<%BANNER%>

Loss of CSF2 Epigenetic Regulation through Aberrant GM-CSF Induced STAT5 Signaling Contributes to Type 1 Diabetic Myeloi...

Permanent Link: http://ufdc.ufl.edu/UFE0022834/00001

Material Information

Title: Loss of CSF2 Epigenetic Regulation through Aberrant GM-CSF Induced STAT5 Signaling Contributes to Type 1 Diabetic Myeloid Cell Dysfunction
Physical Description: 1 online resource (63 p.)
Language: english
Creator: Garrigan, Erin
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: autoimmunity, cytokine, diabetes, gmcsf, macrophage, monocyte, stat5
Medicine -- Dissertations, Academic -- UF
Genre: Medical Sciences thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Defective myeloid differentiation and activation contribute to immunopathogenesis of autoimmune type 1 diabetes (T1D) in humans and non-obese (NOD) mice by retarding their antigen presenting cell function, altering their cytokine production, and promoting a pro-inflammatory microenvironment through aberrant expression of prostaglandin synthase, PGS2 (COX2). Monocytes of at-risk/T1D humans and macrophages of the NOD mouse have elevated autocrine GM-CSF production and persistent signal transduction/activator of transcription (STAT5) phosphorylation. GM-CSF activates STAT5 binding at its own gene promoter within the Idd4.3 region, as well as at the enhancer of the COX2 gene, Ptgs2. Since STAT5 acts as adaptor protein for either deacetylase or acetylase enzymes mediating epigenetic chromatin modification, we hypothesize that the loss of Csf2 gene regulation in autoimmune cells is mediated by changes in epigenetic control perpetuated by GM-CSF activating persistent STAT5 then also genes such as Ptgs2. Our specific aims are to characterize epigenetic modification at regulatory sites upstream of the Csf2 and Ptgs2 genes mediated by STAT5 binding and determine if such changes are related to GM-CSF-induced 1) STAT5 function and 2) Csf2/Ptgs2 gene expression in autoimmune myeloid cells, and 3) the genetic sequence of regulatory regions diabetes susceptibility loci. A small volume flow cytometric analysis was developed to measure STAT5Ptyr levels in T1D patients and NOD mice for use as a potentially clinically useful biomarker assay for autoimmune T1D susceptibility. On average, T1D patient monocytes STAT5Ptyr levels were two fold higher than healthy controls. However when these data are analyzed by gender, the results showed a statistically significant female bias for this biomarker in the subject group. Our phenotypic and genotypic studies of the NOD mouse and subsequent congenic derivatives of this model narrow diabetes susceptibility attributed to Chr. 11 and GM-CSF overproduction with persistent STAT5 phosphorylation to a small (200bp) un-transcribed regulatory region within the Csf2 promoter. In the NOD, this region contains a loss of STAT6 binding site (potential anti-inflammatory mediator) and a gain of a STAT5 binding site. Homologous sites were not found in the T1D patient versus healthy control, suggesting alternate regulatory mechanisms contributing to the persistent STAT5 phosphorylation in human autoimmune monocytes. Our results support the potential of epigenetic control as a mechanism underlying the chromosomal dysregulation seen in autoimmune diseases and indicate that epigenetic control of gene expression can provide a new avenue for discovery of potential prevention/therapeutic intervention targets.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Erin Garrigan.
Thesis: Thesis (M.S.)--University of Florida, 2008.
Local: Adviser: Atkinson, Mark A.
Local: Co-adviser: Litherland, Sally A.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022834:00001

Permanent Link: http://ufdc.ufl.edu/UFE0022834/00001

Material Information

Title: Loss of CSF2 Epigenetic Regulation through Aberrant GM-CSF Induced STAT5 Signaling Contributes to Type 1 Diabetic Myeloid Cell Dysfunction
Physical Description: 1 online resource (63 p.)
Language: english
Creator: Garrigan, Erin
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: autoimmunity, cytokine, diabetes, gmcsf, macrophage, monocyte, stat5
Medicine -- Dissertations, Academic -- UF
Genre: Medical Sciences thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Defective myeloid differentiation and activation contribute to immunopathogenesis of autoimmune type 1 diabetes (T1D) in humans and non-obese (NOD) mice by retarding their antigen presenting cell function, altering their cytokine production, and promoting a pro-inflammatory microenvironment through aberrant expression of prostaglandin synthase, PGS2 (COX2). Monocytes of at-risk/T1D humans and macrophages of the NOD mouse have elevated autocrine GM-CSF production and persistent signal transduction/activator of transcription (STAT5) phosphorylation. GM-CSF activates STAT5 binding at its own gene promoter within the Idd4.3 region, as well as at the enhancer of the COX2 gene, Ptgs2. Since STAT5 acts as adaptor protein for either deacetylase or acetylase enzymes mediating epigenetic chromatin modification, we hypothesize that the loss of Csf2 gene regulation in autoimmune cells is mediated by changes in epigenetic control perpetuated by GM-CSF activating persistent STAT5 then also genes such as Ptgs2. Our specific aims are to characterize epigenetic modification at regulatory sites upstream of the Csf2 and Ptgs2 genes mediated by STAT5 binding and determine if such changes are related to GM-CSF-induced 1) STAT5 function and 2) Csf2/Ptgs2 gene expression in autoimmune myeloid cells, and 3) the genetic sequence of regulatory regions diabetes susceptibility loci. A small volume flow cytometric analysis was developed to measure STAT5Ptyr levels in T1D patients and NOD mice for use as a potentially clinically useful biomarker assay for autoimmune T1D susceptibility. On average, T1D patient monocytes STAT5Ptyr levels were two fold higher than healthy controls. However when these data are analyzed by gender, the results showed a statistically significant female bias for this biomarker in the subject group. Our phenotypic and genotypic studies of the NOD mouse and subsequent congenic derivatives of this model narrow diabetes susceptibility attributed to Chr. 11 and GM-CSF overproduction with persistent STAT5 phosphorylation to a small (200bp) un-transcribed regulatory region within the Csf2 promoter. In the NOD, this region contains a loss of STAT6 binding site (potential anti-inflammatory mediator) and a gain of a STAT5 binding site. Homologous sites were not found in the T1D patient versus healthy control, suggesting alternate regulatory mechanisms contributing to the persistent STAT5 phosphorylation in human autoimmune monocytes. Our results support the potential of epigenetic control as a mechanism underlying the chromosomal dysregulation seen in autoimmune diseases and indicate that epigenetic control of gene expression can provide a new avenue for discovery of potential prevention/therapeutic intervention targets.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Erin Garrigan.
Thesis: Thesis (M.S.)--University of Florida, 2008.
Local: Adviser: Atkinson, Mark A.
Local: Co-adviser: Litherland, Sally A.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022834:00001


This item has the following downloads:


Full Text

PAGE 1

1 LOSS OF CSF2 EPIGENETIC REGULATION THROUGH ABERRANT GM CSF INDUCED STAT5 SIGNALING CONTRIBUTES TO TYPE 1 DIABETIC MYELOID CELL DYS FUNCTION By ERIN LEE GARRIGAN A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2008

PAGE 2

2 2008 E rin L ee G arrigan

PAGE 3

3 To Sally, Mom, Dad and Gran

PAGE 4

4 ACKNOWLEDGMENTS It is with deep est gratitude that I acknowledge my family, friends and colleagues who contributed profoundly to my professional and personal lives. I wish to sincerely thank my mentors Dr. Sally Litherland and Dr. Mark Atkinson for their continued su pport and generosity of time, guidance materials and insight Without their patience encouragement of creativity and open exchange of thoughts and ideas this degre e would not have been possible. In addition to my mentors I wish to gratefully acknowledge my advisory committe e members, Dr. Ammon Peck and Dr. Peter McGuire Their advice and support has been invaluable throughout this process. I would like to thank the former members of the Litherland Lab, especially Fedi Seydel Bryan Stutevoss and Nicole Belkin T ogether we made a great team, shared many laughs and learned a great deal from one another I would like to extend a special thank you to the members of the Atkinson Lab for welcoming me fully, teaching me and lending a hand when I needed it Words cant express my g ratitude to my family. Theyve supported me in every way throughout the years and I wouldnt be where I am or who I am today without them.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ...............................................................................................................4 TABLE OF CONTENTS .................................................................................................................5 LIST OF TABLES ...........................................................................................................................7 LIST OF FIGURES .........................................................................................................................8 ABSTRACT ...................................................................................................................................10 CHAPTER 1 INTRODUCTION ..................................................................................................................12 Autoimmune Disease Susceptibility .......................................................................................12 Type 1 Diabetes (T1D): An Autoimmune Disorder ...............................................................12 T Lymphocyte Tolerance and T1D .................................................................................13 Regulatory T Ce lls and Inflammation in T1D .................................................................14 Myeloid APC Dysfunction in T1D .........................................................................................14 Study Background ..................................................................................................................15 2 M ATERIALS AND METHODS ...........................................................................................19 Human Sample Collection and Preparation ............................................................................19 Small Volume Flow Cytometric Analysis of STAT5Ptyr in Peripheral Blood Monocytes ...19 Antibody Conjugation .....................................................................................................20 Flow Cytometry Analysis of Activated STAT5 ..............................................................20 Deconvolution Microscopy ....................................................................................................21 Animal Models .......................................................................................................................21 Bone Marrow and Tissue Collection, Cell Culture and Sample Preparation .........................22 Mouse and Human Cell Processing ........................................................................................22 Chromatin Immunoprecipitation (ChIP) Analysis of STAT5 Binding at the Csf2 Promoter ..............................................................................................................................23 Polymerase Chain Reaction .............................................................................................24 Real Time PCR ................................................................................................................24 Luminex and ELISA ...............................................................................................................24 Sequence Analysis of Csf2 Gene Promoter and Ptgs2 G ene E nhancer .................................25 Mouse Sequence Analysis ...............................................................................................25 Human Sequence Analysis ..............................................................................................25

PAGE 6

6 3 A CTIVATED STAT5 L EVELS IN HUMAN T1D PATIENTS, NON AUTOIMMUNE CONTROLS AND AT RISK INDIVIDUALS ......................................................................29 Human Flow Cytometry .........................................................................................................29 Chromatin Immunoprecipitation ( ChIP ) Analysis of Human PBMC ....................................29 4 NOD M OUSE M ODEL A NALYSIS .....................................................................................33 Increased GMCSF E xpression and STAT5 P hosphorylation in NOD B one M arrow C ells and P eritoneal M acrophages ......................................................................................33 Enhanced STAT5 Binding on the Csf2 Gene Promoter in NOD Macrophages & Bone Marrow Cells .......................................................................................................................33 Sequence Analysis of Csf2 Promoter Region Defines STAT5 Binding Site Polymorphisms ....................................................................................................................34 Bicongenic B6.NODC11bxC1tb M ice H ave a Macrophage Islet Infiltration ........................35 N onObese Diabeti c GM CSF and STAT5 Phenotypes Segregate With the NOD CSF2 Promoter Not the CSF2 Gene .............................................................................................35 Effects of NOD CSF2 Promoter Polymorphisms on GM CSF Induced STAT5 Binding at the PTGS2 Enhancer .......................................................................................................36 5 H UMAN SEQUENCE ANALYSIS .......................................................................................51 Gene Sequence Analysis: GM CSF ( CSF2 ) ...........................................................................51 Gene Sequence Analysis: COX2 (PTGS2) .............................................................................51 6 D ISCUSSION .........................................................................................................................55 LIST OF REFERENCES ...............................................................................................................59 BIOGRAPHICAL SKETCH .........................................................................................................63

