CHEMERIN AND IL-17 IN , OBESITY, AND METABOLISM

A DISSERTATION SUBMITTED TO THE PROGRAM IN IMMUNOLOGY AND THE COMMITTEE ON GRADUATE STUDIES OF STANFORD UNIVERSITY IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

Luis Alejandro Zúñiga

November 2010

© 2011 by Luis Alejandro Zuniga. All Rights Reserved. Re-distributed by Stanford University under license with the author.

This work is licensed under a Creative Commons Attribution- Noncommercial 3.0 United States License. http://creativecommons.org/licenses/by-nc/3.0/us/

This dissertation is online at: http://purl.stanford.edu/qf352vw3320

ii I certify that I have read this dissertation and that, in my opinion, it is fully adequate in scope and quality as a dissertation for the degree of Doctor of Philosophy.

Eugene Butcher, Primary Adviser

I certify that I have read this dissertation and that, in my opinion, it is fully adequate in scope and quality as a dissertation for the degree of Doctor of Philosophy.

Ajay Chawla

I certify that I have read this dissertation and that, in my opinion, it is fully adequate in scope and quality as a dissertation for the degree of Doctor of Philosophy.

Patricia Jones

I certify that I have read this dissertation and that, in my opinion, it is fully adequate in scope and quality as a dissertation for the degree of Doctor of Philosophy.

Olivia Martinez

Approved for the Stanford University Committee on Graduate Studies. Patricia J. Gumport, Vice Provost Graduate Education

This signature page was generated electronically upon submission of this dissertation in electronic format. An original signed hard copy of the signature page is on file in University Archives.

iii

iv

ABSTRACT

Inflammation is characterized by the influx and activation of leukocytes to sites of tissue damage. Upon recruitment, immune cells become activated and can secrete pro- inflammatory cytokines and chemoattractants that can modulate cellular responses and leukocyte trafficking. However, if inflammation occurs inappropriately, it can lead to unwanted tissue damage to the host or can affect metabolic processes (e.g. glucose metabolism or bone metabolism). Evidence indicates inflammation also occurs in during obesity and that adipose tissue leukocytes secrete cytokines that can contribute to systemic, chronic inflammation. The mechanisms that regulate inflammation and mediate its effects on metabolism are not fully understood. To help clarify what factors mediate inflammation, we sought to determine the role that the receptor, CMKLR1, may play in mediating immune cell trafficking to inflamed tissue. We found CMKLR1 is expressed by murine peritoneal , which can migrate to chemerin in vitro. CMKLR1 was upregulated by TGFβ and downregulated by pro-inflammatory cytokines and TLR ligands. Furthermore, we found that CMKLR1 exacerbated the progression of experimental murine autoimmune encephalomyelitis

(EAE). Chemerin was expressed in inflamed central nervous system (CNS) tissue, and microglial cells and CNS-associated myeloid dendritic cells expressed CMKLR1 in mice with EAE. Additionally, we discovered CCRL2 as a novel receptor for chemerin and determined it is a non-signaling decoy receptor which may bind and present chemerin to

CMKLR1. We found that CCRL2 is expressed by mouse mast cells and can enhance

v

tissue swelling and tissue leukocyte infiltration during mast cell-dependent passive

cutaneous anaphylaxis. In an attempt to uncover a role for chemerin/CMKLR1 in obesity

development or obesity-related inflammation, we discovered an unexpected role for the

pro-inflammatory cytokine, IL-17, in metabolism. IL-17 was upregulated in

obese adipose tissue by T cells. IL-17 KO mice were more susceptible to diet-induced

obesity. IL-17 inhibited adipocyte development and insulin stimulated glucose uptake by

. Furthermore, young IL-17 deficient mice were more insulin sensitive than their WT counterparts. Overall, our data indicate a role for chemerin and chemerin receptors in leukocyte migration and function in inflammatory diseases, and suggest that

IL-17 regulates metabolic responses associated with obesity.

vi

ACKNOWLEDGEMENTS

I feel as though I could write an entire book in order to thank all the people who have walked, ran with, or carried me along the path that eventually led to here. I’m infinitely fortunate to have the support from gifted colleagues, great friends, and a wonderful family. You each have my deepest gratitude.

Eugene – Thank you for allowing me to pursue the line of research I was fortunate enough to have followed, despite it being well outside of your realm of expertise. I cannot begin to express how much your support during my time here means to me. Your drive pushed me through a number of low periods and your mentorship helped me to understand how a great scientist should approach a problem. I am extremely honored to have been your student.

I owe a great debt of appreciation to Brian Zabel. Thank you for taking me under your wing and investing the time and patience to mentor me, even when it sometimes came at the cost of losing precious reagents. You are an amazing scientist, a model teacher, and a true friend.

To my committee members, Pat Jones, Olivia Martinez, and Ajay Chawla – Your direction and encouragement were pivotal in my development as a scientist. Thank you for taking the time out of your busy lives to help guide my research and help teach me how to approach my research with different perspectives.

I couldn’t have made it this far without the help of a number of extremely talented collaborators. Wen-Jun Shen and Barbara-Joyce Shaikh – You have witnessed my ups

vii

and downs, given me so much advice and support, and throughout it all, you have been wonderful friends. Rick Kraemer, Yueh-Hsiu Chien, and Dan Cua – Thank you for your support and guidance, you are all exemplary mentors and exceptional scientists. I can’t forget Sarah Gaffen, Jaya Goswami, Christina Meyer, Xun and Su – Thank you all for taking the time to offer your input and lend a hand with my research.

Angelo Mao, ColinThom, Sofia Andrade, Ekaterina Pyatnova, and Andrew

Richards – You are some of the most talented young adults I have ever had the fortune to meet and I couldn’t have come this far without you. I am confident you will all be very successful in your future endeavors. Thank you for taking my research and making it yours as well, and allowing me the honor to mentor you.

I must thank Maureen Panganiban, the Immunology program’s mother hen. The program would fall apart without you. I’m grateful for everything you have ever done for me. Who else would have offered to lend their own money when my stipend didn’t come on time. The program is lucky to have such a dedicated administrator such as you.

To all my lab mates and extended lab mates, present and past, I owe an extreme amount of gratitude. Thank you to Big Reem, Aida Habtezion, Husein Hadeiba, Carsten

Alt, Reverend Takao, Gudrun Debes, Hekla Sigmundsdottir, Alison Holzer, Raymond

Kwan, Ed O’Hara, Francis Lin, Derek Lindsey, Jean Chen, Tracy Staton-Winslow, Carrie

Arnold, Tohru Sato, Ken Youngman, Linh Nguyen and Katharina Lahl (C4-111!!), Yoe-

Sik Bae, Russell Pachynski, Junliang Pan, Dora Friedman, Ryan Huss, Justin Monnier,

Nicole Lazarus, Cecilia Operup, Jian Zhang, Jean Jang, Evelyn Resurreccion, Lusijah

viii

Rott, Sarah Adler, Bari Nazario (cookies!), Lordes Magalhaes, Orr Sharpe, and Dan

Winer for your technical help, scientific input, and valuable friendship.

Of course, how could I have made it this far without the support of my friends and classmates? You've seen me at my lowest and were there to help me up. You’ve seen me at my best and were there to celebrate. You have made my time here so memorable. I owe special thanks to Eric Mabery, Jon Jones, Ana Older-Aguilar, Emily Deal, Martin

Vandermeer, Cornell Wells, Ian Wiseman, Jarred Caldwell, Mike Alonso, Michael

“Blade” Wong, Oliver Crespo, Xianne Leong, Christina Swanson, Ming Cheah, Michael

“Birney” Birnbaum, Wen-Qi Ho, Ching Ding, Azim Raza Khan, Marc Bruce, Ruizhu

Zheng, and Yann Chong Tan. Thank you for accepting me for who I am, flaws and all. I couldn’t ask for better friends!

My family has had a huge influence on my progress at Stanford. All my uncles, aunts, cousins (cuz-os!), nephews, and nieces have given me so much support over the years. I have to admit, it feels good when Alexander introduces me as the Stanford cousin to his friends, when uncle Tim asks me probing questions about my research, when

Grandmother Clara calls just to check up on me, or when Marnie gives me a hug out of the blue and tells me how proud she is to have me as a son-in-law. My family has given me so many reasons to continue along my way and have pride in my accomplishments. I cannot thank you all enough for your love and support!

To my brothers, Joshua and Noah – I am so proud to have you as siblings. You are exceptional workers, great friends, and wonderful fathers. Noah, you are one of the

ix

wittiest people I know and you always bring joy to the people around you. Josh, you are

passionate about everything you do and always strive to be your best. Both of you are

inspirations to me and just knowing you makes me a better person.

Tina – Thank you for all your words of wisdom and guidance over the years. You

always seem to know exactly what to say to place things in a positive perspective. I’m glad you came into our family, as you have become a wonderful friend, and more importantly, a loving mother.

To my father – It is difficult for me to put into words the impact you’ve had on me, not only as a person, but as a scientist. You are my hero. You always did what you had to do in order to support your family and without that support, I certainly would not have made it this far. Your passion for knowledge just for the sake of knowledge is one of the things that inspired me the most to pursue a career in science. Thank you for doing the best you could for us and allowing me the freedom to find my own way.

Finally, to my beautiful wife Jennifer and unborn son, Luis – Jennifer, you are my biggest supporter and best friend. Before I met you, I was just drifting along in life with no clear direction. You inspired me to continue with my education and become better than I was. You give me strength, clarity, and purpose and I would not be where I am today without you. Thank you for becoming a part of my life and for giving me endless support. Luis, even though you haven’t graced us with your presence yet, you have been an inspiration for me to continue working hard in order to be a father you can look up to.

I can’t wait to meet you. I love you both and look forward to starting our family together.

x

DEDICATION

The author wishes to dedicate this dissertation to his mother, Cynthia Ann Zúñiga.

I wish you could have seen all your “little turkey’s” accomplishments, I think you would have been proud. Although you are not here, you have always been with me, and will always be with me.

xi

TABLE OF CONTENTS

Chapter 1 : Introduction ...... 1 Obesity: A brief history ...... 1 Obesity, metabolism, and inflammation ...... 4 Adipose tissue leukocyte infiltration ...... 8 Adipose tissue T cells ...... 14 Chemerin and chemerin receptors in inflammation, obesity, and metabolism ...... 16 Significance...... 21 References ...... 24 Chapter 2 : -like Receptor 1 Expression by Macrophages in vivo: Regulation by TGF-beta and TLR Ligands ...... 34 Summary ...... 35 Introduction ...... 37 Materials and Methods ...... 40 Results ...... 47 A mCMKLR1-specific mAb stains early DC progenitors in vitro, but not most blood or tissue DC ...... 47 mCMKLR1 is selectively expressed by murine macrophages ...... 51 Chemerin is a functional ligand for mouse DC progenitors and in vivo peritoneal macrophages ...... 56 Regulation of mCMKLR1 by M1 vs. M2 stimuli ...... 59 Ex vivo human ascites macrophages express huCMKLR1 and are chemerin- responsive ...... 61 Discussion ...... 64 Acknowledgements ...... 69 References ...... 70

xii

Chapter 3 : Chemokine-Like Receptor-1 Expression by Central Nervous System-Infiltrating Leukocytes and Involvement in a Model of Autoimmune Demyelinating Disease ...... 73 Summary ...... 74 Introduction ...... 75 Materials and Methods ...... 77 Supplemental Materials and Methods ...... 82 Results ...... 85 Attenuation of clinical EAE in CMKLR1 KO mice ...... 85 Reduced histological EAE in CMKLR1 KO mice ...... 88 CNS inflammatory lesions of CMKLR1 KO mice contain fewer F4/80+ cells ....91 Lymphocyte proliferation and cytokine production ...... 91 Induction of EAE by adoptive transfer ...... 94 Peritoneal responses ...... 96 Analysis of CNS mononuclear cells by flow cytometry ...... 96 Chemerin, a natural ligand for CMKLR1, is up-regulated in the CNS of mice with EAE ...... 100 Discussion ...... 101 Acknowledgements ...... 107 Supplemental Figures ...... 108 References ...... 111 Chapter 4 : Mast Cell-Expressed Orphan Receptor CCRL2 Binds Chemerin and is Required for Optimal Induction of IgE-Mediated Passive Cutaneous Anaphylaxis ...... 114 Summary ...... 116 Introduction ...... 117 Materials and Methods ...... 123 Supplemental Materials and Methods ...... 128

xiii

Results ...... 137 mCCRL2-specific mAbs selectively stain mouse mast cells ...... 137 CCRL2 and mast cell phenotype and function ...... 141 Mast cell-expressed CCRL2 is required for optimal induction of IgE- dependent passive cutaneous anaphylaxis ...... 146 CCRL2 binds chemerin...... 152 CCRL2 does not support chemerin-driven signal transduction ...... 158 CCRL2 does not internalize chemerin ...... 161 Discussion ...... 166 Acknowledgements ...... 172 Supplemental Figures ...... 173 References ...... 184 Chapter 5 : IL-17 regulates adipogenesis, glucose homeostasis, and obesity...... 188 Summary ...... 189 Introduction ...... 190 Materials and Methods ...... 193 Results ...... 201 IL-17 is expressed by γδ T cells in adipose tissue ...... 201 Increased number and frequency of IL-17-producing T cells in obese inguinal adipose tissue ...... 205 IL-17 deficient mice are more susceptible to diet-induced obesity ...... 208 IL-17 treatment inhibits lipid loading and disrupts adipocyte-associated expression in 3T3-L1 adipocytes ...... 211 IL-17 inhibits glucose uptake in vitro and impairs glucose and insulin metabolism in young mice ...... 218

Adipose tissue γδ17 cells are replaced with β TCR positive IL-17 producing T cells in TCRδ KO mice ...... 223 Discussion ...... 226

xiv

Acknowledgements ...... 235 References ...... 236 Chapter 6 : Final Discussion ...... 241

xv

LIST OF FIGURES

Number Page

Chapter 1

Figure 1. Obesity prevalence in the United States between 1960-2006 ...... 5 Figure 2. Adipose tissue leukocyte infiltration and inflammation model ...... 13 Figure 3. Proteolytic cleavage of chemerin ...... 17

Chapter 2 Figure 1. Anti-mCMKLR1 monoclonal Ab stains progenitors, but not most Dendritic cells in vivo...... 50 Figure 2. Freshly isolated macrophages express mCMKLR1...... 54 Figure 3. Rare circulating blood macrophages express mCMKLR1...... 55 Figure 4. mCMKLR1+ leukocytes are chemerin-responsive...... 58 Figure 5. Effects of TLR ligands and cytokines on CMKLR1 expression on mouse macrophages...... 60 Figure 6. Human macrophages express huCMKLR1 and are chemerin-responsive...... 63

Chapter 3 Figure 1. Reduced clinical EAE in CMKLR1 KO mice...... 87 Figure 2. Reduced histological EAE in CMKLR1 KO mice...... 90 Figure 3. Recall proliferation and cytokine responses of lymphocytes from CMKLR1 KO mice...... 93 Figure 4. Induction of EAE by adoptive transfer of MOG-reactive lymphocytes...... 95 Figure 5. Detection of CMKLR1+ cells in CNS of mice with EAE...... 98 Figure S1. Proliferation and cytokine production by CMKLR1 KO T cells...... 108 Figure S2. Peritoneal macrophage responses in CMKLR1 KO mice...... 109 Figure S3. Chemerin transcripts are up-regulated in the CNS of mice with EAE...... 110

xvi

Figure S4. Staining of mouse leukocytes with anti-CMKLR1 mAbs...... 110

Chapter 4 Figure 1. Mast cells express mCCRL2...... 139 Figure 2. CCRL2 KO mice...... 143 Figure 3. BMCMCs from mCCRL2 KO and WT mice display similar functional responses in vitro...... 144 Figure 4. Mast cell-expressed mCCRL2 is required for maximal tissue swelling and numbers of dermal leukocytes in passive cutaneous anaphylaxis...... 149 Figure 5. Histologic features of IgE-dependent PCA reactions in WT BMCMC- vs. KO BMCMC-engrafted KitW-sh/Wsh mice...... 150 Figure 6. CCRL2 binds chemerin...... 157 Figure 7. Chemerin:CCRL2 binding does not trigger intracellular calcium mobilization or ...... 160 Figure 8. CCRL2 can increase local chemerin concentrations...... 165 Figure 9. Proposed model of presentation of chemerin by CCRL2 to CMKLR1...... 167 Figure S1. Blood lymphocytes, BM neutrophils, and peritoneal macrophages do not detectably express mCCRL2...... 173 Figure S2. mCCRL2 is upregulated on macrophages activated by specific cytokines and/or TLR ligands...... 174 Figure S3. mCCRL2 KO mice display a normal contact hypersensitivity response to FITC...... 175 Figure S4. mCCRL2 is dispensable for maximal tissue swelling in high dose IgE- mediated passive cutaneous anaphylaxis...... 176 Figure S5. Histologic features of high dose IgE-dependent PCA reactions in WT BMCMC- vs. KO BMCMC-engrafted KitW-sh/Wsh mice...... 177 Figure S6. mCCRL2/L1.2 transfectants do not migrate to CCL2, CCL5, CCL7, or CCL8 in in vitro transwell chemotaxis...... 178

xvii

Figure S7. Lack of heterologous displacement of chemerin by other chemoattractants. 179 Figure S8. Radioligand binding competition...... 180 Figure S9. Chemerin and/or CCL2 do not trigger intracellular calcium mobilization in CCRL2/HEK293 transfectants...... 181 Figure S10. CCRL2 amino-terminal sequence alignment...... 182 Figure S11. Freshly isolated peritoneal mast cells do not express CMKLR1...... 182 Figure S12. mRNA expression of mCCRL2...... 183

Chapter 5 Figure 1. IL-17 expression by adipose tissue leukocyte subsets...... 203 Figure 2. Diet-induced differences in IL-17 expression by tissue T cell subsets...... 206 Figure 3. Enhanced susceptibility of IL-17 KO mice to dietary obesity...... 209 Figure 4. IL-17 inhibits adipogenesis in 3T3-L1 preadipocytes...... 213 Figure 5. IL-17 inhibits induction of adipocyte ...... 216 Figure 6. IL-17 inhibits glucose uptake in vitro and improves metabolic parameters before the onset of obesity...... 219 Figure 7. Obesity with age reverses protection from metabolic syndrome conferred by IL-17 deficiency...... 221 Figure 8. Evaluation of diet-induced obesity, glucose tolerance, insulin tolerance, and adipose tissue leukocyte cytokine expression in γδ T cell deficient mice. ....224

Chapter 6 Figure 1. Proposed model of chemerin and IL-17 in tissue inflammation, obesity, and metabolism...... 250

xviii

LIST OF TABLES

Number Page

Chapter 3

Table I. Clinical EAE in actively immunized WT and CMKLR1 KO mice ...... 87 Table II. Histological EAE in actively immunized WT and CMKLR1 KO mice ...... 89

xix

CHAPTER 1 : INTRODUCTION

Obesity: A brief history

When ancestors to modern day humans inhabited the Earth 200,000 years ago (1),

the ability to store energy in the form of body fat during periods of abundance was

beneficial for survival. Storing extra energy in a form that could be easily carried and used allowed humans to survive times of famine. Those who were more efficient at converting surplus calories to fat during times of caloric excess may have had a selective advantage over those who were not when food became scarce (2).

As humans advanced, they became more capable hunters and gatherers. Weapons facilitated the hunt for larger, nutrition-rich prey, while tools were used to more efficiently collect and process food (3). The use of fire for cooking allowed humans to chew, digest, and tolerate tough or potentially toxic items (4). Such advancement led to the consumption of more diverse dietary sources, thus increasing potential food supply and decreasing the chances of malnutrition and death. Evidence suggests foraging peoples had the most nutrition-rich diet throughout human history (5). However, before the advent of agriculture approximately 10,000 years ago (6), early humans were relatively lean compared to their modern day counterparts. This may have been in part due to a highly active hunter/gatherer lifestyle and meager to modest availability of food

(5, 7). As such, obesity was likely a rarity among early humans and is theorized to have been viewed as a sign of wealth, prosperity, and fertility (8). Evidence for this theory may be seen in the various prehistoric “Venus figurines” excavated from numerous

1

archaeological sites from areas inhabited by ancient humans. These figurines, ranging from 20,000 to 40,000 years old, depict female forms and may represent some of the oldest known examples of figurative art. Although the Venus figurines span thousands of years in age, making them seemingly unrelated to one another, a strikingly common characteristic among them is the representation of enlarged features and, controversially, apparent obesity (9, 10, 11, 12, 13, 14).

The introduction and development of agriculture marked a significant turning point in human history, as it led to the establishment of early and sustainable civilizations. Farming crops and animals allowed humans to produce food in abundant quantities, thus making foraging obsolete. With less need to travel large distances to gather food, humans were able to settle and establish villages, towns, and cities.

Nevertheless, food stores were not impervious to depletion; cycles of abundance and famine were still present as crops were susceptible to pests and failure while animals were susceptible to plague and death. Additionally, agriculture led to a reduction in the types of food humans ate, thus restricting diet and limiting nutritional exposure (6). Fossil evidence indicates that nutritional deficiencies were common among farming civilizations throughout history (15, 16). As agriculture and commerce developed, and social classes stratified, access to a nutrition-rich diet was more common among the wealthy (17). In some cultures, food was associated with wealth, and obesity was a sign of prosperity and affluence (18).

2

Since the introduction of agriculture, obesity prevalence has risen as

advancements made harvesting crops or tending animals easier, thus making food more

available and affordable. Indeed, modern farming techniques allow industrialized nations

to harvest and process nutrient-dense food with much less human effort compared to earlier farming practices. In developed and developing societies, where calorically-rich foods are easily obtainable to the masses, obesity is no longer a hallmark of wealth and is present at all socio-economic levels (19). Ironically, studies indicate obesity prevalence is now higher among lower social classes (20). A myriad of factors are likely to contribute to the rapid rise in obesity prevalence. One popular, yet controversial theory proposed by

James Neel in 1963 suggests that selective pressures favored propagation of “thrifty genes” that facilitated energy hoarding in the form of fat stores in ancestral humans. This

“thrifty phenotype” was well suited for periods of famine. However, in the face of abundant food supply and an ever increasing sedentary life style, a thrifty phenotype in modern humans may be conducive to the development of obesity and its negative impacts on human health (e.g. type II diabetes) (21). Another theory, recently proposed by John

Speakman, suggests that the absence of selective pressures, possibly due to human dominance in the animal kingdom, on genes predisposing to obesity may have lead to genetic drift. These so-called “drifty genes” resulted in random mutations which could either increase or decrease obesity predisposition, and may explain person-to-person variability in the development of obesity (20).

3

Obesity, metabolism, and inflammation

The first known reference to obesity as a danger to human health is thought to

have come from ancient Greece, approximately 450 years BC. The Greek physician

Hippocrates wrote at length about nutrition and its benefit to health. However, he also commented on the dangers of over-eating, indicating that death was more common among obese individuals than in lean individuals (8, 22). In his writings, Hippocrates stated: “Corpulence is not only a disease itself, but the harbinger of others.”

Obesity is one of several factors including , hyperglycemia, high blood pressure, and dyslipidaemia, collectively known as the metabolic syndrome, which are associated with an increased risk of developing cardiovascular heart disease and/or (23). The World Health Organization considers obesity and its associated health risks to be a global epidemic (24). In fact, in the United States alone, obesity prevalence is at a historic high, with approximately 1/3 of adults (25) (Fig. 1A) and 1/5 of children (26) (Fig. 1B) classified as clinically obese. If one considers overweight, then nearly 70% of U.S. adults are either overweight or obese (25) (Fig. 1A). Worldwide, it is estimated that approximately 1.6 billion adults are overweight, with 400 million of them being clinically obese. By 2015, these numbers are projected to rise to 2.3 billion and 700 million, respectively. The World Health Organization estimates obesity may account for

2-7% of total health care costs in developed countries, however, obesity comorbidities are not accounted for in this estimate, and thus obesity-related health care costs are likely

4

much higher. To illustrate, cardiovascular disease is not only strongly linked to obesity, but it is the world’s leading cause of death, claiming 17 million lives each year (24, 27).

Figure 1. Obesity prevalence in the United States between 1960-2006 Data obtained from the National Health and Nutrition Examination Survey between 1960 - 2006 (28) and (26) were used. (A) Obesity, overweight, and healthy weight prevalence in U.S. adults between 20-74 years old (*data for healthy weight from 2007-2008 is estimated from overweight and obese data). (B) Overweight prevalence in U.S. adolescents between 6-19 years old (**data for 2007-2008 is not yet determined). 5

In addition to metabolic and cardiovascular sequelae, obesity is also linked to an

increased risk of developing certain diseases such as osteoarthritis (29), fibromyalgia (30,

31), asthma (32), and various cancers (33, 34). Interestingly, obesity may also predispose

to development of certain autoimmune diseases such as psoriasis (35) and multiple

sclerosis (MS) (36, 37). To illustrate, a recent study found that females that were obese

during adolescence were twice as likely to have developed MS as compared to adolescent

women who were considered lean. Adult onset obesity, however, did not appear to affect

MS risk (36). Another study showed that within a study group of United States veterans

suffering from multiple sclerosis, approximately 40% were overweight and

approximately 20% were obese, indicating overweight or obesity was present in over

60% of MS veteran patients (38). Moreover, recent studies assessing the effect of diet-

induced obesity (DIO) on experimental autoimmune encephalomyelitis (EAE), a mouse

model of multiple sclerosis, indicate obese mice are more susceptible to developing EAE

than lean controls. The mechanisms behind the link between DIO and EAE severity are

still unknown, but the authors noted that DIO was associated with an increase in the

presence of peripheral lymphoid and central nervous system CD4+ interleukin-17 (IL-17)

expressing helper T cells (TH17 cells) during EAE (37). Another study found genetically

or diet-induced obese mice produced more IL-17 in the peritoneal cavity upon zymosan-

induced peritonitis compared to lean controls. Obese mice had a significantly higher

percentage of IL-17 expressing neutrophils recruited to the peritoneal cavity after

peritonitis induction and fat from obese peritoneal cavities expressed higher levels of IL-

6

17 mRNA. The authors concluded that neutrophil expression of IL-17 was increased in obese mice during zymosan-induced peritonitis and may contribute to inflammatory responses (39).

IL-17 is a pro-inflammatory cytokine expressed by T cells in response to infection and also during the establishment and progression of autoimmune diseases. Although IL-

17 is typically thought to be primarily expressed by CD4 T cells, studies indicate that other cell types can express IL-17, including CD8 helper T cells, natural killer T cells, natural killer cells, and γδ T cells. IL-17 is able to induce expression of other pro- inflammatory cytokines (e.g. L-6, G-CSF, GM-CSF, IL-1β, TNF-α) and (e.g.

IL-8, GRO-α, MCP-1) from non-T cells (e.g. fibroblasts, epithelial cells, macrophages).

IL-17 signaling may result in immune cell recruitment and activation, anti-microbial /peptide synthesis, cellular proliferation, and tissue remodeling (40).

A growing number of studies indicate IL-17 expression is positively correlated with obesity, whether it be systemic expression (41), peripheral tissue and lymphoid T cell expression (37), and, surprisingly, adipose tissue T cell expression (Chapter 5). These studies situate IL-17 as part of a growing list of pro-inflammatory cytokines and chemokines that are systemically upregulated during obesity. Pro-inflammatory cytokines that share a similar expression pattern as IL-17 during obesity include IL-6, IL-8, IL-18,

MIF, and TNFα (42) (Fig. 2). The upregulation of these cytokines has led to the hypothesis that obesity is associated with chronic, sub-clinical systemic inflammation. It is also believed that this systemic inflammation may contribute to the metabolic

7

syndrome and various diseases linked to obesity (43). For instance, IL-6 is a cytokine that has a pro-inflammatory role in several diseases including atherosclerosis (44), systemic lupus erythematosus (45), rheumatoid arthritis (46), and, importantly, diabetes (47). Not only is IL-6 upregulated during obesity, but chronic exposure to IL-6 is thought to induce insulin resistance both in vitro and in vivo, potentially contributing to type II diabetes development (48). Interestingly, IL-6 appears to have anti-obesogenic properties in mice:

IL-6 KO mice are more susceptible to mature-onset obesity (49). TNFα is similar to IL-6 in that it too may impair insulin sensitivity while having cachectic (induction of weight loss) properties (50).

Adipose tissue leukocyte infiltration

Adipose tissue is not merely a site for energy storage: it is a major endocrine organ able to express a number of factors capable of affecting metabolic and immunological processes. Although adipose tissue greatly contributes to the systemic expression of pro-inflammatory cytokines during obesity, the majority of adipose tissue cytokine expression takes place in the stromal vascular fraction (SVF) (42). The SVF is comprised of non-adipocyte cells, including fibroblasts (51), preadipocytes (52), endothelial cells (53), stem cells (54), dendritic cells (55), macrophages (56), neutrophils

(57), mast cells (58), B-cells (59), and T cells (60). Immune cells make up a significant portion of the SVF and numerous studies show obese fat contains significantly more immune cells than lean fat (56, 61, 62, 63, 64, 65). It is believed that recruitment and

8

possible activation of immune cells to obese fat may heavily contribute to systemic inflammation observed in obesity (66) (Fig. 2).

The mechanisms by which fat recruits leukocytes are not fully understood, however, it is believed that tissue necrosis in obese fat may play a role in initiating and exacerbating immune cell recruitment (67, 68). Adipocytes undergo hypertrophy as they collect excess free fatty acids and incorporate them into triglycerides, which are the major form of fatty acid storage in adipocytes. As the main role for adipocytes is energy storage, the triglyceride laden lipid droplet in mature adipocytes comprises the majority of their cellular volume. Adipocytes are unilocular, with the nucleus flattened by the lipid droplet and located near the periphery of the cell (69). The lipid droplet is sequestered from the cytoplasm by perilipin, a protein that coats the lipid droplet, and its size is regulated by various enzymes responsible for lipid metabolism such as hormone sensitive lipase (HSL) and fatty acid binding protein (FABP) (70). Under excessive caloric intake and load, adipocyte hypertrophy leads to extensive cellular remodeling: as the lipid droplet expands, it pushes the nucleus and organelles of the cell closer to the periphery. It is thought that the hypertrophic and mechanical stresses on the various organelles in obese adipocytes may impair proper cellular function and metabolism, leading to cell death. Impairment of the function of the endoplasmic reticulum, particularly the expression and trafficking of , may occur during adipocyte hypertrophy (71, 72).

Hypoxic (low oxygen) stress may also lead to adipocyte necrosis. As obesity progresses, occurs in adipose tissue, which is necessary for adipose tissue expansion.

9

However, despite the formation of new blood vessels, adipose tissue is poorly

vascularized and studies revealed obese adipose tissue is more hypoxic than lean adipose

tissue. As adipocytes expand, their size may preclude them from efficient oxygen

exposure due to limits on oxygen diffusion and/or adipocyte distance from the

vasculature (73). Finally, oxidative stress may be a factor in obese adipose tissue necrosis

as excess NADPH, a metabolite important for lipogenesis, may donate an electron to

- molecular oxygen through NADPH oxidase, generating a superoxide ion (O2 ). This free

radical may then react with proteins, thus oxidizing them and potentially interfering with

proper enzymatic functions (74).

Necrosis leads to the release of danger signals – self proteins with damage

associated molecular patterns (e.g. heat shock proteins), intracellular metabolites (e.g.

uric acid), or DNA/nucleosome/chromatin complexes – that are able to activate innate

tissue resident immune cells. These tissue resident immune cells are usually macrophages

or dendritic cells which can detect danger signals via receptors on their cell surfaces.

Upon detection, the immune cells become activated where they can then release pro- inflammatory cytokines which act upon nearby vessel endothelial cells, inducing them to express chemokines that can recruit non-tissue resident immune cells to the site of necrosis (75). It is speculated that obese adipose tissue may be associated with higher concentrations of certain heat shock proteins (e.g. HSP 60), possibly released by necrotic adipocytes, which may activate macrophages and may induce expression of pro-

10

inflammatory cytokines and chemokines from preadipocytes and mature adipocytes (76,

77, 78, 79).

Chemoattractants and their receptors are critically important for the recruitment of immune cells to fat. Indeed, compared to lean fat, obese fat expresses significantly higher levels of chemoattractants such as IL-8, MCP-1 (CCL2), CCL3, CCL5, CCL7, and

CCL11. Receptors for these chemoattractants are also upregulated in obese fat (42, 64,

80). The majority of published studies on immune cell recruitment to adipose tissue focus primarily on macrophages. A number of studies have demonstrated that obese fat contains significantly more adipose tissue macrophages (ATM) than lean fat, and the recruitment of ATMs may be partially mediated by receptor/ligand pairs, including

CCR2/MCP-1 (64) (CCL7 and CCL11 are also potential CCR2 ligands in fat (80)) and

IL-8/CXCR2 (81). The vast majority of macrophages in obese adipose tissue are found surrounding necrotic adipocytes, a phenomenon which is relatively rare in lean adipose tissue. Necrotic cell associated macrophages form what are known as “crown-like structures” – a mass of macrophages that, when detected histologically, resemble a halo or crown surrounding sites of cell death. Eventually, these macrophages fuse to form multinucleated giant cells. These macrophages appear to be activated, as indicated by the upregulation of the activation marker MAC-2, and appear to scavenge lipid that is released by dead adipocytes. The lipid that is released by necrotic cells may participate in macrophage activation and subsequent secretion of pro-inflammatory cytokines (68, 82).

11

As mentioned earlier, CCR2 is important for macrophage recruitment to fat.

Deficiency of CCR2 leads to reduced macrophage content in adipose tissue and improved insulin sensitivity, presumably due to the lack of activated macrophages in obese adipose tissue (64, 83, 84). Furthermore, overexpression of MCP-1 in adipose tissue leads to increased ATM content and worse insulin resistance (83, 85). It is clear that ATMs play an important role in contributing to systemic inflammation during obesity, however more recent studies indicate T cells may also be recruited to fat and may contribute to local and systemic inflammation as well.

12

Figure 2. Adipose tissue leukocyte infiltration and inflammation model From top left, following clockwise – Under excess caloric load, lean adipose tissue expands (hypertrophy) as it incorporates fatty acids. Hypertrophic, hypoxic, endoplasmic reticulum, or oxidative stresses lead to adipose tissue necrosis. Necrotic adipose tissue releases danger signals, stimulating surrounding cells to release chemoattractants into the microenvironment. Stimulated vascular endothelial cells release chemoattractants (e.g. MCP-1), recruiting leukocytes to damaged tissue. Recruited cells are activated (possibly via altered self-proteins, lipid ligands, or danger signals) and secrete pro-inflammatory cytokines. Secreted cytokines can enter the blood stream, increasing systemic inflammation.

13

Adipose tissue T cells

A collection of studies published together, in 2009, revealed an intricate balance

between various T cell subsets in fat and suggested they can play an active role in

moderating adipose tissue inflammation. A study by Nishimura et al. focused on the role

of CD8 T cells in fat inflammation and found CD8 T cells are recruited to fat before

ATMs. Antibody mediated depletion of CD8 T cells attenuated ATM recruitment,

adipose tissue inflammation, and obesity-related insulin resistance. It was found that

adipose CD8 T cells can interact with adipose tissue to facilitate monocyte differentiation

to macrophages in vitro and this effect was significantly enhanced with CD8 T cells

isolated from obese fat. These T cells could also secrete chemokines (IP-10, MCP-1,

MCP-3, RANTES) that could potentially recruit macrophages to the fat (63). Along with

this study, Winer et al. demonstrated that CD4 T cells populate obese adipose tissue, but

they appear to have dual roles. The authors indicate a high CD4 TH1 (IFNγ expressing) to regulatory CD4 T cell (Foxp3 expressing) ratio and a higher TH1 to TH2 (GATA-3

expressing) ratio compared to lean adipose tissue. TH1 cells are generally considered pro-

inflammatory while regulatory T cells and TH2 cells can attenuate inflammation. The

authors speculate that as obesity progresses, TH1 cell content increases in adipose tissue,

shifting the adaptive immune cell repertoire in fat to a more pro-inflammatory phenotype.

Surprisingly, the authors found a limited Vα segment usage in the T cell receptors of TH1

cells from fat, suggesting a clonal expansion of a finite number of CD4 T cells (65).

Although not fully explored, these data also suggest a potential autoimmune component

14

in the adaptive immune cell response in obese fat. Finally, similar to the Winer studies,

Feuerer et al. found decreased regulatory T cell (Tregs) content in adipose tissue from

obese fat compared to lean fat. They also found CD4 T cells (both conventional and

Tregs) from fat had restricted Vα usage compared to CD4 T cells from lymph nodes.

Ablating Tregs in lean mice lead to insulin resistance while in situ expansion of Tregs

mildly reversed insulin resistance in obese mice (86).

Previous studies have also shown that other types of T cells are either recruited to,

expand in, or are retained in obese fat. Work from Caspar-Bauguil et al. demonstrates

that the percent of natural killer cells decreases as total T cells increases in obese fat,

possibly reflecting the shifting immune cell repertoire previously described. Intriguingly,

the percentage of γδ T cells significantly increases in obese fat (61, 62). γδ+ T cells are a

minor T cell subset that often responds to self-ligands (87, 88, 89) and participates in

immunoregulation (90). Relatively little is known about the activation of these cells, but

it is thought that they may participate in recognizing danger signals from stressed tissue

and can act as rapid initiators of inflammation. Most known γδ TCR ligands are self-

determinants associated with stress responses (e.g. T10/T22 and MICA/B) (91) and

studies demonstrate that γδ T cells are important for establishing inflammation and are

responsible for IL-17A expression during the first several days of infection or during the

onset of autoimmune disease (92, 93). It is unclear whether adipose tissue γδ T cells can

participate in systemic inflammation during obesity.

15

Taken together, it becomes clear that the immune response in fat that occurs during obesity is complex; however chemoattractants and cytokines are critically

important in contributing to immune cell recruitment and to the metabolic syndrome (Fig.

2).

Chemerin and chemerin receptors in inflammation, obesity, and metabolism

Adipogenesis and adipocyte function are dependent on genetic and environmental factors, and are regulated by the action of (e.g. or ),

hormones, lipids, sugar metabolites, cytokines, and chemoattractants (49, 69, 94, 95). In

addition to their roles in immune cell trafficking, certain chemoattractants play important

roles in directly regulating adipocyte function, development, or metabolism. One

noteworthy chemoattractant recently discovered to be important in adipocyte biology is

chemerin.

Chemerin is the product of the TIG-2 gene (tazarotene-induced gene-2), which

was originally found to be expressed at high levels in non-lesional psoriatic skin and at

low levels in lesional psoriatic skin. The transcript was found to be upregulated in

response to the anti-psoriatic retinoid tazarotene in psoriatic lesions (96). The chemerin protein, however, was isolated from ascitic fluid (ovarian carcinoma), inflamed synovial fluid, hemofiltrate, and normal serum (97, 98, 99). Chemerin, a heparin binding protein, initially exists in its pro-form, which is 163 amino acids long. Cleavage of prochemerin, by serine proteases of inflammatory, coagulation, and fibrinolytic cascades, results in the

16

loss of the last 6-11 C-terminal amino acids (100) (Fig. 3). This cleavage leads to a potent increase in chemoattractant activity, resulting in the increased migration of CMKLR1 bearing cells (macrophages and pDCs) to chemerin (101, 102).

Figure 3. Proteolytic cleavage of chemerin Proteolytic cleavage at multiple sites in the carboxyl-terminal region of chemerin results in activation of the chemoattractant. The terminal 14 residues of human pro-form chemerin are shown; the orthologous cleavage site reported for hamster chemerin is indicated. CHO CM, Chinese hamster ovary cell line conditioned media. Cleavage sites and cleavage agents as reported from (98, 99, 101, 102, 103). Adapted from Zabel BA, Zuniga L, Ohyama T, Allen SJ, Cichy J, Handel TM, Butcher EC. Chemoattractants, extracellular proteases, and the integrated host defense response. Exp Hematol. 2006. Exp Hematol. 2006 Aug;34 (8):1021-32. (100)

CMKLR1 is a G-protein coupled receptor (GPCR) that was initially identified

through its homology with several chemoattractant receptors, including the C5a and

formyl-Met-Leu-Phe receptors (~35-40% ). The receptor is believed

to adopt the typical seven-transmembrane α-helix arrangement common among GPCRs

(104, 105, 106). Work by us and others demonstrate human CMKLR1 is specifically 17

expressed on in vitro monocyte derived dendritic cells, ex vivo macrophages, circulating

plasmacytoid dendritic cells (pDCs), and ascitic fluid macrophages (97, 107, 108)

(Chapter 2). Chemerin was only recently discovered to be a natural ligand for CMKLR1

(97, 98, 99). An ortholog of CMKLR1, with approximately 80% identity with its human

counterpart, exists in mice (108, 109). We report that in mouse, CMKLR1 is expressed

by peritoneal macrophages, which can migrate to chemerin in vitro, and by a small

population of blood borne macrophages. On these cells, it is upregulated by TGFβ (108)

and dexamethasone (unpublished data), and it is downregulated by pro-inflammatory

cytokines and TLR ligands (108). Although chemerin expression is upregulated either

systemically in certain inflammatory diseases (110) or in inflamed tissues (111, 112), its

role in inflammation is unclear. A report from our group suggests chemerin may

exacerbate disease progression in EAE via CMKLR1 (112) (Chapter 3). In contrast, chemerin may have an anti-inflammatory role in LPS-induced lung injury (113).

An additional receptor for chemerin, named C-C -like 2

(CCRL2), was very recently discovered by our group (114) (Chapter 4). CCRL2 is a

GPCR with homology to various members of the CC chemokine receptor subfamily.

Typical members of this family carry a consensus DRYLAIV sequence motif in their second cytoplasmic loop, which is important for G-protein binding and signaling (115).

In contrast, CCRL2 contains an atypical sequence in place of the DRYLAIV motif

(QRYLVFL in human CCRL2; QRYRVSF in mouse CCRL2), suggesting it may not bind G-proteins. Mouse CCRL2 (mCCRL2) RNA is upregulated in mouse macrophages,

18

astrocytes, and microglia in response to LPS (116, 117) and is expressed by astrocytes, microglia, and macrophages in the central nervous system during the onset of EAE (118).

Human CCRL2 (huCCRL2) is expressed by circulating T cells, neutrophils, monocytes, and monocyte derived macrophages and dendritic cells (DC). Activation of these cells can lead to the upregulation of huCCRL2 (119). Synovial fluid neutrophils from rheumatoid arthritis patients express huCCRL2 and its expression is upregulated on blood neutrophils treated with LPS or TNFα (120).

In addition to discovering chemerin as a natural ligand for CCRL2, we found that

CCRL2 is expressed by mouse mast cells and can enhance tissue swelling and tissue leukocyte infiltration during mast cell-dependent passive cutaneous anaphylaxis.

Interestingly, chemerin binding to CCRL2 did not elicit common responses associated with functional GPCR signaling including calcium flux, receptor internalization, or leukocyte migration. However, CCRL2 is able to bind to and increase local concentrations of bioactive chemerin (114).

Collaborative work from the laboratory of Christopher Sinal and our group recently discovered an unexpected role for chemerin and CMKLR1 in adipocyte function and metabolism. A screen for chemerin and CMKLR1 mRNA revealed high expression of both genes in white adipose tissue, with the majority of expression found in the adipocyte fraction. Both chemerin and CMKLR1 expression were increased upon adipocyte differentiation, and CMKLR1 expressing L1.2 cells could migrate to adipocyte conditioned media, but not to preadipocyte conditioned media, while CMKLR1 deficient

19

control cells did not migrate to either media, suggesting mature adipocytes expressed

chemerin, and may also express proteases that can activate chemerin. Knockdown of

chemerin and CMKLR1 in preadipocytes significantly impaired adipogenesis and adipocyte . However, knockdown of chemerin in mature adipocytes did not alter their morphology, suggesting chemerin signaling is important for adipogenesis

and may not play a role in maintaining the differentiated state of mature adipocytes. The

data implicates chemerin as a novel able to modulate adipocyte maturation

(95).

Subsequent studies from the Sinal group have revealed systemic chemerin levels

were elevated in obese mice deficient for either leptin or the leptin receptor.

Administration of chemerin in these and diet-induced obese mice worsened glucose

intolerance and decreased glucose tissue uptake in leptin receptor deficient mice (121).

Leptin is a satiety factor that can act on the nervous system to modulate feeding behavior.

Mice deficient for leptin or the leptin receptor are hyperphagic and develop obesity much

faster than control mice (122). The role for chemerin in metabolism is unclear – one

report indicates chemerin enhances in vitro insulin stimulated glucose uptake in mouse

adipocytes (123) while another report indicates chemerin impairs in vitro insulin

stimulated glucose uptake in human skeletal muscle cells (124). Although chemerin

protein is associated with obesity (121, 125, 126, 127, 128, 129), it is interesting to note

that CCRL2 mRNA expression appears to be upregulated while CMKLR1 transcript expression appears to be downregulated in white adipose tissue from obese leptin and

20

leptin receptor deficient mice, compared to control mice. Of further interest, in the same mice, CMKLR1 expression is significantly higher in skeletal muscle compared to controls (121). The in vivo expression discrepancies of CMKLR1 between adipose tissue and skeletal muscle may help explain the apparent difference in chemerin responses during in vitro insulin stimulated glucose uptake between adipocytes and skeletal muscle cells. It is unclear what role, if any, CCRL2 has in adipocyte biology. It is attractive to postulate that CCRL2 may act to bind chemerin and present it to CMKLR1 on the surface of adipocytes or may facilitate macrophage trafficking to or within fat.

