Investigating the Clinical Validity of CUB and Zona-Pellucida-Like Domain-Containing Protein 1 (CUZD1) in Malignant and Non- Malignant Human Diseases

Investigating the Clinical Validity of CUB and Zona Pellucida- like Domain-Containing Protein 1 (CUZD1) in Malignant and Non-Malignant Human Diseases

Sofia Farkona

A thesis submitted in conformity with the requirements for the degree of Doctorate of Philosophy Department of Laboratory Medicine and Pathobiology University of Toronto

Investigating the Clinical Validity of CUB and Zona-Pellucida-like Domain-Containing Protein 1 (CUZD1) in Malignant and Non- Malignant Human Diseases

Sofia Farkona

Doctor of Philosophy

Laboratory Medicine and Pathobiology

University of Toronto

2018 Abstract

CUB and zona pellucida-like domain-containing protein 1 (CUZD1) has been previously shown to be specifically expressed in normal pancreas and was proposed as a candidate biomarker for pancreatic related disorders. Due to the lack of specific reagents and techniques, its levels in tissues and biological fluids have not been extensively examined. We generated mouse monoclonal antibodies against recombinant CUZD1 and used them for the development of an enzyme-linked immunosorbent assay (ELISA). Analysis of various human extracts showed that

CUZD1 is measured in high levels in pancreas and at much lower (but detectable) levels in several other tissues. Analysis of biological fluids showed that CUZD1 is detected exclusively in pancreatic juice.

CUZD1 has been previously linked to diseases (such as pancreatitis, ovarian cancer and IBD) but it is currently unknown if the expression levels of this antigen are elevated in any of the aforementioned or other disorders. Analysis of a large number of serum samples from patients with various malignant and benign disorders showed that CUZD1 levels were elevated in patients with ovarian cysts but not ovarian cancer. ii

CUZD1 is a pancreas-specific protein but it is unclear if its expression is elevated in malignant conditions of the pancreas. IHC staining of pancreatic ductal adenocarcinoma (PDAC) and acinar cell carcinoma (ACC) tissue sections revealed that CUZD1 protein was highly expressed in ACC but not in PDAC.

CUZD1 is one of the targets of pancreatic autoantibodies (PABs) which have been emerged as possible biomarkers for Inflammatory Bowel Disease (IBD). Data assessing the diagnostic significance of CUZD1 autoantibodies in patients with IBD are scarce, mainly due to the lack of high throughput techniques for their detection. We developed an ELISA targeting CUZD1 autoantibodies and used it to analyze 200 serum samples from IBD patients and 129 patients assessed for various autoimmune diseases (vADs). CUZD1 autoantibodies were detected in 16% of CrD patients in 9% of UC patients and in less than 5% of patients being tested for vADs.

In conclusion, this thesis encompasses the development and validation of analytical techniques targeting CUZD1 antigen and CUZD1 autoantibodies. These tools can facilitate future investigations aiming to delineate the role of CUZD1 in physiology and pathobiology.

iii

Acknowledgments

First and foremost, I would like to thank my supervisors Dr. Eleftherios Diamandis and Dr. Ivan Blasutig for their continuous support and guidance throughout my PhD. I cannot thank you both enough for giving me the opportunity and freedom to pursue my scientific interests and for the patience and mentorship throughout my PhD.

I am most thankful to colleagues who contributed to this work and to all members of ACDC laboratory for supporting me and making my time here enjoyable. I will never forget the time we shared together.

I would also like to extend my gratitude and thanks to the members of my PhD advisory committee, Dr. Yousef and Dr. Schmitt Ulms for volunteering their time and for providing invaluable advice and feedback that led to completion of this exciting work. Additionally, I am thankfully to our collaborators Dr. Bogdanos, Dr. Ruckert and Dr. Serra who shared my enthusiasm and who graciously provided me with samples and resources. I would also like to thank the examiners of this thesis Dr. Charames, Dr. Chandran and Dr. Kavsak for reading it and giving useful feedback during the final exam.

I should thank the members of the Department of Laboratory Medicine and Pathobiology, including Dr. Harry Elsholtz, Sue Sarju and Rama Ponda for administrative help in helping scheduling the defense. I would like to acknowledge the funding I received from the LMP department.

Last but not least, none of this work would have been possible without the love and support of my family. Thank you to my brother, Peter, and more importantly my parents, Markos and Eutychia, for always believing me and encouraging me to follow my dreams. I know that I would not be where I am today if it wasn’t for you. I am eternally grateful, and this thesis is dedicated to you.

Table of Contents

Acknowledgments...... iv

Table of Contents ...... v

List of Abbreviations ...... ix

List of Tables ...... xiv

List of Figures ...... xv

Chapter 1 | Introduction ...... 1

1.1 Pancreas ...... 2

1.1.1 General information ...... 2

1.1.2 History of the pancreas ...... 2

1.1.3 Structure of the pancreas ...... 2

1.1.4 Development ...... 4

1.1.5 Development and Molecules...... 5

1.1.6 Physiology of the pancreas ...... 8

1.1.7 Islet-acinar axis/interaction ...... 14

1.1.8 Common diseases...... 16

1.2 Inflammatory Bowel Disease ...... 22

1.2.1 General information (classification, symptoms and prevalence) ...... 22

1.2.2 Causes ...... 23

1.2.3 Diagnosis...... 24

1.3 CUZD1 ...... 27

1.3.1 Genomic location and structure ...... 27

1.3.2 CUZD1 endogenous expression ...... 31

1.3.3 Cellular localization of CUZD1 ...... 34

1.3.4 Proteins similar to CUZD1 ...... 36

1.3.5 Function of CUZD1 (based on experiments) ...... 38

1.3.6 CUZD1 autoantibodies in patients with IBD ...... 40

1.4 Rationale and Objectives ...... 42

1.4.1 Rationale ...... 42

1.4.2 Hypothesis...... 42

1.4.3 Objectives ...... 42

2 Chapter 2 | Novel immunoassays for detection of CUZD1 autoantibodies in serum of patients with inflammatory bowel diseases...... 44

2.1 Introduction ...... 45

2.2 Materials and Methods ...... 47

2.2.1 Patients ...... 47

2.2.2 Production of recombinant CUZD1 ...... 47

2.2.3 Selected Reaction Monitoring (SRM) ...... 50

2.2.4 Chromatographic and MS conditions ...... 51

2.2.5 Data analysis ...... 51

2.2.6 Development of a fluorometric anti-CUZD1 autoantibodies immunoassay ...... 51

2.2.7 Statistics ...... 52

2.3 Results ...... 52

2.3.1 Production and purification of recombinant CUZD1 ...... 52

2.3.2 Anti-CUZD1 immunoassays ...... 55

2.3.3 Anti-CUZD1 antibodies in IBD patients ...... 57

2.3.4 Anti-CUZD1 antibodies in vADs ...... 64

2.4 Discussion ...... 64

3 Chapter 3 | Generation of monoclonal antibodies and development of an immunofluorometric assay for the detection of CUZD1 in tissues and biological fluids ...... 67

3.1 Introduction ...... 68

3.2 Materials and Methods ...... 69

3.2.1 Monoclonal antibody production ...... 69

3.2.2 Screening for immunogen-reacting clones by an IgG capture ELISA ...... 69

3.2.3 Expansion of hybridomas and purification of monoclonal antibodies ...... 70

3.2.4 ELISA development...... 70

3.2.5 Immunohistochemistry ...... 71

3.2.6 Tissue extracts and biological fluids ...... 71

3.2.7 SDS PAGE (& MS/MS) and Western blotting of Protein tissue extracts and biological fluids ...... 72

3.2.8 Fractionation of biological fluids containing CUZD1 ...... 73

3.2.9 MS/MS of bands and gel filtration fractions...... 73

3.3 Results ...... 74

3.3.1 ELISA development...... 74

3.3.2 Immunohistochemical analysis of pancreatic tissue using the in-house generated monoclonal antibodies ...... 77

3.3.3 Western Blot analysis of tissues and biological fluids...... 78

3.3.4 Fractionation of tissue extracts and biological fluids containing CUZD1 ...... 80

3.4 Discussion ...... 82

4 Chapter 4 | Investigating the clinical utility of CUB and zona pellucida-like domain- containing protein 1 (CUZD1) in malignant and non-malignant human diseases ...... 85

4.1 Introduction ...... 86

4.2 Materials and Methods ...... 87

vii

4.2.1 Patients ...... 87

4.2.2 Analysis of samples by using in-house developed ELISA ...... 89

4.2.3 Immunohistochemistry ...... 90

4.3 Results ...... 91

4.3.1 Analysis of serum samples by using in-house developed ELISA ...... 91

4.3.2 Immunohistochemical analysis of IBD, ACC and PDAC ...... 95

4.4 Discussion ...... 97

5 Chapter 5 | Summary and Future Directions ...... 102

5.1 Summary of key findings ...... 103

5.2 Future directions ...... 108

References or Bibliography ...... 110

viii

List of Abbreviations

AC, adenylate cyclase

Ach, acetylcholine

AP, acute pancreatitis

ASCA, anti-Saccharomyces cerevisiae antibodies

ATP adenosine triphosphate bFGF, basic fibroblast growth factor bHLH, basic helix-loop-helix

BN-PAGE, blue native polyacrylamide gel electrophoresis

BRCA2, breast cancer 2 early onset gene

BSA, bovine serum albumin cAMP, cyclic adenosine monophosphate cbCUB, calcium-binding complement subcomponents C1s and C1r, Uegf, Bmp1

CCK, cholecystokinin

CDCPs, CUB domain containing proteins

CDKN2A/p16INK4A, cyclin-dependent kinase inhibitor 2A

CFTR, cystic fibrosis transmembrane conductance regulator

CoA, acetyl-coenzyme A

(COX)-2, cyclooxygenase

CP, chronic pancreatitis ix

CrD, Crohn’s disease

CUZD1, CUB and zona pellucida-like domain-containing protein 1

DMBT1, Deleted in malignant brain tumors

ELISA: enzyme-linked immunosorbent assay

EPGN, epigen

EPR,epiregulin

EST, expressed sequence tag

FAE, follicle-associated epithelium

5-FU, 5-fluorouracil

FPG, fasting plasma glucose

GDM, gestational diabetes

GLUT2, glucose transporter 2

GFP, green fluorescent protein

GHSR, growth hormone secretagogue receptor

GP2, glycoprotein 2

G-6-P, glucose-6-phosphate

GPI, glycosyl phosphoinositol

EGF, epidermal growth factor domain

ERG1, estrogen regulated gene 1

Gαq G, heteromeric G

HES1, Hes Family BHLH Transcription Factor 1

HRP, horseradish peroxidase

HSP, heat shock protein

IBD, Inflammatory Bowel Diseases

IFN, interferon

IIF, indirect immunofluorescence

LC-MS/MS, liquid chromatography-tandem mass spectrometry

LKB1/STK11, serine/threonine–protein kinase 11

MASP2-MBL, mannan-binding lectin-associated serine protease 2

MODY, maturity onset diabetes of the young

Ngn3, Neurogenin 3

NOD2, nucleotide binding oligomerization domain containing 2

Nrg1, neuregulin

OGTT, oral glucose tolerance test

ORF, open reading frame

PABs, Pancreatic autoantibodies pANCA, perinuclear anti-neutrophil cytoplasmic antibodies

PanINs, Pancreatic intraepithelial neoplasias

PCPEs, procollagen proteinases enhancers

PDAC, pancreatic ductal adenocarcinoma

PDX-1, pancreatic duodenal homeobox

PKA, protein kinase A

PKC, protein kinase C

PLC, phospholipase C

PP, pancreatic polypeptide

PP, peyer’s patches

SHH, Sonic hedgehog

SIRS, systemic inflammatory response syndrome

SMAD4/DPC4, mothers against decapentaplegic homolog 4

SOCs, store operated channels

SP-D, surfactant protein-D

TGFβ, transforming growth factor-beta

TGFR-3, transforming growth factor β receptor type 3

PTF1a, transcription factor 1a

TEM, transmission electron microscopy

Th17, type 17 helper T-cell

T1DM, type 1 diabetes mellitus

T2DM, type 2 diabetes mellitus

TP53, tumour protein p53

UEA, europaeus agglutinin I

xii

UC, ulcerative colitis

UO-44, uterine-ovarian specific gene 44

UTCZP, uterine motif zona pellucida vADs: various autoimmune diseases.

VIP, vasoactive intestinal polypeptide

WHO, World Health Organization

ZP, zona pellucida

xiii

List of Tables

2.1 Main demographic, clinical and laboratory characteristics of the patients with Crohn’s disease (CD), ulcerative colitis (UC) and control individuals included in the present study.

4.1 Characterization of samples.

xiv

List of Figures

1.1 Macroscopic anatomy of the human pancreas.

1.2 CUZD1 protein features including the CUB and ZP domain positioning within the protein.

1.3 Some of the diverse modular proteins harbouring cbCUBs

1.4 A general mechanism for assembly of ZP domain proteins.

1.5 Amino acid comparison between human CUZD1 and DMBT1.

2.1 Production of CUZD1 recombinant protein.

2.2 CUZD1 quantification in culture supernatant by SRM.

2.2 Purification of the protein fragment corresponding to the extracellular region of CUZD1 from supernatant of Expi293F cells by AIEX-FPLC.

2.3 ELISA assay cutoff values.

2.4 Analysis of serum samples.

2.5 Contingency table displaying the prevalence of IgG or IgA autoantibodies or both, in patients with A) Crohn’s disease (CrD). B) Ulcerative Colitis (UC).

2.6 Contingency table displaying the prevalence of autoantibodies in IBD patients, as determined by ELISA or indirect immunofluorescence (IIF).

2.7 Analysis of 129 serum samples from patients previously tested for autoimmune diseases.

3.1 Α) Expression of CUZD1 in adult tissue extracts from males and females. Β)

Quantification of CUZD1 in biological fluids.

3.2 Immunohistochemical localization of CUZD1 protein in paraffin-embedded pancreatic tissues.

3.3 Western blot showing detection of CUZD1 protein in pancreas related samples.

3.4 HPLC separation on a size exclusion/gel filtration column of pancreatic tissue extract (A) pancreatic juice (B) and recombinant CUZD1 (C).

4.1 Detection of CUZD1 protein levels in serum sample set A.

4.2 Detection of CUZD1 protein levels in serum sample set B.

4.3 Detection of CUZD1 protein levels in serum sample set D.

4.4 Immunohistochemical localization of CUZD1 protein in paraffin-embedded intestinal tissues from patients with inflammatory bowel disease (IBD).

4.5 Immunohistochemical localization of CUZD1 protein in paraffin-embedded pancreatic tissues from patients with acinar cell carcinoma (ACC)

4.6 Immunohistochemical localization of CUZD1 protein in paraffin-embedded pancreatic tissues from patients with pancreatic ductal adenocarcinoma (PDAC).

xvi

Chapter 1 | Introduction

1 2

1.1 Pancreas

1.1.1 General information

The pancreas is a complex organ that plays a key role in digestion, utilization and storage of energy substrates. It is located in the abdominal cavity and it is composed of a mixture of endocrine and exocrine cells functioning as both an endocrine and an exocrine gland. The endocrine cell components produce and secrete several important hormones for the maintenance of glucose homeostasis. The major function of the exocrine portion is to secrete pancreatic juice containing digestive enzymes responsible for normal digestion and absorption of daily foodstuffs (and finally assimilation of nutrients into the body) and is tightly regulated by the endocrine system.

1.1.2 History of the pancreas

The pancreas was first described by the Greek anatomist and surgeon Herophilus, who was born in 336 BC in Chalcedon. A few hundred years after Herophilus, Rufus of Ephesus, another surgeon and anatomist, gave the organ the name pancreas which literally means all-flesh or all- meat, presumably because of the absence of bone or cartilage in it and its fleshy composition 1, 2.

1.1.3 Structure of the pancreas

The human pancreas is a retroperitoneal organ of the upper abdomen, weighing 50-100 g and measuring 14-18 cm in length 3. Anatomically, the pancreas borders with other abdominal organs such as the spleen, stomach, duodenum and colon and, macroscopically, it can be divided into three parts, termed the head, body and tail 4 (figure 1.1). The head region of the pancreas is a C- shaped part situated within the first loop of the duodenum 4 and it surrounds two blood vessels, the superior mesenteric vein and artery. The body, the largest part of the pancreas, is located underneath the stomach, it crosses with the superior mesenteric artery and vein, abdominal aorta, inferior vena cava and portal vein 4. The tail region borders on the hilum of the spleen 5.

Blood supply to the pancreas is provided by the superior posterior pancreaticoduodenal artery, the inferior pancreaticoduodenal artery and the splenic artery 5. The superior and inferior pancreaticoduodenal arteries, which are pancreatic branches of the gastroduodenal and superior mesenteric artery 6, respectively, supply the head of the pancreas 4. The splenic artery is a branch of the celiac artery and supplies the pancreatic body and tail 4. Venous drainage of the pancreas

3 occurs by way of the splenic vein, which lies posterior to the pancreas while the blood from the head of the pancreas is drained through the portal system or vein 7, 8. Lymphatic drainage from the head of the pancreas is mediated via the pancreaticosplenic, pancreaticoduodenal, subpyloric, and hepatic lymph nodes. The pancreatic body and tail drain into the celiac lymph nodes, superior mesenteric, and para-aortic and aortocaval lymph nodes 4, 6. The pancreas is richly innervated by sympathetic, parasympathetic, and afferent fibers 5. The exocrine portion/component of the pancreas, accounts for most of the cellular mass of the

Figure 1.1 |Macroscopic anatomy of the human pancreas. Human pancreas consists of the head, the body, and the tail and it neighbors with other organs of the abdomen such as the spleen, stomach, duodenum and colon (not shown in the figure). (Reproduced from 4)

4 pancreas 4. Microscopically, the exocrine cells are arranged in lobules with a branched network of tubules 9, composed of grape-like clusters called acini 4, 5, 9. Each acinus consists of highly orientated, pyramidal-shaped acinar cells packed with membrane-bound secretory granules which contain digestive enzymes that are exocytosed into the lumen of the acinus or intercalated duct 10. From there, the acinar cell secretions flow into larger intralobular ducts which then converge into the main pancreatic duct and drains directly into the duodenum 4, 5.

The islets of Langerhans, which comprise the endocrine portion of the pancreas, constitute 1–4% by volume of the total pancreatic volume 4. The islets are of spherical shape, composed of a few to several thousand endocrine cells and their main function is the secretion of several pancreatic peptide hormones which are responsible for glucose homeostasis 4, 5. There are at least 5 types of polypeptide-hormone- secreting endocrine cells which comprise the islet. The insulin-secreting beta cells constitute 50–70% of cells in islets while glucagon-secreting alpha cells make up 20– 40% of the islet cell population 4. Delta cells and PP cells contribute 10% of cells and they release somatostatin and pancreatic polypeptide, respectively. Epsilon cells are present in less than 1% of the cells and they secrete ghrelin 4, 5.

1.1.4 Development

Pancreas development has been demonstrated to depend on epitheliomesenchymal interactions 5, 11-13. The embryonic pancreas in vertebrates originates from two primordial buds of the primitive gut epithelium and in humans is first apparent at 4 weeks of gestation. The dorsal bud extends into the dorsal mesentery to eventually form the head, body and tail of the pancreas 12, 13. The ventral bud is smaller than the dorsal bud, it arises directly adjacent to hepatic diverticulum and it forms the posterior part of the head and the uncinate process 5, 11-13.

At an early phase, around the seventh week of gestation in humans, because of both gut rotation and elongation of the dorsal and the ventral stalk, the ventral bud fuses with the distal portion of the dorsal bud 5, 12. This coalescence results in the formation of the main pancreatic duct, termed the duct of Wirsung, which is the main conduit of the pancreas running the entire length of the organ 14. The proximal end of the dorsal pancreatic duct persists as the smaller (accessory) pancreatic duct termed duct of Santorini 14.

1.1.5 Development and Molecules

Pancreas develops in response to signals generated in the adjacent mesodermal tissues including notochord, aorta and cardiac mesoderm/notochord 12, 15. Signals from the notochord include among others transforming growth factor-beta (TGFβ) and the notochord-produced morphogens, activin β and basic fibroblast growth factor (bFGF, also known as FGF2) 5, 12. Sonic hedgehog (SHH) is an activator of the Hedgehog signaling pathway 12, whose signaling was shown to inhibit pancreatic formation at the onset of organogenesis 5. Studies by Kim et al showed that notochord removal from early chicken embryos in vitro prohibited appropriate dorsal pancreas development and inhibited transcription of pancreatic specific genes such as pancreatic duodenal homeobox (PDX-1) 16. In vitro experiments carried out on cultures of early chick endoderm showed that notochord proximity to the endoderm could suppress SHH expression 16. In a similar manner, deleting notochord in explanted chick embryo cultures resulted in SHH expression ectopically in the pancreatic region, with failure of pancreatic development 16, 17. Additionally, implantation of an ectopic additional notochord could lead in suppression of endodermal SHH, accompanied by ectopic induction of cell-shape changes similar to early pancreas formation 16, 17.

A variety of transcription factors controls both the initiation and maintenance of well- coordinated gene expression patterns to achieve proper cell differentiation during pancreatic development 5, 12, 18. PDX-1, one of the most heavily studied transcription factors, plays a pivotal role in the growth and differentiation of the pancreatic buds 19. It is first expressed at E8.5 (10 somites) in pancreatic progenitors of the mouse foregut and it diverts cells away from the duodenal, hepatic and bile duct fates 12, 19, 20. In the early pancreas, PDX-1 expression is detected within the developing pancreatic epithelium, but then, as the pancreas matures, its expression becomes restricted to a subset of acinar and endocrine cells 20, 21. In adult tissue, PDX-1 is mainly found in pancreatic polypeptide (PP) and insulin expressing cells and PDX-1 null mice fail to form a pancreas 14, 20, 21. An early pancreatic dorsal bud, which formed only a few insulin and glucagon-expressing cells, was detected in these but a ventral bud was not seen 22. Similarly, exon-2 deleted mutant mice demonstrated the formation of an early abortive dorsal bud, with only glucagon-expressing cells and metaplasia of the duodenal mucosa into a bile duct phenotype associated with duodenal obstruction 23, 24. PDX-1 mutations in humans were shown to result in pancreatic agenesis in homozygous null mutants, and insulin insufficiency in heterozygous mutants 25. Zebrafish with PDX-1 null mutations are also apancreatic 26. These data

6 demonstrate that PDX-1 plays a pivotal role in the early phases of pancreas formation. It was more recently shown that PDX-1 function is required during midpancreatic development for both islet and acinar differentiation and that in adults PDX-1 has a critical part in beta-cell function, including regulation of insulin expression as its temporal inactivation in adult pancreas results in severe loss of islet area and diabetes 19, 27.

Pancreas specific transcription factor 1a (PTF1a) is a second molecule whose deficiencies are noted to result in pancreatic agenesis 28. Originally thought to be an exocrine specific transcription factor/described as part of a large heterotrimeric transcriptional regulator complex which regulates acinar enzyme gene expression, it is now recognized as another regulator of general pancreatic organogenesis 28. PTF1a expression is detected slightly later than the expression of PDX-1, at E9.5, in cells of the foregut endoderm destined to generate dorsal and ventral pancreas, and unlike PDX-1, PTF1a is not expressed in other parts of the foregut. By lineage tracing analysis, it was shown that, PTF1a-expressing progenitor cells give rise to all pancreatic cell types. Suppression of PTF1a directs pancreatic progenitors towards intestinal cell fate 28, 29. At later stages, its expression continues mainly in epithelial and acinar cells, supporting the idea that this transcription factor has an exocrine-specific role in midpancreatic development 28, 30. Knockdown studies in mice and zebrafish confirm the crucial role of PTF1a in pancreas specification. Similarly, to PDX-1 null mutants, PTF1a deficient mice develop with only an aborted dorsal pancreatic bud, and a minuscule ventral bud, which actually originates from the bile duct. PTF1a knockdown Xenopus and zebrafish embryos do not harbor exocrine pancreas 31. Whereas, the role of PTF1a in exocrine development is clear, the role (if any) in endocrine development remains elusive. While exocrine pancreas does not form in PTF1a null mutant mice, endocrine cells are generated and migrate through the mesenchyme to populate the spleen 28. Although PTF1a expression does not rely on PDX-1 expression and PDX-1 expression is not dependent on PTF1a expression it has been demonstrated that progenitor cells co-expressing PDX-1 and PTF1a are directed to adopt a pancreatic fate. Loss of HES1 expression, and thus defect in Notch signaling 32, in mice results in ectopic pancreas formation 33, including endocrine, exocrine and ductal pancreatic cells in PDX-1 expressing stomach, duodenum and common bile duct, and, according to lineage tracing analysis, the ectopic pancreas formation is accomplished by misexpression of PTF1a 33. Similarly, it has been demonstrated that combined ectopic expression of PDX-1 and PTF1a transforms posterior endoderm into exocrine and

7 endocrine tissues in Xenopus suggesting that cells simultaneously expressing PDX-1 and PTF1a are embryonic pancreatic stem cells.

Neurogenin 3 (Ngn3) is a basic helix-loop-helix (bHLH) transcription factor that is transiently expressed in the developing pancreas and specific for endocrine development/key driver of endocrine differentiation 34. Ngn3 is firstly expressed in the early pancreatic epithelium at E9, and peaks at E15.5, a time frame that corresponds to the wave of endocrine differentiation while its expression decreases by E17.5 and it is almost undetectable in adult pancreas 12, 35. Ngn3 seems to be a key driver of endocrine differentiation 34. Ngn3 is necessary and sufficient for endocrine development. Transgenic mice in which Ngn3 has been eliminated lack all endocrine cell types 34. On the contrary, overexpression of Ngn3 induces precocious endocrine differentiation 34. Forced overexpression of Νgn3 and at an inappropriately early time in the developing mouse pancreas led to pancreas characterized solely by an endocrine phenotype composed primarily of beta cells 36. When embryonic endoderm derived from chicks was electroporated with an Ngn3 expression vector, the intestine was induced to express glucagon and somatostatin but not insulin 36. Later studies by Schwitzgebel et al., demonstrated that PDX- 1-Νgn3 ectopic expression results in the generation of mainly glucagon-producing alpha cells with a very small number of insulin-positive cells present, suggesting that for the determination of insulin-producing cells, additional signals beyond plain Ngn3 are necessary 37. Ngn3 expression is repressed by notch-mediated intracellular signaling on close/proximal epithelial cells. Notch is a transmembrane receptor, which when activated by its transmembrane ligand Delta, promotes the expression of HES-1 thereby resulting in suppression of the expression of Ngn3. Blocking Notch receptor activation in early pancreatic progenitors leads to high Ngn3 expression, which evokes early endocrine cell differentiation at the expense of pancreatic cell proliferation. For this reason, activation of Notch signaling is crucial to prohibit premature pancreatic progenitor-cell differentiation, thereby allowing subsequent proliferation and morphogenesis of the pancreatic progenitor cells.

Downstream target genes of Ngn3 include endocrine-specific transcription factors Nkx2.2, Pax4, Nkx6.1 and PDX-1 whose transient expression drives Ngn3-positive cells towards a β-cell phenotype and Brn4, Nkx6.2 and Pax6 which direct cells towards an α-cell phenotype 5, 34. Changes in the expression of these regulatory genes will result in disruption of the balance of pancreatic hormone-expressing cells developed 5. Nkx2.2, in particular, represents a key

8 regulator of islet specification. In Nkx2.2 mutant animals, alpha- and beta-cell development is severely damaged whereas a ghrelin-expressing cell population is found augmented 38.

1.1.6 Physiology of the pancreas

Pancreatic acinar cells

Pancreatic acinar cells constitute the majority of the exocrine portion of the pancreas. They are classical polarized pyramid-shaped epithelial cells, accounting for 82% of the total volume of the pancreas 39 and they are morphologically and functionally specialized for the exocrine secretion of fluid and enzyme containing vesicles across the apical membrane 10. Pancreatic acinar cell polarization is preserved by tight and adherens gap junctions to adjacent cells 40. Within each acinar cell the protein containing vesicles called zymogen granules are polarized at the apical side of the acinar cells and the nucleus is found at the basal pole 10, 40. These proteins include mainly the zymogen granules precursors, trypsinogen, chymotrypsinogen and procarboxypeptidases, active digestive enzymes such as pancreatic lipase, amylase and nuclease as well as other secretory products such as co-lipase 10. Following its release into the duodenum, inactive trypsinogen is cleaved and thereby transformed into active trypsin by enterokinase which is localized/located on/to the apical membrane of enterocyte 41. In the duodenum trypsin further catalyzes the cleavage of trypsinogen into trypsin and activates other precursor proteases into their respective active forms, such as chymotrypsinogen into chymotrypsin. The activation of proteases in the duodenum rather than within the pancreas safeguards the organ from autodigestion, preventing the development of pancreatitis 42. The opening of calcium-dependent chloride channels found on the apical membrane of acinar cells releases chloride ions into the acinar lumen followed by an influx of sodium ions into the lumen through tight junctions between acinar cells 43. Sodium chloride in the acinar lumen osmotically pulls in water creating net unidirectional fluid flow 43. In regards with enzyme secretion, upon cell stimulation, zymogen granules fuse with the apical plasma membrane resulting in the release of the granule content into the acinar lumen 40. Since the lumen is continuous with the duct, the granule content flows through the ducts and out into the duodenum.