PAGE 7

7 LIST OF TABLES Table page 21 Characteristics of patient, control and at risk samples collected for flow cytometric and sequencing analysis .....................................................................................................26 22 ChIP buffer compositions ..................................................................................................28

PAGE 8

8 LIST OF FIGURES Figure page 21 Chromosomes 11 and 1 on mouse strains derived through congenic breeding to isolate regions of the Idd4.3 diabetes susceptibility locus as well as the Ptgs2 enhancer region, respectively. ............................................................................................27 31 Flow cytometric analysis of STAT5 levels in human healthy controls, at risk individuals and T1D patients. ............................................................................................31 32 Comparison of STAT5Ptyr expres sion in controls and T1D patients evaluated by gender. ................................................................................................................................31 33 Chromatin Immunoprecipitation analysis of STAT5 binding at various regions upstream of CSF2 and PTGS2 in human peripheral blood monocytes (PBMC) with and without GM CSF stimulation. ...................................................................................32 41 Granulocyte Macrophage Colony Stimulating Factor Production and STAT5 Phosphorylation are aberrantly high in NOD Mouse Myeloid Cells.. ...............................39 42 Macrophage Chromatin Immunoprecipitation (ChIP) Analysis shows STAT5 binding within the promoter region upstream of the gene which encodes for GM CSF, Csf2 .. .........................................................................................................................40 43 Chromatin Immunoprecipitation analysis (ChIP) of GM CSF induced STAT5 binding upstream at multiple sites within the Csf2 promoter involves DNA secondary structure. .............................................................................................................................41 44 Sequence Analysis of Csf2 Promoter region and definition of the STAT5 binding site polymorphisms involved in NOD myeloid cell phenotypes and chromosome 11 diabetes susceptibility. .......................................................................................................43 45 The GM CSF Production by Congenic Mouse Bone Marrow Cells and Peritoneal Macrophages. .....................................................................................................................44 46 The STAT5 Phosphorylation by Congenic Mous e Monocytes and Peritoneal Macrophages. .....................................................................................................................45 47 Histology of Pancreas Tissues of B6.NOD and NOD.L subcongenic mice. Congenic mouse pancreas slices were stained with H and E (C, D, and E) Immunostaining was also performed against CD3 and F4/80 (F). ....................................46 48 Chromatin Immunoprecipitation (ChIP) analysis of STAT5 binding at the Csf2 promoter regions in con genic mice. .................................................................................48 49 Chromatin Immunoprecipitation (ChIP)analysis of STAT5 binding on the Ptgs2 enhancer in GM CSF stimulated congenic mouse bone marrow and macrophages. ........49

PAGE 9

9 410 The PGS2/COX2 expression and PGE2 production in congenic mouse bone marrow and macrophages with and without GM CSF stimulation. ................................................50 51 Granulocyte Macrophage Colony Stimulating Factor gene ( CSF2 ) map. ........................53 52 Sequence alignment of the GM CSF promoter region from healthy control and T1D DNA shows i ntact STAT5 and STAT6 binding sites. .......................................................53 53 Prostaglandin Synthase 2 gene ( PTGS2 ) map. Above is a schematic representation of the enhancer region upstream of the COX 2 gene, PTGS2. .......................................54 54 Sequence alignment of the COX 2 enhancer region from healthy control and T1D DNA shows intact STAT5 and STAT6 binding sites. .......................................................54

PAGE 10

10 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science LOSS OF CSF2 EPIGENETIC REGULATION THROUGH ABERRANT GM CSF INDUCED STAT5 SIGNALING CONTRIBUTES TO TYPE 1 DIABETIC MYELOID CELL DYS FUNCTION By Erin Lee Garrigan December 2008 Chair: Mark Atkinson Cochair: Sally Litherland Major: Medical Sciences Defective myeloid differentiation and activation contribute to immunopathogenesis of autoimmune type 1 diabetes (T1D) in humans and nonobese (NOD) mice by retarding their antigen presenting cell function, altering their cytokine production, and promoting a proinflammatory microenvironment through aberrant expression of prostaglandin synthase, PGS2 (COX2 ). Monocytes of at risk/T1D humans and macrophages of the NOD mouse have elevated autocrine GM CSF production and persistent signal transduction/activator of transcription (STAT5) phosphorylation. GM CSF activates STAT5 binding at its own gene promoter wit hin the Idd4.3 region, as well as at the enhancer of the COX2 gene, Ptgs2 Since STAT5 acts as adaptor protein for either deacetylase or acetylase enzymes mediating epigenetic chromatin modification, we hypothesize that the loss of Csf2 gene regulation in autoimmune cells is mediated by changes in epigenetic control perpetuated by GM CSF activating persistent STAT5 then also genes such as Ptgs2 Our specific aims are to characterize epigenetic modification at regulatory sites upstream of the Csf2 and Ptgs2 genes mediated by STAT5 binding and determine if such changes are related to GMCSF induced 1) STAT5 function and 2) Csf2/Ptgs2

PAGE 11

11 gene expression in autoimmune myeloid cells, and 3) the genetic sequence of regulatory regions diabetes susceptibility loci. A sm all volume flow cytometric analysis was developed to measure STAT5Ptyr levels in T1D patients and NOD mice for use as a potentially clinically useful biomarker assay for autoimmune T1D susceptibility. On average, T1D patient monocytes STAT5Ptyr le vels were two fold higher than healthy controls H owever when these data are analyzed by gender, the results showed a statistically significant female bias for this biomarker in the subject group Our phenotypic and genotypic studies of the NOD mouse and subsequent congenic derivatives of this model narrow diabetes susceptibility attributed to Chr. 11 and GM CSF overproduction with persistent STAT5 phosphorylation to a small (200bp) untranscribed regulatory region within the Csf2 promoter. In the NO D, this region contains a loss of STAT6 binding site (potential antiinflammatory mediator) and a gain of a STAT5 binding site. Homologous sites were not found in the T1D patient versus healthy control, suggesting alternate regulatory mechanisms contribut ing to the persistent STAT5 phosphorylation in human autoimmune monocytes. Our results support the potential of epigenetic control as a mechanism underlying the chromosomal dysregulation seen in autoimmune diseases and indicate that epigenetic control of gene expression can provide a new avenue for discovery of potential prevention/therapeutic intervention targets.

PAGE 12

12 CHAPTER 1 INTRODUCTION Autoimmune Disease Susceptibility Autoimmunity is the condition in which the immune system fails to recogniz e the difference between the bodys own tissues and a foreign matter, and thus can become intolerant of its own components ( 1) When tolerance is lost, the immune cells respond to the bodys own tissue as it would to a foreign invader, often resulting in a self directed immune response which can damage the bodys organs and tissues. Autoimmunity arises from a complex mix of environmental, genetic and immunological factors. Some autoimmune responses may arise as a result of infection or exposure to other fo reign elements leading to the development of antibodies that have cross reactivity with self tissue. Type 1 Diabetes ( T1D ) : An Autoimmune Disorder Diabetes is a disorder generally characterized by the bodys inability to respond to insulin signaling, ei ther through loss of insulin production, resistance in the insulin receptor, or from a failure to adequately use insulin. Insulin is a necessary hormone that facilitates the uptake of glucose contained in food from the blood into the cells around the body. Improperly regulated glucose levels can lead to tissue damage and can become life threatening. Diabetes affects an estimated 20.8 million children and adults in the United States ( 2) Onset of T1D results from a complex blend of genetic predisposition and one or many environmental trigger(s). Two twin studies, conducted in Denmark and the United States observed the presence of islet and beta cell auto antibodies in monozygotic and dyzygotic twins where one or both siblings was afflicted with insulin dependent diabetes ( 3, 4) While the studies differed in their conclusions of which was more important in diabetes etiology, genetic or

PAGE 13

13 environmental triggers, they serve to highlight the multi factorial nature of diabetes pathogenesis. Prior to the onset of clinical symptoms of T1D, an early signal of autoimmunity is the appearance of circulating immune cells and antibodies directed against islet and beta cell antigens ( 5, 6) T Lymphocyte Tolerance and T1D Type 1 Diabetes Mellitus is classified as an au toimmune disorder based on the destruction of insulin producing beta cells in the pancreas by the immune system. The immune mediated beta cell destruction involves cells of both lymphoid and myeloid origin, including CD4+ and CD8+ T lymphocytes and Antige n Presenting Cells (APC), and it may also be influenced by an inflammatory cytokine microenvironment. Random rearrangements of germline sequence contribute to antigenic diversity within the T Cell Receptor (TCR) but may also give rise to auto reactive T lymphocytes. During central tolerance of healthy individ uals, the potentially auto reactive T cells are negatively selected when they recognize self peptides presented in the context of self MHC ( 7) Auto reactive T cells may become ane rgized by contacti ng an antigen MHC complex in the absence of co stimulatory molecules, thus adding an additional level of regulation of potentially immunopathogenic T cells ( 8) Errors in T cell tolerance may promote the escape of autoreactive T cells which can contribu te to the onset of autoimmunity. Serreze et al. (1993) asserted that abnormal maturation of bone marrow derived APC in nonobsese diabetic ( NOD ) mice are affected by both MHC and nonMHC linked diabetes susceptibility genes He found that the NOD has APC that are unable during tolerance ( 9) The result s of these defects are that NOD APC function at suboptimal levels and are very poor at inducing and maintaining self tolerance.

PAGE 14

14 Regulatory T Cells and Inflammation in T1D NOD mice that exhibit homozygous expression of the MHC allotype H 2g7, the MHC region linked to diabetes susceptibility. This MHC molecule has been shown to present auto reactive effector T cells. NOD APC expressin g H 2g7 molecules are known to preferentially activate these auto reactive T cells but poorly activate regulatory T cells I nteraction of NOD derived T cell precursors with APC that present an MHC haplotype other than H 2g7 prevents diabetogenesis ( 10 12) D efects in the regulation of auto reactive T cells which destroy insul pancreatic cells have been implicated in diabetic pathogenesis. In addition, there is growing evidence that defects in the maturation of antigen presenting cells also contribute to the development of diabetes ( 13) Moreover, errors in cytok ine signaling between these immune cells lead to a cellular miscommunication, contributing to both de regulation of T cell ontogeny and peripheral APC dysfunction, which allows the autoimmune responses of T1D to progress unchecked. Chronic inflammation is a common phenotype in many autoimmune diseases and it is thought to play a role in loss of tolerance induction by APC ( 14) Myeloid APC Dysfunction in T1D Myeloid APC development is advanced by a succession of cytokine signals that push cells d own differentiation paths to become different cell types. For example, interleukin 3 (IL 3) activates autocrine GMCSF (granulocyte macrophage colony stimulating factor), which in turn activates G CSF (granulocyte colony stimulating factor) to promote gra nulocyte differentiation. IL 3, in conjunction with GM CSF also stimulates MCSF (macrophage colony stimulating factor), to promote mono cyte and macrophage development ( 15) In this cascade, GMCSF is a multifunctional cytokine ex erting different influences in myeloid cells differentiation and mature monocyte/macrophage/ dendritic cell functionality.

