultures Comprise a Heterogeneous Population of CD11c(+)MHCII(+) Macrophages and Dendritic Cells. Immunity 42: 1197-1211. 368. Stein, C. A., and Y. C. Cheng. 1993. Antisense oligonucleotides as therapeutic agents--is the bullet really magical? Science 261: 1004-1012. 369. Crooke, S. T., and C. F. Bennett. 1996. Progress in antisense oligonucleotide therapeutics. Annu. Rev. Pharmacol. Toxicol. 36: 107-129. 370. Keppler, D., R. Lesch, W. Reutter, and K. Decker. 1968. Experimental hepatitis induced by D-galactosamine. Exp. Mol. Pathol. 9: 279-290. 371. Ghosn, E. E., A. A. Cassado, G. R. Govoni, T. Fukuhara, Y. Yang, D. M. Monack, K. R. Bortoluci, S. R. Almeida, L. A. Herzenberg, and L. A. Herzenberg. 2010. Two physically, functionally, and developmentally distinct peritoneal macrophage subsets. Proc. Natl. Acad. Sci. U. S. A. 107: 2568-2573. 372. Cassado, A. d. A., M. R. D'Imperio Lima, and K. R. Bortoluci. 2015. Revisiting mouse peritoneal macrophages: heterogeneity, development, and function. Frontiers in immunology 6: 225. 373. Kraal, G., M. Breel, M. Janse, and G. Bruin. 1986. Langerhans' cells, veiled cells, and interdigitating cells in the mouse recognized by a monoclonal antibody. The Journal of experimental medicine 163: 981-997. 374. Swiggard, W. J., A. Mirza, M. C. Nussenzweig, and R. M. Steinman. 1995. DEC-205, a 205-kDa protein abundant on mouse dendritic cells and thymic epithelium that is detected by the monoclonal antibody NLDC-145: purification, characterization, and N-terminal amino acid sequence. Cell. Immunol. 165: 302-311. 375. Inaba, K., W. J. Swiggard, M. Inaba, J. Meltzer, A. Mirza, T. Sasagawa, M. C. Nussenzweig, and R. M. Steinman. 1995. Tissue distribution of the DEC-205 protein that is detected by the monoclonal antibody NLDC-145. I. Expression on dendritic cells and other subsets of mouse leukocytes. Cell. Immunol. 163: 148-156. 376. Witmer-Pack, M. D., W. J. Swiggard, A. Mirza, K. Inaba, and R. M. Steinman. 1995. Tissue distribution of the DEC-205 protein that is detected by the monoclonal antibody NLDC-145. II. Expression in situ in lymphoid and nonlymphoid tissues. Cell. Immunol. 163: 157-162.

281

377. Kato, M., K. J. McDonald, S. Khan, I. L. Ross, S. Vuckovic, K. Chen, D. Munster, K. P. MacDonald, and D. N. Hart. 2006. Expression of human DEC- 205 (CD205) multilectin receptor on leukocytes. Int. Immunol. 18: 857-869. 378. Li, J., F. Ahmet, M. O'Keeffe, M. H. Lahoud, W. R. Heath, and I. Caminschi. 2017. CD14 is not involved in the uptake of synthetic CpG oligonucleotides. Mol. Immunol. 81: 52-58. 379. East, L., and C. M. Isacke. 2002. The mannose receptor family. Biochim. Biophys. Acta 1572: 364-386. 380. Nitta, T., and H. Suzuki. 2016. Thymic stromal cell subsets for T cell development. Cell. Mol. Life Sci. 73: 1021-1037. 381. Vremec, D., and K. Shortman. 1997. Dendritic cell subtypes in mouse lymphoid organs: cross-correlation of surface markers, changes with incubation, and differences among thymus, spleen, and lymph nodes. J. Immunol. 159: 565-573. 382. Wu, L., and K. Shortman. 2005. Heterogeneity of thymic dendritic cells. Semin. Immunol. 17: 304-312. 383. Brocker, T., M. Riedinger, and K. Karjalainen. 1997. Targeted expression of major histocompatibility complex (MHC) class II molecules demonstrates that dendritic cells can induce negative but not positive selection of thymocytes in vivo. The Journal of experimental medicine 185: 541-550. 384. Breel, M., R. Mebius, and G. Kraal. 1988. The interaction of interdigitating cells and lymphocytes in vitro and in vivo. Immunology 63: 331-337. 385. Jenkinson, W. E., K. Nakamura, A. J. White, E. J. Jenkinson, and G. Anderson. 2012. Normal T cell selection occurs in CD205-deficient thymic microenvironments. PloS one 7: e53416. 386. Small, M., and G. Kraal. 2003. In vitro evidence for participation of DEC-205 expressed by thymic cortical epithelial cells in clearance of apoptotic thymocytes. Int. Immunol. 15: 197-203. 387. Shrimpton, R. E., M. Butler, A. S. Morel, E. Eren, S. S. Hue, and M. A. Ritter. 2009. CD205 (DEC-205): a recognition receptor for apoptotic and necrotic self. Mol. Immunol. 46: 1229-1239. 388. Geisler, F., and R. E. Leube. 2016. Epithelial Intermediate Filaments: Guardians against Microbial Infection? Cells 5. 389. Zhang, S. S., C. G. Park, P. Zhang, S. S. Bartra, G. V. Plano, J. D. Klena, M. Skurnik, B. J. Hinnebusch, and T. Chen. 2008. Plasminogen activator Pla of Yersinia pestis utilizes murine DEC-205 (CD205) as a receptor to promote dissemination. J. Biol. Chem. 283: 31511-31521. 390. Hatsukari, I., P. Singh, N. Hitosugi, D. Messmer, E. Valderrama, S. Teichberg, W. Chaung, E. Gross, H. Schmidtmayerova, and P. C. Singhal. 2007. DEC- 205-mediated internalization of HIV-1 results in the establishment of silent infection in renal tubular cells. J. Am. Soc. Nephrol. 18: 780-787.