The intake of food is the primary stimulus for the regulation of pancreatic exocrine secretion and it is believed to act via neural and hormonal systems 10, 44. The major regulatory molecules include acetylcholine (ACh) released from vagal postganglionic neurons, cholecystokinin (CCK)

9 released from intestinal endocrine cells as well as some paracrine/endocrine molecules such as secretin, vasoactive intestinal polypeptide (VIP) and angiotensin II 10. These secretagogues elicit downstream signalling transduction pathways by activating receptors localized in the basolateral plasma membrane 45. The receptors for CCK and ACh, muscarinic and CCKB receptors respectively, by binding to heteromeric G (Gαq G) proteins 46, use the phospholipase C (PLC)/Ca2+ signal transduction pathway and eventually lead to enzyme secretion from pancreatic acinar cells 10. On the other hand, binding of VIP and secretin to their respective receptors activate Gαs G-proteins which stimulates adenylate cyclase (AC), resulting in the production of cyclic adenosine monophosphate (cAMP) and the activation of protein kinase A (PKA). Finally, PKA and protein kinase C (PKC) prompts zymogen granules fusion with the apical plasma membrane, driving the secretion of the vesicles contents into the acinar lumen 47.

It has long been known that exocytosis is induced by a secretagogue-induced elevation in cytosolic/intracellular Ca2+ levels 43, 45. Voltage operated Ca2+ channels, common to nerves and muscles, are not present in pancreatic acinar cell. For this reason, the calcium rise in this cell type does not depend on extracellular influx but on release of calcium from intracellular stores 48. It has been demonstrated by experimental data that the initial secretory response to the binding of ACh or CCK on the basolateral aspects of the acini is independent of extracellular Ca2+ 48. However, because prolonged secretion depletes intracellular Ca2+, a surge in extracellular Ca2+ is required to maintain Ca2+ signaling and it is achieved through Ca2+ influx via membrane channels referred to as store operated channels (SOCs) 49, 50. Low concentration of CCK or ACh induces calcium spikes near the secretory pole of the acinar cells which selectively activate chloride currents resulting in fluid secretion 49, 50. On the contrary, increased stimulation levels caused by high concentration of secretagogue induces global calcium waves spreading over the entire cell, which activate both chloride current and zymogen granule fusion leading in this way to both fluid and enzyme secretion 47, 48.

Duct cells

Duct cells represent a minority of the exocrine portion of the adult pancreas 51. They comprise less than 10% of the total glandular tissue mass and they create a branching tree of epithelial cells which narrow in diameter from the main pancreatic duct through consecutive formation of interlobular, intralobular and intercalated cells 52. They have a crucial role beyond their forming

10 an epithelial lining of branched tubes that delivers enzymes from acini into the duodenum; they produce a sodium bicarbonate and mucin rich alkaline fluid that is secreted into the duodenum and neutralizes the acidic chyme entering from the stomach 52. Bicarbonate supplies an optimal pH for the activation of pancreatic digestive enzymes securing the normal physiological digestion of the food and inhibits injury to the duodenal mucosa or peptic ulcer disease 53. Defect of the duct cells thus, leads to maldigestion and malabsorbtion 51. Centroacinar cells, which have several ductal features/traits and are considered as the terminals of the ductal tree, link the duct ends with the acinar cells. Digestive enzyme produced by acinar cells are initially drained into the intercalated ducts, followed by intralobular ducts, and finally an interlobular duct 52. Intralobular ducts merge into the main pancreatic duct which together with the bile duct, opens into the duodenum at the ampulla of Vater 52. The sphincter of Oddi, and its contraction simultaneously controls bile and pancreatic juices entering the duodenum 52. The major producers of bicarbonate are cells in the intercalated and intralobular ducts 52. These ducts are lined with principal cells of the simple epithelium that possess the morphological hallmarks associated with active transport: abundant mitochondria, apical microvilli and elaboration of the basolateral plasma membrane. In the largest branches of the network, goblet cells, which contribute to mucus production, are intermingled with ductal cells and form about 2% of the structure 52.

The concentration of bicarbonate in pancreatic juice is approximately 120-140 mmol/L which is far in excess of the normal concentration in plasma and extracellular fluids (approximately 24 mmol/L). Acinar enzymatic secretion and pancreatic ductal bicarbonate are boosted after the consumption of food 53. Secretion of bicarbonate-rich fluid by ductal epithelium is primarily stimulated by the peptide hormone secretin secreted by the gut endocrine cells (D-cells) in response to the arrival of gastric chyme and vasoactive intestinal peptide (VIP) 53. Although CCK and ACh are not direct agonists of ductal bicarbonate secretion some studies have indicated that they could potentiate the stimulatory effects of secretin on duct epithelial cells 52.

Secretion of high levels of bicarbonate into the ductal fluid is directly controlled by several ion channels, prominently the chloride channel called cystic fibrosis transmembrane conductance regulator (CFTR) 54, specifically expressed on the apical membrane of ductal cells.

According to the CFTR-mediated bicarbonate secretion model, CO2 enters the cell from the basolateral side by passive diffusion and carbonic anhydrase, which is expressed within the duct cell, catalyzes the formation of carbonic acid from CO2 and H2O and the subsequent dissipation into bicarbonate and protons 55, 56. The protons are transported out of the cell through a Na+/H+ exchanger on the basolateral membrane. The driving force for the antiporter is provided by the Na+ pump which pumps Na+ out of the cell, thus providing a driving force for the Na+/H+ exchanger with the establishment of a concentration gradient of Na+ ions 55. Basolateral K+ channels maintain a hyperpolarized basolateral membrane while bicarbonate is secreted out in exchange for Cl− via a disulfonic stilbene-sensitive pathway located in the apical membrane 56. The accumulated Cl− recycling depends critically on the regulation and activity of the CFTR channel whose open state probability is increased by cAMP 55.

Ductal cells of the pancreas have been proposed to be a source of pancreatic stem cells 51. It has been shown that, similar to islet cells, duct cells harbor some proliferative properties, even after the morphogenesis of pancreas has been fully completed. When the pancreas is injured by duct ligation, cellophane wrapping, pancreatectomy, genetically targeted destruction by interferon (IFN) or streptozotocin-induced destruction of beta cells, there is some increase in the mitotic activity of the duct cells and limited regeneration of the organ 57. A subset of cells co-express several ductal markers and the developmental marker PDX-1 57. These findings led to the hypothesis that, pancreatic duct cells serve as progenitors or stem cells for the generation of new pancreatic cells. It has been further shown that in vitro culture of preparations enriched in ducts from normal human pancreas or non-obese diabetic mice are capable of generating insulin- secreting cells and reverse diabetes in vivo. Also, human ductal cells have been found to initiate endocrine cell differentiation, in vitro, upon expression of several specific transcriptional factors 58. However, since there is a lack of specific duct markers a clear interpretation of these data is compromised 44.

Islet cells

On the contrary to the vast exocrine portion of the pancreas, the endocrine portion comprises only about 1-2% of the total mass of the organ. Nevertheless, islets of Langerhans play a fundamental role in glucose control and metabolism within the body. In spite of their small cell population, islet cells are abundantly vascularized receiving approximately 10-15% of the total

12 pancreatic blood flow 44. Four main cell types are detected within the islets, each producing and secreting a distinct primary protein product. Comprising about 80% of total islet tissue, insulin- secreting beta cells are the dominant cell type, and are located centrally and throughout the pancreas 59. Glucagon-secreting alpha cells comprise almost 20% of islet tissue, and are located mainly in the periphery, in the tail, body and superior part of the head of the pancreas. Somatostatin-secretin delta cells are located between these two cell types, and are a minority of the total islet cells population 59. Pancreatic polypeptide-secreting-F cells or PP-cells are detected mostly in islets of the posterior lobe and receive a different blood supply 59. Blood flow travels from the center of the islet to the periphery, thereby allowing insulin to finely control glucagon and somatostatin release. The endocrine part of the pancreas is richly innervated by the autonomic nervous system 44, 59.

The peptide hormone insulin reduces blood glucose levels by promoting its absorption by peripheral tissues such as adipose and skeletal muscle and repressing glucose production by the liver 59. Insulin, similarly to other hormone proteins, it is initially synthesized as a larger molecule called preproinsulin. Preproinsulin harbors a signal peptide which directs its folding and movement/transportation through the Golgi but is removed during its insertion into the endoplasmic reticulum. Proinsulin, which consists of two polypeptide chains named A and B, linked by a connective peptide known as C peptide, is converted to insulin when in the endoplasmic reticulum C peptide is cleaved by specific endopeptidase known as prohormone convertase 60. Removal of C peptide enables the exposure of the end of the insulin chain allowing its interaction with insulin receptors 60. After they are packed into secretory granules in the Golgi apparatus of beta cells, the metabolically inert C peptide and mature insulin are released in equimolar amounts into the blood by exocytosis.

Glucose is the principal secretagogue for insulin but amino acids ingested with a meal, volatile fatty acids, ketone bodies, glucagon, certain gut hormones, and vagal β-adrenergic stimulation also cause release 59. On the contrary, insulin secretion is inhibited by fasting, exercise and somatostatin as well as by enhanced α-adrenergic activity. Insulin enters the beta cells through a membrane-bound glucose transporter, gucose transporter 2 (GLUT2), which has a low affinity for glucose, thereby allowing a graded response. Once in beta cells, glucose is phosphorylated to glucose-6-phosphate (G-6-P) by an enzyme called glucokinase (hexokinase IV) which leads to the chain reaction of glycolysis resulting in the production of acetyl-coenzyme A (CoA) and

13 adenosine triphosphate (ATP) via the citric acid cycle 44, 59. The concomitant rise in the ratio of ATP to ADP within the cell inhibits an ATP-sensitive potassium channel leading to a decrease of K+ efflux, an overall membrane depolarization of the cell and influx of calcium ions through the plasma membrane 59. Finally, in the presence of Ca2+, insulin-containing secretory vesicles attached to microtubules are expelled, thereby releasing insulin into the blood circulation 61. Pancreatic beta cells are also known to synthesize amylin. Amylin, which is a 37 amino acid peptide, is stored and co-secreted with insulin in a ratio of one to one from secretory granules and it is thought to regulate gastric motility, renal resorption, and metabolic actions 62. Formation of islet amyloid deposits from amylin has been reported to be developed only in a few species namely humans, primates, and cats and are believed to act as a barrier to the diffusion of nutritive substances and glucose, thereby reducing beta cell mass and insulin secretion, and overall acting as a key pathophysiological factor in the pathogenesis of diabetes 62, 63.

Alpha cells comprise up to 20% of the human islet cells and their main function is to produce peptide hormone glucagon. Glucagon acts in opposition to insulin by promoting output rather than input of glucose through stimulation of de novo synthesis of both glycogenolysis and gluconeogenesis 44. Glucagon release is activated when the levels of glucose in the bloodstream are too low, but inhibited by hyperglycaemic conditions. Increased concentration of aminoacids, epinephrine in the plasma and vagal activation further induce glucagon secretion while somatostatin suppresses its release in a paracrine fashion 44. Glucagon counters the effect of insulin on glucose homeostasis by binding to the protein-coupled glucagon receptor which leads to the activation of PKA. PKA in turn, phosphorylates various enzymes resulting in the release of glucose-1-phosphate from glycogen polymers 59. Alpha cells have also been reported to produce ghrelin, a ligand of the growth hormone secretagogue receptor (GHSR) 64, which stimulates appetite and food intake, while boosting fat mass deposition and weight gain through its actions on several organ systems including the central nervous and gastro-intestinal systems 65. Recently it has been suggested that the ghrelin-secreting cells are ontogenetically and morphogenetically distinct from alpha-cells and beta-cells and they have designated epsilon cells 66.

Delta cells of the pancreas contribute a relative few cell population of the islet and their main role is to produce somatostatin 44. Somatostatin is also synthesized in the GI tract, in the brain as well as in other tissue and its secretion is induced by nutrients, gastrointestinal hormones,

14 glucagon, and neurotransmitters while insulin inhibits it 44, 59. Somatostatin acts as an inhibitor on both pancreatic exocrine secretion, particularly digestive enzyme release via its mediation of cholecystokinin and secretin and on insulin and glucagon secretion by the islet cells, thereby significantly reducing blood glucose levels in humans and diabetic subjects 44, 59, 67.

PP cells are dispersed throughout the islets, primarily concentrated in the dorsal part of the head of the pancreas, and they comprise less than 1% of islet cell population 44. PP cells are known to secrete pancreatic polypeptide (PP) in a process which is induced by food ingestion, mainly due to protein and fat content, cholinergic stimulation, and hypoglycaemia and suppressed by glucose 44. PP has been reported to induce gastric acid secretion, to inhibit exocrine and endocrine pancreatic secretion, to mediate relaxation of gallbladder motility, and regulation of migrating motor complex of gut smooth muscle 44. Furthermore, it has been shown that increased concentration of PP in the blood are related to a rise in intra-abdominal fat accumulation and thus insulin resistance in humans 68.

1.1.7 Islet-acinar axis/interaction

The pancreas is the only organ in the body where the endocrine and exocrine components are intermixed within the parenchyma. Because there are no basal membranes or compartmentalization capsules for different cell types in the pancreas, the islets of Langerhans are distributed within the exocrine pancreas 44, 69. This unusual anatomical arrangement accommodates the interrelationship between the endocrine and exocrine pancreas mediated by islet-derived hormones, other humoral factors as well as neurotrasmitters released by the nerves innervating the pancreas 70. The close interaction between the endocrine and exocrine components of the pancreas has been demonstrated by the fact that streptozotocin-induced islet destruction in experimental animals led not only in anatomically recognized atrophy but also in a functional reduction in the exocrine portion of the pancreas/exocrine insufficiency 71. Reduced pancreatic exocrine function is also well appreciated in insulin-dependent diabetic patients 72.

Acini in the immediate vicinity of islets, called peri-insular acini, differ morphologically from those located further distant, called tele-insular acini 69. The first ones consist of larger sized cells, possess/harbor larger nuclei and more abundant zymogen granules and their pattern of pancreatic enzymes differ 69, 73, 74. According to previous studies high local concentrations of pancreatic hormones in the region around the islets justify this “halo” phenomenon 69. Wallgren

15 et al observed that the duck exocrine cells possessing the largest nuclei were in contact with beta cells 75 and Kramer et al showed that elimination of beta cells led to the disappearance of this distinct morphological difference 76. These data were argued by later studies showing/demonstrating persistent halos following streptozotocin-induced islet destruction in the same species 77.

The islet-acinar portal system demonstrates a unique anatomic organization that aids to the interaction of several proteins secreted by the islets with the surrounding acini. Briefly, pancreatic intralobular arteries divide into the islets through vas afferents, which then form capillary plexus within the islets. The efferent vessels extend to the surrounding exocrine pancreas to assemble the insulo-acinar portal system 44. In humans (and other species including monkey, pig, cattle, rabbit, dog, cat) the majority of the islets are intralobular in location and emit exclusively islet-acinar portal vessels and only few islets transmit efferent vessels that extend to the interlobular veins. All the efferent islet blood flows into the acinar capillary before leaving the pancreas; in doing so, no blood from the intralobular islets drains directly into veins without passing through the exocrine portion of the pancreas 70. On the contrary, in rats and mouse some venous vessels from the islets drain into both islet-acinar portal vessels and intralobular veins and the interlobular arteries drain into the interlobular veins 70.

The interrelationship/interaction between the endocrine and endocrine pancreas is mediated by islet derived bioactive agents such as insulin and can be exemplified by its effects on exocrine secretion 44, and more specifically on amylase secretion 78.

Insulin, a major and one of the most well characterized regulators of endocrine pancreatic function, is secreted by beta islets cells and plays two important roles. Systemically, it facilitates glucose absorption in tissue and locally on the exocrine pancreas it affects pancreatic growth and exocrine function 79. Insulin has been shown to bind to its own receptor on the acinar secretion resulting in stimulation and potentiation of amylase secretion in response to peptide hormone CKK, ACh or carbamylcholine 79. This is the case for both exogenously supplied and endogenous insulin released by glucose 70.

Endogenous somatostatin is secreted by the pancreatic delta cells and its action on exocrine secretion is inhibitory 80. Somatostatin acts both as neurotransmitter and as a hormone but it is not certain which of the two roles leads to its inhibitory function on the exocrine cells 70. It has

16 been suggested that somatostatin acts as hormone by inhibiting insulin release and that the stimulatory effect on insulin is mediated by CCK and cerulein-stimulated amylase secretion 70. Another proposed mechanism involves binding of somatostatin to its receptors on the acinar cell resulting in a reduction of intracellular cyclic AMP and subsequent Ca2+ signaling 70. Somatostatin functions as a neurotransmitter by controlling vagal and sympathetic pathways or through the SST receptor 2 in the dorsal vagal complex (DVC) 81. Indirectly, it can also inhibit pancreatic secretion through the intrapancreatic cholinergic mechanism by inhibiting acetylcholine release in the peripheral nerve terminals 70.

The role of glucagon on the exocrine secretion is still controversial. Although previous in vivo studies have shown that glucagon inhibits pancreatic secretory response to CCK, secretin or meals, investigators that used extracted glucagon applied to mouse, guinea pig or pancreatic acinar cells reported a stimulatory effect 70. The latter effects were attributed to contaminants present in an impure preparation of natural glucagon.

1.1.8 Common diseases

Pancreatitis

Pancreatitis is an inflammatory condition of the pancreas which, in humans, may be manifested as a pain in the upper abdomen expanding to the back, nausea, and vomiting 82. Acute pancreatitis (AP) is a serious abdominal inflammation characterized by pancreatic tissue edema, acinar cell necrosis, hemorrhage and inflammation of the damaged gland 83. AP is more frequent in Caucasian than in Asian populations. While the general incidence rate is about 300 per million people annually, in the USA and UK it is 700–800 and 150–420 per million annually 84, respectively, and only 106–205 per million annually in Japan 85. Although there are discrepancies among the published studies there has been reported a rise in the number of people with AP, probably because of life-style changes and increased exposure to risk factors in recent decades 86. The spectrum of the disease varies from edematous pancreatitis, which is the mildest form, to severe acute pancreatitis which is characterized by a massive destruction of pancreatic parenchyma and infiltration of leukocytes with potentially fatal complications 10.

Alcoholism and biliary obstruction account for 35% and 45% of AP cases, respectively 86. Remaining pancreatitis cases are often drug-induced or idiopathic cases 10. The main characteristic of alcoholism-induced AP is the formation of protein-related plugs in bile ducts

17 which results in the activation of proteolytic enzymes and overall in pancreatic autodigestion. In gallstone obstruction-induced AP, gallstone forming leads to clog of Ampulla of Vater, and thus regression of biliary and pancreatic juices into pancreatic ducts which eventually causes pancreatic inflammation 86.

An early and important step of acute pancreatitis is the premature activation of trypsinogen to trypsin within the pancreas leading to autodigestion of the gland 87. This results in the release of active pancreatic enzymes into the bloodstream which induces the production of inflammatory cytokines by neutrophils, macrophages, and lymphocytes 88. The release of these molecules initiates an inflammatory cascade that leads to the systemic inflammatory response syndrome (SIRS) which may further develop into acute respiratory distress syndrome and multiorgan dysfunction syndrome 89.

Some patients may progressively develop chronic pancreatitis (CP) which is preceded by recurrent bouts of AP. However, AP and CP are two distinct diseases with discrete characteristics in regards with pancreatic morphology and clinical outcome. The pancreas of CP patients usually bear atrophy and apoptosis while patients with AP usually experience pancreatic swelling 86. Furthermore, in AP the defect is restricted to the exocrine pancreas which gradually results in exocrine insufficiency while in advanced stage of CP endocrine malfunction is manifested in the form of pancreatic fibrosis 86.

Although there has been progress in our understanding of the pathogenesis of AP in recent years promising and effective therapy is still not available 86. For patients who develop mild edematous to moderate pancreatitis the therapeutic approaches include analgesic (opiate) administration for pain relief, intra venous fluids for volume depletion and fasting until the symptoms of acute inflammation subside 10.

Patients with severe AP are usually admitted to an intensive care unit. First-line treatment modalities for clinical AP include enteral/parentenal nutrition, surgical removal of necrotic tissue and some surgical manipulations such as cholecystectomy 90. Meanwhile, gallstone-induced severe acute pancreatitis can be managed with endoscopic retrograde cholangiopancreatography 10. According to clinical studies, the administration of antibiotics has a positive effect in the prevention of infected pancreatic necrosis 91. Due to the involvement of inflammatory cells, inflammatory mediators and cytokines in the pathophysiology of acute pancreatitis, therapeutic schemes that make use of cytokine antagonism and antioxidants may emerge as additional tools

18 for the treatment of acute pancreatitis 10. Additional potential therapeutic schemes against AP include cyclooxygenase (COX)-2 inhibition, substance P antagonism, and heat shock protein (HSP) activation 86.

Pancreatic cancer

Pancreatic cancer is a deadly malignant disease, which constitutes the fourth leading cause of cancer mortality 92. There are several types of pancreatic cancer, tumors that arise from the pancreatic exocrine tissue (ductal adenocarcinoma and acinar cell carcinoma), the pancreatic endocrine tissue (insulinoma, glucagonoma, gastrinoma and VIPoma) as well as atypical neoplasms (lymphomas, mesotheliomas and sarcomas) 93. Accounting for approximately 80% of all pancreatic cancers, pancreatic ductal adenocarcinoma (PDAC) is the most frequent histological type and for this reason the term “pancreatic cancer” usually refers to this type 94. Despite the arrival of novel agents, the clinical outcome of PDAC is still poor. Only 25% will survive for one year while only 5% survive for 5 years 95. Pancreatic intraepithelial neoplasias (PanINs) are well established as the most common precursors of PDAC 96. It has been documented that the progression from PanIN to invasive pancreatic cancer occurs over many years or even decades 96, 97.

Owing to the lack of specific symptoms in the disease's early stages, PDAC is usually not diagnosed until it has reached an advanced stage frequently when the tumors have metastasized 95. These symptoms diverge according to where the tumor is localized in the pancreas. Most of the people that are diagnosed with pancreatic cancer experience unexplained weight loss, nausea, vomiting or a feeling of weakness 98. The diagnosis of pancreatic cancer is relied on a combination of the clinical symptoms, medical imaging techniques such as endoscopic ultrasound or computed tomography, examination of biopsies and laboratory findings 98. CA19- 9, the most commonly used tumor marker in PDAC, is frequently found in higher levels in the sera of patients with PDAC 99. However, it is neither sensitive nor specific enough to be used as biomarker for the screening or diagnosis of PDAC.

Pancreatic cancer occurs more frequently beyond the age of 50 and the incidence is higher in males. Risk factors for pancreatic cancer include cigarette smoking, family history, a diet rich in fat and meat, exposure to chemicals, and a history of diabetes mellitus, chronic pancreatitis and gastric surgery 92, 100. The genetic factors/alterations that characterize PDAC promote

19 tumorigenesis and determine the disease phenotype. One of those factors that frequently define patients with PDAC is re-activation of telomerase expression and telomere dysfunction 101. Also, Shh and Notch signaling pathways, normally involved in pancreatic organogenesis and development, have been reported to be involved in the initiation of PanIN and development of tumor growth 102. Mutations in oncogenes and tumor suppressor genes that result in tumor formation have been also documented 92. KRAS oncogene mutation, which is an early initiation event in PDAC tumorigenesis promoting cell proliferation, differentiation and survival, occurs in 95% of patients with PDAC 92. Genetic alterations in tumor suppression genes such as cyclin- dependent kinase inhibitor 2A (CDKN2A/p16INK4A), tumour protein p53 (TP53), mothers against decapentaplegic homolog 4 (SMAD4), serine/threonine–protein kinase 11 (LKB1/STK11) and the breast cancer 2 early onset gene (BRCA2) are also associated with PDAC formation 92.

Pancreatic surgery has been markedly improved and the related postoperative morbidity and mortality have been reduced considerably over the last three decades. However, surgical operation can be offered only in around 10-25% of patients with pancreatic cancer and this is one of the reasons why the prognosis for patients with pancreatic cancer remains poor 86. The resectability of the tumor is determined from a variety of factors such as the location of the tumor and the extent of its spread as well as involvement of the venous or arterial blood vessels. Chemotherapy is usually offered to most patients as an adjuvant therapy postoperatively or as a neoadjuvant therapy before surgery to eliminate neoplastic cells 86. The major schemes for pancreatic cancer include gemcitabine, 5-fluorouracil (5-FU), cisplatin, oxaliplatin, mitomycin C, and doxorubicin in mono- or combination modalities 86.

At the same time, molecularly targeted therapies, which attack cells with specific genetic characteristics, thereby inhibiting specific pathways for the growth and progression of malignant tumors have also been employed. The EGFR tyrosine inhibitor erlotinib, following published data showing that it increased median survival, was approved in 2005 by FDA for its usage in combination with gemcitabine for treating locally advanced, unresectable, or metastatic pancreatic cancer 103. Additionally, Kras and p53 tumor suppressor gene mutations, which are detected in about 70% of pancreatic cancer have been prevalent targets in preclinical gene therapy studies 103.

Previous studies have proved the importance of NF-κB in tumorigenesis and particularly its pivotal role in the induction and manifestation of chemo-resistance in pancreatic cancer cells 104. For this reason, inhibition of NF-κB activity has been considered a useful strategy for increasing sensitivity towards cytostatic drug treatment. However, pharmacological inhibition of NF-κB appeared to be insufficient, probably because solid tumors appear to be protected from induction of apoptosis and NF-kB inhibition had been suggested to be used as chemo-sensitizing adjuvant in combination with other cytotoxic agents 104.

Cancer immunotherapy, the treatment that harnesses the patient’s immune system to fight cancer, is now emerging as an important addition to conventional therapies 92. Among the several immunotherapeutic approaches that are investigated for their potential to eradicate tumor cells as sole agents or to improve tumor response to previously used therapies are anticancer vaccines, adoptive cell transfer and chimeric receptors, and blockade of immune checkpoints 92.

Diabetes

Diabetes Mellitus refers to a group of metabolic disorders in which there is prolonged high blood glucose levels 105.

The World Health Organization (WHO) distinguishes among normoglycemia, impaired fasting glycaemia, impaired glucose tolerance (IGT) and diabetes mellitus based on fasting plasma glucose (FPG) levels and 2-h oral glucose tolerance test (OGTT) results as follows: normoglycemia is FPG <6.1 mmol/l (110 mg/dl) and 2-h OGTT <7.8 mmol/l (140 mg/dl); impaired fasting glycaemia is FPG ≥6.1(≥110) and <7.0(<126) and 2-h OGTT < 7.8 mmol/l (140 mg/dl); impaired glucose tolerance (IGT) is FPG <7.0 mmol/l (<126 mg/dl) and 2-h OGTT ≥7.8 mmol/l (≥140 mg/dl) and <11.1 mol/l (<200 mg/dl); and diabetes mellitus is FPG ≥7.0 mmol/l (≥126 mg/dl) or 2-h OGTT ≥11.1 mmol/l (≥200 mg/dl) 106.

Common symptoms of untreated diabetes include increased urination, referred to as polyuria, increased thirst, referred to as polydipsia and increased hunger, referred to as polyphagia86. Untreated diabetes is often related with serious chronic complications including nephropathy, which can cause end-stage renal failure, retinopathy which can lead to blindness and neuropathy which may result in diabetic foot 86.

Based on its specific pathogenesis, diabetes mellitus is classified to 3 major types: type 1 diabetes mellitus (T1DM), type 2 diabetes mellitus (T2DM) and gestational diabetes (GDM). Per the WHO records, the two main types, namely T1DM and T2DM, comprise approximately 10 and 90% of diabetics, respectively 105.

T1DM is one of the most common metabolic disorders in children and young adults which results from an absolute deficiency of insulin secretion from pancreatic β-cells. In most cases TD1M is of immune-mediated nature showing characteristics of autoimmune diseases, such as T cell-involved attacks which lead to the destruction of beta cells 107. T1DM pathogenesis is believed to involve genetic, epigenetic and environmental factors. Although several gene loci are known to increase the risk of diabetes, including the HLA-DR and -DQ genes, which account for approximately 40-50% of the disease risk 108, the dramatic rise in the incidence of the disease that has been observed in the past few decades cannot be attributed merely to genetic factors 107. This is further supported by the fact that fewer than 10% of individuals who are genetically predisposed to T1DM de facto develop the disease 107. Environmental factors that may influence the risk of developing T1DM are diet, birth delivery mode, hygiene, and antibiotic usage 107. Because these environmental risk factors modulate the composition of the gut microbiome it has been speculated that microbiota constitute a link between those factors and disease promotion 107.

T2DM is characterized by insulin resistance in skeletal muscle and adipose tissue, which may be combined with relatively reduced insulin secretion. The etiology of T2DM involves both genetic and environmental factors. In regards with the latter ones, diet, lack of physical activity and obesity are known to be involved 86. In the case of the genetic factors, according to epidemiological studies, having a diabetic first-degree relative causes a fourfold increase in one’s risk of developing diabetes 86. Additionally, the concordance rate for type 2 diabetes in monozygotic twins is higher than that of dizygotic twins, demonstrating than genes, more than environmental or maternal factors are responsible for the excess risk 86. Also, maturity onset diabetes of the young (MODY), an autosomal dominant genetic disease leading to metabolic abnormality, is one form of T2DM which has been linked to dysfunction of a single gene 86.