PAGE 15

15 GM CSF regulates the proliferation and differentiation of hematopoietic stem cells but it also helps to direct mature cell function ( 16, 17) GM CSF is thought to be involved in autoimmunity both through its regulation of the stages of maturation of macrophages and granulocytes, but also for its role in activating the inflammatory process in mature macrophages ( 14, 17) Through its stimulation of Jak2/STAT5 signaling, GM CSF activates STAT5 (by phosphorylation with Jak2) to upregulate expression of the dual functional enzyme, prostaglandin synthase/cyclooxygenase 2 (PGS2/COX2), and subsequent production of Prost a glandin E2, a key lipid mediato r in the inflammatory process ( 18, 19) Finally GM CSF stimulates the gene expression of Interleukin 10 (IL 10), which down regulates the signaling cas cade ( 20) Study Background Chase et al (1979) initially observed that T1D patients had high levels of PGE2 in their plasma, indicative of a chronic pro inflammatory state ( 21) Monocytes from Type 1 Diabetic human patients exhibited a high production of COX 2 (cyclooxygenase 2), also known as PGS2 (prostaglandin synthase 2), an enzyme involved in the producti on of proinflammatory prostaglandins specifically Prostaglandin E2 (PGE2). Ptgs2 is the gene that encodes the COX2 enzyme. No genetic abnormalities were observed in the coding region of the NOD and TID patient Ptgs2 gene, nor in the five prime promoter region s or th e three prime poly A tail. Since the COX2 gene appeared to be normal healthy control humans, it was necessary to investigate an alternate explanation for the aberrant COX2 expression in human T1D monocytes, so regulation of Ptgs2 was examined. PGE2 production by COX2 is important not only in promoting inflammation but also in blocking toleragenic APC maturation as well as inhibiting the signal for activation induced cell death (AICD) from the APC from being received by the T cell. COX2 activ ity in T1D patient peripheral blood monocytes was resistant to suppression by IL 10; whereas, COX2 expression was suppressed by the addition of IL 10 to control human monocytes

PAGE 16

16 cultures ( 18) In addition GM CSF, a strong activator of IL 10 and COX2 expre ssion in macrophages, was found to be abnormally high in individuals with or at risk for T1D These findings suggested that both activation and suppression of COX2 in individuals at risk or with T1D may be the result of the disruption of these cytokines regulation of the Ptgs2 gene expression. In nonautoimmune myeloid cells, GM CSF stimulation alone is not sufficient to induce COX2 expression. Yamaoka et al (1998) reported that normally, activation of COX2 expression requires LPS and GM CSF stimulation and it is accomplished through Jak2/STAT5 signal transduction ( 20) However, in diabetic monocytes and macrophages, GM CSF stimulation alone is adequate to activate COX2 expression ( 22) This activation makes COX2 resistant to IL 10 suppression even tho ugh GM CSF expression in these same cells is completely IL 10 sensitive. These findings suggest that some other component in COX2 activation is also abnormally regulated and that the expression of GM CSF itself is aberrantly regulat ed in NOD and T1D myeloid cells GM CSF acts on gene expression through the activation of transcription factors such as PI3K, MAPK, and the JAK/STAT pathway ( 14) Work by Yamaoka et al indicates that GM CSF activation of the Jak2/STAT5 signaling pathway is critical to its influence on the Ptgs2 expression ( 23) STAT5 is an effect or molecule in epigenetic regulation ( 24, 25) After phosphorylation (by Janus Kinase), two STAT5 molecules dimerize and translocate to the nucleus where it binds DNA at the motif TTCNNNGAA ( 18) I n addition to a DNA binding domain, the STAT5 dimer contains an acetylase and de acetylase binding site, making it an adaptor molecule to facilitate the opening and closing of chromatin to influence gene expression based on the primary signal transduction. Epigenetic regulation entails the modification of

PAGE 17

17 histones and the DNA itself with small molecules suc h as acetyl, methyl, ubiquitin and sumo groups ( 2628) These modifications alter the topological structure of chromatin, allowing for the opening and closing of gaps in the nucleosome cores and making the DNA sequences in the region accessible or inaccessible to DNA modification, replication, repair, and transcriptional enzymes. These rapid and inheritable changes in chromatin structure are potentially crucial to cytokine induced gene r egulatory f unctions To examine whether GM CSF overproduction was affecting COX2 expression through JAK2 STAT5 signaling, Litherland et al (2005) looked at Jak2 and STAT5 function in T1D human and NOD mouse myeloid cell s. The results of these studies showed that STAT5 was persistently phosphorylated in un activated T1D patient peripheral blood monocytes and NOD mouse monocytes and macrophages ( 13, 22) Moreover, GM CSF was found to stimulate its own production (via the JAK2/ST AT5 pathway). These data suggest that overproduction of GM CSF in autoimmune cells may contribute to abnormal autoimmune monocytes and macrophage development and activation as well as the abnormal COX2 expression in autoimmune T1D patients and NOD m ice ( 29) This aberrant expression of COX2 in T1D patients and its subsequent production of PGE2 may assist in the establishment of an inflammatory microenvironment which may promote loss of beta cell specific auto reactive T cell tolerance and successive activation, advancing the tissue damage observed in Type 1 Diabetes. We have observed increased COX2 expression and PGE2 production by COX2 in un activated autoimmune monocytes, as well as elevated GMCSF production, increased STAT5 phosphorylation and IL 10 resistant COX2 activity in these cells. Because STAT5 is strongly activated by GMCSF in T1D patients, we asked how GM CSF activation of STAT5 is affecting COX2 expression and potentially disease pathology.

PAGE 18

18 Abnormal regulation of STAT5 could indic ate that STAT5 (after having been activated by an initial GM CSF signal) is feeding back to stimulate the continuous production of GM CSF (through activation it its gene, Csf2 ) as well as affecting expression of other genes (e.g. Ptgs2 ) that are controlled through the same signaling pathway. To test this hypothesis, we analyzed both monocytes in T1D human blood samples and monocyte s and macrophages from NOD and congenic mice. Within the human system, we determined whether the observed phenotypes (elevated GM CSF, persistent STAT5 phosphorylation and increased COX2 expression) were correlative enough to be useful as biomarkers for susceptibility screening among known at risk populations. Through a study of mouse strains derived through congenic breeding, w e investigated the genetic input in these three phenotypes independently as well as their potential interactions in T1D immunopathogenesis.

PAGE 19

19 CHAPTER 2 MATERIALS AND METHOD S Human Sample Collection and Preparation Patient blood samples are obtained from healthy volunteers (age range 4 to 46 years ; Table 2 1) T1D patients, and their immediate relatives, through collaboration with Dr. Michael Clare Salzler M.D and Dr. Mark Atkinson. Samples are collected with informed consent and under IR B (human blood samples) approved protocol number 3721996. Human peripheral blood was initially processed on Ficoll gradients and the monocyte layer (PBMC) was separated from whole blood with Ficoll gradient centrifugation. Later, whole blood was used dir ectly in the analyses. The PBMC layer was collected, washed and counted for viable cells. The monocytes were plated on tissue culture dishes and fed with fresh sterile medium alone or with 1000U/ml of GM CSF (Biosource). Cultures were then processed according to protocol detailed in the Mouse and Human Cell Processing section. Small Volume Flow Cytometric Analysis of STAT5Ptyr in Peripheral Blood Monocytes Fluorescence Activated Cell Sorting (FACS), also known as flow cytometry, was used to determine the presence of tyrosine phosphorylated STAT5 (STAT5Ptyr) in the myeloid cells. Intracellular flow cytometry was carried out with phosphate modification specific anti STAT5 Ptyr monoclonal antibodies that were conjugated with APC (chemical name) or PE (chemical name). Surface staining with anti CD11bFITC/PE was used to identify mouse myeloid cells or with CD14 PE to identify human peripheral blood monocytes.

PAGE 20

20 Antibody Conjugation Anti Phosphorylated STAT5A/B antibody ( STAT5Ptyr, Upstate) and Normal Mo use IgG antibody (Upstate) were conjugated with PhycoLink Activated Allophycocyanin (APC) or PE (Prozyme, Glyko) for use in fluorescent immunohistochemical analyses (e.g. FACS, deconvolution microscopy, etc.). The conjugation was accomplished according t o the protocol from the PhycoLink Allophycocyanin (APC) Conjugation Kit (Product code PJ25K). The final conjugated antibody concentrations were measured by Bradford Assay by comparison to a BSA/Bradford dye comparative standard curve measured at 655nm. D ye conjugation was calculated from Absorbance at 655nm relative to an APC/PE dye standard. Stock antibody concentrations were diluted to 1mg/ml (STAT5Ptyr APC) and 0.01mg/ml (mouse IgG APC) in PBS based on optimum on optimum fluorescence when calibrated w ith human peripheral blood monocytes. F low Cytometry Analysis of Activated STAT5 Blood samples were collected by ven i puncture or by finger prick from Type 1 diabetic patients as well as non autoimmune control subjects under IRB approved protocols. Sampl es were processed within 45 minutes of collection. In keeping with the goal of developing a small volume biomarker assay, 150l of whole blood was incubated for 10 minutes with anti human CD14 (a macrophage marker) (BD Pharmingen) and unlabeled anti mouse IgG (Sigma). Cells were then fixed and made permeable with BD cytofix/cytoperm (BD Biosciences) and incubated for 10 minutes. Cells are then split into two equal volumes, washed with saponin buffer (containing BSA, Sodium Azide, saponin and PBS in ddH2O ) and incubated for one hour in 3l of 1mg/ml STAT5Ptyr APC or 1l 0.01mg/ml mouse IgG APC. After incubation cells are again washed with saponin buffer and stored in an isotonic buffer solution until flow cytometric analysis.

PAGE 21

21 Deconvolution Microscop y Cell samples prepared for Flow Cytometry were also examined by microscopy to confirm activated STAT5 staining. Samples were adhered to a charged slide by centrifugation in a cytospin centrifuge. The specimens were counter stained with 1ng/ml DAPI and w ashed and held at 4C in the dark until analysis. Deconvolution microscopy (Olympus IMT/DeltaVision Deconvolution) was used to determine the subcellular location of STAT5Ptyr in a set of 20 30 0.01micron optical slices per image of isolated cells labeled for STAT5 Ptyr. Animal Models Four to thirty week old male and female mice (The Jackson Laboratory, Bar Harbor, ME) were used for all studies. NOD, C57BL/6, NOD.LC11, B6.NODC11 and B6.NODC1 mice were maintained in the University of Florida Pathology SP F mouse colony in microisolator cages with food and water ad libium throughout the study. NOD background congenic mice (NOD.LC11) and its sub congenic derivatives (NOD.LC11e and NOD.LC11b) were originally developed at the University of Virginia specific p athogen free (SPF) colony and were a generous gift from our collaborator, Marcia McDuffie, MD ( 30) ( F igure 21) The B6.NODC11b and B6.NODC1tb strains were derived for this study from B6.NODC11 and B6.NODC1 congenic mice which are housed at the Universit y of Florida College of Medicine Pathology Department SPF colony, These B6 background strains were originally derived by Yui and Wakeland ( 31) Sub congenic strains of the B6.NODC11 and NOD.LC11 mice and the B6.NODC11bxC1tb bi congenic strain were bred fro m these strains. Genotyping was performed from tissue collected by tail or ear biopsy at the time of weaning to confirm genetic intervals using PCR analysis ( 20, 32) All procedures were conducted according to IACUC approved protocols B083 and D754.

