The Pennsylvania State University the Graduate School

The Pennsylvania State University

The Graduate School

Department of Cell and Molecular Biology

MEPRIN METALLOPROTEASES MODULATE THE

HOST RESPONSE TO ESCHERICHIA COLI

A Dissertation in

Cell and Molecular Biology

Renee E. Yura

Submitted in Partial Fulfillment of the Requirements for the Degree of

Doctor of Philosophy

August 2008

The dissertation of Renee E. Yura was reviewed and approved* by the following:

Judith S. Bond Distinguished Professor and Chair of Biochemistry and Molecular Biology Dissertation Advisor Chair of Committee

Sarah K. Bronson Associate Professor of Cellular and Molecular Physiology

Robert G. Levenson Professor of Pharmacology

Andrea Manni Professor of Medicine

W. Brian Reeves Adjunct Professor of Medicine Special Member

Henry J. Donahue Professor of Orthopaedics and Rehabilitation Director of the Cell and Molecular Biology Graduate Program

*Signatures are on file in the Graduate School

Abstract

Meprin metalloproteases, composed of α and/or β subunits, consist of membrane- bound and secreted forms that are abundantly expressed in kidney and intestinal epithelial cells. They are also expressed in the skin and in certain populations of leukocytes. Meprins have been implicated in several inflammatory diseases, such as renal ischemia, diabetic nephropathy, and inflammatory bowel disease, indicating a role for these enzymes in modulation of the immune response. Prior to the initiation of this work, extensive information about the structure and in vitro behavior of the meprins was known, but little was known about their in vivo roles in immune modulation. These studies are the first to demonstrate a relationship between the inflammatory response and the meprins in both systemic and bladder challenge models.

The aim of this work was to determine the role of meprins in the host response to gram-negative uropathic Escherichia coli (E. coli). Initial studies demonstrated marked increases in meprin α expression in the urine of women with active urinary tract infections

(UTIs), implicating meprin involvement in the host response to bacterial infections. To examine further the role of meprins in the host response to UTI, meprin α knockout (αKO) and wild-type (WT) mice were challenged with a transurethral inoculation of uropathic E. coli. In this localized model of inflammation, bladder myeloperoxidase (MPO) activity, bladder weight, and bladder permeability were significantly less in αKO compared to WT mice after induction of UTI. These data indicate that meprin A is pro-inflammatory, contributing to leukocyte infiltration, edema, and epithelial damage. To determine whether the action of meprin A was the result of a direct interaction with E. coli, extensive in vitro studies were carried out. Meprin A did not decrease the binding of E. coli to bladder cells in

iii

culture, nor did it degrade the pili that are responsible for the attachment of E. coli to the bladder epithelium. No evidence of direct interactions between meprin A and E. coli has been found, indicating that the differential response observed in meprin αKO mice in comparison to WT is indirect and likely involves the immune system.

A large component of the immune response to E. coli is directed toward lipopolysaccharide (LPS) in the bacterial cell wall. Therefore, the responses of meprin αKO, meprin β knockout (βKO), meprin αβ double knockout (αβKO), and WT mice after intraperitoneal (i.p.) LPS challenge were examined. Genotype-specific differences in response to LPS were observed as early as 1 h after challenge with 2.5 mg/kg i.p. E. coli

LPS. Meprin αKO mice displayed a decreased systemic response to LPS compared to WT mice and meprin βKO mice, as indicated by lower blood urea nitrogen (BUN) levels, lower serum TNFα levels, and less severe hypothermia, implicating meprin in modulation of the host immune response to endotoxin. These data are consistent with those from UTI studies, corroborating the proposition that meprin A is pro-inflammatory, contributing to the LPS- induced renal injury, loss of body temperature regulation, and stimulation of cytokine increase. The LPS response of the meprin αβKO mice consisted of a prolonged elevation of

BUN levels and more dramatic hypothermia than the other three genotypes. These results indicate that the meprin β subunit attenuates the pro-inflammatory actions of the meprin α subunit, and that the balance between meprin α subunits and meprin β subunits is important in determining the outcome of the host response. Serum cytokine profiles were genotype- specific and confirmed a role for the meprins in the early phases of the host response, with markedly lower levels of TNFα, IL-1β, and CCL2/MCP-1 in the meprin αKO mice and higher levels of CCL3/MIP-1α in the meprin αβKO mice versus WT mice, within 3 h after

LPS challenge.

Meprin αKO mice were also hyporesponsive to localized LPS administration to the bladder, with decreased bladder edema, leukocyte infiltration, and bladder permeability, again implicating meprin A in inflammation. Increased bladder permeability in WT mice in contrast to meprin αKO mice challenged intravesically with LPS or E. coli can be attributed to the capability of meprin A to disrupt tight junctions, because active, soluble meprin A is capable of disrupting the junctional complexes in a kidney cell culture model without demonstrating cytotoxicity. The results from the LPS bladder challenge were congruent with those from the E. coli challenge, confirming that LPS is a major factor in the host response to

E. coli challenge. Meprins’ role in the modulation of the host response may be attributed to activation or degradation of cytokines/chemokines and/or leukocyte migration through extracellular matrix and/or remodeling of the tight junctions between cells. Taken together, the data indicate a role for meprin A in the modulation of the immune response to infection, that the meprin protease isoforms have distinct functions in modulation of the host response to LPS, and that meprin A contributes to the pathogenesis of endotoxicity.

Table of Contents

List of Figures xii

List of Tables xv

List of Abbreviations xvi

Acknowledgements xx

Chapter 1. Introduction 1

Proteases 1

Metzincins 2

1.1 Metzincin metalloproteases share structural similarities 2

1.2 Metzincin metalloproteases have roles in inflammation and immunity 4

Meprin metalloproteases 6

1.3 Meprin classification 6

1.4 Meprin domain structure and oligomerization 7

1.5 Meprin tissue expression 12

1.6 Meprin substrate specificities 14

1.7 Meprin activation and inhibition 14

1.8 Meprins in disease and inflammation 15

Escherichia coli 19

1.9 The host response to E. coli 19

1.10 Urinary tract infection 20

1.11 Bacterial pili 20

1.12 Host defense mechanisms 21

1.13 The host response to LPS 23

Experimental aims 25

Chapter 2. Materials and methods 29

Animal models 29

2.1 Mice 29

2.2 Intravesical challenge with E. coli or LPS 30

2.3 Intraperitoneal challenge with LPS 31

2.4 Anesthesia 31

2.5 Elicitation and isolation of peritoneal exudate cells 32

Assays 32

2.6 Myeloperoxidase activity 32

2.7 Assessment of bladder permeability 33

2.8 Blood collection and temperature monitoring 33

2.9 Quantification of blood urea nitrogen, creatinine, and nitrate/nitrite 34

2.10 Quantification of serum TNFα and cytokines/chemokines 35

2.11 Quantification of total protein 35

E. coli and pili 36

2.12 Growth conditions and pili enrichment 36

2.13 Agglutination assays 36

2.14 Cleavage of recombinant pili 37

2.15 Antimicrobial assays 38

2.16 Interaction of meprin with bacteria 38

2.17 Whole bladder binding assays 38

vii

Cell culture 39

2.18 Maintenance of cells 39

2.19 E. coli binding assays 39

2.20 Meprin treatment of kidney cells 40

2.21 Immunocytochemistry, confocal microscopy, and extraction of tight junction

proteins 40

2.22 Cytotoxicity assays 41

Meprins 42

2.23 Meprin purification, activation, and activity assays 42

2.24 Western blot analysis 43

2.25 Urinary dot blot development 43

Human specimen processing 44

Statistical analysis and graphing software 44

Chapter 3. Meprin A modulates the host response to E. coli bladder challenge 45

Results 45

3.1 Human urine studies 45

3.1.1 Meprin expression is increased in urine from women with active UTIs 45

3.1.2 Development of a dot blot assay expedites human sample processing 47

3.2 Murine model of UTI 50

3.2.1 Meprin αKO mice have less leukocyte infiltration into the bladder

after UTI challenge 50

3.2.2 Meprin αKO mice have less bladder edema after UTI challenge 50

viii

3.2.3 Meprin αKO mice display less bladder permeability after UTI

challenge 56

3.2.4 Bacterial counts in bladders and kidneys are variable in both genotypes 56

3.3 Exploring direct interaction between E. coli and meprins 58

3.3.1 Meprins do not interfere with E. coli agglutination of red blood cells 58

3.3.2 Meprins do not decrease the binding of E. coli to cultured bladder cells 62

3.3.3 Binding of E. coli to WT and meprin αKO bladders is equivalent 64

3.3.4 Recombinant pili are resistant to meprin cleavage 64

3.3.5 Meprins and E. coli viability 66

3.3.6 Meprin and E. coli activation assays 66

3.3.7 Urine from meprin αKO animals possesses decreased bactericidal

capability 66

Discussion 68

3.4 Urinary meprin expression in healthy and UTI afflicted women 70

3.5 Meprin A has a pro-inflammatory role in the E. coli-challenged bladder 71

3.6 Meprin A contributes to increased bladder permeability and edema after

E. coli challenge 72

3.7 Correlation between bacterial counts and severity of urinary tract disease 73

3.8 Persistence of E. coli in the urinary tract 74

3.9 Pros and cons of the mouse model of ascending UTI 75

Concluding remarks 76

Chapter 4. Meprin modulates the host response to lipopolysaccharide 78

Results 78

4.1 Response to systemic LPS 78

4.1.1 Meprin A contributes to LPS-induced renal injury 78

4.1.2 Meprin A contributes to LPS-induced hypothermia 82

4.1.3 Meprin A contributes to LPS-induced elevation of serum

nitrate/nitrite levels 82

4.1.4 Serum cytokine profiles are significantly different in WT and KO

animals 85

4.1.5 Serum TNFα levels are significantly different in WT and KO animals 92

4.2 Response to bladder LPS 92

4.2.1 Meprin A contributes to the leukocyte response and edema in a model

of bladder inflammation 92

4.2.2 Meprin A contributes to increases in bladder permeability after an LPS

bladder challenge 94

4.2.3 Meprin A disrupts the tight junctions between kidney cells 97

4.2.4 Less leukocyte migration into the peritoneal cavity in meprin αKO mice 100

Discussion 100

4.3 The role of TNFα and IL-1β in kidney damage 102

4.4 Cytokine-mediated hypothermia 103

4.5 Cytokine levels and iNOS release 103

4.6 The role of cytokines in inflammatory disease models 104

4.7 LPS-induced inflammation and leukocyte recruitment 105

4.8 Barrier function of the bladder and disruption of tight junctions 106

Concluding remarks 108

Chapter 5. General Discussion 110

5.1 The inflammatory response in systemic versus localized and infectious versus

noninfectious challenge models 111

5.2 Meprin participation in the inflammatory response: acute versus chronic 112

5.3 Endotoxin and IBD 116

5.4 Meprins modulate barrier function 117

5.5 Meprins in wound healing 118

5.6 Meprin protein-protein interactions 119

5.7 Proposed role for meprins in organ-specific and systemic immune challenges 121

5.7.1 Bladder challenge 121

5.7.2 Systemic challenge 126

5.8 Closing 127

Appendix 1: Ketamine anesthesia attenuates the inflammatory response in mice 128

Appendix 2: A comparison of blood pressures in meprin βKO and WT mice 140

Appendix 3: The effect of estrogen on meprin β expression in breast cancer cells 147

Reference List 158

List of Figures

Figure 1. The five classes of proteases 3

Figure 2. Meprin isoforms: classification and oligomerization 8

Figure 3. Domain structure of meprin subunits α and β 9

Figure 4. Electron micrographs of meprin oligomers 11

Figure 5. Host defense mechanisms against invading pathogens 22

Figure 6. Lipopolysaccharide 24

Figure 7. Urinary meprin expression is increased during active UTI 48

Figure 8. Dot blots of human and mouse urine samples 49

Figure 9. Meprin αKO mice have less leukocyte infiltration after UTI challenge 51

Figure 10. Meprin αKO mice have less bladder inflammation after UTI challenge 52

Figure 11. Meprin αKO mice have less bladder edema after UTI 55

Figure 12. Less bladder permeability in meprin αKO bladders after UTI 57

Figure 13. Bladder and kidney bacterial colonization was measured 48 h after UTI

induction 59

Figure 14. Meprins do not decrease the binding of E. coli to bladder cells in culture 63

Figure 15. Recombinant pili are resistant to cleavage by meprins 65

Figure 16. Meprins are resistant to activation by uropathic E. coli 67

Figure 17. Urine from meprin αKO mice demonstrates impaired bactericidal properties 69

Figure 18. BUN levels in meprin KO versus WT mice after LPS challenge 79

Figure 19. BUN levels are similar in meprin αKO mice and WT mice challenged

with 5mg/kg i.p. LPS 81

xii

Figure 20. Body temperature change in meprin KO versus WT mice after LPS

challenge 83

Figure 21. Serum nitrate/nitrite levels in KO versus WT mice after LPS challenge 84

Figure 22. Serum cytokine levels in KO compared to WT mice after LPS challenge 86

Figure 23. TNFα levels are significantly different between genotypes after LPS

challenge 93

Figure 24. Less bladder MPO activity and edema in meprin αKO versus WT

after transurethral administration of LPS 95

Figure 25. Less bladder permeability in meprin αKO mice after LPS challenge 96

Figure 26. Meprin A disrupts the tight junctions of kidney cells in culture 98

Figure 27. No evidence that meprin directly cleaves tight junction proteins

ZO-1 or occludin 99

Figure 28. Less cell infiltration into the peritoneal cavities of meprin αKO mice 101

Figure 29. Meprins in disease and inflammation 114

Figure 30. Consequences of the host response to LPS 122

Figure 31. Proposed role of meprin in bladder inflammation 124

Figure 32. Organ bacterial counts 48 h after high load E. coli challenge 131

Figure 33. Organ bacterial counts 7 days after high load E. coli challenge 132

Figure 34. Organ bacterial counts 48 h after low load E. coli challenge 134

Figure 35. Pilot UTI experiment utilizing isofluorane anesthesia 135

Figure 36. Ketamine anesthesia attenuates LPS-induced acute renal failure 137

Figure 37. Diagram of blood pressure equipment set-up 141

Figure 38. Principles of blood pressure measurement 143

xiii

Figure 39. Blood pressures in meprin βKO and WT mice 145

Figure 40. Meprin β gene is hormone responsive 148

Figure 41. Summary of exponentiated (raw) mRNA data 154

Figure 42. Zymography of MCF7 and MTR1 conditioned media 155

xiv

List of Tables

Table 1. MMP-deficient mice and their associated immune/inflammatory phenotypes 5

Table 2. Meprins are implicated in multiple diseases 17

Table 3. Human urine chemistry 46

Table 4. Meprins do not interfere with bacterial agglutination of RBCs 60

Table 5. Serum cytokine levels in KO compared to WT mice after LPS challenge 88

Table 6. Inflammatory diseases 113

Table 7. Mouse physiological data 144

Table 8. Summary of transformed mRNA data 153

List of Abbreviations

ADAM a disintegrin and metalloprotease

ANOVA analysis of variance

AP-1 activating protein 1

ARF acute renal failure

BALF bronchoalveolar lavage fluid

BBB blood brain barrier

BMP-1 bone morphogenic protein 1

BSA bovine serum albumin

BUN blood urea nitrogen

C cytosolic domain

CAM cell adhesion molecule

CCL CC chemokine ligand

CFU colony forming units

COX-2 cyclooxygenase-2

CXCL CXC chemokine ligand

DBP diastolic blood pressure

DC dendritic cell

DN diabetic nephropathy

DSS dextran sulfate sodium

E. coli Escherichia coli

ECM extracellular matrix

EDTA ethylene diamine tetraacetic acid

xvi

EGF epidermal growth factor

ELISA enzyme-linked immunosorbent assay

EMB eosin methylene blue

ENaC epithelial sodium channel

FBS fetal bovine serum

Gal(α1-4)Gal globotriasylceramide

GMCSF granulocyte-macrophage colony stimulating factor

H&E hematoxylin and eosin

HEK human embryonic kidney

HR heart rate

HTAB hexadecyltrimethylammonium bromide

I inserted domain i.p. intraperitoneal

IBD inflammatory bowel disease

ICAM-1 β2-intracellular adhesion molecule 1

IFN interferon

IL interleukin iNOS inducible nitric oxide synthase

Ki inhibition constant

KO knockout

LB Luria Bertani

LDH lactate dehydrogenase

LPS lipopolysaccharide

xvii

MAM meprin A5 protein tyrosine phosphatase μ

MBP mannan-binding protein; mean blood pressure

MCP-1 monocyte chemoattractant protein 1

MDCK Madin Darby canine kidney

MEM Modified Eagle’s Medium

MeOH methanol

MIP-1α/β macrophage inflammatory protein 1 α/β

MMP matrix metalloprotease

MPO myeloperoxidase

MTR MCF7 tamoxifen resistant

MTT 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide

NaCl sodium chloride

NaFl sodium fluorescein

NCX sodium calcium exchanger

OD optical density

PAGE polyacrylamide gel electrophoresis

PBS phosphate buffered saline

PCR polymerase chain reaction

PEA-3 polyoma enhancer activator 3

PEC peritoneal exudate cells

PMSF phenylmethylsulphonyl fluoride

RANTES regulated on activation normal T cell expressed and secreted

RNA ribonucleic acid

xviii

S signal sequence

SBP systolic blood pressure

SDS sodium dodecyl sulfate

SE standard error

Serpin serine protease or peptidase inhibitor

STI soybean trypsin inhibitor

TACE tumor necrosis factor α (TNFα) converting enzyme

TAPI-2 TNF a protease inhibitor 2

TBS tris buffered saline

TBS-T tris buffered saline with 0.1% Tween-20

TECK thymus expressed chemokine

TG thioglycollate

TGF transforming growth factor

TIMP tissue inhibitor of metalloproteases

TLD tolloid protein

TLR-4 Toll-like receptor 4

TM transmembrane domain

TNFα tumor necrosis factor α

TRAF tumor necrosis factor receptor associated factor

UTI urinary tract infection

WT wild-type

xix

Acknowledgements

Writing the dissertation, summarizing several years of work in just a few hundred pages, is a task that manages to be daunting and invigorating all at once. The writing process, combined with the pressure and sleep deprivation, inevitably leads to some self- reflection, a bit of philosophical waxing, and the creation of the acknowledgements section.

The dissertation, with its crisp, white pages bound in glossy, faux black leather is stunning and impressive on the shelf, but it does not capture the truly defining moments in one’s graduate career. The late timepoints, the excitement felt when data are actually reproducible, the hours spent on data analysis and interpretation, the brainstorming and idea sharing, the blood (from mouse bites), sweat, and tears—these are the things that are remembered, the events that help to mold a naïve student into a well-rounded researcher.

These experiences are the result of personal and professional interactions with a special group of mentors, colleagues, and friends who deserve recognition and thanks.

First and foremost, I would like to express gratitude to my mentor and thesis advisor

Dr. Judith S. Bond. Dr. Bond gave me a great deal of scientific freedom, which allowed me to create a body of work that I could truly call my own. She taught me the importance of collaboration and networking with the scientific community and provided me with the opportunity to share my work at national and international meetings.

I would also like to thank Dr. S. Gaylen Bradley for what he refers to as his constant

“nagging.” His continued interest in my work, encouragement regardless of success or failure, and brilliant ideas always kept me motivated both professionally and personally.

My Graduate Committee members, Dr. Robert Levenson, Dr. Sarah Bronson, Dr.

Andrea Manni, and Dr. Brian Reeves, were always helpful and supportive throughout the

years as my thesis work changed paths and took shape. I am especially grateful to Dr.

Reeves for the fruitful collaboration that exists between our laboratories. Many of my experiments would not have been possible without the scientific equipment and technical lessons provided by Dr. Ganesan Ramesh from the Reeves laboratory.

My technical expertise and dissertation work were also greatly enriched by collaboration with Dr. David Antonetti. Dr. Antonetti and I established a successful collaboration resulting from his brainchild, the Graduate Student Forum. He and his graduate student, Dr. Jeffrey Sundstrom, contributed a great deal of time as well as reagents to teaching me the basics of immunocytochemistry and confocal microscopy. Additional assistance on the confocal microscope was provided by Dr. Alistair Barber and the always patient Rhona Ellis.

The mouse model of UTI, a foundation of this dissertation work, would not have been possible without the help and collaboration of Dr. Rodney Welch at the University of

Wisconsin. Dr. Welch’s laboratory staff, including Andrew Anfora, Brian Haugen, and

Shahaireen Pellett, spent a great deal of time and care in planning their experiments so that I could observe, learn, and practice all steps of the 48 hour model in my 8 hour visit to their lab. I would also like to thank Dr. Welch for keeping track of me as my flights were delayed, canceled, delayed, and canceled, and for finally picking me up 7 hours later than expected at the airport. On a side note, avoid connections at Chicago’s O’Hare airport at all costs!

I express my appreciation to past and present members of the Bond lab, especially Dr.

Gail Matters and Dr. Susan Ishmael, who taught me many of my fundamental laboratory skills during my rotation and my first year in the lab, respectively. My colleagues in the

Bond lab are a special group of people that make working in the lab a pleasure. I would also

xxi

like to acknowledge Ge Jin, Jialing Bao, and Dr. Sanjita Banerjee, who I consider not only colleagues but also good friends. Sanjita and I became fast friends from early on in our graduate careers, and I cannot imagine attending a scientific conference without having her as my hotel roommate and sightseeing partner.

Finally, I want to recognize the constant support and encouragement that I received and continue to receive from my parents, Marsha and Steve Dusheck, and my husband,

Christopher Yura. I have to thank my father; I seem to have inherited his creative mind, and it has helped me to blossom in science. At the same time, I think that I inherited my optimism and patience from my mother, who managed to teach small children for 35 years without losing those traits…or her mind. In addition to supporting me by moving to Hershey with me after college, my husband has helped me physically (recording data and numbering countless Eppendorf tubes late into the evening), artistically (designing me an unforgettable cell junction T-shirt to wear to my candidacy exam), and emotionally. Chris has unfaltering faith in me and always assures me that I will achieve great things. My gratitude to him is beyond words.

xxii

Chapter 1. Introduction

Proteases

Proteolysis plays a crucial role in a wide array of biological processes, for example, regulatory functions such as digestion to produce available nutrients, hydrolysis to remove unneeded proteins, removal of signal peptides of proteins directed to the secretory pathway, and generation of active enzymes via removal of pro-sequences. Proteases are diverse in structure and function and are abundant in the mammalian genome (∼2% of the human genome is proteases). During embryogenesis and development, proteases participate in cleavage and release of growth factors and extracellular matrix (ECM) remodeling, thereby contributing to cell proliferation and differentiation. Proteases are also involved in hormone processing, cytokine activation and degradation, bone formation, signal transduction, blood coagulation, and apoptosis. Proteases contribute to the pathology of multiple diseases, including arthritis and Alzheimer’s disease. During cancer progression, proteases modulate the tumor microenvironment, playing a role in angiogenesis, invasion, and metastasis (1).

Protease expression is dynamic, and their roles and activation status vary depending on the type of cell and stage of development (1). While most proteases are non-specific and capable of cleaving a wide variety of substrates, their selectivity is maintained via a high level of regulation. Proteases are regulated at the transcriptional and post-transcriptional levels, and their catalytic activity is tightly controlled through tissue-specific localization, compartmentalization, concentration, and pro-enzyme activation. For example, the cathepsins are lysosomal cysteine proteases that are activated by the low pH in the lysosomes. The membrane-bound matrix metalloproteases (MMPs), such as MMP14, are an example of compartmentalization, wherein their anchorage in the membrane maintains a high

concentration of catalytic activity locally (2). Protease activity is also regulated by endogenous inhibitors such as serpins (serine protease or peptidase inhibitors) and TIMPs

(tissue inhibitors of metalloproteases), which are specific for MMPs and some proteases containing a disintegrin and a metalloprotease domain (ADAMs). While the number of proteases identified in the human genome is approximately 600, there are less than 200 known protease inhibitors, and endogenous inhibitors for many proteases have not been identified.

Proteases are grouped into five classes according to their mechanism of action and active site residues: serine, threonine, cysteine, aspartic, and metalloproteases (Figure 1).

The first three classes of proteases cleave peptide bonds by covalent catalysis, in which a powerful nucleophile in the catalytic triad of the active site becomes temporarily covalently modified. In contrast, the aspartic and metalloproteases coordinate a water molecule to attack the peptide bond. The majority of metalloproteases contains a zinc ion in the active site that activates the water molecule, with four to five ligands to coordinate the zinc. The metalloproteases are the most abundant protease class, with over 200 members identified thus far. The metalloproteases have diverse functions and are present in a wide variety of tissues and species, but their conserved binding site motifs and evolutionary origins permit their organization into smaller subgroups.

Metzincins

1.1 Metzincin metalloproteases share structural similarities

The zinc metalloproteases are divided into 15 clans and over 50 families based on structure and evolutionary relationships (3). The metzincin superfamily or MA(M) subclan

Figure 1. The five classes of proteases. Proteases are grouped into five classes based on their mechanism of action and active site residues: serine, threonine, cysteine, aspartic, and metalloproteases. Several examples of each protease class are indicated under each heading.

Recently, a sixth class of proteases, glutamic acid proteases, has also been identified. Thus far, the glutamic acid proteases have only been identified in bacteria (e.g., aspergilloglutamic peptidase) (http://merops.sanger.ac.uk/).

is comprised of four distinct evolutionary families: astacins, ADAMs/reprolysins, serralysins, and MMPs/matrixins (4). While there is little or no amino acid homology among the families, there is striking three-dimensional similarity (4). The metzincins are distinguished by a characteristic ‘met turn’ beneath the active zinc binding site and a highly conserved zinc binding motif HEXXHXXGXXH (5).

1.2 Metzincin metalloproteases have roles in inflammation and immunity

Metalloproteases of the metzincin superfamily have roles in many stages of life and in multiple processes, such as angiogenesis, metastasis, blood pressure regulation, digestion, and ECM remodeling. Of particular interest to this thesis are their roles in the immune system, as the metzincins have been implicated in many aspects of the inflammatory response and immune modulation, including wound healing and epithelial repair, activation and degradation of chemokines and cytokines, leukocyte migration and activation, and ectodomain shedding (2,6). Shedding is an important regulatory mechanism for many membrane-bound immune molecules, including cytokines, cytokine receptors, and adhesion molecules. Greater than 40 cell-surface proteins (e.g., TGFβ, TNFα) are known to undergo ectodomain shedding by proteolytic cleavage, and the majority of shedding activity is mediated by metalloproteases of the metzincin superfamily (1,7). Metalloproteases have also been implicated as determinative factors in bacterial infections, in part attributed to the degradation of ECM and to modulation of interleukin levels (8).

Much of what is currently known regarding the role of MMPs in inflammatory processes has been gleaned from studies with MMP knockout (KO) mouse models (Table 1).

For example, several MMPs are expressed during wound healing, and MMP3 KO mice are

Table 1. MMP-deficient mice and their associated immune/inflammatory phenotypes.

MMP KO mouse models display unique phenotypes in various challenge models. These data

indicate functions for the MMPs in tissue repair, host defense, angiogenesis, tumor

progression, and inflammation. *, indicates phenotypes that were reversed following transplantation of WT bone marrow, suggesting that the effect observed in KO mice was

caused by the lack of MMP in inflammatory cells; ND, not determined.

Table adapted from Parks, W. C., Wilson, C. L., and Lopez-Boado, Y. S. (2004) Nat. Rev.

Immunol. 4, 617-629.

defective in wound repair of the epidermis (9). MMP9-deficient mice demonstrate markedly decreased leukocyte recruitment after E. coli peritonitis challenge in spite of locally increased chemokine concentrations (8). In the mouse intestine, MMP7 activates a class of antimicrobial peptides called defensins, and MMP7-deficient mice demonstrate an impaired ability to kill enteric pathogens (2).

Several lines of evidence also implicate immune roles for the ADAMs, via cell-cell interactions, modification of ECM, and shedding activity (10). For example, ADAM19 expression is enhanced during the maturation of both dendritic cells (DCs) and T cells (11).

Data indicate that ADAM15 is a mediator of rheumatoid arthritis, angiogenesis, and intestinal inflammation (10). TNFα converting enzyme (TACE), arguably the best known of the ADAMs, releases TNFα from cells, and TACE inhibition has been shown to be efficacious in attenuating inflammatory disease models (e.g., LPS-induced airway inflammation, rheumatoid arthritis). TACE also mediates the shedding of intercellular adhesion molecule 1 (ICAM-1), a cell adhesion molecule that promotes leukocyte adhesion and transmigration during inflammation (7). While multiple studies have documented participation of MMPs and ADAMs in the host response to an immune system challenge, the function of the meprin metalloproteases in modulation of the immune response is not well- defined (12).

Meprin metalloproteases

1.3 Meprin classification

Meprins are members of the “astacin family of metalloproteases,” named for the first family member to be biochemically characterized, a digestive protease from the crayfish

(13). Astacin metalloproteases have been identified in a diverse set of organisms, from flavobacteria to humans, and include carp nephrosin, xenopus hatching enzyme, bone morphogenic protein 1 (BMP-1), and tolloid protein (TLD). The protease domains of astacins share many features, including a signature 18 amino acid sequence and a zinc- binding motif. The majority of astacin family members possess signal and pro-sequences, as well as epidermal growth factor (EGF)-like domains (13). Many of the larger family members also contain several additional non-catalytic domains, which often promote protein- protein or protein-substrate interactions.

1.4 Meprin domain structure and oligomerization

The meprins are unique in the astacin family in that they are the only known astacin members to form homo- and hetero-oligomers of two evolutionarily related subunits, α and β

(13) (Figure 2). They are also unique in that the subunits form disulfide-linked dimers that are capable of associating non-covalently to form tetramers and higher order oligomers. The meprin α and β subunits are highly glycosylated, multidomain proteases that contain a signal sequence (S), pro-sequence, protease domain, meprin A5 protein tyrosine phosphatase μ

(MAM) domain, tumor necrosis factor receptor associated factor (TRAF) domain, EGF-like domain, transmembrane (TM) domain, and a cytosolic (C) domain (14) (Figure 3). The protease domain (the catalytic astacin domain) contains the active site zinc, whereas the

MAM and TRAF domains are essential for proper folding and transport of the stable, mature enzyme (15). The meprin α subunit contains an additional inserted (I) domain between the

EGF and TRAF domains that is not present in the β subunit. Thus, while the meprin β subunit is a Type 1 membrane-bound protein, the meprin α subunit is proteolytically

Figure 2. Meprin isoforms: classification and oligomerization. Meprins, which exist in both membrane-bound and secreted forms, form oligomers that are unique in the astacin family of proteases. Dimers of membrane-bound β subunits are referred to as meprin B, whereas tetramers of α and β subunits are referred to as heteromeric meprin A and are membrane-bound due to the presence of the β subunit. The soluble form of meprin, homomeric meprin A, is capable of forming large multimers up to 8 MDa in size.

Figure 3. Domain structure of meprin subunits α and β. The meprins are multidomain

proteases that contain a signal sequence (S), pro-sequence, protease domain, meprin A5

protein tyrosine phosphatase μ (MAM) domain, tumor necrosis factor receptor associated

factor (TRAF) domain, epidermal growth factor-like (EGF) domain, transmembrane (TM)

domain, and a cytosolic (C) domain. The meprin α subunit contains an additional inserted (I)

domain between the EGF and TRAF domains. Thus, while the meprin β subunit is a Type 1

membrane-bound protein, the meprin α subunit is proteolytically processed at the I domain

(black arrow) during biosynthesis and loses its transmembrane domain.

processed at the I domain during biosynthesis and loses its transmembrane domain (16).

Therefore, meprin α is only present at the membrane by virtue of association with the membrane-bound meprin β subunit.