Significance

Adipose tissue was once considered a site which sole purpose was for energy storage. However, it is becoming abundantly clear that obesity and inflammation are intimately linked. The immune response that occurs in obese adipose tissue exhibits classic signs of inflammation, characterized by leukocyte infiltration, activation, and secretion of inflammatory cytokines and chemokines (130). With its rising prevalence throughout the world, and its link to the development of metabolic, cardiovascular, and inflammatory diseases, obesity has become a topic of much scrutiny and research.

Therapeutic approaches to treating obesity may have other beneficial effects in preventing or ameliorating related diseases. However, the mechanisms that regulate the interplay between the and adipose tissue, and vice versa, are poorly understood.

21

The original objective of this work was to elucidate the expression of CMKLR1 in vivo and further define the role of chemerin in inflammation. The study presented in chapter 2 was designed to assess the expression, regulation, and functional responses of leukocyte expressed CMKLR1. Work described in chapter 3 was performed to determine if CMKLR1 and/or chemerin contributed to experimental autoimmune encephalomyelitis. Data presented in chapter 4 was generated from experiments focused on characterizing CCRL2 binding to chemerin and determining the role of CCRL2 in passive cutaneous anaphylaxis. During the course of these studies, our laboratory collaborated in clarifying the role of CMKLR1 and chemerin in adipocyte development.

We, unexpectedly, also found a role for IL-17 in adipocyte biology and metabolism – studies which are described in chapter 5. The work presented here was either significantly contributed to by my work and work product (Chapters 2,3 and 4) or was primarily of my personal efforts (Chapter 5).

Taken together, we have shown that chemerin may contribute to leukocyte trafficking to inflamed tissue via CMKLR1, a process which may be facilitated by

CCRL2 – a decoy receptor that can bind chemerin. CCRL2 may act to present chemerin to CMKLR1. Furthermore, chemerin and CMKLR1 may participate in adipose tissue development, a process which could contribute to systemic inflammation during the progression of obesity. We have also shown that the proinflammatory cytokine, IL-17, can negatively regulate adipocyte development and glucose homeostasis. Overall, these

22

studies situate chemerin, CMKLR1, CCRL2, and IL-17 as possible targets for therapeutic intervention in treating inflammation, obesity, and metabolic disorders.

23

References

1. Pakendorf, B., and M. Stoneking. 2005. Mitochondrial DNA and human evolution. Annu Rev Genomics Hum Genet 6:165-183. 2. Stinson, Bogin, Huss-Ashmore, and O'Rourke. 2000. Human Biology: An Evolutionary, Biocultural Perspective. Wiley-Liss, New York. 3. Ambrose, S. H. 2001. Paleolithic technology and human evolution. Science 291:1748-1753. 4. Weiner, S., Q. Xu, P. Goldberg, J. Liu, and O. Bar-Yosef. 1998. Evidence for the use of fire at zhoukoudian, china. Science 281:251-253. 5. Eaton, S. B., and M. Konner. 1985. Paleolithic nutrition. A consideration of its nature and current implications. N Engl J Med 312:283-289. 6. Diamond, J. 2002. Evolution, consequences and future of plant and animal domestication. Nature 418:700-707. 7. Cordain, L., S. B. Eaton, J. B. Miller, N. Mann, and K. Hill. 2002. The paradoxical nature of hunter-gatherer diets: meat-based, yet non-atherogenic. Eur J Clin Nutr 56 Suppl 1:S42-52. 8. Haslam, D. 2007. Obesity: a medical history. Obes Rev 8 Suppl 1:31-36. 9. Szombathy, J. 1909. Die Aurignacienschichten in Löss von Willendorf. Korrespondenzblatt der Deutschen Gesellschaft für Anthropologie, Ethnologie, und Urgeschichte XL:85-88 10. Králík, M., V. Novotný, and M. Oliva. 2002. Fingerprint on the Venus of Dolní Věstonice I. Anthropologie 40:107–113. 11. Conard, N. J. 2009. A female figurine from the basal Aurignacian of Hohle Fels Cave in southwestern Germany. Nature 459:248-252. 12. McDonnell, A. 2009. Ancient ivory figurine deserves a more thoughtful label. Nature 459:909. 13. Mellars, P. 2009. Archaeology: Origins of the female image. Nature 459:176-177. 14. Shewan, L. 2006. The Venus figurines. Int J Cardiol 113:439. 15. Ruff, C. B., E. Trinkaus, and T. W. Holliday. 1997. Body mass and encephalization in Pleistocene Homo. Nature 387:173-176. 16. Panter-Brick, Layton, and Rowley-Conwy. 2001. Hunter-Gatherers: An Interdisciplinary Perspective. Cambridge University Press, Cambridge. 17. Strickland, and Shetty. 1998. Human Biology and Social Inequality. Cambridge University Press, Cambridge. 18. ME, D. 1999. Nutrition and politics in prehistory. Annu Rev Anthropologie:1-25. 19. Prentice, A. M. 2006. The emerging epidemic of obesity in developing countries. Int J Epidemiol 35:93-99. 20. Speakman, J. R., H. Walker, L. Walker, and D. M. Jackson. 2005. Associations between BMI, social strata and the estimated energy content of foods. Int J Obes (Lond) 29:1281-1288.

24

21. Neel, J. V. 1962. Diabetes mellitus: a "thrifty" genotype rendered detrimental by "progress"? Am J Hum Genet 14:353-362. 22. Lavie, C. J., and R. V. Milani. 2003. Obesity and cardiovascular disease: the hippocrates paradox? J Am Coll Cardiol 42:677-679. 23. Alberti, K. G., P. Zimmet, and J. Shaw. 2006. Metabolic syndrome--a new world- wide definition. A Consensus Statement from the International Diabetes Federation. Diabet Med 23:469-480. 24. http://www.who.int/dietphysicalactivity/publications/facts/obesity/en/. 25. Flegal, K. M., M. D. Carroll, C. L. Ogden, and L. R. Curtin. 2010. Prevalence and trends in obesity among US adults, 1999-2008. JAMA 303:235-241. 26. Ogden, C. L., M. D. Carroll, L. R. Curtin, M. M. Lamb, and K. M. Flegal. 2010. Prevalence of high in US children and adolescents, 2007-2008. JAMA 303:242-249. 27. http://www.who.int/mediacentre/factsheets/fs311/en/index.html. 28. http://www.cdc.gov/nchs/data/hus/hus09.pdf#072. 29. Reijman, M., H. A. Pols, A. P. Bergink, J. M. Hazes, J. N. Belo, A. M. Lievense, and S. M. Bierma-Zeinstra. 2007. Body mass index associated with onset and progression of osteoarthritis of the knee but not of the hip: the Rotterdam Study. Ann Rheum Dis 66:158-162. 30. Neumann, L., E. Lerner, Y. Glazer, A. Bolotin, A. Shefer, and D. Buskila. 2008. A cross-sectional study of the relationship between body mass index and clinical characteristics, tenderness measures, quality of life, and physical functioning in fibromyalgia patients. Clin Rheumatol 27:1543-1547. 31. Shapiro, J. R., D. A. Anderson, and S. Danoff-Burg. 2005. A pilot study of the effects of behavioral weight loss treatment on fibromyalgia symptoms. J Psychosom Res 59:275-282. 32. Beuther, D. A. 2010. Recent insight into obesity and asthma. Curr Opin Pulm Med 16:64-70. 33. Brown, K. A., and E. R. Simpson. 2010. Obesity and breast cancer: progress to understanding the relationship. Cancer Res 70:4-7. 34. Percik, R., and M. Stumvoll. 2009. Obesity and cancer. Exp Clin Endocrinol Diabetes 117:563-566. 35. Yosipovitch, G., A. DeVore, and A. Dawn. 2007. Obesity and the skin: skin physiology and skin manifestations of obesity. J Am Acad Dermatol 56:901-916; quiz 917-920. 36. Munger, K. L., T. Chitnis, and A. Ascherio. 2009. Body size and risk of MS in two cohorts of US women. Neurology 73:1543-1550. 37. Winer, S., G. Paltser, Y. Chan, H. Tsui, E. Engleman, D. Winer, and H. M. Dosch. 2009. Obesity predisposes to Th17 bias. Eur J Immunol 39:2629-2635. 38. Khurana, S. R., A. M. Bamer, A. P. Turner, R. V. Wadhwani, J. D. Bowen, S. L. Leipertz, and J. K. Haselkorn. 2009. The prevalence of overweight and obesity in veterans with multiple sclerosis. Am J Phys Med Rehabil 88:83-91. 25

39. Pini, M., and G. Fantuzzi. 2009. Enhanced production of IL-17A during zymosan- induced peritonitis in obese mice. J Leukoc Biol. 40. Gaffen, S. L. 2008. An overview of IL-17 function and signaling. Cytokine 43:402-407. 41. Sumarac-Dumanovic, M., D. Stevanovic, A. Ljubic, J. Jorga, M. Simic, D. Stamenkovic-Pejkovic, V. Starcevic, V. Trajkovic, and D. Micic. 2009. Increased activity of interleukin-23/interleukin-17 proinflammatory axis in obese women. Int J Obes (Lond) 33:151-156. 42. Fain, J. N. 2006. Release of interleukins and other inflammatory cytokines by human adipose tissue is enhanced in obesity and primarily due to the nonfat cells. Vitam Horm 74:443-477. 43. de Ferranti, S., and D. Mozaffarian. 2008. The perfect storm: obesity, adipocyte dysfunction, and metabolic consequences. Clin Chem 54:945-955. 44. Dubinski, A., and Z. Zdrojewicz. 2007. [The role of interleukin-6 in development and progression of atherosclerosis]. Pol Merkur Lekarski 22:291-294. 45. Tackey, E., P. E. Lipsky, and G. G. Illei. 2004. Rationale for interleukin-6 blockade in systemic lupus erythematosus. Lupus 13:339-343. 46. Nishimoto, N. 2006. Interleukin-6 in rheumatoid arthritis. Curr Opin Rheumatol 18:277-281. 47. Kristiansen, O. P., and T. Mandrup-Poulsen. 2005. Interleukin-6 and diabetes: the good, the bad, or the indifferent? Diabetes 54 Suppl 2:S114-124. 48. Kim, J. H., R. A. Bachmann, and J. Chen. 2009. Interleukin-6 and insulin resistance. Vitam Horm 80:613-633. 49. Wallenius, V., K. Wallenius, B. Ahren, M. Rudling, H. Carlsten, S. L. Dickson, C. Ohlsson, and J. O. Jansson. 2002. Interleukin-6-deficient mice develop mature-onset obesity. Nat Med 8:75-79. 50. Ryden, M., and P. Arner. 2007. Tumour necrosis factor-alpha in human adipose tissue -- from signalling mechanisms to clinical implications. J Intern Med 262:431-438. 51. Lin, C. S., Z. C. Xin, C. H. Deng, H. Ning, G. Lin, and T. F. Lue. 2010. Defining adipose tissue-derived stem cells in tissue and in culture. Histol Histopathol 25:807-815. 52. Sul, H. S. 2009. Minireview: Pref-1: role in adipogenesis and mesenchymal cell fate. Mol Endocrinol 23:1717-1725. 53. Curat, C. A., A. Miranville, C. Sengenes, M. Diehl, C. Tonus, R. Busse, and A. Bouloumie. 2004. From blood monocytes to adipose tissue-resident macrophages: induction of diapedesis by human mature adipocytes. Diabetes 53:1285-1292. 54. Zeve, D., W. Tang, and J. Graff. 2009. Fighting fat with fat: the expanding field of adipose stem cells. Cell Stem Cell 5:472-481. 55. Knight, S. C. 2008. Specialized perinodal fat fuels and fashions immunity. Immunity 28:135-138. 56. 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. 26

57. Elgazar-Carmon, V., A. Rudich, N. Hadad, and R. Levy. 2008. Neutrophils transiently infiltrate intra-abdominal fat early in the course of high-fat feeding. J Lipid Res 49:1894-1903. 58. 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. 59. Duffaut, C., J. Galitzky, M. Lafontan, and A. Bouloumie. 2009. Unexpected trafficking of immune cells within the adipose tissue during the onset of obesity. Biochem Biophys Res Commun 384:482-485. 60. Lumeng, C. N., I. Maillard, and A. R. Saltiel. 2009. T-ing up inflammation in fat. Nat Med 15:846-847. 61. Caspar-Bauguil, S., B. Cousin, M. Andre, M. Nibbelink, A. Galinier, B. Periquet, L. Casteilla, and L. Penicaud. 2006. Weight-dependent changes of immune system in adipose tissue: importance of leptin. Exp Cell Res 312:2195-2202. 62. Caspar-Bauguil, S., B. Cousin, A. Galinier, C. Segafredo, M. Nibbelink, M. Andre, L. Casteilla, and L. Penicaud. 2005. Adipose tissues as an ancestral immune organ: site-specific change in obesity. FEBS Lett 579:3487-3492. 63. Nishimura, S., I. Manabe, M. Nagasaki, K. Eto, H. Yamashita, M. Ohsugi, M. Otsu, K. Hara, K. Ueki, S. Sugiura, K. Yoshimura, T. Kadowaki, and R. Nagai. 2009. CD8+ effector T cells contribute to macrophage recruitment and adipose tissue inflammation in obesity. Nat Med 15:914-920. 64. 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. 65. Winer, S., Y. Chan, G. Paltser, D. Truong, H. Tsui, J. Bahrami, R. Dorfman, Y. Wang, J. Zielenski, F. Mastronardi, Y. Maezawa, D. J. Drucker, E. Engleman, D. Winer, and H. M. Dosch. 2009. Normalization of obesity-associated insulin resistance through immunotherapy. Nat Med 15:921-929. 66. Neels, J. G., and J. M. Olefsky. 2006. Inflamed fat: what starts the fire? J Clin Invest 116:33-35. 67. 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. 68. Strissel, K. J., Z. Stancheva, H. Miyoshi, J. W. Perfield, 2nd, J. DeFuria, Z. Jick, A. S. Greenberg, and M. S. Obin. 2007. Adipocyte death, adipose tissue remodeling, and obesity complications. Diabetes 56:2910-2918. 69. Feve, B. 2005. Adipogenesis: cellular and molecular aspects. Best Pract Res Clin Endocrinol Metab 19:483-499. 70. Carmen, G. Y., and S. M. Victor. 2006. Signalling mechanisms regulating . Cell Signal 18:401-408. 27

71. Ozcan, U., Q. Cao, E. Yilmaz, A. H. Lee, N. N. Iwakoshi, E. Ozdelen, G. Tuncman, C. Gorgun, L. H. Glimcher, and G. S. Hotamisligil. 2004. Endoplasmic reticulum stress links obesity, insulin action, and type 2 diabetes. Science 306:457-461. 72. Hotamisligil, G. S. 2005. Role of endoplasmic reticulum stress and c-Jun NH2- terminal kinase pathways in inflammation and origin of obesity and diabetes. Diabetes 54 Suppl 2:S73-78. 73. Wood, I. S., F. P. de Heredia, B. Wang, and P. Trayhurn. 2009. Cellular hypoxia and adipose tissue dysfunction in obesity. Proc Nutr Soc 68:370-377. 74. Furukawa, S., T. Fujita, M. Shimabukuro, M. Iwaki, Y. Yamada, Y. Nakajima, O. Nakayama, M. Makishima, M. Matsuda, and I. Shimomura. 2004. Increased oxidative stress in obesity and its impact on metabolic syndrome. J Clin Invest 114:1752-1761. 75. Rock, K. L., E. Latz, F. Ontiveros, and H. Kono. 2010. The sterile inflammatory response. Annu Rev Immunol 28:321-342. 76. Gulden, E., T. Marker, J. Kriebel, V. Kolb-Bachofen, V. Burkart, and C. Habich. 2009. Heat shock protein 60: evidence for receptor-mediated induction of proinflammatory mediators during adipocyte differentiation. FEBS Lett 583:2877-2881. 77. Gulden, E., S. Mollerus, J. Bruggemann, V. Burkart, and C. Habich. 2008. Heat shock protein 60 induces inflammatory mediators in mouse adipocytes. FEBS Lett 582:2731-2736. 78. Habich, C., K. Baumgart, H. Kolb, and V. Burkart. 2002. The receptor for heat shock protein 60 on macrophages is saturable, specific, and distinct from receptors for other heat shock proteins. J Immunol 168:569-576. 79. Marker, T., J. Kriebel, U. Wohlrab, and C. Habich. 2010. Heat shock protein 60 and adipocytes: characterization of a ligand-receptor interaction. Biochem Biophys Res Commun 391:1634-1640. 80. Huber, J., F. W. Kiefer, M. Zeyda, B. Ludvik, G. R. Silberhumer, G. Prager, G. J. Zlabinger, and T. M. Stulnig. 2008. CC chemokine and CC chemokine receptor profiles in visceral and subcutaneous adipose tissue are altered in human obesity. J Clin Endocrinol Metab 93:3215-3221. 81. Neels, J. G., L. Badeanlou, K. D. Hester, and F. Samad. 2009. Keratinocyte- derived chemokine in obesity: expression, regulation, and role in adipose macrophage infiltration and glucose homeostasis. J Biol Chem 284:20692-20698. 82. 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. 83. Kanda, H., S. Tateya, Y. Tamori, K. Kotani, K. Hiasa, R. Kitazawa, S. Kitazawa, H. Miyachi, S. Maeda, K. Egashira, and M. Kasuga. 2006. MCP-1 contributes to macrophage infiltration into adipose tissue, insulin resistance, and hepatic steatosis in obesity. J Clin Invest 116:1494-1505. 84. Lumeng, C. N., J. B. DelProposto, D. J. Westcott, and A. R. Saltiel. 2008. Phenotypic switching of adipose tissue macrophages with obesity is generated by spatiotemporal differences in macrophage subtypes. Diabetes 57:3239-3246. 28

85. Kamei, N., K. Tobe, R. Suzuki, M. Ohsugi, T. Watanabe, N. Kubota, N. Ohtsuka- Kowatari, K. Kumagai, K. Sakamoto, M. Kobayashi, T. Yamauchi, K. Ueki, Y. Oishi, S. Nishimura, I. Manabe, H. Hashimoto, Y. Ohnishi, H. Ogata, K. Tokuyama, M. Tsunoda, T. Ide, K. Murakami, R. Nagai, and T. Kadowaki. 2006. Overexpression of monocyte chemoattractant protein-1 in adipose tissues causes macrophage recruitment and insulin resistance. J Biol Chem 281:26602-26614. 86. Feuerer, M., L. Herrero, D. Cipolletta, A. Naaz, J. Wong, A. Nayer, J. Lee, A. B. Goldfine, C. Benoist, S. Shoelson, and D. Mathis. 2009. Lean, but not obese, fat is enriched for a unique population of regulatory T cells that affect metabolic parameters. Nat Med 15:930-939. 87. Shin, S., R. El-Diwany, S. Schaffert, E. J. Adams, K. C. Garcia, P. Pereira, and Y. H. Chien. 2005. Antigen recognition determinants of gammadelta T cell receptors. Science 308:252-255. 88. Groh, V., A. Steinle, S. Bauer, and T. Spies. 1998. Recognition of stress-induced MHC molecules by intestinal epithelial gammadelta T cells. Science 279:1737-1740. 89. Havran, W. L., Y. H. Chien, and J. P. Allison. 1991. Recognition of self antigens by skin-derived T cells with invariant gamma delta antigen receptors. Science 252:1430- 1432. 90. Girardi, M. 2006. Immunosurveillance and immunoregulation by gammadelta T cells. J Invest Dermatol 126:25-31. 91. Thedrez, A., C. Sabourin, J. Gertner, M. C. Devilder, S. Allain-Maillet, J. J. Fournie, E. Scotet, and M. Bonneville. 2007. Self/non-self discrimination by human gammadelta T cells: simple solutions for a complex issue? Immunol Rev 215:123-135. 92. Roark, C. L., P. L. Simonian, A. P. Fontenot, W. K. Born, and R. L. O'Brien. 2008. gammadelta T cells: an important source of IL-17. Curr Opin Immunol 20:353- 357. 93. Jensen, K. D., X. Su, S. Shin, L. Li, S. Youssef, S. Yamasaki, L. Steinman, T. Saito, R. M. Locksley, M. M. Davis, N. Baumgarth, and Y. H. Chien. 2008. Thymic selection determines gammadelta T cell effector fate: antigen-naive cells make interleukin-17 and antigen-experienced cells make . Immunity 29:90- 100. 94. Fajas, L. 2003. Adipogenesis: a cross-talk between cell proliferation and cell differentiation. Ann Med 35:79-85. 95. Goralski, K. B., T. C. McCarthy, E. A. Hanniman, B. A. Zabel, E. C. Butcher, S. D. Parlee, S. Muruganandan, and C. J. Sinal. 2007. Chemerin, a novel adipokine that regulates adipogenesis and adipocyte metabolism. J Biol Chem 282:28175-28188. 96. Nagpal, S., S. Patel, H. Jacobe, D. DiSepio, C. Ghosn, M. Malhotra, M. Teng, M. Duvic, and R. A. Chandraratna. 1997. Tazarotene-induced gene 2 (TIG2), a novel retinoid-responsive gene in skin. J Invest Dermatol 109:91-95. 97. Zabel, B. A., A. M. Silverio, and E. C. Butcher. 2005. Chemokine-like receptor 1 expression and chemerin-directed chemotaxis distinguish plasmacytoid from myeloid dendritic cells in human blood. J Immunol 174:244-251. 29

98. Wittamer, V., J. D. Franssen, M. Vulcano, J. F. Mirjolet, E. Le Poul, I. Migeotte, S. Brezillon, R. Tyldesley, C. Blanpain, M. Detheux, A. Mantovani, S. Sozzani, G. Vassart, M. Parmentier, and D. Communi. 2003. Specific recruitment of antigen- presenting cells by chemerin, a novel processed ligand from human inflammatory fluids. J Exp Med 198:977-985. 99. Meder, W., M. Wendland, A. Busmann, C. Kutzleb, N. Spodsberg, H. John, R. Richter, D. Schleuder, M. Meyer, and W. G. Forssmann. 2003. Characterization of human circulating TIG2 as a ligand for the orphan receptor ChemR23. FEBS Lett 555:495-499. 100. Zabel, B. A., L. Zuniga, T. Ohyama, S. J. Allen, J. Cichy, T. M. Handel, and E. C. Butcher. 2006. Chemoattractants, extracellular proteases, and the integrated host defense response. Exp Hematol 34:1021-1032. 101. Wittamer, V., B. Bondue, A. Guillabert, G. Vassart, M. Parmentier, and D. Communi. 2005. Neutrophil-mediated maturation of chemerin: a link between innate and adaptive immunity. J Immunol 175:487-493. 102. Zabel, B. A., S. J. Allen, P. Kulig, J. A. Allen, J. Cichy, T. M. Handel, and E. C. Butcher. 2005. Chemerin activation by serine proteases of the coagulation, fibrinolytic, and inflammatory cascades. J Biol Chem 280:34661-34666. 103. Busmann, A., M. Walden, M. Wendland, C. Kutzleb, W. G. Forssmann, and H. John. 2004. A three-step purification strategy for isolation of hamster TIG2 from CHO cells: characterization of two processed endogenous forms. J Chromatogr B Analyt Technol Biomed Life Sci 811:217-223. 104. Gantz, I., Y. Konda, Y. K. Yang, D. E. Miller, H. A. Dierick, and T. Yamada. 1996. Molecular cloning of a novel receptor (CMKLR1) with homology to the chemotactic factor receptors. Cytogenet Cell Genet 74:286-290. 105. Owman, C., S. J. Lolait, S. Santen, and B. Olde. 1997. Molecular cloning and tissue distribution of cDNA encoding a novel chemoattractant-like receptor. Biochem Biophys Res Commun 241:390-394. 106. Samson, M., A. L. Edinger, P. Stordeur, J. Rucker, V. Verhasselt, M. Sharron, C. Govaerts, C. Mollereau, G. Vassart, R. W. Doms, and M. Parmentier. 1998. ChemR23, a putative chemoattractant receptor, is expressed in monocyte-derived dendritic cells and macrophages and is a coreceptor for SIV and some primary HIV-1 strains. Eur J Immunol 28:1689-1700. 107. Vermi, W., E. Riboldi, V. Wittamer, F. Gentili, W. Luini, S. Marrelli, A. Vecchi, J. D. Franssen, D. Communi, L. Massardi, M. Sironi, A. Mantovani, M. Parmentier, F. Facchetti, and S. Sozzani. 2005. Role of ChemR23 in directing the migration of myeloid and plasmacytoid dendritic cells to lymphoid organs and inflamed skin. J Exp Med 201:509-515. 108. Zabel, B. A., T. Ohyama, L. Zuniga, J. Y. Kim, B. Johnston, S. J. Allen, D. G. Guido, T. M. Handel, and E. C. Butcher. 2006. Chemokine-like receptor 1 expression by macrophages in vivo: Regulation by TGF-beta and TLR ligands. Exp Hematol 34:1106- 1114. 30

109. Methner, A., G. Hermey, B. Schinke, and I. Hermans-Borgmeyer. 1997. A novel G protein-coupled receptor with homology to neuropeptide and chemoattractant receptors expressed during bone development. Biochem Biophys Res Commun 233:336-342. 110. Weigert, J., F. Obermeier, M. Neumeier, J. Wanninger, M. Filarsky, S. Bauer, C. Aslanidis, G. Rogler, C. Ott, A. Schaffler, J. Scholmerich, and C. Buechler. 2010. Circulating levels of chemerin and adiponectin are higher in ulcerative colitis and chemerin is elevated in Crohn's disease. Inflamm Bowel Dis 16:630-637. 111. Albanesi, C., C. Scarponi, D. Bosisio, S. Sozzani, and G. Girolomoni. 2010. Immune functions and recruitment of plasmacytoid dendritic cells in psoriasis. Autoimmunity 43:215-219. 112. Graham, K. L., B. A. Zabel, S. Loghavi, L. A. Zuniga, P. P. Ho, R. A. Sobel, and E. C. Butcher. 2009. Chemokine-like receptor-1 expression by central nervous system- infiltrating leukocytes and involvement in a model of autoimmune demyelinating disease. J Immunol 183:6717-6723. 113. Luangsay, S., V. Wittamer, B. Bondue, O. De Henau, L. Rouger, M. Brait, J. D. Franssen, P. de Nadai, F. Huaux, and M. Parmentier. 2009. Mouse ChemR23 is expressed in dendritic cell subsets and macrophages, and mediates an anti-inflammatory activity of chemerin in a lung disease model. J Immunol 183:6489-6499. 114. Zabel, B. A., S. Nakae, L. Zuniga, J. Y. Kim, T. Ohyama, C. Alt, J. Pan, H. Suto, D. Soler, S. J. Allen, T. M. Handel, C. H. Song, S. J. Galli, and E. C. Butcher. 2008. Mast cell-expressed orphan receptor CCRL2 binds chemerin and is required for optimal induction of IgE-mediated passive cutaneous anaphylaxis. J Exp Med 205:2207-2220. 115. Nibbs, R. J., S. M. Wylie, I. B. Pragnell, and G. J. Graham. 1997. Cloning and characterization of a novel murine beta chemokine receptor, D6. Comparison to three other related macrophage inflammatory protein-1alpha receptors, CCR-1, CCR-3, and CCR-5. J Biol Chem 272:12495-12504. 116. Shimada, T., M. Matsumoto, Y. Tatsumi, A. Kanamaru, and S. Akira. 1998. A novel lipopolysaccharide inducible C-C chemokine receptor related gene in murine macrophages. FEBS Lett 425:490-494. 117. Zuurman, M. W., J. Heeroma, N. Brouwer, H. W. Boddeke, and K. Biber. 2003. LPS-induced expression of a novel chemokine receptor (L-CCR) in mouse glial cells in vitro and in vivo. Glia 41:327-336. 118. Brouwer, N., M. W. Zuurman, T. Wei, R. M. Ransohoff, H. W. Boddeke, and K. Biber. 2004. Induction of glial L-CCR mRNA expression in spinal cord and brain in experimental autoimmune encephalomyelitis. Glia 46:84-94. 119. Migeotte, I., J. D. Franssen, S. Goriely, F. Willems, and M. Parmentier. 2002. Distribution and regulation of expression of the putative human chemokine receptor HCR in leukocyte populations. Eur J Immunol 32:494-501. 120. Galligan, C. L., W. Matsuyama, A. Matsukawa, H. Mizuta, D. R. Hodge, O. M. Howard, and T. Yoshimura. 2004. Up-regulated expression and activation of the orphan chemokine receptor, CCRL2, in rheumatoid arthritis. Arthritis Rheum 50:1806-1814.

31

121. Ernst, M. C., M. Issa, K. B. Goralski, and C. J. Sinal. 2010. Chemerin exacerbates glucose intolerance in mouse models of obesity and diabetes. Endocrinology 151:1998- 2007. 122. Meister, B. 2000. Control of food intake via leptin receptors in the hypothalamus. Vitam Horm 59:265-304. 123. Takahashi, M., Y. Takahashi, K. Takahashi, F. N. Zolotaryov, K. S. Hong, R. Kitazawa, K. Iida, Y. Okimura, H. Kaji, S. Kitazawa, M. Kasuga, and K. Chihara. 2008. Chemerin enhances insulin signaling and potentiates insulin-stimulated glucose uptake in 3T3-L1 adipocytes. FEBS Lett 582:573-578. 124. Sell, H., J. Laurencikiene, A. Taube, K. Eckardt, A. Cramer, A. Horrighs, P. Arner, and J. Eckel. 2009. Chemerin is a novel adipocyte-derived factor inducing insulin resistance in primary human skeletal muscle cells. Diabetes 58:2731-2740. 125. Parlee, S. D., M. C. Ernst, S. Muruganandan, C. J. Sinal, and K. B. Goralski. 2010. Serum chemerin levels vary with time of day and are modified by obesity and tumor necrosis factor-{alpha}. Endocrinology 151:2590-2602. 126. Bozaoglu, K., D. Segal, K. A. Shields, N. Cummings, J. E. Curran, A. G. Comuzzie, M. C. Mahaney, D. L. Rainwater, J. L. VandeBerg, J. W. MacCluer, G. Collier, J. Blangero, K. Walder, and J. B. Jowett. 2009. Chemerin is associated with metabolic syndrome phenotypes in a Mexican-American population. J Clin Endocrinol Metab 94:3085-3088. 127. Sell, H., A. Divoux, C. Poitou, A. Basdevant, J. L. Bouillot, P. Bedossa, J. Tordjman, J. Eckel, and K. Clement. 2010. Chemerin correlates with markers for fatty liver in morbidly obese patients and strongly decreases after weight loss induced by bariatric surgery. J Clin Endocrinol Metab 95:2892-2896. 128. Ress, C., A. Tschoner, J. Engl, A. Klaus, H. Tilg, C. F. Ebenbichler, J. R. Patsch, and S. Kaser. 2010. Effect of bariatric surgery on circulating chemerin levels. Eur J Clin Invest 40:277-280. 129. Wang, L. Y., L. Wei, H. Y. Yu, Y. Zhang, and W. P. Jia. 2009. [Relationship of serum Chemerin to obesity and type 2 diabetes mellitus]. Zhonghua Yi Xue Za Zhi 89:235-238. 130. Kindt, T., G. RA, and O. BA. 2007. Kuby Immunology. W.H. Freeman and Company, New York.

32

33

CHAPTER 2 : CHEMOKINE-LIKE RECEPTOR 1 EXPRESSION BY

MACROPHAGES IN VIVO: REGULATION BY TGF-BETA AND TLR

LIGANDS

Brian A Zabel1, Takao Ohyama1, Luis Zuniga1, Ji-Yun Kim1, Brent Johnston2, Samantha

J. Allen3, David G. Guido1, Tracy M. Handel3, Eugene C. Butcher1

1Laboratory of Immunology and Vascular Biology, Department of Pathology, Stanford

University School of Medicine, Stanford, CA 94305, and Center for Molecular Biology and Medicine, Veterans Affairs Palo Alto Health Care System, Palo Alto, CA 94304.

2Departments of Microbiology & Immunology and Pediatrics, Dalhousie University,

Halifax, Nova Scotia B3H 1X5.

3Skaggs School of Pharmacy and Pharmaceutical Sciences, University of California, San

Diego, La Jolla, CA 92093.

Reproduced from:

Exp Hematol. 2006 Aug;34 (8):1106-14.

Used by copyright permission of Elsevier.

34

Summary

Objective. Chemokine-like receptor 1 (CMKLR1) is expressed by human antigen presenting cells and binds to chemerin, a proteolytically-activatable chemoattractant.

Here we assessed the expression of mCMKLR1 on mouse leukocytes, focusing on ex

vivo dendritic cells (DC) and macrophages. mCMKLR1-expressing cells were evaluated

for functional responses to chemerin. We examined the regulation of mCMKLR1

expression by exposure to toll-like receptor (TLR)-ligands and cytokines. Finally, we

evaluated ex vivo human ascites macrophages for huCMKLR1 expression and chemerin-

responsiveness.

Methods. A novel anti-mCMKLR1 monoclonal antibody was generated to assess

mCMKLR1 expression by mouse leukocytes using flow cytometry. Mouse bone marrow-

derived DC precursors, mouse peritoneal macrophages, and human ascites leukocytes

were examined in functional assays (in vitro chemotaxis and intracellular calcium

mobilization).

Results. During DC differentiation from bone marrow, mCMKLR1 is upregulated early

and then diminishes with time in culture. Most DC in vivo do not detectably express the

receptor. In contrast, freshly isolated F4/80+CD11b+ mouse serosal macrophages express

mCMKLR1, bind a fluorescently-labeled chemerin peptide, and display calcium

signaling and migration to the active ligand. Interestingly, macrophage mCMKLR1 is

suppressed by pro-inflammatory cytokines and TLR ligands, whereas treatment with

TFGβ upregulates the receptor. A small population of blood borne F4/80+CD11b+

35

macrophages also expresses mCMKLR1. Freshly isolated macrophages from human ascites fluid express CMKLR1 and are chemerin-responsive, as well.

Conclusion. The conserved expression of CMKLR1 by macrophages in mouse and man, coupled with the stimuli-specific regulation of CMKLR1, may reflect a critical role for

CMKLR1:chemerin in shaping the nature (either pro-inflammatory or suppressive) in macrophage-mediated immune responses.

36

Introduction

Resident tissue macrophages are present in the basement membrane of virtually

every epithelial and endothelial surface in the body (1). This system-wide positioning of

macrophages is ideal for their role as “rapid responder” leukocytes in providing immune

protection against microorganisms, and also for regulating peripheral immune responses.

Macrophages execute a spectrum of distinct effector functions that depend on the nature of the activating stimuli (1, 2, 3). “Classical” activation by microbes or toll-like receptor

(TLR) ligands (LPS, CpG, polyI:C) and IFNγ generates microbicidal macrophages (also known as “M1 macrophages”) that secrete large amounts of nitric oxide and proinflammatory cytokines (IL-1, IL-6, TNFα, IFNγ, and IL-12), and display anti-tumor cytotoxicity (4, 5, 6, 7). In contrast, alternative activation by cytokines such as IL-4, IL-

13, or IL-10 in the absence of TLR ligands or other “danger signals” generates “M2 macrophages”, characterized by MHC class II downregulation and secretion of immune- suppressive cytokines such as IL-10 and TGFβ (3). This range of effector functions permits macrophage to act as “immune-interpreters”, ready to establish an appropriate, situation-dependent immune response.

Chemoattractant receptors modulate the activity of macrophages, and control their homing properties and their responses to immune stimuli (2). The recently de-orphaned chemoattractant receptor, chemokine-like receptor 1 (CMKLR1) has been studied in the human, where it is selectively expressed by circulating plasmacytoid but not myeloid dendritic cells (DC) (8), a finding supported by recent microarray RNA analysis (9).

37

Chemerin, a CMKLR1 protein ligand, circulates in an immature pro-form, and is rapidly activated by serine proteases of the coagulation, fibrinolytic, and inflammatory cascades via cleavage of inhibitory carboxyl-terminal amino acids (10, 11). Active chemerin is a potent chemoattractant for CMKLR1-expressing cells. Chemerin can therefore act as a molecular translator of tissue damage or bleeding to alert CMKLR1+ cells via activation of intracellular signaling pathways and recruitment.

In this report, we generated a novel monoclonal antibody (mAb) specific for the murine CMKLR1 ortholog and found that while mCMKLR1 was upregulated early during DC differentiation from bone marrow, freshly isolated CD11c+ mouse dendritic cells (including plasmacytoid DC) were essentially negative for receptor expression.

Mouse serosal macrophages, however, expressed high levels of CMKLR1, and bound a bioactive chemerin peptide. We also identified mCMKLR1 expression on the small population of circulating blood F4/80+CD11b+ macrophages. mCMKLR1+ peritoneal macrophages migrated detectably to chemerin, although the major population chemotaxed only inefficiently in standard transwell assays, whether to chemerin or to chemokines CXCL12 or CCL2. A subset of F4/80loCD11b+ peritoneal macrophages, however, migrated well to chemerin. In addition, we identified stimuli-specific modulation of receptor expression: mouse macrophages treated with pro-inflammatory cytokines and TLR ligands downregulated CMKLR1 expression, while treatment with

TGFβ upregulated the receptor. Freshly isolated human ascites macrophages expressed

CMKLR1 and were chemerin-responsive, therefore establishing the relevance of the

38

mouse as a model for studying human macrophage CMKLR1:chemerin interactions. The conserved expression of CMKLR1 on macrophages and its preferential upregulation on

M2 macrophages may reflect the preservation of a critical role for chemerin in the macrophage response to non-infectious sites of tissue bleeding or injury.

39

Materials and Methods

Antibodies and reagents

Anti-human -CD3, -CD11c, -CD14, -CD16, -CD19, -CD20, -CD56, -CD123, -HLADR,

and anti-mouse -CD11b, -CD11c, -CD14, -CD19, -B220, -F4/80, -Gr1, -IA/E class II, -

Ly6C, -TCRβ dye-linked mAb were obtained from eBioscience (San Diego, CA, USA),

BD PharMingen (San Diego, CA, USA), and Serotec (Raleigh, NC, USA). Anti-rat

phycoerythrin (human and mouse adsorbed) was purchased from BD Pharmingen,

purified Fc block (mouse anti-mouse CD16.2/32.2) was purchased from Caltag

(Burlingame, CA, USA), and mouse IgG, rat IgG, and goat serum were purchased from

Sigma (St. Louis, MO, USA). CXCL12, CCL21, IL-4, GM-CSF, and Flt-3 ligand were

purchased from R&D Systems (Minneapolis, MN, USA). CMFDA, Fluo-4-

acetoxymethyl (AM), and Pluronic acid F-127 (reconstituted in DMSO) were purchased

from Molecular Probes (Eugene, OR, USA). Phosphothioated CpG oligonucleotides (12)

were purchased from Qiagen (Valencia, CA, USA), polyI:C was purchased from Sigma.

LPS (E.coli O11:B4-derived) was purchased from List Biologicals (Campbell, CA,

USA), TNFα and IFNγ from Roche (Penzberg, Germany), and TGFβ1 and TGFβ2 from

R&D Systems. The FAM-labeled bioactive chemerin peptide (FAM-chemerinpep , a.a.

145-157 (Y145 F149), NH2-YHSFFFPGQFAFS-COOH) was purchased from Phoenix

Pharmaceuticals (Belmont, CA), while the corresponding unlabeled bioactive chemerin peptide (chemerinpep) was synthesized by the Stanford Protein and Nucleic Acid

40

Biotechnology Facility (Stanford University, Stanford, CA, USA). Complete and

incomplete Freund’s adjuvant (CFA and IFA) were purchased from Sigma.

Mammalian expression vector construction and generation of stable cell lines

The coding region of mCMKLR1 was amplified from genomic DNA with an engineered

N-terminal hemagglutinin (HA) tag, and cloned into pcDNA3 (Invitrogen, Carlsbad, CA,

USA). Transfectants of mCMKLR1 were generated and stable lines selected in the

murine pre-B lymphoma cell line L1.2 as described (13). Transfected cells were in some

cases treated with 5 mM n-butyric acid (Sigma) for 24 h before experimentation (14). huCMKLR1, chemerin, and empty vector L1.2 transfectants were generated as previously described (8).

Chemerin expression and purification using baculovirus

The “serum form” of chemerin with the sequence NH2-ADPELTE…FAPHHHHHHHH-

COOH was expressed using baculovirus-infected insect cells, as previously described

(10). Carboxypeptidase A (Sigma) cleavage was used to remove the His8 tag, leaving the resulting protein with the sequence NH2-ADPELTE…FAPH-COOH, where the underlined residues are non-native. The protein was lyophilized and checked for purity using electrospray mass spectrometry.

Generating the anti-CMKLR1 mAb BZ194

The immunizing amino-terminal CMKLR1 peptide with the sequence NH2-

DSGIYDDEYSDGFGYFVDLEEASPWC-COOH (corresponding to residues 8-32 of

CMKLR1, with a non-native carboxyl-terminal cysteine to facilitate conjugation to

41

keyhole limpet hemocyanin, (KLH)) was synthesized by the Stanford Protein and Nucleic

Acid Biotechnology Facility and conjugated to KLH according to the manufacturer’s

specifications (Pierce Biotechnology, Rockford, IL, USA). Wistar Furth rats (Charles

River, Wilmington, MA, USA) were immunized with the mCMKLR1 peptide/KLH

conjugate first emulsified in CFA, and then subsequently in IFA. Hybridomas producing

anti-mCMKLR1 mAb were subcloned, and specificity was confirmed by reactivity with

mouse but not human CMKLR1 transfectants. An ELISA-based assay (BD Pharmingen)

was used to determine the IgG2aκ isotype of the resulting rat anti-mouse CMKLR1 mAb, designated BZ194.

Harvesting mouse leukocytes

The Veterans Affairs Palo Alto Health Care System Institutional Animal Care and Use

Committee, Palo Alto, CA, and the Stanford University Administrative Panel on

Laboratory Animal Care, Stanford, CA, approved all animal experiments. C57Bl/6 mice were obtained from Taconic (Oxnard, CA, USA). To harvest blood leukocytes, mice were given a fatal overdose of anesthesia (ketamine/xylazine) as well as an i.p. injection of heparin (100 units, Sigma). Mouse blood was collected by cardiac puncture. Up to 1 mL of blood was added to 5 mL of 2 mM EDTA in PBS, and 6 mL of 2% dextran T500

(Amersham Biosciences, Piscataway, NJ, USA) was added to crosslink red blood cells.

The mixture was incubated for 1 hr at 37°C, the supernatant was removed and pelleted, and the cells were resuspended in 5 mL red blood cell lysis buffer (Sigma) and incubated at RT for 5 min. The cells were pelleted, and resuspended for use in cell staining. Pleural

42

cavity leukocytes were obtained by removing the lungs and heart and rinsing the chest cavity with PBS. , thymus, and spleen cells were harvested by direct

crushing of the organs over wire mesh, with or without pre-incubating with collagenase

D, followed by lysis of red blood cells as needed. Bone marrow cells were harvested by

flushing femurs and tibias with media followed by red blood cell lysis. Peritoneal lavage

cells were obtained by i.p. injection of 10 mL PBS, gentle massage of the peritoneal

cavity, and collection of the exudate. For some experiments, 500 μl of peritoneal cells

(2x106 cells/mL) were incubated for 24 hours with either LPS (1 μg/mL), TNFα (10

ng/mL), IFNγ (100 U/mL), polyI:C (20 μg/mL), CpG (10-100 μg/mL), or TGFβ (5

ng/mL).

Harvesting human ascites leukocytes

The Institutional Review Board at Stanford University approved all human subject

protocols. 1.6 L of freshly collected paracentesis fluid (obtained from

immunocompromised patients with end-stage liver disease, either from chronic hepatitis

C and/or alcoholism) was centrifuged for 15 min at 1500 RPM, and the cellular exudate was subjected to hypotonic red blood cell lysis.

Cell sorting and Wright-Giemsa stain

Human ascites leukocytes and mouse peritoneal cells were stained as described and sorted by standard flow cytometric techniques (FACsvantage, BD Biosciences, Mountain

View, CA, USA; flow cytometry was performed at the Stanford University Digestive

Disease Center Core Facility, VA Hospital, Palo Alto, CA, USA). 1-5x104 sorted cells

43

were loaded into cytospin chambers and centrifuged onto glass slides. The slides were

stained with Wright-Giemsa dye by standard automated techniques at the VA Hospital

Hematology Lab (Palo Alto, CA, USA) and examined by light microscopy with a 40x objective.

Generating in vitro cultured bone marrow derived mouse dendritic cell.

Bone marrow was harvested as described and cultured with: 10 ng/mL GM-CSF + 5 ng/mL IL-4; 20 ng/mL GM-CSF + 100 ng/mL Flt-3 ligand; or 100 ng/mL Flt-3 ligand alone for 9 days. Cytokines were replenished every 3 days, and cells were split to maintain a cell density of 5-10 million cells/mL.

In vitro transwell chemotaxis

Chemotaxis media consisted of RPMI +10% FCS. 1x106 cells in 100 μl were added to the top well of 5-um pore transwell inserts (Costar, Corning, NY, USA), and test samples

(600 μl) were added to the bottom well. After 2 hr at 37ºC, migration was assessed by flow cytometric examination of cells that moved into the lower chamber. Polystyrene beads (15.0 μm diameter, Polysciences, Warrington, PA, USA) were added to each well to allow the cell count to be normalized. A ratio was generated and percent input migration was calculated. A pre-determined volume of chemerin conditioned media eliciting >30% CMKLR1/L1.2 transfectant migration (along with an equivalent volume

of empty vector (pcDNA3) L1.2 transfectant conditioned media as a negative control)

was used. The Student’s t-test was used to determine statistical significance.