Despite the classification, there are cases of patients who seem to manifest features that characterize both T2DM and T1DM 86. For example, some patients with severe T2DM may also require insulin therapy to preserve euglycemic control. Furthermore, age is no longer a strict

22 classification criterion. T1DM is most frequently observed at individuals younger than 30, but it can develop at any age 109. On the other hand, while T2DM was believed to develop at adults, it can also occur in younger children 110. What both types have in common is loss of beta-cell mass over time, causing a defect in glucose tolerance, hyperglycemia and overt diabetes 86.

GDM refers to any degree of glucose intolerance with a beginning during pregnancy. It has been suggested that placental hormones produced during pregnancy may decrease sensitivity to insulin 86. However, the exact cause of GDM is still unclear 86.

Insulin injection treatment is still the main treatment of T1DM because glycemic control is vital for preventing chronic complications 86. The major disadvantage of this strategy is that it can lead to hypoglycemia. Novel approaches include immunoregulatory techniques which deal with autoimmune β-cell loss 111. Furthermore, pancreatic or renal-pancreatic transplantation together with long-term immunosuppressive drugs are suggested as a potential strategy for the treatment of T1DM 112. β-cell transplantation, although it has been accomplished experimentally in animals and in human patients, it is not considered practical for clinical use mainly because of the limited availability of islets 113. It is anticipated that stem-cell approaches to islet transplantation may solve this problem in the future since hyperglycemia has been completely reversed in some cases 86.

In regards to T2DM treatment, the major considerations are to improve insulin secretion and/or insulin sensitivity and to reduce chronic complications. Among the preventive strategies are the decrease of risk factors, and enhancement of β-cell function and cell mass 114. Life-style modifications for the reduction of obesity are considered essential and pharmacological interventions, such as oral hypoglycemic drugs are also effective as first line treatment for T2DM treatment 115.

1.2 Inflammatory Bowel Disease

1.2.1 General information (classification, symptoms and prevalence)

An autoimmune disease develops when the immune system, which naturally defends the body from foreign invaders, pathologically attacks healthy cells. Inflammatory Bowel Diseases (IBD) are heterogeneous autoimmune disorders that affect the gastrointestinal tract. The two principal

23 forms of IBD are Crohn’s disease (CrD), which affects any segment of the gastrointestinal tract from the mouth to the anus and ulcerative colitis (UC) which is limited to the colon 116.

While IBD is rarely fatal on its own, the symptoms including diarrhea, fever, fatigue, anemia, weight fluctuations, reduced appetite, joint pain, and abdominal cramping impair the quality of life of the patient 117.

An estimated of 233,000 people live with IBD in Canada, 129,000 with CD and 104,000 with UC, which is equal to prevalence of 1 in 150 117. IBD affects about 1.7 million Americans (estimated 780,000 with CD; estimated 907,000 with UC), with an incidence rate of 10.7 per 100,000 individuals for CD and 12.2 per 100,000 individuals for UC 117. The annual incidence of CD and UC in Europe was reported as 12.7 per 100,000 person-years and 24.3 per 100,000 person-years, respectively, with a reported CD prevalence of 322 per 100,000 persons and UC prevalence of 505 per 100,000 persons 117. Meanwhile, the incidence and prevalence rates of both CD and UC in Asia have been increasing and it is believed that this change reflects the industrialization of these parts of the world 117.

1.2.2 Causes

A variety of hereditary, environmental and lifestyle factors are thought to affect IBD pathology. Although it is now believed that IBD results from an inappropriate inflammatory response to intestinal microbes in a genetically susceptible host, it is still not known how these elements individually and collectively influence the natural history of the disease 116.

In regards with the genetic contribution, the gene nucleotide binding oligomerization domain containing 2 (NOD2), which regulates bacterial populations in the gut 118, autophagy genes 119, and components of the interleukin-23–type 17 helper T-cell (Th17) pathway have been shown to increase 120. Environmental and lifestyle factors include diet enriched in saturated fats and animal protein, stress, anxiety, depression, physical inactivity and disrupted sleep 116. Smoking affects differently the risk of developing the two types of IBD. Current smokers are nearly twice as likely to have CD as non-smokers, but they also seem to be at lower risk of UC 116. Furthermore, it is believed that the exposure of young children to bacteria may decrease IBD risk by enriching microbial diversity in the gut 116. On a similar note, early antibiotic use, which disrupts these communities, constitutes a risk factor 116.

1.2.3 Diagnosis

The diagnosis of IBD mainly relies on endoscopic and histopathological assessment of the inflamed intestine 121. Distinguishing between CrD and UC can be challenging since some patients present with overlapping features of the two conditions 116. Since the clinical management and therapies of CrD and UC can be distinct, especially in the case of surgical intervention, confirmation of the condition in question is crucial 122.

Because of their minimally invasive and economical characteristics, several serological biomarkers have been tested for their usefulness in diagnosis, disease stratification and prognosis of IBD 123. Aberrant immune response against endogenous or microbial antigens in genetically susceptible individuals plays a role in the pathogenesis of IBD and for this reason serum antibodies against microbial antigens or auto-antigens have also emerged as potential biomarkers 123.

Among the various molecules that have been recently proposed as potential serum biomarkers, anti-Saccharomyces cerevisiae antibodies (ASCA) and perinuclear anti-neutrophil cytoplasmic antibodies (pANCA) are the ones that have been studied the most 123. They have been shown to have value in distinguishing patients with CrD from patients with UC, to contribute to the stratification of IBD and to some extent reflect the disease phenotype and disease severities 123. However, they have a limited clinical value in the primary diagnostic work up, primarily because of their poor sensitivity and moderate specificity 123.

Pancreatic autoantibodies (PABs) targeting the exocrine pancreas are encountered in the serum of IBD patients 124 and have been reported to be highly specific for CrD. PAB recognition has relied on indirect immunofluorescence (IIF) of pancreas sections, which gives two distinct patterns: an extracellular droplet-like staining in the acinar lumen and a reticulogranular cytoplasmic pattern of the pancreatic acinar cells 125. The diagnostic and clinical utility of PAB in IBD, as revealed by IIF, have been partly investigated but the published data are not consistent. Some studies have shown that PABs are detected in up to 42% of patients with CrD and in less than 5% of patients with UC or other pathological conditions and healthy controls 126, 127. According to some other studies IgA PAB are present in about 23% of patients with UC and in up to 13% of patients with celiac disease 125. Furthermore, PAB have also been proposed as markers for clinical features of the disease, e.g. microbial load, severity of penetration, site of

25 inflammation, age at disease onset, duration of the disease etc. However, data failing to confirm associations between PABs and clinical characteristics of the disease have also been published 128 and for this reason they cannot be exploited for clinical practice. The inconsistency amongst the reported data may be due to differences associated with the IIF methodology or study cohorts 125.

Following several unsuccessful attempts, the targets of PABs have been recently identified. The pancreatic major glycoprotein 2 (GP2) of the zymogen granule is the target of PABs giving the reticulogranular type pattern while PABs giving the reticulogranular type pattern target CUB and zona pellucida-like domain-containing protein 1 (CUZD1) 125.

The recent development of immunoassays for the measurement of GP2 autoantibodies has provided a large body of information regarding their diagnostic and clinical relevance in patients with IBD. On the other hand, the clinical utility of CUZD1 autoantibodies has been assessed mostly by IIF assays and their (clinical and diagnostic) significance remain unclear. In general, GP2 autoantibodies are detected in approximately 21–45% of patients with CD and in 2–19% in UC 129, 130. Several studies have reported that GP2 autoantibodies/anti-GP2 antibody seropositivity is associated with disease phenotype. More specifically, GP2 autoantibodies are more prevalent in CrD patients with ileocolonic location 131, stricturing behavior 132, perianal and early onset of the disease 132. These data support the assumption that reactivity against GP2 can assist in the diagnosis and stratification for IBD 130.

For several years, the mechanisms responsible for the induction of these autoantibodies and their pathogenic relevance remained unclear as both of their targets, GP2 and CUZD1, were previously believed to be exclusively expressed by pancreatic acinar cells 125, 133. However, a large body of data has been generated which shows that GP2 is also expressed in the apical surface of the M cells of the peyer’s patches (PP) 131, 134. More importantly, GP2 expression has been reported to be elevated in the targeted tissue of patients with CrD compared to patients with UC. Emerging evidence suggests that PP, which are particularly abundant in the distal part of the ileum, to be potential sites of the inflammatory onset in CrD 134. Thus, GP2 expressed at the outer membranes of M cells in the PP appears to be located in the very center of CrD inflammation and explain, at least partly, the previously unaddressed contradiction of pancreatic autoimmunity and intestinal inflammation 134. Respective data regarding CUZD1 expression in

26 the intestine are lacking, raising all together the question of the relevance of these autoantibodies in CrD.

Current data demonstrate that anti-GP2 autoantibodies or GP2 itself might play a pathophysiological role in CrD and that it does not represent merely an epiphenomenon in CrD inflammation 131. Hase et al have demonstrated that GP2 can act as a bacterial uptake receptor for bacterial antigens and, therefore, involved in the mucosal immune response to these particular bacteria (involved in the generation of humoral immune responses to molecules of the intestinal content interacting specifically with GP2 on M cells) 135. In addition, it has been shown that GP2 has a broad(er) pro-phagocytotic ability than previously assumed and that it is a binding partner of the scavenger receptor expressed on endothelial cells I (SREC-I), which is also located on dendritic cells 136. Interaction of GP2 with SREC-I and uptake may have profound effects on antigen clearance and the modulation of the immune responses, as dendritic cells play an important role in the generation of innate and adaptive immunity 136. In summary, it seems that GP2 has an immunomodulating role in innate and adaptive intestinal immunity, the study of which have led to a better understanding of autoimmunity against GP2 in CrD-specific inflammation.

1.3 CUZD1

Our laboratory has published a novel strategy where we mined publicly-available gene and protein bioinformatic databases, to identify tissue-specific proteins that may be useful cancer serum biomarkers 137. CUB and zona pellucida-like domain-containing protein 1 (CUZD1) was identified in this study as a highly pancreas-specific protein and was proposed as a candidate biomarker for pancreatic related disorders.

The gene, which was first cloned from the mouse uteri, was introduced as the uterine motif zona pellucida (UTCZP) and then as its 80% homologous estrogen regulated gene 1 (ERG1) isolated from the uteri of rats 138. CUZD1 was also previously known as the uterine-ovarian specific gene 44 (UO-44) or HuUO-44 specifying the human ortholog 139.

1.3.1 Genomic location and structure

The gene is mapped to chromosome 10 q, a region of frequently lost in many malignant tumor types such as prostate adenocarcinoma, endometrial cancers, glioblastoma multiforme, and small cell lung cancer, suggesting that this chromosomal region harbors several tumor suppressor genes important for suppression of tumorigenesis and cancer progression.

According to a study published in 2004 by Leong et al, CUZD1 is subjected to a complicated array/series of alternative splicing events that yields four splice variants 140, 141. The functional roles of these splice variants are still unknown, but they give rise to three protein isoforms. The largest isoform, approximately 607 aminoacids long (68 kDa), harbors at the N terminal side a secretory signal sequence , two calcium-binding complement subcomponents C1s and C1r, Uegf, Bmp1 (cbCUB) domains, one zona pellucida (ZP) domain and at the C-terminal side a single-spanning transmenbrane region, a short cytoplasmic tail 141 and it is predicted to be heavily glucosylated. The other two isoforms, 326 aminoacids (37 kDa) and 241 aminoacids (28 kDa) long, contain only the ZP domain and the transmembrane region or half of the ZP domain and the transmembrane region, respectively 141. The presence of a signal peptide at the N terminal aminoacid sequence of the largest isoform along with the presence of the transmembrane region suggest that this polypeptide could either localize at the membrane of the cells or, under certain conditions, secrete to the extracellular space/outside the cells. On the other

28 hand, lacking the secretory signal the two smaller isoforms are predicted to be intracellular. The main features of CUZD1 protein are shown in figure 1.2.

Figure 1. 2 | CUZD1 protein features including the CUB and ZP domain positioning within the protein. (Reproduced from 140).

cbCUB and ZP domains, occupy most the extracellular region of the largest isoform. They are found in diverse proteins and are evolutionarily highly conserved, suggesting that these domains may play a role in fundamental cellular processes 142, 143. CUB domains are named from the three proteins in which they were firstly identified, the human complement subcomponents C1r/C1sC1s, the embryonic sea urchin protein Uegf and bone morphogenetic protein-1 (BMP-1) 142. They are 110-residue protein motifs exhibiting an antiparallel β barrel, like those in immunoglobulins 142. While most CUB domain containing proteins (CDCPs) are known to participate in developmental processes such as embryogenesis and organogenesis others are involved complement activation, inflammation and autoimmunity, cell adhesion, cell migration, extracellular matrix degradation, cell signaling, axon guidance and neurotransmission 142. Although CUB domains are found in functionally diverse proteins (figure 1.3), i.e. growth factors, proteases, activators of the complement system, in general their function is to mediate cell adhesion, recognition of substrates or binding partners and interaction with extracellular

29 matrix components 142. For example, procollagen proteinases enhancers (PCPEs) stimulate the activities of PCPs in a substrate specific manner which involves both their CUB domains. cbCUB domain is a particular subset (endowed with Ca binding ability) which has been identified relatively recently from the resolution of the crystal structures of the N-terminal CUB1– epidermal growth factor domain (EGF) segments of human C1s and mannan-binding lectin-associated serine protease 2 (MASP2-MBL) 144. Unlike other CUB domains, cbCUB domains harbor a homologous Ca2+ binding site that underlies conserved binding site mediating ionic interactions with protein ligands similar to the interactions mediated by the low-density lipoprotein receptor family. cbCUB-mediated protein–ligand interactions usually involve multipoint

Figure 1.3 | Some of the diverse modular proteins harbouring cbCUBs

N-terminal ends are on the left-hand side in all cases. Transmembrane domains are represented as black boxes. Signal sequences and cytoplasmic domains are not depicted. The domain nomenclature and symbols used are those defined in the SMART database. (Reproduced from 144).

30 attachment through several cbCUBs, resulting in high-affinity binding through avidity, despite the low affinity of individual interactions 142. Proteins containing cbCUBs have highly specific roles, often involved in major biological functions such as immune defense and development, and it has been suggested that these domains largely contribute to this specialization, by conferring on them the ability to specifically recognize their protein ligands 142.

ZP domains were described first for the eponymous zona pellucida proteins ZP1, ZP2 and ZP3, which are involved in sperm-egg recognition 143. Later they were identified in tectorins, uromodulin and the major zymogen granule membrane protein (GP2), which are all proteins involved in binding 143. In the cytokine signaling co-receptor transforming growth factor β receptor type 3 (TGFR-3), also known as betaglycan, contributes to cytokine ligand recognition. This structural element is a sequence of approximately 260 aminoacids with eight conserved cysteine residues that is located close to the C terminus of the polypeptide and is present in secreted proteins from a wide variety of species from nematodes to mammals 145. ZP domain proteins are often glycosylated, mosaic structures consisting of multiple types of domains and they perform highly diverse functions, ranging from serving as structural components of egg coats, appendicularian mucous houses, and nematode dauer larvae, to serving as mechanotransducers in flies and receptors in mammals and nonmammals. ZP domain proteins are generally present in matrices or filaments 143. This observation is consistent with the function that it has been proposed for them, to drive the polymerization/assembly of the protein it belongs to, in polymeric protein networks 143. In 2008 Litscer et al showed that HPLC-purified mouse ZP2 and ZP3 form high molecular weight homomeric oligomers when they are analyzed/resolved by blue native polyacrylamide gel electrophoresis (BN-PAGE) that separates proteins and multiprotein complexes under non-denaturing conditions. These structures form despite treatment of the ZP-containing proteins with SDS and 8M urea and storage of the glycoproteins at -80 following lyophilzation. Additionally, examination of HPLC-purified ZP2 and ZP3 by transmission electron microscopy (TEM) revealed the presence of long interconnected fibrils composed of contiguous beads. The ZP domain occurs next to a putative transmembrane region and they contain a long hydrophobic sequence segment and a conserved furin cleavage stretch at or near the C-terminal. According to a general mechanism that has been proposed for their assembly, ZP domain containing proteins are proteolytically cleaved by a member of the furin convertase family at the conserved furin cleavage site before they are released to the extracellular space and assemble into high molecular weight structures/fibrils

(figure 1.4) 145. Consistent with this observation, CUZD1 contains a ZP domain next to the transmembrane region at the C-terminal which harbors a putative furin cleavage site close to the C-terminus.

1.3.2 CUZD1 endogenous expression

In his work published in 1998, Kasik reported that the expression of CUZD1 mRNA in mice/the mouse ortholog of CUZD1, UTCZP, is temporo-spatial; it appears in the uterus 6 days prior to birth, it increases over subsequent days to attain maximal levels at 3 days prior to birth and then abruptly decreases during the last 3 days of pregnancy 146. The expression of this mRNA is practically undetectable in non-pregnant uterus or in a variety of adult and fetal tissues. Kazik by searching the expressed sequence tag (EST) database using peptide sequence derived from this UTCZP found that several ESTs originating from human pancreas would exhibit a high degree of similarity to this protein and he speculated that they could represent the human counterpart to UTCZP 146. The overall conclusion of this study was that UTCZP may play an

Figure 1.4 | A general mechanism for assembly of ZP domain proteins. In all ZP containing proteins, the precursor ZP domain is followed by a long hydrophobic region a conserved furin cleavage site, a transmembrane region and an intracellular c-terminal tail. C-terminal processing at the conserved furin cleavage site by a furin convertase releases them to the extracellular space leading to their assembly into fibrils and matrices. (Reproduced from 145).

33 important role in events that transpire during late pregnancy. A different group showed that UTCZP is also expressed in the mouse pancreas. In their study published in 2002 Imamura et al. named this gene Itmap1 and they demonstrated that both in pregnant uterus and pancreas the translated protein is localized to cytoplasmic granules 147.

One year later, Chen et al reported the isolation of a novel gene, 82% homologous with UTCZP, from rat uterus 138. The authors demonstrated that the expression of this gene in the uterus, which was then name estrogen regulated gene 1 (ERG1), is under estrogenic control and that this gene is also highly expressed in the oviduct and that its expression is restricted to the surface epithelium of both tissues 138. ERG1 mRNA levels increased in the in the uteri of ovariectomized or immature rats upon administration of estrogen and are abruptly decreased upon the administration of antiestrogens 138. Additionally, spatio-temporal analyses of ERG1 expression in the endometrium confirm that it is expressed in a highly stage-specific manner in the uterus and oviduct 138. The ERG1 mRNA levels are high on day 1 of pregnancy, declined on day 2, and were practically undetectable from days 3 to 6 of gestation. During the ovarian cycle, ERG1 is expressed in the uterus in a manner that coincides with the estrogen-induced uterine cell proliferation 138. Another important finding of their study was that ERG1 mRNA levels in the uterus are suppressed in the presence of progesterone and they assumed that ERG1 expression in the uterus is regulated by a complex interplay of estrogen and progesterone 138.

In 2001, Huyhn et al, by using differential display techniques, isolated a tamoxifen- and estrogen-induced complementary DNA from rat uterus cDNA library 139. This gene, which was designated as (UO-44) due to its specific expression in rat uterus and ovaries, shares approximately 87% homology with mouse UTCZP (88% at the aminoacid level) and 99% homology with rat ERG-1. Unlike UTCZP which is expressed only in the uterus during late pregnancy, and similarly to the practically identical ERG1, UO-44 is expressed in nonpregnant and pregnant uteri of mature female rats 139. Unlike ERG1, Hyuhn et al. showed UO-44 mRNA is also detected in granulose cells of ovaries 139. This finding was not in accordance with the report previously published by Chen et al. where ERG-1 was not detected in the rat ovary using Northern blotting. The author speculated that the discrepancy between the two studies could be attributed to differences in the stage of the estrous cycle where the ovaries are collected for the analysis or usage of a different part of the ovary for RNA extraction 139. Another similarity between ERG1 and UO-44 is that they both are induced in rat uterus and oviduct by estradiol

34 treatment and their mRNAs are detected specifically at the surface epithelium 139. UO-44 mRNA levels rapidly decreased in the rat uterus after ovariectomy, but they were restored upon administration of estradiol, in a dose dependent manner 139. Likewise, tamoxifen and GH induced the expression of UO-44 in the uterus of ovariectomized and hypophysectomized rats, while pure 139 antiestrogen ICI 182780 (Faslodex) was inhibitory . Three years later the same group reported that the human ortholog of UO-44 (Hu-UO-44), later named CUZD1, was highly expressed in the pancreas. The author attributed the differences in the CUZD1 expression profile between human and rat to the cell type specific nature of UO-44 139. In the same study the author also claimed that CUZD1 protein was specifically detected at the epithelial cells of normal ovarian and ovarian epithelial cancer tissue sections 139.

In a more recent study Mapes et al. demonstrated that the mouse ortholog of CUZD1 (previously named as Itmap1 or UTCZP) is expressed in the mammary gland of mice 148. According to their data which were obtained by performing immunofluorescence (IF) analysis of CUZD1 in the mammary tissue of CUZD1 +/- mice, the protein was detected in the developing ductal epithelium at puberty and in both cytoplasmic and nuclear compartments of the ductal and alveolar epithelial cells during alveologenesis at late pregnancy whereas strong nuclear staining was observed during lactation 148. On the other hand, no detectable expression in mammary tissue of Cuzd1-/- mice was observed during development 148.

1.3.3 Cellular localization of CUZD1

The presence of a secretory signal peptide at the amino-terminal and a trasmembrane region near the carboxy-terminal aminoacid sequence indicate that CUZD1 is a membrane-bound protein. Indeed, independent research groups have experimentally shown the membrane-associated nature of CUZD1 141, 147.

To define the cellular distribution of CUZD1 protein, Huyhn et al transfected human breast cancer MCF7 cells with a mammalian expression vector containing full length CUZD1 cDNA fused to a 6-histidine tag 139. By performing Western Blot analysis using an antibody against the 6-histidine tag, they showed that CUZD1 was present in the plasma membrane-enriched subcellular fractions isolated from the cells transfected with the plasmid encoding CUZD1, whereas it was not detected in the cytosol 139.

In a similar approach, Leong et al transfected ovarian cancer cells ATCCNIH- OVCAR3 with a mammalian expression vector containing the open reading frame (ORF) of HuUO-44D fused to the ORF of green fluorescent protein (GFP) 141. According to their data, CUZD1 expression was observed to the cytosol 24 h post-transfection and was located on cell membrane 48 h after transfection, but excluded from the nucleus 141.

Imamura et al. went one step further by demonstrating the membrane-associated nature of the endogenously expressed mouse ortholog of CUZD1 in pancreas and uterus 147. They utilized in- house generated rabbit polyclonal antiserum raised against CUZD1 to perform immunofluorescence studies in pancreatic and uterine tissue sections from CUZD1 deficient and wild type mice and they observed that the antiserum specifically stained the acinar cell population of the pancreas and the zymogen granules 147. The localization of CUZD1 in the zymogen granules was further confirmed with immunoelectron microscopy. Additionally, they mentioned that in the uterus, the anti-CUZD1 antiserum also stained granular structures in the apical region of the epithelium but they did not provide further details 147. Western Blot analysis of preparations of total membranes and of zymogen granule enriched membranes derived from both wild type and CUZD1 deficient mice showed a positive signal only in the cellular membrane preparations from wild type pancreas and uterus. Similarly, Western Blots analysis of alkali-extracted zymogen granule enriched membranes further demonstrated that CUZD1 is tightly associated with zymogen granule membranes 147.

On the other hand, in their recently published work, Mapes et al. are the first to report cytoplasmic and nuclear localization of CUZD1 148. When mammary epithelial cells, generated to stably over-express recombinant CUZD1 tagged with a FLAG epitope, are grown in the absence of serum, CUZD1 is predominantly localized in the cytoplasm 148. Culture of these cells with media supplemented with serum directs translocation of CUZD1 to the nucleus. Because of the absence of a nuclear localization motif or a DNA binding domain from the CUZD1 structure the authors demonstrated that the nuclear translocation of CUZD1 relies on the association with a transcription factor.

1.3.4 Proteins similar to CUZD1

DMBT1

Deleted in malignant brain tumors (DMBT1) gene is located at 10q25.3-26.1, a locus just upstream of CUZD1 gene, and loss of its expression has been reported in brain, gastrointestinal and lung cancers 149. DMBT1 has been shown to be expressed in different organs and cell lines related to the gastrointestinal and immune system and its structure and location suggest a mucin- like function in the digestive tract and associated glands 150. According to previous studies, DMBT1, alone or in coordination with the lung mucosal collectin surfactant protein-D (SP-D), interacts with/bind and agglutinate various microorganisms, at mucosal surfaces and direct/promote their clearance from the body 150.

The genomic structure of CUZD1 is very similar to parts of DMBT1 gene 140 (figure 1.5). The significant degree of homology between the two genes and the respective protein products is possibly justified by the similarity between the CUB and ZP domains contained in both. The

Figure 1.5 | Amino acid comparison between human CUZD1 and DMBT1. There is a high degree of homology between the two proteins possibly reflecting the similarity between the

CUB and ZP domains contained in both. The comparison has been performed using the BLASTp2 protein-protein programme. Top row: CUZD1; bottom row: DMTB1. (Reproduced from 140). close proximity and structural similarities between DMBT1 and CUZD1, suggest that CUZD1 could be a product of gene duplication or vice versa 141.

It has been reported that when duplicated genes are preserved, the two copies have either distinct functions (neofunctionalisation) or they share the function of the ancestral gene (subfunctionalisation) 151. According to in-house generated unpublished data, where we compared CUZD1 and DMBT1 mRNA expression data generated by the Cancer Cell Line Encyclopedia (CCLE) project 152, CUZD1 and DMBT1 are differentially expressed in human cancer cell lines/ the mRNA expression data of CUZD1 is distinct from that of DMBT1. Based on the above we could hypothesize that CUZD1 and DMBT1 share some common function and their simultaneous expression in a cell is redundant. If this is the case, CUZD1, under certain conditions, could potentially bind and mediate the clearance of microorganisms and thereby also have a mucinous-like function.

GP2

GP2 is reported to be the most abundant protein in pancreatic acinar cells and is not detected in the duct or in the islets of Langerhans 153. It is a heavily glycosylated protein with N-linked carbohydrates, linked to the zymogen granule membrabe via a glycosyl phosphoinositol (GPI) anchor, and it accounts for half of all zymogen granule membrane proteins 154-156. As mentioned in previous chapters, more recently it has been reported that GP2, both at the mRNA and protein levels, is also expressed in Μ cells of follicle-associated epithelium (FAE) of Peyer's patches 125, 131.

One of the proposed functions for GP2 on M cells is to act as a bacterial uptake receptor for bacterial antigens 125, 131. Hase et al have demonstrated that recombinant GP2 can selectively bind to a subset of commensal and pathogenic enterobacteria and that this interaction is mediated by bacterial adhesin FimH 135. Thus, because of the specific expression of GP2 on intestinal M cells, it is believed that it plays a pivotal role in the generation of immune response to molecules of the intestinal content.

GP2 and CUZD1 share more similarities apart from abundantly expressed on the zymogen granules of the pancreatic acinar cells. They are both characterized as membrane proteins, but they are also components of pancreatic juice. The two proteins also have a structural characteristic in common; they harbor a C terminal ZP domain which, as mentioned earlier, functions as conserved module for the polymerization of extracellular proteins. Additionally, it seems that the two proteins have some similar binding capabilities, since it has been recently demonstrated by Komorrowski et al 157, they both interact with the lectins UEA and SBA. Lastly, CUZD1 and GP2 have recently identified as the targets of PABs which are often encountered in patients with IBD 125, 131, 133. This could indicate that the two proteins could also co-localize in the inflamed intestine too, apart from the exocrine pancreas and the pancreatic juice. More specifically, it could be speculated that, following their secretion to the pancreatic juice, and after the migration of this fluid to the inflamed intestine, the two pancreatic proteins arrive to the inflamed intestine and, for an unknown reason, they are targeted by the host’s immune system. In the case of GP2 it has been suggested that ZP domain causes the protein to aggregate in the pancreatic juice and that through this process it might trap bacteria and prevent their adhesion to mucosal cells 131, 157. Based on those structural, biochemical and spatial similarities it is tempting to speculate that GP2 and CUZD1 might also share a common function. More specifically, if GP2 protein interacts with the bacterial-like lectins FimH and both CUZD1 and GP2 bind to lectins UEA and SBA then CUZD1 might also be able to recognize this bacterial region and thereby get entangled with the respective bacteria. Since it also harbors a ZP domain, CUZD1 protein could aggregate in pancreatic juice and, similarly or in coordination with GP2, it could entrap bacteria in the intestine. As in the case of GP2, this interaction could be important for bacterial triggered immune responses by the host.