The meprin subunits differ markedly in their ability to form large oligomers. In vitro studies with recombinant meprin demonstrate that meprin α homodimers can form heterogeneous multimers (ring-, circle-, spiral-, and tube-like structures) containing up to

100 subunits, depending on protein concentration, ionic strength, and activation state (Figure

4A, B). Homomeric meprin A’s ability to form 1-6 MDa complexes allows it the designation

of the largest known secreted protease (17). In contrast, meprin β subunits form dimers

under a wide range of conditions (Figure 4C, D). Meprin αβ heterodimers tend to form

tetramers but not higher oligomers, indicating that the presence of the meprin β subunit

restricts the oligomerization potential of the meprins (Figure 4E, F) (17).

All isoforms of meprin that contain the α subunit are called meprin A. Membrane-

bound forms of meprin include meprin B (dimers of β subunits) and heteromeric meprin A

(non-covalent tetramers of disulfide-linked meprin α and β heterodimers) (Figure 3) (17,18).

However, there is evidence that a portion of heteromeric meprin A is released or shed into

the human intestinal lumen (19). The majority of homomeric isoforms of meprin A

(multimers of α subunits) are secreted from cells. In the meprin αKO mouse, only

membrane-bound dimeric meprin B is present, whereas in the meprin βKO mouse, only

homomeric meprin A is present, and it is secreted from cells as high molecular mass

multimers (17,20).

The meprin subunits are highly glycosylated, with carbohydrates contributing to 20%

of the molecular mass of the meprin α subunit (21). Of the ten potential N-glycosylation

Figure 4. Electron micrographs of meprin oligomers. Representative fields of negatively stained samples of each recombinant rat meprin isoform are shown. (A and B) latent and active homo-oligomeric meprin A, respectively; (C and D) latent and active meprin B, respectively; (E and F) latent and active hetero-oligomeric meprin A, respectively. Figure adapted from Bertenshaw, G. P., Norcum, M. T., and Bond, J. S. (2003) J. Biol. Chem. 278,

2522-2532.

sites on meprin α, nine were found to be glycosylated in recombinant mouse meprin A.

Removal of various combinations of the glycans decreased the stability, hindered disulfide bond formation, and abolished enzymatic activity of the protease (22). These glycosylations, especially in the TRAF domain, may facilitate the interaction of meprins with other extracellular proteins or glycan-binding proteins.

1.5 Meprin tissue expression

Although meprin α and β are 47% identical in amino acid sequence, they are encoded

on different chromosomes and demonstrate differential expression patterns and substrate

specificities (23). The meprin α subunit is encoded on mouse chromosome 17 and human

chromosome 6p, whereas the β subunit is encoded on chromosome 18 in human and mouse

(24,25). Meprins are abundantly expressed in the brush border membranes of kidney

proximal tubule cells and intestinal epithelial cells in mouse, rat, and human (12). Meprins have also been identified in cancer cells and certain populations of leukocytes (26-28).

Meprin expression is tissue and cell specific, and this localization implicates them in the interface between the host and its environment, playing roles in normal and inflammatory conditions.

Meprin expression in the mouse kidney is strain dependent, with different strains of

inbred mice exhibiting different levels of kidney meprin activity. For example, C3H, CBA,

CHI, and AKR mice have approximately 100-fold lower renal meprin activity in comparison

to C57BL/6, DBA, and 129 mice (29,30). Reciprocal F1 hybrids produced by crossing a

normal mouse strain with a meprin-deficient mouse strain demonstrate that meprin deficiency is inherited as an autosomal recessive trait. Although meprin expression in the

adult mouse kidney varies among strains, in situ hybridization studies revealed comparable embryonic expression patterns of meprin α and β (31).

The meprins are highly regulated during fetal growth and development in mice and rats. Meprin α is present in the intestine of the mouse fetus and increases initially after birth.

However, intestinal expression is markedly decreased after weaning. In contrast, meprin β is the predominant isoform present in the mouse intestine after weaning (31). In addition to the kidney and intestine, meprins have been detected in other mouse tissues, including the salivary glands, heart, and mesenteric lymph nodes (28,32).

In the rat kidney, meprin β RNA expression has been identified as early as embryonic day 13, with levels increasing until adulthood (33). In the rat jejunum, meprin β expression is high at postnatal day 4, decreases during the suckling period, and returns to a high level at weaning. Meprin α expression in the jejunum, in contrast, is constant throughout suckling but decreases at weaning. Meprins have also been detected in neuroepithelial cells, the inner ear, brain ventricles, choroids plexus, developing lens, gut, bladder, and ureter in rat embryos

(34).

In humans, meprin α expression has been detected in the kidney, colon, small

intestine, and appendix. Meprin expression in the human kidney appears to be variable, and

there is reason to believe that many humans have low kidney meprin expression (E.E.

Sterchi, personal communication). Recently, independent expression of the meprin α and β

subunits has been identified in the stratum basale and stratum granulosum of human

epidermis, respectively (35). Meprin α and β subunits are also expressed in leukocytes in the

lamina propria of the human intestine (26).

1.6 Meprin substrate specificities

The bulk of the meprin protease is extracellular, and its presence at the cell surface potentially enables proteolysis of growth factors, cytokines, hormones, ECM proteins, and bioactive peptides (12,36,37). Both meprin A and B are capable of hydrolyzing extracellular matrix proteins, including laminin, collagen IV, nidogen-1, and fibronectin, which are essential for tissue remodeling, wound repair, and leukocyte extravasation (36,37). However, the peptide bond specificities of the meprins are substantially different. Meprin α demonstrates a preference for small or hydrophobic residues and is capable of hydrolyzing bradykinin, substance P, angiotensin, parathyroid hormone, and neurotensin (37-40). Gastrin and osteopontin are two of the best substrates for meprin β, which preferentially cleaves peptide bonds flanked by acidic residues, and has also been shown to hydrolyze orcokinin, kinetensin, glucagon, and secretin (37).

Meprin A has been reported to digest the N termini of several cytokines and chemokines in vitro, including CCL3 (MIP-1α), CCL4 (MIP-1β), CCL5 (RANTES), and

CCL2 (MCP-1), which would render them biologically inactive. Meprin B completely degrades the chemokine CCL25 (TECK) and activates pro-IL-1β (41,42). The ability of meprins to mediate cytokine and ECM functionality via proteolytic cleavage implicates potential roles for the meprins in immune processes such as leukocyte regulation.

1.7 Meprin activation and inhibition

Meprins are synthesized as latent pro-enzymes. This allows for concentration and delivery of proteolytic activity at the membrane and extracellularly without the consequence of proteolysis. In the proximal tubule epithelial cells of the mouse kidney, membrane-

associated meprin α is active while the majority of meprin β subunits are in the latent form

(13). Both meprin α and β are active in the rat kidney, as well as in the mouse and human intestines. Although the mechanism of meprin activation in vivo is unknown, in vitro studies have demonstrated that trypsin-like enzymes are capable of activating meprin subunits.

Meprin α and β are likely activated by trypsin in the intestine, and meprin α can also be activated by plasmin (43-45). Recent studies indicate that human kallikrein-related peptidase

4 is capable of activated pro-meprin β in vitro (35).

Meprins, like all metalloproteases, can be inhibited by metal chelators such as EDTA.

The best meprin α inhibitor identified is actinonin, a peptide hydroxamate (46). Actinonin inhibits human meprin α with a Ki of 20 nM and meprin β with a Ki of 1.7 μM (36). The meprins are not inhibited by TIMPs or by inhibitors of the other protease classes (13).

However, they can be inhibited by several broad-range MMP inhibitors (e.g. TAPI-2 and

GM6001) (47). Recently, meprins have been identified as endogenous ligands of mannan- binding protein (MBP), a host defense lectin that plays a role in innate immunity via activation of the complement system. MBP is the first endogenous meprin inhibitor identified, and its ability to decrease the proteolytic activity of meprin has been demonstrated in vitro and in vivo (48). This implies that regulation of meprin activity could be important in the early phases of the host immune response.

1.8 Meprins in disease and inflammation

Mouse models of inflammation indicate a relationship between chemokine expression

and the intensity of the host inflammatory response as well as leukocyte migration from the circulation to sites of inflammation (49). Several lines of evidence support a potential role

for meprins in modulation of these processes in chronic and acute diseases (Table 2). For example, meprins have been implicated in pathological conditions such as acute renal failure and inflammatory bowel disease (IBD) (28,50,51).

Meprin β has been implicated as a candidate gene for diabetic nephropathy in the

Pima Indian population, as significant within-family associations for single nucleotide polymorphisms (SNPs) in the MEP1B gene and the disease have been demonstrated (52).

Mouse models of kidney ischemia/reperfusion (I/R) demonstrate a redistribution of the meprin subunits in the kidney after renal challenge, and mice deficient in meprin β exhibit less injury in renal ischemia/reperfusion (51). In rodent models of hydronephrosis and

Alport’s disease (a genetic disorder in basement membrane proteins that leads to kidney fibrosis), meprin expression is dramatically decreased (53,54). In these systems, the altered turnover of ECM proteins and chemokines resulting from meprin’s downregulation is implicated in the subsequent renal damage. The renal gene expression profile of the meprin

βKO mouse in comparison to WT kidney indicates a potential role for the meprins in immune signaling. For example, TRAF6, toll-like receptor 6 (TLR6), and JUN-N terminal kinase binding protein 1 are significantly increased in meprin βKO mice (20). TRAF6 is an adaptor protein that mediates TLR signaling, which serves important biological functions including host pathogen recognition.

Polymorphisms in the human MEP1A gene have been correlated with Crohn’s disease and ulcerative colitis, and both α and β are expressed in the lamina propria of the human inflamed bowel. Studies examining a dextran sulfate sodium (DSS)-induced model of experimental colitis demonstrate greater susceptibility to injury and inflammation in the meprin αKO mice in comparison to WT (55). During intestinal inflammation, meprin α, but

Disease Meprin Model system References Acute renal failure (ARF) α, β Rodent (50,51,56-58) • Ischemia reperfusion injury (I/R) • Cisplatin nephrotoxicity Diabetic nephropathy (DN) α, β Rodent (52,59-61) • db/db mice • Adriamycin- induced • Puromycin- induced Human • Pima Indian population Hydronephrosis α, β Rodent (53) • Ureteral obstruction Diabetes α, β Rodent (61) • Streptozotocin- induced • db/db mice Inflammatory bowel disease α, β Rodent (26,28) (IBD) • DSS-induced Human • Inflamed bowel • Polymorphisms Breast cancer α, β prime Human cell lines (41,62-64) Colon cancer α Human (27,41,44) • Colorectal tumors • Cell lines Hyperkeratosis α, β Human skin (35)

Table 2. Meprins are implicated in multiple diseases. Meprins have been implicated in chronic and acute diseases of the kidney as well as inflammatory bowel disease and cancer.

Expression of meprin in pathological conditions is complex, as either or both of the meprin subunits may be up or downregulated. The potential role(s) and differential expression patterns of the meprin subunits are dependent on the disease, the organs or tissues involved, and the timepoints that are examined.

not β, are decreased in leukocytes from the mesenteric lymph nodes of mice. Deletion of the meprin β gene also decreased the ability of leukocytes to migrate through matrigel, implicating the membrane-bound meprins in leukocyte transmigration (28).

Both meprin α and meprin β have been implicated in cancer and tumor metastasis.

Evidence indicates differential regulation of meprin in cancer cells, and a unique isoform of meprin β mRNA is expressed in a variety of cancer cell lines (65). Further, the meprin β promoter region contains PEA and AP-1 sites, and meprin β mRNA is upregulated in cancer cells after treatment with phorbol esters (62). Meprin α is also expressed in multiple cultured human cancer cell lines, including small intestine-like Caco-2 cells and MDA-MB-435 breast cancer cells (27,64). MDA-MB-435 cells treated with actinonin, a potent meprin α inhibitor, were less invasive in vitro, implicating meprin α in cell migration and invasion (64).

Escherichia coli

1.9 The host response to E. coli

Microbial pathogenesis involves the interplay between the environment, the nature of the microbe and the state of the host (66). The bacteria must have the ability to overcome multiple host defenses in order to adhere successfully to and invade the tissue of the urinary tract. Many genetic and immunological host factors determine the outcome of introduction and persistence of bacteria in the bladder. E. coli, the most common causative agent of urinary tract infection (UTI), elicits a strong inflammatory response upon colonization of the host bladder, including upregulation of pro-inflammatory cytokines and leukocyte influx, which is critical for timely resolution of the infection (67).

1.10 Urinary tract infection

UTIs are one of the most common bacterial infections for women, and approximately

50% are infected at one point in their lifetime. Over 80% of women who have had a UTI will experience recurrent episodes, and frequent UTIs are common in young women who are otherwise healthy (68). Multiple biologic and behavioral factors have been implicated in susceptibility to recurrent UTI (e.g., antibiotic, spermicide use) (69). The vaginas of women with a history of recurrent UTIs are more likely to be colonized with E. coli than those of women with no history of UTIs. In addition, vaginal mucosal cells from women with a history of UTIs bind more uropathic E. coli than mucosal cells from women with no history of UTIs (70). Analysis of family pedigrees supports a familial genetic predisposition for UTI among young females (69). Taken together, these observations indicate that susceptibility to

UTIs may have a strong genetic component.

1.11 Bacterial pili

E. coli is the primary causative agent of UTIs and accounts for more than 80% of these infections (67). The ability of uropathic E. coli to invade successfully and colonize host cells can be attributed to multiple virulence factors, including hemolysin, toxins, and adhesive fimbriae or pili (71). Pili facilitate bacterial adhesion to the host cell, a first step in the pathogenesis of UTI (72). Pili enable the bacteria to bind to epithelial cells as well as deliver lipopolysaccharide (LPS) and other bacterial products to the host cell (73). Type-1 pili are the most common adhesin molecule and have been identified on greater than 90% of E. coli isolated from UTI patients (74). Type-1 pili bind mannosylated glycoproteins on the bladder urothelium and are enriched during bacterial bladder infection or cystitis. P pili recognize

globotriasylceramides (Gal(α1-4)Gal) and play a critical role in the pathogenesis of kidney infection or pyelonephritis (75). Bacterial adhesins also facilitate bacterial invasion of host cells. Internalization of the bacteria by host cells enhances bacterial survival and allows the bacteria to evade multiple host defenses.

1.12 Host defense mechanisms

The pathogenesis of uropathic E. coli in the bladder is thought to begin with binding of the bacterium to the urothelium by type-1 pili. The host has multiple defense mechanisms in place to defend against the invading bacteria in the bladder (Figure 5). Urine flow helps to wash away weakly adhering bacteria while the urine composition itself (e.g. slightly acidic pH and high osmolarity) is inhibitory to bacterial growth (76,77). The urine also contains antibacterial molecules such as defensins and Tamm-Horsfall protein. Tamm-Horsfall protein, the most abundant protein in urine, serves as an anti-adherence factor by binding to type-1 fimbriated E. coli (74). Defensins are a class of antimicrobial peptides that are capable of directly killing invading pathogens (78). If the bacteria manage to evade these defenses, the host cells respond with cytokine production, inflammation, and exfoliation of bladder epithelial cells. The bladder epithelium normally has an extremely low turnover rate, but infection by uropathic E. coli induces rapid cell exfoliation (67).

The bladder urothelium plays an important role in the host response to E. coli. In addition to exfoliation, it provides barrier function against invading organisms. Tight junctions in the uppermost layer of the bladder maintain the bladder barrier. During cystitis, epithelial cells are lost due to cell swelling and lysis, decreasing barrier function and allowing urea and other irritants to enter the muscle layers of the bladder (79). The bladder

Figure 5. Host defense mechanisms against invading pathogens. UTI is thought to begin by the binding of bacterial to the bladder epithelium by adhesive pili. Urine flow and cell exfoliation are initial attempts to prevent the subsequent adhesion, invasion, and replication of bacteria. If the bacteria successfully make it past the epithelial barrier, the host responds with rapid cell exfoliation and cytokine release to increase leukocyte influx to the site of infection. Figure adapted from Mulvey, M. A., Schilling, J. D., Martinez, J. J., and Hultgren,

S. J. (2000) Proc. Natl. Acad. Sci. U. S. A 97, 8829-8835 and Vander, A., Sherman, J., and

Luciano, D. (2001) Human Physiology: The Mechanisms of Body Function, 8th Ed., 2nd Ed.,

McGraw-Hill Companies, New York, NY.

urothelium also possesses multiple receptors that bind to components of the uropathic bacteria and initiate inflammatory signaling cascades (71). The subsequent release of cytokines guides neutrophil influx, which is essential for bacterial clearance from the urinary tract (80).

1.13 The host response to LPS

Studies have demonstrated that, even in the absence of bacteria, bacterial toxins alone can induce bladder inflammation (81). One such toxin is endotoxin or LPS, a glycolipid component of the gram-negative bacterial cell wall and a potent immune stimulus. It is composed of an O-antigenic polysaccharide, a core region, and the toxic center of LPS, lipid

A (Figure 6) (82). LPS recognition by the host cell involves the CD14 receptor and the pattern recognition molecule TLR4, which is expressed by epithelial cells and immune cells in the urinary tract (83). LPS is taken up by bladder urothelium and is capable of producing an inflammatory response similar to that of a UTI (81). The delivery of LPS by way of bacterial adhesins stimulates the release of pro-inflammatory cytokines such as IL-8, which attracts neutrophils to the site of infection (67,80). LPS is also thought to be involved in the upregulation of iNOS, adhesion molecules, and COX-2, which contribute to the pathology of cystitis (81,84).

During the pathogenesis of UTI, host receptors recognize bacterial components and a highly regulated cascade of events ensues, in which invading organisms are killed and damages are repaired without harming the host (85). In this situation, the LPS response benefits the host and is a critical contributor to the timely resolution of infection. However, if the infection continues to ascend the urinary tract, complications of UTI can include

Figure 6. Lipopolysaccharide. (a) An electron micrograph of E. coli. (b) A schematic representation of the organization of the gram-negative bacterial cell wall and the location of

LPS. (c) The architecture of LPS. The polysaccharide region of LPS is composed of the core region and the O-specific chain, whereas the lipid region contains lipid A. The sugars in the O-specific region are associated with immunological specificity, and the lipid A region is associated with the toxicity of LPS. (d) The primary structure of the lipid A component. The lipid A component contains the region of LPS that is anchored in the membrane and 6-7 saturated fatty acids. GlcN, D-glucosamine; Hep, L-glycero-D-manno-heptose; Kdo, 2-keto-

3-deoxy-octulosonic acid; P, phosphate. Figure adapted from Beutler, B. and Rietschel, E. T.

(2003) Nat. Rev. Immunol. 3, 169-176.

pyelonephritis, intrarenal abscess, and bacteremia with or without septic shock (76). In the

case of a systemic LPS challenge (e.g., bacteremia), cells responsive to bacterial components are activated throughout the body, and an exaggerated immune response is observed (85).

Much of the detrimental inflammatory response observed during sepsis is due to the dysregulation of cytokine production, especially macrophage-secreted TNFα (82,86,87).

Activated neutrophils, which are essential to resolving UTI, are detrimental during sepsis and

contribute to vascular and tissue damage (86). Thus, the balance between protection and

toxicity is delicate, and the ultimate outcome is determined by a variety of factors (82).

Although many of the key components controlling how the host senses and responds

to pathogens have been identified, not all mediators of the LPS response and uropathogenesis

have been fully characterized (76,82). Identification of factors that contribute to modulation

of the host response will provide a better understanding of how the immune system can

resolve infection with maximal damage to the invading pathogens and minimal damage to the

host.

Experimental aims

While much is known about the in vitro capabilities of meprin, less is known about

the in vivo roles of meprins in inflammatory diseases. This project examined the role of

meprin metalloproteases in the host inflammatory response to bacterial challenge. The

hypothesis was that meprins protect the host against bacterial infections either by directly

interacting with bacteria to prevent infection or by affecting the immune response to bacteria.

The effect of meprins on bacterial adhesion and viability as well as in the initiation and

progression of urinary tract infections in humans and mice were investigated. The aims of

this study were to: 1) Characterize meprin expression in human urine during active infection,

2) Determine whether the lack of meprin A in mice alters the host response to E. coli, 3)

Determine whether meprins interact directly with bacteria or bacterial components, and 4)

Determine how meprin A contributes to the host response to LPS challenge.

E.coli continues to be an important pathogen for humans, with the urinary tract as an important target organ. Murine models of UTI have established that there are differences in mouse strain susceptibility to UTI and that this variation is due to the multigenic nature of

UTI (69,88). Many of the genes implicated in vulnerability to UTI are related to the major histocompatibility complex, the area of the genome which encodes the meprin α gene

(89,90). Additionally, C3H/He mice, which express low meprin levels in the adult kidney, are more susceptible to bladder and kidney infection than C57BL/6 mice, which express high meprin levels (88). The increased infection susceptibility of C3H/HeJ mice is largely due a mutation in the TLR4 gene, making C3H/HeJ mice endotoxin resistant and highly susceptible to infection by Gram-negative bacteria. However, C3H/HeOuJ mice, which are

TLR4-sufficient, are also more susceptible to UTI, indicating that factors other than TLR4 play a role in UTI susceptibility in C3H mice (91).

These data led to the hypothesis that meprins play a role in susceptibility to UTI, and women with low levels of urinary meprin would be more susceptible to infections of the urinary tract. Soluble meprin A is abundantly expressed in rodent urine, but nothing was known regarding its expression in human urine. Investigation of this hypothesis will also allow for the characterization of urinary meprin during both healthy and infected states.

Although the “high” and “low” meprin mouse strains differ in their levels of meprin expression, they also differ at multiple locations other than the meprin α locus. Thus,

meprin-deficient mice are a unique resource that allows for characterization of meprin A’s contribution to UTI. It was proposed that meprin αKO mice would be more susceptible to

UTI because meprin A would not be present to participate in the inflammatory response.

However, it could not be assumed that a differential response would be observed in the meprin αKO mice because genetic compensation and redundancy are commonly observed in

KO mouse models. The localization of meprin in the urinary tract and in populations of leukocytes points to the possibility of direct and/or indirect roles for meprins in the inflammatory response to infection.

Soluble meprin A is a large enzyme complex with high proteolytic potential that could potentially interact directly with bacteria via proteolytic cleavage of pili or other bacterial products. Cleavage of bacterial adhesins could prevent the bacteria from binding to the bladder urothelium, a key factor in the initiation of UTI. Further, meprins are highly glycosylated, with glycans contributing to approximately 20% of the molecular weight of the meprin α subunit. Because host/bacteria interactions are mediated by the binding of bacterial pili to sugars on the host cell surface, it is possible that the highly glycosylated meprins could directly bind to bacterial adhesins. Binding of bacteria by meprins could sequester bacteria, readily clearing the urinary tract of bacteria with urine flow.

Limited in vitro data indicate a role for meprins in leukocyte migration and chemokine inactivation. Therefore, it is possible that meprins could indirectly modulate the

UTI response via effects on cytokine gradients, extracellular matrix remodeling, or leukocyte infiltration. An LPS challenge eliminates the other bacterial virulence factors that impact the course of UTI. Therefore, challenging meprin KO mice with LPS would help clarify the role of meprins in the inflammatory response in the absence of active infection.

The host response to inflammatory stimuli is complex, and proteolysis contributes to the initiation, progression, and resolution of the inflammatory response. These studies with the meprin KO mice will greatly expand what is known about meprins role as a potential mediator of the immune response. Identifying factors, such as the meprins, that participate in the response to UTI is important because the current knowledge about the pathogenesis of

UTI cannot explain the diversity of clinical outcomes observed.

Chapter 2: Materials and Methods

Animal models

2.1 Mice

Female mixed background (C57BL/6 and 129x1/Sv) meprin αKO, C57BL/6 meprin

βKO, and mixed background (C57BL/6 and 129x1/Sv) meprin αβKO mice and corresponding WT mice for each genotype were used at 8-9 weeks of age for all experiments.

All mice were maintained in conventional housing and were allowed water and rodent chow ad libidum. C57BL/6 WT controls were bred in the Pennsylvania State University College of

Medicine Animal Facility and crossed with C57BL/6 mice purchased from Jackson

Laboratories at 12 month intervals.

Meprin αKO and meprin βKO animals were derived as previously described (20,55).

Briefly, the Mep1a gene was inactivated by homologous recombination with a targeting vector, Osdupdel-Mep1a, which inserted a 1.2 kb neomycin cassette with an in-frame stop codon into exon 7 (the catalytic domain of the protease). The targeting vector pKO-Mep1b was used to inactivate the Mep1b gene in a similar manner and disrupted exon 7 with a 1.6 kb neomycin resistance gene. The targeting vectors were electroporated into R1 mouse embryonic stem cells, and clones were screened by PCR. Electroporation, selection of targeted ES cells, and subsequent blastocyst injections were performed at the University of

Michigan Transgenic Animal Facility. Chimeric mice were crossed with C57BL/6 mice at the

Pennsylvania State University College of Medicine.

Meprin αKO mice were generated by crossing chimeric mice with C57BL/6 mice to generate F1 mice heterozygous for the meprin α deletion. Heterozygous mice from the F1 generation were crossed to yield WT, meprin αKO, and heterozygous C57BL/6 and

129X1/Sv mixed background mice. Heterozygous mice were crossed with heterozygous mice for two additional matings (three total). After the third heterozygous cross, WT mice were crossed with WT, and meprin αKO mice were crossed with meprin αKO to generate WT or meprin αKO mice, respectively.

Meprin βKO mice were generated by crossing chimeras with C57BL/6 mice to generate F1 mice heterozygous for the meprin β deletion. Heterozygous mice from the F1 generation were subsequently backcrossed to inbred C57BL/6 mice, and this process was repeated for 13 generations. Females and males were alternately bred, and the inbred

C57BL/6 mice were rotated to prevent the introduction of a spontaneous mutation.

Meprin αβKO mice were generated by crossing mixed background (C57BL/6 and

129x1/Sv) meprin αKO mice with C57BL/6 meprin βKO mice. Southern blot analysis and

PCR of tail biopsies were performed to determine genotypes. All animal protocols were approved by the Penn State Institutional Animal Care and Use Committee.

2.2 Intravesical challenge with E. coli or LPS

The protocol for E. coli and LPS bladder infusion was adapted from Saban et al. and

Haugen et al. (81,92). The abdomens of anesthetized meprin αKO and WT mice were gently massaged to empty residual urine. Approximately 107 E. coli CFT073 (ATCC accession #

700928) or 20 μg LPS in 50 μl of sterile PBS was slowly instilled into the bladder using a 0.5

mm polyethylene catheter (Intramedic PE 10) attached to the hub of a 50 μl Hamilton #705 syringe with 30 gauge blunt-tipped needle. Mice were sacrificed by cervical dislocation following anesthesia by isofluorane inhalation. Bladders and kidneys of the E. coli

challenged mice were placed in 500 μl or 1 ml of PBS, respectively and homogenized in

Nasco Whirl-Pak bags with the rolling pressure of a glass bottle. Organ homogenates were serially diluted with PBS, plated on EMB agar, incubated overnight at 37°C, and bacterial colonies were enumerated.

2.3 Intraperitoneal challenge with LPS

Meprin KO and WT mice were challenged i.p. with 2.5 mg/kg body weight LPS (E. coli 0111:B4, purified by gel filtration chromatography, Sigma-Aldrich) or sterile saline.

LPS was suspended in sterile PBS at a concentration of 1 mg/ml and stored in 1 ml aliquots at -20°C. To make up the working stock, 1 ml of 1 mg/ml LPS was diluted 1:4 with sterile

PBS. A 0.5 cc U-100 insulin syringe (Beckton Dickinson) with a 28-gauge needle was used to administer 10 μl LPS per gram body weight.

2.4 Anesthesia

In general, mice were anesthetized by inhalation of isofluorane. For some of the intravesical E. coli and intraperitoneal LPS experiments, ketamine/xylazine (100 mg/kg body weight ketamine; 10 mg/kg body weight xylazine) anesthesia was administered i.p. to the mice. To make up the working stock, 0.5 ml of 100 mg/ml ketamine and 0.25 ml of 20 mg/ml xylazine was added to 4.25 ml sterile PBS. A 0.5 cc U-100 insulin syringe (Beckton

Dickinson) with a 28-gauge needle was used to i.p. administer 10 μl ketamine/xylazine per gram body weight.

2.5 Elicitation and isolation of peritoneal exudate cells

Resident peritoneal exudate cells (PECs) were harvested from untreated mice. To elicit inflammatory PECs, mice were challenged i.p. with 1 ml of sterile thioglycollate (TG) broth using a 27-gauge needle. Mice were euthanized by inhalation of isofluorane. The abdomen was sterilized with 70% ethanol, and a midline incision was made with scissors.

The abdominal skin was retracted with forceps and scissors, leaving the peritoneal wall intact. The peritoneal cavity was washed with 10 ml serum-free RPMI media administered via a 20-gauge needle attached to a 10 ml syringe, inserted into the midline of the peritoneum with the beveled end facing up. Peritoneal fluid was withdrawn slowly using the same needle and syringe (approximately 8 ml recovery per mouse). PECs were counted using a hemacytometer.

Assays

2.6 Myeloperoxidase activity

Mouse bladders were homogenized in 10 volumes of 20 mM potassium phosphate buffer (pH 7.4) and centrifuged at 20,000xg for 20 min at 4°C. The supernatant fluid was discarded and the sediment was suspended in 50 mM potassium phosphate buffer (pH 6.0) containing 0.5% HTAB and 10 mM EDTA. After five 3 second sonications, the preparation was freeze-thawed 3 times and incubated on ice for 20 min. The lysate was centrifuged at

4,000xg for 15 min at 4°C to remove debris, and the supernatant fluid was transferred to a fresh tube. One hundred μl of the supernatant fluid was added to 400 μl of potassium phosphate buffer (pH 6.0) containing 0.167 mg/ml of o-dianisidine dihydrochloride and

0.0005% hydrogen peroxide. The change in absorbance at 460 nm was measured over 3 min,

and myeloperoxidase activity was expressed in MPO units/g bladder tissue, where 1 MPO unit is equal to the cleavage of 1 mmol H2O2 (80).

2.7 Assessment of bladder permeability

The protocol to determine bladder permeability was adapted from Eichel et al. (93).

Two hundred μl of 10 mg/mL sodium fluorescein (NaFl) was instilled into the bladders of anesthetized mice at several timepoints after LPS or E. coli bladder challenge using a 0.5 mm polyethylene catheter (Intramedic PE 10) attached to the hub of a 50 μl Hamilton #705 syringe with 30 gauge blunt-tipped needle. After 15 min, serum samples were collected via tail vein (or from the renal artery for terminal experiments), diluted 1:5 (5 μl serum, 95 μl

PBS), and NaFl concentrations were measured fluorometrically using a 0.1-100 μg/ml standard curve (excitation 494 nm, emission 516 nm).

2.8 Blood collection and temperature monitoring

Body temperatures were monitored at 2, 6, 12, and 24 h via a rectal thermistor (Cole-

Parmer) lubricated with glycerol. Small blood samples were collected from the tail vein into lithium/heparin tubes (Sardstedt), and plasma was isolated by centrifugation at 9,500xg for 8 min at 14°C. Terminal blood samples were collected from the renal artery into 1.5 ml microfuge tubes at the time of necropsy, and serum was isolated by centrifugation in an identical manner.