282

391. Mikulak, J., S. Teichberg, T. Faust, H. Schmidtmayerova, and P. C. Singhal. 2009. HIV-1 harboring renal tubular epithelial cell interaction with T cells results in T cell trans-infection. Virology 385: 105-114. 392. Wang, L., S. Chen, K. L. Vergin, S. J. Giovannoni, S. W. Chan, M. S. DeMott, K. Taghizadeh, O. X. Cordero, M. Cutler, S. Timberlake, E. J. Alm, M. F. Polz, J. Pinhassi, Z. Deng, and P. C. Dedon. 2011. DNA phosphorothioation is widespread and quantized in bacterial genomes. Proc. Natl. Acad. Sci. U. S. A. 108: 2963-2968. 393. Pisetsky, D. S. 2012. The origin and properties of extracellular DNA: from PAMP to DAMP. Clin. Immunol. 144: 32-40. 394. Tian, J., A. M. Avalos, S. Y. Mao, B. Chen, K. Senthil, H. Wu, P. Parroche, S. Drabic, D. Golenbock, C. Sirois, J. Hua, L. L. An, L. Audoly, G. La Rosa, A. Bierhaus, P. Naworth, A. Marshak-Rothstein, M. K. Crow, K. A. Fitzgerald, E. Latz, P. A. Kiener, and A. J. Coyle. 2007. Toll-like receptor 9-dependent activation by DNA-containing immune complexes is mediated by HMGB1 and RAGE. Nature immunology 8: 487-496. 395. Park, B., L. Buti, S. Lee, T. Matsuwaki, E. Spooner, M. M. Brinkmann, M. Nishihara, and H. L. Ploegh. 2011. Granulin is a soluble cofactor for toll-like receptor 9 signaling. Immunity 34: 505-513. 396. Gonzalez, I., J. Romero, B. L. Rodriguez, R. Perez-Castro, and A. Rojas. 2013. The immunobiology of the receptor of advanced glycation end-products: trends and challenges. Immunobiology 218: 790-797. 397. Chuah, Y. K., R. Basir, H. Talib, T. H. Tie, and N. Nordin. 2013. Receptor for advanced glycation end products and its involvement in inflammatory diseases. International journal of inflammation 2013: 403460. 398. Sparvero, L. J., D. Asafu-Adjei, R. Kang, D. Tang, N. Amin, J. Im, R. Rutledge, B. Lin, A. A. Amoscato, H. J. Zeh, and M. T. Lotze. 2009. RAGE (Receptor for Advanced Glycation Endproducts), RAGE ligands, and their role in cancer and inflammation. Journal of translational medicine 7: 17. 399. Rojas, A., H. Figueroa, and E. Morales. 2010. Fueling inflammation at tumor microenvironment: the role of multiligand/RAGE axis. Carcinogenesis 31: 334- 341. 400. Ibrahim, Z. A., C. L. Armour, S. Phipps, and M. B. Sukkar. 2013. RAGE and TLRs: relatives, friends or neighbours? Mol. Immunol. 56: 739-744. 401. Chavakis, T., A. Bierhaus, N. Al-Fakhri, D. Schneider, S. Witte, T. Linn, M. Nagashima, J. Morser, B. Arnold, K. T. Preissner, and P. P. Nawroth. 2003. The pattern recognition receptor (RAGE) is a counterreceptor for leukocyte integrins: a novel pathway for inflammatory cell recruitment. The Journal of experimental medicine 198: 1507-1515. 402. Orlova, V. V., E. Y. Choi, C. Xie, E. Chavakis, A. Bierhaus, E. Ihanus, C. M. Ballantyne, C. G. Gahmberg, M. E. Bianchi, P. P. Nawroth, and T. Chavakis. 2007. A novel pathway of HMGB1-mediated inflammatory cell recruitment that requires Mac-1-integrin. EMBO J. 26: 1129-1139.