1.3.5 Function of CUZD1 (based on experiments)

In 2004 Leong et al were the first to publish data connecting CUZD1 to ovarian cancer 141. Following work that was pursued at the same laboratory which demonstrated that the rat ortholog of CUZD1, UO-44, was specifically expressed in the uterus and ovaries under the regulation of estrogens, Leong reported that CUZD1 mRNA is overexpressed in a majority of ovarian tumors 141. One key experiment that further suggested an important role for CUZD1 in carcinogenesis demonstrated that human UO-44 antisera strikingly inhibit cell attachment and proliferation of NIH-OVCAR3 ovarian cancer cells 141. This observation has led the authors to

39 postulate/speculate that human CUZD1 could be involved in cancer cells attachment and proliferation and that it could possibly promote cell growth in ovarian cancer 141. In a later study, the same group of investigators reported that treatment of ovarian cancer cells with several chemotherapeutic drugs with cytotoxic mode/mechanism of action involving DNA damage resulted in the suppression of CUZD1 expression 158. In accordance with their previous work, they reported that silencing the expression of CUZD1 with the usage of small interfering RNA (siRNAs) led to inhibition of cell growth and proliferation and more importantly to a significant sensitivity to cis-platin treatment in a dose dependent manner 158. Finally, their data demonstrated that overexpression of CUZD1 in ovarian cancer cells conferred resistance to cisplatin chemotherapy 158.

CUZD1 protein has also been linked to pancreatitis. To better understand the mechanisms giving/conferring resistance or sensitivity to pancreatitis Imamura et al analyzed novel acinar cell proteins 147. The mouse ortholog of CUZD1, designated in their study as Itmap, was found to be prominently expressed in pancreatic acinar cells and was particularly localized to zymogen granule membranes 147. They generated CUZD1-deficient mice and they compared the effects of secretagogue- and diet-induced pancreatitis on those and on wild type mice. The authors observed that CUZD1 defect does not modify the size, appearance or the composition of zymogen granules and that CUZD1 deficient mice had significant less trypsinogen activation than CUZD1 wild type mice 147. It was further shown that CUZD1 changes the progress of acute pancreatitis triggered by cerulein hyperstimulation or by a choline-deficient, ethionine- containing (CDE) diet 147. More specifically, while cerulein-induced pancreatitis caused no mortality in either CUZD1 deficient or CUZD1 wild type mice, it did lead to increased severity of pancreatitis manifested by greater edema, a mild increase in necrosis, significantly increased numbers of apoptotic acinar cells and a statistically significant elevation in basal lipase concentration in the CUZD1 deficient mice compared to the wild type mice. In the CDE diet model, the effect of CUZD1 deficiency was larger, as the mortality levels of CUZD1 null mice increased almost 7-fold. The authors concluded that CUZD1 plays a pivotal role in trypsinogen activation and in the severity of pancreatitis 147.

In a more recent study Mapes et al. reported novel findings regarding the in vivo function of the mouse ortholog of CUZD1 in the signaling pathway that are involved in mammary gland development during pregnancy 148. The authors created CUZD1 deficient mice and following the

40 analysis of the spatio-temporal expression of CUZD1 in the mammary glands during development which indicated that the protein is present in the developing ductal epithelium at puberty, at the ductal and alveolar epithelial cells of mammary glands during alveogenesis at late pregnancy and during lactation/that the protein could have a role during mammary gland development, particularly during pregnancy and lactation they observed that although the CUZD1 deficient mice were fertile they produced inadequate amount of milk 148. Comparison of the phenotypic characteristics between the CUZD1-/+ mice and CUZD1 -/- mice collectively revealed an impairment in alveolar differentiation in the latter ones which could be responsible for the deficiency in milk production 148. There was a significant reduction in the number of mammary ductal epithelia of non-pregnant pubertal CUZD1 -/- mice and a dramatic decline in the lactating mammary gland indicating that CUZD1 could be involved in pathways regulating mammary epithelial proliferation 148. Microarray analysis of mammary epithelial cells derived from CUZD1 +/- and CUZD1 -/- mice revealed lower mRNA levels of three members of the EGF family, neuregulin (Nrg1), epiregulin (EPR) and epigen (EPGN) which suggested that CUZD1 regulates the expression of a specific subset of EGF family ligands in the mammary epithelium during late pregnancy and further experiments indicated that CUZD1 regulates the proliferation of mammary epithelial cells by modulating the ErbB signaling pathway 148. In their study they also provided data showing that CUZD1 is a novel regulator of STAT5 signaling in the steroid-induced mammary epithelium 148. Immunoprecipitation of CUZD1 from mammary epithelial cells followed by mass spectrometric and Western Blot analysis demonstrated that it participates in protein complex composed of, among other proteins, JAK1/JAK2 and STAT5 148. Although previous studies have given evidence that CUZD1 is associated with the membranes, here it was shown that CUZD1 is localized in both cytoplasmic and nuclear compartments of mammary epithelial cells and that CUZD1 translocation to the nucleus relies on interaction with transcription factor Pstat5 148.

1.3.6 CUZD1 autoantibodies in patients with IBD

CUZD1 was first identified as an antigenic target of PABs in an abstract in 2008 and more recently in a published article by Komorrowski et al 157. In a series of experiments the authors demonstrated that the supernatant of homogenized human pancreas entirely abolished PAB reactivity 157. To isolate the glycoproteins from the pancreas, they firstly screened various lectins for their ability to immobilize PAB-positive glycoproteins from cell-free pancreas 157. Ulex

41 europaeus agglutinin I (UEA), the most efficient in this task lectin, was then employed as a solid phase reagent to isolate the pancreatic glycoproteins 157. The UEA-eluate, which was able to neutralize both the reticulogranular and the droplet-like pattern in IIF, was subjected to mass spectrometric analysis and GP2 and CUZD1 were identified as the target autoantigens. Both glycoproteins were separately expressed in mammalian cells and the recombinant cells applied as substrates in IIFT 157. The reaction with the GP2 expressing cells was strictly correlated with the droplet pattern, while the autoantibodies generating the reticulogranular pattern stained the CUZD1 producing cells. Finally, absorption experiments demonstrating that CUZD1 can completely neutralize antibody reactivity of PABs associated with the reticulogranular pattern further confirmed that CUZD1 is the only autoantigen related to this pattern 157.

Kommorowski and colleagues by using recombinant cells expressing CUZD1 as a substrate in IIF detected anti-CUZD1 positivity in 19.8% of patients with CrD 157. Anti-CUZD1 antibody reactivity was strongly correlated with the reticulogranular PAB pattern. In their article the authors presented the development of an antigen capture ELISA using monoclonal antibodies against CUZD1 which confirmed this relationship 157. Roggenbuck et al have reported for CUZD1 autoantibodies a prevalence of 29.2 % in patients with IBD and in particular 22.6% and 14.9% of patients with CD and UC, respectively 131. In their published work, Michaels et al in addition to reporting a prevalence for CUZD1 autoantibodies of 25.0 % and 5.9 % for patients with CrD and UC, respectively, they provided information/data regarding their clinical relevance in patients with IBD 159. According to their data anti-CUZD1 positivity is associated with ileocolonic and perianal disease 159. In a more recent study Pavlidis et al followed the same novel cell-based immunofluorescence approach to assess the clinical utility of anti-CUZD1 (and anti- GP2) antibodies in patients with IBD 132. CUZD1 autoantibodies were detected in 21,7 % of CrD and 9.2 % UC patients but, in this study, their prevalence was not correlated with clinical characteristics of the disease 132.

1.4 Rationale and Objectives

1.4.1 Rationale

CUB and zona pellucida like domains-containing protein 1 (CUZD1) is a protein abundantly expressed in the exocrine pancreas. CUZD1 has been previously linked to pancreatitis, ovarian cancer and IBD and due to its specific expression, it has been proposed as a candidate biomarker for pancreatic related disorders. Because of the lack of reagents specifically detecting CUZD1 antigen, it is currently unknown if this antigen can be used for the detection of the aforementioned or other disorders. CUZD1 protein is selectively expressed in the pancreatic acinar cells but it is unknown if its expression is elevated in malignant conditions of the pancreas. In this thesis, I opted to firstly develop analytical technology for CUZD1 protein, validate it and used it to investigate if CUZD1 has a utility for the detection of the aforementioned or other disorders.

CUZD1 has been recently identified as target autoantigens of PABs which have emerged as potential biomarkers for the differential diagnosis of CrD over UC. Due to the lack of high throughput techniques for the detection of CUZD1 autoantibodies, their diagnostic significance in patients with IBD has not been extensively studied. Furthermore, data regarding the expression of CUZD1 antigen in the site of inflammation in IBD are lacking raising the question of the mechanism that evokes the generation of immune response against this pancreatic protein in the context of IBD. In this thesis I investigated if CUZD1 antigen is expressed in inflamed tissue sections from patients with IBD. Furthermore, I developed assays targeting CUZD1 autoantibodies and used them for the analysis of serum samples from patients with IBD.

1.4.2 Hypothesis

I hypothesize that the development of analytical tools for the detection and quantification of CUZD1 protein will allow for the investigation of the clinical utility of CUZD1 as a serum biomarker in malignant and nonmalignant human diseases.

1.4.3 Objectives 1) Develop novel immunoassays for detection of CUZD1 autoantibodies in serum of patients with inflammatory bowel diseases

 Produce recombinant CUZD1 protein and purify it.  Develop ELISAs targeting CUZD1 autoantibodies using recombinant CUZD1 protein as a solid phase antigen.  Validate the ELISAs using samples from patients with IBD, from patients with irrelevant to IBD disorders and healthy donors.

2) Establish the tissue expression pattern of CUZD1 protein and assessing its levels in various biological fluids.  Use purified recombinant protein as an antigen for the generation of monoclonal antibodies targeting CUZD1  Use monoclonal antibodies to immunohistochemically localize CUZD1 protein in normal pancreatic tissue.  Use monoclonal antibodies to develop an ELISA measuring CUZD1 and determine CUZD1 protein levels in a variety of human tissue extracts and biological fluids.

3) Investigating the clinical utility of CUZD1 in malignant and non-malignant human diseases.  Determine CUZD1 protein levels in a large number of serum samples from patients with various malignant and nonmalignant disorders, including PDAC, ovarian cancer, pancreatitis and IBD.  Perform IHC analysis of inflamed tissues from patients with IBD, in order to investigate if CUZD1 protein is expressed in the site of inflammation.  Perform IHC analysis of ACC and PDAC biopsies to investigate if CUZD1 can be detected in any of these malignant conditions of exocrine pancreas.

2 Chapter 2 | Novel immunoassays for detection of CUZD1 autoantibodies in serum of patients with inflammatory bowel diseases

This chapter includes data published in the Clinical Chemistry and Laboratory Medicine: “Novel immunoassays for detection of CUZD1 autoantibodies in serum of patients with inflammatory bowel diseases” by Farkona S., Soosaipillai A., Filippou P., Liaskos C., Bogdanos D.P., Diamandis E.P., Blasutig I.M.

A link to the published paper can be found at

doi: 10.1515/cclm-2016-1120

SF performed the experiments and drafted the manuscript. AS assisted with the development of the methodology. PF assisted with the protein production. CL and DPB provided the samples. EPD and IMB designed the study and edited the manuscript.

44 45

2.1 Introduction

IBD are heterogenous disorders that affect the gastrointestinal tract via an immune-mediated inflammatory process. The two main forms of IBD are UC, which is limited to the colon and CrD, which can affect any segment of the gastrointestinal tract from the mouth to the anus 116.

The diagnosis of IBD is based on clinical evaluation and a combination of endoscopic, histological, laboratory and imaging-based investigations 121. Distinguishing between CrD and UC can be difficult since some patients present with clinical and laboratory findings that are common to both diseases. Since the clinical management and therapies of CrD and UC can be distinct, particularly in the case of surgical intervention, confirmation of the condition in question is crucial 122.

Having the added benefit of being minimally invasive and economical, several serological biomarkers have been tested for their usefulness in diagnosis, disease stratification and prognostication of IBD 123. The abnormal immune response against endogenous or microbial antigens in genetically susceptible individuals is involved in the pathogenesis of IBD, and for this reason serological antibodies have been tested as potential biomarkers 160.

The most widely described antibody markers are ASCA and pANCA. Although they have been shown to have value in distinguishing patients with CrD from patients with UC, and to be related to disease behavior, serum antibodies have a limited clinical value in the primary diagnostic work up of IBD mainly due to their poor sensitivity and moderate specificity 161.

PABs against the acinar cells of the exocrine pancreas are encountered in the serum of IBD patients 124. PAB detection has relied on IIF, which gives two distinct patterns: an extracellular droplet-like staining in the acinar lumen and a reticulogranular cytoplasmic pattern of the pancreatic acinar cells 125. The diagnostic and clinical utility of PAB in IBD, as revealed by IIF, have been partly investigated but the published data are controversial. According to some reports PABs are present in up to 42% of patients with CrD and in less than 5% of patients with UC or other pathological conditions and healthy controls 126, 127. Other studies show that IgA PAB can be encountered in approximately 23% of patients with UC and in up to 13% of patients with celiac disease 125. In addition, some studies found that CrD patients that are positive for PAB have more frequent extra-intestinal manifestations as compared with PAB-negative patients 162,

163. Data failing to report associations between PABs and clinical features of the disease have also been published 128. Discrepancies amongst the reported data may be to due differences related to the IIF methodology, study cohorts or differences in ethnic background 125.

The targets of PABs have been recently identified. The pancreatic major glycoprotein 2 (GP2) of the zymogen granule is the target antigen associated with the droplet type pattern 133 while CUB and zona pellucida-like domain-containing protein 1 (CUZD1) is the target of PABs giving the the reticulogranular type pattern 157. The development of immunoassays using recombinant human GP2 as a solid phase antigen has provided information on the prevalence and diagnostic significance of anti-GP2 antibodies in patients with IBD 125. On the other hand, data assessing the diagnostic significance of CUZD1 autoantibodies in patients with IBD are still scarce, mainly due to the lack of an ELISA) which can be easily used for the proper detection of CUZD1 autoantibodies.

In this study, we describe the development of a novel high throughput ELISA-type immunofluorometric assay for the detection of CUZD1 autoantibodies in the serum of patients with IBD and other pathological conditions. We produced, purified and coated microwell plates with recombinant protein, to identify CUZD1 autoantibodies in human serum and determined appropriate cutoff values for positivity using 91 samples from apparently healthy individuals. Using this assay, we analyzed the serum of 200 patients with IBD (100 patients with CrD and 100 patients with UC), 50 additional control individuals and 129 patients who were assessed for various autoimmune diseases (vADs) as pathological controls.

2.2 Materials and Methods

2.2.1 Patients

All samples were collected with Research Ethics Board (REB) approval. Serum samples from 91 apparently healthy individuals were collected from the Toronto General Hospital Clinical Biochemistry Laboratory. They were analyzed using our immunofluorometric assays, to establish a cutoff for positivity.

Two hundred serum samples from patients with IBD (equally divided between CrD and UC) and 50 serum samples from healthy individuals and patients with diseases unrelated to IBD were provided by one of the co-authors (DPB). Although, these samples had been previously analyzed for CUZD1 autoantibodies by IIF by two investigators (DPB and CL), those data were blinded. Seven unblinded IBD samples with known seropositivity for CUZD1 autoantibodies, were also provided as positive controls. Table 2.1 shows the main demographic, clinical and laboratory characteristics of the patients with Crohn’s disease (CD), ulcerative colitis (UC) and control individuals included in the present study.

Serum samples from 129 patients submitted for testing for various autoantibodies were collected from the Toronto General Hospital Clinical Biochemistry Laboratory. Of the 129 samples, 36 were tested for ANCA (14 anti-MPO positive, 7 anti-PR3 positive, 6 positive for both and 9 negative for both), 29 samples were tested for anti-cyclic citrullinated peptide (anti-CCP; 15 positive and 14 negative), 26 samples tested for celiac antibodies (5 anti-tTG IgA positive, 1 anti-DGP IgG positive and 20 negative), 30 tested for anti-nuclear antibodies (ANA; 30 positive) and 8 samples tested for both ANCA and ANA (1 anti-MPO and –PR3 positive, 1 anti-MPO positive, 3 ANA positive, 1 positive and 2 negative for all three). Samples were stored at -20 °C until analysis.

2.2.2 Production of recombinant CUZD1

Recombinant CUZD1 protein was produced in the Expi293 mammalian protein expression system (ThermoFisher Scientific, Invitrogen, Toronto, Canada). pcDNA 3.4-TOPO plasmid harboring cDNA encoding for the extracellular region of CUZD1 (aminoacids 1- 568 of NCBI GenBank accession number NP_071317) was used for the transient transfection of Expi293F

48 cells. CUZD1 cDNA carries the native secretion signal peptide of the protein and it was synthesized using Gene Art synthesis under optimal conditions (Invitrogen).

Table 2.1 | Main demographic, clinical and laboratory characteristics of the patients with Crohn’s disease (CD), ulcerative colitis (UC) and control individuals included in the present study.

CD (n=100) UC (n=100) Control* (n=50)

Sex (male/female) 43 (43%) /57 (57%) 47 (47%) /53 (53%) 19 (38%) /31(62%)

Age at sample date (mean 38.03 ± 13.91 48.59 ± 15.63 48.54 ± 18.72 years ± SD)

Age at diagnosis (mean years 24.87 ± 11.23 31.56 ± 14.02 ± SD)

Disease duration (mean years 13.16 ± 8.92 17.03 ± 13.77 ± SD)

Location or extent (%) L1: 16 (16%) E1: 11 (11%)

L2: 22 (22%) E2: 25 (25%)

L3: 62 (62%) E3: 64 (64%)

Behaviour (%) B1:49 (49%)

B2: 27 (27%)

B3: 24 (24%)

Perianal: 30 (30%)

Age (%) A1: 17 (17%)

A2: 73 (73%)

A3: 10 (10%)

PAB pattern reticulogranular 9 (9%) /91 (91%) 7 (7%) /93 (93%) 0 (0%) /50 (100%) IgA≠ positive/negative

Anti-CUZD1 by IFA≠≠ 9 (9%) /91 (91%) 7 (7%) /93 (93%) N/A IgA positive/negative

PAB pattern reticulogranular IgG≠ 19 (19%) /81 (81%) 5 (5%) /95 (95%) 0 (0%) /50 (100%) positive/negative

Anti-CUZD1 by IFA≠≠ 19 (19%) /81 (81%) 5 (5%) /95 (95%) N/A IgG positive/negative

*50 control samples (composed of 32 samples from patients with diseases unrelated to IBD and 18 samples from additional healthy individuals).

≠ pancreatic tissue as substrate, ≠≠recombinant cells expressing CUZD1 as a substrate in IIF.

L (location), B (behaviour) and A (age) according to Montreal classification; L1, ileal; L2, colonic; L3, ileocolonic and L4, upper disease modifier; B1, non-stricturing /non-penetrating; B2, stricturing; B3, penetrating behavior; p perianal disease modifier; A1 below 16 y, A2 between 17 and 40 y, A3 above 40 y; E1, Involvement limited to the rectum (that is, proximal extent of inflammation is distal to the rectosigmoid junction) ; E2, Left sided UC (distal UC): Involvement limited to a proportion of the colorectum distal to the splenic flexure; E3, Extensive UC (pancolitis): Involvement extends proximal to the splenic flexure ; SD, standard deviation.

Expi293F cells were maintained according to manufacturer’s guidelines. For transfection, cells were diluted in prewarmed Expi293™ Expression media to a final cell density of 3x106 cells/ml in 25.5 mL of a 125-mL flask. For each mL of transient transfection volume, 1 μg of expression vector and 2.7 μL ExpiFectamine 293 reagent were used, each diluted in 1.5 ml of Optimem. These were combined and incubated at room temperature for 20-30 minutes, then added drop- wise to the 25.5 mL of Expi293F cells. Following a 24-hour incubation, a mixture of enhancers (150 μL of ExpiFectamine™ 293 Transfection Enhancer 1 and 1.5 ml of ExpiFectamine™ 293 Transfection Enhancer 2) was added to each flask (final volume: 30 ml) and the cells were incubated for 6 more days at 125 rpm, at 8% CO2 and 37 °C.

The expression of the protein was tested by Western Blot, using a commercially available anti- CUZD1 peptide antibody (Santa Cruz Biotechnology, Daglas, TX, USA), using media from the culture every 24 hrs during the incubation period. The amount of the recombinant protein was quantified by SRM mass spectrometric analysis 164.

Following protein production, the cell culture supernatant was harvested and concentrated 10- fold. Recombinant CUZD1 was purified using anion exchange chromatography in 50mM Tris-

HCl pH 9 using a MonoQ column at a flow rate of 1mL/min. The column was then briefly washed with the same buffer and recombinant CUZD1 was eluted in 4 mL fractions with a linear gradient from 0 to 500 mM NaCl over 50 min, followed by a linear gradient from 500mM to 1M NaCl over 5 min.

2.2.3 Selected Reaction Monitoring (SRM)

In the first step of method development 4 unique proteotypic peptides were chosen to be tested for the development of a CUZD1 SRM method using SRM atlas (http://www.srmatlas.org). These doubly charge peptides were also confirmed in our in-house LC-MS/MS identification data and were analyzed with the Basic Local Alignment Search Tool (BLAST) to ensure that they were unique to the protein of interest. In silico protein digestion and peptide fragmentation were performed with Skyline software and initially seven transitions, chosen using SRM Atlas, were monitored for each peptide. Transitions with fragment m/z higher than preqursor m/z were preferable, but transitions with lower m/z were not excluded if had high intensity and low noise. The method was tested in supernatant samples derived from the culture of mammalian cells expressing and secreting recombinant CUZD1. Retention times, relative intensities of peptides and the most intense and selective transitions per peptide were recorded at that step. The final SRM method was developed to monitor 3 transitions of the highest intensity and the lowest noise from peptide ASPTSDFASPTYDLIK and its 13C6, 15N2 lysine-labeled counterpart. For each transition dwell time was set to 145 msec and was estimated to measure 15 points per LC peak.

Samples for SRM analysis were processed for trypsin digestion under standard procedures. Total protein of each sample was measured by the Bradford total protein assay, and the volume was adjusted to obtain equal amounts of total protein from the individual samples. Following this, a fixed amount of isotopically labeled peptide was added to each sample. Following reduction and alkylation proteins were digested with trypsin overnight (trypsin to total protein ratio 1:10 (w/w)), acidified with 1% formic acid and centrifuged at maximum speed for 10 minutes. Peptides from the supernatants were purified and extracted using OMIX C18 tips and eluted using 5 uL of 65% acetonitrile solution (0.1% formic acid). Peptides were diluted with 60 uL of water (0.1% formic acid) and SRM was performed as described previously 164.

2.2.4 Chromatographic and MS conditions

Tryptic peptides were separated on a 2-cm trap precolumn [200 um inside diameter (ID), 5 um C18]. Eluted peptides from the trap column were subsequently loaded onto a 15 cm long resolving column (75 um ID, 3 um C18) with an 8-um tip (New objective, Woburn, MA). The trap and resolving columns were operated on the EASY-nLC system (Thermo Fisher, Odense, Denmark) coupled on line to a triple-quadrupole mass spectrometer (TSQ Quantiva, Thermo Scientific, Germany) using a nano-ESI (Thermo Fisher). A three-step 30 min gradient with an injection volume of 18 uL was used. Buffer A contained 0.1% formic acid in water, and buffer B contained 0.1% formic acid in acetonitrile. The eluted peptides were subjected to SRM analysis with polarity set to positive, collision induced dissociation (CID) pressure in the second quadrupole set to 1.5 mTorr and ion transfer tube at 300 °C. The resolution for the first quadrupole was set to 0.2 FWHM and the third quadrupole to 0.7 FWHM.

2.2.5 Data analysis

Raw files recorded for each run were uploaded to Skyline software (Mac Coss Lab Software, Seattle, Wa, USA) for quantification. Peak areas generated by the analyte (light) peptide were empirically verified and normalized by the peak areas generated by the isotopically labeled counterpart (heavy). The ratio AUClight/AUCheavy was multiplied by the amount of isotope- labeled peptide added in the sample to estimate the amount of each native peptide.

2.2.6 Development of a fluorometric anti-CUZD1 autoantibodies immunoassay

White polystyrene microtiter plates were coated overnight at room temperature with 100 uL of purified recombinant CUZD1 in coating buffer (50 mM Tris buffer, pH 7.8) in a final concentration of 1 ng/uL. The wells were then washed x3 with washing buffer (20 mM Tris, 150 mM NaCl, 0.05% Tween-20, pH 7.4) on an automatic washer and blocked for 60 min with 60 g/L BSA. Following washes, the wells were incubated for 2 hours with serum samples diluted 500-fold in Assay Buffer (60 g/L BSA, 1 g/L goat globulin, 0.2 g/L mouse globulin, 10 g/L bovine globulin, 5 mL/L Tween 20, and 37 g/L KCl), washed x3 and incubated for 1 hour at room temperature with 100 uL of ALP-conjugated-anti-human IgG or IgA (Jackson

Immunoresearch), diluted 20,000 or 10,000-fold in Assay Buffer, respectively. After washes, 100 ul of diflunisal phosphate solution (DFP) [0.1 M Tris HCl buffer (pH 9.1) containing 1 mM diflunisal phosphate, 0.1 M NaCl, and 1 mM MgCl2] diluted 20-fold in substrate buffer (0.1 M

Tris, pH 9.1, 0.1 M NaCl, 1 mM MgCl2), was added to each well and incubated for 10 min. The reaction was stopped by adding developing solution (1 M Tris, 0.4 M NaOH, 2 mM TbCl3 and 3 mM EDTA) and incubating for 1 min. Fluorescence was measured with a Victor time-resolved fluorometer as previously described 165. Each assay run was done in parallel with a control run assessing nonspecific binding effects by using albumin as the solid phase antigen instead of recombinant human CUZD1. The ratio of the signal with recombinant CUZD1 versus the signal with BSA was used to determine if samples were positive. Cut off levels were determined by assessing the 91 healthy serum samples collected from the Toronto General Hospital. Seven serum samples, known to be positive for IgG CUZD1 autoantibodies by IIF, were measured with both our IgG and IgA specific assays. Then, they were pooled, diluted 500-fold in 6% BSA and Assay Buffer and used as positive controls in each run. 200 serum samples from patients with IBD and 50 samples from healthy individuals and patients with diseases unrelated to IBD, blinded for their CUZD1 autoantibodies status and 129 serum samples from patients tested for various autoimmune diseases were analyzed using our newly developed immunoassays.

2.2.7 Statistics

Statistical analysis was performed using SPSS software, version 15.0 (SPSS Inc., Chicago, IL). Normal distribution was evaluated using Kolmogoroff-Smirnoff test and by inspecting the histograms. The continuous variables (ELISA ratios) were not normally distributed and the Kruskal-Wallis test was used for comparison between more than two groups of continuous variables while Spearman’s rank correlation coefficient was used to assess the correlations between continuous variables. Fisher’s exact test was used for categorical variables.

2.3 Results

2.3.1 Production and purification of recombinant CUZD1

The cloned cDNA encodes the native secretory signal of the protein and, as anticipated, the recombinant protein was detected and quantified in the culture supernatant by Western Blot

(Figure 2.1) and SRM analysis, respectively. The theoretical molecular mass of the extracellular region of CUZD1 is 65 kDa but the expressed recombinant protein encoded by the cDNA of this region appeared as a band of approximately 70-80 kDa (Figure 2.1) and as diffuse bands of approximately 130-175 kDa, or larger. These bands were confirmed to represent CUZD1 by LC- MS/MS.

Figure 2.1 | Production of CUZD1 recombinant protein. Recombinant CUZD1 was detected by Western blot analysis using an anti-CUZD1 peptide antibody. Detection of the secreted extracellular fragment of CUZD1 in the supernatant of the Expi293F cells at: 24, 48, 72, 96 hours, 6 days (d6) and 7 days (d7) after transfection. CUZD1 protein reached its peak expression 72h after transfection. Lane 1 shows untransfected cell supernatants (unt). Molecular mass markers (KDa) are shown.

Figure 2.2 | CUZD1 quantification in culture supernatant by SRM. CUZD1 protein was quantified in the culture supernatant by comparing the peak areas generated by the analyte peptide ASPTSDFASPTYDLIK (A) to the peak areas generated by its isotopically labeled counterpart (heavy peptide) (B).

The analysis of the supernatant samples revealed that CUZD1 protein reached its peak expression 72 h after transient transfection (Figure 2.1, lanes 2-7). According to the SRM analysis, the yield at 72h was 53 mg/L.

The culture supernatant was applied onto an anion exchange column and CUZD1 was eluted in 200 mM NaCl fractions (Figure 2.2A) which were tested by Western Blot (Figure 2.2B) and Silver staining analysis (Figure 2.2C). Bands pointed by red arrow were confirmed as CUZD1 by LC-MS/MS (Figures 2.2B and C). Total protein was estimated in these fractions by Bradford Protein assay with bovine albumin as a calibrator and CUZD1 concentration was measured by SRM. By subtracting the amount of CUZD1 from the amount of total protein found, we were able to estimate that the approximate percentage of purity of CUZD1 antigen in the 200 mM NaCl fractions, which correspond to fractions 10 and 11, was approximately 50%. According to SRM analysis of the AIEX chromatography fractions, we obtained approximately 6 mg/L of semi-purified CUZD1 in culture.

Figure 2.3 | Purification of the protein fragment corresponding to the extracellular region of CUZD1 from supernatant of Expi293F cells by AIEX-FPLC. A) Elution chromatogram of the proteins present in the supernatant of the transfected cells. The recombinant protein was eluted in 200 mM NaCl fractions (peak inside the red box). B) Eluted CUZD1 was monitored by Western Blot of the FPLC fractions. C) Eluted CUZD1 was monitored by Silver Staining of the FPLC fractions. Bands pointed by red arrow were confirmed as CUZD1 protein by LC-MS/MS. For more discussion, see text.