2.9 Quantification of blood urea nitrogen, creatinine, and nitrate/nitrite

To assess renal function, blood plasma urea nitrogen (BUN) levels were determined by Vitros DT6011 BUN chemistry slides (Orthoclinical Diagnostics). Urea in the serum samples was converted to ammonia via urease. The ammonia reacted with an indicator (N- propyl-4-(2,6-dinitro-4-chlorobenzyl)-quinolonium ethane sulfonate) to form a dye that was measured at 670 nm by the Vitros slide reader. Serum nitrate/nitrite levels were determined using the Nitrate/Nitrite Colorimetric Assay Kit (Cayman Chemical) according to the manufacturer’s instructions, except all reagent volumes were decreased by half. Briefly, nitrate in the samples was converted to nitrite with the addition of nitrate reductase. Addition of the Griess reagents (sulfanilamide and N-(1-naphthyl) ethylenediamine) converted the nitrite to a purple azo compound. The absorbance at 540 nm was measured and nitrate/nitrite concentrations were determined using a nitrate standard curve (0-50 μM). Serum samples were ultrafiltered through pre-rinsed Microcon YM-10 10 kDa microfuge filtration devices

(Millipore) prior to determination of nitrate/nitrite in order to reduce background resulting from hemoglobin.

Plasma and urine creatinine levels were determined by an end-point colorimetric creatinine assay (DZ072B Diazyme Labs). Human urine samples were diluted 1:10 with

PBS prior to determination of creatinine levels, which were used to normalize gel loading sample volumes. Murine plasma samples were analyzed for creatinine to assess kidney function. All creatinine assays were performed in 96 well plates according to the manufacturer’s instructions, except all reagent volumes were decreased by half. Briefly, the assays consisted of coupled enzymatic reactions in which creatinine was converted to creatine, which was in turn converted to sarcosine. The sarcosine was oxidized to produce

hydrogen peroxide, which was quantified by an oxidative coupling reaction with a Trinder reagent and subsequently read at 540 nm absorbance.

2.10 Quantification of serum TNFα and cytokines/chemokines

Serum cytokines (IL-1α, IL-1β, IL-2, IL-3, IL-4, IL-5, IL-6, IL-9, IL-10, IL-12,

CCL2 (MCP-1), IFNγ, TNFα, CCL3 (MIP-1α), GM-CSF, CCL5 (RANTES)) were measured using the ELISA-based Quansys Q-Plex Mouse Cytokine Array (Biolegend, Inc.). Imaging of the array was performed on the EC3 Biochemi imaging system equipped with CCD camera (UVP) and analyzed with Quansys Q-View software. Serum TNFα levels were confirmed using the Quantikine Mouse TNFα ELISA assay (R&D Systems), according to the manufacturer’s directions.

2.11 Quantification of total protein

The protein concentration of samples was determined by Bradford method using the

Coomassie Plus Protein Assay kit (Pierce). The micro test tube protocol was followed according to the manufacturer’s instructions, except reagent volumes were decreased by half.

Briefly, protein in the sample bound to the Coomassie R-250 dye in the reagent, causing a spectral shift from brown to blue color (465 to 610 nm). The blue dye was measured at an absorbance of 595 nm and protein concentrations were determined from a BSA standard curve (2.5-25 μg/ml). The Coomassie Plus assay was selected over the BCA (bicinchoninic acid) protein assay for the analysis of urine samples due to its high tolerance of urea.

E. coli and pili

2.12 Growth conditions and pili enrichment

Uropathic E. coli CFT073 was obtained from ATCC (accession # 700928). For use in the intravesical E. coli challenge, bacteria were grown overnight at 37°C on Luria Bertani

(LB) agar plates. Bacteria were subsequently transferred to static LB broth with a sterile swab and grown at 37°C for 6 days, passaging every two days to enrich for type-1 pili

expression. To prepare E. coli for bladder instillation, cultures were adjusted to OD600 =

0.45-0.52 with LB broth. Bacteria were centrifuged in 1.5 ml microfuge tubes and centrifuged at 3,400xg for 5 min. The supernatant fractions were discarded and bacteria were suspended in 1/10 volume of sterile PBS (1 ml bacteria suspended in 100 μl PBS) to yield approximately 107 E. coli per 50 μl. For high load UTI experiments, cultures were adjusted

8 to OD600= 0.6-0.8 and prepared in a like manner to yield approximately 10 E. coli per 50 μl.

To enrich for P-pili expression, E. coli was grown on trypticase soy agar plates for 24 h at

37°C. The bacteria were carefully scraped from the solid agar and suspended in PBS for

subsequent assays.

2.13 Agglutination assays

E. coli expressing type-1 pili agglutinate guinea pig red blood cells (RBCs) by

binding to the mannosylated glycoproteins expressed on their surface (94). Agglutination of

guinea pig RBCs is prevented and dissociated by the addition of 25 mg/ml mannose. To

confirm expression of type-1 pili, E. coli was cultured in static LB broth for 48 h prior to

centrifugation at 3,400xg for 5 min and suspension in PBS. Three μl of RBCs were diluted with 97 μl PBS. Fifteen μl each of E. coli and diluted guinea pig RBCs were combined on a

microscope slide by mixing with a pipette tip. The slide was inspected immediately and after

10 min at 40x power with a light microscope. E. coli expressing P pili cause agglutination of human RBCs by binding to the Gal(α1-4)Gal expressed on their surface. To confirm expression of P-pili, E. coli was grown on trypticase soy agar plates for 24 h and human

RBCs were utilized in an identical assay.

To determine if meprin prevented agglutination of guinea pig red blood cells, 6 μg active recombinant mouse meprin A was added immediately or pre-incubated for 30 or 60 min with the E. coli or the RBCs prior to starting the agglutination assay. To determine if meprin was capable of dissociating the agglutinated cells, active recombinant mouse meprin

A was added 20 min after the start of the agglutination assay.

2.14 Cleavage of recombinant pili

Recombinant type-1 pili and P-pili from E. coli were a gift of Scott Hultgren

(Washington University School of Medicine). P-pili (2.5 μg) were incubated for 1 h at 37°C with 1.2 μg of active recombinant homomeric mouse or human meprin A. Type-1 pili (3 μg) were incubated with 1 μg active recombinant homomeric mouse or 2 μg human meprin A.

All preparations were subjected to SDS-PAGE under reducing and denaturing conditions and stained with silver. Some preparations were additionally boiled in 0.1 μl 12 N HCl and reducing and denaturing loading buffer for 10 min and neutralized with 0.2 μl 5 M NaOH prior to electrophoresis in order to dissociate the pili into monomeric units.

2.15 Antimicrobial assays

E. coli was prepared in the same manner as for use in bladder challenge experiments.

Ten μl (approximately 106) of E. coli in PBS was added to 150 μl of LB broth, WT mouse

urine, or meprin αKO urine. Samples were incubated overnight at 37°C. The next morning,

the cultures were serially diluted with PBS, and 100 μl each of 10-4, 10-6, and 10-8 dilutions

was plated on EMB agar. Colonies were enumerated the following morning after plates were

incubated at 37°C overnight.

2.16 Interaction of meprin with bacteria

Uropathic bacteria were grown overnight in static trypticase soy broth. Bacteria were

centrifuged at for 10 min at 3,400xg in a tabletop centrifuge and suspended in an equivalent

volume of 50 mM ethanolamine buffer (pH 9). Three 100 μl reactions were prepared (77 μl

CFT073 + 23 μl buffer, 23 μl (2 μg) latent homomeric mouse meprin A + 77 μl buffer, and

77 μl CFT073 + 23 μl (2 μg) latent homomeric mouse meprin A). All reactions were

incubated for 1 h at 37°C and subsequently centrifuged at 3,400xg for 10 min. Ten μl of

each reaction supernatant was assayed for activity against BK+ (see section 2.22, page 41).

Active homomeric mouse meprin A was assayed as a positive control.

2.17 Whole bladder binding assays

Five meprin αKO and WT mice were euthanized with isofluorane and bladders were

excised. Bladders were slit open with a scalpel, washed with PBS, and placed in a 24-well

plate. Type-1 enriched E. coli were diluted to OD600∼0.5, and 10 μl of bacteria per 990 μl of

McCoy’s media (serum-free) were added to each well. Bladders were incubated for 1 h at

37°C. To determine the number of bound bacteria, bladders were washed 5 times with PBS and incubated with 1 ml 0.25% trypsin/EDTA and 10 μl of 10% Triton-X for 5 min at 37°C.

Bladders were transferred to culture tubes and homogenized for 1 min. Serial dilutions were plated on LB and colonies were enumerated after 24 h incubation at 37°C.

Cell culture

2.18 Maintenance of cells

Madine Darby Canine Kidney (MDCK) cells (ATCC accession # CCL-34) were grown in MEM (Invitrogen) and supplemented with Earl’s salts, L-glutamine, sodium bicarbonate, and 10% FBS (Atlas Biologicals) at 37°C and 5% CO2. T24 bladder carcinoma

cells (ATCC accession # HTB-4) were grown in McCoy’s media (Invitrogen) supplemented

with 10% FBS.

2.19 E. coli binding assays

HTB-4 cells were seeded into 24-well plates and grown to confluence. CFT073 E.

coli were enriched for type-1 pili or P pili expression and diluted to an OD600 of 0.5-0.52.

After washing the cells with PBS, 10 μl of bacteria per 990 μl of McCoy’s media (serum- free) were added to each well. Plates were centrifuged at 1,800 rpm for 5 min and subsequently incubated for 5 min at 37°C. To determine the number of bound bacteria, cells

were washed 5 times with PBS and incubated with 0.25% trypsin/EDTA for 5 min at 37°C.

Cells were harvested by pipetting up and down after the addition of 10 μl 10% Triton-X 100.

Serial dilutions were plated on LB and colonies were enumerated after 24 h incubation at

37°C. For E. coli binding assays with meprin, 200-400 ng of latent or active homomeric

mouse or human meprin A was pre-incubated with 1 ml of the bacteria plus McCoy’s media mixture for 30 min at 37°C prior to addition to the wells. All E. coli dilutions were incubated at 37°C for 30 min to account for possible bacterial growth during that time period.

2.20 Meprin treatment of kidney cells

MDCK cells with a passage number of less than 20 were grown in 24 well plates on round glass coverslips for 4 days after reaching confluency. MEM was changed to serum- free MEM approximately 18 h prior to treatment with meprin. Cells were treated with 4

μg/ml active or latent recombinant human homomeric meprin A for 5 h.

2.21 Immunocytochemistry, confocal microscopy, and extraction of tight junction proteins

Immunocytochemistry was performed as described previously (95). MDCK cells were fixed in 1% paraformaldehyde for 10 min, made permeable with 0.2% Triton-X 100 for

10 min, and blocked in 10% goat serum with 0.1% Triton-X 100 for 1 h with shaking. Cells were incubated with mouse anti-occludin (Zymed) at a 1:50 dilution and rat anti-ZO-1 monoclonal antibody (gift from Dr. B. Stevenson) at a dilution of 1:4. After 5 washes in

0.1% Triton-X 100, goat anti-mouse Alexafluor 488, goat anti-rat Alexafluor 647, and

Hoeschst nuclear stain were added at a 1:1,000 dilution. Coverslips were mounted on slides with Aquamount (Polysciences, Inc.). Cells were imaged on Leica TCS SP2 AOBS confocal microscope (512x512 resolution). Images were merged with Adobe Photoshop software.

To extract proteins for Western analysis, protein was extracted from MDCK cells with Stuart’s buffer (100 mM NaCl, 1% Triton X-100, 0.5% sodium deoxycholate, 0.2%

SDS, 2 mM EDTA, 10 mM HEPES (pH 7.5), 1 mM sodium orthovanadate, 10mM NaF, 10 mM sodium pyrophosphate, 1 mM benzamidine), supplemented with 1 complete protease inhibitor tablet and 10 μl of 1 mM LR-mycrocystin per 10 ml buffer. Media were aspirated from the cells, and cells were washed two times on ice with PBS/PMSF (100 μl of 200 mM

PMSF per 100 ml PBS). Seventy-five μl of Stuart’s buffer was added to each 60 mm plate, and the cells were scraped and transferred to a microfuge tube on ice. The process was repeated with an additional 75 μl of buffer. Samples were rocked for 15 min at 4°C, centrifuged for 10 min at 12,000xg, and the supernatant was transferred to a fresh tube.

2.22 Cytotoxicity assays

The CytoTox-ONE Homogeneous Membrane Integrity Assay (Promega) was used according to the manufacturer’s directions. This fluorometric assay measured the release of lactate dehydrogenase (LDH) from cells with damaged membranes. MDCK cells were seeded at a density of 2x105 cells per well in 96 well plates. After 24 h, media were switched

to serum-free and 6-20 μg/ml of latent or active recombinant forms of rat meprin B,

heteromeric rat meprin A (αβ), homomeric human meprin A, or homomeric mouse meprin

A. LDH released into the culture media was measured with coupled enzymatic assay that

resulted in the conversion of resazurin into a fluorescent resorufin product. The percent of

LDH release was determined by calculating the ratio of fluorescence in the experimental

wells to a maximum LDH release control (excitation 560 nm, emission 590 nm).

MTT cytotoxicity assays were also performed. MTT is a yellow tetrazole compound

that is reduced to a purple formazan in the mitochondria of living cells. Thus, the intensity of

the purple coloration in the sample was directly proportional to the percentage of viable cells

present. For the MTT method, in vitro toxicology assay kit (TOX-1, Sigma) was used according to the manufacturer’s directions, except the 96-well plate was assayed at 595 nm instead of 570 nm.

Meprins

2.23 Meprin purification, activation, and activity assays

Recombinant meprin A and meprin B were purified from transformed human embryonic kidney (HEK) 293 cells (ATCC accession # CRL 1573) by methods previously described (47). The latent form of meprin was subjected to limited digestion by trypsin at a ratio of 1:20 (1 trypsin: 20 meprin) to produce the active protease. Meprin A was incubated for 45 min at 37°C while meprin B incubated for 60 min at 37°C. Trypsin activity was inhibited by addition of 6-fold excess soybean trypsin inhibitor (STI) and incubation for 20 min at 25°C. The majority of the trypsin and STI was removed by filtration through G25 resin before use in cell culture or bacterial assays. To make a G25 column, a disposable 1 ml syringe was filled to the top with a suspension of G25 and water. The column was placed in a 15 ml polypropylene tube and centrifuged at 4300xg for 6 min at 4°C to pack the resin.

One hundred μl of sample was added directly to the top of the resin, a 500 μl microfuge tube was placed at the bottom of the syringe to capture the filtrate, and the sample was centrifuged in an identical manner.

Meprin protease activity was measured by fluorometric substrate assay as previously described (96,97). Briefly, meprin A activity was assayed with BK+ (Abz-ARg-Pro-Pro-

Gly-Phe/Ser-Pro-Phe-ARg-Lys(Dnp)-Gly-OH; excitation 320, emission 417) in 50 mM ethanolamine buffer (pH 8.7). Meprin B activity was assayed with OCK+ (Abz-Met-Gly-

Trp-Met/Asp-Glu-Ile-Asp-Lys(Dnp)-Ser-Gly-OH; excitation 326, emission 418) in 20 mM

HEPES, 100 mM NaCl buffer (pH 6.5). Meprin activity was measured at 5 s intervals for

50-60 s using the Time Scan method on the Hitachi F-2000 fluorimeter.

2.24 Western blot analysis

Samples were subjected to 8% PAGE under denaturing and reducing or non-reducing conditions. Protein was transferred to nitrocellulose membranes using the semi-dry Trans-

Blot SD (Bio-Rad). Membranes were blocked in 10% milk in TBS/0.1% Tween-20 (TBS-

T). Membranes were incubated overnight with primary antibody at 4°C with slight agitation in 5% milk in TBS-T. After one 20 min wash in TBS-T, membranes were incubated with a horseradish peroxidase-conjugated secondary antibody for 1 h in 5% milk in TBS-T. The membranes were washed 2 times for 20 min each in TBS-T. Western blots were developed using SuperSignal West Dura Extended Duration Substrate (ThermoScientific).

2.25 Urinary dot blot development

A dot blot protocol was developed to aid in the genotype determination of mice.

Because homomeric meprin A is secreted into the urine of both WT and meprin βKO mice but not into the urine of meprin αKO mice, the presence or absence of immunoreactivity indicates WT or αKO, respectively. The dot blot was also utilized to analyze urine samples from healthy and UTI-infected women. Optimization of the dot blot protocol was performed to determine the minimum amount of time needed to obtain a reliably strong immunoreactive signal. The following procedure resulted in a strong immunoreactive signal in 5 replicates of both 1 μl of WT mouse urine and 1 μl of 0.1 μg/μl recombinant homomeric mouse meprin A.

One to three μl of urine was blotted directly onto nitrocellulose membranes. Nitrocellulose membranes were blocked in 10% milk/TBS-T for 1 h. Membranes were subsequently incubated in 5% milk/TBS-T with primary antibody for 2 h. After three 5 min washes in

TBS-T, blots were incubated in 5% milk/TBS-T with a secondary antibody conjugated to horseradish peroxidase for 1 h. The blots were washed three times for 5 min each in TBS-T and developed with SuperSignal West Pico chemiluminescent substrate (ThermoScientific).

Human specimen processing

Urine samples from healthy women without a history of UTI, healthy women with a history of UTI, and women suffering from an active UTI were collected for analysis. Urine samples were centrifuged at 500xg for 10 min at 4°C to remove debris. Multistix urinalysis strips (Bayer) were used to measure specific gravity, pH, leukocytes, nitrite, protein, glucose, ketones, urobilinogen, bilirubin, and blood/hemoglobin within 4 h after collection.

Supernatant fluid was removed and processed immediately or stored in 10 ml aliquots at

-80°C. Urine samples were concentrated (10 ml) approximately 10 fold in Macrosep concentrators (Pall) with a 1 MDa molecular weight cut-off. Concentrators were soaked in

10% glycerol at least 24 h prior to use to minimize protein binding to concentrator walls. All human protocols were approved by the Penn State Institutional Review Board and Human

Subjects Protection Office.

Statistical analysis and graphing software

PRISM GraphPad statistical software was used to plot figures and data analysis.

Results are expressed as means ± SE; a p-value <0.05 was considered significantly different.

Chapter 3: Meprin A modulates the host response to E. coli bladder challenge

The soluble form of meprin, homomeric meprin A, is a large multimer that can contain up to 100 meprin α subunits and is secreted from the proximal tubule of the kidney into the urinary tract (98). Meprin A is in an ideal location for preventing or lessening infections of the urinary tract. It is secreted as an inactive complex with high proteolytic potential, and can be readily activated by trypsin-like proteases (43,45). Further, renal meprin expression in mice is strain dependent, and low meprin mice (e.g., C3H) are markedly more susceptible to infections of the urinary tract. These data led to the hypothesis that meprins play a role in susceptibility to UTI, and that women with low levels of urinary meprin would be more susceptible to infections of the urinary tract.

Results

3.1 Human urine studies

3.1.1 Meprin expression is increased in urine from women with active UTIs

To examine meprin expression in human urine, voided urine samples from pre- menopausal women with a history of frequent UTIs, without a history of frequent UTIs, and with an active UTI were collected for analysis (Table 3). Infection was confirmed by urine culture on EMB agar and a positive dipstick reading for nitrate/nitrite and/or leukocytes.

Urinary creatinine concentrations were determined and were not significantly different among the three groups. However, total urinary protein concentrations were significantly higher in women with active UTI in comparison to healthy women with or without a history of frequent UTI (p =0.01, p<0.05, respectively; Table 3). This is consistent with the observation that increased urinary protein is frequently correlated with active UTI (99).

Number of subjects Total protein (μg/μl) Creatinine (mg/dl)

Healthy; no history UTI 20 0.06 ± 0.08 170 ± 131

Healthy; frequent UTI 9 0.04 ± 0.03 100 ± 85

Active UTI 14 0.12 ± 0.08 109 ± 88

Table 3. Human urine chemistry. The number of human female subjects and their corresponding UTI status is indicated. Cells were removed from the urine by centrifugation prior to analysis, ensuring that only soluble protein was detected. Urinary protein and creatinine levels were determined by colorimetric assay. Protein concentrations were significantly greater in active UTI urines in comparison to healthy urines, regardless of UTI history (Healthy, frequent UTI versus active UTI, p =0.01. Healthy, no history UTI versus active UTI, p<0.05). Creatinine levels were similar among all three groups.

Quantitative Western blotting showed that meprin A was detectable in the urine of pre-menopausal women at low concentrations, and that concentrations were variable from subject to subject (Figure 7). In contrast, the majority (86%) of women suffering from an active UTI had high (greater than 5.2 ng meprin) or very high (greater than 9.1 ng per 60 μg total urinary protein) levels of meprin protein in the urine. No correlation between the urinary levels of meprin and history of frequent UTI was observed. The data indicate that urinary meprin A protein expression is markedly increased during active UTI infection in women.

3.1.2 Development of a dot blot assay expedites human sample processing

To expedite the analysis of human urine, a dot blot protocol based on the established

Western blot assay was developed. Urine samples were assayed in triplicate or greater to confirm repeatability of the assay, with recombinant human meprin α as a positive control

(Figure 8A). Immunoreactivity was observed in some of the samples, and the presence of immunoreactivity on the dot blots was consistent with Western blot analysis of the same samples.

Because soluble meprin A protein is abundant in WT mouse urine, the dot blot can also be used to confirm quickly the genotypes of WT and meprin αKO mice. Analysis of mouse urine identified immunoreactivity in all spot replicates of urine from WT mice but not meprin αKO mice, confirming specificity of the antibody for meprin α (Figure 8B). The data confirm that the dot blot protocol is as effective in identifying meprin α positive urines as the traditional Western blot protocol. The dot blot procedure decreases the start to finish

Figure 7. Urinary meprin expression is increased during active UTI. Human urine samples were concentrated 10-fold and analyzed by Western blotting for meprin α expression. Meprin levels are low or below detection limits in urine samples from healthy women (31, 32, 33, 34). In contrast, meprin levels are increased in urine samples from women with active UTI (52, 53, 56, 57). Numbers represent randomly assigned patient identification numbers. Western blot is representative of healthy and UTI samples analyzed;

N for each group is indicated in Table 3.

Figure 8. Dot blots of human and mouse urine samples. (A) Five human urine samples were blotted onto nitrocellulose membrane in triplicate. Dot blot analysis revealed immunoreactivity in 2 out of 5 samples, with recombinant human meprin α as a positive control. (B) One WT mouse urine sample and 4 meprin αKO urine samples were blotted five times each onto nitrocellulose membrane. Meprin α immunoreactivity is observed in

WT lanes only, with recombinant mouse meprin α as a positive control. Dot blots are representative of at least 3 independent experiments.

time of the assay by at least 9 h, thereby allowing for more effective time management.

3.2 Murine model of UTI

3.2.1 Meprin αKO mice have less leukocyte infiltration into the bladder after UTI challenge

The data indicate a correlation between meprin α levels in human urine and active

UTI, implicating a role for meprins in the host response to UTI. To test the interpretations from the human clinical studies, both meprin αKO and WT mice were challenged with E. coli (106-107) according to a model of ascending UTI, and several parameters were

monitored to assess the extent of the inflammatory response. Release of MPO, a marker of

neutrophil activation, was measured to determine the degree of neutrophil migration into

infected bladder tissue. Marked neutrophil infiltration into the bladder wall was observed as

early as 12 h in both genotypes after UTI challenge (Figure 9). MPO activity continued to

increase at 24 h in the WT mice, and significantly more MPO activity was observed 48 h

after UTI induction in the bladders of WT mice in comparison to meprin αKO mice. MPO

levels were low in saline inoculated control mice (Figure 9, dashed line). The data indicate

that the presence of meprin A contributes to the duration and extent of leukocyte infiltration

during UTI.

3.2.2 Meprin αKO mice have less bladder edema after UTI challenge

E. coli-challenged (106-107) WT bladders were visibly redder in color and larger than

meprin αKO mouse bladders 48 h after E. coli bladder challenge (Figure 10A, B). WT

bladders were significantly heavier than bladders from meprin αKO mice when corrected to

Figure 9. Meprin αKO mice have less leukocyte infiltration after UTI challenge.

Release of MPO was measured in homogenized bladders 12, 24, and 48 h post-UTI induction to determine the degree of leukocyte migration into infected bladder tissue. MPO activity was significantly lower in the bladders of meprin αKO mice in comparison to their WT counterparts at 48 h; *, p=0.001. Dashed line indicates the average MPO value for saline-

treated control animals. Results are expressed as means of 2 independent experiments, with

n=10-14 mice per genotype, per timepoint.

Figure 10. Meprin αKO mice have less bladder inflammation after UTI challenge.

Mice were sacrificed 48 h post-UTI and bladders were removed and photographed. Bladder sections were stained with H&E. (A) Infected WT bladders were larger and visibly redder in comparison to (B) meprin αKO infected bladders. (C) Bladder edema is apparent at 4X and

20X magnification of WT infected bladders in contrast to WT saline controls. (D) Bladder histology in infected meprin αKO bladders is similar to meprin αKO saline controls.

Pictures are representative of at least 3 independent experiments.

body weight as early as 12 h after E. coli instillation, indicating more extensive edema in the

WT bladders (Figure 11). Bladder weights of saline-treated meprin αKO and WT mice were between 16 and 24 mg. In contrast, bladder weights of E. coli-challenged WT mice ranged from 27 to 48 mg, with an average weight of 34 mg. The marked edema in E. coli- challenged WT bladders was confirmed by histological examination of bladder tissue (Figure

10C, D). The interstitial space between the urothelium and the smooth muscle was widened, creating the impression of a gap between the two layers. The extent of interstitial widening was more marked in challenged WT bladders than in meprin αKO bladders. To determine whether E. coli instilled in the bladder induced kidney damage, indicative of a systemic infection, serum samples were collected and analyzed for blood urea nitrogen (BUN) at 48 h.

All levels were normal (less than 25 mg/dl) and not significantly different between WT and meprin αKO genotypes, confirming that the bacterial challenge did not induce a systemic host response. To determine if this E. coli challenge level could result in bacteremia, randomly chosen blood samples were cultured and were found to be negative for bacterial growth on EMB agar. No genotypic difference in edema was observed when a high load

(108) E. coli challenge was administered to the bladder (bladder to body weight ratio at 12 h:

meprin αKO 0.00139; WT 0.00153; p=0.32). The data indicate an attenuated inflammatory

response to bladder E. coli challenge in mice lacking meprin A in comparison to their WT counterparts.

Figure 11. Meprin αKO mice have less bladder edema after UTI. Mice were necropsied

12, 24, 48, or 120 h after E. coli instillation, and bladder to body weight ratios were determined. Edema was observed as early as 12 h post-UTI and continued to increase at 48 h post-infection. WT bladders were significantly heavier than KO bladders 12, 24, and 48 h. *, p=0.0003; **, p= 0.02; ***, p= 0.004. Dashed lines indicate average value for saline-treated control animals. The data are an average of 2 independent experiments, with n=12 animals per genotype.

3.2.3 Meprin αKO mice display less bladder permeability after UTI challenge

The bladder wall serves as a protective barrier against foreign substances and invading bacteria, and alterations in the integrity of the bladder lining have been implicated in the etiology of bladder infection. To determine the extent of bladder damage inflicted by

E. coli bladder challenge, bladder permeability was determined by measurement of sodium fluorescein (NaFl) leakage from the bladder into the serum. At several timepoints after transurethral E. coli administration, NaFl instilled into the bladder appeared in the blood

(Figure 12). NaFl was detected in the blood from WT and meprin αKO mice as early as 12 h after E. coli challenge. By 12 h, both meprin αKO and WT mice displayed substantial loss of barrier function, and by 24 h the WT mice were affected to a significantly greater extent than the meprin αKO mice. Bladder permeability remained increased at 5 days (120 h) post-

E. coli challenge. Serum NaFl levels of saline treated control WT and meprin αKO mice were below detection limits at all timepoints (-4.23 ± 1.82 μg/ml). The data establish that mice lacking meprin A have less inflammatory damage and subsequently more intact bladder barrier function after LPS bladder challenge compared to their WT counterparts.

3.2.4 Bacterial counts in bladders and kidneys are variable in both genotypes

The number of bacteria in the tissue during active infection can be used as an indicator of the severity of the disease. At 48 h after transurethral inoculation with pathogenic E. coli, bladders and kidney homogenates from WT and meprin αKO mice were cultured on EMB agar, and colonies were enumerated to determine the severity of infection.

Approximately 26% of the E. coli challenged animals had negative organ cultures (bladder and/or kidney counts equal to 0 CFU/ml). This is consistent with the data of Hagberg et al.,

Figure 12. Less bladder permeability in meprin αKO bladders after UTI. Meprin αKO and WT mice were transurethrally administered sodium fluorescein 12, 24, 48, and 120 h post-UTI. After 15 min, serum samples were collected and NaFl levels were measured spectrophotometrically to assess permeability. WT bladders were significantly more permeable to NaFl at 24 and 120 h. *, p= 0.002, ** p= 0.048. Serum NaFl levels of saline treated controls were below detection limits for all genotypes at all timepoints. The data are an average of three independent experiments, with n=13-20 mice per genotype.

which indicated infection in 5.9 out of 10 mice on average (60% infection, 30% no infection)

(100). Both bladder and kidney counts were elevated in meprin αKO mice in contrast to their

WT counterparts, but the difference did not reach statistical significance (p=0.08 meprin αKO vs. WT bladders) due to the high degree of variation in this model system (Figure 13). It was noted, however, that 21% of meprin αKO animals had bladder bacterial counts of 1.5x106

CFU/ml tissue homogenate or greater, whereas none of the WT mice had counts greater than this value. In kidney infection, 40% of meprin αKO mice versus only 8% of WT mice had bacterial counts over 6,500 CFU/ml tissue homogenate. Taken together, the data support the proposition that meprin A has a pro-inflammatory role, as a decreased inflammatory response would allow for persistence of E. coli in meprin αKO mouse bladders and kidneys.

3.3 Exploring direct interaction between E. coli and meprins

3.3.1 Meprins do not interfere with E. coli agglutination of red blood cells

Because the secreted form of meprin is abundantly expressed in mouse urine, the possibility of a direct interaction between meprins and uropathic E. coli was explored.

Uropathic E. coli were enriched for type-1 pili or P-pili expression, which agglutinate guinea pig red blood cells (RBCs) or human RBCs, respectively, by binding sugars on the RBC surface. Thus, the effect of meprins on E. coli’s ability to agglutinate RBCs was examined.

Guinea pig RBCs incubated with recombinant homomeric meprin A for 30 min agglutinated when mixed with type-1 pili-expressing E. coli, indicating that meprin did not remove or bind to a factor on the RBCs that was essential for agglutination (Table 4A). Further, incubation of E. coli with active or latent recombinant homomeric meprin A did not prevent

Figure 13. Bladder and kidney bacterial colonization was measured 48 h after UTI induction. At 48 h after UTI, bladders and kidneys from WT and meprin αKO animals were homogenized, and serial dilutions were plated on EMB agar and incubated at 37°C overnight.

Individual bacterial colonies were counted, and data are expressed as means of colony forming units per ml of homogenized tissue (CFU/ml). Organ cultures from saline treated controls exhibited no colony formation in both genotypes. The data represented were collected from four independent experiments, with n=31-34 mice per genotype.

A. Type-1 pili

RBC Meprin A E. coli Mannose Agglutination

+/t 0 - - - - +/t 0 - +/t 0 - + +/t 0 - +/t 0 +/t 0 - +/t 0 +/t 0 - - - +/t -30 min +/t -30 min - - - +/t -30 min +/t -30 min +/t 0 - + +/t 0 +/t 0 +/t 0 - + +/t 0 +/t -30 min +/t -30 min - + +/t 0 +/t -60 min +/t -60 min - + +/t 0 +/t 0 +/t 0 +/t 0 - +/t 0 +/t -30 min +/t -30 min +/t 0 - +/t 0 +/t -60 min +/t -60 min +/t 0 - +/t 0 - +/t 0 +/t +20 min +/dissoc +30 min +/t 0 +/t +20 min +/t 0 - +

B. P pili

RBC Meprin A E. coli Agglutination

+/t 0 - - - +/t 0 - +/time 0 + +/t 0 +/t 0 - - +/t -30 min +/t -30 min - - +/t -30 min +/t -30 min +/t 0 + +/t 0 +/t 0 +/t 0 + +/t 0 +/t -30 min +/t -30 min + +/t 0 +/t -60 min +/t -60 min + +/t 0 +/t +20 min +/t 0 +

Table 4. Meprins do not interfere with bacterial agglutination of RBCs. Uropathic E. coli are capable of expressing Type-1 and P-pili, which agglutinate guinea pig or human red blood cells (RBCs), respectively. (A) Pre-incubation of active homomeric mouse meprin A with uropathic E. coli expressing type-1 pili did not prevent agglutination of RBCs, and addition of active meprins to previously agglutinated RBCs did not dissociate the clumps.