283

403. Frommhold, D., A. Kamphues, I. Hepper, M. Pruenster, I. K. Lukic, I. Socher, V. Zablotskaya, K. Buschmann, B. Lange-Sperandio, J. Schymeinsky, E. Ryschich, J. Poeschl, C. Kupatt, P. P. Nawroth, M. Moser, B. Walzog, A. Bierhaus, and M. Sperandio. 2010. RAGE and ICAM-1 cooperate in mediating leukocyte recruitment during acute inflammation in vivo. Blood 116: 841-849. 404. Moser, B., D. D. Desai, M. P. Downie, Y. Chen, S. F. Yan, K. Herold, A. M. Schmidt, and R. Clynes. 2007. Receptor for advanced glycation end products expression on T cells contributes to antigen-specific cellular expansion in vivo. J. Immunol. 179: 8051-8058. 405. Chen, Y., E. M. Akirav, W. Chen, O. Henegariu, B. Moser, D. Desai, J. M. Shen, J. C. Webster, R. C. Andrews, A. M. Mjalli, R. Rothlein, A. M. Schmidt, R. Clynes, and K. C. Herold. 2008. RAGE ligation affects T cell activation and controls T cell differentiation. J. Immunol. 181: 4272-4278. 406. Yamamoto, Y., A. Harashima, H. Saito, K. Tsuneyama, S. Munesue, S. Motoyoshi, D. Han, T. Watanabe, M. Asano, S. Takasawa, H. Okamoto, S. Shimura, T. Karasawa, H. Yonekura, and H. Yamamoto. 2011. Septic shock is associated with receptor for advanced glycation end products ligation of LPS. J. Immunol. 186: 3248-3257. 407. Ruan, B. H., X. Li, A. R. Winkler, K. M. Cunningham, J. Kuai, R. M. Greco, K. H. Nocka, L. J. Fitz, J. F. Wright, D. D. Pittman, X. Y. Tan, J. E. Paulsen, L. L. Lin, and D. G. Winkler. 2010. Complement C3a, CpG oligos, and DNA/C3a complex stimulate IFN-alpha production in a receptor for advanced glycation end product-dependent manner. J. Immunol. 185: 4213-4222. 408. Durr, U. H., U. S. Sudheendra, and A. Ramamoorthy. 2006. LL-37, the only human member of the cathelicidin family of antimicrobial peptides. Biochim. Biophys. Acta 1758: 1408-1425. 409. Caminschi, I., D. Vremec, F. Ahmet, M. H. Lahoud, J. A. Villadangos, K. M. Murphy, W. R. Heath, and K. Shortman. 2012. Antibody responses initiated by Clec9A-bearing dendritic cells in normal and Batf3(-/-) mice. Mol. Immunol. 50: 9-17. 410. Gilbert, S. C. 2012. T-cell-inducing vaccines - what's the future. Immunology 135: 19-26. 411. Ewer, K. J., T. Lambe, C. S. Rollier, A. J. Spencer, A. V. Hill, and L. Dorrell. 2016. Viral vectors as vaccine platforms: from immunogenicity to impact. Curr. Opin. Immunol. 41: 47-54. 412. Ura, T., K. Okuda, and M. Shimada. 2014. Developments in Viral Vector- Based Vaccines. Vaccines 2: 624-641. 413. Li, S., M. Rodrigues, D. Rodriguez, J. R. Rodriguez, M. Esteban, P. Palese, R. S. Nussenzweig, and F. Zavala. 1993. Priming with recombinant influenza virus followed by administration of recombinant vaccinia virus induces CD8+ T- cell-mediated protective immunity against malaria. Proc. Natl. Acad. Sci. U. S. A. 90: 5214-5218.