2.3.2 Anti-CUZD1 immunoassays

The assays are based on a solid phase CUZD1 antigen immobilization approach that was carefully optimized for the amounts of reagents used, incubation times and dilution of serum samples.

Since no standards are available for CUZD1 autoantibodies, standard curves could not be defined. Each assay was run in parallel with an assay employing BSA as a solid phase antigen instead of recombinant CUZD1 protein. The ratio between the fluorescence in the presence of recombinant CUZD1 versus the presence of BSA was used as the read out. To establish a cutoff ratio, above which a serum sample would be classified as positive for CUZD1 autoantibodies, 91 serum samples derived from healthy individuals were analyzed with both the IgG and the IgA anti-CUZD1 assays. For the IgG specific assay, 90 of the 91 samples gave a ratio lower or equal to 2, while for the IgA specific assay, all samples gave a ratio lower than 1.5 (Figure 2.3 upper panel bottom panel). Although the cutoffs as determined by using the 99th percentile method were 1.8 for IgG and 1.3 for IgA assay, the values 2 and 1.5 were decided to be used instead, so as to set stricter criteria for positivity.

Each patient sample was run in duplicate to ensure consistent results. The variability observed between duplicates ranged from 0.03-22.15% for IgG and 0.01-23.50% for IgA. A positive control for anti-CUZD1 IgG was created by pooling patient samples and was run in duplicate on each plate. The mean and CV of this control was 13.2 +/- 14.4%.

8 Healthy Controls

uorescence Ratio uorescence Positive Controls

Fl 6

0 0 20 40 60 80 100 Number of Sample

1.6

1.4

1.2

0.8 Healthy Controls

Fluorescence Ratio Fluorescence 0.6

0.4

0.2

0 0 20 40 60 80 100

Number of Sample

Figure 2.4 | ELISA assay cutoff values. Ninety-one serum samples from normal individuals were measured using the developed immunoassay, to establish a reference cut-off ratio. The upper panel shows the values for IgG, while the bottom panel shows the values obtained for the IgA assay. For the IgG assay, a pool of positive serum samples was used as a positive control in every run.

2.3.3 Anti-CUZD1 antibodies in IBD patients

Twenty of 200 samples (10%) from IBD patients were found positive for IgG subtype of CUZD1 autoantibodies. IgA subtype of CUZD1 autoantibodies were detected in 14 out of 200 (7%) IBD samples (Figure 2.4). Of the 50 control samples (composed of samples from patients with diseases unrelated to IBD and samples from additional healthy individuals), none showed IgG antibody reactivity, while one showed borderline reactivity for the IgA class of CUZD1 autoantibodies (2%) (Figure 2.4 upper and bottom panel).

24 22 20 18 16 UC samples 14 12 CrD samples 10 Control samples 8 Positive controls Fluorescence Ratio Fluorescence 6 4 2 0 0 100 200 300

Number of Sample

18 16.5

15 13.5 12 10.5 UC samples 9

Fluorescence Ratio Fluorescence CrD samples 7.5 6 Control samples 4.5 3 1.5 0 0 50 100 150 200 250

Number of Sample Figure 2.5 | Analysis of serum samples. 100 Crohn’s Disease (red markers), 100 Ulcerative Colitis (blue markers) and 50 Control (green markers) samples we analyzed for anti-CUZD1 IgG antibodies (upper panel) or IgA antibodies (bottom panel) using our ELISA.

At least one type of anti-CUZD1 autoantibody was detected in 16% of patients with CrD (Figure 2.5A) and in 9% of patients with UC (p<0.05 Fisher exact test) (Figure 2.5B). IgG isotypes only were detected in 8% of the CrD patients and 3% of the UC patients (Figure 2.5A and B). IgA only isotypes of CUZD1 autoantibodies were found in 2% of CrD patients and in 3% of UC patients (Figure 2.5A and B). Both IgG and IgA CUZD1 autoantibodies were found in 6% of CrD patients (Figure 2.5A) and in 3% of the UC patients (Figure 2.5B).

Crohn’s Disease Samples

Anti-CUZD1 IgA

- + Total

Anti-CUZD1 - 84 2 86

IgG + 8 6 14

Totals 92 8 100

Ulcerative Colitis Disease Samples

Anti-CUZD1 IgA

- + Total

Anti-CUZD1 - 91 3 94

IgG + 3 3 6

Total 94 6 100

Figure 2.6 | A) Contingency table displaying the prevalence of IgG or IgA autoantibodies or both, in patients with A) Crohn’s disease (CrD). B) Ulcerative Colitis (UC).

Our ELISA results, from both the IgG and the IgA specific assay, after unblinding, were significantly correlated with the IIF results. The Spearman correlation coefficient between our data obtained by the IgG assay and the IIF results for the IgG autoantibodies was 0.555 (p< 0.01) for the CrD samples and 0.343 (p<0.01) for the UC samples. The Spearman correlation coefficient between the ratios obtained by the IgA ELISA and the respective IIF data was 0.435 (p< 0.01) for the CrD samples and 0.430 (p<0.01) for the UC samples. IgG CUZD1 autoantibodies were detected in 25 out of 200 IBD samples by the IIF assay and IgA autoantibodies in 6 out of 200 IBD samples. Our immunofluorometric assay and the IIF assay agreed on the IgG seropositivity of 18 samples, out of 27 that were found positive by at least one assay (Figure 2.6A). IgA autoantibodies were detected by both our immunofluorometric assay and the IIF assay in 11 out of 18 samples that were found positive by at least one assay (Figure 2.6B).

We assessed the extent of significance of clinical correlates such as Montreal classification criteria 166 with antiCUZD1 antibody presence. A Kruskal-Wallis test showed that there was no statistical difference in the prevalence of IgG or IgA CUZD1 autoantibodies in patients with different CrD location or with different phenotype of disease and that there was no significant difference in the prevalence of either IgA or IgG autoantibodies in UC patients with different extent of disease. A Mann-Whitney test showed that the prevalence of CUZD1 autoantibodies was not correlated to the duration of disease (of CrD or UC) or age of onset.

Anti-CUZD1 IgG in Inflammatory Bowel Disease by ELISA, IIF or both

Anti-CUZD1 IgG in IBD by IIF

- + Total

- 173 7 180 Anti-CUZD1 IgG in IBD ELISA + 2 18 20

Total 175 25 200

Anti-CUZD1 IgA in Inflammatory Bowel Disease by ELISA, IIF or both

Anti-CUZD1 IgA in IBD by IIF

- + Total

- 182 4 186 Anti-CUZD1 IgA in IBD ELISA + 3 11 14

Total 185 15 200

Figure 2.7: A) Contingency table displaying the prevalence of IgG autoantibodies (upper panel) or IgA autoantibodies (lower panel) in IBD patients, as determined by ELISA or indirect immunofluorescence (IIF). For discussion see text.

5 Samples from patients being tested for 4 autoimmune diseases

Fluorescence Ratio Fluorescence Positive Controls 3

0 0 50 100

Number of Sample

16 14 12

Fluorescence Ratio Fluorescence Samples from patients 10 being tested for 8 autoimmune diseases 6 4 2 0 0 20 40 60 80 100 120

Number of Sample Figure 2.8 | Analysis of 129 serum samples from patients previously tested for autoimmune diseases. These samples were assessed with our ELISA for IgG (upper panel) or IgA (lower panel) CUZD1 autoantibodies.

2.3.4 Anti-CUZD1 antibodies in vADs

Serum samples were obtained from 129 patients that were tested for various autoantibodies (ANA, ANCA, anti-CCP, celiac). The samples were prepared in a similar manner and analyzed by both assays targeting IgG and IgA CUZD1 autoantibodies. IgG isotypes were reproducibly detected in 3 out of 129 autoimmune samples (2.32%) (Figure 2.7) while IgA autoantibodies were reproducibly found in 2 out of 129 autoimmune samples (1.55%) (Figure 2.7). Both patients were ANCA+ ANA-. In total, out of 129 tested samples, 4 were found positive for IgG and /or IgA CUZD1 autoantibodies (3.1%). Of these, 2 were also positive for ANA and 2 were tested negative for anti-CCP.

2.4 Discussion

PAB visualized by IIF, have been reported to be specific markers of Crohn’s disease. GP2 and CUZD1 have been identified as target autoantigens of PAB 125, 133, 157. GP2 is associated with the droplet pattern staining, while CUZD1 is associated with the reticulogranular staining pattern 157.

GP2 is a pancreas residing protein, which has been recently shown to have an immunomodulating role in innate and adaptive intestinal immunity and to be elevated in intestinal biopsy samples of CrD patients 133. The development of novel immunoassays with recombinant GP2 as a solid-phase antigen, for the detection of antibodies to GP2, has paved the way to investigate the prevalence and diagnostic significance of anti-GP2 antibodies in patients with IBD but analogous immunoassays using CUZD1 as solid-phase antigen have not been developed, preventing from simultaneous detection of both autoantibodies for proper antigen- specific PAB detection 133.

CUZD1 is also highly expressed in exocrine pancreas but data regarding CUZD1 expression in the intestine are lacking 125, 133, 157. An anti-CUZD1 antibody immunoassay based on IIF has been developed, but its clinical utility has not been thoroughly assessed 157. Komorowski and colleagues, by using recombinant cells expressing CUZD1 as a substrate in IIF, analyzed sera from 96 patients with CrD and reported anti-CUZD1 antibody positivity in 19.8% of the patients. Anti-CUZD1 antibody reactivity was strongly correlated with the reticulogranular PAB pattern. The authors demonstrated that their antigen capture ELISA using monoclonal antibodies against CUZD1 verified this relationship 157.

More recently, Pavlidis et al assessed the clinical utility of anti-CUZD1 and anti-GP2 by novel cell-based IIF assays in CrD 132. According to their data, anti-CUZD1 antibodies were detected in 21.7% of CrD patients 132, a percentage comparable to that obtained by Komorowski et al 157.

We developed two immunoassays using recombinant CUZD1 for the detection of IgG and IgA isotypes of CUZD1 autoantibodies in human serum. We analyzed 200 serum samples from IBD patients and found CUZD1 autoantibodies in 16% of CrD patients and in 9% of UC patients.

Our ELISA results show a good correlation with IIF results obtained using pancreatic tissue and transfected cells (Figure 2.6). In total, the percentage of agreement between the ELISA and the IIF data is 95.5% for the IgG and 96.5% for the IgA autoantibodies. Combined, our immunofluorometric assay and the IIF assay detected 27 positives for IgG, 18 which were positive by both assays. For IgA antibodies, 18 samples were positive with at least one of the assay and 11 were positive by both (Figure 2.6).

Our data are in accordance with previous studies, showing that CUZD1 autoantibodies are more prevalent in CrD (16%) than in UC (9%) but they demonstrate CUZD1 autoantibody positivity in a slightly lower percentage than previously reported 132, 157. This discrepancy could be attributed to differences in patient cohorts or methodological approaches. Pavlidis et al detected anti- CUZD1 antibodies by IIF using primate pancreatic tissue and HEK-293 overexpressing CUZD1, while Komorowski et al used IIF and two antigen capture ELISAs in which the microplates were firstly coated with monoclonal antibodies against either CUZD1 or GP2 and then loaded separately with CUZD1 and GP2 purified from pancreatic tissue. Our ELISA differs in that it employs mammalian recombinant CUZD1 as a solid phase antigen. Our assay is fast and simple to perform, and it may complement or even replace the laborious detection of anti-CUZD1 by IIF.

Using the clinical characteristics of the patient cohorts, we assessed the clinical significance of CUZD1 autoantibodies. Similar to a previous study 132, CUZD1 autoantibodies were not more prevalent in patients with a particular disease phenotype or with a certain inflamed location. Furthermore, CUZD1 seropositivity was not associated with the duration (CrD or UC) or with age of disease onset.

Our newly developed ELISAs were also used for the analysis of serum samples from patients with suspected autoimmune diseases. We here show that CUZD1 autoantibodies are detected rarely (<5%) in patients being tested for vADs including 26 for celiac disease. This observation is in accordance with previous studies showing that PABs are rarely encountered in patients with non-IBD colorectal disease or diseases not affecting the colon 125.

Our newly developed ELISAs generated data comparable to those previously reported by other groups using different methodologies. Being fast and easy, our assay is suitable for use in large multicentered studies to validate the clinical utility of CUZD1 autoantibodies in IBD patients.

3 Chapter 3 | Generation of monoclonal antibodies and development of an immunofluorometric assay for the detection of CUZD1 in tissues and biological fluids

This chapter includes data published in the Clinical Biochemistry: “Generation of monoclonal antibodies and development of an immunofluorometric assay for the detection of CUZD1 in tissues and biological fluids” by Farkona S., Soosaipillai A., Filippou P., Korbakis D., Serra S., Rückert F., Diamandis E.P., Blasutig I.M.

A link to the published paper can be found at

doi: 10.1016/j.clinbiochem.2017.07.012

SF performed the experiments and drafted the manuscript. AS assisted with the development of the methodology. PF assisted with the protein production. DK assisted with the antibodies production. SS performed the immunohistochemistry studies. FR provided samples. EPD and IMB designed the study and edited the manuscript.

3.1 Introduction

CUZD1 was identified as a highly pancreas-specific protein and was proposed as a candidate biomarker for pancreatic related disorders 137, 167. The information on CUZD1 protein is sparse: The CUZD1 gene, also known as the UO-44 139 and ERG1 138 is mapped/resides at/on chromosome 10q26.13 and encodes for a 608-amino acid polypeptide containing a signal peptide, two CUB domains, one ZP domain, a single-spanning transmenbrane region and a short cytoplasmic tail 141. The protein is predicted to be heavily glycosylated and is highly conserved among species 140. CUZD1 mRNA undergoes a complex series of alternative splicing events that give rise to three protein isoforms. The largest isoform, approximately 607 aminoacids long (68 kDa), contains the entire gene while the other two isoforms, 326 aminoacids (37 kDa) and 241 aminoacids (28 kDa) long, contain only the ZP domain and the transmembrane region or half of the ZP domain and the transmembrane region, respectively 141. CUB and ZP domains occupy the majority of the extracellular region of the largest isoform and are known to be involved ligand binding, oligomerization and cell adhesion 142, 145.

While the exact biological function of CUZD1 function is still unknown, it has been recently linked to IBD as it is one of the two main targets of pancreatic autoantibodies which are novel serological markers for CrD. According to previously published studies generated by using indirect immunofluorescence, pancreatic autoantibodies against CUZD1 are detected in 22-26% of patients with CrD and in 11-15% of patients with ulcerative colitis 132, 134, 157.

Using a novel in house developed ELISA targeting CUZD1 autoantibodies, Farkona et al. was able to detect them in 16% of patients with Crohn’s disease and in 9% of Ulcerative Colitis patients 168.

Previous studies implicate the protein as having a role in cancer progression 141, 158. In 2001, Huynh et al isolated a uterine and ovarian specific, tamoxifen and estrogen-induced rat UO-44 cDNA (the ortholog of human CUZD1) through differential display and cDNA library screening 139. The authors found that the UO-44 mRNA transcript was observed only in the uterus and ovary of rats. In 2004, the same group reported the cloning and characterization of four novel splice variants of the human ortholog of UO-44 (CUZD1) 137. While the authors found in 2001 that the Rat UO-44 was highly expressed in the ovaries and uterus 139, the human ortholog

CUZD1 was found highly expressed in the pancreas 141. In the same study, they also reported the overexpression of CUZD1 in the majority of ovarian tumors. In a more recent study, Mapes et al have shown that in mice CUZD1 is also expressed in mammary ductal and alveolar epithelium and that it has a pivotal role in JAK/STAT5 signaling that regulates mammary gland development during pregnancy 148.

Thus far, CUZD1 protein levels in tissues and biological fluids have not been extensively examined due to the lack of suitable reagents and techniques. Therefore, the purpose of the present study was to develop recombinant protein, specific antibodies, and an ELISA immunoassay to assess CUZD1 expression at the tissue level and its concentration in various biological fluids.

3.2 Materials and Methods

3.2.1 Monoclonal antibody production

Monoclonal antibodies were produced as previously described 169. Briefly, female BALB/c mice were purchased from Jackson laboratories via the Toronto Centre for Phenogenomics (TCP). All animal research was approved by the TCP Animal Care Committee. Mice were injected subcutaneously with 100 μg of human recombinant CUZD1 protein, mixed (1:1) with Sigma Adjuvant System (Sigma-Aldrich). Two subsequent booster injections with 20 μg of antigen in adjuvant were performed at three-week intervals. The final boost was an intraperitoneal injection with 20 μg of antigen in phosphate-buffered saline (137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 1.8 mM KH2PO4). Three days later, the mouse spleen was excised aseptically and homogenized. Extracted spleen cells were fused with NSO murine myeloma cells (5:1 ratio) using polyethylene glycol (Sigma-Aldrich). Successfully fused cells were isolated using selection medium, containing 2 % Gibco® HAT supplement (Thermo Fisher Scientific, Waltham, MA) and 20% fetal bovine serum (Hyclone, GE Healthcare).

3.2.2 Screening for immunogen-reacting clones by an IgG capture ELISA

Seven hundred and fifty hybridoma clones were screened for reactivity against the recombinant protein by using an IgG capture ELISA, as previously described 170.

3.2.3 Expansion of hybridomas and purification of monoclonal antibodies

Following the screening procedure, 21 hybridoma clones shown to strongly react with the recombinant protein were further grown and transferred in serum-free media (Thermo Fisher Scientific), containing 8 mM L-Glutamine. Supernatants were collected and purified using a protein G column (GenScript, Piscataway NJ, USA). Briefly, culture supernatants were diluted two times in binding buffer (20 mM NaH2PO4, 150 mM NaCl, pH 8.0) and loaded onto the column. The column was then washed with the binding buffer and antibodies were eluted with 0.1 M glycine at pH 3.0. The purified antibodies were then screened against both the recombinant protein and pancreatic tissue extracts by the ELISA described in the next section. Clones 275 and 119 were chosen to be used as the coating and the detection antibody in our ELISA, respectively.

3.2.4 ELISA development

White polystyrene microtiter plates were incubated overnight at room temperature with 100 uL of coating antibody solution containing 500 ng of monoclonal anti-CUZD1 antibody (clone 275) diluted in 50 mmol/L Tris buffer (pH 7.8). The plates were then washed three times with the washing buffer (20 mM Tris, 150 mM NaCl, 0.05% Tween-20, pH 7.4). CUZD1 calibrators or samples, diluted in a bovine serum albumin (BSA) solution [60 g/L BSA, 50 mmol/L Tris (pH 7.80) and 0.5 g/L sodium azide], were then pipetted to each well (50 uL/well) along with 50 μL of assay buffer A (60 g/L BSA, 25 mL/L normal mouse serum, 100 mL/L normal goat serum, and 10 g/L bovine IgG in 50 mM Tris, pH 7.8, 0.005% (v/v) Tween-20) and incubated for 2 h with shaking at room temperature. The plates were washed three times with the washing buffer after which 100 uL of biotinylated detection antibody solution containing 200 ug/L anti-CUZD1 monoclonal antibody 119 in assay buffer A was added to each well and incubated for 1 h at room temperature with shaking. The plates were washed six times with the wash buffer. Subsequently, 100 uL of alkaline phosphatase-conjugated streptavidin (50 ng/mL in BSA solution) were added to each well and incubated for 15 minutes with shaking at room temperature. Plates were washed as above and 100 uL of diflunisal phosphate solution [0.1 mol/L Tris-HCl buffer (pH 9.1) containing 1 mmol/L diflunisal phosphate, 0.1 mol/L NaCl and 1 mmol/L MgCl2] was added to each well and incubated for 10 minutes with shaking at room temperature. Developing solution (100 uL, containing 1 mol/L Tris base, 0.4 mol/L NaOH, 2 mmol/L TbCl3, and 3 mmol/L

EDTA) was pipetted into each well and mixed for 1 min. Time-resolved fluorescene was measured with the Wallac EnVision 2103 Multilabel Reader (Perkin Elmer, Waltham, MA, USA).

Recombinant CUZD1 protein, produced and purified as described elsewhere 168, was used to generate the calibration curve. CUZD1 calibrators were prepared by diluting the purified recombinant CUZD1 in the general diluents. These calibrators were used to define the detection limit of the assay. To determine the linearity of the CUZD1 immunoassay, we serially diluted a pancreatic tissue extract in BSA solution and measured the CUZD1 concentration with the standard assay procedure. Recombinant CUZD1 was added to normal sera at different concentrations and measured with the developed CUZD1 immunoassay. Recoveries were calculated after subtraction of the endogenous concentrations.

3.2.5 Immunohistochemistry

A biotin free IHC method was performed on tissue sections using two mouse monoclonal CUZD1-specific antibodies, clone 275 (used as coating antibody in our ELISA) and clone 119 (used as detection antibody in our ELISA). Four m formalin-fixed paraffin-embedded tissue sections were dewaxed in 5 changes of xylene and brought down to water through graded alcohols. Endogenous peroxidase was blocked with 3% hydrogen peroxide. Antigen retrieval or unmasking procedures were applied, if necessary, as follows. The sections were pretreated with citrate buffer (pH 6.1) or Tris-EDTA in a microwave and then incubated overnight with clone 275 (1:100) or for one hour with clone 119 CUZD1-specific antibody (diluted 1:1,000). The detection system used was MACH 4 universal horseradish peroxidase (HRP) polymer system (Intermedico Cat# BC-M4U534). After following kit instructions, color development was performed with freshly prepared DAB (DAKO Cat# K3468). Finally, sections were counterstained lightly with Mayer’s Hematoxylin, dehydrated in alcohols, cleared in xylene and mounted with Permount mounting medium (Fisher cat# SP15-500).

3.2.6 Tissue extracts and biological fluids

All samples were collected with Research Ethics Board (REB) approval. The levels of CUZD1 in human tissues and biological fluids were determined using the CUZD1 immunoassay. Human tissues were stored at -80oC immediately after collection. The extracts were prepared as follows:

Frozen human tissues were pulverized in liquid nitrogen to fine powders and mixed with 1 mL of extraction buffer [50 mmol/L Tris-HCl buffer (pH 8.0) containing 150 mmol/L NaCl, 5 mmol/L EDTA, 10 g/L NP-40 surfactant]. The mixtures were incubated on ice for 30 min with repeated vortex mixing every 10 min and sonicated three times for 10 s, to break up the tissue further. Mixtures were centrifuged at 15,000 g at 4 °C for 30 min, and the supernatants were used as tissue extracts. The biological fluids were obtained as leftovers of samples submitted for routine biochemical testing and stored at -80 °C until use. Prior to freezing, samples were kept according to routine clinical laboratory procedures, and would not have exceeded 1 week of storage at 4oC.

3.2.7 SDS PAGE (& MS/MS) and Western blotting of Protein tissue extracts and biological fluids

Selected protein tissue extracts and biological fluids were subjected to SDS-PAGE and Western Blot analysis. The samples were mixed 1:1 with loading buffer (Laemmli sample buffer, BIORAD) and reducing agent at a final concentration of 0.3M, boiled for 10 min, and cooled on ice for 2 min. Following short centrifugation for 1 min at 17,000 g they were separated by sodium dodecyl sulfate-polyacrylamide gel electroporesis (SDS-PAGE) using the Mini- PROTEAN® Tetra cell system and 4–12% gradient polyacrylamide gels at 200 V for 45 min (Bio-Rad). The PiNK Plus Prestained Protein Ladder (FroggaBio) was used as protein standard for molecular mass determination. Gels were stained by Biosafe Coomassie staining (Invitrogen). For the Western Blot analysis, the proteins were transferred to a polyvinylidene difluoride (PVDF) membrane (Trans-Blot® TurboTM Mini PVDF Transfer Packs from Bio- Rad) with the Trans-Blot® TurboTM Transfer Starter System. The membrane was blocked overnight at 4 °C with 5% milk and further probed with the in-house primary anti-CUZD1 mouse monoclonal antibody (either clone 275 or 119, 0.46 ug/mL or 0.38 ug/mL, respectively) for 1 h at room temperature in 1% milk. Following 4 washes with Tris-buffered saline, 0.1% Tween-20 (TBST) the membrane was incubated with anti-mouse antibody conjugated to HRP (Jackson ImmunoResearch) diluted 1/10,000 in 1% milk for 45 min in room temperature. After 4 washes with TBST, membranes were incubated with ECL western blotting detection reagent and exposed to x-ray film (GE Healthcare).

3.2.8 Fractionation of biological fluids containing CUZD1

200 uL of recombinant CUZD1 (500 ug/L diluted in 6% BSA), a pancreatic tissue extract and pancreatic juice (containing 20 ug/L and 200 ug/L of CUZD1 respectively, diluted in 0.1 M

NaH2PO4 and 0.15 NaCl) were fractionated using gel filtration chromatography, as described elsewhere 170, 171. Briefly the samples were loaded through a 500 uL loop on a silica-based TSKGEL G3000SW gel filtration column (60 cm × 7.5 mm ID), previously equilibrated with the same buffer and connected to an Agilent 1100 series HPLC system. Separation was accomplished during a 60 min run at a flow-rate of 0.5 mL/min. Fractions (0.5 mL) that were collected throughout the run and analyzed for the presence of CUZD1 protein by the ELISA, as described above and by liquid chromatography-tandem mass spectrometry (LC-MS/MS) as described in the next section.

3.2.9 MS/MS of bands and gel filtration fractions

Selected gel filtration fractions were pooled, concentrated approximately 40-fold using Amicon Ultra 0.5 mL centrifugal filters (cutoff 10K), and reduced with the addition of dithiothreitol at a final concentration of 10 mM in 50 mM NH4HCO3 at 60 °C for 30 min. Following their alkylation with the addition of iodoacetamide at a final concentration of 20 mM and incubation in the dark at room temperature for 1 h, the samples were digested with trypsin at a ratio of 1:20 tryspsin to protein concentration, overnight at 37 °C.

Selected bands from the Coomassie stained gel were excised and destained with 100 uL of 100 mM ammonium bicarbonate/acetonitrile (1:1). Each band was then reduced (10 mM dithiothreitol in 100 mM ammonium bicarbonate, pH 8.3) and alkylated (100 mM iodoacetamide in 50 mM ammonium bicarbonate, pH 8.3) before overnight trypsin digestion (0.5 ug of trypsin per band). Peptide fragments were then extracted with 5% formic acid in acetonitrile.

For MS analysis, in all cases, extracted peptides were desalted and concentrated in a solution of 95% water, 5% acetonitrile and 0.1% formic acid using C18 Bond Elut OMIX tips (10 μl; Agilent Technologies, Mississauga, ON, Canada). The tryptic peptides were loaded on a 3.3 cm long C18 (5 μm) Agilent Pursuit trap precolumn (i.d. 75 μm) via EASY-nLC 1000 pump (Thermo Scientific) at 8 μl/min. Mobile phases included 0.1% formic acid in water (buffer A)

74 and 0.1% formic acid in acetonitrile (buffer B). Peptides were separated with a 15-cm long C18 (3 μm) Agilent Pursuit analytical column (i.d. 50 μm) with a 8 μm tip (New Objective, Woburn, MA) with a 60 min gradient elution at 300 nL/min flow rate. A five-step gradient was used: 1% to 14% of buffer B for 1 min, 14% to 40% for 11 min, 40% to 65% for 2 min, 65% to 100% for 1 min, and 100% for 7 min. The liquid chromatography setup was connected to a Q Exactive™ Plus Hybrid Quadrupole-Orbitrap™ Mass Spectrometer (Thermo Scientific) with a nanoelectrospray ionization source (Proxeon Biosystems, Odense, Denmark). Analysis of the eluted peptides was done by tandem mass spectrometry in positive-ion mode as previously described 172. Briefly a full MS1 scan was acquired from 250 to 1500 m/z in the orbitrap mass analyzer (70,000 resolving power at 200 m/z) followed by 12 data-dependent MS2 scans. Raw files were generated with the use of the XCalibur software 3.0.63 (Thermo Fisher) and integrated through Proteome Discoverer Software 1.4 to complete searches against the nonredundant Human SwissProt Uniprot database (version January 2015).

3.3 Results

3.3.1 ELISA development

Recombinant CUZD1 protein has been previously expressed in a mammalian expression system and purified using anion exchange chromatography 168. The recombinant CUZD1 was used as an immunogen in mice for production of monoclonal antibodies. The fusion of murine splenocytes with murine myeloma cells 169 resulted in the generation of 750 IgG-secreting hybridoma colonies, with 21 antibodies strongly reacting with recombinant CUZD1. These 21 antibodies were further screened for reaction with the endogenously expressed CUZD1 protein in a pancreatic tissue extract. Finally, out of the 21 antibodies, two were selected, clone 275 and clone 119, as capture and detection antibodies for our ELISA, respectively, due to their strong and linear response to both the recombinant and the native CUZD1 protein.

Clone-275 was used for coating microtiter plates and clone-119 was labeled with biotin and used for detection in an ELISA-type sandwich assay. Alkaline phosphatase-labeled streptavidin was then used and the activity of alkaline phosphatase was detected by time-resolved fluorometry, as described previously 170. Purified recombinant CUZD1 was diluted in BSA solution to produce six assay calibrators of the following concentrations: 0, 0.5, 2, 5, 20, 100 ug/L. The detection

75 limit, defined as the concentration of analyte (in this case recombinant CUZD1) that can be distinguished from zero with 95% confidence, was 0.5 ug/L and the upper limit of the dynamic range was 100 ug/L. The within run and day-today CVs for the developed ELISA for CUZD1 were < 15 % within the measurement range. Recovery of recombinant CUZD1 added to sera (male and female) was 40-50%, indicating incomplete recovery. Assay linearity was verified using serial dilutions of pancreatic tissue extracts.