(B) Pre-incubation of active homomeric mouse meprin A with uropathic E. coli expressing P pili did not prevent agglutination of RBCs, and addition of active meprins to previously agglutinated RBCs did not dissociate the clumps. All assays were performed in duplicate.

their ability to agglutinate guinea pig RBCs, nor did the addition of meprin dissociate E. coli- agglutinated cells. Congruent data were obtained when identical assays were performed with

E. coli expressing P pili and human RBCs (Table 4B). Neither meprin nor the E. coli caused

RBC lysis. Thus, the data do not support the proposition that meprins directly interfere with the binding of bacterial pili to their respective sugars.

3.3.2 Meprins do not decrease the binding of E. coli to cultured bladder cells

The proposition that active meprin A interferes with the attachment of uropathic

E.coli to bladder urothelial cells was explored. E. coli expressing P-pili were added to HTB-

4 bladder cells in culture, and after incubation, cells were rinsed to remove unattached bacteria. The attached E. coli were enumerated as colony forming units (CFU). P-pili- enriched E. coli pre-incubated with active recombinant homomeric human meprin A bound to bladder epithelial cells 10-23% more than E. coli alone (Figure 14A). Binding assays were also performed with E. coli expressing type-1 pili. A range of 9-22% of the type-1 enriched E. coli added to the wells attached to the cell culture monolayer when serum was present in the media and 4-22% in media without serum (Figure 14B). Neither activated recombinant homomeric meprin A nor pro-meprin A significantly altered bacteria attachment

(p>0.05). In addition, other proteins such as albumin, trypsin, and soybean trypsin inhibitor

(STI) failed to alter bacterial adherence to the cultured cell monolayer. The data indicate that uropathic E. coli bound to culture cells in media with and without serum, and are consistent with the agglutination data in that latent or activated meprin alpha did not interfere attachment.

Figure 14. Meprins do not decrease the binding of E. coli to bladder cells in culture.

HTB-4 bladder cells were incubated with uropathic E. coli, extensively washed, and lysed to

enumerate the extent of bound bacteria. (A) Pre-incubation of E. coli with active or latent

meprin A prior to their addition to the bladder cells did not significantly decrease the number

of adherent bacteria. (B) Less E. coli bind to bladder cells in serum-free media. The data are an average of two independent experiments, and treatments were performed in triplicate

wells.

3.3.3 Binding of E. coli to WT and meprin αKO bladders is equivalent

Whole bladder binding assays were performed to determine whether type-1 pili- expressing E. coli differentially bind to WT versus meprin αKO bladders. Bladder tissue was slit open, washed, and incubated with approximately 2x105 E. coli in serum-free cell culture

media. The number of bacteria that bound to whole bladder tissue was variable and ranged

from 0.13% to 8.2%. No difference in counts of bound E. coli was observed between the

bladder preparations from the two genotypes (n=5; meprin αKO 5,163 ± 2,805; WT 5,530 ±

2,821 E. coli counts/ml bladder tissue; p=0.93).

3.3.4 Recombinant pili are resistant to meprin cleavage

Recombinant P-pili were incubated with active recombinant homomeric mouse or

human meprin A to determine whether the meprins are capable of degrading bacterial pili, and

the reactions were analyzed by Western blotting. Meprin α subunits consistently migrated

with a molecular mass of 80 kDa (Figure 15A). The pattern of bands was unchanged by

incubation of active human or mouse meprin A with P-pili (Figure 15A, lanes 4, 7). To

dissociate the pili into monomeric units, meprin and pili were incubated together as described

previously, but the preparation was subjected to 12N HCl prior to electrophoresis. The strong

acid disrupted the meprin A molecule and no distinct band was observed after electrophoresis

(Figure 15A, lanes 5, 8). No distinct differences between the pili only and the meprin plus pili

preparations were detected. Cleavage assays were also performed with active homomeric

meprin A and recombinant type-1 pili (Figure 15B). Again, no distinct differences between

the pili only and the meprin plus pili preparations were detected. These data indicate that

recombinant bacterial pili are resistant to degradation by meprin metalloproteases.

Figure 15. Recombinant pili are resistant to cleavage by meprins. (A) Active recombinant human or mouse homomeric meprin A incubated with P pili for 1 h at 37°C.

Some samples were boiled in HCl and neutralized with NaOH prior to electrophoresis to dissociate the pili complex. (1) Molecular weight marker; (2) Human meprin A; (3) P pili;

(4) human meprin A + P pili; (5) Human meprin A + P pili, HCl; (6) mouse meprin A; (7) mouse meprin A + P pili; (8) Mouse meprin A + P pili, HCl; (9) P pili, HCl. (B) Active recombinant human or mouse homomeric meprin A incubated with type-1 pili for 1 h at 37C.

(1) Molecular weight marker; (2) Mouse meprin A, HCl; (3) Mouse meprin A; (4 and 5)

Type-1 pili, HCl; (6) Mouse meprin A; (7 and 8) Mouse meprin A + type-1 pili; (9) Human meprin A; (10) Human meprin A + type-1 pili.

3.3.5 Meprins and E. coli viability

The ability of meprins to act on bacteria as inhibitors of growth was examined.

Uropathic E. coli were incubated alone or with 10 μg active recombinant homomeric mouse meprin A for 30 min and subsequently plated on EMB plates. Bacterial counts were equivalent (E. coli alone, 2007 ± 226; E. coli with meprin A, 1825 ±126), indicating that pre- incubation of uropathic E. coli with meprin A does not decrease bacterial viability.

3.3.6 Meprin and E. coli activation assays

The meprins are synthesized as inactive pro-enzymes and their activation by trypsin or trypsin-like enzymes has been demonstrated in vitro. The ability of uropathic E. coli to activate the meprins was examined by incubation with latent recombinant homomeric meprin

A. Uropathic E. coli were not able to activate latent homomeric meprin A (Figure 16).

Enzymatic activity against a fluorogenic substrate (BK+) was observed in the active meprin A reaction, which served as a positive control, but not in the E. coli only reaction or in the latent meprin A reaction. Thus, there is no evidence from these in vitro studies that E. coli can activate latent homomeric meprin A.

3.3.7 Urine from meprin αKO animals possesses decreased bactericidal capability

Urine acts as a first defense against invading organisms due to the low pH, osmolarity, and presence of small peptides (100). Alteration of the bactericidal capabilities of urine can allow invading bacteria to adhere and initiate a UTI. To assess the contribution of meprin to the urine environment, E. coli was cultured overnight in urine from meprin αKO or WT mice. The number of bacterial colonies observed in WT and meprin αKO urine

Figure 16. Meprins are resistant to activation by uropathic E. coli. Meprin enzymatic activity against the fluorogenic substrate BK+ was measured for 50 s. No activity was observed for latent meprin A, E. coli, or latent meprin A incubated with E. coli. BK+ activity was present in the active meprin A positive control, as indicated by the positive slope over time. Activation assays were performed in triplicate.

cultures was significantly less than the number observed in LB broth control cultures, confirming that urine from both genotypes is bactericidal. More E. coli colonies were observed on meprin αKO plates, with the mean viable bacteria in meprin αKO versus WT urine 4.6x106 and 9.2x104, respectively (Figure 17), indicating decreased bactericidal

capabilities in meprin αKO urine. It is unlikely that this effect can be directly attributed to

meprin, as the data from section 3.3.5 indicate that pre-incubation of uropathic E. coli with

active meprin A does not decrease bacterial viability. The data indicate a decrease in

bacterial killing in meprin αKO versus WT urine.

Discussion

These studies are the first to identify a correlation between urinary meprin expression

and active UTI in women and to demonstrate that meprin A has a determinative role in the

host response to bladder challenge with E. coli in mice. An experimental model of UTI was

used to examine the mechanism by which meprin A influences the course of the disease. The

results establish that meprin A contributes to E. coli-induced inflammation, as signified by

leukocyte infiltration and edema. Taken together, the data indicate that the lack of meprin A

leads to attenuation of E. coli-induced inflammatory response. These results were extended

by permeability studies of E. coli-challenged bladders, which confirmed the participation of

meprin A in increased bladder permeability/decreased barrier function during infection. The

role of meprin A in UTI pathogenesis appears to be in modulating the duration and severity

rather than the initiation of the disease. Moreover, the data do not support a direct interaction

between bacteria and the meprins as the cause of the differential inflammatory response.

Figure 17. Urine from meprin αKO mice demonstrates impaired bactericidal

properties. A sample of 150 μl of urine from WT or meprin αKO mice was incubated

overnight at 37°C with 107 uropathic E. coli. Urine reactions were serially diluted onto EMB

plates and colonies were enumerated the following morning. Each plate represents urine

from a different mouse. More E. coli colonies were observed on meprin αKO plates,

indicating a decrease in bacterial killing in meprin αKO versus WT urine. The mean viable

bacteria in KO versus WT urine is 4.6x106 and 9.2x104 per urine reaction, respectively (n=5 mice per group). The photographs are representative of two independent experiments.

3.4 Urinary meprin expression in healthy and UTI afflicted women

The data indicate that the levels of meprin A in the urine of healthy women with no history of UTI and of women with a history of recurrent UTI do not predict susceptibility to

UTI. The observation that increased levels of meprin A was found only during active infection indicates that elevated meprin in the urine is a secondary response that may be a factor in resolution of the disease. These results are in concordance with studies of meprin expression in the intestine, where meprin subunit expression is markedly increased during inflammation as a result of leukocyte infiltration (28).

Mouse urine from all mouse strains normally contains substantial quantities of protein. Soluble homomeric meprin A is one of the proteins found in high amounts into the urine of some mouse strains (e.g., C57BL/6 mice) but not others (e.g., C3H/HeJ mice).

Therefore, there is no correlation between the total amount of protein in the urine and the amount of meprin A in the urine. Urine from healthy humans contains only trace amounts of protein. Thus far, the data indicate that meprin expression is low to undetectable in urine from healthy women. These findings are similar to those observed for the defensins and

Tamm-Horsefall protein, urinary markers with established roles in UTI. For example, urinary levels of the antimicrobial defensins are low in healthy individuals but are increased nearly two-fold during active infection (101). Meprins, like the defensins, are induced in response to infection. The results in mice are parallel to the clinical observations; WT mouse urine with substantial meprin A is bactericidal whereas urine from meprin αKO mice has decreased killing activity. Further, no correlation between low urinary Tamm-Horsefall levels and a predisposition to frequent UTI was observed, although Tamm-Horsefall is known to bind type-1 piliated E. coli (102). It is unlikely that the low meprin levels are the

result of insufficient sensitivity in the Western blotting assay, as quantities as low as 1.2 ng of recombinant homomeric human meprin A have been detected using this protocol (data not shown).

Currently, UTIs are diagnosed based on clinical symptoms and observations such as white blood cells in the urine, with microbiological confirmation coming a day later.

Because urinary meprin levels are low or below detection in healthy urine, elevated levels of meprin A in the urine may ultimately be useful in UTI diagnosis. The meprin A dot blot could confirm the presence of urinary meprin in within hours, allowing a course of antibiotics to be started a day sooner. Meprin detection may also be useful in determining the successfulness of a course of treatment for UTI.

3.5 Meprin A has a pro-inflammatory role in the E. coli-challenged bladder

Upon binding to urothelial cells, E. coli induces a rapid inflammatory response, including activation and recruitment of neutrophils, which are critical to resolving both bladder and kidney infection (80). The infiltration of neutrophils to the site of infection occurs within hours during E. coli bladder colonization and is a hallmark of UTI in human and mouse (73,103). Both WT and meprin αKO bladders have a marked neutrophil influx as early as 12 h after E. coli challenge to the bladder. At this timepoint, infected bladders of both genotypes of mice have equivalent influx of neutrophils, indicating that the presence of meprin is not essential for the initial neutrophil recruitment. At later timepoints the MPO data confirm that fewer leukocytes are present in inflamed meprin αKO bladders after E. coli instillation in contrast to WT animals. This appears to be due to a continued influx of neutrophils into WT bladders, whereas influx into meprin αKO bladders plateaus and remains

essentially unchanged between 12 and 48 h post-UTI. In acute models of inflammation, chemokines such as CCL2 (MCP-1) and CCL3 (MIP-1α) are responsible for recruitment of monocytes and lymphocytes (49). The differential response of meprin genotypes may reflect either the rate at which leukocytes migrate to the site of injury or the modulation of cytokine/chemokine functionality. The observed differences could also be due to the differences in the integrity of the bladder barrier, as alterations in the integrity of the bladder wall permit entry of leukocytes.

3.6 Meprin A contributes to increased bladder permeability and edema after E. coli challenge

An important feature of the bladder is its ability to maintain a permeability barrier between urine and blood constituents, which is accomplished by the tight junctions in the uppermost epithelial cell layer. In cystitis, loss of barrier function causes leaking of urine constituents into the lamina propria of the bladder (104). Within 12 h after E. coli challenge to the bladder, leakage of NaFl from the bladder into the serum is observed, indicative of impaired barrier function in both KO and WT mice. Bladder permeability in the WT continues to increase at 48 h after E. coli inoculation, whereas the barrier of the meprin αKO bladders remains intact. Thus, the increased permeability of WT bladders would explain the significant increase in bladder edema observed in WT in comparison to meprin αKO bladders.

Bladder permeability remained elevated at five days after the initial E. coli insult. This is congruent with other studies, which demonstrate augmented bladder permeability for up to two weeks after an inflammatory challenge (93).

The genotype difference observed in bladder permeability and edema after E. coli

challenge could explain the differential neutrophil response observed in the meprin αKO bladders. Disruption of bladder permeability irritates the bladder wall, results in bladder pain, and exacerbates inflammation. Thus, the pronounced irritation in the WT bladders results in a more robust inflammatory infiltrate as the infection continues to progress.

The relationship between the neutrophil influx and edema is not well understood.

Studies in the mouse lung indicate that neutrophils are responsible for inducing organ edema.

However, neutrophil depletion studies demonstrate edema as an independent pathological marker, occurring even in the absence of neutrophil infiltrate (105). The data indicate that the presence of meprin A is not required for an increase in both edema and neutrophil infiltration but prolongs inflammation and edema in the bladder.

3.7 Correlation between bacterial counts and severity of urinary tract disease

The data indicate that more bacteria were found in the bladder than in the kidney after

E. coli challenge, with bladder counts 25-50 fold greater than kidney counts for both genotypes. These results were expected because the bacteria were perfused into the bladder, primarily entering the kidney by reflux of bladder contents. Although the count differences were not statistically significant due to high variability, there was a trend indicating that the meprin αKO mice were more vulnerable to high E. coli colonization of the bladder and kidneys in comparison to WT mice. It is unlikely that the variation observed can be explained by the mixed genetic background (e.g. varying contributions of the 129 and/or C57 background). The variation was similar among all four, independently performed experiments, and the animals exhibiting the highest bacterial counts did not share the same litter or parents. Variation in susceptibility to infection is a common occurrence in infectious

animal models and is sometimes corrected for by performing a mixed infection utilizing strains of varying virulence (100).

Although some studies have identified a correlation between bacterial counts and severity of infection, the degree to which counts accurately reflect an active E. coli infection has been questioned in studies of mouse and human UTI. For example, an examination of neutrophil recruitment in response to Pseudomonas challenge of the mouse urinary tract did not identify any correlation between neutrophil numbers and bacterial counts in kidney tissue, in spite of the fact that neutrophil response is critical to the resolution of UTI (80,106).

In humans, the correlation between bacterial counts and UTI symptoms is unpredictable, with women frequently exhibiting the symptoms of UTI with low/no bacteria (107). This phenomenon has been attributed to a combination of the phase of bacterial colonization and the stage of host inflammation (e.g., a well-established UTI versus an early-stage UTI) (103).

Further, a study comparing the clinical and laboratory findings between high and low bacterial count UTI in young children identified leukocyte counts and levels of C-reactive protein (a measure of acute phase inflammation) to be comparable in both groups (108).

3.8 Persistence of E. coli in the urinary tract

Bacterial pili have frequently been referred to as the most important determinant of pathogenicity, as they initiate the infection by allowing bacteria to resist the urine flow by binding urothelial cells (67). Data presented here indicate that homomeric meprin A does not bind to or cleave receptors or pili that might be critical for adherence of bacteria to host cells.

This is consistent with the observed low meprin expression levels in healthy human urine, and does not support a preventative role for meprins in UTI. However, this does not preclude

a determinative role for meprin A in limiting invasion or in mobilizing cellular responses to bacterial invasion.

Although the importance of pili for the retention of bacteria in the urinary tract has been established, it is clearly one of many factors that contribute to the virulence of uropathic

E. coli. This is well-illustrated in a study by Hagberg et al., which demonstrated that the addition of adhesins to a commensal E. coli strain does not confer comparable colonization abilities when compared to a uropathic strain (109).

3.9 Pros and cons of the mouse model of ascending UTI

The ascending mouse model of UTI developed by Hagberg et al. is advantageous in that it allows the investigator to study the initial stages of infection and bacterial adhesion

(100). Further, the specificity and intensity of uropathic bacterial adhesion to the mouse urothelial cells in this model is similar to that observed with human urothelial cells. Mice, unlike other animal species often used in UTI models (e.g. rabbits, rats), express globoseries glycolipid receptors in their kidneys. The globoseries glycolipids are the predominant nonacid glycolipid in the human kidney and are recognized for attachment by the majority of uropathogenic E. coli strains, making mice an ideal system for investigating strains with relevance to human UTI. Because lab animals, including mice, have a low degree of susceptibility to UTI, infections of the urinary tract have to be induced. As a result, the endpoints of the infectious model are affected by many artificially imposed parameters, including bacterial load and volume. For example, Hagberg et al. show that the percentage of mice with kidney infection increases as inoculum volume increases, even if total bacterial number remains constant (100). Volumes greater than 50 μl also increased the number of

positive blood cultures. In humans and mice, the ascent of bacteria from bladder to kidney is facilitated by vesicoureteric reflux. However, reflux, as measured by ascent of india ink and bacteria, could be induced by either increasing inoculum volume or pressure in the mouse model of UTI. Thus, careful and consistent administration of the bacterial inoculum is required to minimize mouse to mouse variation of kidney colonization. As with all disease models, the ascending model of mouse UTI has limitations. Following optimization of the aforementioned variables, use of appropriate controls, and technical mastering, the advantages of the informative ascending model of mouse UTI outweigh the limitations.

Concluding remarks

Meprins’ role in the host response to UTI has been shown to involve modulation of neutrophil infiltration, tissue edema, and barrier permeability. Meprin αKO mice exhibit an attenuated inflammatory response to E. coli challenge, implying that the presence of meprin A is pro-inflammatory. In the case of an acute immune challenge such as UTI, the loosening of the bladder barrier and infiltration of leukocytes is beneficial to the host and allows for prompt resolution of infection. Thus, the presence of meprin A in WT mice promotes inflammation and suppresses the persistence of bacteria in the bladder. Data indicate that pili alone are not sufficient to initiate the cytokine release critical to the host defense in the absence of other bacterial factors; the ability of bacterial pili to augment the bladder epithelium’s response to

E. coli was demonstrated, but only occurs in the presence of LPS (110). Further, Linder et al. demonstrate that bacterial clearance from the kidneys is mediated by LPS-induced inflammation, and that defects in this function drastically impair resistance to infection (111).

Studies with the LPS non-responder C3H/HeJ mouse strain indicate that an LPS response is

also required for neutrophil recruitment in the first 48 h after E. coli challenge (112).

Therefore, a lack of evidence for a direct interaction between meprins and bacteria directed attention to an investigation of meprins as indirect modulators of other aspects of the host response, such as the LPS response.

Chapter 4: Meprin modulates the host response to lipopolysaccharide

The inflammatory response caused by E. coli infection is in large part due to LPS, a glycolipid component of the gram negative bacterial cell wall and a potent immune stimulus

(103,112,113). While the data indicate a pro-inflammatory role for meprin A in the host response to E. coli in the urinary tract, there is no evidence indicating a direct interaction between meprins and E. coli. Thus, the hypothesis that the differential response to E. coli challenge in meprin αKO mice is a result of differences in the response to LPS was explored.

The robust host response to LPS includes upregulation and release of pro- inflammatory cytokines, mobilization of neutrophils and macrophages, and severe vasodilation due to nitric oxide release (84,85). The resulting leukocyte accumulation and hypotension cause transient organ ischemia and frequently lead to complications such as shock and multiple organ failure (114). Extensive production of pro-inflammatory mediators such as TNFα are also implicated in induction of hypothermia, a hallmark of the rodent response to endotoxemia at ambient temperatures (115-117).

Results

4.1 Response to systemic LPS

4.1.1 Meprin A contributes to LPS-induced renal injury

To determine whether meprins play a role in endotoxemia, meprin αKO and WT mice were challenged i.p. with 2.5 mg/kg E. coli LPS, and BUN and creatinine levels were measured to determine the extent of renal damage (Fig. 18A) (118,119). In LPS challenged mice, plasma BUN levels of both meprin αKO and WT were significantly higher than saline controls by 6 h, however, the BUN response to LPS treatment was markedly different

Figure 18. BUN levels in meprin KO versus WT mice after LPS challenge. WT and meprin KO mice were injected i.p. with 2.5 mg/kg LPS or saline. Blood serum samples were collected at 2, 6, 12, and 24 h post-injection and BUN levels were determined. (A) Meprin

αKO mice have lower BUN values compared to WT mice, *, p=0.0002 at 6 h, **, p<0.00001 at 12 h, and ***, p=0.04 at 24 h. (B) Meprin αKO mice have lower creatinine values compared to WT mice, *, p<0.001 at 12 h. (C) Meprin βKO mice have similar BUN levels versus WT mice. (D) Meprin αβKO mice have higher BUN versus WT mice at 24 h, #, p<0.0002. BUN data are an average of three independent experiments, with n=16-20 mice per genotype. Creatinine date are an average of two independent experiments, with n=11-12 mice per genotype. Dashed lines indicate saline injected control levels for WT and KO mice.

between meprin αKO and WT genotypes. In LPS challenged WT mice, BUN levels continued to rise until 12 h and were significantly higher compared to BUN levels in meprin

αKO mice at 6, 12, and 24 h. In WT mice, BUN levels began declining but remained elevated above saline controls 24 h post-LPS challenge. For LPS challenged meprin αKO mice, BUN levels peaked at 6 h and declined to baseline saline levels by 12 h. There was no indication of kidney damage in saline-injected meprin αKO or WT control animals at any timepoint, with average BUN levels not exceeding 24 mg/dL. No genotypic differences were observed in response to higher LPS challenges of 5mg/kg or 10mg/kg (Figure 19). Plasma creatinine levels were congruent with BUN levels, and demonstrated a marked difference in response to LPS challenge between meprin αKO and WT mice (Figure 18B). Creatinine levels in LPS challenged meprin αKO mice were considerably lower at 12 h in comparison to WT mice.

In contrast to the meprin αKO animals, meprin βKO, meprin αβKO, and WT mice showed similar BUN profiles for 12 h after challenge with LPS (Figure 18C, D). At 24 h after challenge, BUN levels in meprin βKO and meprin αβKO mice remained elevated whereas the BUN levels of WT mice began to return to normal values. While the BUN levels in the meprin αKO mice were significantly lower than in WT at 6 to 24 h after LPS injection, the BUN levels in the meprin βKO and meprin αβKO mice were the same or greater than their WT counterparts. The data demonstrate a marked difference in kidney damage in response to LPS challenge among the meprin KO mice and their WT counterparts.

This implies that meprin A contributes to LPS-induced kidney damage in the first 12 h when the meprin β subunit is expressed, but not in the absence of the meprin β subunit.

Figure 19. BUN levels are similar in meprin αKO mice and WT mice challenged with

5mg/kg i.p. LPS. WT and meprin αKO mice were injected i.p. with 5 mg/kg LPS. Blood serum samples were collected at 2, 6, and 24 h post-injection and BUN levels were determined. BUN data are an average of two independent experiments, with n=8-9 mice per genotype. Dashed lines indicate saline injected control levels for WT and KO mice.

4.1.2 Meprin A contributes to LPS-induced hypothermia

A strong systemic stimulus such as LPS induces hypothermia in mice and rats (115-

117). Indeed, the body temperatures of all genotypes of LPS-challenged mice decreased several degrees by 2 h after i.p. LPS challenge (Figure 20). WT body temperatures continued to decrease at 6 h, remained low at 12 h, and returned to normal by 24 h. In contrast, body temperatures of the meprin αKO mice began recovery by 6 h, continued to rise at 12 h, and returned to normal by 24 h (Figure 20A). By 12 h, body temperatures of LPS challenged meprin αKO were not significantly different from saline-treated controls. Both meprin βKO and meprin αβKO mice demonstrated marked hypothermia upon LPS challenge (Figure 20B,

C). The hypothermic response in the meprin βKO was marginally greater than that of the

WT, whereas that of the meprin αβKO mice was significantly greater than WT mice at both

12 and 24 h. The body temperature data are congruent with the BUN data and indicate a marked difference in the response to LPS between the meprin KO and WT mice, with significantly less hypothermia in the meprin αKO mice compared to WT and the same or enhanced hypothermia in the meprin βKO and meprin αβKO mice after LPS challenge.

These data indicate that meprin A contributes to the hypothermic response when the meprin

β subunit is expressed, and that meprin β attenuates the hypothermic response. In the absence of all meprin isoforms, hypothermia is more marked, implying that the balance between the meprin isoforms plays a role in controlling the extent of hypothermia.

4.1.3 Meprin A contributes to LPS-induced elevation of serum nitrate/nitrite levels

Serum nitrate/nitrite levels after LPS treatment peak at 12 h in both meprin αKO and

WT mice (Figure 21A). However, nitrate levels decline sharply by 24 h in the meprin αKO

Figure 20. Body temperature change in meprin KO versus WT mice after LPS challenge. WT and meprin KO mice were injected i.p. with 2.5 mg/kg LPS or saline. Body temperatures were monitored via a rectal thermistor at 2, 6, 12, and 24 h post-injection. (A)

Meprin αKO mice have less severe hypothermia versus WT *, p=0.003 at 6h; **, p=0.009 at

12 h. (B) Meprin βKO and WT mice have similar hypothermic response to LPS challenge.

(C) Meprin αβKO mice have more severe hypothermia versus WT #, p=0.004 at 6 h; ##, p<6x10-5 at 12 h; ###, p=0.0002 at 24 h. Dashed lines indicate saline injected control body temperatures. The data are an average of three independent experiments, with n=16-20 mice per genotype.

Figure 21. Serum nitrate/nitrite levels in KO versus WT mice after LPS challenge. WT and meprin KO mice were injected i.p. with 2.5 mg/kg LPS or saline. Blood serum samples were collected at 2, 6, 12, 24, and 48 h post-injection and nitrate/nitrite levels were determined. (A) Meprin αKO mice have lower serum nitrate/nitrite versus WT mice, *, p=0.041 at 24 h. (B) Meprin βKO mice have similar nitrate/nitrite levels versus WT mice.

(C) Meprin αβKO mice have significantly higher serum nitrate/nitrate versus WT mice, #, p=0.002 at 24 h; ##, p=0.02 at 48 h. The data are an average of three independent experiments, with n=13-16 mice per genotype. Nitrate/nitrite levels in saline-injected controls were low or below detection limits.

mice but remain elevated in the WT. Nitrate levels return to baseline by 48 h in both groups.

No significant difference in serum nitrate/nitrite levels was observed between the meprin

βKO and WT animals after LPS challenge (Figure 21B). However, variation within the genotype was large, and the trend indicates higher nitrate/nitrite levels in the meprin βKO mice in comparison to WT mice. Serum nitrate/nitrite levels peak at 24 h for the meprin

αβKO mice, later than the peak for the WT mice. Nitrate/nitrite levels for WT mice returned to baseline by 48 h, whereas those for the meprin αβKO mice remained elevated.

Nitrate/nitrite levels were low or below detection limits in saline-injected controls. The data demonstrate a marked difference between the meprin KO mice in response to LPS, with opposite trends in nitrate/nitrite levels. These data imply that meprins contribute to the rise and fall in nitrate/nitrite levels and particularly show that meprin A plays a role in elevated levels of this inflammatory marker.

4.1.4 Serum cytokine profiles are significantly different in WT and KO animals

In order to profile the cytokine response in the meprin KO mice, serum levels of 16 cytokines were analyzed for the first 3 h in all three meprin KO genotypes after an i.p. LPS challenge. Distinct differences were observed after LPS challenge in the levels of TNFα, IL-

1β, CCL2 (MCP-1), and CCL3 (MIP-1α) (Figure 22). Serum levels of TNFα peaked between 1 and 2 h and then began to decrease in all groups. For the meprin αKO mice,

TNFα levels decreased more rapidly at 2 and 3 h in comparison to WT mice (Figure 22A).

TNFα levels remained elevated at 3 h in the meprin αβKO mice, in contrast to WT counterparts (Figure 22B). No significant differences were observed in serum TNFα between the meprin βKO and WT mice.

Figure 22. Serum cytokine levels in KO compared to WT mice after LPS challenge. WT and meprin KO mice were injected i.p. with 2.5 mg/kg LPS or saline. Blood serum samples were collected at 0.5, 1, 2, and 3 h post-injection and the levels of 16 cytokines were determined by an ELISA-based array. (A) TNFα, IL-1β, and MCP-1 levels were significantly lower in meprin αKO mice after LPS challenge in contrast to WT counterparts.

MIP-1α levels in meprin αKO and WT mice were not significantly different. (B) Meprin

αβKO mice have significantly higher levels of TNFα and MIP-1α. IL-1β and MCP-1 levels were not significantly different. The data are an average of two independent experiments.

Results are expressed as means ± SE, with n=10-12, per genotype, per timepoint. Serum cytokine levels in meprin βKO mice were not significantly different from WT levels.

Levels of IL-1β rose sharply between 1 and 2 h and remained elevated at 3 h in LPS- challenged WT mice. In the meprin αKO mice, IL-1β levels peak at 2 h and begin to decline at 3 h, with significantly lower levels in the serum at 1 h and 3 h versus WT (Fig. 22A).

Similarly, CCL2 (MCP-1) reached a maximum at 2 h after LPS challenge and remained elevated at 3 h in WT mice. However, CCL2 (MCP-1) levels were significantly lower in meprin αKO mice than WT mice at 1 h. No significant differences were observed among the meprin βKO and meprin αβKO mice and their WT counterparts with respect to IL-1β and

CCL2 (MCP-1) profiles.

CCL3 (MIP-1α) levels rose markedly 1 h after LPS challenge in all genotypes, with significantly higher levels in the meprin αβKO mice versus WT mice at 3 h (Figure 22B).

CCL3 (MIP-1α) levels were similar in meprin αKO and meprin βKO mice and their WT counterparts. Although different expression patterns were observed in the levels of the cytokine response to LPS, only a few cytokines showed unambiguous genotypic differences.