FAM-labeled chemerin peptide binding and dissociation by flow cytometry

44

Freshly isolated resident peritoneal cells were kept on ice until use. For flow cytometric

analysis, cells were resuspended at 1x106 cells/mL in modified Iscove’s medium

(Iscove’s medium with 1% heat inactivated bovine calf serum and 2 mM L-glutamine)

and kept at 25°C. Fluorescent data were acquired continuously up to 1024 seconds at 1s

intervals at 25°C under constant stirring (500 rpm) using a FACsScan flow cytometer

(BD Biosciences) and CellQuest software. To identify CD11b+ macrophage population,

mixed peritoneal leukocytes were pre-incubated with CD11b-PerCP mAb for 3 min at

25°C immediately before the start of each sample. The samples were analyzed for 60-120

second to establish basal state, removed from the nozzle to add the stimuli (15 nM FAM-

chemerinpep), then returned to the nozzle with 5-10 seconds of interruption in data acquisition. For dissociation, after allowing FAM-chemerinpep to bind for 8 min at 25°C,

400 fold molar excess unlabeled chemerinpep was added and allowed to compete for the

cell surface binding sites for 5-6 minutes. Mean channel fluorescence over time was analyzed with FlowJo (TreeStar, Ashland, OR, USA) software for CD11b+ (macrophage) and CD11b- population separately and the amount of FAM-chemerinpep bound was

normalized within each population.

Intracellular calcium mobilization

Chemoattractant-stimulated Ca2+-mobilization was performed following Alliance for Cell

Signaling protocol ID PP00000210. Cells (3x106/mL) were loaded with 4 μM Fluo4-AM,

0.16% Pluronic acid F-127 (Molecular Probes) in modified Iscove’s medium (Invitrogen) for 30 minutes at 37°C. The samples were mixed every 10 minutes during loading,

45

washed once, resuspended at 1x106/mL in the same buffer, and allowed to rest in the dark for 30 minutes. Chemoattractant-stimulated change in Ca2+-sensitive fluorescence of

Fluo4 was measured over real-time with a FACsScan flow cytometer (BD Biosciences) at

25°C under stirring condition following the FAM-labeled chemerin peptide binding protocol.

46

Results

A mCMKLR1-specific mAb stains early DC progenitors in vitro, but not most blood or tissue DC

We generated a monoclonal antibody designated BZ194 (IgG2aκ) to mouse

CMKLR1. The antibody was specific to mouse CMKLR1/L1.2 transfectants, with no

cross-reactivity with human CMKLR1/L1.2 transfectants (Fig. 1A). Human CMKLR1

shares 80% amino acid identity and is more homologous to mouse CMKLR1 than any

murine protein. Reactivity with CXCR1-through-6 and CCR1-through-10 was excluded

by lack of staining of blood cell subsets or cultured mouse cells known to express these

receptors (not shown).

Based on previous data indicating CMKLR1 expression on human DC (8, 15, 16,

17), we used the mAb BZ194 to assess the expression of mCMKLR1 by mouse DC. The

addition of IL-4 and GMCSF, Flt-3 ligand and GMCSF, or Flt-3 ligand alone to cultures

of bone marrow progenitors induced mCMKLR1 expression on a large percentage of

CD11c+MHCIIint cells, peaking at day 1 and then decreasing over time in culture (Fig.

1B,C). Interestingly, in the absence of added exogenous cytokine, about 18% of the bone

marrow-derived CD11c+MHCIIint cells expressed mCMKLR1 on day 1 of culturing (Fig.

1B). Rapid cell death precluded extended observations of cultures lacking cytokines. A

distinct population of CD11c+MHCIIhi leukocytes present in the in vitro cultures, likely

to be bone marrow resident DC, did not express mCMKLR1 (Fig. 1B).

47

We next tested freshly isolated CD11c+ mouse dendritic cells for mCMKLR1

expression, and found that DC from blood, spleen, lymph nodes, and bone marrow were essentially negative for mCMKLR1 (Fig. 1D). Because our previous data, and those from

Vermi et al. (16) demonstrated that human plasmacytoid DC express high levels of

CMKLR1, we tested murine blood plasmacytoid DC (defined as B220+CD11c+Ly6C+) for CMKLR1 expression, and found these cells to be negative as well (Fig. 1D).

48

49

Figure 1. Anti-mCMKLR1 monoclonal Ab stains Dendritic cell progenitors, but not most Dendritic cells in vivo. (A) Unlabeled mCMKLR1/L1.2 transfectants were mixed 1:1 with CMFDA-labeled huCMKLR1/L1.2 transfectants, and used to screen for mCMKLR1-specific mAb by flow cytometry. For (B) and (C), total bone marrow leukocytes were incubated for up to 9 days with the indicated cytokines to generate in vitro cultured dendritic cells. DC precursors were identified by staining with CD11c and MHCII. For C, n=3, with mean percentage (+/- S.E.M.) mCMKLR1+ cells displayed. (D) Leukocytes from bone marrow, spleen, lymph nodes, and blood were harvested and mCMKLR1 expression was evaluated on TCRβ-CD19-CD11c+ DC. Blood plasmacytoid DC were defined by specific markers (B220+Ly6C+CD11c+). For (B) and (D), a representative plot of n=3 with similar results is displayed.

50

mCMKLR1 is selectively expressed by murine macrophages

The initial RNA analysis of human CMKLR1 indicated that it was expressed by

macrophages (17). Furthermore, in vitro cultured monocyte-derived human macrophages

(15) and lymph node macrophages (16) were CMKLR1-positive. We therefore isolated

mouse peritoneal cells and stained with the well-defined macrophage markers F4/80 and

CD11b, and found that 100% of F4/80+CD11b+ macrophages were positive for

mCMKLR1 (Fig. 2A). Macrophages from a second serosal surface, the pleural cavity,

were also positive for mCMKLR1 (Fig. 2A).

To independently confirm the specificity of our mAb for mCMKLR1, we used a

fluorescently-labeled bioactive chemerin peptide (FAM-chemerinpep) and performed

binding experiments to identify peritoneal cells that bound chemerin. We found that the

CD11b+, but not the CD11b- subset of peritoneal cells bound FAM-chemerinpep (Fig.

2B). For the CD11b+ cells, we observed a time-dependent increase in peptide binding, which is characteristic of specific ligand:receptor interactions. We also observed limited dissociation of labeled chemerin peptide following addition of unlabelled peptide

(chemerinpep). In contrast, the CD11b- subset of peritoneal cells did not display time-

dependent binding nor dissociation by chemerinpep, which is consistent with non-specific interactions (Fig. 2B). These results indicate that FAM-chemerinpep selectively binds the

same subset of peritoneal cells (CD11b+ macrophages) that stained with anti-

mCMKLR1, confirming our mAb staining results, and implying a direct interaction

between chemerin and mCMKLR1 in vitro.

51

We next evaluated the expression of additional macrophage-associated surface

molecules on the mCMKLR1+ peritoneal macrophages by flow cytometry.

F4/80+CD11b+ peritoneal macrophages expressed low levels of MHCII and the

costimulatory molecules CD80 and CD86, and were positive for the myeloid markers

CD14 and MAC3 (Fig. 2C).

As predicted, sorted mCMKLR1+ cells displayed morphological features typical

of macrophages (Fig. 2D). mCMKLR1+ cells were generally round with ruffled

membranes, an abundance of granular cytoplasm, and centrally-located round or

monocyte-like lobulated nuclei. Thus both traditional morphologic and immunophenotypic analysis indicated selective expression of mCMKLR1 by macrophages.

A small but discrete subset (<2-5% total) of peritoneal cells can be distinguished from the bulk macrophage population by expression of lower levels of F4/80

(CD11b+F4/80lo cells) and by high levels of MHC class II. These cells expressed the

myeloid marker CD14, and low-to-no levels of CD80, CD86, and MAC3; and they also

expressed mCMKLR1 (Fig. 2E).

In examining blood leukocytes for mCMKLR1 expression, we were surprised to

identify a small population (<0.5% total) of circulating F4/80+CD11b+ macrophages

(Fig. 3A). These blood macrophages expressed mCMKLR1, as well as CD14, but not

MHCII, CD80, or CD86, and a subset expressed MAC3 (Fig. 3A,B). In contrast, circulating monocytes (defined as CD11b+7/4+ (18)) were negative for mCMKLR1

52

(Fig. 3C), as were circulating T cells, B cells, and NK cells, as well as bone marrow neutrophils (not shown).

53

Figure 2. Freshly isolated macrophages express mCMKLR1. (A) Peritoneal and pleural cells were harvested, and mCMKLR1 expression was evaluated on F4/80+CD11b+ macrophages. (B) Peritoneal cells were harvested and binding assays were performed using continuous-acquisition flow cytometry. 15 nM FAM-labeled chemerin peptide (FAM-chemerinpep) was added as indicated and binding assessed on CD11b+ vs. CD11b- cells. A 400-fold molar excess of unlabeled chemerin peptide (chemerinpep) was added as indicated to examine the dissociation of labeled ligand. (C) The surface phenotype of F4/80+CD11b+ peritoneal macrophages was 54

examined by flow cytometry. Filled histograms indicate isotype-matched controls, and open histograms represent staining with the indicated specific antibodies. (D) mCMKLR1+ peritoneal cells were sorted, harvested by cytospin, and stained by Wright- Giemsa. Cells were examined by light microscope using a 40x objective. (E) The surface phenotype of F4/80loCD11b+ peritoneal macrophage-like cells was examined by flow cytometry. The quadrant crossbars for each sample were established based on matched isotype control antibody staining. One representative data set of at least three experiments is shown for each part in this figure.

Figure 3. Rare circulating blood macrophages express mCMKLR1. (A) Blood leukocytes were collected and analyzed by flow cytometry. mCMKLR1 expression was assessed on the small population of F4/80+CD11b+ circulating macrophages. The crossbar was set based on isotype control antibody staining, as shown. (B) The surface phenotype of F4/80+CD11b+ blood macrophages was examined by flow cytometry. (C) A cell gate was set on CD11b+7/4+ blood monocytes, and mCMKLR1 staining was assessed. The filled histogram indicates isotype control antibody staining, and the open histogram indicates mCMKLR1 staining. One representative data set of at least three experiments is shown for each part in this figure.

55

Chemerin is a functional ligand for mouse DC progenitors and in vivo peritoneal

macrophages

Bone marrow-derived CD11c+MHCIIint leukocytes harvested after 1 day of

culture with GMCSF and IL-4 displayed significant migration to chemerin conditioned

media (CM), while the CD11c+MHCIIhi cells did not (Fig. 4A). The CD11c+MHCIIhi cells did, however, migrate to a combination of CXCL12 and CCL21, indicating that the cells were capable of responding to a chemotactic gradient. Interestingly, the

CD11c+MHCIIint DC progenitors did not respond to the same combination of CXCL12

and CCL21, perhaps indicating that mCMKLR1+ DC progenitors migrate selectively to chemerin.

We next examined peritoneal mouse macrophages for chemerin-responsiveness.

We first labeled cells with Fluo-4 (19) and monitored intracellular calcium mobilization.

Chemerin induced a calcium flux in CD11b+ but not CD11b- peritoneal cells, indicating the selective activation of intracellular cell signaling in mCMKLR1+ macrophages (Fig.

4B).

In transwell migration assays, F4/80+CD11b+ peritoneal macrophages displayed low overall motility (the background migration was <0.02%). The cells displayed detectable, but inefficient, chemotaxis to all chemoattractants tested, including chemerin,

CCL2, and CXCL12 (Fig. 4C). In contrast, the small population of F4/80loCD11b+ macrophage-like cells displayed a more robust migratory response, with 4-5% migration

56

to chemerin (Fig. 4D). These cells also responded quite well to both CXCL12 and CCL2, registering >25% migration to these chemokines.

57

Figure 4. mCMKLR1+ leukocytes are chemerin-responsive. (A) Total bone marrow-cultured cells from 1-day incubation with GM-CSF and IL-4 were tested for migratory responses in transwell chemotaxis assays. The migrated cells were stained for CD11c and MHCII. Migration was assessed to the following test samples as indicated: CXCL12 (10 nM) and CCL21 (10 nM), chemerin conditioned media (chemerin CM), empty vector conditioned media (vector CM), and no chemoattractant ( (-) no chem), n=3 with mean percentage (+/- S.E.M.) migration displayed. *, p<0.05, **, p<0.005 comparing “ (-) no chem” vs. CXCL12 + CCL21, or vector CM vs. chemerin CM. (B) Peritoneal cells were isolated and loaded with Fluo4- AM, and intracellular calcium mobilization was examined using continuous-acquisition flow cytometry. 70 nM chemerin was added as indicated. A representative data set for at least 3 experiments is shown. For (C) and (D), total peritoneal cells were tested in in vitro transwell chemotaxis assays using varying doses of chemerin, CXCL12 (10 nM), and CCL2 (1.0 nM). After migration, the cells were collected and stained, and percent input migration is displayed (mean +/- S.D. of triplicate wells, *, p<0.05 compared with “ (-) no chem” background migration). (C) F4/80+CD11b+ macrophage migration. (D) F4/80loCD11b+ macrophage-like cell migration.

58

Regulation of mCMKLR1 by M1 vs. M2 stimuli

We next examined the regulation of mCMKLR1 expression by peritoneal mouse

macrophages. Cells were treated with various cytokines and/or TLR ligands for 24 hours

and then examined for mCMKLR1 expression. In general, pro-inflammatory cytokines and TLR-ligands suppressed mCMKLR1 expression (Fig. 5, upper panel). We observed the following hierarchy in mCMKLR1 suppression by TLR ligands: LPS>polyI:C>CpG.

The pro-inflammatory cytokines IFNγ and TNFα caused modest mCMKLR1 suppression as well, which was enhanced by co-incubation with LPS. Interestingly, the immune- suppressive cytokine TGFβ had the opposite effect on receptor expression: overnight

treatment with either TGFβ1 or TGFβ2 resulted in mCMKLR1 upregulation (Fig. 5,

lower panel). Thus, stimuli associated with inducing a “M1” macrophage phenotype

caused mCMKLR1 suppression, while cytokines associated with inducing a “M2”

macrophage phenotype resulted in upregulation.

59

Figure 5. Effects of TLR ligands and cytokines on CMKLR1 expression on mouse macrophages. Freshly isolated peritoneal macrophages were cultured for 24 hours with various stimuli as indicated. Representative histograms are shown from n>3 experiments with similar results.

60

Ex vivo human ascites macrophages express huCMKLR1 and are chemerin-responsive

Due to the discrepancy between DC expression of CMKLR1 in mouse versus man, we asked if macrophage CMKLR1 expression and chemerin-responsiveness were conserved between species. We obtained fresh human ascites fluid from patients undergoing paracentesis, harvested the cells, and stained for huCMKLR1 expression.

Similar to mouse peritoneal cells, we identified a population of cells exhibiting high granularity and large size (as compared with lymphocytes) by flow cytometry (Fig. 6A).

We established a cell gate based on the unique light scatter profile of these cells, and found that they expressed huCMKLR1 (Fig. 6A). These cells, which are absent in a standard mononuclear cell preparation of normal human blood, were also positive for

CD14 and HLADR, which is consistent with the cells being human ascites macrophages

(data not shown). Cells falling in the lymphocyte gate were negative for huCMKLR1. We did, however, identify a small number of ascites DC, and found that the plasmacytoid DC

(Lin-HLADR+CD123+) were positive for huCMKLR1, as opposed to myeloid DC (Lin-

HLADR+CD123-) in the same ascites fluid, which were CMKLR1-negative, consistent with our previous studies of human blood DC subsets (8).

We next sorted CMKLR1+ human ascites cells for cytospin and Wright-Giemsa staining to examine cell morphology. CMKLR1+ cells were generally round with ruffled membranes, an abundance of granular and vacuolated cytoplasm, and round or reniform nuclei, displaying morphological features typical of macrophages (Fig. 6B).

61

Finally, we assessed the ability of chemerin to transmit intracellular signals in

huCMKLR1+ ascites leukocytes. Cells displaying the unique light scatter profile of tissue macrophages mobilized calcium in response to chemerin, while lymphocytes failed to respond, correlating with huCMKLR1 receptor expression (Fig. 6C).

62

Figure 6. Human macrophages express huCMKLR1 and are chemerin-responsive. (A) Leukocytes were harvested from fresh paracentesis samples and analyzed by flow cytometry. A scatter gate consistent with mouse peritoneal macrophages was established, and staining with anti-huCMKLR1 was assessed. Cells falling in a lymphocyte gate were also tested for huCMKLR1 expression. A scatter gate favoring dendritic cells was established, and huCMKLR1 staining was assessed on plasmacytoid DC (defined as Lin- (negative for CD3, CD14, CD16, CD19, CD20, CD56), HLADR+, CD123+) and myeloid DC (Lin-HLADR+CD123-). The filled histogram indicates isotype control antibody staining, and the open histogram indicates mCMKLR1 staining. (B) huCMKLR1+ peritoneal cells were sorted, harvested by cytospin, and stained by Wright- Giemsa. Cells were examined by light microscope using a 40x objective. (C) Human ascites fluid cells were pre-loaded with Fluo4-AM, and intracellular calcium mobilization was examined using continuous-acquisition flow cytometry. 70 nM chemerin was added as indicated. A representative data set for at least 3 experiments is shown for all parts in this figure.

63

Discussion

In this report, we demonstrate that freshly isolated human and mouse serosal macrophages express CMKLR1 and respond to chemerin. In contrast to human blood or ascites fluid plasmacytoid DC and differentiated, monocyte-derived DC, most mature DC in the mouse lack detectable CMKLR1. Moreover, although mCMKLR1 is transiently expressed on bone marrow derived DC, receptor expression diminished as the cells differentiated. The expression of mCMKLR1 on macrophages is regulated by cytokines and TLR ligands, stimuli known to direct the differentiation of specific effector macrophages. The suppression of mCMKLR1 by M1 stimuli, and the reciprocal upregulation of mCMKLR1 by M2-inducing stimuli, may indicate a role for chemerin in the host response to sterile tissue injury.

The trafficking potential of macrophages is known to be extensively reprogrammed in response to cytokine treatment or exposure to pathogens. At the transcriptional level, peritoneal murine macrophages downregulate CCR2 in response to

LPS (20), and upregulate CCR7 in response to IFNγ (21). This may facilitate the recruitment of macrophages to inflammatory sites enriched in CCL2, and then their subsequent efflux to lymph nodes via CCR7 ligands present in the blind openings of draining lymphatics (as demonstrated in DC migration (22) and lymphocyte exit from tissues (23)). Another macrophage-expressed chemoattractant receptor, FPR, is downregulated in response to IL-4 or TGFβ, but upregulated in response to LPS (24), thus displaying the opposite expression profile of macrophage-expressed CMKLR1. The

64

role of CMKLR1, or any other chemoattractant receptor, in macrophage signaling and positioning is therefore dependent on the nature of the stimuli in the local microenvironment. CMKLR1 expression in the resting state is consistent with a role for the receptor in rapid response to tissue injury. Downregulation of CMKLR1 in response to pro-inflammatory stimuli, and upregulation of CCR7, may allow macrophage efflux to lymph nodes following encounter with pathogens, where they can function as antigen presenting cells. Upregulation of CMKLR1 and increased sensitivity to chemerin in response to immune-suppressive stimuli (i.e. TGFβ) is consistent with macrophage localization to tissue sites requiring repair during the resolution phase of inflammation and wound healing. Thus the regulation of CMKLR1 expression is likely integral to macrophage function in response to injury. In this context, the fact that chemerin is activated by various serine proteases of the hemostatic and inflammatory cascades renders it uniquely suited to position macrophages to sites of bleeding, tissue damage, inflammation, and wound healing.

In addition to expression of CMKLR1 by resident serosal macrophage, we identified a rare population of blood borne F4/80+CD11b+ cells that express the receptor at high levels. Based on their light scatter profile, morphology, and lack of expression of the 7/4 antigen, we concluded that these cells were not classically defined blood monocytes but circulating blood macrophages. We also identified a population of

F4/80loCD11b+ macrophage-like cells in the peritoneum that express mCMKLR1. The relationship between these blood and peritoneal mCMKLR1+ cells, and resident

65

peritoneal macrophages remains to be determined, but we speculate that they may be tissue macrophage precursors. The phenotype of the circulating and F4/80loCD11b+ peritoneal macrophages with respect to fundamental macrophage functions is currently under investigation.

Mouse peritoneal macrophages migrate poorly in the transwell chemotaxis system. This may reflect the relatively sessile nature they display in vivo in the absence of additional stimulation. It is possible that macrophage adherence to plastic overrides existent chemotactic or haptotactic signals, and thus renders the cells immobile in vitro.

CMKLR1 is highly expressed on human plasmacytoid DC, so we were surprised to find that mouse plasmacytoid DC were CMKLR1-negative. Mouse and human plasmacytoid DC display many important phenotypic differences. For example, human plasmacytoid DC express the IL3 receptor (CD123) and do not express CD11c, while mouse plasmacytoid DC display the exact opposite surface phenotype (25). Since species divergence between mouse and man occurred ~70 million years ago (26), these differences have evolved relatively recently. Moreover, although murine and human macrophages share CMKLR1 expression, these cells also show dramatic species differences in phenotype and function. Of particular interest is the observation that mouse macrophages have evolved more potent systems for controlling viral infection than human macrophages. For example, nitric oxide production by murine macrophages is more robust than human (27), and it is well established that nitric oxide inhibits viral dissemination in vivo (28). Mouse macrophages express TLR9 (as do human

66

plasmacytoid DC), while human macrophages do not (26). Thus mouse macrophages may subsume some of the roles that human plasmacytoid DC play in combating viral infection.

While our studies show that most myeloid DC and plasmacytoid DC in the mouse are CMKLR1-negative in the steady state, they do not rule out an important role for

CMKLR1:chemerin in the development or dissemination of DC precursors. Moreover, since our study was limited to freshly isolated leukocytes from unmanipulated mice,

CMKLR1 may be upregulated on other leukocyte populations, particularly DC, in response to certain stimuli.

Tissue macrophages are major targets for HIV infection, and, due to their relative resistance to HIV-associated cytotoxicity, represent a crucial reservoir for virus (29).

CMKLR1 was initially described as expressed by monocyte-derived macrophages, and its first characterized function was as a HIV-1 coreceptor (17). Our data shows that

CMKLR1 is also expressed on freshly isolated macrophages. A recent study by Chen et al. (30) showed that treatment of monocyte-derived macrophages with TFGβ increased cell susceptibility for HIV-1 infection. The authors attributed this effect to the observed upregulation of CXCR4 on the TGFβ treated, culture-derived macrophages. Given our data, it is possible that TGFβ-driven upregulation of CMKLR1 also plays an important role in the increased sensitivity of macrophages to HIV-1 infection. This may be physiologically relevant, as HIV infected patients present with increased levels of circulating TGFβ (31).

67

In conclusion, mouse and human serosal macrophages express CMKLR1, an HIV co-receptor and chemoattractant receptor that has the potential to direct macrophage responses to sites of bleeding, tissue injury and repair.

68

Acknowledgements

We thank J. Zabel for helpful discussions. B.A.Z. is supported by National Institutes of

Health Training Grant 5 T32 AI07290-15. This work is supported by National Institutes

of Health Grants AI-59635, AI-47822, and GM-37734; Specialized Center of Research

Grant HL-67674; Digestive Disease Center Grant DK56339; and a Merit Award from the

Veterans Administration to E.C.B. S.J.A. is supported by a postdoctoral fellowship from

the Cancer Research Institute, New York. T.M.H. is supported by grants from the

National Institutes of Health (AI37113-09), the UC Discovery Program (Bio03-10367),

and the UC AIDS Program (1D03-B-005). B.J. holds the Canada Research Chair in

Inflammation and Immunity and is supported by grants from the National Cancer

Institute of Canada/Terry Fox Foundation and Nova Scotia Health Research Foundation.

69

References

1. Hume, D. A., I. L. Ross, S. R. Himes, R. T. Sasmono, C. A. Wells, and T. Ravasi. 2002. The mononuclear phagocyte system revisited. J Leukoc Biol 72:621-627. 2. 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. 3. Gordon, S. 2003. Alternative activation of macrophages. Nat Rev Immunol 3:23- 35. 4. Mantovani, A., S. Sozzani, M. Locati, P. Allavena, and A. Sica. 2002. Macrophage polarization: tumor-associated macrophages as a paradigm for polarized M2 mononuclear phagocytes. Trends Immunol 23:549-555. 5. Krieg, A. M. 2002. CpG motifs in bacterial DNA and their immune effects. Annu Rev Immunol 20:709-760. 6. Cella, M., M. Salio, Y. Sakakibara, H. Langen, I. Julkunen, and A. Lanzavecchia. 1999. Maturation, activation, and protection of dendritic cells induced by double-stranded RNA. J Exp Med 189:821-829. 7. Suzuki, F., R. B. Pollard, S. Uchimura, T. Munakata, and H. Maeda. 1990. Role of natural killer cells and macrophages in the nonspecific resistance to tumors in mice stimulated with SMANCS, a polymer-conjugated derivative of neocarzinostatin. Cancer Res 50:3897-3904. 8. Zabel, B. A., A. M. Silverio, and E. C. Butcher. 2005. Chemokine-like receptor 1 expression and chemerin-directed chemotaxis distinguish plasmacytoid from myeloid dendritic cells in human blood. J Immunol 174:244-251. 9. Lindstedt, M., K. Lundberg, and C. A. Borrebaeck. 2005. Gene family clustering identifies functionally associated subsets of human in vivo blood and tonsillar dendritic cells. J Immunol 175:4839-4846. 10. Zabel, B. A., S. J. Allen, P. Kulig, J. A. Allen, J. Cichy, T. M. Handel, and E. C. Butcher. 2005. Chemerin activation by serine proteases of the coagulation, fibrinolytic and inflammatory cascades. J Biol Chem. 11. Wittamer, V., B. Bondue, A. Guillabert, G. Vassart, M. Parmentier, and D. Communi. 2005. Neutrophil-mediated maturation of chemerin: a link between innate and adaptive immunity. J Immunol 175:487-493. 12. Bauer, M., V. Redecke, J. W. Ellwart, B. Scherer, J. P. Kremer, H. Wagner, and G. B. Lipford. 2001. Bacterial CpG-DNA triggers activation and maturation of human CD11c-, CD123+ dendritic cells. J Immunol 166:5000-5007. 13. Ponath, P. D., S. Qin, T. W. Post, J. Wang, L. Wu, N. P. Gerard, W. Newman, C. Gerard, and C. R. Mackay. 1996. Molecular cloning and characterization of a human eotaxin receptor expressed selectively on eosinophils. J Exp Med 183:2437-2448. 14. Palermo, D. P., M. E. DeGraaf, K. R. Marotti, E. Rehberg, and L. E. Post. 1991. Production of analytical quantities of recombinant proteins in Chinese hamster ovary cells using sodium butyrate to elevate gene expression. J Biotechnol 19:35-47. 70

15. Wittamer, V., J. D. Franssen, M. Vulcano, J. F. Mirjolet, E. Le Poul, I. Migeotte, S. Brezillon, R. Tyldesley, C. Blanpain, M. Detheux, A. Mantovani, S. Sozzani, G. Vassart, M. Parmentier, and D. Communi. 2003. Specific Recruitment of Antigen- presenting Cells by Chemerin, a Novel Processed Ligand from Human Inflammatory Fluids. J Exp Med 198:977-985. 16. Vermi, W., E. Riboldi, V. Wittamer, F. Gentili, W. Luini, S. Marrelli, A. Vecchi, J. D. Franssen, D. Communi, L. Massardi, M. Sironi, A. Mantovani, M. Parmentier, F. Facchetti, and S. Sozzani. 2005. Role of ChemR23 in directing the migration of myeloid and plasmacytoid dendritic cells to lymphoid organs and inflamed skin. J Exp Med 201:509-515. 17. Samson, M., A. L. Edinger, P. Stordeur, J. Rucker, V. Verhasselt, M. Sharron, C. Govaerts, C. Mollereau, G. Vassart, R. W. Doms, and M. Parmentier. 1998. ChemR23, a putative chemoattractant receptor, is expressed in monocyte-derived dendritic cells and macrophages and is a coreceptor for SIV and some primary HIV-1 strains. Eur J Immunol 28:1689-1700. 18. Henderson, R. B., J. A. Hobbs, M. Mathies, and N. Hogg. 2003. Rapid recruitment of inflammatory monocytes is independent of neutrophil migration. Blood 102:328-335. 19. Gee, K. R., K. A. Brown, W. N. Chen, J. Bishop-Stewart, D. Gray, and I. Johnson. 2000. Chemical and physiological characterization of fluo-4 Ca(2+)-indicator dyes. Cell Calcium 27:97-106. 20. Zhou, Y., Y. Yang, G. Warr, and R. Bravo. 1999. LPS down-regulates the expression of chemokine receptor CCR2 in mice and abolishes macrophage infiltration in acute inflammation. J Leukoc Biol 65:265-269. 21. Donnini, A., K. Argentati, R. Mancini, A. Smorlesi, B. Bartozzi, G. Bernardini, and M. Provinciali. 2002. Phenotype, antigen-presenting capacity, and migration of antigen-presenting cells in young and old age. Exp Gerontol 37:1097-1112. 22. Saeki, H., A. M. Moore, M. J. Brown, and S. T. Hwang. 1999. Cutting edge: secondary lymphoid-tissue chemokine (SLC) and CC chemokine receptor 7 (CCR7) participate in the emigration pathway of mature dendritic cells from the skin to regional lymph nodes. J Immunol 162:2472-2475. 23. Debes, G. F., C. N. Arnold, A. J. Young, S. Krautwald, M. Lipp, J. B. Hay, and E. C. Butcher. 2005. Chemokine receptor CCR7 required for T lymphocyte exit from peripheral tissues. Nat Immunol 6:889-894. 24. Dai, Y., J. Major, M. Novotny, and T. A. Hamilton. 2005. IL-4 inhibits expression of the formyl peptide receptor gene in mouse peritoneal macrophages. J Interferon Cytokine Res 25:11-19. 25. Hochrein, H., M. O'Keeffe, and H. Wagner. 2002. Human and mouse plasmacytoid dendritic cells. Hum Immunol 63:1103-1110. 26. Mestas, J., and C. C. Hughes. 2004. Of mice and not men: differences between mouse and human immunology. J Immunol 172:2731-2738.

71

27. Schneemann, M., and G. Schoedon. 2002. Species differences in macrophage NO production are important. Nat Immunol 3:102. 28. Akaike, T., and H. Maeda. 2000. Nitric oxide and virus infection. Immunology 101:300-308. 29. Balestra, E., C. F. Perno, S. Aquaro, S. Panti, A. Bertoli, M. Piacentini, F. Forbici, R. D'Arrigo, R. Calio, and E. Garaci. 2001. Macrophages: a crucial reservoir for human immunodeficiency virus in the body. J Biol Regul Homeost Agents 15:272-276. 30. Chen, S., D. L. Tuttle, J. T. Oshier, H. J. Knot, W. J. Streit, M. M. Goodenow, and J. K. Harrison. 2005. Transforming growth factor-beta1 increases CXCR4 expression, stromal-derived factor-1alpha-stimulated signalling and human immunodeficiency virus-1 entry in human monocyte-derived macrophages. Immunology 114:565-574. 31. Navikas, V., J. Link, B. Wahren, C. Persson, and H. Link. 1994. Increased levels of interferon-gamma (IFN-gamma), IL-4 and transforming growth factor-beta (TGF-beta) mRNA expressing blood mononuclear cells in human HIV infection. Clin Exp Immunol 96:59-63.

72

CHAPTER 3 : CHEMOKINE-LIKE RECEPTOR-1 EXPRESSION BY CENTRAL

NERVOUS SYSTEM-INFILTRATING LEUKOCYTES AND INVOLVEMENT IN

A MODEL OF AUTOIMMUNE DEMYELINATING DISEASE

Kareem L. Graham,* Brian A. Zabel,* Sanam Loghavi,* Luis A. Zuniga,* Peggy

P. Ho†, Raymond A. Sobel,* and Eugene C. Butcher*

*Laboratory of Immunology and Vascular Biology, Department of Pathology, Stanford

University School of Medicine, Stanford, CA 94305 and Center for Molecular Biology and Medicine, Veterans Affairs Palo Alto Health Care System, Palo Alto, CA 94304

†Department of Neurology, Stanford University School of Medicine, Stanford, CA 94305

Reproduced from:

J Immunol. 2009 Nov 15;183 (10):6717-23.

Used by copyright permission of The Journal of Immunology.

73

Summary

We examined the involvement of chemokine-like receptor-1 (CMKLR1) in experimental

autoimmune encephalomyelitis (EAE), a model of human multiple sclerosis. Upon EAE

induction by active immunization with myelin oligodendrocyte glycoprotein amino acids

( 35–55 MOG35–55), microglial cells and CNS-infiltrating myeloid dendritic cells expressed

CMKLR1, as determined by flow cytometric analysis. In addition, chemerin, a natural

ligand for CMKLR1, was up-regulated in the CNS of mice with EAE. We found that

CMKLR1-deficient (CMKLR1 knockout (KO)) mice develop less severe clinical and

histologic disease than their wild-type (WT) counterparts. CMKLR1 KO lymphocytes

proliferate and produce proinflammatory cytokines in vitro, yet MOG35–55-reactive

CMKLR1 KO lymphocytes are deficient in their ability to induce EAE by adoptive

transfer to WT or CMKLR1 KO recipients. Moreover, CMKLR1 KO recipients fail to fully support EAE induction by transferred MOG-reactive WT lymphocytes. The results imply involvement of CMKLR1 in both the induction and effector phases of disease. We conclude that CMKLR1 participates in the inflammatory mechanisms of EAE and

represents a potential therapeutic target in multiple sclerosis.

74

Introduction

Experimental autoimmune encephalomyelitis (EAE) is a widely studied animal

model of multiple sclerosis (MS), an inflammatory demyelinating disease of the CNS of unknown etiology. Early tissue injury in EAE and MS is mediated by myelin Ag-specific

CD4+ T lymphocytes (1) that require expression of α4 integrin for recruitment from the

blood and entry into the CNS parenchyma (2). Blockade of α4 integrin suppresses MS

and EAE (3, 4). However, the development of EAE and MS is marked by tightly

controlled regulation of other adhesion molecules, as well as numerous chemoattractant

receptors and their respective ligands (5) that are thought to help control immune cell

recruitment, microenvironmental positioning, and function within the inflamed CNS. In

addition to CD4+ T cells, MS and EAE lesions contain several other cell types, including

recruited B lymphocytes, CD8+ T lymphocytes, macrophages, and CNS-resident cells

(e.g. astrocytes and microglia) (reviewed in Ref. (6)). Trafficking and chemoattractant

receptors differentially expressed by these populations thus offer a rich potential for

regulating pathogenic CNS inflammation.

Chemokine-like receptor-1 (CMKLR1; also known as ChemR23 or Dez) is a chemoattractant receptor that is expressed by unique subsets of dendritic cells (DC), as

well as tissue-resident macrophages in mice and humans (7, 8, 9). CMKLR1-expressing

cells migrate to chemerin, a proteolytically regulated chemoattractant (9, 10, 11). TLR

ligands and cytokines regulate CMKLR1 expression on ex vivo mouse macrophages (8).

Macrophages and DC are prominent in EAE and MS inflammatory lesions and have

75

critical roles in mediating tissue injury. These cells can contribute to the disease process through multiple mechanisms, including production of proinflammatory cytokines, Ag

processing and presentation to autoreactive T lymphocytes, and production of reactive

oxygen species that directly induce damage to myelin (1, 12, 13). We hypothesized that

CMKLR1 plays a role in inflammatory processes within the CNS. In this report, we

evaluated the role of CMKLR1 in the EAE model of MS.

76

Materials and Methods

Mice

CMKLR1 knockout (KO) mice were obtained from Deltagen and fully backcrossed (nine

generations) onto the C57BL/6 background. C57BL/6 control mice were purchased from

The Jackson Laboratory. Female mice (8–12 wk old) were used in all experiments. All

animal experiments were conducted in accordance with approved Stanford and National

Institutes of Health Institutional Animal Care and Use Committee guidelines.

EAE induction by active immunization

Myelin oligodendrocyte glycoprotein (MOG) peptide amino acids 35–35

(MEVGWYRSPFSRVVHLYRNGK; MOG35–55) was synthesized by the Stanford Protein

and Nucleic Acid Facility (Stanford, CA). For active EAE, MOG35–55 was dissolved in

PBS at 2 mg/ml. The MOG35–55 solution was then emulsified in an equal volume of CFA,

consisting of incomplete Freund’s adjuvant (Difco) plus 4 mg/ml heat-inactivated

Mycobacterium tuberculosis (strain H37 RA; Difco). The emulsion (100 µl) was injected

s.c.; mice were given an i.v. boost of 400 ng pertussis toxin (List Biological Laboratories)

at the time of immunization and again 2 days later. Clinical EAE was assessed daily as

previously described (14).

EAE induction by adoptive transfer

Mice were immunized s.c. with 100 µg MOG35–55 in CFA. Draining lymph nodes (LN)

and spleen cells were harvested 10 days later and resuspended at 5 x 106 cells/ml in

RPMI 1640 supplemented with 10% FBS, penicillin/streptomycin, L-glutamine, sodium

77

pyruvate, nonessential amino acids, and 2-ME. MOG35–55 was added at 10 µg/ml and recombinant murine IL-12 (R&D Systems) at 10 ng/ml. After 4 days in culture (37°C,

8% CO2), lymphocytes were isolated using Lympholyte-Mammal (Cedarlane

Laboratories), pooled, and resuspended in HBSS for transfer. Mice received 1–2 x 107 viable cells i.v. Pertussis toxin (400 ng) was given i.v. immediately after cell transfer and again 2 days later. Transferred lymphocytes were >95% CD3+, as determined by flow cytometric analysis (data not shown).

Generating the anti-CMKLR1 mAb BZ186

CMKLR1 KO mice were immunized via s.c. injection of ~4 x 107 wild-type (WT) peritoneal exudate cells. For the first immunization, cells were suspended in saline and emulsified in CFA; incomplete Freund’s adjuvant was used for two subsequent injections. Hybridomas producing anti-mouse CMKLR1 mAb were subcloned.

Specificity was confirmed by reactivity with mouse CMKLR1 transfectants and lack of reactivity with human CMKLR1 transfectants. ELISA (BD Pharmingen) was used to determine the isotype (IgG1κ) of the resulting mouse anti-mouse CMKLR1 mAb, designated BZ186.

ELISAs

Mice were immunized s.c. with 100 µg MOG35–55 in CFA. After 10 days, draining LN and spleen cells were harvested and resuspended in RPMI 1640 with supplements. Cells were plated at 2 x 105 cells/well in flat-bottom 96-well plates, MOG35–55 peptide was added at various concentrations, and cells were incubated at 37°C, 8% CO2. Culture

78

supernatants were harvested at 72 h and levels of IFN-γ, TNF (both from BD

Pharmingen), and IL-17 (eBioscience) in triplicate wells were determined by sandwich

ELISA according to the manufacturer’s instructions.

Proliferation assays

LN and spleen cells were cultured as described above for 72 h in triplicate wells with a

3 range of MOG35–55 concentrations. [ H]Thymidine was added for the last 18–24 h of

culture and thymidine incorporation was assessed using a β-plate scintillation counter.

CNS mononuclear cell preparation

Mice were perfused through the heart with 30 ml cold PBS. Spinal cords were extracted,

minced, and incubated with HBSS containing 0.2 U of Liberase R1 (Roche), 50 µg/ml

DNase I (Roche), and 25 mM HEPES for 30 min at 37°C. Digested tissue was forced

through stainless steel mesh and mononuclear cells were collected from 30:70%

discontinuous Percoll gradients.

Flow cytometry

mAbs directed against mouse CD3 (145–2C11), CD11c (HL3), CD11b (M1/70), CD19

(1D3), CD45 (30-F11), and CD45R/B220 (RA3–6B2) were from eBioscience or BD

Pharmingen. For flow cytometric analysis of CMKLR1 expression, PE-conjugated rat

anti-mouse IgG1 (BD Pharmingen) was used to detect BZ186 or its isotype control, mAb

DREG200 (mIgG1 with specificity for human L-selectin (CD62L)). Staining buffer

consisted of PBS containing 2% BSA plus 0.1% sodium azide. Data were acquired using

an LSRII flow cytometer and analyzed with FlowJo Software.

79

Histology

Brains and spinal cords were extracted and fixed in 10% buffered formalin. Formalin- fixed tissue was embedded in paraffin and sections were stained with Luxol fast blue-

H&E stain. CNS inflammatory foci (>10 mononuclear cells/focus) in leptomeninges and parenchyma were counted in each mouse sample in a blinded fashion by one of the authors (R.A.S.).

For immunohistochemical analysis, sections of paraffin-embedded brain and spinal cord tissue were subjected to sodium citrate Ag retrieval, followed by immunostaining with anti-F4/80 mAb (Serotec). Biotinylated, species-specific secondary was added (Vector

Laboratories), followed by incubation with ABC reagent (Vector Laboratories). The reactions were developed using diaminobenzidine (Sigma-Aldrich) and visualized by light microscopy.

The incidence of demyelinating lesions in the CNS tissues was assessed using a semiquantitative scoring system. The presence (score = 1) or absence (score = 0) of a demyelinated lesion as indicated by loss of blue staining in Luxol fast blue-stained sections was determined in four white matter regions (posterior columns, anterior columns, and two lateral columns), in two cross-sections of the spinal cord, the brain stem, and cerebellar white matter in each mouse. The maximum total score for each mouse was 10.

Statistics

80

Nonparametric clinical EAE data were analyzed using the Mann-Whitney U test. All parametric data were analyzed using the Student’s t test. Fisher’s exact test was used to compare disease incidence. Values of p less than 0.05 were considered statistically significant.

81

Supplemental Materials and Methods

T cell proliferation and cytokine assays

T cells were isolated from draining LN or spleens of MOG-immunized mice via a CD3+

column (R&D Systems). CD3+ T cell purity was >80% for spleen preparations and

>98% for LN, as determined by flow cytometric analysis (data not shown). For

preparation of antigen presenting cells (APC), red blood cells were removed from the

spleens of naive WT or CMKLR1 KO mice via hypotonic lysis. Spleen cells were then resuspended at 5x107 cells/ml in PBS, followed by incubation with 50 μg/ml mitomycin

C (MMC; Sigma) for 20 min at 37°C. MMC-treated splenocytes were washed

extensively and resuspended in RPMI with supplements. T cells and APC were plated in

96-well flat-bottom plates; each well contained 2.5x105 APC mixed with 5x104 T cells.

MOG35-55 was added at 25 μg/ml and cells were incubated for 72 h at 37°C, 8% CO2. For

assessment of proliferation, [3H]-thymidine was added for the last 18-24 h of culture, and

thymidine incorporation in triplicate wells was assessed using a beta-plate scintillation counter. For cytokine measurements, culture supernatants were harvested at 72 h; levels

of IFNγ (BD Pharmingen) and IL-17 (eBioscience) in triplicate wells were determined

by sandwich ELISA according to the manufacturer’s instructions.

Peritoneal macrophage responses

Peritoneal lavage cells (PLCs) were collected by lavage of the peritoneal cavity with 10

ml sterile HBSS containing 2% FBS. Harvested cells were washed and resuspended in

DMEM containing 10% FBS, penicillin/streptomycin, L-glutamine, sodium pyruvate and

82

nonessential amino acids. Cells were plated in 24-well tissue culture plates at 5x105 cells per well and allowed to adhere for 2 h at 37°C. Non-adherent cells were washed away with PBS, and adherent cells were stimulated with 1 μg/ml lipopolysaccharide (LPS;

Sigma) plus 10 ng/ml recombinant murine IFNγ (eBioscience). Culture supernatants were harvested after 24 h, and levels of IL-6 and TNF were assayed by ELISA. Where indicated, mice were administered 1.0 ml of 3% thioglycollate (Sigma) via i.p. injection

72 h prior to PLC harvest.

Real-time quantitative PCR

Mice were perfused, and spinal cord RNA was extracted using a Stratagene RNA miniprep per the supplier’s instructions. Gene expression was determined by quantitative PCR (QPCR) using an Applied Biosystems 7900HT real-time PCR instrument equipped with a 384-well reaction block. 100-200 ng total RNA was used as template for cDNA synthesis using MMLV Reverse Transcriptase (Applied Biosystems) with oligo dT primers (Invitrogen) according to the supplier’s instructions. The cDNA was diluted 1:12 and amplified by quantitative PCR in triplicate wells using 10 pmols of gene specific primers in a total volume of 10 μl with Power SYBR Green QPCR Master

Mix (Applied Biosystems), according to manufacturer’s instructions. Relative gene expression normalized to cyclophilin A (cycA) was calculated; relative gene expression values (multiplied by 104 to simplify data presentation) are displayed. Primer sequences

were as follows: cycA forward, 5'-GAGCTGTTTGCAGACAAAGTTC-3'; cycA reverse,

5'-CCCTGGCACATGAATCCTGG-3'; chemerin forward, 5'-

83

TACAGGTGGCTCTGGAGGAGTTC-3'; chemerin reverse, 5'-

CTTCTCCCGTTTGGTTTGATTG-3'; forward, 5’-

CGGTCTTCCTGGTGGTGA-3'; cmklr1 reverse: 5'-GCACATGGCCTTCCCGAA-3'.