The newly developed immunoassay was used to assess the levels of CUZD1 protein in a variety of normal human tissues extracts. (Figure 3.1A). All values were corrected for the total protein content of the extracts. Among the normal adult human tissues, pancreatic tissue was found to express the highest levels of CUZD1 protein (300-3,000 ng/mg). Diverse other tissues, express the protein at much lower but detectable levels (up to 2 ng/mg, except one colon extract, 8 ng/mg). We conclude that CUZD1 expression is 100-1,000 fold higher in the pancreas than any other tested tissue.

The concentration of CUZD1 in biological fluids from healthy individuals was also quantified (Figure 3.1B). The highest levels of CUZD1 were found in pancreatic juice (200-10,000 ug/L), whereas CUZD1 levels were below the limit of detection of the assay in amniotic fluid, saliva, follicular fluid, malignant ascites, serum, peritoneal fluid, urine and breast milk samples derived from individuals.

Figure 3.1 | Α) Expression of CUZD1 in adult tissue extracts from males and females; all concentrations were normalized for the total protein and are expressed as ng of CUZD1 per mg of total protein (note that the y-scale changes); Sm intestine; small intestine. Β) Quantification of CUZD1 in biological fluids, Fol fluid; follicular fluid, Am fluid; amniotic fluid, per fluid; peritoneal fluid, PJ; pancreatic juice. The number of unique samples analyzed is indicated in brackets.

3.3.2 Immunohistochemical analysis of pancreatic tissue using the in- house generated monoclonal antibodies

CUZD1 protein was immunohistochemically localized in normal pancreatic tissues, by using clone 275-and clone-119 antibodies. In both cases, and in accordance with previous studies, staining was observed only in the lumen of the acini and in the zymogen granules of the acinar cells, whereas the ductal epithelium and the islets of Langerhans were negative (Figure 3.2).

A B A A

D A D A A A

D A D A

C A D D A D A A A

D D A I D I A D I A A A A I

Figure 3.2 | Immunohistochemical localization of CUZD1 protein in paraffin-embedded pancreatic tissues. The staining was performed by using the monoclonal antibody that is used as the capture reagent (clone 275) in our ELISA (A and B), or the antibody used for detection (clone 119) in our ELISA (C and D). In both cases staining is seen throughout the acinar cell populations, and especially in the zymogen granules, but not in the ductal cells or the islets of Langerhans. A; acini, D; duct, I; islet. For more discussion see text.

3.3.3 Western Blot analysis of tissues and biological fluids

Both clones 275 and 119 were tested for their utility as reagents for Western Blotting analysis. Clone-275 did not recognize CUZD1 protein in any of the tested samples, including recombinant protein. Analysis of selected tissue extracts by using clone-119 antibody as the primary antibody revealed positivity in pancreatic tissue in the form of a strong diffuse band corresponding to a molecular weight of approximately 100 kDa (Figure 3.3). The larger than expected size of CUZD1 (predicted to be 67 kDa) and its diffuse appearance suggests glycosylation and/or oligomerization of CUZD1. The extracellular domains of CUZD1 are known to participate in such events, but this was not further explored. CUZD1 protein was not detected by Western Blotting analysis in other protein extracts mentioned in figure 1 (data not shown).

Selected biological fluids, including two different pancreatic juice samples, were also analyzed by Western Blot analysis. CUZD1 was detected in one sample as two strong bands, the first at a molecular weight of approximately 100 kDa and the second at a molecular weight of approximately 175 kDa (Figure 3.3). Although this was not further investigated, the appearance of those two bands at a molecular weight larger than what is predicted for CUZD1 protein could be attributed to glycosylated and/or oligomerized forms of the protein. CUZD1 was also detected in a second pancreatic juice as two bands, but at different sizes. The first at a molecular weight between 62 and 70 kDa, which is close to the predicted molecular weight of the full-length protein, and the second at a molecular weight between 50 and 60 kDa (Figure 3.3). The latter could be attributed to a cleaved form of CUZD1 protein, but it was not investigated further. The same samples were resolved on SDS PAGE gels and stained with Coomassie.

1 2 3

kDa4

175 1 2 3 4130

95 62

42 Pancreatic Pancreatic Pancreatic Recombinant juice juice tissue protein extract

Figure 3.3 | Western blot showing detection of CUZD1 protein in two different pancreatic juice samples (lane 1 and 2) and in a pancreatic tissue extract (lane 3). Lane 4 shows recombinant CUZD1 protein previously expressed and purified in our lab (12). In lane 1 CUZD1 appears as two bands of approximately 100 and 175 kDa while in lane 2 CUZD1 appears in the form of two bands of approximately 50 and 62-70 kDa. In pancreatic tissue extract, CUZD1 is detected in the form of a strong diffuse band of a molecular weight of approximately 100 kDa and higher. Recombinant protein appears in the form of a diffuse band of a molecular weight of approximately 70 kDa. On the left panel are molecular mass standards. For more discussion see text.

The areas corresponding to the aforementioned molecular weights were excised and analyzed by LC-MS/MS, which confirmed the presence of CUZD1 protein in all bands, except for the one corresponding to a molecular weight between 62 and 70 kDa (Figure 3.3, lane 2). LC-MS/MS analysis of this band revealed an abundance of albumin peptides (data not shown). For this reason failure to detect CUZD1 protein in the aforementioned band could be attributed to the presence of albumin which is the most abundant protein at this molecular weight and could have masked the presence of CUZD1.

3.3.4 Fractionation of tissue extracts and biological fluids containing CUZD1

To determine the molecular mass of CUZD1 detected in the biological fluids and tissue extracts and whether the endogenous protein can be found in various molecular forms, we fractionated CUZD1 of pancreatic tissue extracts and pancreatic juice with gel filtration chromatography and analyzed all fractions with the CUZD1 ELISA. Recombinant CUZD1 expressed in our laboratory 168 was also analyzed. In the cases of both pancreatic tissue extract and pancreatic juice CUZD1 eluted as a single, distinct peak around fraction 21 (elution time 21 min), corresponding to a molecular mass of 670 kDa, whereas in the case of recombinant protein, apart from the main immunoreactive peak at 670 kDa a shoulder peak around 200 kDa was also observed (Figure 3.4). These data are inconsistent with the theoretical molecular mass of free (uncomplexed) CUZD1 and with the bands that were observed during the Western Blot analysis and they suggest that in pancreatic juice and the pancreatic tissue extract, endogenously expressed CUZD1 is part of a large complex. Fractions corresponding to the immunoreactive peaks were pooled, concentrated and analyzed by LC-MS/MS for the identification of CUZD1.

670 kDa 158 kDa 44 kDa 17 kDa

3.5 Pancreatic Tissue 3 2.5 1 2 2 3 1.5 4 CUZD1 (ug/L) CUZD1 1 0.5 A 0 0 10 20 30 40 50 60

Elution time (min)

Pancreatic Juice

20 1

15 2 3 4

CUZD1 (ug/L) CUZD1 10

0 B 0 10 20 30 40 50 60

Elution time (min)

35 Recombinant Protein 30

25 1 2 20 3 4

CUZD1 (ug/L) CUZD1 15 10 5 C 0 0 10 20 30 40 50 60 Elution time (min) Figure 3.4 | HPLC separation on a size exclusion/gel filtration column of pancreatic tissue extract (A) pancreatic juice (B) and recombinant CUZD1 (C). The collected fractions were analyzed for CUZD1 by the developed ELISA. In pancreatic juice and pancreatic tissue extract

82 there is one immunoreactive peak at elution time 20 min corresponding to a molecular weight of 670 kDa. For recombinant protein, apart from the main immunoreactive peak at elution time of 20 min, there is also a shoulder peak at elution time 25 min, corresponding to a 600 kDa protein. For more discussion see text.

The presence of CUZD1 protein was confirmed by MS in the immunoreactive peaks derived from pancreatic juice and recombinant protein. Other proteins that were detected include major glycoprotein 2 (GP2), chymotrypsin-like elastase family member 3A (CELA3A), mucin5B (MUC5B), mucin 6 (MUC6) also known to be abundant in pancreatic tissue and/or pancreatic juice, as well as less pancreatic specific proteins such as alpha-1 antitrypsin, albumin, actin, lysozyme, calmodulin and keratins. MS analysis did not detect CUZD1 protein in the immunoreactive peak from pancreatic tissue sample, probably, due to injection of a 10-fold lower amount of the protein.

3.4 Discussion

Although cloned in 2004 141, human CUZD1 has been poorly studied, so far, particularly at the protein level and its precise biological function and relevance to pathologic conditions remain elusive. This can be partly attributed to the lack of reagents (monoclonal antibodies and immunoassays) sensitive and specific enough to measure the native protein in biological samples. In the present study, recombinant CUZD1 protein, previously produced and purified in- house 168, was used as an immunogen to generate mouse monoclonal antibodies. These antibodies were used to develop a sensitive and specific ELISA and performed immunohistochemical studies and Western Blots.

The ELISA is based on a mouse monoclonal coating/mouse monoclonal detection antibody configuration and is sensitive and specific, detecting CUZD1 concentrations ≥ 0.5 ug/L. The recovery of recombinant CUZD1 from serum was incomplete, with values ranging from 40-50%, suggesting complexation with other serum proteins or degradation. Using this ELISA, CUZD1 protein expression levels were determined in various human tissues and biological fluids.

CUZD1 protein was found to be expressed in high levels in the pancreas (Figure 3.1A). Immunohistochemical staining of CUZD1 in pancreatic tissue with both antibodies showed that the protein is localized to the acinar cells and the lumen of the acini but not to the ductal cells or

83 the islets of Langerhans (Figure 3.2). Although the predominant expression of CUZD1 in the normal pancreas and its localization to the acinar cells has been shown both at the mRNA and at the protein level in previously published articles 141, 147, 157 and Human Protein Atlas (HPA), the role of the protein in this tissue is still unknown. This study reveals expression of CUZD1 protein in lower but detectable levels in several other tissues such as colon, small intestine, adrenal, heart, fallopian tube, liver and others. CUZD1 expression was also examined in a panel of human biological fluids and it was detected exclusively in pancreatic juice, suggesting that although this protein is predicted to be a (single pass type I) transmembrane protein, it is secreted by the pancreatic acinar cells.

Both antibodies that have been used for the development of our ELISA (clones 275 and 119) were also used on Western blots for the detection of CUZD1 in several tissues and biological fluids (Figure 3.1B). The antibody that is used as a “capture” reagent in our ELISA (clone-275) was not able to detect recombinant CUZD1 or native CUZD1 in any of the sources, indicating that this antibody recognizes conformational epitopes of the protein which are disrupted under denaturing and reducing conditions. The detection antibody of our ELISA (clone-119) was able to recognize recombinant protein and native CUZD1 in a pancreatic tissue extract and two different pancreatic juice samples (Figure 3.3). Two bands of approximately 100 and 175 kDa were observed in the first pancreatic juice sample while two bands of approximately 50 and 62- 70 kDa were shown in the second pancreatic juice sample. In the pancreatic tissue extract, CUZD1 is represented by a strong, diffuse band of approximately 100 kDa and higher, while, as previously shown 168, recombinant protein was detected as a diffuse/strong band of a molecular mass of approximately 70 kDa. CUZD1 cDNA predicts a product of 68 kDa and although the size difference was not investigated here, we speculate that the bands of the higher molecular weight could be attributed to post-translational glycosylation and/or oligomerization processes in which both CUB and ZP domain are known to be involved 142, 145 whereas a degraded product of the protein could account for the product of 50 kDa observed in the second pancreatic juice sample.

Gel filtration HPLC of a pancreatic juice and a pancreatic tissue extract revealed that the ELISA detects only one immunoreactive peak corresponding to a molecular mass of approximately 670 kDa (Figure 3.4A and B), much higher than the size predicted for CUZD1 protein and the molecular mass indicated by the bands in our Western blots. These data suggest that under native

84 conditions CUZD1 exists as a complex. Size exclusion of semi-purified recombinant protein showed the same 670 kDa immunoreactive peak and a shoulder peak corresponding to a lower molecular mass but still much higher than the predicted. The recombinant protein with a predicted molecular weight of approximately 65 kDa, according to the cloned cDNA encoding for the 568 aminoacid-long polypeptide, is only a few aminoacids shorter than the largest isoform of the protein and it contains all these elements/domains of the protein that could lead to its polymerization 142, 145. Thus, in the case of native CUZD1, detected in pancreatic juice and pancreatic tissue extract, the presence of an immunoreactive peak could be attributed to (homo) oligopolymerization processes involving both ZP and CUB domains. These findings suggest that under native conditions CUZD1 protein exists in large complexes or aggregates. Although the composition of these complexes was not further investigated in this study, it is tempting to speculate that in the case of pancreatic juice and pancreatic tissue extract, where the native protein is detected, these complexes could consist of polymerized CUZD1 complexed with other ZP-containing proteins, such as GP2, along with other glycoproteins, such as mucins. In the case of the recombinant protein, these complexes could be composed of homopolymerized CUZD1.

In conclusion, this study reports for the first time the development of tools targeting CUZD1 protein. Monoclonal antibodies were generated against recombinant CUZD1 expressed in a mammalian expression system and used for the development of a sensitive ELISA. The newly developed ELISA was used to establish the tissue expression pattern of the protein and to assess its levels in several biological fluids. The monoclonal antibodies were also used in IHC and Western Blot analysis to further characterize CUZD1 in pancreatic tissue and pancreatic juice. These new tools will facilitate future investigations aiming to delineate the role of CUZD1 in physiology and pathobiology.

4 Chapter 4 | Investigating the clinical utility of CUB and zona pellucida-like domain-containing protein 1 (CUZD1) in malignant and non-malignant human diseases

This chapter is based on a manuscript that is in preparation by Farkona S., Serra S., Filippou P., Korbakis D., Molina R., Bogdanos D.P., Diamandis E.P., Blasutig I.M.

SF performed the experiments and drafted the manuscript. SS performed the immunohistochemistry studies. PF assisted with the protein production. DK assisted with the antibodies production. RM and DPB provided samples. EPD and IMB designed the study and edited the manuscript.

4.1 Introduction

The first report on CUB CUZD1 gene was in 1998 when Kasik et al reported its isolation from mouse uterus. The data showed that the expression of mouse CUZD1 mRNA is temporo-spatial; it appears in the uterus during late pregnancy but it is practically undetectable in non-pregnant uterus or in a variety of adult and fetal tissues 146. The mouse ortholog of CUZD1 was also found to be highly expressed in pancreas by a different group and to play a role in the activation of trypsinogen and in the severity of pancreatitis 147.

According to a more recent study CUZD1 is expressed in the mammary gland of mice at different stages of development 148. CUZD1 protein was detected in the developing ductal epithelium at puberty and in both cytoplasmic and nuclear compartments of the ductal and alveolar epithelial cells during alveologenesis at late pregnancy whereas nuclear staining was observed during lactation. One important finding of this study was that CUZD1 deficient mice experienced a significant defect in mammary ductal branching and alveolar differentiation leading to a subsequent impairment in lactation. In addition, it was demonstrated that CUZD1 is part of a protein complex composed of, among others, JAK1/JAK2 and STAT5. The overall conclusion was that CUZD1 plays a pivotal role in prolactin-induced JAK/STAT5 signaling that regulates the expression of critical STAT5 target genes involved in mammary epithelial proliferation and differentiation during alveolar development 148.

CUZD1 has been previously connected to ovarian cancer. Leong et al. reported in 2004 that CUZD1 mRNA is overexpressed in a majority of ovarian tumors and demonstrated that CUZD1 antisera strikingly inhibit cell attachment and proliferation of NIH-OVCAR3 ovarian cancer cells 141. This observation has led the authors to speculate that human CUZD1 could be involved in cancer cells attachment and proliferation and that it could possibly promote cell growth in ovarian cancer.

CUZD1 has been identified as one of the two main target autoantigens of PAB, which have been proposed as candidate serum biomarkers for the detection of IBD 157. The expression of GP2, which is the other antigenic target of PABs has been demonstrated at sites of intestinal inflammation and, due to the development of ELISAs targeting GP2 autoantibodies, their clinical

87 utility has been thoroughly investigated 125, 131, 133. On the contrary, there are no data regarding CUZD1 expression in the intestine and the clinical utility of CUZD1 autoantibodies has not been assessed at the same extent as this of anti-GP2 antibodies. This has been attributed to the lack of appropriate and validated reagents; monoclonal antibodies against CUZD1 antigen and ELISAs targeting CUZD1 autoantibodies. We have previously reported the development of novel immunoassays detecting CUZD1 autoantibodies and we proved that they can be utilized in large scale studies for the investigation of the diagnostic and clinical utility of CUZD1 autoantibodies 131.

Although CUZD1 has been previously linked to diseases (such as pancreatitis 147, ovarian cancer 141 and IBD 157) it is unknown if this antigen can be used as a serum biomarker for the diagnosis of any of these disorders. In humans, CUZD1 is highly expressed in exocrine pancreas 173 and for this reason one could expect that an elevation of this antigen in the serum of patients could be related to pancreatic or even digestion related disorders. Also, the specific expression of CUZD1 in the pancreatic acinar cells could be translated to elevated CUZD1 levels in the serum of patients with the rare condition of acinar cell carcinoma (ACC). At the same time, due to the previously reported association of CUZD1 with ovarian cancer it is interesting to examine if the protein levels are elevated in the serum of ovarian cancer. In this study, we use previously developed and validated ELISAs measuring CUZD1 protein 173 to determine CUZD1 concentration in two cohorts of serum samples from patients with various benign and malignant disorders, one cohort of serum samples from patients with pancreatic and gastrointestinal malignant and benign disorders and one cohort of serum samples from patients with IBD. In addition, we use previously generated and validated monoclonal antibodies raised against recombinant CUZD1 173, to examine if the protein is present in the inflamed intestine tissue of patients with IBD and in the pancreatic tissue of patients with ACC and PDAC.

4.2 Materials and Methods

4.2.1 Patients

In this study four sample cohorts/sets, defined here as set A, B, C and D, were analyzed with the in-house developed ELISA targeting CUZD1 antigen 173. In all cases samples were collected with patient written informed consent and Institutional Review Board approval. Briefly,

88 peripheral blood was collected in serum separation tubes, allowed to clot for 30 minutes at room temperature and then placed on ice. Within 2h of blood draw, samples were centrifuged at 1100– 1300 g for 10 min to pellet the cells. Immediately after centrifugation, the sera were aliquoted into 1-mL cryotubes and stored at -80 °C until analysis.

Sample set A and B were both provided by Dr. Molina (Service of Clinical Biochemistry, Hospital Clínic de Barcelona, Barcelona, Spain). Sample Set A consists of 406 serum samples from patients with various malignancies (breast cancer, digestive cancer, gynecological cancer, leukemia, lung cancer, melanoma, multiple myeloma, neuroendocrine tumor, primary liver cancer, prostate cancer,) other diseases (acute myocardial infarction, autoimmune, dermatological, digestive, gynecological, liver, pneumopathy, renal failure and stroke) and healthy controls.

Sample set B consists of 236 samples from patients with various malignancies (breast cancer, cervix cancer, colorectal cancer, esophagus cancer, gastric cancer, kidney cancer, liver cancer, lung cancer, lymphoma, melanoma, multiple myeloma, oral cavity cancer, ovarian cancer, pancreatic cancer, prostate cancer, thyroid cancer, tongue cancer), other diseases (adenopathies, cardiac insuffuciency, celiac disease, chirrosis, colorectal polyps, HBP, lumbago, myocardial infarction, neck pain, ovarian cyst, pancreatic cyst, pneumonia, prosthesis infection, renal failure, Sjoegren, SLE) and from healthy donors.

Sample set C consisted of 239 serum samples from patients diagnosed with 9 malignant disorders (colorectal carcinoma, common bile duct cancer, pancreatic cyst cancer, gastric cancer, islet cell cancer, pancreatic cancer, papillary cancer, retroperitoneum paraganglioma, serous cystadenoma) and 15 benign (cholangiocellular caricinoma, common bile duct dilatation, common hepatic artery LN swelling and chronic hepatitis, duodenal diverticulum, gall bladder adenomyomatosis, hemangioma in liver, intraductal proliferative lesion with adenomatoid ductal hyperplasia, intraductal proliferative neoplasia, liposarcoma, pancreatic cyst, pancreatitis, parapancreatic LN swelling, pancreatic duct dilatation, others) disorders of the digestive system. These samples were collected at Toronto General Hospital Clinical Biochemistry Laboratory according to standardized protocols mentioned above.

Sample set D consists of 200 serum samples from patients with IBD (equally divided between CrD and UC) and 50 serum samples from healthy individuals and patients with diseases

89 unrelated to IBD. This sample cohort was provided by Dr. Bogdanos (Faculty of Medicine, Department of Rheumatology and Clinical Immunology, School of Health Sciences, University of Thessaly, Larissa, Greece) and have been previously analyzed by using an in-house developed ELISA targeting CUZD1 autoantibodies 168. Table 4.1 shows the characterization of the samples.

Table 4.1 | Characterization of the samples

Number of samples with

Sample Provider Total Malignant Benign Controls Set number of disorders disorders samples

A Dr. Molina 406 196 168 42

B Dr. Mo lina 236 134 71 31

C TGH 239 79 160

D Dr. Bogdanos 250 200 50

4.2.2 Analysis of samples by using in-house developed ELISA

CUZD1 was measured using an in-house developed ELISA as described previously 173. Briefly, 96-well microtiter plates were incubated overnight at room temperature with 100 uL of coating antibody solution containing 500 ng of monoclonal anti-CUZD1 antibody (clone 275) diluted in 50 mmol/L Tris buffer (pH 7.8). After washing with washing buffer (5 mmol/L Tris, 150 mmol/L NaCl, 0.05% Tween-20, pH 7.8), 50 uL of CUZD1 calibrators or samples was pipetted into each well and incubated with 50 μL of assay buffer (6% BSA, 50 mmol/L Tris, 10% goat IgG, 2% mouse IgG, 1% bovine IgG, 0.5 mol/L KCl, pH 7.8) for 2 hours with shaking at room temperature. Following washing, another previously biotinylated, CUZD1-specific monoclonal antibody (clone 119) was added (200 ng/well) and incubated for 1 hour. Unbound biotinylated antibody was washed off and 100 uL of alkaline phosphatase-conjugated streptavidin (50 ng/mL in BSA solution) was added to each well for 15 min. A final washing preceded the addition of

100 uL diflunisal phosphate solution [0.1 mol/L Tris-HCl buffer (pH 9.1) containing 1 mmol/L diflunisal phosphate, 0.1 mol/L NaCl and 1 mmol/L MgCl2] to develop the quantifiable signal and after 10 min developing solution (100 uL, containing 1 mol/L Tris base, 0.4 mol/L NaOH, 2 mmol/L TbCl3, and 3 mmol/L EDTA) was pipetted into each well and mixed for 1 min. Time- resolved fluorescene was measured with the Wallac EnVision 2103 Multilabel Reader (Perkin Elmer, Waltham, MA, USA). All samples were analyzed in singletons. To calculate the within run and day-to-day CVs, quality controls samples were run in the beginning, middle and end of each plate. Quality control samples were made by spiking semi purified CUZD1 recombinant protein in serum at three different final concentrations 100, 20 and 2 ug/L.

4.2.3 Immunohistochemistry

Tissue biopsies from patients with CrD (5), UC (2), PDAC (3) and ACC (4) were provided from UHN biobank (UHNBB). A biotin free IHC method was performed on tissue sections using two mouse monoclonal CUZD1-specific antibodies, clone 275 and clone 119 (used as capture antibody and detection antibody in our ELISA, respectively), as previously reported 173. Four m formalin-fixed paraffin-embedded tissue sections were dewaxed in 5 changes of xylene and brought down to water through graded alcohols. Endogenous peroxidase was blocked with 3% hydrogen peroxide. Antigen retrieval or unmasking procedures were applied, if necessary, as follows. The sections were pretreated with citrate buffer (pH 6.1) or Tris-EDTA in a microwave and then incubated overnight with clone 275 (1:100) or for one hour with clone 119 CUZD1- specific antibody (diluted 1:1,000). The detection system used was MACH 4 universal horseradish peroxidase (HRP) polymer system (Intermedico Cat# BC-M4U534). After following kit instructions, color development was performed with freshly prepared DAB (DAKO Cat# K3468). Finally, sections were counterstained lightly with Mayer’s Hematoxylin, dehydrated in alcohols, cleared in xylene and mounted with Permount mounting medium (Fisher cat# SP15- 500).

4.3 Results

4.3.1 Analysis of serum samples by using in-house developed ELISA

CUZD1 levels were examined in sample set A which is composed of serum samples from patients with 10 different cancer types (breast, digestive, gynecological, leukemia, lung, melanoma, multiple myeloma, neuroendocrine tumor, primary liver, prostate), 9 other diseases including liver diseases, pneumopathy, renal failure, stroke and healthy controls (figure 4. 1). All categories displayed low levels (mean< 5ug/L) of CUZD1 protein. Elevated CUZD1 levels were observed in patients with breast cancer (5 out 33), digestive cancer (2 out of 41), leukemia (1 out of 10), primary liver cancer (2 out 30), acute myocardial infarct (1 out of 19), dermatological diseases (1 out of 13), digestive diseases (1 out of 22), liver diseases (5 out of 33) and stroke 1 out of 14).

In sample set B, CUZD1 levels were assessed in the serum from patients from 16 cancer types (breast, cervical, colorectal, esophageal, gastric, kidney, liver, lung, lymphoma, melanoma, multiple myeloma, oral cavity, ovarian, prostate, thyroid, tongue), from 15 other diseases (adenopathies, cardiac insufficiency, celiac disease, colorectal polyps, HBP, lumbago, myocardial infarction, neck pain, ovarian cysts, pancreatic cyst, pancreatitis, pneumonia, prosthesis infection, renal failure, sjoegren, SLE) and from healthy controls (figure 4.2). Most categories displayed low levels (mean < 5ug/L) of CUZD1 protein apart from the conditions adenopathies, lumbago, ovarian cyst and SLE which exhibited means of 7, 6, 12 and 6 ug/L, respectively. CUZD1 protein levels in the serum of ovarian cysts were significantly higher than those in ovarian cancer (p<0.05 Mann Whitney test). Elevated CUZD1 levels were also observed in one case out of 21 of patients with lung cancer.

Figure 4.1 | Detection of CUZD1 protein levels in serum sample set A.

A scatter plot showing the CUZD1 protein levels in the serum of CUZD1 protein levels in serum of 10 cancer types, 9 types of other diseases and healthy controls. . All categories displayed low levels (mean< 5ug/L) of CUZD1 protein. Higher concentration of CUZD1 protein was observed in isolated cases of patients with primary liver cancer (1 out of 30) and in patients with leukemia (1 out of 10).

In sample set C, CUZD1 levels were determined in 239 samples from patients with various malignant or benign disorders of the digestive system. Although all categories exhibited low

93 levels (mean<5 ug/L) of CUZD1, the antigen was elevated in 1 out of 28 pancreatitis samples, and in 3 out of 44 pancreatic duct (PD) dilatation samples (data not shown).

In sample set D, CUZD1 was measured in samples from patients with CrD (100), UC (100) and a group of control samples (50) composed of samples from healthy donors and from patients with disease non-related to IBD. Out of 100 UC samples, 18 exhibited detectable CUZD1 levels whereas out of 100 CrD samples 29 were found positive for CUZD1 levels (figure 4.3). Seven out 50 control samples were found positive for CUZD1 antigen. (This cohort of samples was previously analyzed by our group using an ELISA targeting CUZD1 autoantibodies.) CUZD1 detectability did not seem to significantly correlate with the CUZD1 autoantibody detectability previously assessed by our group (Fisher’s exact test p=0.058??).

Figure 4.2 | Detection of CUZD1 protein levels in serum sample set B.

A scatter plot of the results for CUZD1 protein levels in serum of 17 cancer types, 18 types of other diseases and healthy controls. Most categories exhibited levels (mean <5 ug/L) of CUZD1

94 protein. Adenopathies, lumbago, ovarian cyst and SLE which exhibited means of 7, 6, 12 and 6 ug/L, respectively. CUZD1 levels in the serum of patients with ovarian cysts was higher than those in patients in ovarian cancer. High CUZD1 concentration was observed in an isolated case of lung cancer. For more information see text.

B o g d a n o s

5 0 U C (1 0 0 )

4 0 C rD (1 0 0 )

) C o n tro ls (5 0 )

L /

g 3 0

(

D 2 0

U C 1 0

) ) ) 0 0 0 0 0 5 (1 (1 ( s C D l U r o C tr n o C

C o n d itio n

Figure 4.3 | Detection of CUZD1 protein levels in serum sample set D. A scatter plot of the results for CUZD1 protein levels in the serum of 100 Crohn’s disease samples, 100 Ulcerative Colitis samples and 50 controls (samples from healthy volunteers and patients with disorders irrelevant to IBD). Twenty nine out of 100 CrD samples, 18 out of 100 UC samples showed detectable CUZD1 levels and 7 out of 50 controls samples exhibited detectable CUZD1 levels. For more information see text.

