Properties of the C5a Receptor on Human Macrophages Vernon Seow Bachelor of Science (Honours Class 1, Microbiology)

A thesis submitted for the degree of Doctor of Philosophy at The University of Queensland in 2014 Institute for Molecular Bioscience

I Abstract

Macrophages (MΦ) are important to host defence and inflammation but plasticity in their gene expression profiles contributes to their functional heterogeneity. C5a is one of the most potent proinflammatory agents generated upon complement activation and binds to a specific receptor C5aR. Overproduction of C5a can contribute to numerous immune and inflammatory conditions. The objective of this thesis was to focus on human macrophage populations differentiated from primary human monocytes, and examine effects of C5a, C5aR agonists and antagonists, and TLR4- C5aR crosstalk on inflammatory profiles of human monocytes and macrophages. Chapter One summarizes roles for C5a activation on human macrophages, briefly surveying C5a, C5aR and some important analogues of C5a that had been developed to study roles for C5a in vivo. This chapter discusses new insights into C5aR-related biological functions and diseases, macrophage polarization and the usefulness of mouse models for human biology in relation to how closely human and mouse macrophages reproduce functional responses. Chapter Two addresses macrophage polarization. Macrophage heterogeneity was previously achieved by differentiating monocytes with either GM-CSF or M-CSF (generating GM- MΦ or M-MΦ), but was mostly studied on murine rather than human cells. This chapter describes a comparison of gene expression in populations of human macrophages (GM-MΦ versus M-MΦ), both in the basal state as well as in response to stimulation by LPS, using a combination of real-time PCR and cDNA microarray analysis, cellular migration, and functional responses. About 1000 genes are differently regulated between unstimulated GM-MΦ versus M-MΦ. Although evidence is presented in this chapter that human GM-MΦ and M-MΦ have distinct pro- and anti-inflammatory responses in human monocyte-derived macrophages, their activation by LPS supported more of a continuum without clear functional distinctions between each population. Chapter Three investigates the effects of C5aR activation on GM-MΦ versus M-MΦ using cDNA microarray. Previous reports only described gene expression in murine, rather than human, macrophages. Of ~40 genes substantially regulated by C5a in GM-MΦ and M-MΦ, 60% were common to both macrophage types. Furthermore, three different functional assays showed that C5a was equipotent on GM-MΦ and M-MΦ. These similarities for C5a-mediated functions could be due to GM-MΦ and M-MΦ having similar C5aR transcriptional and translational expression. A separate microarray experiment was conducted to overview the temporal gene expression profile induced by C5a on M-MΦ, enabling identification of rate-limiting genes as therapeutic targets in C5a-mediated inflammatory disorders. Chapter Four reports a comparative study of three known antagonists (3D53, W54011, JJ47) of C5a action via C5aR on M-MΦ. They were assessed for their relative advantages and

II disadvantages as antagonists of C5aR. Traditionally, drug discovery has focused on optimizing ligand affinity for a specific receptor, enhancing functional potency, and improving drug-like properties for optimal pharmacokinetics and oral bioavailability. However, these optimization processes do not always improve in vivo efficacy. This chapter highlights some important considerations, often neglected in drug development – namely residence time, insurmountable or non-competitive antagonism for drug-receptor interactions. The non-competitive antagonist 3D53 showed superior antagonist activity compared to competitive antagonists W54011 and JJ47 despite inferior oral bioavailability, and this was because 3D53 remained bound to C5aR on macrophages for over 16h whereas compounds W54011 and JJ47 were no longer bound to C5aR after 2h. As a result, 3D53 retained high potency (IC50 20nM) in the face of competition from increasing concentrations of C5a (0.1-300nM), whereas W54011 and JJ47 were ineffective antagonists against concentrations of C5a above 30nM. The benefits of longer residence time were also demonstrated in vivo in rats. Orally administered 3D53 was an effective anti-inflammatory agent for 16h, compared to less than 2h for W54011 and JJ47, in inhibiting C5aR-mediated paw oedema in male Wistar rats. Chapter Five reports crosstalk between TLR4 and C5aR in human macrophages. C5aR is shown to differentially modulate LPS-induced inflammatory responses in primary human monocytes versus macrophages. While C5a enhanced secretion of LPS-induced IL6 and TNF from human monocytes, C5a inhibited these responses in GM-MΦ and M-MΦ. LPS amplified C5a- induced Gαi/c-Raf/MEK/ERK signalling in macrophages but not in monocytes. This synergy was independent of IL10, PI3K, p38, JNK, GM-CSF and M-CSF. C5a-mediated suppression of IL6 and TNF did not compromise but instead enhanced the clearance of Salmonella Typhimurium from macrophages. These findings implicated C5aR as a regulatory switch that modulates TLR4 signalling via the Gαi/c-Raf/MEK/ERK signalling axis in human macrophages but not in human monocytes. Chapter Six summarises all key findings in chapters 2-5, which represent important new knowledge in the field of complement C5a-C5aR and their effects on primary human monocytes and macrophages. Advances made in this thesis are specifically highlighted, discussed in relation to their novelty and significance and differences from the literature, and possible future research directions that build on the findings in this thesis are also outlined in relation to the fields of immunology and pharmacology.

III Declaration by author

This thesis is composed of my original work, and contains no material previously published or written by another person except where due reference has been made in the text. I have clearly stated the contribution by others to jointly-authored works that I have included in my thesis.

I have clearly stated the contribution of others to my thesis as a whole, including statistical assistance, survey design, data analysis, significant technical procedures, professional editorial advice, and any other original research work used or reported in my thesis. The content of my thesis is the result of work I have carried out since the commencement of my research higher degree candidature and does not include a substantial part of work that has been submitted to qualify for the award of any other degree or diploma in any university or other tertiary institution. I have clearly stated which parts of my thesis, if any, have been submitted to qualify for another award.

I acknowledge that an electronic copy of my thesis must be lodged with the University Library and, subject to the General Award Rules of The University of Queensland, immediately made available for research and study in accordance with the Copyright Act 1968.

I acknowledge that copyright of all material contained in my thesis resides with the copyright holder(s) of that material. Where appropriate I have obtained copyright permission from the copyright holder to reproduce material in this thesis.

IV Publications during candidature

Peer-reviewed papers

J. Lim, A. Iyer, L. Liu, J. Y. Suen, R. J. Lohman, V. Seow, M. K. Yau, L. Brown, and D. P. Fairlie. 2013. Diet-induced obesity, adipose inflammation, and metabolic dysfunction correlating with PAR2 expression are attenuated by PAR2 antagonism. FASEB J. 27: 4757-4767.

D. M. Hohenhaus, K. Schaale, K. A. Le Cao, V. Seow, A. Iyer, D. P. Fairlie, and M. J. Sweet. 2013. An mRNA atlas of G protein-coupled receptor expression during primary human monocyte/macrophage differentiation and lipopolysaccharide-mediated activation identifies targetable candidate regulators of inflammation. Immunobiology 218: 1345-1353.

V. Seow, J. Lim, A. Iyer, J. Y. Suen, J. K. Ariffin, D. M. Hohenhaus, M. J. Sweet, and D. P. Fairlie. 2013. Inflammatory responses induced by lipopolysaccharide are amplified in primary human monocytes but suppressed in macrophages by complement protein c5a. J. Immunol. 191: 4308- 4316.

Vesey, D. A., J. Y. Suen, V. Seow, R. J. Lohman, L. Liu, G. C. Gobe, D. W. Johnson, and D. P. Fairlie. 2013. PAR2-induced inflammatory responses in human kidney tubular epithelial cells. Am. J. Physiol. Renal Physiol. 304: F737-750.

Lim, J., A. Iyer, J. Y. Suen, V. Seow, R. C. Reid, L. Brown, and D. P. Fairlie. 2013. C5aR and C3aR antagonists each inhibit diet-induced obesity, metabolic dysfunction, and adipocyte and macrophage signaling. FASEB J. 27: 822-831.

V Publications included in this thesis

Publication citation – incorporated as Chapter 5.

Contributor Statement of contribution Vernon Seow (Candidate) Designed experiments (80%) Wrote the paper (80%) Drs. Junxian Lim, Abishek Iyer, Jacky Y. Suen Designed experiments (15%) Wrote and edited paper (20%) Daniel M. Hohenhaus, Juliana K. Ariffin Designed experiments (5%) Drs. David P. Fairlie and Matthew J. Sweet Reviewed paper (100%)

Publication citation – incorporated as Chapter 6

Contributor Statement of contribution Vernon Seow (Candidate) Designed experiments (20%) Wrote the paper (10%) Drs. Junxian Lim, Abishek Iyer, Designed experiments (70%) Wrote and edited paper (80%) Drs. Robert C. Reid and Jacky Y. Suen Designed experiments (10%) Wrote and edited paper (10%) Drs. David P. Fairlie and Lindsay Brown Reviewed paper (100%)

VI Contributions by others to the thesis

Prof. David Fairlie and Dr. Jacky Suen have contributed scientific guidance and critical review of the projects; Mr. Adam Cotterell has contributed to the experimental design and conducted the in vivo experiments for Chapter 4 on Male Wistar rats. Dr. Robert Reid has contributed with synthesis of C5aR antagonists (3D53 and JJ47) for Chapter 4. Miss Annika Yau have contributed with synthesis of C5aR antagonist W54011 for Chapter 4 and radioligand [125I]-C5a receptor saturation binding assay for Chapter 3. Dr. Katia Nones has contributed with technical advice and data interpretations of cDNA microarray experiment for Chapter 2 and 3.

Statement of parts of the thesis submitted to qualify for the award of another degree

None.

VII Acknowledgements

I would like to acknowledge the involvement of a host of people whom have made these four years possible. First of all, my principal supervisor Prof. David Fairlie whose scientific guidance and direction has kept the research on track. Thank you for your encouragement and patience during the toughest moments of my PhD and also the endless drafts that you have edited that made the completion of this thesis possible. To Dr Jacky Suen, my co-supervisor, thank you for your suggestions that have helped made the project more interesting and satisfying. Your intolerance to stupidity has made me a better scientist and a critical problem solver. I thank you for your friendship and the warmness that your family and you have given to mine. I enjoyed the exchange of knowledge and experience we both had being newly “minted” daddies. The kindness and generosity I have seen in you and Noel will definitely rub on Anabel and Zoe. They are such a lovely pair. To my dearest wife, Stephanie Tay, thank you for being the most understanding and supportive soul mate anyone can ask for. Thank you for dropping everything in Singapore and be with me in Australia. You being here with me has definitely helped made my time doing PhD easier. To my precious Elizabeth (Liz), the day you arrived has made me the happiest person in the world. Although the combination of being a dad and a PhD student was overwhelming at times, I would never trade a second spent with you for anything else. I have learnt so much from you and watching you grow up and become this sweet, intelligent and energetic girl you are now, has given me a whole new perspective to life. To my soon to be born son, James, I look afterward to your arrival in April. To my family members and In-laws in Singapore, thank you for your encouragement and support throughout these years. I love you all. To Dr James Lim, thank you for your friendship and the help that I very much needed in the lab. Your selflessness to volunteer help during weekends is admirable. Thank you for lending a hand whenever I needed one. Here goes: moving houses, multiple trips to the airport, watering my plants, collecting my letters while on holidays, baby sitting Liz (funny things you taught her that makes Stephanie mad), and other random stuff. You are a brilliant researcher and a great friend to have. Wishing you the very best for your career advancement in future. And I am not naming my son James in honour of you. Stephanie and I happen to really like the name. To Dr Annika Yau, thank you for your friendship and learning a listening ear whenever I am down. My time in the lab would not have been this fun and interesting without you. Although being in the crossfire between you and James are sometime quite intense, it has mostly been enjoyable. I must say I enjoyed your Malaysian humour the most. Thank you for your help with the chemistry aspect of my project, synthesizing whatever I requested and performing the binding assay. Really

VIII appreciated you for looking after Liz in the lab during the weekends. You know that she likes you when she addresses you as “Ah Yau”. I would like to thank the rest of the group-Fairlie members for your company and support in the lab, especially to Shiao (the gangster) and Sheila (the “Sheila”) for the times that we celebrated and stressed out to meet milestones datelines together, to Dr Abishek Iyer (Abi) for your friendship, support and advice on the lab experiments and writing, to Dr Robert Reid (Bob) for offering advice and support on the C5a project, to Wei Jun (Mr Daniel) for your friendship and my coffee buddy despite you hating the hot climate, to Adam (Rat guru) for your in vivo support to the project, to Mei (the sufferer) for enduring Shiao’s lameness and the unbelievable amount of noise and hype from Rink (you should seriously stop ingesting caffeine), to Johan (Mr President) for your friendship and always lightening up the mood in the biology lab (stop farting in the TC room - that’s how cells get contaminated), to Praveer (Kuku Bird) for being my gym buddy for a brief period of time (all the best to you in India).

Keywords gpcr, c5a receptor, c5a, tlr4, crosstalk, complement, human macrophages, human monocytes, c5a receptor agonist, c5a receptor antagonist

Australian and New Zealand Standard Research Classifications (ANZSRC)

ANZSRC code: 060110, Receptors and Membrane Biology, 20% ANZSRC code: 060405, Gene Expression, 40% ANZSRC code: 110707, Innate Immunity, 40%

Fields of Research (FoR) Classification

FoR code: 1107, Immunology, 50% FoR code: 1108, Medicinal Microbiology, 50%

IX Table of Contents

Abstract ...... II Declaration by author ...... IV Publications during candidature ...... V Publications included in this thesis ...... VI Contributions by others to the thesis ...... VII Statement of parts of the thesis submitted to qualify for the award of another degree VII Acknowledgements ...... VIII Keywords ...... IX Australian and New Zealand Standard Research Classifications (ANZSRC) ...... IX Fields of Research (FoR) Classification ...... IX Table of Contents ...... X List of Figures ...... XV List of Tables ...... XVIII List of Abbreviations used in this thesis ...... XIX

X Chapter 1 Properties of the C5a Receptor on Human Macrophages

1.1 Abstract ...... 2 1.2 Introduction ...... 3 1.3 The complement system ...... 4 1.4 Complement C5a receptors and signalling properties ...... 6 1.4.1 C5aR ...... 6 1.4.2 C5L2, the second C5a receptor ...... 8 1.5 Complement component C5a ...... 9 1.6 Ligands of C5aR ...... 11 1.6.1 C5aR agonists ...... 12 1.6.2 C5aR antagonists ...... 13 1.7 New insights into C5aR-related biological functions and diseases ...... 18 1.7.1 Interaction of RP S19 Oligomers with C5aR ...... 19 1.7.2 An unsuspected role for C5aR in cancer-related inflammation ...... 20 1.7.3 C5aR crosstalking with TLRs ...... 22 1.7.4 Linking C5aR signalling to obesity and metabolic dysfunction ...... 24 1.8 Mononuclear phagocytic system ...... 25 1.8.1 Macrophage infiltration into adipose tissue ...... 27 1.8.2 Macrophage activation and polarization ...... 29 1.8.3 Understanding human macrophage biology from a mouse perspective ...... 30 1.8.4 Differentiating primary human monocytes with GM-CSF or M-CSF to generate human macrophages ...... 32 1.9 Thesis objectives ...... 34 1.9.1 Chapter two ...... 35 1.9.2 Chapter three ...... 35 1.9.3 Chapter four ...... 35 1.9.4 Chapter five ...... 36 1.9.5 Chapter six ...... 36 1.10 References ...... 37

Chapter 2 Profiling human monocyte-derived GM-CSF and M-CSF differentiated macrophages: In vitro models for ‘M1-like’ and ‘M2-like’ macrophages

XI 2.5.2 Temporal study of monocyte-derived macrophage response to LPS stimulation 80 2.5.3 Transcriptional difference between GM-MΦ and M-MΦ by cDNA microarray analysis ...... 83 2.5.4 Pathway analysis of differentially expressed genes between GM-MΦ and M-MΦ 86 2.5.5 Cellular movement and actin cytoskeleton in GM-MΦ and M-MΦ ...... 87 2.5.6 GM-MΦ and M-MΦ have similar GPCR signalling profile except for Gα12/13 signalling ...... 89 2.5.7 Inflammatory responses in GM-MΦ and M-MΦ ...... 90 2.5.8 Monocyte-derived macrophages cytokine and gene responses to LPS stimulation 92 2.6 Discussion ...... 94 2.7 Conclusions ...... 95 2.8 References ...... 96

Chapter 3 Responses to human C5a by human monocyte-derived macrophages differentiated with either GM-CSF or M-CSF

3.1 Abstract ...... 100 3.2 Introduction ...... 101 3.3 Aims ...... 102 3.4 Methods ...... 103 3.4.1 Reagents ...... 103 3.4.2 Isolation of human monocytes and macrophages ...... 103 3.4.3 Intracellular calcium mobilization assay ...... 103 3.4.4 Receptor saturation binding assay ...... 104 3.4.5 Quantification of ERK1/2 phosphorylation by AlphaScreen Surefire ...... 104 3.4.6 In vitro migration assay (modified Boyden chamber) ...... 105 3.4.7 Enzyme-linked immunosorbent assay (ELISA) ...... 105 3.4.8 RNA extraction and reverse transcription ...... 105 3.4.9 Real-time quantitative polymerase chain reaction (real-time PCR) ...... 105 3.4.10 Hybridization and analysis of human array-based gene expression ...... 107 3.4.11 Statistical analysis ...... 107 3.5 Results ...... 108 3.5.1 C5a has similar agonist potency on GM-MΦ and M-MΦ ...... 108 3.5.2 GM-MΦ and M-MΦ have similar C5aR expression levels ...... 110 3.5.3 C5a-mediated chemotaxis in GM-MΦ and M-MΦ ...... 112 3.5.4 C5a-induced gene expressions on GM-MΦ and M-MΦ ...... 115 3.5.5 Quality control of genes hits: sample variations versus C5a treatments ...... 116 3.5.6 Microarray experiment 1: Genes regulated by C5a on GM-MΦ versus M-MΦ 121 3.5.7 Microarray experiment 2: Temporal profile of genes mediated by C5a on M-MΦ 129 3.6 Discussion ...... 139 3.7 Conclusion ...... 141 3.8 References ...... 142

XII Chapter 4 A Comparison of Three potent antagonists of C5aR: Importance of Receptor Residence Time

4.1 Abstract ...... 150 4.2 Introduction ...... 152 4.3 Aims ...... 156 4.4 Methods ...... 156 4.4.1 Reagents ...... 156 4.4.2 Animals ...... 157 4.4.3 Isolation of human monocytes and macrophages ...... 157 4.4.4 Isolation of rat bone marrow cells and macrophages ...... 158 4.4.5 Intracellular calcium mobilization assay ...... 158 4.4.6 Quantification of ERK1/2 phosphorylation by AlphaScreen Surefire ...... 159 4.4.7 In vitro migration assay (modified Boyden chamber) ...... 159 4.4.8 RNA extraction and reverse transcription ...... 159 4.4.9 Real-time quantitative polymerase chain reaction (real-time qPCR) ...... 160 4.4.10 Statistical analysis ...... 160 4.5 Results ...... 161 4.5.1 3D53, a potent and non-competitive C5aR antagonist ...... 161 4.5.2 3D53 has a longer residence time than W54011 and JJ47 ...... 163 4.5.3 C5a-induced chemotaxis is efficiently blocked by 3D53 but not W54011 and JJ47 164 4.5.4 C5aR-PA is a selective peptidic agonist of C5aR at low micromolar concentration ...... 166 4.5.5 C5aR-induced ERK phosphorylation in human and rat macrophages ...... 168 4.5.6 Antagonism of C5a- and C5aR-PA- induced gene expressions in human and rat macrophages ...... 169 4.5.7 Comparing 3D53, W54011 and JJ47 in C5aR-PA-induced rat paw oedema ... 174 4.6 Discussion ...... 176 4.7 Conclusions ...... 181 4.8 References ...... 182

Chapter 5 Inflammatory responses induced by LPS are amplified in human monocytes but suppressed in human macrophages by C5a

XIII 5.5.1 C5a acting via C5aR differentially modulates LPS-signalling in human CD14+ monocytes versus macrophages...... 196 5.5.2 Temporal study of C5aR/TLR4 crosstalk in HMDMs ...... 197 5.5.3 C5a suppresses LPS-induced IL6 and TNF via Gαi- and MEK-dependent signalling...... 200 5.5.4 C5a and LPS synergize for Raf/MEK/ERK phosphorylation in macrophages 203 5.5.5 C5a and LPS synergise in phosphorylation of ERK1/2 but not p38 or JNK in macrophages spacing ...... 203 5.5.6 C5a-mediated suppression of LPS-induced IL6 and TNF is IL10- and PI3K- independent...... 206 5.5.7 C5a attenuation of LPS-induced IL6 is selectively TLR4-mediated in macrophages ...... 209 5.5.8 C5a enhances clearance of Salmonella Typhimurium from HMDM ...... 211 5.6 Discussion ...... 212 5.7 Conclusions ...... 215 5.8 References ...... 216

Chapter 6 Summary of thesis

6.1 Thesis summary ...... 223 6.2 References ...... 232

XIV List of Figures

Figure 1.1. Schematic representation of the cells involved in innate and adaptive immunity. Adapted from (3)...... 3 Figure 1.2. Schematic diagram of the three main pathways involved in complement activation...... 4 Figure 1.3. G-proteins contain three subunits (α, β and γ), the Gα subunit binds to a GDP molecule...... 7 Figure 1.4. C5a peptide sequence alignment from various species...... 9 Figure 1.5. NMR solution structure of C5a at pH 5.2 (PDB ID code: 1KJS)...... 10 Figure 1.6. Modelled interaction between 3D53 (green) and human C5aR (orange)...... 14 Figure 1.7. C5aR peptidic antagonists. (A) PMX53, (B) PMX205 and (C) JPE-1375...... 15 Figure 1.8. C5aR non-peptidic anatagonists. (A) W54011, (B) NDT9520492 and aniline- substituted tetrahydroquinoline (C) JJ46 and (D) JJ47...... 17 Figure 1.9. Comparative modelling between human C5a and RP S19 protein...... 20 Figure 1.10. Macrophages localize to crown-like structure (CLS) around individual adipocytes, which increase in frequency in obesity. Reference from (239)...... 28 Figure 1.11. Colour wheel of macrophage activation. Reference from (226)...... 30

Figure 2.1. GM-CSF and M-CSF are required for monocytes survival and differentiation. .. 79 Figure 2.2. Macrophage markers CD68 and EMR1 expression levels in GM-MΦ and M-MΦ...... 80 Figure 2.3. Temporal study of IL6, TNF, MCP1 and IL10 cytokine production in LPS activated GM-MΦ and M-MΦ...... 82 Figure 2.4. Heat map representation of genes differentially expressed in GM-MΦ and M- MΦ that were ≥ 5 fold changed...... 85 Figure 2.5. Heatmap representative of an overview for biological functions and processes differentially expressed between GM-MΦ and M-MΦ...... 86 Figure 2.6. GM-MΦ migrates faster than M-MΦ in the Boyden chamber...... 87 Figure 2.7. GM-MΦ has more basal actin polymerization than M-MΦ...... 89 Figure 2.8. Expression analysis of G-protein coupling receptor signalling pathway-linked genes between GM-MΦ and M-MΦ...... 90 Figure 2.9. Expression analysis of human cytokine signalling and production pathway- linked genes between GM-MΦ and M-MΦ...... 91 Figure 2.10. LPS-induced IL6, IL8, TNF and IL10 proteins and gene expressions...... 93

Figure 3.1. C5a-induced intracellular calcium mobilization in HMDM...... 109

XV Figure 3.2. C5a-induced ERK1/2 phosphorylation in HMDM...... 110 Figure 3.3. GM-MΦ and M-MΦ have similar C5AR gene expression...... 111 Figure 3.4. GM-MΦ and M-MΦ have similar cell surface C5aR expressions...... 112 Figure 3.5. C5a is equipotent in mediating chemotaxis on GM-MΦ and M-MΦ...... 114 Figure 3.6. A panel of genes induced by 30 nM C5a after 0.5 h...... 116 Figure 3.7. Hierarchical clustering of samples used for the microarray analysis...... 118 Figure 3.8. mRNA quality and sample assessment report for sample X...... 119 Figure 3.9. Comparison of density plot of intensity between (A) raw and (B) normalized data...... 120 Figure 3.10. Venn diagram comparison between genes hits obtained using 4 donors versus 3 donors treated with 30 nM C5a for 0.5 h in M-MΦ...... 121 Figure 3.11. Venn diagram comparison of genes upregulated by 30 nM C5a for 0.5 h on GM- MΦ versus M-MΦ...... 124 Figure 3.12. Functionally related gene network with highest score constructed from genes commonly upregulated by 30 nM C5a on GM-MΦ and M-MΦ...... 127 Figure 3.13. Correlation between microarray and quantitative real-time PCR FOSB gene expression upregulated by 30 nM C5a on GM-MΦ and M-MΦ...... 128 Figure 3.14. Temporal profile of 147 genes differentially regulated by 30 nM on M-MΦ. 130 Figure 3.15. Four clusters of genes were categorised based on their responses to 30 nM C5a treatments and the duration the event occurred...... 131 Figure 3.16. C5a can affect genes involved in inflammation, metabolism, signaling, chemotaxis and cell cycle on M-MΦ...... 137 Figure 3.17. Correlation between microarray and real-time PCR expression...... 138

Figure 4.1. Mechanism of C5aR antagonism for 3D53, W54011 and JJ47 against human C5a...... 162 Figure 4.3. 3D53 has a longer residence time on human C5aR compared to W54011 and JJ47...... 164 Figure 4.4. Comparative antagonism of human macrophage migration (induced by 3 nM rhC5a) by 0.1 μM and 1 μM 3D53, W54011 and JJ47...... 165 Figure 4.5. C5aR-PA is a selective C5aR peptide agonist at low μM concentration...... 167 Figure 4.6. C5a-induced ERK phosphorylation in human and rat macrophages...... 169 Figure 4.7. Comparative antagonism of C5a-induced gene expressions in human macrophage by 3D53, W54011 and JJ47 at 1h versus 16 h post-treatment...... 171 Figure 4.8. Comparative antagonism of C5aR-PA-induced gene expressions in human macrophage by 3D53, W54011 and JJ47 at 1h versus 16 h post-treatment...... 172 Figure 4.9. Comparing C5aR-PA- and C5a- induced gene expressions in rat macrophage and antagonism of C5a responses by 3D53, W54011 and JJ47 at 1h post-treatment...... 173

XVI Figure 4.10. Evaluating antagonism of 3D53, W54011 and JJ47 on C5aR-PA-induced paw oedema in male Wistar rats (courtesy of Mr Adam Cotterell, University of Queensland)...... 175 Figure 4.11. Amino acid sequence comparison of human and rat C5aR and C5a...... 180

Figure 5.1. C5a acting via C5aR-dose-dependently modulates LPS-induced IL6, TNF and IL10 cytokine production in human monocytes and macrophages (GM-MΦ and M- MΦ) via C5aR...... 197 Figure 5.2. C5AR and TLR4 mRNA levels are increased during differentiation of CD14+ monocytes to macrophages with GM-CSF or M-CSF...... 198 Figure 5.3. Temporal study of C5aR/TLR4 crosstalk in monocytes and macrophages...... 199 Figure 5.4. C5a suppression of LPS-induced IL6 and TNF in macrophages is sensitive to PTX and MEKi (U0126) treatment...... 201 Figure 5.5. Signalling pathway analysis of C5aR/TLR4 crosstalk in monocytes and macrophages...... 202 Figure 5.6. C5a induced Raf/MEK/ERK phosphorylation is amplified in macrophages but not in monocytes by LPS...... 204 Figure 5.7. C5a-induced ERK1/2 phosphorylation is amplified in macrophages but not monocytes by LPS...... 205 Figure 5.8 LPS-induced p38 and JNK phosphorylation is not affected by C5a in macrophages and monocytes...... 206 Figure 5.9. Anti-inflammatory IL10 cytokine suppressed LPS-induced IL6 and TNF production...... 207 Figure 5.10. C5a suppression of LPS-induced IL6 and TNF macrophages is IL10- independent...... 208 Figure 5.11. C5a suppression of LPS-induced IL6 and TNF macrophages is PI3K- independent...... 209 Figure 5.12. Suppression by C5a of inducible IL6 production from macrophages is selective for LPS responses...... 210 Figure 5.13. C5a decreases S. Typhimurium survival in M-MΦ...... 211

Figure 6.1. C5adesArg did not suppress LPS-induced IL6 and TNF production...... 228 Figure 6.2. C5aR antagonist modulates inflammatory responses in RAW264.7 macrophages...... 229 Figure 6.3. Antagonists of C5aR and C3aR can modulate lipid homeostasis in 3T3-L1 adipocytes...... 230

XVII List of Tables

Table 1.1. Disorders associated with complement activation (17)...... 6 Table 1.2. Pathologic conditions involving anaphylatoxin C5a...... 11

Table 2.1. Correlation between microarray and real-time PCR gene expression ...... 85

Table 3.1. The experimental design carried out for the microarray experiment...... 117 Table 3.2. Genes upregulated by 30 nM C5a in either GM-MΦ or M-MΦ after 0.5 h were expressed as fold changes...... 122 Table 3.3. Genes commonly upregulated by 30 nM C5a in GM-MΦ and M-MΦ after 0.5 h were expressed as fold changes...... 123 Table 3.4. Top 5 network functions and associated diseases that were upregulated by C5a in GM-MΦ as analyzed by IPA ...... 125 Table 3.5. Top 5 network functions and associated diseases that were upregulated by C5a in M-MΦ as analyzed by IPA ...... 126 Table 3.6. Inflammatory genes upregulated (black) or downregulated (red) by 30 nM C5a...... 132 Table 3.7. Metabolism- or biosynthesis-related genes upregulated (black) or downregulated (red) by 30 nM C5a...... 133 Table 3.8. Signal transduction- or transcription-related genes upregulated (black) or downregulated (red) by 30 nM C5a...... 134 Table 3.9. Cell cycle-related genes upregulated (black) or downregulated (red) by 30 nM C5a...... 135 Table 3.10. Adhesion- or migration-related genes upregulated (black) or downregulated (red) by 30 nM C5a...... 136

Table 4.1. Reported activity of 3D53, W54011 and JJ47 on human C5aR...... 155 Table 4.2. Comparison of rule-of-five parameters of 3D53, W54011 and JJ47 ...... 155

XVIII List of Abbreviations

125I Iodine (isotope 125) AAM Alternatively activated macrophages Ac Acetyl AGRF Autralian Genome Research Facility ANOVA Analysis of variance AT Adipose tissue ATM Adipose tissue macrophages B cells B-lymphocytes BH4 tetrahydrobiopterin BMC Bone marrow cells BMM Bone marrow-derived macrophages C5a Complement component 5a C5aR C5a receptor C5L2 C5aR-like receptor 2 Ca Calcium CAM Classically activated macrophages cAMP cyclic adenosine monophosphate CD Cluster of differentiaiton CLS Crown-like structure DAG Diacylglycerol DC Dendritic cells DIO Diet-induced obesity DMSO Dimethylsulfoxide DNA Deoxy ribonucleic acid EC Effective concentration ECL Extracellular loop ELISA Enzyme-linked immunosorbent assay ERK Extracellular signal-regulated kinase FBS Fetal bovine serum FLIPR Fluorescent imaging plate reader GDP Guanosine diphosphate GM-CSF Granulocyte-macrophage colony-stimulating factor GPCR G protein coupled receptor GTP Guanosine triphosphate HBSS Hank's balanced salt solution HMC-1 Human mast cell line 1 HMDM Human monocyte-derived macrophages HUVEC Human umbilical vein endothelial cells i.pl Intraaplantar IC Inhibitory concentration IFN Interferon Ig Immunoglobulin

XIX IL Interleukin iNOS inducible NO synthase IPA Ingenuity pathway analysis IPKB Ingenuity pathway knowledge base IκB Nuclear factor kappa-light-polypeptide gene enhancer of B cells inhibitor LPS Lipopolysaccharide M-CSF Macrophage colony-stimulating factor MAC Membrane attack complex MAPK Mitogen-activated protein kinase MDSC Myeloid-derived suppressor cells MGC Multinucleate cells MMEC Mouse microvascular endothelial cells MOI Multiplicity of infection MΦ Macrophages NFκB Nuclear factor kappa-light-chain-enhancer of activated B cells NK Natural killer cells nM Nano molar NMR Nuclear magnetic resonance NO Nitric oxide p.o per oral administration PA Peptidic agonist PAMP Pathogen-associated molecular pattern PBMC Peripheral blood mononuclear cells PCR Polymerase chain reaction PI3K Phosphatidylinositol 3-kinase pM picomolar PTX Pertusis toxin R Receptor RAEC Rat alveolar epithelial cells RNA Ribonucleic acid RT-PCR Reverse transcription PCR SAR Structure activity relationship SEM Standard error of means T cells T-lymphocytes t1/2 Half life TLR Toll-like receptor TNF Tumour necrosis factor µM Micro molar

XX Chapter

1 Properties of the C5a Receptor on Human Macrophages

1 1.1 Abstract

The complement system of around 30 plasma proteins is a central component of immunity, bridging innate and adaptive immune responses to infection and injury. If complement activation is prolonged or dysregulated, it can also turn become destructive to host cells and lead to numerous pathological conditions and diseases. Overproduction or underregulation of the potent proinflammatory and chemotactic protein, complement component C5a, has been implicated in numerous immune and inflammatory conditions. It is specifically recognized by the complement receptor C5aR, a G Protein Coupled Receptor (GPCR) on the surface of immune and other cells, and C5aR is a promising therapeutic target for complement modulation. Indeed GPCRs are the most prevalent signal-transducing cell surface proteins in pharmaceutical research. Of over 900 human GPCRs identified to date, more than one third are potential therapeutic targets for the treatment of diseases. Despite the importance of C5aR to drug discovery and inflammatory diseases and there are no antagonists of this protein in the marketplace. Although C5aR is widely expressed throughout the body, it is highly expressed in myeloid cells such as macrophages. This chapter summarizes the roles of C5a activation on human macrophages, briefly surveying C5a, C5aR and some important small molecule analogues of C5a that had been developed to study roles for C5a in vivo.

2 1.2 Introduction

The immune system incorporates both cellular and molecular mechanisms to distinguish between host cells and foreign organisms or materials (1, 2). There are two main types of immunity: (a) innate immunity (unspecific) and (b) acquired immunity (specific). Innate immunity does not require previous exposure to, or memory of pathogens or tumours and among many components, includes physical barriers, mucous membranes, antimicrobial substances (e.g. lysozyme, peroxidase, histones and other inflammatory mediators), the complement system, phagocytes (e.g. neutrophils, macrophages) and natural killer (NK) cells. On the other hand, specific immunity can be acquired by infection or vaccination (3). Acquired immunity is usually classed into two subtypes: (a) the humoral response where B-lymphocytes (B- cells) produce specific antibodies targeting foreign materials, and (b) cell-mediated responses where T-lymphocytes (T-cells) either help in maturing/priming other immune cells (B cells, dendritic cells (DC) and macrophages) or in destroying infected and/or tumour cells (Figure 1.1).

Figure 1.1. Schematic representation of the cells involved in innate and adaptive immunity. Adapted from (3).

3 1.3 The complement system

The complement system serves as an important effector of both innate and acquired immunity. Comprising over 30 soluble plasma, and insoluble membrane- bound, proteins, the complement system acts in a wide variety of host defence, inflammatory, homeostatic and immune responses (4). Complement can be activated through at least three distinct pathways, (a) the classical pathway – initiated by antigen-antibody complexes; (b) the alternative pathway – initiated by bacteria and yeast cell walls; (c) the lectin pathway – initiated similarly to the classical pathway but in the presence complex polysaccharides and the absence of complement component 1q (C1q) (Figure 1.2). There is also mounting evidence that proteolytic enzymes (e.g. thrombin, trypsin) normally associated with other cascading networks are also be capable of activating complement (5-8).

Figure 1.2. Schematic diagram of the three main pathways involved in complement activation. Classical, alternative and lectin pathways and factors involved in complement activation are shown. Two isoforms of C3 convertases are highlighted in the blue boxes and the two isoforms of C5 convertases are highlighted in the green box. C3a and C5a anaphylatoxins are highlighted in the brown and red boxes respectively (9).

4 The complement system contributes to opsonisation (tagging) and killing of microorganisms and foreign bodies, and the induction of adaptive immunity to protect the host from pathogens and infected or damaged cells. The activation of complement involves a series of initiation, amplification and proteolytic reactions to sequentially activate proteins and enzymes down the cascade (10, 11). The complement system is regulated at multiple levels, both temporally as well as spatially depending on which component is activated. For example, amplification of complement activation via the alternative pathway is regulated by the formation of the short-lived (t1/2 ~90 s) protease complex, C3 convertase (C3bBb) (10). There is also a second C3 convertase,

(C4b2a) that forms when the classical or lectin pathway is activated (t1/2 ~10 min) (12). C3 convertases cleave the third complement component protein (C3) into its active form C3b, exposing its thioester bond to permit covalent attachment to nearby surfaces (13). The process must be tightly regulated to prevent non-specific attack of host cells, either by accelerating the decay of C3 convertases or by degrading existing C3b (to prevent formation of new C3 convertase) (14). C5 convertase is formed when free C3b binds to C3 convertase. C5 convertases can also exist in two isoforms, depending on the origin of C3 convertase. C5 convertases cleave C5 to produce C5b, and release the proinflammatory chemoattractant C5a that can interact with C5a receptors, which are highly expressed on immune cell surfaces. C5b then complexes with C6, C7, C8 and C9 to form the membrane attack complex (MAC) (15) responsible for initiating cell/pathogen lysis. There is a very delicate balance that must be maintained whenever the complement cascade is activated because of its destructive nature. Despite the stringent and multiple levels of self-regulatory mechanisms, excessive complement activation can occur because the rate of activation may exceed the capacity of host regulatory mechanisms (16). In many diseases, while complement activation may not be the primary initial disease-inducing culprit, tissues damage in certain conditions is clearly complement-driven. Some of the disorders known to be associated with prolonged or aberrant complement activation are briefly covered in (Table 1.1) (17).

5 Table 1.1. Disorders associated with complement activation (17).

Classification Disorder Acute Adult respiratory distress syndrome Ischemia-reperfusion injury: Myocardial infarct Lung inflammation Hyperacute rejection (transplantation) Sepsis Cardiopulmonary bypass Burns, wound healing Asthma Restenosis Multiple organ dysfunction syndrome Trauma, hemorrhagic shock Guillain-Barré syndrome Chronic Paroxysmal nocturnal hemoglobinuria Glomerulonephritis Systemic lupus erythematosus Rheumatoid arthritis Infertility Alzheimer’s disease Organ rejection (transplantation) Fibrotic disorders Multiple sclerosis Biomaterials Platelet storage incompatibility Hemodialysis Cardiopulmonary bypass equipment

1.4 Complement C5a receptors and signalling properties

1.4.1 C5aR

C5a acts through a specific receptor named C5a receptor (C5aR/C5AR1 or CD88). C5aR belongs to the seven membrane-spanning receptor superfamily and is functionally coupled to a pertussis toxin (PTX)-sensitive G-protein (18-20). The seven transmembrane receptors are also referred to as G protein-coupled receptors (GPCRs). Most GPCRs undergo conformational changes upon binding to their native agonist ligand, which activates an associated G-protein heterotrimer (a complex of α-, β- and γ-subunits) by exchanging its bound guanosine diphospate (GDP) for a guanosine triphosphate (GTP). The G-protein α-subunit (Gα), together with a bound

6 GTP dissociates from the dimer βγ-subunit and triggers a series of intracellular signals or target proteins directly depending on the subunit type (Figure 1.3).

Figure 1.3. G-proteins contain three subunits (α, β and γ), the Gα subunit binds to a GDP molecule. When a G-protein binds an activated receptor, the GDP-binding subunit undergoes a subtle conformational change, and exchanges its GDP nucleotide by a GTP. GTP- bound Gα subunits have different downstream effects depending which type of Gα is involved (21).

Cellular stimulation of C5aR through pertussis toxin (PTX)-sensitive Gαi2,

Gαi3, or PTX-insensitive Gα16, results in intracellular calcium mobilisation and activation of signalling pathways including phosphatidylinositol 3-kinase (PI3K), diacylglycerol (DAG), mitogen-activated protein kinase (MAPK), extracellular signal-regulated kinase (ERK) and other downstream signalling proteins (22-24). C5a can stimulate the synthesis and release of proinflammatory cytokines such as TNF, IL1β, IL6 and IL8 in human leukocytes (25, 26). C5aR expression is upregulated in mouse microvascular endothelial cells (MMEC) and rat alveolar epithelial cells (RAEC) by inflammatory mediators, such as lipopolysaccharide (LPS) or IL6 (27, 28). C5a has also been shown to exert strong synergistic effects with LPS for TNF, macrophage inflammatory protein-2 (MIP2), cytokine-induced neutrophil chemoattractant-1 (CINC1) and IL1B in RAEC (27). Human umbilical vein endothelial cells (HUVEC) constitutively express C5aR and, when stimulated with C5a at nanomolar (nM) concentrations, increase the mRNA expression of IL8, IL1B and RANTES in a time- and dose- dependent manner (29). C5a strongly decreases IL12 production in human monocytes treated with LPS

7 or interferon gamma (IFNG) (30). The evidence suggests that the C5a-C5aR signalling axis can produce cytokines and chemokines in many immune cell types. However, C5a-C5aR signalling can lead to completely opposite outcomes depending on the cell type. For example, C5a strongly suppresses LPS-induced TNF production in neutrophils but evidently shows the opposite effect in alveolar macrophages (31, 32). Neutrophils increase cytosolic content of IκBα (nuclear factor of kappa light polypeptide gene enhancer in B-cells inhibitor alpha) when stimulated with C5a, which inhibits NF-κB (nuclear factor kappa light chain enhancer of activated B cells) activation (32). On the other hand, NF-κB activation occurred within 15 min after C5a stimulation of monocytes and macrophages (31). The ability of C5a to simultaneously up- and down-regulate NF-κB dependent expression of genes such as TNF, IL6 and IL8, suggests that C5a-C5aR signaling is not a simple “on or off” switch model but is more complex and can be different in different cell types.

1.4.2 C5L2, the second C5a receptor

In 2000, Ohno et al. identified a second C5aR-like GPCR called C5L2 (GPR77) (33). There is a 38% sequence homology between C5aR and C5L2 (34). C5L2 binds to C5a with similar affinity as C5aR, but binds to C5adesArg with higher avidity compared to C5aR (35, 36). However, C5L2 does not fit neatly with the perceived functions of C5a, generating controversy in the literature. C5L2 was initially thought to be a non-signalling receptor, as binding to C5a did not stimulate mobilisation of intracellular calcium, extracellular signal-related kinase (ERK) phosphorylation or receptor internalisation, in contrast to C5aR (33, 37, 38). It has been suggested that the inability of C5L2 to signal is due to a mutation in a highly conserved ‘DRY motif’ (Asp-Arg-Tyr) in many GPCRs, located at amino acid residues 137-139 near the third transmembrane domain. C5L2 has instead a ‘DLC motif’ (Asp-Leu-Cys) and when the leucine was substituted with an arginine residue, C5a could induce a small increase in intracelluluar calcium levels in 293T cells co- expressing Gα16, suggesting that the central leucine is important for G protein uncoupling in C5L2 (36). Mice lacking C5L2 were found to have enhanced responses to both C5a and C5adesArg (38). A rat model of sepsis showed an increase in C5L2 expression and blocking C5L2 with an anti-C5L2 antibody substantially increased the

8 concentration of the inflammatory cytokine IL6 (39). Interestingly, immune complex- induced lung injury was attenuated in C5L2-/- mice (40), whereas another group described the opposite phenotype showing that C5L2-/- mice exhibited intensified immune complex-induced lung injury (38). The role of C5L2 in inflammatory responses is still highly controversial and at this time remains poorly understood.

1.5 Complement component C5a

Activation of the complement system also generates potent chemoattractants and markers for immune clearance that covalently attach to cell surfaces. For example, C3a, C4a and C5a are anaphylatoxins that are released during classical and lectin pathway activation, while C3a and C5a are also released during alternative pathway activation. Among these three polypeptides, C5a is the most potent anaphylatoxin followed by C3a then C4a (41, 42). Besides being the most potent, C5a has also been the most intensively studied of the anaphylatoxins (43). C5a is a polypeptide, which contains 74-79 amino acids depending on the species type (Figure 1.4). The NMR solution structure of hC5a was first solved in 1989, but ten residues at the C-terminus were highly disordered (44). A subsequent NMR structure with better resolutions reported by Zhang et al. showed that the human C5a structure is arranged into an anti- parallel 4-helix bundle (1st helix4-12; 2nd helix18-26; 3rd helix32-39 and 4th helix46-63), being stabilized by three disulphide bonds (Cys21-Cys47, Cys22-Cys54 and Cys34-Cys55) and connected by three loop segments (1st loop13-17, 2nd loop27-33 and 3rd loop40-45) (45). The C-terminal residues69-74 form a helical turn conformation that brings Arg 74 adjacent to Arg 62 (45). This C-terminal helical turn conformation (red) is critical for receptor activation (Figure 1.5).

Figure 1.4. C5a peptide sequence alignment from various species. C5a(s) were aligned using ClustalW2 (46). Conserved residues are colored according to their physicochemical properties (http://www.ebi.ac.uk/tools/clustalw2). Adapted from (47).

C5a anaphylatoxin has many biological properties, including mast cell degranulation, recruitment of immune cells to sites of inflammation (chemoattractants), smooth muscle contraction, vasodilation and stimulation of proinflammatory cytokines such as tumour necrosis factor alpha (TNF), interleukin-6 (IL6) and -1beta (IL1B) (48-52). The activity of C5a can be modulated by the enzyme, carboxypeptidase B. Serum carboxypeptidase B rapidly metabolises C5a to C5adesArg within seconds by removing the C-terminal arginine of C5a, which is important for binding to C5aR (53, 54). C5adesArg has a 10-1000 times reduced potency compared with C5a, depending on the function being measured and the cell type examined (55). The Kd of C5a and C5adesArg binding to C5aR is ~1 and 412- 660 nM respectively on bone marrow cells and monocytes (36, 56). It has now become apparent that uncontrolled or inappropriate production of C5a is specifically implicated in many inflammatory diseases as summarized in Table 1.2. The details and mechanisms underlying these diseases are investigated elsewhere and will not be discussed here.

Figure 1.5. NMR solution structure of C5a at pH 5.2 (PDB ID code: 1KJS). Human C5a is a 74 glycoprotein consisting of four alpha helices (helix 1, cyan; helix 2, orange; helix 3, pink and helix 4, blue) arranged in an anti-parallel orientation. The four-residue loop in the C-terminus (red) is critical for receptor activation (45).

10 Table 1.2. Pathologic conditions involving anaphylatoxin C5a.

Classification Pathologic condition Cardiovascular Fibrosis in hypertension (57) Atherosclerosis in ApoE-/- mice (58) Preclinical atherosclerosis associated with systemic lupus erythematosus (59) Arterial injury in atherosclerosis-prone mice (60) Hemodialysis-associated thrombosis (61) Human coronary lesions (62) Complement and coagulation cascades (63) Arthritis Inflammation in autoimmune arthritis (64) Inflammatory arthritis (65) Rheumatoid arthritis (RA) (66) Autoimmune arthritis (67) Renal Kidney graft survival (68, 69) Tubulointerstitial injury (70) Liver Fulminant hepatic liver failure (71) Pregnancy Preterm delivery (72) Age-related muscular T-cell cytokine release in AMD (73) degeneration Nervous Cerebral arteriovenous malformations (74) system Incisional pain (75) Nociceptive sensitization (76) Mitochondrial functions of PC12 cells (77) Infection Experimental periodontitis (78) Bone loss in P. gingivalis infection (79) Necrotizing enterocolitis (80) S. aureus bacteraemia (81) Cerebral malaria (82) Gram-negative bacteraemia and endotoxic shock (83) NKT and NK function in sepsis (84) Respiratory Chronic obstructive pulmonary disease (85) disease Experimental allergic asthma (86, 87) Modified from (47). CNS, central nervous system; NK, natural killer cell; NKT, natural killer T cell.

1.6 Ligands of C5aR

Although the mission to discover complement therapeutics started over 30 years ago (88), there are still no small molecule drugs in the marketplace for targeting any complement protein. C5aR has been demonstrated to be an important and promising

11 therapeutic target for complement modulation (89). An antibody that blocks the proteolysis of C5 to C5a and C5b (an early precursor to membrane attack complex) has been clinically validated and approved by the U.S. Food and Drug Administration (FDA) for treating paroxysmal nocturnal hemoglobinuria (90). However, proteins can have limited utility as therapeutics due to the expense to produce, difficulties to formulate and can trigger unwanted immunogenic side effects. The failure of the anti- C5 monoclonal antibody (mAb), pexelizumab, in some clinical studies for treating acute myocardial infarction was attributed to poor tissue penetration, leading to poor therapeutic efficacy. On the other hand, unlike proteins or antibodies, drug-like small molecules do not tend to possess these disadvantages for therapeutic invention. The idea of synthesizing the C-terminus of the human C5a as analogues to probe C5aR function started as early as in the 1970s. Peptide analogues of C5a containing residues 1-69, but lacking five residues of the C-terminal domain of native C5a, were discovered to bind to C5aR but failed to elicit activity at C5aR (91). Interestingly, the pentapeptide MQLGR (corresponding to C5a70-74) alone, or in combination with C5a1-69 was also inactive at C5aR (92). These results led to the hypothesis that C5a possesses an internal ‘recognition’ site in addition to the ‘activation’ site contained in the C-terminal region of C5a (91). Since then, intensive efforts has been made to design and develop better analogues of C5a, both agonists and antagonists of C5aR (89).

1.6.1 C5aR agonists

Sequent extension of the pentapeptide ‘MQLGR’ to the octapeptide ‘HKDMQLGR’ resulted in modest binding activity (93). Replacing the His residue with a Phe residue increased potency on C5aR (94). After conducting a series of SAR studies, the decapeptide EP-54 ‘YSFKPMPLdAR’ was later discovered by Sam Sanderson at the University of Nebraska with biological activity, but its agonist activity was still only in the µM concentration range (94). Although this decapeptide was a full agonist on human C5aR and had little or no activity on C5L2 (95), it also bound non-specifically to C3aR at µM concentrations (96). In contrast, EP-67 ‘YSFKDMP(Me)LdAR’ displayed weak agonist activity on C5aR and C3aR but had better activity on C5L2 (97-99). EP-54 and EP-67 were termed “response-selective”

12 as these peptide mimetics showed different responses at C5aR depending on the cell type (98), although these varying responses are probably due to receptor selectivity issues. Earlier structure activity studies led by researchers from Abbott Laboratories in the 1980s, produced analogues with higher affinity for human C5aR, while downsizing the peptide to six residues, N-Methyl-FKPdChaCha-dArg-OH, with nM activity against C5aR (100). Although this hexapeptide agonist was previously reported to bind with low affinity to C5L2 (101), but did not affect C5a binding to C5L2 even at high µM concentrations, more recent studies have shown otherwise (102). The selectivity of this peptide agonist for C5aR over C3aR was not investigated. First generation peptide agonists for C5aR were notorious for also binding to C3aR (96), despite the latter having only 37% sequence identity to C5aR (103). It is surprising that despite this notorious non-specific binding to C3aR, researchers developing C5a agonists have not investigated or disclosed whether or not their reported compounds were truly C5aR selective agonists. This has been a major hurdle in developing better C5aR agonists from uncertain starting points.

1.6.2 C5aR antagonists

Since the discovery of ‘N-Methyl-FKPdChaCha-dArg-OH’, further SAR studies have led to a family of hexapeptide analogues, ‘N-Methyl-FKPdCha(x)-dArg’, which can be agonists or antagonists depending on the properties of the fifth residue (x) (104). It was found that increasing the aromaticity at position (x) could lead to increased antagonist activity at C5aR (104). Interestingly, over a hundred peptide- activated GPCRs recognize ligands with a turn structure and the conformation of this β/γ turn motif may be crucial to the mechanism of receptor activation (105, 106). Binding of C5a to C5aR involves three binding sites (9). Potential agonists and antagonists for C5aR have been designed based on this region (107). The compound known as “3D53”, a hexapeptide containing a cyclic pentapeptide AcF[OPdChaWR] was the first cyclic peptidomimetics C5aR antagonist and proved to be a potent and water-stable full antagonist of C5aR on human neutrophils (108) and most other human cell types expressing C5aR. 3D53 was

13 created by PhD student Allan Wong in the Centre for Drug Design and Development (3D Centre), University of Queensland in 1996. Figure 1.6 shows the molecular interaction between the Arg, dCha, Trp and AcPhe component of the 3D53 binding sites fitting into the binding pockets within the helices of human C5aR. 3D53 has an antagonist potency IC50 20 nM against human C5a (100 nM) (109). A variety of in vitro assays were used to determine the antagonist properties of 3D53, including inhibition assays of myeloperoxidase release from human polymorphonuclear cells (PMNs), C5a-mediated chemotaxis of PMNs, C5a-induced human umbilical artery contractions and E.coli-induced oxidative burst and phagocytosis (107, 110-112). 3D53 was found to be a highly specific inhibitor of C5a mediated effects and had no cross reactivity with other GPCR ligands, such as formylmethionyl-leucyl- phenylalanie (fMLP), leukotriene B4 (LTB4), and platelet-activating factor (PAF) (107). More interestingly, 3D53 does not bind to C5L2, which makes it an invaluable tool to use to study the effects of C5aR without background contamination with C5L2 signals (101). Most importantly, 3D53 does not bind to C3aR (113).

Figure 1.6. Modelled interaction between 3D53 (green) and human C5aR (orange). Molecular modelling showed ligand binding pockets (left: side view, right: top view) with Arg, Trp, dCha and AcPhe components of 3D53 fitting between helices of C5aR. Adapted from Monk et al (89).

Eventually, 3D53 was licensed as PMX53 (Figure 1.7A) for clinical development by the now defunct Promics Ltd (subsequently passed on through a series of company takeovers to Peptech Pty Ltd, Arana Therapeutics Ltd, Cephalon

14 Inc. and then in 2011 to TEVA Pharmaceutical Industries Ltd, which is the world’s largest manufacturer of generic pharmaceuticals (114). PMX53 was safe and well tolerated in Phase I clinical trials, and 3D53 has been the most intensely evaluated small molecule C5aR antagonist in human clinical trials. PMX53 was found to exert long lasting effects even though it had low oral bioavailability (114). What this means is that even a once daily dose was sufficient to maintain an effective therapeutic window in rats (115). A more stable derivative, PMX205 (Figure 1.7B), was later developed with a hydrocinnamic acid group in the N-terminus of PMX53 (116). PMX205 has also been investigated in animal models of Huntington’s disease (117), Alzheimer’s disease (118) and amyotrophic lateral sclerosis (ALS) (119). JPE-1375 (Figure 1.7C) was developed as a linear analogue of PMX53 by substituting the Arg residue with a Phe residue and adding a hydroorotic acid (Hoo) in place of the N-acetyl cap (120). Although JPE-1375 and PMX53 has similar reported potency on C5aR, JPE-1375 exhibited high microsomal stability (121). Despite promising results in renal allograft survival (122), AMD (123), experimental tubulointerstitial fibrosis (124) and atherosclerotic plaque stabilization (60), JPE-1375 was discontinued early in development (125).

H H N N N N H H N O N O HN O O NH NH O O NH NH O O O O O O O NH NH N N H B H A NH NH HN NH2 HN NH2

O O H H N N N NH2 O N H H H O O O N N N O H HN O

O C NH2

Figure 1.7. C5aR peptidic antagonists. (A) PMX53, (B) PMX205 and (C) JPE- 1375.

Although these peptidic antagonists of C5aR were potent antagonists of C5aR, except for the cyclic peptide 3D53, they unfortunately also inherited some disadvantages of peptidic ligands for example, susceptibility to proteolytic cleavage in serum, low oral bioavailability as the result of hydrophilicity, and high molecular weights. To overcome problems associated with peptides, non-peptidic compounds have been sought with more drug-like properties such as higher oral bioavailability, cheaper synthesis, high stability in serum and resistance to enzymatic degradation. W54011 (Figure 1.8A) was developed by Mitsubishi Pharma, and was the result of an optimized series of substituted phenylguanidines and tetrahydronapthalene-based compounds (126). W54011 was reported to be a competitive non-peptidic C5aR antagonist. Although Sumichika et al. did not report the percent oral bioavailability of W54011, the authors claiming that this antagonist was potent and orally active in vivo (126). However, it is very hydrophobic and had problems with species specificity (active in human, cynomolgus monkey and gerbil but not as active in mouse, rat guinea pig, rabbit or dog neutrophils), which complicated pre-clinical studies (89). NDT9520492 (Figure 1.8B) was developed by Neurogen Corp and like

W54011, NDT9520492 was also species-specific with antagonistic activity of IC50 ~300 nM at human and gerbil but not mouse or rat C5aR (127). NGD2000-1 (chemical structure not disclosed), a derivative of NDT9520492 has been tested in phase II clinical trials for asthma and rheumatoid arthritis (89). Although NGD2000-1 showed some promising results, it also targeted cytochrome P450 3A4 that halted all clinical developments (89).

16 O F N O N O N A B N

N N N N H H OH OH O O

C D

Figure 1.8. C5aR non-peptidic anatagonists. (A) W54011, (B) NDT9520492 and aniline-substituted tetrahydroquinoline (C) JJ46 and (D) JJ47.

A series of tetrahydroquinoline compounds were developed by Johnson & Johnson Pharmaceutical Research and Development with modest antagonist activity on C5aR at high µM concentrations (128). The introduction of functional groups on the anilines was a strategy intended to increase interactions between ligand and C5aR binding sites to improve potency (129). The potency of aniline-substituted tetrahydroquinoline compounds such as JJ46 (Figure 1.8C) resulted in 10-100 fold increased in potency to double-digit nanomolar concentrations against 1.5 nM human C5a in a calcium mobilization assay on human monocytic cell line U937. JJ47 (Figure 1.8D) was the most potent antagonist of a series of aniline-substituted tetrahydroquinoline compounds developed by Johnson & Johnson Pharmaceutical Research and Development, with C5aR antagonist activity at single-digit nanomolar concentrations in a calcium mobilization assay with similar conditions. Docking of JJ47 in human C5aR homology model revealed that JJ47 binds into a putative ligand- binding site where the diethylphenyl group may have hydrophobic interactions with residues Ile116 and Val286 of the receptor. These interactions may contribute to the increased in antagonistic activity in the aniline-substituted tetrahydroquinoline compounds.

17 W54011 and JJ47 are both by far the two most potent reported antagonists in literature with IC50 values of 3 (against 0.1 nM human C5a on human neutrophils) and 7 nM (against 1.5 nM human C5a on U937 cells) respectively on human C5aR.

However, W54011 did not progress and perform as well as 3D53 (IC50 values = 20-55 nM against 0.1-100 nM human C5a on human neutrophils). To date, there has been no clinical development of JJ47. C5a has been implicated in numerous immune and inflammatory conditions but, to date, there is still no drug available for treating C5a- C5aR-mediated diseases. As such, 3D53, W54011 and JJ47 have been chosen to uncover important fundamental reasons for the failure of small molecules in clinical trials (Chapter 4).

1.7 New insights into C5aR-related biological functions and diseases

The understanding of complement biology has undergone a significant metamorphosis since its original discovery as a heat-labile principle in serum that “complemented” antibodies in the killing of bacteria (130, 131). A system, which was traditionally described as a “complement” to humoral immunity, is now perceived as a central constituent of innate immunity and adaptive immune responses. Proinflammatory signalling and phagocytosis are essential for complement-mediated defence against infection (132). The “complementary” concept becomes inadequate in light of recent studies demonstrating functions of complement that are essential and “central” to the innate immune response, as well as functions that link innate with adaptive immunity (133). Complement must now be viewed as a system that orchestrates and connects various responses during immune and inflammatory reactions and not merely as a killer of bacteria (134). During activation and amplification, C5a is constantly generated and released into circulation, triggering proinflammatory signalling through C5aR. The expression of C5aR was initially thought to be restricted to myeloid blood cells, including neutrophils, monocytes, macrophages and eosinophils (135). It now appears that the receptors are widely distributed to multiple organs and tissues, including the liver, lungs, kidney, brain and the central nervous system (135-144). This is an indication that the effects of C5aR signaling may be involved in many other biological functions. Some unexpected and

18 recently discovered functions of the C5a-C5aR axis are briefly summarized in the next few sections.

1.7.1 Interaction of RP S19 Oligomers with C5aR

Monocytes and neutrophils make up ~70% of the total white blood cells in circulation. These cells are major phagocytic leukocytes that can migrate to sites of infection by sensing the anaphylatoxin C5a via C5aR. Endotoxic shock or septic shock is an acute diseased state characterized by systemic hypotension (145), pulmonary hypertension (146), depletion of circulating leukocytes and platelets as a result of severe infection and sepsis (147). Most cases of septic shock are caused by bacterial lipopolysaccharides (LPS) released from gram-negative bacteria (148), which in turn can also activate the complement system to generate C5a (149). Not surprisingly, an intravenous administration of C5a, or C5a peptidic agonists in animals also resulted in hypotension and neutropenia, similar to responses induced by LPS (150). The role of C5a in sepsis has already been extensively reviewed elsewhere (52), hence this will not be the focus of this section. Instead, an unsuspected ligand of C5aR, the RP S19 protein, will be briefly described here. Neutrophils can migrate more rapidly and induce vascular plasma leakage upon infiltration (151), whereas monocytes are superior in phagocytosis and antigen presentation but infiltrate tissues more slowly (152). Since these leukocytes migrate differently in response to inflammation, there is no surprise that they responded to the novel chemotactic factor, ribosomal protein S19 (RP S19) differently (153). RP S19, a component of the small subunit of the translation machinery (ribosome), can dimerize by a transglutaminase-catalyzed reaction (154), and is released into the circulation when infected or damaged cells undergo apoptosis. Although there is no molecular relationship between RP S19 and C5a, RP S19 can bind as a dimer to C5aR and induce chemotaxis (155). However, RP S19 does this with pro- and anti-apoptotic dual functions on neutrophils and monocytes respectively, thereby affecting the fate of different immune cells (156). The human C5a is a 74-residue glycosylated peptide consisting of four alpha helices arranged in an anti-parallel orientation (Figure 1.9). Apart from two short β- sheets in the human RP S19 protein, the secondary structures are almost similar

19 between human C5a and RP S19 proteins (Figure 1.9) (157). Although there are no disulfide bond in RP S19, unlike C5a with 3 disulfide bonds, the overall tertiary structure of RP S19 appeared more loosely packed than C5a. Interestingly in human mast cells HMC-1, RP S19 triggers p38 MAPK activation instead of ERK1/2 activation by C5a (158). It seems that this novel regulatory mechanism via C5aR in immunity is in place to prevent excessive tissue destruction by neutrophils, yet to attract monocytes to clear apoptotic cells that released dimerized RP S19 molecules.

Figure 1.9. Comparative modelling between human C5a and RP S19 protein. Human C5a structure (PDB ID: 1CFA) and human RP S19 structure (homology model of Pyrococcus Abyssi, PDB ID: 2V7F). Adapted from (157)

1.7.2 An unsuspected role for C5aR in cancer-related inflammation

An association between chronic inflammation and cancer was initially suspected on the basis of two observations. First, C5a-induced chemotaxis has many similar features to cancer metastases. Second, there was epidemiological data correlating increased incidence of various malignancies in patients suffering from chronic inflammatory diseases (159). Numerous experimental studies have shown that chronic and indolent inflammation can increase the risk of malignant transformation, exacerbate tumour growth, enhance invasion of normal tissue and facilitate metastasis

20 (160, 161). Malignant tumour cells can suppress antitumour immune responses (162, 163), by recruiting myeloid-derived suppressor cells (MDSCs), which in turn induced an antigen-specific CD8+ T cell tolerance response to evade tumour killing mechanisms (164, 165). Although complement activation plays various roles in mediating inflammatory responses, complement promoting the development and progression of malignant tumours was not substantiated for a very long time. Instead, complement activation was interpreted as host defence mechanisms for killing tumour cells (166-169), by analogy to the well-characterized role of complement in microorganism killing. Various components of the complement system can promote the growth of tumours in a mouse model of cervical carcinoma (170). Complement effectors such as anaphylatoxin C5a, can give rise to many proinflammatory mediators (171), these immunological events in turn create favourable microenvironments encouraging tumour growth (160, 172-174). Furthermore, C5aR was aberrantly expressed in various human cancers (175). Cellular cytoskeletal rearrangement and motility, as well as invasiveness of cancers cells, were greatly enhanced by C5a, and this enhancement by C5a were not observed in normal cells (175). Pharmacological blockade of C5aR with peptidic antagonist 3D53 (also commonly referred to as PMX53) in wild-type mice resulted in reduced tumour growth (170). The role of C5aR signaling in tumour development was also supported by reduced tumour growth in C5aR-/- mice (176). Furthermore, C5aR-deficient mice depleted of CD8+ T cells were free of tumour, leading to the conclusion that C5a promoted tumour growth by suppressing T-cell-mediated antitumour responses (170). The recent discovery that complement promotes the growth of malignant tumours (175) has shifted the paradigm in the field of cancer research, although such findings are also consistent with a growing appreciation for the role of inflammation in cancer progression. In light of this discovery, inhibiting the complement system seems to be a promising new approach to treating cancer. MDSCs in cancer patients were strongly believed to be the main culprit for the failure of antitumour immunization (165). Targeting C5aR in combination with antitumour immunization, may improve the efficacy of existing antitumour vaccines.

21 1.7.3 C5aR crosstalking with TLRs

The complement system appears to have evolved as ancient defence machinery against microbial pathogens. Invading pathogens display unique pathogen-associated molecular patterns (PAMPs), which are detected by the host innate immune system via pattern recognition receptors (PRRs) (177, 178). The Toll-Like Receptors (TLRs) and complement are two key component systems of the innate immune response that can recognise PAMPs and facilitate microbial killing and clearance (131, 179). Although both systems have been historically investigated as separate entities, an emerging body of evidence indicates extensive crosstalk between TLR and complement signaling pathways in several in vitro as well as in vivo experimental human and mice models (180). Crosstalk between different signaling pathways in mammalian cells can result in unexpected and unique functional outcomes (181). Some pathogens have evolved to evade or exploit host microbial-killing by targeting C5aR and TLRs specifically. For example, virulence proteins secreted by Staphylococcus aureus bind to C5aR with potent antagonist activity, preventing recruitment of phagocytes to sites of infection (182). Porphyromonas gingivalis exploits the C5aR/TLR2 crosstalk by increasing cAMP production in macrophages, which suppresses macrophage immune function and enhances pathogen survival (183). Clearly, these pathogens have evolutionarily adapted to exploit complement/TLR crosstalk to their advantage. Of 10 known human TLRs, TLR4 responds to LPS from Gram-negative bacterial cell walls. TLR4 can signal through two different pathways dependent on, and modulated by, separate adaptor proteins MyD88 or Toll/IL-1R domain-containing adapter inducing IFN-β (TRIF) (184). LPS is not only known as a typical ligand for TLR4, it is also a complement activator (185-187). Synergistic crosstalk between TLR4 and C5aR was first documented in 1987, when human monocytes co-treated with LPS and C5a were reported to secrete elevated levels of IL-1β compared to either treatment alone (188). While C5aR and TLR4 signalling have been separately investigated in many studies, there is comparatively little reported about their interplay (180). Among recent reports on C5aR/TLR4 crosstalk, C5a has been found to downregulate TLR4-induced IL12 production via PI3K-dependent (189) and PI3K- independent (190) pathways in mouse macrophages; to signal via the PI3K-Akt

22 pathway to enhance LPS-induced IL17F cytokine production (191), and to suppress LPS-induced IL17A and IL23 cytokine production in mouse macrophages (192). C5a- induced suppression production and ERK1/2 phosphorylation. Part of Chapter 5 deals ahead with C5a acting via C5aR and differentially modulating LPS-induced inflammatory responses in primary human monocytes versus macrophages (193). While C5a enhanced secretion of LPS-induced IL6 and TNF from primary human monocytes, C5a inhibited these responses while increasing IL10 secretion in donor-matched human monocyte-derived macrophages differentiated by GM-CSF or M-CSF (193). C5a induced Gαi/c-Raf/MEK/ERK signaling was amplified in macrophages but not in monocytes by LPS. Furthermore, this C5a-mediated suppression of pro-inflammatory cytokines IL6 and TNF in macrophages did not compromise antimicrobial activity; C5a instead enhanced clearance of the gram-negative bacterial pathogen Salmonella enterica serovar Typhimurium from macrophages (193). These findings implicate C5aR as a regulatory switch that modulates TLR4 signaling via the Gαi/c-Raf/MEK/ERK signaling axis in human macrophages but not in human monocytes. The differential effects of C5a are consistent with amplifying monocyte proinflammatory responses to systemic danger signals, but attenuating macrophage cytokine responses (without compromising microbicidal activity) thereby restraining inflammatory responses to localized infections. However, most studies reporting C5a modulation of TLR4 signalling had been performed with murine (183, 189, 190) and few on human cells (193-195). There are important differences in LPS-induced TLR4 signaling between humans and mice (196-198), thus necessitating careful assessment of C5aR/TLR4 crosstalk in human cells. Some interpretations of human disease mechanisms have been made based on extrapolating observations from mouse studies which, although convenient, are often not appropriate (199). Future research is expected to elucidate additional regulatory links between complement and TLR signalings, particularly in human systems, and this is now considered to be pivotal for correctly understanding their precise roles in human health and human diseases.

23 1.7.4 Linking C5aR signalling to obesity and metabolic dysfunction

Modern diets high in carbohydrates and saturated fats are the cause of current global epidemics in obesity, type II diabetes and cardiovascular disease. Obesity is a medical condition in which excess body fat has accumulated to the extent that it may have adverse effects on health. Individuals with metabolic dysfunction may inherit some of these conditions, such as abdominal obesity, glucose and insulin tolerance, elevated plasma triglycerides and cholesterol, and liver and cardiovascular abnormalities (200, 201). Obesity is causally linked to chronic low-grade inflammation, which contributes to the onset of metabolic disorders (202-205). However, the nature and importance of these links with inflammation remained unclear. Immune and metabolic systems are among the most fundamental requirements for survival. Historically, immunity and energy metabolism are considered to be two distinct disciplines. Energy homeostasis is now believed to be regulated and coordinated by signalling molecules and pathways that are common to those in inflammatory networks (206). Recently, specific modes of immunity and energy metabolism have been interrelated at the molecular, cellular, organ and organism levels (206-208). The energy balance and body composition are dependent on energy income and expenditure, which appears to be interrelated and regulated by food intake, fuel utilisation, thermogenesis and adipocyte metabolism (209, 210). Gut microbiota, immune networks, altered nutrient sensing and metabolism of fatty acids in adipose tissue can also contribute to the onset of obesity and metabolic dysfunction (211, 212). Activating the complement network triggers immune responses that use substantial amounts of energy to fight infection. Aberrant immune responses in chronic inflammatory states may exacerbate obesity and metabolic dysfunction (213). Elevated levels of C3 (a precursor to C5 and C5a) can be detected in patients suffering from type II diabetes and insulin resistance compared to healthy individuals (214, 215). Serum levels of C3 correlated with a progressive increase in body mass index (BMI) in subjects with severe, morbid or extreme obesity, and thus could be a potential biomarker for obesity (216). Interestingly, complement proteins C5a and C5aR have energy-conserving roles that are contraindicated in complement biology (217). Diet-induced obesity (DIO)

24 mice had increased C5aR mRNA expression in obese adipose tissue (218). C5aR deficiency in obese mice led to improved insulin sensitivity (218). PGE2 can have paracrine influences on adipocytes through the inhibition of lipolysis (219).

Interestingly, stimulating murine adipocytes with PGE2 also led to increased gene and protein expression of C5aR (217). Antagonism of C5aR can dramatically attenuate obesity, visceral adiposity, adipose inflammation, glucose/insulin tolerance and cardiovascular dysfunction in rats induced by a high-carbohydrate high-fat diet (217). Thereby, this establishes an axis where C5a can exert paracrine control over adipocytes via adipose macrophages that secrete proinflammatory agents such as

PGE2 (217), which are associated with phagocytosing excess lipids and possibly secreting antilipolytic factors to curtail the increase in free fatty acids during lipolysis in obesity (220). Taken together, the C5a-C5aR axis is now thought to contribute to macrophage accumulation and adipose tissue inflammation in obese subjects, thereby leading to the development of insulin resistance and aberrant metabolic functions in adipose tissues. These findings suggest that C5aR may represent a novel target to control the progression of metabolic dysfunction and weight gain in obese individuals.

1.8 Mononuclear phagocytic system

The mononuclear phagocytic system is derived from haematopoietic stem cells located in the bone marrow, a common myeloid progenitor cell that is the precursor of many different cell types, including neutrophils, erythrocytes, eosinophils, basophils, monocytes, macrophages, dendritic cells and mast cells (221). During monocyte development, myeloid progenitor cells CFU-GM (granuolocyte-macrophage colony- forming units) generate monoblasts and sequentially give rise to pro-monocytes and finally monocytes. Monocytes are released from the bone marrow into the blood circulation as macrophages-precursor cells. As monocytes migrate out of the blood circulation, they can differentiate into macrophages either in the steady state or in response to inflammation (222). Long-lived tissue-specific macrophages can be found residing throughout the body including the bone (osteoclasts), alveoli, central nervous system (microglial cells), connective tissue (histiocytes), gastrointestinal tract, liver (Kupffer cells), spleen, adipose and peritoneum (222, 223). And depending on which

25 tissue the macrophages reside in and the stimulant exposure, they can have distinct appearances and different functional phenotypes. Macrophages are important immune effector cells that interplay between innate and adaptive immunity. In 1908, Élie Metchnikoff received the Nobel Prize in Medicine for his discovery of phagocytosis. Back in the early 1900s, macrophages were thought to primarily function as professional phagocytes and their other functions were overlooked for decades (224). By focusing on the immune functions of macrophages, immunologists have ignored their vital homeostatic roles, which are independent of their involvement in immune responses. The monocyte-macrophage axis plays essential roles in inflammation, antimicrobial defence and tissue wound healing but, like a double-edged sword, they are also the main culprits contributing to tissue destruction during some infections and inflammatory diseases (225). Macrophages have remarkable phagocytic capacities. On average, macrophages clear approximately 200 billion erythrocytes each day equating to almost 3 kg of iron and haemoglobin per year being recycled for the host to reuse (226). The host would not survive without this clearance process. Macrophages are also involved in the removal of cellular debris generated during tissue remodelling, and rapidly remove cells that have undergone apoptosis, infection or other recognisable damage. Some of these processes are carried out independent of immune signals and can result in little or no immune responses by unstimulated macrophages (227). However, if cells undergoing apoptosis as the consequence of some insults, infection or disease, then both defence and repair mechanisms are mobilized by the macrophages. Activated macrophages may undergo biochemical, morphological and functional changes and secrete an array of cytokines and chemokines, which in turn trigger inflammatory responses (228). Macrophage activation at sites of inflammation is typically transient with local tissue repair and remodelling in the process of restoring and establishing normal tissue function. However, at sites of resistant infectious agents (tuberculosis, herpes, HIV), or poorly biodegradable tissue irritants (foreign bodies), macrophages remained activated and fused together to form multinucleate giant cells (MGCs) that can persist for weeks or months surrounding the unresolved inflammatory site, forming the crown-like structure (CLS), a hallmark of chronic inflammation (229).

26 1.8.1 Macrophage infiltration into adipose tissue

A key observation which supported the association between obesity and inflammation was the increased number of immune cells infiltration, in particular of macrophages into adipose tissue (230-232). Resident macrophages may constitute as much as 40% of the total cell population present in adipose tissue (230, 233, 234). Adipose tissues of obese human and obese animal models are infiltrated by a large number of macrophages and the severity of systemic inflammation and insulin resistance correlated to immune cells recruitment (205, 230). Weight loss in obese individuals reduced macrophage infiltration and proinflammatory profiles (232). Mirroring the Th1/Th2 nomenclature, polarized macrophages are often referred to as M1 (classically activated) or M2 (alternatively activated). Macrophages that accumulate in the adipose tissue of obese mice expressed markers of an M1 phenotype, whereas adipose tissue macrophages (ATM) from lean mice expressed markers of an M2 phenotype (234). It has been proposed that M1 (also known as classically activated macrophages; CAMs) or M2 (also known as alternatively activated macrophages; AAMs) ATMs can be distinguished by the presence or the absence of CD11c surface marker (234), while others have proposed that the M1/M2 designation for macrophages should be based on the ratio of IL-12 to IL-10, a ‘pro- versus anti- inflammatory’ cytokine productions model (235). M1 or CAM can be activated by LPS and IFN-γ to produce proinflammatory cytokines (TNF and IL6) and expressed inducible nitric oxide synthase (iNOS); to produce reactive oxygen species (ROS) and nitrogen intermediates (236). M2 or AAM on the other hand, can be activated by IL4 or IL13 and expressed arginase 1, CD206 (mannose receptor) and IL4R α-chain that are regulated by peroxisome proliferator-activated receptor gamma (PPARG) (237, 238). The majority of the M1 macrophages were found aggregating around ‘dead’ adipocytes to remove apoptotic cells from adipose tissue (239). Residual lipid droplets released by dead adipocytes were scavenged by the M1 macrophage and formed CLS (Figure 1.10), a hallmark of chronic inflammation. Depletion of M1 marker genes such as TNF (240) and CCR2 (241) or the ablation of CD11c+ cells (242), resulted in normalized insulin sensitivity in insulin- resistant subjects. M2 macrophages on the other hand, were thought to play a role in injured tissue repair and in suppressing the synthesis of proinflammatory cytokines by

27 producing antiinflammatory cytokine IL10 (236). Mice without peroxisome proliferator-activated receptor (PPAR)-γ and –β/δ expression resulted in insulin resistance and impaired M2 function (238, 243). Due to the nature of these macrophages, M1 macrophages were suggested to adapt the proinflammatory immune profile, increasing obesity-induced insulin resistance, whereas M2 macrophages were adapted to an anti-inflammatory immune profile promoting protective properties against it (244).

Figure 1.10. Macrophages localize to crown-like structure (CLS) around individual adipocytes, which increase in frequency in obesity. Reference from (239).

28 1.8.2 Macrophage activation and polarization

Functional polarization of macrophages into M1 and M2 cells is an operationally useful, simplified conceptual framework describing the plasticity of mononuclear phagocytes. However, macrophages have remarkable plasticity that allows them to efficiently respond to environmental signals and change their phenotype, and their physiology can be altered by both innate and adaptive immune responses (245), unlike lymphocyte, where phenotypic changes can be largely “fixed” by chromatin modifications after exposure to polarization cytokines (246, 247). M2- polarized macrophages can be readily induced to express M1-associated genes by subsequent exposure to TLR ligands or IFN-γ (245, 248, 249). M2 designation further extended to include essentially all other types of macrophages (235). A growing body of evidence indicated that the M2 designation encompasses cells with differences in their physiology and biochemistry (250-252). As plasticity seems to be a general property of macrophages, the rigid assignment of M1 or M2 subset to in vivo inflammation and homeostasis does not make sense unless the overall immune response is dominated by one type of T cell response or by a single stimulus (226). Mosser and co-worker proposed that macrophage classification should be based on the fundamental macrophage functions; host defence, wound healing and immune regulation, instead of just an M1 or M2 designation (226). Shown in Figure 1.11, is a classification of macrophages according to their functions as primary colours represented by a colour wheel, rather than a monochromatic linear scale designating M1 or M2 (226). This classification more accurately illustrates how macrophages can evolve to exhibit characteristics that are shared by more than one macrophage population, compared to the rigid designation of M1 versus M2. Furthermore, it classifies classically activated macrophages (CAMs) more closely related to other subsets than previously designated, covering the development of macrophages with shared characteristics rather than in isolation, resulting in a spectrum of macrophages.

Figure 1.11. Colour wheel of macrophage activation. Reference from (226).

1.8.3 Understanding human macrophage biology from a mouse perspective

Mice have been the experimental tools of choice for immunologists and the study of murine immune responses has yielded tremendous insights into the workings of the human immune system. Sequence comparison between the human and mice genomes showed only 300 or so genes are unique to individual species (253). Despite this conservation, it is now known that there are very significant differences between human and mice in development, activation and responses in both the innate and adaptive arms of host immune system. The antimicrobial peptides, defensins, one of the first lines of defence in higher organisms and often the only defence in lower animals (254), is one of the many differences between human and mouse innate immunity. Defensins can be found in abundance in neutrophils of human, rabbit and

30 even in rats, but are not found at all in mice (255, 256). Secondly, TLRs are important players in defence against invading pathogens (257), and recent discoveries and differences between both species have caused immunologists to re-think whether mice are truly suitable models to study human immunity in detail (258). Mammalian TLR3 recognizes dsRNA and, upon binding, induces NF-κΒ activation and production of type I interferons (259). Alexopoulou et al. found that LPS can strongly react with TLR3 in murine (260, 261), but not in human cells (262, 263). Interestingly, there are also important differences between humans and mice in TLR4 signaling induced by LPS (196-198). For the past three decades, nitric oxide (NO) has been recognized as an important player involved in tumours, autoimmune and chronic degenerative diseases, pathogenesis and control of infectious diseases (264). The expression of inducible NO synthase (iNOS) can be readily induced by IFN-γ and LPS in mouse macrophages (265). However, the same inflammatory agents failed to show consistent effects on human macrophages (266). Apparently, there was no NOS activity nor synthesis of tetrahydrobiopterin (BH4) in human macrophages when stimulated with IFN-γ or LPS (267), which was crucial for the stability and function of the iNOS enzyme in murine macrophages (268, 269). This led some investigators to wonder if human macrophages were capable of NO generation (267). During infection and other inflammatory conditions, elevated levels of NO can be detected in human subjects (270-273). Generation of NO was apparent in macrophages from patients with tuberculosis (274, 275), rheumatoid arthritis (276), or malaria (277), but not in peripheral blood mononuclear cells of healthy subjects (267, 278). It was not until Hickman-Davis et al. discovered the ability of human alveolar macrophages to generate NO production by surfactant protein A (279). Sequentially, treatment with other inflammatory mediators, to name a few, IFN-αβ, IL-4 in combination with anti-CD23, chemokines, bacterial components, and the list goes on, have been reported to induce NO production in human macrophages (272, 280-282). The production of NO by human macrophages till to date has remained debateable (283, 284). L-arginine metabolism in macrophages is regulated by the enzyme Arginase I (Arg1) to produce urea and ornithine or citrulline, and the enzyme iNOS to produce NO (285, 286). It has been suggested for mice that M2 can be characterized by the high expression of CD206, Arg1, MgII and IL-10 (222, 237). Interestingly, like iNOS, Arg1 expression was also not expressed by in vitro polarized human macrophages

31 stimulated with IFN-γ or IL-4 respectively, a contrast from their murine counterparts (287). Cell surface markers such as CD11b, CD18 (MAC-1), CD68, MAC-2 and F4/80 or EMR1 (human ortholog of murine F4/80) have been widely used by immunologists to distinguish macrophages from dendritic cells, especially in the field of immunometabolism where macrophages play key roles in the onset of metabolic disorders and obesity (288). However, Hamann et al. has disputed whether EMR1 was not expressed in human macrophages but an eosinophil-specific receptor in human cells unlike its murine counterparts (289). Khazen et al. supported that EMR1- F4/80 was the best macrophage marker not CD14 and CD68, but this conclusion was only drawn based on RAW264.7, a mouse macrophage not human (290). While F4/80 is a specific marker for murine macrophages (291, 292), the absence of F4/80-EMR1 on human macrophages merits further investigation. In summary, murine and human macrophages exhibit distinct differences in responses important for immunodefense and regulating inflammation (197, 283, 287, 293, 294). It would seem that extrapolating data from mice to humans, although convenient, is often not appropriate (199). Indeed it is of great concern that researchers still today are using these inflammatory markers to determine macrophage phenotypes despite vigorous debate and controversy over the usefulness of mouse models for human biology. Perhaps, to understand macrophage biology, if not the whole immune system, investigations should primarily refocus on human cells, then translate novel interactions or regulatory mechanisms underlying such a function and further validate it in murine cells for further pre-clinical studies in mouse models, and not the other way around. Humans are not mice.

1.8.4 Differentiating primary human monocytes with GM-CSF or M-CSF to generate human macrophages

Granulocyte-macrophage colony-stimulating factor (GM-CSF) and macrophage colony-stimulating factor (M-CSF) were first defined by their abilities to generate in vitro colonies of mature myeloid cells from bone-marrow precursor cells and differentiation of these cells (granulocytic and macrophage by GM-CSF; and macrophages by M-CSF) (295, 296). The biological functions of these CSFs are quite distinct from one another, with different tissue distributions in steady state, different

32 structures, and different receptors. M-CSF receptor (CSF1R) is the tyrosine kinase c- Fms (297, 298) and the GM-CSF receptor (CSF2R) comprises two subunits, α (ligand binding) and β (signaling) (299, 300). However, it became apparent that GM-CSF and M-CSF have other functions on account of their activities on mature myeloid cells (301-303). Since GM-CSF and M-CSF played such an important role in modulating myeloid cellular development and maintaining immune homeostasis, it is not surprising that drugs are currently being investigated in clinical trials in a number of inflammatory and autoinflammatory conditions (302, 304). More importantly, GM-CSF and M-CSF can be produced by various cells in response to infection, injury or inflammation, which are the major survival/mitogenic factors for the macrophage lineage with the capacity to activate monocytes and induce differentiation (222, 305, 306). M-CSF is passively produced in vitro by several cell types, especially macrophages, and circulates at detectable levels basally, whereas GM-CSF is usually undetectable but becomes elevated during immune/inflammatory responses in vivo to infection or in vitro in response to LPS or TNF (302, 307). GM- CSF was first isolated from lungs (308) and monocytes differentiated with GM-CSF closely resemble alveolar macrophages (309, 310), or even immature dendritic cells (311, 312). M-CSF, also known as CSF1, was the first hemopoietic growth factors to be isolated (296) and monocytes differentiated with M-CSF closely resemble osteoclasts and peritoneal macrophages (313, 314) important for tissue remodelling/homeostasis (315). Therefore monocytes recruited to sites where CSF is dominant might differentiate into divergent phenotypes of macrophages. Monocytes exposed to either GM-CSF or M-CSF led to the differentiation of macrophages with some similarities in recognition patterns but showed significant variation in phenotypic appearances, cytokine productions, surface expressions, endocytosis and many more biological functions (305, 306, 309, 316-319). ). It is important to realise that there are a range of other differentiating agents in vivo that can also influence the outcome of M1- versus M2-like functional phenotypes. M1 and M2 macrophage phenotypes were described by analogy to Th1 and Th2 lympocytes subsets. Macrophages differentiated in culture in the presence of GM- CSF or M-CSF, were named proinflammatory M1 and anti-inflammatory M2 phenotypes, respectively (317, 320). GM-CSF and M-CSF can influence macrophage phenotypic dependence on type I interferon signalling, at least in murine bone marrow cells (318). Although the functional heterogeneity of CSF-induced human monocyte-

33 derived macrophages has been extensively studies (305, 316, 321, 322) and reviewed (222, 284), very few studies have actually compared the effects of GM-CSF and M- CSF across species, human versus mouse. One study in particular, examined at the gene expression and only 17% of the genes in human monocytes differentiated by human GM-CSF versus M-CSF had a profile that was conserved in murine GM-CSF versus M-CSF differentiated murine monocytes (317). With such a high divergence across species, caution should therefore be exercised when interpreting the effects of CSF on human macrophage lineage cells from studies conducted on analogous murine cells. Since the classification of macrophages follows a continuum, the activation and polarization of macrophages should to be re-evaluated as a continuum (250) instead as separate distinction as previously reported (226).