Evaluation of anti-CMKLR1 mAbs by flow Cytometry

Spleen cells or PLCs were harvested from naïve C57BL/6 mice, and stained with rat anti-

mouse CMKLR1 mAb BZ194 or mouse antimouse CMKLR1 mAb BZ186 and their respective isotype controls. PE-conjugated goat anti-rat IgG (BD Pharmingen) was used to detect BZ194; PE-labeled rat anti-mouse IgG1 was used to detect BZ186. Cells were

then stained with fluorophore-labeled monoclonal antibodies to identify the indicated

leukocyte subsets and data were acquired on a flow cytometer. CMKLR1 expression was analyzed on NK cells (NK1.1+CD3-), T cells (CD3+NK1.1-), B cells (CD19+) and DC

(CD3-CD19-NK1.1-CD11c+) isolated from the spleen; macrophages (F4/80+CD11b+) were taken from the peritoneal cavity. The anti-NK1.1 mAb (PK136) was from eBioscience.

84

Results

Attenuation of clinical EAE in CMKLR1 KO mice

CMKLR1 KO mice develop normally; they are healthy, fertile, and have no overt immune system abnormalities or defects. We evaluated the cellular composition of blood,

thymus, bone marrow, spleen, and LN derived from WT and CMKLR1 KO mice. There were no differences between the two groups with respect to total numbers or percentages

of numerous leukocyte subsets, including T cells, B cells, NK cells,

monocytes/macrophages, granulocytes, or DC in any of the tissues analyzed (data not

shown).

To evaluate the role of CMKLR1 in a model of autoimmune demyelinating

disease, we induced EAE in CMKLR1 KO mice by active immunization. There was no

significant difference between CMKLR1 KO and WT mice with respect to day of disease

onset (Fig. 1 and Table I). Disease incidence was also similar between the two groups: 33

of 33 (100%) WT mice developed clinical EAE vs 30 of 33 (91%) of CMKLR1 KO mice

(Table I). CMKLR1 KO mice did not, however, develop the same severity of acute EAE

as their WT counterparts. The average maximal disease score for WT mice was 3.4,

compared with 2.6 for CMKLR1 KO mice (p < 0.005, as determined by Mann-Whitney U

test; Table I). In addition, clinical EAE in CMKLR1 KO mice was significantly reduced

throughout the chronic phase of disease (Fig. 1). Together, these data indicate that

CMKLR1 KO mice are susceptible to EAE induction by active immunization, but that

85

CMKLR1 is required for maximal acute EAE. Also, CMKLR1 may promote chronic/progressive EAE.

86

Figure 1. Reduced clinical EAE in CMKLR1 KO mice. EAE was induced by active immunization and mice were monitored daily for clinical disease as described in Materials and Methods. Data are pooled from five independent experiments, each consisting of 4–10 mice per group, and are presented as mean clinical score ± SEM vs time. *, p < 0.05, as determined by Mann-Whitney U test.

Table I. Clinical EAE in actively immunized WT and CMKLR1 KO mice

Incidence of Clinical Mean Day of Onset Mean Maximal Score EAEa (SEM) (SEM)

WT 33/33 (100%) 12.2 (0.7) 3.45 (0.2) CMKLR1 30/33 (91%) 13.1 (0.8) 2.60 (0.2)b KO

a Data are pooled from five independent experiments with each experiment consisting of 4–10 mice per group. b p < 0.05, as determined by Mann-Whitney U test. 87

Reduced histological EAE in CMKLR1 KO mice

EAE and MS are characterized by perivascular inflammatory cell infiltrates

located predominantly in the CNS white matter. We evaluated CNS lesions in WT and

CMKLR1 KO mice, analyzing brain and spinal cord tissue harvested either during the

acute or chronic phase of EAE (13 or 46 days post-immunization (p.i.), respectively). As

shown in Table II, WT and CMKLR1 KO mice had similar numbers of parenchymal

inflammatory foci at day 13 p.i. In contrast, CMKLR1 KO mice had significantly fewer

meningeal inflammatory foci at this time point (Fig. 2, A and B). CMKLR1 KO mice with

chronic EAE (day 46 p.i.) also had significantly fewer meningeal and parenchymal

inflammatory lesions than their WT counterparts (Table II and Fig. 2, C and D).

Consistent with reduced inflammation, there was a trend toward reduced demyelination in

CMKLR1 KO mice with chronic EAE (day 46 p.i.). Although the difference did not achieve statistical significance (p = 0.06), mice with the most extensive demyelination

may have died before the end of the experiment (mortality rate was 9 of 33 (27%) in WT

mice vs 2 of 33 (6.1%) in CMKLR1 KO animals). Thus, our data do not exclude a role

for CMKLR1 in the vulnerability of the myelin to demyelination or the capacity of

oligodendrocytes to remyelinate lesions. Demyelination at the early (day 13 p.i.) time point was generally limited to perivascular areas associated with inflammatory cuffs. At

the later time point (day 46 p.i.), demyelinated foci were generally larger and less

inflammatory in both WT and CMKLR1 KO mice (Table II).

88

Table II. Histological EAE in actively immunized WT and CMKLR1 KO mice

Number of Parenchymal Foci Meningeal Foci Demyelination Mice (SEM)a (SEM)a Score (SEM)a

Day 13 p.i. WT 5 98.0 (12.4) 116 (13.3) 7.8 (0.9) CMKLR1 7 85.7 (7.4) 74.7 (5.1)b 8.23 (0.5) KO Day 46 p.i. WT 7 47.4 (9.9) 35.9 (5.4) 5.86 (0.6) CMKLR1 14 21.4 (4.8)b 20.8 (3.8)b 4.23 (0.7)c KO

a Brain and spinal cord tissue was harvested 13 or 46 days after induction of EAE by active immunization. Histological changes were evaluated as described in Materials and Methods. b p < 0.05, as determined by Student’s t test. c The difference in the demyelination score between WT and CMKLR1 KO mice at day 46 p.i. did not reach statistical significance (p = 0.06).

89

Figure 2. Reduced histological EAE in CMKLR1 KO mice. Representative spinal cord sections are shown from actively immunized WT (left panels) or CMKLR1 KO mice (right panels) that were killed at 13 (A and B) or 46 (C and D) days p.i. A, Meningeal and parenchymal mononuclear cell infiltrates typical of acute EAE in the spinal cord of a WT mouse sacrificed on day 13 p.i. B, Less meningeal infiltration in the spinal cord of CMKLR1 KO mouse with EAE at day 13 p.i. C, Typical meningeal and parenchymal mononuclear cell infiltrates in the spinal cord of a WT mouse with chronic EAE. D, Meningitis and mild parenchymal inflammation are present in the spinal cord of a CMKLR1 KO mouse sacrificed at day 46 p.i. Sections from paraffin-embedded WT (E) and CMKLR1 KO (F) brain and spinal cord tissue harvested at day 46 p.i. were subjected to immunostaining with anti-F4/80 mAb. Reactions were developed with diaminobenzidine chromogen and counterstained with hematoxylin. White arrows highlight microglia, black arrows indicate foamy macrophages. Magnification = 160x. Bar = 50 µm. 90

CNS inflammatory lesions of CMKLR1 KO mice contain fewer F4/80+ cells

Because CMKLR1 is expressed by F4/80+CD11b+ mouse macrophages (8), we evaluated CNS tissue from mice with EAE for F4/80+ cells by immunohistochemistry.

Analysis of CNS tissue harvested at day 13 p.i. revealed no qualitative or quantitative

differences in F4/80 staining between WT and CMKLR1 KO mice (data not shown). In contrast, WT mice with chronic EAE had more F4/80+ macrophages in the leptomeninges

than CMKLR1 KO animals (day 46 p.i.). Compared with CMKLR1 KO mice, WT CNS

tissue also contained more foamy macrophages and microglia in parenchymal lesions

(Fig. 2, E and F). Together, these results are consistent with a contribution of CMKLR1

to recruitment of F4/80+ cells to the CNS.

We used flow cytometry to quantify leukocyte subsets isolated from the spinal

cords of WT and CMKLR1 KO mice during the preclinical and acute phases of EAE.

There were no differences between WT and CMKLR1 KO mice with EAE with respect

to absolute numbers or percentages of CD4+ or CD8+ T cells, B cells, CD11b+ microglia,

or CD11c+ DC in the spinal cord (data not shown). Thus, although there is more

inflammation in WT mice, CNS inflammatory lesions in KO and WT mice appear to

recruit similar leukocyte subsets.

Lymphocyte proliferation and cytokine production

To determine whether lymphocyte activation defects contributed to differences in

inflammatory cell infiltration of the CNS, we assessed recall proliferation and cytokine

91

responses of WT and CMKLR1 KO lymphocytes to MOG35–55 peptide. When they were

restimulated with MOG35–55 in vitro, CMKLR1 KO-draining LN cells and splenocytes proliferated at levels comparable to their WT counterparts (Fig. 3A). CMKLR1 KO-

draining LN cells generally produced lower levels of IFN-γ, IL-17, and TNF than WT LN

cells, but these differences did not reach statistical significance in pair-wise comparisons

for most of the MOG35–55 peptide concentrations tested (Fig. 3B, left panels). In contrast,

CMKLR1 KO splenocytes produced IFN-γ, IL-17, and TNF at levels comparable or

superior to their WT counterparts (Fig. 3B, right panels).

To further evaluate the immune activation and Ag-presenting capabilities of WT

and CMKLR1 KO cells, we isolated CD3+ T cells from the draining LN or spleens of

MOG-immunized mice, followed by stimulation with MOG35–55 in presence of

mitomycin C-treated splenocytes. In an effort to distinguish between potential CMKLR1-

related T cell and APC effects, we used all possible combinations of WT and CMKLR1

KO T lymphocytes and APC in these coculture studies. We found that CMKLR1 KO T

cells were fully capable of proliferative and cytokine responses when stimulated with WT

or CMKLR1 KO APC. In addition, CMKLR1 KO splenocytes induced robust proliferation and cytokine production by WT and CMKLR1 KO T cells (supplemental

Fig. 1). Thus, the reduced severity of EAE in CMKLR1 KO mice is not due to intrinsic defects in lymphocyte or APC functional capacity.

92

Figure 3. Recall proliferation and cytokine responses of lymphocytes from CMKLR1 KO mice. Mice were immunized with MOG35–55 emulsified in CFA. After 10 days, spleen cells and draining LN cells were restimulated in vitro with the indicated concentrations of MOG35– 55. A, After 72 h of stimulation, proliferation of LN (left panel) or spleen cells (right panel) was assessed by [3H]thymidine incorporation assay. Data are presented as a stimulation index (mean cpm with Ag ÷ mean cpm without Ag); bars represent SEM. B, Culture supernatants were collected after 72 h and levels of the indicated cytokines produced by LN (left panels) or spleen cells (right panels) were measured by ELISA. *, p < 0.05, as determined by Student’s t test. Data shown are representative of two independent experiments with similar results. 93

Induction of EAE by adoptive transfer

To determine the relative role of CMKLR1 expression on effector leukocytes vs

recipient cells and tissues, we performed adoptive transfer experiments. MOG35–55-

reactive WT or CMKLR1 KO lymphocytes were generated and transferred into WT

recipient mice (WT:WT or KO:WT transfers, respectively). CMKLR1 KO lymphocytes

were much less effective than their WT counterparts at transferring EAE to WT recipients

(Fig. 4). In addition, WT:KO and KO:KO transfers did not produce EAE with the same

frequency as WT:WT transfers. Moreover, histologic disease was less severe when KO mice were either the donors or recipients of MOG-specific lymphocytes (Table III).

These studies indicate that CMKLR1 expression on donor mononuclear cells and on

recipient cells or tissues is required for maximal EAE induction in an adoptive transfer

setting. The failure of CMKLR1 KO recipients to develop EAE upon transfer with WT

MOG-specific lymphocytes, cells that effectively induce EAE in WT recipients, indicates

that CMKLR1 is involved in CNS inflammation during the effector phase of disease.

Furthermore, the inability of MOG-reactive CMKLR1 KO lymphocytes to induce EAE

upon adoptive transfer to WT hosts suggests an unexpected role for CMKLR1 in the

induction phase of disease.

94

Figure 4. Induction of EAE by adoptive transfer of MOG-reactive lymphocytes. EAE was induced in WT or CMKLR1 KO mice by passive transfer of WT or CMKLR1 KO MOG35–55-reactive lymphocytes as described in Materials and Methods. Data shown are pooled from adoptive transfers that were performed at various times (eight independent experiments, n = 1–4 recipient mice per group). Values are presented as mean clinical score vs time.

95

Peritoneal macrophage responses

We next evaluated WT and CMKLR1 KO peritoneal lavage cells (PLCs) for their responses to proinflammatory stimuli. To enrich for macrophages, nonadherent PLCs

were removed by washing. Adherent cells were stimulated in vitro with LPS plus IFN-γ

(LPS/IFN-γ). Resident PLCs derived from CMKLR1 KO mice produced IL-6 and TNF at

levels similar to WT mice upon LPS/IFN-γ stimulation (supplemental Fig. 2A). To elicit inflammatory macrophages, we gave WT and CMKLR1 KO mice an i.p. injection of

thioglycollate. Thioglycollate-elicited peritoneal macrophages from CMKLR1 KO mice

were also competent in their ability to produce IL-6 and TNF in response to stimulation

with LPS/IFN-γ (supplemental Fig. 2B). Thus, CMKLR1 KO resident and inflammatory macrophages are fully capable of mounting effector cytokine responses.

Analysis of CNS mononuclear cells by flow cytometry

CMKLR1 expression has been reported on APCs in mice and humans (7, 8, 9,

10), as well as on human NK cells (15). Using a novel anti-CMKLR1 mAb, we analyzed

CMKLR1 expression on mouse spinal cord mononuclear cells by flow cytometry. An

earlier anti-CMKLR1 mAb, designated BZ194, was generated by immunizing rats with

an amino-terminal CMKLR1 peptide. As reported previously (8), this mAb specifically

stains mouse macrophages but not NK cells, even though NK cells express the receptor.

The mAb used in the present study (BZ186) was generated by immunizing CMKLR1 KO

96

mice with WT peritoneal exudate cells; it stains both mouse macrophages and NK cells

(supplemental Fig. 4).

CMKLR1 was expressed at relatively low levels on microglia, defined as CD3–

CD19–CD11b+CD45low (16), in naive mice (data not shown), as well as in mice with acute EAE (Fig. 5A). Additional characterization of CNS mononuclear cells demonstrated

CMKLR1 expression by a small subset of CNS-infiltrating macrophages (CD45highCD3–

CD19–CD11b+CD11c–; Fig. 5C). CNS-infiltrating CD45highCD3–CD19–CD11b–

CD11cintB220+ plasmacytoid DC (pDC) are CMKLR1– (Fig. 5D). Conversely,

CD45highCD3–CD19–CD11b+CD11chighB220– myeloid DC (mDC) isolated from the spinal cords of EAE mice are CMKLR1+ (Fig. 5E). CMKLR1+ macrophages were also detectable in the draining LN and spleens of mice with preclinical and acute EAE, though at much lower frequency than in the CNS (data not shown). CMKLR1 was not detectable on macrophages or DC derived from the LN or spleen of naive mice (data not shown).

97

Figure 5. Detection of CMKLR1+ cells in CNS of mice with EAE. Mononuclear cells were isolated from the spinal cords of mice with acute EAE as described in Materials and Methods. Cells were stained with anti-mouse CMKLR1 mAb 98

BZ186 or DREG200 isotype control mAb, followed by incubation with PE-conjugated anti-mouse IgG1. Lastly, mAbs directly conjugated to cell surface Ags were added and cells were analyzed by flow cytometry. For all histograms, the red line represents isotype control mAb staining. A, CD3–CD19–CD11b+CD45low microglia were analyzed for expression of CMKLR1. B, Cells were gated as CD45high and cells in the CD3–CD19– gate were analyzed for expression of CD11b and CD11c. Leukocytes were gated on macrophages (R1), pDC (R2), or mDC (R3), which were then analyzed for expression of CMKLR1. C, A subset of CD11c–CD11b+ macrophages (R1 gate) expresses CMKLR1 (indicated by the arrow). D, CD11cintCD11b–B220+ pDC (R2 gate) are CMKLR1- negative. E, CD11chighCD11b+B220– mDC (R3 gate) express CMKLR1. CNS cells were pooled from three mice for analysis; data shown are representative of three independent experiments with similar results. FSC, forward scatter.

99

Chemerin, a natural ligand for CMKLR1, is up-regulated in the CNS of mice with EAE

We also asked whether chemerin, a natural ligand for CMKLR1, is expressed in the CNS. Chemerin mRNA is up-regulated in the spinal cords of mice with EAE

(supplemental Fig. 3). Levels of CMKLR1 transcripts were also higher in spinal cords from mice with acute EAE, although the increase did not achieve statistical significance

(data not shown). Collectively, the data implicate the CNS as a source of active chemerin

for CMKLR1-expressing cells.

100

Discussion

In this report, we show that mice deficient in CMKLR1 develop less severe

clinical EAE than their WT counterparts. Onset of symptoms is not delayed, but clinical

score is significantly reduced in CMKLR1 KO mice, correlating with a significant

reduction in histologically assessed CNS inflammation. CMKLR1 is expressed on a

subset of CNS-infiltrating macrophages in the inflamed CNS, as well as on microglia and

on mDC in mice with EAE. Consistent with reduced inflammation, there was a trend

toward reduced demyelination in CMKLR1 KO mice with chronic EAE (day 46 p.i.).

Together, our findings demonstrate an important contribution of CMKLR1 to EAE

pathogenesis, although the underlying role (s) of receptor expression by microglia, and by

CNS-infiltrating macrophages and mDC during EAE, remain to be clarified.

CMKLR1 directs chemotaxis of macrophages and DC subsets in response to its proteolytically activated ligand, chemerin (7, 8, 9). Thus it is attractive to postulate a direct role for the receptor in recruitment of macrophages and CMKLR1+ DC subsets into

the inflamed CNS, as well as participation of CMKLR1 in cell positioning and cell-to-cell

interactions in inflammatory lesions. Indeed, compared with WT mice, CMKLR1 KO

mice with chronic EAE had fewer F4/80+ macrophages in the CNS, implying reduced

F4/80+ cell recruitment or survival. Because macrophages can amplify the inflammatory

cascade, reduced macrophage numbers could be a primary determinant of the overall

reduced CNS inflammation seen in CMKLR1 KO mice in the current studies. However,

because the relative frequency of different immune cell subsets was generally similar in

101

WT vs KO mice, one cannot exclude the possibility that the reduced macrophage

numbers are secondary to other CMKLR1-dependent phenomena.

Data presented in this report, as well as experiments using radiation bone marrow chimeric mice (K.L.G. and E.C.B., unpublished observations), show that CMKLR1+ mDC and macrophages are recruited to the CNS during EAE. CMKLR1 has been

implicated in DC recruitment from the blood into lymphoid tissues and inflamed skin in

humans (10, 17, 18), and it is likely to contribute to DC recruitment and/or tissue

interactions in EAE as well. In this context, it is interesting that we observed high levels

of expression of CMKLR1 by mDC in acute EAE in the present studies, but no

expression by pDC. This is the first clear demonstration of CMKLR1 expression by

endogenous mouse DC in our hands as, in contrast to clear expression by pDC in humans,

neither we nor others have detected robust CMKLR1 expression by significant

populations of peripheral lymphoid or extralymphoid tissue DC in adult mice (8). Several

groups have probed the contributions of DC to the pathogenesis of EAE. Greter and

colleagues demonstrated that CD11c+ DC present Ag to autoaggressive T cells in vivo,

which facilitates the development of EAE (13). Bailey and colleagues reported that CNS

mDC colocalize with myelin Ag-specific CD4+ T cells within perivascular spaces of the

inflamed CNS. Moreover, mDC appear to be highly efficient at presenting endogenous

myelin Ags and driving Th17 differentiation within the CNS during relapsing EAE (19).

The same group reported that pDC negatively regulate a relapsing-remitting model of

EAE through direct suppression of mDC activity (20). Furthermore, pDC are major

102

sources of type I IFNs (21), and recent studies support a protective role for type I IFNs in

EAE (22, 23). Collectively, the data suggest that the recruitment of mDC, pDC, and their

microenvironmental localization within the CNS may contribute significantly to the pathogenesis of EAE.

We were surprised to find that MOG35–55 reactive CMKLR1 KO lymphocytes are

not fully capable of transferring EAE to WT recipients (Fig. 4 and Table III). Importantly,

CMKLR1 deficiency had no significant defects on lymphocyte cytokine or proliferative

responses assessed in vitro (Fig. 3 and supplemental Fig. 1). This suggests that

lymphocytes generated in a CMKLR1 KO environment may have other functional properties that are not addressed by these in vitro analyses. We hypothesize that

CMKLR1 expression by CNS resident cells is a key determinant of EAE

pathophysiology, because there is no difference between WT and CMKLR1 KO mice

with respect to peripheral immune cell activation. In light of this possibility, we were intrigued to find that microglia express CMKLR1 protein on the cell surface, extending

early reports of CMKLR1 transcript expression by a mouse microglial cell line (24). The

significance of CMKLR1 expression by microglial cells is currently unclear, but

microglial cell activation is critical to EAE pathogenesis (1, 6, 25). CMKLR1 could

influence microglial cell localization or functional activity during CNS inflammation.

Chemerin, a natural ligand for CMKLR1, also binds GPR1 (26), previously an orphan G- protein-coupled receptor that is phylogenetically related to CMKLR1 (27). The relevance

of chemerin:GPR1 interactions in vivo remains to be determined, although the expression

103

of GPR1 mRNA by human fetal and simian adult astrocytes (28) points to possible CNS- related functions. We have also recently identified chemerin as a ligand for the orphan serpentine receptor CCRL2. Chemerin:CCRL2 binding does not trigger cell migration or intracellular calcium mobilization. Rather, CCRL2 appears to concentrate and present bioactive chemerin to CMKLR1-expressing cells (29). We are currently evaluating the role of CCRL2 in EAE.

The regulation of chemerin activity in vivo is incompletely understood. We and others have shown that the chemoattractant ability of chemerin is activated via cleavage of inhibitory C-terminal amino acids by proteases that have roles in the coagulation, fibrinolytic, and inflammatory cascades (7, 9, 11, 30), pathways that are known or likely to be engaged during CNS inflammation in EAE and MS. Proteases involved in these cascades may promote EAE pathogenesis in part via cleavage of chemerin, with subsequent recruitment of or functional impact on CMKLR1-expressing cells. For example, tissue plasminogen activator activates chemerin and is expressed at high levels by astrocytes within the CNS of mice with EAE (31). Transcripts for mast cell tryptase, for which chemerin is also a substrate, are increased in MS lesions (32). Chemerin can be cleaved and activated by neutrophil-derived serine proteases (11, 30). Thus, the activation of chemerin in vivo may be mediated by proteases derived from both recruited and CNS- resident cells in an inflammatory setting. It is possible that chemerin production or activation is up-regulated in a localized manner (e.g., within unique niches of the CNS parenchyma) during an inflammatory response, thus helping recruit CMKLR1+ cells to

104

specialized compartments. Indeed, Lande et al. recently reported chemerin colocalization

with intralesional endothelial cells in the brains of MS patients (33).

In a recent report, Cash and colleagues reported that chemerin-derived C-terminal peptides possess anti-inflammatory properties in a murine model of zymosan-induced

peritonitis (34). These properties were dependent on CMKLR1, as indicated by analysis

of CMKLR1 KO mice. It will be of interest to determine whether similar C-terminal

chemerin peptides are beneficial in other models of inflammatory disease. Our data

suggest that chemerin promotes inflammatory responses during EAE, raising the

possibility that chemerin has disparate roles in peritoneal vs CNS inflammation or in

different inflammatory settings. It is likely that chemerin, like other cytokines and

chemokines, interacts functionally with other mediators in a complex manner to regulate

inflammatory responses in vivo. It is also possible that CMKLR1 functions independently

of chemerin binding in the EAE model.

As mentioned previously, human blood pDC express CMKLR1 (9, 10) and recent

studies show that a subset of pDC in the brains of human MS patients expresses the

receptor as well (33). However, mouse pDC in lymphoid tissues and blood (8), and in normal and inflamed CNS as shown here, do not express detectable CMKLR1,

highlighting the challenges associated with extrapolating observations made in mouse

models to human systems. Although mouse and human pDC share key functions, they

differ significantly in their expression of a number of surface Ags, in addition to

CMKLR1 (21) . CMKLR1 is only weakly expressed, if at all, by circulating human mDC

105

(9, 10) and has not been reported on endogenous mouse mDC, but is transiently induced

to high levels on both mouse bone marrow-derived and human monocyte-derived mDC

generated in response to GM-CSF or Flt-3 ligand in vitro (8). The present finding of

expression by mouse mDC in EAE supports the potential for CMKLR1 to participate in

mDC recruitment and functions in select inflammatory settings. In contrast to its

distinctive patterns of expression by human vs mouse DC subsets, CMKLR1 in both

species is highly expressed by tissue macrophages, including macrophages in the CNS of

mice with EAE.

In summary, our results demonstrate an important contribution of CMKLR1 in the

pathogenesis of EAE, with potential involvement in the regulation of DC dynamics and

macrophage accumulation in CNS inflammation. They point to CMKLR1:chemerin interactions as a potential target for therapy of chronic and/or progressive MS.

106

Acknowledgements

We are grateful to Jane Eaton, Bari Nazario, and Evelyn Resurreccion for skilled technical assistance. We thank Liana Gefter, Rocky Bilhartz, Anthony Slavin, and

Lawrence Steinman for helpful discussions. This work is supported by National Institutes of Health Grants AI-59635, AI-47822, and GM-37734; Specialized Center of Research

Grant HL-67674; and a Merit Award from the Veterans Administration to E.C.B. B.A.Z. is supported by National Institutes of Health Grant AI-079320 and SPARK awards from

Stanford University. K.L.G. was supported by a United Negro College Fund Merck

Postdoctoral Award and a postdoctoral fellowship from the National Multiple Sclerosis

Society.

107

Supplemental Figures

Figure S1. Proliferation and cytokine production by CMKLR1 KO T cells. CMKLR1 KO or WT mice were immunized with MOG35-55 emulsified in CFA. Ten days later, CD3+ T cells were isolated from spleen or draining LN, followed by co- incubation with MMCtreated WT or CMKLR1 KO splenocytes. [All possible T cell:APC combinations were utilized: i)WT:WT; ii) WT:KO; iii) KO:WT; iv) KO:KO.] MOG35- 55 was added at 25 μg/ml. (A) After 72 h of stimulation, proliferation of LN (left panel) or splenic T cells (right panel) was assessed by [3H]-thymidine incorporation assay. Data are presented as a stimulation index; bars represent SEM. (B) Culture supernatants were collected after 72 h, and levels of the indicated cytokines produced by LN (left panels) or splenic T cells (right panels) were measured by ELISA. Asterisks in panel (A) indicate statistically significant differences (P < 0.05) as follows: * WT:WT vs. KO:KO; ** KO:WT vs. KO:KO; and *** WT:WT vs. KO:WT, as determined by analysis of variance (ANOVA), followed by Bonferroni multiple comparisons post test. Data shown are representative of two independent experiments with similar results.

108

Figure S2. Peritoneal macrophage responses in CMKLR1 KO mice. (A) Peritoneal lavage cells (PLCs) were harvested from naïve WT or CMKLR1 KO mice. After allowing PLCs to adhere for 2 h at 37 C, non-adherent cells were washed away and adherent cells were stimulated with LPS/IFNγ for 24 h. Levels of TNF (top panel) and IL-6 (bottom panel) in culture supernatants were measured by ELISA. (B) WT and CMKLR1 KO were given a single i.p. injection of thioglycollate. Mice were killed 72 h later and PLCs were harvested and processed as described in (A). Levels of TNF and IL-6 were measured by ELISA. * P < 0.05, as determined by Student’s t test. Data shown are representative of two independent experiments with similar results.

109

Figure S3. Chemerin transcripts are up-regulated in the CNS of mice with EAE. Spinal cord RNA was extracted from naïve mice (n=9), or from mice induced to develop EAE at day 13 p.i. (n=8). Relative chemerin RNA expression (normalized to cycA) was assessed by real-time QPCR as described in Supplemental Materials and Methods. Values are presented as mean ± s.e.m. * P < 0.05, as determined by Student’s t test.

Figure S4. Staining of mouse leukocytes with anti-CMKLR1 mAbs. Spleen cells or PLCs were harvested from naïve C57BL/6 mice, and stained with rat anti- mouse CMKLR1 mAb BZ194 (A) or mouse anti-mouse CMKLR1 mAb BZ186 (B) and their respective isotype controls. PE-conjugated goat anti-rat IgG (BD Pharmingen) was used to detect BZ194; PE-labeled rat anti-mouse IgG1 was used to detect BZ186. Cells were then stained with fluorophore-labeled monoclonal antibodies to identify the indicated leukocyte subsets and data were acquired on a flow cytometer. CMKLR1 expression was analyzed on NK cell, T cells, B cells and DC isolated from the spleen; macrophages were taken from the peritoneal cavity. Filled histograms represent BZ332 (top panels) or DREG200 (bottom panels isotype control staining. Open histograms indicate staining with BZ194 (top panels) or BZ186 (bottom panels). 110

References

1. Steinman, L. 2001. Multiple sclerosis: a two-stage disease. Nat Immunol 2:762- 764. 2. Baron, J. L., J. A. Madri, N. H. Ruddle, G. Hashim, and C. A. Janeway, Jr. 1993. Surface expression of alpha 4 integrin by CD4 T cells is required for their entry into brain parenchyma. J Exp Med 177:57-68. 3. Yednock, T. A., C. Cannon, L. C. Fritz, F. Sanchez-Madrid, L. Steinman, and N. Karin. 1992. Prevention of experimental autoimmune encephalomyelitis by antibodies against alpha 4 beta 1 integrin. Nature 356:63-66. 4. Rudick, R. A., W. H. Stuart, P. A. Calabresi, C. Confavreux, S. L. Galetta, E. W. Radue, F. D. Lublin, B. Weinstock-Guttman, D. R. Wynn, F. Lynn, M. A. Panzara, and A. W. Sandrock. 2006. Natalizumab plus interferon beta-1a for relapsing multiple sclerosis. N Engl J Med 354:911-923. 5. Karpus, W. J., and R. M. Ransohoff. 1998. Chemokine regulation of experimental autoimmune encephalomyelitis: temporal and spatial expression patterns govern disease pathogenesis. J Immunol 161:2667-2671. 6. Sospedra, M., and R. Martin. 2005. Immunology of multiple sclerosis. Annu Rev Immunol 23:683-747. 7. Wittamer, V., J. D. Franssen, M. Vulcano, J. F. Mirjolet, E. Le Poul, I. Migeotte, S. Brezillon, R. Tyldesley, C. Blanpain, M. Detheux, A. Mantovani, S. Sozzani, G. Vassart, M. Parmentier, and D. Communi. 2003. Specific recruitment of antigen- presenting cells by chemerin, a novel processed ligand from human inflammatory fluids. J Exp Med 198:977-985. 8. Zabel, B. A., T. Ohyama, L. Zuniga, J. Y. Kim, B. Johnston, S. J. Allen, D. G. Guido, T. M. Handel, and E. C. Butcher. 2006. Chemokine-like receptor 1 expression by macrophages in vivo: Regulation by TGF-beta and TLR ligands. Exp Hematol 34:1106- 1114. 9. Zabel, B. A., A. M. Silverio, and E. C. Butcher. 2005. Chemokine-like receptor 1 expression and chemerin-directed chemotaxis distinguish plasmacytoid from myeloid dendritic cells in human blood. J Immunol 174:244-251. 10. Vermi, W., E. Riboldi, V. Wittamer, F. Gentili, W. Luini, S. Marrelli, A. Vecchi, J. D. Franssen, D. Communi, L. Massardi, M. Sironi, A. Mantovani, M. Parmentier, F. Facchetti, and S. Sozzani. 2005. Role of ChemR23 in directing the migration of myeloid and plasmacytoid dendritic cells to lymphoid organs and inflamed skin. J Exp Med 201:509-515. 11. Zabel, B. A., S. J. Allen, P. Kulig, J. A. Allen, J. Cichy, T. M. Handel, and E. C. Butcher. 2005. Chemerin activation by serine proteases of the coagulation, fibrinolytic, and inflammatory cascades. J Biol Chem 280:34661-34666. 12. Serafini, B., B. Rosicarelli, R. Magliozzi, E. Stigliano, E. Capello, G. L. Mancardi, and F. Aloisi. 2006. Dendritic cells in multiple sclerosis lesions: maturation

111

stage, myelin uptake, and interaction with proliferating T cells. J Neuropathol Exp Neurol 65:124-141. 13. Greter, M., F. L. Heppner, M. P. Lemos, B. M. Odermatt, N. Goebels, T. Laufer, R. J. Noelle, and B. Becher. 2005. Dendritic cells permit immune invasion of the CNS in an animal model of multiple sclerosis. Nat Med 11:328-334. 14. Han, M. H., S. I. Hwang, D. B. Roy, D. H. Lundgren, J. V. Price, S. S. Ousman, G. H. Fernald, B. Gerlitz, W. H. Robinson, S. E. Baranzini, B. W. Grinnell, C. S. Raine, R. A. Sobel, D. K. Han, and L. Steinman. 2008. Proteomic analysis of active multiple sclerosis lesions reveals therapeutic targets. Nature 451:1076-1081. 15. Parolini, S., A. Santoro, E. Marcenaro, W. Luini, L. Massardi, F. Facchetti, D. Communi, M. Parmentier, A. Majorana, M. Sironi, G. Tabellini, A. Moretta, and S. Sozzani. 2007. The role of chemerin in the colocalization of NK and dendritic cell subsets into inflamed tissues. Blood 109:3625-3632. 16. Ford, A. L., A. L. Goodsall, W. F. Hickey, and J. D. Sedgwick. 1995. Normal adult ramified microglia separated from other central nervous system macrophages by flow cytometric sorting. Phenotypic differences defined and direct ex vivo antigen presentation to myelin basic protein-reactive CD4+ T cells compared. J Immunol 154:4309-4321. 17. Skrzeczynska-Moncznik, J., K. Wawro, A. Stefanska, E. Oleszycka, P. Kulig, B. A. Zabel, M. Sulkowski, M. Kapinska-Mrowiecka, M. Czubak-Macugowska, E. C. Butcher, and J. Cichy. 2009. Potential role of chemerin in recruitment of plasmacytoid dendritic cells to diseased skin. Biochem Biophys Res Commun 380:323-327. 18. Albanesi, C., C. Scarponi, S. Pallotta, R. Daniele, D. Bosisio, S. Madonna, P. Fortugno, S. Gonzalvo-Feo, J. D. Franssen, M. Parmentier, O. De Pita, G. Girolomoni, and S. Sozzani. 2009. Chemerin expression marks early psoriatic skin lesions and correlates with plasmacytoid dendritic cell recruitment. J Exp Med 206:249-258. 19. Bailey, S. L., B. Schreiner, E. J. McMahon, and S. D. Miller. 2007. CNS myeloid DCs presenting endogenous myelin peptides 'preferentially' polarize CD4+ T(H)-17 cells in relapsing EAE. Nat Immunol 8:172-180. 20. Bailey-Bucktrout, S. L., S. C. Caulkins, G. Goings, J. A. Fischer, A. Dzionek, and S. D. Miller. 2008. Cutting edge: central nervous system plasmacytoid dendritic cells regulate the severity of relapsing experimental autoimmune encephalomyelitis. J Immunol 180:6457-6461. 21. Colonna, M., G. Trinchieri, and Y. J. Liu. 2004. Plasmacytoid dendritic cells in immunity. Nat Immunol 5:1219-1226. 22. Prinz, M., H. Schmidt, A. Mildner, K. P. Knobeloch, U. K. Hanisch, J. Raasch, D. Merkler, C. Detje, I. Gutcher, J. Mages, R. Lang, R. Martin, R. Gold, B. Becher, W. Bruck, and U. Kalinke. 2008. Distinct and nonredundant in vivo functions of IFNAR on myeloid cells limit autoimmunity in the central nervous system. Immunity 28:675-686. 23. Shinohara, M. L., J. H. Kim, V. A. Garcia, and H. Cantor. 2008. Engagement of the type I interferon receptor on dendritic cells inhibits T helper 17 cell development: role of intracellular osteopontin. Immunity 29:68-78. 112

24. Martensson, U. E., J. Bristulf, C. Owman, and B. Olde. 2005. The mouse chemerin receptor gene, mcmklr1, utilizes alternative promoters for transcription and is regulated by all-trans retinoic acid. Gene 350:65-77. 25. Heppner, F. L., M. Greter, D. Marino, J. Falsig, G. Raivich, N. Hovelmeyer, A. Waisman, T. Rulicke, M. Prinz, J. Priller, B. Becher, and A. Aguzzi. 2005. Experimental autoimmune encephalomyelitis repressed by microglial paralysis. Nat Med 11:146-152. 26. Barnea, G., W. Strapps, G. Herrada, Y. Berman, J. Ong, B. Kloss, R. Axel, and K. J. Lee. 2008. The genetic design of signaling cascades to record receptor activation. Proc Natl Acad Sci U S A 105:64-69. 27. Joost, P., and A. Methner. 2002. Phylogenetic analysis of 277 human G-protein- coupled receptors as a tool for the prediction of orphan receptor ligands. Genome Biol 3:RESEARCH0063. 28. Croitoru-Lamoury, J., G. J. Guillemin, F. D. Boussin, B. Mognetti, L. I. Gigout, A. Cheret, B. Vaslin, R. Le Grand, B. J. Brew, and D. Dormont. 2003. Expression of chemokines and their receptors in human and simian astrocytes: evidence for a central role of TNF alpha and IFN gamma in CXCR4 and CCR5 modulation. Glia 41:354-370. 29. Zabel, B. A., S. Nakae, L. Zuniga, J. Y. Kim, T. Ohyama, C. Alt, J. Pan, H. Suto, D. Soler, S. J. Allen, T. M. Handel, C. H. Song, S. J. Galli, and E. C. Butcher. 2008. Mast cell-expressed orphan receptor CCRL2 binds chemerin and is required for optimal induction of IgE-mediated passive cutaneous anaphylaxis. J Exp Med 205:2207-2220. 30. Wittamer, V., B. Bondue, A. Guillabert, G. Vassart, M. Parmentier, and D. Communi. 2005. Neutrophil-mediated maturation of chemerin: a link between innate and adaptive immunity. J Immunol 175:487-493. 31. Teesalu, T., A. E. Hinkkanen, and A. Vaheri. 2001. Coordinated induction of extracellular proteolysis systems during experimental autoimmune encephalomyelitis in mice. Am J Pathol 159:2227-2237. 32. Lock, C., G. Hermans, R. Pedotti, A. Brendolan, E. Schadt, H. Garren, A. Langer- Gould, S. Strober, B. Cannella, J. Allard, P. Klonowski, A. Austin, N. Lad, N. Kaminski, S. J. Galli, J. R. Oksenberg, C. S. Raine, R. Heller, and L. Steinman. 2002. Gene- microarray analysis of multiple sclerosis lesions yields new targets validated in autoimmune encephalomyelitis. Nat Med 8:500-508. 33. Lande, R., V. Gafa, B. Serafini, E. Giacomini, A. Visconti, M. E. Remoli, M. Severa, M. Parmentier, G. Ristori, M. Salvetti, F. Aloisi, and E. M. Coccia. 2008. Plasmacytoid dendritic cells in multiple sclerosis: intracerebral recruitment and impaired maturation in response to interferon-beta. J Neuropathol Exp Neurol 67:388-401. 34. Cash, J. L., R. Hart, A. Russ, J. P. Dixon, W. H. Colledge, J. Doran, A. G. Hendrick, M. B. Carlton, and D. R. Greaves. 2008. Synthetic chemerin-derived peptides suppress inflammation through ChemR23. J Exp Med 205:767-775.

113

CHAPTER 4 : MAST CELL-EXPRESSED ORPHAN RECEPTOR CCRL2 BINDS

CHEMERIN AND IS REQUIRED FOR OPTIMAL INDUCTION OF IGE-

MEDIATED PASSIVE CUTANEOUS ANAPHYLAXIS

Brian A. Zabel1, Susumu Nakae2, Luis Zuniga1, Ji-Yun Kim1, Takao Ohyama1, Carsten

Alt1, Junliang Pan1, Hajime Suto2, Dulce Soler3, Samantha J. Allen4, Tracy M. Handel4,

Chang Ho Song2,5, Stephen J. Galli2,6, Eugene C. Butcher1

1Laboratory of Immunology and Vascular Biology, Department of Pathology, Stanford

University School of Medicine, Stanford, CA 94305, and Center for Molecular Biology and Medicine, Veterans Affairs Palo Alto Health Care System, Palo Alto, CA 94304.

2Department of Pathology, Stanford University School of Medicine, Stanford, CA 94305

3Millennium Pharmaceuticals, Inflammation Department, Cambridge, MA.

4Skaggs School of Pharmacy and Pharmaceutical Sciences, University of California, San

Diego, La Jolla, CA 92093.

5Department of Anatomy, Chonbuk National University Medical School, Jeonju,

Republic of Korea.

6Department of Microbiology & Immunology, Stanford University School of Medicine,

Stanford, CA 94305.

B. Zabel and S. Nakae contributed equally to this work.

114

Reproduced from:

J Exp Med. 2008 September 29; 205 (10): 2207–2220.

Used by copyright permission of The Rockefeller University Press.

115

Summary

Mast cells contribute importantly to both protective and pathologic IgE-dependent

immune responses. Here we show that the mast cell-expressed orphan serpentine receptor mCCRL2 is not required for expression of IgE-mediated, mast cell-dependent passive cutaneous anaphylaxis, but can enhance the tissue swelling and leukocyte infiltrates associated with such reactions in mice. We further identify chemerin as a natural, non- signaling protein ligand for both human and mouse CCRL2. In contrast to other “silent” or professional chemokine interceptors, chemerin binding does not trigger ligand internalization. Rather, CCRL2 is able to bind the chemoattractant and increase local concentrations of bioactive chemerin, thus providing a link between CCRL2 expression and inflammation via the cell-signaling chemerin receptor CMKLR1.

116

Introduction

Leukocyte-expressed orphan heptahelical receptors that share significant homology with known chemoattractant receptors, yet remain uncharacterized with respect to ligand-binding properties and functions, represent excellent candidates for additional regulators of immune cell trafficking and function. Given its phylogenetic homology with members of the CC chemokine receptor subfamily, orphan serpentine receptor chemokine (CC motif) receptor-like 2 (CCRL2, also known as L-CCR [LPS- inducible C-C chemokine receptor related gene] or Eo1 in mice, and HCR [human chemokine receptor], CRAM-A, CRAM-B, or CKRX [chemokine receptor X] in humans) has been identified as a potential leukocyte chemoattractant receptor. However,

CCRL2 possesses an uncharacteristic intracellular loop 2 sequence in place of the

DRYLAIV motif generally found in signaling chemokine receptors (QRYLVFL in huCCRL2; QRYRVSF in mCCRL2), leading us to postulate that it might be an ‘atypical’ silent or non-signaling receptor. From a phylogenetic standpoint, CCRL2 may be unique, as its orthologs are more divergent in sequence that any other mouse-to-man receptor pair in the chemoattractant GPCR subfamily: the sequence identity of mouse and human

CCRL2 is only 51%, compared with ~80% identity between most other receptor orthologs (1, 2, 3)).

mCCRL2 was initially shown to be upregulated at the RNA level in peritoneal macrophages treated with LPS (3). In EAE, a murine model of multiple sclerosis,

CCRL2 RNA is expressed in the spinal column early during the onset of disease by

117

astrocytes, microglia, and infiltrating macrophages (4). Astrocytes and microglia also upregulate mCCRL2 in response to LPS (5). In a model of ovalbumin-induced airway inflammation, infiltrating lung macrophages express CCRL2 RNA after ovalbumin challenge, while the bronchial epithelium is constitutively positive for expression (6). By mAb staining, huCCRL2 is expressed by circulating human T cells, neutrophils, monocytes, CD34+ bone marrow precursors, and monocyte derived macrophages and dendritic cells (DC), and is generally upregulated upon activation of such cells (7).

HuCCRL2 is also expressed on synovial fluid neutrophils (from rheumatoid arthritis patients), and is upregulated on freshly isolated blood neutrophils treated with LPS or

TNFα (8). While there is a report indicating that CCR2 ligands such as CCL2 act as functional ligands for CCRL2 (9), this finding remains controversial (8, 10).

A number of ‘atypical’ serpentine GPCRs that are homologous to chemoattractant receptors bind to chemoattractants but fail to transduce intracellular signals through heterotrimeric G proteins and or support cell migration. This functionally defined receptor subfamily is currently thought to be comprised of three members—D6, DARC

(Duffy antigen receptor for chemokines), and CCX-CKR (ChemoCentryx chemokine receptor) (10, 11). These receptors are also referred to as professional chemokine

“interceptors”, a name that reflects their ability to efficiently internalize bound ligand

(12). These receptors also lack the consensus “DRYLAIV”-related sequence present in the second intracellular loop domain of most chemokine receptors, possibly accounting for their inability to transduce classical intracellular signals (the sequence is DKYLEIV

118

in D6; LGHRLGA in DARC; and DRYWAIT in CCX-CKR). Identifying ligands for

‘silent’ or non-signaling orphan receptors has proven to be particularly challenging, since the assays employed in most GPCR ligand screens generally depend on functional responses, such as intracellular calcium mobilization, cell migration, or transcriptional activation.

D6 is expressed by lymphatic endothelium and by some leukocytes, and binds pro-inflammatory CC-chemokines. It is constitutively internalized (independent of chemokine binding), and can rapidly degrade large quantities of chemokines (13).