PAGE 22

22 Bo ne Marrow and Tissue Collection, Cell Culture and Sample Preparation Control, NOD, subcongenic and bi congenic strain mice were analyzed at weaning, 5 12 weeks of age and at 20 weeks of age. Mice were euthanized by over anesthetizing and cervical disloca tion. Blood monocytes were collected either at the time of tail or ear biopsy or by cardiac puncture after euthanasia, then analyzed for percentage of phosphorylated STAT5 by flow cytometry. Peritoneal macrophages were collected by injecting ice cold RPMI medium supplemented with 10% fetal calf serum and 1% antibiotic/antimycotic mix (Cellgro Mediatech, Herndon, VA) into the peritoneal cavity. The lavage fluid was then withdrawn and washed with cold media by centrifugation. The long bones of the hind l imbs were excised marrow cells were flushed out of the bones using a 30guage needle and syringe filled with cold RPMI medium supplemented with 10% fetal calf serum and 1% antibiotic/antimycotic mix (Cellgro Mediatech, Herndon, VA). The marrow cells and l avage harvested macrophages were washed with cold media then the red blood cells in samples were lysed by incubation in sterile, cold 0.84% NH4Cl buffer. The remaining bone marrow cells and macrophages were then plated on tissue culture dishes and fed with fresh sterile medium alone or with 1000U/ml of GM CSF (Biosource) with or without 2 g/ml antiGM CSF blocking antibodies (Endogen) for 24 hours at 37C/5%CO2. After incubation, 1ml of the cultured cell supernatant was collected and frozen at 80C to mea sure GM CSF concentration by Luminex and ELISA as well as PGE2 by ELISA (GE/Amersham). Cells were then processed according to protocol detailed in the Mouse and Human Cell Processing section. Mouse and Human Cell Processing Cultures were maintained f or 24hr or 48hr at 37C /5%C02, then washed and re supplemented for an additional 24hr or 48hr in culture at 37C /5%C02. An aliquot of cells was taken to confirm phenotypic identification and phosphotyrosine STAT5 analysis by flow

PAGE 23

23 cytometry as previously described ( 22 42) Cells were fixed in situ with 1%(v/v Cf) formaldehyde (methanol free, Sigma Aldrich, St. Louis, MO) added in the remaining media for 10min at 37C, then washed with 1x PBS. Cells were sonicated for 5 seconds in SDS Lysis Buffer ( T able 22) + protease inhibitors (Roche, Indianapolis, IN) to disrupt membranes and shear chromatin to approximately 1000bp fragments then frozen for later analysis with Chromatin Immunoprecipitation (ChIP). Chromatin Immunoprecipitation (ChIP) Analysis of STAT5 Binding at the Csf2 Promoter Frozen cell extracts detailed in the Mouse and Human Cell Processing section were thawed. The samples consisted of four to five million cells from bone marrow cultures or ex vivo peritoneal macrophages and were divided into al iquots for each run of the analysis. The aliquots used for immunoprecipitation (IP) were pre cleared with salmon sperm DNA Protein A agarose beads (Upstate), then incubated overnight at 4C with anti STAT5Ptyr antibodies (Upstate). After incubation, the a ntibodybound chromatin complexes were precipitated using salmon sperm DNA Protein A agarose beads, and washed extensively with a series of increasing stringency buffers (low salt, high salt, LiCl, TE; t able 22 ). A nonspecific antibody control (mouse IgG, UpState) and a sham IP containing no extract were run as negative controls. Total cell and ChIP extract aliquots were dissociated from the beads in 1% SDS, 0.1M Bicarbonate buffer (Fisher Scientific, Atlanta, GA). Sodium chloride was then added to a fi nal concentration of 500mM and the samples incubated 4 hours at 65C to reverse the formaldehyde crosslinks. DNA was purified from these aliquots for PCR amplification of DNA sequences from Csf2 promoter which have been shown to be epigenetic regulatory si tes for inducible Csf2 expression and from the Ptgs2 gene ( 33, 34) In Double ChIP (dbChIP) analyses, chromatin complexes first precipitated with anti STAT5Ptyr antibodies were dissociated from the antibody Protein A agarose beads and then re -

PAGE 24

24 precipitate d with anti histone H3 antibodies (UpState USA) prior to DNA de crosslinking and purification as described above ( 35) Polymerase Chain Reaction ChIP DNA was eluted in a buffer containing 20% SDS, 1M NaHCO3 and water. The DNA/Protein crosslinks were reve rsed (with four hour incubation in NaCl then addition of 0.5M EDTA, 1M Tris HCl and Proteinase K) and the remaining DNA was purified by Phenol/Chloroform Extraction. PCR reactions were set up using Eppendorf Master Mix (2.5x), 2% DMSO (Sigma Aldrich) and primers (IDT) and run on a Mastercy cler (Eppendorf) with cycle protocol; 98C 5min, 94C 30 sec, 5560C (dependent on the primer set used) 30sec, 72C 30sec, for 35 cycles. Amplification products were separated on a 2% agarose gel (SeaKem Fisher Scientific) a nd visualized by ethidium bromide (Fisher Scientific) intercalation. Real Time PCR DNA samples were volume matched to 100ng of their Total DNA aliquots (no IP) in all PCR reactions. DNA isolated from ChIP extracts was used as a template in a reaction containing SYBR Green PCR Master Mix (Applied Biosciences, Foster City, CA or BioRad, Hercules, CA), primers specific for Csf2 promoter region or COX 2 enhancer (IDT), and 2% DMSO (Fisher Scientific). The Real Time PCR reaction was completed in a contin uous fluorescence detector (MJ Research) and amplification quantitation was compared on the basis of R value calculated as R = 2(Non specific Ig ChIP c(t) antiSTAT5Ptyr ChIP c(t)). Statistical data analyses were performed using Prism 4/5 (Graph Pad, San Diego CA). Luminex and ELISA Supernatant media from myeloid cells cultured at 37C/5%CO2 for 24hr and 48hr in was collected for GM CSF Luminex analysis. Luminex quantifies the amount of GM CSF and other cytokines made by the treated cells using fluoresc ently labeled beads coated with antibodies to a

PAGE 25

25 specific cytokine. The amount of cytokine (i.e., bound specific antibody + fluorescent beads) in a sample is quantified by fluorescence detected on a Luminex flow cytometer versus a standard curve for each cy tokine tested. GM CSF was quantified using Luminex (Upstate Beadlyte) and ELISA (BD biosciences) and PGE2 was detected using ELISA (GE/Amersham). Sequence Analysis of Csf2 Gene Promoter and Ptgs2 gene enhancer Mouse Sequence Analysis Approximately 50ng of genomic DNA from each mouse strain was prepared from liver and amplified by PCR using Master Mix (Eppendorf or Roche Biosciences) reagents and primers (3 CTA AAA CAT GTT TCT TGG CTA; 5 AAA TAA GGT CCA GCC CAA TG) designed to amplify the 3 to 969bp s equence upstream of the Csf2 gene. The amplified DNA was gel purified using Qiagen gel extraction reagents and phenol/chloroform extraction (Qiagen, Valencia CA). The amplified fragment was then used as template in a Big Dye PCR amplification reaction (App lied Biosystems) and sequenced using an AB capillary sequence analyzer (Applied Biosciences). ChromasLite and ClustalW freeware were used for the sequence analysis and alignment. Human Sequence Analysis Primers specific to the Csf2 promoter and Ptgs2 enhancer regions in the human genome that are homologous to the regions in the mouse genome were designed using Net Primer. Primers for the region upstream of h P TGS 2 (5GGGGCGAGTAAGGTTAAGAAAGGC; 3 ACATTTAGCGTCCCTGCAAATTCTG Sigma Genosys) select for a region approximately 397 bp in length and include two STAT5 binding sites ( 33) The primers designed to amplify the human GM CSF gene ( Csf2 ) (5 GTGGATTGGAAAGACTTGTTGACTG; 3 TTCACATGCTCCCAGGGCT Sigma Genosys) generate a PCR product of length 1993bp. Human DNA samples were purif i ed using Qiagen Blood and Cell Culture DNA Mini Kit

PAGE 26

26 (Qiagen, Valencia CA) and amplified by PCR using Eppendorf 2.5X Master Mix (Fisher Scientific) and either Csf2 prom oter or Ptgs2 enhancer primers. The Eppendorf thermocycler program f or the GM CSF promoter amplification was 95C 5 minutes, 55.6C 30 seconds, 72C 2 minutes, 25C 30 seconds. The amplification program for the Cox2 enhancer region differs by annealing temperature and elongation time (95C 5 minutes, 57C 30 seconds, 72C 30 seconds, 25C 30 seconds). PCR amplification products were visualized on 1% agarose gel (SeaKem Fisher Scientific) and purified usin g DNA Clean and ConcentratorTM 5 (Zymo Research) then amplified again in a Big Dye PCR amplification reaction (Applied Biosystems) and sequenced using an AB capillary sequence analyzer (Applied Biosciences). Human sequence was analyzed and compared to consensus sequences in VectorNTI. Table 2 1. Characteristics of patient, control and at risk samples collected for flow cytometric and sequencing a nalysis At risk individual for this study is defined as any individual with a familial history of T1D, HLA susceptibility haplotypes and/or autoantibodies associated with T1D. ** Mean disease duration = 6.59 Characteristic Healthy Control (n =24 ) Type 1 Diabetic ** Patient (n = 24) At risk In dividual (n = 11) Gender Male Female 15 9 12 12 3 8 Age Range 4.97 44.89 6.56 35.1 7.42 46.35

PAGE 27

27 Figure 21. Chromosomes 11 and 1 on mouse strains derived through congenic breeding to isolate regions of the Idd4.3 diabetes susceptibility locus as well as the Ptgs2 enhancer region, respectively. The cluster of strains to the left is Chromosome 11 and the clu ster of strains to the right reflects loci on Chromosome 1.

PAGE 28

28 Table 2 2: ChIP buffer compositions Solution Composition SDS l ysis b uffer 1% SDS, 0.5M EDTA, 10mM Tris ; pH 8.1 Low s alt b uffer 0.1% SDS, 1%Triton X100, 2mM EDTA, 2.4mM Tris, 0.15mM NaCl ; pH 8.1 High s alt b uffer 0. 1% SDS, Triton X100, 2mM EDTA, 2,4mM Tris, 0.5mM NaCl ; pH 8.1 LiCl b uffer 0. 25M Lithium Chloride, 1m L IGEPAL CA630, 1% Deoxycholic acid, 16.5mM Tris, 0.5M EDTA; pH 8.1 TE b uffer 1M Tris HCl, 0.5M EDTA ; pH 8.0

PAGE 29

29 CHAPTER 3 ACTIVATED STAT5 LEVE LS IN HUMAN T1D PATI ENTS, NON AUTOIMMUNE CONTROLS AND AT RISK INDIVIDUALS Human Flow Cytometry Human peripheral blo od mononuclear cells from healthy controls, T1D patients and at risk individuals were ana lyzed by flow cytometry to measure the levels of phosphorylated STAT 5. At risk individuals are those with first degree familial relation to a person with T1D or with measured autoantibodies associated with T1D including those against insulin, beta cell a nd GAD antig en. Consistent with 2005 reports by Litherland et al. we found that the percentage of STAT5Ptyr+/CD14+ untreated peripheral blood monocytes was elevated approximately two fold in Type 1 Diabetic patients versus non autoimmune control individuals (*p= 0.0002, Mann Whitney U test) (Figure 3 1). Because T1D patients showed a wide range of levels of activated STAT5 (STAT5Ptyr) we wanted to determine if there was any influence of gender on levels between the sample groups. To this end we anal yzed the flow cytometry data and found a statistically significant gender bias among females with T1D ( *p= 0.0015, Mann Whitney U test) ( Figure 32). This gender bias may influence the u sefulness of this assay as a biomarker for T1D. Sample data was also analyzed by age ; but no correlation was observed between age and STAT5Ptyr levels (data not shown). Ch romatin Immunoprecipitation (ChIP) Analysis of Human PBMC T1D and nonautoimmune control PBMC were analyzed by Chromatin Immunoprecipitation (ChIP) to isolate and identify chromatin in the CSF2 and PTGS2 regulatory regions associated with epigenetic modifications, STAT5 binding and active transcription (Figure 3 3). GM CSF seems to promote site specific STAT5 binding at locations where

PAGE 30

30 epigenetic regula tion occurs within the autoimmune T1D patients When stimulated with GM CSF in vitro we report enhanced STAT5Ptyr binding in autoimmune cells versus healthy control cells within the CSF2 promoter region and PTGS2 enhancer region (Figure 33). Our findings implicate loss of cytokine induced suppression of epigenetic modification in noncoding regulatory regions as a mechanism for promoting the aberrant expression of genes or genetic regions (e.g., Idd loci) in autoimmune diabetes.