1.9 Thesis objectives

G-protein coupled receptors (GPCRs) are the most prevalent signal-transducing cell surface proteins in pharmaceutical research. Of the over 900 human GPCRs identified to date (323, 324), more than one third of human GPCRs are potential therapeutic targets for the treatment of diseases (325). One such GPCR is the complement receptor (C5aR) for C5a. C5aR has been strongly implicated in many inflammatory diseases (17) and demonstrated to be an important and promising therapeutic target for complement modulation (89). Therefore it is desirable to modulate complement activation by using therapeutic interventions such as inhibitors or antibodies. Over the past two decades, C5aR antagonists have been reported but despite intensive efforts to progress an effective C5aR antagonist to market, this has not yet been achieved. Although C5aR is widely expressed throughout the body, it is expressed highly in myeloid cells, such as macrophages (89, 326). Macrophages play critical roles in host defence, inflammation and the maintenance of homeostasis. Macrophages have remarkable plasticity in their gene expression profiles and cell surface phenotypes, depending on which tissue the macrophages reside in and stimulant exposure.

34 1.9.1 Chapter two

Since macrophage phenotypes were proposed to be a continuum (226), without a clear distinction between M1 and M2 subsets, we explored how differential or similar expression of genes affect functions in primary monocytes differentiated in culture with GM-CSF (GM-MΦ) versus M-CSF (M-MΦ). The objective of this chapter is to characterize and validate GM-MΦ versus M-MΦ as suitable in vitro models to study proinflammatory and anti-inflammatory responses in human macrophages.

1.9.2 Chapter three

Effects of C5aR were previously reported in murine but not human cells (326), or macrophage cell-lines that had undergone mutations and did not behave the same as primary cells in vitro and in vivo (327). The objective of this chapter is to re- evaluate the effects of human C5a and gain knowledge on C5aR-mediated responses in primary human macrophages, which will be invaluable to the development of drugs against C5aR-mediated diseases.

1.9.3 Chapter four

The only FDA approved treatment that targets C5a-C5aR interaction is a humanized antibody which actually binds C5, thereby preventing the formation of C5a but also the formation of the membrane attack complex that effects pathogen lysis. C5a has been implicated in numerous immune and inflammatory conditions but to date, there is no drug available for treating C5a-C5aR-mediated diseases. The objectives of this chapter are to conduct a comparative study on three previously reported antagonists of human C5aR (W54011, JJ47 and 3D53), to investigate possible differences in their properties that might account for problems in clinical development of small molecule antagonists in clinical trials, and to choose one antagonist for further studies (Chapter 5) on synergistic effects of C5a on lipopolysaccharide action on macrophages and monocytes.

35 1.9.4 Chapter five

Monocytes and macrophages are important innate immune cells equipped with danger sensing receptors, including complement and Toll-like receptors. However, most studies reporting C5a modulation of TLR4 signalling had been performed with murine rather than human cells. Some interpretations of human disease mechanisms have been made based on extrapolating observations from mouse studies. However, there are important differences in LPS-induced TLR4 signaling between humans and mice, thus necessitating careful reassessment of C5aR/TLR4 crosstalk in human cells. The objective of this chapter is to compare the modulation of human C5a on LPS- induced inflammatory responses in primary human monocytes versus macrophages.

1.9.5 Chapter six

This chapter summarizes key findings reported in chapters 2-5, which represent new knowledge for researchers in the field of complement C5aR and its effects on primary human monocytes and macrophages, and indicates possible future research directions related to immunology and pharmacology.

36 1.10 References

1. McHeyzer-Williams, M. G. 1997. Immune response decisions at the single cell level. Semin. Immunol. 9: 219-227. 2. Abbas, A. K., and C. A. Janeway, Jr. 2000. Immunology: improving on nature in the twenty-first century. Cell 100: 129-138. 3. Bal, S. M., Z. Ding, E. van Riet, W. Jiskoot, and J. A. Bouwstra. 2010. Advances in transcutaneous vaccine delivery: do all ways lead to Rome? J. Control. Release 148: 266-282. 4. Muller-Eberhard, H. J. 1988. Molecular organization and function of the complement system. Annu. Rev. Biochem. 57: 321-347. 5. Amara, U., D. Rittirsch, M. Flierl, U. Bruckner, A. Klos, F. Gebhard, J. D. Lambris, and M. Huber-Lang. 2008. Interaction between the coagulation and complement system. Adv. Exp. Med. Biol. 632: 71-79. 6. Nishimura, T., T. Myles, A. M. Piliponsky, P. N. Kao, G. J. Berry, and L. L. Leung. 2007. Thrombin-activatable procarboxypeptidase B regulates activated complement C5a in vivo. Blood 109: 1992-1997. 7. Huber-Lang, M., J. V. Sarma, F. S. Zetoune, D. Rittirsch, T. A. Neff, S. R. McGuire, J. D. Lambris, R. L. Warner, M. A. Flierl, L. M. Hoesel, F. Gebhard, J. G. Younger, S. M. Drouin, R. A. Wetsel, and P. A. Ward. 2006. Generation of C5a in the absence of C3: a new complement activation pathway. Nat. Med. 12: 682-687. 8. Minta, J. O., D. Man, and H. Z. Movat. 1977. Kinetic studies on the fragmentation of the third component of complement (C3) by trypsin. J. Immunol. 118: 2192-2198. 9. Huber-Lang, M. S., J. V. Sarma, S. R. McGuire, K. T. Lu, V. A. Padgaonkar, E. M. Younkin, R. F. Guo, C. H. Weber, E. R. Zuiderweg, F. S. Zetoune, and P. A. Ward. 2003. Structure-function relationships of human C5a and C5aR. J. Immunol. 170: 6115-6124. 10. Liszewski, M. K., T. C. Farries, D. M. Lublin, I. A. Rooney, and J. P. Atkinson. 1996. Control of the complement system. Adv. Immunol. 61: 201- 283.

37 11. Parker, C. J. 1992. Regulation of complement by membrane proteins: an overview. Curr. Top. Microbiol. Immunol. 178: 1-6. 12. Gigli, I., J. Sorvillo, and L. Halbwachs-Mecarelli. 1985. Regulation and deregulation of the fluid-phase classical pathway C3 convertase. J. Immunol. 135: 440-444. 13. Rooijakkers, S. H., J. Wu, M. Ruyken, R. van Domselaar, K. L. Planken, A. Tzekou, D. Ricklin, J. D. Lambris, B. J. Janssen, J. A. van Strijp, and P. Gros. 2009. Structural and functional implications of the alternative complement pathway C3 convertase stabilized by a staphylococcal inhibitor. Nat. Immunol. 10: 721-727. 14. Kirkitadze, M. D., and P. N. Barlow. 2001. Structure and flexibility of the multiple domain proteins that regulate complement activation. Immunol. Rev. 180: 146-161. 15. Tschopp, J., H. J. Muller-Eberhard, and E. R. Podack. 1982. Formation of transmembrane tubules by spontaneous polymerization of the hydrophilic complement protein C9. Nature 298: 534-538. 16. Carroll, M. V., and R. B. Sim. 2011. Complement in health and disease. Adv Drug Deliv Rev 63: 965-975. 17. Makrides, S. C. 1998. Therapeutic inhibition of the complement system. Pharmacol. Rev. 50: 59-87. 18. Rollins, T. E., S. Siciliano, S. Kobayashi, D. N. Cianciarulo, V. Bonilla- Argudo, K. Collier, and M. S. Springer. 1991. Purification of the active C5a receptor from human polymorphonuclear leukocytes as a receptor-Gi complex. Proc. Natl. Acad. Sci. U. S. A. 88: 971-975. 19. Norgauer, J., G. Dobos, E. Kownatzki, C. Dahinden, R. Burger, R. Kupper, and P. Gierschik. 1993. Complement fragment C3a stimulates Ca2+ influx in neutrophils via a pertussis-toxin-sensitive G protein. Eur. J. Biochem. 217: 289-294. 20. Settmacher, B., C. Rheinheimer, H. Hamacher, R. S. Ames, A. Wise, L. Jenkinson, D. Bock, M. Schaefer, J. Kohl, and A. Klos. 2003. Structure- function studies of the C3a-receptor: C-terminal serine and threonine residues which influence receptor internalization and signaling. Eur. J. Immunol. 33: 920-927.

38 21. Neer, E. J., S. Y. Kim, S. L. Ang, D. B. Bloch, K. D. Bloch, Y. Kawahara, C. Tolman, R. Lee, D. Logothetis, D. Kim, and et al. 1988. Functions of G- protein subunits. Cold Spring Harb. Symp. Quant. Biol. 53 Pt 1: 241-246. 22. Buhl, A. M., N. Avdi, G. S. Worthen, and G. L. Johnson. 1994. Mapping of the C5a receptor signal transduction network in human neutrophils. Proc. Natl. Acad. Sci. U. S. A. 91: 9190-9194. 23. Perianayagam, M. C., V. S. Balakrishnan, B. J. Pereira, and B. L. Jaber. 2004. C5a delays apoptosis of human neutrophils via an extracellular signal- regulated kinase and Bad-mediated signalling pathway. Eur. J. Clin. Invest. 34: 50-56. 24. Perianayagam, M. C., V. S. Balakrishnan, A. J. King, B. J. Pereira, and B. L. Jaber. 2002. C5a delays apoptosis of human neutrophils by a phosphatidylinositol 3-kinase-signaling pathway. Kidney Int. 61: 456-463. 25. Strieter, R. M., K. Kasahara, R. M. Allen, T. J. Standiford, M. W. Rolfe, F. S. Becker, S. W. Chensue, and S. L. Kunkel. 1992. Cytokine-induced neutrophil- derived interleukin-8. Am. J. Pathol. 141: 397-407. 26. Hopken, U., M. Mohr, A. Struber, H. Montz, H. Burchardi, O. Gotze, and M. Oppermann. 1996. Inhibition of interleukin-6 synthesis in an animal model of septic shock by anti-C5a monoclonal antibodies. Eur. J. Immunol. 26: 1103- 1109. 27. Riedemann, N. C., R. F. Guo, V. J. Sarma, I. J. Laudes, M. Huber-Lang, R. L. Warner, E. A. Albrecht, C. L. Speyer, and P. A. Ward. 2002. Expression and function of the C5a receptor in rat alveolar epithelial cells. J. Immunol. 168: 1919-1925. 28. Laudes, I. J., J. C. Chu, M. Huber-Lang, R. F. Guo, N. C. Riedemann, J. V. Sarma, F. Mahdi, H. S. Murphy, C. Speyer, K. T. Lu, J. D. Lambris, F. S. Zetoune, and P. A. Ward. 2002. Expression and function of C5a receptor in mouse microvascular endothelial cells. J. Immunol. 169: 5962-5970. 29. Monsinjon, T., P. Gasque, P. Chan, A. Ischenko, J. J. Brady, and M. C. Fontaine. 2003. Regulation by complement C3a and C5a anaphylatoxins of cytokine production in human umbilical vein endothelial cells. FASEB J. 17: 1003-1014. 30. Wittmann, M., J. Zwirner, V. A. Larsson, K. Kirchhoff, G. Begemann, A. Kapp, O. Gotze, and T. Werfel. 1999. C5a suppresses the production of IL-12

39 by IFN-gamma-primed and lipopolysaccharide-challenged human monocytes. J. Immunol. 162: 6763-6769. 31. Pan, Z. K. 1998. Anaphylatoxins C5a and C3a induce nuclear factor kappaB activation in human peripheral blood monocytes. Biochim. Biophys. Acta 1443: 90-98. 32. Riedemann, N. C., R. F. Guo, K. D. Bernacki, J. S. Reuben, I. J. Laudes, T. A. Neff, H. Gao, C. Speyer, V. J. Sarma, F. S. Zetoune, and P. A. Ward. 2003. Regulation by C5a of neutrophil activation during sepsis. Immunity 19: 193- 202. 33. Ohno, M., T. Hirata, M. Enomoto, T. Araki, H. Ishimaru, and T. A. Takahashi. 2000. A putative chemoattractant receptor, C5L2, is expressed in granulocyte and immature dendritic cells, but not in mature dendritic cells. Mol. Immunol. 37: 407-412. 34. Li, R., L. G. Coulthard, M. C. Wu, S. M. Taylor, and T. M. Woodruff. 2013. C5L2: a controversial receptor of complement anaphylatoxin, C5a. FASEB J. 27: 855-864. 35. Kalant, D., S. A. Cain, M. Maslowska, A. D. Sniderman, K. Cianflone, and P. N. Monk. 2003. The chemoattractant receptor-like protein C5L2 binds the C3a des-Arg77/acylation-stimulating protein. J. Biol. Chem. 278: 11123-11129. 36. Okinaga, S., D. Slattery, A. Humbles, Z. Zsengeller, O. Morteau, M. B. Kinrade, R. M. Brodbeck, J. E. Krause, H. R. Choe, N. P. Gerard, and C. Gerard. 2003. C5L2, a nonsignaling C5A binding protein. Biochemistry (Mosc.) 42: 9406-9415. 37. Cain, S. A., and P. N. Monk. 2002. The orphan receptor C5L2 has high affinity binding sites for complement fragments C5a and C5a des-Arg(74). J. Biol. Chem. 277: 7165-7169. 38. Gerard, N. P., B. Lu, P. Liu, S. Craig, Y. Fujiwara, S. Okinaga, and C. Gerard. 2005. An anti-inflammatory function for the complement anaphylatoxin C5a- binding protein, C5L2. J. Biol. Chem. 280: 39677-39680. 39. Gao, H., T. A. Neff, R. F. Guo, C. L. Speyer, J. V. Sarma, S. Tomlins, Y. Man, N. C. Riedemann, L. M. Hoesel, E. Younkin, F. S. Zetoune, and P. A. Ward. 2005. Evidence for a functional role of the second C5a receptor C5L2. FASEB J. 19: 1003-1005.

40 40. Chen, N. J., C. Mirtsos, D. Suh, Y. C. Lu, W. J. Lin, C. McKerlie, T. Lee, H. Baribault, H. Tian, and W. C. Yeh. 2007. C5L2 is critical for the biological activities of the anaphylatoxins C5a and C3a. Nature 446: 203-207. 41. Cui, L., D. F. Carney, and T. E. Hugli. 1994. Primary structure and functional characterization of rat C5a: an anaphylatoxin with unusually high potency. Protein Sci. 3: 1169-1177. 42. Hugli, T. E. 1981. The structural basis for anaphylatoxin and chemotactic functions of C3a, C4a, and C5a. Crit. Rev. Immunol. 1: 321-366. 43. Sarma, J. V., and P. A. Ward. 2012. New developments in C5a receptor signaling. Cell health and cytoskeleton 4: 73-82. 44. Zuiderweg, E. R., and S. W. Fesik. 1989. Heteronuclear three-dimensional NMR spectroscopy of the inflammatory protein C5a. Biochemistry (Mosc.) 28: 2387-2391. 45. Zhang, X., W. Boyar, M. J. Toth, L. Wennogle, and N. C. Gonnella. 1997. Structural definition of the C5a C terminus by two-dimensional nuclear magnetic resonance spectroscopy. Proteins 28: 261-267. 46. Larkin, M. A., G. Blackshields, N. P. Brown, R. Chenna, P. A. McGettigan, H. McWilliam, F. Valentin, I. M. Wallace, A. Wilm, R. Lopez, J. D. Thompson, T. J. Gibson, and D. G. Higgins. 2007. Clustal W and Clustal X version 2.0. Bioinformatics 23: 2947-2948. 47. Klos, A., E. Wende, K. J. Wareham, and P. N. Monk. 2013. International Union of Pharmacology. LXXXVII. Complement peptide C5a, C4a, and C3a receptors. Pharmacol. Rev. 65: 500-543. 48. Erdei, A., K. Kerekes, and I. Pecht. 1997. Role of C3a and C5a in the activation of mast cells. Exp. Clin. Immunogenet. 14: 16-18. 49. Nilsson, G., M. Johnell, C. H. Hammer, H. L. Tiffany, K. Nilsson, D. D. Metcalfe, A. Siegbahn, and P. M. Murphy. 1996. C3a and C5a are chemotaxins for human mast cells and act through distinct receptors via a pertussis toxin-sensitive signal transduction pathway. J. Immunol. 157: 1693- 1698. 50. Ember, J. A., and T. E. Hugli. 1997. Complement factors and their receptors. Immunopharmacology 38: 3-15.

41 51. Klos, A., A. J. Tenner, K. O. Johswich, R. R. Ager, E. S. Reis, and J. Kohl. 2009. The role of the anaphylatoxins in health and disease. Mol. Immunol. 46: 2753-2766. 52. Ward, P. A. 2004. The dark side of C5a in sepsis. Nat. Rev. Immunol. 4: 133- 142. 53. Bokisch, V. A., and H. J. Muller-Eberhard. 1970. Anaphylatoxin inactivator of human plasma: its isolation and characterization as a carboxypeptidase. J. Clin. Invest. 49: 2427-2436. 54. Huey, R., C. M. Bloor, M. S. Kawahara, and T. E. Hugli. 1983. Potentiation of the anaphylatoxins in vivo using an inhibitor of serum carboxypeptidase N (SCPN). I. Lethality and pathologic effects on pulmonary tissue. Am. J. Pathol. 112: 48-60. 55. Hugli, T. E., and H. J. Muller-Eberhard. 1978. Anaphylatoxins: C3a and C5a. Adv. Immunol. 26: 1-53. 56. Marder, S. R., D. E. Chenoweth, I. M. Goldstein, and H. D. Perez. 1985. Chemotactic responses of human peripheral blood monocytes to the complement-derived peptides C5a and C5a des Arg. J. Immunol. 134: 3325- 3331. 57. Iyer, A., T. M. Woodruff, M. C. Wu, C. Stylianou, R. C. Reid, D. P. Fairlie, S. M. Taylor, and L. Brown. 2011. Inhibition of inflammation and fibrosis by a complement C5a receptor antagonist in DOCA-salt hypertensive rats. J. Cardiovasc. Pharmacol. 58: 479-486. 58. Manthey, H. D., A. C. Thomas, I. A. Shiels, A. Zernecke, T. M. Woodruff, B. Rolfe, and S. M. Taylor. 2011. Complement C5a inhibition reduces atherosclerosis in ApoE-/- mice. FASEB J. 25: 2447-2455. 59. Rua-Figueroa, I., O. Arencibia-Mireles, M. Elvira, C. Erausquin, S. Ojeda, F. Francisco, A. Naranjo, C. Rodriguez-Gallego, I. Garcia-Laorden, J. Rodriguez-Perez, and C. Rodriguez-Lozano. 2010. Factors involved in the progress of preclinical atherosclerosis associated with systemic lupus erythematosus: a 2-year longitudinal study. Ann. Rheum. Dis. 69: 1136-1139. 60. Shagdarsuren, E., K. Bidzhekov, S. F. Mause, S. Simsekyilmaz, T. Polakowski, H. Hawlisch, J. E. Gessner, A. Zernecke, and C. Weber. 2010. C5a receptor targeting in neointima formation after arterial injury in atherosclerosis-prone mice. Circulation 122: 1026-1036.

42 61. Kourtzelis, I., M. M. Markiewski, M. Doumas, S. Rafail, K. Kambas, I. Mitroulis, S. Panagoutsos, P. Passadakis, V. Vargemezis, P. Magotti, H. Qu, T. E. Mollnes, K. Ritis, and J. D. Lambris. 2010. Complement anaphylatoxin C5a contributes to hemodialysis-associated thrombosis. Blood 116: 631-639. 62. Speidl, W. S., S. P. Kastl, R. Hutter, K. M. Katsaros, C. Kaun, G. Bauriedel, G. Maurer, K. Huber, J. J. Badimon, and J. Wojta. 2011. The complement component C5a is present in human coronary lesions in vivo and induces the expression of MMP-1 and MMP-9 in human macrophages in vitro. FASEB J. 25: 35-44. 63. Oikonomopoulou, K., D. Ricklin, P. A. Ward, and J. D. Lambris. 2012. Interactions between coagulation and complement--their role in inflammation. Seminars in immunopathology 34: 151-165. 64. Song, J. J., I. Hwang, K. H. Cho, M. A. Garcia, A. J. Kim, T. H. Wang, T. M. Lindstrom, A. T. Lee, T. Nishimura, L. Zhao, J. Morser, M. Nesheim, S. B. Goodman, D. M. Lee, S. L. Bridges, Jr., P. K. Gregersen, L. L. Leung, and W. H. Robinson. 2011. Plasma carboxypeptidase B downregulates inflammatory responses in autoimmune arthritis. J. Clin. Invest. 121: 3517-3527. 65. Tsuboi, N., T. Ernandez, X. Li, H. Nishi, X. Cullere, D. Mekala, M. Hazen, J. Kohl, D. M. Lee, and T. N. Mayadas. 2011. Regulation of human neutrophil Fcgamma receptor IIa by C5a receptor promotes inflammatory arthritis in mice. Arthritis Rheum. 63: 467-478. 66. Hueber, A. J., D. L. Asquith, A. M. Miller, J. Reilly, S. Kerr, J. Leipe, A. J. Melendez, and I. B. McInnes. 2010. Mast cells express IL-17A in rheumatoid arthritis synovium. J. Immunol. 184: 3336-3340. 67. Hasegawa, K., M. Tamari, C. Shao, M. Shimizu, N. Takahashi, X. Q. Mao, A. Yamasaki, F. Kamada, S. Doi, H. Fujiwara, A. Miyatake, K. Fujita, G. Tamura, Y. Matsubara, T. Shirakawa, and Y. Suzuki. 2004. Variations in the C3, C3a receptor, and C5 genes affect susceptibility to bronchial asthma. Hum. Genet. 115: 295-301. 68. Lewis, A. G., G. Kohl, Q. Ma, P. Devarajan, and J. Kohl. 2008. Pharmacological targeting of C5a receptors during organ preservation improves kidney graft survival. Clin. Exp. Immunol. 153: 117-126.

43 69. Li, Q., Q. Peng, G. Xing, K. Li, N. Wang, C. A. Farrar, L. Meader, S. H. Sacks, and W. Zhou. 2010. Deficiency of C5aR prolongs renal allograft survival. J. Am. Soc. Nephrol. 21: 1344-1353. 70. Bao, L., Y. Wang, M. Haas, and R. J. Quigg. 2011. Distinct roles for C3a and C5a in complement-induced tubulointerstitial injury. Kidney Int. 80: 524-534. 71. Sun, S., Y. Guo, G. Zhao, X. Zhou, J. Li, J. Hu, H. Yu, Y. Chen, H. Song, F. Qiao, G. Xu, F. Yang, Y. Wu, S. Tomlinson, Z. Duan, and Y. Zhou. 2011. Complement and the alternative pathway play an important role in LPS/D- GalN-induced fulminant hepatic failure. PLoS ONE 6: e26838. 72. Gonzalez, J. M., Z. Dong, R. Romero, and G. Girardi. 2011. Cervical remodeling/ripening at term and preterm delivery: the same mechanism initiated by different mediators and different effector cells. PLoS ONE 6: e26877. 73. Liu, B., L. Wei, C. Meyerle, J. Tuo, H. N. Sen, Z. Li, S. Chakrabarty, E. Agron, C. C. Chan, M. L. Klein, E. Chew, F. Ferris, and R. B. Nussenblatt. 2011. Complement component C5a promotes expression of IL-22 and IL-17 from human T cells and its implication in age-related macular degeneration. J. Transl. Med. 9: 1-12. 74. Haque, R., B. Y. Hwang, G. Appelboom, M. A. Piazza, K. Guo, and E. S. Connolly. 2011. Alterations in systemic complement component 3a and 5a levels in patients with cerebral arteriovenous malformations. J. Clin. Neurosci. 18: 1235-1239. 75. Jang, J. H., D. Liang, K. Kido, Y. Sun, D. J. Clark, and T. J. Brennan. 2011. Increased local concentration of complement C5a contributes to incisional pain in mice. J. Neuroinflammation 8: 80. 76. Jang, J. H., J. D. Clark, X. Li, M. S. Yorek, Y. M. Usachev, and T. J. Brennan. 2010. Nociceptive sensitization by complement C5a and C3a in mouse. Pain 148: 343-352. 77. Martinus, R. D., and C. J. Cook. 2011. The effect of complement C5a on mitochondrial functions of PC12 cells. Neuroreport 22: 581-585. 78. Breivik, T., Y. Gundersen, P. Gjermo, S. M. Taylor, T. M. Woodruff, and P. K. Opstad. 2011. Oral treatment with complement factor C5a receptor (CD88) antagonists inhibits experimental periodontitis in rats. J. Periodontal Res. 46: 643-647.

44 79. Liang, S., J. L. Krauss, H. Domon, M. L. McIntosh, K. B. Hosur, H. Qu, F. Li, A. Tzekou, J. D. Lambris, and G. Hajishengallis. 2011. The C5a receptor impairs IL-12-dependent clearance of Porphyromonas gingivalis and is required for induction of periodontal bone loss. J. Immunol. 186: 869-877. 80. Tayman, C., A. Tonbul, H. Kahveci, S. Uysal, B. Koseoglu, M. M. Tatli, and U. Dilmen. 2011. C5a, a complement activation product, is a useful marker in predicting the severity of necrotizing enterocolitis. Tohoku J. Exp. Med. 224: 143-150. 81. von Kockritz-Blickwede, M., S. Konrad, S. Foster, J. E. Gessner, and E. Medina. 2010. Protective role of complement C5a in an experimental model of Staphylococcus aureus bacteremia. J. Innate Immun. 2: 87-92. 82. Patel, S. N., J. Berghout, F. E. Lovegrove, K. Ayi, A. Conroy, L. Serghides, G. Min-oo, D. C. Gowda, J. V. Sarma, D. Rittirsch, P. A. Ward, W. C. Liles, P. Gros, and K. C. Kain. 2008. C5 deficiency and C5a or C5aR blockade protects against cerebral malaria. J. Exp. Med. 205: 1133-1143. 83. Hollmann, T. J., S. L. Mueller-Ortiz, M. C. Braun, and R. A. Wetsel. 2008. Disruption of the C5a receptor gene increases resistance to acute Gram- negative bacteremia and endotoxic shock: opposing roles of C3a and C5a. Mol. Immunol. 45: 1907-1915. 84. Fusakio, M. E., J. P. Mohammed, Y. Laumonnier, K. Hoebe, J. Kohl, and J. Mattner. 2011. C5a regulates NKT and NK cell functions in sepsis. J. Immunol. 187: 5805-5812. 85. Marc, M. M., P. Korosec, M. Kosnik, I. Kern, M. Flezar, S. Suskovic, and J. Sorli. 2004. Complement factors c3a, c4a, and c5a in chronic obstructive pulmonary disease and asthma. Am. J. Respir. Cell Mol. Biol. 31: 216-219. 86. Lajoie, S., I. P. Lewkowich, Y. Suzuki, J. R. Clark, A. A. Sproles, K. Dienger, A. L. Budelsky, and M. Wills-Karp. 2010. Complement-mediated regulation of the IL-17A axis is a central genetic determinant of the severity of experimental allergic asthma. Nat. Immunol. 11: 928-935. 87. Zhang, J., W. Z. Yao, and Y. H. Chen. 2011. [Change in airway anaphylatoxin-complement factors C3a of sputum in patients with chronic obstructive pulmonary disease]. Beijing Da Xue Xue Bao 43: 446-449. 88. Asghar, S. S. 1984. Pharmacological manipulation of complement system. Pharmacol. Rev. 36: 223-244.

45 89. Monk, P. N., A. M. Scola, P. Madala, and D. P. Fairlie. 2007. Function, structure and therapeutic potential of complement C5a receptors. Br. J. Pharmacol. 152: 429-448. 90. Rother, R. P., S. A. Rollins, C. F. Mojcik, R. A. Brodsky, and L. Bell. 2007. Discovery and development of the complement inhibitor eculizumab for the treatment of paroxysmal nocturnal hemoglobinuria. Nat. Biotechnol. 25: 1256- 1264. 91. Chenoweth, D. E., and T. E. Hugli. 1980. Human C5a and C5a analogs as probes of the neutrophil C5a receptor. Mol. Immunol. 17: 151-161. 92. Chenoweth, D. E., B. W. Erickson, and T. E. Hugli. 1979. Human C5a-related synthetic peptides as neutrophil chemotactic factors. Biochem. Biophys. Res. Commun. 86: 227-234. 93. Kawai, M., D. A. Quincy, B. Lane, K. W. Mollison, J. R. Luly, and G. W. Carter. 1991. Identification and synthesis of a receptor binding site of human anaphylatoxin C5a. J. Med. Chem. 34: 2068-2071. 94. Ember, J. A., S. D. Sanderson, S. M. Taylor, M. Kawahara, and T. E. Hugli. 1992. Biologic activity of synthetic analogues of C5a anaphylatoxin. J. Immunol. 148: 3165-3173. 95. Finch, A. M., S. M. Vogen, S. A. Sherman, L. Kirnarsky, S. M. Taylor, and S. D. Sanderson. 1997. Biologically active conformer of the effector region of human C5a and modulatory effects of N-terminal receptor binding determinants on activity. J. Med. Chem. 40: 877-884. 96. Proctor, L. M., T. V. Arumugam, I. Shiels, R. C. Reid, D. P. Fairlie, and S. M. Taylor. 2004. Comparative anti-inflammatory activities of antagonists to C3a and C5a receptors in a rat model of intestinal ischaemia/reperfusion injury. Br. J. Pharmacol. 142: 756-764. 97. Short, A. J., N. J. Paczkowski, S. M. Vogen, S. D. Sanderson, and S. M. Taylor. 1999. Response-selective C5a agonists: differential effects on neutropenia and hypotension in the rat. Br. J. Pharmacol. 128: 511-514. 98. Taylor, S. M., S. A. Sherman, L. Kirnarsky, and S. D. Sanderson. 2001. Development of response-selective agonists of human C5a anaphylatoxin: conformational, biological, and therapeutic considerations. Curr. Med. Chem. 8: 675-684.

46 99. Vogen, S. M., A. M. Finch, S. K. Wadi, J. Thatcher, P. N. Monk, S. M. Taylor, and S. D. Sanderson. 1999. The influence of Lys68 in decepeptide agonists of C5a on C5a receptor binding, activation and selectivity. J. Pept. Res. 53: 8-17. 100. Kawai, M. W., PE.; Luly, JR.; Or, YS. . 1992. New hexa - and heptapeptides are anaphylatoxin antagonists and agonists - for treating inflammatory and immunodeficiency diseases, cancers and severe infections. W. I. P. Organisation, ed. Abbott Laboratories, USA. 101. Otto, M., H. Hawlisch, P. N. Monk, M. Muller, A. Klos, C. L. Karp, and J. Kohl. 2004. C5a mutants are potent antagonists of the C5a receptor (CD88) and of C5L2: position 69 is the locus that determines agonism or antagonism. J. Biol. Chem. 279: 142-151. 102. Scola, A. M., A. Higginbottom, L. J. Partridge, R. C. Reid, T. Woodruff, S. M. Taylor, D. P. Fairlie, and P. N. Monk. 2007. The role of the N-terminal domain of the complement fragment receptor C5L2 in ligand binding. J. Biol. Chem. 282: 3664-3671. 103. Ames, R. S., Y. Li, H. M. Sarau, P. Nuthulaganti, J. J. Foley, C. Ellis, Z. Zeng, K. Su, A. J. Jurewicz, R. P. Hertzberg, D. J. Bergsma, and C. Kumar. 1996. Molecular cloning and characterization of the human anaphylatoxin C3a receptor. J. Biol. Chem. 271: 20231-20234. 104. Konteatis, Z. D., S. J. Siciliano, G. Van Riper, C. J. Molineaux, S. Pandya, P. Fischer, H. Rosen, R. A. Mumford, and M. S. Springer. 1994. Development of C5a receptor antagonists. Differential loss of functional responses. J. Immunol. 153: 4200-4205. 105. Tyndall, J. D., B. Pfeiffer, G. Abbenante, and D. P. Fairlie. 2005. Over one hundred peptide-activated G protein-coupled receptors recognize ligands with turn structure. Chem. Rev. 105: 793-826. 106. Ruiz-Gomez, G., J. D. Tyndall, B. Pfeiffer, G. Abbenante, and D. P. Fairlie. 2010. Update 1 of: Over one hundred peptide-activated G protein-coupled receptors recognize ligands with turn structure. Chem. Rev. 110: PR1-41. 107. Finch, A. M., A. K. Wong, N. J. Paczkowski, S. K. Wadi, D. J. Craik, D. P. Fairlie, and S. M. Taylor. 1999. Low-molecular-weight peptidic and cyclic antagonists of the receptor for the complement factor C5a. J. Med. Chem. 42: 1965-1974.

47 108. Wong, A. K., A. M. Finch, G. K. Pierens, D. J. Craik, S. M. Taylor, and D. P. Fairlie. 1998. Small molecular probes for G-protein-coupled C5a receptors: conformationally constrained antagonists derived from the C terminus of the human plasma protein C5a. J. Med. Chem. 41: 3417-3425. 109. March, D. R., L. M. Proctor, M. J. Stoermer, R. Sbaglia, G. Abbenante, R. C. Reid, T. M. Woodruff, K. Wadi, N. Paczkowski, J. D. Tyndall, S. M. Taylor, and D. P. Fairlie. 2004. Potent cyclic antagonists of the complement C5a receptor on human polymorphonuclear leukocytes. Relationships between structures and activity. Mol. Pharmacol. 65: 868-879. 110. Haynes, D. R., D. G. Harkin, L. P. Bignold, M. J. Hutchens, S. M. Taylor, and D. P. Fairlie. 2000. Inhibition of C5a-induced neutrophil chemotaxis and macrophage cytokine production in vitro by a new C5a receptor antagonist. Biochem. Pharmacol. 60: 729-733. 111. Mollnes, T. E., O. L. Brekke, M. Fung, H. Fure, D. Christiansen, G. Bergseth, V. Videm, K. T. Lappegard, J. Kohl, and J. D. Lambris. 2002. Essential role of the C5a receptor in E coli-induced oxidative burst and phagocytosis revealed by a novel lepirudin-based human whole blood model of inflammation. Blood 100: 1869-1877. 112. Strachan, A. J., I. A. Shiels, R. C. Reid, D. P. Fairlie, and S. M. Taylor. 2001. Inhibition of immune-complex mediated dermal inflammation in rats following either oral or topical administration of a small molecule C5a receptor antagonist. Br. J. Pharmacol. 134: 1778-1786. 113. Scully, C. C., J. S. Blakeney, R. Singh, H. N. Hoang, G. Abbenante, R. C. Reid, and D. P. Fairlie. 2010. Selective hexapeptide agonists and antagonists for human complement C3a receptor. J. Med. Chem. 53: 4938-4948. 114. Kohl, J. 2006. Drug evaluation: the C5a receptor antagonist PMX-53. Curr. Opin. Mol. Ther. 8: 529-538. 115. Morgan, M., A. C. Bulmer, T. M. Woodruff, L. M. Proctor, H. M. Williams, S. Z. Stocks, S. Pollitt, S. M. Taylor, and I. A. Shiels. 2008. Pharmacokinetics of a C5a receptor antagonist in the rat after different sites of enteral administration. Eur. J. Pharm. Sci. 33: 390-398. 116. Delisle Milton, R. C., S. C. Milton, and A. R. Chamberlin. 2011. Improving the Fmoc Solid Phase Synthesis of the Cyclic Hexapeptide Complement C5a Antagonist, PMX205. Int J Pept Res Ther 17: 337-342.

48 117. Woodruff, T. M., J. W. Crane, L. M. Proctor, K. M. Buller, A. B. Shek, K. de Vos, S. Pollitt, H. M. Williams, I. A. Shiels, P. N. Monk, and S. M. Taylor. 2006. Therapeutic activity of C5a receptor antagonists in a rat model of neurodegeneration. FASEB J. 20: 1407-1417. 118. Fonseca, M. I., R. R. Ager, S. H. Chu, O. Yazan, S. D. Sanderson, F. M. LaFerla, S. M. Taylor, T. M. Woodruff, and A. J. Tenner. 2009. Treatment with a C5aR antagonist decreases pathology and enhances behavioral performance in murine models of Alzheimer's disease. J. Immunol. 183: 1375- 1383. 119. Woodruff, T. M., K. J. Costantini, J. W. Crane, J. D. Atkin, P. N. Monk, S. M. Taylor, and P. G. Noakes. 2008. The complement factor C5a contributes to pathology in a rat model of amyotrophic lateral sclerosis. J. Immunol. 181: 8727-8734. 120. Proctor, L. M., T. M. Woodruff, P. Sharma, I. A. Shiels, and S. M. Taylor. 2006. Transdermal pharmacology of small molecule cyclic C5a antagonists. Adv. Exp. Med. Biol. 586: 329-345. 121. Schnatbaum, K., E. Locardi, D. Scharn, U. Richter, H. Hawlisch, J. Knolle, and T. Polakowski. 2006. Peptidomimetic C5a receptor antagonists with hydrophobic substitutions at the C-terminus: increased receptor specificity and in vivo activity. Bioorg. Med. Chem. Lett. 16: 5088-5092. 122. Gueler, F., S. Rong, W. Gwinner, M. Mengel, V. Brocker, S. Schon, T. F. Greten, H. Hawlisch, T. Polakowski, K. Schnatbaum, J. Menne, H. Haller, and N. Shushakova. 2008. Complement 5a receptor inhibition improves renal allograft survival. J. Am. Soc. Nephrol. 19: 2302-2312. 123. Ricklin, D., and J. D. Lambris. 2007. Complement-targeted therapeutics. Nat. Biotechnol. 25: 1265-1275. 124. Boor, P., A. Konieczny, L. Villa, A. L. Schult, E. Bucher, S. Rong, U. Kunter, C. R. van Roeyen, T. Polakowski, H. Hawlisch, S. Hillebrandt, F. Lammert, F. Eitner, J. Floege, and T. Ostendorf. 2007. Complement C5 mediates experimental tubulointerstitial fibrosis. J. Am. Soc. Nephrol. 18: 1508-1515. 125. Qu, H., D. Ricklin, and J. D. Lambris. 2009. Recent developments in low molecular weight complement inhibitors. Mol. Immunol. 47: 185-195. 126. Sumichika, H., K. Sakata, N. Sato, S. Takeshita, S. Ishibuchi, M. Nakamura, T. Kamahori, S. Ehara, K. Itoh, T. Ohtsuka, T. Ohbora, T. Mishina, H. Komatsu,

49 and Y. Naka. 2002. Identification of a potent and orally active non-peptide C5a receptor antagonist. J. Biol. Chem. 277: 49403-49407. 127. Waters, S. M., R. M. Brodbeck, J. Steflik, J. Yu, C. Baltazar, A. E. Peck, D. Severance, L. Y. Zhang, K. Currie, B. L. Chenard, A. J. Hutchison, G. Maynard, and J. E. Krause. 2005. Molecular characterization of the gerbil C5a receptor and identification of a transmembrane domain V amino acid that is crucial for small molecule antagonist interaction. J. Biol. Chem. 280: 40617- 40623. 128. Barbay, J. K., Y. Gong, M. Buntinx, J. Li, C. Claes, P. J. Hornby, G. Van Lommen, J. Van Wauwe, and W. He. 2008. Synthesis and characterization of 5,6,7,8-tetrahydroquinoline C5a receptor antagonists. Bioorg. Med. Chem. Lett. 18: 2544-2548. 129. Gong, Y., J. K. Barbay, M. Buntinx, J. Li, J. V. Wauwe, C. Claes, G. V. Lommen, P. J. Hornby, and W. He. 2008. Design and optimization of aniline- substituted tetrahydroquinoline C5a receptor antagonists. Bioorg. Med. Chem. Lett. 18: 3852-3855. 130. Ricklin, D., G. Hajishengallis, K. Yang, and J. D. Lambris. 2010. Complement: a key system for immune surveillance and homeostasis. Nat. Immunol. 11: 785-797. 131. Walport, M. J. 2001. Complement. First of two parts. N. Engl. J. Med. 344: 1058-1066. 132. Shi, Y., Y. Tohyama, T. Kadono, J. He, S. M. Miah, R. Hazama, C. Tanaka, K. Tohyama, and H. Yamamura. 2006. Protein-tyrosine kinase Syk is required for pathogen engulfment in complement-mediated phagocytosis. Blood 107: 4554-4562. 133. Carroll, M. C. 2004. The complement system in regulation of adaptive immunity. Nat. Immunol. 5: 981-986. 134. Mastellos, D., and J. D. Lambris. 2006. Cross-disciplinary research stirs new challenges into the study of the structure, function and systems biology of complement. Adv. Exp. Med. Biol. 586: 1-16. 135. Wetsel, R. A. 1995. Expression of the complement C5a anaphylatoxin receptor (C5aR) on non-myeloid cells. Immunol. Lett. 44: 183-187. 136. Buchner, R. R., T. E. Hugli, J. A. Ember, and E. L. Morgan. 1995. Expression of functional receptors for human C5a anaphylatoxin (CD88) on the human

50 hepatocellular carcinoma cell line HepG2. Stimulation of acute-phase protein- specific mRNA and protein synthesis by human C5a anaphylatoxin. J. Immunol. 155: 308-315. 137. Fregonese, L., F. J. Swan, A. van Schadewijk, M. Dolhnikoff, M. A. Santos, M. R. Daha, J. Stolk, T. Tschernig, P. J. Sterk, P. S. Hiemstra, K. F. Rabe, and T. Mauad. 2005. Expression of the anaphylatoxin receptors C3aR and C5aR is increased in fatal asthma. J. Allergy Clin. Immunol. 115: 1148-1154. 138. Gasque, P., P. Chan, M. Fontaine, A. Ischenko, M. Lamacz, O. Gotze, and B. P. Morgan. 1995. Identification and characterization of the complement C5a anaphylatoxin receptor on human astrocytes. J. Immunol. 155: 4882-4889. 139. Haviland, D. L., R. L. McCoy, W. T. Whitehead, H. Akama, E. P. Molmenti, A. Brown, J. C. Haviland, W. C. Parks, D. H. Perlmutter, and R. A. Wetsel. 1995. Cellular expression of the C5a anaphylatoxin receptor (C5aR): demonstration of C5aR on nonmyeloid cells of the liver and lung. J. Immunol. 154: 1861-1869. 140. Gasque, P., S. K. Singhrao, J. W. Neal, O. Gotze, and B. P. Morgan. 1997. Expression of the receptor for complement C5a (CD88) is up-regulated on reactive astrocytes, microglia, and endothelial cells in the inflamed human central nervous system. Am. J. Pathol. 150: 31-41. 141. Stahel, P. F., K. Frei, H. P. Eugster, A. Fontana, K. M. Hummel, R. A. Wetsel, R. S. Ames, and S. R. Barnum. 1997. TNF-alpha-mediated expression of the receptor for anaphylatoxin C5a on neurons in experimental Listeria meningoencephalitis. J. Immunol. 159: 861-869. 142. Foreman, K. E., A. A. Vaporciyan, B. K. Bonish, M. L. Jones, K. J. Johnson, M. M. Glovsky, S. M. Eddy, and P. A. Ward. 1994. C5a-induced expression of P-selectin in endothelial cells. J. Clin. Invest. 94: 1147-1155. 143. Van Beek, J., M. Bernaudin, E. Petit, P. Gasque, A. Nouvelot, E. T. MacKenzie, and M. Fontaine. 2000. Expression of receptors for complement anaphylatoxins C3a and C5a following permanent focal cerebral ischemia in the mouse. Exp. Neurol. 161: 373-382. 144. Langeggen, H., E. Johnson, and G. Hetland. 2001. Effects of C5a and FMLP on interleukin-8 production and proliferation of human umbilical vein endothelial cells. Inflammation 25: 83-89.

51 145. Lundberg, C., F. Marceau, and T. E. Hugli. 1987. C5a-induced hemodynamic and hematologic changes in the rabbit. Role of cyclooxygenase products and polymorphonuclear leukocytes. Am. J. Pathol. 128: 471-483. 146. Marceau, F., C. Lundberg, and T. E. Hugli. 1987. Effects of the anaphylatoxins on circulation. Immunopharmacology 14: 67-84. 147. Short, A., A. K. Wong, A. M. Finch, G. Haaima, I. A. Shiels, D. P. Fairlie, and S. M. Taylor. 1999. Effects of a new C5a receptor antagonist on C5a- and endotoxin-induced neutropenia in the rat. Br. J. Pharmacol. 126: 551-554. 148. Shapira, L., W. A. Soskolne, Y. Houri, V. Barak, A. Halabi, and A. Stabholz. 1996. Protection against endotoxic shock and lipopolysaccharide-induced local inflammation by tetracycline: correlation with inhibition of cytokine secretion. Infect. Immun. 64: 825-828. 149. Huber-Lang, M., E. M. Younkin, J. V. Sarma, N. Riedemann, S. R. McGuire, K. T. Lu, R. Kunkel, J. G. Younger, F. S. Zetoune, and P. A. Ward. 2002. Generation of C5a by phagocytic cells. Am. J. Pathol. 161: 1849-1859. 150. Smedegard, G., L. X. Cui, and T. E. Hugli. 1989. Endotoxin-induced shock in the rat. A role for C5a. Am. J. Pathol. 135: 489-497. 151. Wagner, J. G., and R. A. Roth. 2000. Neutrophil migration mechanisms, with an emphasis on the pulmonary vasculature. Pharmacol. Rev. 52: 349-374. 152. Imhof, B. A., and M. Aurrand-Lions. 2004. Adhesion mechanisms regulating the migration of monocytes. Nat. Rev. Immunol. 4: 432-444. 153. Yamamoto, T. 2007. Roles of the ribosomal protein S19 dimer and the C5a receptor in pathophysiological functions of phagocytic leukocytes. Pathol. Int. 57: 1-11. 154. Nishiura, H., S. Tanase, Y. Sibuya, T. Nishimura, and T. Yamamoto. 1999. Determination of the cross-linked residues in homo-dimerization of S19 ribosomal protein concomitant with exhibition of monocyte chemotactic activity. Lab. Invest. 79: 915-923. 155. Nishiura, H., Y. Shibuya, and T. Yamamoto. 1998. S19 ribosomal protein cross-linked dimer causes monocyte-predominant infiltration by means of molecular mimicry to complement C5a. Lab. Invest. 78: 1615-1623. 156. Nishiura, H., H. Nonaka, I. S. Revollo, U. Semba, Y. Li, Y. Ota, A. Irie, K. Harada, J. H. Kehrl, and T. Yamamoto. 2009. Pro- and anti-apoptotic dual

52 functions of the C5a receptor: involvement of regulator of G protein signaling 3 and extracellular signal-regulated kinase. Lab. Invest. 89: 676-694. 157. Nishiura, H., J. Chen, Y. Ota, U. Semba, H. Higuchi, T. Nakashima, and T. Yamamoto. 2010. Base of molecular mimicry between human ribosomal protein S19 dimer and human C5a anaphylatoxin. Int. Immunopharmacol. 10: 1541-1547. 158. Nishiura, H., K. Tokita, Y. Li, K. Harada, T. M. Woodruff, S. M. Taylor, T. K. Nsiama, N. Nishino, and T. Yamamoto. 2010. The role of the ribosomal protein S19 C-terminus in Gi protein-dependent alternative activation of p38 MAP kinase via the C5a receptor in HMC-1 cells. Apoptosis 15: 966-981. 159. Balkwill, F., and A. Mantovani. 2001. Inflammation and cancer: back to Virchow? Lancet 357: 539-545. 160. Coussens, L. M., and Z. Werb. 2002. Inflammation and cancer. Nature 420: 860-867. 161. Mantovani, A., P. Allavena, A. Sica, and F. Balkwill. 2008. Cancer-related inflammation. Nature 454: 436-444. 162. Swann, J. B., and M. J. Smyth. 2007. Immune surveillance of tumors. J. Clin. Invest. 117: 1137-1146. 163. Dong, H., S. E. Strome, D. R. Salomao, H. Tamura, F. Hirano, D. B. Flies, P. C. Roche, J. Lu, G. Zhu, K. Tamada, V. A. Lennon, E. Celis, and L. Chen. 2002. Tumor-associated B7-H1 promotes T-cell apoptosis: a potential mechanism of immune evasion. Nat. Med. 8: 793-800. 164. Sica, A., and V. Bronte. 2007. Altered macrophage differentiation and immune dysfunction in tumor development. J. Clin. Invest. 117: 1155-1166. 165. Kusmartsev, S., S. Nagaraj, and D. I. Gabrilovich. 2005. Tumor-associated CD8+ T cell tolerance induced by bone marrow-derived immature myeloid cells. J. Immunol. 175: 4583-4592. 166. Niculescu, F., H. G. Rus, M. Retegan, and R. Vlaicu. 1992. Persistent complement activation on tumor cells in breast cancer. Am. J. Pathol. 140: 1039-1043. 167. Baatrup, G., N. Qvist, A. Junker, K. E. Larsen, and C. Zimmermann-Nielsen. 1994. Activity and activation of the complement system in patients being operated on for cancer of the colon. Eur. J. Surg. 160: 503-510.

53 168. Ytting, H., J. C. Jensenius, I. J. Christensen, S. Thiel, and H. J. Nielsen. 2004. Increased activity of the mannan-binding lectin complement activation pathway in patients with colorectal cancer. Scand. J. Gastroenterol. 39: 674- 679. 169. Bjorge, L., J. Hakulinen, O. K. Vintermyr, H. Jarva, T. S. Jensen, O. E. Iversen, and S. Meri. 2005. Ascitic complement system in ovarian cancer. Br. J. Cancer 92: 895-905. 170. Markiewski, M. M., R. A. DeAngelis, F. Benencia, S. K. Ricklin-Lichtsteiner, A. Koutoulaki, C. Gerard, G. Coukos, and J. D. Lambris. 2008. Modulation of the antitumor immune response by complement. Nat. Immunol. 9: 1225-1235. 171. Markiewski, M. M., and J. D. Lambris. 2007. The role of complement in inflammatory diseases from behind the scenes into the spotlight. Am. J. Pathol. 171: 715-727. 172. Bhardwaj, N. 2007. Harnessing the immune system to treat cancer. J. Clin. Invest. 117: 1130-1136. 173. Lin, W. W., and M. Karin. 2007. A cytokine-mediated link between innate immunity, inflammation, and cancer. J. Clin. Invest. 117: 1175-1183. 174. Candido, J., and T. Hagemann. 2013. Cancer-related inflammation. J. Clin. Immunol. 33 Suppl 1: S79-84. 175. Nitta, H., Y. Wada, Y. Kawano, Y. Murakami, A. Irie, K. Taniguchi, K. Kikuchi, G. Yamada, K. Suzuki, J. Honda, M. Wilson-Morifuji, N. Araki, M. Eto, H. Baba, and T. Imamura. 2013. Enhancement of human cancer cell motility and invasiveness by anaphylatoxin C5a via aberrantly expressed C5a receptor (CD88). Clin. Cancer Res. 19: 2004-2013. 176. Markiewski, M. M., and J. D. Lambris. 2009. Unwelcome complement. Cancer Res. 69: 6367-6370. 177. Akira, S., S. Uematsu, and O. Takeuchi. 2006. Pathogen recognition and innate immunity. Cell 124: 783-801. 178. Janeway, C. A., Jr., and R. Medzhitov. 2002. Innate immune recognition. Annual review of immunology 20: 197-216. 179. Akira, S., and K. Takeda. 2004. Toll-like receptor signalling. Nat. Rev. Immunol. 4: 499-511. 180. Hajishengallis, G., and J. D. Lambris. 2010. Crosstalk pathways between Toll- like receptors and the complement system. Trends Immunol. 31: 154-163.

54 181. Natarajan, M., K. M. Lin, R. C. Hsueh, P. C. Sternweis, and R. Ranganathan. 2006. A global analysis of cross-talk in a mammalian cellular signalling network. Nat. Cell Biol. 8: 571-580. 182. Postma, B., M. J. Poppelier, J. C. van Galen, E. R. Prossnitz, J. A. van Strijp, C. J. de Haas, and K. P. van Kessel. 2004. Chemotaxis inhibitory protein of Staphylococcus aureus binds specifically to the C5a and formylated peptide receptor. J. Immunol. 172: 6994-7001. 183. Wang, M., J. L. Krauss, H. Domon, K. B. Hosur, S. Liang, P. Magotti, M. Triantafilou, K. Triantafilou, J. D. Lambris, and G. Hajishengallis. 2010. Microbial hijacking of complement-toll-like receptor crosstalk. Sci. Signal. 3: ra11. 184. Rossol, M., H. Heine, U. Meusch, D. Quandt, C. Klein, M. J. Sweet, and S. Hauschildt. 2011. LPS-induced cytokine production in human monocytes and macrophages. Crit. Rev. Immunol. 31: 379-446. 185. Bjornson, A. B., and H. S. Bjornson. 1977. Activation of complement by opportunist pathogens and chemotypes of Salmonella minnesota. Infection and immunity 16: 748-753. 186. Morrison, D. C., and L. F. Kline. 1977. Activation of the classical and properdin pathways of complement by bacterial lipopolysaccharides (LPS). Journal of immunology (Baltimore, Md. : 1950) 118: 362-368. 187. Sladowski, D., A. Kinsner, I. Langezaal, S. Kay, and S. Coecke. 2001. Activation of the complement system as an indicator of pyrogenic reaction to lipopolysaccharide (LPS). Toxicology in vitro : an international journal published in association with BIBRA 15: 339-342. 188. Okusawa, S., C. A. Dinarello, K. B. Yancey, S. Endres, T. J. Lawley, M. M. Frank, J. F. Burke, and J. A. Gelfand. 1987. C5a induction of human interleukin 1. Synergistic effect with endotoxin or interferon-gamma. Journal of immunology (Baltimore, Md. : 1950) 139: 2635-2640. 189. Hawlisch, H., Y. Belkaid, R. Baelder, D. Hildeman, C. Gerard, and J. Kohl. 2005. C5a negatively regulates toll-like receptor 4-induced immune responses. Immunity 22: 415-426. 190. Okazaki, N., K. Hazeki, T. Izumi, K. Nigorikawa, and O. Hazeki. 2011. C5a controls TLR-induced IL-10 and IL-12 production independent of phosphoinositide 3-kinase. J Biochem 149: 265-274.

55 191. Bosmann, M., V. R. Patel, N. F. Russkamp, F. Pache, F. S. Zetoune, J. V. Sarma, and P. A. Ward. 2011. MyD88-dependent production of IL-17F is modulated by the anaphylatoxin C5a via the Akt signaling pathway. FASEB J. 25: 4222-4232. 192. Bosmann, M., J. V. Sarma, G. Atefi, F. S. Zetoune, and P. A. Ward. 2012. Evidence for anti-inflammatory effects of C5a on the innate IL-17A/IL-23 axis. FASEB J. 26: 1640-1651. 193. Seow, V., J. Lim, A. Iyer, J. Y. Suen, J. K. Ariffin, D. M. Hohenhaus, M. J. Sweet, and D. P. Fairlie. 2013. Inflammatory responses induced by lipopolysaccharide are amplified in primary human monocytes but suppressed in macrophages by complement protein c5a. J. Immunol. 191: 4308-4316. 194. Cavaillon, J. M., C. Fitting, and N. Haeffner-Cavaillon. 1990. Recombinant C5a enhances interleukin 1 and tumor necrosis factor release by lipopolysaccharide-stimulated monocytes and macrophages. Eur. J. Immunol. 20: 253-257. 195. Zaal, A., S. N. Lissenberg-Thunnissen, G. van Schijndel, D. Wouters, S. M. van Ham, and A. ten Brinke. 2013. Crosstalk between Toll like receptors and C5a receptor in human monocyte derived DCs suppress inflammatory cytokine production. Immunobiology 218: 175-180. 196. Copeland, S., H. S. Warren, S. F. Lowry, S. E. Calvano, and D. Remick. 2005. Acute inflammatory response to endotoxin in mice and humans. Clin. Diagn. Lab. Immunol. 12: 60-67. 197. Schroder, K., K. M. Irvine, M. S. Taylor, N. J. Bokil, K. A. Le Cao, K. A. Masterman, L. I. Labzin, C. A. Semple, R. Kapetanovic, L. Fairbairn, A. Akalin, G. J. Faulkner, J. K. Baillie, M. Gongora, C. O. Daub, H. Kawaji, G. J. McLachlan, N. Goldman, S. M. Grimmond, P. Carninci, H. Suzuki, Y. Hayashizaki, B. Lenhard, D. A. Hume, and M. J. Sweet. 2012. Conservation and divergence in Toll-like receptor 4-regulated gene expression in primary human versus mouse macrophages. Proc. Natl. Acad. Sci. U. S. A. 109: E944- 953. 198. Ariffin, J. K., and M. J. Sweet. 2013. Differences in the repertoire, regulation and function of Toll-like Receptors and inflammasome-forming Nod-like Receptors between human and mouse. Curr. Opin. Microbiol.

56 199. Seok, J., H. S. Warren, A. G. Cuenca, M. N. Mindrinos, H. V. Baker, W. Xu, D. R. Richards, G. P. McDonald-Smith, H. Gao, L. Hennessy, C. C. Finnerty, C. M. Lopez, S. Honari, E. E. Moore, J. P. Minei, J. Cuschieri, P. E. Bankey, J. L. Johnson, J. Sperry, A. B. Nathens, T. R. Billiar, M. A. West, M. G. Jeschke, M. B. Klein, R. L. Gamelli, N. S. Gibran, B. H. Brownstein, C. Miller- Graziano, S. E. Calvano, P. H. Mason, J. P. Cobb, L. G. Rahme, S. F. Lowry, R. V. Maier, L. L. Moldawer, D. N. Herndon, R. W. Davis, W. Xiao, and R. G. Tompkins. 2013. Genomic responses in mouse models poorly mimic human inflammatory diseases. Proc. Natl. Acad. Sci. U. S. A. 110: 3507-3512. 200. Wellen, K. E., and G. S. Hotamisligil. 2005. Inflammation, stress, and diabetes. J. Clin. Invest. 115: 1111-1119. 201. Van Gaal, L. F., I. L. Mertens, and C. E. De Block. 2006. Mechanisms linking obesity with cardiovascular disease. Nature 444: 875-880. 202. Navab, M., N. Gharavi, and A. D. Watson. 2008. Inflammation and metabolic disorders. Curr. Opin. Clin. Nutr. Metab. Care 11: 459-464. 203. Hotamisligil, G. S. 2006. Inflammation and metabolic disorders. Nature 444: 860-867. 204. Shoelson, S. E., J. Lee, and A. B. Goldfine. 2006. Inflammation and insulin resistance. J. Clin. Invest. 116: 1793-1801. 205. Xu, H., G. T. Barnes, Q. Yang, G. Tan, D. Yang, C. J. Chou, J. Sole, A. Nichols, J. S. Ross, L. A. Tartaglia, and H. Chen. 2003. Chronic inflammation in fat plays a crucial role in the development of obesity-related insulin resistance. J. Clin. Invest. 112: 1821-1830. 206. Iyer, A., D. P. Fairlie, J. B. Prins, B. D. Hammock, and L. Brown. 2010. Inflammatory lipid mediators in adipocyte function and obesity. Nat. Rev. Endocrinol. 6: 71-82. 207. Scrimshaw, N. S., and J. P. SanGiovanni. 1997. Synergism of nutrition, infection, and immunity: an overview. Am. J. Clin. Nutr. 66: 464S-477S. 208. Roth, J., A. L. Szulc, and A. Danoff. 2011. Energy, evolution, and human diseases: an overview. Am. J. Clin. Nutr. 209. Martinez, J. A., and G. Fruhbeck. 1996. Regulation of energy balance and adiposity: a model with new approaches. Rev. Esp. Fisiol. 52: 255-258. 210. Friedman, J. M. 2000. Obesity in the new millennium. Nature 404: 632-634.

57 211. Vijay-Kumar, M., J. D. Aitken, F. A. Carvalho, T. C. Cullender, S. Mwangi, S. Srinivasan, S. V. Sitaraman, R. Knight, R. E. Ley, and A. T. Gewirtz. 2010. Metabolic syndrome and altered gut microbiota in mice lacking Toll-like receptor 5. Science 328: 228-231. 212. Li, P., and G. S. Hotamisligil. 2010. Metabolism: Host and microbes in a pickle. Nature 464: 1287-1288. 213. Gabrielsson, B. G., J. M. Johansson, M. Lonn, M. Jernas, T. Olbers, M. Peltonen, I. Larsson, L. Lonn, L. Sjostrom, B. Carlsson, and L. M. Carlsson. 2003. High expression of complement components in omental adipose tissue in obese men. Obes. Res. 11: 699-708. 214. Weyer, C., P. A. Tataranni, and R. E. Pratley. 2000. Insulin action and insulinemia are closely related to the fasting complement C3, but not acylation stimulating protein concentration. Diabetes Care 23: 779-785. 215. Peake, P. W., A. D. Kriketos, L. V. Campbell, and J. A. Charlesworth. 2005. Response of the alternative complement pathway to an oral fat load in first- degree relatives of subjects with type II diabetes. Int. J. Obes. 29: 429-435. 216. Hernandez-Mijares, A., M. M. Jarabo-Bueno, A. Lopez-Ruiz, E. Sola- Izquierdo, C. Morillas-Arino, and M. L. Martinez-Triguero. 2007. Levels of C3 in patients with severe, morbid and extreme obesity: its relationship to insulin resistance and different cardiovascular risk factors. Int. J. Obes. 31: 927-932. 217. Lim, J., A. Iyer, J. Y. Suen, V. Seow, R. C. Reid, L. Brown, and D. P. Fairlie. 2013. C5aR and C3aR antagonists each inhibit diet-induced obesity, metabolic dysfunction, and adipocyte and macrophage signaling. FASEB J. 27: 822-831. 218. Phieler, J., K. J. Chung, A. Chatzigeorgiou, A. Klotzsche-von Ameln, R. Garcia-Martin, D. Sprott, M. Moisidou, T. Tzanavari, B. Ludwig, E. Baraban, M. Ehrhart-Bornstein, S. R. Bornstein, H. Mziaut, M. Solimena, K. P. Karalis, M. Economopoulou, J. D. Lambris, and T. Chavakis. 2013. The complement anaphylatoxin c5a receptor contributes to obese adipose tissue inflammation and insulin resistance. J. Immunol. 191: 4367-4374. 219. Jaworski, K., M. Ahmadian, R. E. Duncan, E. Sarkadi-Nagy, K. A. Varady, M. K. Hellerstein, H. Y. Lee, V. T. Samuel, G. I. Shulman, K. H. Kim, S. de Val, C. Kang, and H. S. Sul. 2009. AdPLA ablation increases lipolysis and

58 prevents obesity induced by high-fat feeding or leptin deficiency. Nat. Med. 15: 159-168. 220. Kosteli, A., E. Sugaru, G. Haemmerle, J. F. Martin, J. Lei, R. Zechner, and A. W. Ferrante, Jr. 2010. Weight loss and lipolysis promote a dynamic immune response in murine adipose tissue. J. Clin. Invest. 120: 3466-3479. 221. Dzierzak, E., and N. A. Speck. 2008. Of lineage and legacy: the development of mammalian hematopoietic stem cells. Nat. Immunol. 9: 129-136. 222. Gordon, S., and P. R. Taylor. 2005. Monocyte and macrophage heterogeneity. Nat. Rev. Immunol. 5: 953-964. 223. Murano, I., G. Barbatelli, V. Parisani, C. Latini, G. Muzzonigro, M. Castellucci, and S. Cinti. 2008. Dead adipocytes, detected as crown-like structures, are prevalent in visceral fat depots of genetically obese mice. J. Lipid Res. 49: 1562-1568. 224. Gordon, S. 2007. The macrophage: past, present and future. Eur. J. Immunol. 37 Suppl 1: S9-17. 225. Shi, C., and E. G. Pamer. 2011. Monocyte recruitment during infection and inflammation. Nat. Rev. Immunol. 11: 762-774. 226. Mosser, D. M., and J. P. Edwards. 2008. Exploring the full spectrum of macrophage activation. Nat. Rev. Immunol. 8: 958-969. 227. Kono, H., and K. L. Rock. 2008. How dying cells alert the immune system to danger. Nat. Rev. Immunol. 8: 279-289. 228. Duffield, J. S. 2003. The inflammatory macrophage: a story of Jekyll and Hyde. Clin. Sci. (Lond.) 104: 27-38. 229. McNally, A. K., and J. M. Anderson. 2011. Macrophage fusion and multinucleated giant cells of inflammation. Adv. Exp. Med. Biol. 713: 97-111. 230. Weisberg, S. P., D. McCann, M. Desai, M. Rosenbaum, R. L. Leibel, and A. W. Ferrante, Jr. 2003. Obesity is associated with macrophage accumulation in adipose tissue. J. Clin. Invest. 112: 1796-1808. 231. Cancello, R., and K. Clement. 2006. Is obesity an inflammatory illness? Role of low-grade inflammation and macrophage infiltration in human white adipose tissue. BJOG 113: 1141-1147. 232. Cancello, R., C. Henegar, N. Viguerie, S. Taleb, C. Poitou, C. Rouault, M. Coupaye, V. Pelloux, D. Hugol, J. L. Bouillot, A. Bouloumie, G. Barbatelli, S. Cinti, P. A. Svensson, G. S. Barsh, J. D. Zucker, A. Basdevant, D. Langin, and

59 K. Clement. 2005. Reduction of macrophage infiltration and chemoattractant gene expression changes in white adipose tissue of morbidly obese subjects after surgery-induced weight loss. Diabetes 54: 2277-2286. 233. Schenk, S., M. Saberi, and J. M. Olefsky. 2008. Insulin sensitivity: modulation by nutrients and inflammation. J. Clin. Invest. 118: 2992-3002. 234. Lumeng, C. N., J. L. Bodzin, and A. R. Saltiel. 2007. Obesity induces a phenotypic switch in adipose tissue macrophage polarization. J. Clin. Invest. 117: 175-184. 235. Martinez, F. O., A. Sica, A. Mantovani, and M. Locati. 2008. Macrophage activation and polarization. Front. Biosci. 13: 453-461. 236. Mantovani, A., A. Sica, S. Sozzani, P. Allavena, A. Vecchi, and M. Locati. 2004. The chemokine system in diverse forms of macrophage activation and polarization. Trends Immunol. 25: 677-686. 237. Gordon, S. 2003. Alternative activation of macrophages. Nat. Rev. Immunol. 3: 23-35. 238. Odegaard, J. I., R. R. Ricardo-Gonzalez, M. H. Goforth, C. R. Morel, V. Subramanian, L. Mukundan, A. Red Eagle, D. Vats, F. Brombacher, A. W. Ferrante, and A. Chawla. 2007. Macrophage-specific PPARgamma controls alternative activation and improves insulin resistance. Nature 447: 1116-1120. 239. Cinti, S., G. Mitchell, G. Barbatelli, I. Murano, E. Ceresi, E. Faloia, S. Wang, M. Fortier, A. S. Greenberg, and M. S. Obin. 2005. Adipocyte death defines macrophage localization and function in adipose tissue of obese mice and humans. J. Lipid Res. 46: 2347-2355. 240. Uysal, K. T., S. M. Wiesbrock, M. W. Marino, and G. S. Hotamisligil. 1997. Protection from obesity-induced insulin resistance in mice lacking TNF-alpha function. Nature 389: 610-614. 241. Weisberg, S. P., D. Hunter, R. Huber, J. Lemieux, S. Slaymaker, K. Vaddi, I. Charo, R. L. Leibel, and A. W. Ferrante, Jr. 2006. CCR2 modulates inflammatory and metabolic effects of high-fat feeding. J. Clin. Invest. 116: 115-124. 242. Patsouris, D., P. P. Li, D. Thapar, J. Chapman, J. M. Olefsky, and J. G. Neels. 2008. Ablation of CD11c-positive cells normalizes insulin sensitivity in obese insulin resistant animals. Cell Metab. 8: 301-309.

60 243. Kang, K., S. M. Reilly, V. Karabacak, M. R. Gangl, K. Fitzgerald, B. Hatano, and C. H. Lee. 2008. Adipocyte-derived Th2 cytokines and myeloid PPARdelta regulate macrophage polarization and insulin sensitivity. Cell Metab. 7: 485-495. 244. Odegaard, J. I., and A. Chawla. 2011. Alternative macrophage activation and metabolism. Annu. Rev. Pathol. 6: 275-297. 245. Stout, R. D., C. Jiang, B. Matta, I. Tietzel, S. K. Watkins, and J. Suttles. 2005. Macrophages sequentially change their functional phenotype in response to changes in microenvironmental influences. J. Immunol. 175: 342-349. 246. del Pozo, M. A., P. Sanchez-Mateos, M. Nieto, and F. Sanchez-Madrid. 1995. Chemokines regulate cellular polarization and adhesion receptor redistribution during lymphocyte interaction with endothelium and extracellular matrix. Involvement of cAMP signaling pathway. J. Cell Biol. 131: 495-508. 247. Sanchez-Madrid, F., and M. A. del Pozo. 1999. Leukocyte polarization in cell migration and immune interactions. EMBO J. 18: 501-511. 248. Mylonas, K. J., M. G. Nair, L. Prieto-Lafuente, D. Paape, and J. E. Allen. 2009. Alternatively activated macrophages elicited by helminth infection can be reprogrammed to enable microbial killing. J. Immunol. 182: 3084-3094. 249. Modolell, M., I. M. Corraliza, F. Link, G. Soler, and K. Eichmann. 1995. Reciprocal regulation of the nitric oxide synthase/arginase balance in mouse bone marrow-derived macrophages by TH1 and TH2 cytokines. Eur. J. Immunol. 25: 1101-1104. 250. Edwards, J. P., X. Zhang, K. A. Frauwirth, and D. M. Mosser. 2006. Biochemical and functional characterization of three activated macrophage populations. J. Leukoc. Biol. 80: 1298-1307. 251. Forlenza, M., I. R. Fink, G. Raes, and G. F. Wiegertjes. 2011. Heterogeneity of macrophage activation in fish. Dev. Comp. Immunol. 35: 1246-1255. 252. Draijer, C., P. Robbe, C. E. Boorsma, M. N. Hylkema, and B. N. Melgert. 2013. Characterization of macrophage phenotypes in three murine models of house-dust-mite-induced asthma. Mediators Inflamm. 2013: 632049. 253. Waterston, R. H., K. Lindblad-Toh, E. Birney, J. Rogers, J. F. Abril, P. Agarwal, R. Agarwala, R. Ainscough, M. Alexandersson, P. An, S. E. Antonarakis, J. Attwood, R. Baertsch, J. Bailey, K. Barlow, S. Beck, E. Berry, B. Birren, T. Bloom, P. Bork, M. Botcherby, N. Bray, M. R. Brent, D. G.

61 Brown, S. D. Brown, C. Bult, J. Burton, J. Butler, R. D. Campbell, P. Carninci, S. Cawley, F. Chiaromonte, A. T. Chinwalla, D. M. Church, M. Clamp, C. Clee, F. S. Collins, L. L. Cook, R. R. Copley, A. Coulson, O. Couronne, J. Cuff, V. Curwen, T. Cutts, M. Daly, R. David, J. Davies, K. D. Delehaunty, J. Deri, E. T. Dermitzakis, C. Dewey, N. J. Dickens, M. Diekhans, S. Dodge, I. Dubchak, D. M. Dunn, S. R. Eddy, L. Elnitski, R. D. Emes, P. Eswara, E. Eyras, A. Felsenfeld, G. A. Fewell, P. Flicek, K. Foley, W. N. Frankel, L. A. Fulton, R. S. Fulton, T. S. Furey, D. Gage, R. A. Gibbs, G. Glusman, S. Gnerre, N. Goldman, L. Goodstadt, D. Grafham, T. A. Graves, E. D. Green, S. Gregory, R. Guigo, M. Guyer, R. C. Hardison, D. Haussler, Y. Hayashizaki, L. W. Hillier, A. Hinrichs, W. Hlavina, T. Holzer, F. Hsu, A. Hua, T. Hubbard, A. Hunt, I. Jackson, D. B. Jaffe, L. S. Johnson, M. Jones, T. A. Jones, A. Joy, M. Kamal, E. K. Karlsson, D. Karolchik, A. Kasprzyk, J. Kawai, E. Keibler, C. Kells, W. J. Kent, A. Kirby, D. L. Kolbe, I. Korf, R. S. Kucherlapati, E. J. Kulbokas, D. Kulp, T. Landers, J. P. Leger, S. Leonard, I. Letunic, R. Levine, J. Li, M. Li, C. Lloyd, S. Lucas, B. Ma, D. R. Maglott, E. R. Mardis, L. Matthews, E. Mauceli, J. H. Mayer, M. McCarthy, W. R. McCombie, S. McLaren, K. McLay, J. D. McPherson, J. Meldrim, B. Meredith, J. P. Mesirov, W. Miller, T. L. Miner, E. Mongin, K. T. Montgomery, M. Morgan, R. Mott, J. C. Mullikin, D. M. Muzny, W. E. Nash, J. O. Nelson, M. N. Nhan, R. Nicol, Z. Ning, C. Nusbaum, M. J. O'Connor, Y. Okazaki, K. Oliver, E. Overton-Larty, L. Pachter, G. Parra, K. H. Pepin, J. Peterson, P. Pevzner, R. Plumb, C. S. Pohl, A. Poliakov, T. C. Ponce, C. P. Ponting, S. Potter, M. Quail, A. Reymond, B. A. Roe, K. M. Roskin, E. M. Rubin, A. G. Rust, R. Santos, V. Sapojnikov, B. Schultz, J. Schultz, M. S. Schwartz, S. Schwartz, C. Scott, S. Seaman, S. Searle, T. Sharpe, A. Sheridan, R. Shownkeen, S. Sims, J. B. Singer, G. Slater, A. Smit, D. R. Smith, B. Spencer, A. Stabenau, N. Stange- Thomann, C. Sugnet, M. Suyama, G. Tesler, J. Thompson, D. Torrents, E. Trevaskis, J. Tromp, C. Ucla, A. Ureta-Vidal, J. P. Vinson, A. C. Von Niederhausern, C. M. Wade, M. Wall, R. J. Weber, R. B. Weiss, M. C. Wendl, A. P. West, K. Wetterstrand, R. Wheeler, S. Whelan, J. Wierzbowski, D. Willey, S. Williams, R. K. Wilson, E. Winter, K. C. Worley, D. Wyman, S. Yang, S. P. Yang, E. M. Zdobnov, M. C. Zody, and E. S. Lander. 2002. Initial

62 sequencing and comparative analysis of the mouse genome. Nature 420: 520- 562. 254. Ganz, T. 2003. Defensins: antimicrobial peptides of innate immunity. Nat. Rev. Immunol. 3: 710-720. 255. Risso, A. 2000. Leukocyte antimicrobial peptides: multifunctional effector molecules of innate immunity. J. Leukoc. Biol. 68: 785-792. 256. Eisenhauer, P. B., and R. I. Lehrer. 1992. Mouse neutrophils lack defensins. Infect. Immun. 60: 3446-3447. 257. Takeda, K., T. Kaisho, and S. Akira. 2003. Toll-like receptors. Annu. Rev. Immunol. 21: 335-376. 258. Rehli, M. 2002. Of mice and men: species variations of Toll-like receptor expression. Trends Immunol. 23: 375-378. 259. Akira, S. 2001. Toll-like receptors and innate immunity. Adv. Immunol. 78: 1- 56. 260. Alexopoulou, L., A. C. Holt, R. Medzhitov, and R. A. Flavell. 2001. Recognition of double-stranded RNA and activation of NF-kappaB by Toll- like receptor 3. Nature 413: 732-738. 261. Heinz, S., V. Haehnel, M. Karaghiosoff, L. Schwarzfischer, M. Muller, S. W. Krause, and M. Rehli. 2003. Species-specific regulation of Toll-like receptor 3 genes in men and mice. J. Biol. Chem. 278: 21502-21509. 262. Muzio, M., D. Bosisio, N. Polentarutti, G. D'Amico, A. Stoppacciaro, R. Mancinelli, C. van't Veer, G. Penton-Rol, L. P. Ruco, P. Allavena, and A. Mantovani. 2000. Differential expression and regulation of toll-like receptors (TLR) in human leukocytes: selective expression of TLR3 in dendritic cells. J. Immunol. 164: 5998-6004. 263. Visintin, A., A. Mazzoni, J. H. Spitzer, D. H. Wyllie, S. K. Dower, and D. M. Segal. 2001. Regulation of Toll-like receptors in human monocytes and dendritic cells. J. Immunol. 166: 249-255. 264. Bogdan, C., M. Rollinghoff, and A. Diefenbach. 2000. The role of nitric oxide in innate immunity. Immunol. Rev. 173: 17-26. 265. Bogdan, C. 2001. Nitric oxide and the immune response. Nat. Immunol. 2: 907-916.

63 266. Murray, H. W., and R. F. Teitelbaum. 1992. L-arginine-dependent reactive nitrogen intermediates and the antimicrobial effect of activated human mononuclear phagocytes. J. Infect. Dis. 165: 513-517. 267. Schneemann, M., G. Schoedon, S. Hofer, N. Blau, L. Guerrero, and A. Schaffner. 1993. Nitric oxide synthase is not a constituent of the antimicrobial armature of human mononuclear phagocytes. J. Infect. Dis. 167: 1358-1363. 268. Tayeh, M. A., and M. A. Marletta. 1989. Macrophage oxidation of L-arginine to nitric oxide, nitrite, and nitrate. Tetrahydrobiopterin is required as a cofactor. J. Biol. Chem. 264: 19654-19658. 269. Schoedon, G., M. Schneemann, S. Hofer, L. Guerrero, N. Blau, and A. Schaffner. 1993. Regulation of the L-arginine-dependent and tetrahydrobiopterin-dependent biosynthesis of nitric oxide in murine macrophages. Eur. J. Biochem. 213: 833-839. 270. Dykhuizen, R. S., J. Masson, G. McKnight, A. N. Mowat, C. C. Smith, L. M. Smith, and N. Benjamin. 1996. Plasma nitrate concentration in infective gastroenteritis and inflammatory bowel disease. Gut 39: 393-395. 271. Fang, F. C., and A. Vazquez-Torres. 2002. Nitric oxide production by human macrophages: there's NO doubt about it. Am. J. Physiol. Lung Cell. Mol. Physiol. 282: L941-943. 272. Weitzberg, E., and J. O. Lundberg. 1998. Nonenzymatic nitric oxide production in humans. Nitric Oxide 2: 1-7. 273. Ma, S. X., X. Y. Li, T. Sakurai, and M. Pandjaitan. 2007. Evidence of enhanced non-enzymatic generation of nitric oxide on the skin surface of acupuncture points: An innovative approach in humans. Nitric Oxide 17: 60- 68. 274. Nicholson, S., G. Bonecini-Almeida Mda, J. R. Lapa e Silva, C. Nathan, Q. W. Xie, R. Mumford, J. R. Weidner, J. Calaycay, J. Geng, N. Boechat, C. Linhares, W. Rom, and J. L. Ho. 1996. Inducible nitric oxide synthase in pulmonary alveolar macrophages from patients with tuberculosis. J. Exp. Med. 183: 2293-2302. 275. Wang, C. H., C. Y. Liu, H. C. Lin, C. T. Yu, K. F. Chung, and H. P. Kuo. 1998. Increased exhaled nitric oxide in active pulmonary tuberculosis due to inducible NO synthase upregulation in alveolar macrophages. Eur. Respir. J. 11: 809-815.

64 276. St Clair, E. W., W. E. Wilkinson, T. Lang, L. Sanders, M. A. Misukonis, G. S. Gilkeson, D. S. Pisetsky, D. I. Granger, and J. B. Weinberg. 1996. Increased expression of blood mononuclear cell nitric oxide synthase type 2 in rheumatoid arthritis patients. J. Exp. Med. 184: 1173-1178. 277. Anstey, N. M., J. B. Weinberg, M. Y. Hassanali, E. D. Mwaikambo, D. Manyenga, M. A. Misukonis, D. R. Arnelle, D. Hollis, M. I. McDonald, and D. L. Granger. 1996. Nitric oxide in Tanzanian children with malaria: inverse relationship between malaria severity and nitric oxide production/nitric oxide synthase type 2 expression. J. Exp. Med. 184: 557-567. 278. Schneemann, M., and G. Schoeden. 2007. Macrophage biology and immunology: man is not a mouse. J. Leukoc. Biol. 81: 579; discussion 580. 279. Hickman-Davis, J. M., P. O'Reilly, I. C. Davis, J. Peti-Peterdi, G. Davis, K. R. Young, R. B. Devlin, and S. Matalon. 2002. Killing of Klebsiella pneumoniae by human alveolar macrophages. Am. J. Physiol. Lung Cell. Mol. Physiol. 282: L944-956. 280. Han, I. H., S. Y. Goo, S. J. Park, S. J. Hwang, Y. S. Kim, M. S. Yang, M. H. Ahn, and J. S. Ryu. 2009. Proinflammatory cytokine and nitric oxide production by human macrophages stimulated with Trichomonas vaginalis. Korean J. Parasitol. 47: 205-212. 281. Jung, J. Y., R. Madan-Lala, M. Georgieva, J. Rengarajan, C. D. Sohaskey, F. C. Bange, and C. M. Robinson. 2013. The intracellular environment of human macrophages that produce nitric oxide promotes growth of mycobacteria. Infect. Immun. 81: 3198-3209. 282. Stevanin, T. M., J. W. Moir, and R. C. Read. 2005. Nitric oxide detoxification systems enhance survival of Neisseria meningitidis in human macrophages and in nasopharyngeal mucosa. Infect. Immun. 73: 3322-3329. 283. Schneemann, M., and G. Schoedon. 2002. Species differences in macrophage NO production are important. Nat. Immunol. 3: 102. 284. Murray, P. J., and T. A. Wynn. 2011. Obstacles and opportunities for understanding macrophage polarization. J. Leukoc. Biol. 89: 557-563. 285. Wu, G., and S. M. Morris, Jr. 1998. Arginine metabolism: nitric oxide and beyond. Biochem. J. 336 ( Pt 1): 1-17. 286. Di Costanzo, L., G. Sabio, A. Mora, P. C. Rodriguez, A. C. Ochoa, F. Centeno, and D. W. Christianson. 2005. Crystal structure of human arginase I at 1.29-A

65 resolution and exploration of inhibition in the immune response. Proc. Natl. Acad. Sci. U. S. A. 102: 13058-13063. 287. Raes, G., R. Van den Bergh, P. De Baetselier, G. H. Ghassabeh, C. Scotton, M. Locati, A. Mantovani, and S. Sozzani. 2005. Arginase-1 and Ym1 are markers for murine, but not human, alternatively activated myeloid cells. J. Immunol. 174: 6561; author reply 6561-6562. 288. Murray, P. J., and T. A. Wynn. 2011. Protective and pathogenic functions of macrophage subsets. Nat. Rev. Immunol. 11: 723-737. 289. Hamann, J., N. Koning, W. Pouwels, L. H. Ulfman, M. van Eijk, M. Stacey, H. H. Lin, S. Gordon, and M. J. Kwakkenbos. 2007. EMR1, the human homolog of F4/80, is an eosinophil-specific receptor. Eur. J. Immunol. 37: 2797-2802. 290. Khazen, W., P. M'Bika J, C. Tomkiewicz, C. Benelli, C. Chany, A. Achour, and C. Forest. 2005. Expression of macrophage-selective markers in human and rodent adipocytes. FEBS Lett. 579: 5631-5634. 291. Austyn, J. M., and S. Gordon. 1981. F4/80, a monoclonal antibody directed specifically against the mouse macrophage. Eur. J. Immunol. 11: 805-815. 292. McKnight, A. J., A. J. Macfarlane, P. Dri, L. Turley, A. C. Willis, and S. Gordon. 1996. Molecular cloning of F4/80, a murine macrophage-restricted cell surface glycoprotein with homology to the G-protein-linked transmembrane 7 hormone receptor family. J. Biol. Chem. 271: 486-489. 293. Janicke, R. U., D. Sohn, G. Totzke, and K. Schulze-Osthoff. 2006. Caspase-10 in mouse or not? Science 312: 1874. 294. Nathan, C. 2006. Role of iNOS in human host defense. Science 312: 1874- 1875; author reply 1874-1875. 295. Burgess, A. W., and D. Metcalf. 1980. The nature and action of granulocyte- macrophage colony stimulating factors. Blood 56: 947-958. 296. Johnson, G. R., and A. W. Burgess. 1978. Molecular and biological properties of a macrophage colony-stimulating factor from mouse yolk sacs. J. Cell Biol. 77: 35-47. 297. Sherr, C. J., C. W. Rettenmier, R. Sacca, M. F. Roussel, A. T. Look, and E. R. Stanley. 1985. The c-fms proto-oncogene product is related to the receptor for the mononuclear phagocyte growth factor, CSF-1. Cell 41: 665-676. 298. Mancini, A., A. Koch, R. Wilms, and T. Tamura. 2002. c-Cbl associates directly with the C-terminal tail of the receptor for the macrophage colony-

66 stimulating factor, c-Fms, and down-modulates this receptor but not the viral oncogene v-Fms. J. Biol. Chem. 277: 14635-14640. 299. Rosas, M., S. Gordon, and P. R. Taylor. 2007. Characterisation of the expression and function of the GM-CSF receptor alpha-chain in mice. Eur. J. Immunol. 37: 2518-2528. 300. Jubinsky, P. T., A. S. Laurie, D. G. Nathan, J. Yetz-Aldepe, and C. A. Sieff. 1994. Expression and function of the human granulocyte-macrophage colony- stimulating factor receptor alpha subunit. Blood 84: 4174-4185. 301. Hamilton, J. A., E. R. Stanley, A. W. Burgess, and R. K. Shadduck. 1980. Stimulation of macrophage plasminogen activator activity by colony- stimulating factors. J. Cell. Physiol. 103: 435-445. 302. Hamilton, J. A. 2008. Colony-stimulating factors in inflammation and autoimmunity. Nat. Rev. Immunol. 8: 533-544. 303. Whetton, A. D., and T. M. Dexter. 1989. Myeloid haemopoietic growth factors. Biochim. Biophys. Acta 989: 111-132. 304. Metcalf, D. 1991. Control of granulocytes and macrophages: molecular, cellular, and clinical aspects. Science 254: 529-533. 305. Akagawa, K. S. 2002. Functional heterogeneity of colony-stimulating factor- induced human monocyte-derived macrophages. Int. J. Hematol. 76: 27-34. 306. Bailey, M. J., D. C. Lacey, B. V. de Kok, P. D. Veith, E. C. Reynolds, and J. A. Hamilton. 2011. Extracellular proteomes of M-CSF (CSF-1) and GM-CSF- dependent macrophages. Immunol. Cell Biol. 89: 283-293. 307. Hamilton, J. A., and G. P. Anderson. 2004. GM-CSF Biology. Growth Factors 22: 225-231. 308. Burgess, A. W., D. Metcalf, L. G. Sparrow, R. J. Simpson, and E. C. Nice. 1986. Granulocyte/macrophage colony-stimulating factor from mouse lung conditioned medium. Purification of multiple forms and radioiodination. Biochem. J. 235: 805-814. 309. Akagawa, K. S., I. Komuro, H. Kanazawa, T. Yamazaki, K. Mochida, and F. Kishi. 2006. Functional heterogeneity of colony-stimulating factor-induced human monocyte-derived macrophages. Respirology 11 Suppl: S32-36. 310. 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.