DARC is expressed by vascular endothelium and is present on red blood cells. It binds mostly pro-inflammatory chemokines (both CC and CXC types), and its expression on red blood cells can buffer levels of circulating chemokines (14, 15). DARC-bound chemokines on erythrocytes are not presented or transferred to other chemokine receptors

(16). Endothelial cell-expressed DARC may be able to transcytose certain chemokines, although this finding is controversial (17, 18). CCX-CKR mRNA is expressed in many tissues, including lymph node, spleen, placenta, kidney, brain, and by leukocytes. It binds the homeostatic chemokines CCL19, CCL21, and CXCL13, and has been shown to be able to degrade large quantities of CCL19 (19). Genetic deficiencies in D6 or DARC have been associated with exacerbated or dysregulated inflammatory responses, reflecting the importance of interceptors in dampening chemokine-mediated activities

(reviewed in (10)).

119

We and others identified chemerin as a natural non-chemokine chemoattractant ligand for chemokine like receptor-1 (CMKLR1) (20, 21, 22). CMKLR1 is expressed by circulating human plasmacytoid dendritic cells (pDC) but not by circulating or tissue mouse DC (20, 23). In contrast, tissue resident macrophages in both humans and mice express CMKLR1 (23). A recent report by Parolini et al. (24) describes CMKLR1 expression by human NK cells. We and others have demonstrated that activation of chemerin involves proteolytic processing of the carboxyl-terminus and removal of inhibitory amino acids ( (22, 25, 26, 27), reviewed in (28)). Various serine proteases of the coagulation, fibrinolytic, and inflammatory cascades can activate chemerin, including plasmin, factor XII, and neutrophil elastase and cathepsin G (27). Mast cell tryptase, released upon antigen-mediated crosslinking of surface IgE/FcεRI on mast cells, is a particularly potent activator of chemerin (27). We have also demonstrated that chemerin circulates at low nanomolar concentrations in plasma in an immature, inactive pro-form

(28). We hypothesize that chemerin may serve as a nearly ubiquitous ‘humor’ poised to translate tissue damage or bleeding into rapidly generated attractant fields for specialized

CMKLR1-positive cells.

Mast cells, tissue-dwelling derivatives of circulating mast cell progenitors, are anatomically deployed to host-environment boundaries (such as the skin, airways, and alimentary canal) where they can mediate first-line encounters with pathogens and environmental allergens (29). Antigen-mediated crosslinking of surface IgE/FcεRI triggers mast cell degranulation and the release of preformed pro-inflammatory mediators

120

stored in the cells’ cytoplasmic granules, such as histamine, heparin, and proteases (e.g.

tryptases, chymases, and cathepsins), and initiates the synthesis and release of lipid

mediators (e.g. prostaglandin D2 and leukotrienes) and diverse cytokines, chemokines and growth factors (30). Localized IgE-mediated mast cell “anaphylactic-type” degranulation can be modeled in vivo by experimental IgE-dependent passive cutaneous

anaphylaxis (PCA) in mice.

In this report, we generated novel monoclonal antibodies (mAb) specific for the

orphan G-protein coupled receptor (GPCR) mCCRL2, and demonstrated that freshly

isolated mouse peritoneal mast cells selectively express the receptor. mCCRL2 KO mice

displayed no overt phenotype and had normal numbers of mast cells in the tissues

analyzed. When tested in vitro, bone marrow-derived cultured mast cells (BMCMCs) of mCCRL2 KO mouse origin exhibited responses to IgE and specific antigen-mediated crosslinking of FcεRI that were statistically indistinguishable from those of wild type

BMCMCs. In IgE-dependent PCA reactions in vivo, a model of IgE-mediated local anaphylaxis, mast cell-expressed CCRL2 was not required for the development of cutaneous inflammatory responses in mice sensitized with a high dose of antigen-specific

IgE. However, mast cell-expressed CCRL2 was required for the development of optimal cutaneous tissue swelling and leukocyte infiltrates in mice sensitized with a low dose of antigen-specific IgE.

To investigate the mechanism by which CCRL2 may contribute to inflammation, we sought to identify potential CCRL2 ligands. In experiments based on binding rather

121

than functional output, we identified chemerin as a novel, non-signaling protein ligand

for both human and mouse receptors. As opposed to chemokine interceptors that serve as a sink for chemokines upon their ligand-induced internalization, chemerin is not rapidly internalized by CCRL2; rather, CCRL2 concentrates the chemoattractant and increases local chemerin concentrations available to interact with CMKLR1. We propose that

CCRL2 is a novel attractant receptor, serving to focus chemerin localization in vivo and contribute to CMKLR1-mediated processes that in turn regulate pathways leading to increased vascular permeability, tissue swelling, and leukocyte recruitment.

122

Materials and Methods

Animals.

CCRL2 KO mice were obtained from Lexicon (The Woodlands, TX, USA) and

backcrossed 4 generations on the C57BL/6 background. Mast cell deficient KitW-sh/Wsh

mice on the C57BL/6 background (31) were kindly provided by Peter Besmer (Memorial

Sloan-Kettering Cancer Center and Cornell University Graduate School of Medical

Sciences, NY) and WT C57BL/6 mice were obtained from Taconic (Oxnard, CA, USA).

Wistar Furth rats were obtained from Charles River Laboratories (Wilmington, MA,

USA).

Mammalian expression vector construction and generation of stable cell lines.

The coding regions of mCCRL2, huCCRL2, mCRTH2, and huCCR10 were amplified

from genomic DNA with or without an engineered N-terminal hemagglutinin (HA) tag, and cloned into pcDNA3 (Invitrogen, Carlsbad, CA, USA). Transfectants were generated and stable lines selected in the mouse pre-B lymphoma cell line L1.2 or HEK293 cells as described (32). mCMKLR1 and empty vector L1.2 transfectants were generated as previously described (20). Transfected cells were in some cases treated with 5 mM n- butyric acid (Sigma) for 24 h before experimentation (33).

Chemerin expression and purification using baculovirus.

The following carboxyl-terminal His8-tagged proteins were expressed using baculovirus- infected insect cells as previously described (27): “serum form” human chemerin (NH2-

ADPELTE…FAPHHHHHHHH-COOH), “pro-form” human chemerin (NH2-

123

ADPELTE… LPRSPHHHHHH-COOH), and “serum form” mouse chemerin (NH2-

ADPTEPE…FAPHHHHHHHH-COOH). Since certain experiments required non-tagged proteins, the His8-tag was proteolytically removed by treatment with carboxypeptidase A

(Sigma), generating the respective proteins NH2-ADPELTE…FAPH-COOH, NH2-

ADPELTE…RSPH-COOH, and NH2-ADPTEPE…FAPH-COOH, where the underlined residues are non-native. The proteins were lyophilized and checked for purity using

electrospray mass spectrometry.

Generating the anti-mCCRL2 mAbs BZ5B8 and BZ2E3.

The immunizing amino-terminal mCCRL2 peptide with the sequence NH2-

MDNYTVAPDDEYDVLILDDYLDNSC-COOH (corresponding to residues 1-24 of

mCCRL2, with a non-native carboxyl-terminal cysteine to facilitate conjugation to

keyhole limpet hemocyanin, (KLH)) was synthesized by the Stanford Protein and Nucleic

Acid Biotechnology Facility and conjugated to KLH according to the manufacturer’s

specifications (Pierce Biotechnology, Rockford, IL, USA). Wistar Furth rats were

immunized with the mCCRL2 peptide/KLH conjugate first emulsified in CFA, and then

subsequently in IFA. Hybridomas producing anti-mCCRL2 mAbs were subcloned, and

specificity was confirmed by reactivity with mCCRL2 but not other L1.2 receptor

transfectants. An ELISA-based assay (BD Pharmingen) was used to assess the IgG2aκ

isotypes of the resulting rat anti-mouse CCRL2 mAbs, designated BZ5B8 and BZ2E3.

Preparation of bone marrow-derived cultured mast cells (BMCMCs).

124

Mouse femoral BM cells were cultured in 20% WEHI-3 cell conditioned medium

(containing IL-3) for 6-12 weeks, at which time the cells were >98% c-kithigh FcεRIαhigh

by flow cytometry analysis (data not shown).

Passive cutaneous anaphylaxis (PCA) reaction.

Experimental PCA was performed as previously described (34), with minor

modifications. Mice were injected intradermally with 20 μl of either anti-DNP IgE mAb

(H1-ε-26; 5, 50 or 150 ng) in the left ear skin, or vehicle alone (PIPES-HMEM buffer) in the right ear skin. The next day, mice received 200 μl of 1 mg/ml DNP-HSA (200 μg per mouse) intravenously. Ear thickness was measured before and at multiple intervals after

DNP-HSA injection with an engineer’s microcaliper (Ozaki MFG. CO., LTD., Itabashi,

Tokyo, Japan). For BMCMC engraftment experiments, BMCMCs were generated from

either WT or LCCR KO mice, and 1 x 106 cells in 40 μl DMEM were injected into the ear skin of mast cell-deficient KitW-sh/Wsh mice. 6-8 weeks later the mice were subjected to

experimental PCA. After the assay, the mast cells in the ear skin were enumerated in

formalin-fixed, paraffin-embedded, toluidine blue-stained sections to evaluate the extent

of engraftment; in some experiments, formalin-fixed, paraffin-embedded, hematoxylin

and eosin-stained sections were examined to enumerate numbers of leukocytes present in

the dermis. In all histological studies, examination of the slides was performed by an

observer who was not aware of the identity of individual sections.

Chemerin binding assays.

125

For chemerin binding/anti-mCCRL2 mAbs displacement assays, total peritoneal exudate

cells were incubated with various concentrations of chemerin or CCL2 for 5 minutes on ice in binding buffer, washed with PBS, and stained with primary antibodies (either anti- mCMKLR1 BZ2E3 or IgE, + Fc block) for 45 min on ice. The cells were washed in PBS and stained with secondary anti-rat IgG PE or anti-mouse IgE PE (+ goat IgG) for 30 min on ice, washed with PBS, stained with directly conjugated F4/80 and c-kit mAbs, and analyzed by flow cytometry. For radioligand binding assays, radioiodinated chemerin

(residues 21-148, R&D Systems, custom radiolabeling performed by Perkin Elmer) was provided as a kind gift from Dr. Juan Jaen (Chemocentryx, Mountain View, CA). The specific activity of the 125I-labeled chemerin was 97 Ci/g. For competition binding assays,

L1.2 cells transfected with huCCRL2, mCCRL2, or mCMKLR1 were plated into 96-well

plates at 0.5 x106 cells/well. Cells were incubated in binding buffer (Hanks + 0.5% BSA)

for 3 hr at 4°C shaking with 1 nM 125I-chemerin and increasing concentrations of chemerin, His8-tagged chemerin, or peptide (9-aa YFPGQFAFS, corresponding to

chemerin residues 149-157) as competitors. For saturation binding assays, mCCRL2/L1.2

cells were plated at 0.5 x106 cells/well. Nonspecific binding was measured in the

presence of 100 nM cold chemerin. Binding was terminated by washing the cells in saline

buffer, and bound radioactivity was measured. Data were analyzed using Prism

(GraphPad Software). Binding data (triplicate or quadruplicate wells) were fitted to one-

site binding hyperbola for saturation assays, or to a one-site competition curve for

competition assays. For direct chemerin binding immunofluorescence assays, mCCRL2-

126

HA, huCCRL2-HA, mCMKLR1-HA, mCRTH2-HA L1.2 transfectants were incubated

for 30 min on ice with 10 nM His8-tagged serum form human chemerin and the indicated

concentration of untagged chemerin in binding buffer (PBS with 0.5% BSA, 0.02%

azide). The cells were then washed with PBS, and incubated with anti-His6 PE (+ 2%

goat serum) for 30 min on ice, washed and analyzed by flow cytometry. Similar binding experiments were performed on total WT or CCRL2 KO peritoneal exudate cells with the

indicated combinations and concentrations of pro-form or serum-form His8-tagged or

untagged chemerin. Following chemerin binding, the cells were stained with directly

conjugated F4/80 and c-kit mAbs and analyzed by flow cytometry. For competitive

heterologous displacement binding assays, mCCRL2-HA/L1.2 cells were co-incubated

with 10 nM His8-tagged serum form human chemerin and 1000 nM untagged

chemoattractants in binding buffer for 30 min on ice. The cells were then washed with

PBS, and incubated with anti- His6 PE (+ 2% goat serum) for 30 min on ice, and then analyzed by flow cytometry.

Statistics.

The unpaired Student's t-test (2-tailed), Mann Whitney U-test (2-tailed), or ANOVA was

used for statistical evaluation of the results, as indicated.

127

Supplemental Materials and Methods

Antibodies and reagents.

Anti-mouse -CD11b, -CD11c, -CD14, -CD19, -B220, -F4/80, -Gr1, -TCRβ -c-kit, -

CD49b, -TER119 dye-linked mAb were obtained from eBioscience (San Diego, CA,

USA), BD PharMingen (San Diego, CA, USA), and Serotec (Raleigh, NC, USA). Anti-

rat phycoerythrin (human and mouse adsorbed) was purchased from BD Pharmingen,

anti-HIS6 phycoerythrin was purchased from R&D Systems (Minneapolis, MN, USA),

purified Fc block (mouse anti-mouse CD16.2/32.2) was purchased from Caltag

(Burlingame, CA, USA), anti-mCMKLR1 (BZ194) was prepared in house, and mouse

IgG, rat IgG, and goat serum were purchased from Sigma (St. Louis, MO, USA).

CCL2-5,7-9,11,16,17,19-22,25,28; CXCL1,2,3,9,10,12,13,16; IL-4; GM-CSF; and Flt-3

ligand were purchased from R&D Systems. Mouse CCL2 biotinylated fluorokine kit was

purchased from R&D Systems. CMFDA, Fluo-4-acetoxymethyl (AM), and Pluronic acid

F-127 (reconstituted in DMSO) were purchased from Molecular Probes (Eugene, OR,

USA). Bioactive chemerin peptide (YFPGQFAFS) was synthesized by the Stanford PAN facility. Phosphothioated CpG oligonucleotides (35) were purchased from Qiagen

(Valencia, CA, USA). polyI:C and fMLP were purchased from Sigma. LPS (E.coli

O11:B4-derived) was purchased from List Biologicals (Campbell, CA, USA). TNFα and

IFNγ were purchased from Roche (Penzberg, Germany). Complete and incomplete

Freund’s adjuvant (CFA and IFA) were purchased from Sigma. Cytokine levels in culture

128

supernatants were measured by using mouse TNFα and IL-6 BD OptEIA™ ELISA Sets

(BD PharMingen).

RNA expression analysis.

A RNA dot blot array was purchased from BD Clontech and hybridizations were

performed according to the manufacturer’s recommendation. A full-length gel-purified

mCCRL2 cDNA probe was radiolabeled with 32P using RediPrime reagents (Amersham

Biosciences) according to manufacturer’s specifications. For RT-PCR, RNA from the indicated tissues or cells was extracted using a Qiagen RNeasy kit per the supplier’s instructions. Gene expression was determined by quantitative PCR (QPCR) using an

Applied Biosystems 7900HT real-time PCR instrument equipped with a 384-well reaction block. 0.3 – 1.0 μg total RNA was used as template for cDNA synthesis using

MMLV Reverse Transcriptase (Applied Biosystems) with oligo dT primers according to the supplier’s instructions. The cDNA was diluted and amplified by quantitative PCR in triplicate wells using 10 pmols of gene specific primers in a total volume of 10 μL with

Power SYBR Green QPCR Master Mix (Applied Biosystems), according to manufacturer’s instructions. CCRL2 gene expression was normalized to cyclophilin A

(cycA) levels in each tissue, and displayed relative to CCRL2 expression levels detected in the spleen using the 2-ΔΔCT method (36). mCCRL2 5’ primer: ttccaacatcctcctccttg;

mCCRL2 3’ primer: gatgcacgcaacaataccac; cycA 5’ primer: gagctgtttgcagaccaaagttc;

cycA 3’ primer: ccctggcacatgaatcctgg.

129

Cell sorting and Wright-Giemsa stain.

Mouse peritoneal cells were stained as described and sorted by standard flow cytometric

techniques (FACsvantage, BD Biosciences, Mountain View, CA, USA; flow cytometry

was performed at the Stanford University Digestive Disease Center Core Facility, VA

Hospital, Palo Alto, CA, USA). 1-5x104 sorted cells were loaded into cytospin chambers

and centrifuged onto glass slides. The slides were stained with Wright-Giemsa dye by

standard automated techniques at the VA Hospital Hematology Lab (Palo Alto, CA,

USA) and examined by light microscopy with a 40x objective.

β-hexosaminidase release assay.

BMCMCs were sensitized with 10 μg/ml of anti-DNP IgE mAb (H1-ε-26) (37) by

overnight incubation at 37°C. The cells were then washed with Tyrodes buffer (10 mM

HEPES pH 7.4, 130 mM NaCl, 5 mM KCl, 1.4 mM CaCl2, 1 mM MgCl2, 0.1% glucose

and 0.1% bovine serum albumin [fraction V, SIGMA]), and resuspended at 8 x 106

cells/ml. 25 μl of a 2x concentration of stimuli (final 0, 1, 10 and 100 ng/ml 2,4-

dinitrophenyl-conjugated human serum albumin [DNP-HSA; SIGMA] or 0.1 μg/ml PMA

[SIGMA] + 1 μg/ml A23187 calcium ionophore [SIGMA]) were added to the wells of 96

well V-bottom plate (Costar), and then 25 μl of 8 x 106 cells/ml IgE-sensitized BMCMCs

were added and incubated at 37°C for 1 hour. After centrifugation, supernatants were collected. The supernatants from non-stimulated IgE-sensitized BMCMCs treated with 50

μl of 0.5% (v/v) Triton X-100 (SIGMA) were used to determine the maximal (100%) cellular β-hexoaminidase content, to which the experimental samples were normalized. 130

β-hexosaminidase release was determined by enzyme immunoassay with p-nitrophenyl-

N-acetyl-β-D-glucosamine (SIGMA) substrate as follows: 10 μl of culture supernatant were added to the wells of a 96 well flat-bottom plate. 50 μl of 1.3 mg/ml p-nitrophenyl-

N-acetyl-β-D-glucosamine solution in 100 mM sodium citrate (pH 4.5) was added, and the plate was incubated at room temperature for 15-30 minutes. Next, 140 μl of 200 mM glycine (pH 7.0) was added to stop the reaction and the OD405 was determined.

T cell:mast cell co-culture.

For CD3+ T cell purification, a single cell suspension of spleen cells was prepared, and red blood cells were lysed (RBC lysing buffer, Sigma). Spleen cells were incubated with biotinylated anti-mouse B220, Gr-1, CD11b, CD11c, CD49b, Ter119, and c-kit for 20 minutes at 4 °C. The cells were then washed and incubated with streptavidin-beads

(Miltenyi Biotec) for 20 minutes at 4 °C, and washed again and passed through a magnetic cell-sorting column (MACS column; Miltenyi Biotec), yielding >95% CD3+ T cells. T cells were co-cultured with mast cells as described previously (38). BMCMCs were sensitized with 1 μg/ml of anti-DNP IgE mAb at 37 °C overnight. After IgE sensitization, BMCMCs were treated with mitomycin C (Sigma; 50 μg per 107 cells) for

15 minutes at 37 °C. BMCMCs and T cells were suspended in RPMI 1640 media

(Cellgro) supplemented with 50 μM 2-mercaptoethanol (Sigma), 50 μg/ml streptomycin

(Invitrogen), 50 U/ml penicillin (Invitrogen) and 10% heat inactivated FCS (Sigma). T cells (0.25 x 105 cells/well) were plated in a 96 well flat-bottom plate (BD Falcon) coated

with 1 μg/ml anti-mouse CD3 (145-2C11; BD PharMingen) or hamster IgG 131

(eBioscience) (in some experiments, “anti-CD3 (-)” indicates the substitution of control

IgG for anti-CD3), and mitomycin C-treated, IgE-sensitized or non-sensitized BMCMCs

(0.25 x 105 cells/well) in the presence or absence of 5 ng/ml DNP-HSA at 37 °C for 72 hours. Proliferation was assessed by pulsing with 0.25 μCi [3H]-thymidine (Amersham

Bioscience) for 6 hours, harvesting the cells using Harvester 96® Mach IIIM (TOMTEC) and measuring incorporated [3H]-thymidine using a Micro Beta System (Amersham

Bioscience).

FITC-induced contact hypersensitivity (CHS).

FITC-induced CHS was performed as described previously (39) with minor modifications. Mice were shaved on abdomen 2 days before FITC-sensitization. Mice were then treated with 200 μl of 2.0% (w/v) FITC isomer I (SIGMA) suspension in acetone-dibutyl phthalate (1:1). Five days after sensitization with 2.0% FITC, mice were challenged with 40 μl of vehicle alone to the right ear (20 μl to each side) and 0.5% (w/v)

FITC solution to the left ear (20 μl to each side). Each mouse was housed in a separate cage to prevent contact with each other after FITC challenge. Ear thickness was measured before and at multiple intervals after FITC challenge with an engineer’s microcaliper

(Ozaki MFG. CO., LTD., Itabashi, Tokyo, Japan).

Harvesting mouse leukocytes.

The Veterans Affairs Palo Alto Health Care System Institutional Animal Care and Use

Committee, Palo Alto, CA, and the Stanford University Administrative Panel on

Laboratory Animal Care, Stanford, CA, approved all animal experiments. To harvest

132

blood leukocytes, mice were given a fatal overdose of anesthesia (ketamine/xylazine) as

well as an i.p. injection of heparin (100 units, Sigma). Mouse blood was collected by

cardiac puncture. Up to 1 mL of blood was added to 5 mL of 2 mM EDTA in PBS, and 6

mL of 2% dextran T500 (Amersham Biosciences, Piscataway, NJ, USA) was added to

crosslink red blood cells. The mixture was incubated for 1 hour at 37°C, the supernatant

was removed and pelleted, and the cells were resuspended in 5 mL red blood cell lysis

buffer (Sigma) and incubated at RT for 5 minutes. The cells were pelleted, and

resuspended for use in cell staining. Bone marrow cells were harvested by flushing

femurs and tibias with media followed by red blood cell lysis. Peritoneal lavage cells

were obtained by i.p. injection of 10 mL PBS, gentle massage of the peritoneal cavity,

and collection of the exudate. For some experiments, 500 μl of peritoneal cells (2x106 cells/mL) were incubated for 24 hours with either LPS (1 μg/mL), TNFα (10 ng/mL),

IFNγ (100 U/mL), polyI:C (20 μg/mL), CpG (10-100 μg/mL), or TGFβ (5 ng/mL). For mast cell RNA isolation, peritoneal mast cells were enriched for by density centrifugation. Peritoneal exudate cells (~140 million) were harvested from 9 male WT mice >1 y.o. The cells were resuspended in 10 ml of PBS and underlayed with 5 ml

NycoPrep 1.077A (Axis-shield PoC AS, Oslo, Norway). Following centrifugation,

~140,000 high-density mast cells were recovered at the bottom of the tube (along with

~100,000 contaminating red blood cells). For functional assays using primary mast cells, peritoneal lavage was performed using Tyrodes solution, and the cells were kept at room temperature throughout harvest.

133

In vitro transwell chemotaxis.

For migration experiments using cell lines, 2.5x105 cells/100 μl chemotaxis media (RPMI

+ 10% fetal calf serum) were added to the top wells of 5-um pore transwell inserts

(Costar, Corning, NY, USA), and test samples in 600 μl media were added to the bottom wells. After incubating the transwell plates for 1.5 hours at 37ºC, the bottom wells were harvested and flow cytometry was used to assess migration. For primary cell chemotaxis,

1x106 cells/100 μl were added to the top well, and, following a 2 hour incubation at 37ºC,

polystyrene beads (15.0 μm diameter, Polysciences, Warrington, PA, USA) were added

to each well to facilitate normalization of the cell count. The cells were then stained for c-

kit, F4/80, and/or CD11b expression and analyzed by flow cytometry. A ratio was

generated and percent input migration was calculated.

Intracellular calcium mobilization.

Chemoattractant-stimulated Ca2+-mobilization was performed following Alliance for Cell

Signaling protocol ID PP00000210. Cells (3x106/mL) were loaded with 4 μM Fluo4-AM

and 0.16% Pluronic acid F-127 (Molecular Probes) in modified Iscove’s medium

(Iscove’s medium with 1% heat inactivated bovine calf serum and 2 mM L-glutamine,

Invitrogen) for 30 minutes at 37°C. The samples were mixed every 10 minutes during

loading, washed once, resuspended at up to 2x106/mL in the same buffer, and allowed to

rest in the dark for 30 minutes at room temperature. Chemoattractant-stimulated change

in Ca2+-sensitive fluorescence of Fluo-4 was measured over real-time with a FACsScan flow cytometer and CellQuest software (BD Biosciences) at room temperature under 134

constant stirring (500 rpm). Fluorescent data were acquired continuously up to 1024

seconds at 1-second intervals. The samples were analyzed for 45 seconds to establish

basal state, removed from the nozzle to add the stimuli, and then returned to the nozzle.

Mean channel fluorescence over time was analyzed with FlowJo (TreeStar, Ashland, OR,

USA) software. In some experiments, to identify mast cells, mixed peritoneal leukocytes

were pre-incubated with c-kit-PerCP mAb for 3 minutes immediately before the start of

each sample. In other experiments, mCCRL2HA/L1.2 or empty vector pcDNA3/L1.2

cells were loaded for 30 minutes with 1000 nM serum-form chemerin (incubated in

binding buffer on ice), washed 2x with PBS, and resuspended in Iscoves at 2 x106/ml.

500 μl of these chemerin-loaded cells were added to mCMKLR1/L1.2 cells loaded with

Fluo4-AM, and calcium mobilization was evaluated.

Receptor internalization assay. mCMKLR1-HA, huCCRL2-HA, and mCCRL2-HA L1.2 transfectants were incubated with for 15 minutes with 100 nM serum form chemerin at the indicated temperature in cell culture media. The cells were then washed with 200 μl PBS and stained with mouse anti-HA (Covance, Inc) or mIgG1 isotype control, followed by staining with secondary anti-mouse IgG1 PE, fixed, and analyzed by flow cytometry.

Ligand-independent receptor internalization assay. mCMKLR1-HA and mCCRL2-HA L1.2 transfectants were loaded for 30 minutes on ice with primary antibody (anti-HA or mIgG1 isotype control). The cells were washed with

200 μl PBS, incubated for the indicated times at 37°C to allow for receptor

135

internalization, and then placed on ice. The cells were then incubated with secondary anti-mouse IgG1 PE, and analyzed by flow cytometry.

Chemerin internalization assay.

mCCRL2-HA L1.2 transfectants or total peritoneal exudate cells were incubated with 10 nM His8-tagged serum-form chemerin for 30 minutes on ice. The primary cells were also

stained with F4/80 and c-kit mAbs. Secondary anti- His6 PE was added to the cells and

incubated for 30 minutes on ice. The cells were then incubated for the indicated times at

37°C to allow for chemerin internalization. The cells were incubated for 5 minutes on ice

with either PBS or acid wash buffer (0.2 M acetic acid, 0.5 M NaCl), and then analyzed

by flow cytometry. Mast cells were identified by gating on SSChiF4/80-c-kit+ cells.

Chemerin sequestration assay.

2 nM serum form chemerin was incubated with 40 million cells of the indicated

transfectant lines (or media alone) for 15 minutes at 37°C. The cells were removed by

centrifugation, and a volume of the conditioned media equivalent to 0.2 nM chemerin

(barring any sequestration or degradation) was tested in transwell chemotaxis using mCMKLR1-HA/L1.2 cells.

136

Results

mCCRL2-specific mAbs selectively stain mouse mast cells

We generated monoclonal antibodies BZ5B8 and BZ2E3 (IgG2aκ) with reactivity

to the extracellular amino-terminal domain of mCCRL2 (Fig. 1A). The antibodies were

specific to mCCRL2-HA/L1.2 transfectants, displaying no cross-reactivity with other

GPCR/L1.2 transfectants tested (mCMKLR1, huCMKLR1, mCRTH2, huCCRL2, or

mCCR10). Reactivity with CXCR1-through-6 and CCR1-10 was excluded by lack of

staining of blood cell subsets or cultured mouse cells known to express these receptors

(Fig. S1 and data not shown). In agreement with published RNA expression data (3),

peritoneal macrophages treated with LPS upregulated mCCRL2 protein expression;

expression of mCCRL2 was also upregulated in such cells by treatment with TNFα,

IFNγ, or poly:IC (Fig. S2A).

Freshly isolated mouse blood T cells, B cells, NK cells, bone marrow neutrophils,

and resting peritoneal macrophages were all negative for mCCRL2 expression (Fig.

S1A). A small population of highly granular (SSChi), F4/80- c-kit+ leukocytes in the peritoneal cavity, however, uniformly stained for mCCRL2 (Fig. 1B). These cells also expressed the high affinity IgE Fc receptor FcεRI (data not shown). On staining with

Wright-Giemsa stain, sorted F480-CCRL2+ cells displayed intense metachromatic staining of abundant cytoplasmic granules, as did mast cells sorted as F4/80- c-kit+ cells

(Fig. 1C). Thus both traditional morphologic and immunophenotypic analyses indicate

137

that mCCRL2 is constitutively expressed by mast cells, and the expression is surprisingly selective for mast cells in the absence of pathologic stimuli.

Mast cells derived from bone marrow progenitors in vitro (BMCMCs) upregulated expression of mCCRL2 over time in culture. Early mast cell progenitors were negative for mCCRL2, but after >2 months in culture the cells uniformly expressed detectable levels of mCCRL2, albeit the levels were lower than those on peritoneal mast cells in vivo (Fig. 1D). We confirmed RNA expression of mCCRL2 in peritoneal mast cells by real time quantitative RT-PCR (Fig. 1E).

138

Figure 1. Mast cells express mCCRL2. (A) Generation of anti-mCCRL2 specific mAbs. Unlabeled mCCRL2/L1.2 transfectants were mixed 1:1 with CMFDA-labeled CCR10/L1.2 transfectants and used to identify mCCRL2-specific mAbs by flow cytometry. (B) Freshly isolated peritoneal leukocytes were harvested and mCCRL2 expression was evaluated on SSChi F4/80- c-kit+ mast cells. (C) F4/80- mCCRL2+ and F4/80- c-kit+ peritoneal cells were sorted, harvested by cytospin, and stained by Wright-Giemsa. Cells were examined by light microscope using a 40x objective. (D) Bone marrow-derived cultured mast cells (BMCMCs) were generated and stained for mCCRL2 reactivity. (E) The relative RNA expression of mCCRL2 was assessed in mast cells by real time quantitative PCR. The expression data were normalized to Cyclophilin A and displayed relative to mCCRL2 expression in the spleen (set = 1.0). Each bar represents the mean ± SD of triplicate wells; ND, not

139

detectable; RT, reverse transcriptase. One representative data set of the at least 3 experiments, each of which gave similar results, is shown for each part of this figure.

140

CCRL2 and mast cell phenotype and function

We evaluated numbers of mast cells in CCRL2 KO mice in vivo, as well as certain basic functions of CCRL2 KO BMCMC in vitro. Our anti-mCCRL2 mAbs failed to stain peritoneal mast cells from CCRL2 KO mice, confirming the genetic ablation of the gene (Fig. 2A). The mice are fertile, reproduce with the expected Mendelian distribution of KO:heterozygotes:WT and male:female ratios, and display no differences in basal mast cell numbers in the ear skin or in mesenteric windows (Fig. 2B).

CCRL2 KO and WT BMCMCs expressed similar levels of c-kit (CD117) and

FcεRI (data not shown). BMCMCs from WT or mCCRL2 KO mice also displayed similar chemotactic responses to stem cell factor (SCF), indicating no inherent differences in cell migration (Fig. 3A); similar degranulation responses to antigen- mediated IgE/FcεRI crosslinking as assessed by β-hexosaminidase release (Fig. 3B); and similar activation-dependent secretion of cytokines TNFα and IL-6 (Fig. 3C). We recently showed that antigen-mediated IgE/FcεRI crosslinking upregulated expression of several co-stimulatory molecules on BMCMCs (40), however, we did not detect any

CCRL2-dependent differences in CD137 (4-1BB) or CD153 (CD30L) induction (Fig.

3D). BMCMCs stimulated by antigen-mediated IgE/FcεRI crosslinking also can enhance

T cell proliferation (38, 40). While naive T cells proliferated markedly in response to treatment with anti-CD3 and co-incubation with mitomycin C-treated, antigen-specific

IgE stimulated BMCMCs, there were no CCRL2-dependent differences in the ability of

BMCMCs to enhance T cell proliferation (Fig. 3E) or T cell secretion of IFNγ or IL-17 141

(not shown) in the conditions tested. Thus, the presence of absence of CCRL2 on

BMCMCs did not significantly influence the basic mast cell functions tested here.

142

Figure 2. CCRL2 KO mice. (A) mCCRL2 KO mice are deficient in CCRL2 protein expression. Freshly isolated peritoneal leukocytes were harvested and mCCRL2 expression was evaluated on SSChiF4/80-ckit+ mast cells. A representative histogram plot of the at least 3 independent experiments performed, each of which gave similar results, is shown. (B) Enumeration of ear skin and mesenteric mast cells in WT and KO mice. Ear skin: KO (n = 9), WT,het (n = 4,1), 4 sections/mouse. Mesenteric window: KO (n=18), WT,het (n = 9,2).

143

Figure 3. BMCMCs from mCCRL2 KO and WT mice display similar functional responses in vitro. (A) In vitro transwell chemotaxis to stem cell factor (SCF). Four populations of BMCMCs were tested, with duplicate wells for each genotype. The mean ± SEM is displayed. (B-D) BMCMCs were sensitized with DNP-specific IgE and then activated by addition of DNP-HSA. The following parameters were measured: (B) Degranulation (as quantified by β-hexoaminidase release), (C) TNFα and IL-6 secretion, and (D) Upregulation of co-stimulatory molecules CD137 and CD153. (E) BMCMC-stimulated T cell proliferation. Naive T cells were incubated as indicated with anti-CD3, and co- cultured with mitomycin C-treated BMCMCs from WT or CCRL2 KO mice pre- incubated with or without DNP-specific IgE, and tested in the presence or absence of DNP-HSA. Cell proliferation was measured by tritiated thymidine incorporation. For (B), (C), and (D), n = 7 KO, n = 4 WT, mean ± SD. For E, the mean of triplicate 144

measurements ± SD is shown for a representative data set of 3 experiments (each of which gave similar results).

145

Mast cell-expressed CCRL2 is required for optimal induction of IgE-dependent passive

cutaneous anaphylaxis

To search for potential contributions of CCRL2 to pathophysiologic responses in

vivo, we next examined certain inflammatory conditions that are known to involve mast

cells. Mast cells are required for optimal expression of the T cell-mediated contact hypersensitivity (CHS) induced by a protocol employing FITC (fluorescein

isothiocyanate), but not that induced by other protocols employing DNFB (2,4-dinitro-1- fluorobenzene) (39, 41). However, CCRL2 was largely dispensable for the tissue swelling associated with FITC-triggered CHS, as both WT and CCRL2 KO mice developed statistically indistinguishable responses (Fig. S3).

We next examined a mast cell-dependent model of atopic allergy, the IgE- dependent passive cutaneous anaphylaxis (PCA) reaction. Animals sensitized with 150 ng/ear DNP-specific IgE and challenged with antigen (DNP-HSA) i.v. developed strong local inflammatory responses, with no significant difference in the tissue swelling observed in WT vs. CCRL2 KO mice (82 ± 9 vs. 91 ± 9 x 10-2 mm of swelling at 30 min

after antigen challenge, respectively [p>0.05, by Student’s t-test] (Fig. S4A)). However,

when the sensitizing dose of DNP-specific IgE was reduced to 50 ng/ear, the PCA reactions in CCRL2 KO mice were significantly impaired compared to those in WT mice

(42.2 ± 2.8 vs. 24.9 ± 2.7 x 10-2 mm of swelling at 30 min after antigen challenge,

respectively [p<0.005, by Student’s t-test] (Fig. 4A)).

146

To assess the extent to which mCCRL2 expression specifically on mast cells was critical for the defect in IgE-dependent PCA observed in mCCRL2 KO mice, we engrafted mast cell-deficient KitW-sh/Wsh mice intra-dermally in the ear pinnae with either

WT or mCCRL2 KO BMCMCs; 6-8 weeks later, the animals were subjected to IgE-

dependent PCA. Such mast cell engraftment of mast cell-deficient KitW-sh/Wsh or KitW/W-v mice is routinely used to identify the roles of mast cells in biological responses in vivo

(29). There was no difference in the extent of PCA-associated ear swelling between KitW- sh/Wsh mice that had been engrafted with WT vs. mCCRL2 KO BMCMCs when the

animals were sensitized with 50 ng/ear DNP-specific IgE and challenged with i.v. antigen

(19.5 ± 3.6 vs. 19.9 ± 2.6 x 10-2 mm of swelling at 30 min after antigen challenge,

respectively [p>0.05, by Student’s t-test] (Fig. S4B)). Nor was there a significant

difference in the numbers of leukocytes infiltrating the dermis at these sites at 6 h after

antigen challenge (Fig. S4C and S5).

However, when the sensitizing dose of DNP-specific IgE was reduced to 5 ng/ear,

there was a significant reduction in ear swelling responses in mice that had been

engrafted with mCCRL2 KO BMCMCs compared with those that had been engrafted

with WT BMCMCs (12.5 ± 1.2 vs. 8.4 ± 0.8 x 10-2 mm of swelling at 30 min after

antigen challenge, respectively [p<0.01, by Student’s t-test] (Fig. 4B). There were no significant differences in the total number of mast cells detected histologically in WT vs.

CCRL2 KO BMCMC-engrafted ears, thus ruling out any CCRL2-dependent effects on

mast cell engraftment efficiency (Fig. 4C). However, at 6 h after antigen challenge, IgE-

147

dependent PCA reactions in ears that had been engrafted with CCRL2 KO mast cells

contained ~ 50 % fewer leukocytes (predominantly neutrophils and mononuclear cells)

than did reactions in ears that had been engrafted with wild type mast cells [p<0.03 by the

Mann Whitney U-test] (Fig. 4D and Fig. 5). IgE-dependent PCA reactions were

associated with a marked reduction in the numbers of dermal mast cells which could be

identified in histological sections of these sites based on the staining of the cells’ cytoplasmic granules (Fig. 4C), an effect that most likely reflected extensive mast cell degranulation at these sites (42, 43).

We conclude that while mast cell-expressed mCCRL2 is not required for the development of IgE-dependent PCA reactions in vivo, mast cell expression of CCRL2 can significantly enhance the local tissue swelling and leukocyte infiltrates associated with such reactions in mice that have been sensitized with relatively low amounts of antigen-specific IgE.

148

Figure 4. Mast cell-expressed mCCRL2 is required for maximal tissue swelling and numbers of dermal leukocytes in passive cutaneous anaphylaxis. (A) Wild type (WT) or CCRL2 knock out (KO) mice were sensitized by injection of 50 ng anti-DNP IgE into left ear skin (with vehicle injection into right ear skin as the control). The mice were challenged by i.v. injection of DNP-HSA (200 μg/mouse, i.v.) the next day, and ear swelling was measured at the indicated time points, mean ± SEM, n = 3 experiments (a total of 21 KO and 16 WT mice per group), * p <0.005 by ANOVA comparing swelling in WT vs. KO ears sensitized with antigen specific IgE. (B-D) The 149

ears of mast cell deficient KitW-sh/Wsh mice were engrafted with bone marrow-derived cultured mast cells (BMCMCs) from either WT or mCCRL2 KO mice. 6-8 weeks later, the mice were sensitized (5 ng IgE/left ear, with vehicle into the right ear as the control), challenged with specific antigen (200 μg DNP-HSA, i.v.), and assessed for (B) tissue swelling as described in part (A), and for numbers of mast cells (C) or leukocytes (D) per mm2 of dermis. Data shown as mean ± SEM, n = 3 experiments, 15 total mice per group in (B) and the numbers of mice sampled for histological data shown in (C) and (D). * p <0.001 by ANOVA comparing swelling in mCCRL2 KO BMCMC- vs. WT BMCMC- engrafted ears sensitized with antigen specific IgE. (C) Enumeration of mast cells present in the dermis of ear skin in engrafted animals from (B) following elicitation of PCA (IgE) or in vehicle-injected control (vehicle) ears. ** p <0.005 by Student’s t-test vs. values for the vehicle-injected ears in the corresponding WT BMCMC- or KO BMCMC-engrafted KitW-sh/Wsh mice. (D) Numbers of leukocytes per mm2 of dermis, assessed in formalin- fixed, paraffin-embedded, hematoxylin and eosin-stained sections of mice from (B, C). *** p<0.0001 by the Mann Whitney U-test vs. corresponding values for the vehicle- injected ears in WT BMCMC- or KO BMCMC-engrafted KitW-sh/Wsh mice. The numbers over the bars for vehicle-injected mice are the mean values.

Figure 5. Histologic features of IgE-dependent PCA reactions in WT BMCMC- vs. KO BMCMC-engrafted KitW-sh/Wsh mice. Histological sections of ear skin from WT BMCMC-engrafted KitW-sh/Wsh mice (A-C) and KO BMCMC-engrafted KitW-sh/Wsh mice (D-F) from the same group shown in Figure 4D show no evidence of inflammation in ears analyzed 6 h after injection of vehicle (A,D), but evidence of tissue swelling and increased numbers of leukocytes, consisting

150

predominantly of polymorphonuclear leukocytes (some indicated by arrowheads in (C,F) and occasional mononuclear cells (indicated by an arrow in (C)), at 6 h after antigen challenge in both WT BMCMC-engrafted KitW-sh/Wsh mice (B, C) and KO BMCMC- engrafted KitW-sh/Wsh mice (E,F). Hematoxylin and eosin stain; scale bars = 50 μm.

151

CCRL2 binds chemerin

To investigate possible functional roles for CCRL2, we set out to validate/identify

CCRL2 ligands. It was reported that mCCRL2/HEK293 transfectants respond functionally to CCR2 ligands CCL2, CCL5, CCL7, and CCL8 via intracellular calcium mobilization and transwell chemotaxis (9), although this conclusion is controversial (8,

10). These chemokines did not induce migration of mCCRL2/L1.2 transfectants in our in vitro transwell chemotaxis assays (Fig. S6). We also tested a panel of known chemoattractants (CCL11, CCL17, CCL22, CCL25, CCL27, CCL28, CXCL9, and

CXCL13), as well as protein extracts from homogenized mouse tissues (lungs, kidney, liver, brain, and spleen) and found that none stimulated mCCRL2-dependent chemotaxis in our in vitro transwell chemotaxis assays (data not shown). Given the aberrant

“DRYLAIV” motif present in mouse and human CCRL2, we and others postulated that mCCRL2 may act as a “silent” receptor (10), capable of binding chemoattractant (s) but incapable of transducing signals via classical second messengers. That hypothesis is consistent with the negative results obtained in our efforts to induce chemotaxis of mCCRL2/L1.2 transfectants in our in vitro transwell chemotaxis assays.

Although we failed to identify evidence of signaling effects of any of the tested chemoattractants, we were able to identify a high affinity ligand for the receptor: in independent studies in which we were using our anti-CCRL2 mAbs as controls for staining, we serendipitously discovered that chemerin, a protein ligand for signaling receptor CMKLR1 (chemokine-like receptor 1, reviewed in (28)), inhibited the binding

152

of mCCRL2-specific mAbs to mouse peritoneal mast cells. In Figure 6A we illustrate the

potent ability of chemerin to inhibit anti-CCRL2 staining of mouse peritoneal mast cells.

Increasing concentrations of chemerin blocked the binding of anti-mCCRL2 BZ5B8 (Fig.

6A) and BZ2E3 (data not shown) in a dose-dependent manner (EC50 = 21 nM). The effect

was specific to anti-mCCRL2:mCCRL2 interactions, since binding of IgE to FcRεI was unaffected by 1000 nM chemerin (Fig. 6A); and 1000 nM CCL2 had no effect on CCRL2 staining (Fig. 6A).

To confirm the identification of CCRL2 as a chemerin receptor, we performed radioligand-binding studies using iodinated chemerin. Cells were incubated with a fixed concentration of radiolabeled human chemerin plus increasing concentrations of unlabelled chemerin. Chemerin bound specifically to mCCRL2-HA/L1.2 transfectants

(EC50 = 1.6 nM) (Figure 6B), but no binding was detected to untransfected or mCRTH2-

HA-transfected cells (a prostaglandin D2-binding chemoattractant receptor, data not

shown). Furthermore, despite being the most divergent mouse-to-man orthologs in the

chemoattractant receptor subfamily, huCCRL2 also bound specifically to chemerin (EC50

= 0.2 nM) (Figure 6B). The binding affinity of chemerin for CCRL2 was similar to if not slightly better than chemerin binding to the first identified chemerin receptor, mCMKLR1 (EC50 = 3.1 nM) (Fig. 6B). In saturation-binding studies, chemerin bound to

mCCRL2 at a single binding site with a calculated Kd of 1.6 nM (Figure 6C).

We developed an immunofluorescence-based chemerin-binding assay to evaluate

chemerin binding by flow cytometry. Cells were incubated with a fixed concentration of

153

C-terminal His8-tagged serum form human chemerin plus increasing concentrations of untagged chemerin, and anti-His6 PE was used to detect binding. In this assay, chemerin

bound specifically to mCCRL2-HA (EC50 = 45 nM) and huCCRL2 (EC50 = 7 nM) L1.2

transfectants (Figure 6D). Chemerin binding to CCRL2 was not affected by a variety of

other chemoattractants (Fig. S7), and mCRTH2-HA/L1.2 transfectants did not bind to

chemerin (Figure 6D), demonstrating specificity for chemerin:CCRL2 interactions.