PAGE 31

31 Figure 31. Flow cytometric analysis of STAT5 levels in human healthy controls, at risk individuals and T1D patients. Figure 32. Comparison of STAT5Ptyr expression in controls and T1D patients evaluated by gender. control at-risk TID 0 10 20 30 40 50 60N = 28 N = 12 N = 28 *p < 0.05 *p < 0.001*p = 0.0002 (Kruskal-Wallis)% STAT5Ptyr+/CD14+ Cells Control F at-risk F T1D F Control M at-risk M T1D M 0 10 20 30 40 50 ** p = 0.0015 (Mann Whitney t test) ns nsFemale Male%STAT5Ptyr/CD14+ Cells

PAGE 32

32 Figure 33. Chromatin Immunoprecipitation (ChIP) analysis of STAT5 binding at various regions upstream of CSF2 and PTGS2 in human peripheral blood monocytes (PBMC) with and without GM CSF stimulation. Autoimmune (AI) and healthy control (C) PBMC were cultured with (G) and without (0) GM CSF Real Time PCR analysis was performed on the ChIP extracts with primer sets specific for region s A, B, C and D to asses how STAT5Ptyr binding is affected by GM CSF stimulation. Region A represents a n approximately 1kb non transcribed region upstream of CSF2. Region B represents the enhancer region of the COX 2 gene, PTGS. Region C encompasses a 191 bp region within the CSF2 promoter and region D represents a 115 bp region upstream of the gene. AI 0 AI G C 0 C G AI 0 AI G C 0 C G AI 0 AI G C 0 C G AI 0 AI G C 0 C G -3 -2 -1 0 1 2 3 4 5 -181 to +10bp -47 to -162bp CSF2 3-969 PTGS2 Csf2 PRO Csf2 AR A B C D

PAGE 33

33 CHAPTER 4 NOD MOUSE MODEL ANALYSIS Increased GMCSF E xpression and STAT5 P hosphorylation in NOD B one M arrow C ells and P eritoneal M acrophages GM CSF is an important contributing cytokine in both myeloid cell differentiation and in the inflammatory process mediated by m ature myeloid cells. We examined GMCSF and STAT5 levels within immature bone marrow precursor cells and mature peritoneal macrophages of autoimmune NOD mice and non autoimmune C57BL/6 control mice. As was previously reported by the Litherland lab ( 13, 22, 36) NOD bone marrow cells have increased GMCSF expression (*p=0.0450, MannWhitney U test) and high STAT5 phosphorylation compared to C57BL/6 mouse bone marrow cells ( Figure 41 ). However, the STAT5 phosphorylation in NOD bone marrow cells was signifi cantly lower than in more mature cells(*p= 0399, ANOVA, Figure 41b), despite its comparably high GM CSF production ( Figure 41A ). Enhanced STAT5 Binding on the Csf2 Gene Promoter in NOD Macrophages and Bone Marrow Cells To investigate the potential role of STAT5 in the regulation of Csf2 we performed Chromatin Immunoprecipitation (ChIP) with STAT5Ptyr antibody to examine STAT5 associated chromatin. Extracted ChIP DNA was analyzed by real time PCR using primers designed to amplify the first 1000 bp of t he Csf2 promoter region (based on known de acetylase binding sites) ( 33, 34) STAT5 proteins in NOD peritoneal macrophages exhibit strong binding on sequences within the Csf2 promoter region without exogenous GM CSF stimulation. In contrast, STAT5 chromatin binding at the same sequence was decreased by 4.5 logs in C57BL/6 macrophages in real time analysis and undetectable by western blot ( Figure 42). STAT5 binding within the entire promoter region ( 3 to 969 bp) of Csf2 upstream of the transcriptional s tart site and within subsets of that region was analyzed by STAT5Ptyr mediated

PAGE 34

34 ChIP and real time PCR for NOD and C57BL/6 bone marrow cells ( F igure 4 3A, B ). Within the entire promoter region, stimulation with GM CSF produced an overall decrease in STAT5 binding in both the NOD and C57BL/6 mice ( Fig 4 3A right ). However, STAT5 binding within identified epigenetic regulatory regions of the promoter was enhanced in NOD bone marrow cells stimulated with GM CSF compared to unstimulated versus the C57BL/6 GM CSF stimulated bone marrow cells which showed a decrease in STAT5 binding compared to the untreated group ( F igure 4A ,B ). These findings suggest that GM CSF is enhancing STAT5 binding at specific regulatory regions within the Csf2 promoter of NOD but is re ducing STAT5 binding at all sites within the C57BL/6 Csf2 promoter region. These analyses do not define if STAT5 binding at these sites i s synergistic or independent at each individual site Sequence Analysis of Csf2 Promoter Region Defines STAT5 Bindin g Site Polymorphisms In order to examine the genetic components involved in the GM CSF induced STAT5 binding within the Csf2 promoter region we performed a sequence analysis of this region in all congenic strains and compared in a ClustalW sequence align ment analysis (Figure 4 3). The overlapping NOD region of the B6.NODC11b and the NOD.LC11e mouse strains defines the region most likely responsible for the GM CSF and STAT5 phenotypes seen in the NOD (enhanced GM CSF expression (Figure 45) and persistent STAT5 phosphorylation (Figure 46)). This shared NOD.LC11b and NOD.LC11e region contains at least 2 potential STAT5 GAS binding sites ( 33, 34) It is this region of the Csf2 gene promoter and not the coding sequence that defines the contribution to diabetes resistance conferred on the NOD.LC11 by correction of the Idd4.3 loci on an NOD genetic background. However, insulitis in the B6.NODC11b conferred by this promoter region is minor in comparison with the NOD.L11b strain, suggesting that other components of the Idd4.3 region upstream of the Csf2 promoter may also contribute to this disease phenotype. In the more proximal region of the Csf2 promoter, 2 more STAT5

PAGE 35

35 binding sites and one adjacent half site (GAA) are also altered in the NOD and found in the diabetic congenic NOD.LC11e mouse strain and corrected in NOD.LC11b strain mice which the GM CSF and STAT5 phenotypes are normal and the diabetes disease incidence is greatly reduced. B icongenic B6.NODC11bxC1tb M ice H ave a Macrophage Islet Infiltration NOD.LC11 mice develop a strong T cell dominant peri islet insulitis which remains non invasive despite having fully autoimmune phenotypes present in their T cell population [Marcie 2000]. In contrast, Pancreatic tissue was analyzed from B6.NODC11b and B6.NO DC1tb mice shows very little infiltration, mainly limited to peri ductal areas of the pancreas near islets (Figure 4 7 A, B ). However, an invasive infiltration of was seen in the bicongenic B6.NODC11bxC1tb male (Figure 47C ) and female mice (Figure 47D ). This infiltrate was devoid of T cells but showed a modest infiltrate of macrophages within the islet. N OD GM CSF and STAT5 Phenotypes Segregate w ith the NOD CSF2 Promoter No t the CSF2 Gene GM CSF production of congenic mouse bone marrow cells (Figure 45A ) and peritoneal macrophages (Figure 4 5B ), was increased in strains that contained Csf2 promoter region from the NOD, including the NOD.LC11e mouse strain which has the NOD promoter with C57L Csf2 gene sequence, but not in strains where this region was derived from the C57BL/6 or C57L nonautoimmune strains. These data indicate that the Csf2 promoter, not its coding sequence, is responsible for GM CSF overproduction seen in these mice and in the NOD. The STAT5 phosphorylation phenotype was stronger in unactivated monocytes (Figure 46A ) than in macrophages (Figure 4 6B ) of the NOD.LC11e and B6.NODC11b strains compared to the NOD.LC11b or control strains. STAT5 phosphorylation levels in neither the

PAGE 36

36 B6.NODC11b nor NOD.LC11e myeloid cells recreated the le vels observed in the NOD mouse suggesting some other component in or acting on the Idd4.3 region contributes to this phenotype. The amount of STAT5 DNA binding at the Csf2 promoter in ChIP analyses was also defined by the presence of NOD sequence in thi s region (Figure 48). Moreover, exogenous GM CSF treatment of bone marrow cells from the NOD.LC11b restores STAT5 binding within the Csf2 promoter region, but not in bone marrow from other strains containing NOD sequence in this region, and decreased STA T5 binding in bone marrow cells from NOD and strains containing NOD sequence in this region (Figure 48A ). However, untreated ex vivo macrophages from the B6.NODC11b and the NOD.LC11e mice which containing NOD DNA in the Csf2 promoter upstream of the micro satellite insertion in the region had enhanced STAT5 DNA binding at the site (Figure 38B ). These findings suggest that bone marrow cells require specific amount of GM CSF stimulation to activate STAT5 binding on its own genes promoter and promotion of i ts own expression. However, over stimulation (double the amount expressed in NOD macrophages) can block this effect. Effects of NOD CSF2 Promoter Polymorphisms on GM CSF Induced STAT5 Binding at the PTGS2 Enhancer We previously reported that the high GM CS F production of NOD and T1D human myeloid cells could enhance the PGS2/COX2 production of these cells, even when given alone without additional activation stimuli, such as by LPS ( 22, 37) ChIP analysis of STAT5 binding at the Ptgs2 enhancer region show ed that NOD bone marrow myeloid cells treated with GM CSF in vitro have STAT5 bound to this sequence, while control strain myeloid cells do not (Figure 4 8A ). In contrast, GM CSF treatment of NOD peritoneal macrophages diminished the high level of STAT5 bin di ng at the same site in untreated NOD macrophages. Furthermore, GM CSF enhances STAT5 binding at the Ptgs2 enhancer site in control strains C57BL/6 and C57L