67 311. Inaba, K., M. Inaba, N. Romani, H. Aya, M. Deguchi, S. Ikehara, S. Muramatsu, and R. M. Steinman. 1992. Generation of large numbers of dendritic cells from mouse bone marrow cultures supplemented with granulocyte/macrophage colony-stimulating factor. J. Exp. Med. 176: 1693- 1702. 312. Jackson, S. H., C. Alicea, J. W. Owens, C. L. Eigsti, and H. L. Malech. 2002. Characterization of an early dendritic cell precursor derived from murine lineage-negative hematopoietic progenitor cells. Exp. Hematol. 30: 430-439. 313. Naito, M., S. Umeda, K. Takahashi, and L. D. Shultz. 1997. Macrophage differentiation and granulomatous inflammation in osteopetrotic mice (op/op) defective in the production of CSF-1. Mol. Reprod. Dev. 46: 85-91. 314. Xu, W., N. Schlagwein, A. Roos, T. K. van den Berg, M. R. Daha, and C. van Kooten. 2007. Human peritoneal macrophages show functional characteristics of M-CSF-driven anti-inflammatory type 2 macrophages. Eur. J. Immunol. 37: 1594-1599. 315. Stanley, E. R., K. L. Berg, D. B. Einstein, P. S. Lee, F. J. Pixley, Y. Wang, and Y. G. Yeung. 1997. Biology and action of colony--stimulating factor-1. Mol. Reprod. Dev. 46: 4-10. 316. Bender, A. T., C. L. Ostenson, D. Giordano, and J. A. Beavo. 2004. Differentiation of human monocytes in vitro with granulocyte-macrophage colony-stimulating factor and macrophage colony-stimulating factor produces distinct changes in cGMP phosphodiesterase expression. Cell. Signal. 16: 365- 374. 317. Lacey, D. C., A. Achuthan, A. J. Fleetwood, H. Dinh, J. Roiniotis, G. M. Scholz, M. W. Chang, S. K. Beckman, A. D. Cook, and J. A. Hamilton. 2012. Defining GM-CSF- and macrophage-CSF-dependent macrophage responses by in vitro models. J. Immunol. 188: 5752-5765. 318. Fleetwood, A. J., H. Dinh, A. D. Cook, P. J. Hertzog, and J. A. Hamilton. 2009. GM-CSF- and M-CSF-dependent macrophage phenotypes display differential dependence on type I interferon signaling. J. Leukoc. Biol. 86: 411-421. 319. Bhatt, N. Y., T. W. Kelley, V. V. Khramtsov, Y. Wang, G. K. Lam, T. L. Clanton, and C. B. Marsh. 2002. Macrophage-colony-stimulating factor- induced activation of extracellular-regulated kinase involves

68 phosphatidylinositol 3-kinase and reactive oxygen species in human monocytes. J. Immunol. 169: 6427-6434. 320. Mantovani, A., A. Sica, S. Sozzani, P. Allavena, A. Vecchi, and M. Locati. 2004. The chemokine system in diverse forms of macrophage activation and polarization. Trends Immunol. 25: 677-686. 321. Brocheriou, I., S. Maouche, H. Durand, V. Braunersreuther, G. Le Naour, A. Gratchev, F. Koskas, F. Mach, J. Kzhyshkowska, and E. Ninio. 2011. Antagonistic regulation of macrophage phenotype by M-CSF and GM-CSF: implication in atherosclerosis. Atherosclerosis 214: 316-324. 322. Jaguin, M., N. Houlbert, O. Fardel, and V. Lecureur. 2013. Polarization profiles of human M-CSF-generated macrophages and comparison of M1- markers in classically activated macrophages from GM-CSF and M-CSF origin. Cell. Immunol. 281: 51-61. 323. Stevens, R. C., V. Cherezov, V. Katritch, R. Abagyan, P. Kuhn, H. Rosen, and K. Wuthrich. 2013. The GPCR Network: a large-scale collaboration to determine human GPCR structure and function. Nat. Rev. Drug Discov. 12: 25-34. 324. Davenport, A. P., S. P. Alexander, J. L. Sharman, A. J. Pawson, H. E. Benson, A. E. Monaghan, W. C. Liew, C. P. Mpamhanga, T. I. Bonner, R. R. Neubig, J. P. Pin, M. Spedding, and A. J. Harmar. 2013. International Union of Basic and Clinical Pharmacology. LXXXVIII. G protein-coupled receptor list: recommendations for new pairings with cognate ligands. Pharmacol. Rev. 65: 967-986. 325. Zhang, Y., M. E. Devries, and J. Skolnick. 2006. Structure modeling of all identified G protein-coupled receptors in the human genome. PLoS Comput. Biol. 2: e13. 326. Soruri, A., S. Kim, Z. Kiafard, and J. Zwirner. 2003. Characterization of C5aR expression on murine myeloid and lymphoid cells by the use of a novel monoclonal antibody. Immunol. Lett. 88: 47-52. 327. Rao, K. M. 2001. MAP kinase activation in macrophages. J. Leukoc. Biol. 69: 3-10.

69 Chapter

2 Profiling human monocyte-derived GM-CSF- and M-CSF differentiated macrophages: In vitro models for ‘M1-like’ and ‘M2-like’ macrophages

2.1 Abstract ...... 71 2.2 Introduction ...... 72 2.3 Aims ...... 73 2.4 Materials and Methods ...... 74 2.4.1 Reagents ...... 74 2.4.2 Isolation of human monocytes and macrophages ...... 74 2.4.3 RNA extraction and reverse transcription ...... 74 2.4.4 Real-time quantitative polymerase chain reaction (real-time qPCR) ...... 75 2.4.5 Hybridization and analysis of human array-based gene expression ...... 76 2.4.6 Enzyme-linked immunosorbent assay (ELISA) ...... 76 2.4.7 In vitro migration assay (modified Boyden chamber) ...... 76 2.4.8 Actin visualization by fluorescence microscopy ...... 77 2.4.9 Statistical analysis ...... 77 2.5 Results ...... 78 2.5.1 GM-CSF or M-CSF is necessary for macrophage differentiation and survival . 78 2.5.2 Temporal study of monocyte-derived macrophage response to LPS stimulation 80 2.5.3 Transcriptional difference between GM-MΦ and M-MΦ by cDNA microarray analysis ...... 83 2.5.4 Pathway analysis of differentially expressed genes between GM-MΦ and M-MΦ 86 2.5.5 Cellular movement and actin cytoskeleton in GM-MΦ and M-MΦ ...... 87 2.5.6 GM-MΦ and M-MΦ have similar GPCR signalling profile except for Gα12/13 signalling ...... 89 2.5.7 Inflammatory responses in GM-MΦ and M-MΦ ...... 90 2.5.8 Monocyte-derived macrophages cytokine and gene responses to LPS stimulation 92 2.6 Discussion ...... 94 2.7 Conclusions ...... 95 2.8 References ...... 96

70 2.1 Abstract

Macrophages have remarkably plasticity in their gene expression profiles and cell surface phenotypes, depending on which tissue the macrophages reside in and stimulant exposure. These variations give rise to different populations of cells with distinct functions. Granulocyte-macrophage colony-stimulating factor (GM-CSF) or macrophage colony-stimulating factor (M-CSF) is used in cell culture to differentiate primary human monocytes to macrophages (Human Monocyte-Derived Macrophages) into two different populations (GM-MΦ and M-MΦ). Although they do not have distinct responses when activated by LPS, they represent a continuum model with GM-MΦ being more pro-inflammatory than M-MΦ (higher levels of proinflammatory cytokines such as IL6, TNF, MCP1), and M-MΦ being more anti- inflammatory than GM-MΦ (higher levels of anti-inflammatory cytokine IL10). GM- MΦ and M-MΦ exhibit significant differences in gene expression with more than 1000 genes differentially expressed under basal conditions. The IPA software has identified large differences between GM-MΦ and M-MΦ in bio-functions such as cellular movement, haematological development and inflammatory response. Genes such as LEP, CCL2, IL10, CSF1, OLR1, IL1RN, GPR120, MMP10, PPBP and CXCL5 were selected from the list of gene hits and were validated by real-time PCR. The largest functional differences between GM-MΦ and M-MΦ have been further investigated and validated as well. GM-MΦ is not only more proinflammatory in nature, but also more robust to LPS stimulation and can migrate to site of inflammation four times more rapidly than M-MΦ. This chapter has characterized and validated GM-MΦ and M-MΦ as suitable in vitro models to study proinflammatory and anti-inflammatory responses specific for human monocyte-derived macrophages.

71 2.2 Introduction

Monocytes and macrophages are important immune effector cells interplaying between innate and adaptive immunity. Monocytes are released from the bone marrow into the blood circulation as macrophage precursor cells. Monocytes can differentiate into macrophages when they migrate from the circulation into the tissue in the steady state or in response to inflammation. Macrophages can be found residing throughout the body tissues and, depending on which tissue the macrophages reside in and the stimulants they become exposed to, they can have distinct appearances and cell surface phenotypes (1). The monocyte-macrophage axis plays essential roles in immune functions such as inflammation, antimicrobial immune defence and tissue wound healing but, like a double-edged sword, these cells are also the main culprits in tissue destruction during certain infections and inflammatory diseases (2). Granulocyte-macrophage colony-stimulating factor (GM-CSF) and macrophage colony-stimulating factor (M-CSF) were originally defined as hemopoietic growth factors (3). However, these cytokines are also major survival/mitogenic factors for the macrophage lineage, with the capacity to activate and induce monocytic differentiation to macrophage (1). GM-CSF and M-CSF are quite distinct with different tissue distributions in the steady state, different structures and different receptors, with the M-CSF receptor (CSF1R) being the tyrosine kinase c-Fms and the GM-CSF receptor (CSF2R) comprising α (ligand binding domain) and β (signalling domain) subunits (4, 5). M-CSF is passively produced in vitro by several cell types, especially macrophages, and circulates at detectable levels basally, whereas GM-CSF is usually undetectable but becomes elevated during immune/inflammatory responses in vivo to infection or in vitro in response to LPS or TNF (6, 7). GM-CSF was first isolated from lungs (8) and monocytes differentiated with GM-CSF closely resemble alveolar macrophages (9). M-CSF, also known as CSF1, was the first of the hemopoietic growth factors to be isolated and was found to be important for the development, proliferation and differentiation of osteoclasts and peritoneal macrophages (10, 11) and tissue remodelling/homeostasis (12). Therefore, monocytes recruited to sites would differentiate into divergent phenotypes of macrophages depending on the CSF present. M1 and M2 macrophage phenotypes were described by analogy to Th1 and Th2 lymphocyte subsets; macrophages

72 differentiated in culture in the presence of GM-CSF or M-CSF were named proinflammatory M1 and anti-inflammatory M2 phenotypes, respectively (13, 14). The functional heterogeneity of CSF-induced human monocyte-derived macrophages has been reviewed elsewhere (15). GM-CSF and M-CSF can influence macrophage phenotypic dependence on type I interferon signalling, at least in murine bone marrow cells (16). However, only around 17% of the genes in a comparison of human monocytes differentiated by human GM-CSF versus M-CSF had a profile that was conserved in murine bone marrow cells differentiated by GM-CSF versus M-CSF (14). With such a high divergence across species, caution should therefore be exercised when interpreting the effects of a CSF on human macrophage lineage cells from studies conducted on analogous murine cells.

2.3 Aims

This chapter will describe and validate monocyte-derived macrophages differentiated with either GM-CSF (GM-MΦ) or M-CSF (M-MΦ), as models for proinflammatory ‘M1-like’ and antiinflammatory ‘M2-like’ primary human macrophages, to gain insights into the contributions of these macrophages to inflammation. Therefore the specific aims of this chapter were: 1. To identify cell-type specific differences in gene expression profiles of GM- MΦ and M-MΦ by cDNA microarray analysis. 2. To validate the differentially expression of genes by quantitative realtime-PCR analysis. 3. To identify conserved and divergent downstream biological functions for GM- MΦ versus M-MΦ. 4. To examine the contributions and properties of these macrophages basally or in responses to inflammatory stimuli, such as LPS, in a selection of appropriate functional studies.

73 2.4 Materials and Methods

2.4.1 Reagents

Recombinant human GM-CSF and M-CSF were purchased from PeproTech. LPS from Salmonella enterica (serotype Minnesota RE 595) was purchased from Sigma-Aldrich. All cell culture reagents were purchased from Invitrogen and all analytical grade chemical reagents were obtained from Sigma-Aldrich, unless otherwise stated.

2.4.2 Isolation of human monocytes and macrophages

PBMCs were isolated from buffy coats of anonymous donors (Australian Red Cross Blood Service, Kelvin Grove, Queensland, Australia) by density centrifugation using Ficoll-Paque Plus (GE Healthcare) manufacturer’s instructions. Contaminating erythrocytes were removed by repeated washing with ice-cold sterile water. CD14+ MACS® microbeads were used to positively select monocytes from isolated PBMCs according to manufacturer’s instructions (Miltenyi Biotec). CD14+ monocytes were seeded at 1 x 106 cells/mL in IMDM supplemented with 10% FBS, 100 U/mL penicillin, 100 µg/mL streptomycin and 2 mM GlutaMAX at 37 °C in presence of 5% + CO2. To generate human monocyte-derived macophages (HMDM), CD14 monocytes were cultured in the presence of either 10 ng/mL GM-CSF (GM-MΦ) or M-CSF (M-MΦ) for at least 6 days.

2.4.3 RNA extraction and reverse transcription

74 incubated at 70°C for 10 min and then cooled on ice for at least 1 min before being reverse transcripted with Superscript III (Invitrogen) at 50°C for 50 min, then 70°C for 10 min. cDNA samples were stored at -20°C until further use.

2.4.4 Real-time quantitative polymerase chain reaction (real-time qPCR)

Primers were designed using the free web-based software Primer-BLAST National Center for Biotechnology Information (NCBI). cDNA (50 ng) was prepared together with SYBR® Green PCR master mix (Applied Biosystems™) and relevant primers to perform real time-PCR on the ABI PRISM® 7500 Real Time PCR System (Applied Biosystems™). Conditions were used according to the manufacturer’s instructions. Relative gene expression was normalised against 18S rRNA expression and then converted to fold change against control samples. All samples were analysed in duplicates. Primer sequences used are: CD68, 5’-TTTGGGTGAGGCGGTTCAG-3’ and 3’-CCAGTGCTCTCTGCCAGTA-5’; EMR1, 5’- CCACCTGCACCTCTGCGTGTG-3’ and 3’-GGCGCAGCCCATCTTGTTGTC-5’; LEP, 5’-ATGTCCAAGCTGTGCCCATCCAA-3’ and 3’- TTGGAGGAGACTGACTGCGTGTGT-5’; CCL2, 5’- AGCTCGCACTCTCGCCTCCAG-3’ and 3’- GGCATTGATTGCATCTGGCTGAGC-5’; IL10, 5’- TGGAGGAGGTGATGCCCCAAGC-3’ and 3’- AAATCGATGACAGCGCCGTAGC-5’; CSF1, 5’- CTCCAGCCAAGGCCATGAGA-3’ and 3’-CAGCAAGACCAGGATGACACTG- 5’; OLR1, 5’-CCCTTGCTCGGAAGCTGAAT-3’ and 3’- GCTTGCTGGATGAAGTCCTGAA-5’; IL1RN, 5’- GGTACTGCCCGGGTGCTACTTT-3’ and 3’- GGTCGGCAGATCGTCTCTAAAGC-5’; GPR120, 5’- GAGATCTCGTGGGATGTCTCT-3’ and 3’-CCTTGATGCCTTTGTGATCTGT-5’; MMP10, 5’-CCAGGACACAGTTTGGCTCATGCC-3’ and 3’- AATTGGTGCCTGATGCATCTTCTGT-5’; PPBP, 5’- TGGCGAAAGGCAAAGAGGAA-3’ and 3’-TGGGATGAATTCCAGAGGTTGT- 5’; CXCL5, 5’-AGACCACGCAAGGAGTTCAT-3’ and 3’-

75 CTTCAGGGAGGCTACCACTTC-5’; IL6, 5’-GCCCACCGGGAACGAAAGAGA- 3’ and 3’-GACCGAAGGCGCTTGTGGAGAAG-5’; IL8, 5’- ACCACCGGAAGGAACCATCTCACT-3’ and 3’- CTTGGCAAAACTGCACCTTCACAC-5’; TNF, 5’- CCCAGGGACCTCTCTCTAATC-3’ and 3’-ATGGGCTACAGGCTTGTCACT-5’.

2.4.5 Hybridization and analysis of human array-based gene expression

Quality controls were performed on the RNA to check the integrity, purity and concentration of samples. Synthesis of biotin-labelled cRNA, hybridisation to 50-mer probes on the HumanHT-12 v4.0 Expression BeadChip (Illumina®) and detection of hybridised target cRNA were performed by Australian Genome Research Facility (AGRF) according to the Illumina® whole-genome gene expression beadchips manufacturer’s instruction. GeneSpring GX software version 11.5 (Agilent Technologies) and web-based IPA software (Qiagen) was used for the analysis of differences in gene expression between GM-MΦ and M-MΦ. Statistically significant differences between samples with more than 2-fold change were assessed using student t-test.

2.4.6 Enzyme-linked immunosorbent assay (ELISA)

Cells were seeded at 1 x 106 cells/mL and serum-deprived overnight in the incubator prior to treatment. Cell culture supernatants were collected and cytokines level were determined using specific ELISA sets from BD Pharmingen, according to the manufacturer’s instructions.

2.4.7 In vitro migration assay (modified Boyden chamber)

76 initial cell migration, serum free medium was placed in the lower chamber. At the end of the time course study, cells were removed from the upper side of the membranes, and nuclei of migratory cells on the lower side of the membrane were stained with DAPI. And counted under a fluorescence microscope using a x10 objective.

2.4.8 Actin visualization by fluorescence microscopy

HMDMs were seeded (5 x 105 cells/mL) overnight on glass coverslips. Cells were serum deprived for at least 6 h prior to treatment as described in figure legend. After treatment, cells were fixed with 4% formaldehyde and permeabilized with 0.5% Triton-X 100 for 15 min. Cells were stained with FITC-phalloidin (400 nM) (Sigma) and nuclei were stained with Prolong® Gold DAPI (Invitrogen). Images were acquired using a x40 objective.

2.4.9 Statistical analysis

Data were plotted and analysed using GraphPad Prism version 5.0c for Mac OS X (GraphPad Software). Statistically significant differences were assessed using student’s t-test for paired comparison. All values of independent parameters are shown as mean ± SEM of at least three independent experiments, unless otherwise stated. Significance was set at *p < 0.05, **p < 0.01 and ***p < 0.001.

77 2.5 Results

2.5.1 GM-CSF or M-CSF is necessary for macrophage differentiation and survival

Freshly isolated CD14+ human monocytes were obtained from buffy coats of healthy blood donors provided by Australian Red Cross Blood Service. Blood monocytes were seeded at 1 x 106 cells/mL in 10% FBS/IMDM medium alone or in the presence of 10 ng/mL GM-CSF or M-CSF in vitro. In the absence of either GM- CSF or M-CSF, more than 90% of the cell population died after 7 days in culture (Figure 2.1). CSFs were necessary for monocytes survival and differentiation into macrophages. By appearance, monocytes appeared to be less adherent to the plastic surface of culture plates, followed by GM-MΦ (monocytes differentiated with GM- CSF) and M-MΦ (monocytes differentiated with M-CSF) being most adherent. The morphology of the two populations of macrophages was distinct after 6 days; GM- MΦ were round shaped, whereas M-MΦ were more elongated in shape. Macrophages were harvested at the end of the 6 days differentiation by scraping. The expected cell yields by this differentiating protocol varied from 50 to 80%. This difference in cell yields was limited to donor-to-donor variation rather than differences between CSF used in differentiation. High variations were often observed across donors, rather than within the same donor. The mRNA levels of specific macrophage markers, such as CD68 and EMR1, were used to monitor the differentiation process. As expected, CD68 was increased after 4 days of differentiation (Figure 2.2A), and EMR1 after 6 days (Figure 2.2B), in the presence of either GM-CSF or M-CSF. The mRNA expression levels of CD68 and EMR1 between GM-MΦ and M-MΦ were not significantly different at the end of the 7 days differentiation protocol with either GM-CSF or M-CSF (Figure 2.2C, D).

Figure 2.1. GM-CSF and M-CSF are required for monocytes survival and differentiation. Human CD14+ monocytes were seeded at 1 x 106 cells/mL 10% FBS/IMDM in the absence or presence of either differentiating agents (10 ng/mL GM-CSF or M-CSF) for 6 days. The morphologies of cells were monitored by light microscope using a x20 objective on day 0, 2, 4 and 6.

Figure 2.2. Macrophage markers CD68 and EMR1 expression levels in GM-MΦ and M-MΦ. Human CD14+ monocytes were differentiated to HMDMs with either GM-CSF (10 ng/mL) (GM-MΦ, black) or M-CSF (10 ng/mL) (M-MΦ, white) in 10% FBS/IMDM over a 6 day time course. Macrophage marker (A,C) CD68 and (B,D) EMR1 gene expression were monitored during differentiation, in the presence of GM-CSF (black) or in the presence of M-CSF (white) on Days 0, 2, 4 and 6. Genes were detected by quantitative realtime-PCR, normalised against 18S rRNA expression and converted to fold change relative to Day 0. Error bars are means ± SEM of three independent experiments (n=3). *p < 0.05, **p < 0.01 and ***p < 0.001 by student t-test.

2.5.2 Temporal study of monocyte-derived macrophage response to LPS stimulation

To characterise the phenotypic differences between the two subtypes of (GM- MΦ and M-MΦ), we examined their responses to LPS (5 ng/mL) for 24 h in 10% FBS/IMDM medium on day 0, 2, 4 and 6 during the differentiation process, in the presence of either GM-CSF (10 ng/mL) or M-CSF (10 ng/mL).

80 During the course of differentiation, proinflammatory cytokines such as IL6, TNF and MCP1/CCL2 were generally highly produced in GM-MΦ compared to M- MΦ (Figure 2.3A-C). This trend was observed even at day 1, with IL6 and MCP1 being produced in monocytes after exposure to GM-CSF (10 g/mL) for a day compared to monocytes without CSF exposure. By day 2, monocytes in the presence of M-CSF adopted the anti-inflammatory profile of M2-like macrophages by secreting less proinflammatory cytokines compared to monocytes in the presence of GM-CSF. Interestingly, both subtypes of macrophages passively secreted MCP1, but MCP1 is inducible by LPS in GM-MΦ but not in M-MΦ (Figure 2.3C). IL6, TNF and IL10 on the other hand were not passively secreted but LPS-induced only in GM- MΦ and M-MΦ (Figure 2.3A, B, D). Although GM-MΦ and M-MΦ did not have distinct responses when activated by LPS, they represented a continuum model where GM-MΦ being more proinflammatory than M-MΦ and M-MΦ being more anti-inflammatory than GM-MΦ. The data supported the hypothesis that macrophages can be highly heterogeneous in terms of their polarization, and can respond spontaneously according to the microenvironment and external stimuli, and hence these macrophages have high “plasticity”. The responsiveness of GM-MΦ and M-MΦ to LPS was compared and validated in these two subtypes of macrophages. In summary, the data suggested that human monocytes differentiated with GM-CSF leads to the formation of proinflammatory macrophages (M1-likeness), and M-CSF leads to the formation of anti-inflammatory macrophages (M2-likeness).

Figure 2.3. Temporal study of IL6, TNF, MCP1 and IL10 cytokine production in LPS activated GM-MΦ and M-MΦ. Human CD14+ monocytes (grey) were differentiated to HMDMs with either GM- CSF (10 ng/mL) (GM-MΦ, black) or M-CSF (10 ng/mL) (M-MΦ, white) in 10% FBS/IMDM over a 6 day time course. These cells were also treated with LPS (5 ng/mL) for 24 h in 10% FBS/IMDM on Day 0, 2, 4 and 6 to monitor LPS-induced (A) IL6, (B) TNF, (C) MCP1 and (D) IL10 in HMDMs. ELISA was used to detect cytokines secreted in culture media. Error bars are means ± SEM of three independent experiments (n=3). *p < 0.05, **p < 0.01 and ***p < 0.001 by student t-test.

82 2.5.3 Transcriptional difference between GM-MΦ and M-MΦ by cDNA microarray analysis

The previous section described and characterized GM-MΦ and M-MΦ after activated by LPS. In this section, differentially expressed genes between GM-MΦ and M-MΦ at their resting state were analysed globally by cDNA microarray experiments. Four batches of pair-wise donor-matched cells were submitted to AGRF for reading on an Illumina microarray platform. Labelled cRNA were detected by hybridising to 50-mer probes on the HumanHT-12 v4 BeadChip (Illumina®) comprising of 47,231 probes derived from NCBI RefSeq Release 38 (November 7, 2009); of which 28,688 probes corresponded to well-established coding transcripts, 11,121 probes corresponded to provisional coding transcripts, 1,752 probes corresponded to well- established non-coding transcripts and 2,209 probes corresponded to provisional non- coding transcripts. The raw data obtained was quantified, normalised and filtered using the GeneSpring software v11.5. The statistical significant threshold limit was set at p < 0.05 and employed the Benjamini-Hochberg method to control the false discovery rate (17). The cut-off value for fold change was set at two-fold. In order for an outcome to be considered as positive, the hit must pass these two conditions. Using these criteria, the comparison of GM-MΦ and M-MΦ identified more than 1000 hits, of which 568 were upregulated in GM-MΦ against M-MΦ (Appendix I), and 592 were downregulated in GM-MΦ against M-MΦ (Appendix II). Due to the extension list of genes differentially expressed between GM-MΦ and M-MΦ, for simplicity, genes that were ≥ 5 fold altered were shortlisted and represented in the heat map; with red representing genes upregulated in GM-MΦ and green representing genes downregulated in M-MΦ (Figure 2.4).

84 Figure 2.4. Heat map representation of genes differentially expressed in GM-MΦ and M-MΦ that were ≥ 5 fold changed. CD14+ monocytes were seeded at 1x106 cels/mL and differentiated to HMDMs in the presence of 10 ng/mL GM-CSF (GM-MΦ) or 10 ng/mL M-CSF (M-MΦ) for 6 days. cDNA microarray was performed on mRNA extracted from four individual donors at the end of 6 days differentiation, using the Illumina HumanHT-12 v4 BeadChip array. Colour bar shows log2 fold change in gene expression. Heat map representation of normalized signal intensity values for genes that were at least altered by ≥ 5-fold. Red colour represents upregulated genes and green colour represents downregulated genes in GM-MΦ versus M-MΦ. Chosen genes indicated by black arrows are validated by real-time PCR. Full list of genes altered ≥ 2-fold change, refer to (Appendix I and II).

A subset of genes was selected from the heat map and validated by real-time PCR (Table 2.1). All ten genes selected for the PCR experiments correlated to the microarray experiment in terms of the direction of fold changes. However, the magnitude of fold change was mostly higher in PCR than microarray with the exception of LEP, MMP10 and CXCL5. It is apparent that the greater the fold changes, the higher the variation that is observed. This is not surprising as resolution is higher and more sensitive in real-time PCR than in microarray, but most importantly, both experiments adopted different methods to normalize the raw data. Normalization algorithms were used in microarray whereas expression values in PCR were normalized against housekeeping genes, in this case 18S rRNA expression.

Table 2.1. Correlation between microarray and real-time PCR gene expression

Fold change (GM-MΦ vs M-MΦ) Gene Microarray Real-time PCR LEP -34.1 -14.9 CCL2 -5.8 -6.2 IL10 -5.5 -7.8 CSF1 5.5 5.2 OLR1 6.5 6.7 IL1RN 7.7 20.6 GPR120 8.3 10.5 MMP10 49.3 19.7 PPBP 60.2 149.0 CXCL5 69.1 34.3

85 2.5.4 Pathway analysis of differentially expressed genes between GM-MΦ and M-MΦ

The list of differentially expressed genes between GM-MΦ and M-MΦ exposed to LPS were uploaded to online software, Ingenuity Pathway Analysis (IPA). The IPA software was used to analyse the microarray data and detect expression patterns of genes linked to specific biological functions and signalling pathways. The top 4 populated bio-functional categories that were highly modulated by the differentially expressed genes between GM-MΦ and M-MΦ were (i) cellular movement, (ii) haematological system development and functions, (iii) immune cell trafficking and (iv) inflammatory responses (Figure 2.5).

Figure 2.5. Heatmap representative of an overview for biological functions and processes differentially expressed between GM-MΦ and M-MΦ. Visualization of biological functions is represented by the heatmap; z-score ≥ 2 is predicted to increase (orange) and ≤ 2 is predicted to decrease (blue).

The primary aim for this chapter was to compare different populations of human macrophages produced from primary human monocytes using different differentiating agents, GM-CSF versus M-CSF. It was expected that GM-CSF differentiated macrophages would adopt a more proinflammatory profile of gene expression, while M-CSF-differentiated macrophages would be more anti- inflammatory. Interestingly, the majority of the differentially expressed genes did fall into the haematological system development and functions category. In Figure 2.3, GM-MΦ was found to be more robust in LPS signalling than M-MΦ in producing proinflammatory cytokines IL6, TNF and MCP1/CCL2, while M-MΦ produced more

86 anti-inflammatory IL10. Since the inflammatory response category was ranked among the top 4 functional categories affected in these macrophages, it is evident that GM- MΦ and M-MΦ would adopt different immune profiles when activated. The bio- functional categories, cellular movement and inflammatory responses, will be investigated further in the next two sections.

2.5.5 Cellular movement and actin cytoskeleton in GM-MΦ and M-MΦ

Macrophages can migrate actively through the walls of blood vessels into surrounding tissues under inflammatory or non-inflammatory conditions. The basal cellular migration ability of GM-MΦ and M-MΦ were tested in a Boyden chamber in serum free IMDM medium. There was no difference in the number of migrated cells after 3 h. However, M-MΦ took 4 times longer (GM-MΦ, 6 h; M-MΦ, 24 h) to migrate 200 cells to the lower side of the membrane than GM-MΦ (Figure 2.6).

Figure 2.6. GM-MΦ migrates faster than M-MΦ in the Boyden chamber. GM-MΦ and M-MΦ (1 x 105 cells/100 µL) were added to the upper chamber of the Transwell in serum free medium and allowed to migrate through the 5µm porous membrane into the lower chamber. Migrated cells on the lower side of the membrane were stained with DAPI and counted under a fluorescence microscope using a x10 objective. Error bars are means ± SEM of three independent experiments (n=3). ***p < 0.001 by student t-test.

87 These observations support the hypothesis that M1-like or proinflammatory macrophages are more robust in cellular movement and can readily infiltrate inflamed tissues when signal is triggered, whereas M2-like or anti-inflammatory macrophages are less motile and reside within tissues to maintain tissue resolution and homeostasis (18). Cell migration is a complex process that involves the formation of new focal contacts and disruption of existing focal contacts (19). Actin polymerization is necessary in cellular movement and to form lamellipodia (20). Lamellipodia are cytoskeletal protein actin projections on the surface of the cell. GM-MΦ and M-MΦ were stained with phalloidin to detect for polymerized actin. Untreated GM-MΦ has more polymerized actin in the cytosol compared to untreated M-MΦ (Figure 2.7). The formation of lamellipodia was also evident in untreated GM-MΦ but absent in M-MΦ. Rho belongs to a superfamily of small GTPases that can regulate the polymerization of actin to produce lamellipodia. Indeed, the lamellipodia formation seen in GM-MΦ is Rho-dependent because GM-MΦ treated with selective Rho- associated protein kinase inhibitor Y-27632 blocked lamellipodia formation. Rho activator Calpeptin-treated GM-MΦ and M-MΦ led to lamellipodia formation. These results suggest that the presence of lamellipodia on macrophages is mediated by Rho GTPase activation. The presence of lamellipodia, probably due to higher Rho activity, in GM-MΦ but not M-MΦ may explain why GM-MΦ are able to migrate faster than M-MΦ.

Figure 2.7. GM-MΦ has more basal actin polymerization than M-MΦ. GM-MΦ and M-MΦ were pre-treated with either Rho inhibitor Y-27632 (30 µM) or Rho activator Calpeptin (1 unit/mL) for 20 min. After fixation, treated and untreated human macrophages (GM-MΦ and M-MΦ) were stained with 400 nM FITC- phalloidin (green) and nuclei were visualized with DAPI (blue) using a x40 objective.

2.5.6 GM-MΦ and M-MΦ have similar GPCR signalling profile except for

Gα12/13 signalling

The Gα12/13 family is best known for their involvement in the processes of cell proliferation and morphology, such as actin stress fibre and focal adhesion formation. Interactions with Rho guanine nucleotide exchange factors (RhoGEFs) are thought to mediate many of these processes (21). Interestingly, there was no significance difference in upstream GPCR-mediated signalling between GM-MΦ and M-MΦ.

However, Gα12/13 signalling appears to be significantly different between GM-MΦ and M-MΦ (Figure 2.8).

Figure 2.8. Expression analysis of G-protein coupling receptor signalling pathway-linked genes between GM-MΦ and M-MΦ. The x-axis displays selected canonical pathways and the y-axis displays the significance by –log(p-value). P values are calculated using the right-tailed Fisher’s Exact test. The threshold lines denotes the cutoff value for significance, which represent p = 0.05

Activation of Rho, or the regulation of events mediated by Rho, is often taken as evidence of Gα12/13 signalling. The differences in Rho activation between GM-MΦ and M-MΦ, as observed by lamellipodia shown in Figure 2.7, is also reflected in the microarray results when compared for genes related to Gα12/13 signalling pathway in Figure 2.8.

2.5.7 Inflammatory responses in GM-MΦ and M-MΦ

Pain, heat, redness, swelling and loss of function are the classical signs of acute inflammation. Inflammation is a complex array of biological responses of tissues to harmful stimuli, such as pathogens, irritants, or damaged cells. Due to the complex nature of inflammatory responses, the focus of this section on inflammatory responses will be only on selected cytokine production by macrophages (GM-MΦ and M-MΦ) (Figure 2.9).

Figure 2.9. Expression analysis of human cytokine signalling and production pathway-linked genes between GM-MΦ and M-MΦ. The x-axis displays selected canonical pathways and the y-axis displays the significance by –log(p-value). P values are calculated using the right-tailed Fisher’s Exact test. The threshold lines denotes the cutoff value for significance, which represent p = 0.05

IL10-, TREM1-, IL6-, IL8-, IL17- and TNFR2-signalling pathway-linked genes are significantly different between GM-MΦ and M-MΦ. However, most of these cytokines are not produced by macrophages basally, unless triggered by a stimulant such as LPS. IL17 cytokines are primarily produced by Th17 helper cells and are involved in autoimmune diseases (28). Given that IL17-signalling was three

91 fold more abundant in M-MΦ than in GM-MΦ, suggested that M-MΦ may have major contributions to organ-specific autoimmune conditions driven by Th17 cells. Cells of the monocyte-macrophage lineage have been recognized to be heterogeneous (1) and there is considerable interest in their polarization and functional diversity. Since macrophage phenotypes were proposed to be a continuum rather than distinctly separate for M1 and M2 types (18), would GM-MΦ and M-MΦ have distinct responses when activated by LPS or would their responses to LPS be similar?

2.5.8 Monocyte-derived macrophages cytokine and gene responses to LPS stimulation

GM-MΦ and M-MΦ were seeded at 1 x 106 cells/mL in 12-well plate format and left in 5% CO2 incubator at 37 °C overnight for cells to adhere. Cells were serum deprived for at least 8 h prior to treatment. Cells were treated with LPS (5 ng/mL) for 24 h. Untreated and LPS-treated cell culture supernatants were collected and analysed for cytokine productions by ELISA. Cells were also lysed and collected for mRNA extraction. GM-MΦ responded more robustly to LPS-induced proinflammatory cytokine productions such as IL6, IL8 and TNF (Figure 2.10A-C), compared to M- MΦ. Conversely, LPS-induced IL10 production was higher in M-MΦ than GM-MΦ (Figure 2.10D). The robustness of GM-MΦ in responding to LPS stimulation compared to M-MΦ was also reflected at mRNA levels. LPS induced higher expression levels of IL6, IL8 and TNF in GM-MΦ (Figure 2.10E-G) and higher IL10 expression levels in M-MΦ (Figure 2.10H).

Figure 2.10. LPS-induced IL6, IL8, TNF and IL10 proteins and gene expressions. GM-MΦ (black) and M-MΦ (white) were subjected to LPS (5 ng/mL) stimulation for 24 h. LPS-treated and untreated cell culture supernatants were collected to detect cytokine production and total mRNA were isolated from cell lysates at the end of LPS treatment. Genes were detected by real-time PCR, normalised against 18S rRNA expression and converted to fold change relative to control. Error bars are means ± SEM of three independent experiments (n=3). *p < 0.05, **p < 0.01 and ***p < 0.001 by student t-test.

93 2.6 Discussion

GM-CSF and M-CSF, previously known as hemopoietic growth factors, are now better known as mediators involved in regulating the macrophage lineage leading to macrophage heterogeneity. In the field of immunology, macrophages are now routinely categorized in subpopulations according to their gene expression profiles induced by various cytokines and pathogen-derived ligands. Although correlations between murine and human macrophages differentiated with CSFs are as low as 17%, it is appealing that information about macrophage mechanisms and signalling pathways were extrapolated from murine counterparts to understand human macrophage immunity. Although genetic differences between GM-CSF differentiated and M-CSF differentiated human monocyte-derived macrophages have been reported in a few studies (14, 22), others have focused mainly on polarization profiles (23) or dependence on interferon signalling (16). This chapter instead describes the functional heterogeneity of GM-MΦ and M-MΦ in the basal state and in response to LPS and, most importantly, validate these cell populations as in vitro models for gaining insights to human macrophage biology. Cellular migration and inflammatory responses have been identified as the most important biological functions by microarray and pathway analysis, and to be different between human GM-MΦ and M-MΦ. Macrophages are motile and able to migrate actively through the walls of blood vessels into surrounding tissues. The exchange between M1/M2 macrophage populations locally within the tissues can change over time and is associated with pathology (18). Two well-studied examples of pathology as the result of macrophage phenotypic switching are cancer and obesity-induced inflammation. Rho GTPases are important regulators of the actin cytoskeleton and thereby control cellular adhesion and migration. Although GM-CSF stimulation led to increased Rac2 expression, while M-CSF increased RhoA expression during myeloid differentiation (24), the actual cellular migration is not addressed in this study. GM-MΦ has the propensity to migrate more rapidly compared to M-MΦ, in fact 4 times faster in this study. When visualized under the microscope, GM-MΦ appeared to have more lamellipodia formation compared to M- MΦ. The lamellipodia formation is sensitive to Rho GTPase inhibitor, thereby suggesting that the level of Rho activity between GM-MΦ and M-MΦ is different.

94 Macrophages have a ‘plastic’ gene expression phenotype that can change rapidly depending on the microenvironment signals and even lead to a switching of phenotype from M1 to M2 and vice versa (25, 26). In fact, even the presence of serum has been demonstrated to alter macrophage polarization (27). Interestingly, after the 6 days differentiation protocol in the presence of 10% serum, the magnitude of IL6, TNF and IL10 cytokine production was increased by the presence of serum and CSFs, but the correlation between GM-MΦ and M-MΦ was not affected when comparing their responses to LPS (Figure 2.3 versus Figure 2.10). Besides influences of serum in macrophage polarization, the duration of CSF exposure can directly influence gene expressions (14). This reinforced the view that, when studying macrophage biology and in particular macrophage polarization, the culturing conditions and durations need to be taken into account.

2.7 Conclusions

In conclusion, GM-MΦ and M-MΦ exhibit significant differences in gene expression under basal conditions and these differences lead to different signalling outcomes for these cells when challenged with the bacteria-mimicking stimulant LPS. Data generated from cDNA microarray and PCR experiments correlated with each other. The largest functional differences between GM-MΦ and M-MΦ have been further investigated and validated as well. GM-MΦ is not only more proinflammatory in nature, but also more robust to LPS stimulation and can migrate to site of inflammation more rapidly than M-MΦ. This chapter has characterized and validated GM-MΦ and M-MΦ as suitable in vitro models to study proinflammatory and anti- inflammatory responses specific for human monocyte-derived macrophages.

95 2.8 References

1. Gordon, S., and P. R. Taylor. 2005. Monocyte and macrophage heterogeneity. Nat. Rev. Immunol. 5: 953-964. 2. Shi, C., and E. G. Pamer. 2011. Monocyte recruitment during infection and inflammation. Nat. Rev. Immunol. 11: 762-774. 3. Burgess, A. W., and D. Metcalf. 1980. The nature and action of granulocyte- macrophage colony stimulating factors. Blood 56: 947-958. 4. Sherr, C. J., C. W. Rettenmier, R. Sacca, M. F. Roussel, A. T. Look, and E. R. Stanley. 1985. The c-fms proto-oncogene product is related to the receptor for the mononuclear phagocyte growth factor, CSF-1. Cell 41: 665-676. 5. Metcalf, D. 1991. Control of granulocytes and macrophages: molecular, cellular, and clinical aspects. Science 254: 529-533. 6. Hamilton, J. A. 2008. Colony-stimulating factors in inflammation and autoimmunity. Nat. Rev. Immunol. 8: 533-544. 7. Hamilton, J. A., and G. P. Anderson. 2004. GM-CSF Biology. Growth Factors 22: 225-231. 8. Burgess, A. W., D. Metcalf, L. G. Sparrow, R. J. Simpson, and E. C. Nice. 1986. Granulocyte/macrophage colony-stimulating factor from mouse lung conditioned medium. Purification of multiple forms and radioiodination. Biochem. J. 235: 805-814. 9. Akagawa, K. S., I. Komuro, H. Kanazawa, T. Yamazaki, K. Mochida, and F. Kishi. 2006. Functional heterogeneity of colony-stimulating factor-induced human monocyte-derived macrophages. Respirology 11 Suppl: S32-36. 10. Naito, M., S. Umeda, K. Takahashi, and L. D. Shultz. 1997. Macrophage differentiation and granulomatous inflammation in osteopetrotic mice (op/op) defective in the production of CSF-1. Mol. Reprod. Dev. 46: 85-91. 11. Xu, W., N. Schlagwein, A. Roos, T. K. van den Berg, M. R. Daha, and C. van Kooten. 2007. Human peritoneal macrophages show functional characteristics of M-CSF-driven anti-inflammatory type 2 macrophages. Eur. J. Immunol. 37: 1594-1599.

96 12. Stanley, E. R., K. L. Berg, D. B. Einstein, P. S. Lee, F. J. Pixley, Y. Wang, and Y. G. Yeung. 1997. Biology and action of colony--stimulating factor-1. Mol. Reprod. Dev. 46: 4-10. 13. Mantovani, A., A. Sica, S. Sozzani, P. Allavena, A. Vecchi, and M. Locati. 2004. The chemokine system in diverse forms of macrophage activation and polarization. Trends Immunol. 25: 677-686. 14. Lacey, D. C., A. Achuthan, A. J. Fleetwood, H. Dinh, J. Roiniotis, G. M. Scholz, M. W. Chang, S. K. Beckman, A. D. Cook, and J. A. Hamilton. 2012. Defining GM-CSF- and macrophage-CSF-dependent macrophage responses by in vitro models. J. Immunol. 188: 5752-5765. 15. Akagawa, K. S. 2002. Functional heterogeneity of colony-stimulating factor- induced human monocyte-derived macrophages. Int. J. Hematol. 76: 27-34. 16. Fleetwood, A. J., H. Dinh, A. D. Cook, P. J. Hertzog, and J. A. Hamilton. 2009. GM-CSF- and M-CSF-dependent macrophage phenotypes display differential dependence on type I interferon signaling. J. Leukoc. Biol. 86: 411-421. 17. Hochberg, Y., and Y. Benjamini. 1990. More powerful procedures for multiple significance testing. Stat. Med. 9: 811-818. 18. Mosser, D. M., and J. P. Edwards. 2008. Exploring the full spectrum of macrophage activation. Nat. Rev. Immunol. 8: 958-969. 19. Stossel, T. P. 1994. The machinery of cell crawling. Sci. Am. 271: 54-55, 58- 63. 20. Small, J. V., T. Stradal, E. Vignal, and K. Rottner. 2002. The lamellipodium: where motility begins. Trends Cell Biol. 12: 112-120. 21. Buhl, A. M., N. L. Johnson, N. Dhanasekaran, and G. L. Johnson. 1995. G alpha 12 and G alpha 13 stimulate Rho-dependent stress fiber formation and focal adhesion assembly. J. Biol. Chem. 270: 24631-24634. 22. Brocheriou, I., S. Maouche, H. Durand, V. Braunersreuther, G. Le Naour, A. Gratchev, F. Koskas, F. Mach, J. Kzhyshkowska, and E. Ninio. 2011. Antagonistic regulation of macrophage phenotype by M-CSF and GM-CSF: implication in atherosclerosis. Atherosclerosis 214: 316-324. 23. Jaguin, M., N. Houlbert, O. Fardel, and V. Lecureur. 2013. Polarization profiles of human M-CSF-generated macrophages and comparison of M1-

97 markers in classically activated macrophages from GM-CSF and M-CSF origin. Cell. Immunol. 281: 51-61. 24. van Helden, S. F., E. C. Anthony, R. Dee, and P. L. Hordijk. 2012. Rho GTPase expression in human myeloid cells. PLoS ONE 7: e42563. 25. Porcheray, F., S. Viaud, A. C. Rimaniol, C. Leone, B. Samah, N. Dereuddre- Bosquet, D. Dormont, and G. Gras. 2005. Macrophage activation switching: an asset for the resolution of inflammation. Clin. Exp. Immunol. 142: 481-489. 26. Heusinkveld, M., P. J. de Vos van Steenwijk, R. Goedemans, T. H. Ramwadhdoebe, A. Gorter, M. J. Welters, T. van Hall, and S. H. van der Burg. 2011. M2 macrophages induced by prostaglandin E2 and IL-6 from cervical carcinoma are switched to activated M1 macrophages by CD4+ Th1 cells. J. Immunol. 187: 1157-1165. 27. Rey-Giraud, F., M. Hafner, and C. H. Ries. 2012. In vitro generation of monocyte-derived macrophages under serum-free conditions improves their tumor promoting functions. PLoS ONE 7: e42656. 28. Harrington, L. E., R. D. Hatton, P. R. Mangan, H. Turner, T. L. Murphy, K. M. Murphy, and C. T. Weaver. 2005. Interleukin 17-producing CD4+ effector T cells develop via a lineage distinct from the T helper type 1 and 2 lineages. Nat. Immunol. 6: 1123-1132.

98 Chapter

3 Responses to human C5a by human monocyte-derived macrophages differentiated with either GM-CSF or M-CSF

99 3.1 Abstract

In Chapter 2, GM-MΦ and M-MΦ were found to be quite different from each other. These differences were recorded to be the largest in the microarray experiment as calculated by distant matrix algorithms in this chapter. Therefore it was expected that GM-MΦ and M-MΦ, with large genomic differences would respond to C5a differently. C5a exerted similar effects on GM-MΦ and M-MΦ in the calcium release,

ERK phosphorylation and macrophage migration assays. The EC50 values of C5a in these three functional assays between GM-MΦ and M-MΦ were comparable. The transcriptional profile stimulated by C5a included several genes whose products are proinflammatory such as chemokines and cytokines (CCL3, CCL3L1, CCL3L3, CXCL2, IL8, PTGS2 and TNF). C5a also stimulated the expression of a number of transcription factors (ATF3, EGR1, EGR2, EGR3, FOS, FOSB and JUNB) that amplified immune responses and were also likely to be involved in a positive feedback loop resulting in the stimulation of expression of themselves and other proinflammatory genes. FOSB and FOS genes were highly upregulated in human macrophages (GM-MΦ and M-MΦ) when stimulated with C5a. At the same time, C5a can limit immune responses and likely participates in a negative feedback loop to preserve homeostasis and dampen down damage by activating ZFP36, SOCS3, NFKBIZ, DUSP1 and DUSP6 genes. A list of 33 genes that were commonly stimulated by C5a on GM-MΦ and M-MΦ was uploaded for Ingenuity Pathway Analysis (IPA) and top ranking diseases and functions between GM-MΦ and M-MΦ overlapped with major pathways involved in neurological disease, cancer, cell death and survival. A separate microarray experiment was conducted to create an overview of the temporal profile of C5a induced genes and it was found that all genes differentially regulated by C5a could be further categorised into clusters according to their temporal profile, with cluster 1, 2 and 3 featuring genes that were upregulated by C5a at 0.5, 6 and 18 h respectively and cluster 4 featuring genes that were downregulated by C5a at 6 h. By identifying the clusters of genes activated by C5a, potential rate-limiting gene candidates may be identified for potential clinical intervention in C5a-mediated inflammatory disorders.

100 3.2 Introduction

C5a is one of the most potent inflammatory mediators generated upon complement activation through C5 cleavage and usually accompanied by the formation of the membrane attack complex (1). C5a has been show to stimulate proinflammatory cytokines such as tumour necrosis factor alpha (TNF), interleukin-6 (IL6) and -1beta (IL1B) in various cell types (2-6). These inflammatory mediators were produced as the result of increased transcriptional activity induced by C5a, which suggested that transcriptional regulation was important in C5a signalling. Only one study has been reported in the literature to date on the use of cDNA microarray analysis to assess gene expression induced by C5a and this was performed with human umbilical vein endothelial cells (HUVECs) (7). C5a particularly induced gene expression of cell adhesion molecules (E-selectin, ICAM-1, VCAM-1), cytokines/chemokines (VEGFC, IL6) (7). In HUVECs, C5a activated a similar set of genes to those induced by either TNF or LPS (7). The chief receptor for C5a (C5aR) is widely distributed in the human body (8), but especially highly expressed in leukocytes, particularly neutrophils, monocytes and macrophages (ArrayExpress: E-MTAB-513). A high throughput transcriptional profile of RNA from individual human tissues, as well as a mixture of 16 different human tissues, has been reported (Illumina Body Map) by ArrayExpress (http://www.ebi.ac.uk/arrayexpress) (9). Although C5aR signalling pathways have been previously defined in human neutrophils in 1994 (10), the transcriptional regulation of inflammatory genes by C5a in human leukocytes remains to be determined to date. In particular, the C5a-induced gene expression profile in human macrophages has not been reported and was of interest to us in the context of Chapter 2, to build on our knowledge of inflammatory gene expression differences between M1-like and M2-like macrophages (GM-MΦ and M-MΦ). Plasticity is a hallmark of macrophages. When activated by innate recognition or signals from lymphocyte subsets, macrophages can undergo adaptive responses playing critical roles in immunity by selectively participating in the initiation of host defence, inflammation and the maintenance of homeostasis (11). This phenomenon of macrophages has been extensively reviewed by others and therefore will not be discussed here (12, 13). Exposing freshly isolated CD14+ human blood monocytes to

101 different differentiating agents (GM-CSF or M-CSF) for 6 days, gave rise to macrophages with large genomic differences, which then translated to differences in cellular movement, haematological development and inflammatory responses between GM-MΦ and M-MΦ (Chapter 2). Based on the variable levels of C5aR expression in across different human tissue samples, it was considered that macrophages would be a good cell model to investigate C5a-induced gene expression. Thus, it was decided that a cDNA study of C5a-induced gene expression in GM-MΦ and M-MΦ would likely provide valuable insights to C5a-mediated inflammation in inflammatory cells known to be associated with immunity and chronic inflammatory diseases.

3.3 Aims

In Chapter 2, the differently stimulated culturing of primary human monocytes to obtain macrophage-like cells, GM-MΦ and M-MΦ, has been established as producing suitable in vitro models for primary human proinflammatory and anti-inflammatory macrophages, respectively. Therefore, in this chapter, a comparative study on the effects of human C5a on GM-MΦ versus M-MΦ will be reported with the objective of identifying similarities and differences in responses of these different cell types to recombinant human C5a. The specific aims of this chapter were: 1. To determine the potency of C5a on GM-MΦ and M-MΦ in several functional assays, namely, intracellular Ca2+ mobilization, ERK phosphorylation and cell migration. 2. To identify cell-type specific differences in gene expression profiles of GM- MΦ and M-MΦ in response to C5a stimulation by microarray analysis. 3. To identify conserved and divergent downstream biological processes induced by C5a between GM-MΦ and M-MΦ.

102 3.4 Methods

3.4.1 Reagents

Recombinant human GM-CSF and M-CSF were purchased from PeproTech. Recombinant human C5a was purchased from Sino Biological Inc. All cell culture reagents were purchased from Invitrogen and all analytical grade chemical reagents were obtained from Sigma-Aldrich, unless otherwise stated.

3.4.2 Isolation of human monocytes and macrophages

µg/mL streptomycin and 2 mM GlutaMAX at 37 °C in presence of 5% CO2. To generate human monocyte-derived macophages (HMDM), CD14+ monocytes were cultured in the presence of either 10 ng/mL GM-CSF (GM-MΦ) or M-CSF (M-MΦ) for at least 6 days.

3.4.3 Intracellular calcium mobilization assay

HMDMs were seeded at 0.5 x 106 cells/mL (100 µL) in black-wall, clear- bottom 96-well plates (Corning) and incubated at 37 °C in the presence of 5% CO2 overnight for cells to adhere. Cells were washed once with assay buffer (Hank’s buffered salt solution (HBSS), 20 mM HEPES, 2.5 mM probenecid, pH 7.4). 100 µL of dye-loading buffer (12 mL of assay buffer, 1% FBS, 25 µL of Fluo-3AM (Invitrogen) (final concentration = 4 µM), 25 µL 20% pluronic acid) was added to

103 each well and incubated for 1 h at 37°C. Excess loading dye was washed away with assay buffer before adding 50 µL of tested compound. Fluorescence intensity was recorded with on FLIPR Tetra (Molecular Devices) with excitation and emission wavelengths of 485 nm and 520 nm respectively. Calcimycin ionophore A23187 (25 µM) (Sigma) is used as a dye loading positive control.

3.4.4 Receptor saturation binding assay

Binding equilibrium was measured by titrating various concentration of [125I]- C5a (2200 Ci/mmol, Perkin Elmer) with 1 x106 cells/mL in Tris buffer (50 mM Tris,

3 mM MgCl2, 0.1 mM CaCl2, 0.5% (w/v) bovine serum albumin, pH 7.4) for 60 min in a 96-well Nunc round bottom plate at room temperature. Unbound [125I]-C5a was removed by filtration through glass microfiber filter GF/B (Whatman) which had been soaked in 0.6% polyethyleneimine to reduce non-specific binding. The filter was washed 3 times with cold buffer (50 mM Tris) pH 7.4 and bound [125I]-C5a was assessed by scintillation counting on Microbeta counter. Specific [125I]-C5a binding was defined as the difference between total binding and non-specific binding as determined in the presence of 1 µM unlabeled C5a. Bmax and Kd values were calculated by nonlinear regression analysis using GraphPad Prism.

3.4.5 Quantification of ERK1/2 phosphorylation by AlphaScreen Surefire

Cells were seeded at 4000 cells/well in white 384-ProxiPlate (PerkinElmer) overnight. On the following day, media was removed from the cells and were serum- deprived for 2 h to reduce basal ERK1/2 activity, prior to stimulation with C5a over a stated time course. Cells were lysed with Alphascreen SureFire lysis buffer, with shaking for 10 min at room temperature and assayed for phosphor-ERK1/2 (phosphor-Thr202/Tyr204) according to Alphascreen SureFire assays (PerkinElmer) manufacturer’s instructions. At completion of the assay, signal in the wells was measured using Pherastar (BMG Labtech).

104 3.4.6 In vitro migration assay (modified Boyden chamber)

Migration of human macrophages was measured in a modified Boyden chamber migration assay using Transwell inserts with a 5 µm porous membrane (Corning). Cells were loaded into the migration chamber in serum free medium. To initial cell migration, serum free medium was placed in the lower chamber. At the end of the time course study, cells were removed from the upper side of the membranes, and nuclei of migratory cells on the lower side of the membrane were stained with DAPI. And counted under a fluorescence microscope using a x10 objective.

3.4.7 Enzyme-linked immunosorbent assay (ELISA)

3.4.8 RNA extraction and reverse transcription

Total RNA was extracted from cells using an RNeasy Mini Plus kit (Qiagen) and following manufacturer’s instruction. RNA concentrations were determined by measuring absorbance at λ = 260/280 nm on a spectrophotometer. Total RNA (2-10 µg) extracted from cells and random oligo dT primer (1 µg) were initially incubated at 70°C for 10 min and then cooled on ice for at least 1 min before being reverse transcripted with Superscript III (Invitrogen) at 50°C for 50 min, then 70°C for 10 min. cDNA samples were stored at -20°C until further use.

3.4.9 Real-time quantitative polymerase chain reaction (real-time PCR)

105 specific primers to perform real-time PCR on the ABI PRISM® 7500 Real Time PCR System (Applied Biosystems™). Conditions were used according to the manufacturer’s instructions. Relative gene expression was normalised against 18S rRNA expression and then converted to fold change against control samples. All samples were analysed in duplicates. Primer sequences used are: C5AR1, 5’- GCCCAGGAGACCAGAACATGAACTC-3’ and 3’- TCCACAGGGGTGTTGAGGTCCA-5’; CCL3, 5’- CCCACATTCCGTCACCTGCTCAG-3’ and 3’- AGCAGCAAGTGATGCAGAGAACTG-5’; TNF, 5’- TCTGGCCCAGGCAGTCAGATC-3’ and 3’-CACTGGAGCTGCCCCTCAGC-5’; IL8, 5’-ACCACCGGAAGGAACCATCTCACT-3’ and 3’- CTTGGCAAAACTGCACCTTCACAC-5’; PTGS2, 5’- TGGCAGGGTTGCTGGTGGTA-3’ and 3’-TCTGCCTGCTCTGGTCAATGGA-5’; IL1B, 5’-CCTCTTCGAGGCACAAGGCACAA-3’ and 3’- TGGCTGCTTCAGACACTTGAGCAAT-5’; EGR1, 5’- CGAGCACCTGACCGCAGAGTCT-3’ and 3’-GGGGGCAGTCGAGTGGTTTGG- 5’; FOSB, 5’-GACCCTCAGCCCTGGCCTTT-3’ and 3’- CAGCCCCTCACTGCCTTCCT-5’; TLR7, 5’-GAACGCCCTGGCCACAGACA-3’ and 3’-ACAGCCAGGCCTCTCCTTGGT-5’; TLR8, 5’- AGCTGCGCTACCACCTTGAAGA-3’ and 3’- GGCCAATCCCGGATCCCAATCC-5’; FABP5, 5’- GCACCCACCATGGCCACAGT-3’ and 3’-CGCAAAGCTATTCCCACTCCTAGC- 5’; LEP, 5’-GTAGGAATCGCAGCGCCAACG-3’ and 3’- TCCGCACAGGGTTCCCCAATG-5’; ATF3, 5’-CGAGCGGAGCCTGGAGCAAA- 3’ and 3’-GGGGACAGGCAGGGGACGAT-5’; DUSP5, 5’- ATGACCAGGGTGGCCCAGTTGAA-3’ and 3’- CGGGAGACATTCAGCAGGGCTGTG-5’; DUSP6, 5’- CGACGAGAGCAGCAGCGACT-3’ and 3’- TGAAGCCACCTTCCAGGTAGAACG-5’; CD83, 5’- TCACTTGTTTTGCACGGCTACAGA-3’ and 3’- TGGGGAGGTAACTGGGAGAAAAGC-5’; CD276, 5’- ACAGCTGCCTGGTGCGCAAC-3’ and 3’-TGGACCTCCACGGCTCCTGT-5’; CTSL1, 5’-TCCGAGCCGGGTGGACACAGT-3’ and 3’-

106 AGCTGAGGCAATTCCCAGGCAAAA-5’; TIMP3, 5’- AGCTGGAGCCTGGGGGACTG-3’ and 3’-GGCCCGGATCACGATGTCGG-5’; MMP19, 5’-AGGTGGCGCCCGTGGACTA-3’ and 3’- GGGCCAGGACTCTCCCTCAGA-5’.

3.4.10 Hybridization and analysis of human array-based gene expression

Quality controls were performed on the RNA to check the integrity, purity and concentration of samples. Synthesis of biotin-labelled cRNA, hybridisation to 50-mer probes on the HumanHT-12 v4.0 Expression BeadChip (Illumina®) and detection of hybridised target cRNA were performed by Australian Genome Research Facility (AGRF) according to the Illumina® whole-genome gene expression beadchips manufacturer’s instruction. GeneSpring GX software version 11.5 (Agilent Technologies) and web-based Ingenuity Pathway Analysis (IPA) software (Ingenuity Systems, Inc.) was used for the analysis of differences in C5a-mediated gene expression between GM-MΦ and M-MΦ. Statistically significant differences between samples with more than 2-fold change were assessed using student t-test.

3.4.11 Statistical analysis

107 3.5 Results

C5a is generated by the cleavage of C5 during complement activation and it is known to have powerful proinflammatory actions and attracts immune cells to sites of infection or tissue damage. Serum levels of C5a can reach as high as 100 nM in patients with sepsis (14). The local concentration of C5a is expected to be higher at the site of inflammation, but this has not been determined. Human C5a (100 nM) induces significant apoptosis in freshly isolated thymocytes (15). Other lymphoid cells such as B and T cells had similar apoptotic responses when treated with such a concentration of C5a (16). As it is the objective to investigate the acute responses of C5a under normal healthy physiological conditions, it is therefore crucial that HMDM are not over-stimulated or over-stressed as in a chronic or diseased state. The calcium mobilization, ERK phosphorylation and macrophage migration assays will be conducted to determine the optimum concentration of human C5a for stimulating HMDMs (GM-MΦ: macrophages differentiated with GM-CSF; and M-MΦ: macrophages differentiated with M-CSF). This optimized concentration of C5a

(~EC80 values) will be used to treat GM-MΦ and M-MΦ and analysed for the microarray experiment.

3.5.1 C5a has similar agonist potency on GM-MΦ and M-MΦ

C5a can induce intracellular calcium mobilization in a variety of cells upon C5aR activation (17). C5a had similar potency on the C5aRs of GM-MΦ and M-MΦ as determined by Ca2+ mobilization assay (Figure 3.1). Interestingly, GM-MΦ and

M-MΦ not only achieved similar EC50 values, but the maximal responses induced by C5a (300 nM) were also similar in these two subtypes of macrophages (Figure 3.1B). Higher maximal responses are usually indicative of higher receptor expressions. This is also an indication that the number of C5aR receptors expressed on both GM-MΦ and M-MΦ cell surface may be similar.

108

Cells pEC50 ± SEM EC50 (nM) n GM-MΦ 8.6 ± 0.1 2.3 3 M-MΦ 8.9 ± 0.1 1.3 3 Figure 3.1. C5a-induced intracellular calcium mobilization in HMDM. (A) Ca2+ mobilization in GM-MΦ or M-MΦ was induced by various concentrations of recombinant human C5a. Cells treated with calcimycin (25 µM) were used as a dye-loading control. (B) Maximal Ca2+ mobilization induced by C5a (300 nM) in GM-MΦ (black) and M-MΦ (white) were plotted against calcimycin control. Error bars are means ± SEM of three independent experiments (n=3). ***p < 0.001 by student t-test.

Besides calcium mobilization, C5a can also induce ERK phosphorylation (10). Similar to previous calcium release results in Figure 3.1, C5a has similar potency on the C5aR of GM-MΦ and M-MΦ (Figure 3.2). Although GM-MΦ adopted a more proinflammatory profile than M-MΦ, as characterized in Chapter 2, C5a exerts similar effects on both GM-MΦ and M-MΦ based on data collected from calcium release and ERK activation assays, measuring different biological readouts.

109

Cells pEC50 ± SEM EC50 (nM) n GM-MΦ 8.4 ± 0.4 4.2 3 M-MΦ 8.6 ± 0.3 2.8 3 Figure 3.2. C5a-induced ERK1/2 phosphorylation in HMDM. ERK phosphorylation in GM-MΦ and M-MΦ was induced by various concentrations of recombinant human C5a. Cells were treated for 5 min and lyzed to stop C5a- induced reaction. Phosphorylated ERK1/2 was measured by AlphaScreen® SureFire® assays. Error bars are means ± SEM of three independent experiments (n=3).

3.5.2 GM-MΦ and M-MΦ have similar C5aR expression levels

Total mRNA was isolated from GM-MΦ and M-MΦ to measure C5AR gene expression by real-time PCR. The raw data collected was normalized against 18S ribosomal RNA as relative gene expression. There were no significant difference in C5AR expression levels between GM-MΦ and M-MΦ (Figure 3.3). Although there were no differences in C5aR expression at the transcriptional levels between GM-MΦ and M-MΦ, the same trend may not be reflected at the translational levels. A key assumption, and often the biggest flaw commonly made by researchers, was that the levels of mRNA expression correlates with levels of protein expression. This assumption is certainly far from being perfect, and has often been conveniently overlooked (18, 19). The mean correlations between protein and corresponding RNA levels were about 0.2 - 0.3. In other words, very little correlation exists between gene and protein expression.

110 8×10-05 S

8 -05

1 6×10

1 4×10-05 R A

5 2×10-05 C 0 GM-MΦ M-MΦ

Figure 3.3. GM-MΦ and M-MΦ have similar C5AR gene expression. Total mRNA extracted from GM-MΦ (black) and M-MΦ (white) were used to measure C5AR gene expression by real-time PCR. Raw values were then normalised against 18S rRNA expressions. Error bars are means ± SEM of three independent experiments (n=3).

Total protein expression of interest can be easily determined by performing western blots. However, GPCRs have been shown to have enhanced functions in optimized conditions without changing total receptor protein expression levels (20, 21), therefore total protein expression level may not be that useful when ascertaining GPCR functions. The amount of functional receptors must be measured by ligand binding because the total receptor yield does not account for receptors that are either in an inactive form or located intracellularly with no access to external ligand binding. To measure the levels of functionally active C5aR, a saturation-binding assay was used to determine the expression of functional C5aR on the cell surface using 125 radioactive I-C5a. As expected, GM-MΦ and M-MΦ have similar Bmax values (total density of receptors) (Figure 3.4).

111

Cells Bmax (fmol/mg) Kd (pM) GM-MΦ 12.0 117.5 M-MΦ 11.2 87.5

Figure 3.4. GM-MΦ and M-MΦ have similar cell surface C5aR expressions. Receptor saturation assay was performed using 1 x 106 cells/mL (A) GM-MΦ and (B) M-MΦ and incubated with various concentration of 125I-C5a in the absence or in the presence of C5a (1 µM) to determine total and non-specific C5aR binding. (C and D) Specific binding values were calculated by subtracting non-specific binding values from total binding values. Counts were converted to fmol/mg and represented by the scatchard plot. Bmax and Kd values were calculated using nonlinear regression after converting cpm values to fmol/mg values. Error bars are means ± SEM of three independent experiments (n=3).

3.5.3 C5a-mediated chemotaxis in GM-MΦ and M-MΦ

Metabolic dysfunction is associated with increased immune cells (especially macrophages, mast cells and T cells) infiltrating into adipose tissue (22, 23). Macrophages constitute as much as 40% of the cell population in the adipose tissue of

112 obese humans and animals (24). Although macrophages are highly plastic in their responses and largely influenced by their local environment, no one has actually reported whether the chemotactic index between macrophage subtypes are different or not. Since C5a is one of the most potent endogenous agents known to induce chemotaxis of immune cells, especially macrophages, the chemotactic responses between GM-MΦ and M-MΦ were compared in a migration assay herein. The cell migration assay measured the ability of macrophages to migrate across a physical barrier (5 µm porous membrane) over time. Cells that have migrated from the upper chamber across to the lower chamber of the Transwell will be quantified under a microscope. The greater the number of cells counted, the higher the chemotactic potential. The basal cellular migration ability of GM-MΦ and M-MΦ were tested and there were little or no difference between the number of GM-MΦ and M-MΦ cells that migrated across the Tanswell membrane after 3 h in Chapter 2. However, after 6 h, the number of GM-MΦ compared to M-MΦ substantially increased. Therefore, the time point 3 h was chosen to investigate the effects of C5a on macrophage migration in GM-MΦ and M-MΦ in this chapter. Low concentrations of C5a (0.1 nM) did not induce chemotaxis in M-MΦ (Figure 3.5B) but the same concentration of C5a in GM-MΦ induced 2-fold more cells to migrate relative to control (Figure 3.5A). This trend was also observed at a much higher concentration of C5a (EC90 = 3 nM). Despite the magnitude of response seen in GM-MΦ compared to M-MΦ, C5a appeared to exert equipotent chemotactic effects on both subtypes with similar EC50 values at sub nanomolar concentrations (Figure 3.5C, D). In summary, C5a had equipotent effects on both GM-MΦ and M-MΦ in all three functional assays investigated in this study; the calcium mobilization assay, ERK phosphorylation assay and cell migration assay. These results were unexpected, as one would hypothesize or expect that inflammatory responses mediated by C5a would be exacerbated in GM-MΦ; macrophages with a proinflammatory background compared to M-MΦ; macrophages with an anti-inflammatory background. Results from these functional assays supported the theory that macrophages are more likely to be influenced depending on their microenvironment rather than their previous polarization background.

113

Cells pEC50 ± SEM EC50 (nM) n GM-MΦ 9.5 ± 0.2 0.3 3 M-MΦ 9.4 ± 0.2 0.4 3

Figure 3.5. C5a is equipotent in mediating chemotaxis on GM-MΦ and M-MΦ. (A) GM-MΦ: black and (B) M-MΦ: white; were serum-deprived overnight prior to addition to the upper chamber of the Transwell and then allowed to migrate through the 0.5µm porous membrane into the lower chamber containing serum free media with various concentrations of C5a (0.1-30 nM) for 3 h. Migrated cells on the lower side of the membrane were stained with DAPI and counted under a fluorescence microscope using a x10 objective. (C) Chemotactic index is expressed as % against C5a maximal response. (D) The EC50 values were calculated using the Prism software. Error bars are means ± SEM of three independent experiments (n=3). **p < 0.01 and ***p < 0.001 by student t-test.

114 3.5.4 C5a-induced gene expressions on GM-MΦ and M-MΦ

C5a appeared to be equipotent on GM-MΦ and M-MΦ in all three functional assays; calcium release, ERK phosphorylation and macrophage migration. The EC80 value of C5a on GM-MΦ and M-MΦ in the calcium release and ERK phosphorylation assay was calculated to be ~ 30 nM, whereas this value was 10-fold more potent in the migration assay. The migration assay was performed with longer exposure of the cells to C5a (3 h), compared to the calcium release and ERK phosphorylation assays (5 min). This could be the reason for differences in potency between assays; hence in the PCR and later with microarray, C5a will be used at 30 nM for inducing gene expression. The levels of CCL3, TNF, IL8, PTGS2, IL1B and EGR1 mRNA expression were significantly amplified after exposing GM-MΦ and M-MΦ to C5a for 0.5 h, and all 6 genes were more highly expressed in GM-MΦ than M-MΦ (Figure 3.6).

115

Figure 3.6. A panel of genes induced by 30 nM C5a after 0.5 h. GM-MΦ (black) and M-MΦ (white) were serum deprived overnight before treating with 30 nM C5a for 0.5 h. The expression of (A) CCL3, (B) TNF, (C) IL8, (D) PTGS2, (E) IL1B and (F) EGR1 genes was detected by quantitative real-time PCR. The relative expression of genes was normalized against 18S rRNA. Error bars are means ± SEM of three independent experiments (n=3). *p < 0.05, **p < 0.01 and ***p < 0.001 by student t-test.

3.5.5 Quality control of genes hits: sample variations versus C5a treatments

cDNA microarray analysis of C5a-treated human umbilical vein endothelial cells (HUVEC) (7), or C5aR-antagonist-treated tissues from disease murine models of secondary cataract (25) and neuropathic pain (26) have been previously performed. It was very surprising that such an analysis had not been reported in primary human macrophages, or at the very least on other macrophages. Microarray technology has

116 become a widely used tool for the rapid generation of gene expression and functional genomics data, however it is crucial that the conditions and treatments are optimized before performing the microarray experiment to gather good quality data and reproducible information. Given that the magnitude of gene responses induced by 30 nM C5a for 0.5 h on both GM-MΦ and M-MΦ was considerable (Figure 3.6), the signal to noise ratio was considered to be sensitive enough to be able to detect changes in gene expression in the notoriously low sensitivity microarray platform. While primary cells usually give much better gene expression profiles than laboratory cultured immortalized cells, a drawback is that genetic variations between donor samples can be significant, leading to distortion and inflated error rates and associated reproducibility issues. Increasing the population size can help to overcome variably between donors but a large population pool (n ≤ 20), increases technical difficulty and is too expensive for most academic research. As such, four instead of three (n=4) different donor-matched human monocyte-derived macrophages populations were differentiated by GM-CSF (GM-MΦ) and M-CSF (M-MΦ) for analysis by microarray. A total of 24 samples were submitted to Australian Genome Research Facility (AGRF) for reading on an Illumina microarray platform and processing as described in Chapter 2. The experimental design carried out for the microarray experiment was illustrated in Table 3.1. Since the focus of this chapter was on C5a, non C5a-treated samples (*) will not be discussed here. The raw data obtained was quantified, normalised and filtered using the GeneSpring software v11.5.

Table 3.1. The experimental design carried out for the microarray experiment.

Donor 1 Donor 2 GM- MΦ A Untreated B * C C5a G Untreated H * I C5a M-MΦ D Untreated E * F C5a J Untreated K * L C5a Donor 3 Donor 4 GM- MΦ M Untreated N * O C5a S Untreated T * U C5a M-MΦ P Untreated Q * R C5a V Untreated W * X C5a GM-MΦ or M-MΦ was seeded at 2 x 106 cells per well in a 6-well plate format and then serum deprived overnight when cells were adherent. The next day, cells were stimulated with 30 nM C5a for 0.5 h and then lysed cells to extract total mRNA. *Treatments not part of this thesis.

All samples were subjected to hierarchical clustering using pairwise distance matrix analysis by BioConductor software (27, 28). This is a very simple and quick

117 method to visualize and measure how similar or different all the samples are from one another, with clusters being most convergent at the bottom and divergent at the top. Samples that were most divergent were grouped according to the subtype of macrophages (GM-MΦ versus M-MΦ), and their differences were documented in Chapter 2, followed by donor-to-donor variations, and lastly by treatments (Figure 3.7). Although this distribution was not unexpected, it again reinforced the fact that the differences between donors can be greater than the differences between treatments versus control.

Figure 3.7. Hierarchical clustering of samples used for the microarray analysis. The dendrogram represents 24 samples being clustered using the pairwise distance matrix algorithms by BioConductor (28) and ranked with samples most closely related at the bottom and most divergent at the highest.

In the dendrogram, sample X (Donor 4: M-MΦ treated with 30 nM C5a) appeared to be an outlier (Figure 3.7), which was very surprising because all samples had undergone preliminary in-house quality control checks before submission for microarray gene expression. The calculated 28S:18S ratio for sample X was 1.5 which suggested that mRNA were intact and not degraded (Figure 3.8). This value was consistent cross all samples (~1.5-1.7). The concentration of mRNA was also within the manufacture’s recommendation. Given that the mRNA conditions were ideal and properly processed, the reason for this outlier was unlikely due to poor sample preparation. The dendrogram was drawn based on normalized data (Figure

118 3.9B), which was the recommended practice under the assumption that all samples were handled the same.

Start End % of total Name Size Size Area Area [NT] [NT] 18S 1499 1891 56.9 21.5 28S 2868 4093 85 32.1

Overall Results RNA Area: 265.0 RNA Conc: 123 ng/µL rRNA ratio [28S/18S]: 1.5

Figure 3.8. mRNA quality and sample assessment report for sample X. The mRNA quality for sample X was assessed by electrophoresis of total RNA followed by staining with ethidium bromide. The calculated 28S:18S rRNA ratio was 1.5.

However, this was not the case as a comparison between the raw (Figure 3.9A) and the normalized (Figure 3.9B) density plot of intensity showed that the intensity peak of sample X was at least 3-fold lower than the rest of the samples. Using the venn diagram plot, the intersection consisting of 36 genes were commonly regulated by C5a between 4 donors versus 3 donors (without sample X) (Figure 3.10). This result further confirmed that the sample X being flagged as an outlier originated from differences in signal intensity rather than differences in expression. As such, all

119 4 donors will be included in the analysis to achieve better statistic significance compared to 3 donors.

Figure 3.9. Comparison of density plot of intensity between (A) raw and (B) normalized data. Gene expression data measured by HumanHT-12v4 Beadchip are preprocessed using image analyzer to extract expression values from images and scaling algorithms to make expression values comparable across chips.

120

Figure 3.10. Venn diagram comparison between genes hits obtained using 4 donors versus 3 donors treated with 30 nM C5a for 0.5 h in M-MΦ.

Circles represented the ANOVA analysis of C5a-regulated genes obtained by 4 donors (blue) versus 3 donors (omit sample X) (yellow). Numbers inside each compartment represented the number of transcripts that were significant (≥ 2-fold change, p < 0.05) for that effect. The interacting region represents genes consistently upregulated by both 4 donors and 3 donors analysis.

3.5.6 Microarray experiment 1: Genes regulated by C5a on GM-MΦ versus M-MΦ

Labelled cRNA were detected by hybridising to 50-mer probes on the HumanHT-12 v4 BeadChip (Illumina, Inc. San Diego, USA). Out of 47,231 probes, 28,688 probes were detected for well-established coding transcripts, 11,121 probes were detected for provisional coding transcripts, 1,752 probes were detected for well- established non-coding transcripts and 2,209 probes were detected for provisional non-coding transcripts. The raw data obtained was quantified, normalised and filtered using the GeneSpring software v11.5. The statistical significance threshold limit was set at P < 0.05 and the Benjamini-Hochberg method was employed to control the false discovery rate (29). The cut-off value for fold changes was set at 2-fold. In order for an outcome to be considered as a positive hit, it must pass these two conditions. Gene hits with well-established coding transcripts were considered while the rest of

121 the hits were excluded from this analysis. Genes with duplicated or triplicated hits had their respective fold changes averaged. There were a total of 53 genes in GM-MΦ and 40 genes in M-MΦ upregulated after subjecting the cells to 30 nM C5a for 0.5 h. The list of genes uniquely regulated by C5a in either GM-MΦ (20 genes) or M-MΦ (7 genes) is summarized in Table 3.2, and the list of genes commonly regulated (33 genes) by C5a in both GM-MΦ and M- MΦ is summarized in Table 3.3.

Table 3.2. Genes upregulated by 30 nM C5a in either GM-MΦ or M-MΦ after 0.5 h were expressed as fold changes.

GM- Gene Accession Gene M-MΦ# Accession MΦ# CCDC138 2.1 NM_144978.1 CCL14 3.6 NM_032962.2 CCL8 3.1 NM_005623.2 DEFB104A 2.1 NM_080389.1 CD69 2.7 NM_001781.1 HMGN3 2.8 NM_004242.2 CRCP 2.2 NM_001040648.1 MCL1 2.2 NM_021960.3 DISC1 2.2 NM_018662.2 NFASC 2.0 NM_001005387.1 DUSP2 3.9 NM_004418.2 PRUNE2 2.3 NM_138818.2 EBI2 2.3 NM_004951.3 PSORS1C1 2.3 NM_014068.1 ETNK1 2.4 NM_001039481.1 IGFBPL1 2.5 NM_001007563.1 IL1A 4.6 NM_000575.3 JUN 2.3 NM_002228.3 PPM1K 2.1 NM_152542.2 SERPINB2 2.5 NM_002575.1 SLC25A25 2.6 NM_052901.2 SNORA41 2.5 NR_002590.1 TAGAP 3.0 NM_054114.3 TNFSF9 2.2 NM_003811.2 TRK1 2.5 NR_001449.1 TSC22D3 2.2 NM_198057.2 ZNF571 2.4 NM_016536.3 #Fold change annotated in GM-MΦ and M-MΦ columns indicates the difference in gene expression between untreated and C5a-treated macrophages.

Given that the EC50 values of C5a on GM-MΦ and M-MΦ in the functional assays conducted such as ERK phosphorylation, calcium release and macrophage migration were so similar, it was not surprising that the magnitude of genes commonly mediated by C5a were similar as well (Table 3.3).

122 Table 3.3. Genes commonly upregulated by 30 nM C5a in GM-MΦ and M-MΦ after 0.5 h were expressed as fold changes.

Gene GM-MΦ# M-MΦ# Accession ATF3 2.8 2.6 NM_001040619.1 AXUD1 2.6 2.4 NM_033027.2 BTG2 5.3 3.9 NM_006763.2 CCL3 3.2 2.9 NM_002983.1 CCL3L1 3.5 3.5 NM_021006.4 CCL3L3 3.1 2.7 NM_001001437.3 CXCL2 3.5 2.4 NM_002089.3 CYP4B1 3.7 2.2 NM_000779.2 DUSP1 8.4 5.7 NM_004417.2 DUSP6 3.9 3.1 NM_001946.2 EGR1 13.2 13.3 NM_001964.2 EGR2 3.5 3.7 NM_000399.2 EGR3 4.9 5.5 NM_004430.2 FOS 14.4 24.5 NM_005252.2 FOSB 395.7 201.0 NM_006732.1 IER2 4.3 3.2 NM_004907.2 IER3 3.2 2.1 NM_003897.3 IL8 3.7 2.5 NM_000584.2 JUNB 3.5 2.6 NM_002229.2 KLF2 5.6 3.8 NM_016270.2 KLF4 3.5 3.8 NM_004235.3 KLF6 2.2 3.3 NM_001008490.1 MAFF 2.9 2.5 NM_012323.2 NFKBIZ 3.9 2.9 NM_001005474.1 NR4A2 5.3 6.9 NM_006186.2 PHLDA1 3.3 2.5 NM_007350.3 PMAIP1 2.2 3.2 NM_021127.1 PTGS2 4.3 3.4 NM_000963.1 RGS1 4.6 6.3 NM_002922.3 RGS2 2.1 2.2 NM_002923.1 SOCS3 3.1 3.9 NM_003955.3 TNF 4.5 3.6 NM_000594.2 ZFP36 4.4 4.5 NM_003407.2 #Fold change annotated in GM-MΦ and M-MΦ columns indicates the difference in gene expression between untreated and C5a-treated macrophages.

Interestingly, out of all the genes upregulated by C5a in both GM-MΦ and M- MΦ, 33 of them were commonly upregulated (Figure 3.11), which translated into 62% - 83% of the genes identically mediated between GM-MΦ and M-MΦ. To analyze the functional effects of the genes differentially regulated by C5a in GM-MΦ or in M-MΦ, the entire list of genes (Table 3.2 and Table 3.3) was uploaded into IPA in order to generate network, functional and pathway analyses.

123

Figure 3.11. Venn diagram comparison of genes upregulated by 30 nM C5a for 0.5 h on GM-MΦ versus M-MΦ. Circles represented the ANOVA analysis of C5a-regulated genes by GM-MΦ (blue) versus M-MΦ (yellow). Numbers inside each compartment represented the number of transcripts that were significant (≥ 2-fold change, p < 0.05) for that effect. The interacting region represents genes consistently upregulated by both population of human macrophages.

The pathway analysis converted the list of C5a-upregulated genes and each annotated gene was mapped to its corresponding gene object into a set of relevant networks based on the Ingenuity Pathways Knowledge Base (IPKB). The networks were generated from the focus molecules (list of differentially regulated genes uploaded into IPA) and then focus molecules with the most interactions to other focus molecules were then connected together to form a network. Non-focus molecules from IPKB were then added to form a network based upon molecular function consisting of 35 molecules. These genetic networks were then sorted based on the score, which reflects the probability that a collection of genes equal to or greater than the number in a network that could be achieved by chance alone. The network with the highest score for C5a-treated macrophages was 30 for GM-MΦ (Table 3.4) and 34 for M-MΦ (Table 3.5), which composed of 14 genes and 15 genes respectively. The top ranking diseases and functions commonly between GM-MΦ and M-MΦ overlapped with major pathways involved in neurological disease, cancer, cell death and survival and the gene network corresponding to those pathways were depicted in Figure 3.12.