No marked genotypic differences were observed after LPS challenge in respect to IL-2, IL-3,

IL-9, IL-10, IL-12, IFNγ, or CCL5 (RANTES) (Tables 5A-C). Statistical differences between the genotypes were observed for IL-1α, IL-4, IL-5, IL-6 and GM-CSF, but the numerical differences between the meprin KO and WT mice were marginal (Tables 5A-C).

The data demonstrate significantly lower levels of pro-inflammatory cytokines after i.p. LPS challenge in the meprin αKO animals in comparison to WT, indicating a role for meprin A in cytokine modulation.

Table 5A. Cytokine expression in meprin αKO mice after ip LPS

0.5 h 1 h 2 h 3 h

IL-1α WT 0 354±248 747±220 869±283 KO 0 228±142 731±393 505±339

IL-1β WT 0 2593±1265 6300±3020 8896±1919 KO 0 1586±708 5542±3286 4788±3580

IL-2 WT 281±363 285±325 446±272 777±377 KO 100±224 173±366 330±270 559±444

IL-3 WT 1936±751 1032±959 1636±747 3039±864 KO 1538±853 1192±830 1957±1156 2522±843

IL-4 WT 1036±653 555±428 699±276 1141±585 KO 583±669 451±239 703±358 972±547

IL-5 WT 529±528 457±109 777±311 1095±287 KO 255±569 353±68 711±375 726±252

IL-6 WT 3806±2112 12456±1594 12433±2309 11535±1503 KO 4018±2461 12670±1217 12111±2454 11030±3999

IL-9 WT 20425±14226 14612±12700 24681±10734 44218±13519 KO 10906±13897 8332±8833 23210±12769 37260±9866

IL-10 WT 495±932 2936±1297 3333±1592 4673±2183 KO 0 3803±3401 3124±1932 2957±2878

IL-12 WT 423±38 459±88 1213±317 954±183 KO 408±92 477±92 1175±482 975±178

MCP-1 WT 282±690 15842±5309 26838±8036 23079±2912 KO 0 11995±2006 25853±5366 21467±9960

IFNγ WT 11020±3433 6271±5304 7929±3695 15845±4209 KO 10055±3769 8335±5109 10485±6217 13468±3716

TNFα WT 125±306 6809±2172 5260±2523 2358±995 KO 135±190 5789±2048 2417±944 662±1238

MIP-1α WT 0 684±865 7843±8506 16128±4548 KO 0 255±633 8874±6356 16701±4321

GMCSF WT 1453±661 835±774 1326±393 2401±747 KO 1099±892 984±565 1706±606 2339±670

RANTES WT 0 0 870±940 3904±630 KO 0 0 268±357 2689±1634

Table 5B. Cytokine expression in meprin βKO mice after ip LPS

0.5 h 1 h 2 h 3 h

IL-1α WT 0 459±223 1079±166 1032±122 KO 55±42 412±101 891±487 1295±548

IL-1β WT 1300±499 4204±634 7331±1026 7107±790 KO 1412±250 3813±555 6102±1484 7107±790

IL-2 WT 1±254 789±137 1068±182 919±83 KO 121±178 630±226 909±172 899±122

IL-3 WT 643±74 777±381 1556±384 1153±199 KO 535±279 750±346 1637±1136 1180±262

IL-4 WT 467±156 851±128 1072±284 979±165 KO 444±64 723±180 1002±180 933±133

IL-5 WT 107±127 616±109 909±95 963±80 KO 150±129 573±69 817±148 887±119

IL-6 WT 2780±895 21690±2805 19559±2670 18121±942 KO 1075±652 20172±1969 19426±1625 17589±1434

IL-9 WT 11528±4234 27417±8565 31084±4609 24239±4373 KO 10061±1339 24972±4923 30839±8503 29128±7000

IL-10 WT 0 8240±6118 9120±3015 10440±1254 KO 540±3444 6040±1656 7690±3755 7580±2626

IL-12 WT 140±17 352±83 1452±232 1619±42 KO 111±61 373±50 1506±273 1530±88

MCP-1 WT 939±370 23602±6248 36634±4186 35220±1265 KO 1469±916 22409±7336 36855±3807 33895±2712

IFNγ WT 4795±736 6289±1966 8137±2419 6377±1391 KO 4265±2295 5761±2753 14825±10521 8841±3650

TNFα WT 0 12854±4274 7226±3003 3747±667 KO 0 13212±5849 6356±3529 3747±560

MIP-1α WT 0 2187±1951 24088±10318 22899±4146 KO 0 2385±1688 21412±8705 19331±4239

GMCSF WT 349±167 885±313 1665±292 1320±146 KO 298±214 949±490 1614±451 1524±271

RANTES WT 0 0 1526±953 3077±908 KO 0 0 1840±949 3789±618

Table 5C. Cytokine expression in meprin αβKO mice after ip LPS

1 h 2 h 3 h

IL-1α WT 637±332 1389±674 1576±382 KO 280±139 956±343 1201±644

IL-1β WT 3422±306 6437±828 6661±828 KO 3422±499 5990±1045 6214±1958

IL-2 WT 600±173 919±83 1058±194 KO 600±291 899±71 939±320

IL-3 WT 750±112 1583±204 1664±329 KO 831±221 1368±240 1744±662

IL-4 WT 467±104 839±52 816±104 KO 420±128 630±184 839±152

IL-5 WT 421±59 779±74 768±45 KO 421±89 671±106 768±169

IL-6 WT 23128±1212 21690±720 21850±1434 KO 22915±1767 22489±1258 22649±2966

IL-9 WT 25706±6327 34506±4441 35484±3187 KO 24728±2186 35484±4373 39884±7042

IL-10 WT 1860±2761 5820±1968 4280±920 KO 760±1841 6700±3564 3840±2972

IL-12 WT 328±65 1742±76 1775±107 KO 361±89 1624±345 1756±241

MCP-1 WT 20377±7174 40345±1890 40521±2970 KO 27622±4825 40610±3813 40875±5948

IFNγ WT 4969±2109 10249±5541 12537±997 KO 6377±3027 10601±2835 13241±1560

TNFα WT 8761±2109 12138±5541 2417±997 KO 8249±3027 4054±2835 4872±1560

MIP-1α WT 502±543 21511±9072 15764±6476 KO 304±2659 19728±12563 34791±11933

GMCSF WT 681±232 1805±574 1652±300 KO 911±437 2086±821 2265±277

RANTES WT 0 1023±381 4166±806 KO 0 898±663 3999±1180

Tables 5A-C. Serum cytokine levels in KO compared to WT mice after LPS challenge.

WT and meprin KO mice were injected i.p. with 2.5 mg/kg LPS or saline. Blood serum samples were collected at 0.5, 1, 2, and 3 h post-injection and the levels of 16 cytokines were determined by an ELISA-based array. (A) Cytokine expression in meprin αKO and WT mice (B) Cytokine expression in meprin βKO and WT mice (C) Cytokine expression in meprin αβKO and WT mice. The data are an average of two independent experiments.

Results are expressed as means ± SD, with n=10-12, per genotype, per timepoint. Bolded values indicate p <0.05 KO versus WT.

4.1.5 Serum TNFα levels are significantly different in WT and KO animals

To confirm the genotypic differences observed in the levels of TNFα determined by cytokine array, ELISA was performed (Figure 23). TNFα levels rose sharply between 30 and 60 min after LPS challenge for all genotypes. After reaching maximum TNFα levels between 1 and 2 h, the TNFα levels for all genotypes returned to near baseline levels by 3 h.

For the meprin αKO mice, TNFα levels decrease more rapidly at 2 and 3 h versus WT mice

(Figure 23A). This trend was also consistent at later time points after LPS challenge, with decreased TNFα levels in the meprin αKO mice at 12 h (αKO 53 ± 32 pg/ml vs. WT 173 ±

55 pg/ml. p=0.002). For the meprin αβKO mice, TNFα levels are significantly lower than

WT at 2 h. However, they remain elevated at 3 h in αβKO mice, in contrast to WT counterparts (Figure 23C). No significant differences were observed in serum TNFα between the meprin βKO and WT mice (Figure 23B). TNFα levels in saline controls were low or below detection limits for all groups (data not shown). Thus, the ELISA data are congruent with the cytokine array data and confirm that TNFα levels fell more rapidly in mice lacking meprin A but possessing meprin B in comparison to WT mice, whereas TNFα levels remained elevated at 3 h in mice lacking all meprin subunits.

4.2 Response to bladder LPS

4.2.1 Meprin A contributes to the leukocyte response and edema in a model of bladder inflammation

To relate the systemic studies to a specific target organ, meprin αKO and WT mice were challenged with LPS transurethrally, and myeloperoxidase (MPO) activity was measured to determine the extent of leukocyte infiltration into the bladder (120).

Figure 23. TNFα levels are significantly different between genotypes after LPS challenge. TNFα levels were measured in mouse serum 0.5, 1, 2, and 3 h after IP challenge with 2.5 mg/kg LPS. (A) Meprin αKO mice have lower serum TNFα levels versus WT mice,

*, p=0.02 at 2 h; **, p=0.01 at 3 h. (B) Meprin βKO mice have similar TNFα levels versus

WT mice. (C) Meprin αβKO mice have lower serum TNFα levels at 2h and higher TNFα levels at 3 h versus WT, #, p=0.012 at 2 h;##, p= 0.0005 at 3 h. Results are expressed as means ± SE, with five to six mice per genotype, per timepoint. The data are representative of at least two independent experiments.

Substantially more MPO activity was observed 48 h after LPS treatment in WT versus meprin

αKO bladders, indicating an attenuated host response in the meprin αKO mice (Figure 24A).

MPO activity was low in saline controls (0.3±0.18 MPO units/g). In addition, LPS-treated bladders were significantly heavier in WT versus meprin αKO mice as early as 3 h after LPS instillation, indicating more extensive edema in the WT bladders (Figure 24B). To determine whether LPS instilled in the bladder induced kidney damage, serum samples were collected and analyzed for BUN at 24 h. All levels were normal and not significantly different between

WT and meprin αKO genotypes (data not shown), confirming that the LPS challenge did not induce a systemic host response; these results are consistent with a rat model of LPS-induced bladder irritation (121). The data indicate a decreased inflammatory response to bladder LPS challenge in mice lacking meprin A in comparison to WT counterparts.

4.2.2 Meprin A contributes to increases in bladder permeability after an LPS bladder challenge

To determine the extent of bladder damage inflicted by LPS-induced bladder inflammation, bladder permeability after LPS challenge was determined by measurement of

NaFl leakage from the bladder into the serum. At several timepoints after transurethral LPS administration, NaFl instilled into the bladder appeared in the blood (Figure 25). Significantly more NaFl was detected in the blood of WT samples than in meprin αKO samples at 6 h after

LPS challenge. By 12 h, both meprin αKO and WT mice displayed substantial loss of barrier function, with the WT affected to a greater extent than the meprin αKO mice. Serum NaFl levels of saline treated control WT and meprin αKO mice were at or below detection limits at all timepoints (data not shown). The data establish that mice lacking meprin A but possessing

Figure 24. Less bladder MPO activity and edema in meprin αKO versus WT after

transurethral administration of LPS. LPS (20 μg) was instilled into the bladder via catheter. Mice were sacrificed at multiple timepoints, and leukocyte infiltration was measured by MPO assay. (A) Bladders from WT mice had significantly more MPO units/g bladder tissue versus meprin αKO, *, p=0.01 at 24 h; **, p<0.03 at 48 h. (B) Bladder to body weight ratios were higher in WT versus αKO, indicating more bladder edema, *, p=0.007 at 3 h; **, p<0.01 at 6 h; *** p<0.005 at 12 h; #, p<0.02 at 24 h. The data are an average of two independent experiments, with n=8-12 mice per genotype, per timepoint.

Dashed lines indicate average value for saline-treated control animals.

Figure 25. Less bladder permeability in meprin αKO mice after LPS challenge. After

LPS bladder challenge, NaFl was instilled into bladders at 6, 12, 24, or 48 h and serum was analyzed after 15 min. Leakage of NaFl into the serum was significantly higher in WT versus

αKO mice, *, p=0.02. The data are an average of two independent experiments, with n=14 mice per genotype. Serum NaFl levels for saline treated controls were below detection limits.

meprin B have less inflammatory damage and subsequently more intact bladder barrier function after LPS bladder challenge compared to their WT counterparts.

4.2.3 Meprin A disrupts the tight junctions between kidney cells

Because tight junctions are important in the maintenance of bladder barrier function and regulation of bladder permeability, monolayers of MDCK cells were treated with active or latent recombinant human homomeric meprin A to assess its effects on tight junction integrity. Staining for occludin and ZO-1 revealed disruption of tight junctions between

MDCK cells of cultures treated with active meprin A but not in untreated cultures or cultures treated with latent meprin A (Figure 26). LDH release from cell cultures treated with latent or active meprin A was not significantly different from control cells, indicating that meprin is capable of disrupting the tight junctions without having a cytotoxic effect.

To determine if this disruption was due to direct cleavage of ZO-1 or occludin by meprin, protein extracts from meprin-treated MDCK cells were analyzed by Western blotting.

Western analysis with an anti-ZO-1 antibody indicated bands of similar size and intensity in protein extracts from active meprin A-treated versus non-treated control cells (Figure 27A).

An additional, lower molecular weight band of nonspecific immunoreactivity was also present in all lanes, regardless of treatment group. Similar results were obtained when Western analysis was performed on the same extracts using an anti-occludin antibody (Figure 27B).

The data do not provide any evidence that meprins directly cleave occludin or ZO-1.

Figure 26. Meprin A disrupts the tight junctions of kidney cells in culture. Confluent

MDCK cells were treated with active or latent recombinant homomeric meprin A and stained for nuclei (blue), occludin (green), and ZO-1 (purple). Occludin and ZO-1 staining are diminished after treatment with homomeric meprin A while nuclear staining is comparable in all treatment groups. White arrows indicate areas of marked tight junction disruption. (A)

Tight junctions are intact in control cells. (B) Tight junctions are disrupted after treatment with active homomeric meprin A. (C) Tight junctions are intact after treatment with latent homomeric meprin A.

Figure 27. No evidence that meprin directly cleaves tight junction proteins ZO-1 or occludin. (A) Western blot of MDCK cell protein extracts probed with anti-ZO-1 antibody.

(B) Western blot of MDCK cell protein extracts probed with anti-occludin antibody. The size, intensity, or number of immunoreactive bands is not altered in meprin-treated MDCK cell extracts in comparison to control non-treated cells.

4.2.4 Less leukocyte migration into the peritoneal cavity in meprin αKO mice

To determine whether macrophage populations and/or their migration are affected by meprin expression, resident and TG-elicited PECs were isolated from meprin αKO and WT mice and cell numbers were determined. Six h after TG challenge, an increase in the number of PECs was observed in the peritoneal cavities of both meprin αKO and WT mice in comparison to resident cell counts (Figure 28). Significantly more PECs were present in the peritoneal cavities of WT mice in comparison to meprin αKO mice at 12 and 24 h after

TG challenge. The number of resident cells was similar between meprin αKO and WT mice

(dashed line, Figure 28). The data indicate that PECs are capable of migration in the absence of meprin α, as the number of PECs in the TG challenged meprin αKO mice is greater than the number of resident cells present in the peritoneal cavities of non-challenged animals.

However, the significantly lower number of PECs in meprin αKO mice in contrast to WT mice indicates that PEC migration is impaired in the absence of meprin α.

Discussion

These studies with the meprin KO mice are the first to demonstrate that meprin A has a determinative role in host response to challenge with LPS. The data establish that the presence of meprin A contributes to LPS-induced hypothermia, renal damage, and pro- inflammatory cytokine production. In the meprin αβKO mice, which lack all forms of meprin, the LPS-induced kidney damage and hypothermia are more severe than in WT mice.

Taken together, the data indicate that both α and β subunits contribute to the host response to

LPS, and that the absence of meprin A and the presence of meprin B contribute to attenuation of LPS-induced injury. These results were confirmed and extended in an LPS model of

100

Figure 28. Less cell infiltration into the peritoneal cavities of meprin αKO mice. After

TG challenge, mice were necropsied at 6, 12, or 24 h and PECs were isolated by peritoneal lavage. The number of elicited meprin αKO PECs was significantly lower in meprin αKO versus WT mice, *, p=0.006 at 12 h; **, p<0.02. The data are an average of two independent experiments, with n=4-6 mice per genotype. Dashed line indicates an average of resident

PECs from meprin αKO and WT mice.

101

bladder challenge, which highlights participation of meprin A in increased leukocyte infiltration and edema in bladder tissue. The observed breakdown in the bladder barrier can be attributed to the ability of active meprin A to disrupt cellular tight junctions, as bladder permeability after LPS challenge was markedly decreased in mice lacking meprin A.

4.3 The role of TNFα and Il-1β in kidney damage

Kidney damage is mediated by pro-inflammatory cytokines that are released into the circulation in early phases of the LPS response (116). TNFα has direct and indirect cytotoxic effects as a consequence of LPS challenge that have been associated with renal failure, and acts synergistically with IL-1β in the endotoxemic response (86,115,122). Serum levels of

TNFα and IL-1β are lower in mice lacking meprin A but possessing meprin B after LPS challenge. Thus, it is likely that the prolonged elevation of cytokine levels in WT mice contribute to the more extensive kidney damage observed compared to the meprin αKO mice.

Consistent with these results, previous work with C3H/He mice that have low levels of kidney meprin A show less kidney damage when subjected to ischemia reperfusion in contrast to

C57BL/6 mice with high levels of meprin A (50). TNFα and IL-1β levels in the meprin βKO mice and WT mice were the same during the 3 h after LPS administration, and BUN levels were not statistically different. In the mice lacking all meprin subunits, TNFα levels do not return to baseline and are higher than those of WT mice at 3 h after LPS challenge.

Additionally, CCL3 (MIP-1α) levels were markedly elevated in the meprin αβKO mice, perhaps because there was no meprin A present, which is capable of degrading CCL3 (MIP-

1α) (41). CCL3 (MIP-1α) is a potent monocyte chemoattractant that is capable of recruiting macrophages and exacerbating kidney damage, which may explain the prolonged elevation of

102

BUN levels in the meprin αβKO mice (123).

4.4 Cytokine-mediated hypothermia

TNFα possesses both pro- and anti-pyretic properties and, along with IL-1β, modulates the hypothermia observed in mice after LPS challenge (115). Therefore, the attenuated hypothermic response in the mice lacking meprin A but possessing meprin B can be attributed to the lower levels of TNFα and IL-1β, in contrast to WT mice. IL-1β and

TNFα levels are congruent in meprin βKO and WT mice, confirming that the presence of meprin A and meprin B have opposing physiologic consequences. The dramatic hypothermia observed in the meprin αβKO mice in comparison to WT mice can be attributed to higher levels of TNFα and CCL3 (MIP-1α) at 3 h, which play a role in modulation of the febrile response (124,125).

4.5 Cytokine levels and iNOS release

LPS triggers the synthesis of iNOS, which results in an increase in production of nitric oxide, a potent vasodilator and known mediator of organ failure in endotoxemia (122). TNFα and IL-1β have redundant activities, including eliciting nitric oxide production (126,127).

The lower levels of TNFα and IL-1β in serum of mice lacking meprin A but possessing meprin B, as compared to the levels in WT mice after LPS administration, would account for the lower levels of nitrate/nitrite in the meprin αKO mice (127). In the mice lacking all meprin subunits, the nitric oxide response is biphasic, with an initial phase that lags below

WT levels until 12 h and a late phase in which the nitrate/nitrite levels in meprin αβKO mice surpass WT levels by 24 h and remain elevated at 48 h. TNFα levels in meprin αβKO mice

103

are lower at 2 h, in contrast to WT mice. However, by 3 h TNFα and CCL3 (MIP-1α) levels are markedly increased in mice lacking all meprin subunits. CCL3 (MIP-1α) has been implicated in mediation of nitric oxide release by macrophages and could contribute to the increased nitrate/nitrite levels observed in the meprin αβKO mice in the late phase of the nitric oxide response (128).

4.6 The role of cytokines in inflammatory disease models

In inflammatory disease models, a misbalance in metalloprotease levels is observed before the appearance of inflammatory infiltrate, and failure to properly regulate the subsequent inflammatory cascade can lead to sustained TNF levels and impaired tissue repair

(129). A majority of the damage induced by LPS challenge can be attributed to a dysregulation of cytokine levels as a result of the robust inflammatory response (86).

Furthermore, it has been shown that an imbalance in TNFα levels 1-4 h after an initial immune insult results in persistence of biochemical and hemodynamic perturbations for as long as 18 h (130). Therefore, early differential regulation of TNFα, IL-1β, and CCL3 (MIP-

1α) could account for the marked differences observed in BUN levels, hypothermia, and nitrate/nitrite levels between 12 and 24 h after LPS challenge in the different meprin genotypes.

LPS upregulates the synthesis of IL-10, which functions as an antagonist of TNFα, IL-

1, and IL-6 (131). However, IL-10 levels were not significantly different at any of the timepoints examined, ruling out upregulation of anti-inflammatory cytokines as the cause of the meprin αKO hyporesponsiveness. Similarly, IL-6, a cytokine that decreases TNFα levels and protects the host against the adverse effects of LPS, is not different in meprin KO versus

104

WT mice. Although IFNγ, like TNFα, is a pleotropic cytokine involved in modulation of the

LPS response, no significant difference was observed between genotypes in IFNγ levels. This is consistent with the observations that the IFNγ and TNFα pathways are distinct and differentially modulated.

Although cytokine levels obtained from the array data allows conclusions to be

drawn, it is important to note that there are limitations in interpreting ELISA data. For

example, proteases such as MMP2 and MMP9 inactivate chemokines via cleavage or activate

chemokines through release of an active fragment. However, MMP2 and MMP9-deficient

mice demonstrate decreased levels of chemokines in bronchoalveolar lavage fluid (BALF)

during lung inflammation. The authors explain this paradox by comparing ELISA data to the

functional activity of the cytokines as measured by chemotaxis assay. The results are quite

distinct, indicating that while total concentration of chemokines as detected by ELISA is

decreased, their cleaved forms may be more biologically active (132).

4.7 LPS-induced inflammation and leukocyte recruitment

The inflammatory cascade that LPS sets in motion is amplified by leukocyte

recruitment (133). In acute models of inflammation, chemokines such as CCL2 (MCP-1)

and CCL3 (MIP-1α) are expressed at early timepoints and are responsible for recruitment of

monocytes and lymphocytes (49). CCL2 (MCP-1) is secreted in response to LPS as well as

to cytokines such as TNFα and IL-1β and has been shown to modulate the endotoxemic

response in the intestine, lung, and brain (133,134). At 1 h post LPS challenge, CCL2

(MCP-1) levels are moderately lower in the serum of meprin αKO mice. Consistent with

these results, the MPO data confirm that fewer leukocytes are migrating to the inflamed

105

meprin αKO bladders 24 h after LPS instillation in contrast to WT animals. However, LPS- elicited neutrophil migration to the bladder wall was initially the same in mice expressing meprin A and those lacking meprin A, indicating that altered neutrophil migration is not responsible for the early differences in response of WT and meprin αKO mice. The differential response of meprin genotypes may reflect either the rate at which TNFα- producing leukocytes migrate to the site of injury, or the modulation of TNFα functionality.

Current data indicate that the effect is not on synthesis of TNFα, as levels of TNFα are comparable at 1 h after LPS challenge.

At the same time, breakdown of ECM components is also essential for neutrophil influx after an inflammatory challenge. For example, studies have identified a novel neutrophil chemoattractant, PGP, which is released via MMP9 cleavage of collagen. In a model of LPS-induced lung inflammation, PGP enhances neutrophil recruitment while other chemokine levels (e.g., KC, MIP-2) remained unchanged (135). Thus, it is possible that the decreased neutrophil influx observed in the meprin αKO mice is due to cleavage of a yet to be identified meprin substrate.

4.8 Barrier function of bladder and disruption of tight junctions

An important feature of the bladder is its ability to maintain a permeability barrier between urine and blood constituents, which is accomplished by the tight junctions in the uppermost epithelial cell layer (104). Maintenance of proper barrier function is crucial in protection from endotoxemia (133), and LPS can disrupt the barrier function by both direct action and as a secondary effect of the inflammatory response (79,136). Cytokines are capable of modulating the permeability of the bladder epithelium, and TNFα is capable of

106

increasing epithelial permeability in a dose-dependent manner (136). Loss of barrier function causes leaking of urine constituents into the lamina propria of the bladder and higher TNFα levels could explain the significant increase in bladder edema observed in WT in comparison to meprin αKO bladders (104).

Protease alteration of tight junctions during pathological states has been documented, and proteases can alter permeability via direct proteolysis of tight junction and adherens junction complexes (79,137,138). Several lines of evidence implicate metalloproteases in disruption of junctional adhesion proteins, but specific enzymes are not identified (138,139).

For example, the tight junctions of kidney cells treated with TNFα undergo a metalloprotease- dependent cleavage of occludin that results in increased permeability, indicative of impaired barrier function (138). This works provides evidence that the meprins are among the proteases actively involved in modulation of tight junctional proteins, as treatment of kidney cells in culture with active meprin A resulted in disruption of the tight junctions between these cells.

Evidence indicates that proteases derived from both neutrophils and endothelial cells are responsible for partial degradation of the VE-cadherin complex, and endothelial cadherin, which make up adherens junctions between cells (140). Soluble homomeric meprin A, which is abundantly expressed in mouse urine, could be responsible for the differences in barrier integrity observed in the meprin αKO bladders. Meprin α and β subunits are also expressed in non-macrophage populations of WT mouse leukocytes. In the mesenteric lymph node, the meprin β subunit, but not the α subunit, is expressed in normal macrophages (28).

Disruption of tight junctions is frequently associated with edema during inflammatory states (141). Loss of barrier function causes leaking of urine constituents into the lamina

107

propria of the bladder and would explain the significant increase in bladder edema and permeability observed in WT in comparison to meprin αKO bladders (104). The proposed role for meprins in LPS-treated mice is congruent with observations in an ischemic kidney model, where alteration of occludin and ZO-1 distribution results in loss of barrier function and leakage of the glomerular filtrate back across the damaged epithelium (142).

The data indicate that although meprin is capable of disrupting tight junctions, there is no evidence that meprin directly cleaves the tight junction proteins occludin or ZO-1. Several other proteins also contribute to tight junction formation and are necessary for its stability

(e.g., claudins). In addition, the tight junctions are intimately connected to the adherens junctions, and it has been demonstrated that the adherens junctional protein e-cadherin regulates the tight junctions and is required for barrier function (143). Recently, cleavage of e-cadherin by meprin B has been demonstrated in vitro (D. Lottaz, personal communication).

Thus, it is possible that meprin is interacting with or proteolytically cleaving other components of the tight or adherens junctions that were not analyzed.

Concluding remarks

Meprins’ role in the host response to LPS has been shown to involve modulation of serum chemokine levels as well as epithelial barrier function, at least in part attributable to disruption of tight junctions. Meprin αβKO mice, which lack all meprin isoforms, exhibit greater hypothermia and higher BUN and nitric oxide levels than WT or meprin αKO mice, implying that meprin B has a role in attenuating the injurious action of meprin A. These studies are the first to demonstrate that meprin A selectively modulates the levels of TNFα and IL-1β elicited by LPS and that meprin A disrupts tight junctions in vitro and in vivo,

108

leading to increased epithelial permeability, greater leukocyte infiltration, and enhanced bladder edema.

109

Chapter 5: General Discussion

Inflammation is a complex and highly coordinated process. Ideally, the inflammatory response promptly and efficiently mobilizes leukocytes, kills invading microbes, and repairs any incidental damages to healthy host tissue. If inflammation is allowed to proceed in an unchecked fashion, the deleterious results can include accumulation of leukocytes in tissues, fibrosis, organ failure, and possibly death. It is important to note that maintenance of, or return to, a non-inflammatory state is the result of positive actions of effector molecules that actively suppress the inflammatory response as well as the absence of pro-inflammatory stimuli (144). The host response is further complicated by the fact that some mediators of inflammation can serve both pro- and anti-inflammatory roles (e.g., TNF, TGFβ).

Ultimately, a better understanding of the factors that modulate inflammation will lead to interventions that upregulate and downregulate inflammatory responses favoring outcomes that are more beneficial than detrimental to the host. A major step closer toward this goal is the identification of proteases that participate in the host response.

The contribution of proteases to the progression of the inflammatory response has been documented, and it has been proposed that proteases play roles from early establishment of chemokine gradients to the later phases of wound healing and tissue repair. For example, metalloproteases have been implicated in leukocyte recruitment via ECM remodeling as well as in activation of cytokines, whereas mast cell tryptases cleave receptors to promote leukocyte-endothelium interactions (144,145). A balance of protease activity is critical to the proper functioning of the inflammatory response, as proteases participate in highly coordinated cascades with other molecules (146). For example, the mouse chemokine KC

(CXCL1) accumulates on the membrane-bound proteoglycan syndecan-1 of injured cells.

110

MMP7 sheds the ectodomain of syndecan-1, thereby establishing a chemokine gradient and controlling neutrophil influx (2). Thus, the primary focus of this work was to utilize meprin

KO mice to characterize the role of these metalloproteases in the host response to infectious and noninfectious inflammatory challenges.

5.1 The inflammatory response in systemic versus localized and infectious versus noninfectious challenge models

Taken together, the data herein identify meprin A as a modulator of the pro- inflammatory host response in systemic and bladder-specific LPS challenges as well as in an infectious E. coli challenge. In addition to shedding light on meprin’s role in the immune response, these studies also illustrate reasons that the immune system is often referred to as

“a double-edged sword (85,147).” In the infectious model of UTI, meprins contribute to a localized pro-inflammatory response to LPS and other bacterial products that induce controlled leukocyte recruitment to eradicate the pathogenic bacteria. Induction of inflammation after an acute E. coli challenge such as UTI is considered beneficial and necessary for prompt bacterial clearance. In contrast, the augmented inflammatory response to systemic LPS challenge contributes to widespread leukocyte infiltration. In cases of inappropriate inflammatory response (e.g., sepsis), leukocytes localize to specific organs

(e.g., lung, kidney), become trapped, and are not capable of reentering the circulation. Tissue damage, impaired function, and ultimately organ failure can result (148). An excellent example of the duality of the LPS response is found in studies with TNF KO mice. While these mice are protected against the lethal effects of septic shock, they easily succumb to minute infectious challenges (85,149).

111

Thus, whether the host inflammatory response to LPS is a “good thing” or a “bad thing” is clearly dictated by the nature of the microbe, the location of the injury, and the state of the host. Similarly, whether the participation of meprins in inflammation is beneficial or detrimental to the host depends on the situation in question, as demonstrated by this work and other disease models using meprin KO mice. Merging the information gained from this work with previous knowledge about meprins in inflammatory diseases will provide insight into novel roles for meprins in immune system modulation.

5.2 Meprin participation in the inflammatory response: acute versus chronic

A multitude of diseases in which inflammation contributes an important pathological role have been identified (Table 6) (144). Interestingly, meprins have been implicated in several of these conditions, including diabetes, ischemia-reperfusion injury, IBD (Crohn’s disease, ulcerative colitis), and sepsis (Figure 29). In rodent models of hydronephrosis and ischemic and nephrotoxic (folic acid-induced) renal tubular necrosis, meprin β RNA expression is downregulated (53,57). Further, pretreatment with actinonin, a potent meprin

A inhibitor, decreased nephrotoxicity in mice and rats with cisplatin-induced or I/R-induced acute kidney injury, respectively (56,58). The most striking evidence comes from kidney ischemia/reperfusion studies with meprin βKO mice, which demonstrate that WT mice were more susceptible to renal injury than mice deficient in meprin β (51). Thus, the previous renal failure studies in combination with the UTI and LPS data herein implicate a pro- inflammatory role for the meprins during acute inflammatory challenges. However, the same conclusions cannot be drawn from models of chronic disease and inflammation.