Interestingly, we were unable to detect chemerin binding to mCMKLR1-HA/L1.2

transfectants in the immunofluorescence chemerin-binding assay (Figure 6D): this may

reflect inhibition of binding by the C-terminal His8 tag (which would be analogous to the

inhibitory activity of the carboxyl-terminal residues in the chemerin pro-form); or

potentially inaccessibility of the His8 epitope to the detection mAbs when His8-tagged

chemerin is bound to CMKLR1.

In radioligand binding studies, the His8-tag had little effect on the potency of

chemerin binding to mCCRL2 (EC50 = 0.8 nM); however, His8-tagged chemerin bound

with 10-fold less potency to mCMKLR1 (EC50 = 26.3 nM) (Figure S8A). The bioactive

9-mer carboxyl-terminal chemerin peptide (residues 149-157, YFPGQFAFS) was 10-fold

less potent (EC50 = 26.2 nM, Figure 8A) than chemerin protein in binding to CMKLR1.

In CCRL2 binding, however, the bioactive chemerin peptide wan an inefficient

competitor (EC50 could not be determined, Figure S8B).

Thus, the data indicate distinct binding modes for chemerin and CCRL2 vs. chemerin and CMKLR1: the carboxyl-terminal domain of chemerin that is critical for

154

binding to CMKLR1 is relatively uninvolved and unencumbered when chemerin is bound

to CCRL2.

Freshly isolated mouse peritoneal mast cells also bound to chemerin (Figure 6E);

and there was no obvious preference for binding of the pro-form vs. the active serum

form (this was also observed in radioligand binding studies using L1.2 transfectants, data not shown). Moreover, mouse peritoneal mast cells bound both human and mouse chemerin (Figure 6E). Finally, peritoneal mast cells from mCCRL2 KO mice did not bind to chemerin, further confirming the role of CCRL2 in the binding of chemerin to such mast cells (Figure 6E).

155

156

Figure 6. CCRL2 binds chemerin. (A) Chemerin blocks anti-CCRL2 mAb binding. Various concentrations of human chemerin or CCL2 were incubated with total peritoneal mast cells on ice for 5 minutes, followed by incubation with CCRL2 specific mAb BZ2E3 or anti-IgE and detected with secondary anti-rat PE or anti-mouse IgE PE. (B-C) Radiolabeled chemerin binding. (B) Displacement of iodinated chemerin (residues 21-148) binding to mCMKLR1, huCCRL2, and mCCRL2 by full-length chemerin. (C) Saturation binding of 125I- chemerin21-148 to mCCRL2-transfected cells. (D) Immunofluorescence-based chemerin binding. Various concentrations of untagged serum-form chemerin were incubated with mCCRL2-HA, huCCRL2-HA, mCRTH2-HA, or mCMKLR1-HA L1.2 transfectants in the presence of 10 nM His8-tagged serum-form chemerin. Samples were incubated on ice for 30 min. Secondary anti-His6 PE was added to detect levels of bound His8-tagged chemerin, and MFI values are displayed. Mean MFI ± range of duplicate staining wells are shown. (E) Mast cell binding. 1000 nM untagged chemerin isoforms were incubated with total peritoneal cells from either WT or CCRL2 KO mice in the presence or absence of 10 nM His8-tagged chemerin isoforms. Secondary anti- His6 PE was added to detect hi levels of bound His8-tagged chemerin. SSC F4/80-c-kit+ mast cells were analyzed. A representative data set of the 3 (for B, D, and E) or 2 experiments (for A and C) performed, each of which gave similar results, is shown.

157

CCRL2 does not support chemerin-driven signal transduction

Despite high affinity binding to binding to mCCRL2, chemerin failed to trigger intracellular calcium mobilization in mCCRL2/L1.2 transfectants (Fig. 7A). Chemerin triggered a robust calcium flux in cells expressing the chemerin signaling receptor mCMKLR1, confirming its activity (Fig. 7A). mCCRL2-HA/L1.2 transfectants responded to CXCL12 (via endogenous CXCR4), indicating their competence for demonstrating calcium mobilization in this assay (Fig. 7A). Furthermore, although it was reported that CCL2 triggered intracellular calcium mobilization in CCRL2/HEK293 transfectants, in our experiments neither CCL2 nor chemerin functioned as agonists for

CCRL2 in the HEK293 background, either alone or in combination (Figure S9).

Since GPCR function can require cell type-specific cofactors, we wanted to determine whether CCRL2 could mediate chemerin signaling when expressed physiologically on mouse peritoneal mast cells. Chemerin did not trigger intracellular calcium mobilization in freshly isolated mouse peritoneal mast cells, although these cells responded to ATP (via purinoreceptors (44)), indicating their competence in this assay

(Fig. 7B). Furthermore, neither human nor mouse CCRL2-HA/L1.2 transfectants migrated to a range of doses of chemerin in in vitro transwell chemotaxis experiments

(Fig. 7C). Freshly isolated mouse peritoneal mast cells also failed to migrate to chemerin

(Fig. 7D). In comparison, chemerin triggered a robust, dose dependent migratory response in mCMKLR1-HA/L1.2 cells (Fig. 7C). Mouse and human CCRL2/L1.2 cells migrated to CXCL12 and CCL19 (via endogenously expressed CXCR4 and CCR7,

158

respectively), and primary mouse peritoneal mast cells migrated to stem cell factor

(SCF), indicating their ability to demonstrate chemotaxis is this assay (Fig. 7C,D).

Furthermore, CCL2 and chemerin did not synergize with each other to induce a functional migratory response in mCCRL2/L1.2 transfectants in in vitro transwell migration assays (data not show).

159

Figure 7. Chemerin:CCRL2 binding does not trigger intracellular calcium mobilization or chemotaxis. (A) mCCRL2 and mCMKLR1 L1.2 transfectants were loaded with Fluo-4, treated with chemerin and/or CXCL12 at the indicated times, and examined for intracellular calcium mobilization. (B) Mouse peritoneal mast cells were enriched by Nycoprep density centrifugation, loaded with Fura-2 and Fluo-4, and assayed for calcium mobilization. 1000 nM chemerin and 100 nM ATP were added as indicated. (C) mCCRL2-HA, huCCRL2-HA, and mCMKLR1-HA L1.2 transfectants were tested for transwell chemotaxis to various concentrations of chemerin. The mean ± range of duplicate wells is shown. (D) Mouse peritoneal mast cells were assayed for in vitro chemotaxis to various concentrations of SCF and chemerin. Mast cells were identified by gating on SSChi CD11b- c-kit+ cells. The mean ± SD of triplicate wells is shown for an individual experiment. A representative data set of the 3 experiments performed, each of which gave similar results, is shown for all parts of this figure.

160

CCRL2 does not internalize chemerin

Our data place CCRL2 in a class of ‘atypical’ receptors that include D6, DARC, and CCX-CKR, all of which bind to chemoattractants but do not support classical ligand- driven signal transduction (11). These other receptors have recently been termed professional “chemokine interceptors” because they internalize and either degrade and/or transcytose chemokines (reviewed in (10, 11, 12)). To ask whether CCRL2 might have interceptor activity, we assessed the internalization of CCRL2, and of CMKLR1 for comparison, in response to ligand binding. mCMKLR1-HA internalized rapidly (within

15 min) in response to 100 nM chemerin; and this internalization was inhibited by incubation on ice and in the presence of sodium azide (Fig. 8A), confirming that the effect is an active process (not due to chemerin-mediated displacement of the anti-HA mAb). In contrast, under the same conditions, CCRL2 failed to internalize: neither mouse nor human CCRL2, expressed on L1.2 transfectants, underwent ligand-induced internalization (Fig. 8A). Even prolonged incubation with chemerin (2 h at 37°C) failed to significantly reduce surface receptor levels (data not shown).

We also asked whether CCRL2 might undergo constitutive, ligand-independent endocytosis, as has been observed with D6 (13). Cell surface mCCRL2-HA and mCMKLR1-HA were stained with primary anti-HA mAb on ice, washed, and then shifted to 37°C for various times to permit receptor internalization. The cells were then stained with a secondary antibody to detect remaining surface anti-HA mAb. At the 15- min time point, neither mCCRL2 nor mCMKLR1 had undergone substantial ligand-

161

independent endocytosis, similar to CCX-CKR (19) and in contrast to D6 (13). By 60

min there was a noticeable reduction in staining intensity for both human and mouse

receptors, suggesting either low level constitutive endocytosis, receptor turnover, and/or antibody release (Fig. 8B).

Given that chemerin does not trigger mCCRL2 internalization, it is unlikely that chemerin itself is internalized in substantial amounts in CCRL2+ cells (in the absence of

CMKLR1). To confirm this directly, we loaded mCCRL2-HA/L1.2 cells with His8- tagged serum form chemerin and anti-His6 PE on ice, and then shifted the cells to 37°C to

permit internalization. At the indicated time points, the cells were washed with either

PBS or a high salt acid wash buffer that strips bound chemerin from the surface of the

cell (see the zero time point in Fig. 8C). In contrast to D6 and CCX-CKR, where >70% of cell-associated CCL3 (13) and 100% of cell-associated CCL19 (19) became resistant to acid wash within 5 min, respectively, there was essentially no acid resistant cell- associated chemerin at the 5 min time point, and very little at the 60 min time point (Fig.

8C). On the other hand, there was a time-dependent increase in surface bound chemerin

(“PBS wash” in Fig. 8C). Freshly isolated peritoneal mouse mast cells also did not

internalize chemerin (“acid wash” in Fig. 8D). In contrast to mCCRL2/L1.2 transfectants,

however, at the 60 min time point there was a considerable reduction in surface bound chemerin (“PBS wash” in Fig. 8D), suggesting either eventual extracellular degradation or chemerin release. Furthermore, mCCRL2-HA/L1.2 transfectants efficiently bound

162

chemerin from dilute aqueous solutions (Fig. 8E). Thus, it appears that CCRL2 binds and

indeed concentrates chemerin on the cell surface.

Finally, we wanted to assess whether chemerin sequestered by mCCRL2+ cells could trigger a response in mCMKLR1+ cells. We loaded empty vector pcDNA3 or mCCRL2-HA L1.2 cells with chemerin, washed with PBS, added the loaded cells to mCMKLR1/L1.2 responder cells labeled with a calcium sensitive dye, and assessed intracellular calcium mobilization. Chemerin-loaded mCCRL2-HA/L1.2 cells, but not pcDNA3/L1.2 cells, triggered calcium flux in the responder cells (Fig. 8F). Thus, CCRL2 can concentrate bioactive chemerin, which then is available for interaction with

CMKLR1 on adjacent cells.

163

164

Figure 8. CCRL2 can increase local chemerin concentrations. (A) Chemerin does not trigger CCRL2 receptor internalization. mCCRL2-HA, huCCRL2-HA, and mCMKLR1-HA L1.2 transfectants were stained with anti-HA mAb and then incubated with or without 100 nM chemerin for 15 min at the indicated temperatures. (B) mCCRL2 is not rapidly constitutively internalized. mCCRL2-HA and mCMKLR1-HA L1.2 transfectants were incubated with primary anti-HA mAb, incubated for the indicated times at 37°C, and then stained with secondary anti-mIgG1 PE. mCMKLR1 cells incubated with 100 nM serum form chemerin served as a positive control. (C and D) Chemerin is not rapidly internalized. mCCRL2-HA L1.2 transfectants (C) or total peritoneal exudate cells (D) were incubated with 10 nM HIS8-tagged serum form chemerin and anti-His6 PE for 1 h on ice, and then shifted to 37°C. At the indicated time points, the cells were then washed with either PBS or acid wash buffer. Mast cells were identified by gating on SSChi F4/80-c-kit+ cells in (D). (E) CCRL2 can sequester chemerin from solution. 2 nM serum form chemerin was incubated with the indicated transfectant lines (or media alone) for 15 minutes at 37°C. The cells were removed by centrifugation, and the conditioned media was tested in transwell chemotaxis using mCMKLR1HA/L1.2 responder cells. The mean ± SD of triplicate wells for an individual experiment is shown. (F) CCRL2 can increase local concentrations of bioactive chemerin. mCCRL2-HA or empty vector pcDNA3 L1.2 transfectants were pre-loaded with 1000 nM serum-form chemerin and washed with PBS. mCMKLR1/L1.2 loaded with Fluo-4 served as responder cells. The intracellular calcium mobilization in the responder cells was measured over time as loaded cells or purified chemoattractant was added. Note that different scales are used on either side of the broken-axis indicator. A representative data set of the 3 experiments performed, each of which gave similar results, is shown for all parts of this figure.

165

Discussion

The evolutionary conservation of chemerin binding to CCRL2 despite divergence

in overall sequence between mouse and human receptors strongly suggests that the

interaction of chemerin with CCRL2 is physiologically important. The lack of substantial

chemerin internalization by CCRL2 disqualifies the receptor as a professional

chemoattractant interceptor, and instead implicates it as a physiologic mediator for the

concentration and presentation of bioactive chemerin. We hypothesize therefore that

CCRL2 serves a novel role among cell surface chemoattractant receptors, regulating the

bioavailability of chemerin in vivo to fine-tune immune responses mediated via the

signaling receptor CMKLR1.

Chemerin inhibits the binding of mAbs specific to amino-terminal epitopes of mCCRL2, implying that chemerin binds directly to the amino-terminal domain of the receptor. In this regard it is interesting to note that the first 16 amino acids of human and mouse CCRL2 share 81% identity, compared with just 17% shared identity in the remaining 24 amino acids of the amino-terminal domain (Fig. S10). This short conserved amino-terminal sequence may therefore embody a critical chemerin-binding domain.

Furthermore, CCRL2 binds to chemerin in an orientation that permits antibody access to

the short C-terminal His8-tag, suggesting that the critical cell-signaling carboxyl-terminus

of chemerin would also be exposed in the untagged form of the protein. Thus, we

hypothesize that CCRL2 (e.g. expressed by mast cells and activated macrophages) binds

to chemerin in vivo and presents the cell-signaling carboxyl-terminal domain of the

166

chemoattractant to CMKLR1+ cells such as macrophages, pDC, and NK cells (Fig. 9). In

its role as a specialized molecule for concentrating extracellular chemerin, CCRL2 may

operate akin to glycosaminoglycans, which are thought to bind, concentrate, and present

chemokines to leukocytes and facilitate their chemotaxis (45).

Figure 9. Proposed model of presentation of chemerin by CCRL2 to CMKLR1. (A) Chemerin binds to CCRL2 leaving the C-terminal peptide sequence free. The carboxyl-terminal domain of chemerin is critical for transducing intracellular signals and interacts directly with CMKLR1. CCRL2 may thus allow direct presentation of bound chemerin to adjacent CMKLR1-expressing cells. (B) Alternatively, CCRL2 may concentrate the ligand for proteolytic processing by activated mast cells or macrophages, enhancing the local production of the active form that could then act as a chemoattractant following release from the cell surface.

167

In experimental IgE-dependent PCA, a simple in vivo model of an IgE- and mast

cell-dependent acute allergic reaction, CCRL2 deficient mice expressed wild type levels

of tissue swelling when the mice were sensitized with one dose of antigen-specific IgE,

but exhibited significantly reduced tissue swelling associated with PCA reactions elicited in mice sensitized with one third less antigen-specific IgE. This clearly shows that mast cell expression of CCRL2 is not required for elicitation of IgE-dependent PCA, but may

substantially enhance reactions elicited under conditions where IgE may be limiting. PCA

has been widely used to study factors involved in mast cell responses in vivo (46, 47). In

our previous studies, we showed that pro-chemerin is rapidly activated by extracellular

serine proteases, such as secreted mast cell tryptase (27); thus the acute nature of the

PCA reaction is well suited to highlight possible consequences of rapid chemerin

conversion in vivo. CCRL2 expression by tissue mast cells may serve to coat the cells

with the pro-form of chemerin, which would be immediately activated upon mast cell

degranulation and tryptase release (27).

The reason why the tissue swelling associated with IgE-dependent PCA reactions

in CCRL2 KO mice was less than that observed in wild type mice is not yet clear. The

largest differences in the tissue swelling responses of the WT vs. KO mice (or in the mast

cell-deficient mice that had been engrafted with WT vs. KO BMCMCs) were at 30 min

after i.v. antigen challenge, an interval too short to observe substantial recruitment of

circulating leukocytes to the site of the reaction. Perhaps mast cell-bound and then

activated chemerin interacts with CMKLR1+ macrophages or other mononuclear cells 168

that might be present in normal skin, and such cells in turn have effects that can enhance

local vascular permeability. Alternatively, other cells types normally resident in the skin,

such as vascular endothelial cells, may respond to locally activated chemerin. The

reduced leukocyte numbers present in the dermis 6 h after the elicitation of IgE-

dependent PCA reactions in mice whose dermal mast cells lacked CCRL2 indicates that a

lack of mast cell CCRL2 also can diminish the longer term consequences of IgE-

dependent cutaneous reactions, perhaps by multiple mechanisms (which are yet to be

defined). Mast cells themselves are CMKLR1- (Figure S11). Notably, the inflammation associated with relatively strong PCA reactions was not substantially affected by the presence of absence of CCRL2, whether on all cells types (Fig. S4A) or only on skin mast cells (Fig. S4B and C). Thus, whatever contribution mast cell-associated CCRL2 makes to the tissue swelling and leukocyte infiltrates associated with PCA responses appears to enhance tissue sensitivity to suboptimal IgE-dependent mast cell activation.

Treatment of peritoneal macrophages with pro-inflammatory cytokines (IFNγ and

TNFα) or TLR ligands (LPS and poly:IC, but not CpG) upregulates CCRL2 expression (

(3) and Fig. S2A). We observed synergistic effects on receptor upregulation when TNFα or IFNγ treatment was combined with LPS. An analysis of transcription factor binding sites revealed an interferon-stimulated response element (ISRE) upstream of the CCRL2 promoter, which is conserved among mammals and may be responsible for the IFNγ- mediated regulation of gene expression in macrophages (Fig. S2B). LPS and IFNγ did not influence CCRL2 expression on similarly treated peritoneal mast cells (data not shown). 169

The constitutive expression of CCRL2 on freshly isolated mast cells and the regulated

expression of CCRL2 on macrophages indicates that there can be cell type-specific regulation of CCRL2 gene expression. Furthermore, it is interesting to note that CCRL2 and CMKLR1 on macrophages are reciprocally regulated (e.g LPS causes downregulation of CMKLR1 but upregulation of CCRL2 ( (23) and Fig. S2A). Our data indicate that at some point prior to 24 h after activation by TLR ligands or pro- inflammatory cytokines, macrophages can concurrently express both chemerin receptors

CMKLR1 and CCRL2. The functional consequence of this is under investigation.

CCRL2 may play distinct physiologic roles depending on the cellular context of its expression. In a general survey of mouse tissues, mCCRL2 is expressed at the RNA level in the lung, heart, and spleen, with minimal expression in the brain or thymus (Fig.

S12). CCRL2 expression by stromal cells in the lung and heart may serve to maintain a

reservoir of chemerin and to buffer tissue levels of the attractant in vivo, in the same manner that erythrocyte-expressed ‘atypical’ chemokine receptor DARC can serve as a buffer to maintain chemokine levels in the blood (14, 15). We are currently investigating this hypothesis by comparing chemerin protein levels in blood and tissue extracts from

WT and CCRL2 KO mice.

Transfectant-expressed CCRL2 has been reported to bind CCL2 and CCL5,

although this result is controversial (9). CCL2 did not interfere with anti-mCCRL2 mAb

staining in our studies. In addition, using either the binding conditions published in this report for chemerin, or those published by Biber et al. (9) for CCL2, we were unable to

170

confirm the binding of biotinylated CCL2 to CCRL2/L1.2 transfectants (data not shown).

However, it is unclear whether the biotinylated CCL2 reagent available to us is useful for identifying cognate receptor expression, since we observed similar high background

binding of this reagent to T cells in WT and CCR2 KO mice.

The identification of natural ligands for ‘silent’ or non-signaling receptors (such as D6, DARC, and CCX-CKR) has lagged behind the de-orphaning of signaling receptors. This is likely due to an intrinsic bias towards GPCR:ligand screening methods

that rely on measuring functional responses (e.g., calcium mobilization or chemotaxis),

rather than binding per se, to identify ‘hits’. While there are >100 orphan heptahelical

GPCR sequences that have been identified via genome sequencing (48), the phylogenetic

homology of CCRL2 with CC chemokine receptors, and its expression at the RNA level

by LPS activated macrophages, suggested that CCRL2 may be a modulator of leukocyte

trafficking. Indeed, we show herein that the expression of CCRL2 by mast cells can

enhance the numbers of leukocytes in the dermis at sites of IgE-dependent cutaneous

inflammation (Fig. 4D). The identification of chemerin as a non-signaling ligand for

CCRL2 introduces a novel functionality for ‘atypical’ receptors (i.e., concentration and

presentation, as we have shown herein for CCRL2, as opposed to

internalization/degradation). Our findings expand the possible ligand space for ‘atypical’

orphan receptors to include chemoattractants beyond the chemokine family, and provide

a potential link between CCRL2 expression and chemerin-dependent effects on

inflammation that are mediated via the cell-signaling chemerin receptor CMKLR1.

171

Acknowledgements

We thank Mindy Tsai, Kareem Graham, and J. Zabel for helpful discussions. We thank

Dan Dairaghi, Tim Sullivan, Niky Zhao, and Susanna Lewen for assistance with the radioligand binding assays. E.C.B., B.A.Z., T.O., and J.P. are supported by grants from

the National Institutes of Health (AI-59635, AI-47822, and GM-37734); Specialized

Center of Research Grant HL-67674; Digestive Disease Center Grant DK-56339; and a

Merit Award from the Veterans Administration to E.C.B. S.J.G., S.N., and H.S. are supported by grants from the National Institutes of Health (HL-67674, AI-23990, AI-

070813, and CA-72074) to S.J.G. T.M.H. is supported by grants from the National

Institutes of Health (AI37113-09) and the California HIV/AIDS Research Program

(ID06-SD-206). L.Z. is support by a National Institutes of Health pre-doctoral fellowship

(AI073198). J.Y.K. was supported by the Serono Foundation and the Cancer Research

Institute. C.A. was supported by the Deutsche Forschungsgemeinschaft. S.J.A. is supported by a California HIV/AIDS Research Program fellowship award (TF06-SD-

501). The authors have no conflicting financial interests.

172

Supplemental Figures

Figure S1. Blood lymphocytes, BM neutrophils, and peritoneal macrophages do not detectably express mCCRL2. A representative data set of the 3 experiments, each of which gave similar results, is shown.

173

Figure S2. mCCRL2 is upregulated on macrophages activated by specific cytokines and/or TLR ligands. (A) Freshly isolated peritoneal macrophages were cultured for 24 h with various stimuli as indicated. A representative data set of the 3 experiments performed, each of which gave similar results, is shown. (B) The promoter regions of CCRL2 contain interferon- stimulated response element (ISRE) sequences that are conserved across species.

174

Figure S3. mCCRL2 KO mice display a normal contact hypersensitivity response to FITC. Mice were sensitized by application of 2% FITC (suspended in acetone-dibutyl phthalate) to the shaved abdomen. Five d later, the mice were challenged by application of 0.5% FITC to the left ear, or vehicle alone to the right ear. Ear swelling was measured at the indicated time points. N = 7 KO, n = 3 WT, 2 het, mean ± SEM.

175

Figure S4. mCCRL2 is dispensable for maximal tissue swelling in high dose IgE- mediated passive cutaneous anaphylaxis. (A) Mice were sensitized by injection of 150 ng anti-DNP IgE into left ear skin (with vehicle injection into right ear skin as the control). The mice were challenged by i.v. injection of DNP-HSA (200 μg/mouse) the next day, and ear swelling was measured at

176

the indicated time points, mean ± SEM, n = 2 experiments (12 total KO and WT mice per group); NS, not significant (p > 0.05) by ANOVA comparing swelling in WT vs. KO ears sensitized with antigen specific IgE. (B) The ears of mast cell deficient KitW-sh/Wsh mice were engrafted with bone marrow-derived cultured mast cells from either WT or mCCRL2 KO mice. 6-8 weeks later, the mice were sensitized (50 ng IgE), challenged (200 μg DNP-HSA), and monitored as described in (A). mean ± SEM, 5 total mice per group; NS, not significant (p > 0.05) by ANOVA comparing swelling in mCCRL2 KO vs. WT BMCMC reconstituted ears sensitized with antigen specific IgE. (C) Numbers of leukocytes per mm2 of dermis, assessed in formalin-fixed, paraffin-embedded, hematoxylin and eosin-stained sections of mice from (B). *p<0.03 and †† p < 0.01 by the Mann Whitney U-test (2-tailed) vs. the corresponding values for the vehicle-injected ears in the corresponding WT BMCMC- or KO BMCMC-engrafted KitW-sh/Wsh mice, respectively. The numbers over the bars for vehicle-injected mice are the mean values.

Figure S5. Histologic features of high dose IgE-dependent PCA reactions in WT BMCMC- vs. KO BMCMC-engrafted KitW-sh/Wsh mice. Histological sections of ear skin from WT BMCMC-engrafted KitW-sh/Wsh mice (A-C) and KO BMCMC-engrafted KitW-sh/Wsh mice (D-F) from the same group shown in Figure S4C show no evidence of inflammation in ears analyzed 6 h after injection of vehicle (A,D), but evidence of tissue swelling and increased numbers of leukocytes, consisting predominantly of polymorphonuclear leukocytes (some indicated by arrowheads in (C,F) and occasional mononuclear cells (indicated by an arrow in (F)), at 6 h after antigen challenge in both WT BMCMC-engrafted KitW-sh/Wsh mice (B, C) and KO BMCMC- engrafted KitW-sh/Wsh mice (E,F). Hematoxylin and eosin stain; scale bars = 50. 177

Figure S6. mCCRL2/L1.2 transfectants do not migrate to CCL2, CCL5, CCL7, or CCL8 in in vitro transwell chemotaxis. Left panel. Mouse CD11b+ peritoneal cells were used as positive controls to demonstrate functional activity of the chemokines tested. Right panel. mCCRL2/L1.2 cells were tested for chemotactic responses to a range of doses of the indicated chemokines. CCL19/CXCL12 were used as a positive control to demonstrate functional migratory responses by mCCRL2/L1.2 cells (through endogenous expression of CCR7 and CXCR4 by L1.2 cells). A representative experiment (mean ± range of duplicate wells) of the 3 performed, each of which gave similar results, is shown.

178

Figure S7. Lack of heterologous displacement of chemerin by other chemoattractants. mCCRL2/L1.2 transfectants were co-incubated with 10 nM tagged chemerin and 100- fold excess untagged chemoattractants. Secondary anti-His6 PE was used to detect levels of bound chemerin. The horizontal bar at MFI = 77 indicates the fluorescence intensity of cells incubated with tagged chemerin/secondary anti-HIS8 PE in the absence of untagged attractants. The mean ± SEM of n=3 experiments is displayed. * p <0.005 comparing the MFI of cells incubated in the presence of tagged chemerin ± untagged chemerin.

179

Figure S8. Radioligand binding competition. (A) Displacement of iodinated chemerin (residues 21-148) binding to mCMKLR1 and mCCRL2 by His8-tagged chemerin. (B) Displacement of iodinated chemerin (residues 21-148) binding to mCMKLR1 and mCCRL2 by bioactive carboxyl-terminal chemerin peptide (YFPGQFAFS). Regression analysis of the binding in part (B) to mCCRL2 failed 2 2 to fit a curve to the data with R >0.8 (R =0.66); thus the EC50 could not be determined (N.D.). A representative experiment (mean ± SD of triplicate wells) of the 3 performed, each of which gave similar results, is shown for each part.

180

Figure S9. Chemerin and/or CCL2 do not trigger intracellular calcium mobilization in CCRL2/HEK293 transfectants. 181

(A and B) mCCRL2 HEK293 transfectants were loaded with Fluo-4, sequentially treated with chemerin, CCL2, or ionomycin at the indicated times, and examined for intracellular calcium mobilization. (C) THP1 cells were loaded with Fluo-4, treated with CCL2 at the indicated time, and examined for intracellular calcium mobilization. (D) mCMKLR1 HEK293 transfectants were loaded with Fluo-4, treated with chemerin at the indicated time, and examined for intracellular calcium mobilization. A representative data set of the 3 experiments performed, each of which gave similar results, is shown for all parts of this figure.

Figure S10. CCRL2 amino-terminal sequence alignment. The predicted amino-terminal domains of mouse and human CCRL2 were aligned using Clustal W.

Figure S11. Freshly isolated peritoneal mast cells do not express CMKLR1. Freshly isolated peritoneal leukocytes were harvested and mCMKLR1 expression was evaluated on SSChi F4/80- c-kit+ mast cells. One representative data set of the at least 3 experiments, each of which gave similar results, is shown.

182

Figure S12. mRNA expression of mCCRL2. A mouse RNA array was probed with mCCRL2 cDNA.

183

References

1. Fan, P., H. Kyaw, K. Su, Z. Zeng, M. Augustus, K. C. Carter, and Y. Li. 1998. Cloning and characterization of a novel human chemokine receptor. Biochem Biophys Res Commun 243:264-268. 2. DeVries, M. E., A. A. Kelvin, L. Xu, L. Ran, J. Robinson, and D. J. Kelvin. 2006. Defining the origins and evolution of the chemokine/chemokine receptor system. J Immunol 176:401-415. 3. Shimada, T., M. Matsumoto, Y. Tatsumi, A. Kanamaru, and S. Akira. 1998. A novel lipopolysaccharide inducible C-C chemokine receptor related gene in murine macrophages. FEBS Lett 425:490-494. 4. Brouwer, N., M. W. Zuurman, T. Wei, R. M. Ransohoff, H. W. Boddeke, and K. Biber. 2004. Induction of glial L-CCR mRNA expression in spinal cord and brain in experimental autoimmune encephalomyelitis. Glia 46:84-94. 5. Zuurman, M. W., J. Heeroma, N. Brouwer, H. W. Boddeke, and K. Biber. 2003. LPS-induced expression of a novel chemokine receptor (L-CCR) in mouse glial cells in vitro and in vivo. Glia 41:327-336. 6. Oostendorp, J., M. N. Hylkema, M. Luinge, M. Geerlings, H. Meurs, W. Timens, J. Zaagsma, D. S. Postma, H. W. Boddeke, and K. Biber. 2004. Localization and enhanced mRNA expression of the orphan chemokine receptor L-CCR in the lung in a murine model of ovalbumin-induced airway inflammation. J Histochem Cytochem 52:401-410. 7. Migeotte, I., J. D. Franssen, S. Goriely, F. Willems, and M. Parmentier. 2002. Distribution and regulation of expression of the putative human chemokine receptor HCR in leukocyte populations. Eur J Immunol 32:494-501. 8. Galligan, C. L., W. Matsuyama, A. Matsukawa, H. Mizuta, D. R. Hodge, O. M. Howard, and T. Yoshimura. 2004. Up-regulated expression and activation of the orphan chemokine receptor, CCRL2, in rheumatoid arthritis. Arthritis Rheum 50:1806-1814. 9. Biber, K., M. W. Zuurman, H. Homan, and H. W. Boddeke. 2003. Expression of L-CCR in HEK 293 cells reveals functional responses to CCL2, CCL5, CCL7, and CCL8. J Leukoc Biol 74:243-251. 10. Mantovani, A., R. Bonecchi, and M. Locati. 2006. Tuning inflammation and immunity by chemokine sequestration: decoys and more. Nat Rev Immunol 6:907-918. 11. Comerford, I., W. Litchfield, Y. Harata-Lee, R. J. Nibbs, and S. R. McColl. 2007. Regulation of chemotactic networks by 'atypical' receptors. Bioessays 29:237-247. 12. Haraldsen, G., and A. Rot. 2006. Coy decoy with a new ploy: interceptor controls the levels of homeostatic chemokines. Eur J Immunol 36:1659-1661. 13. Weber, M., E. Blair, C. V. Simpson, M. O'Hara, P. E. Blackburn, A. Rot, G. J. Graham, and R. J. Nibbs. 2004. The chemokine receptor D6 constitutively traffics to and from the cell surface to internalize and degrade chemokines. Mol Biol Cell 15:2492-2508. 14. Hadley, T. J., and S. C. Peiper. 1997. From malaria to chemokine receptor: the emerging physiologic role of the Duffy blood group antigen. Blood 89:3077-3091. 184

15. Darbonne, W. C., G. C. Rice, M. A. Mohler, T. Apple, C. A. Hebert, A. J. Valente, and J. B. Baker. 1991. Red blood cells are a sink for , a leukocyte chemotaxin. J Clin Invest 88:1362-1369. 16. Kashiwazaki, M., T. Tanaka, H. Kanda, Y. Ebisuno, D. Izawa, N. Fukuma, N. Akimitsu, K. Sekimizu, M. Monden, and M. Miyasaka. 2003. A high endothelial venule- expressing promiscuous chemokine receptor DARC can bind inflammatory, but not lymphoid, chemokines and is dispensable for lymphocyte homing under physiological conditions. Int Immunol 15:1219-1227. 17. Middleton, J., S. Neil, J. Wintle, I. Clark-Lewis, H. Moore, C. Lam, M. Auer, E. Hub, and A. Rot. 1997. Transcytosis and surface presentation of IL-8 by venular endothelial cells. Cell 91:385-395. 18. Lee, J. S., C. W. Frevert, M. M. Wurfel, S. C. Peiper, V. A. Wong, K. K. Ballman, J. T. Ruzinski, J. S. Rhim, T. R. Martin, and R. B. Goodman. 2003. Duffy antigen facilitates movement of chemokine across the endothelium in vitro and promotes neutrophil transmigration in vitro and in vivo. J Immunol 170:5244-5251. 19. Comerford, I., S. Milasta, V. Morrow, G. Milligan, and R. Nibbs. 2006. The chemokine receptor CCX-CKR mediates effective scavenging of CCL19 in vitro. Eur J Immunol 36:1904-1916. 20. Zabel, B. A., A. M. Silverio, and E. C. Butcher. 2005. Chemokine-like receptor 1 expression and chemerin-directed chemotaxis distinguish plasmacytoid from myeloid dendritic cells in human blood. J Immunol 174:244-251. 21. Meder, W., M. Wendland, A. Busmann, C. Kutzleb, N. Spodsberg, H. John, R. Richter, D. Schleuder, M. Meyer, and W. G. Forssmann. 2003. Characterization of human circulating TIG2 as a ligand for the orphan receptor ChemR23. FEBS Lett 555:495-499. 22. Wittamer, V., J. D. Franssen, M. Vulcano, J. F. Mirjolet, E. Le Poul, I. Migeotte, S. Brezillon, R. Tyldesley, C. Blanpain, M. Detheux, A. Mantovani, S. Sozzani, G. Vassart, M. Parmentier, and D. Communi. 2003. Specific Recruitment of Antigen- presenting Cells by Chemerin, a Novel Processed Ligand from Human Inflammatory Fluids. J Exp Med 198:977-985. 23. Zabel, B. A., T. Ohyama, L. Zuniga, J. Y. Kim, B. Johnston, S. J. Allen, D. G. Guido, T. M. Handel, and E. C. Butcher. 2006. Chemokine-like receptor 1 expression by macrophages in vivo: Regulation by TGF-beta and TLR ligands. Exp Hematol 34:1106- 1114. 24. Parolini, S., A. Santoro, E. Marcenaro, W. Luini, L. Massardi, F. Facchetti, D. Communi, M. Parmentier, A. Majorana, M. Sironi, G. Tabellini, A. Moretta, and S. Sozzani. 2007. The role of chemerin in the colocalization of NK and dendritic cell subsets into inflamed tissues. Blood 109:3625-3632. 25. Wittamer, V., B. Bondue, A. Guillabert, G. Vassart, M. Parmentier, and D. Communi. 2005. Neutrophil-mediated maturation of chemerin: a link between innate and adaptive immunity. J Immunol 175:487-493.

185

26. Busmann, A., M. Walden, M. Wendland, C. Kutzleb, W. G. Forssmann, and H. John. 2004. A three-step purification strategy for isolation of hamster TIG2 from CHO cells: characterization of two processed endogenous forms. J Chromatogr B Analyt Technol Biomed Life Sci 811:217-223. 27. Zabel, B. A., S. J. Allen, P. Kulig, J. A. Allen, J. Cichy, T. M. Handel, and E. C. Butcher. 2005. Chemerin activation by serine proteases of the coagulation, fibrinolytic, and inflammatory cascades. J Biol Chem 280:34661-34666. 28. Zabel, B. A., L. Zuniga, T. Ohyama, S. J. Allen, J. Cichy, T. M. Handel, and E. C. Butcher. 2006. Chemoattractants, extracellular proteases, and the integrated host defense response. Exp Hematol 34:1021-1032. 29. Galli, S. J., S. Nakae, and M. Tsai. 2005. Mast cells in the development of adaptive immune responses. Nat Immunol 6:135-142. 30. Kraft, S., S. Rana, M. H. Jouvin, and J. P. Kinet. 2004. The role of the FcepsilonRI beta-chain in allergic diseases. Int Arch Allergy Immunol 135:62-72. 31. Tono, T., T. Tsujimura, U. Koshimizu, T. Kasugai, S. Adachi, K. Isozaki, S. Nishikawa, M. Morimoto, Y. Nishimune, S. Nomura, and et al. 1992. c-kit Gene was not transcribed in cultured mast cells of mast cell-deficient Wsh/Wsh mice that have a normal number of erythrocytes and a normal c-kit coding region. Blood 80:1448-1453. 32. Ponath, P. D., S. Qin, T. W. Post, J. Wang, L. Wu, N. P. Gerard, W. Newman, C. Gerard, and C. R. Mackay. 1996. Molecular cloning and characterization of a human eotaxin receptor expressed selectively on eosinophils. J Exp Med 183:2437-2448. 33. Palermo, D. P., M. E. DeGraaf, K. R. Marotti, E. Rehberg, and L. E. Post. 1991. Production of analytical quantities of recombinant proteins in Chinese hamster ovary cells using sodium butyrate to elevate gene expression. J Biotechnol 19:35-47. 34. Wershil, B. K., Z. S. Wang, J. R. Gordon, and S. J. Galli. 1991. Recruitment of neutrophils during IgE-dependent cutaneous late phase reactions in the mouse is mast cell-dependent. Partial inhibition of the reaction with antiserum against tumor necrosis factor-alpha. J Clin Invest 87:446-453. 35. Bauer, M., V. Redecke, J. W. Ellwart, B. Scherer, J. P. Kremer, H. Wagner, and G. B. Lipford. 2001. Bacterial CpG-DNA triggers activation and maturation of human CD11c-, CD123+ dendritic cells. J Immunol 166:5000-5007. 36. Livak, K. J., and T. D. Schmittgen. 2001. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 25:402-408. 37. Liu, F. T., J. W. Bohn, E. L. Ferry, H. Yamamoto, C. A. Molinaro, L. A. Sherman, N. R. Klinman, and D. H. Katz. 1980. Monoclonal dinitrophenyl-specific murine IgE antibody: preparation, isolation, and characterization. J Immunol 124:2728- 2737. 38. Nakae, S., H. Suto, M. Kakurai, J. D. Sedgwick, M. Tsai, and S. J. Galli. 2005. Mast cells enhance T cell activation: Importance of mast cell-derived TNF. Proc Natl Acad Sci U S A 102:6467-6472.

186

39. Suto, H., S. Nakae, M. Kakurai, J. D. Sedgwick, M. Tsai, and S. J. Galli. 2006. Mast cell-associated TNF promotes dendritic cell migration. J Immunol 176:4102-4112. 40. Nakae, S., H. Suto, M. Iikura, M. Kakurai, J. D. Sedgwick, M. Tsai, and S. J. Galli. 2006. Mast cells enhance T cell activation: importance of mast cell costimulatory molecules and secreted TNF. J Immunol 176:2238-2248. 41. Takeshita, K., T. Yamasaki, S. Akira, F. Gantner, and K. B. Bacon. 2004. Essential role of MHC II-independent CD4+ T cells, IL-4 and STAT6 in contact hypersensitivity induced by fluorescein isothiocyanate in the mouse. Int Immunol 16:685- 695. 42. Wershil, B. K., T. Murakami, and S. J. Galli. 1988. Mast cell-dependent amplification of an immunologically nonspecific inflammatory response. Mast cells are required for the full expression of cutaneous acute inflammation induced by phorbol 12- myristate 13-acetate. J Immunol 140:2356-2360. 43. Martin, T. R., T. Takeishi, H. R. Katz, K. F. Austen, J. M. Drazen, and S. J. Galli. 1993. Mast cell activation enhances airway responsiveness to methacholine in the mouse. J Clin Invest 91:1176-1182. 44. Bulanova, E., V. Budagian, Z. Orinska, M. Hein, F. Petersen, L. Thon, D. Adam, and S. Bulfone-Paus. 2005. Extracellular ATP induces cytokine expression and through P2X7 receptor in murine mast cells. J Immunol 174:3880-3890. 45. Lau, E. K., S. Allen, A. R. Hsu, and T. M. Handel. 2004. Chemokine-receptor interactions: GPCRs, glycosaminoglycans and viral chemokine binding proteins. Adv Protein Chem 68:351-391. 46. Hata, D., Y. Kawakami, N. Inagaki, C. S. Lantz, T. Kitamura, W. N. Khan, M. Maeda-Yamamoto, T. Miura, W. Han, S. E. Hartman, L. Yao, H. Nagai, A. E. Goldfeld, F. W. Alt, S. J. Galli, O. N. Witte, and T. Kawakami. 1998. Involvement of Bruton's tyrosine kinase in FcepsilonRI-dependent mast cell degranulation and cytokine production. J Exp Med 187:1235-1247. 47. Fleming, T. J., E. Donnadieu, C. H. Song, F. V. Laethem, S. J. Galli, and J. P. Kinet. 1997. Negative regulation of Fc epsilon RI-mediated degranulation by CD81. J Exp Med 186:1307-1314. 48. Wise, A., S. C. Jupe, and S. Rees. 2004. The identification of ligands at orphan G- protein coupled receptors. Annu Rev Pharmacol Toxicol 44:43-66.

187

CHAPTER 5 : IL-17 REGULATES ADIPOGENESIS, GLUCOSE

HOMEOSTASIS, AND OBESITY.

Luis A. Zúñiga†, Wen-Jun Shen‡, Barbara Joyce-Shaikh§, Ekaterina A. Pyatnova†,

Andrew G. Richards†, Colin Thom†, Sofia M. Andrade†, Daniel J. Cua§, Fredric B.

Kraemer‡, Eugene C. Butcher†

†Laboratory of Immunology and Vascular Biology, Department of Pathology, Stanford

University School of Medicine, Stanford, CA 94305, USA, and Center for Molecular

Biology and Medicine, Veterans Affairs Palo Alto Health Care System, Palo Alto, CA

94304, USA.

‡Veterans Affairs Palo Alto Health Care System, Palo Alto, CA 94304, USA, and Division of Endocrinology, Department of Medicine, Stanford University, Stanford, CA 94305,

USA.

§MERCK, Palo Alto, CA 94304, USA.

188

Summary

Inflammatory mediators have the potential to impact a surprising range of diseases,

including obesity and its associated metabolic syndrome. Here we show that the

proinflammatory cytokine IL-17 inhibits adipogenesis, moderates adipose tissue

accumulation, and regulates glucose metabolism in mice. IL-17 deficiency enhances diet-

induced obesity in mice and accelerates adipose tissue accumulation even in mice fed a

low fat diet. In addition to potential systemic effects, IL-17 is expressed locally in

adipose tissue by leukocytes, predominantly by γδ T cells. IL-17 suppresses adipocyte

differentiation from mouse derived 3T3-L1 preadipocytes in vitro, and inhibits

expression of genes encoding pro-adipogenic transcription factors, adipokines, and

molecules involved in lipid and glucose metabolism. IL-17 also acts on differentiated

adipocytes, impairing glucose uptake; and young IL-17 deficient mice show enhanced

glucose tolerance and insulin sensitivity. Our findings implicate IL-17 as a negative

regulator of adipogenesis and glucose metabolism in mice, and show that it delays the

development of obesity.

189

Introduction

Obesity is associated with a spectrum of metabolic abnormalities including insulin

resistance, hyperglycemia, high blood pressure, and dyslipidaemia—collectively known

as the metabolic syndrome—which are important risk factors for cardiovascular disease

and type 2 diabetes (1). The development of obesity and of its associated metabolic abnormalities is dependent on genetic factors, lifestyle, and diet, but the causative connection between obesity and the metabolic syndrome remains incompletely understood. An emerging consensus, however, suggests that inflammation links obesity and metabolic dysfunction.

The development of adipocytes (adipogenesis) and their function in lipid storage and metabolism are regulated in a complex fashion by hormones, lipids, sugar

metabolites (2, 3), and adipokines (e.g. leptin, resistin, adiponectin, and chemerin) (4, 5);

but also by inflammatory cytokines. For instance, IFNγ (6, 7) and IL-6 (8) inhibit

adipogenesis, and TNFα induces lipolysis in mature adipocytes (9). Moreover, circulating

levels of inflammatory proteins including serum amyloid A, IL-6, IL-8, IL-18, MIF and

TNFα are elevated in obese subjects (10) and are thought to contribute to insulin

resistance and to low grade systemic inflammation (11). Indeed, obesity is linked not only to heart disease (12) and diabetes (13, 14, 15), but also to an increased incidence of

inflammatory diseases including psoriasis (16), multiple sclerosis in humans (17) and

EAE in mice (18), and asthma (19). Recent studies implicate the proinflammatory

cytokine IL-17 in several of these diseases as well (20, 21, 22, 23, 24). Interestingly,

190

serum IL-17 is upregulated in obese human patients (25) and obesity is positively

correlated with enhanced IL-17 expression and increased severity of inflammation in IL-

17-dependent mouse models of EAE and trinitrobenzene sulfonic acid-induced colitis

(18). Although these studies suggest a link between obesity, IL-17, and inflammatory

diseases, it is currently unclear whether IL-17 influences metabolism and obesity, or is

merely enhanced by obesity-associated inflammatory changes.