124 Table 3.4. Top 5 network functions and associated diseases that were upregulated by C5a in GM-MΦ as analyzed by IPA

Top Diseases Focus Score Molecules in Network# and Functions Molecules Adaptor protein 1, ATF3, CaMKII, Caspase 3/7, Neurological Cg, Creb, DUSP1, DUSP2, DUSP6, EGR1, Disease, EGR3, Eotaxin, ERK1/2, FOSB, FSH, GC-GCR Cancer, Cell dimer, Hdac, IER2, JINK1/2, JUN, JUNB, 30 14 Death and KLF6, LDL, Lh, MAP2K1/2, Mek, PDGF BB, Survival PHLDA1, PI3K (family), PMAIP1, PTPase, SERPINB2, SWI-SNF, TCF, thyroid hormone receptor Inflammatory Akt, Calcineurin protein(s), CCL3, CCL8, Response, CCL3L1/CCL3L3, CXCL2, cyclooxygenase, Cellular elastase, Fc gamma receptor, Fcer1, Gm-csf, Movement, Growth hormone, HLA-DR, IER3, Ifn, Ifn Hematological 20 10 gamma, Ige, Ikb, IKK (complex), IL1, IL17a System dimer, IL1A, lymphotoxin-alpha1-beta2, NFkB Development (family), NfkB-RelA, NfkB1-RelA, NFKBIZ, and Function Nr1h, Oas, SAA, SOCS3, Tlr, Tnf receptor, TNFSF9, TSC22D3 AGTPBP1, ATF4, ATP5I, beta-estradiol, BMP2, BTBD10, CCDC138, CDK5, CREB1, CREB5, Cellular CSRNP1, CYP4B1, DDIT3, DLEU1, ELK3, Movement, ETNK1, FGF4, FOSL1, HTATIP2, JUNB, Cell Cycle, 20 10 JUND, MAFB, miR-17-5p (and other miRNAs Gene w/seed AAAGUGC), miR-450b-3p (and other Expression miRNAs w/seed UGGGAUC), NUCB2, PPM1K,

SERPINB2, SLC25A25, TAGAP, TBP, TGM2, TMSB10/TMSB4X, tretinoin, UBC, ZNF571 ADCY, ADRB, BCR (complex), BTG2, calpain, Cell-To-Cell Cbp/p300, CD69, Cyclin E, ERK, estrogen Signaling and receptor, IFN Beta, IgG, Igm, IL8, IL12 Interaction, (complex), IL12 (family), Immunoglobulin, Cellular Interferon alpha, JUN/JUNB/JUND, KLF4, Growth and 14 8 MAFF, MEF2, MHC CLASS I (family), NADPH Proliferation, oxidase, Nfat (family), P38 MAPK, Pro- Cellular inflammatory Cytokine, PTGS2, Rap1, RGS2, Development sphingomyelinase, STAT5a/b, TMSB4, TSH,

ZFP36 aldosterone, amino acids, CaMKII, CCR4, CDK5, Cell Signaling, chemokine, CHRM3, CHRM4, Ck2, CRCP, Molecular DDIT3, DISC1, Endothelin, EPO, ethanol, Transport, GNRH1, GNRH, Gpcr, GPR183, Gsk3, Histone Vitamin and 10 6 h3, Histone h4, IGFBPL1, Mapk, MC4R, miR- Mineral 1231-3p (and other miRNAs w/seed GCCCUGU), Metabolism MOS, NPS, Pka, PLC, PTH, RGS1, SFTPA1,

STAT, TNF #Focus molecules were represented in the network as bolded genes.

125 Table 3.5. Top 5 network functions and associated diseases that were upregulated by C5a in M-MΦ as analyzed by IPA

Top Diseases Focus Score Molecules in Network# and Functions Molecules Adaptor protein 1, Ap1, ATF3, CaMKII, Caspase Neurological 3/7, Cg, Creb, DUSP1, DUSP6, EGR1, EGR3, Disease, ERK1/2, FOSB, FSH, Hdac, IER2, Cancer, Cell 34 15 Immunoglobulin, JUNB, KLF4, KLF6, LDL, Lh, Death and MAP2K1/2, MCL1, Mek, MIR101, Pdgf Survival (complex), PDGF BB, PHLDA1, PMAIP1,

PTGS2, PTPase, RGS2, TCF, Vegf Inflammatory Akt, CCL3, CCL14, CCL3L1/CCL3L3, Response, CXCL2, DEFB104A/DEFB104B, elastase, Fc Cellular gamma receptor, Fcer1, Gm-csf, Growth Movement, hormone, HLA-DR, IER3, Ifn, IFN Beta,Ifn Hematological 21 10 gamma, Ige, IgG, IL1, IL8, IL12 (complex), System Interferon alpha, KLF2, NFkB (family), NfkB- Development RelA, NfkB1-RelA, Notch, Nr1h, Oas, PI3K and Function (family), Pro-inflammatory Cytokine, SAA, SOCS3, Tlr, ZFP36 ADCY, ANKS1A, ATF4, ATF7, BMP2, CAPN2, Cancer, CDK5, CSRNP1, CYP4B1, DAB2, DDIT3, Cellular FAM50A, FOSL1, GRM5, JUNB, JUND, Development, KCNIP3, KDM2A, KRT15, MC1R, miR-1207-5p Cellular 11 6 (and other miRNAs w/seed GGCAGGG), Growth and NBPF10 (includes others), NFASC, NMI, PKP2, Proliferation PRUNE2, PSORS1C1, RAP1GAP, SLC19A1, SUPT6H, TMSB10/TMSB4X, TPR, tretinoin, TRPV4, UBC 26s Proteasome, amino acids, caspase, CDK5, Cell chemokine, Ck2, Focal adhesion kinase, GNRH1, Morphology, Gsk3, HISTONE, Histone h3, HMGN3, Hsp70, Cell-To-Cell IL12 (family), Mapk, Mmp, MMP23B, MOS, Signaling and 9 5 Neurotrophin, NFkB (complex), NFKBIZ, Interaction, NR4A2, PDGF-AA, PI3K (complex), Pka, PLC, Inflammatory Ras homolog, RGS1, RNA polymerase II, SNCG, Response STAT, TEC/BTK/ITK/TXK/BMX, TFIIA, TNF,

Ubiquitin AMPK, BCR (complex), BTG2, Calcineurin Cell Cycle, protein(s), Cbp/p300, CD3, Collagen type I, Cellular Cyclin E, cyclooxygenase, EGR2, ERK, estrogen Development, receptor, FOS, Histone h4, Hsp27, IKK 6 4 Embryonic (complex), Integrin, Jnk, JUN/JUNB/JUND, Development MAFF, NADPH oxidase, Nfat (family), Nos, P38 MAPK, Pkc(s), Rac, Rap1, Ras, Rb, Rock, Sapk, STAT5a/b, TCR, Tgf beta, TSH #Focus molecules were represented in the network as bolded genes.

The functional pathway mapping analysis also suggested that C5a activation in human macrophages is also involved in a variety of cellular and pathological conditions (score ≥ 10), such as inflammatory responses, cellular movement, cellular

126 development, haematological system development and function, cellular growth and proliferation based on matching C5a-induced gene expression to pathways.

Figure 3.12. Functionally related gene network with highest score constructed from genes commonly upregulated by 30 nM C5a on GM-MΦ and M-MΦ. Genes commonly upregulated by 30 nM C5a on GM-MΦ and M-MΦ were uploaded into Ingenuity Pathway Analysis (IPA) software. Nodes displayed in red represented genes upregulated by C5a and the magnitude of expression was reflected by the intensity of the colour with darker red indicating higher levels of expression.

The magnitude of expression was reflected by the intensity of the colour, with darker red indicating higher levels of expression. Most of the C5a-upregulated genes detected by microarray were not more than 25-fold changed except for the expression of FOSB gene, which was 396 and 201 fold altered by C5a in GM-MΦ and M-MΦ respectively. cDNA microarray provides an unprecedented capacity for whole genome profiling. However, the quality of gene expression data obtained from

127 microarray can vary greatly with platforms and procedures used (30). Although the list of genes differentially expressed between control and treated samples detected in the microarray experiment was already validated by quantitative real-time PCR previously in Chapter 2, the magnitude of fold change in the FOSB gene appeared to be oddly high, unlike the FOS gene (~15- to 25-fold change) which was the closest related gene. As suspected, the expression of FOSB was found to be out of the sensitivity range when validated by real-time PCR (Figure 3.13). Although the fold change values for FOSB gene were inaccurate, the differences between GM-MΦ and M-MΦ were consistent showing higher expression in GM-MΦ against M-MΦ across two platforms.

Figure 3.13. Correlation between microarray and quantitative real-time PCR FOSB gene expression upregulated by 30 nM C5a on GM-MΦ and M-MΦ. The expression of FOSB gene induced by 30 nM C5a for 30 min in GM-MΦ (black) and M-MΦ (white) were detected by microarray and quantitative real-time PCR. Error bars are means ± SEM of three-four independent experiments (n=4, microarray; n=3, PCR). *p < 0.05 and **p < 0.01 by student t-test.

The transcriptional program stimulated by C5a included several genes whose products are proinflammatory such as chemokines and cytokines (CCL3, CCL3L1, CCL3L3, CXCL2, IL8, PTGS2 and TNF). C5a also stimulated the expression of a number of transcription factors (ATF3, EGR1, EGR2, EGR3, FOS, FOSB and JUNB) that amplified immune responses and were also likely to have a positive feedback

128 loop resulting in the stimulation of expression of themselves and other proinflammatory genes. FOSB and FOS genes were highly upregulated in human macrophages (GM-MΦ and M-MΦ) when stimulated with C5a. The FOS gene family consists of 4 members, namely FOS, FOSB, FOSL1 and FOSL2. Besides inflammation, FOS proteins had also been implicated as a regulator of cell proliferation, differentiation and transformation by dimerising with proteins of the JUN family to form the transcription factor complex activating protein-1 (AP-1) (31). C5a activation in human macrophages upregulated genes such as ZFP36 can interfere with TNF production by destabilizing its mRNA (32), and also cytokine signaling suppressors (SOCS3, NFKBIZ, DUSP1 and DUSP6) (33, 34). These genes were simultaneously activated by C5a to prevent excessive damage that could be deleterious by dampening down proinflammatory responses and restore homeostasis. In contrast to the number of genes upregulated by C5a, the number of genes downregulated by C5a was very modest in GM-MΦ with only 1 gene (CLEC1B, -2.2 fold change), and in M-MΦ with 5 genes (SNORD99, -3.6 fold change; BTRC, -2.3 fold change; SNORA33, -2.2 fold change; CDC14C, -2.1 fold change and PTPRU, -2 fold change). Therefore, a separate microarray experiment was conducted to overview the temporal profile of C5a-induced gene expression in macrophages extending the observation beyond 0.5 h stimulation.

3.5.7 Microarray experiment 2: Temporal profile of genes mediated by C5a on M-MΦ

The temporal study with different time points (0, 0.5, 6 and 18 h) was conducted in M-MΦ only, not in GM-MΦ (since majority of the upregulated genes were commonly expressed between GM-MΦ and M-MΦ) from three different healthy blood donors. Moreover, macrophages can undergo dynamic changes during different phases of the immune response. For example, in physiological conditions, proinflammatory macrophages mediate tissue damage and initiate inflammatory responses but at the later stage when the initiate inflammatory wave subsides, the antiinflammatory macrophages will begin wound healing and initiate haemostasis (12, 35). These temporal changes on macrophages have been observed in several other

129 pathological conditions and diseases (36-38). As such, it was suspected that C5a would exert a greater effect on M-MΦ than GM-MΦ at later time points. There were a total of 147 genes significantly regulated by 30 nM C5a on M- MΦ (Figure 3.14). As expected, at 0.5 h time point post C5a stimulation, all genes induced were upregulated in M-MΦ. At 6 h time point, a number of genes that were not responsive to C5a stimulation at earlier time point (0.5 h) were either downregulated or upregulated. Majority of the genes that were upregulated by C5a at 0.5 h, returned to baseline expression by 6 h time point. These differentially regulated genes can be further categorised into clusters according to their temporal profile, with cluster 1, 2 and 3 featuring genes that were upregulated by 30 nM C5a at 0.5, 6 and 18 h respectively and cluster 4 featuring genes that were downregulated by 30 nM C5a at 6 h (Figure 3.15).

Figure 3.14. Temporal profile of 147 genes differentially regulated by 30 nM on M-MΦ. Genes were represented by colors according to their temporal profile (yellow: upreguated genes after 0.5 h, orange: upregulated genes after 6 h and blue: downregulated after 6 h). All genes represented in the profile plot were at least 2-fold change and p < 0.05 by ANOVA.

130

Figure 3.15. Four clusters of genes were categorised based on their responses to 30 nM C5a treatments and the duration the event occurred. Cluster 1, 2 and 3 featured genes that were upregulated ≥ 2-fold change at 0.5 h (yellow), 6 h (orange) and 18 h (pale yellow) respectively. Cluster 4 featured genes that were downregulated ≥ 2-fold change at 6 h. All genes are p < 0.05 by ANOVA.

3.5.7.1 Inflammatory genes regulated by C5a

C5a plays an essential role in inflammation and innate immunity against infectious disease and Table 3.6 lists inflammation-related genes regulated by C5a. C5a induces acute inflammatory responses in a variety of disease models (39, 40). Important inflammatory mediators such as IL1B (6 fold, 0.5 h), PTGS2 (3.3-fold, 0.5 h) and TNF (2.6 fold, 0.5 h) were upregulated as the result of C5a stimulation. C5a anaphylatoxin is a potent chemotaxin and upregulated chemokines such as CCL8 (2.5 fold, 0.5 h), CCL3 (2.9-fold, 0.5 h), CCL3L1 (2.6-fold, 0.5 h), CCL4L2 (2.5-fold, 0.5 h), CCL3L3 (2.4-fold, 0.5 h), CCL7 (6.1-fold, 6 h), CCL2 (2.2-fold, 0.5 h and 2.4-fold, 6 h) while CXCR4 was downregulated (-2.4-fold, 6 h) in the microarray experiment. IL1B mRNA expression peaks at 0.5 h and was quickly reduced to basal expression level when IL1RN was upregulated (2-fold, 6 h), which seems to suggest a direct

131 feedback mechanism in M-MΦ to prevent hyper-stimulation by C5a via IL1B expression.

Table 3.6. Inflammatory genes upregulated (black) or downregulated (red) by 30 nM C5a.

Gene Name 0.5h 6h 18h Interleukin 1, beta 6.0 1.4 (-1.3) Prostaglandin-endoperoxide synthase 2 3.3 1.2 1.0 Heparin-binding EGF-like growth factor 2.5 (-1.1) (-1.4) Chemokine (C-C motif) ligand 8 2.5 1.3 (-1.1) Chemokine (C-C motif) ligand 3 2.9 1.0 (-1.4) Tumor necrosis factor 2.6 1.4 (-1.0) Chemokine (C-C motif) ligand 3-like 1 2.6 (-1.1) (-1.1) Chemokine (C-C motif) ligand 4-like 2 2.5 1.2 (-1.4) Zinc finger protein 36 2.4 (-1.2) (-1.2) Chemokine (C-C motif) ligand 3-like 3 2.4 (-1.0) (-1.3) Chemokine (C-C motif) ligand 7 1.3 6.1 (-1.0) Interleukin 1 receptor antagonist 1.1 2.0 (-0.2) Tumor necrosis factor superfamily, member 21 (-1.2) 2.6 (-1.0) Heparan sulfate (glucosamine) 3-O-sulfotransferase 1 1.0 2.4 1.1 Chemokine (C-C motif) ligand 2 2.2 2.4 1.2 Heparanse (-1.1) (-4.7) (-1.4) Toll-like receptor 7 (-1.1) (-2.9) (-1.2) Leptin (-1.2) (-2.2) (-1.5) Arachidonate 15-lipoxygenase, type B 1.0 (-2.6) (-1.7) Transient receptor potential cation channel, subfamily M, member 2 1.1 (-2.5) (-1.1) Chemokine (C-X-C motif) receptor 4 (-1.4) (-2.4) (-1.5) Hydrogen voltage-gated channel 1 (-1.2) (-2.4) (-1.1) Toll-like receptor 8 (-1.1) (-2.2) (-1.0)

3.5.7.2 Metabolism- or biosynthesis-related genes regulated by C5a

The immune system requires a steady intake of nutrients to maintain proper functions. The relationship between nutrition and immunity is complex. From this microarray analysis, more than 15 metabolism- or biosynthesis-related genes were significantly regulated as the result of C5a stimulation (Table 3.7). C5a has not been shown to play a direct role in metabolic homeostasis before. Interestingly, fatty acid transporters (FABP5 and FABP5L2) and amino acid transporters (SLC7A1 and SLC7A5) were found to be more than 2-fold upregulated in M-MΦ stimulated with C5a and this may be an important observation in relation to possible effects of C5a on

132 adiposity. On the other hand, sugar transporter (SLC45A4) and iron transporter (SLC40A1) genes were more than 2-fold downregulated. CTSL1 and CTPS are important enzymes for intracellular protein catabolism and biosynthesis of phospholipids and nucleic acids respectively. CTSL1 and CTPS mRNA levels were upregulated (more than 2-fold) by C5a stimulation in HMDM, while the genes encoding for enzymes responsible for lipid metabolism (PPAP2B, ALOX15B and LYPLAL1) and protein metabolism (DPEP2) were downregulated (more than 2-fold).

Table 3.7. Metabolism- or biosynthesis-related genes upregulated (black) or downregulated (red) by 30 nM C5a.

Gene Name 0.5h 6h 18h Solute carrier family 7, member 1 (-1.1) 2.5 (-1.1) Solute carrier family 7, member 5 1.1 2.3 (-1.1) Fatty acid binding protein 5-like 2 1.2 2.1 (-1.3) Fatty acid binding protein 5 1.1 2.2 (-1.2) Cathepsin L1 1.0 2.2 1.1 CTP synthase 1.0 2.0 1.1 Leptin (-1.2) (-2.2) (-1.5) Chromosome 5 open reading frame 13 1.0 (-2.7) (-1.1) Arachidonate 15-lipoxygenase, type B 1.0 (-2.6) (-1.7) Phosphatidic acid phosphatase type 2B (-1.2) (-2.7) (-1.4) Solute carrier family 45, member 4 (-1.1) (-2.5) (-1.2) Transient receptor potential cation channel, subfamily M, member 2 1.1 (-2.5) (-1.1) Solute carrier family 40, member 1 (-1.0) (-2.2) (-1.2) Lysophospholipase-like 1 (-1.1) (-2.1) (-1.1) Dipeptidase 2 (-1.2) (-2.0) (-1.3) Thymidylate synthetase (-1.1) (-1.1) 2.2

3.5.7.3 Signal transduction or transcription related genes regulated by C5a

Genes associated with signal transductions and transcriptions are listed in Table 3.8. Epidermal growth factor response genes, EGR3 (7.3-fold), EGR1 (4.3- fold) and EGR2 (4.2-fold) were upregulated in M-MΦ after stimulating with C5a for 0.5 h. The inducible zinc finger transcription factors EGR1, EGR2 and EGR3, participate in the transcriptional regulation of genes in controlling biological rhythm, by activating the transcription of target genes whose products are required for mitogenesis and differentiation (41, 42). Stress response gene, ATF3, was also upregulated (5.3-fold) in M-MΦ stimulated with C5a. These early C5a-induced

133 responses were quickly suppressed back to basal expression levels after 6 h. EGR1 and ATF3 are immediate early response genes that are associated with many inflammatory conditions and can be activated after C5a stimulation (41, 43).

Table 3.8. Signal transduction- or transcription-related genes upregulated (black) or downregulated (red) by 30 nM C5a.

Gene Name 0.5h 6h 18h Early growth response 3 7.3 1.1 (-1.0) Pleckstrin homology-like domain, family A, member 1 5.7 2.2 (-1.4) FBJ murine osteosarcoma viral oncogene homolog B 5.9 (-1.3) (-1.3) Activating transcription factor 3 5.3 (-1.3) (-1.5) Early growth response 1 4.3 1.0 (-1.5) Early growth response 2 4.2 (-1.1) (-1.9) Dual specificity phosphatase 5 3.0 1.4 (-1.1) Dual specificity phosphatase 6 2.4 1.9 1.1 BTG family, member 2 2.8 (-1.2) (-1.5) Zinc finger protein 36 2.4 (-1.2) (-1.2) Kruppel-like factor 6 2.2 (-1.1) (-1.1) Polymerase I and transcript release factor (-1.1) 2.4 (-1.2) Megakaryocyte-associated tyrosine kinase 1.2 2.2 1.1 Filamin B, beta (-1.2) 2.6 (-1.4) Transmembrane protein 158 1.0 2.4 (-1.3) MARCKS-like 1 1.0 2.4 1.1 DEAD (Asp-Glu-Ala-Asp) box polypeptide 21 1.0 2.0 (-1.2) P8 protein (-1.1) (-2.4) (-1.3) D site of albumin promotor binding protein (-1.0) (-2.4) 1.3 SLIT-ROBO Rho GTPase activating protein 3 (-1.0) (-2.4) (-1.3) Pellino homolog 2 (-1.2) (-2.2) (-1.1) AT rich interactive domain 3A 1.1 (-2.1) (-1.1) Zinc finger protein 467 1.0 (-2.1) (-1.2) Zinc finger protein 581 1.0 (-2.1) (-1.2) Ubiquitin-like with PHD and ring finger domains 1 (-1.0) 1.2 2.5

3.5.7.4 Cell cycle-related genes regulated by C5a

Agonists such as TNF, LPS and C5a can activate pro-apoptotic and anti- apoptotic genes in endothelial cells (7, 44, 45). C5a induces proliferation, migration and vessel formation in endothelial cells (46). More than 20 cell cycle-related genes was found to be mediated by C5a (Table 3.9). DUSP5 (3-fold, 0.5 h) and DUSP6 (2.4-fold, 0.5 h) were upregulated in M-MΦ by C5a stimulation. The DUSP family negatively regulate members of the mitogen-activated protein (MAP) kinase

134 superfamily (MAPK/ERK, SAPK/JNK, p38), which are associated with cellular proliferation and differentiation.

Table 3.9. Cell cycle-related genes upregulated (black) or downregulated (red) by 30 nM C5a.

Gene Name 0.5h 6h 18h Pleckstrin homology-like domain, family A, member 1 5.7 2.2 (-1.4) FBJ murine osteosarcoma viral oncogene homolog B 5.9 (-1.3) (-1.3) Early growth response 1 4.3 1.0 (-1.5) Dual specificity phosphatase 5 3.0 1.4 (-1.1) AXIN1 up-regulated 1 3.1 1.1 (-1.1) Dual specificity phosphatase 6 2.4 1.9 1.1 Tumor necrosis factor 2.6 1.4 (-1.0) CD83 molecule 2.4 1.0 (-1.4) Myeloid-associated differentiation marker 2.5 1.7 (-1.2) Kruppel-like factor 4 2.4 (-1.5) (-1.2) Myeloid cell leukemia sequence 1 2.3 1.0 (-1.2) GDNF family receptor alpha 2 1.0 4.0 (-1.1) CD276 molecule 1.1 2.5 (-1.2) Tumor necrosis factor superfamily, member 21 (-1.2) 2.6 (-1.0) Lamin A/C 1.4 2.5 1.0 Cathepsin L1 1.0 2.2 1.1 CTP synthase 1.0 2.0 1.1 Matrix metallopeptidase 19 1.5 2.2 1.5 Bcl2 modifying factor (-1.3) (-3.1) (-1.5) Chromosome 5 open reading frame 13 1.0 (-2.7) (-1.1) HtrA serine peptidase 4 (-1.2) (-2.6) (-1.7) Yippee-like 3 1.2 (-2.6) (-1.0) AT rich interactive domain 3A 1.1 (-2.1) (-1.1) B-cell CLL/lymphoma 11A (-1.2) (-2.2) (-1.5) Fas apoptotic inhibitory molecule 3 (-1.2) (-1.6) (-2.0)

FOSB was also upregulated (5.9 fold, 0.5 h) in M-MΦ when stimulated with C5a. Interestingly, CD83 was upregulated (2.4-fold) in C5a-stimulated HMDM. CD83 is one of the best-known and characterised maturation markers for human dendritic cells (DCs). The CD83 protein is preformed intracellularly in monocytes, macrophages and dendritic cells and is stably expressed on activated dendritic cells (47). C5a can induce monocyte recruitment and differentiation into dendritic cells by TNF and prostaglandin E2-dependent mechanisms (48). It is however not known if C5a differentiates macrophages to dendritic cells. CD276 is a costimulatory B7 molecule that signals through CD28 and may participate in the priming of T cells (49). Unstimulated antigen presenting cells are largely B7 negative. CD276 is upregulated

135 when DCs macrophages, B cells and Langerhans’ cells are activated. CD276 was upregulated (2.5-fold) in M-MΦ after stimulating with C5a for 6 h.

3.5.7.5 Adhesion- or migration-related genes regulated by C5a

Chemotaxis plays a major role in macrophage accumulation at the site of an infection or tissue damage. C5a can manifest chemotactic activity in macrophages, and some adhesion- or migration-related genes regulated by C5a in M-MΦ are listed in Table 3.10. FAM129B (3.7-fold), TIMP3 (2.2-fold) and MMP19 (2.2-fold) were upregulated in M-MΦ when stimulated with C5a.

Table 3.10. Adhesion- or migration-related genes upregulated (black) or downregulated (red) by 30 nM C5a.

Gene Name 0.5h 6h 18h Family with sequence similarity 129, member B 1.1 3.7 (-1.0) TIMP metallopeptidase inhibitor 3 1.2 2.2 (-1.5) Neuropilin 1 (-1.2) 2.3 1.1 Matrix metallopeptidase 19 1.5 2.2 1.5 Purinergic receptor P2Y, G-protein coupled, 8 (-1.1) (-2.9) (-1.5) Plasminogen activator, urokinase 1.0 (-2.4) (-1.2) Adhesion molecule, interacts with CXADR antigen 1 (-1.1) (-2.1) (-1.6) Polycystic kidney disease 2-like 1 (-1.1) (-2.1) (-2.1)

Although FAM129B did not regulate cell growth and division, the deletion of FAM129B in melanoma cells significantly impaired B-raf/MEK/Erk-dependent invasion into extracellular matrix. MMP19 encodes for the metalloproteinase-19 enzyme that was found to be involved in the breakdown of extracellular matrix, whereas TIMP3 encodes for the tissue inhibitor of metalloproteinase-3. It appeared that C5a might direct cellular movement to the target site by regulating the expression of MMP19 and TIMP3 in M-MΦ.

3.5.7.6 Validation by real-time PCR

Of all 147 genes (Appendix III), 18 genes (IL1B, TNF, CCL3, PTGS2, TLR7, TLR8, FABP5, LEP, FOSB, ATF3, EGR1, DUSP5, DUSP6, CD83, CD276, CTSL1, TIMP3 and MMP19) were selected for validation by real-time PCR. These genes

136 represented genes activated by C5a and are strongly implicated in inflammation, metabolism, signalling and chemotaxis in M-MΦ (Figure 3.16).

Figure 3.16. C5a can affect genes involved in inflammation, metabolism, signaling, chemotaxis and cell cycle on M-MΦ.

The directions of fold change were consistent in 17 out of 18 genes, with LEP gene being the only exception, validated with PCR against the result in the microarray (Figure 3.17). Although the magnitude of fold change for selected genes were mostly higher in real-time PCR experiments compared to that of microarray, this is not surprising as resolution is higher and more sensitive in real-time PCR. C5a has been long known to exert potent proinflammatory effects on immune cells. Here, genes that were drastically regulated by a single exposure of 30 nM C5a within the first 6 h returned to basal state after 18 h. It is expected that C5a can attribute very different profiles depending on whether it is an acute versus chronic exposure to C5a. Human macrophages (GM-MΦ and M-MΦ) will undergo a single exposure of C5a at 30 nM for 30 min in order to investigate the acute responses of C5a on primary human macrophages.

137 IL1B TNF CCL3 30 4 10 e e e g g g 3 8 n n 20 n a a a 2 6 h h 10 h 1 4 C C C

d 0 2 d d l l 0 l o o o -1 0 F F -10 F -2 -2 Microarray qRT-PCR Microarray qRT-PCR Microarray qRT-PCR

PTGS2 15 TLR7 TLR8

e 2 2 e e g g g

n 10 1 n n a

a 0 a

h 0 5 h h C C C

-1 -2 d d d l

0 l l -2 o o o F -5 F -4 F -3 Microarray qRT-PCR Microarray qRT-PCR Microarray qRT-PCR

FABP5 FOSB 6 LEP 8

e 4 e e g g 4 g 6 n n

n 2 a a 2 a 4 h h h 0 C C C

0 2

d d d l l -2 l -2 0 o o o F F -4 F -4 -2 Microarray qRT-PCR Microarray qRT-PCR Microarray qRT-PCR

ATF3 EGR1 DUSP5 15 10 15 e e e g g g

n 10 n n 10

a a 5 a

h 5 h h 5 C C C 0 d d d

l 0 l l 0 o o o

F -5 F -5 F -5 Microarray qRT-PCR Microarray qRT-PCR Microarray qRT-PCR

DUSP6 CD83 CD276 6 4 4 e e e g g g n n n a a a 4 2 2 h h h C C C

2 0 0 d d d l l l o o o F F F 0 -2 -2 Microarray qRT-PCR Microarray qRT-PCR Microarray qRT-PCR

TIMP3 MMP19 CTSL1 4 4 e

8 e e g g g

6 n n

n 3 a

2 a a

4 h h h 2 C C C

2 0 d d d l l 0 l 1 o o o F F -2 -2 F 0 Microarray qRT-PCR Microarray qRT-PCR Microarray qRT-PCR

Figure 3.17. Correlation between microarray and real-time PCR expression. The set of column obtained by microarray (black) and real-time PCR (white), with three columns representing the duration of 30 nM C5a treatment in M-MΦ (1st column, 0.5 h; 2nd column, 6 h; 3rd column, 18 h). The dotted lines across the graph represented the 2-fold change threshold. Error bars represent mean ± SEM of three donors (n=3).

138 3.6 Discussion

The complement system can be activated via three separate pathways to promote inflammatory responses, eliminate invading pathogens and enhance immune responses. Under physiological conditions, the concentrations of C5a are relatively low, promoting beneficial effects on the priming of neutrophils and monocytes as well as activating endothelial cells to support efficient innate immune responses against infection (6). However, C5a is like a double-edge sword in also causing immune paralysis of phagocytic cells due to excessive complement activation resulting in an overproduction of C5a (50). High levels of C5a can be detected in the serum of patients suffering with sepsis and can reach as high as 100 nM (14). Although the local concentration of C5a has not been determined, it is expected to be higher at the sites of infection and inflammation. Human C5a above 100 nM concentration has been show to induce apoptosis in freshly isolated thymocytes (15). Other lymphoid cells such as B and T cells had similar apoptotic responses when treated with such a concentration of C5a (16). With this in mind, C5a treatment concentrations for human primary macrophages (GM-MΦ and M-MΦ) had been optimized to be 30 nM, to investigate how C5a would physiologically exert its biological effects on these macrophages without activating death signals. C5a was previously reported to stimulate the synthesis and release of proinflammatory cytokines such as TNF and IL1β from human leukocytes, not specifically human macrophages (51). A common mistake in studying mRNA expression is that gene expression is informative in the prediction of protein expression, but this is more often than not the case (18, 52) and it is certainly not the case in this study here on human macrophages differentiated with GM-CSF or M-CSF from human CD14+ monocytes. C5a was found to readily upregulate IL1B and TNF mRNA expression in both human macrophages (GM-MΦ and M-MΦ) measured by microarray and PCR. However, their corresponding proteins were not detectable in the culture media of C5a-treated macrophages (data not shown). This observation is consistent with work reported by others (53, 54). C5a provided a transcriptional (gene) but not translational (protein), signal for IL1β and TNF in human peripheral blood mononuclear cells (54). So not surprisingly, majority of the genes regulated by

139 C5a in human macrophages were involved in the gene transcriptional network (Figure 3.12). Transcriptional regulation plays an important role in both differentiation and homeostasis (55, 56). Two of the major classes of regulatory elements are involved in transcriptional regulation by extracellular signals, the AP-1/TRE (57) and the ATF/CRE (58). The Fos gene family consists of 4 members namely, FOS, FOSB, FOSL1 and FOSL2. These genes encoded for the leucine zipper proteins that can dimerize with proteins of the JUN family (c-Jun, JunB and JunD) to form the transcription factor complex AP-1 or phorbol 12-O-tetradecanoate 13-actate (TPA) responsive element (TRE). The ATF/CRE element (TGACGTCA) was defined as the activating transcription factor (ATF) binding site or the cAMP responsive element (CRE). The ATF/CRE site is recognised by ATF or CRE binding protein (CREB) (59). ATF3 gene encodes a member of the activating transcription factor/cAMP responsive element binding protein (ATF/CREB) family, which is usually induced by physiological stresses. The Fos/Jun and ATF/CREB families were first thought to have distinct sets of transcription factors that share the same leucine zipper motif, and have different DNA binding specificities. This theory was later proven to be wrong (60). Members of the ATF/CREB family can form selective cross-family heterodimers with members of the Fos/Jun family to have distinguishable binding specificities in response environmental stimuli. This diversity increases the complexity in transcriptional regulation, which C5a can evidently manipulate to maintain balance between inflammatory responses and homeostasis. Interestingly, human umbilical vein endothelial cells could also activate EGR1 and ATF3 when stimulated with C5a but only after 2 h (7), as opposed to 0.5 h in human macrophages discovered in this study. Perhaps it is the different chronological pattern of gene activation observed in different cell types upon C5a stimulation that is key to the differential outcome of C5a-mediated effects across different cell types. The toll-like receptors and complement related proteins are the key innate defence systems in the body and the two systems were previously thought to function as separate entities. However recently, Hajishengallis et al reviewed the complement system and noted that TLR crosstalk reinforces innate immunity or regulates excessive inflammation, through synergistic or antagonistic interactions (61). This crosstalk is still not well investigated and there are many questions about the possible interplay between complement activation and TLR activation. Surprisingly, TLR7 and

140 TLR8 genes were downregulated in M-MΦ when stimulated with C5a. Chapter 5 has described the differential crosstalk observed between C5aR and TLR4 on human monocytes versus macrophages. It might be worth investigating if this TLR4/C5aR crosstalk could be extended further to TLR7 or TLR8. Understanding this interplay and its relationship to innate and adaptive immunity may uncover important new aspects of immunological stimulation and control as well as lead to validation of new targets for therapeutic regulation.

3.7 Conclusions

Macrophages have a ‘plastic’ gene expression phenotype that can change rapidly depending on the microenvironment and even lead to a switching of phenotype from M1 to M2 and vice versa (62). This is supported in Chapter 2 where GM-MΦ and M-MΦ were found to be quite different in their genotypes from each other. Their genotypic differences were shown to be greater than C5a treatment in the microarray experiment as calculated by distant matrix algorithms (Figure 3.7) and it is reasonable to expect that these two macrophage subtypes with large genomic differences (Chapter 2) would respond differently when stimulated by C5a (Chapter 3). However, in this chapter it was found that genes stimulated by C5a were found to be at least 62% identical between GM-MΦ and M-MΦ. C5a was also found to exert similar effects on GM-MΦ and M-MΦ in different functional assays - calcium release, ERK phosphorylation and macrophage migration. More importantly, this study has provided important clues to the temporal role of C5a-induced gene expression in primary human macrophages. By identifying the cluster of genes activated by C5a, potential rate-limiting gene candidates may be identified for potential clinical intervention in C5a-mediated inflammatory disorders. Some of the genes found in this chapter to be highly responsive to C5a stimulation in human macrophages were used to compare three reported C5a ligands for their underlying antagonist mechanisms in Chapter 4.

141 3.8 References

1. Hugli, T. E. 1981. The structural basis for anaphylatoxin and chemotactic functions of C3a, C4a, and C5a. Crit. Rev. Immunol. 1: 321-366. 2. Erdei, A., K. Kerekes, and I. Pecht. 1997. Role of C3a and C5a in the activation of mast cells. Exp. Clin. Immunogenet. 14: 16-18. 3. Nilsson, G., M. Johnell, C. H. Hammer, H. L. Tiffany, K. Nilsson, D. D. Metcalfe, A. Siegbahn, and P. M. Murphy. 1996. C3a and C5a are chemotaxins for human mast cells and act through distinct receptors via a pertussis toxin-sensitive signal transduction pathway. J. Immunol. 157: 1693- 1698. 4. Ember, J. A., and T. E. Hugli. 1997. Complement factors and their receptors. Immunopharmacology 38: 3-15. 5. Klos, A., A. J. Tenner, K. O. Johswich, R. R. Ager, E. S. Reis, and J. Kohl. 2009. The role of the anaphylatoxins in health and disease. Mol. Immunol. 46: 2753-2766. 6. Ward, P. A. 2004. The dark side of C5a in sepsis. Nat. Rev. Immunol. 4: 133- 142. 7. Albrecht, E. A., A. M. Chinnaiyan, S. Varambally, C. Kumar-Sinha, T. R. Barrette, J. V. Sarma, and P. A. Ward. 2004. C5a-induced gene expression in human umbilical vein endothelial cells. Am. J. Pathol. 164: 849-859. 8. Haviland, D. L., R. L. McCoy, W. T. Whitehead, H. Akama, E. P. Molmenti, A. Brown, J. C. Haviland, W. C. Parks, D. H. Perlmutter, and R. A. Wetsel. 1995. Cellular expression of the C5a anaphylatoxin receptor (C5aR): demonstration of C5aR on nonmyeloid cells of the liver and lung. J. Immunol. 154: 1861-1869. 9. Rustici, G., N. Kolesnikov, M. Brandizi, T. Burdett, M. Dylag, I. Emam, A. Farne, E. Hastings, J. Ison, M. Keays, N. Kurbatova, J. Malone, R. Mani, A. Mupo, R. Pedro Pereira, E. Pilicheva, J. Rung, A. Sharma, Y. A. Tang, T. Ternent, A. Tikhonov, D. Welter, E. Williams, A. Brazma, H. Parkinson, and U. Sarkans. 2013. ArrayExpress update--trends in database growth and links to data analysis tools. Nucleic Acids Res. 41: D987-990.

142 10. Buhl, A. M., N. Avdi, G. S. Worthen, and G. L. Johnson. 1994. Mapping of the C5a receptor signal transduction network in human neutrophils. Proc. Natl. Acad. Sci. U. S. A. 91: 9190-9194. 11. Mosser, D. M., and J. P. Edwards. 2008. Exploring the full spectrum of macrophage activation. Nat. Rev. Immunol. 8: 958-969. 12. Sica, A., and A. Mantovani. 2012. Macrophage plasticity and polarization: in vivo veritas. J. Clin. Invest. 122: 787-795. 13. Akagawa, K. S. 2002. Functional heterogeneity of colony-stimulating factor- induced human monocyte-derived macrophages. Int. J. Hematol. 76: 27-34. 14. Solomkin, J. S., M. K. Jenkins, R. D. Nelson, D. Chenoweth, and R. L. Simmons. 1981. Neutrophil dysfunction in sepsis. II. Evidence for the role of complement activation products in cellular deactivation. Surgery 90: 319-327. 15. Riedemann, N. C., R. F. Guo, I. J. Laudes, K. Keller, V. J. Sarma, V. Padgaonkar, F. S. Zetoune, and P. A. Ward. 2002. C5a receptor and thymocyte apoptosis in sepsis. FASEB J. 16: 887-888. 16. Hotchkiss, R. S., K. W. Tinsley, P. E. Swanson, R. E. Schmieg, Jr., J. J. Hui, K. C. Chang, D. F. Osborne, B. D. Freeman, J. P. Cobb, T. G. Buchman, and I. E. Karl. 2001. Sepsis-induced apoptosis causes progressive profound depletion of B and CD4+ T lymphocytes in humans. J. Immunol. 166: 6952-6963. 17. Monk, P. N., A. M. Scola, P. Madala, and D. P. Fairlie. 2007. Function, structure and therapeutic potential of complement C5a receptors. Br. J. Pharmacol. 152: 429-448. 18. Guo, Y., P. Xiao, S. Lei, F. Deng, G. G. Xiao, Y. Liu, X. Chen, L. Li, S. Wu, Y. Chen, H. Jiang, L. Tan, J. Xie, X. Zhu, S. Liang, and H. Deng. 2008. How is mRNA expression predictive for protein expression? A correlation study on human circulating monocytes. Acta Biochim Biophys Sin (Shanghai) 40: 426- 436. 19. Gry, M., R. Rimini, S. Stromberg, A. Asplund, F. Ponten, M. Uhlen, and P. Nilsson. 2009. Correlations between RNA and protein expression profiles in 23 human cell lines. BMC Genomics 10: 365. 20. Sarramegna, V., P. Demange, A. Milon, and F. Talmont. 2002. Optimizing functional versus total expression of the human mu-opioid receptor in Pichia pastoris. Protein Expr. Purif. 24: 212-220.

143 21. Andre, N., N. Cherouati, C. Prual, T. Steffan, G. Zeder-Lutz, T. Magnin, F. Pattus, H. Michel, R. Wagner, and C. Reinhart. 2006. Enhancing functional production of G protein-coupled receptors in Pichia pastoris to levels required for structural studies via a single expression screen. Protein Sci. 15: 1115- 1126. 22. Lumeng, C. N., J. L. Bodzin, and A. R. Saltiel. 2007. Obesity induces a phenotypic switch in adipose tissue macrophage polarization. J. Clin. Invest. 117: 175-184. 23. Liu, J., A. Divoux, J. Sun, J. Zhang, K. Clement, J. N. Glickman, G. K. Sukhova, P. J. Wolters, J. Du, C. Z. Gorgun, A. Doria, P. Libby, R. S. Blumberg, B. B. Kahn, G. S. Hotamisligil, and G. P. Shi. 2009. Genetic deficiency and pharmacological stabilization of mast cells reduce diet-induced obesity and diabetes in mice. Nat. Med. 15: 940-945. 24. Weisberg, S. P., D. McCann, M. Desai, M. Rosenbaum, R. L. Leibel, and A. W. Ferrante, Jr. 2003. Obesity is associated with macrophage accumulation in adipose tissue. J. Clin. Invest. 112: 1796-1808. 25. Suetsugu-Maki, R., N. Maki, T. P. Fox, K. Nakamura, R. Cowper Solari, C. R. Tomlinson, H. Qu, J. D. Lambris, and P. A. Tsonis. 2011. A complement receptor C5a antagonist regulates epithelial to mesenchymal transition and crystallin expression after lens cataract surgery in mice. Mol. Vis. 17: 949-964. 26. Griffin, R. S., M. Costigan, G. J. Brenner, C. H. Ma, J. Scholz, A. Moss, A. J. Allchorne, G. L. Stahl, and C. J. Woolf. 2007. Complement induction in spinal cord microglia results in anaphylatoxin C5a-mediated pain hypersensitivity. J. Neurosci. 27: 8699-8708. 27. Slonim, D. K., and I. Yanai. 2009. Getting started in gene expression microarray analysis. PLoS Comput. Biol. 5: e1000543. 28. Gentleman, R. C., V. J. Carey, D. M. Bates, B. Bolstad, M. Dettling, S. Dudoit, B. Ellis, L. Gautier, Y. Ge, J. Gentry, K. Hornik, T. Hothorn, W. Huber, S. Iacus, R. Irizarry, F. Leisch, C. Li, M. Maechler, A. J. Rossini, G. Sawitzki, C. Smith, G. Smyth, L. Tierney, J. Y. Yang, and J. Zhang. 2004. Bioconductor: open software development for computational biology and bioinformatics. Genome Biol. 5: R80. 29. Hochberg, Y., and Y. Benjamini. 1990. More powerful procedures for multiple significance testing. Stat. Med. 9: 811-818.

144 30. Morey, J. S., J. C. Ryan, and F. M. Van Dolah. 2006. Microarray validation: factors influencing correlation between oligonucleotide microarrays and real- time PCR. Biol. Proced. Online 8: 175-193. 31. Gavala, M. L., L. M. Hill, L. Y. Lenertz, M. R. Karta, and P. J. Bertics. 2010. Activation of the transcription factor FosB/activating protein-1 (AP-1) is a prominent downstream signal of the extracellular nucleotide receptor P2RX7 in monocytic and osteoblastic cells. J. Biol. Chem. 285: 34288-34298. 32. Carballo, E., W. S. Lai, and P. J. Blackshear. 1998. Feedback inhibition of macrophage tumor necrosis factor-alpha production by tristetraprolin. Science 281: 1001-1005. 33. Patterson, K. I., T. Brummer, P. M. O'Brien, and R. J. Daly. 2009. Dual- specificity phosphatases: critical regulators with diverse cellular targets. Biochem. J. 418: 475-489. 34. Babon, J. J., N. J. Kershaw, J. M. Murphy, L. N. Varghese, A. Laktyushin, S. N. Young, I. S. Lucet, R. S. Norton, and N. A. Nicola. 2012. Suppression of cytokine signaling by SOCS3: characterization of the mode of inhibition and the basis of its specificity. Immunity 36: 239-250. 35. Gordon, S. 2003. Alternative activation of macrophages. Nat. Rev. Immunol. 3: 23-35. 36. Chawla, A., K. D. Nguyen, and Y. P. Goh. 2011. Macrophage-mediated inflammation in metabolic disease. Nat. Rev. Immunol. 11: 738-749. 37. Gordon, S., and P. R. Taylor. 2005. Monocyte and macrophage heterogeneity. Nat. Rev. Immunol. 5: 953-964. 38. Stout, R. D., C. Jiang, B. Matta, I. Tietzel, S. K. Watkins, and J. Suttles. 2005. Macrophages sequentially change their functional phenotype in response to changes in microenvironmental influences. J. Immunol. 175: 342-349. 39. Mulligan, M. S., E. Schmid, B. Beck-Schimmer, G. O. Till, H. P. Friedl, R. B. Brauer, T. E. Hugli, M. Miyasaka, R. L. Warner, K. J. Johnson, and P. A. Ward. 1996. Requirement and role of C5a in acute lung inflammatory injury in rats. J. Clin. Invest. 98: 503-512. 40. Johnson, K. J., and P. A. Ward. 1974. Acute immunologic pulmonary alveolitis. J. Clin. Invest. 54: 349-357. 41. Thiel, G., and G. Cibelli. 2002. Regulation of life and death by the zinc finger transcription factor Egr-1. J. Cell. Physiol. 193: 287-292.

145 42. Kumbrink, J., K. H. Kirsch, and J. P. Johnson. 2010. EGR1, EGR2, and EGR3 activate the expression of their coregulator NAB2 establishing a negative feedback loop in cells of neuroectodermal and epithelial origin. J. Cell. Biochem. 111: 207-217. 43. Kawauchi, J., C. Zhang, K. Nobori, Y. Hashimoto, M. T. Adachi, A. Noda, M. Sunamori, and S. Kitajima. 2002. Transcriptional repressor activating transcription factor 3 protects human umbilical vein endothelial cells from tumor necrosis factor-alpha-induced apoptosis through down-regulation of p53 transcription. J. Biol. Chem. 277: 39025-39034. 44. Hu, X., E. Yee, J. M. Harlan, F. Wong, and A. Karsan. 1998. Lipopolysaccharide induces the antiapoptotic molecules, A1 and A20, in microvascular endothelial cells. Blood 92: 2759-2765. 45. Dent, A. L., F. H. Vasanwala, and L. M. Toney. 2002. Regulation of gene expression by the proto-oncogene BCL-6. Crit. Rev. Oncol. Hematol. 41: 1-9. 46. Kurihara, R., K. Yamaoka, N. Sawamukai, S. Shimajiri, K. Oshita, S. Yukawa, M. Tokunaga, S. Iwata, K. Saito, K. Chiba, and Y. Tanaka. 2010. C5a promotes migration, proliferation, and vessel formation in endothelial cells. Inflamm. Res. 59: 659-666. 47. Cao, W., S. H. Lee, and J. Lu. 2005. CD83 is preformed inside monocytes, macrophages and dendritic cells, but it is only stably expressed on activated dendritic cells. Biochem. J. 385: 85-93. 48. Soruri, A., J. Riggert, T. Schlott, Z. Kiafard, C. Dettmer, and J. Zwirner. 2003. Anaphylatoxin C5a induces monocyte recruitment and differentiation into dendritic cells by TNF-alpha and prostaglandin E2-dependent mechanisms. J. Immunol. 171: 2631-2636. 49. Chapoval, A. I., J. Ni, J. S. Lau, R. A. Wilcox, D. B. Flies, D. Liu, H. Dong, G. L. Sica, G. Zhu, K. Tamada, and L. Chen. 2001. B7-H3: a costimulatory molecule for T cell activation and IFN-gamma production. Nat. Immunol. 2: 269-274. 50. Flierl, M. A., H. Schreiber, and M. S. Huber-Lang. 2006. The role of complement, C5a and its receptors in sepsis and multiorgan dysfunction syndrome. J. Invest. Surg. 19: 255-265. 51. Okusawa, S., K. B. Yancey, J. W. van der Meer, S. Endres, G. Lonnemann, K. Hefter, M. M. Frank, J. F. Burke, C. A. Dinarello, and J. A. Gelfand. 1988.

146 C5a stimulates secretion of tumor necrosis factor from human mononuclear cells in vitro. Comparison with secretion of interleukin 1 beta and interleukin 1 alpha. J. Exp. Med. 168: 443-448. 52. Miyamoto, S., A. Ochiai, N. Boku, A. Ohtsu, M. Tahara, S. Yoshida, H. Okabe, T. Takechi, and M. Fukushima. 2001. Discrepancies between the gene expression, protein expression, and enzymatic activity of thymidylate synthase and dihydropyrimidine dehydrogenase in human gastrointestinal cancers and adjacent normal mucosa. Int. J. Oncol. 18: 705-713. 53. Geiger, T., C. Rordorf, N. Galakatos, B. Seligmann, R. Henn, J. Lazdins, and K. Vosbeck. 1992. Recombinant human C5a induces transcription but not translation of interleukin-1 beta mRNA in human monocytes. Res. Immunol. 143: 117-123. 54. Schindler, R., J. A. Gelfand, and C. A. Dinarello. 1990. Recombinant C5a stimulates transcription rather than translation of interleukin-1 (IL-1) and tumor necrosis factor: translational signal provided by lipopolysaccharide or IL-1 itself. Blood 76: 1631-1638. 55. de la Brousse, F. C., and S. L. McKnight. 1993. Glimpses of allostery in the control of eukaryotic gene expression. Trends Genet. 9: 151-154. 56. Hill, C. S., and R. Treisman. 1995. Transcriptional regulation by extracellular signals: mechanisms and specificity. Cell 80: 199-211. 57. Angel, P., M. Imagawa, R. Chiu, B. Stein, R. J. Imbra, H. J. Rahmsdorf, C. Jonat, P. Herrlich, and M. Karin. 1987. Phorbol ester-inducible genes contain a common cis element recognized by a TPA-modulated trans-acting factor. Cell 49: 729-739. 58. Lee, K. A., T. Y. Hai, L. SivaRaman, B. Thimmappaya, H. C. Hurst, N. C. Jones, and M. R. Green. 1987. A cellular protein, activating transcription factor, activates transcription of multiple E1A-inducible adenovirus early promoters. Proc. Natl. Acad. Sci. U. S. A. 84: 8355-8359. 59. Maekawa, T., H. Sakura, C. Kanei-Ishii, T. Sudo, T. Yoshimura, J. Fujisawa, M. Yoshida, and S. Ishii. 1989. Leucine zipper structure of the protein CRE- BP1 binding to the cyclic AMP response element in brain. EMBO J. 8: 2023- 2028.

147 60. Hai, T., and T. Curran. 1991. Cross-family dimerization of transcription factors Fos/Jun and ATF/CREB alters DNA binding specificity. Proc. Natl. Acad. Sci. U. S. A. 88: 3720-3724. 61. Hajishengallis, G., and J. D. Lambris. 2010. Crosstalk pathways between Toll- like receptors and the complement system. Trends Immunol. 31: 154-163. 62. Porcheray, F., S. Viaud, A. C. Rimaniol, C. Leone, B. Samah, N. Dereuddre- Bosquet, D. Dormont, and G. Gras. 2005. Macrophage activation switching: an asset for the resolution of inflammation. Clin. Exp. Immunol. 142: 481-489.

148 Chapter

4 A Comparison of Three potent antagonists of C5aR: Importance of Receptor Residence Time

149 4.1 Abstract

Activation of complement produces a cascading network of plasma and membrane proteins that functions to rapidly identify, tag, lyse and quickly eliminate foreign and infectious organisms, as well as damaged cells, from the host, thereby contributing to immune defence. When the foreign stimulus cannot be removed quickly, the acute immune response is prolonged (sometimes for years) and complement proteins, such as C5a and its receptors, may be excessively expressed. Uncontrolled complement activation can cause many inflammatory and autoimmune disorders, leading to chronic inflammatory diseases and cancer. Potential therapeutics have been developed to target several complement proteins, but the only FDA approved treatment that targets C5a-C5aR interaction is a humanized antibody which actually binds C5 thereby preventing the formation of C5a as well as formation of the membrane attack complex that effects pathogen lysis. In this present study, three previously reported antagonists of human C5aR (W54011, JJ47 and 3D53) have been compared, with the goal of choosing one for further studies (Chapter 5) on synergistic effects of C5a on lipopolysaccharide action on macrophages and monocytes. All three antagonists were confirmed here as being potent antagonists (IC50 < 50 nM against

0.1 nM human C5a). However, IC50 for a competitive (‘surmountable’) antagonist is dependent on the concentration of the competing agonist, so both W54011 and JJ47 became antagonists at only µM concentrations when the concentration of C5a increased above 30 nM concentrations. By contrast, 3D53 is a non-competitive (“insurmountable”) antagonist that maintains its potency at low nM concentrations against 0.1 to 300 nM C5a. 3D53 had a much longer duration of action (t1/2 = 22 h) on human macrophages than W54011 (t1/2 = 1.2) or and JJ47 ( t1/2 = 0.6 h). 3D53 was not only superior to W54011 and JJ47 in an assay measuring intracellular calcium release, but also in a cell migration (chemotaxis) assay, where antagonist-treated macrophages were exposed to C5a for 16 h. Furthermore, orally administered 3D53 was an effective anti-inflammatory agent for up to 16h, compared to ≤ 2h for W54011 and JJ47, in inhibiting C5aR-mediated paw oedema in male Wistar rats, even though 3D53 reportedly has low oral bioavailability and is rapidly cleared from blood while W54011 and JJ47 are reportedly much more orally bioavailable. The results highlight the fact that receptor affinity, antagonist potency, drug-likeness and oral

150 bioavailability of a ligand for its receptor are insufficient to define the effectiveness and duration of biological action. Rather, the lifetime of the binary receptor-ligand complex can be even more important in dictating drug efficacy in the cellular and organismal context. Based on the findings herein, 3D53 was selected for further study in Chapter 5.

151 4.2 Introduction

Complement component C5a, is a potent chemotactic factor for neutrophils and other leukocytes (1). C5a exerts these activities by binding to a G-protein coupled C5a receptor (C5aR) (2), with a second receptor C5L2 of uncertain function also known (3), although this has largely been reported to be non-signalling and may play a secondary control role on levels of C5aR being accessible to ligands. C5aR expression has been traditionally thought to be limited primarily to myeloid blood cells, including neutrophils, monocytes, macrophages and eosinophils but recent evidence suggests it is more widely distributed to organs, including the liver, kidney and the central nervous system than previously thought (4-7). The effects of C5a- C5aR interaction may also be involved in other biological functions besides immunity and inflammation. C5aR signalling has recently been linked to the development of obesity and metabolic dysfunction in rats (8) and may be important in cancer development (9-11). Clearly, C5a is not just another agent regulating inflammation and chemtaxis, although these are perhaps the best-documented functions to date. Complement operates in a non-specific manner, attacking both foreign as well as host cells if the latter are tagged for destruction. Complement activation is usually tightly regulated by host self-regulatory mechanisms during normal physiology condition. However, excessive complement activation can occur when the extent of activation exceeds the capacity of host mechanisms to regulate it, leading to an overproduction of C5a in circulation and, eventually, to many inflammatory and autoimmune disorders (12). As such, it is desirable to modulate complement activation by using therapeutic interventions such as inhibitors or antibodies. Antibodies that block the proteolysis of C5 to C5a and C5b (an early precursor to membrane attack complex) has been clinically validated and approved by the U.S. Food and Drug Administration (FDA) for treating paroxysmal nocturnal hemoglobinuria (13). However, protein as effective inhibitors are expensive to produce, difficult to formulate and can potentially trigger unwanted immunogenic side effects. The failure of anti-C5 mAb pexelizumab in clinical studies for treating acute myocardial infarction was found to be due to poor tissue penetration, leading to poor therapeutic efficacy. On the other hand, unlike proteins or antibodies, drug-like small molecules do not possess these disadvantages in therapeutic invention.

152 Although the mission towards complement therapeutics started over 30 years ago (14), there are still no small molecule drugs in the marketplace for targeting complement. C5aR has been demonstrated to be an important and promising therapeutic target for complement modulation (2). Efforts in developing C5aR antagonists have resulted in the discovery of different types of ligands, including antibodies, peptides, peptidomimetics and non-peptidic small molecules (2). Over the past two decades, a few potent small molecule C5aR antagonists have been reported with activity in various animal models and in humans (15). Despite such intensive efforts to develop an effective C5aR antagonist to market, this has not yet been achieved with pre-clinical and early stage clinical trials being thwarted by unexpected problems. Traditionally, drug discovery has focused on optimizing ligand affinity for a specific receptor, enhancing functional potency and selectivity, and improving drug-like properties for optimal pharmacokinetics and oral bioavailability. However, these optimization processes have not always led to optimal in vivo efficacy (16). Consequently, advancing drugs based on their simple in vitro potencies and affinities may not be sufficient high chances of clinical success and more sophisticated thinking needs to be brought to traditional drug design and development approaches to advancing compounds to market. This chapter highlights an important consideration, often overlooked or not given enough importance, in drug development – namely residence time for drug- receptor interactions. Here, three potent C5aR antagonists described in literature were chosen for a comparative study to re-evaluate their intrinsic properties as antagonists on C5aR. The compounds were 3D53 (17), W54011 (18) and JJ47 (19).

153

The original cyclic peptide scaffold 3D53 was designed in our group from linear peptides, such as compound C089 (20), which was derived from the C-terminus of C5a through many years of peptide research at Abbott and then Merck. After many structure-activity relationship (SAR) studies, 3D53 was created by PhD student Allan Wong in the Centre for Drug Design and Development (3D Centre), University of Queensland in 1996. It was the first cyclic peptidomimetics C5aR antagonist and proved to be a potent and water-stable full antagonist of C5aR on human neutrophils (21). Eventually, 3D53 was licensed as PMX53 for clinical development by the now defunct Promics Ltd (subsequently passed on through a series of company takeovers to Peptech Pty Ltd, Arana Therapeutics Ltd, Cephalon Inc. and then in 2011 by TEVA Pharmaceutical Industries Ltd, which is the world’s largest manufacturer of generic pharmaceuticals (22). PMX53 was safe and well tolerated in Phase I clinical trials, and 3D53 has been the most intensely evaluated small molecule C5aR antagonist in human clinical trials. W54011 was developed by Mitsubishi Pharma, and was the result of an optimized series of substituted phenylguanidines and tetrahydronapthalene-based compounds (18). W54011 was reported to be a competitive non-peptidic C5aR antagonist. Although Sumichika et al. did not report the percent oral bioavailability of W54011, the authors claimed that this antagonist was potent and orally active in vivo

154 (18). However, it is very hydrophobic and had problems with species specificity (active in human, cynomolgus monkey and gerbil but not as active in mouse, rat guinea pig, rabbit or dog neutrophils), which complicated pre-clinical studies (2). JJ47 was the most potent antagonist of a series of aniline-substituted tetrahydroquinoline compounds developed by Johnson & Johnson Pharmaceutical Research and Development, with C5aR antagonist activity at single-digit nanomolar concentrations in a calcium mobilization assay (19). Since W54011 and JJ47 are non- peptidic antagonists of C5aR, they were not expected to have the disadvantages of peptidic ligands, for example, susceptibility to proteolytic cleavage in serum, low oral bioavailability as the result of hydrophillicity, and high molecular weights. Based on reported IC50 values, 3D53 appeared to be the least potent C5aR antagonist among the three antagonists, as well as with the least drug-like properties (Table 4.1). Furthermore, the former two compounds obey the rule-of-five (23, 24) prediction for oral bioavailability (hydrogen bond donors (HBD) ≤ 5, hydrogen bond acceptors (HBA) ≤ 10, LogP < 5, MW < 500), whereas 3D53 does not (Table 4.2).

Table 4.1. Reported activity of 3D53, W54011 and JJ47 on human C5aR.

Human C5a Cell-types Oral Antagonist Compound concentration used bioavailability IC50 (nM)+ (nM) (%) in rats 55 (18) 0.1 Human 3D53 1 (22) 20 (25) 100 neutrophils Human W54011 3 (18) 0.1 N.D.* neutrophils JJ47 7 (19) 1.5 U937 21# (19) +Values are determined by calcium release versus stated hC5a concentration. *Presumably high, values not disclosed by the authors. #Pharmacokinetic study was derived from a represented compound of the same series.

Table 4.2. Comparison of rule-of-five parameters of 3D53, W54011 and JJ47

3D53 W54011 JJ47 HBD 12 0 2 HBA 10 4 4 CLogP 3.2 7.4 6.6 MW 896.1 456.6 458.7 Values that are bolded do not obey the rule-of-five prediction for oral bioavailability.

155 By these criteria, 3D53 would be expected to be least drug-like, least effective in vitro, and least effective given orally to animals. In summary, 3D53 is least drug- like of these antagonists.

4.3 Aims

The overall objective of this chapter was to select the most appropriate C5aR antagonist for studies in Chapter 5. Three reported antagonists (3D53, W54011 and JJ47) with quite different chemical structures, comparable antagonist potency at low nanomolar concentrations, and oral activity in animals have been re-examined here for functional potency, drug-like potential and oral efficacy. The specific aims for this chapter were: 1. To re-assess the reported antagonism by 3D53, W54011 and JJ47 using C5a- induced calcium release in human macrophages. 2. To investigate and compare the mechanisms of antagonism for the three antagonists. 3. To compare properties of these antagonists using different in vitro assays. 4. To compare these antagonists in C5a-induced inflammation in an animal model, thereby potentially relating in vitro observations to in vivo outcomes.

4.4 Methods

4.4.1 Reagents

Recombinant human M-CSF was purchased from PeproTech (USA). Recombinant human C5a was purchased from Sino Biological Inc (China). C5aR antagonists (3D53 (26), W54011 (18) and JJ47 (19)) and C5a peptide agonist Ac-Phe- Lys-Pro-D-Cha-Cha-D-Arg-OH (modified with an acetyl-capped on the N-terminus) (27) were synthesized and characterized (analytical HPLC, mass spectroscopy and nuclear magnetic resonance spectroscopy) in-house as described (18, 19, 26). All cell culture reagents were purchased from Invitrogen (Australia) and all analytical grade

156 chemical reagents were obtained from Sigma-Aldrich (Australia), unless otherwise stated.

4.4.2 Animals

The in vivo work was undertaken, on my behalf, by Mr Adam Cotterell (Institute for Molecular Bioscience, University of Queensland). Male Wistar rats (6-8 wk, 250 ± 50 g) were bred at the Animal Resources Centre (Canning Vale, WA, Australia) and housed post-shipment at the Australian Institute for Bioengineering and Nanotechnology (University of Queensland, Brisbane, QLD, Australia). Animals were maintained in a 12-h light-dark cycle according to the standard of holding facility with food and water provided. Male Wistar rats were given 3D53 (5 mg/kg, n=3/group) or W54011/JJ47 (30 mg/kg, n=2/group) (p.o. via gavage in olive oil, 500 µL) prior to C5aR peptide agonist (C5aR-PA) administration. Control animals received only olive oil (500 µL p.o.). After specific time-points, C5aR-PA was administered into the right hind paw pad (350 µg in saline, 100 µL, i.pl). The left hind paw acted as a control, receiving saline only (100 µL). Paw thickness and widths were measured at 1 h post-injection using digital callipers (World Precision Instruments, Sarasota, FL, USA) and swelling was expressed as area (mm2; thickness x width) and plotted as percentage change from baseline of each individual paw. The animal ethics committee of The University of Queensland approved all experiments performed in this study.

4.4.3 Isolation of human monocytes and macrophages

Peripheral blood mononuclear cells (PBMCs) were isolated from buffy coats of anonymous donors (Australian Red Cross Blood Service, Kelvin Grove, Queensland, Australia) by density centrifugation using Ficoll-Paque Plus (GE Healthcare) manufacturer’s instructions. Contaminating erythrocytes were removed by repeated ice-cold sterile water. CD14+ MACS® microbeads were used to positively select monocytes from isolated PBMCs according to manufacturer’s instructions (Miltenyi Biotec). CD14+ monocytes were seeded at 1 x 106 cells/mL in

157 IMDM supplemented with 10% FBS, 100 U/mL penicillin, 100 µg/mL streptomycin and 2 mM GlutaMAX at 37 °C in presence of 5% CO2. To generate human monocyte-derived macrophages (HMDM), CD14+ monocytes were cultured in the presence 10 ng/mL M-CSF for at least 6 days.

4.4.4 Isolation of rat bone marrow cells and macrophages

Bone marrow cells (BMC) were aseptically eluted from excised rat femurs by injecting vigorously with RPMI medium through the bone marrow cavity. Contaminating erythrocytes were removed by repeated ice-cold sterile water. BMC were seeded at 1 x 106 cells/mL in RPMI supplemented with 10% FBS, 100 U/mL penicillin, 100 µg/mL streptomycin and 2 mM GlutaMAX at 37 °C in presence of 5%

CO2. To generate rat bone marrow-derived macrophages (BMM), BMC were cultured in the presence 10 ng/mL M-CSF for at least 6 days.

4.4.5 Intracellular calcium mobilization assay

HMDM were seeded at 0.5 x 106 cells/mL (100 µL) in black-wall, clear- bottom 96-well plates (Corning) and incubated at 37 °C in the presence of 5% CO2 overnight for cells to adhere. Cells were washed once with assay buffer (Hank’s buffered salt solution (HBSS), 20 mM HEPES, 2.5 mM probenecid, pH 7.4). 100 µL of dye-loading buffer (12 mL of assay buffer, 1% FBS, 25 µL of Fluo-3AM (Invitrogen) (final concentration = 4 µM), 25 µL 20% pluronic acid) was added to each well and incubated for 1 h at 37°C. Excess loading dye was washed away with assay buffer before adding 50 µL of tested compound. Fluorescence intensity was recorded with on FLIPR Tetra (Molecular Devices) with excitation and emission wavelengths of 485 nm and 520 nm respectively. Calcimycin ionophore A23187 (25 µM) (Sigma) is used as a dye loading positive control.

158

4.4.6 Quantification of ERK1/2 phosphorylation by AlphaScreen Surefire

(phosphor-Thr202/Tyr204) according to Alphascreen SureFire assays (PerkinElmer) manufacturer’s instructions. At completion of the assay, signal in the wells was measured using Pherastar (BMG Labtech).

4.4.7 In vitro migration assay (modified Boyden chamber)

4.4.8 RNA extraction and reverse transcription

159 transcripted with Superscript III (Invitrogen) at 50°C for 50 min, then 70°C for 10 min. cDNA samples were stored at -20°C until further use.

4.4.9 Real-time quantitative polymerase chain reaction (real-time qPCR)

Gene-specific primers were designed using the free web-based software Primer-BLAST National Center for Biotechnology Information (NCBI). Quantitative real-time PCR was performed using an ABI Prism 7900 real-time PCR system (Applied Biosystems), cDNA (50 ng), SYBR® Green PCR master mix (Applied Biosystems) and gene-specific primers. Conditions were used according to the manufacturer’s instructions. Relative gene expression was normalised against 18S rRNA expression and then converted to fold change against control samples. All samples were analysed in duplicates. Primer sequences used are: ATF3, 5’- CGAGCGGAGCCTGGAGCAAA-3’ and 3’-GGGGACAGGCAGGGGACGAT-5’; TNF, 5’-TCTGGCCCAGGCAGTCAGATC-3’ and 3’- CACTGGAGCTGCCCCTCAGC-5’; IL1B, 5’- CCTCTTCGAGGCACAAGGCACAA-3’ and 3’- TGGCTGCTTCAGACACTTGAGCAAT-5’; CCL3, 5’- CCCACATTCCGTCACCTGCTCAG-3’ and 3’- AGCAGCAAGTGATGCAGAGAACTG-5’; PTGS2, 5’- TGGCAGGGTTGCTGGTGGTA-3’ and 3’-TCTGCCTGCTCTGGTCAATGGA-5’; 18S, 5’-ACCACGGGTGACGGGGAATC-3’ and 3’- CCGGGTCGGGAGTGGGTAAT-5’.

4.4.10 Statistical analysis

160 4.5 Results

4.5.1 3D53, a potent and non-competitive C5aR antagonist

3D53, W54011 and JJ47 were reported to be potent human C5aR ligands with low nanomolar antagonism activity (18, 19, 25). However, antagonism was measured under different assay conditions, making it is difficult to draw comparisons from previously reported data. For example, 3D53 and W54011 were evaluated with human neutrophils, whereas JJ47 was evaluated with human monocytic cell line U937.

The reported IC50 values were also derived from experiments using different concentrations of human C5a, whereas this parameter is usually dependent on the concentration of agonist being measured against. A comparative study between 3D53, W54011 and JJ47 was therefore conducted here, so that these antagonists could be directly compared and ranked for antagonist properties. Since these compounds are structurally different, the mechanism of the antagonism for 3D53, W54011 and JJ47 against human C5a was also investigated in a C5a-induced intracellular calcium release assay using human macrophages. The concentration-response curves for Ca2+ mobilization induced by C5a were determined in the presence of increasing concentrations of the three antagonists (Figure 4.1). A reduction of the maximal C5a responses was observed as concentrations of 3D53 increased, but there was no sideways rightward displacement of the curve typical of a ‘competitive’ antagonist, suggesting that 3D53 is an insurmountable C5aR antagonist (Figure 4.1A). However, W54011 and JJ47 both shifted rightward the concentration-response curves for C5a-induced calcium release in human macrophages, without depressing the maximal responses, indicating that both were competitive antagonists (Figure 4.1B, C). Furthermore, the Schild plot or pA2 analysis revealed that the slope of the regression was not 1.0 for 3D53 (Figure 4.1D), which by definition is classified to be non-competitive (28). On the other hand, the slope of regression for W54011 (Figure 4.1E) and JJ47 (Figure 4.1F) was ~1, consistent with W54011 and JJ47 being competitive antagonists with antagonist IC50 dependent on the concentration of C5a being measured against. Although these observations were consistent with previously reported data for 3D53 (25) and

161 W54011 (18) on neutrophils, the mechanism of antagonism for JJ47 was previously unknown, nor were Schild plots reported previously.

Figure 4.1. Mechanism of C5aR antagonism for 3D53, W54011 and JJ47 against human C5a. Concentration dependent responses of C5a with treatment of antagonist (A) 3D53, (B) W54011 and (C) JJ47 at various concentration (0 nM, ; 3 nM, ¢; 10 nM, p; 30 nM, ¿; 100 nM, Â; 300 nM, q; 1000 nM, ê) with C5a (300 nM) as 100% response in human macrophages. Schild plots for antagonists (D) 3D53, (E) W54011 and (F) JJ47 against human recombinant C5a. Calculated pA2 values for 3D53, W54011 and JJ47 are 8.3, 8.6 and 8.3 respectively. Error bars are means ± SEM of three independent experiments (n=3).

Since IC50 values for the insurmountable antagonist 3D53 were independent of the concentration of agonist, in this case the C5a concentration, 3D53 still maintained its antagonistic potency at low nanomolar concentrations even in the presence of concentrations of C5a as high 300 nM (Figure 4.2A). On the other hand, W54011 lost its nM antagonist activity at 100 nM C5a (Figure 4.2B), with the even less potent JJ47 losing its nM antagonist potency at just 10 nM C5a (Figure 4.2C). Although these antagonists were previously reported to be potent activity at low nM concentrations, clearly these reported IC50 values are only of this magnitude at very low concentrations of C5a. W54011 and JJ47 are not truly potent antagonists under changing physiological conditions, where local concentrations of C5a are expected to be higher than 0.1 nM during inflammation (29, 30).

162

C5a 3D53 W54011 JJ47

(nM) pIC50 ± SEM IC50 (nM) pIC50 ± SEM IC50 (nM) pIC50 ± SEM IC50 (nM) 300 7.8 ± 0.1 17 5.8 ± 0.9 1456 6.0 ± 0.9 1071 100 7.9 ± 0.1 14 6.5 ± 0.4 350 6.2 ± 0.4 632 30 7.8 ± 0.2 15 7.3 ± 0.2 55 6.3 ± 0.4 509 10 8.0 ± 0.3 11 7.6 ± 0.2 24 6.9 ± 0.2 122 3 8.5 ± 0.5 3 8.1 ± 0.3 9 8.0 ± 0.2 10 1 8.5 ± 1.0 3 8.4 ± 0.6 4 8.8 ± 0.7 2

Figure 4.2. Antagonist potencies (IC50) and dependence on C5a concentration. Inhibitory responses of antagonist (A) 3D53, (B) W54011 and (C) JJ47 against various concentration of C5a (300 nM, ; 100 nM, ¢; 30 nM, p; 10 nM, ¿; 3 nM, q; 1 nM, ê) with C5a (300 nM) inducing 100% relative response in human macrophages. Calculated pIC50 ± SEM and IC50 values of the three C5aR antagonists were summarized in the table against concentration of C5a. Error bars are means ± SEM of three independent experiments (n=3).

4.5.2 3D53 has a longer residence time than W54011 and JJ47

Residence time in biochemistry is usually determined by measuring ‘on’ and ‘off’ rates for affinity, but this is more difficult to measure on cell surfaces. An indication of residence was instead determined for the three C5aR antagonists by incubating human macrophages in a saturating concentration of each antagonist (1 µM) for one hour, the cells washed to remove unbound ligands, and the treated cells were challenged with C5a at EC50 concentration (3 nM), as determined in chapter 3, and observed at the stated times. If the ligand was still associated with the receptor, C5a did not trigger intracellular calcium release and there was no fluorescence response. It appeared that competitive antagonists W54011 and JJ47 had very short half-lives for duration of antagonist action, 1.2 h and 0.6 h respectively (Figure 4.3A).

On the other hand, 3D53 had a much longer duration of action, t1/2 = 18.2 h (Figure 4.3B).

163

Half-life t ± SEM C5aR Antagonist 1/2 (h) 3D53 18.2 ± 4.1 W54011 1.2 ± 0.3 JJ47 0.6 ± 0.2

Figure 4.3. 3D53 has a longer residence time on human C5aR compared to W54011 and JJ47. Cells were pre-treated with 1 µM C5aR antagonist () 3D53, (p) W54011 and () JJ47 for 1 h. Excess unbound antagonists were removed by washing. (A, B) Antagonist residence time on C5aR was determined by subjecting treated cells to 3 nM C5a during the stated duration post antagonist-treatment. Calculated half-life for 3D53, W54011 and JJ47 are 18.2 h, 1.2 h and 0.6 h respectively. Error bars are means ± SEM of three independent experiments (n=3).

4.5.3 C5a-induced chemotaxis is efficiently blocked by 3D53 but not W54011 and JJ47

Insurmountable antagonism implies that the off-rate of the antagonist from the receptor-binding site is so slow that a true equilibrium between the agonist and antagonist on the receptor cannot be attained in the short timeframe of the calcium release assay. Therefore, a more stringent assay was performed to report on the surmountability of the antagonists. The chemotaxis migration assay, in which cells migrate along an agonist concentration gradient due to agonist-receptor interaction, is a more stringent test than the calcium release assay. This is because antagonist-treated macrophages are now exposed to C5a for 16 h continuously instead of for 5 min at different intervals. Antagonists with a fast off-rate will be inefficient in this assay because, once the antagonist has dissociated, C5a immediately activates the cells and

164 induce migration. Human macrophages were stimulated with C5a at EC80 concentration (3 nM) to induce cell migration (as described in Chapter 3). While 3D53 was able to block C5a-induced migration, W54011 and JJ47 were not even at 1 µM concentrations (Figure 4.4). C5a is one of the most potent endogenous agents known to induce chemotaxis of immune cells. Since the antagonists W54011 and JJ47 failed to block chemotaxis, a primary function of C5a, of macrophages, one may wonder if these antagonists can be functional in an in vivo setting.

6 *** *** 4 *** 2

Chemotactic index 0 C5a (3 nM) - + + + + + + + Antagonist (µM) - - 0.1 1 0.1 1 0.1 1 3D53 W54011 JJ47

Figure 4.4. Comparative antagonism of human macrophage migration (induced by 3 nM rhC5a) by 0.1 µM and 1 µM 3D53, W54011 and JJ47. Human macrophages were serum-deprived overnight prior to addition to the upper chamber of the Transwell and then allowed to migrate through the 0.5µm porous membrane into the lower chamber containing serum free media with or without 3 nM C5a for 16 h. For antagonism of C5aR, cells were treated with 3D53 (0.1 or 1 µM), W54011 (0.1 or 1 µM) or JJ47 (0.1 or 1 µM) for 1 h. Unbound antagonists were washed away prior to stimulation with 3 nM human recombinant C5a. Migrated cells on the lower side of the membrane were stained with DAPI and counted under a fluorescence microscope using a x10 objective. Chemotactic index is expressed as fold change against control. Error bars are means ± SEM of three independent experiments (n=3). ***p < 0.001 by student t-test.

165 4.5.4 C5aR-PA is a selective peptidic agonist of C5aR at low micromolar concentration

The activities of C5a can be modulated by the enzyme carboxypeptidase, which rapidly metabolizes C5a to C5adesArg in serum within seconds by removing the C-terminal arginine that is important for binding to C5aR (31). C5adesArg has a 10-1000 times reduced potency compared to C5a depending the function being measured (32). It is technically not feasible to establish a C5aR-induced inflammatory rat model using recombinant C5a protein due to this degradation, not to mention too costly. Interestingly, only 8 residues of the C-terminus of C5a are responsible for its agonist activity on C5aR, although the remainder of C5a is crucial for high affinity binding and thus agonist potency (33). A previously reported C5aR agonist peptide, Phe-Lys-Pro-DCha-Cha-DArg-OH, was developed based on the C-terminus of C5a (34). Peptides are generally not stable because proteolytic enzymes recognize their random or strand like conformations in water (35), but by incorporating D-residues, N-methyl substitutents or acetylating the N-terminus increases stability to proteolytic degradation (36, 37). Ac-FKP-dChaCha-dR-OH (referred to now as C5aR-PA), is an agonist of C5aR and was used here as a cheaper and more feasible tool to induce C5aR-mediated paw oedema in rats, and for comparing the effectiveness of 3D53, W54011 and JJ47 as antagonists of C5aR-induced paw oedema in rats. C5aR-PA was previously reported to be selective for C5aR over C5L2 (27), but it was unclear whether it non-specifically bound to C5aR. First generation peptide agonists for C5aR were notorious for also binding to C3aR (38), despite the latter having only 37% sequence identity to C5aR (39). The desensitization calcium assay was setup to investigate the selectivity of C5aR-PA for C5aR, over other GPCRs capable of also inducing calcium release upon activation. The first addition of rhC5a at 1 µM was sufficient to desensitize all C5aR present on human macrophages (Figure 4.5A). These cells were then no longer responsive to additional C5a, as shown by the absence of a second spike in the fluorescence corresponding to a calcium signal. However, a second spike was present when these C5aR-desensitized cells were exposed to 1 µM C3a, suggesting that other GPCRs can still be readily activated upon stimulation if the second compound is not specific for C5aR (Figure 4.5B).

166

Agonist pEC50 ± SEM EC50 (nM) C5a 8.7 ± 0.1 2.2 C5aR-PA 7.1 ± 0.1 82.9

Figure 4.5. C5aR-PA is a selective C5aR peptide agonist at low µM concentration.

Desensitization assays were performed with 1 µM C5a at t0 to desensitize C5aR on human macrophages. (A) Cells were challenged with 1 µM C5a at t920 to confirm C5aR desensitization. (B) Cells were challenged with 1 µM C3a at t920 to confirm that the cells is still capable of inducing calcium release when other receptors are activated. (C) Cells were challenged with 1 µM C5aR-PA at t920 to confirm ligand selectivity to C5aR. (D) Calcium release in macrophages was induced by various concentration of (¢) C5aR-PA or () C5a. (E) For antagonism of C5aR, cells were pretreated with 3D53, W54011 and JJ47 (1 µM, 30 min) prior to 300 nM C5aR-PA stimulation. Cells treated with calcimycin (25 µM) were used as a dye-loading control. Error bars are means ± SEM of four independent experiment (n=4). ***p < 0.001 by student t-test.

167 This was an important control because there was a possibility that the cells could have been depleted of cytosolic calcium released from the ER as the result, these cells are no longer able to trigger calcium release upon activation of other GPCRs. The desensitization assay revealed that the C5aR-PA was selective for C5aR at 1 µM concentration (Figure 4.5C), but not above 10 µM as shown by a second calcium signal spike (Figure 4.5D). It is possible that C5aR-PA may therefore bind non-specifically to C3aR and other GPCRs at concentrations above 1 µM on human macrophages. C5aR-PA treatment above 1 µM on human macrophages should thus be avoided to trigger C5aR specific responses. C5aR-PA was also ~40-fold less active than C5a in the calcium release assay on human macrophages (Figure 4.5E). C5aR- PA is more stable in serum, which means that it is also more biologically active and will be inducing a stronger and sustainable C5aR-mediated response in vivo compared to its native C5a protein. C5aR-PA mimics the C-terminus of the native C5a protein, which is responsible for the binding and activation of C5aR signalling. Given that 3D53, W54011 and JJ47 were able to block C5aR-PA-induced calcium release in human macrophages (Figure 4.5F), these antagonists likely bind similarly to C5aR- PA. Although there is no structural evidence in the literature describing where and how these antagonists bind, the results here shed some insights on which pockets within the C5aR helical-transmembrane domain these antagonists might interact.

4.5.5 C5aR-induced ERK phosphorylation in human and rat macrophages

After validating C5aR-PA as a suitable agonist substitute for the native C5a protein, as a tool to examine C5aR mediated responses in human macrophages, this section was aimed at determining whether C5aR-PA was mimicking the effect of C5a in inducing C5aR-mediated responses in rats, as shown earlier in vitro for human macrophages. To monitor C5aR activation, C5aR-PA and human C5a (up to 300 nM) were first examined on rat macrophages differentiated from rat bone marrow cells. However, neither agonist was able to induce calcium release (data not shown). Another pathway related to GPCR signalling, ERK phosphorylation, was therefore monitored since human C5a able to induce ERK phosphorylation in rat macrophages

168 (Figure 4.6). Although human C5a was ~15-fold less potent on rat C5aR compared to human C5aR on macrophages in the ERK phosphorylation assay.

Species of pEC ± SEM EC (nM) Macrophages 50 50 Human 8.6 ± 0.3 2.8 Rat 7.3 ± 0.4 41.8

Figure 4.6. C5a-induced ERK phosphorylation in human and rat macrophages. ERK1/2 phosphorylation in Human-MΦ (¢) and Rat-MΦ () was induced by various concentrations of recombinant human C5a. Cells were treated for 5 min and lyzed to stop C5a-induced reaction. Phosphorylated ERK1/2 was measured by AlphaScreen® SureFire® assays. Error bars are means ± SEM of three independent experiments (n=3).

4.5.6 Antagonism of C5a- and C5aR-PA- induced gene expressions in human and rat macrophages

A panel of genes (ATF3, TNF, IL1B, CCL3 and PTGS2) was next chosen to monitor C5aR-induced mRNA expression upon activation by C5a and C5aR-PA. C5a substantially amplified these 5 genes after 30 min exposure, as evidenced in microarray studies in chapter 3. Both human and rat macrophages were treated with

C5a and C5aR-PA (at EC90 concentration) to induce C5aR-mediated gene expression.

EC90 values of C5a and C5aR-PA were calculated from the concentration-dependent

169 curves for calcium release (Figure 4.5D) and ERK phosphorylation (Figure 4.6) assays for human or rat macrophages. All three antagonists efficiently blocked C5a- induced ATF3, TNF, IL1B and PTGS2 gene expressions at 1 h post-antagonism (Figure 4.7A-C,E), and W54011 and JJ47 to a lesser extent blocked C5a-induced CCL3 gene expressions (Figure 4.7D). However, at 16 h post-antagonism, while 3D53 still antagonized C5a-mediated gene expressions, W54011 and JJ47 did not (Figure 4.7F-J). This is consistent with the antagonists W54011 and JJ47 having much shorter durations of action due to much shorter residence times on C5aR, being completely displaced from the receptor well within 16 h. As expected, the magnitude of gene expression induced by 10 nM C5a (Figure 4.7) and 300 nM C5aR-PA (Figure 4.8) appeared to be very similar, thereby reaffirming the accuracy and reproducibility of the EC90 values translated from the calcium release assay to PCR. Mirroring the results obtained from the three antagonists against C5a, all three antagonists also efficiently blocked C5aR-PA- induced gene expressions ATF3, TNF, IL1B and CCL3 (Figure 4.8A-D), and again W54011 and JJ47 to a lesser extent blocked C5aR-PA-induced PTGS2 gene expression (Figure 4.8E). Both W54011 and JJ47 were ineffective at blocking C5aR- PA-induced gene expression after 16 h post-antagonism, whereas 3D53 maintained its antagonist effect even at this time point (Figure 4.8F-J). The effectiveness of 3D53, W54011 and JJ47 in blocking C5a-induced and C5aR-PA-induced gene expression thus correlated well with their respective residence times inferred by the results of Figure 4.3 reported for human macrophages. The magnitude of response for gene expression induced by human C5a and C5aR-PA on rat macrophages was surprisingly more modest (Figure 4.9A-E), even though these macrophages were treated at EC90 concentrations as determined in the ERK phosphorylation assay (Figure 4.6). Since C5aR-PA did not manage to induce a large response in rat macrophages, antagonism of 3D53 and W54011 and JJ47 were measured against 100 nM C5a. At 1 h post-antagonism, all three antagonists blocked gene expression induced by human C5a on rat macrophages (Figure 4.9F-J). Most importantly, all three antagonists were able to recognize C5aR expressed on the cell surface of the rat macrophages and blocked responses mediated by human C5a. Therefore, it was feasible to compare 3D53, W54011 and JJ47 in vivo using rats as the choice of animal model.

170

Figure 4.7. Comparative antagonism of C5a-induced gene expressions in human macrophage by 3D53, W54011 and JJ47 at 1h versus 16 h post-treatment. Human macrophages were treated with 10 nM C5a and cells are lysed after 30 min to stop the reaction. For antagonism of C5aR, cells were pretreated with 1 µM 3D53, W54011 or JJ47 for 1 h. Excess unbound antagonists were removed by washing and subjected to C5a treatment after (A-E) 1 h or (F-J) 16 h post-treatment with antagonist. The expressions of (A,F) ATF3, (B,G) TNF, (C,H) IL1B, (D,I) CCL3 and (E,J) PTGS2 were detected by quantitative real-time PCR, normalized against 18S rRNA and converted to fold change relative to control (untreated). Error bars are means ± SEM of three independent experiments (n=3). **p < 0.01 and ***p < 0.001 by student t-test.