112

Table 6. Inflammatory diseases. Controlling inflammation is a major goal in a diverse array of diseases. In some cases, the inflammation causes as much or more damage than the invading microbes. Table adapted from Nathan, C. (2002) Nature 420, 846-852.

113

Figure 29. Meprins in disease and inflammation. The meprins have been implicated in acute inflammatory as well as chronic diseases. In all cases, the inflammatory response plays important roles in the progression, severity, and resolution of the disease. The work herein focused on the role of meprins in acute inflammation and shed light on how meprins could mediate early events in the inflammatory response while previous data demonstrated meprins’ participation in chronic conditions. Taken together, the data implicate meprins in the initiation, progression, and resolution of many diseases that have an inflammatory component. Asterisks (*, **) expand on two aspects of the inflammatory response.

114

In acute models of inflammation, the host response to an immune challenge such as

LPS has a limited timecourse. Perturbations in the immune system are evident within hours, with resolution of inflammation and return to pre-challenge conditions in 48 to 96 h. In contrast, chronic models of inflammation may take days to weeks for induction of inflammation, with no resolution for months after an initial insult. In streptozotocin-induced diabetes in rats, a chronic model of kidney dysfunction, meprin β is upregulated in the kidney

(61). Additional studies from the same authors revealed differential expression of meprin A in mice with chronic diabetic nephropathy and in rats with streptozotocin-induced diabetes at

52 weeks. In both cases, decreased meprin A expression in the kidney was correlated with more severe renal injury (59). The authors conclude that the role of meprin A in the

pathogenesis of renal damage may depend on the acuity versus chronicity of the disease

process. This is consistent with models of hydronephrosis and adriamycin-induced

nephropathy, which implicate decreased meprin levels in chronic nephropathy and fibrosis

(53,60). Taken together, these studies imply a protective role for meprins in models of

chronic kidney disease.

A protective role for meprin A has also been proposed in a model of DSS-induced

IBD. Oral administration of 3.5% DSS in drinking water over a period of several days

induces chronic colitis in mice. In this model, meprin αKO mice demonstrate greater

susceptibility to injury and inflammation in comparison to WT. Unlike the models of kidney

disease, which implicate only the tissue-expressed meprins in the progression of renal

pathologies, data from the IBD model indicates a role for both intestinal meprins and

leukocytic meprins. Crisman et al. identified expression of both meprin α and β in

leukocytes isolated from the mesenteric lymph node in mice, and Sun et al. recently

115

confirmed meprin α expression in human peripheral mononuclear leukocytes (28, Q. Sun, personal communication). Banerjee et al. suggest that leukocytic meprins modulate pro- inflammatory cytokine cascades while intestinal meprins contribute to reepithelialization and restoration of the damaged intestinal lining (55). Thus, in addition to performing diverse functions in chronic and acute conditions, it is also likely that meprins play multiple roles within a single pathology.

5.3 Endotoxin and IBD

At first glance, the connection between the acute response to LPS that was explored in the current work and the chronic inflammation observed in IBD is not immediately apparent. However, data indicate that an insufficient ability to recognize LPS in the intestine can contribute to a dysfunctional mucosal immune response in the gut (150). This is especially interesting because meprin A has been implicated in the LPS response as well as in IBD. In the DSS model of colitis, TLR4 KO mice, which are hyporesponsive to LPS, demonstrated earlier and more severe rectal bleeding in comparison to DSS-challenged WT mice (150,151). The authors conclude that a defective innate immune response ultimately contributes to an abnormal immune response to normal intestinal flora. This is consistent with other IBD research that suggests a link between chronic intestinal inflammation and an exaggerated mucosal immune response to commensal flora. Thus, it is reasonable to propose that meprin is involved in host self-recognition. In addition to meprin’s role in modulating the LPS response and IBD inflammation, the pattern of meprin expression is consistent with a potential role in host self-recognition. For example, unlike the majority of MMPs, which are expressed in response to a stimulus, meprins are constitutively expressed in many tissues,

116

implying a regulatory role during non-inflammatory states. Further, meprin is expressed in the skin and the gut, interfaces between the host and the environment that are continually exposed to microbes. Meprins are highly expressed in the kidneys, which are prone to trapping immune complexes, and are coded for near the MHC genes, which are also involved in self-tolerance (144,152).

5.4 Meprins modulate barrier function

Meprins have been implicated in other homeostatic processes that are disturbed

during acute and chronic inflammatory disease. For example, meprin A deficient mice had

more intact barrier function after LPS or E. coli challenge to the bladder and WT bladders

compared to meprin αKO are significantly more permeable and edematous within hours after

an inflammatory challenge. Meprin A also disrupted tight junctions between kidney cells in

culture. Taken together, the data implicate meprin A in early stages of barrier disruption

Maintenance and repair of epithelial barrier function is important during the pathogenesis of

IBD as well as UTI. In contrast to the UTI model, the epithelial barrier of the intestine was

still disrupted 5, 6, and 7 after DSS-induced colitis, and the damage was comparable in

meprin αKO and WT mice. However, a recovery study at day 10 post-DSS demonstrated

markedly more intact intestinal barrier in WT mice in comparison to meprin αKO mice (55).

These findings are similar to a model of acute cerebral ischemia, which implicates

dual roles for MMP9 in the acute and chronic phases of the insult. MMP9 deficient mice are

protected during an acute ischemic challenge, with more intact tight junctions in the blood

brain barrier (BBB) and less brain edema. As the brain recovers, MMP9 is beneficial,

contributing to the re-growth of vascular structures and remapping of axonal connections

117

(153-155). Taken together, the DSS data indicate that meprin does not play a role in maintenance of the intestinal barrier, but is important in repair, as the presence of meprin A expedites the healing of the damaged intestine. The UTI data implicates meprins in compromised bladder integrity and edema. The data are consistent with the idea that meprins are capable of playing multiple roles within the same condition and throughout the progression of the inflammatory response.

5.5 Meprins in wound healing

The final stage of the inflammatory response is the repair of host tissues that were damaged by the initial insult or the host response itself. Many proteases are required for reepithelialization, and their ability to relax cellular junctions (e.g., via shedding of cadherin) facilitates cell migration (2). The observations from the previous section invite a closer look at meprins in wound healing. The current data indicate a role for meprins in disrupting epithelial barriers and promoting leukocyte migration via loosening of cell contacts.

Previous data implicate meprins in breakdown and turnover of ECM proteins. These proteolytic processes are equally vital to wound healing as they are to the early infiltration of neutrophils. Initial experiments in the laboratory have identified meprin protein expression in human wound fluid (T. Kieffer, personal communication). Further, preliminary wound healing studies with meprin αKO and WT mice demonstrated delayed healing in meprin A deficient mice (P. Erlich, personal communication). Additional studies will help to reveal meprin’s contribution to host repair.

118

5.6 Meprin protein-protein interactions

Although the majority of this body of work focused on the role of meprin A, the contribution of meprin B cannot be ignored. The meprin αKO mice are in fact meprin B mice, as meprin B is still present at the membrane in the absence of the meprin α subunit. In addition, the LPS response of the meprin αβKO mice is more severe than that of the meprin

αKO mice, implying that meprin B may have a protective function independent of meprin A.

The meprins have many attributes of receptors, with adhesion and EGF-like domains.

Meprins also possess a TRAF domain, which has been implicated in protein-protein interactions among the oligomerized meprin subunits. TRAF domains are present in many proteins and mediate self-associations as well as interactions with receptors and adaptor proteins (156). The meprin β subunit is unique from the meprin α subunit by the fact that it is membrane bound and possesses a cytoplasmic tail. The cytoplasmic domain contains two potential protein kinase C phosphorylation sites and a calmodulin kinase phosphorylation site

(13,157). The cytoplasmic domain is also highly positively charged, which may allow meprin B to interact with cytoskeletal or other cytoplasmic proteins.

Litovchick et al. identified a transient interaction between the cytoplasmic protein

OS-9 and the cytoplasmic tail of meprin β upon its export from the endoplasmic reticulum

(158). Additionally, an interaction between the cytoplasmic tail of meprin β and villin has recently been identified in the mouse kidney (E.M. Ongeri, personal communication). Villin is an actin binding protein that is involved in the organization stabilization of the actin cytoskeleton (159). Cell adhesion systems, such as the tight junctions, are critical participants in the reorganization of the actin cytoskeleton during cell adhesion and migration

(160,161). Interestingly, studies have demonstrated that stabilization of the microtubule

119

network attenuates disruption of the epithelial barrier and edema in LPS-induced lung injury

(162). It is possible that in the absence of meprin A, meprin B can differentially interact with cytoskeletal proteins, enhancing stabilization. Thus, although pro-inflammatory meprin A is not able to participate in the LPS response of the αβKO mice, the absence of B could account for the more severe phenotype.

The current work identified a role for meprins in the immune response via modulation of tight junctions as well as cytokine levels that resulted in increased edema and leukocyte infiltration in WT mice. However, the leukocyte response that follows an LPS challenge is also mediated by cellular adhesion molecules (CAMs) (e.g., integrins), which participate in rearrangement of the actin cytoskeleton. CAMs are present in the plasma membrane near tight junctions and are attached to the cytoskeleton via adaptor proteins (161,163). CAMs enable efficient binding of infiltrating leukocytes to endothelial cells, and their presence is required for migration, adhesion, and activation of PMNs (116). For example, when challenged with LPS, ICAM-1-deficient mice are hyporesponsive and demonstrate decreased accumulation of PMNs in the liver (116). Although it has been established that CAMs are important for the maintenance of the epithelium, how they are recruited as well as how alterations in these processes can lead to pathologic states is unclear (161). Clearly, the possibility of an interaction between the meprins and CAMs cannot be ruled out. These data present the possibility that meprin B might participate in interactions that are independent of its proteolytic activity. This is consistent with the fact that meprin B is inactive in the normal mouse kidney.

120

5.7 Proposed role for meprins in organ-specific and systemic immune challenges

The consequences of LPS administration both systemically and within the bladder are outlined in Figure 30A. The current data indicate a potential role for meprin A at multiple steps along these pathways and allow for the development of models illustrating meprin’s roles in both bladder-specific and systemic inflammatory challenges (Figure 30B).

5.7.1 Bladder challenge

In the WT mouse bladder, cytokines and chemokines are released from urothelial cells and resident macrophages. The established chemokine gradient attracts peripheral leukocytes that migrate through the tight junctions between blood vessel cells, bladder tissue, and eventually enter the lumen of the bladder. The loosened tight junctions and leukocyte influx disrupt the barrier of the bladder, increasing permeability between the bladder contents and surrounding blood vessels and contributing to tissue edema (Figure 31A). Infiltrating leukocytes maintain the chemokine gradient and help to destroy invading bacteria. Because there is no evidence of meprin expression in bladder tissue and the urinary meprin A is in latent form, it is likely that the absence of leukocytic meprin A is responsible for the decreased bladder permeability and edema observed in the meprin αKO mice (Figure 31B). In the meprin αKO bladder, the infiltrating cells lack meprin A and cannot migrate through the bladder tissue as easily. This is consistent with observations by Sun et al., which demonstrate that monocyte numbers are decreased in the peripheral blood and increased in the bone marrow of meprin deficient mice in comparison to WT mice, indicating a decreased mobility in the KO cells (Q. Sun, personal communication). Because of decreased leukocyte infiltration, the bladder’s barrier is more intact and tissue edema is kept at a minimum. At the same time,

121

122

Figure 30. Consequences of the host response to LPS . (A) LPS binds receptors on target cells, and pro-inflammatory cytokines are released. In the bladder-specific challenge, cytokine levels are increased in the surrounding tissue, whereas in the systemic challenge cytokine levels are increased in the tissue and blood serum. This cytokine release activates resident macrophages, and additional macrophages and neutrophils are recruited from the peripheral blood. In the bladder, leukocyte migration is facilitated by loosening of the tight junctions between the cells and establishment of chemokine gradients. Although the disruption of the tight junctions is ultimately beneficial to the host by aiding in the resolution of infection, increases in tissue permeability and tissue edema can lead to pain and localized tissue damage.

In the systemic challenge, leukocyte infiltration is widespread and accumulates in organs, while increased serum cytokine levels and iNOS levels dilate the blood vessels. Accumulating leukocytes cannot be effectively circulated out of the kidney due to decreased blood pressure, and transient organ ischemia results. All of these factors contribute to the observed hypothermia and renal damage observed within hours after a systemic LPS challenge. (B) The data implicate roles for meprins along multiple steps within the host response. Cytokine array data from the systemic LPS challenge implicate meprin A in modulation of serum cytokine levels and renal failure (A and D). Data from the bladder-specific E. coli and LPS challenges indicate that meprins facilitate leukocyte migration into the bladder and are capable of disrupting the tight junctions therein (B and C).

123

124

Figure 31. Proposed role of meprin in bladder inflammation. (A) In the WT mouse bladder, a chemokine gradient attracts peripheral leukocytes that migrate through the tight junctions in order to resolve the active infection. The loosened tight junctions and leukocyte influx disrupt the barrier of the bladder, increasing permeability between the bladder contents and surrounding blood vessels and contributing to tissue edema. (B) In the meprin αKO bladder, the infiltrating cells lack meprin A and cannot migrate through the bladder tissue as easily. A decreased leukocyte infiltrate results in a more intact barrier in the bladder, which in turn minimizes the flow of toxic urine components into the bladder wall and subsequent edema. At the same time, bacterial counts are consequently elevated due to decreased leukocyte infiltration into the bladder.

125

bacterial counts are consequently elevated due to decreased leukocyte infiltration into the bladder.

5.7.2 Systemic challenge

Meprin’s role(s) in the systemic LPS challenge are complicated by the presence of both renal and leukocytic meprins. It is possible that meprin A in the kidney contributes to the elevated BUN and creatinine levels observed in the WT mice during a systemic LPS challenge, as has been demonstrated in several previously mentioned models of acute renal injury.

However, data from kidney transplant studies support the hypothesis that LPS-mediated kidney damage is controlled from inside and outside of the kidney (164,165). At 1 h post-LPS challenge, cytokine levels are comparable among the genotypes, but are markedly decreased in the meprin αKO by 2 and 3 h. It is possible that meprins are affecting the migration of cytokine-producing leukocytes. Resident tissue macrophages and mast cells release the first burst of pro-inflammatory cytokines, including TNFα. After the first hour, cytokine production by infiltrating cells and synthesis of new cytokines is necessary to perpetuate the inflammatory response. As meprins are implicated in leukocyte migration, it is reasonable to propose that a decrease in infiltrating cells is responsible for the attenuated LPS response. It is also likely that meprins are affecting the levels of cytokines directly. This is consistent with in vitro meprin data that demonstrate activation and degradation of cytokines and chemokines

(41,42,55). These scenarios are not mutually exclusive and could be occurring simultaneously, depending on the tissue and the timecourse of the disease.

The systemic LPS data with the meprin αβKO mice implicate a role for meprin B in attenuation of the LPS response, as these mice display a more severe response to LPS than

126

their WT counterparts. While meprin A decreases pro-inflammatory cytokine levels and leukocyte infiltration in the acute phases of the inflammatory response, meprin B could keep the response in check by stabilizing the junctions between cells and by modulating cytokine levels independently of meprin A. The severity of the response in the αβKO mouse is enhanced at later timepoints, implying that meprin B and/or A contribute to the return of the system to the pre-inflammatory state. As has been demonstrated in the previously mentioned disease models, the meprins likely have a role in repair of tissue and resolution after the initial insult.

5.8 Closing

In summary, this work identified a pro-inflammatory role for meprin A in the early stages of the host inflammatory response to LPS and E. coli. This study is the first to demonstrate a role for meprins in modulation of the permeability and tight junctions of the bladder epithelium, whose integrity plays an important role in the acute stages of inflammation.

The data discussed herein illustrate novel roles for the meprin metalloproteases and calls for further study of the meprin KO mice to expand our knowledge of meprins’ contributions to the initiation and resolution of inflammation. Further insight into the roles of these unique metalloproteases will increase our understanding of the diverse and important functions of proteolysis in inflammation and immunity.

127

Appendix 1: Ketamine anesthesia attenuates the inflammatory response in mice

Characterization of the meprin αKO and WT mouse response to UTI requires the use of anesthesia for insertion of the catheter. The use of either inhaled anesthetic (e.g., isofluorane) or i.p. administered ketamine/xylazine for this purpose has been reported in UTI literature (81,121,166). Inhaled isofluorane anesthesia was utilized for the first experiment, a pilot study performed by R.A. Welch and colleagues at the University of Wisconsin-

Madison. When the UTI model was implemented in our laboratory, ketamine/xylazine was used due to its ease of administration and previous use in the laboratory. The data obtained using the different anesthetics were not congruent with those of the pilot study, leading to the hypothesis that ketamine anesthesia was suppressing the inflammatory response.

Methods

Anesthesia

Ketamine/xylazine (100 mg/kg body weight ketamine; 10 mg/kg body weight xylazine) anesthesia was administered i.p. to the mice prior to intravesical E. coli and intraperitoneal LPS. To make up the working stock, 0.5 ml of 100 mg/ml ketamine and 0.25 ml of 20 mg/ml xylazine was added to 4.25 ml sterile PBS. A 0.5 cc U-100 insulin syringe

(Beckton Dickinson) with a 28-gauge needle was used to i.p. administer 10 μl ketamine/xylazine per gram body weight.

Intravesical challenge with E. coli

The protocol for E. coli bladder infusion was adapted from Saban et al. and Haugen et al. (81,92). The abdomens of anesthetized meprin αKO and WT mice were gently

128

massaged to empty residual urine. Approximately 106 (low load) or 108 (high load) E. coli

CFT073 (ATCC accession # 700928) or 20 μg LPS in 50 μl of sterile PBS was slowly

instilled into the bladder using a 0.5 mm polyethylene catheter (Intramedic PE 10) attached

to the hub of a 50 μl Hamilton #705 syringe with 30 gauge blunt-tipped needle. Mice were

sacrificed by cervical dislocation following anesthesia by isofluorane inhalation. Bladders

and kidneys of the E. coli challenged mice were placed in 500 μl or 1 ml of PBS,

respectively and homogenized in Nasco Whirl-Pak bags with the rolling pressure of a glass

bottle. Organ homogenates were serially diluted with PBS, plated on EMB agar, incubated

overnight at 37°C, and bacterial colonies were enumerated.

Intraperitoneal challenge with LPS

Meprin αKO and WT mice were challenged i.p. with 2.5 mg/kg body weight LPS (E.

coli 0111:B4, purified by gel filtration chromatography, Sigma-Aldrich) or sterile saline.

LPS was suspended in sterile PBS at a concentration of 1 mg/ml and stored in 1 ml lots at -

20°C. To make up the working stock, 1 ml of 1 mg/ml LPS was diluted 1:4 with sterile PBS.

A 0.5 cc U-100 insulin syringe (Beckton Dickinson) with a 28 gauge needle was used to administer 10 μl LPS per gram body weight.

Assessment of body temperature and kidney function

Body temperatures were monitored via a rectal thermistor (Cole-Parmer) lubricated with glycerol. Small blood samples were collected from the tail vein into lithium/heparin tubes (Sardstedt), and plasma was isolated by centrifugation at 9,500xg for 8 min at 14°C.

Terminal blood samples were collected from the renal artery into 1.5 ml microfuge tubes at

129

the time of necropsy, and serum was isolated by centrifugation in an identical manner. To assess renal function, blood plasma urea nitrogen (BUN) levels were determined by Vitros

DT6011 BUN chemistry slides (Orthoclinical Diagnostics).

Results

High E. coli burden in the bladder is associated with increased incidence of systemic infection

In 6 independent experiments, a total of 46 meprin αKO and 47 WT mice were challenged intravesically with approximately 108 uropathic E. coli. At 48 h after transurethral

inoculation with pathogenic E. coli, bladders and kidney homogenates from WT and meprin

αKO mice were cultured on EMB agar, and colonies were enumerated to determine the

severity of infection. Bladder and kidney E. coli counts were similar in meprin αKO and WT

mice (Figure 32). To determine if the E. coli spread to the blood stream, blood samples were

collected from a sampling of the infected mice and cultured on EMB plates. Of the 18 WT

blood samples tested, seven were positive for E. coli growth in comparison to seven out of the

24 meprin αKO blood samples tested (39% vs. 29%). Further, 13 E. coli-challenged WT

mice and 6 meprin αKO mice died prior to the 48 h experimental endpoint (28% vs. 13%).

Taken together, the data indicate a higher percentage of systemic infection and death in E.

coli-challenged WT mice in comparison to meprin αKO mice.

An E. coli challenge with a day 7 endpoint was performed to determine if the resolution of infection at later timepoints was different between genotypes. Bladder and

kidney bacterial counts were elevated in WT mice in contrast to meprin αKO mice, but the

difference did not reach statistical significance (Figure 33). Three of the nine E. coli-

130

Figure 32. Organ bacterial counts 48 h after high load E. coli challenge. Bacterial counts in bladders and kidneys were similar in meprin αKO and WT mice 48 h after E. coli challenge. Six independent experiments were performed, and counts from each experiment are indicated by a different color on the graph (n=46-47 per genotype). Bladder bacterial counts were higher than kidney bacterial counts, and a great deal of variation was observed in all groups.

131

Figure 33. Organ bacterial counts 7 days after high load E. coli challenge. There was a trend for higher bladder and kidney bacterial counts in WT versus meprin αKO mice 7 days after high load E. coli challenge. Bladder bacterial counts were higher than kidney bacterial counts, and a great deal of variation was observed in all groups.

132

challenged meprin αKO mice and six of the ten WT mice died. Systemic infection and body temperatures were also monitored. Body temperatures of mice with systemic infections (as indicated by positive blood culture) were markedly lower than those of mice without systemic infections (systemic, 32.1°C; non-systemic, 37.4°C; p<0.00001). There were no genotypic differences in body temperature observed within the systemic or non-systemic infection groups.

Low E. coli burden in the bladder decreases systemic and kidney infections

In an attempt to minimize the number of systemic infections and mortality associated with the high E. coli burden, a UTI experiment with a lower bacterial load (∼106 E. coli) was performed. All mice survived until the necropsy timepoint, and no positive blood cultures were detected. In contrast to the high load UTI experiments, kidney infection was low or absent in the 106 E. coli-challenged mice. Bladder bacterial counts were elevated in WT mice

in contrast to meprin αKO mice 48 h after UTI challenge, but the difference did not reach

statistical significance (Figure 34). This trend is consistent with that observed in the day 7

high burden UTI experiment (Figure 33).

Ketamine/xylazine anesthesia attenuates the severity LPS-induced renal failure

The trend in bacterial counts (increased in E. coli-challenged WT versus meprin αKO

mouse organs) was opposite of that observed in the initial experiment performed by R.A.

Welch and colleagues (Figure 35). A notable difference between these experiments was the

type and administration of anesthesia used, implying that this could account for the

inconsistent experimental results. In order to determine if ketamine/xylazine anesthesia was

133

Figure 34. Organ bacterial counts 48 h after low load E. coli challenge. There was a trend for higher bladder and kidney bacterial counts in WT versus meprin αKO mice 48 h after low load E. coli challenge. Bladder bacterial counts were higher than kidney bacterial counts, and a great deal of variation was observed in all groups.

134

Figure 35. Pilot UTI experiment utilizing isofluorane anesthesia. There was a trend for higher bladder and kidney bacterial counts in meprin αKO mice versus WT 48 h after E. coli challenge when isofluorane anesthesia was used. Bladder bacterial counts were higher than kidney bacterial counts, and a great deal of variation was observed in all groups. WT bladder counts were 16,250 ± 7,473, and meprin αKO bladder counts were 811,600 ± 485,200. WT kidney counts were 680 ± 663, and meprin αKO kidney counts were 3,767 ± 3,098. Data are expressed as mean ± SEM, with n=5 for both genotypes. This experiment was performed by the R.A. Welch laboratory.

135

capable of influencing the inflammatory response, WT mice were anesthetized with ketamine/xylazine followed by a 2.5 mg/kg i.p. LPS challenge or i.p. LPS challenge only. At

6 and 12 h, BUN levels were significantly lower in WT mice that were anesthetized with ketamine/xylazine prior to LPS challenge in comparison to LPS-challenged mice without prior anesthesia administration (Figure 36A). The trend was consistent 24 h post-LPS challenge, although statistical significance was not reached at this timepoint.

Body temperatures were monitored at 2, 6, 12, and 24 h after LPS challenge (Figure

36B). Temperatures were significantly lower in anesthetized LPS-challenged mice in comparison to unanesthetized mice at 2 h. Similarly, body temperatures were lower in anesthetized saline control mice in comparison to unanesthetized saline control mice. The mice were unconscious up to 3 h post-ketamine/xylazine anesthesia, which would account for the hypothermia at the early timepoint. At 6 and 12 h, body temperatures increased in both LPS-challenged groups but remained markedly lower in comparison to saline-treated control mice. No differences were observed between the body temperatures of anesthetized and unanesthetized LPS-challenged mice at 6, 12, or 24 h. The data indicate that administration of ketamine/xylazine anesthesia prior to an i.p. LPS challenge attenuates LPS- induced renal failure as measured by serum BUN levels.

Discussion

The data imply that WT mice are more susceptible to systemic infection and death after a high load E. coli challenge to the bladder. Although no apparent differences in bladder bacterial counts were observed in high load experiments, a trend for higher bacterial

136

Figure 36. Ketamine anesthesia attenuates LPS-induced acute renal failure. (A) BUN

levels were significantly lower in WT mice that were anesthetized with prior to LPS

challenge in comparison to LPS-challenged mice without prior anesthesia administration. *,

p=0.0005; **, p=0.02. (B) Body temperatures were significantly lower in anesthetized LPS- challenged and saline mice in comparison to unanesthetized mice at 2 h. No differences were observed between the body temperatures of anesthetized and unanesthetized LPS-challenged mice at 6, 12, or 24 h. The data are an average of 2 independent experiments; n=6-10 mice, per genotype, per timepoint.

137

counts in WT versus meprin αKO mouse bladders was observed when the E. coli challenge level was decreased. This is similar to observations from other studies, which demonstrate that phenotypic differences observed at a bacterial challenge of 106-7 are no longer present at

higher challenge levels (167). The data also indicate that decreasing the bacterial challenge

in the UTI model minimizes systemic infection and mortality observed at a higher challenge

level.

Systemic infection because of E. coli dissemination into the blood stream would

produce a host response similar to that observed in an i.p. LPS challenge model. Thus, it is

probable that the meprin αKO mice were more protected from sepsis-induced death due to

their hyporesponsiveness to a systemic LPS challenge, as discussed in Chapter 4. However,

any conclusions drawn from these data are complicated by the inclusion of ketamine/xylazine

anesthesia used in these experiments.

The data demonstrate that administration of ketamine/xylazine anesthesia prior to

administration of LPS decreases the severity of subsequent renal failure, indicating that

ketamine anesthesia suppresses inflammation. The data also depicted opposing trends in

bacterial colonization of bladder tissue in E. coli-challenged WT and meprin αKO mice

depending on the type of anesthesia utilized (Chapter 3 versus Appendix 1). The LPS data

herein are congruent with that from several other studies, which implicate ketamine

anesthesia in suppression of TNFα release, attenuation of gastric and hepatic injury, and

decreased mortality in mouse models of endotoxin shock (168-171). A decrease in LPS-

induced pro-inflammatory cytokines release (e.g., IL-6) has also been demonstrated in vitro

(172,173). In spite of the substantial body of literature demonstrating ketamine’s anti-

inflammatory effects, this anesthesia is frequently utilized in mouse models of UTI (81,121).

138

Somewhat more troubling is the fact that in several manuscripts, the type of anesthesia used in the UTI model is not mentioned at all (80,112,174,175). These experiments illustrate the importance of clearly defining all variables in a model disease system and underscore the profound effects that parameters such as anesthesia and challenge load can have on experimental outcomes.

139

Appendix 2: A comparison of blood pressures in meprin βKO and WT mice

A genome-wide microarray study comparing normotensive Wistar-Kyoto rats to three substrains of spontaneously hypertensive (SH) rats indicates a 2.4-fold decrease in meprin β mRNA expression in the kidney of SH rats in contrast to normotensive Wistar-Kyoto rats

(176). The differential meprin β expression is consistent across all substrains over multiple time periods, implicating a role for meprin in blood pressure regulation due to the hypertensive alleles shared among the SH substrains. Because SH rats have less meprin β

expression than normotensive rats, the hypothesis that meprin βKO mice would exhibit

higher blood pressure than WT mice was tested.

Methods

Female C57BL/6 meprin βKO mice backcrossed for 13 generations and corresponding WT mice were used at 7 months of age for all experiments. Blood pressure measurements were obtained via the Non-Invasive Blood Pressure Monitor NIBP-8 tail cuff apparatus, which was controlled by Windows-based application software (Columbus

Instruments). The animals were placed one at a time in a ventilated restraint chamber, and their tails were warmed to 37°C with a thermostatically controlled air heater for 10 min.

Because little evidence of tail blood flow is observed below room temperatures of 25°C, tail warming maintains continuous blood flow for accurate blood pressure measurements. An occlusion cuff was placed at the base of the mouse’s tail to occlude the tail artery (Figure 37).

A sensor cuff monitored the tail artery pulse and pressure during the inflation and deflation of the occlusion cuff. The occlusion cuff and sensor cuff pressures were set to 165 mmHg and 60 mmHg, respectively. Blood pressure readings were taken automatically every 90 s,

140

Figure 37. Diagram of blood pressure equipment set-up. The mouse restraint allows for placement of the occlusion and sensing cuff on the mouse’s tail and minimizes movement.

The cuff pressures and tail warmer are controlled via computer software. Figure adapted from Columbus Instruments user manual (www.colinst.com).

141

with 15-20 total blood pressure readings per animal. Systolic blood pressure (SBP), diastolic blood pressure (DBP), mean blood pressure (MBP), and heart rate (HR) measurements were recorded by the Columbus Instruments computer software (Figure 38). The blood pressure readings were taken at the same time of day for five consecutive days to allow the mice to acclimate to the procedure. Data from day 5 were used for statistical analysis.

Results

The blood pressure readings obtained from all animal were within normal physiological limits (Table 7). The data indicate that the SBP and MBP were significantly higher in meprin βKO mice in comparison to WT mice (Figure 39). DBP and HR were comparable between meprin βKO and WT mice.

Discussion

The data indicate increased blood pressure in meprin βKO mice in comparison to WT mice, indicating a role for the meprin β subunit in blood pressure regulation. Several lines of evidence implicate meprins in blood pressure regulation. For example, meprin α is capable of cleaving the vasodilator bradykinin. Additionally, meprin β is associated with angiotensin converting enzyme (ACE) in the proximal tubules of the kidney (59). ACE enhances vasoconstriction and is a component of the renin-angiotensin system (RAS). The RAS plays critical roles in long-term blood pressure control by regulating the amount of fluid in the blood and the width of the blood vessels (177).

The kidneys also play important roles in long-term blood pressure regulation by balancing urinary sodium and water excretion with daily intake. Epithelial Na+ channels

142

Figure 38. Principles of blood pressure measurement. As occlusion cuff pressure

increases, the sensor cuff monitors the diminishing amplitude of the tail arterial pulsation.

Once the point of blood flow blockage is reached, the corresponding occlusion cuff pressure is recorded as the SBP. The occlusion cuff pressure at the moment that the pulsation signal begins to diminish is recorded as the DBP. To determine this point, the software calculates the average signal level (line La) and the envelope slope (line Ls). The point of intersection of the two lines designates the DBP. Mean blood pressure is calculated from the measured

SBP and DBP using the following formula: MBP = 1/3 X (SBP - DBP) + DBP. Figure adapted from Columbus Instruments user manual (www.colinst.com).