Adipocytes produce a number of cytokines in obesity, but adipose tissue (AT) is

also infiltrated by immune cells that produce inflammatory and regulatory factors. For

example, macrophages accumulate in obese adipose tissue, where they secrete TNFα, IL-

6, IL-1, MCP-1, and MIP-1α (26, 27, 28). Cytokine producing T cells also infiltrate

adipose tissue. Interestingly, a shift from TH2 (IL-4-producing) and regulatory T cells

(Tregs) to a pathogenic TH1 (IFNγ-producing) T cell infiltration occurs during the

progression of diet-induced obesity (7). While pathogenic TH1 cells and CD8 T cells increase with obesity, Tregs, which have positive effects on metabolic parameters in part

through production of IL-10, decline in frequency (7, 29, 30). The cytokines produced by

infiltrating leukocytes, including T cells, are thought to play significant pathogenic and

regulatory roles in adipose biology and metabolism (31). γδ+ T cells, a minor T cell

subset that often responds to self-ligands (32, 33, 34) and participates in

immunoregulation (35), also increases significantly in inguinal adipose tissue during

obesity (36, 37); but cytokine expression by adipose γδ T cells has not been examined.

191

In this report, we show that T cells in adipose tissue produce IL-17, and we present evidence that IL-17 is an important regulator of adipogenesis and glucose metabolism. In adipose tissue, IL-17 is produced primarily by γδ T cells, and the majority of adipose tissue γδ T cells express IL-17. IL-17 acts on preadipocytes and adipocytes to inhibit adipogenesis and moderate lipid and glucose uptake, and IL-17 deficient mice develop more severe adult onset obesity and display altered glucose homeostasis. Our results identify IL-17 as an adipose tissue-associated cytokine that regulates adipocyte biology and metabolism.

192

Materials and Methods

Animals.

Animal studies were performed in accordance to NIH guidelines and in accordance to

guidelines set forth by Stanford University. Animal protocols were approved by the VA

Institutional Animal Care and Use Committee. Age matched, male mice were used for all studies. C57BL/6NCr IL-17 deficient and wild-type C57BL/6NCr (National Cancer

Institute) control mice were housed in accordance to NIH guidelines. C57BL/6J TCRδ

KO and age matched WT control mice were purchased from Jackson Laboratories. Mice

were maintained on a twelve-hour light/dark cycle and given free access to food and

water unless otherwise indicated. 6-8 week old mice were fed either a 10% fat or a 60%

fat diet (Teklad) and weighed weekly. Food was also weighed weekly. Internal core temperatures were measured with a Traceable® Thermometer and probe (Fisher

Scientific). Body composition analysis was determined by dual energy x-ray absorptiometry (DEXA) using a Discovery model DEXA scanner adapted for rodent imaging (Hologic). Calibration was performed before each set of measurements. The animals were anaesthetized (i.p. injection of ketamine/xylazine) prior to scanning, and data was obtained according to manufacturer’s protocols. Mice were fasted for 18 hours prior to sacrifice. Tissues were collected, weighed as indicated, and processed for leukocyte isolation. Blood was collected via cardiac puncture, allowed to clot for 30 minutes, centrifuged for 10 minutes at 5000 RPM, and the serum fraction was isolated and stored at -80°C for later analyses.

193

Tissue and cell isolation.

Spleen and inguinal lymph nodes were carefully excised from fatty deposits and passed through 40 μm cell strainers (Fisher) yielding single cell suspensions. Lymph nodes were

first removed from inguinal adipose depots. Inguinal and epididymal adipose depots were carefully excised, weighed, minced and placed into DMEM with 10 mg/mL bovine serum albumin (BSA) Fraction V (Sigma-Aldrich), centrifuged at 4°C for 5 minutes at 1500

RPM, and the adipose fraction was then digested with 0.03 mg/mL Blendzyme 3 (Roche) in the presence of 50 U/mL DNase I (Sigma-Aldrich) in digestion buffer (153 mM NaCl,

5.6 mM KCl, 2.3 mM CaCl2, 2.6 mM MgCl2, 15 mM HEPES, 1% (w/v) BSA, pH 7.4)

for 1 hour at 37°C. Digested samples were passed through a 100 μm cell strainer and

floating cells were collected as the adipocyte fraction and pelleted cells as the SVC

fraction. Isolated cells were resuspended in 40% Percoll (GE) and underlaid with 75%

Percoll. Gradients were centrifuged for 20 minutes at room temperature and the cellular

interface was harvested. Cells were resuspended in cDMEM, counted using a

hemacytometer and trypan blue, and were treated as indicated below.

Fluorescence Activated Cytometry.

1-2 million isolated cells were stimulated in the presence of brefeldin A (eBioscience) for

either 4 hours with 40 ng/mL phorbol 12-myristate 13-acetate and 1 μg/mL ionomycin

(Sigma-Aldrich) or for 8-16 hours without exogenous stimulation. Cells were then

stained with fluorochrome tagged antibodies against cell surface markers (CD3ε, CD4,

CD8α, NK1.1, β TCR, and/or γδ TCR) and intracellular cytokines as indicated (IFNγ and

194

IL-17), following recommended procedures (eBioscience). Data was collected on an

LSRII cytometer using FACSDiva (Becton, Dickinson) software, followed by FlowJo

(Treestar) analysis. Plots are represented using the bi-exponential transformation function

to allow visualization of events close to or below the axes.

In vitro 3T3-L1 adipocyte cultures.

3T3-L1 preadipocytes were seeded in 12, 24, or 48-well culture plates and incubated at

37°C in DMEM with 10% BCS. Two days post-confluency (day 0), media was replaced

with DMEM supplemented with 10% fetal bovine serum (FBS), pen/strep, 400 ng/mL dexamethasone (Sigma-Aldrich), 0.5 mM 3-isobutyl-1-methylxanthine (IBMX, Sigma-

Aldrich), and 0.07 mg/mL bovine insulin (Sigma-Aldrich) to induce differentiation in the presence or absence of a range of recombinant mouse IL-17 (R and D Systems) concentrations. Differentiation media and IL-17 was replaced 48 hours later. 96 hours after the start of differentiation (day 4), the media was removed and replaced with lipid loading media (DMEM plus 10% FBS with pen/strep). Cells were kept in lipid loading media between 2-14 days, with fresh changes of media every 2-3 days. Cells were harvested for RNA, stained for lipid, or used for functional assays as indicated.

Western Blot Analysis

Two day, post-confluent 3T3-L1 preadipocytes were serum starved in DMEM + 1% fatty acid free bovine serum albumin (FAF BSA) for 4-8 hours. Cells were then left untreated or were stimulated with either 1 μg/mL phorbol myristate acetate (PMA) or 100 ng/mL mouse rIL-17 (R and D Systems) in DMEM + 1% FAF BSA for 5, 10, 15, or 30 minutes.

195

Immediately following, cells were washed twice with ice cold PBS and harvested in

RIPA buffer supplemented with protease and phosphatase inhibitors (Thermo Scientific).

Cells were homogenized with a needle and syringe and protein concentration was measured using the BCA method (Pierce). 10 μg of total protein was separated on a 4-

20% SDS-polyacrylamide gel (Biorad) and then transferred to a PVDF membrane

(Amersham). Membranes were blocked for 2 hours at room temperature with 5% nonfat dry milk in PBS and were then incubated overnight at 4°C with the primary antibody in

5% milk in PBS. All primary antibodies were purchased from Cell Signaling Technology and used at a 1:1000 dilution (phospho-Erk1/2 #4370, total Erk1/2 #9102, phospho-AKT

#4058, total AKT #9272, phospho-NF-κB #3033, total NF-κB #4764).

Membranes were then washed with PBS + 0.5% Tween 20 and incubated with goat anti- rabbit secondary antibody conjugated to horseradish peroxidase (Sigma) diluted 1:2000 in 5% milk for 2 hours at room temperature. Membranes were then washed and exposed via enhanced chemiluminescence (Perkin Elmer). Blots were then stripped with stripping buffer (Thermo Scientific) per manufacturer’s recommendations and re-probed with a different primary antibody.

Oil-Red-O staining.

The media from 2 day lipid loaded 3T3-L1 adipocytes was removed from each well and cells were washed with PBS, then distilled water, and fixed with 4% paraformaldehyde for 30 minutes at room temperature. Fixative was aspirated and cells were washed with

PBS and then distilled water. Water was then aspirated and the cells coated with a

196

saturated Oil-Red-O solution in 60% isopropanol. The cells were incubated at room temperature for 50 minutes, after which the staining solution was removed and the cells washed with water followed by a 70% ethanol wash. Ethanol was removed and replaced with PBS followed by light microscopy and imaging.

Triglyceride and Protein Quantification.

Total lipid was extracted from mature 3T3-L1 adipocytes with 3:2 hexane:isopropanol mix. Extracted lipid was then dried under a vacuum and resuspended in absolute ethanol.

Total triglyceride was measured using a triglyceride quantification kit (Sigma) per the supplier’s recommendations. After lipid extraction, total protein from culture wells was resuspended with 1N NaOH. Total protein was determined using BCA reagent (Pierce) following manufacturers recommendations.

Proliferation Assays.

Two day, post-confluent 3T3-L1 preadipocytes (96-well plates) were treated with differentiation media for 24 hours in the presence or absence of a range of IL-17 concentrations. [3H]Thymidine was added for the last 18 hours of culture and thymidine incorporation was assessed using a β-plate scintillation counter.

QPCR.

RNA from cultures was extracted using a Ribopure (Ambion) kit per the supplier’s instructions. Gene expression was determined by quantitative PCR (QPCR) using an

Applied Biosystems 7900HT real-time PCR instrument. Total RNA (5 µg) was subjected to treatment with DNase (Roche Molecular Biochemicals) according to the

197

manufacturer’s instructions to eliminate possible genomic DNA contamination. DNase-

treated total RNA was reverse-transcribed using Superscript II (Invitrogen Life

Technologies) according to manufacturer’s instructions. Primers for peroxisome

proliferative activated receptor γ (PPARγ, Forward: 5’-TCGCTGATGCACTGCCTATG-

3’, Reverse: 5’-GAGAGGTCCACAGAGCTGAAT-3’), Perilipin (Forward: 5’-

ACACTCTCCGGAACACCATC-3’, Reverse: 5’-CCCTCCCTTTGGTAGAGGAG-3’),

Hormone Sensitive Lipase (HSL, Forward: 5’-GCTTGGTTCAACTGGAGAGC-3’,

Reverse: 5'-GCCTAGTGCCTTCTGGTCTG-3’), glucose transporter type 4 (GLUT4,

Forward: 5’-ACTCTTGCCACACAGGCTCT-3', Reverse: 5’-

AATGGAGACTGATGCGCTCT-3’), adipose triglyceride lipase (ATGL, Forward: 5’-

AACACCAGCATCCAGTTCAA-3’, Reverse: 5’-GGTTCAGTAGGCCATTCCTC-3’), and chemokine-like receptor 1 (CMKLR-1, Forward: 5’-

CAAGCAAACAGCCACTACCA-3’, Reverse: 5’-TAGATGCCGGAGTCGTTGTAA-

3’) were designed using Primer Express (Applied Biosystems). Other primers, for which sequences are unavailable, were obtained commercially from Applied Biosystems. Real- time RT-PCR on 10 ng of cDNA from each sample was performed using either of two

methods. In the first method, two gene-specific unlabeled primers were used at 400 nM in

a PerkinElmer SYBR Green real-time quantitative PCR assay using an ABI 5700

instrument (Applied Biosystems). In the second method, two unlabeled primers at 900 nM

each were used with 250 nM FAM-labeled probe (Applied Biosystems) in a TaqMan

real-time quantitative PCR on an ABI 7700 sequence detection system. The absence of

198

genomic DNA contamination was confirmed using primers that recognize the genomic

region of the CD4 promoter. levels were measured in a separate reaction and

used to normalize the data. Using the mean cycle threshold (Ct) value for ubiquitin and

the gene of interest for each sample, the equation 1.8 e (Ct ubiquitin – Ct gene of interest)

x 104 was used to obtain the normalized values.

Glucose uptake.

Differentiated 3T3-L1 adipocytes were washed with PBS and serum starved for 4 hours

in DMEM with 0.5% fatty-acid free (FAF)-BSA. Cells were washed and changed to

DMEM without glucose with 0.5% FAF-BSA with or without 100 ng/mL IL-17 for 1

hour. 1 μM insulin was added to the cells, as indicated, for 15 minutes before adding labelling cocktail containing 2 μCi 3H-deoxyglucose and 5 mM cold deoxyglucose. After

15 minutes of incubation, media was aspirated, cells were washed twice with ice cold

PBS and harvested in Tris-EDTA with 0.1% Triton X-100. Cells were lysed by shaking and vigorous vortexing. Cell lysates were centrifuged and aliquots of the supernatants were assayed for 3H-deoxyglucose in scintillation fluid.

Glucose tolerance and insulin sensitivity tests.

Fasted mice were injected, i.p., with D-glucose dissolved in 0.9% sterile saline (1 or 1.5 g

D-glucose/kg body weight, Sigma-Aldrich) or with human insulin (1 U insulin/kg body

weight, Lilly). Blood glucose concentrations were measured before and 20, 40, 60, and

120 minutes after injection with a Precision Xtra glucometer (Abbott).

Serum Cytokine Analysis.

199

Serum from fasted mice was collected and assayed for insulin, IL-6, and leptin using a

Milliplex Map mouse serum adipokine panel kit (Millipore), per the manufacturer’s recommendations. Data was collected using a Luminex 200 System (Luminex).

Statistics.

Data are expressed as the mean of replicate measurements or mean normalized values between multiple experiments +/- SEM. For comparisons between two groups, statistical significance was evaluated using the Student's t-test (2-tailed, α=0.05) assuming unequal variance. For comparisons across multiple groups (either within a given day/tissue or across multiple days/tissues), one way analysis of variance was used (Holm-Sidak method; SigmaStat 3.0, SPCC Inc.).

200

Results

IL-17 is expressed by γδ T cells in adipose tissue

To determine if adipose tissue contained IL-17-producing T cells, we initially studied leukocytes from the stromal vascular fraction of inguinal (subcutaneous) and epididymal (visceral) adipose depots from mice fed a high fat diet for 18 weeks.

Leukocytes were stimulated with phorbol 12-myristate 13-acetate (PMA) and ionomycin, in the presence of brefeldin A, for 4 hours to induce intracellular accumulation of expressed cytokines. The stimulated cells were stained with antibodies to CD3, CD4,

CD8, and γδ TCR to identify T cell subsets, followed by intracellular staining for IL-17 or interferon γ (IFNγ). Lymphocytes from inguinal lymph node and spleen were analyzed

for comparison.

We found that IL-17-producing T cells are present in adipose tissue (Fig. 1A), and

the great majority of these are CD4-, CD8- T cells that express the γδ T cell receptor (γδ17

cells) (Fig. 1B and data not shown). Although γδ T cells represent only ~4-11% of CD3+

T cells in these tissues, γδ17 T cells comprised between 70-90 % of IL-17 producing

CD3+ cells in inguinal adipose tissue, and between 80-90% in epididymal adipose tissue.

IFNγ is expressed by adipose tissue infiltrating T cells in obesity. We therefore

compared the expression of IL-17 with that of IFNγ in each of the T cell subsets (Fig 1C).

Conventional CD4 and CD8 T cells in adipose tissue contained a high frequency of IFNγ-

producing TH1 and TC1 cells, respectively, but, consistent with recent reports (7), TH17

and TC17 cells were rare. In contrast, the majority of adipose tissue γδ T cells stimulated

201

with PMA and ionomycin expressed IL-17, but relatively few γδ T cells expressed IFNγ, particularly in the subcutaneous (inguinal) adipose tissue. A smaller fraction of IL-17- producing CD3+ T cells are found among the CD4- CD8- and γδ TCR- compartment, which comprise NK T cells and so called double negative αβ T cells; but these are much less frequent than γδ17 T cells in normal mice. Expression of IL-17 and IFNγ is mutually exclusive in all these populations.

To determine whether γδ T cells from adipose tissue spontaneously secrete IL-17, or whether secretion requires exogenous stimulation, tissue leukocytes were incubated in medium in the presence of brefeldin A only. Under these conditions, expression of IL-17 was detectable by flow cytometry in a subset of γδ T cells from both adipose depots of mice fed a high fat diet for 16-18 weeks (Fig. 1D). The frequency of spontaneous IL-17 expression by inguinal adipose tissue γδ T cells increased with age in obese mice as well

(Fig 1E), but remained relatively rare among CD4 or CD8 T cells. In the absence of stimulation, IFNγ was expressed by few, if any, γδ, CD4 or CD8 T cells from secondary lymphoid tissue or adipose tissue from high fat fed mice at 16-18 weeks or 12 months

(data not shown).

202

Figure 1. IL-17 expression by adipose tissue leukocyte subsets. For A-C, leukocytes were isolated from spleen, inguinal lymph node (ILN), inguinal adipose tissue (Ing. AT), and epididymal adipose tissue (Epi. AT) from male mice fed a high fat diet for 18 weeks. Cells were stimulated with phorbol myristate acetate (PMA) and ionomycin for 4 hours in the presence of brefeldin A. Intracellular expression of IL- 17 and IFNγ was assessed on the indicated T cell subsets, defined by staining for CD3, CD4, CD8, and γδ TCR. (A) Evaluation of CD3 and IL-17 expression by tissue leukocytes. (B) Expression of γδ TCR, CD4, or CD8 by CD3+IL-17+ cells gated in (A). (C) Stimulated expression of IL-17 and IFNγ by various tissue T cell subsets. (D-E) Evaluation of spontaneous IL-17 expression: Cells were isolated from male mice fed a high fat diet for 16-18 weeks (D) or 12 months (E) and were cultured for 8-16 hours in medium supplemented with brefeldin A without exogenous stimulation, followed by

203

intracellular staining for IL-17. Data are representative of 2-5 experiments with similar results.

204

Increased number and frequency of IL-17-producing T cells in obese inguinal adipose

tissue

Commitment of effector/memory cells to production of particular cytokines is

usually assessed by short term non-specific stimulation with PMA and ionomycin. We

therefore evaluated PMA and ionomycin stimulated IL-17 expression by T cells in

adipose tissue of mice fed either a low fat (LF; 10% fat) or high fat (HF; 60% fat) diet for

18 weeks. Mice were 6 weeks old at the beginning of the study. Upon sacrifice, the

average weight for mice fed a LF diet was 36.0+/- 0.9 grams (SEM, n = 5) while mice fed

a HF diet weighed an average of 48.8 +/- 2.4 grams (SEM, n = 5). As expected, the

difference in weight between groups primarily reflected adipose tissue mass differences,

as shown by dual energy X-ray absorptiometry (DEXA) analysis (data not shown). Flow

cytometric analysis revealed an increase in the percentage of IL-17-producing T cells in

obese inguinal adipose tissue (Fig. 2A); and this increase was primarily attributable to an

increased frequency and number of γδ17 T cells per gram of adipose tissue (Fig. 2B, C).

There was a significant increase in TH17 cell frequency and number in inguinal adipose tissue as well, but γδ17 cells were consistently 7-8 times more abundant in both LF and

HF inguinal adipose tissue. In contrast to inguinal adipose tissue, epididymal adipose

tissue in obese mice displayed a reduction in the frequency of IL-17-producing T cells,

but this reflected primarily an increase in the number of CD8+ T cells (data not shown), consistent with a recent report demonstrating increased CD8+ T cell content in obese

visceral adipose tissue (30). The actual number per gram of γδ17, TH17 and TC17 cells in

205

epididymal adipose tissue was similar in LF and HF mice (Fig. 2B, C). As in inguinal

adipose depots, γδ T cells were the main producers of IL-17 in epididymal adipose tissue.

No significant differences in IL-17-producing T cells or T cell subsets were observed in

spleens or inguinal lymph nodes (ILN) between groups.

Figure 2. Diet-induced differences in IL-17 expression by tissue T cell subsets. 6 week old male C57Bl/6NCr mice were fed a low fat (LF, 10% fat) or high fat (HF, 60% fat) diet for 3 months. Spleen, inguinal lymph node (ILN), inguinal adipose tissue (Ing. AT), and epididymal adipose tissue (Epi. AT) were harvested, adipose tissue was weighed and tissue cells were analyzed for T cell content and stimulated intracellular IL- 17 as in Figure 1. 5 mice per group were analyzed, and mean values are given with SEM. (A) Percent of tissue CD3+ T cells expressing IL-17. (*p<0.003 vs. LF Ing. AT; **p<0.04 vs. HF Epi. AT; Holm-Sidak multiple comparison test). The mean representation of CD4, CD8, and γδ T cells among the IL-17-expressing pool is indicated within the bars. (B) Frequency of IL-17+ cells of the indicated phenotype, expressed as a percentage of the total CD3+ T cell pool. (γδ: *p<0.004 vs. LF Ing. AT, **p<0.04 vs. HF Epi. AT; CD4: *p<0.001; Holm-Sidak multiple comparison test) (c) Number of recovered IL-17+ cells of the indicated T cell phenotype per gram adipose tissue. (γδ: *p<0.02 vs. LF Ing. AT; CD4: *p<0.001 vs. LF Ing. AT; Student’s T-test). Data are from individual experiments, and are representative of 3 experiments with similar results.

206

207

IL-17 deficient mice are more susceptible to diet-induced obesity

To determine the effect of IL-17 on the development of obesity, we fed 6-7 week old IL-17 wildtype (WT) and IL-17 knockout (KO) mice either a LF or HF diet for 14-18 weeks and measured body mass weekly. IL-17 KO mice reached a significantly higher weight than their wild type cohorts over time in both the low fat and high fat conditions

(Fig. 3A and B, respectively). DEXA analysis indicated differences in body mass were mainly attributable to body adipose tissue composition (Fig. 3C and D). Food consumption and body temperatures were measured and were not detectably different between the groups.

208

Figure 3. Enhanced susceptibility of IL-17 KO mice to dietary obesity. 6-7 week old IL-17 WT (n = 25-30 per group) and IL-17 KO (n = 14-26 per group) male mice were fed a low fat (LF, 10% fat) (A) or high fat (HF, 60% fat) (B) diet and their weights were measured over time. Data are pooled from three independent experiments, each consisting of 4–10 mice per group, and are presented as mean body mass +/- SEM (*p<0.05 between IL-17 WT and KO groups at the indicated time-points; Student’s T- test). Mice from (A) and (B) were analyzed for body adipose tissue mass (C) and lean mass (D) content by dual energy X-ray absorptiometry (DEXA ) after 3 months of 209

feeding. Results are expressed as mean tissue mass +/- SEM. Significance was determined using Student’s T-test, and P-values are represented between IL-17 WT and KO groups. Data for (C) and (D) are from a single experiment, and are representative of 3 individual experiments with similar results.

210

IL-17 treatment inhibits lipid loading and disrupts adipocyte-associated gene expression in 3T3-L1 adipocytes

Based on the expression of IL-17 by adipose tissue T cells and the effect of IL-17 deficiency on diet-induced obesity, we hypothesized that IL-17 might act directly on adipocytes or their progenitors and alter their biology. RT-PCR confirmed that primary adipocytes (Fig. 4A) and 3T3-L1 preadipocytes (data not shown) express IL-17 receptor

A (IL-17RA) mRNA. To assess functional responses, we asked if IL-17 could activate downstream signaling molecules (ERK, Akt and NF-κB) in 3T3-L1 preadipocytes. 2-day post-confluent 3T3-L1 cells were serum starved for 4-8 hours followed by treatment with

IL-17 or PMA. NF-κB p65 was phosphorylated at time zero, and its phosphorylation was not significantly or consistently altered by IL-17 treatment over the time course (data not shown). However, IL-17 treatment led to phosphorylation of AKT and ERK1/2 within 5 minutes (Fig. 4C). The timing of signaling through these pathways helps determines the outcome of adipogenic differentiation and gene expression in a complex and context- dependent manner (38). We conclude that adipocytes express functional IL-17 receptors.

To determine if IL-17 signaling could influence adipocyte differentiation, we exposed 3T3-L1 preadipocytes to adipogenic conditions either in the presence or absence of IL-17. After two days under differentiating conditions, we allowed the resulting cells to take up and accumulate lipid from loading media. The presence of IL-17 during differentiation significantly inhibited subsequent lipid uptake as assessed by triglyceride

(TG) analysis or staining for neutral lipids with Oil-Red-O (Fig. 4C,D). IL-17 did not

211

inhibit the proliferative burst (mitotic clonal expansion) induced by addition of the differentiation media (39) as measured by [3H]thymidine incorporation assays (Fig. 4E) and did not show any evidence of cell toxicity as measured by alamar blue cytotoxicity assays (data not shown).

212

Figure 4. IL-17 inhibits adipogenesis in 3T3-L1 preadipocytes. (A) Expression of IL-17 receptor A mRNA in adipose tissue. Primary inguinal adipose depot adipocyte RNA was isolated and assayed for IL-17RA expression via reverse transcriptase-PCR. E. coli RNA used as a negative control and mouse embryonic fibroblast (MEF) RNA as a positive control. Genomic DNA contamination was assessed in samples without reverse transcriptase added (No RT). (B) IL-17 activates signalling pathways in preadipocytes. Confluent, 4-8 hour serum-starved 3T3-L1 preadipocytes were treated with 100 ng/mL IL-17 for times indicated before being harvested for protein. Phosphorylated and total ERK and AKT were detected by Western blot analysis. (C and D) IL-17 treatment during preadipocyte differentiation inhibits lipid accumulation. Day 0 213

3T3-L1 preadipocytes were treated with differentiation media with or without IL-17 for 2 days. Day 2 Media was replaced with lipid loading media supplemented with insulin only and incubated for 2 days. Finally, day 4 media was replaced with lipid loading media only and allowed to load lipid for an addition two days (C) or six days (D). (C) Two day lipid loaded cells were treated with hexane to extract total lipid and assayed for triglyceride (TG) and protein content (n=6 replicates). Undifferentiated (Undiff) 3T3-L1 cells were grown alongside differentiated cells and are included for comparison. TG data are represented as mg TG per mg protein +/- SEM (*p<0.02 vs. indicated conditions; Holm-Sidak multiple comparison test). Data are representative of 5 independent experiments. (D) Six day post-induction lipid loaded cells were treated with formalin and stained with Oil-Red-O. Representative images for each condition are shown. (E) 3T3-L1 preadipocyte proliferative response to IL-17. Day 0 3T3-L1 preadipocytes were treated with differentiation media with or without IL-17 for 24 hours and assessed for [3H]thymidine incorporation (n=6 replicates). Data are presented as mean counts per minute +/- SEM (p<0.001 vs. all other conditions). Data are representative of 2 independent experiments.

214

In addition to its effect on lipid accumulation, IL-17 suppressed the induction of adipocyte-associated genes in response to differentiation medium, as assessed by qPCR analysis. As expected, in the absence of IL-17, 3T3-L1 cells differentiated well, leading to upregulation of mRNAs for pro-adipogenic transcription factors (C/EBPα, PPARγ), adipocyte related cytokines (leptin, resistin, adipsin, adiponectin, chemerin, and the chemerin receptor CMKLR1), and genes involved in lipid (fatty acid binding protein 4, hormone sensitive lipase, perilipin, adipose triglyceride lipase) and glucose (glucose transporter-4) metabolism (Fig. 5): IL-17 treatment, however, significantly impaired the expression of most of these genes (Fig. 5, A-G). In contrast, IL-17 treatment induced IL-6 mRNA expression, an effect which was rapidly reversed upon removal of IL-17 (Fig.

5H). CMKLR1, resistin, adiponectin, and hormone sensitive lipase (HSL) displayed non- significant downregulated expression with IL-17 treatment.

215

Figure 5. IL-17 inhibits induction of adipocyte genes. On day 0 3T3-L1 preadipocytes were treated with differentiation media with (+) or without (-) 100 ng/mL IL-17 for 2 days (day 2 samples). Media was then replaced with lipid loading media supplemented with insulin only and incubated for 2 days (day 4 samples). Finally, on day 4, media was replaced with lipid loading media only and allowed to load lipid for two more days (day 6). RNA was isolated, for qPCR analysis, either 2, 4, or 6 days after differentiation media was added. CT values were normalized to ubiquitin B values within each sample. To allow pooling of data from different experiments, results for each gene were then normalized to values from 5 day post- confluent undifferentiated 3T3-L1 control cells (included in each experiment). Results are presented as mean gene expression relative to undifferentiated control cells from three independent experiments +/- SEM. Significance was determined using Student’s T-test, and p-values are represented between IL-17 treated and untreated conditions. (PPARγ = peroxisome proliferator-activated receptor γ; CEBPα = CCAAT/enhancer binding protein α; GLUT4 = glucose transporter-4; FABP4 = fatty acid binding protein 4; ATGL = adipose tissue triglyceride lipase; IL-6 = ; CMKLR1 = chemokine-like receptor 1; HSL = hormone sensitive lipase) 216

217

IL-17 inhibits glucose uptake in vitro and impairs glucose and insulin metabolism in

young mice

We next asked if IL-17 could modulate metabolic processes in mature adipocytes.

We focused on glucose uptake, a key indicator of altered metabolism in metabolic syndrome and diabetes. To evaluate glucose uptake in vitro, differentiated 3T3-L1 adipocytes were allowed to load lipid for 10-14 days. Following 4 hour starvation, cells were cultured in glucose-free medium for 1 hour in the presence or absence of 100 ng/mL

IL-17 prior to addition of a labeling cocktail containing [3H]-deoxyglucose with and

without insulin. Insulin stimulation alone led to a 2.5 fold increase in glucose uptake. IL-

17 substantially inhibited this insulin effect, reducing glucose uptake to near basal levels.

IL-17 alone did not significantly alter basal glucose uptake in the absence of insulin (Fig.

6A).

To determine if IL-17 might have a role in glucose metabolism in vivo, we evaluated glucose homeostasis in IL-17 deficient mice. IL-17 KO mice had slightly elevated fasting glucose levels compared to their WT counterparts (Fig 6B and C, insets). In a standard glucose tolerance test, IL-17 KO mice displayed improved glucose clearance compared to IL-17 WT mice, and they were more sensitive to insulin-induced hypoglycemia (Fig. 6B and C). Interestingly, KO mice also had lower basal insulin (Fig.

6D) and IL-6 levels (Fig. 6E), along with higher adiponectin (Fig. 6F). Serum leptin was similar between groups (Fig. 6G). Together, these results implicate IL-17 in the homeostatic regulation of glucose metabolism.

218

Figure 6. IL-17 inhibits glucose uptake in vitro and improves metabolic parameters before the onset of obesity. (A) In vitro glucose uptake: Mature 3T3-L1 adipocytes were allowed to load lipid for 10- 14 days. Cells were serum starved for 4 hours, then washed and switched to glucose free media with or without 100 ng/mL IL-17 for one hour. 1 μM insulin was added to some wells, as indicated, for 15 minutes before adding [3H]-deoxyglucose. After 15 minutes additional incubation, cells were washed and assayed for radiolabeled deoxyglucose content. Results are expressed as mean DPM +/- SEM (*p<0.03 vs. basal, **p<0.001 vs. all other treatments; Holm-Sidak multiple comparison test). (B) and (C) In vivo glucose and insulin challenge. 10 week old low fat diet fed IL-17 WT (open circles, n=9-10) and KO (closed circles, n=9) male mice were fasted for 6 hours and injected i.p. with either 1.5 g glucose/kg body weight (B) or 1 U insulin/kg body weight (C). Blood glucose was measured before and after injection at the times indicated. Results are expressed as mean % initial glucose +/- SEM (*p<0.05 between IL-17 WT and KO groups at indicated time- points; Student’s T-test). Insets represent basal (fasting) glucose levels for each assay, expressed as mean total glucose (mg/dL) +/- SEM (*p<0.04 vs. KO; Student’s T-Test). For (D-G), 10 week old low fat fed IL-17 WT (n=9-10) and KO (n=9) male mice were fasted for 6 hours, and serum was collected and assayed for insulin (D), IL-6 (E), adiponectin (F), and leptin (G). Results are expressed as mean analyte concentration +/- SEM (*p<0.05 vs. KO, **p<0.03 vs. KO; Student’s T-test). Data are from individual experiments, and are representative of 3 experiments with similar results.

219

To determine whether IL-17-dependent effects on metabolism were still

significant in older mice, we performed glucose (Fig. 7A) and insulin (Fig. 7C) tolerance

tests on mice fed a LF or HF diet for 14-16 weeks (~ 6 months of age at the time of the

tests). As expected, mice on the HF diet displayed impaired glucose tolerance and

increased insulin resistance compared to mice fed the LF diet. Surprisingly however, no

significant differences between WT and IL-17 KO cohorts were observed: if anything,

IL-17 deficiency increased rather than decreased insulin tolerance, although this did not

achieve statistical significance. Due to the fact that obese mice generally have

significantly higher basal glucose than lean mice (Fig. 7D, WT mice), we found it

necessary to fast older mice for a longer period of time (18 hours) before performing

glucose tolerance tests in order to gather data that was within the detection limits of the

glucometer used (<500 mg/dL). 18 hour fasted mice showed no significant differences in

basal glucose levels; however, WT mice fed a HF diet trended towards higher basal glucose (Fig. 7B). In contrast to young mice, 6 hour fasted HF IL-17 KO mice had lower

basal glucose than HF WT mice (Fig. 7D). At this age, IL-17 deficient mice fed either

diet were only modestly more obese than their WT cohorts (Fig. 7E). We hypothesize

that the metabolic influence of IL-17 is compromised or overwhelmed by other

pathogenic or physiologic regulatory mechanisms once obesity is established.

220

Figure 7. Obesity with age reverses protection from metabolic syndrome conferred by IL-17 deficiency. For (A) and (C), IL-17 WT (closed circles, n=5 per group) and KO (open circles, n=4-5 per group) mice fed a LF or a HF diet for 14-18 weeks were fasted for 18 hours and injected i.p. with 1 g glucose/kg body weight (A) or were fasted for 6 hours and injected i.p. with 1 U insulin/kg body weight (C). Blood glucose was measured before and after injection at the indicated time points. Results are expressed as mean % initial glucose +/- SEM (*p<0.05 between IL-17 WT and KO groups at indicated time-points). Fasting blood glucose for the GTT (B) and ITT (D) are represented as mean total glucose (mg/mL) +/- SEM (*p<0.05 vs. all other conditions; Holm-Sidak multiple comparison test). (E) Body mass from mice described in (A-D). Results are expressed as mean body

221

mass +/- SEM (Student’s T-test). Data are from individual experiments, and are representative of 2 experiments with similar results.

222

Adipose tissue γδ17 cells are replaced with β TCR positive IL-17 producing T cells in

TCRδ KO mice

To determine whether γδ T cells are essential for the IL-17 effects on adipose

tissue and metabolism, we analyzed TCRδ deficient mice (γδ KO). We hypothesized that

γδ T cell deficiency would mirror IL-17 deficiency, significantly enhancing diet-induced

obesity and reversing the IL-17 inhibition of glucose metabolism in young mice. Unlike

IL-17 deficient mice, however, γδ KO mice showed no significant difference from WT

mice in weight gain (Fig. 8A). In some cohorts, young γδ KO mice, like IL-17 KO mice,

displayed higher basal glucose than their WT cohort, but this was not consistent.

Moreover, γδ KO mice did not differ from WT mice in glucose or insulin tolerance (Fig.

8B and C). Thus γδ T cell deficiency does not recapitulate the metabolic effects of IL-17

deficiency, suggesting the potential for compensatory sources of IL-17.

In fact, flow cytometry of adipose CD3+ T cells revealed a significant population

of IL-17-producing lymphocytes in the adipose tissue of γδ KO mice, despite the

expected absence of γδ TCR+ cells (Fig. 8D and E). These IL-17-expressing T cells are β

+ TCR , and include infrequent TH17 and TC17 cells, but are predominantly CD4 and CD8

negative. They also lack the natural killer cell marker NK1.1. Phenotypically similar IL-

17-expressing T cells are rare but present in WT adipose tissue (Fig. 8D and E). We

conclude that γδ T cell deficiency leads to a compensatory accumulation of a distinct

CD4-CD8- β TCR+ IL-17-expressing adipose tissue population, potentially representing

223

expansion or recruitment of a normally rare double negative IL-17-producing T cell

subset.

Figure 8. Evaluation of diet-induced obesity, glucose tolerance, insulin tolerance, and adipose tissue leukocyte cytokine expression in γδ T cell deficient mice. (A) 6-8 week old C57BL/6J WT (n = 5 per group) and TCRδ KO (n = 5 per group) male mice were fed a high fat (HF, 60% fat) diet and their weights were measured over time. Results are expressed as mean body mass +/- SEM. (B) and (C) In vivo glucose and insulin challenge, respectively. 8-10 week old C57BL/6J WT (WT, open circles, n = 5 per group) and TCRδ KO (KO, closed circles, n = 5 per group) male mice were fed a 60% fat diet. Mice were fasted for 6 hours and injected i.p. with either 1.5 g glucose/kg body weight (B) or 1 U insulin/kg body weight (C). Blood glucose was measured before and after injection at the times indicated. Results are expressed as mean % initial glucose +/- SEM (*p<0.05 between WT and KO groups at indicated time-points; Student’s T-test). Insets represent basal (fasting) glucose levels for each assay, expressed as mean total glucose (mg/dL) +/- SEM (*p<0.04 vs. KO; Student’s T-Test). (D) and (E) Leukocytes were isolated from inguinal adipose tissue (Ing. AT) or epididymal adipose tissue (Epi.

224

AT) from C57BL/6J WT (WT, n = 4-5 per group) and TCRδ KO (γδ KO, n = 4-5 per group) male mice fed a high fat diet for 16-20 weeks. Cells were stimulated with PMA and ionomycin for 4 hours in the presence of brefeldin A. Intracellular expression of IL- 17 was assessed on the indicated T cell subsets, defined by staining for CD3, CD4, CD8, NK1.1, β TCR, and γδ TCR. (D) IL-17+ T cells from the epididymal adipose tissue of WT and TCRδ KO animals were evaluated for expression of β-TCR, γδ TCR, CD4, and CD8. IL-17+ events are represented in black while total CD3+ T cells are represented in gray. Isotype control staining for IL-17 is shown for comparison with anti-IL-17. (E) Total IL-17 expressing T cell load in adipose tissue expressed as mean total IL-17+ T cells per gram of adipose tissue +/- SEM (n = 4-5 per group). Columns also indicate the contribution of specific IL-17+ T cell subsets per gram of adipose tissue, as indicated.

225

Discussion

We have defined a role for IL-17, a T cell-expressed cytokine, in the regulation of body weight, adipocyte differentiation and insulin and glucose homeostasis. We show that in normal mice, IL-17 is expressed by γδ T cells in white adipose tissue, and that it can act directly on adipocytes and their progenitors to impair adipocyte differentiation, glucose and lipid uptake, and insulin sensitivity.

Our findings implicate IL-17 in the regulation of adipogenesis, and in the metabolic functions of mature differentiated adipocytes. IL-17 treatment during in vitro differentiation of adipocytes from precursors significantly reduced the frequency of cells capable of accumulating lipid, consistent with a reduction in mature adipocyte numbers or impairment of adipocyte function. Moreover, IL-17 inhibited the induction of multiple genes whose expression is characteristic of mature adipocytes, including adipokines

(adipsin), and genes involved in lipid (fatty acid binding protein 4, perilipin, adipose triglyceride lipase) and glucose metabolism (glucose transporter-4). Our in vitro findings with mouse derived 3T3-L1 preadipocytes are consistent with recent studies demonstrating IL-17 inhibition of adipogenesis using human mesenchymal stem cells

(40). IL-17 did not significantly inhibit insulin dependent fatty-acid uptake or enhance lipolysis with mature 3T3-L1 adipocytes (data not shown). Inhibition of adipocyte differentiation by IL-17 may contribute to its attenuation of adipose tissue accumulation in LF and HF diet-fed mice.

226

The specific mechanisms by which IL-17 regulates adipogenesis and adipocyte

metabolism will require further study, but a number of mechanisms are likely to

participate. IL-17 induces expression of IL-6 in preadipocytes and, in previous studies, in

fibroblasts (41); and IL-6 is reduced in the serum of IL-17 KO mice. IL-6 is known to induce insulin resistance in vitro and in vivo (15, 42), and it reduces mature-onset obesity in mice (43). IL-17 also inhibited the induction of genes encoding transcription factors

(C/EBPα, PPARγ) that are essential for efficient adipocyte differentiation (44). Inhibition of these genes would negatively impact the differentiation program. IL-17 treatment of preadipocytes also triggered signaling pathways implicated in adipogenesis and adipocyte functions, as evidenced by ERK1/2 and Akt phosphorylation. These pathways regulate in vitro adipogenesis in a complex fashion depending on the stage of differentiation, the cellular environment, and the kinetics of pathway activation (45, 46, 47): As an example, early ERK1/2 activation is important for preadipocyte proliferation, but chronic activation inhibits preadipocyte differentiation to mature adipocytes (38). Inhibition of specific genes required for mature adipocyte function could clearly contribute to IL-17’s effects. For example, reduced expression of the lipid transporter, fatty acid binding protein 4 (FABP4, also known as aP2), is expected to inhibit lipid accumulation (48); and reduced expression of the glucose transporter (glucose transporter 4) may contribute to

IL-17 moderation of glucose uptake. We found no evidence that IL-17 could impair the mitotic clonal expansion that occurs within the first 18 hours of 3T3-L1 differentiation

227

(39), nor could we detect any toxic effect on cell metabolism or survival, indicating the anti-adipogenic effects of IL-17 are unlikely due to reduced cell viability.

Our in vivo studies suggest that IL-17 contributes significantly to systemic glucose homeostasis as well. IL-17 deficiency enhances glucose tolerance and insulin sensitivity in young mice. These systemic effects mirror the improved glucose uptake and insulin responses of adipocytes in the absence of IL-17 in vitro. IL-17 deficiency is also associated with modestly higher fasting glucose levels. Basal hyperglycemia is a hallmark of insulin resistance, a feature of the metabolic syndrome associated with obesity (49). However, the young WT and KO mice used in these glucose and insulin challenges were lean, with no significant differences in mass, indicating that IL-17 contributes to systemic glucose homeostasis even before the onset of obesity. It is interesting to note that IL-17 deficiency was also associated with a significant reduction in serum insulin. Reduced basal insulin levels are often indicative of improved insulin sensitivity, which may contribute to the more efficient glucose metabolism observed; but the modest fasting hyperglycemia observed in IL-17 deficient mice suggests that control of basal insulin secretion may also be affected.

It is important to note that although IL-17 deficient mice display improved metabolic responses before the onset of obesity, they are also more susceptible to accumulating greater adipose tissue mass than WT mice over time. A similar phenomenon is seen in both mice and humans treated with thiazolidinedione PPARγ agonists (e.g. rosiglitazone, pioglitazone, and troglitazone) (50). Mammals treated with

228

thiazolidinediones also, paradoxically, experience improved insulin sensitivity along with enhanced adipose tissue mass accumulation, however the mechanisms that mediate these effects are not fully known. One hypothesis suggests that the pro-adipogenic properties of

glitazones may facilitate adipocyte differentiation, leading to a greater mature adipocyte pool that can act to take up systemic glucose, thus lowering systemic glucose levels and improving glucose tolerance (50). Additionally, thiazolidinediones have anti- inflammatory properties, which may help reduce systemic inflammation in obese patients, thus lowering systemic levels of pro-inflammatory cytokines that can contribute to obesity associated insulin resistance (51). IL-17 deficiency may have similar effects in that it enhances adipocyte differentiation, leading to increased adipose tissue mass accumulation over time, and reduces systemic inflammation, resulting in improved metabolic parameters before the onset of obesity.

Although young IL-17 deficient mice have improved metabolic responses compared to WT mice, protection from the metabolic syndrome in IL-17 KO mice was lost upon the development of age-associated obesity. ~6 month old IL-17 KO mice fed either a low fat or high fat diet were modestly more obese and had similarly impaired

GTT and ITT compared to WT controls. The reasons for this phenomenon are unclear; however it may reflect an advance of other mechanisms of obesity-related inflammation that overwhelm the effects of IL-17 deficiency. As mentioned above, in WT mice, IL-17 may contribute to systemic inflammation before the onset of obesity, thus predisposing them to metabolic syndrome compared to IL-17 deficient mice. IL-17 may not only

229

inhibit adipogenesis, but may also contribute to immune cell recruitment, as IL-17 can

induce the expression of various chemokines important for leukocyte tissue infiltration

(e.g. MCP-1, MCP-3, IP-10, MIG) (52), which may further exacerbate adipose tissue

inflammation. However, as obesity progresses, IL-17 KO mice suffer from accelerated

adipose tissue accumulation and obesity. It is likely that the larger adipose tissue deposits

in IL-17 KO mice still experience the inflammation-inducing insults that occur in WT

adipose tissue (e.g. necrosis, hypoxia) (53, 54, 55, 56), followed by immune cell

infiltration; adipose tissue-associated inflammatory factors (e.g. TNFa, IL-1β, IL-8)

secreted from IL-17 deficient adipose tissue leukocytes could contribute to systemic

insulin resistance, despite the lack of IL-17. Moreover, we cannot rule out possible age-

related changes in systemic responses to IL-17.

We also found higher serum adiponectin in young mice deficient for IL-17,

suggesting that the cytokine inhibits adiponectin expression in vivo as it does in vitro.