171

Figure 4.8. Comparative antagonism of C5aR-PA-induced gene expressions in human macrophage by 3D53, W54011 and JJ47 at 1h versus 16 h post-treatment. Human macrophages were treated with 300 nM C5aR-PA and cells are lysed after 30 min to stop the reaction. For antagonism of C5aR, cells were pretreated with 1 µM 3D53, W54011 or JJ47 for 1 h. Excess unbound antagonists were removed by washing and subjected to C5aR-PA treatment after (A-E) 1 h or (F-J) 16 h post- treatment with antagonist. The expressions of (A,F) ATF3, (B,G) TNF, (C,H) IL1B, (D,I) CCL3 and (E,J) PTGS2 were detected by quantitative real-time PCR, normalized against 18S rRNA and converted to fold change relative to control (untreated). Error bars are means ± SEM of three independent experiments (n=3). *p < 0.05, **p < 0.01 and ***p < 0.001 by student t-test.

172

Figure 4.9. Comparing C5aR-PA- and C5a- induced gene expressions in rat macrophage and antagonism of C5a responses by 3D53, W54011 and JJ47 at 1h post-treatment. Rat macrophages were treated with (A-E) 0.1 µM C5a or (F-J) 0.3 µM C5aR-PA and cells are lysed after 30 min to stop the reaction. For antagonism of C5aR, cells were pretreated with 1 µM 3D53, W54011 or JJ47 for 1 h. Excess unbound antagonists were removed by washing and subjected to C5a treatment after (A-E) 1 h post- treatment with antagonist. The expressions of (A,F) ATF3, (B,G) TNF, (C,H) IL1B, (D,I) CCL3 and (E,J) PTGS2 were detected by quantitative real-time PCR, normalized against 18S rRNA and converted to fold change relative to control (untreated). Error bars are means ± SEM of four independent experiments (n=4). *p < 0.05, **p < 0.01 and ***p < 0.001 by student t-test.

173 4.5.7 Comparing 3D53, W54011 and JJ47 in C5aR-PA-induced rat paw oedema

The experiments in this section were carried out by research assistant Mr Adam Cotterell (University of Queensland), who performed all the Wistar rat experiments here under my direction and at my request. Since C5aR-PA did not induce large enough responses in the PCR experiments on rat macrophages but C5a did, a pilot study was conducted to determine whether C5aR-PA was able to induce an inflammatory response in male Wistar rats. The left hind paw was given 100 µL saline as baseline control to the right hind paw which was given 350 µg C5aR-PA in 100 µL saline. Both hind paw thickness are recorded over 4 h duration. The maximal swelling induced by 350 µg C5aR-PA was recorded at 1 h and started to subside after 4 h post-injection (Figure 4.10A). Higher amount of C5aR-PA (1 mg and 2 mg) did not further increase the magnitude of paw swelling, therefore the optimized amount of C5aR-PA injected per paw was 350 µg. To antagonize C5aR-PA-induced paw swelling, 5 mg/kg and 10 mg/kg 3D53 were given orally with 500 µL olive 1 h prior to C5aR-PA injection. Treatment with 3D53 at 5 mg/kg reduced paw swelling by ~30% (Figure 4.10B), with no greater effect at 10 mg/kg, and the reduction in swelling was sustained after 2 h post-injection of agonist. On the other hand, W54011 at 5 mg/kg did not block C5aR-PA-induced paw swelling (Figure 4.10C). Even at 10 mg/kg, W54011 did not appear to have an effect until 1h post-injection but the antagonist effect was quickly lost by 2 h post-injection (Figure 4.10C). W54011 has been reported to block C5a-induced gerbil neutropenia with concentrations ranging 3-30 mg/kg (18). Since W54011 at 10 mg/kg did not appear to have much effect, the dose was increased to 30 mg/kg for the next study. No in vivo data for JJ47 was available, so rats were treated at the same dose as W54011 (30 mg/kg). Thus the comparison in vivo experiment involved Wistar rats given 3D53 at 5 mg/kg/p.o. (Figure 4.10C), W54011 and JJ47 at 30 mg/kg/p.o. (Figure 4.10D,E), all 1 h prior to intraplantar injection into rear paws with the inflammogen, 350µg C5aR-PA. At 30 min post-injection, W54011 was more effective than 3D53 and JJ47 with 35% reduction in paw swelling (Figure 4.10E). However, W54011 lost its

174 antagonist activity after 2 h. Even though 3D53 was only given at 5 mg/kg, compared to W54011 and JJ47 at 30 mg/kg, the antagonist effect on paw swelling was maximally sustained from 2-16 hours (Figure 4.10D). Interestingly, JJ47 only caused 15% reduction in paw swelling at 30 min (Figure 4.10F), its antagonist effect (albeit weak) lasting for 2 h, or 4 times longer than W54011. Clearly, 3D53 was a far superior antagonist of C5aR-mediated paw inflammation than the other two compounds.

Figure 4.10. Evaluating antagonism of 3D53, W54011 and JJ47 on C5aR-PA- induced paw oedema in male Wistar rats (courtesy of Mr Adam Cotterell, University of Queensland). Optimization for C5aR-PA-induced paw oedema in male Wistar rats were performed by (A) intraplantar (i.pl.) administration of C5aR-PA (0.35-2 mg per paw in 100 µL of saline control). Paw swelling (% area change from baseline) was recorded over 4 h duration. Both (B) 3D53 and (C) W54011 (5 and 10 mg/kg, p.o. in 500 µL olive oil per rat) given orally 1 h prior 350 µg C5aR-PA injection (n=2 per group). To determine the in vivo residence time of (D) 3D53 (5 mg/kg, p.o., n=3 per group), (E) W54011 (30 mg/kg, p.o., n=2 per group) and (F) JJ47 (30 mg/kg, p.o., n=2 per group) were given orally 1 h prior 350 µg C5aR-PA injection at the specific time-points post- antagonism. Error bars are means ± SEM normalized to maximal swelling of saline control. *p < 0.05, **p < 0.01 and ***p < 0.001 by student t-test.

175 4.6 Discussion

Complement activation ultimately leads to the generation of the potent chemoattractant C5a, which interacts with the receptors C5aR and C5L2, activating important signaling pathways that are crucial for pathogenic defence and maintaining immune homeostasis. However, when C5a is excessively generated as the result of an overly active immune system, one strategy to control its effects without compromising formation of C5b and the pathogen-removing membrane attack complex, is to bind a masking ligand (antagonist) to C5aR (or possibly C5L2). One approved therapeutic for modulating complement is an anti-C5 antibody, but this prevents formation of C5b as well as C5a and thus compromises formation of C5b- derived membrane attack complex (MAC). Besides disabling the most important function of the complement system, formation of MAC, biologics are extremely expensive with poor bioavailability and poor tissue penetration, rendering it impossible for treating chronic inflammatory diseases, which requires repeated dosing. Intensive efforts have been made to develop C5aR antagonist for over two decades but the success rate of a candidate drug moving to the pre-clinical development phase has been disappointingly low. Three previously reported antagonists of C5aR, 3D53, W54011 and JJ47 have been re-evaluated here. Although all antagonists were reported to have low nanomolar antagonist properties against C5a under a defined set of conditions, results in this chapter have shown the danger of making such as conclusion and uncovered important reasons for the failure of W54011 and JJ47 in clinical trials.

First, a key point often overlooked is that IC50 is a concentration-dependent term. Thus reports of IC50 for antagonists can only be compared against the same concentration of agonist on the same cell type under identical experimental conditions. Thus, the differences shown in this chapter for the three antagonist compared are important. The cyclic peptide 3D53 was an insurmountable antagonist of C5aR on human macrophages, and its IC50 was virtually independent of the concentration of C5a (0.1nM to 300 nM); whereas W54011 and JJ47 were competitive and surmountable antagonists of C5aR on human macrophages, and their IC50 values varied by three orders of magnitude in the C5a concentration range 0.1 nM to 300 nM (Figure 4.1, 4.2). Serum levels of C5a can reach as high as 100 nM in patients with

176 sepsis (29). The local concentration of C5a however, is expected to be even higher at sites of inflammation, but this is yet to be determined. These results raise the question as to whether W54011 and JJ47, and other surmountable antagonists can be effective therapeutic agents to treat C5a-mediated diseases. A second key point concerns residence time on the receptor. Human macrophages were treated with each of the antagonists, and then subsets of these cell populations were treated with rhC5a at different times to compare the duration of action of each of the antagonists. A striking difference was observed in duration of antagonist action, with 3D53 action being an order of magnitude longer than for

W54011 and JJ47 (ca. t1/2 20h vs 2h). This difference is a reflection of a much longer residence time on the receptor C5aR for 3D53, which does not bind at all to C5L2 (27). The interpretation of this data is that 3D53 would require once a day administration, whereas W54011 and JJ47 would be required on a 1-2 hourly basis, not to mention in larger quantities because they would be rendered ineffective as the serum or local tissue concentration of C5a increases. Clearly, a non-competitive antagonist has advantages over competitive antagonists. The important properties conferred by longer residence time are not widely appreciated and often overlooked in drug design and development. Residence time is beginning to make some impact on some drug discovery programs, the recently reported being a prime example of why longer residence time is preferred as there were already twenty kinase inhibitors in advanced clinical trials or in man (40). Increasing residence time for drugs can overcome less optimal pharmacokinetic profiles that really only report on plasma levels of circulating drugs. The third important point concern a very misleading pharmacokinetic measurement F% or oral bioavailability, which is currently considered the most important property in drug design and development. When a drug is administered intravenously, bioavailability is 100%. Since drugs that are administered orally or via other routes have to cross membrane barriers before exerting effects in systemic circulation, their oral bioavailability is usually much lower. F% is the fraction of an orally administered drug that reaches systemic circulation against an intravenously administered drug. A high F% means that a lower dose is necessary to achieve a desired therapeutic effect and therefore potentially reduces the risks of side effects or off-target effects and toxicity. Drugs with poor F% on the other hand can result in low oral efficacy, higher dose required and higher variability between individuals, leading

177 to unpredictable drug responses. The current method used to determine oral bioavailability is liquid chromatography-mass spectrometry (LC-MS). The concept is the theory that blood levels of a drug correlate with efficacy, but this can be fundamentally incorrect if the drug targets cell surface receptors like GPCRs. For example, the reported F% value for 3D53 is 1% (22), while the reported F% values for W54011 and JJ47 are as high as 21% (18, 19). If these values are truly an estimate of the amount of drug exerting its effect on its targeted receptor, 3D53 will be an inferior antagonist compared to W54011 and JJ47 based on F%. However, in vivo studies and data from clinical trials reported so far have proven otherwise. Blood collected from treated animals is centrifuged, and only the plasma is kept for analysis. Since it is chemically not possible to separate and isolate drugs bound to all targeted cell receptors, blood cells and unwanted serum proteins are discarded before loading to the LC-MS in order to reduce background noise. As shown earlier,

3D53 has a very slow relative off-rate (t1/2 = 18.2 h) compared to W54011 (t1/2 = 1.2 h) and JJ47 (t1/2 = 0.6 h). It is expected that the majority of 3D53 is bound for a long time to C5aR expressed throughout the body, including in organs and tissues. Since the current protocol only determines oral bioavailability based on drugs present in plasma, this is a major flaw for drugs that have a fast on-rate (fast clearance from blood) and a slow off-rate. Thus the reported F% for 3D53 is clearly an underestimate for the actual amount of bioavailable drug, and ignores the efficiency of oral 3D53 in quickly targeting C5aR. A fast clearance rate from plasma combined with and a low F%, are commonly interpreted as resulting in low efficacy, a conclusion that is clearly wrong as shown here for 3D53. The affinity of a ligand for its receptor does not, per se, define the effectiveness and duration of biological action. Rather, it is the lifetime of the binary receptor-ligand complex that in part dictates the effect in the cellular and organismal context. If the drug is circulating in the blood, unbound to its targeted receptor, it is not exerting its effect and this needs to be considered more carefully in drug optimization studies. W54011 has been previously reported to inhibit C5a-induced calcium release in human, cynomolgus monkey and gerbils neutrophils, but not in mice, rats, guinea pigs, rabbits and dogs neutrophils (18). However, in this study, W54011 clearly had antagonist effect against C5a-mediated gene expression in rat macrophages (Figure 4.9) as well as in vivo in the Wistar rat models of C5aR-PA-induced paw oedema (Figure 4.10E). W54011 has also been shown to be a C5aR antagonist in mouse

178 dendritic cells (41). These results suggest that the interpretation made by Sumichika et. al., that W54011 has only a narrow cross-species reactivity may not be true, at least not for rats and mice. The correct interpretation from their results should be that human C5a only triggers calcium mobilization in human, cynomolgus monkey and gerbils neutrophils, but not in mouse, rats, guinea pigs, rabbits or dogs neutrophils. In this study, human C5a failed to stimulate calcium release, but did stimulated ERK phosphorylation and amplified C5a-related gene expression, in Wistar rat macrophages. Peptide analogues of the C5a anaphylatoxin were based on 8 residues at the C- terminus of human C5a but only had modest binding activity (33). Peptide agonist YSFKPMPLdAR was later discovered by Sam Sanderson at the University of Nebraska with improved biological activity, but still in the high µM range (42). Although this decapeptide was a full agonist on C5aR (43), it also bound non- specifically to C3aR at µM concentrations (38). Structure activity studies led by researchers from Abbott Laboratories over a decade in the 1980s, produced analogues with better affinity to C5aR, while downsizing the peptide to six residues, N-Methyl- FKPdChaCha-dArg-OH, with nM activity against C5aR (44). Although this hexapeptide agonist was previously reported to bind with low affinity to C5L2 (45), but did not affect C5a binding to C5L2 even at high µM concentrations, separate studies have now shown otherwise (27). By swapping the N-methyl to an acetyl cap, C5aR-PA mimicked the native C5a protein more closely as determined by NMR spectroscopy (unpublished data by Dr Martin Stoermer). Acetylating peptides can result in increased metabolic stability and improved in vivo efficacy (36, 37, 46). Human C5aR is 64.9% identical in amino acid sequence to rat C5aR (Figure 4.11A). Despite the high similarity between human and rat C5aRs, there are a handful of non-conserved residues within the putative transmembrane region of C5aRs where C5a or ligands are expected to bind. A follow-up study on W54011 investigated why this antagonist exerted its antagonism across limited species and found that the underlying reason was due to the Trp (W) residue in the transmembrane domain V (47). Results from this study and others (41), have demonstrated that W54011 in fact can exert its antagonistic effects in mice and rat models, thereby refuting claims for a species-specific effect of W54011.

179

Figure 4.11. Amino acid sequence comparison of human and rat C5aR and C5a. Sequence alignment between human and rat (A) C5aRs and (B) C5a anaphylatoxins were performed using online software ClustalW2. Consensus symbols (*) indicates fully conserved residue, (:) indicates residues with strongly similar properties, (.) indicates residues with weakly similar properties. Residues with no symbols are not similar. Putative transmembrane domains are indicated by solid black lines (48). Residues previously reported responsible for agonistic activity of C5a are indicated by black open box (33). Genbank ID for human and rat C5aRs are gi:4502509 and gi:16758418; for C5a are gi:227968182 and gi:913781 respectively. Sequence similarity score between human and rat C5aRs: 64.9% and C5a: 63.5%.

Although C5aR-PA had full agonist activity on human macrophages, it appeared to be a partial agonist on rat macrophages. C5aR-PA did not induce calcium release in rat macrophages, but did induce paw oedema in rat models. Scola et. al., have previously identified that human recombinant C5a did not bind very well to rat C5aR but rat recombinant C5a on the other hand, bind efficiently to human C5aR as well as rat C5aR (27). Since C5aR-PA was designed based on human C5aR, it is expected to exert similar effects as human C5a. Even though human and rat C5a are 63.5% identical, the eight residues at C-terminal that are responsible for C5a agonist responses are conserved except for one residue (Figure 4.11B). Since human C5a can

180 activate gene expressions and ERK phosphorylation, but not calcium release, in rat macrophages, could this residue (a single change from Gln to Leu) be important for activating calcium release? If so, would not it be also possible to develop a C5aR biased ligands that could block certain C5a-mediated signalling pathways while leaving other pathways functional and responsive to C5a?

4.7 Conclusions

This chapter is a comparative study of three reported antagonists of C5aR, with the goal of selecting one of three known antagonists of C5a action to probe C5a function in monocytes and macrophages in Chapter 5. This chapter has demonstrated that the residence time of an antagonist measured in vitro in cells translates into, and is vitally important for, in vivo efficacy. Drug discovery has been traditionally focused on optimizing for receptor affinity, drug potency, receptor selectivity and drug-like properties to maximise oral bioavailability and pharmacokinetics for a given compound series. However, these properties do not necessarily maximise in vivo efficacy (16), and this chapter has highlighted this point. At the molecular level, drug action is usually effected when the drug is bound to its molecular target, in this case a GPCR. In principle, the longer that the receptor-antagonist complex is maintained intact, the longer the intended antagonist effect that is produced. This fundamental property of a ligand, residence time on the receptor, is not a new concept but perhaps has not been appreciated to the extent it needs to be. Both the pharmaceutical drug industry and academic researchers in the field of drug discovery face tremendous challenges in taking new and better drugs to the market. However, perhaps the biggest problem faced is a less realised one, have we lost sight of the importance of basic biology? Time is of the essence in drug development, but this is also especially true in drug efficacy where residence time of a drug on its specific receptor can also be critical in determining the duration of drug action

181 4.8 References

1. Hugli, T. E. 1981. The structural basis for anaphylatoxin and chemotactic functions of C3a, C4a, and C5a. Crit. Rev. Immunol. 1: 321-366. 2. Monk, P. N., A. M. Scola, P. Madala, and D. P. Fairlie. 2007. Function, structure and therapeutic potential of complement C5a receptors. Br. J. Pharmacol. 152: 429-448. 3. Okinaga, S., D. Slattery, A. Humbles, Z. Zsengeller, O. Morteau, M. B. Kinrade, R. M. Brodbeck, J. E. Krause, H. R. Choe, N. P. Gerard, and C. Gerard. 2003. C5L2, a nonsignaling C5A binding protein. Biochemistry (Mosc.) 42: 9406-9415. 4. Buchner, R. R., T. E. Hugli, J. A. Ember, and E. L. Morgan. 1995. Expression of functional receptors for human C5a anaphylatoxin (CD88) on the human hepatocellular carcinoma cell line HepG2. Stimulation of acute-phase protein- specific mRNA and protein synthesis by human C5a anaphylatoxin. J. Immunol. 155: 308-315. 5. Van Beek, J., M. Bernaudin, E. Petit, P. Gasque, A. Nouvelot, E. T. MacKenzie, and M. Fontaine. 2000. Expression of receptors for complement anaphylatoxins C3a and C5a following permanent focal cerebral ischemia in the mouse. Exp. Neurol. 161: 373-382. 6. Wetsel, R. A. 1995. Expression of the complement C5a anaphylatoxin receptor (C5aR) on non-myeloid cells. Immunol. Lett. 44: 183-187. 7. Haviland, D. L., R. L. McCoy, W. T. Whitehead, H. Akama, E. P. Molmenti, A. Brown, J. C. Haviland, W. C. Parks, D. H. Perlmutter, and R. A. Wetsel. 1995. Cellular expression of the C5a anaphylatoxin receptor (C5aR): demonstration of C5aR on nonmyeloid cells of the liver and lung. J. Immunol. 154: 1861-1869. 8. Lim, J., A. Iyer, J. Y. Suen, V. Seow, R. C. Reid, L. Brown, and D. P. Fairlie. 2013. C5aR and C3aR antagonists each inhibit diet-induced obesity, metabolic dysfunction, and adipocyte and macrophage signaling. FASEB J. 27: 822-831. 9. Nitta, H., Y. Wada, Y. Kawano, Y. Murakami, A. Irie, K. Taniguchi, K. Kikuchi, G. Yamada, K. Suzuki, J. Honda, M. Wilson-Morifuji, N. Araki, M. Eto, H. Baba, and T. Imamura. 2013. Enhancement of human cancer cell

182 motility and invasiveness by anaphylatoxin C5a via aberrantly expressed C5a receptor (CD88). Clin. Cancer Res. 19: 2004-2013. 10. Markiewski, M. M., R. A. DeAngelis, F. Benencia, S. K. Ricklin-Lichtsteiner, A. Koutoulaki, C. Gerard, G. Coukos, and J. D. Lambris. 2008. Modulation of the antitumor immune response by complement. Nat. Immunol. 9: 1225-1235. 11. Markiewski, M. M., and J. D. Lambris. 2009. Unwelcome complement. Cancer Res. 69: 6367-6370. 12. Makrides, S. C. 1998. Therapeutic inhibition of the complement system. Pharmacol. Rev. 50: 59-87. 13. Rother, R. P., S. A. Rollins, C. F. Mojcik, R. A. Brodsky, and L. Bell. 2007. Discovery and development of the complement inhibitor eculizumab for the treatment of paroxysmal nocturnal hemoglobinuria. Nat. Biotechnol. 25: 1256- 1264. 14. Asghar, S. S. 1984. Pharmacological manipulation of complement system. Pharmacol. Rev. 36: 223-244. 15. Qu, H., D. Ricklin, and J. D. Lambris. 2009. Recent developments in low molecular weight complement inhibitors. Mol. Immunol. 47: 185-195. 16. Kola, I., and J. Landis. 2004. Can the pharmaceutical industry reduce attrition rates? Nat. Rev. Drug Discov. 3: 711-715. 17. March, D. R., L. M. Proctor, M. J. Stoermer, R. Sbaglia, G. Abbenante, R. C. Reid, T. M. Woodruff, K. Wadi, N. Paczkowski, J. D. Tyndall, S. M. Taylor, and D. P. Fairlie. 2004. Potent cyclic antagonists of the complement C5a receptor on human polymorphonuclear leukocytes. Relationships between structures and activity. Mol. Pharmacol. 65: 868-879. 18. Sumichika, H., K. Sakata, N. Sato, S. Takeshita, S. Ishibuchi, M. Nakamura, T. Kamahori, S. Ehara, K. Itoh, T. Ohtsuka, T. Ohbora, T. Mishina, H. Komatsu, and Y. Naka. 2002. Identification of a potent and orally active non-peptide C5a receptor antagonist. J. Biol. Chem. 277: 49403-49407. 19. Gong, Y., J. K. Barbay, M. Buntinx, J. Li, J. V. Wauwe, C. Claes, G. V. Lommen, P. J. Hornby, and W. He. 2008. Design and optimization of aniline- substituted tetrahydroquinoline C5a receptor antagonists. Bioorg. Med. Chem. Lett. 18: 3852-3855. 20. Konteatis, Z. D., S. J. Siciliano, G. Van Riper, C. J. Molineaux, S. Pandya, P. Fischer, H. Rosen, R. A. Mumford, and M. S. Springer. 1994. Development of

183 C5a receptor antagonists. Differential loss of functional responses. J. Immunol. 153: 4200-4205. 21. Wong, A. K., A. M. Finch, G. K. Pierens, D. J. Craik, S. M. Taylor, and D. P. Fairlie. 1998. Small molecular probes for G-protein-coupled C5a receptors: conformationally constrained antagonists derived from the C terminus of the human plasma protein C5a. J. Med. Chem. 41: 3417-3425. 22. Kohl, J. 2006. Drug evaluation: the C5a receptor antagonist PMX-53. Curr. Opin. Mol. Ther. 8: 529-538. 23. Leeson, P. 2012. Drug discovery: Chemical beauty contest. Nature 481: 455- 456. 24. Lipinski, C. A., F. Lombardo, B. W. Dominy, and P. J. Feeney. 2001. Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. Adv Drug Deliv Rev 46: 3-26. 25. Finch, A. M., A. K. Wong, N. J. Paczkowski, S. K. Wadi, D. J. Craik, D. P. Fairlie, and S. M. Taylor. 1999. Low-molecular-weight peptidic and cyclic antagonists of the receptor for the complement factor C5a. J. Med. Chem. 42: 1965-1974. 26. Reid, R. C., G. Abbenante, S. M. Taylor, and D. P. Fairlie. 2003. A convergent solution-phase synthesis of the macrocycle Ac-Phe-[Orn-Pro-D- Cha-Trp-Arg], a potent new antiinflammatory drug. J. Org. Chem. 68: 4464- 4471. 27. Scola, A. M., A. Higginbottom, L. J. Partridge, R. C. Reid, T. Woodruff, S. M. Taylor, D. P. Fairlie, and P. N. Monk. 2007. The role of the N-terminal domain of the complement fragment receptor C5L2 in ligand binding. J. Biol. Chem. 282: 3664-3671. 28. Arunlakshana, O., and H. O. Schild. 1959. Some quantitative uses of drug antagonists. British journal of pharmacology and chemotherapy 14: 48-58. 29. Solomkin, J. S., M. K. Jenkins, R. D. Nelson, D. Chenoweth, and R. L. Simmons. 1981. Neutrophil dysfunction in sepsis. II. Evidence for the role of complement activation products in cellular deactivation. Surgery 90: 319-327. 30. Speidl, W. S., M. Exner, J. Amighi, S. P. Kastl, G. Zorn, G. Maurer, O. Wagner, K. Huber, E. Minar, J. Wojta, and M. Schillinger. 2005. Complement

184 component C5a predicts future cardiovascular events in patients with advanced atherosclerosis. Eur. Heart J. 26: 2294-2299. 31. Bokisch, V. A., and H. J. Muller-Eberhard. 1970. Anaphylatoxin inactivator of human plasma: its isolation and characterization as a carboxypeptidase. J. Clin. Invest. 49: 2427-2436. 32. Hugli, T. E., and H. J. Muller-Eberhard. 1978. Anaphylatoxins: C3a and C5a. Adv. Immunol. 26: 1-53. 33. Kawai, M., D. A. Quincy, B. Lane, K. W. Mollison, J. R. Luly, and G. W. Carter. 1991. Identification and synthesis of a receptor binding site of human anaphylatoxin C5a. J. Med. Chem. 34: 2068-2071. 34. Higginbottom, A., S. A. Cain, T. M. Woodruff, L. M. Proctor, P. K. Madala, J. D. Tyndall, S. M. Taylor, D. P. Fairlie, and P. N. Monk. 2005. Comparative agonist/antagonist responses in mutant human C5a receptors define the ligand binding site. J. Biol. Chem. 280: 17831-17840. 35. Ruiz-Gomez, G., J. D. Tyndall, B. Pfeiffer, G. Abbenante, and D. P. Fairlie. 2010. Update 1 of: Over one hundred peptide-activated G protein-coupled receptors recognize ligands with turn structure. Chem. Rev. 110: PR1-41. 36. Adessi, C., and C. Soto. 2002. Converting a peptide into a drug: strategies to improve stability and bioavailability. Curr. Med. Chem. 9: 963-978. 37. Drapeau, G., S. Brochu, D. Godin, L. Levesque, F. Rioux, and F. Marceau. 1993. Synthetic C5a receptor agonists. Pharmacology, metabolism and in vivo cardiovascular and hematologic effects. Biochem. Pharmacol. 45: 1289-1299. 38. Proctor, L. M., T. V. Arumugam, I. Shiels, R. C. Reid, D. P. Fairlie, and S. M. Taylor. 2004. Comparative anti-inflammatory activities of antagonists to C3a and C5a receptors in a rat model of intestinal ischaemia/reperfusion injury. Br. J. Pharmacol. 142: 756-764. 39. Ames, R. S., Y. Li, H. M. Sarau, P. Nuthulaganti, J. J. Foley, C. Ellis, Z. Zeng, K. Su, A. J. Jurewicz, R. P. Hertzberg, D. J. Bergsma, and C. Kumar. 1996. Molecular cloning and characterization of the human anaphylatoxin C3a receptor. J. Biol. Chem. 271: 20231-20234. 40. Sanderson, K. 2013. Irreversible kinase inhibitors gain traction. Nat. Rev. Drug Discov. 12: 649-651.

185 41. Peng, Q., K. Li, N. Wang, Q. Li, E. Asgari, B. Lu, T. M. Woodruff, S. H. Sacks, and W. Zhou. 2009. Dendritic cell function in allostimulation is modulated by C5aR signaling. J. Immunol. 183: 6058-6068. 42. Ember, J. A., S. D. Sanderson, S. M. Taylor, M. Kawahara, and T. E. Hugli. 1992. Biologic activity of synthetic analogues of C5a anaphylatoxin. J. Immunol. 148: 3165-3173. 43. Finch, A. M., S. M. Vogen, S. A. Sherman, L. Kirnarsky, S. M. Taylor, and S. D. Sanderson. 1997. Biologically active conformer of the effector region of human C5a and modulatory effects of N-terminal receptor binding determinants on activity. J. Med. Chem. 40: 877-884. 44. Kawai, M. W., PE.; Luly, JR.; Or, YS. . 1992. New hexa - and heptapeptides are anaphylatoxin antagonists and agonists - for treating inflammatory and immunodeficiency diseases, cancers and severe infections. W. I. P. Organisation, ed. Abbott Laboratories, USA. 45. Otto, M., H. Hawlisch, P. N. Monk, M. Muller, A. Klos, C. L. Karp, and J. Kohl. 2004. C5a mutants are potent antagonists of the C5a receptor (CD88) and of C5L2: position 69 is the locus that determines agonism or antagonism. J. Biol. Chem. 279: 142-151. 46. Arispe, N., J. C. Diaz, and M. Flora. 2008. Efficiency of histidine-associating compounds for blocking the alzheimer's Abeta channel activity and cytotoxicity. Biophys. J. 95: 4879-4889. 47. Waters, S. M., R. M. Brodbeck, J. Steflik, J. Yu, C. Baltazar, A. E. Peck, D. Severance, L. Y. Zhang, K. Currie, B. L. Chenard, A. J. Hutchison, G. Maynard, and J. E. Krause. 2005. Molecular characterization of the gerbil C5a receptor and identification of a transmembrane domain V amino acid that is crucial for small molecule antagonist interaction. J. Biol. Chem. 280: 40617- 40623. 48. Baldwin, J. M. 1993. The probable arrangement of the helices in G protein- coupled receptors. EMBO J. 12: 1693-1703.

186 Chapter

5 Inflammatory responses induced by LPS are amplified in human monocytes but suppressed in human macrophages by C5a

5.1 Abstract ...... 188 5.2 Introduction ...... 189 5.3 Aims ...... 191 5.4 Materials and Methods ...... 192 5.4.1 Reagents ...... 192 5.4.2 Isolation of human monocytes and macrophages ...... 192 5.4.3 Quantitative cytokine ELISA ...... 193 5.4.4 Isolation of mRNA and RT-PCR ...... 193 5.4.5 Gene expression analysis by real-time PCR ...... 193 5.4.6 Quantification of MEK1, ERK1/2, p38 and JNK phosphorylation by AlphaScreen Surefire ...... 194 5.4.7 Western blot analysis ...... 194 5.4.8 Bacterial clearance assay ...... 194 5.4.9 Statistical analysis ...... 195 5.5 Results ...... 196 5.5.1 C5a acting via C5aR differentially modulates LPS-signalling in human CD14+ monocytes versus macrophages...... 196 5.5.2 Temporal study of C5aR/TLR4 crosstalk in HMDMs ...... 197 5.5.3 C5a suppresses LPS-induced IL6 and TNF via Gαi- and MEK-dependent signalling...... 200 5.5.4 C5a and LPS synergize for Raf/MEK/ERK phosphorylation in macrophages 203 5.5.5 C5a and LPS synergise in phosphorylation of ERK1/2 but not p38 or JNK in macrophages spacing ...... 203 5.5.6 C5a-mediated suppression of LPS-induced IL6 and TNF is IL10- and PI3K- independent...... 206 5.5.7 C5a attenuation of LPS-induced IL6 is selectively TLR4-mediated in macrophages ...... 209 5.5.8 C5a enhances clearance of Salmonella Typhimurium from HMDM ...... 211 5.6 Discussion ...... 212 5.7 Conclusions ...... 215 5.8 References ...... 216

187 5.1 Abstract

Monocytes and macrophages are important innate immune cells equipped with danger sensing receptors, including complement and Toll-like receptors. Complement protein C5a, acting via C5aR, is shown here to differentially modulate LPS-induced inflammatory responses in primary human monocytes versus macrophages. While C5a enhanced secretion of LPS-induced IL6 and TNF from primary human monocytes, C5a inhibited these responses while increasing IL10 secretion in donor- matched human monocyte-derived macrophages differentiated by GM-CSF or M-CSF.

C5a induced Gαi/c-Raf/MEK/ERK signaling was amplified in macrophages but not in monocytes by LPS. Accordingly, the Gαi inhibitor PTX and MEK inhibitor U0126 blocked the inhibition by C5a of LPS-induced IL6 and TNF production from macrophages. This synergy was independent of IL10, PI3K, p38, JNK and the differentiating agent. Since C5a did not inhibit IL6 production from macrophages induced by other TLR agonists that are selective for TRIF (poly I:C) or MyD88 (imiquimod), C5a selectively regulates LPS-mediated responses. Furthermore, this C5a-mediated suppression of pro-inflammatory cytokines IL6 and TNF in macrophages did not compromise antimicrobial activity; C5a instead enhanced clearance of the gram-negative bacterial pathogen Salmonella enterica serovar Typhimurium from macrophages. These findings implicate C5aR as a regulatory switch that modulates TLR4 signaling via the Gαi/c-Raf/MEK/ERK signaling axis in human macrophages but not in human monocytes. The differential effects of C5a are consistent with amplifying monocyte proinflammatory responses to systemic danger signals, but attenuating macrophage cytokine responses (without compromising microbicidal activity) thereby restraining inflammatory responses to localized infections.

188 5.2 Introduction

Complement is an important cascading network of over 30 soluble plasma, and insoluble membrane-bound, proteins that cooperate in innate immunity in the recognition, opsonization, destruction, and removal of pathogens and infected or damaged cells. Complement activation by classical, alternative, lectin pathways and other routes is tightly controlled, producing the lytic membrane attack complex (MAC), as well as opsonins such as C3b and proinflammatory anaphylatoxins C3a, C4a and C5a. Elevated plasma levels of C5a or its receptor are associated with inflammatory diseases such as rheumatoid arthritis, inflammatory bowel diseases, respiratory distress syndrome, ischemia-reperfusion injury and sepsis (1). C5a binds to and signals through the receptor C5aR (CD88), a class A rhodopsin-like G protein- coupled receptor (2), and to a second receptor C5L2 which does not couple G proteins, may be non-signalling but may also modulate C5aR (3, 4). C5aR couples to the pertussis toxin-sensitive Gαi, and in some cells like monocytes to the pertussis toxin- insensitive Gα16, and recruits β-arrestins-1,2 to the plasma membrane (1, 5). C5a activates downstream several signalling pathways, including Akt, MAPK/ERK kinase (MEK), phosphatidylinositol 3-kinase (PI3K), phosphokinase A (PKA), phospholipase C (PLC) and NF-κB (1). The Toll-like receptors (TLRs) are an important family of type I transmembrane proteins that respond to infectious organisms and danger signals by initiating inflammatory responses, activating microbial killing and clearance mechanisms, and priming the adaptive immune response (6). Of ten known human TLRs, TLR4 responds to lipopolysaccharide (LPS) from Gram-negative bacterial cell walls. TLR4 can signal through two different pathways dependent upon, and modulated by, separate adaptor proteins MyD88 or TRIF (7). MyD88 activates IRAKs (IL1R-associated kinases) and TRAF6 (TNF-receptor-associated factor 6), as well as transcription factors NF-κB, AP-1 and IRF5 further downstream. TRIF signals the induction of type I IFNs by recruiting TRAF3 and RIP1 to activate transcription factor IRF3, NF-κB and AP-1. Both MyD88 and TRIF activate AP-1 via downstream MAPK activation (8), although NF-κB and MAPK activation by TRIF occurs later than for MyD88 (9).

189 Crosstalk between different signalling pathways in mammalian cells can result in unexpected and unique functional outcomes (10). Some pathogens have evolved to evade or exploit host microbial-killing by targeting C5aR and TLRs specifically. For example, virulence proteins secreted by Staphylococcus aureus bind to C5aR with potent antagonist activity, preventing recruitment of phagocytes to sites of infection (11). Porphyromonas gingivalis exploits the C5aR/TLR2 crosstalk by increasing cAMP production in macrophages, which suppresses macrophage immune function and enhances pathogen survival (12). Clearly, these pathogens have evolutionarily adapted to exploit complement/TLR crosstalk to their advantage. While C5aR and TLR4 signalling have been separately investigated in many studies, there is comparatively little reported about their interplay (13). Among recent reports on C5aR/TLR4 crosstalk, C5a has been found to downregulate TLR4-induced IL12 production via PI3K-dependent (14) and PI3K-independent (15) pathways in mouse macrophages; to signal via the PI3K-Akt pathway to enhance LPS-induced IL17F cytokine production (16), and to suppress LPS-induced IL17A and IL23 cytokine production in mouse macrophages (17). C5a-induced suppression of LPS-inducible IL17A/IL23 axis was reported to occur via enhanced IL10 production and ERK1/2 phosphorylation. However, most studies reporting C5a modulation of TLR4 signalling had been performed with murine rather than human cells (18, 19). Some interpretations of human disease mechanisms have been made based on extrapolating observations from mouse studies. However, there are important differences in LPS- induced TLR4 signaling between humans and mice (20-22), thus necessitating careful assessment of C5aR/TLR4 crosstalk in human cells. Monocytes and macrophages are important cellular mediators of innate immunity and are targets for C5a and LPS. Some features of LPS signalling are also distinctly different between human monocytes and macrophages (23). Monocytes act as danger sensors in the circulation and typically generate a more rapid, heightened inflammatory response than macrophages to danger signals. For example, LPS alone can trigger inflammasome activation in monocytes, whereas an additional activation signal such as ATP is required in macrophages (24). The present study compares effects of C5a on LPS responses in primary human monocytes versus donor-matched macrophages derived through differentiating monocytes with either GM-CSF or M- CSF. Surprisingly, we found that C5a differentially modulated LPS responses depending upon the cell type, that C5a activated the Gαi/c-Raf/MEK/ERK axis to

190 selectively downregulate pro-inflammatory cytokine production from macrophages but not monocytes, and that effects on macrophages do not compromise but instead stimulate pathogen killing. Thus, it is proposed that C5a has distinctly different regulatory effects on myeloid cell functions during localized versus systemic inflammatory responses. In the present study, crosstalk between C5aR and TLR4 was investigated in human monocytes and human monocyte-derived macrophages (HMDM) using the respective ligands C5a and LPS.

5.3 Aims

While C5aR and TLR4 signalling had been separately investigated in many studies (10), there was comparatively little reported about their interplay. Since there had been some conflicting reports about the interplay between LPS and C5a in mouse vs human cell studies (20-23), we sought to clarify this interplay for human cells. Therefore, we sought to discover whether complement protein C5a, acting via C5aR, differentially modulated lipopolysaccharide (LPS)-induced inflammatory responses in primary human monocytes versus human macrophages. The specific aims of this chapter were: 1. To determine if LPS-induced responses are differentially modulated by C5a in human monocytes versus macrophages 2. To identify which signalling pathways are involved in the C5aR/TLR4 crosstalk in monocytes versus macrophages 3. To find out where signalling cascades intersect in monocytes and macrophages to mediate C5aR/TLR4 crosstalk

The results discussed in this chapter have also been accepted in the Journal of Immunology (Appendix IV).

191 5.4 Materials and Methods

5.4.1 Reagents

Recombinant human GM-CSF, M-CSF and IL10 were purchased from PeproTech. LPS from Salmonella enterica (serotype Minnesota RE 595), U0126 monothanolate, wortmannin and pertussis toxin (PTX) were purchased from Sigma- Aldrich. Imiquimod R837 (IMQ) and polyinosinic-polycytidylic acid (poly(I:C)) were purchased from InvivoGen. Human C5a was purchased from Sino Biological Inc. Goat neutralizing antibody against human IL10 (AB-217-NA) was purchased from R&D Systems. C5aR antagonist (3D53 also licensed as PMX53) (1, 25) was synthesized and characterized (analytical HPLC, mass spectroscopy and NMR spectroscopy) in-house as described (26). Antibodies against phosphorylated c-Raf and total c-Raf were purchased from Cell Signaling Technology. All cell culture reagents were purchased from Invitrogen and all analytical grade chemical reagents were obtained from Sigma-Aldrich, unless otherwise stated.

5.4.2 Isolation of human monocytes and macrophages

PBMCs were isolated from buffy coats of anonymous donors (Australian Red Cross Blood Service, Kelvin Grove, Queensland, Australia) by density centrifugation using Ficoll-Paque Plus (GE Healthcare) manufacturer’s instructions. Contaminating erythrocytes were removed by repeated ice-cold sterile water. CD14+ MACS® microbeads were used to positively select monocytes from isolated PBMCs according to the manufacturer’s instructions (Miltenyi Biotec). CD14+ monocytes were seeded at 1 x 106 cells/mL in IMDM supplemented with 10% FBS, 100 U/mL penicillin, 100 µg/mL streptomycin and 2 mM GlutaMAX at 37°C in presence of 5% CO2. To generate human monocyte-derived macophages (HMDM), CD14+ monocytes were cultured in the presence of either 10 ng/mL GM-CSF (GM-MΦ) or M-CSF (M-MΦ) for at least 6 days.

192 5.4.3 Quantitative cytokine ELISA

Cells were seeded at 1 x 106 cells/mL and serum-deprived overnight in the incubator prior to treatment. LPS and C5a treatments were pre-mixed prior to 24 h stimulation. Cell culture supernatants were collected and cytokines level were determined using specific ELISA sets from BD Pharmingen, according to the manufacturer’s instructions.

5.4.4 Isolation of mRNA and RT-PCR

Total RNA was extracted from cells using an RNeasy Mini Plus kit (Qiagen), according to the manufacturer’s instructions. Total cellular RNA (2-10 µg) was reverse transcribed to cDNA using oligo dT primer (1 µg) and Superscript III (Invitrogen), according to the manufacturer’s instructions. cDNA samples were stored at -20°C until further use.

5.4.5 Gene expression analysis by real-time PCR

Gene-specific primers were designed using the free web-based software Primer-BLAST (National Center for Biotechnology Information). Quantitative real time-PCR was performed using an ABI PRISM® 7900 Real Time PCR System (Applied Biosystems™), cDNA (50 ng), SYBR® Green PCR master mix (Applied Biosystems™) and gene-specific primers. Relative gene expression was normalised against 18S rRNA expression and then converted to fold change against control samples. All samples were analysed in duplicate. Primer sequences used are: CD68, 5’-TTTGGGTGAGGCGGTTCAG-3’ and 3’-CCAGTGCTCTCTGCCAGTA-5’; 18S, 5’-ACCACGGGTGACGGGGAATC-3’ and 3’-CCGGGTCGGGAGTGGGTAAT-5’.

193 5.4.6 Quantification of MEK1, ERK1/2, p38 and JNK phosphorylation by AlphaScreen Surefire

Ser218/Ser222), phosphor-ERK1/2 (phosphor-Thr202/Tyr204), phosphor-p38

(phosphor-Thr180/Tyr182) and phosphor-JNK (phosphor-Thr183/Tyr185) according to Alphascreen SureFire assays (PerkinElmer) manufacturer’s instructions. At completion of the assay, signal in the wells was measured using Pherastar (BMG

Labtech).

5.4.7 Western blot analysis

Samples were homogenized in 50 mM Tris (pH 7.4), 1.7 mM SDS and protease inhibitor cocktail (Roche Applied Science). Equal amounts of total cell lysates (10 µg) were loaded and subjected to denaturing SDS-PAGE, transferred to PVDF membranes, and proteins were detected using antibodies against phosphorylated and total C-Raf protein (Cell Signalling Technology). Bands were visualized by an ECL method and quantified by densitometry measurements of phospho c-Raf normalized with total c-Raf against control.

5.4.8 Bacterial clearance assay

Intracellular survival of Salmonella enterica serovar Typhimurium (S. Typhimurium) strain SL1344 within HMDMs was monitored in gentamicin exclusion assays (27). M-MΦ were seeded at 200,000 cells/well in penicillin-streptomycin (P/S)

194 free media containing M-CSF (10 ng/mL). On the following day, cells were infected with S. Typhimurium at a multiplicity of infection (MOI) of 100. One hour post- infection, cells were washed with gentamicin (200 µg/mL) in P/S-free media and further incubated (1 h) to kill extracellular bacteria in the same media. Additional incubations beyond this 2 h time point were carried out in media containing 20 µg/mL gentamicin. At specific time points, cells were washed twice and lysed with 0.01% Triton X-100 in PBS, after which lysates were cultured overnight on LB agar. Colony counts were performed to assess intracellular bacterial loads. The MOI was also confirmed by plating out inoculum on LB agar plates and performing colony counts. Treatments with C5a (30 nM) were conducted by co-treatment with bacteria upon infection. For antagonism assays, cells were pretreated with the C5a inhibitor 3D53 (1 µM) 30 min prior to infection. 3D53 was also added back following the washing with gentamicin-containing media.

5.4.9 Statistical analysis

195 5.5 Results

5.5.1 C5a acting via C5aR differentially modulates LPS-signalling in human CD14+ monocytes versus macrophages.

In light of recent discoveries linking crosstalk between Toll-like receptors and the complement system as possible control points for regulation of host defence (13), this study compared C5aR/TLR4 crosstalk between human monocytes and macrophages, given their different roles in responding to inflammatory stimuli. C5a decreased LPS-induced secreted levels of IL6 and TNF in both GM-MΦ and M-MΦ, but increased the levels of the same proinflammatory cytokines in monocytes (Figure. 5.1A, B). On the other hand, C5a enhanced LPS-induced IL10 production from macrophages, but had no effect on this response in monocytes (Figure. 5.1C). This suggests that the regulatory effects of C5aR on TLR4 depend upon cellular context, for example differentiation status. Although C5aR mediates C5a-dependent responses, a second C5a receptor, C5L2, has also been reported to exhibit anti-inflammatory functions (3, 28). However, the biological roles of C5L2 in inflammation are highly controversial. Our group has previously developed 3D53, a potent, insurmountable and selective antagonist of C5aR that does not bind C5L2 (29, 30). 3D53 blocked the immunomodulatory effects of C5a on TLR4 signalling in both macrophages and monocytes (Figure. 5.1D-F), thus demonstrating that the observed responses of C5a were mediated via C5aR, rather than C5L2.

196

Figure 5.1. C5a acting via C5aR-dose-dependently modulates LPS-induced IL6, TNF and IL10 cytokine production in human monocytes and macrophages (GM- MΦ and M-MΦ) via C5aR. (A-C) Human CD14+ monocytes (grey) and HMDMs (GM-MΦ, black; M-MΦ, white) were serum-deprived for at least 8 h before treatment with LPS (5 ng/mL) in the absence or presence of various concentration of C5a (0.03-300 nM) for 24 h. (D- F) For antagonism of C5aR, cells were pre-treated with 3D53 (1 µ , 30 min) prior to C5a and LPS stimulation. ELISA was used to detect levels of secreted (A, D) IL6, (B, E) TNF and (C, F) IL10 in culture media. Error bars are means ± SEM of three independent experiments (n=3). *p < 0.05, **p < 0.01 and ***p < 0.001 by student t- test.

5.5.2 Temporal study of C5aR/TLR4 crosstalk in HMDMs

To further characterise the phenotypic switch in C5a responsiveness during monocyte to macrophage differentiation, time course experiments were performed. As expected, mRNA levels of the macrophage marker CD68 were increased after 4 days of differentiation in the presence of either GM-CSF or M-CSF (Figure. 5.2A).

197

Figure 5.2. C5AR and TLR4 mRNA levels are increased during differentiation of CD14+ monocytes to macrophages with GM-CSF or M-CSF. (A) CD68, (B) C5AR and (C) TLR4 gene expression were monitored during differentiation of CD14+ monocytes, in the presence of GM-CSF (black) or in the presence of M-CSF (white) on Days 0, 2, 4 and 6. (D) C5AR and (E) TLR4 gene mRNA levels were also measured when treated with either C5a (30 nM) or LPS (5 ng/mL) overnight. Genes were detected by real-time PCR, normalised against 18S rRNA expression and converted to fold change relative to Day 0. Error bars are means ± SEM of three independent experiments (n=3) against control (Day 0). *p < 0.05, **p < 0.01 and ***p < 0.001 by student t-test.

The mRNA expression levels of C5AR and TLR4 are increased during monocyte to macrophage differentiation (Figure 5.2B, C). Levels of C5AR and TLR4 mRNA were not affected by C5a treatment, but LPS decreased C5AR expression by ~30% (Figure 5.2D, E), consistent with a previous study (31). The suppressive effect of C5a on LPS-induced cytokine production was apparent by day 4 for TNF, and by

198 day 6 for IL6 (Figure. 5.3B, C). Increased IL10 production was also observed by day 4 (Figure. 5.3D).

Figure 5.3. Temporal study of C5aR/TLR4 crosstalk in monocytes and macrophages. Human CD14+ monocytes were differentiated to HMDMs with either GM-CSF (10 ng/mL) (GM-MΦ, black) or M-CSF (10 ng/mL) (M-MΦ, white) in 10% FBS/IMDM over a 6 d time course. Macrophage marker These cells were also treated with LPS (5 ng/mL) ± C5a (30 nM) for 24 h in 10% FBS/IMDM on Day 0, 2, 4 and 6 to monitor suppression by C5a of LPS-induced IL6 and TNF in HMDMs. ELISA was used to detect (A) IL6, (B) TNF and (C) IL10 secreted in culture media. Error bars are means ± SEM of three independent experiments (n=3). *p < 0.05, **p < 0.01 and ***p < 0.001 by student t-test.

199 5.5.3 C5a suppresses LPS-induced IL6 and TNF via Gαi- and MEK- dependent signalling.

C5a-dependent C5aR signalling involves coupling to Gαi proteins (5, 32). PTX, which specifically interferes with Gαi protein signalling, was used to investigate whether the blockade of Gαi signalling could reverse inhibition of TLR4 responses by C5a. As expected, all modulatory effects of C5a on LPS-induced cytokine production from monocytes and macrophages were sensitive to PTX (Figure. 5.4A-C). These results confirmed involvement of C5aR- and not C5L2-signalling in this crosstalk, as C5aR but not C5L2 employs Gαi signalling (1, 3). We also pharmacologically targeted downstream signalling of C5aR signalling and found that the MEK inhibitor U0126, a selective inhibitor of the ERK pathway (33), prevented immunomodulation by C5a of LPS-induced IL6 and TNF production in macrophages, but not monocytes (Figure. 5.4A, B). Interestingly, although the literature suggested that ERK1/2 is involved in TLR-induced IL10 production from human macrophages (15, 34), it was found here that C5a enhanced LPS-mediated IL10 production in an ERK1/2- independent, but Gαi-dependent, manner (Figure. 5.4C). Similarly, LPS-induced IL- 10 production from monocytes, as well as GM-MΦ and M-MΦ, was ERK- independent (Figure. 5.4C). In monocytes, the enhancement of LPS-induced IL6 production by C5a was altered by the presence of MAPK p38- and PLCβ- inhibitors, SB203580 and U73122 respectively, whilst the enhancement of LPS-induced TNF production was affected by PKA inhibitor PKI-14-22 (Figure. 5.5A, D). This effect was not observed in macrophages (Figure. 5.5B-C, E-F). Collectively, these data suggest that C5a- dependent ERK1/2 phosphorylation selectively suppresses LPS-induced IL6 and TNF production in macrophages.

200

Figure 5.4. C5a suppression of LPS-induced IL6 and TNF in macrophages is sensitive to PTX and MEKi (U0126) treatment. To inhibit Gαi-dependent interactions, cells were pre-treated with PTX (200 ng/mL, overnight); or for MEK1/2 inhibition, U0126 (1 µM, 30 min), prior to LPS (5 ng/mL) ± C5a (30 nM) stimulation of monocytes (grey) and HMDMs (GM-MΦ, black; M- MΦ, white). ELISA was used to detect (A) IL6, (B) TNF and (C) IL10 secreted cytokine relative to the LPS response (100%) at 24 h post-stimulation. Error bars are means ± SEM of three independent experiments (n=3). *p < 0.05, **p < 0.01 and ***p < 0.001 by student t-test.

201

Figure 5.5. Signalling pathway analysis of C5aR/TLR4 crosstalk in monocytes and macrophages. Human CD14+ monocytes (grey) and HMDMs (GM-MC5aR/TLR4 crosstalk in monocytes and macrophages.nM) ± LPS (5 ng/mL) for 24 h in 10% FBS/IMDM. To inhibit (i) Gαi-dependent interactions, cells were pre-treated overnight with PTX (200 ng/mL); (ii) p38-dependent interactions, cells were pretreated for 30 min with SB203580 (600 nM), (iii) PLC-dependent interactions, cells were pretreated for 30 min with U73122 (1 µM) and (iv) PKA-dependent interactions, cells were pretreated for 30 min with PKI 14-22 (100 nM), prior to LPS (5 ng/mL) ± C5a (30 nM) stimulation of human monocytes (grey) and HMDMs (GM-MΦ, black; M-MΦ, white) for 24 h. ELISA was used to detect (A-C) IL6 and (D-F) TNF cytokine with LPS as 100% response. Error bars are means ± SEM of three independent experiments (n=3). **p < 0.01 and ***p < 0.001 by student t-test.

202 5.5.4 C5a and LPS synergize for Raf/MEK/ERK phosphorylation in macrophages

The data for PTX and U0126 inhibitors showed that C5a modulated suppression of LPS-mediated inflammatory responses was dependent on Gαi- coupling and MAPK ERK1/2 phosphorylation. Therefore additional checkpoints in the signalling pathway between Gαi signalling and upstream of ERK phosphorylation (Gαi/Raf/MEK/ERK) signalling cascade were investigated to further map the crosstalk signalling mechanisms between C5aR and TLR4. Interestingly, C5a phosphorylated c-Raf only in macrophages (both GM-MΦ and M-MΦ) but not in monocytes (Figure. 5.6A). Co-treatment of GM-MΦ or M-MΦ with LPS and C5a synergistically promoted c-Raf phosphorylation. This synergy was not observed in monocytes. LPS triggered MEK1 phosphorylation in macrophages (Figure. 5.6B). Although C5a and LPS each induced MEK1 phosphorylation in monocytes and macrophages, the synergistic effect of LPS and C5a co-treatment was only observed in macrophages. Consistent with this, co-treatment of GM-MΦ or M- MΦ with LPS and C5a synergistically promoted ERK1/2 phosphorylation, whereas this synergy was not observed for monocytes (Figure. 5.6C).

5.5.5 C5a and LPS synergise in phosphorylation of ERK1/2 but not p38 or JNK in macrophages spacing

Since C5a plus LPS potentiate ERK1/2 in macrophages, the temporal profile of ERK1/2 activation together with other MAPK (p38 and JNK) in response to C5a versus C5a plus LPS were investigated next. C5a induced a transient peak in ERK1/2 phosphorylation within 5 min in monocytes (Figure. 5.7A) and macrophages (Figure. 5.7B, C), but as previously observed (Figure. 5.6C), LPS amplified this response only in macrophages. LPS at 5 ng/mL concentration did not stimulate ERK1/2 phosphorylation in HMDM, however, at a higher concentration (1 µg/mL), ERK1/2 phosphorylation was detected after 20 min (Figure. 5.7D).

203

Figure 5.6. C5a induced Raf/MEK/ERK phosphorylation is amplified in macrophages but not in monocytes by LPS. To quantify c-Raf, MEK1 and ERK1/2 phosphorylation, monocytes (grey) GM-MΦ (black) and M-MΦ (white) were treated with LPS (5 ng/mL) ± C5a (30 nM) for 5 min, after which cells were lysed. (A) C-Raf phosphorylation was detected by western blot and probed with antibodies against phospho c-Raf and total c-Raf. Bands were visualized by an ECL method and quantified by densitometry measurements of phospho c-Raf normalized with total c-Raf against control. Levels of phosphorylated (B) MEK1 and (C) ERK1/2 were measured by AlphaScreen® SureFire® assays. Error bars are means ± SEM of three independent experiments (n=3). **p < 0.01 and ***p < 0.001 by student t-test.

204

Figure 5.7. C5a-induced ERK1/2 phosphorylation is amplified in macrophages but not monocytes by LPS. To quantify (A-C) ERK1/2, cells (monocytes, GM-Mon is amplified in macrophages but nong/mL) ± C5a (30 nM) over 10 min, (D) over 30 min, after which cells were lysed. Phosphorylated ERK1/2 was measured by AlphaScreen® SureFire® assays. Treatments by C5a, LPS or both were represented by broken lines, dotted lines and solid lines, respectively. Error bars are means ± SEM of three independent experiments (n=3). **p < 0.01 and ***p < 0.001 by student t-test.

C5a appeared to have little or no effect on p38 (Figure. 5.8A-C) and JNK phosphorylation (Figure. 5.8D-F), and did not amplify LPS-mediated activation of these signalling responses in monocytes and macrophages. Although LPS at 5 ng/mL concentration did not stimulate ERK1/2 phosphorylation in HMDM, LPS did induce p38 and JNK phosphorylation, at the same concentration. Although there was no synergistic effect for p38 phosphorylation in monocytes, the C5a-mediated amplification of LPS-induced IL6 production (Figure. 5.5A), but not TNF production (Figure. 5.5B), from these cells was affected by the p38 inhibitor SB203580. Furthermore, SB203580 did not block the suppression by C5a of LPS-induced IL6 and TNF production in macrophages (Figure. 5.5B-C, E-F).

205

Figure 5.8 LPS-induced p38 and JNK phosphorylation is not affected by C5a in macrophages and monocytes. To quantify (A-C) p38 and (D-F) JNK phosphorylation, cells (monocytes, GM-MΦ and M-MΦ) were treated with LPS (5 ng/mL) ± C5a (30 nM) over 10 min, after which cells were lysed. Phosphorylated p38 and JNK were measured by AlphaScreen® SureFire® assays. Treatments by C5a, LPS or both were represented by broken lines, dotted lines and solid lines, respectively. Error bars are means ± SEM of three independent experiments (n=3). **p < 0.01 and ***p < 0.001 by student t-test.

5.5.6 C5a-mediated suppression of LPS-induced IL6 and TNF is IL10- and PI3K-independent.

C5a has been previously reported to enhance LPS-induced IL10 cytokine production from macrophages (15), and has also been reported to suppress IL17A and IL23 by enhancing IL10 production (17). Consistent with such a mechanism, C5a upregulated LPS-induced IL10 production from macrophages, but did not alter this response in monocytes (Figure. 5.1C). However, the ability of the MEK inhibitor to block C5a-mediated suppression of IL6 and TNF production, without affecting enhancement of IL10 production in macrophages, implied that the inhibitory effects on IL6 and TNF were IL10-independent (Figure. 5.4C). Although, LPS-induced IL6 and TNF production was suppressed with increasing concentrations of IL10 (Figure. 5.9A, B) which is consistent with literature indicating that IL10 attenuates IL6 and TNF production from activated macrophages (34, 35), a neutralizing antibody against

206 human IL10 (10 µg/mL) completely blocked the effect of exogenously added IL10 (2 ng/mL) in macrophages (Figure. 5.9C, D) and its addition prior to LPS stimulation, resulted in increased TNF production in M-MΦ.

Figure 5.9. Anti-inflammatory IL10 cytokine suppressed LPS-induced IL6 and TNF production. (A,B) M-Mi-inflammatory IL10 cytokine supprng/mL) with various concentration of recombinant human IL10 (0.05-10 ng/mL) for 24 h. (C-D) To block IL10 activity, neutralizing anti-IL10 (10 n IL10 (0.05-10 F production.n in exogenng/mL) ± IL10 (2 ng/mL) 24 h exposure. (A,C) IL6 production and (B,D) TNF were detected by ELISA. Error bars are means ± SEM of three independent experiments (n=3). *p < 0.05 and **p < 0.01 by student t-test against positive control (LPS alone) or otherwise indicated.

However, C5a still suppressed IL6 and TNF production in the presence of the neutralizing antibody against human IL10 (Figure 5.10A, B), thus demonstrating that the immunomodulation by C5a was independent of IL10. Therefore, C5a-enhanced LPS-induced IL10 production does not account for the inhibition of inducible IL6 and TNF production by C5a.

207

Figure 5.10. C5a suppression of LPS-induced IL6 and TNF macrophages is IL10-independent. To determine dependency of IL10, GM-MΦ (black) and M-MΦ (white) were pretreated with neutralizing antibody against hIL10 (10 µg/mL, 30 min) prior to LPS (5 ng/mL) ± C5a (30 nM) stimulation, after which (A) IL6 and (B) TNF production were monitored at 24 h post-stimulation. Error bars are means ± SEM of three independent experiments (n=3). *p < 0.05, **p < 0.01 and ***p < 0.001 by student t- test.

Apart from IL10, the immunomodulation by C5a was also found to be independent of LPS-induced PI3K signalling in GM-MΦ (Figure. 5.11A, B) and in M-MΦ (Figure. 5.11C, D). The presence of the pan PI3K inhibitor wortmannin at the highest concentration (100 nM) did not abolish C5a-mediated suppression of LPS- inducible IL6 and TNF release, although wortmannin on its own appears to enhance LPS-induced IL6 and TNF production from macrophages at 100 nM concentration.

208

Figure 5.11. C5a suppression of LPS-induced IL6 and TNF macrophages is PI3K-independent. To determine dependency of PI3K signalling, wortmannin (1-100 nM, pre-treated for 1 h) prior to LPS (5 ng/mL) ± C5a (30 nM) stimulation, after which (A,C) IL6 and (B,D) TNF production were monitored at 24 h post-stimulation. Error bars are means ± SEM of three independent experiments (n=3). *p < 0.05, **p < 0.01 and ***p < 0.001 by student t-test.

5.5.7 C5a attenuation of LPS-induced IL6 is selectively TLR4-mediated in macrophages

TLR4 signalling involves two major pathways, the Mal/MyD88- and the TRAM/TRIF-dependent pathways (7). To determine whether C5a affected responses through multiple TLRs in primary human macrophages, poly(I:C) (TLR3 agonist signalling via TRIF) and IMQ (TLR7 agonist signalling via MyD88) were used. In contrast to TLR4 activation by LPS, C5a amplified poly(I:C)-induced (TRIF- dependent) IL6 production from GM-MΦ (Figure. 5.12B), whereas it did not induce detectable levels of this cytokine in monocytes and M-MΦ (Figure. 5.12A, C).

209

Figure 5.12. Suppression by C5a of inducible IL6 production from macrophages is selective for LPS responses. Human CD14+ monocytes (grey) and HMDMs (GM-MΦ, black; M-MΦ, white) were treated with (i) LPS (TLR4 agonist, 5 ng/mL) ± C5a (30 nM) stimulation, (ii) Poly(I:C) (TLR3 agonist, 10 µg/mL) ± C5a (30 nM) stimulation or (iii) Imiquimod (IMQ) (TLR7 agonist, 10 iquimodod3 agonisnM) stimulation for 24 h in 10% FBS/IMDM. ELISA was used to detect (A-C) IL6 and (D-F) IL10 secreted in culture media. Error bars are means ± SEM of three independent experiments (n=3). *p < 0.05, **p < 0.01 and ***p < 0.001 by student t-test.

In the case of IMQ-induced (MyD88-dependent) IL6 production, C5a enhanced this response in human monocytes, as well as in both human macrophage populations. C5a also increased poly(I:C)-induced and IMQ-induced IL10 production in monocytes and M-MΦ (Figure. 5.12D, F), but did not modulate this response in GM-MΦ (Figure. 5.12E). These results suggest that suppression of inducible IL6 production by C5a in GM-MΦ and M-MΦ was selective for TLR4 signalling.

210 5.5.8 C5a enhances clearance of Salmonella Typhimurium from HMDM

There is evidence in support of C5aR/TLR crosstalk being exploited by some pathogens for host subversion (12, 35). Therefore, the impact of C5a on the ability of HMDM to clear S. Typhimurium was examined in collaboration with a colleague, Ms Juliana Arrafin (IMB). As expected, C5a alone did not have any bactericidal effect on S. Typhimurium (Figure. 5.13A). Instead, C5a promoted the clearance of this bacterial pathogen from M-MΦ, an effect that was reversed by the C5aR antagonist 3D53 (Figure. 5.13B).

Figure 5.13. C5a decreases S. Typhimurium survival in M-MΦ. (A) Growth of S. Typhimurium was monitored ± C5a (100 nM) over 7 h. (B) M-Ma (100 nMC5a (100 nMred Typhimurium (MOI 100) ± C5a (30 nM). For antagonism of C5aR, M-MΦ were pre-treated with 3D53 (1 µM) for 30 min prior to C5a stimulation. Intra-macrophage survival was assayed at 2 h and 8 h post-infection. Graph shows a representative of three independent experiments. Error bars are means ± SEM of experimental triplicates. **p < 0.01 and ***p < 0.001 by student t-test. Courtesy of Juliana Arrafin (IMB).

In cytotoxicity assays monitoring LDH release, infection assay supernatants harvested 8h post-infection showed no differences in macrophage viability irrespective of treatment (DMSO, SLI344, hC5a, 3D53, or combinations), thereby indicating that reduced bacterial loads upon C5a treatment were not a result of increased macrophage death. Thus, C5a enhances human macrophage antimicrobial responses, while reducing production of key pro-inflammatory cytokines from these cells.

211

5.6 Discussion

Most pathogens trigger both complement and TLR activation, and crosstalk between these two systems could potentially impact on inflammatory and antimicrobial responses in vivo in infectious disease. However, studies on C5aR/TLR crosstalk have typically focused on monocytes or macrophages separately. No previously reported study has directly compared the impact of crosstalk on the functional responses of primary human monocytes versus donor-matched human macrophages. The present study has found that the myeloid differentiation state is an important variable in determining the impact of C5aR/TLR4 crosstalk on inflammatory responses of human monocytes and macrophages. New findings here suggest that C5a modulates TLR4 signalling in macrophages to favour pathogen clearance, while simultaneously limiting excessive inflammatory responses. This control may be important during localized infections, whereas C5a detection by monocytes likely acts as an alarm signal that triggers or amplifies inflammatory responses during systemic infections. The mechanisms of C5aR/TLR4 crosstalk are complex and evidently species- and cell type-dependent. In human monocytes, C5aR signalling has been reported to both downregulate (36) and upregulate (37) TLR-induced IL12 production. Although those studies were performed on similar IFNγ-primed human blood monocytes, the second signal used to stimulate the cells was different. C5a suppressed LPS-induced IL12 production, but upregulated S. aureus Cowan I-induced IL12 production. Although not addressed in either study, the variations reported are likely due to different TLRs being activated and differentially affected by C5aR signalling. The current study emphasizes that C5a suppression of IL6 production is specific to LPS/TLR4 signalling in human macrophages, since IL6 production is actually boosted by C5aR crosstalk with other TLRs, such as TLR3 and TLR7. Similarly, in mouse macrophages, C5a was reported to selectively crosstalk with TLR4, but not TLR3, in suppressing IL27 (p28) production (38). C5a is thus not only a potent chemotactic agent, attracting immune cells to site of infection or injury, but it also differentially modulates host immune responses, possibly tailoring them to match a particular type of infection.

212 The differentiating agents GM-CSF and M-CSF have differential effects on macrophage lineage populations, which contribute to their functional heterogeneity (39). GM-MΦ are often considered as ‘M1-like’ macrophages, while M-MΦ are classed as ‘M2-like’ macrophages, according to their cytokine and gene expression profiles (40, 41). Whilst LPS-induced inflammatory cytokine production was clearly enhanced in GM-MΦ compared to M-MΦ in the present study, C5a exerted a similar suppressive effect on inflammatory responses in both human macrophage populations. Thus, the phenotypic switch in C5a responsiveness during monocyte to macrophage differentiation occurred irrespective of the differentiating agent. In the future, it will be of interest to determine whether other macrophage polarizing factors (e.g. IL-4, IgG) influence the ability of C5a to suppress TLR4-inducible inflammatory mediator production. Although the basal levels of C5AR and TLR4 increased during the course of differentiating monocytes to macrophages, the differences in the expression levels are unlikely to be the cause for the phenotypic switch in the crosstalk. Firstly, C5a treatments did not alter C5AR expression levels. Secondly, C5a had opposing effects on inducible cytokine production when the macrophages were challenged with Poly(I:C) or Imiquimod, instead of LPS. The effects of C5a on TLR4 responses were PTX-sensitive in both monocytes and macrophages, suggesting a Gαi-sensitive mechanism. C3a can also bind to a PTX-sensitive receptor, C3aR that can recruit Goi-coupling when activated. Others have shown that C3a acting on IFNγ-primed macrophages from C5aR-/- mice (14), as well as other Gαi-specific ligands acting on IFNγ-primed human monocytes (42), can negatively regulate LPS-induced IL12 production. Also, PI3K signalling has been implicated in C5a regulation of LPS signalling. The pan PI3K inhibitor wortmannin was reported by la Sala et al to block the suppression by C5a of IL12p70 production (42), but this was not the case in the study performed by Hawlisch et al (14). In our study, wortmannin treatment had no impact on C5a-mediated suppression of LPS- induced IL6 and TNF production in human macrophages. This suggests that the crosstalk is not only tightly regulated at multiple checkpoints in the signalling cascade, but that it is also highly dependent on the activation status of the cells. As shown previously in human neutrophils (32,43), C5a can activate the Raf/MEK/ERK signalling cascade either directly via Gαi activation or indirectly via Gβϒ/PLCβ/PKC pathways. Here, it was found that C5a-mediated suppression of LPS-induced IL6 and TNF was independent of PLCβ-signalling (Supp. Fig. 2).

213 Although C5a rapidly phosphorylated ERK1/2 in both monocytes and macrophages (Fig X), consistent with the literature (14, 15), only in macrophages did C5a increase c-Raf phosphorylation in the presence of LPS, leading to further downstream increase in phosphorylation of MEK and ERK. Since, LPS and C5a did not synergize to alter other MAPKs (p38 and JNK) in monocytes and macrophages, the differences in monocytes and macrophages are attributed to the activation of Gαi/c-Raf/MEK/ERK pathway by C5a with LPS in macrophages versus monocytes. Low concentrations of LPS failed to stimulate ERK phosphorylation. However, only high concentrations of LPS triggered delayed ERK1/2 activation (20 min post- stimulation) by comparison with the C5a response, which occured within 5 min. The MAP3K tumour progression locus 2 (TPL2) is essential for activation of ERK1/2 by multiple TLRs, including TLR4, in macrophages (44). Consistent with this, it was observed that LPS triggered MEK1 phosphorylation in macrophages independently of activation of another MAP3K, c-Raf. ERK activation can produce different results, depending upon the stimulus and cell types, especially if comparing primary cells versus laboratory cultured cell lines (45). C5a enhanced ERK1/2 activation in rat neutrophils, and this was associated with an increase in IL6 production (46). This contrasts with our present findings for human macrophages. Complement-mediated ERK1/2 activation also inhibited IL12 production in human monocytes and mouse macrophages (14, 42), whereas the opposite was observed in mouse dendritic cells (DC) (47). Such reported variations in effects of C5a on LPS responses reinforce the view that observations of complement-mediated effects on TLR signalling are likely to not only be cell- and species-specific, but also to vary with the inflammatory readout being monitored.

214 5.7 Conclusions

In summary, this study has identified C5aR as a regulatory switch that modulates LPS/TLR4 signalling via the Gαi/c-Raf/MEK/ERK signalling axis in primary human macrophages, but not in monocytes. Extrapolating data from mouse to human, although convenient, is often not appropriate (48). Regulatory mechanisms involved in C5aR/TLR4 crosstalk that were previously reported in mouse studies may not be necessarily be conserved in humans. Differential effects of C5a on monocytes versus macrophages have been demonstrated here, and are consistent with the need to either amplify monocyte proinflammatory responses during systemic danger recognition but to attenuate macrophage proinflammatory cytokine responses (without compromising microbicidal activity) during localized infections. These differences between immune cell responses and control mechanisms might conceivably be targeted for therapeutic benefit. For example, many therapies, including chemotherapy, radiation, cancer vaccines and small-molecule drugs, already utilise immune cell-specific treatment strategies (49), but the rational basis for targeting specific cell types frequently lacks a level of understanding necessary to optimise such treatments. Studies such as this one have the potential to unlock secrets relating to interacting human immunoreceptors and control mechanisms that are not obvious in studies targeting one particular receptor on a specific human immune cell type, let alone on immune cells from other mammalian species.

215 5.8 References

1. Monk, P. N., A. M. Scola, P. Madala, and D. P. Fairlie. 2007. Function, structure and therapeutic potential of complement C5a receptors. Br. J. Pharmacol. 152: 429-448. 2. Gerard, N. P., and C. Gerard. 1991. The chemotactic receptor for human C5a anaphylatoxin. Nature 349: 614-617. 3. Okinaga, S., D. Slattery, A. Humbles, Z. Zsengeller, O. Morteau, M. B. Kinrade, R. M. Brodbeck, J. E. Krause, H. R. Choe, N. P. Gerard, and C. Gerard. 2003. C5L2, a nonsignaling C5A binding protein. Biochemistry (Mosc.) 42: 9406- 9415. 4. Chen, N. J., C. Mirtsos, D. Suh, Y. C. Lu, W. J. Lin, C. McKerlie, T. Lee, H. Baribault, H. Tian, and W. C. Yeh. 2007. C5L2 is critical for the biological activities of the anaphylatoxins C5a and C3a. Nature 446: 203-207. 5. Skokowa, J., S. R. Ali, O. Felda, V. Kumar, S. Konrad, N. Shushakova, R. E. Schmidt, R. P. Piekorz, B. Nurnberg, K. Spicher, L. Birnbaumer, J. Zwirner, J. W. Claassens, J. S. Verbeek, N. van Rooijen, J. Kohl, and J. E. Gessner. 2005. Macrophages induce the inflammatory response in the pulmonary Arthus reaction through G alpha i2 activation that controls C5aR and Fc receptor cooperation. J. Immunol. 174: 3041-3050. 6. Akira, S., and K. Takeda. 2004. Toll-like receptor signalling. Nat. Rev. Immunol. 4: 499-511. 7. Rossol, M., H. Heine, U. Meusch, D. Quandt, C. Klein, M. J. Sweet, and S. Hauschildt. 2011. LPS-induced cytokine production in human monocytes and macrophages. Crit. Rev. Immunol. 31: 379-446. 8. Chang, L., and M. Karin. 2001. Mammalian MAP kinase signalling cascades. Nature 410: 37-40. 9. Kawai, T., O. Adachi, T. Ogawa, K. Takeda, and S. Akira. 1999. Unresponsiveness of MyD88-deficient mice to endotoxin. Immunity 11: 115- 122. 10. Natarajan, M., K. M. Lin, R. C. Hsueh, P. C. Sternweis, and R. Ranganathan. 2006. A global analysis of cross-talk in a mammalian cellular signalling network. Nat. Cell Biol. 8: 571-580.

216 11. Postma, B., M. J. Poppelier, J. C. van Galen, E. R. Prossnitz, J. A. van Strijp, C. J. de Haas, and K. P. van Kessel. 2004. Chemotaxis inhibitory protein of Staphylococcus aureus binds specifically to the C5a and formylated peptide receptor. J. Immunol. 172: 6994-7001. 12. Wang, M., J. L. Krauss, H. Domon, K. B. Hosur, S. Liang, P. Magotti, M. Triantafilou, K. Triantafilou, J. D. Lambris, and G. Hajishengallis. 2010. Microbial hijacking of complement-toll-like receptor crosstalk. Sci. Signal. 3: ra11. 13. Hajishengallis, G., and J. D. Lambris. 2010. Crosstalk pathways between Toll- like receptors and the complement system. Trends Immunol. 31: 154-163. 14. Hawlisch, H., Y. Belkaid, R. Baelder, D. Hildeman, C. Gerard, and J. Kohl. 2005. C5a negatively regulates toll-like receptor 4-induced immune responses. Immunity 22: 415-426. 15. Okazaki, N., K. Hazeki, T. Izumi, K. Nigorikawa, and O. Hazeki. 2011. C5a controls TLR-induced IL-10 and IL-12 production independent of phosphoinositide 3-kinase. J Biochem 149: 265-274. 16. Bosmann, M., V. R. Patel, N. F. Russkamp, F. Pache, F. S. Zetoune, J. V. Sarma, and P. A. Ward. 2011. MyD88-dependent production of IL-17F is modulated by the anaphylatoxin C5a via the Akt signaling pathway. FASEB J. 25: 4222-4232. 17. Bosmann, M., J. V. Sarma, G. Atefi, F. S. Zetoune, and P. A. Ward. 2012. Evidence for anti-inflammatory effects of C5a on the innate IL-17A/IL-23 axis. FASEB J. 26: 1640-1651. 18. Cavaillon, J. M., C. Fitting, and N. Haeffner-Cavaillon. 1990. Recombinant C5a enhances interleukin 1 and tumor necrosis factor release by lipopolysaccharide- stimulated monocytes and macrophages. Eur. J. Immunol. 20: 253-257. 19. Zaal, A., S. N. Lissenberg-Thunnissen, G. van Schijndel, D. Wouters, S. M. van Ham, and A. ten Brinke. 2013. Crosstalk between Toll like receptors and C5a receptor in human monocyte derived DCs suppress inflammatory cytokine production. Immunobiology 218: 175-180. 20. Copeland, S., H. S. Warren, S. F. Lowry, S. E. Calvano, and D. Remick. 2005. Acute inflammatory response to endotoxin in mice and humans. Clin. Diagn. Lab. Immunol. 12: 60-67.

217 21. Schroder, K., K. M. Irvine, M. S. Taylor, N. J. Bokil, K. A. Le Cao, K. A. Masterman, L. I. Labzin, C. A. Semple, R. Kapetanovic, L. Fairbairn, A. Akalin, G. J. Faulkner, J. K. Baillie, M. Gongora, C. O. Daub, H. Kawaji, G. J. McLachlan, N. Goldman, S. M. Grimmond, P. Carninci, H. Suzuki, Y. Hayashizaki, B. Lenhard, D. A. Hume, and M. J. Sweet. 2012. Conservation and divergence in Toll-like receptor 4-regulated gene expression in primary human versus mouse macrophages. Proc. Natl. Acad. Sci. U. S. A. 109: E944- 953. 22. Ariffin, J. K., and M. J. Sweet. 2013. Differences in the repertoire, regulation and function of Toll-like Receptors and inflammasome-forming Nod-like Receptors between human and mouse. Curr. Opin. Microbiol. 23. Dinakar, C., A. Malur, B. Raychaudhuri, L. T. Buhrow, A. L. Melton, M. S. Kavuru, and M. J. Thomassen. 1999. Differential regulation of human blood monocyte and alveolar macrophage inflammatory cytokine production by nitric oxide. Ann. Allergy. Asthma. Immunol. 82: 217-222. 24. Netea, M. G., C. A. Nold-Petry, M. F. Nold, L. A. Joosten, B. Opitz, J. H. van der Meer, F. L. van de Veerdonk, G. Ferwerda, B. Heinhuis, I. Devesa, C. J. Funk, R. J. Mason, B. J. Kullberg, A. Rubartelli, J. W. van der Meer, and C. A. Dinarello. 2009. Differential requirement for the activation of the inflammasome for processing and release of IL-1beta in monocytes and macrophages. Blood 113: 2324-2335. 25. Finch, A. M., A. K. Wong, N. J. Paczkowski, S. K. Wadi, D. J. Craik, D. P. Fairlie, and S. M. Taylor. 1999. Low-molecular-weight peptidic and cyclic antagonists of the receptor for the complement factor C5a. J. Med. Chem. 42: 1965-1974. 26. Reid, R. C., G. Abbenante, S. M. Taylor, and D. P. Fairlie. 2003. A convergent solution-phase synthesis of the macrocycle Ac-Phe-[Orn-Pro-D-Cha-Trp-Arg], a potent new antiinflammatory drug. J. Org. Chem. 68: 4464-4471. 27. Trieu, A., N. Bokil, J. A. Dunn, T. L. Roberts, D. Xu, F. Y. Liew, D. A. Hume, K. J. Stacey, and M. J. Sweet. 2009. TLR9-independent effects of inhibitory oligonucleotides on macrophage responses to S. typhimurium. Immunol. Cell Biol. 87: 218-225. 28. Bamberg, C. E., C. R. Mackay, H. Lee, D. Zahra, J. Jackson, Y. S. Lim, P. L. Whitfeld, S. Craig, E. Corsini, B. Lu, C. Gerard, and N. P. Gerard. 2010. The

218 C5a receptor (C5aR) C5L2 is a modulator of C5aR-mediated signal transduction. J. Biol. Chem. 285: 7633-7644. 29. Scola, A. M., A. Higginbottom, L. J. Partridge, R. C. Reid, T. Woodruff, S. M. Taylor, D. P. Fairlie, and P. N. Monk. 2007. The role of the N-terminal domain of the complement fragment receptor C5L2 in ligand binding. J. Biol. Chem. 282: 3664-3671. 30. Lim, J., A. Iyer, J. Y. Suen, V. Seow, R. C. Reid, L. Brown, and D. P. Fairlie. 2013. C5aR and C3aR antagonists each inhibit diet-induced obesity, metabolic dysfunction, and adipocyte and macrophage signaling. FASEB J. 27: 822-831. 31. Raby, A. C., B. Holst, J. Davies, C. Colmont, Y. Laumonnier, B. Coles, S. Shah, J. Hall, N. Topley, J. Kohl, B. P. Morgan, and M. O. Labeta. 2011. TLR activation enhances C5a-induced pro-inflammatory responses by negatively modulating the second C5a receptor, C5L2. Eur. J. Immunol. 41: 2741-2752. 32. Buhl, A. M., N. Avdi, G. S. Worthen, and G. L. Johnson. 1994. Mapping of the C5a receptor signal transduction network in human neutrophils. Proc. Natl. Acad. Sci. U. S. A. 91: 9190-9194. 33. Favata, M. F., K. Y. Horiuchi, E. J. Manos, A. J. Daulerio, D. A. Stradley, W. S. Feeser, D. E. Van Dyk, W. J. Pitts, R. A. Earl, F. Hobbs, R. A. Copeland, R. L. Magolda, P. A. Scherle, and J. M. Trzaskos. 1998. Identification of a novel inhibitor of mitogen-activated protein kinase kinase. J. Biol. Chem. 273: 18623- 18632. 34. Chanteux, H., A. C. Guisset, C. Pilette, and Y. Sibille. 2007. LPS induces IL-10 production by human alveolar macrophages via MAPKinases- and Sp1- dependent mechanisms. Respir. Res. 8: 71. 35. Liang, S., J. L. Krauss, H. Domon, M. L. McIntosh, K. B. Hosur, H. Qu, F. Li, A. Tzekou, J. D. Lambris, and G. Hajishengallis. 2011. The C5a receptor impairs IL-12-dependent clearance of Porphyromonas gingivalis and is required for induction of periodontal bone loss. J. Immunol. 186: 869-877. 36. Wittmann, M., J. Zwirner, V. A. Larsson, K. Kirchhoff, G. Begemann, A. Kapp, O. Gotze, and T. Werfel. 1999. C5a suppresses the production of IL-12 by IFN- gamma-primed and lipopolysaccharide-challenged human monocytes. J. Immunol. 162: 6763-6769. 37. Karp, C. L., A. Grupe, E. Schadt, S. L. Ewart, M. Keane-Moore, P. J. Cuomo, J. Kohl, L. Wahl, D. Kuperman, S. Germer, D. Aud, G. Peltz, and M. Wills-Karp.

219 2000. Identification of complement factor 5 as a susceptibility locus for experimental allergic asthma. Nat. Immunol. 1: 221-226. 38. Bosmann, M., M. D. Haggadone, M. R. Hemmila, F. S. Zetoune, J. V. Sarma, and P. A. Ward. 2012. Complement activation product C5a is a selective suppressor of TLR4-induced, but not TLR3-induced, production of IL-27(p28) from macrophages. J. Immunol. 188: 5086-5093. 39. Bailey, M. J., D. C. Lacey, B. V. de Kok, P. D. Veith, E. C. Reynolds, and J. A. Hamilton. 2011. Extracellular proteomes of M-CSF (CSF-1) and GM-CSF- dependent macrophages. Immunol. Cell Biol. 89: 283-293. 40. Fleetwood, A. J., T. Lawrence, J. A. Hamilton, and A. D. Cook. 2007. Granulocyte-macrophage colony-stimulating factor (CSF) and macrophage CSF-dependent macrophage phenotypes display differences in cytokine profiles and transcription factor activities: implications for CSF blockade in inflammation. J. Immunol. 178: 5245-5252. 41. Lacey, D. C., A. Achuthan, A. J. Fleetwood, H. Dinh, J. Roiniotis, G. M. Scholz, M. W. Chang, S. K. Beckman, A. D. Cook, and J. A. Hamilton. 2012. Defining GM-CSF- and macrophage-CSF-dependent macrophage responses by in vitro models. J. Immunol. 188: 5752-5765. 42. la Sala, A., M. Gadina, and B. L. Kelsall. 2005. G(i)-protein-dependent inhibition of IL-12 production is mediated by activation of the phosphatidylinositol 3-kinase-protein 3 kinase B/Akt pathway and JNK. J. Immunol. 175: 2994-2999. 43. Jiang, H., Y. Kuang, Y. Wu, A. Smrcka, M. I. Simon, and D. Wu. 1996. Pertussis toxin-sensitive activation of phospholipase C by the C5a and fMet- Leu-Phe receptors. J. Biol. Chem. 271: 13430-13434. 44. Gantke, T., S. Sriskantharajah, M. Sadowski, and S. C. Ley. 2012. IkappaB kinase regulation of the TPL-2/ERK MAPK pathway. Immunol. Rev. 246: 168- 182. 45. Rao, K. M. 2001. MAP kinase activation in macrophages. J. Leukoc. Biol. 69: 3-10. 46. Riedemann, N. C., R. F. Guo, T. J. Hollmann, H. Gao, T. A. Neff, J. S. Reuben, C. L. Speyer, J. V. Sarma, R. A. Wetsel, F. S. Zetoune, and P. A. Ward. 2004. Regulatory role of C5a in LPS-induced IL-6 production by neutrophils during sepsis. FASEB J. 18: 370-372.