PAGE 37

37 (Figure 4 9 B ). ChIP analyses of the NOD chromosome 11 B6.NODC11b congenic mice which have NOD genet ic sequence at the Csf2 promoter but not at the Ptgs2 region. E ven though these cells had strong binding without stimulation; like in the NOD, exogenous GM CSF treatment of bone marrow and macrophage cells did not enhance their STAT5 binding at the Ptgs2 e nhancer (Figure 4 9 A, B ). Furthermore, macrophages but not bone marrow cells from NOD.LC11e mice that have the NOD Csf2 promoter, C57L Csf2 coding region, and NOD Chromosome 1 in their genetic make up, show enhanced STAT5 binding as seen in the NOD with GM CSF treatment. Congenic B6.NODC1tb mice that have the Ptgs2 containing chromosome 1 region from the NOD on an otherwise C57BL/6 genetic background, show no binding of STAT5 at this site in bone marrow cells without GM CSF treatment; but have strong STAT5 binding at the Ptgs2 enhancer in their macrophages comparable to the NOD without exogenous GM CSF stimulation. NOD.LC11b mouse myeloid cells which have NOD chromosome 1; and therefore, NOD sequence at the Ptgs2 enhancer, do not exhibit STAT5 binding at t his site until stimulated with exogenous GM CSF in culture. T hese findings suggest that GM CSF differentially regulates STAT5 interactions with chromosome 1 chromatin while influencing myeloid cell maturation and matur e macrophage activation. Also, induc tion of STAT5 binding on Csf2 regulatory sequences on Chromosome 11 enhance s this Chromosome 1 sequence interaction in the NOD. As a con firmation that STAT5 binding may influence the interaction between these two genetic regions the same GMCSF, STAT5, an d PGE2 phenotypes were measured in myeloid cells from sub congenic mice bred for the NOD Ptgs2 containing region (Chromosome 1 telomeric, C1t and C1tb) on a C57BL/6 genetic background (B6.NODC1tb strain). We also looked at the phenotypes of myeloid cells from bi congenic C57BL/6 genetic background mice that have both the NOD Ptgs2 containingregion on chromosome 1 (C1tb) and the Csf2 promoter

PAGE 38

38 re gion on chromosome 11(C11b) ( B6.NODC11bxC1tb strain) (Figures 4 2 47). Both of the B6.NODC1tb and the bi congenic B6.NODC11bxC1tb mice strains show STAT5 binding at the Ptgs2 enhancer when stimulated with GMCSF (Figure 49 ) However, only the bi congenic mice had STAT5 binding at both Ptgs2 enhancer and Csf2 promoter sites (Figures 48 and 49), as well as incr eased STAT5 phosphorylation and enhanced GM CSF production without GM CSF stimulation. These results indicate that GM CSF induced STAT5 binding at these 2 regulatory sites is associated with the expression of Csf2 and GM CSF stimulated expression of Ptgs2 in the NOD. We measured PGE2 production in cultures of congenic mouse bone marrow and macrophage as an indicator of COX 2 expression, and we found that PGE2 production was greatly enhanced in the NOD.LC11b and the B6.NODC11bxC1tb mice (Figure 410A, B ). These findings are consistent with a mechanistic model where STAT5 activated by GMCSF is binding to regulatory sequences to promote the expression of both Csf2 and Ptgs2 and suggest that polymorphisms in the Csf2 promoter region affects the abnormal regulation of both genes in the NOD.

PAGE 39

39 A B Figure 41. Granulocyte Macrophage Colony Stimulating Factor Production and STAT5 Phosphorylation are aberrantly high in NOD Mouse Myeloid Cells. A ) Four to five million bone marro w cells and adherenceisolated peritoneal macrophages were cultured without supplementation for 24hr at 37C/5%CO2. The p values listed were obtained from MannWhitney U test analysis of the data. Patterned bars indicate the mean GMCSF production from NOD samples and open bars the mean of C57BL/6 samples. Error bars represent SEM. B) Ex vivo myeloid cells from NOD and C57BL/6 mice (peritoneal macrophages, peripheral blood, and bone marrow cells) were collected and fixed within 4hr of collection and then ana lyzed for phosphorylated STAT5 by intracellular flow cytometry. The p values listed were obtained from MannWhitney U test analysis of the data. Patterned bars indicate the mean %STAT5Ptyr+/CD11b+ cells detected in NOD samples and open bars the mean of C57BL/6 samples. Error bars represent SEM. The p values listed are from pair wise (Student t or Mann Whitney U) or group wise ANOVA analyses. NOD C57BL/6 NOD C57BL/6 0 0 50 100 150 200 *p=0.0004 *p=0.0450 BONE MARROW CELLS PERITONEAL MACROPHAGESpg GM-CSF/million cells NOD C57BL/6 NOD C57BL/6 NOD C57BL/6 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 *p<0.0001 *p=0.0005 ns p=0.25 *p=0.0399 PERITONEAL MACROPHAGES PERIPHERAL BLOOD MONOCYTES BONE MARROW CELLS%STAT5PTYR+/CD11B+ CELLS *p= 0.0399

PAGE 40

40 Figure 42. Macrophage Chromatin Immunoprecipitation (ChIP) Analysis shows STAT5 binding within the promot er region upstream of the gene which encodes for GM CSF, Csf2 Inset: Western blot analysis of STAT5 proteins isolated from ChIP assay in NOD and C57BL/6 mice. Key: 5P= ChIP anti STAT5 precipitated DNA, T= total cellular DNA from un precipitated fixed cell extracts, Ig = ChIP nonspecific mouse IgG precipitated DNA, W= DNAfree water control. Patterned bars (log of mean R values) and gel represent data obtained from 3 independent runs of each strain. Error bars represent SEM. Published work by Federica Sey del et al. (2008) ( 29) NOD PMAC C57BL/6 PMAC 1.01000 1.01001 1.01002 1.01003 1.01004 1.01005 1.01006 1.01007LOG R

PAGE 41

41 Figure 43. Chromatin Immunoprecipitation (ChIP) a nalysis of GM CSF i nduced STAT5 binding upstream at m ultiple sites within the Csf2 promoter involves DNA s econdary s tructure. Four million cell cultures of NOD and C56BL/6 mouse bone marrow cells (BM) were grown for 24hr in the presence (GM or G) or absence (0) of 1000U/ml GM CSF before being fixed and extracted for ChIP analysis. Aliquots of 100ng of total DNA extracted from ChIP protein chromatin complexes precipitated with antiSTAT5 antibodies were amplified using primers to potential epigenetic modification sites within the Csf2 promoter region ( 33, 34) A ) Real time PCR analysis using specific primers to amplify and identify Csf2 pr omoter regions previously identified as epigenetic control sites for Csf2 gene expression. Non patterned bars indicate cells without treatment (0) and hatched bars indicate cells treated with GM CSF (G). Data NOD BM 0 NOD BM G C57BL/6 BM 0 C57BL/6 BM G NOD BM 0 NOD BM G C57BL/6 BM 0 C57BL/6 BM G NOD BM 0 NOD BM G C57BL/6 BM 0 C57BL/6 BM G NOD BM 0 NOD BM G C57BL/6 BM 0 C57BL/6 BM G NOD BM 0 NOD BM G C57BL/6 BM 0 C57BL/6 BM G 1 2 3 4 6 10 16 25 40 63 100 158 251 398 631 1000REGION A REGION B REGION F REGION H REGION I -47 to -162bp -181 to -281bp -355 to -429bp -507 to -630bp +17 to 157bp LOG R B NOD BM 0 NOD BM G C57BL/6 BM 0 C57BL/6 BM G NOD BM 0 NOD BM G C57BL/6 BM 0 C57BL/6 BM G 0 20 40 60 80 STAT5 BINDING AT STAT5 BINDING DEACETYLASE BINDING WITHIN THE REGIONS A,B,F,H,I INTACT WITHIN CSF2 PROMOTER CSF2 PROMOTER +17 TO -630bp -3 to -969bp *p= 0.0001 ANOVA ns; p= 0.2998 ANOVA **p=0.0011 *p=0.0015 **p=0.0001 *p=0.0021 R ns: p = 0.2988 ANOVA A

PAGE 42

42 representative of 23 sample sets. B ) Non patte rned bars indicate cells without treatment (0) and hatched bars indicate cells treated with GM CSF (G). Data representative of two (combined A I) and three (Promoter 3 to 969bp) sample sets. The p values indicate one way ANOVA analysis (above graphs) or Mann Whitney U tests (on graph) of pairwise comparisons

PAGE 43

43 Figure 44. Sequence Analysis of Csf2 Promoter r egion and definition of the STAT5 binding s ite polymorphisms involved in NOD m yeloid c ell phenotypes and c hromosome 11 diabetes s usceptib ility. Yellow boxes represent STAT binding sites. Pink highlight is a microsatellite insertion.

PAGE 44

44 NOD C57BL/6 C57L B6.NODC11 B6.NODC11b NOD.LC11 NOD.LC11b NOD.LC11e B6.NODC1tb B6.NODC11bXC1tb 0 10 20 30 40 50 60 70 80 90 100 110 120*p=0.0002; ANOVA ns ns *p=0.0081; ANOVA *p=0.0005; ANOVA pg GM-CSF/million cells Figure 45. The GM CSF Production by Congenic Mouse Bone Marrow Cells and Peritoneal Macrophages. A) GM CSF production in mous e bone marrow cells. B) GM CSF production in mouse peritoneal macrophages. A B

PAGE 45

45 NOD C57BL/6 B6.NODC11b NOD.LC11b NOD.LC11e B6.NOD C11B/C1TB 0 5 10 15 20 25 30***p<0.0001; ANOVA *p=0.004 p=0.002 *p=0.0008 *p<0.0001 *p=0.0001 *p=0.007 ns *p=0.0062 *p=0.0072%STAT5Ptyr+CD11b+ cells NOD C57BL/6 B6.NODC11b NOD.LC11b NOD.LC11e B6.NODC1tb C11BXC1TB 0 10 20 30 40***p<0.0001; ANOVA *p=0.0065 ns ns *p=0.0003 *p=0.0071 *p=0.0033 *p=0.0338%STAT5Ptyr+CD11b+ cells *p=0.0071 ns Figure 46. The STAT5 Phosphorylation by Congenic Mouse Monocytes and Peri t oneal Macrophages. A) Phosphorylation of STA T5 at the tyrosine residue (STAT5Ptyr) in mouse monocytes. B) STAT5Ptyr in mouse peritoneal macrophages. A B

PAGE 46

46 A B6.NODC11b B B6.NODC1tb C B6.NODC11bxC1tb Female peri islet Female peri ductal Fem aleperi islet Maleperi islet Male peri ductal Maleinvasive Figure 47. Histology of Pancreas Tissues of B6.NOD and NOD.L subcongenic mice. Congenic mouse pancreas slices were stained with H and E (A, B, C ). Immunostaining was also performed against CD3 and F4/80 ( D ).

PAGE 47

47 D CD3 staining on pancreas CD3staining on pancreas C D3staining on p ancreas Lymph node CD3 Islet F4/80 staining Islet F4/80 staining Positive control 40x F4/80 staining 100x F4/80 staining 100x F4/80 staining 100x F4/80 staining Figure 47. Continued.