4.3.2 Immunohistochemical analysis of IBD, ACC and PDAC

In-house generated monoclonal antibodies targeting CUZD1 (clone 275 and 119) 173 were used to stain colon or small intestine paraffin-embedded tissue section from 5 patients with CrD, 2 patients with UC, 4 patients with ACC and 3 patients with PDAC. In all IBD cases, clone 275 detected CUZD1 in the neuroendocrine cells (NC) present at the bottom of the intestinal crypts (figure 4.4). In addition, clone 275 stained positive for CUZD1 NC in the crypts of the normal glands. In IBD cases, clone 119 occasionally stained plasma cells of the inflammatory infiltrate. Clone 119 detected CUZD1 protein in plasma cells in normal intestine as well. Three of four paraffin-embedded tissue sections from patients with ACC showed a strong positive staining in nests of ACC with both antibodies (clone 275 and 119) (figure 4.5). The normal acini were weakly positive while the adjacent small pancreatic ducts were not reactive for CUZD1. One out of four ACC biopsies did not demonstrate immunoreactivity/positive staining for CUZD1 with either antibodies. CUZD1 antibodies did not react with any of the 3 PDAC cases (figure 4.6).

A B

Figure 4.4 | Immunohistochemical localization of CUZD1 protein in paraffin-embedded intestinal tissues from patients with inflammatory bowel disease (IBD). In this figure, sections from patients with Ulcerative Colitis (UC) are shown but the staining pattern is very similar to patients with Crohn’s disease (CrD). A) (In IBD) Clone 275 stained the cells (NC) present at the bottom of the intestinal crypts. (This is not a specific feature since NC stain positive also in normal intestine). B) (In IBD,) clone 119 occasionally stained plasma cells of the inflammatory infiltrate. (This is not a specific feature since plasma cells stain positive also in normal intestine).

A B

Figure 4.5 | Immunohistochemical localization of CUZD1 protein in paraffin-embedded pancreatic tissues from patients with Acinar cell carcinoma (ACC). A) Staining with clone 275. Nests of acinic cell carcinoma (ACC) are strongly immunoreactive for CUZD1, with both antibodies, while the normal acini (A) are weakly positive and the adjacent small pancreatic ducts (D) are negative. B) Staining with clone 119. The cytoplasm of the ACC cells is strongly immunoreactive for CUZD1, with both antibodies; the Islets (IS) are entirely negative.

Figure 4.6 | Immunohistochemical localization of CUZD1 protein in paraffin-embedded pancreatic tissues from patients with pancreatic ductal adenocarcinoma (PDAC). In this figure it is shown staining performed with clone 119. Pancreatic ductal adenocarcinoma (PDAC) and normal ducts (D) are negative while the normal acini (A) are weakly positive.

4.4 Discussion

CUZD1 is a relatively poorly studied protein that is abundantly expressed in exocrine pancreas 173. Previously, the mouse ortholog of CUZD1 has been associated with acute pancreatitis 147, while human CUZD1 has been associated with ovarian cancer 141 and CUZD1 autoantibodies have more recently been linked to IBD 157. Because of the lack of specific reagents and tools selectively targeting CUZD1, it is currently not known if this antigen possesses clinical utility as a serum biomarker for the diagnosis of the aforementioned and other diseases.

For this reason, the previously in-house developed and validated ELISA measuring CUZD1 was used for the analysis of four serum sample cohorts from patients with various malignant and benign disorders. The specific expression of CUZD1 in exocrine pancreas 173 and its previously reported association with CrD and UC 157 prompted us to investigate if it can be immunohistochemically localized in tissue sections from patients with malignant disorders of the pancreas and tissue biopsies derived from the site of inflammation from patients with IBD.

CUZD1 levels were firstly determined in a cohort composed of 406 serum samples from patients with 10 different cancer types and 9 different benign diseases (cohort A). Analysis of this cohort revealed the sporadic elevation of CUZD1 antigen in a few samples of cancer types such as breast cancer (5/33), digestive cancer (2/41), leukemia (1/10), primary liver cancer (2/30) and in other non-cancer conditions such as acute myocardial infarction (1/19), dermatological diseases (1/13), digestive diseases (1/22), liver diseases (5/33) and stroke (1/14). These data do not provide evidence that CUZD1 antigen can be used as serological biomarker of any of these disorders, but they do evoke questions regarding the mechanism behind the particularly high levels in some cases.

CUZD1 concentration was assessed in a second cohort, which consists of 236 serum samples from patients with 16 malignant diseases and 15 different benign diseases (cohort B). The most interesting finding was the significant elevated levels of CUZD1 in patients with ovarian cysts in comparison to patients with ovarian cancer. However, the number of samples in each category is small and further studies are warranted to test if CUZD1 protein has a utility for the differential diagnosis of patients with ovarian cancer over patients with ovarian cysts. This analysis also showed the occasional elevated levels of CUZD1 protein in patients with lung cancer (2/21), lymphoma (1/2), adenopathies (1/1), lumbago (1/4), pneumonia (1/2) and SLE (1/3).

CUZD1 is a pancreatic specific protein and for this reason we hypothesized that it could be related to pancreatic or in gastrointestinal related malignant or benign disorders. CUZD1 antigen was thus measured in a third cohort composed of 239 samples from patients with malignant or benign disorders of the digestive system (cohort C). CUZD1 protein was elevated in three cases of patients with MPD dilatation and in one case of patients with chronic pancreatitis. The number of samples used in this study in each of these categories is small, and for this reason these data do not support that CUZD1 antigen could be used as a biomarker for the diagnosis of MPD

99 dilatation or chronic pancreatitis. Since both chronic pancreatitis and MPD dilatation, are associated with changes in pancreatic parenchyma and pancreatic duct an elevation of the pancreatic acini-specific CUZD1 in the serum of patients with these disorders is not surprising.

CUZD1 concentration was determined in a final sample set composed of serum samples from patients with IBD and control samples from healthy donors and patients with diseases irrelevant to IBD (cohort D). CUZD1 antigen was detected in 29% of CrD samples, 18 % in patients with UC and in 14% of control samples. These data were compared with the results generated from a recent study 168, where the same cohort of IBD samples and controls were analyzed with an in- house generated ELISA targeting CUZD1 autoantibodies. In the previous study CUZD1 autoantibodies were identified in 16% of CrD samples, 9% of UC samples and in 2% of control samples. Although there were cases of serum samples were both CUZD1 antigen and CUZD1 autoantibodies were present, CUZD1 detectability did not seem to correlate with the CUZD1 autoantibody detectability and interestingly in only one out of three samples with high CUZD1 concentration, CUZD1 autoantibodies were also present. Similarly, high titers of CUZD1 autoantibodies did not necessarily translate to high concentration of antigen. This paradox can be explained in multiple ways. The presence of autoantibodies in the serum of a patient without the presence of the autoantigen can happen if once the autoantibodies are generated, they bind to CUZD1 antigen rendering it undetectable by our assay. The lack of autoantibodies in the serum where CUZD1 antigen is detected could be due to the suppression of the host’s immune system. Alternatively, since the mechanism behind the generation of CUZD1 autoantibodies is unknown it can be speculated that the generation of those autoantibodies could be induced by the presence of non-common CUZD1 protein isoforms which are not detected by the ELISA developed here.

CUZD1, along with GP2, is one of the main antigenic targets of PABs, which have been proposed as candidate biomarkers for the differential expression of CrD over UC. While both proteins are abundantly expressed in pancreatic acinar cells, GP2 has been also shown to be expressed in the apical surface of the M cells of the Peyer’s patches which are the original site of inflammation in CrD. More importantly, GP2 expression has been reported to be elevated in the targeted tissue of patients with CrD compared to patients with UC. These findings have explained, at least partly, the previously unresolved contradiction pancreatic autoimmunity and intestinal inflammation. To the best of our knowledge, data regarding CUZD1 expression in the intestine and especially at the site of inflammation in the course of CrD are lacking, raising all

100 together the question of the relevance of these autoantibodies in CrD. For this reason, one of the goals of this study was to investigate if CUZD1 protein is expressed in the inflamed tissue biopsies from patients with IBD. We used in-house generated and validated monoclonal antibodies (clone 275 and 119) to immunohistochemically localize CUZD1 in biopsies from patients with CrD and IBD. Clone 275 stained positive for CUZD1 neuroendocrine cells (NC) present at the bottom of the intestinal crypts but also NC in the crypts of the normal glands. Clone 119 sporadically reacted with plasma cells of the inflammatory infiltrate and plasma cells in normal intestine. These data do not provide information in regards to the mechanisms responsible for the induction of CUZD1 autoantibodies since it cannot be implied that the selective expression of CUZD1 in the inflamed tissue evokes autoimmunity against this antigen.

It is now well documented that CUZD1 protein is selectively and abundantly expressed in exocrine pancreas, specifically in the zymogen granules of acinar cells. For this reason, in this study, monoclonal antibodies previously generated and validated in-house were used to stain pancreatic tissue biopsies from patients with ACC and, for comparison, from patients with PDAC. In the first case, both antibodies strongly stained the cancerous tissue arisen from the acinar cells while the cancerous tissue derived from the ductal region was negative. These findings correlate well with previous report showing the absence of CUZD1 protein from the pancreatic duct. Analogous to PSA, which is specifically expressed in prostate tissue/produced almost exclusively in the prostate and represents the most well-known biomarker for prostate cancer, CUZD1 could be a serum biomarker for acinar cell carcinoma. For this reason in the future it will be interesting to test if CUZD1 protein levels are elevated in matched serum samples from patients with ACC, whose biopsies were positive for CUZD1, compared to matched serum samples from patients with PDAC whose tissue sections were negative for the protein.

In conclusion, in this study, previously developed and validated reagents and assays were utilized to investigate the clinical utility of CUZD1 in malignant and non-malignant diseases. An in-house developed ELISA measuring CUZD1 was employed for the analysis of over than 1000 clinical samples. The most interesting finding was that CUZD1 levels seemed to be elevated in patients with ovarian cysts, in comparison with patients with ovarian cancer. This observation needs further analysis in a larger set of samples. The immunohistochemical analysis of PDAC and ACC biopsies revealed that CUZD1 is highly expressed in ACC while it is not detected in

101

PDAC cases. If the differential expression of CUZD1 in ACC over PDAC is translated into higher CUZD1 levels in the matched serum samples, CUZD1 could be a potential serum biomarker for the diagnosis of ACC.

102

5 Chapter 5 | Summary and Future Directions

103

5.1 Summary of key findings

CUZD1 is a relatively poorly studied protein abundantly and primarily expressed in normal pancreas137, 173 which has been previously proposed as a potential biomarker for pancreatic related disorders 137. Owing to the lack of analytical reagents and assays targeting CUZD1, its levels in tissues and biological fluids have not been widely assessed. For this reason, one of the main goals of this study was to develop and validate analytical tools targeting CUZD1 and use them to establish the tissue expression pattern of this antigen and to assess its levels in a variety of biological fluids.

CUZD1 has been previously associated with a few disorders such as pancreatitis 147, ovarian cancer141, 158 and IBD157 but it has not been investigated if its protein levels are higher in the serum of patients with the aforementioned or other disorders. It was thus of clinical interest to use the in-house developed analytical technology to assess CUZD1 levels in large number of serum samples from patients with a variety of malignant and benign disorders.

CUZD1 protein is selectively expressed in the pancreatic acinar cells147 but it is elusive if its expression is elevated in malignant conditions of the pancreas. We thus investigated if CUZD1 protein is expressed in biopsies from patients with ACC and PDAC.

CUZD1 is one of the main targets of PABs which have recently been arisen as potential biomarkers for the differential diagnosis of CrD over UC 125, 157. The diagnostic utility of CUZD1 autoantibodies in patients with IBD is not clear, mainly because of the lack of high throughput technology aiming their detection. An additional aim of this study was thus to develop and validate novel immunoassays for the detection of CUZD1 autoantibodies on the serum of IBD patients.

Additionally, data regarding the expression of CUZD1 antigen at the site of inflammation in IBD are lacking. For this reason, we investigated if CUZD1 protein can be detected in biopsies from patients with CrD and UC.

There were three main goals of my thesis: 1) develop and validate novel immunoassays for the detection of CUZD1 autoantibodies on the serum of IBD patients; 2) establish the tissue expression pattern of CUZD1 and assess its levels in various biological fluids; 3) Investigating

104 the clinical utility of CUZD1 as a serum-based biomarker in malignant and non-malignant human diseases.

An overview of the main techniques and key findings from the three parts of the study are summarized below:

1) Development of novel immunoassays for detection of CUZD1 autoantibodies in serum of patients with inflammatory bowel diseases

 The extracellular region of CUZD1 protein was produced in a mammalian protein expression system and purified using AIEX chromatography.  Development of ELISAs targeting IgA and IgG CUZD1 autoantibodies using recombinant CUZD1 protein as a solid phase antigen. Each assay was run in parallel with an assay employing BSA as a solid phase antigen instead of recombinant CUZD1 protein. The ratio between the fluorescence in the presence of recombinant CUZD1 versus the presence of BSA was used as the read out.  To establish a cutoff ratio, above which a serum sample would be classified as positive for CUZD1 autoantibodies, 91 serum samples derived from healthy individuals were analyzed with both the IgG and the IgA anti-CUZD1 assays. The values 2 and 1.5 were used as cutoffs for assays IgG and IgA, respectively.  Serum samples from 100 patients with CrD, 100 patients with UC, 129 patients assessed for vADs and 50 control individuals were analyzed. CUZD1 autoantibodies were detected in 12.5% IBD patients including 16% of patients with CrD and in 9% of patients with UC (CrD vs UC, p<0.05) compared to 3.1% patients suspected of having vADs (CrD vs ADs, p<0.05 UC vs ADs, p=0.08). CUZD1 autoantibody positivity was not found to be related to disease location, age of disease onset, or disease phenotype.  These data agree well with standard IIF techniques and can be utilized in multi-centre studies to investigate the diagnostic and clinical utility of CUZD1 autoantibodies.

2) Establishing the tissue expression pattern of CUZD1 protein and assessing its levels in various biological fluids.

105

 Purified recombinant protein was used as an antigen for the generation of highly specific monoclonal antibodies targeting CUZD1.  Monoclonal antibodies were used for the IHC analysis of normal pancreatic tissue which, consistent with previous studies, showed CUZD1 is localized to the acinar cells and the lumen of the acini.  Monoclonal antibodies were used for the development of an ELISA measuring CUZD1 which was employed for the analysis of a panel of human tissue extracts and biological fluids. In accordance with published mRNA data CUZD1 protein was detected in high levels in normal pancreatic tissue and in much lower levels in other tissues. In the biological fluids tested, CUZD1 expression was detected exclusively in pancreatic juice.  The analysis of gel filtration chromatography-derived fractions of pancreatic tissue extract, pancreatic juice and recombinant CUZD1 by using the in-house developed ELISA suggested that the protein exists in high molecular weight protein complexes.

3) Investigating the clinical utility of CUZD1 in malignant and non-malignant human diseases.

 CUZD1 protein levels were measured in a large number of serum samples from patients with various malignant and non-malignant diseases. CUZD1 levels seemed to be elevated in patients with ovarian cysts, in comparison with patients with ovarian cancer (p<0.05). However, the number of patients per category was small and a larger set of samples is warranted for confirmation.  The IHC analysis of inflamed tissues form patients with UC and CrD detected CUZD1 in neuroendocrine cells NC present at the bottom of the intestinal crypts and occasionally in plasma cells of the inflammatory infiltrate. However, the staining detected CUZD1 in the same types of cells in normal tissue as well.  IHC staining of ACC and PDAC biopsies revealed that CUZD1 is highly expressed in ACC tissue while it is not expressed in PDAC tissues.

Overall, in this study we aimed to investigate the clinical utility of CUZD1 in malignant and non-malignant human diseases. In our attempts we discovered that CUZD1 levels seemed to be

106 elevated in patients with ovarian cysts, in comparison with patients with ovarian cancer and that CUZD1 protein is highly expressed in biopsies from patients with ACC while it is completely absent in PDAC tissues. Additionally, we showed that CUZD1 exists in high molecular weight protein complexes. Finally, we described the development of novel immunoassays for the detection of CUZD1 autoantibodies in the serum of patients with IBD. One of the main strengths of our study is that we developed and validated reagents and assays selectively targeting CUZD1 antigen and CUZD1 autoantibodies. This is of paramount importance because the lack of such technology has previously hampered research on CUZD1 antigen and CUZD1 autoantibodies. Thus, by developing and validating these analytical tools not only we obtained several novels findings, but we also paved the way for further studies on this protein.

As with most studies we recognize that there are several limitations of our research approach. Our reagents (antibodies) were raised against (the extracellular part of) recombinant protein and not against endogenously expressed protein; If the recombinant protein is not identical to the native/endogenously expressed protein, then the antibodies that are raised against the recombinant may recognize epitopes found only in the recombinant and not in the endogenously expressed protein. As such, the monoclonal antibodies and thereby our ELISA too would not react with the endogenously expressed protein. To account for this we have expressed the extracellular region of CUZD1, which accounts for the majority of the protein in a mammalian expression system to ensure for proper posttranslational modifications and thereby folding of our protein. As shown in the 3rd chapter of this thesis, our findings in regards to the expression of native CUZD1 are in agreement with the literature and with numerous expression databases which report that CUZD1 is expressed both at the mRNA and protein level in human pancreas. It seems that, despite using recombinant protein for the development of our reagents and assays, we succeeded in our goal to develop analytical antibodies capable of targeting the most common isoform of CUZD1.

For the development of immunoassays targeting CUZD1 autoantibodies (2nd chapter) we used recombinant CUZD1 protein which was just 50 % pure. This could raise the question whether the contaminants or the other proteins present to the solution where CUZD1 protein is diluted, could cause false-positive reactions in the newly developed immunoassays. To address this, we performed a control experiment in which we coated our wells with supernatant from untransfected mammalian cells and analyzed a few patient samples that produced a positive

107 result in our ELISAs. None of these samples produced a positive result in the control assay. We are also aware that the agreement between the data generated by our newly developed immunoassays and IIF methodologies do not agree 100%. Yet, the data are comparable to those previously reported by other groups using IIF, these immunoassays provide an initial platform for fast and easy detection of CUZD1 autoantibodies and after their essential optimization we believe that the agreement will be higher.

CUZD1 has been proposed as biomarker for pancreas related disorders, for ovarian cancer based on its tissue specificity and previously published papers associating CUZD1 with ovarian cancer, respectively. For this reason, in the 4th chapter of this thesis we tested if CUZD1 antigen can pass through a “qualification phase” or in other words if it can be detected in the serum of patients with the aforementioned disorders. We are aware that during this phase the candidate biomarker is usually tested in a binary manner; if it can distinguish the two conditions, the disease and the non-disease. Additionally, for this process tens of gold standard human serum or plasma samples (of reduced biological variation) are required. CUZD1 antigen levels were assessed, instead, in sample sets composed of patients with various malignant and benign disorders including pancreatic cancer and ovarian cancer in an effort to simultaneously evaluate the sensitivity and specificity of this antigen as a candidate marker for the detection of the aforementioned disorders. Although the process that we followed is not exactly the one that has been previously described 174, it has provided us some valuable information. CUZD1 protein was not found to be elevated in patients with pancreatic cancer, while it was shown to be elevated in patients with ovarian cysts in comparison to patients with ovarian cancer. Because the number of samples tested per category is small these findings will further be evaluated in a larger set of samples.

108

5.2 Future directions

In this thesis IHC staining of ACC and PDAC tissue sections showed that CUZD1 is expressed in ACC cases while it is completely absent in PDAC. The analysis of a larger sample of ACC biopsies would be desirable to confirm this finding. As a next step it would be valuable to examine whether the difference of CUZD1 expression between ACC and PDAC tissues is also observed in matched serum samples. If this is the case, CUZD1 protein could serve as a candidate serum biomarker for the differential diagnosis of ACC over PDAC.

It was also shown that CUZD1 protein levels are elevated in the serum of individuals with ovarian cysts as compared to patients with ovarian cancer. However, the number of patients analyzed in each case is small and CUZD1 will need to be further evaluated as potential biomarker for the differential diagnosis of ovarian cysts over ovarian cancer in a larger set of samples.

As previously mentioned CUZD1 protein is one of the antigenic targets of PABs which are often encountered in patients with IBD. It has been previously questioned what evokes the generation of autoimmunity against CUZD1 and whether this, under normal conditions, pancreatic specific protein is detected in the site of intestinal inflammation in the context of IBD. In this work, IHC localization of CUZD1 was performed in affected tissue biopsies from patients with CrD and UC and it showed positive staining in the NC present at the bottom of the intestinal crypts and occasionally in plasma cells of the inflammatory infiltrate. Staining was also observed in the same cell types in normal intestine and as such, it cannot be claimed that the reason behind the generation of immune defense against CUZD1 is simply its expression in the diseased/affected tissue. Therefore, further studies are needed to investigate the mechanism responsible for the production of CUZD1 autoantibodies in IBD.

CUZD1 is a poorly studied protein and in spite of previous studies hinting on its multifunctionality, its common function is still elusive. Since most cellular functions are executed by protein complexes, characterization of the molecular partners of a protein has become a critical part of analyzing its biological function, next to knocking down its expression by RNA interference or studying its subcellular localization. CUZD1 is a transmembrane protein, harboring ectodomains known to be involved in ligand binding and oligomerization processes.

109

As such immunoprecipitation of the CUZD1 protein complex and identifying CUZD1 binding partners may reveal its common biological function in pancreas.

In summary, this thesis encompasses the development and validation of analytical techniques targeting CUZD1 antigen and CUZD1 autoantibodies. The tools targeting CUZD1 protein have facilitated an initial effort to assess the clinical utility of this antigen as a biomarker and they can be exploited for future investigations aiming to investigate the role of CUZD1 in physiology and pathobiology. The assays detecting CUZD1 autoantibodies can be employed to thoroughly investigate their diagnostic and clinical applicability in the context of IBD.

References or Bibliography

1. Busnardo AC, DiDio LJ, Tidrick RT, Thomford NR. History of the pancreas. Am J Surg 146, 539-50 (1983). 2. Fitzgerald PJ. Medical anecdotes concerning some diseases of the pancreas. Monogr. Pathol. 21, 1-29 (1980). 3. Rahier J, Goebbels RM, Henquin JC. Cellular composition of the human diabetic pancreas. Diabetologia 24, 366-71 (1983). 4. Dolensek J, Rupnik MS, Stozer A. Structural similarities and differences between the human and the mouse pancreas. Islets 7, e1024405 (2015). 5. Leung PS. Overview of the Pancreas. The Renin-Angiotensin System: Current Research Progress in The Pancreas: The RAS in the Pancreas. Dordrecht: Springer Netherlands; 2010. p. 3-12. 6. Nina W. Flay FSG. Pancreas, Anatomy. Encyclopedia of Gastroenterology 25-9 (2004). 7. Meyers MA, Charnsangavej C, Oliphant M. Patterns of Spread of Disease from the Pancreas. Meyers' Dynamic Radiology of the Abdomen: Normal and Pathologic Anatomy. New York, NY: Springer New York; 2011. p. 259-74. 8. Mikami Y, Otsuka A, Unno M. Surgical Vascular Anatomy and Histology. In: Beger HG, Matsuno S, Cameron JL, Rau BM, Sunamura M, Schulick RD, editors. Diseases of the Pancreas: Current Surgical Therapy. Berlin, Heidelberg: Springer Berlin Heidelberg; 2008. p. 19-28. 9. Watanabe T, Yaegashi H, Koizumi M, Toyota T, Takahashi T. The lobular architecture of the normal human pancreas: a computer-assisted three-dimensional reconstruction study. Pancreas 15, 48- 52 (1997). 10. Leung PS, Ip SP. Pancreatic acinar cell: its role in acute pancreatitis. Int. J. Biochem. Cell Biol. 38, 1024-30 (2006). 11. Ahlgren U, Pfaff SL, Jessell TM, Edlund T, Edlund H. Independent requirement for ISL1 in formation of pancreatic mesenchyme and islet cells. Nature 385, 257-60 (1997). 12. Edlund H. Pancreatic organogenesis--developmental mechanisms and implications for therapy. Nat Rev Genet 3, 524-32 (2002). 13. Wessells NK, Cohen JH. Early Pancreas Organogenesis: Morphogenesis, Tissue Interactions, and Mass Effects. Dev. Biol. 15, 237-70 (1967). 14. Cano DA, Hebrok M, Zenker M. Pancreatic development and disease. Gastroenterology 132, 745-62 (2007). 15. Edlund H. Pancreas: how to get there from the gut? Curr. Opin. Cell Biol. 11, 663-8 (1999). 16. Kim SK, Hebrok M, Melton DA. Notochord to endoderm signaling is required for pancreas development. Development 124, 4243-52 (1997). 17. Hebrok M, Kim SK, Melton DA. Notochord repression of endodermal Sonic hedgehog permits pancreas development. Genes Dev 12, 1705-13 (1998). 18. Edlund H. Transcribing pancreas. Diabetes 47, 1817-23 (1998). 19. Kaneto H, Matsuoka TA, Miyatsuka T, Kawamori D, Katakami N, Yamasaki Y, et al. PDX-1 functions as a master factor in the pancreas. Frontiers in bioscience : a journal and virtual library 13, 6406-20 (2008). 20. Guz Y, Montminy MR, Stein R, Leonard J, Gamer LW, Wright CV, et al. Expression of murine STF- 1, a putative insulin gene transcription factor, in beta cells of pancreas, duodenal epithelium and pancreatic exocrine and endocrine progenitors during ontogeny. Development 121, 11-8 (1995). 21. Ohlsson H, Karlsson K, Edlund T. IPF1, a homeodomain-containing transactivator of the insulin gene. EMBO J. 12, 4251-9 (1993). 22. Offield MF, Jetton TL, Labosky PA, Ray M, Stein RW, Magnuson MA, et al. PDX-1 is required for pancreatic outgrowth and differentiation of the rostral duodenum. Development 122, 983-95 (1996).

110 111

23. Holland AM, Hale MA, Kagami H, Hammer RE, MacDonald RJ. Experimental control of pancreatic development and maintenance. Proc. Natl. Acad. Sci. U. S. A. 99, 12236-41 (2002). 24. Brissova M, Shiota M, Nicholson WE, Gannon M, Knobel SM, Piston DW, et al. Reduction in pancreatic transcription factor PDX-1 impairs glucose-stimulated insulin secretion. J Biol Chem 277, 11225-32 (2002). 25. Stoffers DA, Zinkin NT, Stanojevic V, Clarke WL, Habener JF. Pancreatic agenesis attributable to a single nucleotide deletion in the human IPF1 gene coding sequence. Nat Genet 15, 106-10 (1997). 26. Yee NS, Yusuff S, Pack M. Zebrafish pdx1 morphant displays defects in pancreas development and digestive organ chirality, and potentially identifies a multipotent pancreas progenitor cell. Genesis 30, 137-40 (2001). 27. Sander M, German MS. The beta cell transcription factors and development of the pancreas. Journal of molecular medicine (Berlin, Germany) 75, 327-40 (1997). 28. Krapp A, Knofler M, Ledermann B, Burki K, Berney C, Zoerkler N, et al. The bHLH protein PTF1- p48 is essential for the formation of the exocrine and the correct spatial organization of the endocrine pancreas. Genes & Development 12, 3752-63 (1998). 29. Kawaguchi Y, Cooper B, Gannon M, Ray M, MacDonald RJ, Wright CV. The role of the transcriptional regulator Ptf1a in converting intestinal to pancreatic progenitors. Nat Genet 32, 128-34 (2002). 30. Masui T, Swift GH, Hale MA, Meredith DM, Johnson JE, Macdonald RJ. Transcriptional autoregulation controls pancreatic Ptf1a expression during development and adulthood. Mol. Cell. Biol. 28, 5458-68 (2008). 31. Afelik S, Chen Y, Pieler T. Combined ectopic expression of Pdx1 and Ptf1a/p48 results in the stable conversion of posterior endoderm into endocrine and exocrine pancreatic tissue. Genes Dev 20, 1441-6 (2006). 32. Jensen J, Pedersen EE, Galante P, Hald J, Heller RS, Ishibashi M, et al. Control of endodermal endocrine development by Hes-1. Nat Genet 24, 36-44 (2000). 33. Apelqvist A, Li H, Sommer L, Beatus P, Anderson DJ, Honjo T, et al. Notch signalling controls pancreatic cell differentiation. Nature 400, 877-81 (1999). 34. Gradwohl G, Dierich A, LeMeur M, Guillemot F. neurogenin3 is required for the development of the four endocrine cell lineages of the pancreas. Proceedings of the National Academy of Sciences 97, 1607-11 (2000). 35. Villasenor A, Chong DC, Cleaver O. Biphasic Ngn3 expression in the developing pancreas. Developmental dynamics : an official publication of the American Association of Anatomists 237, 3270-9 (2008). 36. Gittes GK. Developmental biology of the pancreas: a comprehensive review. Dev. Biol. 326, 4-35 (2009). 37. Schwitzgebel VM, Scheel DW, Conners JR, Kalamaras J, Lee JE, Anderson DJ, et al. Expression of neurogenin3 reveals an islet cell precursor population in the pancreas. Development 127, 3533-42 (2000). 38. Prado CL, Pugh-Bernard AE, Elghazi L, Sosa-Pineda B, Sussel L. Ghrelin cells replace insulin- producing beta cells in two mouse models of pancreas development. Proc. Natl. Acad. Sci. U. S. A. 101, 2924-9 (2004). 39. Bolender RP. Stereological study of pancreatic exocrine cells. Adv Cytopharmacol 2, 99-106 (1974). 40. Low JT, Shukla A, Thorn P. Pancreatic acinar cell: new insights into the control of secretion. Int J Biochem Cell Biol 42, 1586-9 (2010). 41. Yule DI. Pancreatic acinar cells: molecular insight from studies of signal-transduction using transgenic animals. Int J Biochem Cell Biol 42, 1757-61 (2010).