Adiponectin is expressed exclusively by mature adipocytes, and its serum concentration

is negatively correlated with obesity (57, 58). Thus its enhanced expression in lean IL-17

KO mice is consistent with a direct role for IL-17 in adipose tissue in vivo. Together with

our in vitro studies, our results suggest that IL-17’s primary influence is in the systemic

regulation of glucose homeostasis prior to development of obesity; and the cytokine has a

beneficial effect in delaying adipose tissue accumulation, likely reflecting local effects on adipocyte precursors.

230

Our discovery of γδ17 T cells in adipose tissue complements recent analyses of cytokine-producing conventional (αβ TCR-expressing) T cell subsets. As confirmed in

our data, while TH17 cells are rare in adipose tissue, IFNγ producing TH1 cells are abundant (6, 7). γδ T cells were observed in adipose tissue as a minor T cell subset that increases significantly in inguinal adipose tissue during obesity (36, 37). Consistent with this, our studies demonstrate an increase in inguinal adipose tissue γδ T cell content (data not shown), however this increase in γδ T cells was not observed in visceral, epididymal adipose tissue. Visceral and subcutaneous adipose tissue also differ in their content of

CD4 helper and regulatory T cells and in the alterations seen in these conventional T cell populations in obesity (7). The reasons for this distinction are not known, but may reflect the fundamental differences in the biology of these adipose tissue deposits (36, 59).

Previous studies have described significant differences in T cell frequencies between inguinal (subcutaneous) vs. epididymal (visceral) adipose tissue. For example, it was shown that, among lymphocytes, the frequency of NK cells is reduced in the epididymal adipose tissue, but not the inguinal adipose tissue, in animals fed a HF diet compared to animals fed a LF diet. Furthermore, αβ T cell frequency is significantly reduced in the inguinal adipose tissue, but not the epididymal adipose tissue, in animals fed a HF diet (36, 37). We also observed significant differences in the two adipose tissue depots: although in both sites γδ T cells are the major IL-17 producing T cell subset. The frequency and absolute number of γδ17 T cells per gram of fat increased substantially with high fat feeding in the inguinal adipose tissue pads, but not in the epididymal

231

adipose deposits. γδ T cells as a whole also increase in frequency in subcutaneous but not

visceral adipose tissue during obesity, as confirmed by our studies (data not shown) and

in previous studies (36, 37). It is not clear why these two adipose tissue depots differ in

γδ17 and other T cell content, but the disparities may reflect differences in the state or extent of spontaneous vs. obesity-induced inflammation between the two adipose tissue depots, or to their differential development or functions in metabolism. Obese visceral adipose tissue is thought to be the main adipose tissue depot responsible for contributing to the metabolic syndrome, and is generally considered to be more inflamed than subcutaneous adipose tissue. Indeed, in obese mice, both depots show an increase in the percentage of T cells that express IFNγ (6, 7), a proinflammatory cytokine that worsens insulin tolerance in vivo (6), but only obese visceral adipose tissue shows a decrease in anti-inflammatory regulatory T cells (29). Interestingly, our data show that, although γδ and γδ17 T cell infiltration increases with high fat feeding in subcutaneous inguinal (but

not in epididymal) adipose tissue, visceral epididymal adipose tissue in fact contains a

high number of γδ17 T cells per gram under both dietary conditions (low or high fat), comparable to that reached in inguinal adipose tissue on the HF diet. Thus the inflammatory infiltration and dietary responses of visceral vs. subcutaneous adipose tissue differ in a complex way that may contribute to their differential roles in metabolism. However, several studies indicate subcutaneous adipose tissue, in particular deep subcutaneous adipose tissue (60), may also contribute to the development of the metabolic syndrome (60, 61, 62).

232

A small but detectable subset of adipose tissue γδ T cells, from both inguinal and

epididymal depots harvested from mice fed a HF diet for 16-18 weeks, spontaneously

expressed IL-17 in vitro without requiring experimental stimulation. These cells were

absent in adipose tissue from mice fed a LF diet. After a prolonged period of high fat

feeding (1 year), the frequency of spontaneous expression by inguinal adipose tissue γδ17

T cells reached surprisingly high levels (up to 20 percent of γδ cells). Although the basis for this remains unclear, the spontaneous expression of IL-17 by a subset of γδ T cells suggests that γδ17 cells may be activated in the adipose tissue by yet unknown ligands.

Most known γδ TCR ligands are self-determinants associated with stress responses (e.g.

T10/T22 and MICA/B) (63), and one possibility is that adipocyte distress response

factors can act as ligands for γδ T cell receptors. Indeed, obese adipose tissue is

associated with increased tissue hypoxia, and adipocyte necrosis (53, 54, 55, 56).

Emerging studies demonstrate that γδ T cells are important for establishing inflammation

and are responsible for IL-17 expression during the first several days of infection or

during the onset of autoimmune disease (64, 65). Thus the local secretion of IL-17 by γδ

T cells may not only reflect low level inflammatory changes associated with adipocyte

stress, but may also help orchestrate the progression of inflammatory changes that

exacerbate the metabolic syndrome. In this context it is intriguing that, in the absence of

γδ T cells (in TCRδ KO mice), IL-17 production in adipose tissue appears to be taken on

by a phenotypically distinct, predominantly CD4-CD8- β TCR+ population, consistent

with a fundamental significance of the cytokine in adipocyte biology. This observation is

233

also consistent with a similar replacement of γδ T cells with αβ T cells in the skin of

TCRδ KO mice (66).

We conclude that IL-17 participates in the complex interplay between inflammation and metabolism, with systemic effects on glucose homeostasis and a negative regulatory role in adipogenesis and adipocyte function.

234

Acknowledgements

We thank Y. Chien (Department of Microbiology and Immunology, Stanford University)

for insightful scientific discussion, both J. Kolls (Department of Pediatrics, University of

Pittsburgh, Children’s Hospital of Pittsburgh) and Y. Iwakura (Institute of Medical

Science, University of Tokyo) for graciously supplying us with IL-17 knock-out mice,

and S. Adler (Department of Pathology, Stanford University), B. Nazario (University of

California, Santa Cruz), and E. Murphy (MERCK, Palo Alto) for expert technical assistance. This work was supported by grants from the National Institutes of Health

(Grants DK-084647, AI-47822, AI-079320, AI-072618, and AG-028908), funds from the

Specialized Center of Research (HL-67674), and the FACS Core facility of the Stanford

Digestive Disease Center (DK-56339), a Merit Award from the Department of Veterans

Affairs, and a National Institutes of Health pre-doctoral fellowship (AI-073198).

235

References

1. Olufadi, R., and C. D. Byrne. 2008. Clinical and laboratory diagnosis of the metabolic syndrome. J Clin Pathol 61:697-706. 2. Fajas, L. 2003. Adipogenesis: a cross-talk between cell proliferation and cell differentiation. Ann Med 35:79-85. 3. Feve, B. 2005. Adipogenesis: cellular and molecular aspects. Best Pract Res Clin Endocrinol Metab 19:483-499. 4. Lago, F., C. Dieguez, J. Gomez-Reino, and O. Gualillo. 2007. Adipokines as emerging mediators of immune response and inflammation. Nat Clin Pract Rheumatol 3:716-724. 5. Goralski, K. B., T. C. McCarthy, E. A. Hanniman, B. A. Zabel, E. C. Butcher, S. D. Parlee, S. Muruganandan, and C. J. Sinal. 2007. Chemerin, a novel adipokine that regulates adipogenesis and adipocyte metabolism. J Biol Chem 282:28175-28188. 6. Rocha, V. Z., E. J. Folco, G. Sukhova, K. Shimizu, I. Gotsman, A. H. Vernon, and P. Libby. 2008. Interferon-gamma, a Th1 cytokine, regulates fat inflammation: a role for adaptive immunity in obesity. Circ Res 103:467-476. 7. Winer, S., Y. Chan, G. Paltser, D. Truong, H. Tsui, J. Bahrami, R. Dorfman, Y. Wang, J. Zielenski, F. Mastronardi, Y. Maezawa, D. J. Drucker, E. Engleman, D. Winer, and H. M. Dosch. 2009. Normalization of obesity-associated insulin resistance through immunotherapy. Nat Med 15:921-929. 8. Ohsumi, J., S. Sakakibara, J. Yamaguchi, K. Miyadai, S. Yoshioka, T. Fujiwara, H. Horikoshi, and N. Serizawa. 1994. Troglitazone prevents the inhibitory effects of inflammatory cytokines on insulin-induced adipocyte differentiation in 3T3-L1 cells. Endocrinology 135:2279-2282. 9. Gasic, S., B. Tian, and A. Green. 1999. Tumor necrosis factor alpha stimulates lipolysis in adipocytes by decreasing Gi protein concentrations. J Biol Chem 274:6770- 6775. 10. Fain, J. N. 2006. Release of interleukins and other inflammatory cytokines by human adipose tissue is enhanced in obesity and primarily due to the nonfat cells. Vitam Horm 74:443-477. 11. Ferrante, A. W., Jr. 2007. Obesity-induced inflammation: a metabolic dialogue in the language of inflammation. J Intern Med 262:408-414. 12. Govindarajan, G., M. A. Alpert, and L. Tejwani. 2008. Endocrine and metabolic effects of fat: cardiovascular implications. Am J Med 121:366-370. 13. Steppan, C. M., S. T. Bailey, S. Bhat, E. J. Brown, R. R. Banerjee, C. M. Wright, H. R. Patel, R. S. Ahima, and M. A. Lazar. 2001. The hormone resistin links obesity to diabetes. Nature 409:307-312. 14. Kern, P. A., G. B. Di Gregorio, T. Lu, N. Rassouli, and G. Ranganathan. 2003. Adiponectin expression from human adipose tissue: relation to obesity, insulin resistance, and tumor necrosis factor-alpha expression. Diabetes 52:1779-1785.

236

15. Kern, P. A., S. Ranganathan, C. Li, L. Wood, and G. Ranganathan. 2001. Adipose tissue tumor necrosis factor and interleukin-6 expression in human obesity and insulin resistance. Am J Physiol Endocrinol Metab 280:E745-751. 16. Sterry, W., B. E. Strober, and A. Menter. 2007. Obesity in psoriasis: the metabolic, clinical and therapeutic implications. Report of an interdisciplinary conference and review. Br J Dermatol 157:649-655. 17. Munger, K. L., T. Chitnis, and A. Ascherio. 2009. Body size and risk of MS in two cohorts of US women. Neurology 73:1543-1550. 18. Winer, S., G. Paltser, Y. Chan, H. Tsui, E. Engleman, D. Winer, and H. M. Dosch. 2009. Obesity predisposes to Th17 bias. Eur J Immunol 39:2629-2635. 19. Plumb, J., R. Brawer, and N. Brisbon. 2007. The interplay of obesity and asthma. Curr Allergy Asthma Rep 7:385-389. 20. Hashmi, S., and Q. T. Zeng. 2006. Role of interleukin-17 and interleukin-17- induced cytokines interleukin-6 and interleukin-8 in unstable coronary artery disease. Coron Artery Dis 17:699-706. 21. Asarch, A., O. Barak, D. S. Loo, and A. B. Gottlieb. 2008. Th17 cells: A new therapeutic target in inflammatory dermatoses. J Dermatolog Treat:1-9. 22. Pene, J., S. Chevalier, L. Preisser, E. Venereau, M. H. Guilleux, S. Ghannam, J. P. Moles, Y. Danger, E. Ravon, S. Lesaux, H. Yssel, and H. Gascan. 2008. Chronically inflamed human tissues are infiltrated by highly differentiated Th17 lymphocytes. J Immunol 180:7423-7430. 23. Dong, C. 2008. TH17 cells in development: an updated view of their molecular identity and genetic programming. Nat Rev Immunol 8:337-348. 24. Chen, Z., and J. J. O'Shea. 2008. Th17 cells: a new fate for differentiating helper T cells. Immunol Res 41:87-102. 25. Sumarac-Dumanovic, M., D. Stevanovic, A. Ljubic, J. Jorga, M. Simic, D. Stamenkovic-Pejkovic, V. Starcevic, V. Trajkovic, and D. Micic. 2009. Increased activity of interleukin-23/interleukin-17 proinflammatory axis in obese women. Int J Obes (Lond) 33:151-156. 26. Zeyda, M., D. Farmer, J. Todoric, O. Aszmann, M. Speiser, G. Gyori, G. J. Zlabinger, and T. M. Stulnig. 2007. Human adipose tissue macrophages are of an anti- inflammatory phenotype but capable of excessive pro-inflammatory mediator production. Int J Obes (Lond) 31:1420-1428. 27. 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. 28. 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. 29. Feuerer, M., L. Herrero, D. Cipolletta, A. Naaz, J. Wong, A. Nayer, J. Lee, A. B. Goldfine, C. Benoist, S. Shoelson, and D. Mathis. 2009. Lean, but not obese, fat is

237

enriched for a unique population of regulatory T cells that affect metabolic parameters. Nat Med 15:930-939. 30. Nishimura, S., I. Manabe, M. Nagasaki, K. Eto, H. Yamashita, M. Ohsugi, M. Otsu, K. Hara, K. Ueki, S. Sugiura, K. Yoshimura, T. Kadowaki, and R. Nagai. 2009. CD8+ effector T cells contribute to macrophage recruitment and adipose tissue inflammation in obesity. Nat Med 15:914-920. 31. Lumeng, C. N., I. Maillard, and A. R. Saltiel. 2009. T-ing up inflammation in fat. Nat Med 15:846-847. 32. Shin, S., R. El-Diwany, S. Schaffert, E. J. Adams, K. C. Garcia, P. Pereira, and Y. H. Chien. 2005. Antigen recognition determinants of gammadelta T cell receptors. Science 308:252-255. 33. Groh, V., A. Steinle, S. Bauer, and T. Spies. 1998. Recognition of stress-induced MHC molecules by intestinal epithelial gammadelta T cells. Science 279:1737-1740. 34. Havran, W. L., Y. H. Chien, and J. P. Allison. 1991. Recognition of self antigens by skin-derived T cells with invariant gamma delta antigen receptors. Science 252:1430- 1432. 35. Girardi, M. 2006. Immunosurveillance and immunoregulation by gammadelta T cells. J Invest Dermatol 126:25-31. 36. Caspar-Bauguil, S., B. Cousin, A. Galinier, C. Segafredo, M. Nibbelink, M. Andre, L. Casteilla, and L. Penicaud. 2005. Adipose tissues as an ancestral immune organ: site-specific change in obesity. FEBS Lett 579:3487-3492. 37. Caspar-Bauguil, S., B. Cousin, M. Andre, M. Nibbelink, A. Galinier, B. Periquet, L. Casteilla, and L. Penicaud. 2006. Weight-dependent changes of immune system in adipose tissue: importance of leptin. Exp Cell Res 312:2195-2202. 38. Bost, F., M. Aouadi, L. Caron, and B. Binetruy. 2005. The role of MAPKs in adipocyte differentiation and obesity. Biochimie 87:51-56. 39. Fajas, L., J. C. Fruchart, and J. Auwerx. 1998. Transcriptional control of adipogenesis. Curr Opin Cell Biol 10:165-173. 40. Shin, J. H., D. W. Shin, and M. Noh. 2009. Interleukin-17A inhibits adipocyte differentiation in human mesenchymal stem cells and regulates pro-inflammatory responses in adipocytes. Biochem Pharmacol 77:1835-1844. 41. Hwang, S. Y., J. Y. Kim, K. W. Kim, M. K. Park, Y. Moon, W. U. Kim, and H. Y. Kim. 2004. IL-17 induces production of IL-6 and IL-8 in rheumatoid arthritis synovial fibroblasts via NF-kappaB- and PI3-kinase/Akt-dependent pathways. Arthritis Res Ther 6:R120-128. 42. Rotter, V., I. Nagaev, and U. Smith. 2003. Interleukin-6 (IL-6) induces insulin resistance in 3T3-L1 adipocytes and is, like IL-8 and tumor necrosis factor-alpha, overexpressed in human fat cells from insulin-resistant subjects. J Biol Chem 278:45777- 45784. 43. Wallenius, V., K. Wallenius, B. Ahren, M. Rudling, H. Carlsten, S. L. Dickson, C. Ohlsson, and J. O. Jansson. 2002. Interleukin-6-deficient mice develop mature-onset obesity. Nat Med 8:75-79. 238

44. Rosen, E. D. 2005. The transcriptional basis of adipocyte development. Prostaglandins Leukot Essent Fatty Acids 73:31-34. 45. Tang, Q. Q., T. C. Otto, and M. D. Lane. 2003. Mitotic clonal expansion: a synchronous process required for adipogenesis. Proc Natl Acad Sci U S A 100:44-49. 46. Hu, E., J. B. Kim, P. Sarraf, and B. M. Spiegelman. 1996. Inhibition of adipogenesis through MAP kinase-mediated phosphorylation of PPARgamma. Science 274:2100-2103. 47. Bhattacharya, I., and A. Ullrich. 2006. Endothelin-1 inhibits adipogenesis: role of phosphorylation of Akt and ERK1/2. FEBS Lett 580:5765-5771. 48. Furuhashi, M., and G. S. Hotamisligil. 2008. Fatty acid-binding proteins: role in metabolic diseases and potential as drug targets. Nat Rev Drug Discov 7:489-503. 49. Cornier, M. A., D. Dabelea, T. L. Hernandez, R. C. Lindstrom, A. J. Steig, N. R. Stob, R. E. Van Pelt, H. Wang, and R. H. Eckel. 2008. The metabolic syndrome. Endocr Rev 29:777-822. 50. Larsen, T. M., S. Toubro, and A. Astrup. 2003. PPARgamma agonists in the treatment of type II diabetes: is increased fatness commensurate with long-term efficacy? Int J Obes Relat Metab Disord 27:147-161. 51. Ceriello, A. 2008. Thiazolidinediones as anti-inflammatory and anti-atherogenic agents. Diabetes Metab Res Rev 24:14-26. 52. Gaffen, S. L., J. M. Kramer, J. J. Yu, and F. Shen. 2006. The IL-17 cytokine family. Vitam Horm 74:255-282. 53. 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. 54. 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. 55. Hosogai, N., A. Fukuhara, K. Oshima, Y. Miyata, S. Tanaka, K. Segawa, S. Furukawa, Y. Tochino, R. Komuro, M. Matsuda, and I. Shimomura. 2007. Adipose tissue hypoxia in obesity and its impact on adipocytokine dysregulation. Diabetes 56:901-911. 56. Rausch, M. E., S. Weisberg, P. Vardhana, and D. V. Tortoriello. 2008. Obesity in C57BL/6J mice is characterized by adipose tissue hypoxia and cytotoxic T-cell infiltration. Int J Obes (Lond) 32:451-463. 57. Lara-Castro, C., Y. Fu, B. H. Chung, and W. T. Garvey. 2007. Adiponectin and the metabolic syndrome: mechanisms mediating risk for metabolic and cardiovascular disease. Curr Opin Lipidol 18:263-270. 58. Arita, Y., S. Kihara, N. Ouchi, M. Takahashi, K. Maeda, J. Miyagawa, K. Hotta, I. Shimomura, T. Nakamura, K. Miyaoka, H. Kuriyama, M. Nishida, S. Yamashita, K. Okubo, K. Matsubara, M. Muraguchi, Y. Ohmoto, T. Funahashi, and Y. Matsuzawa. 1999. Paradoxical decrease of an adipose-specific protein, adiponectin, in obesity. Biochem Biophys Res Commun 257:79-83. 239

59. Hamdy, O., S. Porramatikul, and E. Al-Ozairi. 2006. Metabolic obesity: the paradox between visceral and subcutaneous fat. Curr Diabetes Rev 2:367-373. 60. Monzon, J. R., R. Basile, S. Heneghan, V. Udupi, and A. Green. 2002. Lipolysis in adipocytes isolated from deep and superficial subcutaneous adipose tissue. Obes Res 10:266-269. 61. Tulloch-Reid, M. K., R. L. Hanson, N. G. Sebring, J. C. Reynolds, A. Premkumar, D. J. Genovese, and A. E. Sumner. 2004. Both subcutaneous and visceral adipose tissue correlate highly with insulin resistance in african americans. Obes Res 12:1352-1359. 62. Kelley, D. E., F. L. Thaete, F. Troost, T. Huwe, and B. H. Goodpaster. 2000. Subdivisions of subcutaneous abdominal adipose tissue and insulin resistance. Am J Physiol Endocrinol Metab 278:E941-948. 63. Thedrez, A., C. Sabourin, J. Gertner, M. C. Devilder, S. Allain-Maillet, J. J. Fournie, E. Scotet, and M. Bonneville. 2007. Self/non-self discrimination by human gammadelta T cells: simple solutions for a complex issue? Immunol Rev 215:123-135. 64. Roark, C. L., P. L. Simonian, A. P. Fontenot, W. K. Born, and R. L. O'Brien. 2008. gammadelta T cells: an important source of IL-17. Curr Opin Immunol 20:353- 357. 65. Jensen, K. D., X. Su, S. Shin, L. Li, S. Youssef, S. Yamasaki, L. Steinman, T. Saito, R. M. Locksley, M. M. Davis, N. Baumgarth, and Y. H. Chien. 2008. Thymic selection determines gammadelta T cell effector fate: antigen-naive cells make interleukin-17 and antigen-experienced cells make interferon gamma. Immunity 29:90- 100. 66. Jameson, J. M., G. Cauvi, D. A. Witherden, and W. L. Havran. 2004. A keratinocyte-responsive gamma delta TCR is necessary for dendritic epidermal T cell activation by damaged keratinocytes and maintenance in the epidermis. J Immunol 172:3573-3579.

240

CHAPTER 6 : FINAL DISCUSSION

Inflammation is critically important for clearing pathogens from the body during

an infection and for initiating tissue repair. Leukocyte infiltration to damaged tissue is a

hallmark of inflammation and is mediated by chemoattractants. Upon recruitment to

damaged tissue, immune cells become activated and secrete acute phase proteins (e.g.

cytokines) that can have local and systemic effects (1). Modification of energy

homeostasis is one such effect. Tissue damage and inflammation lead to the production of

acute phase proteins, either from the liver or activated immune cells, that have the ability

to induce energy release from glycogen (liver), amino acid (muscle), and fatty acid

(adipose tissue) stores, resulting in hyperglycemia and hyperlipidemia. Presumably, this

release of energy and protein resources facilitates tissue and immune cell turnover and

function. However, prolonged exposure to acute phase proteins can lead to muscle and fat

wasting, known as cachexia (2, 3). Indeed, components of the inflammatory response (IL-

6, IL-1, TNFα, IL-8, and IFNγ) are not only associated with cachexia, but some are

known cachectic/anorexogenic factors (4, 5, 6, 7, 8). Cancer and infection associated

wasting is thought to be mediated partly by properties of the aforementioned pro-

inflammatory cytokines, which can induce anorexic feeding behavior and/or catabolism

in adipose tissue and muscle (9, 10). Thus, it is well known that disease associated inflammation can affect systemic lipid and glucose metabolism.

Just as inflammation is associated with altered metabolism, disorders of metabolism are correlated with altered inflammatory and immune status. As described

241

before, in addition to metabolic complications, obesity is now linked to predisposition of

various inflammatory diseases (e.g. asthma, psoriasis, and multiple sclerosis). The

interplay between disease related inflammation and metabolism, and adipose tissue

inflammation and disease is complex. However, the inflammation that occurs in damaged

tissue and obese adipose tissue are much alike and are probably regulated by similar

mechanisms.

Chemoattractants and their receptors are critically important for leukocyte

recruitment to damaged tissue and obese adipose tissue (11, 12). Using ex vivo and in

vitro assays, we discovered a potential role for the chemoattractant chemerin in the

chemotaxis of macrophages via the chemerin receptor, CMKLR1. Resident peritoneal

cavity macrophages expressed high levels of CMKLR1 under homeostatic conditions, suggesting CMKLR1 may play a role in normal tissue trafficking of macrophages (13)

(Chapter 2). Furthermore, work from our group and others demonstrate the presence of

bioactive chemerin in inflamed tissue fluids (13, 14, 15) and the upregulation of either

systemic or tissue chemerin expression during inflammatory disease (16, 17, 18, 19),

suggesting chemerin may participate in the trafficking of immune cells to inflamed tissue.

In support of this, a number of studies have shown the colocalization of CMKLR1

expressing immune cells to sites of bioactive chemerin (18, 19, 20), most notably in a

pro-inflammatory setting. Consistent with these observations, we found CMKLR1

positive myeloid dendritic cells (mDCs) and macrophages were recruited to inflamed

central nervous system tissue from mice with EAE (21) (Chapter 3). Chemerin was

242

upregulated in inflamed CNS and, compared to wild type controls, mice deficient for

CMKLR1 developed less severe EAE, accompanied with a lower immune cell CNS

infiltrate. Adoptive transfer of MOG35-55 reactive CMKLR1 deficient lymphocytes to

either WT or CMKLR1 deficient animals failed to induce robust disease compared to WT

mice that received WT MOG35-55 reactive lymphocytes. These data suggest chemerin

may be involved in the recruitment of mDCs to inflamed CNS tissue (Figure 1).

Our studies also revealed expression of CMKLR1 on microglia in inflamed CNS,

suggesting microglia can bind and potentially respond to chemerin. One possibility is that

chemerin can influence microglial activation or function. Microglia are resident CNS

tissue phagocytes and antigen presenting cells which are important for

immunosurveilance (22). As one speculative example, a recent report demonstrated that

synthetic chemerin peptides, through CMKRL1, can significantly enhance phagocytosis

by peripheral macrophages (23). A similar role for CMKLR1 in the CNS could contribute

to antigen presentation and inflammation during EAE and could help explain reduced

disease in CMKLR1-deficient mice.

Although various reports indicate chemerin is important for recruiting immune

cells to inflamed tissue (13, 14, 17, 19, 20, 21, 24, 25), suggesting it has a pro-

inflammatory role in disease development, a handful of reports suggest chemerin or its

derivatives have anti-inflammatory effects. Cash et al. showed that synthetic chemerin peptides can ameliorate zymosan-induced peritonitis in a CMKLR1 dependent manner

(26), presumably via enhancing macrophage phagocytosis, allowing increased clearance

243

of apoptotic cells, and via inhibition of macrophage activation, thus reducing

inflammation (23). An independent report from Luangsay et al. (27) found mature

chemerin could attenuate inflammation in LPS induced murine lung injury in a CMKLR1

dependent manner. In these studies, co-administration of chemerin and LPS led to

increased macrophage recruitment and decreased neutrophil infiltration to the lungs

compared to LPS alone. There was also a reduction in pro-inflammatory cytokine tissue expression with chemerin administration. The mechanism responsible for the anti- inflammatory effect of chemerin in these disease models is still unclear. Although both

groups came to a similar conclusion, Luangsay could not demonstrate any functional

responses to the chemerin derivatives reported by Cash, thus the findings are conflicting

and somewhat controversial.

Our discovery of chemerin binding to CCRL2 further adds to the complexity of

chemerin’s role in inflammation (28) (Chapter 4). Our studies implicate CCRL2 as a non- signaling receptor for chemerin. The carboxyl-terminus contains the bioactive motif in chemerin, responsible for CMKLR1 chemotaxis (29). CCRL2 seems able to bind chemerin in such a way that the carboxyl-terminus of chemerin is exposed. Consistent with this, we discovered CCRL2+ L1.2 cells, but not CCRL2- L1.2 cells, could bind

chemerin and subsequently induce calcium flux in CMKLR1+ L1.2 responder cells. In

vivo studies indicate CCRL2 has a functional role in tissue inflammation. Mice deficient

for CCRL2 had similar ear swelling in a model of robust IgE-dependent passive

cutaneous anaphylaxis (PCA). However, when the amount of antigen specific IgE used

244

for inducing PCA was reduced to attenuate the severity of PCA, CCRL2 deficient mice

had less ear swelling compared to WT animals. This suggests that CCRL2 is dispensable

for the induction of IgE-dependent PCA, but may play a role in mediating the inflammation that follows induction of IgE-dependent PCA. Supporting this hypothesis, skin mast cells expressed high levels of CCRL2 and could bind chemerin, ex vivo, but did

not express CMKLR1. Studies using adoptive masT cell transfer to masT cell deficient

mice revealed mast cells lacking CCRL2 elicited fewer infiltrating leukocytes during

IgE-dependent PCA, suggesting CCRL2 plays a role in leukocyte trafficking to inflamed

skin. It is interesting to note that CCRL2 expression can be induced on macrophages with

TLR ligands and pro-inflammatory cytokines. CCRL2 and CMKLR1 are also inversely

regulated by the same stimuli (13, 28). Thus, it appears as though pro-inflammatory

stimuli upregulate CCRL2 while down-regulating CMKLR1, making it reasonable to

postulate that macrophages can express both CCRL2 and CMKLR1 at some point upon

exposure to pro-inflammatory signals. Given these data, we hypothesize that CCRL2

(expressed by mast cells or macrophages) may present bioactive chemerin to CMKLR1

expressing cells, acting in such a way as to help locally concentrate chemerin and

possibly mediate CMKLR1 expressing immune cell trafficking to inflamed skin or elicit

responses (e.g. cytokine secretion) from CMKLR1 expressing cells (Figure 1). Further

research is needed to confirm these postulates.

In addition to mediating leukocyte trafficking, chemerin was found to have a non- trafficking role in the development of adipocytes. Goralski et al., with the help of our

245

group, found the expression of chemerin and CMKLR1 by adipocytes and determined

chemerin has a proadipogenic effect via CMKLR1 (30). Interestingly, while participating

in this collaborative work, we found a previously unknown role for the pro-inflammatory

cytokine, IL-17, in the negative regulation of adipogenesis (Chapter 5).

Similar to chemerin, IL-17 is also upregulated in inflamed tissue. However, IL-17

is expressed by various T cell subsets depending on the stage or type of inflammation.

For instance, during EAE, IL-17 is expressed by γδ T cells and may potentiate IL-17

expression by conventional CD4 T cells during disease progression (31). Although CD4

T cells are typically considered the primary cell type that can express IL-17 during

inflammation, a number of studies have shown that γδ T cells can greatly contribute to

IL-17 expression, or in some instances, be the dominant cell type that expresses IL-17.

To illustrate, in several models of infection, γδ T cells vastly dominate the IL-17

expression response and are thought to rapidly establish inflammation as an early defense against pathogens (32). As mentioned earlier, compared to lean adipose tissue, obese adipose tissue is associated with increased leukocyte infiltration and is a source of

increased pro-inflammatory cytokine secretion, making it an important endocrine organ

that can increase systemic inflammation and may predispose to the development of

various pro-inflammatory diseases. The inflammation that occurs in obese adipose tissue

is similar to that found in other models of tissue damage and is typified by infiltration of

macrophages, B-cells, and T cells (33). Consistent with the theory that obese adipose

tissue is associated with inflammation, T cell expression of IL-17 was upregulated in

246

obese adipose tissue compared to lean adipose tissue. γδ T cells dominated T cell IL-17

expression in fat.

In addition to establishing an association between IL-17 expression and obesity,

we also found IL-17 deficient animals were more susceptible to diet induced obesity

(DIO). Although IL-17 deficient animals had higher body fat mass yet similar lean mass

compared to WT animals, there were no differences in food intake or body temperature

between WT and KO groups, thus we do not know if susceptibility to DIO in KO animals

is due to increased energy uptake and/or decreased energy expenditure. More sensitive

means of measuring energy usage or caloric uptake may help elucidate what mechanisms

are involved in this phenomenon. We did, however, find IL-17 could inhibit adipogenesis

in vitro, indicating IL-17 may have an anti-adipogenic effect in vivo. It is possible that T cell IL-17 expression in obese fat can attenuate adipocyte differentiation, leading to a decrease in mature adipocytes and reduced reservoir for lipid storage (Figure 1).

We also found IL-17 was capable of inhibiting insulin stimulated glucose uptake in mature 3T3-L1 adipocytes, in vitro, suggesting it could modulate glucose homeostasis.

Consistent with this theory, young, IL-17 deficient animals were more insulin sensitive and had better glucose tolerance than WT animals before any significant differences in

body mass was detectable between genotypes, indicating IL-17 may regulate adipose or

systemic glucose homeostasis in vivo. Although leukocytes expressed high levels of IL-

17 in adipose tissue, implicating fat as a possible source for IL-17, we cannot rule out

other sources of IL-17 (e.g. liver, muscle, gut). Adipose tissue, however, is a likely

247

source for IL-17 production as systemic IL-17 is higher in obese mice than in lean mice

(34). In contrast to young mice, older obese IL-17 deficient animals had similar insulin

and glucose tolerance to WT animals. Why IL-17 deficient animals lose protection

against insulin resistance is still unclear. It is possible that because IL-17 deficient

animals become more obese that WT animals, other inflammatory cytokines can perform

similar functions as IL-17 in exacerbating the metabolic syndrome. Indeed, a number of

pro-inflammatory cytokines expressed by obese adipose tissue have a negative impact on

glucose and lipid homeostasis (35).

Our results, along with those from Goralski et al. (30, 36), show that IL-17 and

chemerin influence adipogenesis, and IL-17 affects diet-induced obesity and metabolism.

Moreover, these factors are produced within obese adipose tissue and fat-derived IL-17

or chemerin have the potential to exacerbate systemic inflammatory disease. Consistent

with this hypothesis, obese mice display a higher TH17 cell bias (34) and have higher circulating levels of chemerin (37). Obese mice also develop exacerbated trinitrobenzene sulfonic acid-induced colitis, EAE, and zymosan-induced peritonitis (34, 38). In the case of zymosan-induced peritonitis, obese mice produce significantly higher levels of IL-17 in the peritoneal cavity (38), a tissue site that harbors large adipose tissue depots. Thus, although speculative, obese adipose derived IL-17 may predispose to autoimmunity or exacerbate acute tissue injury. Furthermore, obese adipose tissue may provide systemically high levels of chemerin, which could contribute to tissue leukocyte

248

trafficking upon proteolysis during tissue inflammation. Further studies are necessary to address these possibilities.

The work reported here defines chemerin and its receptors as significant contributing factors in inflammatory disease, and reveals IL-17 as a novel adipose tissue associated cytokine with metabolism altering properties. Given the link between inflammation and obesity, both chemerin and IL-17 represent potential therapeutic targets

for the amelioration or treatment of inflammatory disease and the metabolic syndrome.

249

Figure 1. Proposed model of chemerin and IL-17 in tissue inflammation, obesity, and metabolism. Tissue damage (skin and CNS) or necrosis (adipose tissue) induces tissue inflammation and leads to the recruitment and/or activation of immune cells. Specifically, chemerin from inflamed tissue may be activated and mediate the recruitment of CMKLR1 expressing immune cells from the vasculature (e.g. macrophages, NK cells, or pDCs). CCRL2 may bind activated chemerin, thus enhancing local tissue chemerin concentrations, and promote trafficking of CMKLR1+ immune cells to damaged tissue. Chemerin may also interact with immune cells, such as microglia in the CNS, to alter effector functions, or with non-immune cells, such as adipocytes, to moderate cell development. Activated immune cells (recruited or resident) may secrete IL-17, promoting inflammation. Furthermore, obese adipose tissue derived IL-17 may act 250

locally in fat to inhibit adipogenesis and glucose metabolism, or may act systemically to exacerbate peripheral tissue inflammation. Blue lines indicate chemerin tissue effects, while red lines indicate IL-17 tissue effects.

251

References 1. Cray, C., J. Zaias, and N. H. Altman. 2009. Acute phase response in animals: a review. Comp Med 59:517-526. 2. Goldstein, S. A., and D. H. Elwyn. 1989. The effects of injury and sepsis on fuel utilization. Annu Rev Nutr 9:445-473. 3. Jeschke, M. G., C. C. Finnerty, G. A. Kulp, R. Przkora, R. P. Mlcak, and D. N. Herndon. 2008. Combination of recombinant human growth hormone and propranolol decreases hypermetabolism and inflammation in severely burned children. Pediatr Crit Care Med 9:209-216. 4. Wallenius, V., K. Wallenius, B. Ahren, M. Rudling, H. Carlsten, S. L. Dickson, C. Ohlsson, and J. O. Jansson. 2002. Interleukin-6-deficient mice develop mature-onset obesity. Nat Med 8:75-79. 5. Debonis, D., M. M. Meguid, A. Laviano, Z. J. Yang, and J. R. Gleason. 1995. Temporal changes in meal number and meal size relationship in response to rHu IL-1 alpha. Neuroreport 6:1752-1756. 6. Opara, E. I., A. Laviano, M. M. Meguid, and Z. J. Yang. 1995. Correlation between food intake and CSF IL-1 alpha in anorectic tumor bearing rats. Neuroreport 6:750-752. 7. Sonti, G., S. E. Ilyin, and C. R. Plata-Salaman. 1996. Anorexia induced by cytokine interactions at pathophysiological concentrations. Am J Physiol 270:R1394- 1402. 8. Plata-Salaman, C. R. 2000. Central nervous system mechanisms contributing to the cachexia-anorexia syndrome. Nutrition 16:1009-1012. 9. Argiles, J. M., S. Busquets, M. Toledo, and F. J. Lopez-Soriano. 2009. The role of cytokines in cancer cachexia. Curr Opin Support Palliat Care 3:263-268. 10. Delano, M. J., and L. L. Moldawer. 2006. The origins of cachexia in acute and chronic inflammatory diseases. Nutr Clin Pract 21:68-81. 11. Wilson, E. H., W. Weninger, and C. A. Hunter. 2010. Trafficking of immune cells in the central nervous system. J Clin Invest 120:1368-1379. 12. Anderson, E. K., D. A. Gutierrez, and A. H. Hasty. 2010. Adipose tissue recruitment of leukocytes. Curr Opin Lipidol 21:172-177. 13. Zabel, B. A., T. Ohyama, L. Zuniga, J. Y. Kim, B. Johnston, S. J. Allen, D. G. Guido, T. M. Handel, and E. C. Butcher. 2006. Chemokine-like receptor 1 expression by macrophages in vivo: Regulation by TGF-beta and TLR ligands. Exp Hematol 34:1106- 1114. 14. Wittamer, V., J. D. Franssen, M. Vulcano, J. F. Mirjolet, E. Le Poul, I. Migeotte, S. Brezillon, R. Tyldesley, C. Blanpain, M. Detheux, A. Mantovani, S. Sozzani, G. Vassart, M. Parmentier, and D. Communi. 2003. Specific recruitment of antigen- presenting cells by chemerin, a novel processed ligand from human inflammatory fluids. J Exp Med 198:977-985.

252

15. Kutzleb, C., A. Busmann, M. Wendland, and E. Maronde. 2005. Discovery of novel regulatory peptides by reverse : spotlight on chemerin and the RF- amide peptides metastin and QRFP. Curr Protein Pept Sci 6:265-278. 16. Weigert, J., F. Obermeier, M. Neumeier, J. Wanninger, M. Filarsky, S. Bauer, C. Aslanidis, G. Rogler, C. Ott, A. Schaffler, J. Scholmerich, and C. Buechler. 2010. Circulating levels of chemerin and adiponectin are higher in ulcerative colitis and chemerin is elevated in Crohn's disease. Inflamm Bowel Dis 16:630-637. 17. Parolini, S., A. Santoro, E. Marcenaro, W. Luini, L. Massardi, F. Facchetti, D. Communi, M. Parmentier, A. Majorana, M. Sironi, G. Tabellini, A. Moretta, and S. Sozzani. 2007. The role of chemerin in the colocalization of NK and dendritic cell subsets into inflamed tissues. Blood 109:3625-3632. 18. Skrzeczynska-Moncznik, J., K. Wawro, A. Stefanska, E. Oleszycka, P. Kulig, B. A. Zabel, M. Sulkowski, M. Kapinska-Mrowiecka, M. Czubak-Macugowska, E. C. Butcher, and J. Cichy. 2009. Potential role of chemerin in recruitment of plasmacytoid dendritic cells to diseased skin. Biochem Biophys Res Commun 380:323-327. 19. Vermi, W., E. Riboldi, V. Wittamer, F. Gentili, W. Luini, S. Marrelli, A. Vecchi, J. D. Franssen, D. Communi, L. Massardi, M. Sironi, A. Mantovani, M. Parmentier, F. Facchetti, and S. Sozzani. 2005. Role of ChemR23 in directing the migration of myeloid and plasmacytoid dendritic cells to lymphoid organs and inflamed skin. J Exp Med 201:509-515. 20. Skrzeczynska-Moncznik, J., A. Stefanska, B. A. Zabel, M. Kapinska-Mrowiecka, E. C. Butcher, and J. Cichy. 2009. Chemerin and the recruitment of NK cells to diseased skin. Acta Biochim Pol 56:355-360. 21. Graham, K. L., B. A. Zabel, S. Loghavi, L. A. Zuniga, P. P. Ho, R. A. Sobel, and E. C. Butcher. 2009. Chemokine-like receptor-1 expression by central nervous system- infiltrating leukocytes and involvement in a model of autoimmune demyelinating disease. J Immunol 183:6717-6723. 22. Drew, P. D., P. D. Storer, J. Xu, and J. A. Chavis. 2005. Hormone regulation of microglial cell activation: relevance to multiple sclerosis. Brain Res Brain Res Rev 48:322-327. 23. Cash, J. L., A. R. Christian, and D. R. Greaves. 2010. Chemerin peptides promote phagocytosis in a ChemR23- and Syk-dependent manner. J Immunol 184:5315-5324. 24. Albanesi, C., C. Scarponi, S. Pallotta, R. Daniele, D. Bosisio, S. Madonna, P. Fortugno, S. Gonzalvo-Feo, J. D. Franssen, M. Parmentier, O. De Pita, G. Girolomoni, and S. Sozzani. 2009. Chemerin expression marks early psoriatic skin lesions and correlates with plasmacytoid dendritic cell recruitment. J Exp Med 206:249-258. 25. Lande, R., V. Gafa, B. Serafini, E. Giacomini, A. Visconti, M. E. Remoli, M. Severa, M. Parmentier, G. Ristori, M. Salvetti, F. Aloisi, and E. M. Coccia. 2008. Plasmacytoid dendritic cells in multiple sclerosis: intracerebral recruitment and impaired maturation in response to interferon-beta. J Neuropathol Exp Neurol 67:388-401.

253

26. Cash, J. L., R. Hart, A. Russ, J. P. Dixon, W. H. Colledge, J. Doran, A. G. Hendrick, M. B. Carlton, and D. R. Greaves. 2008. Synthetic chemerin-derived peptides suppress inflammation through ChemR23. J Exp Med 205:767-775. 27. Luangsay, S., V. Wittamer, B. Bondue, O. De Henau, L. Rouger, M. Brait, J. D. Franssen, P. de Nadai, F. Huaux, and M. Parmentier. 2009. Mouse ChemR23 is expressed in dendritic cell subsets and macrophages, and mediates an anti-inflammatory activity of chemerin in a lung disease model. J Immunol 183:6489-6499. 28. Zabel, B. A., S. Nakae, L. Zuniga, J. Y. Kim, T. Ohyama, C. Alt, J. Pan, H. Suto, D. Soler, S. J. Allen, T. M. Handel, C. H. Song, S. J. Galli, and E. C. Butcher. 2008. Mast cell-expressed orphan receptor CCRL2 binds chemerin and is required for optimal induction of IgE-mediated passive cutaneous anaphylaxis. J Exp Med 205:2207-2220. 29. Zabel, B. A., L. Zuniga, T. Ohyama, S. J. Allen, J. Cichy, T. M. Handel, and E. C. Butcher. 2006. Chemoattractants, extracellular proteases, and the integrated host defense response. Exp Hematol 34:1021-1032. 30. Goralski, K. B., T. C. McCarthy, E. A. Hanniman, B. A. Zabel, E. C. Butcher, S. D. Parlee, S. Muruganandan, and C. J. Sinal. 2007. Chemerin, a novel adipokine that regulates adipogenesis and adipocyte metabolism. J Biol Chem 282:28175-28188. 31. Sutton, C. E., S. J. Lalor, C. M. Sweeney, C. F. Brereton, E. C. Lavelle, and K. H. Mills. 2009. Interleukin-1 and IL-23 induce innate IL-17 production from gammadelta T cells, amplifying Th17 responses and autoimmunity. Immunity 31:331-341. 32. Roark, C. L., P. L. Simonian, A. P. Fontenot, W. K. Born, and R. L. O'Brien. 2008. gammadelta T cells: an important source of IL-17. Curr Opin Immunol 20:353- 357. 33. Caspar-Bauguil, S., B. Cousin, S. Bour, L. Castiella, L. Penicaud, and C. Carpene. 2009. Adipose tissue lymphocytes: types and roles. J Physiol Biochem 65:423-436. 34. Winer, S., G. Paltser, Y. Chan, H. Tsui, E. Engleman, D. Winer, and H. M. Dosch. 2009. Obesity predisposes to Th17 bias. Eur J Immunol 39:2629-2635. 35. Fain, J. N. 2006. Release of interleukins and other inflammatory cytokines by human adipose tissue is enhanced in obesity and primarily due to the nonfat cells. Vitam Horm 74:443-477. 36. Goralski, K. B., and C. J. Sinal. 2009. Elucidation of chemerin and chemokine- like receptor-1 function in adipocytes by adenoviral-mediated shRNA knockdown of gene expression. Methods Enzymol 460:289-312. 37. Ernst, M. C., M. Issa, K. B. Goralski, and C. J. Sinal. 2010. Chemerin exacerbates glucose intolerance in mouse models of obesity and diabetes. Endocrinology 151:1998- 2007. 38. Pini, M., and G. Fantuzzi. 2009. Enhanced production of IL-17A during zymosan- induced peritonitis in obese mice. J Leukoc Biol.

254