PAGE 48

48 NOD GM-CSF C57BL/6 GM-CSF NOD.LC11B GM-CSF NOD.LC11E GM-CSF B6.NODC11B GM-CSF B6.NODC1TB GM-CSF B6.NODC11BXC1TB GM-CSF 0 1 2 3 4 5 6 7 8 9 10 50 100 150 200 250 300 350 400 450 500STRAIN/ TREATMENTR NOD C57BL/6 NOD.LC11B NOD.LC11E B6.NODC11B B6.NODC1TB B6.NODC11BXC1TB 1.010-02 1.010-01 1.01000 1.01001 1.01002 1.01003 1.01004 1.01005 1.01006 1.01007STRAIN/ TREATMENTR Figure 48. Chromatin Immunoprecipitation (ChIP) analysis of STAT5 binding at the Csf2 promoter regions in congenic mice. A) Quantitative real time PCR analysis of mouse ChIP samples with and without GM CSF stimulation in mo use bone marrow cells B) Quantitative real time PCR of activated STAT5 binding at the Csf2 promoter in mouse peritoneal macrophages without GM CSF stimulation A B

PAGE 49

49 NOD GM-CSF C57BL/6 GM-CSF NOD.LC11B GM-CSF NOD.LC11E GM-CSF B6.NODC11B GM-CSF B6.NODC1TB GM-CSF B6.NODC11BXC1TB GM-CSF 0.1 1 10 100STRAIN / TREATMENTR NOD C57BL/6 C57L NOD.LC11B NOD.LC11E B6.NODC11B B6.NODC1TB B6.NODC11BXC1TB 0.1 1 10 100 1000 10000 100000 1000000 1.01007STRAIN / TREATMENTR Figure 49. Chromatin Immunopreci pitation (ChIP) a nalysis of STAT5 binding on the Ptgs2 e nhancer in GM CSF s timulated c ongenic m ouse bone m arrow and m acrophages. STAT5 binding at the Ptgs2 enhancer in congenic mouse bone marrow cells (A) and peritoneal macrophages (B). A B

PAGE 50

50 NOD NOD + GM C57BL/6 C57BL/6 + GM NOD.LC11B NOD.LC11B +GM NOD.LC11E NOD.LC11E + GM B6.NODC11B B6.NODC11B +GM B6.NODC11B/C1TB B6.NODC11B/C1TB + GM 0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85pg PGE2/million cells NOD C57BL/6 NOD.LC11B NOD.LC11E B6.NODC11B B6.NODC1TB B6.NODC11BXC1TB 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22pg PGE2/million cells Figure 410. The PGS2/COX2 e xpression and PGE2 production in c ongenic m ouse monocytes and m acrophages with and without GM CSF s timulation. A) PGE2 production in mouse monocytes. B) PGE2 produc tion in mouse macrophages. A B

PAGE 51

51 CHAPTER 5 HUMAN SEQUENCE ANALYSIS Given the STAT5 binding site changes in the promoter/enhancer regions of GM CSF and COX2 genes observed in the NOD mouse model, we postulated that the same binding site changes might occur in humans. To that end, we sequenced the promoter regions of homolog GM CSF and COX2 genes in human type 1 diabetics (T1D) as well as healthy controls to look for STAT5 binding site changes (Figures 51; 5 3) Gene Sequence Analysis: GM CSF ( CSF2 ) U pon assembly of the sequencing contigs obtained from four T1D patients and four healthy control samples, a STAT5 binding site (TTCN3GAA) was noted at 1015 relative to the transcriptional start site. Another STAT5 binding site was located at 962 in the r everse orientation (AAGN3CTT) 53 bases downstream of the above noted STAT5 site. Considering the traditional orientation independent mechanism of enhancers (3527033), it was not surprising to find a STAT site in this context. In addition, a STAT6 binding site (TTCN4GAA) was located 73 downstream of the STAT5 site at 942 (Figure 52). Both the STAT5 and STAT6 binding sites are identical between the T1D, controls and the published genomic sequence. Gene Sequence Analysis: COX2 (PTGS2) We also evaluated the PTGS2 enhancer region of 10 T1D patients and 12 healthy controls. Primers specifically designed to amplify the non transcribed region approximately 2kb upstream of the CSF2 were created based on GAS sequences (Interferon Gamma Activated Sequence) repo rted by Yamaoka et al ( 20, 38) (Figure 53 ) Upon examining a 397bp sequence within the enhancer region of PTGS2 we confirmed the location of the GAS motif sequences and found one to be a preferential STAT5 binding site (TTCN3GAA) and the other a prefer ential STAT6 binding site (TTCN4GAA). The STAT5 binding site, located at 757 relative to the

PAGE 52

52 transcriptional start site, is intact within the sequences of the T1D, controls and the published genomic sequence (Figure 5 4) The STAT6 binding site was obs erved at 713, approximately 44bp upstream of the STAT5 binding site, and was also intact (Figure 5 4) Additionally, a C/G heterozygote polymorphism 8 bases downstream of the STAT5 binding site was observed in 3 of the analyzed samples. Of the 22 total samples analyzed, the heterozygosity was found in both T1D and control samples (Figure 54 )

PAGE 53

53 Figure 51. Granulocyte Macrophage Colony Stimulating Factor gene ( CSF2 ) map. The above is a schemati c representation of the promoter region upstream of the GM CSF gene, CSF2. The location of the amplified region is delineated by forward and reverse primers, FP1,2,3 and RP1,2,3, respectively. Also defined are the STAT5 and STAT6 binding sites within th e region. Figure 52. Sequence alignment of the GM CSF promoter region from healthy control and T1D DNA shows intact STAT5 and STAT6 binding sites. STAT5 Binding Site STAT5 Binding Site STAT6 Binding Site Controls Patients

PAGE 54

54 Figure 53. Prostaglandin Synthase 2 gene ( PTG S2 ) map T he enhancer region upstream of the COX2 gene, PTGS2. The location of the amplified region is delineated by forward and reverse primers, FP1 and RP1, respectively. Also defined are the STAT5 and STAT6 binding sites within the region. Figure 54. Sequence alignment of the COX2 enhancer region from healthy control and T1D DNA shows intact STAT5 and STAT6 binding sites. STAT6 Binding Site STAT5 Binding Site STAT5 Binding Site Controls Controls Patients

PAGE 55

55 CHAPTER 6 DISCUSSION The inability of APC to effectively induce tolerance and regulate effect or cell function is t hought to play a contributing role in the immunopathology of T1D (11). Chronic inflammation and aberrant myeloid differe ntiation and activation may be factors in this loss of APC tolerance. The proinflammatory cytokine GM CSF is required f or two signals in myeloid cells: differentiation into mature monocytes, macrophages and dendritic cells, capable of toleragenic activity, and as an inflammatory signal in mature monocytes and macrophages to activate production of proinflammatory cytokines and prostanoi ds. Along with M CSF, GM CSF is found in abundance at sites of inflammation and localized autoimmunity (such as in rheumatoid arthritis) ( 16). It has been suggested that during inflammation, GM CSF, in concert with M CSF and G CSF form an important networ k of communication between myeloid cells and adjacent cells which may activate mature myeloid cell populations to produce pro inflammatory mediators ( 37, 38). Our previous data showed that GM CSF is expressed in abnormally high levels in NOD mice and T1D patient autoimmune myeloid cells and this subsequently activates transcription factors STAT5A and STAT5B which mediate epigenetic regulation ( 20, 23). Along with our finding that in the NOD mouse, GM CSF feeds back to regulate its own gene expression, it is possible that in autoimmune cells a positive autocrine/paracrine feedback loop is established which promotes maintenance of the pro inflammatory microenvironment characteristic of autoimmune disease. Our ChIP binding studies within the Csf2 promoter r egion suggest that there is a critical level of GM CSF stimulus needed to influence STAT5 binding within the region. We suggest that in the NOD mouse, the persistent presence of activated STAT5 and its subsequent abnormal binding at the Csf2 promoter medi ate enhanced expression of the gene and increased production

PAGE 56

56 of GM CSF. B6.NOD C11b and NOD.LC11e congenic mice share an NOD derived region within the Csf2 promoter that encompasses approximately 1kb to 250bp away from the TATA box of Csf2 gene. These mice recreate the three observed autoimmune phenotypes A proposed mechanism for this aberrant epigenetic regulation lies in the function of STAT5 as a n adapter for de acetylase binding and activity ( 39) Prolonged de acetylase act i vity could promote prot racted assembly and stabilization of transcriptional machinery and thus enhanced transcription of Csf2 leading to over production of GM CSF. STAT5 binding data from the bi congenic mouse (NOD background at both the Ptgs2 enhancer on Chr. 1 and Csf2 promot er on Chr 11) suggest that there is interaction between Chromosomes 1 and 11, mediated by GM CSF. STAT5 binding this site affects not only GM CSF expression, accompanied by its continued activation of STAT5, but also affects GM CSF stimulation of Ptgs2 ex pression, apparently through stimulating STAT5 binding at the Ptgs2 gene enhancer. We suggest a model by which abnormal GM CSF regulation contribute s to persistent inflammation via overproduction of the inflammatory mediator enzyme COX 2, with enhanced G M CSF inflam matory cytokine production would perpetuate the pro inflammatory cycle. In conjunction with preliminary histology data from the bi congenic mouse, showing macrophage infiltrate into the islet in the absence of T cell infiltrate, we propose tha t the proinflammatory macrophages and monocytes may migrate into the islet tissue and create an inflammatory microenvironment that may facilitate the migration and activation of destructive effect or cells. In order to further investigate this claim, comp arative histology data is needed to look at the presence of macrophages in pancreatic tissue of non autoimmune mice. Within the NOD mouse Csf2 promoter we observed a loss of STAT6 binding site (and a potential lo ss of anti inflammatory mediation through IL 4 suppression ) and gains of a STAT5

PAGE 57

57 binding site which may in part explain abnormal STAT5 binding within the promoter region. Homologous sites were not found in the T1D patient versus healthy control, suggesting alternate regulatory mechanisms contributi ng to the persistent STAT5 phosphorylation in autoimmune monocytes. However, we did observe a binding site pattern within both the Ptgs2 enhancer and Csf2 promoter in humans that consists of two STAT5 binding sites and one STAT6 binding. In order to under stand how STAT family mole cular interactions impact transcription of these genes and possibly other it may be valuable to perform a genome wide assessment of this particular binding pattern. The Ptgs2 enhancer sequence lies between the I L 10 gene and the Ptgs2 it is feasible that GM CSF regulation of the expression of both Csf2 and Ptgs2 may be affected by the enhanced STAT5 binding at the site. A future direction for this study would be to expand sequence analysis in m ouse and human samples to inclu de the IL 10 gene as well as the non transcribed region between the IL 10 and COX 2 genes in order to look for additional regulatory elements that may contribute to altered transcription of these two genes An insulator sequence element, which acts to modulate gene expression between two adjacent genes with differential expression patterns, could affect transcriptional regulation when located between an enhancer and a promoter by blocking gene expression stimulated by the enhancer ( 40, 41) It is also po ssible that a microsatellite insert within the non transcribed region could affect bi directional enhancer binding depending on where it is added, thus changing the expression patterns of the surrounding genes Based on research demonstrating three shared phenotypes between human Type 1 Diabetic patients and autoimmune NOD mice; persistent STAT5 phosphorylation, C OX2 over expression and GM CSF over production ( 20, 23), we developed a small volume blood drop

PAGE 58

58 assay of activated STAT5 levels within human peripheral blood monocytes to determine the reproducibility of the phenotypic findings for potential use as a minimally invasive T1D biomarker. Assay validity for T1D is uncertain for the entire population, although it may be useful in females as a biomarker for T1D or as an early indicator of autoimmunity. STAT5 is activated by estrogen and progesterone in addition to other hormones so that may be a contributing factor in the usefulness of this assay as a biomarker in studies of gender bias seen in autoimmunity Our studies in humans were limited to analysis of peripheral blood monocytes. Gaipa et al. reported that bone marrow precursor cells of j uvenile myelomonocytic leukemia patients exhibited elevated levels of STAT5 in response to GM CSF stimulation ( 39). In future studies it may be valuable to investigate the functional assays of the project within the context of autoimmune bone marrow precursor cells and autoimmune macrophages. This may also be useful in elucidating the mechanisms driving the observed phenotypes within the NOD mouse and human T1D patient given that the human sequence data did not produce a direct homolog to the mouse binding site changes within the GM CSF gene promoter region.