143

Measurement Published values WT Meprin βKO

Heart rate (HR) 652 ± 25 754 ± 29 746 ± 11 Systolic blood pressure (SBP) 120 ± 30 128 ± 4 141 ± 4 Diastolic blood pressure (DBP) 90 ± 30 52 ± 4 61 ± 3 Mean blood pressure (MBP) 100 ± 30 77 ± 4 88 ± 3

Table 7. Mouse physiological data. The published values and the average WT and meprin

βKO values for mouse hemodynamic parameters are indicated. A typical mouse HR can range between 328 and 780. HR is measured in beats/min; SBP, DBP, and MBP are measured in mmHg; n=7-8 per genotype. Published values were obtained from Columbus

Instruments user manual (www.colinst.com).

144

Figure 39. Blood pressures in meprin βKO and WT mice. Systolic, diastolic, and mean blood pressures measurements were recorded via tail cuff. SBP and MBP were significantly higher in meprin βKO mice in comparison to WT mice, *, p=0.03; **, p=0.05. DBP and HR were comparable between genotypes. The data are an average of 2 independent experiments, with n=7-8 mice per genotype.

145

(ENaCs) help to maintain salt and volume homeostasis in the kidney, and abnormalities in the regulation of epithelial sodium channels are known to be involved in the pathogenesis of salt-sensitive hypertension (177). Recent studies have identified an interaction between

ENaC and the cytoplasmic tail of meprin β (A. Garcia-Caballero, personal communication).

Further, co-expression of ENaC and meprin β in xenopus oocytes increased amiloride- sensitive Na+ currents approximately two fold. Proteolytic cleavage, binding, or aggregation of ENaC subunits by meprin B could result in an alteration of channel properties, such as open probability. It is possible that meprin βKO are impaired in their ability to regulate properly ENaCs in the kidney, resulting in increased blood pressure.

Excessive NaCl ingestion or NaCl retention by the kidneys leads to expansion of

plasma volume (178). Data indicate that the accumulation of sodium raises cytosolic calcium concentrations via the sodium calcium exchanger (NCX) and could lead to hypertension

(179). Interestingly, NCX1 mRNA is increased 2.3 fold in the kidneys of meprin βKO mice

(180). Data indicate that transgenic mice over-expressing NCX1 in smooth muscles have increased blood pressure and were hypersensitive to salt-induced hypertension (179).

Additionally, heterozygous NCX1-deficient mice demonstrated lowered sensitivity to salt- induced hypertension (179). Taken together, the data implicate a role for the meprins in mediation of sodium channels in the kidney, thereby contributing to the maintenance of blood pressure. Monitoring the blood pressure of meprin βKO and WT mice subjected to a

high-salt diet would help to clarify meprin’s contribution to blood pressure regulation.

146

Appendix 3: The effect of estrogen on meprin β expression in breast cancer cells

A novel mRNA isoform of meprin β, meprin β’, with promoter elements that are specific to cancer cells has been identified in mouse and human cancer cell lines (62,63). In human cancer cells, meprin β’ is transcribed from a novel transcription start site, which results in a larger transcript with an altered 5’ untranslated region without altering the coding sequence. In mouse cancer cell lines, the β’ transcripts are due to the use of alternative promoter and splicing. This produces a transcript with a novel 5’ un-translated region and encodes a pro-meprin β with an altered signal peptide and pro-peptide (181).

Several lines of evidence imply that meprin β expression is hormone responsive.

Therefore, the hypothesis that meprin β is estrogen responsive and estrogen-treated cells will have higher meprin β mRNA expression was tested. Previous work identified several putative binding sites known to influence cancer cell gene expression, including activator protein-1 (AP-1) and polyoma enhancer activator 3 (PEA3) elements, in the meprin β promoter region (63). In addition, three putative half-consensus estrogen response elements

(EREs) are also present. Meprin β mRNA expression is significantly higher in the estrogen- sensitive MCF7 cells versus estrogen-independent MCF7 tamoxifen resistant (MTR) cells

(G. Matters, personal communication) (Figure 40A). Further, treatment with the glucocorticoid dexamethasone upregulates meprin β expression in primary colon and pancreatic cancer cell lines (Yura, unpublished observations) (Figure 40B).

147

Figure 40. Meprin β gene is hormone responsive. A) A comparison of meprin α and meprin β mRNA levels among MTR1, MTR3, MTR5, and MCF7 breast cancer cells demonstrated significantly higher meprin β in the MCF-7 cells versus MTR1, MTR3, and

MTR5. Meprin α mRNA levels are similar among the four cell lines (Figure courtesy of

Gail Matters). B) Meprin β mRNA is increased in dexamethasone-treated primary colon cancer SW480 cells and pancreatic cancer BxPC-3 cells. C, control; T, dexamethasone- treated; Lad, molecular size marker; GAPDH, glyceraldehyde phosphate dehydrogenase.

148

Methods

Cell culture and estrogen/anti-estrogen treatments

MCF7, MTR1, MTR3, and MTR 5 cells were obtained from M.D. Planas-Silva.

MCF7 cells were maintained in DMEM supplemented with 5% FBS and Penn/Strep and incubated with 37°C and 5% CO2. MTR1 cells were maintained in Phenol red-free DMEM

supplemented with 5% charcoal-stripped serum, 4 mM glutamine, and 1 μM Tamoxifen ((Z)-

1-(p-Dimethylaminoethoxyphenyl)-1,2-diphenyl-1-butene; TAM). To passage, cells were

rinsed two times with sterile PBS followed by a 5 min incubation with 0.25% trypsin/1 mM

EDTA.

Three days prior to the start of experiments, the media for all cell types were replaced

with Phenol red-free DMEM supplemented with 5% charcoal-stripped serum. TAM was not

included in the media unless it was the designated treatment group. Confluent MCF7 cells

were treated with 1 μM TAM (anti-estrogen), 100 nM ICI 182, 780 ((7a,17b)-7-[9-

[(4,4,5,5,5pentafluoropentyl)Sulfinyl]nonyl]estra-1,3,5(10)-triene-3,17-diol; Fulvestrant; ICI)

(anti-estrogen), 5 nM estrogen, or vehicle control (ethanol) in Phenol red-free DMEM supplemented with 5% charcoal stripped serum for 4 days. The indicated treatment was added on day 0. Tissue culture media were collected, and media were replaced on day 2.

Media were collected again on day 4. Collected media from treated and non-treated MCF7

cells and non-treated MTR cells were used for zymogram analysis. MCF7 cells were also

lysed at various timepoints to isolate RNA for real-time PCR analysis.

149

RNA isolation and real-time PCR

RNA was isolated from breast cancer cells according to the RNeasy kit protocol

(Quiagen). Briefly, cells were lysed directly in the culture dish in the presence of guanidine isothiocyanate buffer and homogenized in the QIAshredder spin column (Quiagen). Ethanol was added to the samples, which were subsequently applied to RNeasy mini columns. RNA bound to the silica-gel membrane in the column, and the column was washed to remove contaminants. RNA was eluted from the column with 30 μl of RNase-free water and the quality of the RNA was verified in the Functional Genomics Core Facility at the

Pennsylvania State University using the RNA 6000 Bioanalyzer (Agilent technologies).

The cDNA needed for real-time PCR reactions was synthesized according to the

Superscript III First Strand Synthesis protocol (Invitrogen). Briefly, 1 μg RNA was used for each cDNA reaction. Ten μl of 5X reverse transcriptase (RT) buffer, 0.5 μl of RNAase inhibitor, 2.5 μl 0.1 M DTT, 2.5 μl 10 mM dNTP mix, 1.25 μl 100 μm random hexamers, and 0.6 μl Superscript III were included in each reaction, and the total volume was brought up to 50 μl with RNAase-free water. The PCR conditions were as follows: 25°C for 5 min,

50°C for 45 min, 85°C for 5 min, and a hold cycle at 4°C. All reactions were also RNase- treated by adding 1 μl RNase H and incubating at 37°C for 20 min.

Real-time PCR was performed in an ABI 7700 sequence detector (Applied

Biosciences) using the QuantiTect SYBR-Green PCR kit. Fifty μl reactions were set up in

96 well plates and Conditions used were 50°C 2 min, 95°C 10 min (95°C 15 s, 60°C 1 min) for 40 cycles followed by a dissociation curve. Relative fold differences among the groups were determined by the standard curve method (dilutions 1:1, 1:4, 1:16, 1:64, 1:256). Data were normalized to GAPDH levels and the results are expressed as fold expression.

150

Treatments were performed in triplicate and the experiment was performed two times.

Zymography

Five μl of non-reducing SDS-PAGE sample buffer was added to 20 μl of conditioned cell culture media and incubated at 37°C for 1 h. Twenty μl of each sample was loaded into the wells of 10% zymogram gels and electrophoresed (Bio-Rad). The zymogram gels were soaked in 2.5% Triton-X 100 for 1 h at RT with gentle agitation to wash out the SDS. The gels were rinsed with three, 5 min washes of de-ionized water and incubated in 1X refolding buffer overnight at 37°C with gentle agitation (10X refolding buffer: 0.5M Tris-HCl, 2M

NaCl, 55% CaCl2, 0.67% (w/v) polyoxyethylene 23 lauryl ether (Brij35), pH 7.6). The gels

were rinsed with three, 5 min washes and stained with Coomassie Brilliant Blue (0.5% R-250

in 40% MeOH / 10% HOAc) for 1 h. The gels were destained for several hours with 40%

MeOH / 10% HOAc until the presence of clear bands could be detected visually.

Statistical Analysis

Real-time PCR data were analyzed by Dr. Vern Chinchilli using a nested analysis of variance (ANOVA). The data were transformed the data via the natural logarithm.

Therefore, the p-value is the result of testing the null hypothesis that the natural logarithm of the ratio (of comparing the effect to control) is zero. The adjusted geometric mean for the ratio can be obtained by exponentiation of the estimate. For example, estrogen at day #1, exp

(0.2858) = 1.33, the estimated geometric mean of the ratio.

151

Results

Meprin β RNA expression is increased in MCF-7 cells treated with anti-estrogen

To determine if meprin β mRNA was upregulated by estrogen treatment, MCF7 breast cancer cells were treated with estrogen, or anti-estrogens (TAM or ICI) for 4 days, and meprin β mRNA was determined at day 1-4. The data indicate that meprin β mRNA levels were significantly greater (p<0.05) for all treatment groups compared to control cells with one exception except for one (estrogen, day 2) (Table 8). When the estimated values were exponentiated to determine the fold difference over control, the data demonstrate that the differences observed on days 1-3, while statistically significant, are not dramatic (less than

1.5 fold) (Figure 41). Meprin β mRNA expression is the same or slightly higher than control for all timepoints. By day 3, meprin β mRNA expression in tamoxifen-treated MCF7 cells increases slightly and is approximately two-fold higher than control cells by day 4. The most striking difference was observed in ICI-treated MCF7 cells, which displayed 2.8-fold higher meprin β mRNA expression in comparison to control cells on day 4 (Figure 41).

Zymogram profiles differ among estrogen-dependent and estrogen-independent breast cancer cell lines

Zymograms have been used to detect metalloprotease activities and the gelatinase activity of meprins has been demonstrated previously (43). Therefore, to determine if meprin protein was secreted by breast cancer cells, conditioned media samples from estrogen and anti-estrogen treated MCF7 cells were analyzed by zymography. The zymogram profiles are equivalent among control, estrogen, TAM, and ICI-treated MCF7 cells (Figure 42). Pro-

152

Treatment Day Estimate Std Error P-value

Estrogen 1 0.2858 0.05762 p = 0.0002

Estrogen 2 0.0791 0.05762 p = 0.1901

Estrogen 3 0.3779 0.05762 p < 0.0001

Estrogen 4 0.3076 0.05762 p < 0.0001

Tamoxifen 1 0.1765 0.05117 p = 0.0036

Tamoxifen 2 0.2222 0.05117 p = 0.0006

Tamoxifen 3 0.4941 0.05117 p < 0.0001

Tamoxifen 4 0.7068 0.05117 p < 0.0001

ICI 1 0.3410 0.04170 p < 0.0001

ICI 2 0.3753 0.04170 p < 0.0001

ICI 3 0.6227 0.04170 p < 0.0001

ICI 4 1.0607 0.04170 p < 0.0001

Table 8. Summary of transformed mRNA data. The data were transformed for statistical analysis via the natural logarithm. Therefore, the p-value is the result of testing the null hypothesis that the natural logarithm of the ratio is zero; p-value is comparison of experimental group to control cells. The adjusted geometric mean for the ratio can be obtained by exponentiation of the estimate.

153

Meprin β mRNA expression in MCF-7 cells treated with estrogen or anti-estrogens

3.5 3 2.5 estrogen 2 tamoxifen 1.5 ICI 1 0.5 vehicle only control only vehicle

Fold difference over over difference Fold 0 1234 Time (days)

Figure 41. Summary of exponentiated (raw) mRNA data. The adjusted geometric means for the ratios were obtained by exponentiation of the estimate values. Meprin β mRNA levels at day 4 were higher in ICI- and tamoxifen-treated MCF7 cells in comparison to non- treated MCF7 cells. The bar graph is an average of all timecourses and PCRs.

154

Figure 42. Zymography of MCF7 and MTR1 conditioned media. A) Pro-MMP2 and

pro-MMP9 activity was observed in conditioned media from control and TAM, ICI, and

estrogen treated MCF7 cells; CSS, charcoal stripped serum control cells. B) Pro-MMP2 and

pro-MMP9 was detected in conditioned media from untreated MTR1, MTR3, MTR5, and

MCF7 cells. The active form of MMP2 was observed in MTR1 and MTR5 conditioned

media. Active MMP2 was not observed in MTR3 or MCF7 cell media. No meprin activity

was observed in any samples.

155

MMP9 and pro-MMP2 (the gelatinases) were detected in all media samples. Meprin protein expression could not be detected on the zymograms.

Conditioned media samples from non-treated estrogen-independent MTR1, MTR3, and MTR5 cells were also analyzed. Pro-MMP9 and pro-MMP2 activity was similar in

MCF7, MTR1, MTR3, and MTR5 conditioned media. In addition, active MMP2 was detected in conditioned media from MTR1 and MTR5 cells (Figure 42B).

Discussion

The data indicate that meprin β mRNA expression is increased in MCF7 cells after treatment with anti-estrogens. There is no evidence to support the original hypothesis that estrogen upregulates β mRNA expression, and mRNA levels were similar in estrogen-treated and control cells. TAM and ICI treatments are toxic to breast cancer cells. By day 4 of anti- estrogen treatment, it is likely that many of the cells were dying. Because proteases are upregulated during cell death, this could account for the increased meprin β mRNA levels observed during day 3 and 4 of anti-estrogen treatment. Meprin protein was not detected in conditioned media by zymography. This could be because the meprin β subunit is membrane bound, and soluble forms of meprin may not be present in MCF7 cells.

The zymogram data demonstrate that MMP2 is activated in MTR1 and MTR5 cells, but not MTR3 or MCF7 cells. Further, preliminary data revealed that the growth of MTR1 cells was sensitive to treatment with an MMP2 inhibitor (M.D. Planas-Silva, personal communication). It is possible that the activation of MMP2 promotes estrogen-independent growth in the MTR1 and MTR5 cell lines. Although all MTR cells are estrogen-independent and tamoxifen-resistant, each was derived separately from the MCF7 line (M.D. Planas-

156

Silva, personal communication). Thus, it is likely that MTR3 cells achieved estrogen- independence via an alternate mechanism. Further studies would help to determine whether the activation of MMP2 is dependent on the estrogen receptor and how estrogens or anti- estrogens affect MMP2 activation.

157

Reference List

1. Chang, C., and Z. Werb. 2001. The many faces of metalloproteases: cell growth, invasion, angiogenesis and metastasis. Trends Cell Biol. 11:S37-S43.

2. Parks, W. C., C. L. Wilson, and Y. S. Lopez-Boado. 2004. Matrix metalloproteinases as modulators of inflammation and innate immunity. Nat. Rev. Immunol. 4:617-629.

3. Rawlings, N. D., F. R. Morton, C. Y. Kok, J. Kong, and A. J. Barrett. 2008. MEROPS: the peptidase database. Nucleic Acids Res. 36:D320-D325.

4. Stocker, W., and W. Bode. 1995. Structural features of a superfamily of zinc- endopeptidases: the metzincins. Curr. Opin. Struct. Biol. 5:383-390.

5. Bode, W., F. Grams, P. Reinemer, F. X. Gomis-Ruth, U. Baumann, D. B. McKay, and W. Stocker. 1996. The metzincin-superfamily of zinc-peptidases. Adv. Exp. Med. Biol. 389:1-11.

6. McCawley, L. J., and L. M. Matrisian. 2001. Matrix metalloproteinases: they're not just for matrix anymore! Curr. Opin. Cell Biol. 13:534-540.

7. Tsakadze, N. L., S. D. Sithu, U. Sen, W. R. English, G. Murphy, and S. E. D'Souza. 2006. Tumor necrosis factor-alpha-converting enzyme (TACE/ADAM-17) mediates the ectodomain cleavage of intercellular adhesion molecule-1 (ICAM-1). J. Biol. Chem. 281:3157-3164.

8. Renckens, R., J. J. Roelofs, S. Florquin, A. F. de Vos, H. R. Lijnen, V. C. van't, and P. T. van der. 2006. Matrix metalloproteinase-9 deficiency impairs host defense against abdominal sepsis. J. Immunol. 176:3735-3741.

9. Bullard, K. M., L. Lund, J. S. Mudgett, T. N. Mellin, T. K. Hunt, B. Murphy, J. Ronan, Z. Werb, and M. J. Banda. 1999. Impaired wound contraction in stromelysin- 1-deficient mice. Ann. Surg. 230:260-265.

10. Charrier-Hisamuddin, L., C. L. Laboisse, and D. Merlin. 2008. ADAM-15: a metalloprotease that mediates inflammation. FASEB J. 22:641-653.

11. Becherer, J. D., and C. P. Blobel. 2003. Biochemical properties and functions of membrane-anchored metalloprotease-disintegrin proteins (ADAMs). Curr. Top. Dev. Biol. 54:101-123.

12. Bond, J. S., G. L. Matters, S. Banerjee, and R. E. Dusheck. 2005. Meprin metalloprotease expression and regulation in kidney, intestine, urinary tract infections and cancer. FEBS Lett. 579:3317-3322.

13. Bond, J. S., and R. J. Beynon. 1995. The Astacin Family of Metalloendopeptidases. Protein Science 4:1247-1261.

158

14. Villa, J. P., G. P. Bertenshaw, J. E. Bylander, and J. S. Bond. 2003. Meprin proteolytic complexes at the cell surface and in extracellular spaces. Biochem. Soc. Symp.53-63.

15. Tsukuba, T., and J. S. Bond. 1998. Role of the COOH-terminal domains of meprin A in folding, secretion, and activity of the metalloendopeptidase. J. Biol. Chem. 273:35260-35267.

16. Marchand, P., J. Tang, G. D. Johnson, and J. S. Bond. 1995. COOH-terminal proteolytic processing of secreted and membrane forms of the alpha subunit of the metalloprotease meprin A. Requirement of the I domain for processing in the endoplasmic reticulum. J. Biol. Chem. 270:5449-5456.

17. Bertenshaw, G. P., M. T. Norcum, and J. S. Bond. 2003. Structure of homo- and hetero-oligomeric meprin metalloproteases. Dimers, tetramers, and high molecular mass multimers. J. Biol. Chem. 278:2522-2532.

18. Ishmael, F. T., M. T. Norcum, S. J. Benkovic, and J. S. Bond. 2001. Multimeric structure of the secreted meprin A metalloproteinase and characterization of the functional protomer. J. Biol. Chem. 276:23207-23211.

19. Eldering, J. A., J. Grunberg, D. Hahn, H. J. Croes, J. A. Fransen, and E. E. Sterchi. 1997. Polarised expression of human intestinal N-benzoyl-L-tyrosyl-p-aminobenzoic acid hydrolase (human meprin) alpha and beta subunits in Madin-Darby canine kidney cells. Eur. J. Biochem. 247:920-932.

20. Norman, L. P., W. Jiang, X. Han, T. L. Saunders, and J. S. Bond. 2003. Targeted disruption of the meprin beta gene in mice leads to underrepresentation of knockout mice and changes in renal gene expression profiles. Mol. Cell Biol. 23:1221-1230.

21. Tang, J., and J. S. Bond. 1998. Maturation of secreted meprin alpha during biosynthesis: role of the furin site and identification of the COOH-terminal amino acids of the mouse kidney metalloprotease subunit. Arch. Biochem. Biophys. 349:192- 200.

22. Ishmael, S. S., F. T. Ishmael, A. D. Jones, and J. S. Bond. 2006. Protease domain glycans affect oligomerization, disulfide bond formation, and stability of the meprin A metalloprotease homo-oligomer. J. Biol. Chem. 281:37404-37415.

23. Jiang, W., C. M. Gorbea, A. V. Flannery, R. J. Beynon, G. A. Grant, and J. S. Bond. 1992. The alpha subunit of meprin A. Molecular cloning and sequencing, differential expression in inbred mouse strains, and evidence for divergent evolution of the alpha and beta subunits. J. Biol. Chem. 267:9185-9193.

24. Jiang, W. P., P. M. Sadler, N. A. Jenkins, D. J. Gilbert, N. G. Copeland, and J. S. Bond. 1993. Tissue-Specific Expression and Chromosomal Localization of the Alpha-Subunit of Mouse Meprin-A. Journal of Biological Chemistry 268:10380- 10385.

159

25. Gorbea, C. M., P. Marchand, W. P. Jiang, N. G. Copeland, D. J. Gilbert, N. A. Jenkins, and J. S. Bond. 1993. Cloning, Expression, and Chromosomal Localization of the Mouse Meprin Beta-Subunit. Journal of Biological Chemistry 268:21035- 21043.

26. Lottaz, D., D. Hahn, S. Muller, C. Muller, and E. E. Sterchi. 1999. Secretion of human meprin from intestinal epithelial cells depends on differential expression of the alpha and beta subunits. Eur. J. Biochem. 259:496-504.

27. Lottaz, D., C. A. Maurer, D. Hahn, M. W. Buchler, and E. E. Sterchi. 1999. Nonpolarized secretion of human meprin alpha in colorectal cancer generates an increased proteolytic potential in the stroma. Cancer Res. 59:1127-1133.

28. Crisman, J. M., B. Zhang, L. P. Norman, and J. S. Bond. 2004. Deletion of the mouse meprin beta metalloprotease gene diminishes the ability of leukocytes to disseminate through extracellular matrix. J. Immunol. 172:4510-4519.

29. Beynon, R. J., and J. S. Bond. 1983. Deficiency of a kidney metalloproteinase activity in inbred mouse strains. Science 219:1351-1353.

30. Bond, J. S., J. D. Shannon, and R. J. Beynon. 1983. Certain mouse strains are deficient in a kidney brush-border metallo-endopeptidase activity. Biochem. J. 209:251-255.

31. Kumar, J. M., and J. S. Bond. 2001. Developmental expression of meprin metalloprotease subunits in ICR and C3H/He mouse kidney and intestine in the embryo, postnatally and after weaning. Biochim. Biophys. Acta 1518:106-114.

32. Craig, S. S., C. Mader, and J. S. Bond. 1991. Immunohistochemical localization of the metalloproteinase meprin in salivary glands of male and female mice. J. Histochem. Cytochem. 39:123-129.

33. Stuart, R. O., K. T. Bush, and S. K. Nigam. 2001. Changes in global gene expression patterns during development and maturation of the rat kidney. Proc. Natl. Acad. Sci. U. S. A 98:5649-5654.

34. Spencer-Dene, B., P. Thorogood, S. Nair, A. J. Kenny, M. Harris, and B. Henderson. 1994. Distribution of, and a putative role for, the cell-surface neutral metalloendopeptidases during mammalian craniofacial development. Development 120:3213- 3226.

35. Becker-Pauly, C., M. Howel, T. Walker, A. Vlad, K. Aufenvenne, V. Oji, D. Lottaz, E. E. Sterchi, M. Debela, V. Magdolen, H. Traupe, and W. Stocker. 2007. The alpha and beta subunits of the metalloprotease meprin are expressed in separate layers of human epidermis, revealing different functions in keratinocyte proliferation and differentiation. J. Invest Dermatol. 127:1115-1125.

36. Kruse, M. N., C. Becker, D. Lottaz, D. Kohler, I. Yiallouros, H. W. Krell, E. E.

160

Sterchi, and W. Stocker. 2004. Human meprin alpha and beta homo-oligomers: cleavage of basement membrane proteins and sensitivity to metalloprotease inhibitors. Biochem. J. 378:383-389.

37. Bertenshaw, G. P., B. E. Turk, S. J. Hubbard, G. L. Matters, J. E. Bylander, J. M. Crisman, L. C. Cantley, and J. S. Bond. 2001. Marked differences between metalloproteases meprin A and B in substrate and peptide bond specificity. J. Biol. Chem. 276:13248-13255.

38. Kenny, A. J., and J. Ingram. 1987. Proteins of the kidney microvillar membrane. Purification and properties of the phosphoramidon-insensitive endopeptidase ('endopeptidase-2') from rat kidney. Biochem. J. 245:515-524.

39. Butler, P. E., M. J. McKay, and J. S. Bond. 1987. Characterization of meprin, a membrane-bound metalloendopeptidase from mouse kidney. Biochem. J. 241:229- 235.

40. Yamaguchi, T., H. Kido, and N. Katunuma. 1992. A membrane-bound metalloendopeptidase from rat kidney. Characteristics of its hydrolysis of peptide hormones and neuropeptides. Eur. J. Biochem. 204:547-552.

41. Norman, L. P., G. L. Matters, J. M. Crisman, and J. S. Bond. 2003. Expression of meprins in health and disease. Curr. Top. Dev. Biol. 54:145-166.

42. Herzog, C., G. P. Kaushal, and R. S. Haun. 2005. Generation of biologically active interleukin-1beta by meprin B. Cytokine 31:394-403.

43. Becker, C., M. N. Kruse, K. A. Slotty, D. Kohler, J. R. Harris, S. Rosmann, E. E. Sterchi, and W. Stocker. 2003. Differences in the activation mechanism between the alpha and beta subunits of human meprin. Biol. Chem. 384:825-831.

44. Rosmann, S., D. Hahn, D. Lottaz, M. N. Kruse, W. Stocker, and E. E. Sterchi. 2002. Activation of human meprin-alpha in a cell culture model of colorectal cancer is triggered by the plasminogen-activating system. J. Biol. Chem. 277:40650-40658.

45. Grunberg, J., E. Dumermuth, J. A. Eldering, and E. E. Sterchi. 1993. Expression of the alpha subunit of PABA peptide hydrolase (EC 3.4.24.18) in MDCK cells. Synthesis and secretion of an enzymatically inactive homodimer. FEBS Lett. 335:376-379.

46. Wolz, R. L., and J. S. Bond. 1995. Meprins A and B. Methods Enzymol. 248:325-345.

47. Bylander, J. E., G. P. Bertenshaw, G. L. Matters, S. J. Hubbard, and J. S. Bond. 2007. Human and mouse homo-oligomeric meprin A metalloendopeptidase: substrate and inhibitor specificities. Biol. Chem. 388:1163-1172.

48. Hirano, M., B. Y. Ma, N. Kawasaki, K. Okimura, M. Baba, T. Nakagawa, K. Miwa, N. Kawasaki, S. Oka, and T. Kawasaki. 2005. Mannan-binding protein blocks the

161

activation of metalloproteases meprin alpha and beta. Journal of Immunology 175:3177-3185.

49. Ajuebor, M. N., A. M. Das, L. Virag, R. J. Flower, C. Szabo, and M. Perretti. 1999. Role of resident peritoneal macrophages and mast cells in chemokine production and neutrophil migration in acute inflammation: evidence for an inhibitory loop involving endogenous IL-10. J. Immunol. 162:1685-1691.

50. Trachtman, H., E. Valderrama, J. M. Dietrich, and J. S. Bond. 1995. The role of meprin A in the pathogenesis of acute renal failure. Biochem. Biophys. Res. Commun. 208:498-505.

51. Bylander, J., Q. Li, G. Ramesh, B. Zhang, W. B. Reeves, and J. S. Bond. 2008. Targed disruption of the meprin metalloproteinase beta gene protects against renal ischemia/reperfusion injury in mice. Am. J. Physiol Renal Physiol 294:F480-F490.

52. Red Eagle, A. R., R. L. Hanson, W. Jiang, X. Han, G. L. Matters, G. Imperatore, W. C. Knowler, and J. S. Bond. 2005. Meprin beta metalloprotease gene polymorphisms associated with diabetic nephropathy in the Pima Indians. Hum. Genet. 118:12-22.

53. Ricardo, S. D., J. S. Bond, G. D. Johnson, J. Kaspar, and J. R. Diamond. 1996. Expression of subunits of the metalloendopeptidase meprin in renal cortex in experimental hydronephrosis. Am. J. Physiol 270:F669-F676.

54. Sampson, N. S., S. T. Ryan, D. A. Enke, D. Cosgrove, V. Koteliansky, and P. Gotwals. 2001. Global gene expression analysis reveals a role for the alpha 1 integrin in renal pathogenesis. J. Biol. Chem. 276:34182-34188.

55. Banerjee, S. 2008. Dissertation: Meprin metalloproteases modulate the intestinal host response.

56. Carmago, S., S. V. Shah, and P. D. Walker. 2002. Meprin, a brush-border enzyme, plays an important role in hypoxic/ischemic acute renal tubular injury in rats. Kidney Int. 61:959-966.

57. Kieran, N. E., P. P. Doran, S. B. Connolly, M. C. Greenan, D. F. Higgins, M. Leonard, C. Godson, C. T. Taylor, A. Henger, M. Kretzler, M. J. Burne, H. Rabb, and H. R. Brady. 2003. Modification of the transcriptomic response to renal ischemia/reperfusion injury by lipoxin analog. Kidney Int. 64:480-492.

58. Herzog, C., R. Seth, S. V. Shah, and G. P. Kaushal. 2007. Role of meprin A in renal tubular epithelial cell injury. Kidney Int. 71:1009-1018.

59. Mathew, R., S. Futterweit, E. Valderrama, A. A. Tarectecan, J. E. Bylander, J. S. Bond, and H. Trachtman. 2005. Meprin-alpha in chronic diabetic nephropathy: interaction with the renin-angiotensin axis. Am. J. Physiol Renal Physiol 289:F911- F921.

162

60. Sadlier, D. M., S. B. Connolly, N. E. Kieran, S. Roxburgh, D. P. Brazil, L. Kairaitis, Y. Wang, D. C. Harris, P. Doran, and H. R. Brady. 2004. Sequential extracellular matrix-focused and baited-global cluster analysis of serial transcriptomic profiles identifies candidate modulators of renal tubulointerstitial fibrosis in murine adriamycin-induced nephropathy. J. Biol. Chem. 279:29670-29680.

61. Trachtman, H., R. Greenwald, S. Moak, J. Tang, and J. S. Bond. 1993. Meprin activity in rats with experimental renal disease. Life Sci. 53:1339-1344.

62. Matters, G. L., and J. S. Bond. 1999. Expression and regulation of the meprin beta gene in human cancer cells. Mol. Carcinog. 25:169-178.

63. Matters, G. L., and J. S. Bond. 1999. Meprin B: transcriptional and posttranscriptional regulation of the meprin beta metalloproteinase subunit in human and mouse cancer cells. APMIS 107:19-27.

64. Matters, G. L., A. Manni, and J. S. Bond. 2005. Inhibitors of polyamine biosynthesis decrease the expression of the metalloproteases meprin alpha and MMP-7 in hormone-independent human breast cancer cells. Clin. Exp. Metastasis 22:331-339.