220 47. Baruah, P., I. E. Dumitriu, T. H. Malik, H. T. Cook, J. Dyson, D. Scott, E. Simpson, and M. Botto. 2009. C1q enhances IFN-gamma production by antigen-specific T cells via the CD40 costimulatory pathway on dendritic cells. Blood 113: 3485-3493. 48. Seok, J., H. S. Warren, A. G. Cuenca, M. N. Mindrinos, H. V. Baker, W. Xu, D. R. Richards, G. P. McDonald-Smith, H. Gao, L. Hennessy, C. C. Finnerty, C. M. Lopez, S. Honari, E. E. Moore, J. P. Minei, J. Cuschieri, P. E. Bankey, J. L. Johnson, J. Sperry, A. B. Nathens, T. R. Billiar, M. A. West, M. G. Jeschke, M. B. Klein, R. L. Gamelli, N. S. Gibran, B. H. Brownstein, C. Miller-Graziano, S. E. Calvano, P. H. Mason, J. P. Cobb, L. G. Rahme, S. F. Lowry, R. V. Maier, L. L. Moldawer, D. N. Herndon, R. W. Davis, W. Xiao, and R. G. Tompkins. 2013. Genomic responses in mouse models poorly mimic human inflammatory diseases. Proc. Natl. Acad. Sci. U. S. A. 110: 3507-3512. 49. Sheridan, C. 2013. Industry pursues co-stimulatory receptor immunomodulators to treat cancer. Nat. Biotechnol. 31: 181-183.

221 Chapter

6 Summary of thesis

6.1 Thesis summary ...... 223 6.2 References ...... 232

222 6.1 Thesis summary

Macrophages are immune cells produced by the differentiation of monocytes in tissues. Macrophages were first discovered in the early 1900s and were thought to assume a defensive role in innate immunity by their remarkable phagocytic capacities to engulf and destroy invading microbes. However, in the 1970s macrophages were also found to regulate adaptive immunity via the activation and proliferation of lymphocytes (1). Since then, the roles of macrophages have extended beyond being just primarily as phagocytes (2). The roles of macrophages are now known to be important in many aspects of physiology, not just immunity. Therefore, it is not surprising that macrophages and complement can also cooperate closely in mounting immune responses. The anaphylatoxin C5a is an extremely potent proinflammatory peptide produced during activation of the complement protein network by infection, injury or metabolic disturbance. The principal receptor for C5a (C5aR) has been strongly implicated in many inflammatory diseases and C5aR antagonists have shown promise as therapeutic agents for modulating C5a-associated inflammatory conditions, at least in animal models. No C5aR antagonist has yet progressed in human trials to a marketable drug (3). Significant challenges need to be overcome in order to bring a drug to market, not the least of which is obtaining a better understanding of how C5a is involved in normal and aberrant physiology. Therefore, in part this research program sought to investigate properties of human C5a on human monocytes and their differentiated forms, macrophages, to further profile the role of C5a and its receptor C5aR in immunity. Chapter One summarizes the current status of complement C5a and C5aR research in relation to inflammation and macrophages. This chapter not only serves as a general overview on the current status of the field, it also identifies important information gaps and obstacles that are currently faced by researchers in the field. For example, one of the biggest assumptions made by many researchers has been that mechanisms underlying C5aR signalling in murine cells can be extrapolated to human diseases, even though there are crucial differences in signalling mechanisms between human and murine cells (4). Unfortunately, C5aR activation has often been studied using murine rather than human cells. For example, the only cDNA microarray analysis of gene expression induced by C5a-treated human cells was performed on

223 non-immune cells (5), even though the majority of effects mediated by C5a have to date been associated with immune cells, macrophages in particular (6). Another problem is that human macrophages have remarkable plasticity in functional and phenotypic responses upon activation, which makes the analysis of human macrophages responses to C5a quite complex. Depending on the activation stage of the macrophages, the same stimulant can trigger different outcomes (7). Moreover, MAPK signalling by primary and cultured cells are different (8). The combination of these kinds of differences has made understanding human immunology and inflammatory diseases mediated by macrophages extraordinarily complex. Although the primary objective of this thesis was not focused on the complexity of human macrophages, some assessment of the effect of the monocyte-macrophages lineage was made in Chapter Two, where different populations of macrophages were studied following differentiation of primary human monocytes with GM-CSF or M- CSF to macrophages (MΦ). GM-CSF and M-CSF have been used before as differentiating agents for murine bone marrow cells to generate different populations of murine macrophages, termed M1 and M2 respectively according to genomic analysis (9) and extracellular secretory proteomics (10). In this thesis chapter, the focus was on gene expression profiling human macrophages (GM-MΦ versus M-MΦ) in response to LPS, the major component of the outer membrane of Gram-negative bacteria. We were interested in comparing effects of stimulating gene expression in these different cell populations to better understand macrophage plasticity in immunity. Rather than exhibiting distinct gene expression profiles, the different human macrophage populations represented more of a continuum model rather than distinctive genotypes/phenotypes (M1 or M2). These observations challenge current dogma that human macrophages exhibit either distinctly pro- or anti-inflammatory phenotypes depending on the differentiation state, instead suggesting much less functional diversity than previously thought. In collaboration with Mr. Daniel Hohenhaus and Associate Professor Matthew Sweet the GPCR repertoire of primary human macrophages (GM-MΦ versus M-MΦ) was profiled using TaqMan Low Density Array (TLDA), with the finding that only 15 of 380 GPCRs were differentially expressed between these cell populations in the basal state (11). This small difference was similarly identified in the cDNA microarray profiling reported in Chapter 2, implying that even though the primary cells have huge genetic variations from different batches of human macrophages,

224 these variations did not compromise the analysis of GM-MΦ versus M-MΦ, as demonstrated on two different platforms (cDNA microarray and TLDA). Importantly, these studies also revealed that LPS signalling could influence GPCR expression in human macrophages, suggesting that novel mechanisms involved in Toll-Like receptors (TLRs) and GPCRs signalling exist that are important for human macrophages response. As mentioned earlier, the only previously reported cDNA microarray analysis of gene expression induced by C5a-treated human cells was performed on non- immune cells (5). Therefore, Chapter Three assessed the responses of human macrophages (GM-MΦ versus M-MΦ) to human C5a. C5a had similar potency on both populations of macrophages (GM-MΦ and M-MΦ) in three different functional assays (calcium release, ERK phosphorylation and macrophage migration), despite the large genomic differences found between GM-MΦ and M-MΦ in their basal state in Chapter 2. C5a also induced similar genomic changes on GM-MΦ and M-MΦ as analysed globally by microarray. The majority of genes induced by human C5a belongs to proinflammatory cytokines and chemokines (CCL3, CCL3L1, CXCL2, IL8, PTGS2 and TNF), and a number of transcription factors (ATF3, EGR1, EGR2, EGR3, FOS, FOSB and JUNB). A total of 33 genes commonly induced by human C5a after 30 min on GM-MΦ and M-MΦ were analysed by Ingenuity Pathway Analysis and the majority of the associated overlapping signalling pathways were involved in neurological disease, cancer, cell death and survival. Since all of the C5a-induced genes on human macrophages were upregulated, it was important to understand what C5a also did to human macrophages after 30 min. A separate microarray experiment was conducted to overview the temporal gene expression profile of the effects of C5a on human macrophages. Interestingly, C5a induced a set of genes related to metabolism and biosynthesis after 6 h, suggesting that C5a has important roles not only in regulating immunity and inflammation, but also in cellular metabolism as well. Chapter Four reports a comparative study of the effects of three potent C5aR antagonists described in the literature (3D53, W54011, JJ47), and re-evaluated their relative antagonist potential on human C5aR on human macrophages and in a simple rat model of inflammation (paw oedema induced by C5aR agonist). Traditionally, drug discovery focuses on optimizing ligand affinity for a specific receptor, enhancing functional potency and selectivity, and improving drug-like properties for optimal pharmacokinetics and oral bioavailability. These optimization processes have

225 not always led to optimal in vivo efficacy. The reported oral bioavailability for 3D53 is much lower than for W54011 and JJ47. However, a single oral dose of 3D53 inhibited C5aR-mediated paw oedema in male Wistar rats for at least 16 hours, whereas this effect lasted less than 2 hours for orally administered W54011 or JJ47. This difference was traced here to a much longer residence time (very slow off rate) on human C5aR for 3D53 than for W54011 and JJ47. This property was also reflected in vitro in that 3D53 is a non-competitive, insurmountable, antagonist of human C5aR, and retains its high potency (IC50 ~20 nM), independent of the C5a concentration (0.1 nM to 300 nM), whereas W54011 and JJ47 are both competitive antagonists that are only potent (IC50 < 20 nM) antagonists against low C5a concentrations (0.1 nM) but become ineffective against C5a concentrations above 30 nM. Chapter 4 highlighted the fact that receptor affinity, antagonist potency, drug-likeness and oral bioavailability of a ligand for its receptor are sometimes insufficient for defining the effectiveness and duration of biological action of a drug candidate. Rather, the residence time of ligand on the receptor is much more important in dictating the drug efficacy in both cellular and organismal contexts. Too often, researchers seek IC50 values for putative antagonists without considering the mechanism of antagonism, thus IC50 values which are dependent on concentration of the competing agonist may be meaningless in vivo where physiological conditions are constantly changing depending on the severity of disease, unlike the controlled conditions of in vitro assays. A further demonstration of the value of long receptor residence time was highlighted by an in vivo study in Chapter 4. Abbott Laboratories developed a relatively potent C5aR agonist in the 1980s (12), namely Phe-Lys-Pro-DCha-Cha-DArg-OH, which was reported to be selective for C5aR over C5L2 (13). This peptide was used in this chapter to induce a C5aR- mediated paw oedema in male Wistar rats, which was challenged by each of the three antagonists compared above. Mirroring the effects on macrophages in vitro, 3D53 was able to inhibit induction of this paw oedma for at least 16 h after a single oral dose (5 mg/kg), whereas orally delivered W54011 (30 mg/kg) and JJ47 (30 mg/kg) were effective for less than 2 h in suppressing the onset of the rat paw oedema. Interestingly, 3D53 at 5 mg/kg p.o. was unable to fully block (~ 40%) the paw oedema induced by the C5aR peptidic agonist and a higher dose (10 mg/kg, p.o.) did not increase inhibition. This suggested that the inflammatory response induced by the peptidic agonist might not only be mediated through C5aR activation. Evidence of its

226 non-selectivity was further provided via a calcium de-sensitization assay, in which macrophages primed by C5a to desensitise C5aR still responded after 5 minutes to other non-C5a agonists. The C5aR-mediated in vivo model is only as good as the agonist triggering the C5aR response. Clearly the Abbott agonist is not as receptor selective as previously thought and, since C5a itself is too expensive and unstable in vivo to use, there is scope for developing a more selective small molecule agonist for C5aR. When a more selective C5aR agonist is available, this paw oedema experiment should be repeated to further prove that 3D53 is effective in specifically abolishing C5aR-mediated inflammation in vivo. Chapter Five reports the novel observation that human C5a, acting via C5aR, modulates responses of human macrophages but not human monocytes to LPS (14). C5a suppressed the LPS-induced production of proinflammatory cytokines IL6 and TNF in macrophages, while promoting more efficient killing of Salmonella Typhimurium. The suppressive effect of human C5a of LPS-induced IL6 and TNF production on human macrophages acted independent of the differentiating agents,

GM-CSF and M-CSF. Rather, C5a induced Gαi/c-Raf/MEK/ERK signalling was amplified in human macrophages but not in human monocytes by LPS. However, there were marked differences between the signalling mechanisms activated by C5a and LPS, via receptor C5aR/TLR4 crosstalk on human macrophages, compared to a previous report for murine macrophages (15). This crosstalk to suppress proinflammatory cytokines was not dependent on IL10 or PI3K for human macrophages, contrary to previously reported findings for mouse macrophages. This emphasises the danger in extrapolating signalling mechanisms in murine macrophages to human macrophages. C5a can also bind to a second receptor, C5L2 (16). The involvement of C5L2 in the C5aR/TLR4 crosstalk was ruled out here through the use of a C5aR-selective antagonist that does not bind to C5L2 (17), the blockade of C5aR signalling abolishing C5a-mediated effects on LPS signalling. However, it was recently reported that activation of C5aR in human macrophages led to heteromer formation with C5L2 to regulate C5aR signalling, a process that was enhanced by C5a, but not by C5adesArg (18). Besides C5a, C5adesArg was also shown to suppress LPS-induced IL6 and TNF in a dose-dependent manner on human macrophages (18). This finding is inconsistent with our preliminary results, as C5adesArg did not have an effect on LPS-induced IL6 and TNF production on human macrophages (Figure 6.1). Nevertheless, given the differential effect of C5a on LPS signalling between

227 human monocytes and human macrophages, it would be useful to determine whether this C5aR/C5L2 heteromer formation occurs in human monocytes.

Figure 6.1. C5adesArg did not suppress LPS-induced IL6 and TNF production. M-MΦ were serum deprived for at least 8 h before treatment with LPS (5 ng/mL) in the absence or presence of C5adesArg (300 nM) for 24 h. ELISA was used to detect levels of secreted (A) IL6 and (B) TNF in culture media. Error bars are means ± SEM of two independent experiments (n=2). n.s (not significant) by student t-test.

C5a and macrophages are also involved in other important physiological processes (19). Although not a major part of this thesis, I did conduct additional experiments briefly summarized here on the role of C5a and macrophages in metabolism and obesity. Activating complement proteins in blood triggers immune responses and requires energy to fight infection. Further experiments were performed in collaboration with Drs. Junxian Lim and Abishek Iyer, to identify new energy conserving roles for complement protein C5a and C5aR on diet-induced obesity on male Wistar rats (Appendix V). These roles challenged the field of complement biology and suggest that complement signalling through these proteins and receptors plays previously unknown roles in regulating metabolism and energy homeostasis. Rats fed on high-energy diet for 16 weeks had higher complement activation in blood serum. Many of the metabolic disorders associated with the high-energy diet on rats were attenuated by daily oral administration of selective small molecular antagonists of C5aR (3D53). Interestingly, C5a enhanced PGE2 production in murine macrophages RAW264.7, and also PGE2, on the other hand, increased both gene and protein expressions of C5aR on murine mature adipocytes 3T3-L1 (Figure 6.2). These results suggested that C5a acted directly on immune cells, which in turn exerted paracrine control over adipocytes via macrophages in the adipose tissues.

228

Figure 6.2. C5aR antagonist modulates inflammatory responses in RAW264.7 macrophages. (A) Enhancement in PGE2 secretion on macrophages RAW264.7 by human C5a (100 nM) can be inhibited by pre-treatment with C5aR antagonist (10 µM 3D53). (B) PGE2 (2 µg/mL) upregulates protein and mRNA levels of C5aR in 3T3-L1 adipocytes. Expression of mRNA was normalised against 18S rRNA and fold change was calculated relative to untreated control. Error bars are mean ± SEM of three independent experiments (n=3). **p < 0.01 by student t-test.

Although C3adesArg (also known as acylation stimulating protein, ASP) was previously reported to activate adipocytes via C5L2, for triglyceride synthesis and fatty acid uptake in 3T3-L1 adipocytes (20), the same role for C5a activating adipocytes has not previously been reported. The development of adipogenesis requires uptake of excess nutrients, such as glucose and fatty acids, and indeed C5a was found to be able to stimulate uptake of BODIPY-fatty acid (FA) and 2- deoxyglucose (2-DG) (Figure 6.3) by 3T3-L1 adipocytes. Pre-treatment with the C5aR specific antagonist 3D53 inhibited C5a-induced uptake of FA and 2-DG suggesting that such responses were specific to C5aR and not via C5L2 signalling (21). Consistent with this finding, another group of researchers reported the same observation, supporting the role of C5aR in obesity and adipose tissue inflammation although their diet-induced obesity was studied in mice instead of rats (22).

229

Figure 6.3. Antagonists of C5aR and C3aR can modulate lipid homeostasis in 3T3-L1 adipocytes. Like insulin (50 nM), C5aR antagonist (10 µM 3D53) can inhibit C5a-induced (100 nM) uptake of (A) BODIPY-fatty acid (FA) and (B) 2-deoxyglucose (2-DG) in 3T3- L1 adipocytes. Error bars are mean ± SEM of three independent experiments (n=3). *p < 0.05 and **p < 0.01 by student t-test.

In conclusion, the thesis has highlighted some important considerations when investigating human macrophages and monocytes for immunoregulatory effects. It is clear that the differentiation status of such immune cells in vitro can give misleading information in regards to genotype and functional phenotype, when compared either (a) in vitro between human cell populations, (b) in vitro with murine cell populations, or (c) in vivo with responses in animal (and presumably human) models of healthy versus diseased physiology. Moreover, the effects of C5a mediated responses can also be interpreted differently between macrophages of different species, of different differentiation status, and in response to different stimuli. The development of more effective C5aR antagonists has been handicapped by some of these differences and may be overcome in the future by giving more consideration to (a) murine versus

230 human species differences, (b) differences in responses in primary cells compared with cultured cell lines, (c) cell types relevant to a specific diseased state, (d) molecular mechanisms of antagonism that can affect duration and efficacy of treatment, and (e) the nature of the cell type, degree of receptor expression, stimulatory effects on receptor expression and crosstalk with other receptors. The characterisation of two populations of human macrophages GM-MΦ and M-MΦ in this thesis and responses mediated by human C5a have so far resulted in publications in (i) the Journal of Immunology (Appendix IV) and (ii) Immunobiology (Appendix V) (11, 14), while the work on metabolism and C5a has been published in FASEB J (Appendix VI) (21).

231 6.2 References

1. Unanue, E. R. 1978. The regulation of lymphocyte functions by the macrophage. Immunol. Rev. 40: 227-255. 2. Gordon, S. 2007. The macrophage: past, present and future. Eur. J. Immunol. 37 Suppl 1: S9-17. 3. Monk, P. N., A. M. Scola, P. Madala, and D. P. Fairlie. 2007. Function, structure and therapeutic potential of complement C5a receptors. Br. J. Pharmacol. 152: 429-448. 4. Copeland, S., H. S. Warren, S. F. Lowry, S. E. Calvano, and D. Remick. 2005. Acute inflammatory response to endotoxin in mice and humans. Clin. Diagn. Lab. Immunol. 12: 60-67. 5. Albrecht, E. A., A. M. Chinnaiyan, S. Varambally, C. Kumar-Sinha, T. R. Barrette, J. V. Sarma, and P. A. Ward. 2004. C5a-induced gene expression in human umbilical vein endothelial cells. Am. J. Pathol. 164: 849-859. 6. Markiewski, M. M., and J. D. Lambris. 2007. The role of complement in inflammatory diseases from behind the scenes into the spotlight. Am. J. Pathol. 171: 715-727. 7. Porcheray, F., S. Viaud, A. C. Rimaniol, C. Leone, B. Samah, N. Dereuddre- Bosquet, D. Dormont, and G. Gras. 2005. Macrophage activation switching: an asset for the resolution of inflammation. Clin. Exp. Immunol. 142: 481-489. 8. Rao, K. M. 2001. MAP kinase activation in macrophages. J. Leukoc. Biol. 69: 3-10. 9. Fleetwood, A. J., T. Lawrence, J. A. Hamilton, and A. D. Cook. 2007. Granulocyte-macrophage colony-stimulating factor (CSF) and macrophage CSF-dependent macrophage phenotypes display differences in cytokine profiles and transcription factor activities: implications for CSF blockade in inflammation. J. Immunol. 178: 5245-5252. 10. Bailey, M. J., D. C. Lacey, B. V. de Kok, P. D. Veith, E. C. Reynolds, and J. A. Hamilton. 2011. Extracellular proteomes of M-CSF (CSF-1) and GM-CSF- dependent macrophages. Immunol. Cell Biol. 89: 283-293. 11. Hohenhaus, D. M., K. Schaale, K. A. Le Cao, V. Seow, A. Iyer, D. P. Fairlie, and M. J. Sweet. 2013. An mRNA atlas of G protein-coupled receptor

232 expression during primary human monocyte/macrophage differentiation and lipopolysaccharide-mediated activation identifies targetable candidate regulators of inflammation. Immunobiology 218: 1345-1353. 12. Kawai, M. W., PE.; Luly, JR.; Or, YS. . 1992. New hexa - and heptapeptides are anaphylatoxin antagonists and agonists - for treating inflammatory and immunodeficiency diseases, cancers and severe infections. W. I. P. Organisation, ed. Abbott Laboratories, USA. 13. Otto, M., H. Hawlisch, P. N. Monk, M. Muller, A. Klos, C. L. Karp, and J. Kohl. 2004. C5a mutants are potent antagonists of the C5a receptor (CD88) and of C5L2: position 69 is the locus that determines agonism or antagonism. J. Biol. Chem. 279: 142-151. 14. Seow, V., J. Lim, A. Iyer, J. Y. Suen, J. K. Ariffin, D. M. Hohenhaus, M. J. Sweet, and D. P. Fairlie. 2013. Inflammatory responses induced by lipopolysaccharide are amplified in primary human monocytes but suppressed in macrophages by complement protein c5a. J. Immunol. 191: 4308-4316. 15. Bosmann, M., J. V. Sarma, G. Atefi, F. S. Zetoune, and P. A. Ward. 2012. Evidence for anti-inflammatory effects of C5a on the innate IL-17A/IL-23 axis. FASEB J. 26: 1640-1651. 16. Cain, S. A., and P. N. Monk. 2002. The orphan receptor C5L2 has high affinity binding sites for complement fragments C5a and C5a des-Arg(74). J. Biol. Chem. 277: 7165-7169. 17. Scola, A. M., A. Higginbottom, L. J. Partridge, R. C. Reid, T. Woodruff, S. M. Taylor, D. P. Fairlie, and P. N. Monk. 2007. The role of the N-terminal domain of the complement fragment receptor C5L2 in ligand binding. J. Biol. Chem. 282: 3664-3671. 18. Croker, D. E., R. Halai, D. P. Fairlie, and M. A. Cooper. 2013. C5a, but not C5a-des Arg, induces upregulation of heteromer formation between complement C5a receptors C5aR and C5L2. Immunol. Cell Biol. 91: 625-633. 19. Mastellos, D., and J. D. Lambris. 2006. Cross-disciplinary research stirs new challenges into the study of the structure, function and systems biology of complement. Adv. Exp. Med. Biol. 586: 1-16. 20. Cui, W., M. Lapointe, D. Gauvreau, D. Kalant, and K. Cianflone. 2009. Recombinant C3adesArg/acylation stimulating protein (ASP) is highly

233 bioactive: a critical evaluation of C5L2 binding and 3T3-L1 adipocyte activation. Mol. Immunol. 46: 3207-3217. 21. Lim, J., A. Iyer, J. Y. Suen, V. Seow, R. C. Reid, L. Brown, and D. P. Fairlie. 2013. C5aR and C3aR antagonists each inhibit diet-induced obesity, metabolic dysfunction, and adipocyte and macrophage signaling. FASEB J. 27: 822-831. 22. Phieler, J., K. J. Chung, A. Chatzigeorgiou, A. Klotzsche-von Ameln, R. Garcia-Martin, D. Sprott, M. Moisidou, T. Tzanavari, B. Ludwig, E. Baraban, M. Ehrhart-Bornstein, S. R. Bornstein, H. Mziaut, M. Solimena, K. P. Karalis, M. Economopoulou, J. D. Lambris, and T. Chavakis. 2013. The complement anaphylatoxin c5a receptor contributes to obese adipose tissue inflammation and insulin resistance. J. Immunol. 191: 4367-4374.

234 Appendices

Appendix I

All genes up-regulated ≥ 2 fold change in GM-MΦ versus M-MΦ

Appendix I

Gene FC p (Corr) Accession Gene FC p (Corr) Accession TACSTD2 98.2 7.8E-05 NM_002353.1 UBE2C 8.8 1.3E-03 NM_181803.1 CEACAM8 90.1 6.1E-06 NM_001816.2 HRASLS3 8.5 2.6E-02 NM_007069.2 CXCL5 87.3 1.5E-04 NM_002994.3 IGFBP2 8.4 5.6E-03 NM_000597.2 CXCL5 69.1 7.8E-05 NM_002994.3 MUCL1 8.3 3.6E-03 NM_058173.2 PPBP 60.2 1.7E-05 NM_002704.2 C1RL 8.3 1.2E-04 NM_016546.1 S100P 57.8 4.0E-05 NM_005980.2 GPR120 8.3 1.6E-03 NM_181745.2 CCL24 55.2 9.9E-04 NM_002991.2 NR3C1 8.2 6.0E-04 NM_001018077.1 MMP10 49.3 9.8E-05 NM_002425.1 NCAPG 8.2 6.2E-03 NM_022346.3 MARCO 38.6 3.8E-06 NM_006770.3 DLGAP5 8.1 3.3E-04 NM_014750.3 IL1F8 33.1 5.1E-05 NM_014438.3 ADAMTSL4 8.1 6.1E-06 NM_025008.3 FCER2 31.5 3.8E-06 NM_002002.3 ECSCR 8.1 1.0E-05 NM_001077693.2 RETN 30.2 5.6E-06 NM_020415.2 SYNC1 7.9 4.0E-04 NM_030786.1 IL17RB 29.2 1.5E-03 NM_018725.3 DLGAP5 7.9 9.5E-04 NM_014750.3 KIFC3 27.6 2.5E-04 NM_005550.2 VCAN 7.7 2.8E-02 NM_004385.2 CA12 27.4 2.4E-05 NM_001218.3 C11ORF82 7.7 7.8E-05 NM_145018.2 CCL23 25.6 3.2E-05 NM_005064.3 IL1RN 7.7 5.6E-03 NM_173843.1 FGD5 23.4 7.6E-06 NM_152536.2 BIRC5 7.7 1.2E-03 NM_001168.2 ALDH1A2 23.3 6.0E-05 NM_170697.1 MTE 7.6 1.5E-04 NM_175621.2 PPARG 22.9 8.9E-04 NM_015869.4 PKIA 7.5 2.2E-04 NM_181839.1 CES1 21.4 2.5E-04 NM_001266.4 NDP 7.4 1.8E-02 NM_000266.1 CES4 20.8 2.4E-04 NR_003276.1 MRC1L1 7.3 7.2E-05 NM_001009567.1 STAC 20.7 3.8E-06 NM_003149.1 KRT79 7.3 2.5E-04 NM_175834.2 C19ORF59 20.1 2.2E-05 NM_174918.2 C20ORF123 7.1 1.9E-02 XM_938433.2 CHDH 19.7 2.4E-03 NM_018397.3 HMMR 7.1 1.5E-03 NM_012484.1 ALDH1A2 19.3 1.3E-04 NM_170696.1 SERPINA1 7.0 7.2E-05 NM_001002236.1 PCOLCE2 19.0 7.6E-04 NM_013363.2 ALOX5 7.0 5.4E-04 XM_001127464.1 MOBKL2B 18.4 3.8E-06 NM_024761.3 LOC731954 6.8 3.8E-04 XR_015662.2 FAM23B 18.4 7.6E-06 NM_001013629.1 CSF1 6.7 7.2E-06 NM_172212.1 COL22A1 17.5 1.1E-02 NM_152888.1 RNASE2 6.7 1.6E-04 NM_002934.2 GLDN 17.2 6.7E-06 NM_181789.2 HBEGF 6.6 5.0E-04 NM_001945.1 MT1H 15.2 7.5E-05 NM_005951.2 CENPM 6.6 4.6E-03 NM_001002876.1 CCL23 14.9 5.1E-05 NM_145898.1 OLR1 6.5 4.8E-03 NM_002543.3 ADAMTSL4 13.4 5.2E-05 NM_025008.3 RGNEF 6.4 7.6E-04 NM_001080479.1 TREM1 12.7 6.1E-06 NM_018643.2 C1QA 6.4 2.3E-04 NM_015991.1 LOC645638 12.7 4.0E-05 XR_040455.1 CDC45L 6.3 1.7E-02 NM_003504.3 IL1F8 12.7 1.5E-04 NM_173178.1 MS4A14 6.3 5.5E-04 NM_001079692.1 OR8G5 12.7 1.6E-03 NM_001005198.1 ARAP3 6.2 2.7E-04 NM_022481.5 SPINK1 11.7 4.8E-03 NM_003122.2 NEDD9 6.1 2.1E-04 NM_006403.2 MT1G 11.6 4.0E-05 NM_005950.1 ASPM 6.1 1.3E-03 NM_018136.3 AGRP 11.3 2.7E-04 NM_007316.1 CA12 6.0 2.5E-03 NM_001218.3 HSD11B1 11.1 2.5E-05 NM_181755.1 TK1 6.0 6.5E-03 NM_003258.2 HSD11B1 10.7 2.4E-05 NM_005525.2 RAP1GAP 5.9 4.2E-03 NM_002885.1 UBE2C 10.5 1.7E-03 NM_181800.1 CTSG 5.9 1.1E-03 NM_001911.2 CPE 10.5 1.4E-02 NM_001873.1 NEK2 5.9 8.6E-04 NM_002497.2 CES1 10.5 4.2E-04 NM_001025195.1 CHI3L2 5.9 5.1E-03 NM_004000.2 CFH 10.2 6.1E-05 NM_001014975.1 ZBTB7C 5.8 1.5E-03 NM_001039360.1 HSD11B1 9.8 2.5E-05 NM_181755.1 VSIG4 5.8 3.5E-04 NM_007268.2 EMR3 9.7 7.2E-05 NM_032571.2 TTC39B 5.8 1.2E-02 NM_152574.1 CFH 9.6 3.8E-05 NM_001014975.1 LDLRAP1 5.8 9.3E-05 NM_015627.2 APM-1 9.4 1.2E-03 XM_113971.4 FANCE 5.8 4.8E-06 NM_021922.2 HES2 9.4 1.2E-03 NM_019089.3 CCL18 5.6 1.6E-03 NM_002988.2 KIAA0101 9.3 1.0E-02 NM_014736.4 CSF1 5.5 6.7E-06 NM_172212.1 CSF1 9.2 7.2E-06 NM_000757.3 CDC2 5.4 1.1E-02 NM_001786.2 TOP2A 9.0 4.1E-03 NM_001067.2 TCEA3 5.4 1.2E-02 NM_003196.1 KIF20A 9.0 8.4E-04 NM_005733.1 FAIM2 5.3 3.8E-02 NM_012306.2 ZDHHC19 8.9 8.5E-05 NM_001039617.1 CTSW 5.2 2.1E-03 NM_001335.3 Appendix I

Gene FC p (Corr) Accession Gene FC p (Corr) Accession PITX1 5.2 3.8E-05 NM_002653.3 FOLR3 4.0 2.1E-03 NM_000804.2 CDC2 5.2 6.5E-03 NM_001786.2 CAMP 4.0 1.8E-03 NM_004345.3 APOBEC3B 5.1 8.0E-03 NM_004900.3 VNN2 3.9 3.0E-02 NM_078488.1 LOC644680 5.1 6.5E-03 XM_932300.1 APOL4 3.9 7.6E-04 NM_030643.3 TEX14 5.1 2.5E-02 NM_031272.3 C20ORF127 3.9 1.7E-03 NM_080757.1 IL7R 5.0 1.6E-03 NM_002185.2 WWTR1 3.9 5.6E-03 NM_015472.3 KIF11 5.0 1.8E-03 NM_004523.2 MAPRE3 3.9 6.0E-04 NM_012326.2 ANKRD29 5.0 3.4E-03 NM_173505.2 FAM113B 3.9 2.2E-03 NM_138371.1 CDC20 4.9 3.0E-03 NM_001255.2 SH2B2 3.9 2.9E-04 NM_020979.2 TCF7L2 4.9 2.2E-04 NM_030756.2 TM7SF4 3.9 2.1E-04 NM_030788.2 ITGB3 4.9 7.6E-03 NM_000212.2 LOC653481 3.9 8.2E-03 XM_927594.1 HMMR 4.8 1.2E-03 NM_012485.1 C6ORF105 3.8 1.8E-03 NM_032744.1 IL1F5 4.8 1.9E-02 NM_173170.1 RHBDD2 3.8 4.5E-04 NM_001040456.1 AURKB 4.8 5.1E-03 NM_004217.2 HS.126245 3.8 4.5E-04 CR743148 C12ORF59 4.7 1.8E-03 NM_153022.1 MRC1 3.8 7.6E-04 NM_002438.2 S1PR4 4.7 1.5E-04 NM_003775.2 HPGD 3.8 2.0E-03 NM_000860.3 IL1A 4.7 3.5E-02 NM_000575.3 ZBTB8A 3.8 4.9E-04 NM_001040441.1 ANLN 4.6 3.7E-03 NM_018685.2 SOBP 3.8 1.8E-03 NM_018013.3 KIFC1 4.6 3.8E-03 NM_002263.2 CTNNAL1 3.8 1.5E-03 NM_003798.2 ABCG2 4.6 2.9E-02 NM_004827.2 VASP 3.7 1.5E-04 NM_003370.3 CTNNAL1 4.6 2.2E-03 NM_003798.1 GPD1 3.7 7.6E-03 NM_005276.2 RRM2 4.6 2.0E-02 NM_001034.1 C15ORF21 3.7 9.3E-05 NM_001005267.1 SPIN4 4.6 1.1E-03 NM_001012968.2 TMCC3 3.7 2.9E-04 NM_020698.1 TYMS 4.6 2.2E-02 NM_001071.1 AQP3 3.7 8.0E-04 NM_004925.3 SVIL 4.5 7.6E-04 NM_003174.3 DDO 3.7 7.6E-04 NM_004032.2 SERPINA1 4.5 5.1E-05 NM_001002235.1 IL3RA 3.7 1.7E-02 NM_002183.2 AQP12A 4.5 4.0E-05 NM_198998.1 CLEC12A 3.7 3.3E-04 NM_201623.2 LOC730081 4.4 4.2E-03 XR_041261.1 SERPINA1 3.7 9.3E-05 NM_001002236.1 MT1A 4.4 2.9E-04 NM_005946.2 IL1F10 3.7 9.6E-03 NM_173161.1 MT2A 4.4 2.0E-04 NM_005953.2 SPN 3.6 4.2E-04 NM_003123.3 CES1 4.4 1.8E-03 XM_945034.1 KAL1 3.6 2.6E-02 NM_000216.2 CLEC12A 4.4 2.1E-04 NM_138337.4 KBTBD11 3.6 3.8E-04 NM_014867.1 S100A13 4.3 6.0E-05 NM_001024211.1 PKIA 3.6 5.3E-03 NM_006823.2 GINS2 4.3 8.0E-03 NM_016095.1 NRGN 3.6 6.5E-03 NM_006176.1 IL1RN 4.3 4.2E-03 NM_173843.1 IL1RN 3.6 8.9E-03 NM_173842.1 BUB1 4.3 7.4E-03 NM_004336.2 RHBDD2 3.6 6.4E-04 NM_001040456.1 MPZL2 4.3 1.0E-02 NM_005797.2 DEFB1 3.6 9.3E-03 NM_005218.3 SPINT1 4.2 3.1E-03 NM_003710.3 CLEC12A 3.6 1.5E-04 NM_138337.4 IL7R 4.2 1.6E-03 XM_937367.1 CCND2 3.6 1.1E-02 NM_001759.2 CKAP2L 4.2 1.0E-02 NM_152515.2 APCDD1L 3.5 9.9E-03 NM_153360.1 RGNEF 4.2 2.2E-03 XM_932949.2 LAPTM4B 3.5 5.5E-03 NM_018407.4 GPX3 4.2 1.6E-03 NM_002084.3 CD1B 3.5 3.7E-02 NM_001764.2 CPM 4.2 1.6E-03 NM_001874.3 MS4A14 3.5 1.1E-03 NM_032597.3 MT1E 4.1 3.0E-03 NM_175617.3 LOC728744 3.5 2.0E-02 XM_001128342.1 NDC80 4.1 3.7E-03 NM_006101.1 PAQR5 3.5 4.1E-03 NM_017705.2 ALOX5AP 4.1 1.0E-05 NM_001629.2 SNAI3 3.5 5.9E-05 NM_178310.1 LGALS3BP 4.1 1.5E-03 NM_005567.2 GCHFR 3.5 3.8E-04 NM_005258.2 CCNA2 4.1 2.9E-03 NM_001237.2 ITGA6 3.5 4.0E-03 NM_000210.2 MTMR11 4.1 1.1E-03 NM_181873.2 GALNT12 3.5 2.9E-03 NM_024642.3 HS.116870 4.1 8.3E-04 BG776012 LOC400174 3.5 1.1E-03 XR_016147.1 CCNB2 4.1 1.5E-03 NM_004701.2 ITGA6 3.4 2.4E-03 NM_000210.2 LOC650261 4.1 1.2E-02 XM_939353.1 CDT1 3.4 2.3E-02 NM_030928.2 FCHO1 4.0 1.1E-03 NM_015122.1 ALAS1 3.4 6.0E-05 NM_000688.4 PRSS21 4.0 4.4E-03 NM_144956.1 PRC1 3.4 8.7E-03 NM_199413.1 CCND2 4.0 9.2E-03 NM_001759.2 CENPN 3.4 8.5E-05 NM_018455.3 FLJ21986 4.0 1.0E-02 NM_024913.3 CLEC1A 3.4 2.1E-02 NM_016511.2 Aug-99 4.0 1.8E-03 NM_019106.4 FCGR1B 3.4 2.8E-03 NM_001004340.1 Appendix I

Gene FC p (Corr) Accession Gene FC p (Corr) Accession ARAP3 3.4 1.8E-03 NM_022481.5 P2RX1 2.9 3.6E-03 NM_002558.2 TMEM173 3.3 5.2E-04 NM_198282.1 HS.4892 2.9 1.2E-02 AF131834 CALCOCO1 3.3 3.8E-04 NM_020898.1 CLEC6A 2.9 5.1E-03 NM_001007033.1 PROS1 3.3 5.8E-04 NM_000313.1 CLEC1B 2.9 8.5E-03 NM_016509.3 ANKRD29 3.3 2.7E-02 NM_173505.2 UGP2 2.9 5.2E-05 NM_006759.3 MYOZ1 3.3 1.2E-03 NM_021245.2 KCNE1 2.9 2.9E-03 NM_000219.2 SORT1 3.3 1.7E-04 NM_002959.4 ECHDC2 2.9 3.7E-02 NM_018281.2 GPRIN3 3.3 8.0E-04 NM_198281.2 NMB 2.9 5.2E-04 NM_021077.3 CENPA 3.3 1.5E-02 NM_001042426.1 ADAMTSL4 2.9 7.3E-03 NM_019032.4 TREML3 3.3 1.2E-02 XM_929970.2 Aug-99 2.9 1.7E-03 NM_019106.4 MT1F 3.3 9.2E-03 NM_005949.2 MFGE8 2.9 1.5E-02 NM_005928.1 SLC44A2 3.3 1.7E-04 NM_020428.2 FCGR1C 2.8 2.4E-03 NM_001128589.1 KIF2C 3.3 9.9E-04 NM_006845.2 SEMA4B 2.8 1.6E-03 NM_198925.1 FCGR1A 3.3 1.6E-03 NM_000566.2 HRK 2.8 1.7E-02 NM_003806.1 CDKN3 3.3 3.6E-04 NM_005192.2 NLRP2 2.8 4.9E-02 NM_017852.1 S100A16 3.2 4.4E-03 NM_080388.1 CENPM 2.8 1.0E-02 NM_001002876.1 PPARD 3.2 8.5E-05 NM_006238.2 HIST2H2AB 2.8 1.4E-03 NM_175065.2 PODXL 3.2 1.5E-02 NM_001018111.1 LOC646674 2.8 3.7E-03 XR_042303.1 NLRP12 3.2 2.0E-02 NM_144687.1 GSN 2.8 1.7E-04 NM_198252.2 TMEM44 3.2 2.4E-03 NM_138399.3 CLEC5A 2.8 5.1E-03 NM_013252.2 CTAGE5 3.2 6.1E-04 NM_203355.1 LOC647357 2.8 1.3E-03 XR_001274.1 MCM10 3.2 3.4E-02 NM_018518.3 C17ORF87 2.8 1.6E-03 NM_207103.2 LOC729652 3.2 1.8E-02 XR_040489.1 TSPAN32 2.8 9.9E-03 NM_005705.4 C7ORF58 3.2 2.5E-03 NM_001105533.1 PPIC 2.8 2.2E-02 NM_000943.4 CDCP1 3.2 1.2E-02 NM_178181.1 TSPAN32 2.8 1.1E-03 NM_005705.4 LOC100131391 3.2 1.5E-03 XM_001726689.1 ABCB9 2.8 1.1E-03 NM_019624.2 CHST11 3.1 7.1E-03 NM_018413.3 LOC401002 2.8 1.8E-03 XR_018284.1 LOC641972 3.1 1.6E-03 XM_935742.1 F5 2.8 3.6E-03 NM_000130.4 PPP1R13B 3.1 1.4E-02 NM_015316.2 JUP 2.7 1.3E-03 NM_021991.1 CDCP1 3.1 2.7E-03 NM_178181.1 FAM89A 2.7 1.6E-02 NM_198552.1 TGFA 3.1 1.7E-03 NM_003236.1 UGP2 2.7 4.5E-04 NM_006759.3 FBXO15 3.1 1.6E-03 NM_152676.1 TMEM44 2.7 2.3E-02 NM_138399.3 GJB2 3.1 3.1E-03 NM_004004.4 TMPO 2.7 1.5E-03 NM_003276.1 FBP1 3.1 9.6E-05 NM_000507.2 TPX2 2.7 9.8E-03 NM_012112.4 SPOCD1 3.1 4.8E-03 NM_144569.4 STMN1 2.7 2.6E-02 NM_005563.3 PROCR 3.1 1.1E-02 NM_006404.3 LOC649276 2.7 6.3E-03 XM_938336.1 OR8G1 3.1 2.2E-02 NM_001002905.1 LOC645726 2.7 3.0E-03 XR_018230.2 CEP55 3.0 1.1E-02 NM_018131.3 PKMYT1 2.7 9.0E-03 NM_182687.1 PROK1 3.0 2.3E-04 NM_032414.2 SSBP3 2.7 5.1E-03 NM_018070.3 LOC647108 3.0 1.4E-02 XM_934378.1 FGFRL1 2.7 3.9E-04 NM_021923.3 MGAT3 3.0 4.5E-02 NM_002409.4 SLC39A3 2.7 4.0E-05 NM_213568.1 EMR3 3.0 1.3E-03 NM_032571.3 BUB1B 2.7 3.3E-04 NM_001211.4 MAP1LC3A 3.0 2.0E-03 NM_181509.1 IL1RAP 2.7 3.7E-03 NM_134470.2 CBS 3.0 5.1E-03 NM_000071.1 LOC441019 2.7 4.5E-03 XM_498969.2 MT1H 3.0 2.1E-02 NM_005951.1 ACOX1 2.7 3.9E-03 NM_004035.4 MGC26733 3.0 3.8E-04 NM_144992.3 NUSAP1 2.7 4.1E-03 NM_018454.5 ABCB9 3.0 2.4E-03 NM_019624.2 SYNC 2.6 9.7E-03 NM_030786.2 C11ORF21 3.0 1.6E-02 NM_001142946.1 ABAT 2.6 1.4E-02 NM_000663.3 FCGR1B 2.9 2.5E-03 NM_001017986.1 ACVRL1 2.6 2.9E-03 NM_000020.1 COL8A2 2.9 4.0E-05 NM_005202.1 QSOX1 2.6 2.3E-04 NM_002826.4 SPN 2.9 2.7E-04 NM_001030288.1 CDCP1 2.6 7.7E-03 NM_178181.1 SSPN 2.9 1.3E-03 NM_005086.3 CPM 2.6 4.4E-02 NM_001005502.1 HES2 2.9 3.5E-03 NM_019089.3 KIF11 2.6 1.1E-02 NM_004523.2 AOC3 2.9 2.2E-02 NM_003734.2 BHLHB3 2.6 2.7E-04 NM_030762.1 CACNG1 2.9 1.3E-03 NM_000727.2 MCF2L2 2.6 3.9E-02 NM_015078.2 GPT 2.9 5.6E-03 NM_005309.1 PCGF2 2.6 4.4E-03 NM_007144.2 LMNB1 2.9 9.2E-03 NM_005573.2 LOC389049 2.6 4.2E-03 XR_017334.2 Appendix I

Gene FC p (Corr) Accession Gene FC p (Corr) Accession CDC42EP3 2.6 3.7E-02 NM_006449.3 CD276 2.4 1.7E-03 NM_025240.2 SH2D3C 2.6 7.9E-03 NM_170600.1 HS.563550 2.4 1.9E-02 BQ184411 FAM64A 2.6 6.7E-03 NM_019013.1 HPCAL1 2.4 4.7E-03 NM_134421.1 ABCG1 2.6 1.8E-02 NM_016818.2 E2F2 2.4 2.9E-02 NM_004091.2 QPCT 2.6 1.2E-03 NM_012413.3 ACTA2 2.4 1.3E-02 NM_001613.1 COL8A2 2.6 3.2E-04 NM_005202.1 OIP5 2.4 3.1E-02 NM_007280.1 SAV1 2.6 8.9E-04 NM_021818.2 HS.561747 2.4 1.5E-04 BU935198 TCN1 2.6 2.5E-02 NM_001062.3 S100A13 2.4 4.4E-03 NM_005979.2 AMICA1 2.6 7.6E-03 NM_153206.1 LPCAT2 2.4 2.7E-04 NM_017839.3 LOC100132510 2.6 3.6E-04 XM_001725190.1 POLD1 2.4 5.5E-04 NM_002691.1 LOC653204 2.6 1.9E-03 XM_926469.1 RAB17 2.4 3.2E-02 NM_022449.1 OIP5 2.6 1.2E-02 NM_007280.1 DNASE2B 2.4 2.9E-03 NM_021233.2 VDR 2.6 4.7E-03 NM_000376.2 LOC729057 2.4 2.7E-03 XR_042044.1 CFH 2.6 1.7E-03 NM_001014975.1 PELI3 2.4 7.8E-03 NM_145065.1 LOC441241 2.6 8.4E-04 XM_935516.1 LOC730546 2.4 9.5E-03 XM_001126287.1 ALAS1 2.5 2.1E-04 NM_199166.1 DENND5A 2.4 4.0E-03 NM_015213.2 RASL11A 2.5 5.7E-03 NM_206827.1 USP30 2.4 1.2E-03 NM_032663.2 PRKAR1A 2.5 4.3E-04 NM_212471.1 TGM2 2.3 9.2E-03 NM_004613.2 CRABP2 2.5 3.1E-02 NM_001878.2 ABHD14A 2.3 1.4E-02 NM_015407.3 PPIC 2.5 3.8E-02 NM_000943.4 CLEC11A 2.3 2.2E-03 NM_002975.2 HS.173957 2.5 3.0E-03 AW967735 C14ORF106 2.3 2.2E-03 NM_018353.3 C14ORF72 2.5 5.9E-03 XM_944937.2 LOC1001290342.3 1.2E-03 XM_001720357.1 LOC387841 2.5 4.4E-03 XM_932678.1 BACE1 2.3 4.2E-03 NM_138972.2 SPC24 2.5 6.6E-03 NM_182513.1 C17ORF47 2.3 2.5E-02 NM_001038704.1 ALOX5 2.5 1.7E-02 NM_000698.2 C5ORF4 2.3 7.9E-03 NM_032385.3 SIGLEC11 2.5 2.5E-02 NM_052884.1 C17ORF87 2.3 1.1E-02 NM_207103.1 KIAA0922 2.5 2.5E-03 NM_015196.2 LOC729259 2.3 2.2E-03 XR_037945.1 FAM19A3 2.5 1.4E-02 NM_182759.2 C5ORF27 2.3 2.7E-03 XR_040299.1 NDUFV3 2.5 1.0E-03 NM_001001503.1 NUAK2 2.3 1.5E-03 NM_030952.1 FAM89A 2.5 2.2E-03 NM_198552.1 MGC33556 2.3 2.0E-02 NM_001004307.1 SPINT1 2.5 3.1E-03 NM_003710.3 PEX14 2.3 1.9E-03 NM_004565.2 C1QB 2.5 4.2E-02 NM_000491.3 NFIX 2.3 2.6E-02 NM_002501.2 UHRF1 2.5 3.2E-02 NM_001048201.1 ZNF609 2.3 5.9E-03 NM_015042.1 SLC22A16 2.5 3.7E-03 NM_033125.2 LIPN 2.3 1.0E-02 NM_001102469.1 TNFSF13 2.5 1.5E-02 NM_172088.1 FCGR2B 2.3 3.5E-03 NM_004001.3 PDE1B 2.5 3.5E-02 NM_000924.2 ABAT 2.3 8.0E-04 NM_000663.3 NEXN 2.5 2.1E-03 NM_144573.3 HS.187499 2.3 1.9E-02 BC063871 EPR1 2.5 3.1E-02 NR_002219.1 FAM89A 2.3 8.8E-03 NM_198552.1 TMPO 2.5 3.2E-04 NM_003276.1 FLJ40453 2.3 4.7E-02 NM_001007542.1 PDE3B 2.4 4.8E-03 NM_000922.2 BCL2A1 2.3 6.6E-03 NM_004049.2 CDCA3 2.4 1.7E-02 NM_031299.3 SMAD7 2.3 4.2E-03 NM_005904.2 C9ORF127 2.4 1.5E-03 NM_016446.2 TSPAN18 2.3 8.5E-03 NM_130783.3 CDS1 2.4 1.5E-03 NM_001263.2 RAPGEFL1 2.3 3.9E-02 NM_016339.2 PSD3 2.4 2.5E-04 NM_015310.3 B3GALTL 2.3 5.6E-04 NM_194318.3 PLK1 2.4 1.9E-02 NM_005030.3 ATP6V1E2 2.3 1.3E-03 NM_080653.3 PIK3CG 2.4 2.1E-04 NM_002649.2 CD300LF 2.3 8.5E-04 NM_139018.2 FMNL3 2.4 3.7E-02 NM_198900.2 TCEB2 2.3 2.8E-03 NM_007108.2 IL8RB 2.4 2.3E-03 NM_001557.2 LOC1001291692.3 2.9E-03 XM_001716328.1 SMAD7 2.4 1.1E-03 NM_005904.2 JUP 2.3 2.1E-03 NM_002230.1 MSI2 2.4 2.1E-02 NM_170721.1 MND1 2.3 3.0E-02 NM_032117.2 KIF23 2.4 1.1E-03 NM_004856.4 ST6GALNAC4 2.3 1.5E-03 NM_175039.3 NUSAP1 2.4 2.3E-03 NM_018454.5 NUP210 2.3 2.4E-02 NM_024923.2 RAPGEF3 2.4 4.2E-02 NM_006105.3 LIMK2 2.3 7.0E-03 NM_001031801.1 PPARG 2.4 7.6E-04 NM_138712.3 PPP2R1B 2.3 3.2E-03 NM_181699.2 LBR 2.4 3.8E-04 NM_002296.2 LOC93349 2.2 9.3E-03 NM_138402.3 FAM135B 2.4 3.7E-03 NM_015912.3 PLAUR 2.2 2.2E-03 NM_002659.2 P2RY12 2.4 2.3E-02 NM_022788.3 PLEKHG4 2.2 2.2E-03 NM_015432.2 Appendix I

Gene FC p (Corr) Accession Gene FC p (Corr) Accession TUT1 2.2 2.2E-03 NM_022830.1 AK1 2.1 5.6E-03 NM_000476.1 PFTK1 2.2 2.9E-03 NM_012395.2 CENPA 2.1 1.7E-02 NM_001042426.1 MYO1G 2.2 3.7E-02 NM_033054.1 HLA-DQB2 2.1 1.8E-03 NM_182549.1 LNPEP 2.2 2.2E-02 NM_175920.3 LOC1001313642.1 6.2E-03 XM_001725122.1 CTXN1 2.2 1.8E-03 NM_206833.2 DENND5A 2.1 6.5E-03 NM_015213.2 RTN2 2.2 4.5E-03 NM_206901.1 FAM89A 2.1 2.0E-02 XM_939093.1 SSBP3 2.2 5.4E-03 NM_018070.3 CENPK 2.1 3.8E-02 NM_022145.3 GSN 2.2 3.7E-02 NM_000177.3 CA11 2.1 1.1E-02 NM_001217.3 ABHD2 2.2 1.6E-03 NM_152924.3 RECK 2.1 3.1E-03 NM_021111.1 CD276 2.2 1.3E-02 NM_025240.2 KIAA0564 2.1 6.8E-04 NM_015058.1 CCNB1 2.2 1.6E-03 NM_031966.2 MPO 2.1 2.3E-02 NM_000250.1 ACOT1 2.2 9.7E-03 NM_001037161.1 IGSF8 2.1 1.9E-03 NM_052868.2 EFCAB4A 2.2 2.2E-02 NM_173584.3 RPS6KA1 2.1 2.5E-04 NM_002953.3 NDST1 2.2 2.0E-03 NM_001543.3 FCGR2B 2.1 6.9E-03 XM_938851.1 LY6E 2.2 1.0E-02 NM_002346.1 ITPKB 2.1 9.0E-04 NM_002221.2 BACE1 2.2 1.2E-02 NM_138971.2 TRAF3IP2 2.1 7.0E-03 NM_147686.1 MGC12538 2.2 7.1E-03 XR_040874.1 TBC1D10C 2.1 2.0E-02 NM_198517.2 NCF1 2.2 3.8E-02 NM_000265.4 C1ORF210 2.1 1.2E-02 NM_182517.1 KCNQ1 2.2 1.2E-02 NM_000218.2 LOC284757 2.1 1.1E-02 NM_001004305.1 SDCCAG8 2.2 1.7E-03 NM_006642.2 HELLS 2.1 1.4E-02 NM_018063.3 GK 2.2 1.1E-03 NM_203391.1 SLC25A40 2.1 1.6E-03 NM_018843.2 MYD88 2.2 3.7E-04 NM_002468.3 DOK2 2.1 1.6E-02 NM_003974.2 SLC27A3 2.2 7.3E-04 NM_024330.1 TRIP13 2.1 1.2E-02 NM_004237.2 EML4 2.2 5.4E-04 NM_019063.2 SEL1L3 2.1 7.5E-03 NM_015187.3 SPINK6 2.2 5.1E-03 NM_205841.2 FPR2 2.1 3.9E-02 NM_001005738.1 PTTG3P 2.2 9.2E-03 NR_002734.1 METTL9 2.0 1.5E-02 NM_016025.3 CSPG5 2.2 9.0E-04 NM_006574.3 LOC152586 2.0 1.8E-02 XM_929404.1 ACE 2.2 1.5E-02 NM_152830.1 KILLIN 2.0 3.9E-02 NM_001126049.1 VASH1 2.2 4.8E-04 NM_014909.3 ANKRD28 2.0 5.4E-03 NM_015199.2 ICAM2 2.2 4.5E-02 NM_001099786.1 HS.518527 2.0 1.3E-02 XM_374071 C13ORF3 2.2 2.9E-03 NM_145061.3 CSTA 2.0 7.6E-04 NM_005213.3 ABHD2 2.2 2.1E-03 NM_152924.3 CHAF1A 2.0 3.7E-02 NM_005483.2 KIAA0564 2.2 3.7E-03 NM_015058.1 HVCN1 2.0 1.7E-03 NM_001040107.1 LOC642456 2.2 1.1E-03 XM_925964.1 PTPN22 2.0 9.6E-04 NM_015967.3 PTPN22 2.2 1.9E-03 NM_012411.2 ACO1 2.0 4.7E-04 NM_002197.1 LIN7B 2.2 1.6E-03 NM_022165.2 PHLDA3 2.0 1.1E-02 NM_012396.3 S100A13 2.2 1.6E-03 NM_001024212.1 NT5C3L 2.0 2.0E-02 NM_052935.2 GK 2.2 2.7E-03 NM_000167.3 ABCG1 2.0 4.8E-02 NM_016818.2 TSGA14 2.2 2.0E-04 NM_018718.1 LOC652616 2.0 9.7E-03 XM_942152.1 C12ORF37 2.2 2.7E-03 XM_001134418.1 UFD1L 2.0 3.3E-02 NM_001035247.1 MYCBP 2.1 2.2E-03 NM_012333.3 MCM7 2.0 1.7E-02 NM_182776.1 HVCN1 2.1 1.7E-03 NM_032369.2 ITGA5 2.0 4.8E-03 NM_002205.2 C1S 2.1 2.7E-02 NM_001734.2 CD276 2.0 3.7E-03 NM_001024736.1 PLAUR 2.1 1.1E-03 NM_001005376.1 LOC1001334592.0 2.3E-03 XM_001720054.1 DENND3 2.1 7.3E-03 NM_014957.2 GAS1 2.0 5.5E-03 NM_002048.1 FAM100A 2.1 1.6E-03 NM_145253.2 CKLF 2.0 4.3E-04 NM_001040139.1 ZMYND15 2.1 1.2E-03 NM_032265.1 HS.191602 2.0 3.2E-02 AI821625 MPZL2 2.1 4.5E-03 NM_144765.1 EVL 2.0 4.7E-03 NM_016337.2 NEK6 2.1 5.4E-03 NM_014397.3 ST6GALNAC4 2.0 1.5E-03 NM_175039.3 SLC11A1 2.1 5.4E-03 NM_000578.2 DHCR24 2.0 2.4E-02 NM_014762.3 CTAGE6 2.1 3.7E-03 XM_936776.2 CCDC102B 2.0 1.5E-02 NM_001093729.1 PTPN22 2.1 3.7E-03 NM_015967.3 AKR1C3 2.0 3.9E-02 NM_003739.4 NEUROG2 2.1 8.9E-04 NM_024019.2 SLC9A1 2.0 1.6E-03 NM_003047.2 SLC25A29 2.1 1.2E-02 NM_001039355.1 LOC730254 2.0 2.2E-02 XM_001724977.1 LAPTM4B 2.1 6.8E-03 NM_018407.4 INSIG1 2.0 2.1E-02 NM_198336.1 FPR2 2.1 2.6E-02 NM_001462.3 PTCRA 2.0 3.4E-02 NM_138296.2 RASAL2 2.1 6.3E-03 NM_170692.1 KIF15 2.0 1.3E-02 NM_020242.1 Appendix II

All genes down-regulated ≥ 2 fold change in GM-MΦ versus M-MΦ

Appendix II

Gene FC p (Corr) Accession Gene FC p (Corr) Accession SEPP1 -89.9 1.1E-05 NM_005410.2 HPSE -6.9 3.3E-04 NM_006665.3 M160 -52.6 7.8E-06 NM_174941.3 GNG2 -6.9 3.9E-04 NM_053064.3 ADORA3 -45.3 4.5E-06 NM_001081976.1 ZNF467 -6.9 3.4E-04 NM_207336.1 LEP -34.1 7.8E-06 NM_000230.1 HPSE -6.9 1.0E-04 NM_006665.2 CACNA2D3 -34.0 3.9E-06 NM_018398.2 SRGAP3 -6.8 2.8E-04 NM_001033117.1 STAB1 -30.1 1.0E-05 NM_015136.2 DEPDC6 -6.7 4.6E-04 NM_022783.1 SERPINB2 -25.6 1.0E-05 NM_002575.1 EBI3 -6.7 1.4E-03 NM_005755.2 SERPINB2 -23.5 3.3E-05 NM_002575.1 ARL4C -6.6 1.1E-04 NM_005737.3 RNASE1 -22.2 1.7E-03 NM_198232.1 DEPDC6 -6.6 5.0E-05 NM_022783.1 CD163L1 -21.3 4.0E-05 NM_174941.4 RARRES1 -6.6 4.3E-05 NM_206963.1 SLC40A1 -19.7 5.3E-06 NM_014585.3 CUL9 -6.4 2.1E-05 NM_015089.2 IGF1 -19.0 6.8E-05 NM_000618.2 GAMT -6.4 3.3E-05 NM_000156.4 FLJ14213 -17.9 3.9E-06 NM_024841.3 NOV -6.3 2.3E-04 NM_002514.2 SERPINF1 -16.6 3.6E-05 NM_002615.4 TCF4 -6.1 3.4E-04 NM_003199.2 LGMN -16.2 3.4E-05 NM_001008530.1 ASIP -6.0 6.9E-05 NM_001672.2 CADM1 -15.7 3.1E-05 NM_014333.3 GATM -6.0 5.7E-04 NM_001482.2 ADORA3 -15.4 1.2E-04 NM_000677.3 HS.197143 -5.9 4.7E-05 AK126405 TMEM119 -15.4 4.2E-04 NM_181724.1 ADAM8 -5.8 9.7E-04 NM_001109.3 OLFML2B -15.2 8.6E-04 NM_015441.1 CCL2 -5.8 9.2E-03 NM_002982.3 SGPP2 -15.2 6.9E-04 XM_938742.1 SLC9A9 -5.6 4.7E-05 NM_173653.1 DLL1 -14.8 3.9E-06 NM_005618.3 BIRC3 -5.6 1.4E-03 NM_001165.3 SLC40A1 -14.7 6.2E-05 NM_014585.4 RASAL1 -5.6 1.7E-04 NM_004658.1 COL23A1 -14.2 2.1E-05 NM_173465.2 EDNRB -5.6 7.8E-06 NM_000115.1 BAALC -14.1 9.2E-06 NM_001024372.1 PTGDS -5.5 2.8E-04 NM_000954.5 RNASE1 -14.1 3.4E-03 NM_198235.1 MTSS1 -5.5 5.6E-05 NM_014751.4 HS3ST1 -13.4 2.6E-04 NM_005114.2 IL10 -5.5 5.4E-04 NM_000572.2 RBP1 -12.3 5.7E-04 NM_002899.2 TPST1 -5.5 2.0E-02 NM_003596.2 RAB3IL1 -12.1 8.6E-04 NM_013401.2 DHDH -5.5 3.3E-03 NM_014475.3 CD38 -11.9 1.0E-04 NM_001775.2 MARCKS -5.4 6.4E-05 NM_002356.5 TNFRSF9 -11.5 6.3E-04 NM_001561.4 LPAR5 -5.3 3.8E-03 NM_020400.4 FZD2 -11.5 1.7E-04 NM_001466.2 CXCL12 -5.3 6.0E-03 NM_199168.2 ADAMDEC1 -11.3 2.1E-05 NM_014479.2 MTSS1 -5.3 1.4E-05 NM_014751.2 TPSAB1 -11.3 3.4E-02 NM_003294.3 RGS20 -5.2 3.1E-03 NM_170587.1 CCL8 -11.0 6.5E-03 NM_005623.2 NCKAP1 -5.1 1.3E-03 NM_013436.3 BAALC -10.9 1.2E-04 NM_024812.2 PDK4 -5.1 3.2E-03 NM_002612.3 LILRB5 -10.2 3.1E-03 NM_006840.3 GSDMA -5.1 5.1E-03 NM_178171.4 EBI2 -9.4 7.3E-05 NM_004951.3 DYRK3 -5.1 1.5E-04 NM_003582.2 LGMN -9.3 1.1E-04 NM_001008530.1 VCAM1 -5.1 4.5E-03 NM_001078.2 CLEC10A -9.2 2.8E-03 NM_006344.2 ADA -5.0 1.2E-04 NM_000022.2 SCIN -9.1 1.3E-05 NM_033128.1 ATP8A1 -5.0 1.1E-05 NM_006095.1 FAM174B -8.9 5.3E-06 NM_207446.2 NT5DC2 -5.0 1.6E-05 NM_022908.1 IL18BP -8.3 2.1E-05 NM_173042.2 CD320 -5.0 4.3E-04 NM_016579.2 SLC35F2 -8.2 3.8E-03 NM_017515.3 TIFAB -5.0 7.3E-04 NM_001099221.1 EBI2 -8.2 3.8E-05 NM_004951.3 C1ORF128 -5.0 1.3E-05 NM_020362.3 C10ORF116 -8.2 8.6E-04 NM_006829.2 TTYH2 -5.0 1.4E-04 NM_032646.5 GPR183 -8.0 1.9E-05 NM_004951.4 TSPAN10 -5.0 1.1E-04 NM_031945.3 ADAMDEC1 -7.9 2.6E-05 NM_014479.2 MMP25 -5.0 1.3E-03 NM_022468.4 SDS -7.8 2.2E-04 NM_006843.2 COBLL1 -5.0 7.7E-04 NM_014900.3 CMKLR1 -7.8 6.8E-05 NM_004072.1 KCNJ5 -4.9 1.3E-02 NM_000890.3 TMEM37 -7.8 6.7E-03 NM_183240.2 HS.561679 -4.9 1.2E-04 DA830074 RARRES1 -7.4 3.6E-05 NM_002888.2 PNMA3 -4.9 3.4E-03 NM_013364.4 KITLG -7.4 5.4E-06 NM_000899.3 TPCN1 -4.9 8.6E-04 NM_017901.3 HTR2B -7.3 8.7E-04 NM_000867.3 CDH23 -4.8 8.9E-04 NM_052836.1 P2RY6 -7.3 3.4E-04 NM_176796.1 KLHL13 -4.7 3.3E-04 NM_033495.2 CCDC81 -7.2 4.0E-05 NM_021827.3 NCKAP1 -4.7 5.1E-04 NM_205842.1 CLEC10A -7.0 2.2E-03 NM_006344.2 LY9 -4.7 1.3E-03 NM_002348.2 KITLG -7.0 4.1E-05 NM_000899.3 PDPN -4.7 6.4E-03 NM_001006625.1 Appendix II

Gene FC p (Corr) Accession Gene FC p (Corr) Accession PRDM1 -4.6 9.0E-05 NM_001198.2 ETV5 -3.6 1.0E-03 NM_004454.1 IGF1 -4.6 1.6E-03 NM_000618.2 LY9 -3.6 8.5E-03 NM_001033667.1 STARD13 -4.6 1.6E-03 NM_178006.1 FBN2 -3.6 4.7E-02 NM_001999.3 FARP1 -4.6 3.2E-04 NM_005766.2 FOXRED2 -3.6 2.1E-03 NM_024955.4 RARRES1 -4.5 6.9E-05 NM_002888.2 FOLR2 -3.6 1.4E-03 NM_000803.3 BIRC3 -4.5 6.8E-04 NM_001165.3 CXCL12 -3.5 2.8E-02 NM_000609.4 LINCR -4.5 5.4E-04 NM_001080535.1 CD200R1 -3.5 3.7E-03 NM_138806.3 RIMS3 -4.5 1.8E-03 NM_014747.2 ABCB6 -3.5 1.5E-03 NM_005689.1 PLTP -4.5 4.2E-04 NM_182676.1 NINL -3.5 3.1E-03 NM_025176.4 ASNS -4.4 4.5E-03 NM_133436.1 PTGIR -3.5 4.8E-04 NM_000960.3 LGI2 -4.4 3.9E-03 NM_018176.2 MS4A6A -3.5 2.3E-03 NM_152851.1 C6ORF192 -4.4 2.1E-03 NM_052831.2 IER3 -3.4 1.3E-02 NM_003897.3 OSGIN1 -4.4 1.6E-03 NM_182980.2 CXCL10 -3.4 1.4E-02 NM_001565.2 ABCC5 -4.3 5.2E-04 NM_001023587.1 FUCA1 -3.4 3.3E-04 NM_000147.3 HS.547277 -4.3 3.3E-04 W91949 ABCA3 -3.4 1.6E-03 NM_001089.1 PLTP -4.2 1.8E-03 NM_006227.2 FLJ77644 -3.4 6.8E-04 XM_001133074.2 HTRA1 -4.2 4.2E-03 NM_002775.3 SEMA4A -3.4 3.4E-04 NM_022367.2 CMTM8 -4.2 9.7E-04 NM_178868.3 SEPP1 -3.3 2.8E-03 NM_005410.2 SIGLEC15 -4.2 6.8E-04 NM_213602.1 C14ORF78 -3.3 5.8E-03 XM_001132404.1 HS3ST2 -4.2 1.1E-02 NM_006043.1 MPEG1 -3.3 3.4E-04 XM_166227.6 SNORD99 -4.2 2.8E-04 NR_003077.1 ABCC5 -3.3 4.0E-04 NM_001023587.1 PRDM1 -4.1 3.4E-04 NM_001198.2 NCF4 -3.3 3.4E-04 NM_013416.2 HES4 -4.1 3.3E-03 NM_021170.2 AHNAK2 -3.3 1.7E-03 NM_138420.2 ALOX15B -4.1 5.9E-03 NM_001141.2 RCN3 -3.3 6.3E-04 NM_020650.2 OAF -4.1 9.4E-04 NM_178507.2 BNIP3 -3.3 1.6E-03 NM_004052.2 QPRT -4.1 5.9E-05 NM_014298.3 CDK6 -3.3 5.1E-04 NM_001259.5 CLEC10A -4.1 1.3E-02 NM_182906.2 HSPA6 -3.3 3.9E-03 NM_002155.3 HGF -4.1 2.8E-04 NM_000601.4 SEZ6L2 -3.3 4.9E-02 NM_201575.1 PSAT1 -4.0 2.6E-02 NM_021154.3 CHST13 -3.3 5.5E-04 NM_152889.1 SNAPC1 -4.0 1.6E-04 NM_003082.2 EYA2 -3.3 6.0E-04 NM_172112.1 BNC2 -4.0 9.3E-03 NM_017637.5 NQO1 -3.2 3.6E-03 NM_000903.2 IL18BP -4.0 5.6E-05 NM_173042.2 BIRC3 -3.2 1.0E-03 NM_182962.1 MYO7A -4.0 8.9E-04 NM_000260.2 DPP4 -3.2 1.0E-03 NM_001935.3 ADM -4.0 1.7E-03 NM_001124.1 NIPAL4 -3.2 2.9E-02 NM_001099287.1 RN5S9 -3.9 2.0E-02 NR_023371.1 PPM1L -3.2 1.4E-03 NM_139245.2 MS4A6A -3.9 4.4E-03 NM_022349.2 HS.259679 -3.2 1.9E-02 AW956608 SEMA3A -3.9 7.4E-03 NM_006080.2 CCNA1 -3.2 4.6E-03 NM_003914.2 DDIT4L -3.9 2.5E-05 NM_145244.2 IL1R2 -3.2 3.0E-02 NM_173343.1 TMEM26 -3.9 4.5E-03 NM_178505.5 SULT1C2 -3.2 4.6E-02 NM_001056.3 RGL1 -3.9 2.8E-05 NM_015149.3 SLC2A5 -3.2 3.3E-03 NM_003039.1 MEF2C -3.9 2.8E-03 NM_002397.2 CDH23 -3.2 2.0E-03 NM_022124.3 GAL3ST4 -3.9 2.5E-03 NM_024637.4 FBLN5 -3.2 3.4E-04 NM_006329.2 ALOX15B -3.9 2.0E-02 NM_001141.2 HS.560357 -3.1 1.3E-02 AI539492 CCL5 -3.8 4.2E-03 NM_002985.2 ADAM28 -3.1 4.9E-04 NM_014265.4 ABCB4 -3.8 3.0E-03 NM_018850.2 ZBTB46 -3.1 3.4E-04 NM_025224.2 CHCHD6 -3.8 1.3E-03 NM_032343.1 CD200R1 -3.1 1.4E-03 NM_138940.2 MS4A6A -3.8 2.0E-03 NM_152851.1 ALOX15B -3.1 2.4E-02 NM_001039131.1 KCNJ5 -3.8 1.1E-02 NM_000890.3 ZSWIM4 -3.1 7.1E-03 NM_023072.1 KIR2DL4 -3.8 1.4E-03 NM_002255.3 LTC4S -3.1 3.3E-03 NM_145867.1 CDH23 -3.8 2.7E-03 NM_052836.1 LEP -3.1 5.1E-03 NM_000230.1 P2RY8 -3.8 1.9E-03 NM_178129.3 OKL38 -3.1 5.9E-04 NM_013370.2 DYRK3 -3.7 2.2E-03 NM_003582.2 ABP1 -3.1 1.8E-02 NM_001091.2 DNM3 -3.7 3.1E-03 NM_015569.2 ABHD12 -3.1 5.2E-04 NM_015600.3 TMEM163 -3.7 2.6E-03 NM_030923.3 ETV5 -3.1 8.3E-04 NM_004454.1 MEGF6 -3.7 2.8E-04 NM_001409.3 Feb-05 -3.1 1.6E-03 NM_138396.4 UBR4 -3.7 6.7E-04 NM_020765.2 C5ORF20 -3.1 2.1E-03 NM_130848.2 PCSK6 -3.7 5.1E-03 NM_138325.2 FOXRED2 -3.0 1.4E-03 NM_024955.4 Appendix II

Gene FC p (Corr) Accession Gene FC p (Corr) Accession MGC16384 -3.0 1.1E-04 XM_001726755.1 IL4I1 -2.7 3.4E-04 NM_172374.1 LAT -3.0 2.3E-02 NM_001014987.1 CLIC2 -2.7 1.2E-03 NM_001289.4 CCDC28B -3.0 4.3E-03 NM_024296.2 SAMD4A -2.7 5.6E-03 NM_015589.3 PIP -3.0 8.2E-03 NM_002652.2 S100A12 -2.7 9.0E-03 NM_005621.1 C14ORF45 -3.0 1.1E-03 NM_025057.1 C9ORF98 -2.7 1.1E-02 NM_152572.2 C20ORF26 -3.0 1.9E-02 NM_015585.2 CMBL -2.7 4.5E-02 NM_138809.3 RHBDL2 -3.0 7.3E-04 NM_017821.3 FNDC5 -2.7 3.1E-03 NM_153756.1 REPS2 -3.0 1.5E-02 NM_004726.2 PKD2L1 -2.7 1.7E-02 NM_016112.2 RASSF2 -3.0 3.1E-05 NM_170774.1 USP46 -2.7 1.3E-04 NM_022832.2 HS.40061 -3.0 1.4E-04 BX112120 SEZ6L2 -2.7 1.3E-02 NM_201575.1 MYO10 -3.0 4.8E-04 NM_012334.1 SGTB -2.7 1.1E-02 NM_019072.2 NAV2 -3.0 8.4E-04 NM_145117.3 ZKSCAN1 -2.7 3.7E-03 NM_003439.1 MME -2.9 3.1E-02 NM_000902.3 LOC728499 -2.7 1.8E-04 XM_001127568.1 HS.579631 -2.9 1.6E-03 BU536065 ABCC5 -2.7 7.0E-04 NM_005688.2 OLFML3 -2.9 4.8E-04 NM_020190.2 GNGT2 -2.7 7.0E-04 NM_031498.1 SLC35F2 -2.9 5.9E-03 NM_017515.3 CXCL9 -2.7 3.5E-02 NM_002416.1 TSC22D3 -2.9 1.4E-03 NM_004089.3 H2AFY2 -2.7 1.5E-04 NM_018649.2 DDB1 -2.9 1.9E-03 XM_943551.1 SPATA7 -2.7 3.1E-03 NM_018418.2 LOC100170939 -2.9 9.6E-03 NR_024054.1 LOC100190986 -2.7 5.0E-03 NR_024456.1 SNORA33 -2.9 2.8E-04 NR_002436.1 LOC143666 -2.7 4.5E-03 XM_001127524.1 TSPAN33 -2.9 1.4E-03 NM_178562.2 SNORA61 -2.7 1.8E-04 NR_002987.1 TRAF3IP3 -2.9 2.6E-03 NM_025228.1 LOC100132564 -2.6 6.9E-03 XM_001713808.1 ENC1 -2.9 1.7E-03 NM_003633.1 MAP4K1 -2.6 2.0E-03 NM_001042600.1 FERMT2 -2.9 1.0E-02 NM_006832.1 ZNF774 -2.6 6.5E-03 NM_001004309.1 LOC653879 -2.8 7.4E-03 XM_936226.1 PTGFRN -2.6 4.3E-02 NM_020440.2 FAM125B -2.8 8.9E-04 NM_033446.1 FAM114A1 -2.6 3.9E-03 NM_138389.1 BASP1 -2.8 1.4E-03 NM_006317.3 ZNF296 -2.6 1.6E-04 NM_145288.1 MFSD2 -2.8 9.0E-03 NM_032793.2 TRAF5 -2.6 1.0E-02 NM_004619.3 SLMO1 -2.8 4.3E-03 NM_006553.2 UCHL1 -2.6 2.9E-02 NM_004181.3 TSPAN4 -2.8 6.0E-03 NM_001025234.1 KLF11 -2.6 1.3E-02 XM_001129527.1 LAT -2.8 2.6E-02 NM_001014987.1 PGBD5 -2.6 2.5E-02 NM_024554.2 REPS2 -2.8 9.4E-03 NM_004726.2 MCOLN2 -2.6 7.3E-03 NM_153259.2 DYSF -2.8 2.5E-02 NM_003494.2 PTK6 -2.6 1.8E-03 NM_005975.2 ZFP36L1 -2.8 1.3E-04 NM_004926.2 SNORA41 -2.6 6.5E-03 NR_002590.1 SPATA7 -2.8 4.4E-03 NM_001040428.2 LOC729779 -2.6 2.6E-02 XR_019592.2 HS.536451 -2.8 1.3E-03 BX105743 SLAMF7 -2.6 9.0E-03 NM_021181.3 TRAF1 -2.8 2.1E-02 NM_005658.3 CYP26B1 -2.6 4.3E-03 NM_019885.2 SLMO1 -2.8 3.5E-04 NM_006553.2 SMPDL3A -2.6 5.6E-03 NM_006714.2 CCL5 -2.8 1.4E-02 NM_002985.2 TMOD2 -2.6 9.1E-04 NM_014548.3 RASSF2 -2.8 5.0E-05 NM_170773.1 ACRBP -2.6 2.1E-04 NM_032489.2 BCL2 -2.8 9.6E-03 NM_000633.2 TOM1L2 -2.6 7.2E-03 NM_001082968.1 TMOD2 -2.8 1.5E-03 NM_001142885.1 LOC339843 -2.6 1.6E-03 XR_016598.2 OSBPL7 -2.8 2.6E-03 NM_145798.2 PPM1K -2.6 3.7E-03 NM_152542.2 C9ORF91 -2.8 2.6E-04 NM_153045.2 STARD13 -2.6 2.3E-04 NM_178008.1 HS.171481 -2.8 2.6E-03 CN484989 SLC46A1 -2.6 1.4E-03 NM_080669.3 SMPDL3A -2.8 3.9E-02 NM_006714.2 ABCC4 -2.6 5.8E-04 NM_005845.2 TRAF5 -2.8 7.9E-04 NM_004619.3 FCGR2A -2.6 3.3E-04 NM_021642.2 HS.562504 -2.8 8.2E-04 AW364673 RNU1-3 -2.5 2.6E-02 NR_004408.1 PAQR7 -2.8 7.6E-03 NM_178422.4 SLAIN1 -2.5 1.9E-03 NM_001040153.2 ITSN1 -2.8 8.5E-03 NM_003024.2 SAMSN1 -2.5 9.9E-05 NM_022136.3 KIR2DL3 -2.8 4.5E-02 NM_014511.3 NBPF11 -2.5 3.5E-03 NM_001101663.1 REPS2 -2.8 1.1E-02 NM_001080975.1 CD300A -2.5 2.4E-02 NM_007261.2 CFP -2.8 1.3E-02 NM_002621.1 NFKB2 -2.5 2.5E-03 NM_001077493.1 PLXNC1 -2.8 1.2E-03 NM_005761.1 ETS2 -2.5 1.2E-02 NM_005239.4 LAMA5 -2.7 1.8E-03 NM_005560.3 EPHB2 -2.5 9.1E-03 NM_004442.6 GCLM -2.7 1.1E-02 NM_002061.2 HS.434957 -2.5 4.1E-03 BC004287 GAMT -2.7 7.6E-04 NM_000156.4 KCNMB4 -2.5 2.9E-03 NM_014505.4 Appendix II

Gene FC p (Corr) Accession Gene FC p (Corr) Accession TSHZ3 -2.5 1.1E-03 NM_020856.2 FCGR2A -2.3 2.0E-04 NM_021642.2 ABCC4 -2.5 3.3E-04 NM_005845.2 LOC728772 -2.3 1.9E-03 XM_001133074.1 CNRIP1 -2.5 5.7E-04 NM_001111101.1 GVIN1 -2.3 9.7E-04 XM_495863.3 PCDH10 -2.5 2.4E-03 NM_020815.1 ZDHHC14 -2.3 3.8E-03 NM_024630.2 SAMSN1 -2.5 9.8E-04 NM_022136.3 SULT1A3 -2.3 3.7E-03 NM_177552.2 RAB33A -2.5 3.5E-02 NM_004794.2 SULT1A4 -2.3 9.3E-04 NM_001017391.1 SNORA24 -2.5 1.9E-02 NR_002963.1 CTTN -2.3 2.3E-02 NM_005231.2 PORCN -2.5 2.3E-03 NM_022825.2 VNN1 -2.3 4.3E-03 NM_004666.1 CRHBP -2.5 2.5E-03 NM_001882.3 PHF16 -2.3 7.9E-04 NM_001077445.1 LOC153561 -2.5 1.7E-03 NM_207331.2 WWC3 -2.3 3.5E-03 NM_015691.2 CXCL13 -2.5 1.6E-02 NM_006419.1 LY9 -2.3 3.9E-03 NM_002348.2 AOAH -2.5 1.0E-03 NM_001637.1 EHD1 -2.3 1.0E-02 NM_006795.2 HLX -2.5 1.5E-03 NM_021958.2 TRIM47 -2.3 5.7E-04 NM_033452.2 LOC100134361 -2.5 2.6E-03 XM_001726827.1 CD163 -2.3 3.5E-02 NM_004244.4 CYGB -2.5 1.7E-02 NM_134268.3 GGA3 -2.3 3.9E-03 NM_138619.1 SCARNA9 -2.5 4.9E-03 NR_002569.1 ABHD12 -2.3 6.5E-04 NM_001042472.1 GAS7 -2.5 1.0E-03 NM_201433.1 GPR132 -2.3 3.0E-03 NM_013345.2 ZC4H2 -2.5 1.3E-03 NM_018684.1 IVNS1ABP -2.3 1.1E-04 NM_006469.4 KLRD1 -2.4 3.2E-02 NM_002262.2 AMDHD2 -2.3 9.3E-04 NM_015944.2 S1PR1 -2.4 1.6E-03 NM_001400.4 CCDC88C -2.3 2.1E-02 NM_001080414.2 SNORA16A -2.4 2.7E-03 NR_003035.1 TRPM2 -2.3 1.8E-02 NM_001001188.3 TRIM16 -2.4 1.0E-02 NM_006470.3 SDC3 -2.3 1.8E-03 NM_014654.2 ZC3H12C -2.4 1.5E-04 NM_033390.1 MAP3K14 -2.3 1.3E-03 NM_003954.2 AP1B1 -2.4 1.4E-03 NM_145730.1 LMOD3 -2.3 4.4E-02 NM_198271.2 LOC644150 -2.4 9.5E-03 XM_933686.1 TNFRSF8 -2.3 6.1E-03 NM_152942.2 KATNAL1 -2.4 1.0E-03 NM_001014380.1 PEG3 -2.3 1.3E-02 NM_006210.1 PTGFRN -2.4 3.8E-02 NM_020440.2 KIAA1328 -2.3 3.4E-04 NM_020776.1 HS.222909 -2.4 1.7E-03 AL117578 WDR74 -2.3 2.1E-02 NM_018093.1 PLIN2 -2.4 7.4E-03 NM_001122.2 SULT1C2 -2.3 3.0E-02 NM_001056.3 ARPM1 -2.4 2.8E-02 NM_032487.3 C17ORF91 -2.3 1.4E-03 NM_001001870.1 MAGED2 -2.4 3.7E-03 NM_177433.1 CPT1A -2.3 8.6E-04 NM_001031847.1 SNHG12 -2.4 4.6E-04 NR_024127.1 KLF6 -2.3 3.9E-04 NM_001008490.1 TSPAN4 -2.4 5.5E-03 NM_001025238.1 BCL2 -2.3 1.2E-02 NM_000633.2 NOL3 -2.4 5.3E-03 NM_003946.3 ELOVL6 -2.3 6.3E-03 NM_024090.1 ABLIM3 -2.4 1.9E-02 NM_014945.2 LOC606724 -2.3 3.1E-03 NR_002454.2 ZNF571 -2.4 1.3E-02 NM_016536.3 LOC100133840 -2.3 2.2E-02 XM_001714384.1 C6ORF85 -2.4 3.8E-03 NM_021945.4 HECTD2 -2.3 1.5E-03 NM_182765.2 PGM2L1 -2.4 8.6E-04 NM_173582.3 BCORL1 -2.3 1.2E-03 NM_021946.2 IL6R -2.4 2.1E-02 NM_000565.2 GDPD1 -2.3 2.5E-02 NM_182569.1 C13ORF31 -2.4 7.5E-04 NM_153218.1 ITSN1 -2.2 3.7E-03 NM_001001132.1 KIAA1147 -2.4 1.4E-03 XM_001130020.1 HS.569162 -2.2 3.3E-02 AV681673 CD180 -2.4 5.0E-03 NM_005582.1 IL6R -2.2 1.0E-02 NM_000565.2 SDCCAG3 -2.4 8.6E-04 NM_001039708.1 CYBASC3 -2.2 2.6E-03 NM_153611.3 PID1 -2.4 2.3E-03 NM_017933.3 AFF1 -2.2 2.9E-04 NM_005935.1 APPL2 -2.4 1.0E-03 NM_018171.3 MYL5 -2.2 4.1E-03 NM_002477.1 UBQLN3 -2.4 5.1E-03 NM_017481.2 GALIG -2.2 3.9E-04 NM_194327.1 SNCA -2.4 1.8E-03 NM_000345.2 MEI1 -2.2 3.0E-02 NM_152513.3 CACHD1 -2.4 1.9E-03 NM_020925.2 KRBA1 -2.2 3.7E-03 NM_032534.2 LRRC3 -2.4 1.0E-03 NM_030891.3 MAGED2 -2.2 2.2E-03 NM_201222.1 RNU1G2 -2.4 3.6E-02 NR_004426.1 EPB41 -2.2 5.1E-03 NM_203343.1 TRIM2 -2.4 3.6E-02 NM_015271.2 MESDC1 -2.2 1.5E-02 NM_022566.1 PPM1H -2.3 4.2E-04 NM_020700.1 ABHD12 -2.2 2.8E-04 NM_001042472.1 ARMCX1 -2.3 6.4E-03 NM_016608.1 FAM7A1 -2.2 5.8E-03 XM_931257.1 P2RY5 -2.3 1.3E-03 NM_005767.4 ACPP -2.2 1.7E-02 NM_001099.2 C17ORF38 -2.3 4.0E-03 NM_001010855.1 EPPB9 -2.2 1.1E-02 NM_015681.2 DLG4 -2.3 4.0E-03 NM_001365.2 LOC644860 -2.2 2.4E-02 XR_039515.1 ARL4A -2.3 1.0E-03 NM_001037164.1 LOC85389 -2.2 3.3E-03 NR_001453.1 Appendix II