PAGE 59

59 LIST OF REFERENCES 1. Rose, N. R. 2004. Autoimmune disease 2002: an overview. J. Investig. Dermatol. Symp. Proc. 9: 14. 2. JDRF. Type 1 Diabetes: Dedicated to Finding a Cure: Juvenile Diabetes Research Foundation International. 2008: 3. Petersen, J. S., K. O. Kyvik, P. J. Bingley, E. A. Gale, A. Green, T. Dyrberg, and H. BeckNielsen. 1997. Population based study of prevalence of islet cell autoantibodies in monozygotic and dizygotic Danish twin pairs with insulin dependent diabetes mellitus. BMJ 314: 15751579. 4. Redondo, M. J., M. Rewers, L. Yu, S. G arg, C. C. Pilcher, R. B. Elliott, and G. S. Eisenbarth. 1999. Genetic determination of islet cell autoimmunity in monozygotic twin, dizygotic twin, and nontwin siblings of patients with type 1 diabetes: prospective twin study. BMJ 318: 698702. 5. Pirot, P., A. K. Cardozo, and D. L. Eizirik. 2008. Mediators and mechanisms of pancreatic beta cell death in type 1 diabetes. Arq. Bras. Endocrinol. Metabol. 52: 156 165. 6. Pihoker, C., L. K. Gilliam, C. S. Hampe, and A. Lernmark. 2005. Autoantibodies in diabet es. Diabetes 54 Suppl 2 : S52 61. 7. Finkel, T. H., J. C. Cambier, R. T. Kubo, W. K. Born, P. Marrack, and J. W. Kappler. 1989. The thymus has two functionally distinct populations of immature alpha beta + T cells: one population is deleted by ligation of a lpha beta TCR. Cell 58: 10471054. 8. Ramsdell, F. and B. J. Fowlkes. 1990. Clonal deletion versus clonal anergy: the role of the thymus in inducing self tolerance. Science 248: 13421348. 9. Serreze, D. V. 1993. Autoimmune diabetes results from genetic de fects manifest by antigen presenting cells. FASEB J. 7: 10921096. 10. Serreze, D. V. and E. H. Leiter. 1991. Development of diabetogenic T cells from NOD/Lt marrow is blocked when an allo H 2 haplotype is expressed on cells of hemopoietic origin, but not on thymic epithelium. J. Immunol. 147: 12221229. 11. Serreze, D. V. and E. H. Leiter. 1988. Defective activation of T suppressor cell function in nonobese diabetic mice. Potential relation to cytokine deficiencies. J. Immunol. 140: 38013807. 12. Ucker, D S., J. Meyers, and P. S. Obermiller. 1992. Activation driven T cell death. II. Quantitative differences alone distinguish stimuli triggering nontransformed T cell proliferation or death. J. Immunol. 149: 15831592. 13. Litherland, S. A., K. M. Grebe, N. S. Belkin, E. Paek, J. Elf, M. Atkinson, L. Morel, M. J. Clare Salzler, and M. McDuffie. 2005. Nonobese diabetic mouse congenic analysis reveals chromosome 11 locus contributing to diabetes susceptibility, macrophage STAT5 dysfunction, and granulocyte macr ophage colony stimulating factor overproduction. J. Immunol. 175: 45614565.

PAGE 60

60 14. Hamilton, J. A. 2002. GM CSF in inflammation and autoimmunity. Trends Immunol. 23: 403408. 15. Litherland, S. A. 2007. Cytokine Input in Monocyte and Macrophage Development. 16. Wheadon, H., P. J. Roberts, M. J. Watts, and D. C. Linch. 1999. Changes in signal transduction downstream from the granulocyte macrophage colonystimulating factor receptor during differentiation of primary hemopoietic cells. Exp. Hematol. 27: 10771086. 17. Hamilton, J. A., E. R. Stanley, A. W. Burgess, and R. K. Shadduck. 1980. Stimulation of macrophage plasminogen activator activity by colonystimulating factors. J. Cell. Physiol. 103: 435445. 18. Yamaoka, K., T. Otsuka, H. Niiro, H. Nakashima, Y. Tanaka, S. Nagano, E. Ogami, Y. Niho, N. Hamasaki, and K. Izuhara. 1999. Selective DNA binding activity of interleukin 10 stimulated STAT molecules in human monocytes. J. Interferon Cytokine Res. 19: 679 685. 19. Mertz, P. M., D. L. DeWitt, W. G. Stetler S tevenson, and L. M. Wahl. 1994. Interleukin 10 suppression of monocyte prostaglandin H synthase 2. Mechanism of inhibition of prostaglandin dependent matrix metalloproteinase production. J. Biol. Chem. 269: 2132221329. 20. Yamaoka, K., T. Otsuka, H. Niiro, Y. Arinobu, Y. Niho, N. Hamasaki, and K. Izuhara. 1998. Activation of STAT5 by lipopolysaccharide through granulocyte macrophage colony stimulating factor production in human monocytes. J. Immunol. 160: 838 845. 21. Chase, H. P., R. L. Williams, and J. D upont. 1979. Increased prostaglandin synthesis in childhood diabetes mellitus. J. Pediatr. 94: 185189. 22. Litherland, S. A., T. X. Xie, K. M. Grebe, A. Davoodi Semiromi, J. Elf, N. S. Belkin, L. L. Moldawer, and M. J. Clare Salzler. 2005. Signal transduc tion activator of transcription 5 (STAT5) dysfunction in autoimmune monocytes and macrophages. J. Autoimmun. 24: 297310. 23. Yagisawa, M., K. Saeki, E. Okuma, T. Kitamura, S. Kitagawa, H. Hirai, Y. Yazaki, F. Takaku, and A. Yuo. 1999. Signal transduction pathways in normal human monocytes stimulated by cytokines and mediators: comparative study with normal human neutrophils or transformed cells and the putative roles in functionality and cell biology. Exp. Hematol. 27: 10631076. 24. Darnell, J. E.,Jr. 199 7. STATs and gene regulation. Science 277: 16301635. 25. Schindler, C. and J. E. Darnell Jr. 1995. Transcriptional responses to polypeptide ligands: the JAKSTAT pathway. Annu. Rev. Biochem. 64: 621651. 26. Liu, L., Y. Li, and T. O. Tollefsbol. 2008. Gene environment interactions and epigenetic basis of human diseases. Curr. Issues Mol. Biol. 10 : 2536. 27. Hebbes, T. R., A. W. Thorne, and C. Crane Robinson. 1988. A direct link between core histone acetylation and transcriptionally active chromatin. EMBO J. 7: 13951402.

PAGE 61

61 28. Landsberger, N. and A. P. Wolffe. 1997. Remodeling of regulatory nucleoprotein complexes on the Xenopus hsp70 promoter during meiotic maturation of the Xenopus oocyte. EMBO J. 16: 43614373. 29. Seydel, F., E. Garrigan, B. Stutevoss, N Belkin, B. Makadia, J. Carter, J. D. Shi, A. Davoodi Semiromi, M. McDuffie, and S. A. Litherland. 2008. GM CSF Induces STAT5 Binding at Epigenetic Regulatory Sites within the Csf2 Promoter of Nonobese Diabetic (NOD) Mouse Myeloid Cells. 30. McDuffie, M. 2000. Derivation of diabetes resistant congenic lines from the nonobese diabetic mouse. Clin. Immunol. 96: 119130. 31. Yui, M. A., K. Muralidharan, B. MorenoAltamirano, G. Perrin, K. Chestnut, and E. K. Wakeland. 1996. Production of congenic mouse str ains carrying NOD derived diabetogenic genetic intervals: an approach for the genetic dissection of complex traits. Mamm. Genome 7: 331334. 32. Morel, L. 2004. PCR Technique and Congenic Mouse Breeding. 33. Chen, X., J. Wang, D. Woltring, S. Gerondakis, and M. F. Shannon. 2005. Histone dynamics on the interleukin2 gene in response to T cell activation. Mol. Cell. Biol. 25: 32093219. 34. Ito, K., P. J. Barnes, and I. M. Adcock. 2000. Glucocorticoid receptor recruitment of histone deacetylase 2 inhibits i nterleukin 1beta induced histone H4 acetylation on lysines 8 and 12. Mol. Cell. Biol. 20 : 68916903. 35. Yan, C., H. Wang, Y. Toh, and D. D. Boyd. 2003. Repression of 92kDa type IV collagenase expression by MTA1 is mediated through direct interactions wit h the promoter via a mechanism, which is both dependent on and independent of histone deacetylation. J. Biol. Chem. 278: 23092316. 36. Litherland, S. A., T. X. Xie, K. M. Grebe, Y. Li, L. L. Moldawer, and M. J. Clare Salzler. 2004. IL10 resistant PGS2 expression in at risk/Type 1 diabetic human monocytes. J. Autoimmun. 22: 227233. 37. Ehret, G. B., P. Reichenbach, U. Schindler, C. M. Horvath, S. Fritz, M. Nabholz, and P. Bucher. 2001. DNA binding specificity of different STAT proteins. Comparison of in vi tro specificity with natural target sites. J. Biol. Chem. 276: 66756688. 38. Imada, K. and W. J. Leonard. 2000. The JakSTAT pathway. Mol. Immunol. 37: 111. 39. Rascle, A., J. A. Johnston, and B. Amati. 2003. Deacetylase activity is required for recruitm ent of the basal transcription machinery and transactivation by STAT5. Mol. Cell. Biol. 23: 41624173. 40. Burgess Beusse, B., C. Farrell, M. Gaszner, M. Litt, V. Mutskov, F. Recillas Targa, M. Simpson, A. West, and G. Felsenfeld. 2002. The insulation of genes from external enhancers and silencing chromatin. Proc. Natl. Acad. Sci. U. S. A. 99 Suppl 4: 1643316437.

PAGE 62

62 41. Wallace, J. A. and G. Felsenfeld. 2007. We gather together: insulators and genome organization. Curr. Opin. Genet. Dev. 17: 400407. 42. Rumore Maton, B., J. Elf, N. Belkin, B. Stutevoss, F. Seydel, E. Garrigan, and S. A. Litherland. 2008. M CSF and GM CSF Regulation of STAT5 Activation and DNA Binding in Myeloid Cell Differentiation is Disrupted in Nonobese Diabetic (NOD) Mice. Clinical and De velopmental Immunology

PAGE 63

63 BIOGRAPHICAL SKETCH Erin Garrigan was born in Tarpon Springs, Florida in 1984 to Patrick and Deborah Garriga n. After attending Palm Harbor University High School in Palm Harbor, Florida she graduated from the International Baccalau reate Pro gram in 2003, and continued on to the University of Florida. In s pring 2005, Erin joined the lab of Dr. Sally Litherland in the Department of Pathology, Immunology and Laboratory Medicine, where she worked as a research assistant until enrolling i n graduate school in 200 7. Erin Garrigan graduated from the University of Florida in 2007 with a Bachelor of Science degree in i nterdisciplinary s tudies with a focus on biochemistry and m olecular b i ology and a minor in nutrition. She conducted her m aster s research in Dr. Mark Atkinsons lab under the comentorship of Drs. Sall y Litherland and Mark Atkinson. Erin received her m aster s degree in Medical Sciences from the University of Florida in December 2008.