284

414. Gilbert, S. C., J. Schneider, C. M. Hannan, J. T. Hu, M. Plebanski, R. Sinden, and A. V. Hill. 2002. Enhanced CD8 T cell immunogenicity and protective efficacy in a mouse malaria model using a recombinant adenoviral vaccine in heterologous prime-boost immunisation regimes. Vaccine 20: 1039-1045. 415. Kutzler, M. A., and D. B. Weiner. 2008. DNA vaccines: ready for prime time? Nature reviews. Genetics 9: 776-788. 416. Li, L., and N. Petrovsky. 2016. Molecular mechanisms for enhanced DNA vaccine immunogenicity. Expert review of vaccines 15: 313-329. 417. Schneider, J., S. C. Gilbert, T. J. Blanchard, T. Hanke, K. J. Robson, C. M. Hannan, M. Becker, R. Sinden, G. L. Smith, and A. V. Hill. 1998. Enhanced immunogenicity for CD8+ T cell induction and complete protective efficacy of malaria DNA vaccination by boosting with modified vaccinia virus Ankara. Nat. Med. 4: 397-402. 418. Sullivan, N. J., A. Sanchez, P. E. Rollin, Z. Y. Yang, and G. J. Nabel. 2000. Development of a preventive vaccine for Ebola virus infection in primates. Nature 408: 605-609. 419. McShane, H., R. Brookes, S. C. Gilbert, and A. V. Hill. 2001. Enhanced immunogenicity of CD4(+) t-cell responses and protective efficacy of a DNA- modified vaccinia virus Ankara prime-boost vaccination regimen for murine tuberculosis. Infect. Immun. 69: 681-686. 420. Gudmundsdotter, L., C. Nilsson, A. Brave, B. Hejdeman, P. Earl, B. Moss, M. Robb, J. Cox, N. Michael, M. Marovich, G. Biberfeld, E. Sandstrom, and B. Wahren. 2009. Recombinant Modified Vaccinia Ankara (MVA) effectively boosts DNA-primed HIV-specific immune responses in humans despite pre- existing vaccinia immunity. Vaccine 27: 4468-4474. 421. McConkey, S. J., W. H. Reece, V. S. Moorthy, D. Webster, S. Dunachie, G. Butcher, J. M. Vuola, T. J. Blanchard, P. Gothard, K. Watkins, C. M. Hannan, S. Everaere, K. Brown, K. E. Kester, J. Cummings, J. Williams, D. G. Heppner, A. Pathan, K. Flanagan, N. Arulanantham, M. T. Roberts, M. Roy, G. L. Smith, J. Schneider, T. Peto, R. E. Sinden, S. C. Gilbert, and A. V. Hill. 2003. Enhanced T-cell immunogenicity of plasmid DNA vaccines boosted by recombinant modified vaccinia virus Ankara in humans. Nat. Med. 9: 729-735. 422. Harrington, L. E., R. Most Rv, J. L. Whitton, and R. Ahmed. 2002. Recombinant vaccinia virus-induced T-cell immunity: quantitation of the response to the virus vector and the foreign epitope. J. Virol. 76: 3329-3337. 423. Schneider, J., S. C. Gilbert, C. M. Hannan, P. Degano, E. Prieur, E. G. Sheu, M. Plebanski, and A. V. Hill. 1999. Induction of CD8+ T cells using heterologous prime-boost immunisation strategies. Immunol. Rev. 170: 29-38. 424. de Mare, A., A. J. Lambeck, J. Regts, G. M. van Dam, H. W. Nijman, H. Snippe, J. Wilschut, and T. Daemen. 2008. Viral vector-based prime-boost immunization regimens: a possible involvement of T-cell competition. Gene Ther. 15: 393-403.