112

42. Pandol SJ, Saluja AK, Imrie CW, Banks PA. Acute pancreatitis: bench to the bedside. Gastroenterology 132, 1127-51 (2007). 43. Petersen OH. Stimulus-secretion coupling: cytoplasmic calcium signals and the control of ion channels in exocrine acinar cells. J Physiol 448, 1-51 (1992). 44. Leung PS. Physiology of the Pancreas. The Renin-Angiotensin System: Current Research Progress in The Pancreas: The RAS in the Pancreas. Dordrecht: Springer Netherlands; 2010. p. 13-27. 45. Williams JA. Regulation of acinar cell function in the pancreas. Curr Opin Gastroenterol 26, 478- 83 (2010). 46. Wu SV, Harikumar KG, Burgess RJ, Reeve JR, Jr., Miller LJ. Effects of cholecystokinin-58 on type 1 cholecystokinin receptor function and regulation. Am J Physiol Gastrointest Liver Physiol 295, G641-7 (2008). 47. Wasle B, Edwardson JM. The regulation of exocytosis in the pancreatic acinar cell. Cell Signal 14, 191-7 (2002). 48. Matthews EK, Petersen OH, Williams JA. Pancreatic acinar cells: acetylcholine-induced membrane depolarization, calcium efflux and amylase release. J. Physiol. 234, 689-701 (1973). 49. Thorn P, Lawrie AM, Smith PM, Gallacher DV, Petersen OH. Local and global cytosolic Ca2+ oscillations in exocrine cells evoked by agonists and inositol trisphosphate. Cell 74, 661-8 (1993). 50. Thorn P, Lawrie AM, Smith PM, Gallacher DV, Petersen OH. Ca2+ oscillations in pancreatic acinar cells: spatiotemporal relationships and functional implications. Cell calcium 14, 746-57 (1993). 51. Grapin-Botton A. Ductal cells of the pancreas. Int J Biochem Cell Biol 37, 504-10 (2005). 52. Hootman SR, de Ondarza J. Overview of pancreatic duct physiology and pathophysiology. Digestion 54, 323-30 (1993). 53. Poulsen JH, Fischer H, Illek B, Machen TE. Bicarbonate conductance and pH regulatory capability of cystic fibrosis transmembrane conductance regulator. Proc Natl Acad Sci U S A 91, 5340-4 (1994). 54. Welsh MJ, Anderson MP, Rich DP, Berger HA, Denning GM, Ostedgaard LS, et al. Cystic fibrosis transmembrane conductance regulator: a chloride channel with novel regulation. Neuron 8, 821-9 (1992). 55. Novak I, Greger R. Properties of the luminal membrane of isolated perfused rat pancreatic ducts. Effect of cyclic AMP and blockers of chloride transport. Pflugers Arch 411, 546-53 (1988). 56. Stuenkel EL, Machen TE, Williams JA. pH regulatory mechanisms in rat pancreatic ductal cells. Am J Physiol 254, G925-30 (1988). 57. Bonner-Weir S, Sharma A. Pancreatic stem cells. J Pathol 197, 519-26 (2002). 58. Heremans Y, Van De Casteele M, in't Veld P, Gradwohl G, Serup P, Madsen O, et al. Recapitulation of embryonic neuroendocrine differentiation in adult human pancreatic duct cells expressing neurogenin 3. J Cell Biol 159, 303-12 (2002). 59. Engelking LR. Physiology of the endocrine pancreas. Semin Vet Med Surg (Small Anim) 12, 224-9 (1997). 60. Steiner DF, Rubenstein AH. Proinsulin C-peptide--biological activity? Science 277, 531-2 (1997). 61. Schuit FC, Huypens P, Heimberg H, Pipeleers DG. Glucose sensing in pancreatic beta-cells: a model for the study of other glucose-regulated cells in gut, pancreas, and hypothalamus. Diabetes 50, 1- 11 (2001). 62. Wookey PJ, Lutz TA, Andrikopoulos S. Amylin in the periphery II: An updated mini-review. ScientificWorldJournal 6, 1642-55 (2006). 63. Johnson KH, O'Brien TD, Betsholtz C, Westermark P. Islet amyloid, islet-amyloid polypeptide, and diabetes mellitus. N Engl J Med 321, 513-8 (1989). 64. Date Y, Murakami N, Toshinai K, Matsukura S, Niijima A, Matsuo H, et al. The role of the gastric afferent vagal nerve in ghrelin-induced feeding and growth hormone secretion in rats. Gastroenterology 123, 1120-8 (2002).

113

65. De Vriese C, Delporte C. Ghrelin: a new peptide regulating growth hormone release and food intake. Int J Biochem Cell Biol 40, 1420-4 (2008). 66. Andralojc KM, Mercalli A, Nowak KW, Albarello L, Calcagno R, Luzi L, et al. Ghrelin-producing epsilon cells in the developing and adult human pancreas. Diabetologia 52, 486-93 (2009). 67. Wahren J. Influence of somatostatin on carbohydrate disposal and absorption in diabetes mellitus. Lancet 2, 1213-6 (1976). 68. Tong J, Utzschneider KM, Carr DB, Zraika S, Udayasankar J, Gerchman F, et al. Plasma pancreatic polypeptide levels are associated with differences in body fat distribution in human subjects. Diabetologia 50, 439-42 (2007). 69. von Schonfeld J, Goebell H, Muller MK. The islet-acinar axis of the pancreas. Int J Pancreatol 16, 131-40 (1994). 70. Barreto SG, Carati CJ, Toouli J, Saccone GT. The islet-acinar axis of the pancreas: more than just insulin. Am J Physiol Gastrointest Liver Physiol 299, G10-22 (2010). 71. Patel R, Yago MD, Manas M, Victoria EM, Shervington A, Singh J. Mechanism of exocrine pancreatic insufficiency in streptozotocin-induced diabetes mellitus in rat: effect of cholecystokinin- octapeptide. Mol Cell Biochem 261, 83-9 (2004). 72. Chey WY, Shay H, Shuman CR. External Pancreatic Secretion in Diabetes Mellitus. Ann Intern Med 59, 812-21 (1963). 73. Hellman B, Wallgren A, Petersson B. Cytological characteristics of the exocrine pancreatic cells with regard to their position in relation to the islets of Langerhans. A study in normal and obese- hyperglycaemic mice. Acta Endocrinol (Copenh) 39, 465-73 (1962). 74. Murakami T, Fujita T, Taguchi T, Nonaka Y, Orita K. The blood vascular bed of the human pancreas, with special reference to the insulo-acinar portal system. Scanning electron microscopy of corrosion casts. Arch Histol Cytol 55, 381-95 (1992). 75. Wallgren A, Hellman B. Influence of the islet A and B cells on the exocrine pancreatic tissue in the duck. Acta Anat (Basel) 48, 137-41 (1962). 76. Kramer MF, Tan HT. The peri-insular acini of the pancreas of the rat. Z Zellforsch Mikrosk Anat 86, 163-70 (1968). 77. Malaisse-Lagae F, Ravazzola M, Robberecht P, Vandermeers A, Malaisse WJ, Orci L. Exocrine pancreas: evidence for topographic partition of secretory function. Science 190, 795-7 (1975). 78. Mueller MK, Scheck T, Dreesmann V, Miodonski A, Goebell H. GIP potentiates CCK stimulated pancreatic enzyme secretion: correlation of anatomical structures with the effects of GIP and CCK on amylase secretion. Pancreas 2, 106-13 (1987). 79. Saito A, Williams JA, Kanno T. Potentiation of cholecystokinin-induced exocrine secretion by both exogenous and endogenous insulin in isolated and perfused rat pancreata. J Clin Invest 65, 777-82 (1980). 80. Boden G, Sivitz MC, Owen OE, Essa-Koumar N, Landor JH. Somatostatin suppresses secretin and pancreatic exocrine secretion. Science 190, 163-5 (1975). 81. Liao Z, Li ZS, Lu Y, Wang WZ. Microinjection of exogenous somatostatin in the dorsal vagal complex inhibits pancreatic secretion via somatostatin receptor-2 in rats. Am J Physiol Gastrointest Liver Physiol 292, G746-52 (2007). 82. Pandiri AR. Overview of exocrine pancreatic pathobiology. Toxicol Pathol 42, 207-16 (2014). 83. Chan YC, Leung PS. Angiotensin II type 1 receptor-dependent nuclear factor-kappaB activation- mediated proinflammatory actions in a rat model of obstructive acute pancreatitis. J Pharmacol Exp Ther 323, 10-8 (2007). 84. Banks PA. Epidemiology, natural history, and predictors of disease outcome in acute and chronic pancreatitis. Gastrointest Endosc 56, S226-30 (2002).

114

85. Sekimoto M, Takada T, Kawarada Y, Hirata K, Mayumi T, Yoshida M, et al. JPN Guidelines for the management of acute pancreatitis: epidemiology, etiology, natural history, and outcome predictors in acute pancreatitis. J Hepatobiliary Pancreat Surg 13, 10-24 (2006). 86. Leung PS. Common Pancreatic Disease. The Renin-Angiotensin System: Current Research Progress in The Pancreas: The RAS in the Pancreas. Dordrecht: Springer Netherlands; 2010. p. 29-51. 87. Singh VP, Saluja AK, Bhagat L, van Acker GJ, Song AM, Soltoff SP, et al. Phosphatidylinositol 3- kinase-dependent activation of trypsinogen modulates the severity of acute pancreatitis. J Clin Invest 108, 1387-95 (2001). 88. Hammer HF. An update on pancreatic pathophysiology (do we have to rewrite pancreatic pathophysiology?). Wien Med Wochenschr 164, 57-62 (2014). 89. Norman J. The role of cytokines in the pathogenesis of acute pancreatitis. Am J Surg 175, 76-83 (1998). 90. Makola D, Krenitsky J, Parrish CR. Enteral feeding in acute and chronic pancreatitis. Gastrointest Endosc Clin N Am 17, 747-64 (2007). 91. Pederzoli P, Bassi C, Vesentini S, Campedelli A. A randomized multicenter clinical trial of antibiotic prophylaxis of septic complications in acute necrotizing pancreatitis with imipenem. Surg Gynecol Obstet 176, 480-3 (1993). 92. Fokas E, O'Neill E, Gordon-Weeks A, Mukherjee S, McKenna WG, Muschel RJ. Pancreatic ductal adenocarcinoma: From genetics to biology to radiobiology to oncoimmunology and all the way back to the clinic. Biochim Biophys Acta 1855, 61-82 (2015). 93. Hruban RH, Adsay NV. Molecular classification of neoplasms of the pancreas. Hum Pathol 40, 612-23 (2009). 94. Lewis BC. Development of the pancreas and pancreatic cancer. Endocrinol Metab Clin North Am 35, 397-404, xi (2006). 95. Siegel R, Ward E, Brawley O, Jemal A. Cancer statistics, 2011: the impact of eliminating socioeconomic and racial disparities on premature cancer deaths. CA Cancer J Clin 61, 212-36 (2011). 96. Distler M, Aust D, Weitz J, Pilarsky C, Grutzmann R. Precursor lesions for sporadic pancreatic cancer: PanIN, IPMN, and MCN. Biomed Res Int 2014, 474905 (2014). 97. Guo J, Xie K, Zheng S. Molecular Biomarkers of Pancreatic Intraepithelial Neoplasia and Their Implications in Early Diagnosis and Therapeutic Intervention of Pancreatic Cancer. Int J Biol Sci 12, 292- 301 (2016). 98. Ryan DP, Hong TS, Bardeesy N. Pancreatic adenocarcinoma. N Engl J Med 371, 1039-49 (2014). 99. Bond-Smith G, Banga N, Hammond TM, Imber CJ. Pancreatic adenocarcinoma. BMJ 344, e2476 (2012). 100. Dunphy EP. Pancreatic cancer: a review and update. Clin J Oncol Nurs 12, 735-41 (2008). 101. Campa D, Mergarten B, De Vivo I, Boutron-Ruault MC, Racine A, Severi G, et al. Leukocyte telomere length in relation to pancreatic cancer risk: a prospective study. Cancer Epidemiol Biomarkers Prev 23, 2447-54 (2014). 102. Miyamoto Y, Maitra A, Ghosh B, Zechner U, Argani P, Iacobuzio-Donahue CA, et al. Notch mediates TGF alpha-induced changes in epithelial differentiation during pancreatic tumorigenesis. Cancer Cell 3, 565-76 (2003). 103. Danovi SA, Wong HH, Lemoine NR. Targeted therapies for pancreatic cancer. Br Med Bull 87, 97- 130 (2008). 104. Sebens S, Arlt A, Schafer H. NF-kappaB as a molecular target in the therapy of pancreatic carcinoma. Recent Results Cancer Res 177, 151-64 (2008). 105. American Diabetes A. Diagnosis and classification of diabetes mellitus. Diabetes Care 32 Suppl 1, S62-7 (2009). 106. Vijan S. In the clinic. Type 2 diabetes. Ann Intern Med 152, ITC31-15; quiz ITC316 (2010).

115

107. Gulden E, Wong FS, Wen L. The gut microbiota and Type 1 Diabetes. Clin Immunol 159, 143-53 (2015). 108. Pociot F, McDermott MF. Genetics of type 1 diabetes mellitus. Genes Immun 3, 235-49 (2002). 109. Eisenbarth GS. Update in type 1 diabetes. J Clin Endocrinol Metab 92, 2403-7 (2007). 110. Song SH, Hardisty CA. Early-onset Type 2 diabetes mellitus: an increasing phenomenon of elevated cardiovascular risk. Expert Rev Cardiovasc Ther 6, 315-22 (2008). 111. Phillips BE, Giannoukakis N, Trucco M. Dendritic cell mediated therapy for immunoregulation of type 1 diabetes mellitus. Pediatr Endocrinol Rev 5, 873-9 (2008). 112. Stratta RJ, Alloway RR. Pancreas transplantation for diabetes mellitus: a guide to recipient selection and optimum immunosuppression. BioDrugs 10, 347-57 (1998). 113. Shapiro AM, Ricordi C, Hering BJ, Auchincloss H, Lindblad R, Robertson RP, et al. International trial of the Edmonton protocol for islet transplantation. N Engl J Med 355, 1318-30 (2006). 114. Reimann M, Bonifacio E, Solimena M, Schwarz PE, Ludwig B, Hanefeld M, et al. An update on preventive and regenerative therapies in diabetes mellitus. Pharmacol Ther 121, 317-31 (2009). 115. Torgerson JS, Hauptman J, Boldrin MN, Sjostrom L. XENical in the prevention of diabetes in obese subjects (XENDOS) study: a randomized study of orlistat as an adjunct to lifestyle changes for the prevention of type 2 diabetes in obese patients. Diabetes Care 27, 155-61 (2004). 116. Abraham C, Cho JH. Inflammatory bowel disease. N. Engl. J. Med. 361, 2066-78 (2009). 117. Petkau JM, Eksteen B. Selective biologics for ulcerative colitis and Crohn's disease - clinical utility of vedolizumab. Biologics : targets & therapy 10, 33-52 (2016). 118. Cho JH, Weaver CT. The genetics of inflammatory bowel disease. Gastroenterology 133, 1327-39 (2007). 119. Hampe J, Franke A, Rosenstiel P, Till A, Teuber M, Huse K, et al. A genome-wide association scan of nonsynonymous SNPs identifies a susceptibility variant for Crohn disease in ATG16L1. Nat Genet 39, 207-11 (2007). 120. Barrett JC, Hansoul S, Nicolae DL, Cho JH, Duerr RH, Rioux JD, et al. Genome-wide association defines more than 30 distinct susceptibility loci for Crohn's disease. Nat Genet 40, 955-62 (2008). 121. Mowat C, Cole A, Windsor A, Ahmad T, Arnott I, Driscoll R, et al. Guidelines for the management of inflammatory bowel disease in adults. Gut 60, 571-607 (2011). 122. Burger D, Travis S. Conventional medical management of inflammatory bowel disease. Gastroenterology 140, 1827-37 e2 (2011). 123. Lewis JD. The utility of biomarkers in the diagnosis and therapy of inflammatory bowel disease. Gastroenterology 140, 1817-26 e2 (2011). 124. Stocker W, Otte M, Ulrich S, Normann D, Finkbeiner H, Stocker K, et al. Autoimmunity to pancreatic juice in Crohn's disease. Results of an autoantibody screening in patients with chronic inflammatory bowel disease. Scand. J. Gastroenterol. Suppl. 139, 41-52 (1987). 125. Bogdanos DP, Rigopoulou EI, Smyk DS, Roggenbuck D, Reinhold D, Forbes A, et al. Diagnostic value, clinical utility and pathogenic significance of reactivity to the molecular targets of Crohn's disease specific-pancreatic autoantibodies. Autoimmun Rev 11, 143-8 (2011). 126. Joossens S, Vermeire S, Van Steen K, Godefridis G, Claessens G, Pierik M, et al. Pancreatic autoantibodies in inflammatory bowel disease. Inflamm. Bowel Dis. 10, 771-7 (2004). 127. Seibold F, Scheurlen M, Muller A, Jenss H, Weber P. Impaired pancreatic function in patients with Crohn's disease with and without pancreatic autoantibodies. J. Clin. Gastroenterol. 22, 202-6 (1996). 128. Seibold F, Weber P, Jenss H, Wiedmann KH. Antibodies to a trypsin sensitive pancreatic antigen in chronic inflammatory bowel disease: specific markers for a subgroup of patients with Crohn's disease. Gut 32, 1192-7 (1991). 129. Zhou G, Song Y, Yang W, Guo Y, Fang L, Chen Y, et al. ASCA, ANCA, ALCA and Many More: Are They Useful in the Diagnosis of Inflammatory Bowel Disease? Dig. Dis. 34, 90-7 (2016).

116

130. Bonneau J, Dumestre-Perard C, Rinaudo-Gaujous M, Genin C, Sparrow M, Roblin X, et al. Systematic review: new serological markers (anti-glycan, anti-GP2, anti-GM-CSF Ab) in the prediction of IBD patient outcomes. Autoimmun Rev 14, 231-45 (2015). 131. Roggenbuck D, Reinhold D, Werner L, Schierack P, Bogdanos DP, Conrad K. Glycoprotein 2 antibodies in Crohn's disease. Adv. Clin. Chem. 60, 187-208 (2013). 132. Pavlidis P, Komorowski L, Teegen B, Liaskos C, Koutsoumpas AL, Smyk DS, et al. Diagnostic and clinical significance of Crohn's disease-specific pancreatic anti-GP2 and anti-CUZD1 antibodies. Clin. Chem. Lab. Med. 54, 249-56 (2016). 133. Roggenbuck D, Hausdorf G, Martinez-Gamboa L, Reinhold D, Buttner T, Jungblut PR, et al. Identification of GP2, the major zymogen granule membrane glycoprotein, as the autoantigen of pancreatic antibodies in Crohn's disease. Gut 58, 1620-8 (2009). 134. Roggenbuck D, Bogdanos DP, Conrad K. Loss of tolerance to one or two major targets in Crohn’s disease or just cross reactivity. J Crohns Colitis (2012). 135. Hase K, Kawano K, Nochi T, Pontes GS, Fukuda S, Ebisawa M, et al. Uptake through glycoprotein 2 of FimH+ bacteria by M cells initiates mucosal immune response. Nature 462, 226-30 (2009). 136. Holzl MA, Hofer J, Kovarik JJ, Roggenbuck D, Reinhold D, Goihl A, et al. The zymogen granule protein 2 (GP2) binds to scavenger receptor expressed on endothelial cells I (SREC-I). Cell. Immunol. 267, 88-93 (2011). 137. Prassas I, Chrystoja CC, Makawita S, Diamandis EP. Bioinformatic identification of proteins with tissue-specific expression for biomarker discovery. BMC Med. 10, 39 (2012). 138. Chen D, Xu X, Zhu LJ, Angervo M, Li Q, Bagchi MK, et al. Cloning and uterus/oviduct-specific expression of a novel estrogen-regulated gene (ERG1). J. Biol. Chem. 274, 32215-24 (1999). 139. Huynh H, Ng CY, Lim KB, Ong CK, Ong CS, Tran E, et al. Induction of UO-44 gene expression by tamoxifen in the rat uterus and ovary. Endocrinology 142, 2985-95 (2001). 140. Liaskos C, Rigopoulou EI, Orfanidou T, Bogdanos DP, Papandreou CN. CUZD1 and anti-CUZD1 antibodies as markers of cancer and inflammatory bowel diseases. Clin. Dev. Immunol. 2013, 968041 (2013). 141. Leong CT, Ng CY, Ong CK, Ng CP, Ma ZS, Nguyen TH, et al. Molecular cloning, characterization and isolation of novel spliced variants of the human ortholog of a rat estrogen-regulated membrane- associated protein, UO-44. Oncogene 23, 5707-18 (2004). 142. Bork P, Beckmann G. The CUB domain. A widespread module in developmentally regulated proteins. J. Mol. Biol. 231, 539-45 (1993). 143. Bork P, Sander C. A large domain common to sperm receptors (Zp2 and Zp3) and TGF-beta type III receptor. FEBS Lett. 300, 237-40 (1992). 144. Gaboriaud C, Gregory-Pauron L, Teillet F, Thielens NM, Bally I, Arlaud GJ. Structure and properties of the Ca(2+)-binding CUB domain, a widespread ligand-recognition unit involved in major biological functions. Biochem. J. 439, 185-93 (2011). 145. Jovine L, Darie CC, Litscher ES, Wassarman PM. Zona pellucida domain proteins. Annu. Rev. Biochem. 74, 83-114 (2005). 146. Kasik JW. A cDNA cloned from pregnant mouse uterus exhibits temporo-spatial expression and predicts a novel protein. Biochem. J. 330 ( Pt 2), 947-50 (1998). 147. Imamura T, Asada M, Vogt SK, Rudnick DA, Lowe ME, Muglia LJ. Protection from pancreatitis by the zymogen granule membrane protein integral membrane-associated protein-1. J. Biol. Chem. 277, 50725-33 (2002). 148. Mapes J, Li Q, Kannan A, Anandan L, Laws M, Lydon JP, et al. CUZD1 is a critical mediator of the JAK/STAT5 signaling pathway that controls mammary gland development during pregnancy. PLoS genetics 13, e1006654 (2017).

117

149. Mollenhauer J, Wiemann S, Scheurlen W, Korn B, Hayashi Y, Wilgenbus KK, et al. DMBT1, a new member of the SRCR superfamily, on chromosome 10q25.3-26.1 is deleted in malignant brain tumours. Nat Genet 17, 32-9 (1997). 150. Madsen J, Tornoe I, Nielsen O, Lausen M, Krebs I, Mollenhauer J, et al. CRP-ductin, the mouse homologue of gp-340/deleted in malignant brain tumors 1 (DMBT1), binds gram-positive and gram- negative bacteria and interacts with lung surfactant protein D. Eur. J. Immunol. 33, 2327-36 (2003). 151. Innan H. Population genetic models of duplicated genes. Genetica 137, 19-37 (2009). 152. Barretina J, Caponigro G, Stransky N, Venkatesan K, Margolin AA, Kim S, et al. The Cancer Cell Line Encyclopedia enables predictive modelling of anticancer drug sensitivity. Nature 483, 603-7 (2012). 153. Hoops TC, Ivanov I, Cui Z, Colomer-Gould V, Rindler MJ. Incorporation of the pancreatic membrane protein GP-2 into secretory granules in exocrine but not endocrine cells. J. Biol. Chem. 268, 25694-705 (1993). 154. Ronzio RA, Kronquist KE, Lewis DS, MacDonald RJ, Mohrlok SH, O'Donnell JJ, Jr. Glycoprotein synthesis in the adult rat pancreas. IV. Subcellular distribution of membrane glycoproteins. Biochim. Biophys. Acta 508, 65-84 (1978). 155. Havinga JR, Strous GJ, Poort C. Biosynthesis of the major glycoprotein associated with zymogen- granule membranes in the pancreas. Eur. J. Biochem. 133, 449-54 (1983). 156. LeBel D, Beattie M. The major protein of pancreatic zymogen granule membranes (GP-2) is anchored via covalent bonds to phosphatidylinositol. Biochemical and Biophysical Research Communications 154, 818-23 (1988). 157. Komorowski L, Teegen B, Probst C, Aulinger-Stocker K, Sina C, Fellermann K, et al. Autoantibodies against exocrine pancreas in Crohn's disease are directed against two antigens: the glycoproteins CUZD1 and GP2. J Crohns Colitis 7, 780-90 (2013). 158. Leong CTC, Ong CK, Tay SK, Huynh H. Silencing expression of UO-44 (CUZD1) using small interfering RNA sensitizes human ovarian cancer cells to cisplatin in vitro. Oncogene 26, 870-80 (2006). 159. Michaels MA, Jendrek ST, Korf T, Nitzsche T, Teegen B, Komorowski L, et al. Pancreatic Autoantibodies Against CUZD1 and GP2 Are Associated with Distinct Clinical Phenotypes of Crohn's Disease. Inflamm. Bowel Dis. 21, 2864-72 (2015). 160. Bossuyt X. Serologic markers in inflammatory bowel disease. Clin. Chem. 52, 171-81 (2006). 161. Papp M, Lakatos PL. Serological studies in inflammatory bowel disease: how important are they? Curr Opin Gastroenterol 30, 359-64 (2014). 162. Lakatos PL, Altorjay I, Szamosi T, Palatka K, Vitalis Z, Tumpek J, et al. Pancreatic autoantibodies are associated with reactivity to microbial antibodies, penetrating disease behavior, perianal disease, and extraintestinal manifestations, but not with NOD2/CARD15 or TLR4 genotype in a Hungarian IBD cohort. Inflamm. Bowel Dis. 15, 365-74 (2009). 163. Klebl FH, Bataille F, Huy C, Hofstadter F, Scholmerich J, Rogler G. Association of antibodies to exocrine pancreas with subtypes of Crohn's disease. Eur. J. Gastroenterol. Hepatol. 17, 73-7 (2005). 164. Drabovich AP, Dimitromanolakis A, Saraon P, Soosaipillai A, Batruch I, Mullen B, et al. Differential diagnosis of azoospermia with proteomic biomarkers ECM1 and TEX101 quantified in seminal plasma. Sci. Transl. Med. 5, 212ra160 (2013). 165. Christopoulos TK, Diamandis EP. Enzymatically amplified time-resolved fluorescence immunoassay with terbium chelates. Anal. Chem. 64, 342-6 (1992). 166. Silverberg MS, Satsangi J, Ahmad T, Arnott ID, Bernstein CN, Brant SR, et al. Toward an integrated clinical, molecular and serological classification of inflammatory bowel disease: report of a Working Party of the 2005 Montreal World Congress of Gastroenterology. Can. J. Gastroenterol. 19 Suppl A, 5A-36A (2005). 167. Prassas I, Brinc D, Farkona S, Leung F, Dimitromanolakis A, Chrystoja CC, et al. False biomarker discovery due to reactivity of a commercial ELISA for CUZD1 with cancer antigen CA125. Clin. Chem. 60, 381-8 (2014).

118

168. Farkona S, Soosaipillai A, Filippou P, Liaskos C, Bogdanos DP, Diamandis EP, et al. Novel immunoassays for detection of CUZD1 autoantibodies in serum of patients with inflammatory bowel diseases. Clin. Chem. Lab. Med. 55, 1574-81 (2017). 169. Korbakis D, Prassas I, Brinc D, Batruch I, Krastins B, Lopez MF, et al. Delineating monoclonal antibody specificity by mass spectrometry. J. Proteomics 114, 115-24 (2015). 170. Filippou P, Korbakis D, Farkona S, Soosaipillai A, Karakosta T, Diamandis EP. A new enzyme- linked immunosorbent assay (ELISA) for human free and bound kallikrein 9. Clin. Proteomics 14, 4 (2017). 171. Shaw JL, Grass L, Sotiropoulou G, Diamandis EP. Development of an immunofluorometric assay for human kallikrein 15 (KLK15) and identification of KLK15 in tissues and biological fluids. Clin. Biochem. 40, 104-10 (2007). 172. Karakosta TD, Soosaipillai A, Diamandis EP, Batruch I, Drabovich AP. Quantification of Human Kallikrein-Related Peptidases in Biological Fluids by Multiplatform Targeted Mass Spectrometry Assays. Mol. Cell. Proteomics 15, 2863-76 (2016). 173. Farkona S, Soosaipillai A, Filippou P, Korbakis D, Serra S, Ruckert F, et al. Generation of monoclonal antibodies and development of an immunofluorometric assay for the detection of CUZD1 in tissues and biological fluids. Clin. Biochem. (2017). 174. Rifai N, Gillette MA, Carr SA. Protein biomarker discovery and validation: the long and uncertain path to clinical utility. Nat. Biotechnol. 24, 971-83 (2006).

119

120