Gene FC p (Corr) Accession Gene FC p (Corr) Accession RRAS2 -2.2 7.8E-03 NM_012250.3 SNCA -2.1 9.2E-03 NM_007308.1 HS.311428 -2.2 2.2E-02 AK024261 WWP1 -2.1 4.8E-04 NM_007013.3 LILRB2 -2.2 1.3E-02 NM_001080978.1 SNORD104 -2.1 1.3E-02 NR_004380.1 LILRB5 -2.2 2.3E-03 NM_006840.3 GCNT1 -2.1 3.1E-02 NM_001097635.1 DACT3 -2.2 4.0E-03 NM_145056.1 HEBP2 -2.1 1.0E-02 NM_014320.2 ARRDC3 -2.2 1.5E-03 NM_020801.1 CKB -2.1 9.9E-03 NM_001823.3 SMCR5 -2.2 6.9E-03 NR_024007.1 GRIN3A -2.1 4.1E-03 NM_133445.1 ARMCX1 -2.2 6.1E-03 NM_016608.1 PON2 -2.1 5.6E-03 NM_000305.2 HABP4 -2.2 3.1E-03 NM_014282.1 HINT3 -2.1 1.4E-03 NM_138571.4 CCDC24 -2.2 3.0E-02 NM_152499.1 ANGPTL6 -2.1 3.6E-02 NM_031917.2 RHOB -2.2 1.1E-03 NM_004040.2 TSC22D3 -2.1 5.7E-04 NM_004089.3 ITGA4 -2.2 3.0E-03 NM_000885.4 SRGAP3 -2.1 9.9E-03 NM_001033117.1 C3ORF54 -2.2 6.5E-03 NM_203370.1 HS.527535 -2.1 2.2E-02 AK026716 FLJ33630 -2.2 6.6E-03 NR_015360.1 RB1 -2.1 1.1E-03 NM_000321.2 MTUS1 -2.2 4.8E-02 NM_001001924.1 CRYBB1 -2.1 2.2E-02 NM_001887.3 GAS6 -2.2 3.2E-02 NM_000820.1 C7ORF41 -2.1 2.6E-02 NM_152793.2 ARHGEF12 -2.2 5.1E-04 NM_015313.1 HS.561735 -2.1 2.8E-03 AL528570 GNB5 -2.2 4.6E-04 NM_006578.3 TOM1L2 -2.1 9.5E-03 NM_001082968.1 LOC729196 -2.2 1.0E-03 XM_001129638.1 HS.163752 -2.1 2.9E-02 AA204695 SGPP2 -2.2 5.1E-03 XM_001128702.1 HS.544637 -2.1 2.7E-02 AW276479 GPR175 -2.2 4.5E-04 NM_016372.1 C11ORF63 -2.1 3.1E-02 NM_199124.1 EMB -2.2 1.1E-02 NM_198449.1 SHROOM4 -2.1 4.7E-02 NM_020717.2 SULT1A3 -2.2 4.8E-04 NM_003166.3 ATP11C -2.1 2.4E-02 NM_173694.3 LOC100128309 -2.2 1.9E-03 XM_001717579.1 KLF6 -2.1 2.6E-03 NM_001008490.1 USE1 -2.2 5.6E-03 NM_018467.3 HSF4 -2.1 7.6E-03 NM_001040667.1 LILRB2 -2.2 6.8E-04 NM_001080978.1 SHMT1 -2.1 4.4E-02 NM_004169.3 OXER1 -2.2 7.1E-03 NM_148962.4 PRRG4 -2.1 4.1E-03 NM_024081.4 KIAA1147 -2.2 3.2E-04 NM_001080392.1 HS.575812 -2.1 3.1E-03 DB296040 HECTD2 -2.2 1.3E-02 NM_173497.1 INPP4B -2.1 1.8E-03 NM_003866.1 IL17RD -2.2 3.4E-02 NM_001080973.1 HEYL -2.1 8.8E-03 NM_014571.3 HS.542265 -2.2 3.4E-02 AI863214 TRIM32 -2.1 8.5E-03 NM_012210.3 KIAA1430 -2.2 5.8E-03 NM_020827.1 SDCCAG3 -2.1 1.0E-03 NM_001039708.1 TRIP10 -2.2 1.9E-02 NM_004240.2 TCF23 -2.1 3.8E-02 NM_175769.1 PLAU -2.2 5.5E-03 NM_002658.2 SNORA70 -2.1 7.9E-03 NR_000011.1 C5ORF30 -2.2 5.8E-03 NM_033211.2 SLAMF6 -2.1 1.3E-02 NM_052931.3 C14ORF145 -2.2 2.2E-02 NM_152446.2 LYSMD3 -2.1 3.3E-02 NM_198273.1 DHRS9 -2.1 1.0E-02 NM_005771.3 PCSK6 -2.1 2.6E-04 NM_138320.1 DDIT4 -2.1 1.8E-02 NM_019058.2 LOC647493 -2.1 9.9E-03 XM_942864.1 SNORD52 -2.1 3.3E-02 NR_002742.1 MFI2 -2.1 8.8E-03 NM_033316.2 AP2A2 -2.1 2.1E-03 NM_012305.2 ZSCAN12 -2.1 2.7E-03 NM_001039643.1 PPP2R2B -2.1 4.0E-02 NM_181676.1 SNX16 -2.1 1.7E-02 NM_152837.1 WAC -2.1 9.0E-03 NM_100264.1 WWOX -2.1 3.4E-02 NM_016373.1 TESC -2.1 2.7E-02 NM_017899.2 CXCL12 -2.1 2.0E-02 NM_000609.4 EPB41L5 -2.1 1.2E-03 NM_020909.2 GCDH -2.1 4.6E-03 NM_013976.2 ICOSLG -2.1 3.1E-02 NM_015259.4 LOC643757 -2.1 1.9E-03 XM_931780.1 ANGPTL4 -2.1 2.6E-02 NM_139314.1 ERCC6 -2.1 1.7E-02 NM_000124.1 LOC644474 -2.1 4.0E-02 XM_930098.1 BLOC1S2 -2.1 6.1E-03 NM_001001342.1 CNRIP1 -2.1 4.9E-03 NM_015463.2 SATB2 -2.1 1.4E-03 NM_015265.1 LSR -2.1 1.0E-02 NM_205835.2 LOC731789 -2.0 5.1E-04 XM_001722670.1 HLX -2.1 1.0E-02 NM_021958.2 IRAK2 -2.0 3.7E-02 NM_001570.3 AQP11 -2.1 1.0E-02 NM_173039.1 SLC45A4 -2.0 1.5E-02 XM_933796.2 HS.19339 -2.1 6.3E-04 BC035116 PCTK3 -2.0 3.2E-02 NM_212503.1 ST8SIA4 -2.1 3.5E-03 NM_005668.3 PTGS2 -2.0 3.3E-02 NM_000963.1 GCLM -2.1 3.0E-03 NM_002061.2 LOC100129269 -2.0 3.8E-02 XM_001719843.1 RAB3IP -2.1 3.8E-03 NM_001024647.2 ATP11C -2.0 4.9E-03 NM_173694.3 SAGE1 -2.1 2.7E-03 NM_018666.2 EEPD1 -2.0 1.1E-03 NM_030636.2 LOC727935 -2.1 2.2E-02 XM_001126418.1 SLC45A4 -2.0 1.5E-02 NM_001080431.1 Appendix II

Gene FC p (Corr) Accession CHRNA5 -2.0 4.6E-02 NM_000745.2 SLC12A2 -2.0 5.1E-03 NM_001046.2 LOC647008 -2.0 6.5E-03 XM_929992.1 DTNA -2.0 7.4E-03 NM_032981.2 LOC644739 -2.0 2.0E-02 XM_933679.1 ATP8B4 -2.0 3.9E-03 NM_024837.2 TJP1 -2.0 3.0E-02 NM_003257.3 MYL5 -2.0 3.0E-03 NM_002477.1 CORO1A -2.0 6.5E-03 NM_007074.2 FXYD1 -2.0 2.5E-02 NM_005031.3 LRP8 -2.0 2.2E-02 NM_001018054.1 IGF2BP3 -2.0 1.8E-02 NM_006547.2 SQLE -2.0 8.9E-03 NM_003129.3 RAP2B -2.0 4.8E-02 NM_002886.2 PCDHGC4 -2.0 1.1E-02 NM_032406.1 TNFSF8 -2.0 3.6E-04 NM_001244.2 LOC100133923 -2.0 6.9E-04 XM_001714921.1 ASMTL -2.0 3.5E-02 NM_004192.2 RIMKLB -2.0 1.3E-02 NM_020734.2 DPRXP4 -2.0 9.9E-03 NR_002221.1 HS.552025 -2.0 1.4E-02 BM716742 IVNS1ABP -2.0 3.1E-04 NM_006469.4 Appendix III

All genes regulated ≥ 2 fold change by 30 nM C5a on M-MΦ

Appendix III

Gene 0.5 h 6 h 18 h Accession Gene 0.5 h 6 h 18 h Accession EGR3 7.5 -1 1.1 NM_004430.2 LMNA 1.4 1.3 2.6 NM_005572.3 IL1B 6 -1.3 1.4 NM_000576.2 LMNA 1.4 1 2.5 NM_005572.3 FOSB 5.9 -1.3 -1.3 NM_006732.1 CCL7 1.3 -1 6.1 NM_006273.2 PHLDA1 5.7 -1.4 2.2 NM_007350.3 TUBB6 1.3 1.1 2.3 XM_940079.1 ATF3 5.3 -1.5 -1.3 NM_001040619.1 TPM4 1.3 1 2.5 NM_003290.1 EGR1 4.3 -1.5 1 NM_001964.2 MATK 1.2 1.1 2.2 NM_139355.2 EGR2 4.2 -1.9 -1.1 NM_000399.2 IL7R 1.2 1.2 2.2 XM_937367.1 PTGS2 3.3 1 1.2 NM_000963.1 FABP5L2 1.2 -1.3 2.1 XM_001721172.1 AXUD1 3.1 -1.1 1.1 NM_033027.2 YPEL3 1.2 -1 -2.6 NM_031477.4 CCL3L1 3 -1.4 1 NM_021006.4 TIMP3 1.2 -1.5 2.2 NM_000362.4 DUSP5 3 -1.1 1.4 NM_004419.3 IL1RN 1.1 -1.8 2.1 NM_173843.1 CCL3L1 3 -1.2 1.1 NM_021006.4 MGC4677 1.1 1.1 2.1 NM_052871.3 CCL3 2.9 -1.4 1 NM_002983.1 FAM129B 1.1 -1 3.7 NM_022833.2 BTG2 2.8 -1.5 -1.2 NM_006763.2 TPM4 1.1 1.1 2.2 NM_003290.1 CCL3L1 2.6 -1.1 -1.1 NM_021006.4 MATK 1.1 -1.1 2.1 NM_139354.2 TNF 2.6 -1 1.4 NM_000594.2 FABP5 1.1 -1.2 2.2 NM_001444.1 KLF6 2.5 -1.1 -1.1 NM_001008490.1 ALOX15B 1.1 -1.8 -2.1 NM_001141.2 CCL8 2.5 -1.1 1.3 NM_005623.2 ARID3A 1.1 -1.1 -2.1 NM_005224.2 MYADM 2.5 -1.2 1.7 NM_138373.3 ARMET 1.1 -1 2 NM_006010.2 HBEGF 2.5 -1.4 -1.1 NM_001945.1 IL1RN 1.1 -1.6 2 NM_173842.1 CCL4L2 2.5 -1.4 1.2 NM_207007.2 CD276 1.1 -1.2 2.5 NM_001024736.1 KLF6 2.4 -1.1 1 NM_001008490.1 CBX7 1.1 -1 -2.3 NM_175709.2 KLF4 2.4 -1.2 -1.5 NM_004235.3 HAMP 1.1 -1.6 -2 NM_021175.2 ZFP36 2.4 -1.3 -1.2 NM_003407.2 SLC7A5 1.1 -1.1 2.3 NM_003486.5 DUSP6 2.4 1.1 1.9 NM_001946.2 LOC387934 1.1 -1.3 2.1 XM_937508.2 CCL3L3 2.4 -1.3 -1 NM_001001437.3 FAM129B 1.1 1 3.2 NM_001035534.1 CD83 2.4 -1.4 1 NM_001040280.1 TRPM2 1.1 -1.1 -2.5 NM_001001188.3 NR4A2 2.3 1 -1.2 NM_006186.2 LOC642956 1.1 -1.4 2.1 XM_938166.3 PTGER2 2.3 -1 1 NM_000956.2 ZNF581 1 -1.2 -2.1 NM_016535.3 MCL1 2.3 -1.2 1 NM_021960.3 TMEM158 1 -1.3 2.4 NM_015444.2 KLF4 2.3 -1.3 -1.4 NM_004235.3 MARCKSL1 1 1.1 2.4 NM_023009.4 LOC728835 2.2 -1.3 1.2 XM_001133190.1 TUBB2A 1 -1 2 NM_001069.2 KLF2 2.2 -1.4 -1.7 NM_016270.2 DDX21 1 -1.2 2 NM_004728.2 CCL2 2.2 1.2 2.4 NM_002982.3 HS3ST1 1 1.1 2.4 NM_005114.2 KLF6 2.2 -1.1 -1.1 NM_001300.4 C5orf13 1 -1.1 -2.7 NM_004772.1 MCL1 2.2 -1 1.1 NM_021960.3 NDRG1 1 -1.2 -2 NM_006096.2 CD83 2.1 -1.5 -1.2 NM_004233.3 BOP1 1 1.1 2.2 NM_015201.3 IER2 2.1 -1.3 -1.1 NM_004907.2 ALOX15B 1 -1.7 -2.6 NM_001039131.1 DUSP6 2.1 1 1.5 NM_022652.2 PLAU 1 -1.2 -2.4 NM_002658.2 C13orf15 2.1 -1.4 -1.1 NM_014059.2 GFRA2 1 -1.1 4 NM_001495.4 SRF 2.1 -1.1 1.2 NM_003131.2 ZNF467 1 -1.2 -2.1 NM_207336.1 PPP1R15A 2 -1.2 -1.1 NM_014330.2 CTSL1 1 1.1 2.2 NM_001912.3 FILIP1L 2 -1.1 1.3 NM_014890.2 CTPS 1 1.1 2 NM_001905.1 GBP1 1.6 -1.1 2 NM_002053.1 UHRF1 -1 2.5 1.2 NM_001048201.1 MMP19 1.5 1.5 2.2 NM_001032360.1 GOLGA7B -1 -1.2 2.2 NM_001010917.1 Appendix III

Gene 0.5 h 6 h 18 h Accession Gene 0.5 h 6 h 18 h Accession SLC29A1 -1 -1.3 -2.2 NM_001078174.1 TNFRSF21 -1.2 -1 2.6 NM_014452.3 LOC100133893 -1 -1.3 -2.3 XM_001724850.1 HTRA4 -1.2 -1.7 -2.6 NM_153692.2 LOC732415 -1 1.3 2.3 XM_001716577.1 SLC40A1 -1.2 -1.3 -2.1 NM_014585.4 PFKP -1 1.2 2 NM_002627.3 BCL11A -1.2 -1.5 -2.2 NM_022893.2 SLC40A1 -1 -1.2 -2.2 NM_014585.3 FLNB -1.2 -1.4 2.6 NM_001457.1 DBP -1 1.3 -2.4 NM_001352.2 PPAP2B -1.2 -1.4 -2.7 NM_177414.1 SRGAP3 -1 -1.3 -2.4 NM_001033117.1 FAIM3 -1.2 -2 -1.6 NM_005449.3 NRIP3 -1.1 -1.1 2.3 NM_020645.1 CXCR4 -1.2 -1.1 -2.1 NM_003467.2 PRKCA -1.1 1 2.2 NM_002737.2 BMF -1.3 -1.5 -3.1 NM_033503.3 AMICA1 -1.1 -1.6 -2.1 NM_153206.1 ABCC5 -1.3 -1.2 -2 NM_001023587.1 SLC7A1 -1.1 -1.1 2.5 NM_003045.3 CXCR4 -1.4 -1.5 -2.4 NM_001008540.1 TYMS -1.1 2.2 -1.1 NM_001071.1 MIR1914 -2.4 -1.8 -2.2 NR_031735.1 HPSE -1.1 -1.4 -4.7 NM_006665.2 PKD2L1 -1.1 -2.1 -2.1 NM_016112.2 LYPLAL1 -1.1 -1.1 -2.1 NM_138794.1 C6orf192 -1.1 -1.3 -2.1 NM_052831.2 BMF -1.1 -1.3 -2.4 NM_001003943.1 ANG -1.1 -1.2 -2.1 NM_001097577.1 ALPK1 -1.1 -1.1 -2 NM_025144.2 TLR7 -1.1 -1.2 -2.9 NM_016562.3 RPP40 -1.1 1 2.2 NM_006638.2 P8 -1.1 -1.3 -2.4 NM_012385.1 P2RY8 -1.1 -1.3 -2.9 NM_178129.3 PTRF -1.1 -1.2 2.4 NM_012232.3 SLC45A4 -1.1 -1.2 -2.5 XM_933796.2 TOR3A -1.1 -1.3 -2 NM_022371.3 C15orf52 -1.1 -1.4 -2.2 NM_207380.1 HVCN1 -1.1 -1.3 -2.1 NM_032369.2 IQCK -1.1 -1.3 -2.1 NM_153208.1 LOC100133866 -1.1 -1.3 -2.1 XM_001719715.1 SLC29A1 -1.1 -1.4 -2.4 NM_001078174.1 PPAP2B -1.1 -1.3 -2.5 NM_003713.3 KLHDC8B -1.1 -1.2 -3.2 NM_173546.1 TLR8 -1.1 -1 -2.2 NM_138636.2 LOC100129550 -1.1 -1.2 -2.4 NR_024618.1 TOM1 -1.1 -1.2 -2 NM_005488.1 LOC644496 -1.1 -1 2.1 XR_039005.1 DPEP2 -1.2 -1.3 -2 NM_022355.1 NUPR1 -1.2 -1.2 -2.3 NM_001042483.1 EPOR -1.2 1 -2 NM_000121.2 P2RY13 -1.2 -1.4 -2.1 NM_023914.2 PELI2 -1.2 -1.1 -2.2 NM_021255.2 HVCN1 -1.2 -1.1 -2.4 NM_001040107.1 LEP -1.2 -1.5 -2.2 NM_000230.1 NRP1 -1.2 1.1 2.3 NM_003873.4 Appendix IV

Vernon Seow, Junxian Lim, Abishek Iyer, Jacky Y. Suen, Juliana K. Ariffin, Daniel M. Hohenhaus, Matthew J. Sweet and David P. Fairlie. Inflammatory responses induced by lipopolysaccharide are amplified in primary human monocytes but suppressed in macrophages by complement protein C5a. J Immunol. (2013) Oct;191(8):4308-16.

The Journal of Immunology

Inﬂammatory Responses Induced by Lipopolysaccharide Are Ampliﬁed in Primary Human Monocytes but Suppressed in Macrophages by Complement Protein C5a

Vernon Seow, Junxian Lim, Abishek Iyer, Jacky Y. Suen, Juliana K. Arifﬁn, Daniel M. Hohenhaus, Matthew J. Sweet, and David P. Fairlie

Monocytes and macrophages are important innate immune cells equipped with danger-sensing receptors, including complement and Toll-like receptors. Complement protein C5a, acting via C5aR, is shown in this study to differentially modulate LPS-induced inflammatory responses in primary human monocytes versus macrophages. Whereas C5a enhanced secretion of LPS-induced IL-6 and TNF from primary human monocytes, C5a inhibited these responses while increasing IL-10 secretion in donor-matched human monocyte-derived macrophages differentiated by GM-CSF or M-CSF. Gai/c-Raf/MEK/ERK signaling induced by C5a was amplified in macrophages but not in monocytes by LPS. Accordingly, the Gai inhibitor pertussis toxin and MEK inhibitor U0126 blocked C5a inhibition of LPS-induced IL-6 and TNF production from macrophages. This synergy was independent of IL-10, PI3K, p38, JNK, and the differentiating agent. Furthermore, C5a did not inhibit IL-6 production from macrophages induced by other TLR agonists that are selective for Toll/IL-1R domain–containing adapter inducing IFN-b (polyinosinic-polycytidylic acid) or MyD88 (imiquimod), demonstrating selectivity for C5a regulation of LPS responses. Finally, suppression of proinflammatory cytokines IL-6 and TNF in macrophages did not compromise antimicrobial activity; instead, C5a enhanced clearance of the Gram-negative bacterial pathogen Salmonella enterica serovar Typhimurium from macrophages. C5aR is thus a regulatory switch that modulates TLR4 signaling via the Gai/c-Raf/MEK/ERK signaling axis in human macrophages but not monocytes. The differential effects of C5a are consistent with amplifying monocyte proinflammatory responses to systemic danger signals, but attenuating macrophage cytokine responses (without compromising microbicidal activity), thereby restraining inflammatory responses to localized infections. The Journal of Immunology, 2013, 191: 4308–4316.

omplement is an important cascading network of .30 reperfusion injury, and sepsis (1). C5a binds to and signals through soluble plasma and insoluble membrane-bound proteins the receptor C5aR (CD88), a class A G protein–coupled rhodopsin- C that cooperate in innate immunity in the recognition, like receptor (2), and to a second receptor C5L2, which does not opsonization, destruction, and removal of pathogens and infected couple G proteins and may be nonsignaling, but may also mod- or damaged cells. Complement activation by classical, alternative, ulate C5aR (3, 4). C5aR couples to the pertussis toxin (PTX)–sen- and lectin pathways and other routes is tightly controlled, pro- sitive Gai and in some cells such as monocytes to the PTX-insensitive ducing the lytic membrane attack complex, as well as opsonins Ga16, and it recruits b-arrestins-1,2 to the plasma membrane (1, 5). such as C3b and proinflammatory anaphylatoxins C3a, C4a, and C5a activates several downstream signaling pathways, including C5a. Elevated plasma levels of C5a or its receptor are associated Akt, MEK, PI3K, phosphokinase A, phospholipase C (PLC), and with inflammatory diseases such as rheumatoid arthritis, inflam- NF-kB(1). matory bowel diseases, respiratory distress syndrome, ischemia– The TLRs are an important family of type I transmembrane proteins that respond to infectious organisms and danger signals by initiating inflammatory responses, activating microbial killing and Institute for Molecular Bioscience, University of Queensland, Brisbane, Queensland clearance mechanisms, and priming the adaptive immune response 4072, Australia (6). Of 10 known human TLRs, TLR4 responds to LPS from Received for publication May 23, 2013. Accepted for publication August 10, 2013. Gram-negative bacterial cell walls. TLR4 can signal through two This work was supported by Australian National Health and Medical Research different pathways dependent on, and modulated by, separate Council Grants 1030169 and 1028423, Senior Principal Research Fellowship 1027369 (to D.P.F.), and honorary Senior Research Fellowship 1003470 (to M.J.S.), adaptor proteins MyD88 or Toll/IL-1R domain–containing adapter as well as by Australian Research Council Grants DP130100629 and DP1093245, inducing IFN-b (TRIF) (7). MyD88 activates IL-1R–associated Future Fellowship FT100100657 (to M.J.S.), and Federation Fellowship FF0668733 kinases (IRAKs) and TNFR-associated factor 6 (TRAF6), as well (to D.P.F.). as transcription factors NF-kB, AP-1, and IFN regulatory factor 5 Address correspondence and reprint requests to Prof. David P. Fairlie, Institute for Molecular Bioscience, University of Queensland, Brisbane, QLD 4072, Australia. (IRF5) further downstream. TRIF signals the induction of type I E-mail address: [email protected] IFNs by recruiting TRAF3 and receptor interacting protein 1 (RIP1) The online version of this article contains supplemental material. to activate transcription factor IFN regulatory factor 3 (IRF3), NF-kB, Abbreviations used in this article: GM-Mf, macrophages generated from monocytes and AP-1. Both MyD88 and TRIF activate AP-1 via downstream cultured in GM-CSF; HMDM, human monocyte–derived macrophage; IMQ, imiqui- MAPK activation (8), although NF-kB and MAPK activation by mod; M-Mf, macrophages generated from monocytes cultured in M-CSF; MOI, multiplicity of infection; PLC, phospholipase C; poly(I:C), polyinosinic-polycytidylic TRIF occurs later than for MyD88 (9). acid; PTX, pertussis toxin; S. Typhimurium, Salmonella enterica serovar Typhimu- In the present study, crosstalk between C5aR and TLR4 was rium; TRAF, TNFR-associated factor; TRIF, Toll/IL-1R domain–containing adapter investigated in human monocytes and human monocyte–derived inducing IFN-b. macrophages (HMDMs) using the respective ligands C5a and Copyright Ó 2013 by The American Association of Immunologists, Inc. 0022-1767/13/$16.00 LPS. Crosstalk between different signaling pathways in mam- www.jimmunol.org/cgi/doi/10.4049/jimmunol.1301355 The Journal of Immunology 4309 malian cells can result in unexpected and unique functional out- whereas an additional activation signal such as ATP is required comes (10). Some pathogens have evolved to evade or exploit host in macrophages (24). The present study compares effects of C5a microbial-killing by targeting C5aR and TLRs specifically. For on LPS responses in primary human monocytes versus donor- example, virulence proteins secreted by Staphylococcus aureus matched macrophages derived through differentiating monocytes bind to C5aR with potent antagonist activity, preventing recruit- with either GM-CSF or M-CSF. Surprisingly, we found that C5a ment of phagocytes to sites of infection (11). Porphyromonas differentially modulated LPS responses depending upon the cell gingivalis exploits the C5aR/TLR2 crosstalk by increasing cAMP type, that C5a activated the Gai/c-Raf/MEK/ERK axis to selec- production in macrophages, which suppresses macrophage im- tively downregulate proinflammatory cytokine production from mune function and enhances pathogen survival (12). Clearly, these macrophages but not monocytes, and that effects on macrophages pathogens have evolutionarily adapted to exploit complement/ do not compromise but instead stimulate pathogen killing. Thus, it TLR crosstalk to their advantage. Although C5aR and TLR4 is proposed that C5a has distinctly different regulatory effects on signaling have been separately investigated in many studies, there myeloid cell functions during localized versus systemic inflam- is comparatively little reported about their interplay (13). Among matory responses. recent reports on C5aR/TLR4 crosstalk, C5a has been found to downregulate TLR4-induced IL-12 production via PI3K-dependent Materials and Methods (14) and PI3K-independent (15) pathways in mouse macrophages; Reagents to signal via the PI3K-Akt pathway to enhance LPS-induced IL-17F cytokine production (16); and to suppress LPS-induced IL-17A and Recombinant human GM-CSF, M-CSF, and IL-10 were purchased from IL-23 cytokine production in mouse macrophages (17). C5a-induced PeproTech. LPS from Salmonella enterica (serotype Minnesota RE 595), U0126 monothanolate, wortmannin, and PTX were purchased from suppression of the LPS-inducible IL-17A/IL-23 axis was reported Sigma-Aldrich. Imiquimod R837 (IMQ) and polyinosinic-polycytidylic to occur via enhanced IL-10 production and ERK1/2 phosphoryla- acid (poly(I:C)) were purchased from InvivoGen. Human C5a was pur- tion. However, most studies reporting C5a modulation of TLR4 chased from Sino Biological. Goat neutralizing Ab against human IL-10 signaling had been performed with murine rather than human cells (AB-217-NA) was purchased from R&D Systems. C5aR antagonist (3D53 also licensed as PMX53) (1, 25) was synthesized and characterized (an- (18, 19). Some interpretations of human disease mechanisms have alytical HPLC, mass spectroscopy, and nuclear magnetic resonance spec- been made based on extrapolating from observations of mouse troscopy) in-house as described (26). Abs against phosphorylated c-Raf studies. However, there are important differences in LPS-induced and total c-Raf were purchased from Cell Signaling Technology. All cell TLR4 signaling between humans and mice (20–22), necessitating culture reagents were purchased from Invitrogen, and all analytical-grade careful assessment of C5aR/TLR4 crosstalk in human cells. chemical reagents were obtained from Sigma-Aldrich, unless otherwise stated. Monocytes and macrophages are important cellular mediators of innate immunity and are targets for C5a and LPS. Some features Isolation of human monocytes and macrophages of LPS signaling are also distinctly different between human monocytes and macrophages (23). Monocytes act as danger sensors in the PBMCs were isolated from buffy coats of anonymous donors (Australian Red Cross Blood Service, Kelvin Grove, QLD, Australia) by density circulation and typically generate a more rapid, heightened inflam- centrifugation using Ficoll-Paque Plus (GE Healthcare) following the matory response than do macrophages to danger signals. For exam- manufacturer’s instructions. Contaminating erythrocytes were removed by ple, LPS alone can trigger inflammasome activation in monocytes, repeated ice-cold sterile water. CD14+ MACS MicroBeads (Miltenyi

FIGURE 1. C5a acting via C5aR dose-dependently modulates LPS-induced IL-6, TNF, and IL-10 cytokine production in human monocytes and macrophages (GM-Mf and M-Mf) via C5aR. (A–C) Human CD14+ monocytes (gray) and HMDMs (GM-Mf, black; M-Mf, white) were serum-deprived for at least 8 h before treatment with LPS (5 ng/ml) in the absence or presence of various concentration of C5a (0.03–300 nM) for 24 h. (D–F) For antagonism of C5aR, cells were pretreated with 3D53 (1 mM, 30 min) prior to C5a and LPS stimulation. ELISA was used to detect levels of secreted (A, D)IL-6,(B, E) TNF, and (C, F)IL-10 in culture media. Error bars are means 6 SEM of three independent experiments (n = 3). *p , 0.05, **p , 0.01, ***p , 0.001 by Student t test. 4310 C5aR AND TLR4 CROSSTALK IN MONOCYTES VERSUS MACROPHAGES

Biotec) were used to positively select monocytes from isolated PBMCs Results according to the manufacturer’s instructions. CD14+ monocytes were 3 6 C5a acting via C5aR differentially modulates LPS signaling in seeded at 1 10 cells/ml in IMDM supplemented with 10% FBS, 100 + U/ml penicillin, 100 mg/ml streptomycin, and 2 mM GlutaMAX at 37˚C human CD14 monocytes versus macrophages + in presence of 5% CO2. To generate HMDMs, CD14 monocytes were In light of recent discoveries linking crosstalk between TLRs and cultured in the presence of either 10 ng/ml GM-CSF (GM-Mf) or M-CSF (M-Mf) for at least 6 d. the complement system as possible control points for regulation of host defense (13), this study compared C5aR/TLR4 crosstalk Quantitative cytokine ELISA between human monocytes and macrophages, given their different Cells were seeded at 1 3 106 cells/ml and serum-deprived overnight in the roles in responding to inflammatory stimuli. C5a decreased incubator prior to treatment. LPS and C5a treatments were premixed prior LPS-induced secreted levels of IL-6 and TNF in both GM-Mf to 24 h stimulation. Cell culture supernatants were collected and cytokine and M-Mf, but it increased the levels of the same proinflammatory levels were determined using specific ELISA sets from BD Pharmingen, cytokines in monocytes (Fig. 1A, 1B). Alternatively, C5a enhanced according to the manufacturer’s instructions. LPS-induced IL-10 production from macrophages, but it had no Isolation of mRNA and RT-PCR effect on this response in monocytes (Fig. 1C). This suggests that the regulatory effects of C5aR on TLR4 depend on cellular context, Total RNA was extracted from cells using an RNeasy Mini Plus kit (Qiagen), according to the manufacturer’s instructions. Total cellular RNA for example differentiation status. Although C5aR mediates C5a- (2–10 mg) was reverse transcribed to cDNA using oligo(dT) primer (1 mg) dependent responses, a second C5a receptor, C5L2, has also been and SuperScript III (Invitrogen), according to the manufacturer’s instruc- reported to exhibit anti-inflammatory functions (3, 28). However, tions. cDNA samples were stored at 220˚C until further use. Gene expression analysis by real-time PCR Gene-specific primers were designed using the free Web-based software Primer-BLAST (National Center for Biotechnology Information). Quan- titative real time-PCR was performed using an ABI Prism 7900 real-time PCR system (Applied Biosystems), cDNA (50 ng), SYBR Green PCR Master Mix (Applied Biosystems), and gene-specific primers. Relative gene expression was normalized against 18S rRNA expression and then converted to fold change against control samples. All samples were analyzed in duplicate. Primer sequences used are: CD68,59-TTTGGGTGAGGCGGTTCAG-39 and 39-CCAGTGCTCTCTGCCAGTA-59; 18S,59-ACCACGGGTGACGGGGA- ATC-39 and 39-CCGGGTCGGGAGTGGGTAAT-59. Quantification of MEK1, ERK1/2, p38, and JNK phosphorylation Cells were seeded at 4000 cells/well in white ProxiPlate-384 (PerkinElmer) overnight. On the following day, cells were serum-deprived for 2 h prior to stimulation with C5a (30 nM), LPS (5 ng/ml), or both over a stated time course. Alphascreen SureFire assays (PerkinElmer) were performed according to the manufacturer’s instructions. Western blot analysis Equal amounts of total cell lysates (10 mg) were subjected to SDS-PAGE, transferred to polyvinylidene difluoride membranes, and proteins were detected using appropriate Abs according to the manufacturer’s instructions.

Bacterial clearance assay Intracellular survival of Salmonella enterica serovar Typhimurium (S. Typhimurium) strain SL1344 within HMDMs was monitored in gentamicin exclusion assays (27). M-Mf were seeded at 200,000 cells/well in penicillin/streptomycin-free media containing M-CSF (10 ng/ml). On the following day, cells were infected with S. Typhimurium at a multiplicity of infection (MOI) of 100. One hour after infection, cells were washed with gentamicin (200 mg/ml) in penicillin/streptomycin-free media and further incubated (1 h) to kill extracellular bacteria in the same media. Additional incubations beyond this 2 h time point were carried out in media containing 20 mg/ml gentamicin. At specific time points, cells were washed twice and lysed with 0.01% Triton X-100 in PBS, after which lysates were cultured overnight on Luria–Bertani agar. Colony counts were performed FIGURE 2. Temporal study of C5aR/TLR4 crosstalk in monocytes and to assess intracellular bacterial loads. The MOI was also confirmed by macrophages. Human CD14+ monocytes were differentiated to HMDMs plating out inoculum on Luria–Bertani agar plates and performing colony with either GM-CSF (10 ng/ml) (GM-Mf, black) or M-CSF (10 ng/ml) counts. Treatments with C5a (30 nM) were conducted by cotreatment with (M-Mf, white) in 10% FBS/IMDM during a 6-d time course. Macrophage bacteria upon infection. For antagonism assays, cells were pretreated with A the C5a inhibitor 3D53 (1 mM) 30 min prior to infection. 3D53 was also marker ( ) CD68 gene expression was monitored during differentiation in added back following the washing with gentamicin-containing media. the presence of GM-CSF (black) or in the presence of M-CSF (white) on days 0, 2, 4, and 6. Genes were detected by quantitative RT-PCR, normalized Statistical analysis against 18S rRNA expression, and converted to fold change relative to day 0. These cells were also treated with LPS (5 ng/ml) with and without C5a (30 Data were plotted and analyzed using GraphPad Prism version 5.0c for Mac OS X (GraphPad Software). Statistically significant differences were nM) for 24 h in 10% FBS/IMDM on days 0, 2, 4, and 6 to monitor sup- assessed using a Student t test for paired comparison. All values of inde- pression by C5a of LPS-induced IL-6 and TNF in HMDMs. ELISA was used pendent parameters are shown as means 6 SEM of at least three independent to detect (B)IL-6,(C)TNF,and(D) IL-10 secreted in culture media. Error experiments, unless otherwise stated. A p value , 0.05 was considered bars are means 6 SEM of three independent experiments (n =3).*p , 0.05, statistically significant. **p , 0.01, ***p , 0.001 by Student t test. The Journal of Immunology 4311 the biological roles of C5L2 in inflammation are highly contro- targeted downstream signaling of C5aR signaling and found that versial. Our group has previously developed 3D53, a potent, in- the MEK inhibitor U0126, a selective inhibitor of the ERK pathway surmountable, and selective antagonist of C5aR that does not bind (33), prevented immunomodulation by C5a of LPS-induced IL-6 C5L2 (1, 29, 30). 3D53 blocked the immunomodulatory effects of and TNF production in macrophages, but not monocytes (Fig. 3A, C5a on TLR4 signaling in both macrophages and monocytes (Fig. 3B). In monocytes, the enhancement of LPS-induced IL-6 pro- 1D–F), thus demonstrating that the observed responses of C5a duction by C5a was altered by the presence of MAPK p38 and were mediated via C5aR rather than C5L2. PLCb inhibitors SB203580 and U73122, respectively, whereas the enhancement of LPS-induced TNF production was affected by Temporal study of C5aR/TLR4 crosstalk in HMDMs phosphokinase A inhibitor PKI 14-22 (Supplemental Fig. 2A, 2D). To further characterize the phenotypic switch in C5a responsive- Collectively, these data suggest that C5a-dependent ERK1/2 phos- ness during monocyte to macrophage differentiation, time course phorylation selectively suppresses LPS-induced IL-6 and TNF experiments were performed. As expected, mRNA levels of the production in macrophages. Interestingly, although the literature macrophage marker CD68 were increased after 4 d of differenti- suggests that ERK1/2 is involved in TLR-induced IL-10 production ation in the presence of either GM-CSF or M-CSF (Fig. 2A). The from human macrophages (15, 34), we found in the present study suppressive effect of C5a on LPS-induced cytokine production that C5a enhanced LPS-mediated IL-10 production in an ERK1/2- was apparent by day 4 for TNF, and by day 6 for IL-6 (Fig. 2B, independent, but Gai-dependent, manner (Fig. 3C). Similarly, 2C). Increased IL-10 production was also observed by day 4 (Fig. LPS-induced IL-10 production from monocytes, as well as from 2D). The suppressive effects of C5a on LPS-induced IL-6 and GM-Mf and M-Mf, was ERK-independent (Fig. 3C). TNF production in macrophages correlated with increased C5AR C5a and LPS synergize for Raf/MEK/ERK phosphorylation in and TLR4 expression during monocyte to macrophage differentia- macrophages tion (Supplemental Fig. 1A, 1B). Levels of C5AR and TLR4 mRNA were not affected by C5a treatment, but LPS decreased C5AR ex- The data for PTX and U0126 inhibitors showed that C5a-modulated pression by ∼30% (Supplemental Fig. 1C, 1D), consistent with a suppression of LPS-mediated inflammatory responses was de- previous study (31). pendent on Gai coupling and MAPK ERK1/2 phosphorylation. Therefore, we further investigated additional checkpoints in the C5a suppresses LPS-induced IL-6 and TNF via Gai- and signaling pathway between Gai signaling and upstream of ERK MEK-dependent signaling phosphorylation (Gai/Raf/MEK/ERK) signaling cascade to fur- C5a-dependent C5aR signaling involves coupling to Gai proteins ther map the crosstalk signaling mechanisms between C5aR and (5, 32). We used PTX, which specifically interferes with Gai TLR4. Interestingly, C5a phosphorylated c-Raf only in macro- protein signaling, to investigate whether the blockade of Gai phages (both GM-Mf and M-Mf) but not in monocytes (Fig. 4A). signaling would reverse inhibition by C5a of TLR4 responses. As Co-treatment of GM-Mf or M-Mf with LPS and C5a synergis- expected, all modulatory effects of C5a on LPS-induced cytokine tically promoted c-Raf phosphorylation. This synergy was not production from monocytes and macrophages was sensitive to observed in monocytes. LPS triggered MEK1 phosphorylation in PTX (Fig. 3). These results further confirmed involvement of macrophages (Fig. 4B). Although C5a and LPS each induced C5aR and not C5L2 signaling in this crosstalk, as C5aR, but not MEK1 phosphorylation in monocytes and macrophages, the syn- C5L2, employs Gai signaling (1, 3). We also pharmacologically ergistic effect of LPS and C5a cotreatment was only observed in

FIGURE 3. C5a suppression of LPS-induced IL-6 and TNF in macrophages is sensitive to PTX and

MEKi (U0126) treatment. To inhibit Gai-dependent interactions, cells were pretreated with PTX (200 ng/ml, overnight) or, for MEK1/2 inhibition, U0126 (1 mM, 30 min) prior to LPS (5 ng/ml) with and without C5a (30 nM) stimulation of monocytes (gray) and HMDMs (GM-Mf, black; M-Mf, white). ELISA was used to detect (A) IL-6, (B) TNF, and (C) IL-10 secreted cytokines relative to the LPS response (100%) at 24 h after stimulation. Error bars are means 6 SEM of three independent experiments (n =3).*p , 0.05, **p , 0.01, ***p , 0.001 by Student t test. 4312 C5aR AND TLR4 CROSSTALK IN MONOCYTES VERSUS MACROPHAGES

C5a-mediated amplification of LPS-induced IL-6 production (Supplemental Fig. 2A), but not TNF production (Supplemental Fig. 2B), from these cells was affected by the p38 inhibitor SB203580. Furthermore, SB203580 did not block the suppression by C5a of LPS-induced IL-6 and TNF production in macrophages (Supplemental Fig. 2B, 2C, 2E, 2F). C5a-mediated suppression of LPS-induced IL-6 and TNF is IL-10 and PI3K independent C5a has been previously reported to enhance LPS-induced IL-10 cytokine production from macrophages (15), and it has also been reported to suppress IL-17A and IL-23 by enhancing IL-10 production (17). Consistent with such a mechanism, C5a upregulated LPS-induced IL-10 production from macrophages, but it did not alter this response in monocytes (Fig. 1C). However, the ability of the MEK inhibitor to block C5a-mediated suppression of IL-6 and TNF production, without affecting enhancement of IL-10 production in macrophages, implied that the inhibitory effects on IL-6 and TNF were IL-10–independent (Fig. 3C). We therefore examined this hypothesis. LPS-induced IL-6 and TNF production was suppressed with increasing concentrations of IL-10 (Supple- mental Fig. 3A, 3B). This is consistent with literature indicating that IL-10 attenuates IL-6 and TNF production from activated macrophages (34, 35). A neutralizing Ab against human IL-10 (10 mg/ml) completely blocked the effect of exogenously added IL-10 (2 ng/ml) in macrophages (Supplemental Fig. 3C, 3D), and its addition prior to LPS stimulation resulted in increased TNF production in GM-Mf and M-Mf (Fig. 6B). However, C5a still suppressed IL-6 and TNF production in the presence of the neutralizing Ab against human IL-10 (Fig. 6A, 6B), thus demon- FIGURE 4. C5a induced Raf/MEK/ERK phosphorylation is amplified in strating that the immunomodulation by C5a was independent of macrophages but not in monocytes by LPS. To quantify c-Raf, MEK1, and IL-10. Therefore, C5a-enhanced LPS-induced IL-10 production ERK1/2 phosphorylation, monocytes (gray) GM-Mf (black), and M-Mf does not account for the inhibition of inducible IL-6 and TNF (white) were treated with LPS (5 ng/ml) with and without C5a (30 nM) for production by C5a. Apart from IL-10, we also found that the 5 min, after which cells were lysed. (A) c-Raf phosphorylation was detected immunomodulation by C5a was also independent of LPS-induced by Western blot and probed with Abs against phosphorylated c-Raf and total PI3K signaling in GM-Mf (Fig. 6C, 6D) and in M-Mf (Fig. 6E, c-Raf. Bands were visualized by an ECL method and quantified by densi- 6F). The presence of the pan-PI3K inhibitor wortmannin at the tometry measurements of phosphorylated c-Raf normalized with total c-Raf highest concentration (100 nM) did not abolish C5a-mediated B C against control. Levels of phosphorylated ( )MEK1and( ) ERK1/2 were suppression of LPS-inducible IL-6 and TNF release, although measured by AlphaScreen SureFire assays. Error bars are means 6 SEM of wortmannin on its own appears to enhance LPS-induced IL-6 and three independent experiments (n =3).*p , 0.05, **p , 0.01, ***p , 0.001 by Student t test. TNF production from macrophages at 100 nM concentration. C5a attenuation of LPS-induced IL-6 is selectively TLR4 macrophages. Consistent with this, cotreatment of GM-Mf or mediated in macrophages M-Mf with LPS and C5a synergistically promoted ERK1/2 phosphorylation, whereas this synergy was not observed for TLR4 signaling involves two major pathways, the Mal/MyD88- monocytes (Fig. 4C). and the TRAM/TRIF-dependent pathways (7). To determine whether C5a affected responses through multiple TLRs in primary C5a and LPS synergize for phosphorylation of ERK1/2 but not human macrophages, poly(I:C) (TLR3 agonist signaling via TRIF) p38 or JNK in macrophages and IMQ (TLR7 agonist signaling via MyD88) were used. In con- Because C5a together with LPS potentiates ERK1/2 in macro- trast to TLR4 activation by LPS, C5a amplified poly(I:C)-induced phages, we next investigated the temporal profile of ERK1/2 ac- (TRIF-dependent) IL-6 production from GM-Mf (Fig. 7B), tivation together with other MAPKs (p38 and JNK) in response to whereas it did not induce detectable levels of this cytokine in C5a versus C5a plus LPS. C5a induced a transient peak in ERK1/2 monocytes and M-Mf (Fig. 7A, 7C). In the case of IMQ-induced phosphorylation within 5 min in monocytes (Fig. 5A) and mac- (MyD88-dependent) IL-6 production, C5a enhanced this response rophages (Fig. 5B, 5C), but as we previously observed (Fig. 4C), in monocytes, as well as in both macrophage populations. C5a LPS amplified this response only in macrophages. C5a appeared also increased poly(I:C)-induced and IMQ-induced IL-10 pro- to have little or no effect on p38 and JNK phosphorylation (Fig. duction in monocytes and M-Mf (Fig. 7D, 7F), but it did not 5D–I) and did not amplify LPS-mediated activation of these sig- modulate this response in GM-Mf (Fig. 7E). These results sug- naling responses. Although LPS at 5 ng/ml concentration induced gest that suppression of inducible IL-6 production by C5a in GM- p38 and JNK phosphorylation, at the same concentration, LPS did Mf and M-Mf is selective for TLR4 signaling. not stimulate ERK1/2 phosphorylation in HMDMs; however, at a higher concentration (1 mg/ml), ERK1/2 phosphorylation was C5a enhances clearance of S. Typhimurium from HMDMs detected after 20 min (Supplemental Fig. 2G). Although there was There is evidence in support of C5aR/TLR crosstalk being exploited no synergistic effect for p38 phosphorylation in monocytes, the by some pathogens for host subversion (12, 35). Therefore, the The Journal of Immunology 4313

FIGURE 5. C5a-induced ERK1/2 phosphorylation is ampliﬁed in macrophages but not monocytes by LPS. To quantify (A–C) ERK1/2, (D–F) p38, and (G–I) JNK phosphorylation, cells (monocytes, GM-Mf, and M-Mf) were treated with LPS (5 ng/ml) with and without C5a (30 nM) for 10 min, after which cells were lysed. Phos- phorylated ERK1/2, p38, and JNK were measured by AlphaScreen SureFire assays. Treatments by C5a, LPS, or both are represented by broken lines, dotted lines, and solid lines, respectively. Error bars are means 6 SEM of three independent experiments (n = 3). **p , 0.01, ***p , 0.001 by student t test.

impact of C5a on the ability of HMDMs to clear S. Typhimurium the clearance of this bacterial pathogen from M-Mf, an effect that was examined. As expected, C5a alone did not have any bacte- was reversed by the C5aR antagonist 3D53 (Fig. 8B). In cytotoxicity ricidal effect on S. Typhimurium (Fig. 8A). Instead, C5a promoted assays monitoring lactate dehydrogenase release, infection assay

FIGURE 6. C5a suppression of LPS-induced IL-6 and TNF in macrophages is IL-10– and PI3K-independent. To determine dependency of IL-10, GM-Mf (black) and M-Mf (white) were pretreated with neutralizing Ab against human IL-10 (10 mg/ml, 30 min), or for dependency of PI3K signaling, wortmannin (1– 100 nM, pretreated for 1 h) prior to LPS (5 ng/ml) with and without C5a (30 nM) stimulation, after which (A, C, E) IL-6 and (B, D, F) TNF production were monitored at 24 h after stimulation. Error bars are means 6 SEM of three independent experiments (n = 3). *p , 0.05, **p , 0.01, ***p , 0.001 by Student t test. 4314 C5aR AND TLR4 CROSSTALK IN MONOCYTES VERSUS MACROPHAGES

FIGURE 7. Suppression by C5a of inducible IL-6 production from macrophages is selective for LPS responses. Human CD14+ monocytes (gray) and HMDMs (GM-Mf, black; M-Mf, white) were treated with 1) LPS (TLR4 agonist, 5 ng/ml) with and without C5a (30 nM) stimulation, 2) poly(I:C) (TLR3 agonist, 10 mg/ml) with and without C5a (30 nM) stimulation, or 3) IMQ (TLR7 agonist, 10 mg/ml) with and without C5a (30 nM) stimulation for 24 h in 10% FBS/IMDM. ELISA was used to detect (A–C) IL-6 and (D–F) IL-10 secreted in culture media. Error bars are means 6 SEM of three independent experiments (n = 3). *p , 0.05, **p , 0.01, ***p , 0.001 by Student t test.

supernatants harvested 8 h after infection showed no differences responses of human monocytes and macrophages. New findings in in macrophage viability irrespective of treatment (DMSO, SLI344, this study suggest that C5a modulates TLR4 signaling in macro- hC5a, 3D53, or combinations), thereby indicating that reduced phages to favor pathogen clearance while simultaneously limiting bacterial loads upon C5a treatment were not a result of increased excessive inflammatory responses. This control may be important macrophage death. Thus, C5a enhances human macrophage antimi- during localized infections, whereas C5a detection by monocytes crobial responses while reducing production of key proinflammatory likely acts as an alarm signal that triggers or amplifies inflammatory cytokines from these cells. responses during systemic infections. The mechanisms of C5aR/TLR4 crosstalk are complex and Discussion evidently species- and cell type–dependent. In human monocytes, Most pathogens trigger both complement and TLR activation, C5aR signaling has been reported to both downregulate (36) and and crosstalk between these two systems could potentially affect upregulate (37) TLR-induced IL-12 production. Although those inflammatory and antimicrobial responses in vivo in infectious studies were performedonsimilarIFN-g–primed human blood disease. However, studies on C5aR/TLR crosstalk have typically monocytes, the second signal used to stimulate the cells was focused on monocytes or macrophages separately. No previously different. C5a suppressed LPS-induced IL-12 production but up- reported study has directly compared the impact of crosstalk on the regulated S. aureus Cowan I–induced IL-12 production. Although functional responses of primary human monocytes versus donor- not addressed in either study, the variations reported are likely due matched human macrophages. The present study has found that to different TLRs being activated and differentially affected by the myeloid differentiation state is an important variable in de- C5aR signaling. The present study emphasizes that C5a suppression termining the impact of C5aR/TLR4 crosstalk on inflammatory of IL-6 production is specific to LPS/TLR4 signaling in human

FIGURE 8. C5a decreases S. Typhimurium survival in M-Mf.(A)GrowthofS. Typhimurium was monitored with and without C5a (100 nM) for 7 h. (B) M-Mf were infected with S. Typhimurium (MOI of 100) with and without C5a (30 nM). For antagonism of C5aR, M-Mf were pretreated with 3D53 (1 mM) for 30 min prior to C5a stimulation. Intramacrophage survival was assayed at 2 and 8 h after infection. Graph shows a representative of three independent experiments. Error bars are means 6 SEM of experimental triplicates. **p , 0.01, ***p , 0.001 by Student t test. The Journal of Immunology 4315 macrophages, because IL-6 production is actually boosted by parison with the C5a response, which occurs within 5 min. The C5aR crosstalk with other TLRs, such as TLR3 and TLR7. Similarly, MAP3K tumor progression locus 2 is essential for activation of in mouse macrophages, C5a was reported to selectively crosstalk with ERK1/2 by multiple TLRs, including TLR4, in macrophages (44). TLR4, but not TLR3, in suppressing IL-27 (p28) production (38). Consistent with this, we observed that LPS triggered MEK1 phos- C5a is thus not only a potent chemotactic agent, attracting immune phorylation in macrophages independently of activation of another cells to site of infection or injury, but it also differentially modulates MAP3K, c-Raf. ERK activation can produce different results, host immune responses, possibly tailoring them to match a particular depending on the stimulus and cell types, especially when com- type of infection. paring primary cells versus laboratory cultured cell lines (45). C5a The differentiating agents GM-CSF and M-CSF have differential enhanced ERK1/2 activation in rat neutrophils, and this was as- effects on macrophage lineage populations, which contribute to sociated with an increase in IL-6 production (46). This contrasts their functional heterogeneity (39). GM-Mf are often considered with our present findings for human macrophages. Complement- as “M1-like” macrophages, whereas M-Mf are classed as “M2-like” mediated ERK1/2 activation also inhibited IL-12 production in macrophages, according to their cytokine and gene expression human monocytes and mouse macrophages (14, 42), whereas the profiles (40, 41). Whereas LPS-induced inflammatory cytokine opposite was observed in mouse dendritic cells (47). Such re- production was clearly enhanced in GM-Mf compared with M-Mf ported variations in effects of C5a on LPS responses reinforce the in the present study, C5a exerted a similar suppressive effect on view that observations of complement-mediated effects on TLR inflammatory responses in both human macrophage populations. signaling are likely to not only be cell- and species-specific, but Thus, the phenotypic switch in C5a responsiveness during monocyte also to vary with the inflammatory readout being monitored. to macrophage differentiation occurred irrespective of the differ- In summary, in this study we have identified C5aR as a regula- entiating agent. In the future, it will be of interest to determine tory switch that modulates LPS/TLR4 signaling via the Gai/c-Raf/ whether other macrophage-polarizing factors (e.g., IL-4, IgG) MEK/ERK signaling axis in primary human macrophages, but not influence the ability of C5a to suppress TLR4-inducible inflam- in monocytes. Extrapolating data from mice to humans, although matory mediator production. Although the basal levels of C5AR convenient, is often not appropriate (48). Regulatory mechanisms and TLR4 increased during the course of differentiating mono- involved in C5aR/TLR4 crosstalk that were previously reported cytes to macrophages, the differences in the expression levels are in mouse studies may not be conserved in humans. Differential unlikely to be the cause of the phenotypic switch in the crosstalk. effects of C5a on monocytes versus macrophages have been dem- First, C5a treatments did not alter C5AR expression levels. Sec- onstrated in this study, consistent with a need to either amplify ond, C5a had opposing effects on inducible cytokine production monocyte proinflammatory responses during systemic danger when the macrophages were challenged with poly(I:C) or IMQ, recognition or attenuate macrophage proinflammatory cytokine instead of LPS. responses (without compromising microbicidal activity) during The effects of C5a on TLR4 responses were PTX-sensitive in both localized infections. Many therapies, including chemotherapy, ra- monocytes and macrophages, suggesting a Gai-sensitive mecha- diation, cancer vaccines, and small-molecule drugs, already use nism. C3a can also bind to a PTX-sensitive receptor, C3aR, that can immune cell–specific treatment strategies (49), but the rational recruit GaI coupling when activated. Others have shown that C3a basis for targeting specific cell types frequently lacks a level of acting on IFN-g–primed macrophages from C5aR2/2 mice (14), as understanding necessary to optimize such treatments. Studies such well as other Gai-specific ligands acting on IFN-g–primed human as this one have the potential to unlock secrets relating to inter- monocytes (42), can negatively regulate LPS-induced IL-12 pro- acting human immunoreceptors and control mechanisms that are duction. Also, PI3K signaling has been implicated in C5a regula- not obvious in studies targeting one particular receptor on a spe- tion of LPS signaling. The pan-PI3K inhibitor wortmannin was cific human immune cell type, let alone on immune cells from reported by la Sala et al. (42) to block the suppression by C5a of other mammalian species. IL-12p70 production, but this was not the case in the study performed by Hawlisch et al. (14). In our study, wortmannin treatment Acknowledgments had no impact on C5a-mediated suppression of LPS-induced IL-6 We acknowledge the Australian Red Cross for generously providing buffy and TNF production in human macrophages under the conditions coats for human monocyte isolation. examined. This suggests that the crosstalk is not only tightly regulated at multiple checkpoints in the signaling cascade, but that it is Disclosures also highly dependent on the activation status of the cells. The authors have no financial conflicts of interest. As shown previously in human neutrophils, C5a can activate the Raf/MEK/ERK signaling cascade either directly via Gai activation References or indirectly via Gbϒ/PLCb/PKC pathways (32, 43). We found 1. Monk, P. N., A. M. Scola, P. Madala, and D. P. Fairlie. 2007. Function, structure and that C5a-mediated suppression of LPS-induced IL-6 and TNF was therapeutic potential of complement C5a receptors. Br.J.Pharmacol.152: 429–448. independent of PLCb signaling (Supplemental Fig. 2). Although 2. Gerard, N. P., and C. Gerard. 1991. The chemotactic receptor for human C5a anaphylatoxin. Nature 349: 614–617. C5a rapidly phosphorylated ERK1/2 in both monocytes and mac- 3. Okinaga, S., D. Slattery, A. Humbles, Z. Zsengeller, O. Morteau, M. B. Kinrade, rophages, consistent with the literature (14, 15), only in macrophages R. M. Brodbeck, J. E. Krause, H. R. Choe, N. P. Gerard, and C. Gerard. 2003. C5L2, a nonsignaling C5A binding protein. Biochemistry 42: 9406–9415. did C5a increase c-Raf phosphorylation in the presence of LPS, 4. Chen, N. J., C. Mirtsos, D. Suh, Y. C. Lu, W. J. Lin, C. McKerlie, T. Lee, leading to a further downstream increase in phosphorylation of H. Baribault, H. Tian, and W. C. Yeh. 2007. C5L2 is critical for the biological MEK and ERK. Because LPS and C5a did not synergize to alter activities of the anaphylatoxins C5a and C3a. Nature 446: 203–207. 5. Skokowa, J., S. R. Ali, O. Felda, V. Kumar, S. Konrad, N. Shushakova, other MAPKs (p38 and JNK) in monocytes and macrophages, we R. E. Schmidt, R. P. Piekorz, B. Nu¨rnberg, K. Spicher, et al. 2005. Macrophages attribute the differences in monocytes and macrophages to the induce the inflammatory response in the pulmonary Arthus reaction through activation of the G /c-Raf/MEK/ERK pathway by C5a with LPS Gai2 activation that controls C5aR and Fc receptor cooperation. J. Immunol. 174: ai 3041–3050. in macrophages versus monocytes. 6. Akira, S., and K. Takeda. 2004. Toll-like receptor signalling. Nat. Rev. Immunol. 4: Low concentrations of LPS failed to stimulate ERK phospho- 499–511. 7. Rossol, M., H. Heine, U. Meusch, D. Quandt, C. Klein, M. J. Sweet, and rylation. However, only high concentrations of LPS triggered S. Hauschildt. 2011. LPS-induced cytokine production in human monocytes and delayed ERK1/2 activation (20 min after stimulation) by com- macrophages. Crit. Rev. Immunol. 31: 379–446. 4316 C5aR AND TLR4 CROSSTALK IN MONOCYTES VERSUS MACROPHAGES

8. Chang, L., and M. Karin. 2001. Mammalian MAP kinase signalling cascades. domain of the complement fragment receptor C5L2 in ligand binding. J. Biol. Nature 410: 37–40. Chem. 282: 3664–3671. 9. Kawai, T., O. Adachi, T. Ogawa, K. Takeda, and S. Akira. 1999. Unrespon- 30. Lim, J., A. Iyer, J. Y. Suen, V. Seow, R. C. Reid, L. Brown, and D. P. Fairlie. siveness of MyD88-deficient mice to endotoxin. Immunity 11: 115–122. 2013. C5aR and C3aR antagonists each inhibit diet-induced obesity, metabolic 10. Natarajan, M., K. M. Lin, R. C. Hsueh, P. C. Sternweis, and R. Ranganathan. dysfunction, and adipocyte and macrophage signaling. FASEB J. 27: 822–831. 2006. A global analysis of cross-talk in a mammalian cellular signalling net- 31. Raby, A. C., B. Holst, J. Davies, C. Colmont, Y. Laumonnier, B. Coles, S. Shah, work. Nat. Cell Biol. 8: 571–580. J. Hall, N. Topley, J. Ko¨hl, et al. 2011. TLR activation enhances C5a-induced 11. Postma, B., M. J. Poppelier, J. C. van Galen, E. R. Prossnitz, J. A. van Strijp, pro-inflammatory responses by negatively modulating the second C5a receptor, C. J. de Haas, and K. P. van Kessel. 2004. Chemotaxis inhibitory protein of C5L2. Eur. J. Immunol. 41: 2741–2752. Staphylococcus aureus binds specifically to the C5a and formylated peptide 32. Buhl, A. M., N. Avdi, G. S. Worthen, and G. L. Johnson. 1994. Mapping of the receptor. J. Immunol. 172: 6994–7001. C5a receptor signal transduction network in human neutrophils. Proc. Natl. 12. Wang, M., J. L. Krauss, H. Domon, K. B. Hosur, S. Liang, P. Magotti, M. Triantafilou, Acad. Sci. USA 91: 9190–9194. K. Triantafilou, J. D. Lambris, and G. Hajishengallis. 2010. Microbial hijacking of 33. Favata, M. F., K. Y. Horiuchi, E. J. Manos, A. J. Daulerio, D. A. Stradley, complement-Toll-like receptor crosstalk. Sci. Signal. 3: ra11. W. S. Feeser, D. E. Van Dyk, W. J. Pitts, R. A. Earl, F. Hobbs, et al. 1998. 13. Hajishengallis, G., and J. D. Lambris. 2010. Crosstalk pathways between Toll- Identification of a novel inhibitor of mitogen-activated protein kinase kinase. J. like receptors and the complement system. Trends Immunol. 31: 154–163. Biol. Chem. 273: 18623–18632. 14. Hawlisch, H., Y. Belkaid, R. Baelder, D. Hildeman, C. Gerard, and J. Ko¨hl. 2005. 34. Chanteux, H., A. C. Guisset, C. Pilette, and Y. Sibille. 2007. LPS induces IL-10 C5a negatively regulates Toll-like receptor 4-induced immune responses. Im- production by human alveolar macrophages via MAPKinases- and Sp1-dependent munity 22: 415–426. mechanisms. Respir. Res. 8: 71. 15. Okazaki, N., K. Hazeki, T. Izumi, K. Nigorikawa, and O. Hazeki. 2011. C5a 35. Liang, S., J. L. Krauss, H. Domon, M. L. McIntosh, K. B. Hosur, H. Qu, F. Li, controls TLR-induced IL-10 and IL-12 production independent of phosphoino- A. Tzekou, J. D. Lambris, and G. Hajishengallis. 2011. The C5a receptor impairs sitide 3-kinase. J. Biochem. 149: 265–274. IL-12-dependent clearance of Porphyromonas gingivalis and is required for in- 16. Bosmann, M., V. R. Patel, N. F. Russkamp, F. Pache, F. S. Zetoune, J. V. Sarma, duction of periodontal bone loss. J. Immunol. 186: 869–877. and P. A. Ward. 2011. MyD88-dependent production of IL-17F is modulated by 36. Wittmann, M., J. Zwirner, V. A. Larsson, K. Kirchhoff, G. Begemann, A. Kapp, the anaphylatoxin C5a via the Akt signaling pathway. FASEB J. 25: 4222–4232. O. Go¨tze, and T. Werfel. 1999. C5a suppresses the production of IL-12 by IFN- 17. Bosmann, M., J. V. Sarma, G. Atefi, F. S. Zetoune, and P. A. Ward. 2012. Ev- g-primed and lipopolysaccharide-challenged human monocytes. J. Immunol. idence for anti-inflammatory effects of C5a on the innate IL-17A/IL-23 axis. 162: 6763–6769. FASEB J. 26: 1640–1651. 37. Karp, C. L., A. Grupe, E. Schadt, S. L. Ewart, M. Keane-Moore, P. J. Cuomo, 18. Cavaillon, J. M., C. Fitting, and N. Haeffner-Cavaillon. 1990. Recombinant C5a J. Ko¨hl, L. Wahl, D. Kuperman, S. Germer, et al. 2000. Identification of com- enhances interleukin 1 and tumor necrosis factor release by lipopolysaccharide- plement factor 5 as a susceptibility locus for experimental allergic asthma. Nat. stimulated monocytes and macrophages. Eur. J. Immunol. 20: 253–257. Immunol. 1: 221–226. 19. Zaal, A., S. N. Lissenberg-Thunnissen, G. van Schijndel, D. Wouters, 38. Bosmann, M., M. D. Haggadone, M. R. Hemmila, F. S. Zetoune, J. V. Sarma, and S. M. van Ham, and A. ten Brinke. 2013. Crosstalk between Toll like receptors P. A. Ward. 2012. Complement activation product C5a is a selective suppressor and C5a receptor in human monocyte derived DCs suppress inflammatory cy- of TLR4-induced, but not TLR3-induced, production of IL-27(p28) from mac- tokine production. Immunobiology 218: 175–180. rophages. J. Immunol. 188: 5086–5093. 20. Copeland, S., H. S. Warren, S. F. Lowry, S. E. Calvano, and D. Remick; In- 39. Bailey, M. J., D. C. Lacey, B. V. de Kok, P. D. Veith, E. C. Reynolds, and flammation and the Host Response to Injury Investigators. 2005. Acute in- J. A. Hamilton. 2011. Extracellular proteomes of M-CSF (CSF-1) and GM-CSF- flammatory response to endotoxin in mice and humans. Clin. Diagn. Lab. dependent macrophages. Immunol. Cell Biol. 89: 283–293. Immunol. 12: 60–67. 40. Fleetwood, A. J., T. Lawrence, J. A. Hamilton, and A. D. Cook. 2007. 21. Schroder, K., K. M. Irvine, M. S. Taylor, N. J. Bokil, K. A. Le Cao, Granulocyte-macrophage colony-stimulating factor (CSF) and macrophage CSF- K. A. Masterman, L. I. Labzin, C. A. Semple, R. Kapetanovic, L. Fairbairn, et al. dependent macrophage phenotypes display differences in cytokine profiles and 2012. Conservation and divergence in Toll-like receptor 4-regulated gene ex- transcription factor activities: implications for CSF blockade in inflammation. J. pression in primary human versus mouse macrophages. Proc. Natl. Acad. Sci. Immunol. 178: 5245–5252. USA 109: E944–E953. 41. Lacey, D. C., A. Achuthan, A. J. Fleetwood, H. Dinh, J. Roiniotis, G. M. Scholz, 22. Ariffin, J. K., and M. J. Sweet. 2013. Differences in the repertoire, regulation and M. W. Chang, S. K. Beckman, A. D. Cook, and J. A. Hamilton. 2012. Defining function of Toll-like receptors and inflammasome-forming Nod-like receptors GM-CSF- and macrophage-CSF-dependent macrophage responses by in vitro between human and mouse. Curr. Opin. Microbiol. 16: 303–310. models. J. Immunol. 188: 5752–5765. 23. Dinakar, C., A. Malur, B. Raychaudhuri, L. T. Buhrow, A. L. Melton, 42. la Sala, A., M. Gadina, and B. L. Kelsall. 2005. Gi-protein-dependent inhibition M. S. Kavuru, and M. J. Thomassen. 1999. Differential regulation of human of IL-12 production is mediated by activation of the phosphatidylinositol 3- blood monocyte and alveolar macrophage inflammatory cytokine production by kinase-protein 3 kinase B/Akt pathway and JNK. J. Immunol. 175: 2994–2999. nitric oxide. Ann. Allergy Asthma Immunol. 82: 217–222. 43. Jiang, H., Y. Kuang, Y. Wu, A. Smrcka, M. I. Simon, and D. Wu. 1996. Pertussis 24. Netea, M. G., C. A. Nold-Petry, M. F. Nold, L. A. Joosten, B. Opitz, J. H. van der toxin-sensitive activation of phospholipase C by the C5a and fMet-Leu-Phe Meer, F. L. van de Veerdonk, G. Ferwerda, B. Heinhuis, I. Devesa, et al. 2009. receptors. J. Biol. Chem. 271: 13430–13434. Differential requirement for the activation of the inflammasome for processing 44. Gantke, T., S. Sriskantharajah, M. Sadowski, and S. C. Ley. 2012. IkB kinase and release of IL-1b in monocytes and macrophages. Blood 113: 2324–2335. regulation of the TPL-2/ERK MAPK pathway. Immunol. Rev. 246: 168–182. 25. Finch, A. M., A. K. Wong, N. J. Paczkowski, S. K. Wadi, D. J. Craik, 45. Rao, K. M. 2001. MAP kinase activation in macrophages. J. Leukoc. Biol. 69: D. P. Fairlie, and S. M. Taylor. 1999. Low-molecular-weight peptidic and cyclic 3–10. antagonists of the receptor for the complement factor C5a. J. Med. Chem. 42: 46. Riedemann, N. C., R. F. Guo, T. J. Hollmann, H. Gao, T. A. Neff, J. S. Reuben, 1965–1974. C. L. Speyer, J. V. Sarma, R. A. Wetsel, F. S. Zetoune, and P. A. Ward. 2004. 26. Reid, R. C., G. Abbenante, S. M. Taylor, and D. P. Fairlie. 2003. A convergent Regulatory role of C5a in LPS-induced IL-6 production by neutrophils during solution-phase synthesis of the macrocycle Ac-Phe-[Orn-Pro-D-Cha-Trp-Arg], sepsis. FASEB J. 18: 370–372. a potent new antiinflammatory drug. J. Org. Chem. 68: 4464–4471. 47. Baruah, P., I. E. Dumitriu, T. H. Malik, H. T. Cook, J. Dyson, D. Scott, 27. Trieu, A., N. Bokil, J. A. Dunn, T. L. Roberts, D. Xu, F. Y. Liew, D. A. Hume, E. Simpson, and M. Botto. 2009. C1q enhances IFN-g production by antigen- K. J. Stacey, and M. J. Sweet. 2009. TLR9-independent effects of inhibitory specific T cells via the CD40 costimulatory pathway on dendritic cells. Blood oligonucleotides on macrophage responses to S. typhimurium. Immunol. Cell 113: 3485–3493. Biol. 87: 218–225. 48. Seok, J., H. S. Warren, A. G. Cuenca, M. N. Mindrinos, H. V. Baker, W. Xu, 28. Bamberg, C. E., C. R. Mackay, H. Lee, D. Zahra, J. Jackson, Y. S. Lim, D. R. Richards, G. P. McDonald-Smith, H. Gao, L. Hennessy, et al; Inflammation P. L. Whitfeld, S. Craig, E. Corsini, B. Lu, et al. 2010. The C5a receptor (C5aR) and Host Response to Injury, Large Scale Collaborative Research Program. C5L2 is a modulator of C5aR-mediated signal transduction. J. Biol. Chem. 285: 2013. Genomic responses in mouse models poorly mimic human inflammatory 7633–7644. diseases. Proc. Natl. Acad. Sci. USA 110: 3507–3512. 29. Scola, A. M., A. Higginbottom, L. J. Partridge, R. C. Reid, T. Woodruff, 49. Sheridan, C. 2013. Industry pursues co-stimulatory receptor immunomodulators S. M. Taylor, D. P. Fairlie, and P. N. Monk. 2007. The role of the N-terminal to treat cancer. Nat. Biotechnol. 31: 181–183. Appendix V

Daniel M. Hohenhaus, , Kolja Schaale, Kim-Anh Le Cao, Vernon Seow, Abishek Iyer, David P. Fairlie, and Matthew J. Sweet. An mRNA atlas of G protein-coupled receptor expression during primary human monocyte/macrophage differentiation and lipopolysaccharide-mediated activation identifies targetable candidate regulators of inflammation. Immunobiology (2013) 218: 1345-1353.

Immunobiology 218 (2013) 1345–1353

Contents lists available at ScienceDirect

Immunobiology

jou rnal homepage: www.elsevier.com/locate/imbio

An mRNA atlas of G protein-coupled receptor expression during

primary human monocyte/macrophage differentiation and

lipopolysaccharide-mediated activation identiﬁes targetable

candidate regulators of inﬂammation

a,b a,b a,c a

Daniel M. Hohenhaus , Kolja Schaale , Kim-Anh Le Cao , Vernon Seow ,

a a,b a,b,

Abishek Iyer , David P. Fairlie , Matthew J. Sweet ∗

Institute for Molecular Bioscience, The University of Queensland, Brisbane, Qld 4072, Australia

Australian Infectious Disease Research Centre, The University of Queensland, Brisbane, Qld 4072, Australia

Queensland Facility for Advanced Bioinformatics, The University of Queensland, Brisbane, Qld 4072, Australia

a r t i c l e i n f o a b s t r a c t

Article history: G protein-coupled receptors (GPCRs) are among the most important targets in drug discovery. In this

Received 10 April 2013

study, we used TaqMan Low Density Arrays to proﬁle the full GPCR repertoire of primary human

Received in revised form 4 July 2013

macrophages differentiated from monocytes using either colony stimulating factor-1 (CSF-1/M-CSF)

Accepted 6 July 2013

(CSF-1 M␾) or granulocyte macrophage colony stimulating factor (GM-CSF) (GM-CSF M␾). The over-

Available online 15 July 2013

all trend was a downregulation of GPCRs during monocyte to macrophage differentiation, but a core set

of 10 genes (e.g. LGR4, MRGPRF and GPR143) encoding seven transmembrane proteins were upregulated,

Keywords:

irrespective of the differentiating agent used. Several of these upregulated GPCRs have not previously

G protein-coupled receptor

Infection been studied in the context of macrophage biology and/or inﬂammation. As expected, CSF-1 M␾ and GM-

Inflammation CSF M␾ exhibited differential inflammatory cytokine profiles in response to the Toll-like Receptor (TLR)4

Lipopolysaccharide agonist lipopolysaccharide (LPS). Moreover, 15 GPCRs were differentially expressed between these cell

Macrophage populations in the basal state. For example, EDG1 was expressed at elevated levels in CSF-1 M␾ versus

Monocyte GM-CSF M␾, whereas the reverse was true for EDG6. 101 GPCRs showed differential regulation over an

Toll-like receptor

LPS time course, with 65 of these proﬁles being impacted by the basal differentiation state (e.g. GPRC5A,

GPRC5B). Only 14 LPS-regulated GPCRs showed asynchronous behavior (divergent LPS regulation) with

respect to differentiation status. Thus, the differentiation state primarily affects the magnitude of LPS-

regulated expression, rather than causing major reprogramming of GPCR gene expression proﬁles. Several

GPCRs showing differential proﬁles between CSF-1 M␾ and GM-CSF M␾ (e.g. P2RY8, GPR92, EMR3) have

not been widely investigated in macrophage biology and inﬂammation. Strikingly, several closely related

GPCRs displayed completely opposing patterns of regulation during differentiation and/or activation (e.g.

EDG1 versus EDG6, LGR4 versus LGR7, GPRC5A versus GPRC5B). We propose that selective regulation of

GPCR5A and GPCR5B in CSF-1 M␾ contributes to skewing toward the M2 macrophage phenotype. Our

analysis of the GPCR repertoire expressed during primary human monocyte to macrophage differenti-

ation and TLR4-mediated activation provides a valuable new platform for conducting future functional

Introduction

The human genome encodes more than 800 seven

Abbreviations: CSF-1, colony stimulating factor-1; CSF-1 M␾, monocytes differ-

transmembrane-containing proteins, of which more than 300

entiated to HMDM with CSF-1; GM-CSF, granulocyte macrophage colony stimulating

factor; GM-CSF M␾, monocytes differentiated to HMDM with GM-CSF; GPCR, G are non-olfactory receptors (Alexander et al., 2011). The major-

protein-coupled receptor; HMDM, human monocyte-derived macrophage; HSC, ity of these receptors transduce signals, at least in part, by

hematopoietic stem cell; IL, interleukin; LPS, lipopolysaccharide; TLDA, TaqMan low

coupling to heterotrimeric G proteins. Thus, the term G protein-

density arrays; TLR, Toll-like receptor; TNF, tumor necrosis factor.

coupled receptor (GPCR) is often used interchangeably with

∗ Corresponding author at: Institute for Molecular Bioscience, University of

seven transmembrane-containing receptor, although this is not

Queensland, St Lucia 4072, Australia. Tel.: +61 7 3346 2082; fax: +61 7 3346 2101.

E-mail address: [email protected] (M.J. Sweet). always the case. For simplicity, we hereafter refer to all seven

1346 D.M. Hohenhaus et al. / Immunobiology 218 (2013) 1345–1353

transmembrane-containing proteins as GPCRs. GPCRs have been present at elevated levels during inﬂammatory responses, and that

sub-divided according to the GRAFS classiﬁcation system into the GM-CSF-differentiated macrophages have a hyper-inﬂammatory

glutamate, rhodopsin, adhesion, frizzled/taste2 and secretin GPCR phenotype. For example, in in vitro experiments, mouse bone

sub-families (Fredriksson et al., 2003). The International Union of marrow-derived macrophages differentiated with GM-CSF display

Basic and Clinical Pharmacology Committee on Receptor Nomen- elevated LPS-triggered inﬂammatory responses, as compared to

clature and Drug Classiﬁcation alternatively classiﬁes GPCRs into those differentiated with CSF-1 (Fleetwood et al., 2007). Increased

class A (rhodopsin), class B (secretin), class C (glutamate), frizzled, pro-inﬂammatory responses have also been reported with primary

and other 7TM proteins. GPCRs are the most prevalent signal- human monocytes differentiated into macrophages with GM-CSF

transducing proteins on the cell surface, and respond to a diverse versus CSF-1 (Lacey et al., 2012; Sierra-Filardi et al., 2010; Verreck

array of environmental cues to control numerous physiological and et al., 2004). Furthermore, many studies have demonstrated patho-

pathological processes. They represent targets for approximately logical roles for this cytokine in inﬂammatory diseases (Hamilton,

one third of all pharmaceuticals (Overington et al., 2006). With 2008; Hamilton and Anderson, 2004), and in acute inﬂamma-

the recent emergence of solved three dimensional structures for tory models such as LPS-induced lung inﬂammation (Bozinovski

several class A GPCRs (Venkatakrishnan et al., 2013), this family et al., 2002). Thus, GM-CSF M␾ can be considered as representative

of cell surface proteins promises many new opportunities for of inﬂammatory macrophages. In contrast, CSF-1 M␾ are proba-

development of therapeutic agents. bly reasonable cellular surrogates of tissue resident macrophages

Macrophages are key innate immune cells that occupy dis- under homeostatic conditions, since CSF-1 is constitutively present

tinct anatomical locations within every tissue of the body, and are in vivo during homeostasis.

recruited in further numbers to speciﬁc tissues when homeostasis We previously documented the constitutive and LPS-regulated

is dysregulated (Hume, 2008). As danger sentinels, macrophages GPCR repertoire of mouse macrophages (Lattin et al., 2008). This

initiate inﬂammatory responses to enable a coordinated physio- analysis identiﬁed several members of the P2RY family of GPCRs

logical response to limit tissue damage and initiate and coordinate that were enriched in macrophages and/or regulated by LPS. More

repair processes. Macrophages recognize danger through several recently however, we identiﬁed widespread divergence in LPS-

families of pattern recognition receptors such as the Toll-like regulated gene expression between primary human and mouse

receptors (TLRs) that detect both pathogen-associated molecular macrophages (Schroder et al., 2012). Others have also reported

patterns during infections, as well as host-derived damage- human versus mouse differences in the basal gene expression

associated molecular patterns that are indicative of cell stress programs of monocyte subsets (Ingersoll et al., 2010), and in CSF-

and/or damage during acute and chronic diseases. For example, 1 and GM-CSF differentiated macrophages (Lacey et al., 2012).

TLR4 recognizes lipopolysaccharide (LPS) from Gram-negative bac- Such differences may contribute to the very different transcrip-

terial cell walls (Rossol et al., 2011), as well as several host-derived tomic responses in mouse models versus human inﬂammatory

factors including HMGB1 (Yang et al., 2010), oxidized LDL and beta- diseases (Seok et al., 2013). In light of such species differences and

amyloid peptide (Stewart et al., 2010). TLRs primarily act, though the need to understand GPCR expression proﬁles and functions

not exclusively, by regulating the expression of a suite of genes that in normal and pathological immune responses, we conducted a

control inﬂammation, antimicrobial responses and antigen presen- detailed expression analysis of the non-sensory GPCR repertoire

tation. expressed during primary human monocyte to macrophage dif-

Recent evidence suggests that tissue resident macrophage ferentiation and their activation by the TLR4 agonist, LPS. Our

populations can arise through both hematopoietic stem cell (HSC)- analysis has identiﬁed a small set of GPCRs that are upregulated

dependent and -independent processes. The latter pathway occurs during monocyte to macrophage differentiation, as well as numer-

independently of the c-Myb transcription factor and involves the ous GPCRs showing differential expression patterns between CSF-1

development of some tissue macrophage populations from embry- M␾ and GM-CSF M␾. Our data provide insights into the differential

onic macrophages in the yolk sac (Schulz et al., 2012). The former inﬂammatory proﬁles of these macrophage populations, and iden-

pathway involves the cytokine-mediated differentiation of HSC tiﬁes several GPCRs not previously associated with inﬂammation

into circulating monocytes, and of these cells into tissue resi- and/or macrophage biology (e.g. LGR4, MRGPRF). This information

dent and inﬂammatory macrophages. Several cytokines, including is expected to be a valuable resource for future functional analysis

IL-3, GM-CSF, CSF-1 and IL-34, direct monocyte/macrophage dif- of human macrophage-expressed GPCRs.

ferentiation from HSC. CSF-1 and IL-34 are expressed widely, but

differentially, in adult mouse tissues (Wei et al., 2010), and are

required for the differentiation of many tissue resident macrophage Materials and methods

populations (Wang et al., 2012; Yoshida et al., 1990). During acute

and chronic inﬂammation, several other cytokines including tumor Cell culture and reagents

necrosis factor (TNF) (Sade-Feldman et al., 2013), interferon-␥

(Schroder et al., 2004) and GM-CSF (Hamilton, 2008) can also All studies using primary human cells were approved by the

inﬂuence myeloid development. In terms of macrophage differ- University of Queensland Medical Research Ethics Committee.

entiation, GM-CSF is probably the most widely studied of these Human CD14 monocytes were isolated from buffy coat provided

factors. Comparisons between GM-CSF- and CSF-1-differentiated by the Australian Red Cross Blood Service using MACS CD14

macrophages are often made in the context of classically activated positive selection kits (Miltenyi Biotech), according to the manu-

“M1” macrophages and alternatively activated “M2” macrophages, facturer’s instructions. CD14 monocytes were cultured overnight

respectively (Hamilton, 2008; Hamilton and Achuthan, 2013). M1 on 10 cm TC plates (Nunc) prior to RNA extraction. Human

macrophages are associated with pro-inﬂammatory and micro- monocyte-derived macrophages (HMDM) were generated by 7-

bicidal responses, and are usually generated in vitro using IFN-␥ day in vitro culture of CD14 monocytes in the presence of either

plus LPS. In contrast, M2 macrophages are associated with anti- 1 10 U/mL recombinant human CSF-1 (Chiron) (CSF-1 M␾) or

inﬂammatory responses, tissue remodeling and immunoregulation 10 ng/mL recombinant human GM-CSF (Peprotech) (GM-CSF M␾).

and are typically generated in vitro using IL-4 or IL-13 (Biswas Both CD14 monocytes and HMDMs were maintained in Iscove’s

et al., 2012). The extensive phenotypic diversity and plasticity modiﬁed Dulbecco’s medium (Life Technologies), supplemented

of macrophages means that this classiﬁcation system is probably with 10% heat-inactivated fetal calf serum (Life Technologies),

overly simplistic, but it is certainly the case that GM-CSF is typically 20 U/mL penicillin (Life Technologies), 20 ␮g/mL streptomycin (Life

D.M. Hohenhaus et al. / Immunobiology 218 (2013) 1345–1353 1347

Technologies) and 2 mM l-glutamine (Life Technologies). On day 6, status. All heatmaps were displayed with Euclidian distance and

CSF-1 M␾ and GM-CSF M␾ were replated on 10 cm TC plates in the Ward linkage.

presence of fresh CSF-1 or GM-CSF (10 10 cells per plate in 10 mL

complete media). On day 7, these cells were treated with 10 ng/mL SYBR Green qPCR

Salmonella enterica serovar minnesota LPS (Sigma, L9764) for 1, 4,

8 or 24 h, or were left untreated, after which RNA was extracted. RNA was prepared using Qiagen Rneasy mini-kits (Qiagen).

Culture supernatants were prepared from the same samples. cDNA was generated with Superscript III reverse transcriptase

(Life Technologies). Levels of mRNA for individual genes were

determined by quantitative real-time PCR using SYBR Green (Life

cDNA generation and TaqMan low density arrays

Technologies), gene speciﬁc primers, and an ABI Prism 7900 HT

sequence detection system (Life Technologies). CT values were

RNA extraction was performed using RNeasy mini kits (Qiagen),

calculated from ampliﬁcation plots and gene expression was

as per the manufacturer’s instructions. Genomic DNA contami-

expressed relative to the internal control, hypoxanthine phospho-

nation of puriﬁed RNA was removed by treatment with DNaseI

ribosyltransferase 1 (HPRT1), using the !CT method. Statistical

(Life Technologies). cDNA was generated from 2 ␮g of extracted

analysis was performed using GraphPad Prism 5© software. Sig-

RNA using Superscript III (Life Technologies). Samples were ana-

niﬁcance was calculated by one-way ANOVA, with a Bonferroni

lyzed on pre-designed 384-well Human GPCR TaqMan Low Density

correction applied for multiple comparisons. Primer sequences are

Arrays (TLDA) (Life Technologies, Cat No. 4367785). These arrays

contained within Supplementary data.

contained validated primer/probe sets for 367 GPCRs and 14

endogenous controls (18S, ACTB, B2M, GAPDH, GUSB, HMBS,

ELISAs

HPRT1, IPOB, PGK1, POLR2A, PPIA, RPLPO, TBP and TFRC). 250 ng of

cDNA was loaded per port on each TLDA card. TLDA were run on a

Levels of speciﬁc cytokines (IL-6, IL-10 and TNF) in HMDM

Life Technologies 7900HT fast real-time PCR instrument, according

cell culture supernatants at 24 h post-stimulation were measured

to the manufacturer’s instruction.

using sandwich ELISAs (Becton Dickinson), according to the man-

ufacturer’s instructions. Statistical analysis was performed using

TLDA data analysis GraphPad Prism 5© software. Signiﬁcance was calculated by one-

way ANOVA, with a Bonferroni correction applied for multiple

Data were analyzed using RQ manager version 1.2 (Life Tech- comparisons.

nologies) and Integromics StatMiner software suite version 4.2

(Integromics). CT thresholds were set in the linear phase of each Immunoblotting

genes ampliﬁcation plot using RQ manager version 1.2 and were

kept consistent across samples. The Genorm algorithm, built into Day 6 CSF-1 M␾ and GM-CSF M␾ were plated overnight in 6-

the StatMiner software, was applied to determine the most sta- well plates (1 10 cells/well in 2 mL complete media). On day 7,

ble endogenous controls, with 18S, ACTB, GAPDH, HPRT1, PGK1 these cells were treated with 10 ng/mL S. enterica serovar min-

and PPIA selected for data normalization. Only genes with a mean nesota LPS (Sigma, L9764) for 1, 4, 8, 24 or 48 h, or were left

C 32 for at least one sample were included in StatMiner analyses. untreated. Cells were lysed with 66 mM Tris–HCl (pH 7.4)/2%

T ≤

HDAC3 and MATK, which were included on the commercially avail- SDS, containing 1 mM sodium vanadate (Sigma), 1 mM sodium

able arrays and were not annotated as control genes, were excluded pyrophosphate (Sigma), 1 mM sodium molybdite (Sigma), 10 mM

from further analysis, since they do not encode seven transmem- sodium ﬂuoride (Sigma) and 1 Complete EDTA-free protease

brane receptors. Subsequent statistical analysis was performed inhibitor cocktail (Roche). Proteins were resolved by SDS-PAGE

using the R statistical software. Principal component analysis and using pre-cast NuPAGE Bis-Tris 4–12% gels (Life Technologies) and

hierarchical clustering were used as unsupervised approaches to transferred to methanol-activated Immobilon-P PDVF membranes

cluster the data. In both cases, a multilevel approach was applied to (Millipore). Membranes were blocked and probed with rabbit

take into account repeated measures for individual donors (Liquet anti-EDNRB (1:1000) (Abcam) and rabbit anti-GAPDH (1:10,000)

et al., 2012). To identify GPCRs showing statistically signiﬁcant dif- (Trevigen). Membranes were probed with Goat HRP-linked anti-

ferential expression between monocytes and either CSF-1 M␾ or rabbit IgG secondary antibody (1:2500) (Cell Signaling). HRP was

GM-CSF M␾, a linear mixed model, followed by Benjamini and detected using the ECL Plus system (GE Healthcare) and Fuji ﬁlm.