285

425. Yang, Z. Y., L. S. Wyatt, W. P. Kong, Z. Moodie, B. Moss, and G. J. Nabel. 2003. Overcoming immunity to a viral vaccine by DNA priming before vector boosting. J. Virol. 77: 799-803. 426. Sumida, S. M., D. M. Truitt, M. G. Kishko, J. C. Arthur, S. S. Jackson, D. A. Gorgone, M. A. Lifton, W. Koudstaal, M. G. Pau, S. Kostense, M. J. Havenga, J. Goudsmit, N. L. Letvin, and D. H. Barouch. 2004. Neutralizing antibodies and CD8+ T lymphocytes both contribute to immunity to adenovirus serotype 5 vaccine vectors. J. Virol. 78: 2666-2673. 427. McCoy, K., N. Tatsis, B. Korioth-Schmitz, M. O. Lasaro, S. E. Hensley, S. W. Lin, Y. Li, W. Giles-Davis, A. Cun, D. Zhou, Z. Xiang, N. L. Letvin, and H. C. Ertl. 2007. Effect of preexisting immunity to adenovirus human serotype 5 antigens on the immune responses of nonhuman primates to vaccine regimens based on human- or chimpanzee-derived adenovirus vectors. J. Virol. 81: 6594-6604. 428. Kitamura, D., J. Roes, R. Kuhn, and K. Rajewsky. 1991. A B cell-deficient mouse by targeted disruption of the membrane exon of the immunoglobulin mu chain gene. Nature 350: 423-426. 429. Schreiber, R. D., L. J. Old, and M. J. Smyth. 2011. Cancer immunoediting: integrating immunity's roles in cancer suppression and promotion. Science 331: 1565-1570. 430. Mittal, D., M. M. Gubin, R. D. Schreiber, and M. J. Smyth. 2014. New insights into cancer immunoediting and its three component phases--elimination, equilibrium and escape. Curr. Opin. Immunol. 27: 16-25. 431. Brice, G. T., C. Dobano, M. Sedegah, M. Stefaniak, N. L. Graber, J. J. Campo, D. J. Carucci, and D. L. Doolan. 2007. Extended immunization intervals enhance the immunogenicity and protective efficacy of plasmid DNA vaccines. Microbes and infection 9: 1439-1446. 432. Weiss, W. R., A. Kumar, G. Jiang, J. Williams, A. Bostick, S. Conteh, D. Fryauff, J. Aguiar, M. Singh, D. T. O'Hagan, J. B. Ulmer, and T. L. Richie. 2007. Protection of rhesus monkeys by a DNA prime/poxvirus boost malaria vaccine depends on optimal DNA priming and inclusion of blood stage antigens. PloS one 2: e1063. 433. Park, H. Y., A. Light, M. H. Lahoud, I. Caminschi, D. M. Tarlinton, and K. Shortman. 2013. Evolution of B cell responses to Clec9A-targeted antigen. J. Immunol. 191: 4919-4925. 434. See, P., C. A. Dutertre, J. Chen, P. Gunther, N. McGovern, S. E. Irac, M. Gunawan, M. Beyer, K. Handler, K. Duan, H. R. B. Sumatoh, N. Ruffin, M. Jouve, E. Gea-Mallorqui, R. C. M. Hennekam, T. Lim, C. C. Yip, M. Wen, B. Malleret, I. Low, N. B. Shadan, C. F. S. Fen, A. Tay, J. Lum, F. Zolezzi, A. Larbi, M. Poidinger, J. K. Y. Chan, Q. Chen, L. Renia, M. Haniffa, P. Benaroch, A. Schlitzer, J. L. Schultze, E. W. Newell, and F. Ginhoux. 2017. Mapping the human DC lineage through the integration of high-dimensional techniques. Science.

286

435. Tel, J., D. Benitez-Ribas, S. Hoosemans, A. Cambi, G. J. Adema, C. G. Figdor, P. J. Tacken, and I. J. de Vries. 2011. DEC-205 mediates antigen uptake and presentation by both resting and activated human plasmacytoid dendritic cells. Eur. J. Immunol. 41: 1014-1023.

287

Minerva Access is the Institutional Repository of The University of Melbourne

Author/s: Li, Jessica

Title: Using dendritic cell receptors to enhance immunity

Date: 2017

Persistent Link: http://hdl.handle.net/11343/192947

File Description: Using dendritic cell receptors to enhance immunity

Terms and Conditions: Terms and Conditions: Copyright in works deposited in Minerva Access is retained by the copyright owner. The work may not be altered without permission from the copyright owner. Readers may only download, print and save electronic copies of whole works for their own personal non-commercial use. Any use that exceeds these limits requires permission from the copyright owner. Attribution is essential when quoting or paraphrasing from these works.