65. Jiang, W., J. M. Kumar, G. L. Matters, and J. S. Bond. 2000. Structure of the mouse metalloprotease meprin beta gene (Mep1b): alternative splicing in cancer cells. Gene 248:77-87.

66. Finlay, B. B., and S. Falkow. 1989. Common Themes in Microbial Pathogenicity. Microbiological Reviews 53:210-230.

67. Mulvey, M. A. 2002. Adhesion and entry of uropathogenic Escherichia coli. Cell Microbiol. 4:257-271.

68. Hooton, T. M. 2001. Recurrent urinary tract infection in women. Int. J. Antimicrob. Agents 17:259-268.

69. Finer, G., and D. Landau. 2004. Pathogenesis of urinary tract infections with normal female anatomy. Lancet Infect. Dis. 4:631-635.

70. Schaeffer, A. J., J. M. Jones, and J. K. Dunn. 1981. Association of vitro Escherichia coli adherence to vaginal and buccal epithelial cells with susceptibility of women to recurrent urinary-tract infections. N. Engl. J. Med. 304:1062-1066.

71. Chowdhury, P., S. H. Sacks, and N. S. Sheerin. 2004. Minireview: functions of the renal tract epithelium in coordinating the innate immune response to infection. Kidney Int. 66:1334-1344.

72. Geerlings, S. E., R. Meiland, E. C. van Lith, E. C. Brouwer, W. Gaastra, and A. I. Hoepelman. 2002. Adherence of type 1-fimbriated Escherichia coli to uroepithelial cells: more in diabetic women than in control subjects. Diabetes Care 25:1405-1409.

163

73. Mulvey, M. A., J. D. Schilling, J. J. Martinez, and S. J. Hultgren. 2000. Bad bugs and beleaguered bladders: interplay between uropathogenic Escherichia coli and innate host defenses. Proc. Natl. Acad. Sci. U. S. A 97:8829-8835.

74. Pak, J., Y. B. Pu, Z. T. Zhang, D. L. Hasty, and X. R. Wu. 2001. Tamm-Horsfall protein binds to type 1 fimbriated Escherichia coli and prevents E. coli from binding to uroplakin Ia and Ib receptors. Journal of Biological Chemistry 276:9924-9930.

75. Leach, J. L., S. A. Garber, A. A. Marcon, and P. A. Prieto. 2005. In vitro and in vivo effects of soluble, monovalent globotriose on bacterial attachment and colonization. Antimicrob. Agents Chemother. 49:3842-3846.

76. Russo, T. A., U. B. Carlino, A. Mong, and S. T. Jodush. 1999. Identification of genes in an extraintestinal isolate of Escherichia coli with increased expression after exposure to human urine. Infect. Immun. 67:5306-5314.

77. Schwan, W. R., J. L. Lee, F. A. Lenard, B. T. Matthews, and M. T. Beck. 2002. Osmolarity and pH growth conditions regulate fim gene transcription and type 1 pilus expression in uropathogenic Escherichia coli. Infect. Immun. 70:1391-1402.

78. Selsted, M. E., and A. J. Ouellette. 2005. Mammalian defensins in the antimicrobial immune response. Nat. Immunol. 6:551-557.

79. Lewis, S. A. 2000. Everything you wanted to know about the bladder epithelium but were afraid to ask. Am. J. Physiol Renal Physiol 278:F867-F874.

80. Haraoka, M., L. Hang, B. Frendeus, G. Godaly, M. Burdick, R. Strieter, and C. Svanborg. 1999. Neutrophil recruitment and resistance to urinary tract infection. Journal of Infectious Diseases 180:1220-1229.

81. Saban, M. R., H. Hellmich, N. B. Nguyen, J. Winston, T. G. Hammond, and R. Saban. 2001. Time course of LPS-induced gene expression in a mouse model of genitourinary inflammation. Physiol Genomics 5:147-160.

82. Beutler, B., and E. T. Rietschel. 2003. Innate immune sensing and its roots: the story of endotoxin. Nat. Rev. Immunol. 3:169-176.

83. Daubeuf, B., J. Mathison, S. Spiller, S. Hugues, S. Herren, W. Ferlin, M. Kosco- Vilbois, H. Wagner, C. J. Kirschning, R. Ulevitch, and G. Elson. 2007. TLR4/MD-2 monoclonal antibody therapy affords protection in experimental models of septic shock. Journal of Immunology 179:6107-6114.

84. Zhang, C., L. M. Walker, and P. R. Mayeux. 2000. Role of nitric oxide in lipopolysaccharide-induced oxidant stress in the rat kidney. Biochem. Pharmacol. 59:203-209.

85. Van Amersfoort, E. S., T. J. Van Berkel, and J. Kuiper. 2003. Receptors, mediators, and mechanisms involved in bacterial sepsis and septic shock. Clin. Microbiol. Rev.

164

16:379-414.

86. Glauser, M. P., G. Zanetti, J. D. Baumgartner, and J. Cohen. 1991. Septic shock: pathogenesis. Lancet 338:732-736.

87. Calandra, T., J. D. Baumgartner, G. E. Grau, M. M. Wu, P. H. Lambert, J. Schellekens, J. Verhoef, and M. P. Glauser. 1990. Prognostic values of tumor necrosis factor/cachectin, interleukin-1, interferon-alpha, and interferon-gamma in the serum of patients with septic shock. Swiss-Dutch J5 Immunoglobulin Study Group. J. Infect. Dis. 161:982-987.

88. Hopkins, W. J., J. E. Elkahwaji, D. M. Heisey, and C. J. Ott. 2003. Inheritance of susceptibility to induced Escherichia coli bladder and kidney infections in female C3H/HeJ mice. J. Infect. Dis. 187:418-423.

89. Bond, J. S., R. J. Beynon, J. F. Reckelhoff, and C. S. David. 1984. Mep-1 gene controlling a kidney metalloendopeptidase is linked to the major histocompatibility complex in mice. Proc. Natl. Acad. Sci. U. S. A 81:5542-5545.

90. Bond, J. S., K. Rojas, J. Overhauser, H. Y. Zoghbi, and W. Jiang. 1995. The structural genes, MEP1A and MEP1B, for the alpha and beta subunits of the metalloendopeptidase meprin map to human chromosomes 6p and 18q, respectively. Genomics 25:300-303.

91. Hopkins, W., A. Gendron-Fitzpatrick, D. O. McCarthy, J. E. Haine, and D. T. Uehling. 1996. Lipopolysaccharide-responder and nonresponder C3H mouse strains are equally susceptible to an induced Escherichia coli urinary tract infection. Infect. Immun. 64:1369-1372.

92. Haugen, B. J., S. Pellett, P. Redford, H. L. Hamilton, P. L. Roesch, and R. A. Welch. 2007. In vivo gene expression analysis identifies genes required for enhanced colonization of the mouse urinary tract by uropathogenic Escherichia coli strain CFT073 dsdA. Infect. Immun. 75:278-289.

93. Eichel, L., K. Scheidweiler, J. Kost, J. Shojaie, E. Schwarz, E. Messing, and R. Wood. 2001. Assessment of murine bladder permeability with fluorescein: validation with cyclophosphamide and protamine. Urology 58:113-118.

94. Martinez, J. J., M. A. Mulvey, J. D. Schilling, J. S. Pinkner, and S. J. Hultgren. 2000. Type 1 pilus-mediated bacterial invasion of bladder epithelial cells. EMBO J. 19:2803-2812.

95. Harhaj, N. S., A. J. Barber, and D. A. Antonetti. 2002. Platelet-derived growth factor mediates tight junction redistribution and increases permeability in MDCK cells. J. Cell Physiol 193:349-364.

96. Bertenshaw, G. P., J. P. Villa, J. A. Hengst, and J. S. Bond. 2002. Probing the active sites and mechanisms of rat metalloproteases meprin A and B. Biol. Chem. 383:1175-

165

1183.

97. Villa, J. P., G. P. Bertenshaw, and J. S. Bond. 2003. Critical amino acids in the active site of meprin metalloproteinases for substrate and peptide bond specificity. J. Biol. Chem. 278:42545-42550.

98. Beynon, R. J., S. Oliver, and D. H. L. Robertson. 1996. Characterization of the soluble, secreted form of urinary meprin. Biochemical Journal 315:461-465.

99. Lejeune, B., R. Baron, B. Guillois, and D. Mayeux. 1991. Evaluation of a screening test for detecting urinary tract infection in newborns and infants. J. Clin. Pathol. 44:1029-1030.

100. Hagberg, L., I. Engberg, R. Freter, J. Lam, S. Olling, and E. C. Svanborg. 1983. Ascending, unobstructed urinary tract infection in mice caused by pyelonephritogenic Escherichia coli of human origin. Infect. Immun. 40:273-283.

101. Hiratsuka, T., M. Nakazato, T. Ihi, T. Minematsu, N. Chino, T. Nakanishi, A. Shimizu, K. Kangawa, and S. Matsukura. 2000. Structural analysis of human beta- defensin-1 and its significance in urinary tract infection. Nephron 85:34-40.

102. Reinhart, H., N. Obedeanu, T. Hooton, W. Stamm, and J. Sobel. 1990. Urinary excretion of Tamm-Horsfall protein in women with recurrent urinary tract infections. J. Urol. 144:1185-1187.

103. Franz, M., and W. H. Horl. 1999. Common errors in diagnosis and management of urinary tract infection. I: pathophysiology and diagnostic techniques. Nephrol. Dial. Transplant. 14:2746-2753.

104. Lavelle, J. P., G. Apodaca, S. A. Meyers, W. G. Ruiz, and M. L. Zeidel. 1998. Disruption of guinea pig urinary bladder permeability barrier in noninfectious cystitis. Am. J. Physiol 274:F205-F214.

105. Hocking, D. C., P. G. Phillips, T. J. Ferro, and A. Johnson. 1990. Mechanisms of pulmonary edema induced by tumor necrosis factor-alpha. Circ. Res. 67:68-77.

106. Mittal, R., S. Chhibber, S. Sharma, and K. Harjai. 2004. Macrophage inflammatory protein-2, neutrophil recruitment and bacterial persistence in an experimental mouse model of urinary tract infection. Microbes. Infect. 6:1326-1332.

107. Ferry, S. A., S. E. Holm, H. Stenlund, R. Lundholm, and T. J. Monsen. 2004. The natural course of uncomplicated lower urinary tract infection in women illustrated by a randomized placebo controlled study. Scand. J. Infect. Dis. 36:296-301.

108. Kanellopoulos, T. A., P. J. Vassilakos, M. Kantzis, A. Ellina, F. Kolonitsiou, and D. A. Papanastasiou. 2005. Low bacterial count urinary tract infections in infants and young children. Eur. J. Pediatr. 164:355-361.

166

109. Hagberg, L., R. Hull, S. Hull, S. Falkow, R. Freter, and E. C. Svanborg. 1983. Contribution of adhesion to bacterial persistence in the mouse urinary tract. Infect. Immun. 40:265-272.

110. Schilling, J. D., M. A. Mulvey, C. D. Vincent, R. G. Lorenz, and S. J. Hultgren. 2001. Bacterial invasion augments epithelial cytokine responses to Escherichia coli through a lipopolysaccharide-dependent mechanism. J. Immunol. 166:1148-1155.

111. Linder, H., I. Engberg, I. Mattsbybaltzer, and C. S. Eden. 1988. Natural-Resistance to Urinary-Tract Infection Determined by Endotoxin-Induced Inflammation. Fems Microbiology Letters 49:219-222.

112. Shahin, R. D., I. Engberg, L. Hagberg, and E. C. Svanborg. 1987. Neutrophil recruitment and bacterial clearance correlated with LPS responsiveness in local gram- negative infection. J. Immunol. 138:3475-3480.

113. Shahin, R., L. Hagberg, I. Engberg, and C. Svanborg-Eden. 1987. Role of LPS responsiveness in urinary tract infection. Adv. Exp. Med. Biol. 216A:685-689.

114. Jonassen, T. E., M. Graebe, D. Promeneur, S. Nielsen, S. Christensen, and N. V. Olsen. 2002. Lipopolysaccharide-induced acute renal failure in conscious rats: effects of specific phosphodiesterase type 3 and 4 inhibition. J. Pharmacol. Exp. Ther. 303:364-374.

115. Leon, L. R. 2002. Invited review: cytokine regulation of fever: studies using gene knockout mice. J. Appl. Physiol. 92:2648-2655.

116. Lowell, C. A., and G. Berton. 1998. Resistance to endotoxic shock and reduced neutrophil migration in mice deficient for the Src-family kinases Hck and Fgr. Proc. Natl. Acad. Sci. U. S. A 95:7580-7584.

117. Blanque, R., C. Meakin, S. Millet, and C. R. Gardner. 1996. Hypothermia as an indicator of the acute effects of lipopolysaccharides: comparison with serum levels of IL1 beta, IL6 and TNF alpha. Gen. Pharmacol. 27:973-977.

118. Ishiguro, K., K. Kadomatsu, T. Kojima, H. Muramatsu, S. Matsuo, K. Kusugami, H. Saito, and T. Muramatsu. 2001. Syndecan-4 deficiency increases susceptibility to kappa-carrageenan-induced renal damage. Lab Invest 81:509-516.

119. Ramesh, G., S. R. Kimball, L. S. Jefferson, and W. B. Reeves. 2007. Endotoxin and cisplatin synergistically stimulate TNF-alpha production by renal epithelial cells. Am. J. Physiol Renal Physiol 292:F812-F819.

120. Warren, J. W. 1994. Interstitial cystitis as an infectious disease. Urol. Clin. North Am. 21:31-39.

121. Olsson, L. E., M. A. Wheeler, W. C. Sessa, and R. M. Weiss. 1998. Bladder instillation and intraperitoneal injection of Escherichia coli lipopolysaccharide up-

167

regulate cytokines and iNOS in rat urinary bladder. J. Pharmacol. Exp. Ther. 284:1203-1208.

122. Knotek, M., B. Rogachev, W. Wang, T. Ecder, V. Melnikov, P. E. Gengaro, M. Esson, C. L. Edelstein, C. A. Dinarello, and R. W. Schrier. 2001. Endotoxemic renal failure in mice: role of tumor necrosis factor independent of inducible nitric oxide synthase. Kidney Int. 59:2243-2249.

123. Keepers, T. R., L. K. Gross, and T. G. Obrig. 2007. Monocyte chemoattractant protein 1, macrophage inflammatory protein 1 alpha, and RANTES recruit macrophages to the kidney in a mouse model of hemolytic-uremic syndrome. Infect. Immun. 75:1229-1236.

124. Minano, F. J., A. Fernandez-Alonso, R. D. Myers, and M. Sancibrian. 1996. Hypothalamic interaction between macrophage inflammatory protein-1 alpha (MIP-1 alpha) and MIP-1 beta in rats: a new level for fever control? J. Physiol 491 ( Pt 1):209-217.

125. Minano, F. J., M. Sancibrian, and R. D. Myers. 1991. Fever induced by macrophage inflammatory protein-1 (MIP-1) in rats: hypothalamic sites of action. Brain Res. Bull. 27:701-706.

126. Kielian, T., E. D. Bearden, A. C. Baldwin, and N. Esen. 2004. IL-1 and TNF-alpha play a pivotal role in the host immune response in a mouse model of Staphylococcus aureus-induced experimental brain abscess. J. Neuropathol. Exp. Neurol. 63:381-396.

127. Kuemmerle, J. F. 1998. Synergistic regulation of NOS II expression by IL-1 beta and TNF-alpha in cultured rat colonic smooth muscle cells. Am. J. Physiol 274:G178- G185.

128. Brandonisio, O., M. A. Panaro, I. Fumarola, M. Sisto, D. Leogrande, A. Acquafredda, R. Spinelli, and V. Mitolo. 2002. Macrophage chemotactic protein-1 and macrophage inflammatory protein-1 alpha induce nitric oxide release and enhance parasite killing in Leishmania infantum-infected human macrophages. Clin. Exp. Med. 2:125-129.

129. Armaka, M., M. Apostolaki, P. Jacques, D. L. Kontoyiannis, D. Elewaut, and G. Kollias. 2008. Mesenchymal cell targeting by TNF as a common pathogenic principle in chronic inflammatory joint and intestinal diseases. J. Exp. Med. 205:331-337.

130. Aderka, D., P. Sorkine, S. Abu-Abid, D. Lev, A. Setton, A. P. Cope, D. Wallach, and J. Klausner. 1998. Shedding kinetics of soluble tumor necrosis factor (TNF) receptors after systemic TNF leaking during isolated limb perfusion. Relevance to the pathophysiology of septic shock. J. Clin. Invest 101:650-659.

131. Berg, D. J., R. Kuhn, K. Rajewsky, W. Muller, S. Menon, N. Davidson, G. Grunig, and D. Rennick. 1995. Interleukin-10 is a central regulator of the response to LPS in murine models of endotoxic shock and the Shwartzman reaction but not endotoxin tolerance. J. Clin. Invest 96:2339-2347.

168

132. Greenlee, K. J., D. B. Corry, D. A. Engler, R. K. Matsunami, P. Tessier, R. G. Cook, Z. Werb, and F. Kheradmand. 2006. Proteomic identification of in vivo substrates for matrix metalloproteinases 2 and 9 reveals a mechanism for resolution of inflammation. J. Immunol. 177:7312-7321.

133. Turler, A., N. T. Schwarz, E. Turler, J. C. Kalff, and A. J. Bauer. 2002. MCP-1 causes leukocyte recruitment and subsequently endotoxemic ileus in rat. Am. J. Physiol Gastrointest. Liver Physiol 282:G145-G155.

134. Fuentes, M. E., S. K. Durham, M. R. Swerdel, A. C. Lewin, D. S. Barton, J. R. Megill, R. Bravo, and S. A. Lira. 1995. Controlled recruitment of monocytes and macrophages to specific organs through transgenic expression of monocyte chemoattractant protein-1. J. Immunol. 155:5769-5776.

135. Weathington, N. M., A. H. van Houwelingen, B. D. Noerager, P. L. Jackson, A. D. Kraneveld, F. S. Galin, G. Folkerts, F. P. Nijkamp, and J. E. Blalock. 2006. A novel peptide CXCR ligand derived from extracellular matrix degradation during airway inflammation. Nat. Med. 12:317-323.

136. Lewis, S. A., J. R. Berg, and T. J. Kleine. 1995. Modulation of epithelial permeability by extracellular macromolecules. Physiol Rev. 75:561-589.

137. Bannerman, D. D., M. Sathyamoorthy, and S. E. Goldblum. 1998. Bacterial lipopolysaccharide disrupts endothelial monolayer integrity and survival signaling events through caspase cleavage of adherens junction proteins. J. Biol. Chem. 273:35371-35380.

138. Bojarski, C., J. Weiske, T. Schoneberg, W. Schroder, J. Mankertz, J. D. Schulzke, P. Florian, M. Fromm, R. Tauber, and O. Huber. 2004. The specific fates of tight junction proteins in apoptotic epithelial cells. J. Cell Sci. 117:2097-2107.

139. Tanaka, Y., K. Irie, T. Hirota, T. Sakisaka, H. Nakanishi, and Y. Takai. 2002. Ectodomain shedding of nectin-1alpha by SF/HGF and TPA in MDCK cells. Biochem. Biophys. Res. Commun. 299:472-478.

140. Johnson-Leger, C., M. urrand-Lions, and B. A. Imhof. 2000. The parting of the endothelium: miracle, or simply a junctional affair? J. Cell Sci. 113 ( Pt 6):921-933.

141. Hirase, T., S. Kawashima, E. Y. Wong, T. Ueyama, Y. Rikitake, S. Tsukita, M. Yokoyama, and J. M. Staddon. 2001. Regulation of tight junction permeability and occludin phosphorylation by Rhoa-p160ROCK-dependent and -independent mechanisms. J. Biol. Chem. 276:10423-10431.

142. Lee, D. B., E. Huang, and H. J. Ward. 2006. Tight junction biology and kidney dysfunction. Am. J. Physiol Renal Physiol 290:F20-F34.

143. Tunggal, J. A., I. Helfrich, A. Schmitz, H. Schwarz, D. Gunzel, M. Fromm, R. Kemler, T. Krieg, and C. M. Niessen. 2005. E-cadherin is essential for in vivo

169

epidermal barrier function by regulating tight junctions. EMBO J. 24:1146-1156.

144. Nathan, C. 2002. Points of control in inflammation. Nature 420:846-852.

145. Malik, M., C. S. Bakshi, K. McCabe, S. V. Catlett, A. Shah, R. Singh, P. L. Jackson, A. Gaggar, D. W. Metzger, J. A. Melendez, J. E. Blalock, and T. J. Sellati. 2007. Matrix metalloproteinase 9 activity enhances host susceptibility to pulmonary infection with type A and B strains of Francisella tularensis. J. Immunol. 178:1013- 1020.

146. Elkington, P. T., C. M. O'Kane, and J. S. Friedland. 2005. The paradox of matrix metalloproteinases in infectious disease. Clin. Exp. Immunol. 142:12-20.

147. Kobayashi, K. S., and R. A. Flavell. 2004. Shielding the double-edged sword: negative regulation of the innate immune system. J. Leukoc. Biol. 75:428-433.

148. Yipp, B. G., G. Andonegui, C. J. Howlett, S. M. Robbins, T. Hartung, M. Ho, and P. Kubes. 2002. Profound differences in leukocyte-endothelial cell responses to lipopolysaccharide versus lipoteichoic acid. J. Immunol. 168:4650-4658.

149. Marino, M. W., A. Dunn, D. Grail, M. Inglese, Y. Noguchi, E. Richards, A. Jungbluth, H. Wada, M. Moore, B. Williamson, S. Basu, and L. J. Old. 1997. Characterization of tumor necrosis factor-deficient mice. Proc. Natl. Acad. Sci. U. S. A 94:8093-8098.

150. Fukata, M., K. S. Michelsen, R. Eri, L. S. Thomas, B. Hu, K. Lukasek, C. C. Nast, J. Lechago, R. Xu, Y. Naiki, A. Soliman, M. Arditi, and M. T. Abreu. 2005. Toll-like receptor-4 is required for intestinal response to epithelial injury and limiting bacterial translocation in a murine model of acute colitis. Am. J. Physiol Gastrointest. Liver Physiol 288:G1055-G1065.

151. Rakoff-Nahoum, S., J. Paglino, F. Eslami-Varzaneh, S. Edberg, and R. Medzhitov. 2004. Recognition of commensal microflora by toll-like receptors is required for intestinal homeostasis. Cell 118:229-241.

152. Raulet, D. H., and R. E. Vance. 2006. Self-tolerance of natural killer cells. Nat. Rev. Immunol. 6:520-531.

153. Mori, T., X. Wang, T. Aoki, and E. H. Lo. 2002. Downregulation of matrix metalloproteinase-9 and attenuation of edema via inhibition of ERK mitogen activated protein kinase in traumatic brain injury. J. Neurotrauma 19:1411-1419.

154. Asahi, M., X. Wang, T. Mori, T. Sumii, J. C. Jung, M. A. Moskowitz, M. E. Fini, and E. H. Lo. 2001. Effects of matrix metalloproteinase-9 gene knock-out on the proteolysis of blood-brain barrier and white matter components after cerebral ischemia. J. Neurosci. 21:7724-7732.

155. Asahi, M., K. Asahi, J. C. Jung, G. J. del Zoppo, M. E. Fini, and E. H. Lo. 2000. Role

170

for matrix metalloproteinase 9 after focal cerebral ischemia: effects of gene knockout and enzyme inhibition with BB-94. J. Cereb. Blood Flow Metab 20:1681-1689.

156. Zapata, J. M., K. Pawlowski, E. Haas, C. F. Ware, A. Godzik, and J. C. Reed. 2001. A diverse family of proteins containing tumor necrosis factor receptor-associated factor domains. J. Biol. Chem. 276:24242-24252.

157. Hahn, D., A. Pischitzis, S. Roesmann, M. K. Hansen, B. Leuenberger, U. Luginbuehl, and E. E. Sterchi. 2003. Phorbol 12-myristate 13-acetate-induced ectodomain shedding and phosphorylation of the human meprinbeta metalloprotease. J. Biol. Chem. 278:42829-42839.

158. Litovchick, L., E. Friedmann, and S. Shaltiel. 2002. A selective interaction between OS-9 and the carboxyl-terminal tail of meprin beta. J. Biol. Chem. 277:34413-34423.

159. Friederich, E., K. Vancompernolle, D. Louvard, and J. Vandekerckhove. 1999. Villin function in the organization of the actin cytoskeleton. Correlation of in vivo effects to its biochemical activities in vitro. J. Biol. Chem. 274:26751-26760.

160. Miyoshi, J., and Y. Takai. 2008. Structural and functional associations of apical junctions with cytoskeleton. Biochim. Biophys. Acta 1778:670-691.

161. Miyoshi, J., and Y. Takai. 2005. Molecular perspective on tight-junction assembly and epithelial polarity. Adv. Drug Deliv. Rev. 57:815-855.

162. Mirzapoiazova, T., I. A. Kolosova, L. Moreno, S. Sammani, J. G. Garcia, and A. D. Verin. 2007. Suppression of endotoxin-induced inflammation by taxol. Eur. Respir. J. 30:429-435.

163. Takeichi, M. 1991. Cadherin cell adhesion receptors as a morphogenetic regulator. Science 251:1451-1455.

164. Cunningham, P. N., Y. Wang, R. Guo, G. He, and R. J. Quigg. 2004. Role of Toll- like receptor 4 in endotoxin-induced acute renal failure. J. Immunol. 172:2629-2635.

165. Cunningham, P. N., H. M. Dyanov, P. Park, J. Wang, K. A. Newell, and R. J. Quigg. 2002. Acute renal failure in endotoxemia is caused by TNF acting directly on TNF receptor-1 in kidney. J. Immunol. 168:5817-5823.

166. Johnson, D. E., C. V. Lockatell, R. G. Russell, J. R. Hebel, M. D. Island, A. Stapleton, W. E. Stamm, and J. W. Warren. 1998. Comparison of Escherichia coli strains recovered from human cystitis and pyelonephritis infections in transurethrally challenged mice. Infect. Immun. 66:3059-3065.

167. Mo, L., X. H. Zhu, H. Y. Huang, E. Shapiro, D. L. Hasty, and X. R. Wu. 2004. Ablation of the Tamm-Horsfall protein gene increases susceptibility of mice to bladder colonization by type 1-fimbriated Escherichia coli. Am. J. Physiol Renal Physiol 286:F795-F802.

171

168. Suliburk, J. W., K. S. Helmer, E. A. Gonzalez, E. K. Robinson, and D. W. Mercer. 2005. Ketamine attenuates liver injury attributed to endotoxemia: role of cyclooxygenase-2. Surgery 138:134-140.

169. Suliburk, J. W., E. A. Gonzalez, S. D. Kennison, K. S. Helmer, and D. W. Mercer. 2005. Differential effects of anesthetics on endotoxin-induced liver injury. J. Trauma 58:711-716.

170. Koga, K., M. Ogata, I. Takenaka, T. Matsumoto, and A. Shigematsu. 1994. Ketamine suppresses tumor necrosis factor-alpha activity and mortality in carrageenan- sensitized endotoxin shock model. Circ. Shock 44:160-168.

171. Takenaka, I., M. Ogata, K. Koga, T. Matsumoto, and A. Shigematsu. 1994. Ketamine suppresses endotoxin-induced tumor necrosis factor alpha production in mice. Anesthesiology 80:402-408.

172. Sun, J., Z. Q. Zhou, R. Lv, W. Y. Li, and J. G. Xu. 2004. Ketamine inhibits LPS- induced calcium elevation and NF-kappa B activation in monocytes. Inflamm. Res. 53:304-308.

173. Kawasaki, T., M. Ogata, C. Kawasaki, J. Ogata, Y. Inoue, and A. Shigematsu. 1999. Ketamine suppresses proinflammatory cytokine production in human whole blood in vitro. Anesth. Analg. 89:665-669.

174. Engel, D., U. Dobrindt, A. Tittel, P. Peters, J. Maurer, I. Gutgemann, B. Kaissling, W. Kuziel, S. Jung, and C. Kurts. 2006. Tumor necrosis factor alpha- and inducible nitric oxide synthase-producing dendritic cells are rapidly recruited to the bladder in urinary tract infection but are dispensable for bacterial clearance. Infect. Immun. 74:6100- 6107.

175. Frendeus, B., G. Godaly, L. Hang, D. Karpman, and C. Svanborg. 2001. Interleukin-8 receptor deficiency confers susceptibility to acute pyelonephritis. J. Infect. Dis. 183 Suppl 1:S56-S60.

176. Hinojos, C. A., E. Boerwinkle, M. Fornage, and P. A. Doris. 2005. Combined genealogical, mapping, and expression approaches to identify spontaneously hypertensive rat hypertension candidate genes. Hypertension 45:698-704.

177. Schild, L. 2004. The epithelial sodium channel: from molecule to disease. Rev. Physiol Biochem. Pharmacol. 151:93-107.

178. Blaustein, M. P., J. Zhang, L. Chen, and B. P. Hamilton. 2006. How does salt retention raise blood pressure? Am. J. Physiol Regul. Integr. Comp Physiol 290:R514- R523.

179. Iwamoto, T., S. Kita, J. Zhang, M. P. Blaustein, Y. Arai, S. Yoshida, K. Wakimoto, I. Komuro, and T. Katsuragi. 2004. Salt-sensitive hypertension is triggered by Ca2+ entry via Na+/Ca2+ exchanger type-1 in vascular smooth muscle. Nat. Med. 10:1193-

172

1199.

180. Norman, L. P. 2003. Dissertation: Disruption of the metalloprotease meprin beta gene in mice.

181. Jiang, W., and B. Le. 2000. Structure and expression of the human MEP1A gene encoding the alpha subunit of metalloendopeptidase meprin A. Arch. Biochem. Biophys. 379:183-187.

173

Renee E. Yura Vitae

Education 1998-2000 LaSalle University, Philadelphia, PA

2000-2002 B.S. in Biology, Summa Cum Laude King’s College, Wilkes-Barre, PA

2002-2008 Ph.D. in Cell and Molecular Biology, Molecular Medicine Option Pennsylvania State University College of Medicine, Hershey, PA

Honors and Awards 1998 Full Tuition University Scholarship, LaSalle University 2000 Member National Society of Collegiate Scholars, LaSalle University 2000 Moreau Scholarship, King’s College 2001 Member ΑΕΔ National Premedical Honor Society, King’s College 2002 Regina Award for Outstanding Biology Student, King’s College 2002 University Graduate Fellowship, Pennsylvania State University 2004, 2007 Travel Fellowship, American Society for Biochemistry and Molecular Biology 2006 Travel Fellowship, Metzincin Metalloproteases in Health and Disease International Conference, Switzerland 2007 Travel Fellowship, 5th General Meeting of the International Proteolysis Society, Greece 2007 Department of Health Tobacco Graduate Research Supplement Award

Publications Bond JS, Matters GL, Banerjee S, and Dusheck RE (2005) Meprin metalloprotease expression and regulation in kidney, intestine, urinary tract infections and cancer. FEBS Letters 579:3317-3322

Yura RE, Bradley SG, Ramesh G, Antonetti DA, Reeves WB, and Bond JS. Meprin A metalloprotease modulates the host response to lipopolysaccharide. In preparation

Yura RE, Bradley SG, Welch RA, and Bond JS. A role for meprin metalloproteases in urinary tract infection. In preparation

Published abstracts 2004, 2006, 2007 Experimental Biology Meetings (IUBMB/ASBMB); Poster 2006 Metzincin Metalloproteases in Health and Disease International Conference, Switzerland; Poster 2006 ACNP Annual Meeting 2006, Hollywood, FL; Poster 2007 Innoventure 2007, Hershey, PA; Poster 2007 5th General Meeting of the IPS, Greece; Poster 2007 American Society of Hematologists Conference, Atlanta; Poster