Hochberg multiple correction, was employed. Genes showing 5

≥

fold differences between these populations were also captured to Results

identify genes of likely biological signiﬁcance, even though some

may not have shown statistically signiﬁcant differences for the The GPCR repertoire during primary human macrophage

sample size analyzed. To identify GPCRs showing statistically sig- differentiation and activation

niﬁcant changes in gene expression in response to LPS, one-way

Welch t-tests followed by Benjamini and Hochberg multiple cor- To generate an atlas of the non-olfactory GPCR repertoire dur-

rection were applied for each time point (Benjamini and Hochberg, ing human macrophage differentiation and activation, we used

1995). As above, genes showing 5 fold differences in LPS-regulated Human GPCR TLDA (Life Technologies) to analyze donor-matched

≥ + +

gene expression at any single time point were also captured. In CD14 monocytes, macrophages differentiated from CD14 mono-

addition, a statistical analysis was performed using repeated meas- cytes with either CSF-1 (CSF-1 M␾) or GM-CSF (GM-CSF M␾), and

ures ANOVA to include time (h LPS stimulation), differentiation CSF-1 M␾ and GM-CSF M␾ activated with LPS for 1, 4 or 24 h. Of

status (CSF-1 M␾ or GM-CSF M␾) and the interaction between the 367 GPCRs on the TLDA, 123 GPCRs were expressed in at least

time and differentiation status as ﬁxed effects, and the individ- one of the above conditions at a level above the selected threshold

ual donors as a random effect. This analysis enabled us to capture (C 32) (Supplementary data). Consistent with prior expectation,

T ≤

genes showing statistically signiﬁcant LPS-regulated proﬁles, and principal component analysis demonstrated that individual treat-

the impact of differentiation status on these proﬁles. Genes iden- ments from the 3 different donors (biological replicates) clustered

tified as having a significant interaction were further described as together, that the monocyte profile diverged substantially from the

having an “asychronous behavior” across time and differentiation various macrophage proﬁles, and that the 4 and 24 h LPS-treated

1348 D.M. Hohenhaus et al. / Immunobiology 218 (2013) 1345–1353

Fig. 1. Unsupervised clustering of the GPCR expression repertoire. Principal component analysis (A) and hierarchical clustering analysis (B) of all ﬁltered data (at least one

condition showing a mean C 32).

T ≤

macrophage populations were clearly differentiated from the cor- interestingly, TNF was not (Fig. 3F). As expected, LPS-induced IL-10

responding control and 1 h LPS-treated macrophage populations and CCL2 mRNA levels were reduced in GM-CSF M␾ in comparison

(Fig. 1A). These conclusions were also supported by unsupervised to CSF-1 M␾ (Fig. 3G, H). These ﬁndings thus validate our approach

hierarchical clustering analysis (Fig. 1B). Fig. S1 again shows that for examining GPCR expression proﬁles in CSF-1 M␾ versus GM-CSF

monocytes diverge substantially from the macrophage populations M␾ for the speciﬁc samples used in the TLDA proﬁling.

in terms of their GPCR proﬁle. Collectively, the above ﬁndings indi- We next asked which GPCRs were differentially expressed in

cate that inherent donor variability does not confound analysis of unstimulated CSF-1 M␾ versus GM-CSF M␾. 15 GPCRs showed

this data set, and provide conﬁdence in the robustness of the data differential expression proﬁles for CSF-1 M␾ versus GM-CSF M␾

as a whole. (Fig. 2A, Supplementary data). Fig. 4A shows those GPCRs with

5 fold difference between unstimulated CSF-1 M␾ and GM-CSF

≥

M␾, some of which are not statistically signiﬁcant for the sam-

GPCR expression patterns during human monocyte to

ple size, but are likely to be biologically signiﬁcant. We again used

macrophage differentiation

independent SYBR Green PCR on additional donors to conﬁrm ele-

vated EDNRB and EDG1 mRNAs in CSF-1 M␾ versus GM-CSF M␾

We next identiﬁed GPCRs differentially expressed during

(Fig. 4B, C), as well as increased HRH4 and EDG6 expression in

human monocyte to macrophage differentiation. Of the 60 GPCRs

GM-CSF M␾ versus CSF-1 M␾ (Fig. 4D, E). Such GPCRs could con-

showing statistically signiﬁcant differential expression patterns

ceivably contribute to the differing inﬂammatory proﬁles of these

between monocytes and HMDM, 38 were different between mono-

two macrophage populations.

cytes and HMDM irrespective of the differentiation stimulus, 6

differed only between monocytes and CSF-1 M␾, and 16 differed

only between monocytes and GM-CSF M␾ (Fig. 2A). Thus, the

GPCR proﬁling of primary human macrophage LPS-mediated

majority of changes in GPCR expression during macrophage differ-

activation

entiation occur irrespective of the differentiation stimulus. Fig. 2B

illustrates those GPCRs showing 5 fold difference between mono-

≥ Our initial unsupervised clustering approaches demonstrated

cytes and CSF-1 M␾, whilst Fig. 2C highlights those GPCRs with a

that LPS had little effect on GPCR expression at 1 h post-LPS stim-

5 fold difference between monocytes and GM-CSF M␾. For GPCRs

≥ ulation, but had profound effects at 4 and 24 h post-stimulation

that were regulated during monocyte to macrophage differentia-

(Fig. 1A, B). We next used two distinct approaches to capture

tion irrespective of the differentiating agent, there was broadly a

individual LPS-regulated GPCRs. Firstly, we performed statistical

trend for downregulation of expression. That is, 28 of 38 GPCRs

analysis at individual time points (Supplementary data). Secondly,

were downregulated during differentiation to both CSF-1 M␾ and

we used a repeated measures analysis to capture genes showing

GM-CSF M␾, while only 10 GPCRs were upregulated (Supplemen-

divergent proﬁles over the LPS time course (Supplementary data).

tary data). We used SYBR Green qPCR, incorporating additional

The latter analysis identiﬁed 101 GPCRs that were signiﬁcantly

donor samples, to further conﬁrm that LGR4 (Fig. 2D), novel in the

regulated by LPS, of which 65 showed differing LPS proﬁles for

context of macrophage biology and inﬂammation, was upregulated

CSF-1 M␾ versus GM-CSF M␾ (Fig. 5A). However, only 16 genes

during macrophage differentiation.

displayed asynchronous behavior (i.e. a signiﬁcant interaction in

the repeated measures ANOVA) when comparing differentiation

GM-CSF M" display enhanced pro-inﬂammatory responses as status (CSF-1 versus GM-CSF-1) versus LPS regulation (Fig. 5A),

compared to CSF-1 M" thus indicating that the major effect of different differentiation

conditions was to alter the magnitude of the LPS response, rather

Other groups have reported that GM-CSF M␾ display enhanced than to cause a divergence in the LPS proﬁle. Fig. 5B highlights

inﬂammatory cytokine proﬁles by comparison to CSF-1 M␾ (Lacey all LPS-regulated genes affected by differentiation status, whilst

et al., 2012; Sierra-Filardi et al., 2010; Verreck et al., 2004). We ﬁrst Fig. 5C shows the 16 genes displaying asynchronous behavior. We

conﬁrmed these previous observations for the speciﬁc macrophage incorporated additional donor samples and LPS time points to inde-

populations used in the TLDA analysis. GM-CSF M␾ displayed pendently validate LPS-regulated expression of individual GPCRs

enhanced LPS-induced levels of secreted IL-6 (Fig. 3A) and TNF (P2RY5, LGR4) (Fig. 5D, E), including examples of LPS-regulated

(Fig. 3B), as well as reduced levels of secreted IL-10 (Fig. 3C), by genes that were substantially affected by differentiation status

comparison to CSF-1 M␾. Similarly, LPS-induced mRNA levels of IL- (EDNRB, GPRC5A) (Fig. 5F, G). The latter are likely to contribute to

6 and IL-12p40 were increased in GM-CSF M␾ (Fig. 3D, E), although differential inﬂammatory proﬁles of CSF-1 M␾ versus GM-CSF M␾.

D.M. Hohenhaus et al. / Immunobiology 218 (2013) 1345–1353 1349

Fig. 2. GPCR mRNA regulation during macrophage differentiation. (A) Venn diagram showing statistically signiﬁcant differential expression of GPCR-encoding genes during

human monocyte to macrophage differentiation (for both CSF-1 M␾ and GM-CSF M␾, as well as CSF-1 M␾ versus GM-CSF M␾). GPCRs showing 5 fold difference in gene

≥

expression for CSF-1 M␾ versus donor matched monocytes (B) and GM-CSF M␾ versus donor matched monocytes (C) (log10 scale). (D) Independent qPCR validation (SYBR

Green) of upregulated LGR4 mRNA expression during human monocyte to macrophage differentiation (mean of 5 independent donors + SEM).

The elevated expression of EDNRB in CSF-1 M␾, as well as downreg- macrophages (M1: IFN-␥ plus LPS; M2: IL-4). Although these con-

ulation by LPS in both macrophage populations, was also conﬁrmed ditions were not directly comparable to our own, using TLDA we

at the protein level (Fig. 5H). identiﬁed 119 GPCRs that were differentially expressed during

human macrophage differentiation and/or activation. Several of

Discussion these GPCRs (e.g. MRGPRF, LGR4, GPR143) were not identiﬁed in

the former study. More recently, Lacey et al. (2012) used microar-

rays to identify differential gene expression programs in human

Previous studies have used microarray expression proﬁling to

and mouse CSF-1 M␾ versus GM-CSF M␾. As with our recent anal-

provide important insights into gene regulation and biological

ysis of LPS-regulated gene expression (Schroder et al., 2012), this

responses during differentiation of primary human monocytes to

study found substantial divergence at the individual gene level

macrophages and/or macrophage polarization (Biswas et al., 2006;

when comparing mouse versus human macrophage populations.

Irvine et al., 2009; Lacey et al., 2012; Li et al., 2007; Liu et al.,

Although Lacey et al. (2012) did not analyze LPS-regulated gene

2008; Martinez et al., 2006; Sierra-Filardi et al., 2010). Nonetheless,

expression in their microarray analysis, they did identify several

issues such as sensitivity for weakly expressed genes, incomplete

GPCRs differentially expressed between CSF-1 M␾ and GM-CSF

transcriptome coverage and probe failure have limited the capac-

M␾. The data on these GPCRs were generally consistent with our

ity of microarrays to capture truly global gene expression data. RNA

own. Our study also complements a very recent study that used

sequencing can potentially overcome such issues, as can TLDA for

TLDA to analyze ion channel and GPCR expression in human alve-

analysis of individual gene families. Martinez et al. (2006) used

olar macrophages versus various other macrophage populations

microarrays to identify 53 GPCRs differentially expressed between

(Groot-Kormelink et al., 2012). The primary aim of that study was

monocytes, differentiated cells (3 day, 7 day) and/or polarized

1350 D.M. Hohenhaus et al. / Immunobiology 218 (2013) 1345–1353

Fig. 3. Inﬂammatory responses are elevated in GM-CSF M␾ versus CSF-1 M␾. (A–C) Secreted LPS-induced cytokine production (A: IL-6; B: TNF; C: IL-10) from CSF-1 M␾ and

GM-CSF M␾ used in TLDA analysis (mean of 3 independent donors + SEM; *P < 0.05, **P < 0.01). (D–H) LPS-regulated mRNA expression for genes encoding cytokines in CSF-1

M␾ and GM-CSF M␾ used in TLDA analysis (D: IL6; E: IL12p40; F: TNF; G: IL10; H: CCL2) (mean of 3 independent donors + SEM; *P < 0.05, **P < 0.01, ***P < 0.001).

Fig. 4. GPCR mRNA regulation during macrophage polarization. (A) GPCRs showing 5 fold difference in gene expression for CSF-1 M␾ versus donor matched GM-CSF M␾

≥

(log10 scale). (B–E) Independent qPCR validation (SYBR Green) of differential EDNRB (B), EDG1 (C), HRH4 (D) and EDG6 (E) mRNA expression between CSF-1 M␾ and GM-CSF

M␾ (mean of 5 independent donors + SEM).

D.M. Hohenhaus et al. / Immunobiology 218 (2013) 1345–1353 1351

to determine the validity of various cellular models as surrogates for

alveolar macrophages, so it did not compare proﬁles of CSF-1 M␾

versus GM-CSF M␾, nor did it examine LPS-regulated GPCR expres-

sion. Thus, our data provide a new resource of GPCR expression

patterns during human macrophage differentiation and activation.

In addition to identifying broad trends such as downregulation of a

substantial set of GPCRs during differentiation from monocytes to

macrophages, our approach has identiﬁed sets of GPCRs that were

(1) upregulated during macrophage differentiation; (2) differen-

tially expressed between unstimulated CSF-1 M␾ and GM-CSF M␾;

(3) LPS-regulated; and (4) LPS-regulated and impacted upon by the

differentiating agent.

Although the regulation and function of individual GPCRs, par-

ticularly chemokine receptors, during monocyte to macrophage

differentiation has been widely studied, our analysis captured sev-

eral others that have not been extensively studied in the context

of macrophage biology or inﬂammation. For example, LGR4 mRNA

was upregulated in both CSF-1 M␾ and GM-CSF M␾, as com-

pared to monocytes (Fig. 2D). The secreted R-spondin proteins

are LGR4 ligands that potentiate Wnt/␤-catenin signaling (Carmon

et al., 2011; Glinka et al., 2011; Ruffner et al., 2012). Some stud-

ies have shown that Wnt signaling, acting via Frizzled receptors,

regulates macrophage inﬂammatory responses (Blumenthal et al.,

2006; Schaale et al., 2011). However, to our knowledge, no previ-

ous studies have examined LGR4 functions in this cell type. LGR4

is essential for several developmental processes, including epithe-

lial cell proliferation and paneth cell differentiation in the intestine

(Mustata et al., 2011). This requirement is likely to underpin the

increased susceptibility and severity of Lgr4− − mice to dextran sul-

fate sodium-induced inﬂammatory bowel disease (Liu et al., 2013).

Our data suggest that LGR4 expression in macrophages may also

act to modulate inﬂammatory pathways. As with LGR4, MRGPRF

was strongly upregulated in CSF-1 M␾ and GM-CSF M␾ versus

monocytes (115.7 fold and 422.4 fold, respectively; Fig. 2B, C, Sup-

plementary data). This gene belongs to the MAS-related GPCRs that

are expressed in sensory neurons (Dong et al., 2001) and have been

linked to inhibition of pathological pain (Guan et al., 2010). MRG-

PRF has yet to be deorphanized, and very little is known about

the regulation or function of this speciﬁc family member. One

study showed that its expression was signiﬁcantly downregulated

in a mouse model of intestinal inﬂammation (Avula et al., 2011).

Similarly, we observed that in addition to its robust upregulation

during monocyte to macrophage differentiation, LPS downregu-

lated its expression in macrophages (Supplementary data). We are

not aware of any studies examining MRGPRF expression and/or

function in macrophages. Other genes encoding GPCRs that were

upregulated during macrophage differentiation and have not been

studied in this context include GPR143, GPR125 and GPR85, amongst

others (Fig. 2B, C).

Given the differing inﬂammatory cytokine proﬁles of CSF-

1 M␾ and GM-CSF M␾, GPCRs showing differential expression

between these two cell populations have the potential to mod-

ulate inﬂammatory pathways. We identiﬁed 15 such GPCRs in

the basal state (Fig. 2A), as well as many others when comparing

across an LPS time course (Fig. 5A). The set of GPCRs associated

with elevated expression in unstimulated CSF-1 M␾ versus GM-CSF

M␾ generally reﬂects downregulated expression during GM-CSF-

mediated differentiation of monocytes into macrophages. One

HTR2B

Fig. 5. The inﬂuence of differentiating agent on LPS-regulated GPCR mRNA striking exception to this was that was weakly expressed

expression. (A) Venn diagram showing statistically signiﬁcant LPS-regulated by both monocytes and GM-CSF M␾, but was upregulated during

GPCR-encoding genes in CSF-1 M␾ and/or GM-CSF M␾, including a comparison

of differentiation status versus LPS regulation. All genes indicated are statistically

signiﬁcant when considering the entire proﬁle (con, 1 h LPS, 4 h LPS, 24 h LPS).

Semi-supervised hierarchical clustering of all statistically signiﬁcant LPS-regulated

of LPS-regulated genes, including those affected by differentiation status (D: LGR4; E:

GPCR-encoding genes affected by differentiation status (B) and all statistically

P2RY5; F: EDNRB; G: GPRC5a) (mean of 5 independent donors + SEM). (H) EDNRB pro-

signiﬁcant genes showing asynchronous behavior when comparing differentiation

tein expression levels, as assessed by immunoblotting, over a 48 h LPS time course

status versus LPS regulation (C). (D–G) Independent qPCR validation (SYBR Green)

in CSF-1 M␾ and GM-CSF M␾. GAPDH protein levels are shown as a loading control.

1352 D.M. Hohenhaus et al. / Immunobiology 218 (2013) 1345–1353

differentiation with CSF-1 (Supplementary data). This GPCR, a Acknowledgments

receptor for serotonin, was recently shown to be selectively

expressed by M2 macrophages and to skew macrophage polar- This work was funded by an Australian National Health and

ization toward this phenotype (de las Casas-Engel et al., 2013). Medical Research Council (NHMRC) Project Grant (ID APP1030169).

Conversely, HRH4, which encodes a histamine receptor, was essen- MJS is supported by an ARC Future Fellowship (FT100100657)

tially undetectable in monocytes and CSF-1 M␾, but was robustly and an honorary Australian NHMRC Senior Research Fellowship

upregulated during differentiation with GM-CSF (Fig. 4D, Supple- (1003470). We thank Dr. Stephen Rudd from the Queensland Facil-

mentary data). Consistent with this, Simon et al. (2012) reported ity for Advanced Bioinformatics for useful discussions relating to

that CD1a dendritic cells, differentiated from human monocytes this project.

with GM-CSF plus IL-4, expressed this receptor. There has been

intense interest in HRH4 as an inﬂammation target (Walter et al.,

Appendix A. Supplementary data

2011), and a recent study linked this receptor to LPS-induced

TNF production from macrophages in vivo (Cowden et al., 2013).

Supplementary data associated with this article can be

Several other GPCRs that were selectively enriched in CSF-1 M␾

found, in the online version, at http://dx.doi.org/10.1016/j.imbio.

(e.g. P2RY8, GPR92) or GM-CSF M␾ (e.g. EMR3) have not previ-

2013.07.001.

ously been linked to polarized macrophage phenotypes (unlike

HTR2B and HRH4), and clearly warrant further investigation in this

context. References

Not surprisingly, LPS regulated the expression of a substantial

Alexander, S.P., Mathie, A., Peters, J.A., 2011. Guide to receptors and channels (GRAC),

number of genes encoding GPCRs over a 24 h time course (101

5th edition. Br. J. Pharmacol. 164 (Suppl. 1), S1–S324.

genes). Many of these were also regulated during differentiation Avula, L.R., Buckinx, R., Alpaerts, K., Costagliola, A., Adriaensen, D., Van Nassauw,

L., Timmermans, J.P., 2011. The effect of inﬂammation on the expression and

or polarization of macrophages. For example, LGR4 and MRGPRF

distribution of the MAS-related gene receptors MrgE and MrgF in the murine

were upregulated during monocyte to macrophage differentia-

ileum. Histochem. Cell. Biol. 136, 569–585.

tion (Fig. 2B–D), and were also downregulated by LPS in HMDM Barta, P., Van Pelt, C., Men, T., Dickey, B.F., Lotan, R., Moghaddam, S.J., 2012. Enhance-

(Fig. 5D, Supplementary data). Similarly, EDNRB was expressed at ment of lung tumorigenesis in a Gprc5a Knockout mouse by chronic extrinsic

airway inﬂammation. Mol. Cancer 11, 4.

elevated levels in unstimulated CSF-1 M␾ as compared to GM-

Benjamini, Y., Hochberg, Y., 1995. Controlling the false discovery rate: a practical

CSF M␾ (Fig. 4B), and was also downregulated by LPS in both cell and powerful approach to multiple testing. J. R. Stat. Soc. B 57, 289–300.

populations (Fig. 5F, H). Again, numerous GPCR-encoding genes Biswas, S.K., Chittezhath, M., Shalova, I.N., Lim, J.Y., 2012. Macrophage polarization

and plasticity in health and disease. Immunol. Res. 53, 11–24.

upregulated by LPS (e.g. GPR31, GPR64) or downregulated by LPS

Biswas, S.K., Gangi, L., Paul, S., Schioppa, T., Saccani, A., Sironi, M., Bottazzi, B., Doni,

GPR126 GPR143

(e.g. , ) are yet to be studied in macrophage biology A., Vincenzo, B., Pasqualini, F., Vago, L., Nebuloni, M., Mantovani, A., Sica, A., 2006.

and/or TLR signaling. These represent obvious candidates for future A distinct and unique transcriptional program expressed by tumor-associated

macrophages (defective NF-kappaB and enhanced IRF-3/STAT1 activation).

functional analysis.

Blood 107, 2112–2122.

One clear signature arising from our proﬁling data was that

Blumenthal, A., Ehlers, S., Lauber, J., Buer, J., Lange, C., Goldmann, T., Heine, H., Brandt,

divergent regulation of closely related family members was com- E., Reiling, N., 2006. The Wingless homolog WNT5A and its receptor Frizzled-

5 regulate inﬂammatory responses of human mononuclear cells induced by

mon. For example, EDG1 and EDG6, which both encode GPCRs

microbial stimulation. Blood 108, 965–973.

that sense sphingosine-1-phosphate, showed opposing regula-

Bozinovski, S., Jones, J.E., Vlahos, R., Hamilton, J.A., Anderson, G.P., 2002.

tion in CSF-1 M␾ versus GM-CSF M␾ (Fig. 4C, E). Similarly, LGR4 Granulocyte/macrophage-colony-stimulating factor (GM-CSF) regulates lung

innate immunity to lipopolysaccharide through Akt/Erk activation of NFkappa

and LGR7 were downregulated and upregulated by LPS, respec-

B and AP-1 in vivo. J. Biol. Chem. 277, 42808–42814.

tively (Fig. 5D, Supplementary data), and divergent LPS proﬁles

Carmon, K.S., Gong, X., Lin, Q., Thomas, A., Liu, Q., 2011. R-spondins function as

were also apparent for ADORA2A versus ADORA2B (Supplemen- ligands of the orphan receptors LGR4 and LGR5 to regulate Wnt/beta-catenin

tary data), and GPCR5A versus GPCR5B (Fig. 5G, Supplementary signaling. Proc. Natl. Acad. Sci. U.S.A. 108, 11452–11457.

Cowden, J.M., Yu, F., Challapalli, M., Huang, J.F., Kim, S., Fung-Leung, W.P., Ma, J.Y.,

data). The functional signiﬁcance of such opposing regulation is

Riley, J.P., Zhang, M., Dunford, P.J., Thurmond, R.L., 2013. Antagonism of the his-

not always clear, but it is tempting to speculate that coordinated tamine H receptor reduces LPS-induced TNF production in vivo. Inﬂamm. Res.,

changes in expression of related GPCRs may provide a molecu- http://dx.doi.org/10.1007/s00011-013-0612-5.

de las Casas-Engel, M., Dominguez-Soto, A., Sierra-Filardi, E., Bragado, R.,

lar switch to control macrophage inﬂammatory pathways. At least

Nieto, C., Puig-Kroger, A., Samaniego, R., Loza, M., Corcuera, M.T., Gomez-

in the case of GPRC5A and GPRC5B, emerging evidence supports Aguado, F., Bustos, M., Sanchez-Mateos, P., Corbi, A.L., 2013. Serotonin skews

this possibility. Gprc5a− − mice develop exacerbated lung inﬂam- human macrophage polarization through HTR2B and HTR7. J. Immunol. 190,

2301–2310.

mation that is accompanied by ampliﬁed NF-␬B activation and

Deng, J., Fujimoto, J., Ye, X.F., Men, T.Y., Van Pelt, C.S., Chen, Y.L., Lin, X.F., Kadara, H.,

pro-inﬂammatory cytokine production (Barta et al., 2012; Deng

Tao, Q., Lotan, D., Lotan, R., 2010. Knockout of the tumor suppressor gene Gprc5a

et al., 2010), whereas Gprc5b− − mice were protected from diet- in mice leads to NF-kappaB activation in airway epithelium and promotes lung

inﬂammation and tumorigenesis. Cancer Prev. Res. (Phila) 3, 424–437.

induced obesity, and displayed reduced inﬂammatory cytokine

Dong, X., Han, S., Zylka, M.J., Simon, M.I., Anderson, D.J., 2001. A diverse family of

production and NF-␬B activation (Kim et al., 2012). For this gene

GPCRs expressed in speciﬁc subsets of nociceptive sensory neurons. Cell 106,

pair, LPS regulation was much more striking in CSF-1 M␾ as com- 619–632.

Fleetwood, A.J., Lawrence, T., Hamilton, J.A., Cook, A.D., 2007. Granulocyte-

pared to GM-CSF M␾, with GPRC5A being induced and GPRC5B

macrophage colony-stimulating factor (CSF) and macrophage CSF-dependent

repressed in CSF-1 M␾ (Fig. 5G, Supplementary data). On the

macrophage phenotypes display differences in cytokine proﬁles and transcrip-

basis of the above knock-out mouse studies, this regulation would tion factor activities: implications for CSF blockade in inﬂammation. J. Immunol.

be predicted to constrain LPS-induced inﬂammatory responses in 178, 5245–5252.

Fredriksson, R., Lagerstrom, M.C., Lundin, L.G., Schioth, H.B., 2003. The G-protein-

CSF-1 M␾, whereas this control mechanism would not exist in

coupled receptors in the human genome form ﬁve main families. Phylogenetic

GM-CSF M␾. Hence, selective GPRC5A and GPRC5B regulation in analysis, paralogon groups, and ﬁngerprints. Mol. Pharmacol. 63, 1256–1272.

CSF-1 M␾ hints that these receptors may contribute to skewing Glinka, A., Dolde, C., Kirsch, N., Huang, Y.L., Kazanskaya, O., Ingelﬁnger, D., Boutros,

M., Cruciat, C.M., Niehrs, C., 2011. LGR4 and LGR5 are R-spondin receptors medi-

of M2 M␾ polarization. This possibility warrants further investi-

ating Wnt/beta-catenin and Wnt/PCP signalling. EMBO Rep. 12, 1055–1061.

gation. In summary, our detailed expression analysis of the GPCR Groot-Kormelink, P.J., Fawcett, L., Wright, P.D., Gosling, M., Kent, T.C., 2012. Quanti-

repertoire during monocyte to macrophage differentiation and tative GPCR and ion channel transcriptomics in primary alveolar macrophages

and macrophage surrogates. BMC Immunol. 13, 57.

LPS-mediated activation provides a wealth of candidate GPCRs as

Guan, Y., Liu, Q., Tang, Z., Raja, S.N., Anderson, D.J., Dong, X., 2010. Mas-related G-

uncharacterized, but potentially targetable, regulators of inﬂam-

protein-coupled receptors inhibit pathological pain in mice. Proc. Natl. Acad. Sci. mation. U.S.A. 107, 15933–15938.

D.M. Hohenhaus et al. / Immunobiology 218 (2013) 1345–1353 1353

Hamilton, J.A., 2008. Colony-stimulating factors in inﬂammation and autoimmunity. Schroder, K., Hertzog, P.J., Ravasi, T., Hume, D.A., 2004. Interferon-gamma:

Nat. Rev. Immunol. 8, 533–544. an overview of signals, mechanisms and functions. J. Leukoc. Biol. 75,

Hamilton, J.A., Achuthan, A., 2013. Colony stimulating factors and myeloid cell biol- 163–189.

ogy in health and disease. Trends Immunol. 34, 81–89. Schroder, K., Irvine, K.M., Taylor, M.S., Bokil, N.J., Le Cao, K.A., Masterman, K.A., Labzin,

Hamilton, J.A., Anderson, G.P., 2004. GM-CSF biology. Growth Factors 22, 225–231. L.I., Semple, C.A., Kapetanovic, R., Fairbairn, L., Akalin, A., Faulkner, G.J., Baillie,

Hume, D.A., 2008. Differentiation and heterogeneity in the mononuclear phagocyte J.K., Gongora, M., Daub, C.O., Kawaji, H., McLachlan, G.J., Goldman, N., Grimmond,

system. Mucosal Immunol. 1, 432–441. S.M., Carninci, P., Suzuki, H., Hayashizaki, Y., Lenhard, B., Hume, D.A., Sweet,

Ingersoll, M.A., Spanbroek, R., Lottaz, C., Gautier, E.L., Frankenberger, M., Hoffmann, M.J., 2012. Conservation and divergence in Toll-like receptor 4-regulated gene

R., Lang, R., Haniffa, M., Collin, M., Tacke, F., Habenicht, A.J., Ziegler-Heitbrock, L., expression in primary human versus mouse macrophages. Proc. Natl. Acad. Sci.

Randolph, G.J., 2010. Comparison of gene expression proﬁles between human U.S.A. 109, E944–E953.

and mouse monocyte subsets. Blood 115, e10–e19. Schulz, C., Gomez Perdiguero, E., Chorro, L., Szabo-Rogers, H., Cagnard, N., Kierdorf,

Irvine, K.M., Andrews, M.R., Fernandez-Rojo, M.A., Schroder, K., Burns, C.J., Su, S., K., Prinz, M., Wu, B., Jacobsen, S.E., Pollard, J.W., Frampton, J., Liu, K.J., Geissmann,

Wilks, A.F., Parton, R.G., Hume, D.A., Sweet, M.J., 2009. Colony-stimulating factor- F., 2012. A lineage of myeloid cells independent of Myb and hematopoietic stem

1 (CSF-1) delivers a proatherogenic signal to human macrophages. J. Leukoc. Biol. cells. Science 336, 86–90.

85, 278–288. Seok, J., Warren, H.S., Cuenca, A.G., Mindrinos, M.N., Baker, H.V., Xu, W., Richards,

Kim, Y.J., Sano, T., Nabetani, T., Asano, Y., Hirabayashi, Y., 2012. GPRC5B activates D.R., McDonald-Smith, G.P., Gao, H., Hennessy, L., Finnerty, C.C., Lopez, C.M.,

obesity-associated inﬂammatory signaling in adipocytes. Sci. Signal. 5, ra85. Honari, S., Moore, E.E., Minei, J.P., Cuschieri, J., Bankey, P.E., Johnson, J.L., Sperry,

Lacey, D.C., Achuthan, A., Fleetwood, A.J., Dinh, H., Roiniotis, J., Scholz, G.M., J., Nathens, A.B., Billiar, T.R., West, M.A., Jeschke, M.G., Klein, M.B., Gamelli, R.L.,

Chang, M.W., Beckman, S.K., Cook, A.D., Hamilton, J.A., 2012. Deﬁning GM-CSF- Gibran, N.S., Brownstein, B.H., Miller-Graziano, C., Calvano, S.E., Mason, P.H.,

and macrophage-CSF-dependent macrophage responses by in vitro models. J. Cobb, J.P., Rahme, L.G., Lowry, S.F., Maier, R.V., Moldawer, L.L., Herndon, D.N.,

Immunol. 188, 5752–5765. Davis, R.W., Xiao, W., Tompkins, R.G., 2013. Genomic responses in mouse models

Lattin, J.E., Schroder, K., Su, A.I., Walker, J.R., Zhang, J., Wiltshire, T., Saijo, K., Glass, poorly mimic human inﬂammatory diseases. Proc. Natl. Acad. Sci. U.S.A. 110,

C.K., Hume, D.A., Kellie, S., Sweet, M.J., 2008. Expression analysis of G protein- 3507–3512.

coupled receptors in mouse macrophages. Immunome Res. 4, 5. Sierra-Filardi, E., Vega, M.A., Sanchez-Mateos, P., Corbi, A.L., Puig-Kroger, A., 2010.

Li, J., Pritchard, D.K., Wang, X., Park, D.R., Bumgarner, R.E., Schwartz, S.M., Liles, W.C., Heme oxygenase-1 expression in M-CSF-polarized M2 macrophages contributes

2007. cDNA microarray analysis reveals fundamental differences in the expres- to LPS-induced IL-10 release. Immunobiology 215, 788–795.

sion proﬁles of primary human monocytes, monocyte-derived macrophages, Simon, T., Gogolak, P., Kis-Toth, K., Jelinek, I., Laszlo, V., Rajnavolgyi, E., 2012. His-

and alveolar macrophages. J. Leukoc. Biol. 81, 328–335. tamine modulates multiple functional activities of monocyte-derived dendritic

Liquet, B., Le Cao, K.A., Hocini, H., Thiebaut, R., 2012. A novel approach for biomarker cell subsets via histamine receptor 2. Int. Immunol. 24, 107–116.

selection and the integration of repeated measures experiments from two Stewart, C.R., Stuart, L.M., Wilkinson, K., van Gils, J.M., Deng, J., Halle, A., Rayner, K.J.,

assays. BMC Bioinform. 13, 325. Boyer, L., Zhong, R., Frazier, W.A., Lacy-Hulbert, A., El Khoury, J., Golenbock, D.T.,

Liu, H., Shi, B., Huang, C.C., Eksarko, P., Pope, R.M., 2008. Transcriptional diversity Moore, K.J., 2010. CD36 ligands promote sterile inﬂammation through assembly

during monocyte to macrophage differentiation. Immunol. Lett. 117, 70–80. of a Toll-like receptor 4 and 6 heterodimer. Nat. Immunol. 11, 155–161.

Liu, S., Qian, Y., Li, L., Wei, G., Guan, Y., Pan, H., Guan, X., Zhang, L., Lu, X., Zhao, Y., Venkatakrishnan, A.J., Deupi, X., Lebon, G., Tate, C.G., Schertler, G.F., Babu,

Liu, M., Li, D., 2013. Lgr4 gene deﬁciency increases susceptibility and severity M.M., 2013. Molecular signatures of G-protein-coupled receptors. Nature 494,

of dextran sodium sulfate-induced inﬂammatory bowel disease in mice. J. Biol. 185–194.

Chem. 288, 8794–8803. Verreck, F.A., de Boer, T., Langenberg, D.M., Hoeve, M.A., Kramer, M., Vaisberg,

Martinez, F.O., Gordon, S., Locati, M., Mantovani, A., 2006. Transcriptional proﬁling E., Kastelein, R., Kolk, A., de Waal-Malefyt, R., Ottenhoff, T.H., 2004. Human

of the human monocyte-to-macrophage differentiation and polarization: new IL-23-producing type 1 macrophages promote but IL-10-producing type 2

molecules and patterns of gene expression. J. Immunol. 177, 7303–7311. macrophages subvert immunity to (myco)bacteria. Proc. Natl. Acad. Sci. U.S.A.

Mustata, R.C., Van Loy, T., Lefort, A., Libert, F., Strollo, S., Vassart, G., Garcia, M.I., 2011. 101, 4560–4565.

Lgr4 is required for Paneth cell differentiation and maintenance of intestinal Walter, M., Kottke, T., Stark, H., 2011. The histamine H(4) receptor: targeting inﬂam-

stem cells ex vivo. EMBO Rep. 12, 558–564. matory disorders. Eur. J. Pharmacol. 668, 1–5.

Overington, J.P., Al-Lazikani, B., Hopkins, A.L., 2006. How many drug targets are Wang, Y., Szretter, K.J., Vermi, W., Gilﬁllan, S., Rossini, C., Cella, M., Barrow, A.D.,

there? Nat. Rev. Drug Discov. 5, 993–996. Diamond, M.S., Colonna, M., 2012. IL-34 is a tissue-restricted ligand of CSF1R

Rossol, M., Heine, H., Meusch, U., Quandt, D., Klein, C., Sweet, M.J., Hauschildt, S., required for the development of Langerhans cells and microglia. Nat. Immunol.

2011. LPS-induced cytokine production in human monocytes and macrophages. 13, 753–760.

Crit. Rev. Immunol. 31, 379–446. Wei, S., Nandi, S., Chitu, V., Yeung, Y.G., Yu, W., Huang, M., Williams, L.T., Lin, H.,

Ruffner, H., Sprunger, J., Charlat, O., Leighton-Davies, J., Grosshans, B., Salathe, A., Stanley, E.R., 2010. Functional overlap but differential expression of CSF-1 and

Zietzling, S., Beck, V., Therier, M., Isken, A., Xie, Y., Zhang, Y., Hao, H., Shi, X., Liu, D., IL-34 in their CSF-1 receptor-mediated regulation of myeloid cells. J. Leukoc.

Song, Q., Clay, I., Hintzen, G., Tchorz, J., Bouchez, L.C., Michaud, G., Finan, P., Myer, Biol. 88, 495–505.

V.E., Bouwmeester, T., Porter, J., Hild, M., Bassilana, F., Parker, C.N., Cong, F., 2012. Yang, H., Hreggvidsdottir, H.S., Palmblad, K., Wang, H., Ochani, M., Li, J., Lu, B., Cha-

R-Spondin potentiates Wnt/beta-catenin signaling through orphan receptors van, S., Rosas-Ballina, M., Al-Abed, Y., Akira, S., Bierhaus, A., Erlandsson-Harris,

LGR4 and LGR5. PLoS One 7, e40976. H., Andersson, U., Tracey, K.J., 2010. A critical cysteine is required for HMGB1

Sade-Feldman, M., Kanterman, J., Ish-Shalom, E., Elnekave, M., Horwitz, E., Baniyash, binding to Toll-like receptor 4 and activation of macrophage cytokine release.

M., 2013. Tumor necrosis factor-alpha blocks differentiation and enhances Proc. Natl. Acad. Sci. U.S.A. 107, 11942–11947.

suppressive activity of immature myeloid cells during chronic inﬂammation. Yoshida, H., Hayashi, S., Kunisada, T., Ogawa, M., Nishikawa, S., Okamura, H., Sudo,

Immunity 38, 541–554. T., Shultz, L.D., Nishikawa, S., 1990. The murine mutation osteopetrosis is in the

Schaale, K., Neumann, J., Schneider, D., Ehlers, S., Reiling, N., 2011. Wnt signaling in coding region of the macrophage colony stimulating factor gene. Nature 345,

macrophages: augmenting and inhibiting mycobacteria-induced inﬂammatory 442–444.

responses. Eur. J. Cell. Biol. 90, 553–559. Appendix VI

Junxian Lim, Abishek Iyer, Jacky Y. Suen, Vernon Seow, Robert C. Reid, Lindsay Brown, and David P. Fairlie. C5aR and C3aR antagonists each inhibit diet-induced obesity, metabolic dysfunction, and adipocyte and macrophage signaling. FASEB J. (2013) 27: 822-831.

The FASEB Journal • Research Communication

C5aR and C3aR antagonists each inhibit diet-induced obesity, metabolic dysfunction, and adipocyte and macrophage signaling

Junxian Lim,*,1 Abishek Iyer,*,1 Jacky Y. Suen,* Vernon Seow,* Robert C. Reid,* Lindsay Brown,†,2 and David P. Fairlie*,2 *Institute for Molecular Bioscience, The University of Queensland, Brisbane, Queensland, Australia; and †Faculty of Sciences, University of Southern Queensland, Toowoomba, Queensland, Australia

ABSTRACT Mammalian survival depends on metab- Modern diets high in carbohydrates and saturated olizing nutrients, storing energy, and combating infec- fats are producing a global epidemic in obesity, type tion. Complement activation in blood triggers energy- II diabetes, and cardiovascular disease. Metabolic depleting immune responses to fight infections. Here dysfunction predisposes people to these conditions we identify surprising energy-conserving roles for com- and is characterized by abdominal obesity, glucose plement proteins C5a and C3a and their receptors, and insulin intolerance, elevated plasma triglycerides C5aR and C3aR, roles that are contraindicated in and cholesterol, and liver and cardiovascular abnor- complement biology. Rats fed a high-carbohydrate malities (1, 2). Obesity and metabolic dysfunction high-fat diet developed obesity, visceral adiposity, adi- are now associated with a state of chronic low-grade pose inflammation, glucose/insulin intolerance, and inflammation, but the nature and importance of this cardiovascular dysfunction that correlated with in- association remain unclear (3–5). Energy deficiency creased plasma C3a, adipose C5aR, and C3aR. These in through malnutrition or starvation impairs immune vivo changes were dramatically attenuated by receptor- responses in mammals, while nutrient overload in- selective antagonists of either C5aR (5 mg/kg/d p.o.) duces inflammatory responses through cellular stress or C3aR (30 mg/kg/d p.o.), which both reduced pro- in mitochondria, endoplasmic reticulum, and other inflammatory adipokines and altered expression of organelles, with chronic inflammatory diseases often inflammatory genes in adipose tissue. In vitro C5a and associated with premature or more severe metabolic C3a (100 nM) exhibited novel insulin-like effects on dysfunction (6). Metabolic defects, such as insulin 3T3-L1 adipocytes, promoting energy conservation by resistance in muscle and lipid accumulation in liver, increasing glucose and fatty acid uptake while inhibit- can occur even in preobese states without obvious ing cAMP signaling and lipolysis, and induced PGE2 systemic inflammation (7); whereas a single high- release from macrophages, effects all blocked by each carbohydrate, high-fat meal induces oxidative and respective antagonist (10 ␮M). These studies reveal inflammatory responses in healthy lean people (8). important new links between complement signaling Obesity and metabolic dysfunction are also induced and metabolism, highlight new complement func- by altered nutrient sensing, gut microbiota, immune tions on adipocytes and in adipose tissue, demon- networks, and metabolism of fatty acids in adipose strate how aberrant immune responses may exacer- tissue (9, 10). These observations suggest that energy bate obesity and metabolic dysfunction, and show homeostasis may be regulated and coordinated by that targeting C3aR or C5aR with antagonists is a new signaling molecules and pathways that are common strategy for treating metabolic dysfunction.—Lim, J., to those in inflammatory networks. Inflammatory Iyer, A., Suen, J. Y., Seow, V., Reid, R. C., Brown, L., mediators that make important contributions to in- Fairlie, D. P. C5aR and C3aR antagonists each inhibit nate and adaptive immunity are the complement diet-induced obesity, metabolic dysfunction, and adi- proteins, an ancient network of plasma proteins that pocyte and macrophage signaling. FASEB J. 27, effect recognition, opsonization, destruction, and 822–831 (2013). www.fasebj.org removal of pathogens and infected or damaged cells. Key Words: adipose inflammation ⅐ complement ⅐ GPCR ⅐ metabolism ⅐ immunity 1 These authors contributed equally to this work. 2 Correspondence: Institute for Molecular Bioscience, The University of Queensland, Brisbane, QLD 4072, Australia. Abbreviations: C3aR, C3a receptor; C3aRa, C3a receptor an- E-mail: D.P.F., [email protected]; L.B., lindsay.brown@ tagonist; C5aR, C5a receptor; C5aRa, C5a receptor antagonist; usq.edu.au cAMP, cyclic adenosine monophosphate; CS, cornstarch; FBS, doi: 10.1096/fj.12-220582 fetal bovine serum; GPCR, G-protein-coupled receptor; HCHF, This article includes supplemental data. Please visit http:// high-carbohydrate high-fat; PGE2,prostaglandinE2 www.fasebj.org to obtain this information.

822 0892-6638/13/0027-0822 © FASEB Activating the complement network triggers immune MATERIALS AND METHODS responses that use substantial amounts of energy to fight infection. Animal studies Among products of complement activation are the proteins C3a and C5a that bind to specific G-protein- Male Wistar rats were bred at The University of Queensland coupled receptors (GPCRs; C3aR and C5aR) on Biological Resources Facility. All experimental procedures surfaces of immune cells and induce chemotaxis, were approved by the Animal Experimental Ethics Commit- tee of The University of Queensland, under the guidelines of inflammatory signaling, and immunological defense the National Health and Medical Council of Australia. The against infection and injury (11). These plasma pro- 8–9 wk old male Wistar rats were provided ad libitum access to teins have also recently been found to have nonim- either a cornstarch (CS) or a high-carbohydrate, high-fat munological roles, such as in hematopoiesis, tissue (HCHF) diet for 16 wk. The CS and HCHF diets were regeneration, angiogenesis, and lipid metabolism described previously (20). HCHF-group rats further received (12). Complement factors B, H, and C3 are also drinking water supplemented with 25% fructose, while the CS group received normal drinking water for the 16 wk (20). secreted by adipose tissue (13, 14), and plasma C3 C5aR antagonist (C5aRa; 5 mg/kg/d dissolved in 10% etha- concentrations have been reported to correlate with nol/distilled water), also known as Ac-(cyclo-2,6)-Phe-[Orn- obesity, type II diabetes, cardiovascular disease, and Pro-dCha-Trp-Arg] or 3D53 (21, 22), and C3aR antagonist postprandial serum triglycerides (15–17). C3 is pro- (C3aRa; 30 mg/kg/d dissolved in 10% ethanol/distilled teolytically cleaved to the chemotactic proinflamma- water), also known as N2-[(2,2-diphenylethoxy)acetyl]-l-argi- tory protein C3a and then to a derivative C3a-desArg, nine or SB290157 (23) were administered daily by oral gavage to HCHF-treated male Wistar rats, starting after wk 8 of the which has been controversially linked to adipose study protocol. Control HCHF-fed rats received equal stimulation and C5L2 binding (18, 19). Neither C3a amounts of 10% ethanol/distilled water as oral gavage vehi- nor C5a has been reported to regulate adipocytes, cle. Dual energy X-ray absorptiometric measurements were adipose function, lipid homeostasis, or metabolic performed as described previously (20). Oral glucose toler- dysfunction, and no C3aR or C5aR antagonists have ance (2 g/kg body weight) and insulin tolerance tests (0.3 been examined in metabolic syndrome. IU/kg body weight) were performed as described previously Here we report novel discoveries for receptor- (20). Plasma lipids and enzymes concentrations were measured by the Veterinary Pathology Service of The University of selective antagonists of C3aR and C5aR, using these Queensland. Systolic blood pressure (24) and changes in compounds to identify and characterize important cardiovascular structure and function (20, 24) were made as previously unknown roles for C3a and C5a in lipid/ described after 16 wk. Collagen analysis was performed in carbohydrate metabolism and energy homeostasis, excised perfused hearts using picrosirius red staining as attenuate diet-induced obesity and metabolic dys- described previously (20, 24). function in rats, and dissect mechanisms by which these complement proteins modulate adipocyte and Cell studies adipose function. Obesity, visceral adiposity, and metabolic dysfunction were induced in rats fed for 16 The 3T3-L1 [European Collection of Cell Cultures (ECACC); Sigma-Aldrich, Castle Hill, NSW, Australia] preadipocytes wk on a diet high in carbohydrates and saturated fats. were maintained in DMEM supplemented with 10% fetal Those rats given a selective antagonist of C3aR or bovine serum (FBS), 10 U/ml penicillin, 10 U/ml streptomy- C5aR during the last 8 wk of the diet had substantially cin, and 10 mM glutamine. Differentiation was induced 2 d reduced adiposity and body weight and significantly after confluence in DMEM/10% FBS supplemented with 0.35 altered adipose and liver metabolism. Mechanisms ␮M insulin, 0.5 mM 3-isobutyl-1-methylxanthine, and 0.25 ␮M were studied in vivo, ex vivo,andin vitro to reveal for dexamethasone. After 3 d, the differentiating medium was replaced with postdifferential medium consisting of DMEM/ the first time that C3a and C5a act directly on 10% FBS with 0.35 ␮M insulin. Fully differentiated 3T3-L1 adipocytes and indirectly, via macrophages, to pro- adipocytes (Ն90% of which showed phenotypic appearance mote obesity, adiposity, and energy conservation. by accumulation of multiple lipid droplets) were then refed C3a and C5a were found to exert insulin-like func- every 2 d with DMEM/10% FBS. Primary human monocyte- tions on adipocytes by stimulating their uptake of derived macrophages were purified from buffy coat using fatty acid and glucose, while inhibiting cyclic adeno- Ficoll-Paque Plus (GE Healthcare, Castle Hill, NSW, Austra- lia) density centrifugation, and CD14ϩ monocytes were pos- sine monophosphate (cAMP) stimulation and lipoly- itively selected using CD14ϩ MACS magnetic beads (Miltenyi sis. These new metabolic roles for C3a and C5a in Biotech, Auburn, CA, USA). CD14ϩ cells were then differen- regulating energy storage and utilization in adipose tiated with 10 ng/ml of macrophage-colony stimulating factor and liver contrast starkly with the intensive energy for 7 d. RAW264.7 cells [American Type Culture Collection consuming roles of other complement components (ATCC), Manassas, VA, USA] were maintained in DMEM that fight infection. Since the compounds used here supplemented with 10% FBS, 10 U/ml penicillin, 10 U/ml are potent and selective antagonists of human C3aR streptomycin, and 10 mM glutamine. and C5aR, such antagonism may also be therapeuti- Real-time PCR cally useful for the prevention and treatment of diet-induced obesity, metabolic dysfunction, and car- All samples were homogenized in Qiazol (Qiagen, Doncaster, diovascular diseases in humans. This is the first VIC, Australia). RNA was extracted from the homogenate report of an antagonist for either of these receptors according to the RNeasy mini kit (Qiagen) manufacturer’s exerting an influence on obesity and adiposity. instructions. Primers (Supplemental Table S1) were designed

TARGETING C5AR AND C3AR IN OBESITY AND METABOLISM 823 using the Primer-Blast online-based software (www.ncbi.nlm. phages were incubated with 125I-C3a or 125I-C5a and various nih.gov/tools/primer-blast/). PCR was run on an ABI PRISM concentrations of C3a, C5a, C3aRa, and C5aRa for 1 h at 7500 (Applied Biosystems, Melbourne, VIC, Australia) as room temperature. Radioactivity uptake was determined by described previously (25). scintillation counting on a ␤-counter.

Immunoblot analysis Compounds and materials

For serum samples, equal loading volumes were concentrated C5aRa (3D53) and C3aRa (SB290157) were synthesized and using Amicon 30-kDa Ultra 0.5 ml centrifugal filter devices characterized in-house. Human recombinant C5a, PGE2, and (Millipore, Billerica, MA, USA) before being subjected to ␤-actin antibody were purchased from Sigma-Aldrich. Human SDS-PAGE. Retroperitoneal adipose tissue and 3T3-L1 adi- serum C3a was purchased from Calbiochem (San Diego, CA, pocytes were homogenized on ice. Equal amounts of proteins USA). Anti-C3a and HRP-linked rabbit anti-chicken antibod- were loaded and separated on denaturing SDS-PAGE, trans- ies were purchased from Abcam (Cambridge, MA, USA). ferred, and detected using appropriate antibodies. Densito- Anti-C3aR and anti-C5aR antibodies were purchased from metric measurements of the immunoblot were analyzed using Santa Cruz Biotechnology (Santa Cruz, CA, USA). ImageJ 1.40e software (U.S. National Institutes of Health, Bethesda, MD, USA). Statistical analysis Enzyme-linked immunosorbent assay (ELISA) Data were analyzed using GraphPad Prism 5.0 (GraphPad, Retroperitoneal adipose tissue (150 mg) was incubated in San Diego, CA, USA). Pairwise comparisons between treat- KRBH for 24 h at 37°C. Explants were collected, and adipo- ments were assessed by 2-tailed Student’s t test, and changes kine levels were determined using specific ELISA sets from in body weight differences were analyzed by 2-way ANOVA BD Pharmingen (San Jose, CA, USA). RAW264.7 cells were with Bonferroni posttests. All data and error bars are ex- seeded at a density of 1 ϫ 106 cells/ml and treated with pressed as means Ϯ se. Values of P Ͻ 0.05 were considered various agents for 24 h, and culture supernatant was collected statistically significant. for prostaglandin E2 (PGE2) analysis using monoclonal kits from Sapphire Bioscience (Sydney, NSW, Australia). RESULTS cAMP analysis C3aRa and C5aRa prevent and treat diet-induced cAMP concentrations were determined using a cAMP-Glo assay (Promega, Madison, WI, USA) according to the manu- obesity in rats facturers’ instructions. Briefly, 3T3-L1 adipocytes were seeded at a density of 5 ϫ 105 cells/ml and allowed to adhere Since C3 may be a biomarker for obesity (15–17), we overnight in DMEM/10% FBS. Culture media were then hypothesized that antagonists of C3aR or C5aR that removed, and cells were washed with PBS supplemented with prevent proinflammatory actions of two downstream 2 mM 3-isobutyl-1-methylxanthine (Sigma-Aldrich) and 200 products (C3a and C5a) of complement activation ␮M 4-(3-butoxy-4-methoxybenzyl)imidazolidin-2-one (Ro 20- might attenuate diet-induced metabolic dysfunction in 1724; Sigma-Aldrich). Compounds were added at specified rats. Notably, C3a and C5a are not required to generate concentration and incubated at room temperature for 10 min and then lysed. the membrane attack complex that enables complement-mediated lysis of bacteria or damaged cells, and Fatty acid and glucose uptake so blocking the action of C3a and C5a does not block a key feature of host immunity. Fatty acid uptake was measured using the QBT Fatty Acid Here we study a C3aRa, SB290157 (23), that binds on Uptake Assay Kit (Molecular Devices, Sunnyvale, CA, USA). 5 immune cells specifically to C3aR (but not C5aR or The 3T3-L1 adipocytes were seeded at 5 ϫ 10 cells and C5L2; ref. 27 and Supplemental Fig. S1A), and a serum-starved for 1 h. Cells were exposed to various agents for 30 min at 37°C. For the inhibition assay, cells were pretreated C5aRa, 3D53 (21, 22), that binds on immune cells with C3aRa or C5aRa for 30 min prior to addition of specifically to C5aR (not C5L2 or C3aR; refs. 28, 29 and compounds. Glucose uptake was performed using a nonra- Supplemental Fig. S1B). The two antagonists were used dioisotope enzymatic assay as described previously (26). to elucidate regulation of adiposity and metabolic Briefly, cells were serum-starved for 2 h and pretreated with function in rats fed a HCHF diet. Relative to rats fed a C3aRa or C5aRa for 30 min, followed by stimulation with low-fat CS diet, those receiving the HCHF diet for 16 wk various agents for 1 h. After incubation, 2-deoxyglucose (0.5 become obese, gaining body weight and total fat mass, mM) was added and incubated for 30 min. especially abdominal or visceral fat (Fig. 1)together Lipolysis with classic signs of metabolic syndrome (20). The concentrations of plasma C3a and expression of C3aR and 3T3-L1 adipocytes were incubated in KRBH supplemented C5aR in adipose tissue were also elevated by HCHF with 0.1% glucose and 2.5% fatty acid-free BSA. Glycerol feeding, positively correlating with increased adiposity as released was quantified using coupled enzyme reactions from well as symptoms of metabolic dysfunction (Figs. 1and2) Free Glycerol Reagent (Sigma-Aldrich). and local adipose tissue inflammation (Fig. 3). After 16 Competitive receptor binding wk of HCHF feeding, many of the metabolic indicators (Figs. 1 and 2) were attenuated or reversed by daily oral Radioligand receptor binding was performed as described administration from wk 8–16 with the C3aRa (30 previously (27). Briefly, human monocyte-derived macro- mg/kg/d) or the C5aRa (5 mg/kg/d), with marked

824 Vol. 27 February 2013 The FASEB Journal ⅐ www.fasebj.org LIM ET AL. Figure 1. Antagonists C3aRa and C5aRa modulate diet-induced obesity and adiposity in rats. A) Cumulative percentage weight gain for HCHF-fed rats (nϭ10) with or without daily oral treatment from wk 8–16 with C3aRa (nϭ8) or C5aRa (nϭ8). B) Body mass composition for CS- vs. HCHF-fed rats with or without C3aRa or C5aRa treatment, assessed by dual X-ray emission spectroscopy (nϭ5). C) Visceral adiposity index for rats from different treatment groups (nϭ6). D) Plasma C3a concentrations from rats, determined by immunoblot analysis; each lane represents serum from one individual animal (nϭ4). E, F) Protein and mRNA level of C3aR (E) and C5aR (F) in adipose tissue from different treatment groups (nϭ4). Optical density of bands was measured using ImageJ software. Error bars are means Ϯ se.*P Ͻ 0.05, **P Ͻ 0.01, ***P Ͻ 0.001.

prevention in total fat mass, body weight gain (wk 8–16 treatment from wk 8-16, with plasma C3a being com- HCHF, 21Ϯ1%; ϩC3aRa, 12Ϯ2%; ϩC5aRa, 13Ϯ3%), parable to that in CS-fed rats (Fig. 1D). Despite not visceral fat deposition, plasma C3a, and adipose C3aR detecting plasma C5a (below detection limits) in and C5aR (Fig. 1). Antagonist treatment did not result HCHF-fed rats, comparable metabolic improvements in appetite suppression (Supplemental Fig. S2). This is were observed in rats treated with C5aRa or C3aRa the ﬁrst report of an antagonist for either of these (Figs. 1 and 2). We note that C5a is not normally receptors inﬂuencing obesity and adiposity. detectable in human plasma because it is readily converted to C5adesArg by carboxypeptidases. Metabolic C3aRa and C5aRa inhibit metabolic dysfunction in rats parameters that were elevated in rats fed the HCHF vs. the CS diet include impaired glucose and insulin The increase in plasma C3a concentration induced by tolerance (Fig. 2B, C), elevated plasma lipids and liver HCHF feeding over 16 wk was prevented by C3aRa enzymes (Fig. 2D, E), increased systolic blood pressure

TARGETING C5AR AND C3AR IN OBESITY AND METABOLISM 825 Figure 2. In vivo responses of antagonists C3aRa and C5aRa on the regulation of metabolic parameters elevated in HCHF- vs. CS-fed rats. A) Fasting blood glucose concentrations in rats of different groups (nϭ6). B) Oral glucose tolerance test (OGTT) in rats of different groups (nϭ6). C) Insulin tolerance test (ITT) in rats of different groups (nϭ6). D) Plasma concentrations for nonesteriﬁed fatty acids (NEFA), triglycerides, and total cholesterol in rats of different groups (nϭ5–10). E) Plasma liver enzymes for alanine transaminase (ALT), aspartate transaminase (AST), lactate dehydrogenase (LDH), and alkaline phosphatase (ALP) in rats of different groups (nϭ5–10). F–H) Echocardiographic and histological characterization of cardiac structure (F) and function (G, H) of rats in different groups (nϭ6–10). Error bars are means Ϯ se.*P Ͻ 0.05, **P Ͻ 0.01, ***P Ͻ 0.001.

(Supplemental Fig. S2C), and abnormalities in cardiac attenuated by treatment with C3aRa or C5aRa (Fig. structure (e.g., increased collagen deposition in the left 2D), including cardiac ﬁbrosis as measured by collagen ventricle of the heart) and function (Fig. 2F–H and deposition (Supplemental Fig. S2D). Treatment with Supplemental Fig. S2D). All these parameters, except C3aRa also displayed a modest increase in fasting blood elevated triglycerides and systolic blood pressure, were glucose concentration in these rats (Fig. 2A).

826 Vol. 27 February 2013 The FASEB Journal ⅐ www.fasebj.org LIM ET AL. Figure 3. Ex vivo analysis of inflammatory profiles in rat adipose tissue. A) Real-time PCR gene expression of adipokines in adipose tissue in rats of different groups. B) Protein expression of obesity-related proinflammatory adipokines from explants of adipose tissue. Error bars are means Ϯ se.*P Ͻ 0.05, **P Ͻ 0.01, ***P Ͻ 0.001.

C3aRa and C5aRa inhibit adipose inflammation Pretreatment with C3aRa and C5aRa abolished this stimulatory effect (Fig. 4A, B). Due to limitations on Obesity is usually accompanied by dysregulation of examining rat adipocytes, we investigated whether genes involved in metabolism, inflammation, and cel- PGE2 could alter C3aR and C5aR expression in 3T3-L1 lular stress in adipose tissue (30–32). We investigated murine adipocytes. We found that stimulating adi- several genes that may be influenced by HCHF feeding pocytes with PGE2 also led to increased gene and and modulated by C3aRa and C5aRa in rats. In retro- protein expression of C3aR and C5aR (Fig. 4C, D). peritoneal adipose tissue, the expression of leptin (Lep) PGE2 has been implicated as having a paracrine influ- and macrophage-specific gene (Emr1) was altered in ence on adipocytes through inhibition of lipolysis (37). HCHF rats. Treatment with either C3aRa or C5aRa Our findings demonstrate that C3a and C5a may per- attenuated these changes. Similarly, expression of pro- petuate the underlying immune response in obesity by inflammatory adipokines, such as interleukin-6 (Il6), tu- stimulating macrophages to secrete inflammatory and mor necrosis factor ␣ (Tnfa), and prostaglandin-endoper- antilipolytic mediators, such as PGE2, that increase oxide synthase 2 (Ptgs2), in retroperitoneal adipose tissue responsiveness to C3a and C5a, and may also control was increased by HCHF feeding and attenuated by the influx/efflux of macrophages into adipose tissue dur- antagonists (Fig. 3A). Given that adipokine-mediated ing weight loss. inflammatory responses are mechanistically linked to insulin resistance and metabolic dysfunction (33–35), C3aRa and C5aRa modulate lipid homeostasis in we also examined the secretory profile of obesity- 3T3-L1 adipocytes related inflammatory adipokines such as monocyte chemotactic protein-1 (MCP-1), IL-6, and TNF-␣. Ex To investigate whether adipogenesis can alter the in- vivo analysis of retroperitoneal adipose tissue explants volvement of C3a/C3aR and C5a/C5aR, we induced from HCHF rats showed elevated secretion of these 3T3-L1 preadipocytes to undergo adipogenesis. We proinflammatory adipokines, which were attenuated by found that C3ar and C5ar mRNA levels became elevated in vivo treatment over wk 8–16 with either C3aRa or on differentiation of 3T3-L1 preadipocytes to mature C5aRa (Fig. 3B). These results clearly demonstrate adipocytes, although only C5aR protein expression decreased adiposity and improved metabolic and in- increased on maturation while C3aR protein expres- flammatory status through antagonist blockade of C3aR sion remained unchanged (Supplemental Fig. S3). This and C5aR in vivo. suggests that C5a, if not both C3a and C5a, might be important for regulating adipogenesis by recruiting C3aRa and C5aRa ameliorate inflammatory responses preadipocytes. in murine macrophages Since development of adipogenesis and adipocyte hypertrophy involves the uptake of excess nutrients, Since C3a and C5a act directly on immune cells, we such as glucose and fatty acids, we investigated whether investigated whether these proteins might also exert C3a and C5a can modulate lipid homeostasis. We paracrine control over adipocytes via macrophages that found that C3a and C5a were able to stimulate uptake are particularly associated with phagocytosing excess of 2-deoxyglucose (Fig. 5A) and BODIPY fatty acid (Fig. lipids and possibly secreting antilipolytic factors to 5B) by 3T3-L1 adipocytes. Notably, pretreatment with curtail the increase in free fatty acids during lipolysis in C3aRa and C5aRa blocked the uptake of 2-deoxyglu- obesity (36). We stimulated RAW264.7 murine macro- cose and BODIPY fatty acid, returning them to basal phages with C3a or C5a, and found particularly en- levels suggestive of selective actions through their re- hanced secretion of PGE2 from these macrophages. spective receptors (Fig. 5A, B). C3aR and C5aR are

TARGETING C5AR AND C3AR IN OBESITY AND METABOLISM 827 Figure 4. C3aRa and C5aRa modulate inﬂammatory responses in RAW264.7 macrophages. A, B) PGE2 secretion is enhanced by C3a (A) and C5a (B), but inhibited by pretreatment with C3aRa (A) or C5aRa (B). C, D) Up-regulation of protein and mRNA

expression levels of C3aR (C) and C5aRa (D) in 3T3-L1 adipocytes in the presence of PGE2. Error bars are means Ϯ se.*P Ͻ 0.05, **P Ͻ 0.01, ***P Ͻ 0.001.

primarily Gi-coupled receptors in immune cells, and, C5a and their receptors in adipocyte biology, energy since cyclic AMP (cAMP) is vital for dictating lipolysis conservation, diet-induced adiposity, and metabolic (38), we examined the capacity of C3a and C5a to dysfunction. Oral administration over 8 wk with selec- decrease cAMP concentrations in adipocytes (Fig. 5C). tive antagonists C3aRa and C5aRa (Supplemental Fig. Both C3a and C5a reduced cAMP concentrations, and S1) ameliorated classic symptoms of metabolic syn- pretreatment with pertussis toxin attenuated both C3a- drome in diet-induced obese rats, with marked reduc- and C5a-mediated inhibition of cAMP (Supplemental tion in body weight, visceral fat, and adiposity, and Fig. S4), establishing that both C3aR and C5aR are improvements in glucose and insulin intolerance, adipose Gi-coupled receptors in adipocytes. Further, we demon- tissue inflammation, obesity-associated adipokines, and strated that both C3a and C5a suppressed isoprotere- certain lipid and cardiovascular abnormalities (Figs. 1–3 nol-induced lipolysis measured as glycerol release (Fig. and Supplemental Fig. S2). The mechanism for these in 5D). Consistent with the above findings, treatment with vivo effects was traced to previously unknown signaling either C3aRa or C5aRa restored both cAMP concentra- roles for C3a and C5a in promoting energy conservation tions and lipolysis in adipocytes. Taken together, these and regulating energy homeostasis in adipose tissue results suggest that both C3a and C5a are important (Fig. 3). Receptor-selective antagonists were used here promoters of energy conservation in adipose tissue, to identify two conceptually new mechanisms for in- both promoting energy storage through uptake of volvement of C3a and C5a in metabolic function, 2-deoxyglucose and fatty acids into adipocytes, while at the same time retarding energy utilization by inhibiting namely stimulating uptake of fatty acid and glucose by cAMP stimulation and lipolysis. The antagonists were adipocytes, while at the same time inhibiting fat burn- able to inhibit these effects that contribute to adiposity off by inhibiting cAMP stimulation and lipolysis in in vitro, and thus to inhibit diet-induced obesity in vivo. adipocytes (Fig. 5). It is known that activating complement proteins in blood triggers immune responses that require energy to fight infection, but here we identify DISCUSSION for the first time new energy-conserving roles for complement proteins, C3a and C5a, and their receptors, This study reveals important new findings indicating C3aR and C5aR, that are contraindicated in known involvement of complement anaphylatoxins C3a and complement biology.

828 Vol. 27 February 2013 The FASEB Journal ⅐ www.fasebj.org LIM ET AL. Figure 5. C3aRa and C5aRa modulate glucose or lipid homeostasis in 3T3-L1 adipocytes. A) Like insulin (3 nM), C3a stimulates uptake of 2-deoxyglucose (2-DG) in 3T3-L1 adipocytes but was inhibited by pretreatment with C3aRa (10 ␮M). C5a (100 nM) had no effect (not shown). B) Like insulin (50 nM), C5a stimulates uptake of fatty acids by 3T3-L1 adipocytes, which was inhibited by pretreatment with C5aRa (10 ␮M). C3a (100 nM) had no effect (not shown). C) C3a and C5a (100 nM) inhibit cAMP stimulated by forskolin in 3T3-L1 adipocytes, while pretreatment with C3aRa or C5aRa (10 ␮M) normalizes cAMP concentrations to levels induced by forskolin. D) C3a and C5a (100 nM) inhibit isoproterenol-induced glycerol release from 3T3-L1 adipocytes, while pretreatment with C3aRa or C5aRa (10 ␮M) normalizes glycerol release to isoproterenol-levels.

Notably, C3a and C5a are not directly required compromised by excessive fat uptake or by paracrine during host immunity to generate the membrane attack influences from immune cells that infiltrate adipose complex that enables complement-mediated lysis of tissue and secrete proinflammatory agents. Adipocytes bacteria or damaged cells, and so blocking the action of and immune cells in adipose tissue release many bioac- C3a and C5a on their respective cell surface GPCRs tive mediators, including adipokines and inflammatory does not block host immunity. Complement signaling cytokines that, when aberrantly expressed, result in through these proteins clearly plays previously un- metabolic dysfunction (4). Paracrine signaling through known and important roles in regulating metabolism PGE2 secreted by macrophages (Fig. 4) constitutes a and energy homeostasis through direct action on adi- paracrine loop by which these complement mediators pocytes and indirect action through other nonadipose aid further recruitment of immune cells into adipose cells (e.g., macrophages) on adipocytes. tissue, thereby perpetuating local adipose inflamma- Both C3a and C5a were shown herein to act in vitro tion and proliferative adipocyte precursor cells that directly on adipocytes, altering carbohydrate and lipid induce adipogenesis in response to increased demands storage, transport, and utilization, including adipogen- for energy storage in obesity. While the C3aRa and esis, lipogenesis, glucose uptake, and lipolysis (Fig. 5). C5aRa used here have certain anti-inflammatory prop- Our conclusions are offered partial support from a erties in animal models of inflammatory diseases (11, recent gene expression study demonstrating that C3aR 27, 41), other anti-inflammatory agents, such as salicy- may be a candidate gene for predicting obesity and lates, have shown negligible or only very minor benefits transcriptional networking related to diabetes and ath- in vivo in metabolic syndrome (42). erosclerosis in mice (39). Furthermore, mice with diet- Although complement proteins are formed in re- induced obesity exhibited increased expression of adi- sponse to infection through energy-depleting immune pose tissue C3aR, while C3aRϪ/Ϫ knockout mice have defense mechanisms, certain byproducts of the activa- reduced macrophage infiltration in adipose tissue with tion of the complement protein network, such as C3a improved insulin responsiveness (40). However, mech- and C5a, play roles in metabolism and evidently pro- anisms associating the C3aRϪ/Ϫ phenotype with im- mote energy conservation. This occurs directly through proved adiposity and insulin responsiveness were in- adipocytes by promoting fatty acid and glucose uptake conclusive and attributed to macrophage responses while at the same time inhibiting lipolysis, and indi- (40), rather than to adipocytes and immune cells in rectly through macrophages by stimulating release of combination in adipose tissue as described herein. PGE2, which is also known to regulate cAMP stimula- Although adipocytes have evolved to specifically reg- tion in adipocytes. Metabolic homeostasis is important ulate energy storage and utilization, their functions are for an optimally functioning immune system, since

TARGETING C5AR AND C3AR IN OBESITY AND METABOLISM 829 initiation and maintenance of immunity is metaboli- R. E., and Gewirtz, A. T. (2010) Metabolic syndrome and altered cally expensive and is impaired during both energy gut microbiota in mice lacking toll-like receptor 5. Science 328, 228–231 deficiency and surplus (5). Energy regulation through 10. Li, P., and Hotamisligil, G. S. (2010) Metabolism: host and storage, transport, and metabolism may be intimately microbes in a pickle. Nature 464, 1287–1288 coordinated with immune defense and pathogen-sens- 11. Monk, P. N., Scola, A. M., Madala, P., and Fairlie, D. P. (2007) ing mechanisms in an organism. Recent evidence sug- Function, structure and therapeutic potential of complement C5a receptors. Br. J. Pharmacol. 152, 429–448 gests that nutrient and pathogen sensing systems in 12. Ricklin, D., Hajishengallis, G., Yang, K., and Lambris, J. D. humans may have evolved together, possibly to coordi- (2010) Complement: a key system for immune surveillance and nate efficient management of energy homeostasis, met- homeostasis. Nat. Immunol. 11, 785–797 abolic function, and immune defense (6). Taken to- 13. Choy, L. N., Rosen, B. S., and Spiegelman, B. M. (1992) Adipsin and an endogenous pathway of complement from adipose cells. gether, our results indicate an important link between J. Biol. Chem. 267, 12736–12741 energy homeostasis and innate immunity, raising the 14. Zhang, J., Wright, W., Bernlohr, D. A., Cushman, S. W., and specter of aberrant immune responses (producing C3a Chen, X. (2007) Alterations of the classic pathway of complement in adipose tissue of obesity and insulin resistance. Am. J. and C5a and overexpressing C3aR and C5aR) exacer- Physiol. Endocrinol. Metab. 292, E1433–1440 bating metabolic dysfunction and vice versa. The results 15. Engstrom, G., Hedblad, B., Janzon, L., and Lindgarde, F. (2005) suggest that targeting key signaling components of the Weight gain in relation to plasma levels of complement factor 3: complement system, and by implication perhaps other results from a population-based cohort study. Diabetologia 48, 2525–2531 immune components, may be of therapeutic benefit for 16. Phillips, C. M., Goumidi, L., Bertrais, S., Ferguson, J. F., Field, treating diet-induced obesity, metabolic dysfunction, M. R., Kelly, E. D., Peloso, G. M., Cupples, L. A., Shen, J., and cardiovascular disease. Ordovas, J. M., McManus, R., Hercberg, S., Portugal, H., Lairon, D., Planells, R., and Roche, H. M. (2009) Complement compo- The authors thank the Australian Research Council (grant nent 3 polymorphisms interact with polyunsaturated fatty acids to modulate risk of metabolic syndrome. Am. J. Clin. Nutr. 90, DP1093245), Federation Fellowship FF0668733, the Australian 1665–1673 National Health and Medical Research Council (grant 631534), 17. Van Oostrom, A., Alipour, A., Plokker, T. W. M., Sniderman, and Senior Principal Research Fellowship APP1027369 (to D.F.) A. D., and Cabezas, M. C. (2007) The metabolic syndrome in for funding, and the Institute for Molecular Bioscience and relation to complement component 3 and postprandial lipemia The University of Queensland for Ph.D. scholarships (to J.L. in patients from an outpatient lipid clinic and healthy volun- and V.S.). Author contributions: J.L., A.I., and D.F. wrote the teers. Atherosclerosis 190, 167–173 manuscript with important editorial input from L.B.; J.L. and 18. Weyer, C., Tataranni, P. A., and Pratley, R. E. (2000) Insulin A.I. contributed most research results with help on gene action and insulinemia are closely related to the fasting com- expression and cell regulation from J.S. and V.S., all working plement C3, but not acylation stimulating protein concentra- in the laboratories of D.F. and L.B.; R.R. synthesized and tion. Diabetes Care 23, 779–785 characterized the key antagonists.; J.L., A.I., L.B., and D.F. 19. Johswich, K., Martin, M., Thalmann, J., Rheinheimer, C., Monk, P. N., and Klos, A. (2006) Ligand specificity of the anaphyla- designed the experiments. The authors declare no conflicts toxin C5L2 receptor and its regulation on myeloid and epithe- of interest. lial cell lines. J. Biol. Chem. 281, 39088–39095 20. Panchal, S. K., Poudyal, H., Iyer, A., Nazer, R., Alam, A., Diwan, V., Kauter, K., Sernia, C., Campbell, F., Ward, L., Gobe, G., REFERENCES Fenning, A., and Brown, L. (2011) High-carbohydrate high-fat diet-induced metabolic syndrome and cardiovascular remodel- 1. Wellen, K. E., and Hotamisligil, G. S. (2005) Inflammation, ing in rats. J. Cardiovasc. Pharmacol. 57, 611–624 stress, and diabetes. J. Clin. Invest. 115, 1111–1119 21. Finch, A. M., Wong, A. K., Paczkowski, N. J., Wadi, S. K., Craik, 2. Van Gaal, L. F., Mertens, I. L., and De Block, C. E. (2006) D. J., Fairlie, D. P., and Taylor, S. M. (1999) Low-molecular- Mechanisms linking obesity with cardiovascular disease. Nature weight peptidic and cyclic antagonists of the receptor for the 444, 875–880 complement factor C5a. J. Med. Chem. 42, 1965–1974 3. Shoelson, S. E., Lee, J., and Goldfine, A. B. (2006) Inflammation 22. Wong, A. K., Finch, A. M., Pierens, G. K., Craik, D. J., Taylor, and insulin resistance. J. Clin. Invest. 116, 1793–1801 S. M., and Fairlie, D. P. (1998) Small molecular probes for 4. Iyer, A., Fairlie, D. P., Prins, J. B., Hammock, B. D., and Brown, G-protein-coupled C5a receptors: conformationally constrained L. (2010) Inflammatory lipid mediators in adipocyte function antagonists derived from the C terminus of the human plasma and obesity. Nat. Rev. Endocrinol. 6, 71–82 protein C5a. J. Med. Chem. 41, 3417–3425 5. Hotamisligil, G. S. (2006) Inflammation and metabolic disor- 23. Ames, R. S., Lee, D., Foley, J. J., Jurewicz, A. J., Tornetta, M. A., ders. Nature 444, 860–867 Bautsch, W., Settmacher, B., Klos, A., Erhard, K. F., Cousins, 6. Hotamisligil, G. S., and Erbay, E. (2008) Nutrient sensing and R. D., Sulpizio, A. C., Hieble, J. P., McCafferty, G., Ward, K. W., inflammation in metabolic diseases. Nat. Rev. Immunol. 8, 923– Adams, J. L., Bondinell, W. E., Underwood, D. C., Osborn, R. R., 934 Badger, A. M., and Sarau, H. M. (2001) Identification of a 7. Seppala-Lindroos, A., Vehkavaara, S., Hakkinen, A. M., Goto, T., selective nonpeptide antagonist of the anaphylatoxin C3a recep- Westerbacka, J., Sovijarvi, A., Halavaara, J., and Yki-Jarvinen, H. tor that demonstrates antiinflammatory activity in animal mod- (2002) Fat accumulation in the liver is associated with defects in els. J. Immunol. 166, 6341–6348 insulin suppression of glucose production and serum free fatty 24. Iyer, A., Fenning, A., Lim, J., Le, G. T., Reid, R. C., Halili, M. A., acids independent of obesity in normal men. J. Clin. Endocrinol. Fairlie, D. P., and Brown, L. (2010) Antifibrotic activity of an Metab. 87, 3023–3028 inhibitor of histone deacetylases in DOCA-salt hypertensive rats. 8. Cani, P. D., Amar, J., Iglesias, M. A., Poggi, M., Knauf, C., Br. J. Pharmacol. 159, 1408–1417 Bastelica, D., Neyrinck, A. M., Fava, F., Tuohy, K. M., Chabo, C., 25. Suen, J. Y., Gardiner, B., Grimmond, S., and Fairlie, D. P. (2010) Waget, A., Delmee, E., Cousin, B., Sulpice, T., Chamontin, B., Profiling gene expression induced by protease-activated recep- Ferrieres, J., Tanti, J. F., Gibson, G. R., Casteilla, L., Delzenne, tor 2 (PAR2) activation in human kidney cells. PLoS One 5, N. M., Alessi, M. C., and Burcelin, R. (2007) Metabolic endo- e13809 toxemia initiates obesity and insulin resistance. Diabetes 56, 26. Yamamoto, N., Kawasaki, K., Kawabata, K., and Ashida, H. 1761–1772 (2010) An enzymatic fluorimetric assay to quantitate 2-deoxyg- 9. Vijay-Kumar, M., Aitken, J. D., Carvalho, F. A., Cullender, T. C., lucose and 2-deoxyglucose-6-phosphate for in vitro and in vivo Mwangi, S., Srinivasan, S., Sitaraman, S. V., Knight, R., Ley, use. Anal. Biochem. 404, 238–240

830 Vol. 27 February 2013 The FASEB Journal ⅐ www.fasebj.org LIM ET AL. 27. Scully, C. C., Blakeney, J. S., Singh, R., Hoang, H. N., Ab- 36. Kosteli, A., Sugaru, E., Haemmerle, G., Martin, J. F., Lei, J., benante, G., Reid, R. C., and Fairlie, D. P. (2010) Selective Zechner, R., and Ferrante, A. W., Jr. (2010) Weight loss and hexapeptide agonists and antagonists for human complement lipolysis promote a dynamic immune response in murine adi- C3a receptor. J. Med. Chem. 53, 4938–4948 pose tissue. J. Clin. Invest. 120, 3466–3479 28. Scola, A. M., Higginbottom, A., Partridge, L. J., Reid, R. C., 37. Jaworski, K., Ahmadian, M., Duncan, R. E., Sarkadi-Nagy, E., Woodruff, T., Taylor, S. M., Fairlie, D. P., and Monk, P. N. Varady, K. A., Hellerstein, M. K., Lee, H. Y., Samuel, V. T., (2007) The role of the N-terminal domain of the complement Shulman, G. I., Kim, K. H., de Val, S., Kang, C., and Sul, H. S. fragment receptor C5L2 in ligand binding. J. Biol. Chem. 282, (2009) AdPLA ablation increases lipolysis and prevents obesity 3664–3671 induced by high-fat feeding or leptin deficiency. Nat. Med. 15, 29. Van Lith, L. H., Oosterom, J., Van Elsas, A., and Zaman, G. J. 159–168 (2009) C5a-stimulated recruitment of beta-arrestin2 to the 38. Kawamura, M., Jensen, D. F., Wancewicz, E. V., Joy, L. L., Khoo, nonsignaling 7-transmembrane decoy receptor C5L2. J. Biomol. J. C., and Steinberg, D. (1981) Hormone-sensitive lipase in Screen. 14, 1067–1075 differentiated 3T3-L1 cells and its activation by cyclic AMP- 30. Laaksonen, D. E., Lakka, H. M., Salonen, J. T., Niskanen, L. K., dependent protein kinase. Proc. Natl. Acad. Sci. U. S. A. 78, Rauramaa, R., and Lakka, T. A. (2002) Low levels of leisure-time 732–736 physical activity and cardiorespiratory fitness predict develop- 39. Yang, X., Deignan, J. L., Qi, H., Zhu, J., Qian, S., Zhong, J., ment of the metabolic syndrome. Diabetes Care 25, 1612–1618 Torosyan, G., Majid, S., Falkard, B., Kleinhanz, R. R., Karlsson, 31. Liu, J., Divoux, A., Sun, J., Zhang, J., Clement, K., Glickman, J., Castellani, L. W., Mumick, S., Wang, K., Xie, T., Coon, M., J. N., Sukhova, G. K., Wolters, P. J., Du, J., Gorgun, C. Z., Doria, Zhang, C., Estrada-Smith, D., Farber, C. R., Wang, S. S., van Nas, A., Libby, P., Blumberg, R. S., Kahn, B. B., Hotamisligil, G. S., A., Ghazalpour, A., Zhang, B., Macneil, D. J., Lamb, J. R., and Shi, G. P. (2009) Genetic deficiency and pharmacological Dipple, K. M., Reitman, M. L., Mehrabian, M., Lum, P. Y., stabilization of mast cells reduce diet-induced obesity and Schadt, E. E., Lusis, A. J., and Drake, T. A. (2009) Validation of diabetes in mice. Nat. Med. 15, 940–945 candidate causal genes for obesity that affect shared metabolic 32. Nishimura, S., Manabe, I., Nagasaki, M., Eto, K., Yamashita, H., pathways and networks. Nat. Genet. 41, 415–423 Ohsugi, M., Otsu, M., Hara, K., Ueki, K., Sugiura, S., Yoshimura, 40. Mamane, Y., Chan, C. C., Lavallee, G., Morin, N., Xu, L. J., K., Kadowaki, T., and Nagai, R. (2009) CD8ϩ effector T cells Huang, J. Q., Gordon, R., Thomas, W., Lamb, J., Schadt, E. E., contribute to macrophage recruitment and adipose tissue in- Kennedy, B. P., and Mancini, J. A. (2009) The C3a anaphyla- flammation in obesity. Nat. Med. 15, 914–920 toxin receptor is a key mediator of insulin resistance and 33. Hotamisligil, G. S., Arner, P., Caro, J. F., Atkinson, R. L., and functions by modulating adipose tissue macrophage infiltration Spiegelman, B. M. (1995) Increased adipose tissue expression of and activation. Diabetes 58, 2006–2017 tumor necrosis factor-alpha in human obesity and insulin resis- 41. Proctor, L. M., Arumugam, T. V., Shiels, I., Reid, R. C., Fairlie, tance. J. Clin. Invest. 95, 2409–2415 D. P., and Taylor, S. M. (2004) Comparative anti-inflammatory 34. Kanda, H., Tateya, S., Tamori, Y., Kotani, K., Hiasa, K., Kitazawa, activities of antagonists to C3a and C5a receptors in a rat model R., Kitazawa, S., Miyachi, H., Maeda, S., Egashira, K., and of intestinal ischaemia/reperfusion injury. Br. J. Pharmacol. 142, Kasuga, M. (2006) MCP-1 contributes to macrophage infiltra- 756–764 tion into adipose tissue, insulin resistance, and hepatic steatosis 42. Yuan, M., Konstantopoulos, N., Lee, J., Hansen, L., Li, Z. W., in obesity. J. Clin. Invest. 116, 1494–1505 Karin, M., and Shoelson, S. E. (2001) Reversal of obesity- and 35. Xu, H., Barnes, G. T., Yang, Q., Tan, G., Yang, D., Chou, C. J., diet-induced insulin resistance with salicylates or targeted dis- Sole, J., Nichols, A., Ross, J. S., Tartaglia, L. A., and Chen, H. ruption of Ikkbeta. Science 293, 1673–1677 (2003) Chronic inflammation in fat plays a crucial role in the development of obesity-related insulin resistance. J. Clin. Invest. Received for publication September 13, 2012. 112, 1821–1830 Accepted for publication October 22, 2012.

TARGETING C5AR AND C3AR IN OBESITY AND METABOLISM 831